A Study of Tarka and Its Role in Indian Logic

INDUCTIVE REASONING
A Study of Tarka and Its Role

in Indian Logic
By
SITANSUSEKHAR BAG CHI,
M.A.,
LL.B., D.Litt. (CAL.)
Professor, Mithila Institute of Post-Graduate Studies and Research

in Sanskri Learning, Darbhanga; Formerly Lecturer,
Calcutta
University
and
Maharaja Manindra
Chandra College, Calcutta.
1953
PREFACE
The p resent work w:\s submitted as
of the Calcutta
University
n d was
:\
thesis for the D. Litt. degr ee
unanimously approved by the
Board of Adjudicators in 1949'
One of the chief incentives for em barking upon this arduous work
was
and
to place materials at the hands
of the schol ar s of the E uro p e a n
American universities, w h i ch would enable them
unbiased appraisal of
Indian philosop hers.
the flight
to make an
of logical speculations recorded by
I thought that my cherished desire was bound
to remain unrealized if the volume was not speed ily published.

So with this object always in view, I first approa ched the autho
rities of the Univer sity of C a l cutta and then the Education
Depart4
ment of the Government of West Benga l , but only to reap the result of
my fruitless endeavo u r s.
I sent my work to the press even at the risk of bei ng involved in

fi n ancial difficulties, and I
seen the light of t he day.
am
glad to see t h at this humble work
has
Before I finish, I must exp re ss my sen se of
p rofound gratitude to Prof. Anukul Chandra Mookerjee of the Allaha4

bad University for the unstinted appre c iation and encouragement which
he ungrudgingly e xtend ed twords me.
My heartiest tha nks are due
to Prof. JatiIcoomar Mookerjee, who, young in age but

scholarship, has rendered
A cold4blooded
log icia n
origi nal writing in ma ny
invaluable
assistance
in
numerous
aged in
ways.
as he is, he has compelled me to rev ise my

places
by h is unsparing
and devastating
CritICIsm.
It is
matter of pr o found regret to me that Sri Ksitindranath
Ba nerjee B.A., LL.B., and his son Sri Visvanath Banerjee are no l onger
in the land of the l i vi ng to see the result of my researches.
Lastly
I must express my sense of ir redeem ab le debts which lowe to Maha4

raja Bhupendrachandra
Sinha B A. of S usan g,
.
Banerjee M.A., LL. B., Advocate,
to Sri: Binay aknath
Calcutta High Court,
and to
Dr.
Jyotirmoya Ba nerj ee M.B., D.P.H., D.T.M., for the keen appreciation
and constant encouragement which I have received from them in the
course of my scholastic pursuita debt which is too deep and personal

for formal acknowledgement
I shall be failing in my duty if I do not handsomely express my
thanks to Sri Kamalakanta Dalai, Manager, Calcutta Oriental Press
Ltd. for the lively interest which he took In the publication of my
work.
Calcutta, October 1953.
SITANSUSEKHAR BAGCHI
CONTENTS
(THE NUMBERS INDICATE PAGES)
INTRODUCTION
PRONUNCIATION
ABBREVIATIONS
xxvii
xxix
CHAPTER I
THE NATURE AND UTILITY OF REASONING
1-15
Enquiry regarding the importance and the possibility of the
knowledge of the universal proposition, 1. Consideration of the
Naiyyika's theory, 2. Implication of the universal proposition
unfolded, 3. Nature of doubt explained, 3. Reasoning (tarka)
serves to eliminate doubt and thereby helps the determination
of truth, 4. Reasoning as defined in the NS, 4. Vtsyyana's
view on the services of reasoning, 4-5. Vacaspati's, and Uddyotakara's view on the services of reasoning explained, 4-10.
Elucidation of Udayana's difference from Vacaspati with regard
to the actual results that are achieved by reasoning, 9-15. Bhatta
Vadindra's elucidation of Udayana's position, io, Mallintha's
confusion, i j n .
CHAPTER II
THE POSITION OF SRIVALLABHATHE AUTHOR OF THE
NYYALlLlVATl
16-30
SriVallabha's tentative suggestion regarding the result of the

application of reasoning, I 6 - I 8 . His elaborate criticism of the
position of Udayana, Vacaspati and Jayanta Bhatta, 16-18.
Srvallabha shows that reasoning cannot be regarded as the
necessary function of inference, 18. Refutation of the view of
Vacaspati and Udayana who have asserted reasoning to be the
function of a cognitive instrument, 18-19. Srivallabha's recognition of the necessity of the second type of reasoning (yyaptigrhaka tarka), 20, and his conclusion that the result of the
application of reasoning is the removal of doubt about the

necessity of the relation between the probans and the probandum, 20-23. Bhagiratha Thakkura's criticism of rvallabha's
rejection of Udayana's position, 23, and the elucidation of his
contention that reasoning of the first type (yisaya-parisodhaka
tarka) is of service even in inference by eliminating enquiry
about the opposite possibility as maintained by Udayana, 23.
His making out reasoning as a subsequent sub-organ of the
instrument of inference, 26, and his exposition of Udayana's
position shown to be more loyal than that of Vardhamna, 26.
His agreement and difference with Rucidatta, 26-30.
Rucidatta, his difference in substantial respects from the views
of Srivallabha and Bhagiratha Thakkura, and the modus operandi
of reasoning as shown by him differs from what is assumed by
Bhagiratha, 29. His contention that reasoning is always an
antecedent sub-organ and has two different functions elaborated, 30.
CHAPTER III
GANGELA'S SPECULATIONS ON REASONING
31-53
Gagesa's enquiry into the means of arriving at the knowledge
of universal concomitance, 32, and his elucidation of the condition of induction, 32. His elaboration of two kinds of doubt,
32-33, and their elimination, 33-34. Answers to the objection that the foundational concomitance of reasoning is bound
to entail regresses ad infinitum, 34-35 Gagesa's explanation of the modus operandi of reasoning, 35-36. Doubt is not
infinitely possible and consequently the appeal to reasoning for
rebutting doubt is not necessary ad infinitum, 37. Elucidation and criticism of the contention of Snharsa, 33-41. Admission of the possibility of reasoning being correct or incorrect, right or wrong, 43, Consideration of the views of the
older Naiyyikas as recorded by Gagesa on the problem of
infinite regress as the inevitable consequence of the application
of reasoning, 43-45. Gagesa's allusion to the views of an older
school of logicians who have sought to solve the problem of
infinite recurrence of doubt by appeal to psychological evidence, 45, and his criticism of this psychological solution, 46.
View of this older school of logicians recorded and rejected by
Sriharsa, 47-48. Reference to a class of Mimmsists who held
chat observation of concomitance alone was sufficient to prove
the necessity and universality of the relation between two
terms, 48, and its refutation, 48-50, resume of the views as
given by Sakara Misra, 50-51. Evaluation of the solution
offered by Gagesa of the problem of infinite regress urged by
Sriharsa, 51-53.
CHAPTER IV
THE LOGICAL STATUS OF REASONING
54-76
SECTION I
54-7 2
Unbroken tradition among ttie Naiyyikas that reasoning is

not an instrument of knowledge, 54. The Mdhvas, the Rmnujyas, and the Jainas have accorded to reasoning the status
of a cognitive instrument, 54. Vtsyyana, Uddyotakara, and
Vcaspati regard reasoning rather as a helpful condition and
an auxiliary to an independent instrument, 54. Udayana's
attempt to establish the subordinate character and role of
reasoning from a different line of approach, 54, and his assertion
that reasoning is not knowledge, but an assumption and an
imaginary imposition, 54. Difference of hypothetical assumption
from inference elucidated, 55, and the wide divergence of
opinion regarding the nature of reasoning elaborated, 56.
Allusion to the view of a class of philosophers,' who
regard assumption as different from doubt and certitude, and
think it to be a species of cognition, 56. Assumption whether
a species of cognition or not discussed, 57. The difference
between judgment in general, and reasoning, hypothetical
judgment and assumption which mean one and the same thing
amplified, 58. Characteristics which distinguish assumption
from knowledge or judgment considered, 58-59. Me Taggart's
view incidentally referred to, 58-59. Criticism of the different

classes of Naiyayikas' view on the nature of assumption, 60.
Udayana's attempt to place assumption under the head of
erroneous knowledge criticised, 60. Assumption distinguished from knowledge proper, and error, 60-61. Error and
assumption are shown to be cases of ascription, 6 1 . McTaggart's
view referred to, 62. The processes of meditation recommended in the Upanisads and the later Paurnic literature are
shown to be cases of conscious ascription, 62. Two members
of reasoning as a hypothetical judgment elucidated, 62.
Definition of assumption by the neo-logician, 62, and it cannot
be confounded with authentic knowledge, 62. Reasoning is a
complex judgment or rather a compound of assumptions, 63-64.
Unanimity amongst the Naiyyikas that reasoning is not an
independent cognitive instrument, 64-72.
SECTION II
THE LOGICAL STATUS OF REASONING AS ELUCIDATED IN

THE NYYAKANDAL1 AND IN THE VY0MAVAT1
72-76
Sridhara's elucidation of the position of the older Naiyyika

regarding the nature and the logical status of reasoning, 72.
His refutation of older Naiyyika's position, 73. His conclusion that reasoning is an instrument for proving the impossibility of the position advocated by the opponent and as such
has all the characteristics of inference elaborated, 73-75 His
view that the validity of reasoning as an independent organ
cannot be impeached amplified, 73.
Vyomasivcrya, his arguments regarding the nature and the
logical status of reasoning are shown to be substantially in
agreement with those of Srldhara, 75. His contention that
reasoning is not an invalid assumption, 75, and his conclusion
that reasoning must be an organ of knowledge, and as such
must be subsumed either under perception or inference elucidated, 75-76.
CHAPTER V
THE STATUS OF REASONING IN MADHVA'S SCHOOL AS
ELUCIDATED IN T]HE PRAMNAPADDHAT1 BY JAY ATIRTHA
77-91
Jayatlrtha's view on reasoning as a species of inference and
as such an independent organ of knowledge elucidated, 79.
His classification of inference, 79. Elaboration of his
view that
reasoning is a variety of inference for
refutation of a position, 80. His definition of reasoning,
79.
Commentator Janrdana Bhatta's elucidation of this
definition, 80.
Elaboration of his view that the defects
of a categorical inference, are not the defects of a hypothetical inference, which is called a reductio ad absurdum,
86. Elucidation of his contention that reasoning consists
of reductio ad absurdum as only one half and the negative
implication as the other half; the two together make a completed whole, and it is only as such a whole that it functions
as a logical organ, Sj.
Criticism of the Naiyyika's
objections, 87. His conclusion that reasoning is an accredited
organ of knowledge and is subsumed under inference, 87.
His partial agreement with Vacaspti Misra, 88.
His
difference from the Naiyyikas for regarding reasoning
as an independent organ irrespective of its bearing upon
another organ elaborated, S3. His agreement with the
Naiyyikas in so far as reasoning is regarded as a helping
factor, 89. His view that reasoning is a case of hypothetical inference and the latter is as much possessed of cogency
as categorical inference is accepted to be, 89-90. His complete agreement with the Jaina logicians, with this exception
that whereas the latter regard it as an extra kind of organ, the
former have subsumed it under inference, 9 1 .
CHAPTER VI
REASONING AS ELUCIDATED IN THE NYYASUDH BY
JAYATIRXHA
92-106
Jayatrtha's elaborate dissertation on reasoning and his survey of

the views of the writers of the Nyya school beginning with the
author of the Nyyastra and coming down to Udayancrya,
92. His attempt to prove reasoning to be a variety of inference, 92. Elucidation of a debater's twofold task, and the way for
the accomplishment of the objectives, 93. Elaboration of the
twofold classification of inference for refutation, 95. Elucidation of the nature of reasoning, 95-98. The essential
condition of inference is the same in categorical and
hypothetical inference known as reasoning, 99. Allusion
to the view of Madhva who considers that reasoning is in the
same position as casuistry and sophisim, 104-5.
Venkatantha's view is a faint echo of Jayatrtha's conclusion,
106.
CHAPTER VII
REASONING AS ELUCIDATED IN THE
BY VYASATRT]HA
TARKATNDAVA
107-12 5
Vysatirtha's endorsement of the conclusion of Jayatirtha, 107.

His elaborate criticism of the conception of reasoning as a case of
conscious ascription, as has been set forth by the Naiyyikas
in their traditional definitions, 108. His definition of reasoning, 109. Refutation of Naiyayika's view that the minor
premise or the conclusion of reasoning is false, 111. Vindication of Jayatrtha's position that reasoning has the same
logical apparatus as inference and as such should be classed
with it, 113-14. Vysatirtha's exposition of the nature of
reasoning, 114-19. His view that reasoning as defined and
interpreted by him is both formally and materially valid
and that again even independently of the negative implication, 119-20.
His defence of Jayatrtha's position, 119.
His exposition of reasoning leaving no room for doubt
that it is to be classed with inference, and as such must
be admitted to be a valid logical organ, 120.
The
attempt of the Naiyayika to equate reasoning with the
act of deliberate identification, which is involved in the case of
XV
images of deities is shown to be due to the reluctance of facing

an unwelcome situation, 121. His observation that reasoning
is resorted to with a view to effective rebuttal of doubt, 122.
Criticism of the contention of Gagesa that reasoning is a
species of cognition different from the other recognised species,
122. His admission of the possibility of the two kinds of
doubt, antecedent and consequent to the knowledge of the
premises, 123. His classification of reasoning under two heads
in conformity with the twofold division of doubt, antecedent
and consequent to the knowledge of the premises, 124. His
admission of the two conventional types of reasoning, 124. His
criticism of the exposition of the nature of service rendered
by reasoning according to the Naiyyikas, 124-25.
His
consideration of the different views of the utility of reasoning,
125.
Evaluation of his definition of reasoning and his
exposition of its nature and service, 125.
CHAPTER
VIII
THE JAINA CONCEPTION OF REASONING
126-150
Elucidation, of Jaina logician's contention that reasoning is a

separate organ of cognition and its content is universal
concomitance, 126-27.
Elaboration of the conditions of
reasoning which are observation and non-observation' of the
co-occurrence of two terms 126-27. Elucidation of the Jaina
logician's view that reasoning is an entirely different species
of cognition which can be subsumed neither under perception,
nor inference, nor any other kind of knowledge, 126.
Reasoning is a mental act, which though it arises from the
perceptual cognition of the present data, has jurisdiction
over past, present and future, as such it is a different type of
cognition both in respect of form as well as matter, 126.
The relation of necessary concomitance holds between two
terms in their essential character, 134. Amplification of
his view that the existence of pure universals or pure
particulars, are the fictions of abstract logic, 134, and
that the dichotomy between universal and particular is false,
134 H's position that the necessary concomitance holds

between two reals which are by their very constitution
inseparable from each other explained, 135. Reals are neither
universals nor particulars, but constitute a third type, which
is distinct from both of them severally and jointly, 135.
Negative concomitance subsist between two reals taken in
their universal reference, 135. Negation is not an abstract
non-entity, but a positive entity, 135. Naiyyika's objections
that reasoning cannot be an organ of knowledge, and the organ
of necessary concomitance is perception and not reasoning
answered, 139-43. Madhva's attempt to equate reasoning
with inference, and also the position of a class of Buddhists,
who call reasoning prasangnumna refuted, 143-44. The
Buddhist position on reasoning elucidated and examined,
144-49.
His agreement with other realistic schools of
philosophers, such as the Naiyyikas, the Mmmsists,
particularly of the school of Prabhkara, and his opinion
that there is no ground a priori or a posteriori for the
condemnation of conceptual thought as inherently a vicious
construction of human intellect elucidated, 144. Refutation
of the Buddhist's denunciation of conceptual thought, 145.
Elaboration of the nature of perceptual cognition, 148.
Reasoning shown to possess the criterion of originality,
agreement and verification which are the tests of an organ of
knowledge, 150. Possession of the third criterion of truth by
reasoning, namely, its capacity to rebut doubt and misconception elucidated, 150. Agreement with rdhara, Vyomasivcrya
and the whole school of Mdhva philosophers, who have
unanimously protested against the absurdity involved in the
denial of validity of reasoning, and in having at the same
time recourse to it as the instrument for the rebuttal of
doubt, 150.
CHAPTER
CLASSIFICATION OF REASONING
IX
151-183
Section I
151-156
Udayana's fivefold classification of reasoning, 151. Varadarja's and Visvantha's endorsement o Udayana's classification,
151. Visvantha's repudiation of other varieties of reasoning, 151. Tarka used as the symbol of the number six, 152.
Udayana's acceptance of this usage, although has only spoken
of the five species of reasoning, 152. The sixth variety of
reasoning given by Sriharsa, 152. Not the slightest suggestion
of classification of reasoning in the Nyyastra, Bhasya>
Varttika, and the Ttparyaftkat 152. Udayana seems to
speak of the fivefold classification for the first time in the
tmatattvaviveka, 152. Enumeration of reductio ad absurdum
as an additional variety of reasoning by Udayancrya is not
logically justifiable, 152. Vamsdhara Misra's acceptance of
Sriharsa's classification of reasoning, 155. Reductio ad
absurdum is not a particular variety of reasoning according to
Bhaglratha and Raghuntha, 153. Sakara Misra's acceptance
of the fivefold classification of reasoning as propounded by
Udayana, 154, and his calling the fifth variety emergence
of an absurdity other than that involved in the aforesaid
cases, 154. Vekatanatha's acceptance of Udayana's classification of reasoning, 154. His innovation of pure reductio
ad absurdum, 154. Division of the pure reductio ad absurdum
into two sub-varieties in the Prajaparitrana, 154. Vekatanatha's enumeration of three other varieties, 154. Srlnivasa's
addition of two further varieties, 154. Srharsa's enumeration of eleven varieties of reasoning, 155, which are regarded
by Sakara Misra as simulations of reasoning, 155, but Is
dittoed almost totidem verbis in the Vadivinoda, 155.
SECTION II
DEFINITIONS OF THE DIFFERENT VARIETIES OF REASONING 156-83

(1) Self-dependence, (2) Mutual dependence, (3) Vicious
circle, and (4) The vicious infinite explained, 156-60.
Srlharsa's enumeration of two infinite series viz. regressus ad
infinitum and the progresses ad infiniturn, 160. Their salient

features elaborated, 160-63. Tu quoque argument elucidated, 163-64. its striking example in the Nyayavarttika, 164.
Sriharsa's contention that the upshot of this type of reasoning
is the demonstration of the equality of both the parties, 165,
and his allusion to Kumrila, 165. His difference from
Uddyotakara, 165-66. Absence of decisive proof i.e. of
crucial evidence elucidated, 166-69.
Empirical induction
explained, 170-73. Criticism of the enumeration of the types
of reasoning as given in the Sarvadarsanasamgraha, 173.
Complexity and Simplicity elucidated, 175-76. Consideration
of the cogency of initial presentation, 178. Elucidation of
contradiction, 178.
Elaboration of Impertinence alias
Impudence, 180. Sriharsa's recognition of Impertinence, asa
distinct type of re duetto ad absurdum and his allusion to
Mahimabhatta in his support, 181. Equalisation urged by
Venkatantha and his commentator Srlnivscrya as a distinct
type of reductio ad absurdum. Elucidation of Double noo.se,
182-83.
CHAPTER X
CONDITIONS AND FALLACIES OF REASONING
SECTION I
184-204
184-91
Enunciation of conditions and fallacies of reasoning by Udayana,

184. Their codification by Varadarja and Snnivscrya,
184. Enumeration of the five conditions of reasoning, 184.
Enunciation of the five fallacies of reasoning, 185. (1) Mutual contradiction elucidated and illustrated, 187 and is the
self-same fallacy as countermanded probans, 1S7. (2) The
lack of logical sanction explained and illustrated, 187-88.
(3) The establishment of an issue endorsed by the opponent,
explained and illustrated, 188-89* (4) To be helpful (in the
proof of the opponent's thesis) elucidated and illustrated,
189-90. (5) Failure to culminate in the proof of the opposite
expounded with illustration, 190. Sriharsa's enunciation of
XIX
another variety of the fallacy of reasoning, viz. the unreality of

the subject, 190-91.
SECTION II
FALLACIOUS SIMULATIONS OF REASONING
191-204
(i)( Simulation of self-dependence explained and illustrated,

191-93. Definition of a universal subsistent, 194. Criticism of
the assertion of a group of Nyya-Vaisesika philosophers that
self-dependence in respect of subsistence is not a fallacy, 196.
Elaborate consideration of the definition of a universal subsistent
as propounded by the Naiyyika, 196. Madhusdana Sarasvat's declaration that self-dependence in respect of subsistence
is equally repugnant to logical sense with self-dependence in
respect of genesis and cognition, 197. His amendment of the
Naiyayika's definition of a universal subsistent, 197, which
meets the requirements of the situation, 197, and is shown to
apply to the disputed instances, knowability and the like
197-200. The position of the author of the Nyayamrta denying self-dependence in respect of subsistence as a case of
fallacy shown to be untenable, 200. Rmarudra's adoption of
the same line of defence as that of Madhusdana SarasvatI and
Gauda Brahmnanda, 200. (2) Simulation of mutual dependence explained and illustrated, 200-1. (3) Simulation of
vicious circle elucidated, 201. (4) Simulation of the vicious
infinite explained and illustrated, 201-2. (5) Simulation of
reductio ad absurdum elucidated, 202-3. (6) Simulation of
contradiction explained, 203. (7) Simulation of recrimination
or tu qiioque elucidated, 203. Simulation of the additional
varieties of reasoning, 204.
CHAPTER XI
REASONING QUA INSTRUMENT OF REFUTATION
205-14
Assignation of reason in a syllogistic argument liable to be

vitiated by fourfold defect enumerated, 205. All the four
varieties of defect of argument are shown to be nothing but

cases of fallacious reason, 205-6. Consideration of the omission
of the classification of reasoning on the part of the author of
the Nyayastra and Vtsyyana, 206-7. akara Misra's
non-recognition of contradiction as a variety of reductio ad
absurdum is in conformity with the views of Udayana and
Gagesa, 207. Attempt to refute the charge of infinite
regress urged by rharsa against the foundational concomitance of reasoning by Udayana and Gagesa by the plea that
the ultimate solvent of all doubt is contradiction, 207.
Allusion to Sakara Misra's contention that contradiction as
the third variety of censure called contradiction of the thesis,
207.
Consideration of the legitimate procedure in the
demonstration of defect in the opponent's argument, 208-9.
Elaboration of close affinity and difference between inference
and reasoning, 209-10. The fallacies of inference and the
fallacies of reasoning are shown to be essentially the same,
2 i o . The fallacies of reasoning can be subsumed-under the
logical fallacy, 210-12. Naiyayika's enumeration of the three
fallacies of reason, 212, are shown to be possible also in
reasoning, 212-14. Close kinship of the fallacies of reasoning
with those of inference creates a presumption in favour of
the position that reasoning is rather a case of inference, 214.
CHAPTER XII
A CRITIQUE OF REASONING
215-29
Sriharsa's charges against the definition of reasoning elucidated,

216-17.
Criticism of the amendments of the definition
of reasoning as sponsored by the Naiyyika, 217-22. The
denial of the status of reasoning to hypothetical assertion
on the ground of its failure to lead to the opposite conclusion
is shown to be unjustifiable, 222. Sriharsa's conclusion that as
all the possible definitions of reasoning are vitiated by one
defect or another, it cannot be an acceptable logical category,
216-29.
XXi
CHAPTER XIII
SPECIES OF REASONING CRITICISED
SECTION I
230-76
230-51
(1) Criticism of Self-dependence, 233-41. (2) Mutual dependence exposed to fatal objections, 241-43. (3) Criticism of
Vicious circle, 243-51.
SECTION II
251-76
(4) Contradiction, defined, 251-52, and subjected to scrutiny,

252-62. Nature of opposition between being and non-being
elucidated, 253-54. Definitions of contradiction found on examination to be absurd, 254-56. Logician's failure to explain
the nature of opposition, 257. The concept of opposition is
shown to be not capable of being derived from experience,
257-62. Opposition makes itself felt when we ideally try to
bring being and non-being together, 262, and as such is
a priori and ideal and never actual, 262. (5) Criticism of The
vicious infinite, 262-6J.
(6) Tu quoque or recrimination
elaborated and criticised,
263-67.
Logician's reliance
on tu quoque as necessitating an identical solution is shown to
be absolutely unjustified by facts, 268*65. The untenability
of ttt qnoque as a form of logical argument, 269-76.
CHAPTER XIV
REASONING QUA ORGAN OF INDUCTION
277-304
Elucidation of. the nature of universal and necessary concomitance between the probans and the probandum as the pivotal
basis of inference, 278. Criticism of the Naiyayika's contention that the concomitance of the probans and the probandum is guaranteed by reasoning, 278-79. Srlharsa's charge of
regressiis ad injinitum against the foundational concomitance
of reasoning elaborated, 279. Logician's retort to Sriharsa's
criticism, 279. Sriharsa's defence, 275-80.
Naiyayika's
defensive arguments elaborated, 280-82. Naiyayika's concept

of causality criticised by Srlharsa, 282-86. Elaboration of the
Vedantist's contention that the suspicion of a plurality of causes
makes all causal inferences precarious, 285-87. Naiyayika's contention that homogeneity of effect can be accounted for only by
the homogeneity of the cause is criticised by Sriharsa, 286-88*
Vedantist's position that even heterogeneous causes can produce a homogeneous effect explained and illustrated, 287.
Naiyayika's reply, 287-88. Vctulscs counter-reply, 288.
Logician's rejoinder, 289.
Vedantist's elaboration of the
possibility of plurality of causes, 289-94. Vedantist's refutation of Naiyayika's advocacy of validity of inference by
appeal to contradiction, 294-96. Further criticism of the concept of contradiction by Sriharsa, 295-96. Refutation of logician's contention that contradiction is the final solvent of doubt,
296. Vedantist's position that doubt is the condition and
support o contradiction and so there can be no antagonism
between the two elaborated, 296. Naiyayika's plea that*the
knowledge of contradiction or the memory-impression of contradiction will play the role of crucial evidence and will be
destructive of doubt elucidated, 296-27, and refutated by Sriharsa, 297-98. Contradiction or the knowledge of it cannot be
established by a cognitive organ or by a rednctio ad absurdum
explained, 299. Doubt of basal concomitance of reasoning
is shown to entail the vicious infinite series, 299.
Udayana's reply, 300. Vedantist's retort, 300. Vedantist's
detailed refutation of the logician's contention that the final
solvent of doubt is self-contradiction, which is the limit of it,
300-2. Srlharsa's facetious observation on Udayana's contention, 301. Vedantist's own position elaborated, 302-4.
Concluding remarks, 304.
I. INDEX OF SUBJECTS
II. INDEX OF WORKS
305
308'
III. INDEX OF AUTHORS
309
IVa INDEX OF SANSKRIT WORDS
309:
INTRODUCTION
In the course of my study of Jaina Nyya I found that the Jainas
[uve accorded a position of supreme importance to tarka as an instrument of inductive knowledge. Its criticism of the Nyya conception
of tarka stimulated my curiosity to study the Nyya literature on the
subject. I found that the fund of speculations in Nyya far exceeded in extent and in quality the speculations of the Jainas. The
wealth of materials proved too strong a temptation for me and I
thought that if I could give a critical representation of the thoughts
of the philosophers, whose activities were spread out over several
centuries, it might be regarded as an humble contribution to the
present stock of knowledge. The difficulty of language and the
abstruseness of the arguments in the original texts are real hurdles
in the way of a research scholar on Indian philosophy. But he survey
of the contour and also of the fundamental topics of Indian philosophy
has been completed by the works of Sir Sarvapalli Radhakrishnan and
of the late Professor S. N . Dasgupta, whose achievements will remain
immortal in the annals of the cultural history of India in recent times.
The immortal work of the late Mahmahopdhyya Dr S. C. Vidybhsana and the illuminating articles and papers of Mahmahopdhyya
Dr Gopinath Kaviraj, M.A., D. Litt., have also served to narrow down
the scope of future workers on Indian philosophy. What now remains
is the study of individual problems. I adopted for my researches the
problem of tarka, inductive reasoning, which I have rendered for the
sake of brevity simply as reasoning. The difficulties of the original
presentation in Sanskrit are too familiar to require to be stressed. I
have perhaps set myself a difficult task-perhaps too difficult for my
equipment. But the faith sustains me that scholars would make
handsome allowance for the difficulties and obscurities of the
materials and they would encourage me if I succeed in throwing spots
of light on a tangled problem, which is hardly studied in the original
even by the orthodox students of tols. I am perhaps guilty of
imprudence for choosing a subject which is enormously stiff and
whose promise of success is problematic. I was rather drawn into
it and must be excused if I have been enterprising, which is the

privilege of youth.
The present dissertation represents the strenuous labour of full
six years involved in the study of the writings of the philosophers
of various schools. I now present the results of my study of reasoning
in all its aspects, which must be admitted to have tremendous logical
and epistemological value. The importance of the subject of my
research may be gathered from the fact that Vardhamna has accorded
almost co-equal status to reasoning and inference, as the dual organ of
the proof of God's existence. Udayana has compared these two to the
feet of God. The aptness of the imagery is derived from the fact
that reasoning and inference together constitute the evidence of Divine
Reality1. I have made my study as comprehensive and as intensive
as lay in my power.
The present dissertation starts from the
Nyyastra and comes down to the latest exponents of the school of
neo-logicians, and embraces a comprehensive treatment of the
Vaisesika, Mdhva and Jaina schools of thought.
The representation of the speculations of the Mdhva school of
philosophy on this topic will, I hope, be regarded as particularly
interesting and the views of Vysatlrtha will be appreciated as highly
original. I must in this connexion place on record my deep gratitude
to the late lamented Professor S. N . Dasgupta, who allowed me to
make a copy of the chapter dealing with reasoning from the manuscript
of the Tarkatndava of Vyasatlrtha, which was in his possession.
It is a matter of genuine gratification that this important work of
Vysatlrtha is being published by the Mysore University in parts.
But the part dealing with anumana (deductive reasoning) and tarka
(inductive reasoning) is still unpublished. had the good fortune
to make a study of this chapter in the manuscript and I have recorded
the results of my study of the chapter on tarka (inductive reasoning)
in the present work.
Lastly, I have made a dispassionate study of the speculations of
rharsa on the problem. The last three chapters are devoted to the
elucidation of Stlharsa's subtle analysis. I have bestowed close refleci Isasyai'sa nivesitah padayuge. ...NKu. ch. I. i. padayuge, padyate
jayate 'nene' ti vyutpatty pratyyakayuge pramnatarkarpe.
NKuP, p. 2.
tion and hard thinking on this part, which has proved too exacting at
times. But I have not allowed myself to be deterred by the obvious
difficulties of Sriharsa's analysis. I have sought to represent to the
modern philosophers of Europe and America, who may favour my
work with their perusal, what the astounding intellect of Sriharsa has
achieved. It will be admitted, I trust, that the contributions of the
Indian mind in philosophy are not inferior to those of Europeans.
Notwithstanding what has been done, the Indian mind, in the field
of philosophy at any rate, has not been thoroughly represented to the
modern world. It may not be a vain hope on my part that the
scholars of the world will not refuse it the study that it deserves. The
present condition of the world is not favourable or conducive to
philosophical speculation. People, including the custodians of culture,
are too much obsessed with economic worries. But this must be a
passing phase. Men cannot live in perpetual tension.
With the
advent of true peace and economic self-sufficiency our academies are
bound tp recover their wonted vigour and I rest in the hope that then
a thorough and alLout endeavour will be made to recover and to
reappraise the cultural heritage of India. The professional philosophers
of Europe had a cheap estimate of India's philosophies. The writings
of Sir Sarvapalli Radhakrishnan, late Prof. S. N . Dasgupta, Dr Satkari
Mookerjee and others must have proved the absurdity of these opinions. I am only following in their foot-steps and the present attempt
of mine, which gives a survey of the logical speculations of India on
a particular problem spread over at least 1500 years, s inspired by
the same faith which has characterized my predecessors. I have had
to study almost all the important problems of logic and some 03
metaphysics in the course of my dissertation and I hope that a modern
student will get a fairly comprehensive idea of India's philosophical
thought from this work.
I have made the study of tarka as complete and thorough as the
available data have made possible.
It will show that no problem,
whatever may be its position in the whole scheme of philosophy, has
received desultory attention from the philosophers of India. Indian
philosophers have been thorough-going all along in their study of
problems and the differences of philosophers have only served to shed
a flood of light upon all the aspects and bearings of each. This shows
that the Indian mind has never been languid or indolent in the past.
The cultural integrity and independence of India were not affected
in the least by the loss of political freedom prior to British rule. The
cause of this self-possession and balanced attitude must be found in
the fact that the Brahmins, who were the custodians and trustees of
the national culture, never ceased to dedicate their lives to the cultivation of their national literature and stood immune from the temptations of political career. It is rather in the British period that the
majority of the best intellects were weaned away from the study
of Sanskrit to the cultivation of English. The knowledge of English
culture was a great boon, but the total ignorance and neglect of the
nation's past cultural achievements almost entirely enshrined in Sanskrit has made the cultural make-up of the Indian graduates of modern
universities one-sided and deficient in balance. Imperfect knowledge
is responsible for imperfect sympathy and however much we may
regret the apathy arid studious indifference to India's past culture on
the part of the alumni of the modern universities, we must not bow
to this position of affairs as the inevitable dispensation of fate. It
appeared that the most convenient and expedient way to attract the
sympathetic attention of cultured Indians to the merits of India's
thought was to make a dispassionate presentment of the same in
English which is now occupying the position of India's cultural
language in the place of Sanskrit. The present dissertation is an
attempt in that direction.
The speculations on reasoning should be regarded by a modern
student of logic as entirely original contributions of Indian logicians.
I have tried to make them available to the academies of the modern
world as faithfully as possible. The elaborate foot-notes will go to prove
how far I have been successful. I have been progressively convinced
of the love of thoroughness and clarity of the old thinkers of India
and I have felt that comprehensive treatment of their subtle reflections
may appear to be too technical and too elaborate. But I have elected
to run the risk and I think that we must not be afraid of the temper of
the time, which may either be frayed by these subtle distinctions or
may fight shy of these logical exercises. It has however been rubbed
in on my mind by my revered guru, Dr Satkan Mookerjee, under

whose guidance I have worked all along, that the present-day Sanskrit scholars of India owe it to the world to lay bare the Indian mind
in all its subtle workings before the votaries of truth. The importance
of philosophical speculation does not, in my humble judgment, lie so
much in the truth achieved as in the attempts at the finding of the
same. Inspired by this ideal I have ventured to represent the thinking of India on what ought to be regarded as an important problem.
It is no doubt a study of only one problem of India's logic, but the
future edifice of Indian culture can be re-built only on the foundation
laid by the thorough study of its individual problems. Regarded
from this angle of view the present venture may be looked upon as a
contribution toward the fulfilment of the consummation, to be achieved in future by means of the accumulated labours of hosts of scholars.
PRONUNCIATION
SCHEME OF TRANSLITERATION OF SANSKRIT ALPHABET
Vowels
W a,
W 5,
i,
^ j,
IT
e,
| i,
ft
ai,
3 u,
3T ,
gft o,
m\ au,
Consonants
$ k, ^ kh,
*r g,
\ gh,
% ,
^ c,
sr j ,
Tft jh,
3^ ,
^ ch,
t,
5" d,
% dh,
*rr n, .
c t,
v\ th,
ch,
^ d,
q" dh,
^ n,
Tp,
T7 ph,
cf b,
*r bh,
q; m,
5 v,
^r s,
q^ s,
^ s,
rh or m,
: h.
h,
^j r,
ABBREVIATIONS
AS.
TV.
TVP.
TVD.
Bib. Ind.
ChSS
com.
CSS.
Din.
Ed.
K.
KhKh.
=
=
=
=
=
=
=
=
=
=
=
=
Laksanvali
M.
MB.
MS,
n
NB.
NK.
NKu.
NKuP.
NKuPM.
NL.
NLP.
NLPV.
NM.
NP.
NPa.
NS.
NSP.
Nsu.
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
Advairasiddhi with Laghucandrik (NSP)

tmatattvaviveka (Bib. lnd.)
tmatattvavivekapraksik (Bib. Ind.)
tmatattvavivekadidhiti (Bib. Ind.)
Bibliotheca Indica
Chowkhamba Sanskrit Series
commentary
Calcutta Sanskrit Series
Dinakarl with com. Rmarudri (NSP)
Edited by
Kiranavall (ChSS)
Khandanakhandakhdya with the com. of nandaprna (Ed. Laxman Sastri, ChSS)
(ChSS)
Muktvall (NSP)
Mahbhsya
Manuscript
foot-note
Nyyabhsya (CSS)
NyyakandalT (VSS)
Nyyakusumnjali (ChSS)
Nyayakusumajalipraksa (ChSS)
Nyayakusumjalipraksamakaranda (ChSS)
Nyayallvat (ChSS)
Nyyalilvatlpraksa (ChSS)
Nyyalllvatipraksavivrti (ChSS)
Nyayamanjarl (VSS)
Nyyaparisuddhi with the com. of Srinivs'a (ChSS)
Nyyaparisista (CSS)
Nyyastra (CSS)
Nirnaya Sagar Press
Niyayasudha (Ed. T. R. Krishnamacharya, Kumbliakonam)
jSJV.
NVTT.
NVTTP.
NVTTPP.
=
=
=
=
PNT.
PP.
PSED.
PPT.
RS.
==
Sakuntal
=
Sahityadarpana=
SDS.
=
SVR.
STV.
SI.
SV.
TCM.
TR.
=
=
=
=
=
=
TT. MS.
VV.
VSS.
VVr.
=
=
=
Nyyavrttika (CSS)
Nyyavrttikattpatyatika (CSS)
Nyyavrtcikactparyaukpansuddhi (Bib. Ind.)
Nyyavrttikatcparyatikparisuddhipraksa
(Bib.
Ind.)
Pramnanayatattvlloklamkra (Motilal Ladhaji,
Poona)
Pramnapaddhati (Ed. T. R. Krishnamacharya,
Kumbhakonam)
THE PRACTICAL SANSKRIT-ENGLISH
DICTIONARY (Apte) (Third Edition, 1924)
Pramnapaddhatitlk of Janrdana Bhatta (Ed. T.
R. Krishnamacharya, Kumbhakonam)
Rasasra (Ed. MM. Gopinach Kaviraj)
(Vani Vilas Press)
(Ed. Gurunath.. Vidynidhi)
Sarvadarsanasamgraha (Ed. M M . Vasudev Sastri
Abhyankar, Poona)
Syadvdaratnkara (Motilal Ladhaji, Poona)
Smkhyatattvavibhkara (ChSS)
Siokavrttika (ChSS)
Taccvacintmani with Gddhari (ChSS)
Trkikaraksa with the com. of Maliintha (Medical Hall Press, Benares)
Ms. of the Tarkatndava (belonging to the late
Prof. S. N . Dasgupta)
Vdivinoda (Ed. M M . Ganganath }ha)
Vizianagram Sanskrit Series
Vyomavatvrtti (ChSS)
CHAPTER I
ia the Jama epistemoiogy a definitive place has been accorded
co tarka as an independent source of knowledge of the universal
and necessary character of a connexion between two terms. In
fact inference is possible if there is knowledge of universal proposition
at its back. Now the problem of paramount importance and extreme
difficulty at the same time, which logic or rather the philosophy of
loîc has to solve, is the problem of the knowledge of the universal
proposition. The validity of the conclusion ultimately rests on the
validity of the universal proposition. But how can such knowledge
be acquired ? What is its source ? Certainly it cannot be entirely
empirical in character. Human experience severally and even collectively has its obvious limitations in that it is confined to a limited
number of instances in which the two phenomena are found to be
mutually associated.
But the universal proposition which is the
foundational datum and starting point of inference is categorically
universal ia its reference.
Even the relation of humanity and
mortality or of fire and smoke could not be ascertained as valid for
all times and all places if we were to depend entirely on the evidence
of our experience. Even if it is allowed that the knowledge of
one individual includes in its scope the knowledge of all the individuals of the class in so far as they embody the class-character
which is directly envisaged in the individual in question, yet the
character of necessity of the relation between the two terms which
is presupposed in the universal proposition would not be explained.
It has been contended by a late school of Naiyyikas that when a
person perceives an individual, which belongs to a class, he perceives
the individual as possessed of the class-character or the universal ;
and through his experience of the universal, he comes to have a
sort of direct intuition of all the individuals possessing that universal
as a necessary factor of their being. The fact that when a man
perceives a coin and realizes its value, he has no doubt that a second
INDUCTIVE REASONING
[CHAP.
coin of che same class possesses the same value. This proves that
through the knowledge of one individual coin and its value he has
realized the value of all the coins belonging to that category. For
instance, the knowledge that a particular rupee possesses the value
of sixteen annas is not confined to that individual coin only, but
extends to all the coins of the same denomination irrespective of
the differences that distinguish individual coins from one another.
It should be noted that this apparently simple cognition of a rupee
and its value is not a cognition of an isolated fact. In the first
place, the cognition of one rupee is to be understood on the contention of the Naiyyika as equivalent to the cognition of all rupees.
In the second place, the cognition of the value of an individual
coin as consisting of sixteen annas is also a cognition of such value
of all such coins. In the third place, the connexion of all the coins
with such value in all its cases is also cognised as a universal fact
through the cognition of such connexion in one rupee.
Now a question arises. Are the.;cognition of all the individual
coins and the cognition of the value of all the individual coins and
the cognition of the connexion of each individual coin with each
individual value three separate cognitions, or one cognition from
the very start ? It seems obvious that the third cognition which
takes stock of the connexion between all the individuals of the class
and their value is the result of the preceding two cognitions. Though
the same process may be at work in all these three instances so
far as the universal reference of the cognitions in question is taken
into account, it seems equally obvious that something more is needed
to account for the element of necessity which goes with the third
cognition. By knowing the value of one rupee we come to have
a necessary belief that all rupees possess the same value. Even if
it is admitted that the cognition of one individual carries with it
the cognition of all the individuals, that can at most account for
the element of universality and not the element of necessity both
of which however are presupposed in the universal proposition in
order that it may serve as the condition o: inference.
We started with an enquiry into the possibility of the knowledge
of the universal proposition and we have considered the theory
,i
of the Naiyyikas that the cognition of one individual of a class

irives insight into all the individuals of that class through the
medium of the universal that characterizes each and all of them.
We have adumbrated that even if we admit the Naiyyika's contention to be valid, the belief in the necessity of the connexion of two
facts, which is posited in the universal proposition,
c | u s s e s of
requires something more to justify it. A cognition o an individual
is at most the cognition of a contingent fact and however much its
scope may be extended it cannot rise above the level of its assertory
character.
In other words, such a cognition can enable us to
assert that A is B, It cannot guarantee that A must be B.
The implication of the universal proposition which serves as the
major premise is not exhausted by the element of the universality,
but the element of necessity is also included in its fold. Now, this
element of necessity, which can be satisfied only by an unwavering
assurance of the impossibility of its being otherwise, cannot be
afforded by any ordinary cognition, perceptual or the like.
Certitude of knowledge is capable of being attained by the
elimination of contrary possibilities. In other words, knowledge is
hindered from attaining the level of certitude if there is a doubt
about the possibility of its being otherwise.
Doubt means the
oscillation of the mind between two or more alternatives, which can
be removed only by the demonstration of the conflicting alternatives
as impossible, and the consequent determination of the knowledge
to only one alternative. It is a question of paramount importance as
to what is the instrument that makes such elimination of the contrary
alternatives possible.
It is common knowledge that even such
accredited instruments as perception and inference are thwarted by
the emergence of doubt from yielding; their ultimate deliverances.
In perception of a particular entity standing erect before our tyes
the mind is perplexed by a doubt whether it is a tree or a human
being that is being perceived. Unless the doubt is removed knowledge
cannot take place. And even in cases of perception where there
is no initial doubt, its validity is liable to be called in question by
the arousal of a doubt on the analogy of a false experience that
happened in the past. Take, for instance, the perception of a
INDUCTIVE REASONING
[CHAP.
stream of water in the desert, which may be veridical. But even

this veridical perception is assailed by a doubt of its authenticity
on the analogy of past experiences which proved to be false. Of
course, verification furnishes the necessary corrective in cases where
ic is possible. But in cases which are not capable of verification
we have to discover reasons for determining the validity of the
belief. The issue can be briefly put as follows. Are the so-called
sources of knowledge capable of giving us authentic knowledge
or only tentative beliefs, which require to be proved as valid before
they can be accepted as knowledge proper? The answer of Indian
philosophers has been that doubts, whether frustrating the operation
of the instruments of knowledge or assailing the validity of the
resultant cognition, are rather of the nature of accidents to be
looked upon as thwarting agents. They cannot be construed as
evidence of intrinsic inefficiency of the instruments of knowledge
or of the inherent invalidity of their resultant deliverances. Doubt
has to be overcome in order that knowledge may occur or become
effective. How can the removal of the doubt be secured? Herein
tarka comes in request. We propose to render henceforward tarka
as reasoning, though we are not absolutely sure of the justice and
adequacy of the English equivalent. But the definition of tarka
as propounded by the traditional philosophers will also apply to the
term 'reasoning', and this will sufficiently delimit the connotation of
the latter and thus make undesirable extension impossible.
That reasoning {tarka) serves to eliminate doubt and thereby
to help the determination of truth is recognised even by the
author of the Nyyastra. Reasoning has been defined to be an
intellectual act which contributes to the determination of truth by
means of adducing logical grounds (in favour of one of the alternative
possibilities), when the reality is not known in its proper
character1.
Vtsyyana observes that when two contradictory
alternatives present themselves, as equally possible with regard to
a particular subject of enquiry and the mind is made to oscillate
i
avijtatattvc'rthc kiiranopapattitas tattvajfmrtham filias tarkah.

NS, L i. 40.
!]
between them, the consideration
of
reasons
alternative helps to resolve the deadlock.

of
doubt resulting
in the
elimination
uucontradicted alternative is taken
m support of one
With the
elimination
of one of the terms the
to be the truth 1 .
To revert to
our old example of the tree and the man, the object lying ahead is
determined
to be a human being if the characteristic features, e.g.
hands and feet, are cognised.
This cognition of attributes, which are
possible only in one of the two conflicting
alternatives, serves to
clinch the issue, and we at once conclude that the object ahead is
a human being.
It is obvious that both the alternatives cannot be
true because of their mutual
incompatibility.
the process consists in the recognition

alternatives being true.
and
subsequent
of
So the first step in
the impossibility
The second step lies
of
both
in the search for
discovery of logical grounds which are possible
only in one alternative.
The final result is the determination of
the subject of dispute being this (a human
being) and not that
(a tree).
Now, reasoning has not been regarded as an independent
ment of knowledge
by
the Naiyyikas.
instru-
But nevertheless, they
acknowledge the efficacy of reasoning as contributory to the determination of truth.
Reasoning, it is argued, consists in the differentiation
of the possible alternatives and
consideration
which serve to distinguish one
alternative
of logical grounds,
from
the remaining
possibilities as the most reasonable, and accordingly determines our

preference for this.
Reasoning, thus, helps to facilitate the operation
of the relevant instrument of knowledge with regard to one of the

alternative data by putting it on a sound logical basis.
determination of the nature of the datum is effected
instrument, perception or inference
The actual
by the relevant
etc., and reasoning only
paves
the way for its free activity .
1 atha jijnsitasya vastuno vyhatau clharmau vibhagena vimrsati kimsvicl

ittham hosvin nettham iti. vimrsyamnayor <lharmaytr ekam kranopapatty
'niijniti, sambhavaty asmin kranam pramnam hetur iti kranopapatty sycl
evam etan netarad iti.
NB, pp. 320-21.
2 NV, pp. 326-27.
INDUCTIVE REASONING
[CHAP.
Vacaspati Misra is in complete agreement with Vtsyfiyana and

Uddyotakara with regard to the status of reasoning as an auxiliary
faccor to che independent instruments of knowledge1. In the preceding paragraph we have considered the view o Uddyotakara that
reasoning helps to determine between two competing alternatives on
the superior strength of one of them. Reasoning, thus, indicates that
one of the alternatives is logically impossible and by the method
of elimination the remaining alternative is the possible truth
Vacaspati lays stress upon this aspect of elimination as the true
contribution of reasoning.
Reasoning is thus the redactio ad
absurdum of the competing possibilities. It comes into play only
when there arises a deadlock on account of the presentation of
two contradictorily opposed possibilities as claimants for truth.
Reasoning enables us to conclude, in the first instance, that both \
these alternatives are not logically possible. In the second instance, i
by putting forward considerations against one of the alternatives it i
enables us to reject what is shown to be logically absurd. The
determination of truth is automatically accomplished by the relevant
instrument of knowledge which was hitherto kept in check by the
doubt arising from the prima facie claim for truth by two conflicting
alternatives.
So according to Vacaspati reasoning serves to
demonstrate the absurdity of the truth-claim of the false competitors
and this demonstration of absurdity subserves the purpose of the
instrument of knowledge. Reasoning comes to be requisitioned when
there is an enquiry of truth and subsequently the appearance of
conflicting possibilities creates a doubt about the real issue. Here
reasoning comes to the aid of the cognitive instrument by removing
the doubt. The removal of doubt is effected by the consideration of
overwhelming grounds, which prove the absurdity of the false
alternatives.
The determination of truth, as has been observed
before, is effected by the relevant cognitive instrument and reasoning
has nothing to contribute to the positive finding. Its services and
functions are thus of a negative character. Reasoning helps the
i The instruments of knowledge, according to the Naiyyika are only fourj
viz. perception, inference, comparison and verbal testimony.

,f rhr false alternative and this rejection
fCJccciou or cnc ^^
of the contradictory
1
ative is interpreted as its positive service towards the determination of truth 1 .

Udayana endorses the position ot Vacaspati that reasoning serves
eliminate the opposite issue by the demonstration of absurd
equences, which follow from the opponent's position as a matter
f logical necessity. To take an example, let us suppose that the
$<; tic questions the causal efficiency of water for allaying the pain
1 jrsc
How can it be proved that the sceptic's position is wrong?
Udayana would resort to reasoning. He would argue, *lf water did
allav thirst, nobody would drink water when thirsty. As a matter
of fact, people drink water to allay thirst and nobody has been
rjkapointed in his expectation'*. Here as elsewhere reasoning assumes
the form of a reductio ad absurdum. In the instance cited, the
tiductio ad absurdum serves to demonstrate that the doubt or denial
hvould entail rejection of an admitted truth. Another variety of it
^consists in the enforced admission of an issue, contradicted by an
Accredited instrument of valid-rcognition. Thus, if one were to argue
thac drinking of water causes burning sensation, the absurdity of the
position can be demonstrated by an actual experiment. One drinks
water and is satisfied that it does not produce any such effect. The
position of the opponent is thus invalidated by the necessity of
admitting a fact contradicted by valid cognition. Reasoning thus is
of two types according as it makes the admission of an absurdity or
the rejection of an accepted truth inevitable2. In spite of the fact
that reasoning in the shape of reductio ad absurdum enables us to
reject a false conclusion, it is not regarded as a form of inference
or an independent source of knowledge by the exponents of the Nyfiya
school of philosophy. Reasoning is regarded as an instrument of
1 yasmin visaye pramnam pravartitum udyatam tadviparyaysankyam na
tavat pravartatc, na yvad anisc.patty viparyaysank* paniyatc, taclapanaya eva
ca svavisaye pramnasambhava iti do'papattir iti vykhyyate.
NVTT, p. 321.
2 anistani ca dvividham prmyriikapantygo prmnikaparigrahas ca. yatha
yady udakam pipsduhkham na samayet na piyeta, yadi ca tad eva param an tar
dalict, tadS avisistam mam api dahct.
ATV, p. 538.
INDUCTIVE REASONING
[CHAP.
negative value in that it only serves to show the absurdity of the

opposite alternative and has no direct bearing upon the determination
o positive truth. The positive truth is determined by an accredited
cognitive instrument. The resultant rejection of the false issue is
made possible by the demonstration o the absurdity o the opponent's
position which is made the ground ior the inference o such absurdity.
But though reasoning satisfies the formal conditions o inference inasmuch as it is based upon the necessary connexion between the
opponent's position, which serves the purpose of the logical ground,
and the absurd issue, which follows as a necessary consequence, it is noc
regarded as a means of proof of an actual situation. The inference
is of a hypothetical character and hypothetical inference is noc accepted
as a valid instrument by Indian logicians. To return to our old
example, "if water did not allay thirst, no man would drink it for
the purpose".. The logical ground is furnished by the antecedent
clause of the hypothetical proposition. The argucr knows that it
is false. He is absolutely sure that water allays thirst and the denial
of it is untrue. The untruth of the position is demonstrated by the
bsurd consequence which follows from it. The consequence follows
a
as a matter of undeniable logical necessity, because it is necessarily
bound up with the condition. Indian logicians believe only in the
material truth of the conditions of inference. Although the necessary
relarion between the ground and the consequent in the hypothetical
proposition is a material truth, the ground is only assumed for the
sake of argument. The ground 'the incapacity to allay thirst' does
not belong as a matter of actual fact to the subject, 'water', and
thus the minor premise in the proposed syllogism is false. But it is
admitted purely for the sake of argument and it serves as a condition
for the demonstration of the absurdity inherent in the opponent's
opposition. Though its services are valued, reasoning as a kind of
hypothetical argument is not admitted to be an accredited source of
valid knowledge inasmuch as it lacks material truth 1 . We shall devote five separate chapters to the consideration of this problem and we
i NVTT, p. 54; vide also NV, p. 32S.
,]
shall find that Indian logicians are by no means agreed in their views
of che logical value of this important organ of philosophical
speculation.1
We now propose to resume our discussion of reasoning as an
active organ for the resolution of doubt, which has been found to create
a logical deadlock. Vacaspati maintains, we have found, that reasoning serves to remove the doubt which the presentation of conflicting
alternatives entails. There is hardly any room for doubt that reasoning contributes to the elimination of doubt and thus removes the
obstacle to the operation of the relevant instrument of knowledge2.
Let us take a concrete example. The Naiyyika holds that word is
perishable, but the Mlmnisist maintains that it is eternal. Each
advances reasons in support of his position and the conflict of the
philosophers is bound to create a doubt about the nature of word,
and thus a deadlock emerges. Here reasoning is called into requisition and by the consideration of the logical consequences of the'
grounds advanced by the philosophers, it helps us to see whether the.
particular position is in conflict with the established convictions of
mankind. If the logical pursuit of the consequences of the ground
advanced by one philosopher ends in demonstrating
the
incompatibility of the results with accepted truths, his view is
rejected. This is the broad survey of the way hi which
reasoning works and- of the ultimate results achieved with
its help. But philosophers are not satisfied with a broad view
of things and they endeavour to assess the exact contribution of
each factor. Udayana, though he is a commentator of Vaeaspati,
does not agree with him with regard to the actual results that are
achieved by reasoning, Vacaspati thinks that the service of reasoning
lies in the elimination of doubt. But Udayana here differs. Udayana
thinks that doubt is due to the incapacity to find out conclusive
reasons in favour of one of the conflicting alternatives, and these
reasons are nothing but the characteristics which belong to one
1 Vide1 chapters iv-viii.
2 anistam prasajya viparyayasy aiva sksn nivarttanc
pramnasy
bhyanujntasya visodhitc visaye pramnam apratyham pravartate.
NVTT, p. 321.
lo
INDUCTIVE REASONING
[CHAP.
and one only of the alternatives. To revert to our old example

of the doubt about the tree or man, it is certain that the objecc
lying in the front can be determined to be a human being only by
the knowledge of such characteristics, as hands and feet, which can
belong to a human being and not to a tree. But does reasoning
give us the knowledge of the crucial evidence? No, it is perception which supplies these data and the knowledge of the data, hands
and feet, removes the doubt1.
Udayana's position has been elucidated by Bhatta Vadindra.
Doubt, it is argued by him, is due to the lack of knowledge of the
specific characteristics of the real entity, which becomes the object
of doubt owing to the presentation of such common characteristics
which can belong to a human being and to a tree with equal degree
of likelihood.
The presentation of common features is not by
itself the cause of doubt, but only when it is accompanied by the
lack of knowledge of the specific characteristics. Even when it is
determined to be a human being, the common characteristics, which
it shares with trees, such as straightness, height etc. are also perceived,
but the doubt ceases by reason of the knowledge of hands and feet,
which are the specific characteristics of a human being. It follows,
therefore, that doubt is due to the absence of knowledge of specific
characteristics, and it can be removed only by such specific knowledge.
But as reasoning does not yield such specific knowledge, we have
to depend upon perception or the like for the elimination of doubt2.
Udayana for this reason abandons the position maintained by
Vcaspati and seeks to account for the bearing of reasoning upon
the determination of truth in a different manner. Reasoning is
incompetent to remove doubt as the contents of doubt and reasoning
are not opposed. The opposition of judgments is due to the opposition of their concents (yisayavirodha). Now, reasoning undoubtedly
comes into play when there is doubt. It seems paradoxical to
assert that reasoning fails to remove doubt, though doubt is the
occasion for the emergence of reasoning. It is not a fact that
1 KV, p, 260.
2 RS, p. 94.
,i
Udayana is not conscious of this' broad truth that reasoning has a

n^ on the elimination of doubt. But Udayana is too minute a
thinker not to recognise che absence of opposition between reasoning
and doubt in respect of their contents and hence probes deeper into
the problem. For instance, a man entertains doubt of the existence
of fire n a place where he observes smoke. The doubt is of the
form "is the hill possessed of fire or not?" Here reasoning arises
and expresses itself in the form of a hypothetical proposition, namely,
"If the hill did not possess fire, it would be devoid of smoke".
Certainly, the two propositions (doubt and reasoning) are not on a par
in respect of the opposition of their contents. The subject-matter of
doubt is the existence or non-existence of fire, and reasoning asserts
that the absence of fire would entail as a necessary consequence the
absence of smoke. Now, the absence of the smoke is not opposed to
the presence of fire. Absence of fire is opposed to presence of fire
and absence of smoke is opposed to presence of smoke. So the
consequence, namely, the alleged absence of smoke being not antagonistic to the presence of fire, the reasoning in question cannot
logically rebut the doubt of the presence or absence of fire which
made the appeal to reasoning necessary. What is then the instrument for the removal of doubt? The answer is, it is the knowledge
of the specific characteristic of one alternative. In the instance cited
above, the perception of smoke as the concomitant of fire is the instrument of the removal of doubt of the presence of fire. Smoke is a
specific characteristic of fire, being its necessary concomitant. Doubt
of fire was engendered by the absence of knowledge of smoke as the
necessary concomitant of fire. But as soon as the knowledge of
smoke-as-concomitant-with-fire takes place the condition of doubt is
removed. It has been stated that the cognition of common characteristics and the absence of cognition of the specific characteristic of one
of the alternatives are the conditions of doubt. As has been observed
before, the cognition of common characteristics occasions doubt only,
when it is accompanied by the lack of the cognition of a specific
characteristic, and not when the latter is present. So it is the latter
condition that is the direct cause of doubt. In the particular case
under consideration, the recognition of smoke as the concomitant of
i2
INDUCTIVE REASONING
[CHAP.
fire removes the absence of the cognition of this fact, which is the
specific characteristic of fire, which is the subject-matter of doubt.
Thus, doubt of fire is removed not by the knowledge of absence of
smoke entailed as a consequence by reasoning, but by the recognition
of smoke as the concomitant of fire, which is entirely different from
what is entailed by reasoning.1
But now a question arises. If reasoning has no bearing upon
the elimination of doubt which is the raison d'etre of its coming into
operation, what is then the nature of its service? Udayana answers
the problem in the following manner. He thinks that reasoning
serves to remove the desire for knowledge of the opposing alternative
and not doubt. Though doubt is a necessary condition of reasoning,
it is not its universal condition. Reasoning is requisitioned even
when there is no doubt. For instance, a hungry man sets to eat a
plate of food and a friend warns him that the food is poisoned and
will prove fatal if consumed. Now, this warning, if logically stated,
should be expressed as follows: "If you take the food in question,
you will die", which is the form of the statement in which the reasoning is necessarily expressed. The warning here is a reductio ad
absutdum which is the usual form of reasoning. Here no question
of doubt arises, but nevertheless the employment of reasoning is effective. What is its effect? The answer is that reasoning here removes
the hungry man's desire for the consumption of the food under consideration by showing the undesirable consequences that would necessarily follow the act. That reasoning serves to demonstrate an . undesirable consequence of the acceptance of an alternative course of action
has been stressed by Vcaspati Misra. Udayana accepts this finding
of Vacaspati; but whereas the latter makes such demonstration of an
undesirable contingency the instrument for the elimination of doubt,
Udayana, in conformity with the instance of the poisoned food
where reasoning comes into play even in the absence of doubt, asserts
that the actual result produced by reasoning is the removal of desire
for the opposite course of action only. Even where doubt furnishes
the occasion for resort to reasoning, the service of reasoning consists
i
See RS, p. 92.
,j
13
in the removal of desire for the opposite alternative, though removal

of doubt ensues as a remote consequence. In the case of doubt about
che existence or non-existence of fire, reasoning serves to remove the
desire for -knowledge of the non-existence of fire by demonstration
The
of ehe undesirable consequence, namely, the absence of smoke.
doubt of the possibility of the non-existence of fire is removed by
the knowledge of fire through the knowledge of the presence of
smoke-as-concomitant-of-fire.1
A dimtulty has been raised in this connexion. If opposition of
contents be the determinant of logical opposition between two propositions, the opposition of reasoning to the desire for knowledge cannot
possibly be regarded as a case of logical opposition. If the opposition
in question be looked upon as a case of causal opposition as illustrated
by the opposition of a disturbing noise to study, there is no reason for
combating the view of Vcaspati Misra regarding the opposition of
reasoning to doubt, if observation of concomitance in presence and
absence between reasoning and "cessation of desire is appealed to as
proof of the opposition, the same can be availed of by an adherent of
Vacaspati's view. That doubt is removed, if there is reasoning as an
antecedent fact, is a matter of observation.
Udayana's objection was
to the effect that there was no logical opposition between doubt and
reasoning.
But he gives up the plea of logical opposition when he
makes reasoning the condition of the removal of desire. Besides desire
has no logical value. If the opposition be shifted to one between
reasoning and the judgment which is the cause of desire, it may be
given a logical character.
The judgment 'poisoned food is fatal to
human life' is the cause of the cessation of desire for such food. Now,
what is desired is, in the first place, something which is desirable on.
its own account, for instance pleasure and the like; and, in the second
place, a thing is desired on account of its consequences which are
1 aniyatakctyupaniptc niyatakctiparigrahasya yadyarthatvt
katham
tarhi pramnam apy anugrhniyt aniyatecchviccheden ety uktam. n,a hi kurym
na kurym ity aniyatacikirsah kart kicic karoti. aniyatccchvicchcdam katham
kuryd iti cet? anistpattinivrttatvcl asya. yath savisnnabhaksanotlyatasya
savisam idam yadi bhaksayisyast marisyasi ty anistpattcr na bhaksayeyam iti
tatsmaranam eve cchvicchittih. KV, p. 260.
i4
INDUCTIVE REASONING
[CHAP.
desirable for their own sake.

Thus wealth is desired because it is a
means of securing happiness. Similarly aversion or cessation of desire
has reference to a thing, which is undesirable on its own account, or
which is an instrument of an undesirable consequence. The relation
between cessation of desire and knowledge of an undesirable consequence is thus necessary and universal. When reasoning demonstrates
an undesirable consequence to follow from a belief or a course of action
based upon such belief, the opposition of the latter to desire for entertaining such a belief or consequential action is only an instance of this
universal truth. So the opposition is not on a par with the opposition
between a frustrating agent and an expected happening. The opposition of reasoning with doubt as Vacaspati maintains would be a case
of specific causality which was not deduced from a general proposition.
But the opposition of the untoward consequence, shown by
reasoning with desire for the belief or the action proceeding from such
a belief, is one which is deducible from the general proposition enunciated above, viz. 'that which is undesirable per se or produces an
undesirable consequence is the object of aversion.1 Thus Udayana's
contention seems to be more cogent than that of Vacaspati Misra.
What is then the cause of the removal of doubt according to
Udayana?
The question can be answered by an analysis of the conditions of doubt. To take a concrete example, let us suppose that a
man sees smoke issuing from a hill, who would naturally infer that
smoke was due to fire.
But meanwhile a doubt about the necessary
relation between fire and smoke crops up and prevents the inference.
The doubt is of the form that the smoke in question may be possible
in the absence of fire. Here reasoning comes to the aid. The man
argues that if smoke be not the effect of fire as it is not known to be
the effect of anything else, the smoke would be an uncaused effect
which is an absurdity. The realization of this absurdity removes the
desire for belief in a fireless smoke.
In other words, the belief or
rather the desire for belief in the possibility of fireless smoke which
created a deadlock is eliminated by reasoning and on the removal of
this impediment, the natural inference that the hill is on fire takes
place without a hitch. And the inference of fire removes the doubt.
Thus according to Udayana doubt is always eliminated by definite
,j
15
knowledge which accrues from the relevant conditions of knowledge

called pramna and the reasoning is called into request to eliminate
an obstacle, which always is m the form of a diversion caused by an
enquiry after a false possibility1.
1 We have taken sufficient pains to elucidate Udayana's position here which

deserves scrutiny firstly on account of its originality, and secondly because it has
not been properly represented even by authors of repute. Varadarja, die author
of the Trkikaraksa, first correctly states in, the Krika tiie position of Udayana
that reasoning is of service in that it eliminates desire for the opposite belief.
But in his comments he surrenders this position and asserts that reasoning is the
instrument for the removal of doubt of the contrary possibility, which is the
position of Vacaspati MLsra and has been elaborated by Srivallabha, the author
of the Nyyalllvutt. It is painful to observe that Mallinfitha, the commentator of
the Tarkikaraksa, has made the confusion worse confounded. Vide Trkikaraksa,
Chapter I, pp. 196-97.
CHAPTER II
THE POSITION OF SRTVALLABHATHE AUTHOR OF THE
NYYAULVATi
After Udayana Srlvallabha has devoted elaborate consideration to
the problem as to how reasoning comes to the aid of an accredited
instrument of knowledge. Srivallabha comes to the conclusion that
reasoning resolves doubt which frustrates the operation of an instrument of knowledge. Udayana's position has been subjected to an
elaborate criticism. To start with, rvallabha tentatively suggests
that the application of reasoning may result in one of these three
consequences, namely (i) che cessation of desire for belief in the
opposite (yipaksajijasanivrtti): (2) cessation of the doubt about the
opposite (sankinivrttt): (3) the weakening of the opposite alternative
(tannynata). Now the first alternative is not entertainable. There
is absolutely no necessity for the cessation of such desire AS it does
not and cannot operate as an obstacle to the operation of an instrument
of knowledge1, in fact desire has no bearing upon the resultant
knowledge which follows as a matter of necessity when the conditions
of such knowledge are present. For instance, a man desiring to
perceive a jar perceives in fact a piece of linen because it is there.
The desire for knowledge of a different-thing here does not prevent
the knowledge of another thing. Nor can it be maintained that
such a desire would be an obstacle in the case of inference, because
the combined knowledge of the major premise and the minor premise
is all that is necessary for inference, and desire for one or the other
kind of knowledge has nothing to do with it. It might be maintained
with some plausibility that desire for the opposite knowledge serves to
frustrate the enquiry for the proband urn, the knowledge of which is
achieved by inference. So here reasoning comes into request and by
eliminating such obstructive desire makes the realization of inference
1 nanu kim etasya phalam? vipaksajijasimivrttih saiikfmivrttir v tannynat
v? ndyah. anumnd eva tatsicldhch.
NL, pp. 514-15.
n|
THE POSITION OF SRIVALLABHA
t;
possible. But this advocacy of the utility of reasoning does not seem
to be based upon an undeniable logical or psychological necessity.
The immediate cause of inference is always the combined knowledge
of the major and the minor premises and this has nothing to do with
desire. The position could be accepted if it were shown to be the
case that desire for knowledge of the proband urn was the necessary
condition of inference. If this were the case, desire for the opposite
knowledge would certainly operate as a bar to the materialization of
inference. But such a desire cannot be made out as the universal and
che necessary condition of inference, because there is such a thing as
unpremeditated inference, as in the case of the inference of clouded
sky from hearing a thunder-clap. So reasoning cannot be of use by
che removal of the desire for knowledge of the opposite, which is the
position of Udayana1.
The second alternative is also rejected by Srivallabha. The doubt
about the probandum is resolved by inference. If a reasoning were
co effect the removal of doubt there would be no necessity for resorting
to inference. Nor can reasoning be supposed to effect the certitude
of the removal of doubt achieved by inference. For inference is not
only competent to eliminate doubt, but also to produce an assurance
of such removal2.
The third alternative is equally unworthy of acceptance. Even
if it is granted that reasoning serves to weaken the possibility of the
false alternative being the truth, the result would be at most a
heightened presumption for the true alternative. But this would not
amount to certitude unless the possibility of the weakened alternative
is removed with absolute certainty. This can be done only by inference. If inference is necessary for the elimination of the attenuated
doubt about the false alternative, it can with equal efficiency eliminate
even the stronger doubt. Such being the case, the necessity of reasoning for the diminution of the strength of doubt is not only problematic, but also superfluous. In this way the three consequences alleged
to be the results of the application of reasoning are rejected by Sr1 Vide sufra, p. 12.
2 na dvitiyah. anumnavaiyarthyt. niscayrtham tad iti cct? na. tata cva
sanknivrtter api siddheh.
NL, pp. 516-17.
18
INDUCTIVE REASONING
[CHAP.
vallabha1. It deserves to be noted that che first view rejected by

Srvallabha is the position of Udayana, which we have explained already
at length2. The second view criticized is the position of Vficaspati
Misra. Vacaspati has explicitly stated that a cognitive instrument
is thwarted from producing certitude of its object, if the doubt of the
contradictory of the object of proof crops up in the mind. But when
this doubt is resolved by reasoning by demonstration of absurdities in
the opposite possibility, then the cognitive instrument in question
succeeds in achieving its objective3. The third view seems to be
representative of Jayanta Bhatta's position. Jayanta Bhatta has stated
that reasoning serves to produce a strong presumption in support of
the probandum. The implication of this strong presumption is the
diminution of the probability of the opposite alternative, which is
criticized in the Nyyatilava4'.
Srlvailabha shows further that reasoning cannot be regarded as the
necessary function of inference. Inference is certainly an instrument
of knowledge. An instrument is defined to be that which exercises a
function or operation. Function or operation is again defined to be
one the presence of which is necessary for the production of the effect,
for which the instrument is resorted to. It follows that an instrument
cannot produce its effect if it is destitute of exercise of the same
function. An example will make the position clear. An axe is an
instrument for cutting a tree. But it can produce the effect only
when it is operated by a person with a firm grasp of the hand. This
manual exercise is the operation of the axe. Certainly without this
operation the axe cannot produce severance of the tree. Inference,
as an instrument, is nothing but the probans found in the minor term
understood as necessarily related to the probandum. The understanding in question is the operation of this instrument. It is a perversion
of truth to regard reasoning as an operation of such an instrument,
1 na trtiyah. nyiiay atiyny v anumnanivartaniyatvena tannynatvpdanavaiyarthyt.
NL, p. 517.
2 Vide sufra, pp. 7-15,
3 Vide sufra, p. 6.
4 sandigdhc'rtlic'nyatarapaksnukfdakranadarsant tasmin snmbhvanpratyayas tarkah.
NM, p.. 8.
19
since the absence of reasoning does not thwart the production of the
effect, namely, the knowledge of the conclusion, provided the understanding of the necessary relation of the probans with the probaruUim
and of the subsistence of the same probans in the subject is there.
Ao-ain, reasoning in the absence of such an understanding is not
competent to produce the knowledge of the conclusion. So reasoning
cannot have any efficiency as a function of the instrument of inference. This applies mutatis mutandis to the other instruments of
knowledge1. It is worthy of notice that the criticism of this view is
directed against Vacaspati Misra and Udayana who have asserted
reasoning to be the function of a cognitive instrument 2 .
After having criticized the views sponsored by the previous
thinkers, rivaliabha emphatically maintains that reasoning has no
bearing upon the subject-matter of cognitive instruments by means
of elimination of doubt regarding it. He is positive that if an instrument is at work doubt of the opposite (yiparyayasanka) or a diversion
of enquiry after the opposite (yipaksajijsa) cannot prevent materialization of the resultant knowledge. In other words, "'he thinks Vacaspati and Udayana to be in tfte wrong, and he rejects the possibility
of reasoning serving as a condition for elimination of obstruction
of the subject-matter, which is called visayaparisodhaka tarka3.
Reasoning is conventionally classed under two heads, namely, (1)
that which removes an obstacle to the clear understanding of
the subject-matter (yisayaparisodbaka); (2) that which is conducive to the knowledge of the universal concomitance between
the probans and the probandum (yyapatigrabakaf. The first variety
of reasoning sponsored by Vacaspati and his followers is rejected by
Srivallabha.
According to him reasoning is of service in the elimination of doubt about the infallibility of the relation between two
terms. Take, for instance, the relation of smoke and fire. One may
1
2
3
4
NL, pp. 517-18.

NVTT, p. 54; vide also KV, p. 260.
NL, p. 518.
tarkas ca dvividho visayapsrisodhako vyptigrhakas ce ti.
Din, p. $\\ vide also NLP, p. 5.
20
INDUCTIVE REASONING
[CHAP.
naturally doubt that the observed association of fire and smoke is an

accidental coincidence. Now this doubt is resolved by reasoning in
the following way Smoke is an event and as such muse have a cause
and consequently must be related by an indissoluble bond with such
a cause.
What can be the cause o smoke? The application of the
Joint Method proves chat smoke has fire for its cause, and so the
relation between the two cannot be accidental.
Reasoning takes a
hypothetical form, namely, 'if smoke were not necessarily concomitant
with fire, it could not be its effect.' If it were not the effect of anything else than fire, nor again of fire, it would be an uncaused event,
which is an absurdity. Thus, reasoning by demonstrating the
absurdity of the opposite possibility helps the determination of the
observed relation between the probans and the probandum as an
invariable and essential case of ontological necessity1.
rlvallabha's views on reasoning can be summed up in brief as
follows.
Of the two types of reasoning sponsored by the ancients,
Srlvallabha only recognises the necessity of the second type. It is
worthy of remark that Srlvallabha has exercised a profound influence
on Gagesa, the author of the Tattvacintamani, who has also acknowledged the necessity of the second type of reasoning only. Now the
reason for rejection of the first type consists in the consideration that
doubt about the subject-matter cannot be an obstacle to the deduction
of the conclusion.
It has been shown that desire for the opposite
knowledge does not obstruct the process of inference from leading to
the conclusion (sadhyajna): Doubt.about the probandum also cannot
operate as a bar to inference, if the knowledge of the probans-asconcomitant-with-the-probandum is at its back.
In one word, doubt
cannot frustrate the realization of inference, if its conditions are at
work. And if there be any deficiency in the cause, which is the sum
of conditions, then also the failure of inference will be due to such
deficiency irrespective of the presence or absence of doubt. The
later Naiyayikas call this doubt about the probandum as doubt regarding the object of knowledge (grabyasamsaya).
Now this doubt
about the object, which is to be established by inference, is asserted
i
NL, p. 518.
nj
THE POSITION OF SRVALLABHA
21
co possess no obscructive efficiency, and this is rather an endorsement of

che position maintained by Srvallabha. rivallabha's position seems
co be based upon strong foundation. Let us elucidate it by a concrete
example. Suppose, a man has a doubt about a tyrant whether he is
liable to die or not.
Suppose, again, that the arguer is convinced of
the truth beyond the shadow of doubt that all men are mortal and
that he recognises that the tyrant is a human being. The doubt about
the mortality of the tyrant in question will not prevent him from
reaching the indubitable conclusion, "The tyrant is mortal." The
doubt is resolved by the result of inference and, though unresolved
before, it could not frustrate the operation of inference, rvallabha
accordingly asserts that there is no necessity for appeal to any other
aaent save and except the conditions and resultant of inference for the
removal of doubt or of a desire for knowledge of the contrary, simply
because they do not interpose any obstacle in the way of inference.
But though he rejects the first type of reasoning, which is supposed
by the previous thinkers to be the necessary condition of the elimination of doubt or a contrary desire, on the ground of its incapacity,
he still admits the necessity of reasoning as a condition of the elimination of doubt about the necessary and universal concomitance1.
But a question pertinently arises in this connexion. Necessary
concomitance (vypti) is nothing but the impossibility of the connexion
being contingent.
But the doubt of contingency of the relation
between the probans and the proband urn may be of the nature of the
doubc about the object of knowledge. The problem is how and by
what means the knowledge of the necessary relation, in one word,
induction, Is secured. The desired object of knowledge is the absence
of the fallibility and the contingency of the relation. And so a doubt
about this absence of contingency is on-a par with the doubt about
the objective of knowledge, which has been declared by Srivaliabha
to be devoid of all obstructive capacity regarding the realization of the
resultant knowledge sought to be achieved.
This doubt, therefore,
being absolutely innocuous, does not require the service of an agent
for its elimination. So reasoning of the second type also being on the
1 NL, p. 517.
22
INDUCTIVE REASONING
[CHAP.
same footing with the first cype should be equally rejected by

Srlvallabha1.
The truth of the matter regarding the capacity of doubt for
obstruction or otherwise is to be determined by the evidence of experience. Now, an initial doubt about a truth does not prevent the
realization of the truth by the pursuance of the relevant conditions of
proof. There may be a doubt about the infallibility of the concomitance between any two terms, e.g. between smoke and fire. This
initial doubt, far from being an obstacle to the pursuit of the condition of knowledge of concomitance, namely, the observation of concomitance in agreement and difference, is rather an incentive to the
adoption of the procedure. But the doubt, that may arise regarding
the validity of the knowledge of the necessary concomitance reached
by the pursuit of the Joint Method, is certainly an impediment.
Thus, for instance, the necessary concomitance of smoke and fire may
be known in the form, "Smoke is not concomitant with what is notfire, that is to say, vth what is different from fire." But the truth
of this knowledge may be assailed by the doubt about the possibility
of the occurrence of smoke in a fireless place. Now, such a doubt
can be resolved only by a reductio ad absurdum of such a possibility.
Here reasoning is of service. Moreover, the assurance of the necessary relation may be secured only by a lack of sufficient scrutiny based
on the empirical testimony of repeated experience. Thus ordinary
people aie convinced of the truth of the concomitance of many things,
about which a philosopher comes to entertain grave doubts. The
vulgar masses are not disturbed by any doubt about their belief in
the concomitance of the day with the particular movement of the sun
in the horizon3. But the scientist has dismissed this belief as
untrue on the basis of data which are not accessible to an uneducated
1 NL, p. 518.
2 There is a widespread belief that sneezing on the eve of journey is an
evil omen or that the number thirteen is unlucky. Such beliefs in the concomitance of the iacts in question may be regarded as superstitious. And the
doubt about their truth is a real doubt which can be removed only by the application of reasoning, favourable or unfavourable, resulting in the elimination of
the doubt by consideration of pros and cons. The efficiency of reasoning will be
evident in the resultant conviction of the truth or falsity of the beliefs.
n|
23
person. Thus there may be many accepted generalizations which

are liable to be called in question owing to the advance of knowledge.
The truth or otherwise of such accepted generalizations can thus be
tested by the help o: reasoning. The question can be decided by a
division of the types of doubt. The initial doubt about the
truth of a universal proposition, which has not yet been proved, is not
an obstacle to the realization of the truth or otherwise of the universal
proposition in question. This is called in the terminology of the
Naiyyika doubt about the matter to be known (grbyasamsaya).
Such doubt is not an impediment to the future realization of the
truth. Buc a doubt about an accepted belief, upon which a man
may base a conclusion, is certainly an obstacle, for the elimination of
which reasoning is called in request. Thus the position maintained
by ^rlvailabha in his Nyyalllvafi is entirely different from that of
Udayana, and it has been found that a doubt as the occasion of reasoning is a real obstacle1.
Bhagiratha Thakkura, the sub-commentator of the Nyyalilavati,
after elucidating the position of the original author, has propounded
his independent views on the necessity of the first type of reasoning.
Bhagiratha does not think that Stivaliabha's summary rejection of
Udayana's position is absolutely infallible. He asserts that reasoning
of the first type is of service even in inference by eliminating enquiry
about the opposite possibility as maintained by Udayana. it is
common knowledge that the expected cognition does not take place
if a disturbing agent crops up, though the full conditions of the cognition in question are at work. The favourite example with the
Naiyyika of a disturbing agent is thought of a woman. Even the
combined knowledge of the premises does not result in the inference of
the conclusion, if a person happens to think of a woman. That the
thought of a woman or such like things obstruct the actualization of
the result is due to the fact that such thought, or desire for thought,
relates to an object, which s opposed to the object of the expected
knowledge. The truth of the matter is the principle that thought of
a contrary object is a disturbing condition, which frustrates the actua1 NL, pp. 517-19.
24
INDUCTIVE REASONING
[CHAP.
iization of the expected result. The thought of the opposite of the

probandum would thus thwart the process of inference even after the
combined knowledge of the premises1. It may be contended that
there can be no such diversion of enquiry after the knowledge of the
existence of the probans in the subject known to be necessarily concomitant with the probandum (-paramarsa). Now a diversion of the
attention is supposed to be caused by doubt about the contrary possibility. To revert to our old example, let us suppose that a man perceives smoke in a hill and remembers that smoke is necessarily
concomitant with fire. Is it possible that a man should come to have
doubt regarding the possibility of absence o fire? A doubt of the
kind can be resolved only by the knowledge of the necessary relation
of the probans and the probandum. But such knowledge is ex hypothesi present at the time. How can then there be any doubt about
the opposite alternative? The impossibility o doubt makes impossible
the diversion of enquiry towards the opposite possibility, since the
absence of the cause makes the absence of the effect inevitable. So
reasoning has no raison d'etre in such a situation, as there is neither
doubt nor diversion of thought produced by doubt, for the elimination
of which reasoning might be held in request. If there were any doubt
before, it would be expelled by the knowledge of the major and the
minor premises. Indeed this has been the argument of Srivallabha for
rejection of this type of reasoning2.
Bhaglratha observes in reply that though doubt about the contrary
possibility does not spontaneously arise after the combined knowledge
of the premises and an antecedent doubt of the kind has been removed
by the latter^ yet the possibility of doubt, due to the operation of an
opposite cognitive instrument, is not precluded. It is not universally
true that the knowledge of the combined premises necessarily culminates in the knowledge of the conclusion.
Were it so, contradicted
1 vastuto visayaparisodhakasy'pi jijnsnivrttidvr upayogah. kminijijiir
ssthalc viroclhivisayakajijnstvona pratibandhakatvakalpant. sdhybhvajijnsy api pratibandhakatvt.
NLPV, pp. 518-19.
2 na ca parmarsd eva sanknivrttir iti mlanivrttau jijnsy api nivrtteh
kirn tarkena iti vcyam?
Ibid., p. 519.
25
probans [bcidba) or countermanded probans (satpratipaksa) could not be

included in the list of fallacies.
Now, a fallacy is that which thwarts
the realization of inference, in other words, the knowledge of the
conclusion.
Let us return to our previous example of contradicted
probans. Suppose a person is going to argue that fire should be cold,
because it is a substance, and all substances such as earth, water etc.
are cold. The knowledge of the universal proposition, "All substances
are cold," and of the minor premise, "Fire is a substance," is there.
Suppose, again, that immediately after, the man touches the fire and
feels it to be hot.
The perception of hot fire prevents the inference
of coldness.
Here the perceptual cognition directly thwarts the
inference of coldness in fire without annulling the major or the minor
premise. It is not denied that the contradiction of the probandum
may result in the subsequent conviction that the relation between the
probans and the probandum is neither universal nor necessary. The
case is the same with the countermanding reason which is known
immediately after the combined knowledge of the premises. Suppose,
a man thinks in the strain as follows. Impalpable things are imperishable and word is impalpable.
Suppose again, that he realizes that
all products are perishable and word is a product. Now the subsequent knowledge would directly prevent the inference that word is
imperishable without falsifying the major and the minor premises.
Though as a matter of fact the major premise comes to b realized
as false, the knowledge of the falsity is rather the result and not the
condition of the knowledge of contradiction1.
Let us evaluate the results of the logical enquiry conducted in the
preceding paragraphs.
It has been found that the combined knowledge of the premises fails to lead to the conclusion, if it has been
thwarted by knowledge of the opposite truth. What is true of knowledge of the opposite is also true of doubt about the opposite, with
this difference that knowledge of the opposite permanently precludes
the knowledge of the conclusion, whereas doubt or diversion of enquiry*
1 svrasikasanknivmv api viroclhiprainanasaclbhavanibanclhanasankayah
sulabhatay jijnsy api sulabhatvc.
NLPV, p. 519.
z6
INDUCTIVE REASONING
produced by it operates as a temporary setback. The removal of the

diversion is effected by reasoning, and with the removal of this impediment the previous knowledge of the major and the minor premises
leads to the knowledge oE che conclusion.
The interpretation of
Bhaglratha Tliakkura seems to be the most faithful and loyal exposition of Udayana's position. Bhaglratha thus makes out that reasoning serves as an auxiliary to the condition of inference after it has set
to work; that is, in other words, reasoning is regarded as a subsequent
sub-organ (uttarnga) of the instrument of inference. Bhaglratha thus
substantiates the claim of the first type of reasoning in opposition to
Srivallabha who repudiated it.
It is obvious that the defence of
Bhaglratha has succeeded only to a limited extent. The first type of
reasoning is shown to have a necessity of its own only in cases where
an opposite mode of knowledge creates doubt about the object of
inference and subsequently produces a diversion. The necessity of
appeal to the first type of reasoning is thus limited to rare and exceptional occasions and is not universally operative. It may not be irrelevant to observe that Bhaglratha's exposition of Udayana's position is
more loyal than that of Vardhamna, who holds that reasoning serves
to eliminate not only an opposite enquiry but also the doubt which
is the condition of it1.
Though he adduces reasons in support of his
position, it is undisputabie that this exposition runs counter to the
position maintained by Udayana.
Rucidattas -position
Rucidatta, an astute logician of Mithil noted for his originality of
views and critical insight, has propounded his views on the necessity
and function of reasoning in his work called Makaranda^ a gloss on
Vardhamna's commentary on the Nyayakusumanjali,
Rucidatta's
interpretation differs in substantial respects from the views of rivallabha and Bhaglratha Thakkura, though his agreement with the latter is
pronounced.
Udayana maintains that doubt about the contrary of
probandum is the cause of enquiry about the latter. For instance,
i tajjanakasamsayasahite'ti sesah. anyath jijnsvicchetle'pi samsayasattve
vyptyagraht.
NKuP, ch. I l l p. 23.
27
suppose, that a man is in doubt whether the hill is possessed of fire

or not.
The doubt about the negative alternative, namely, the
absence of fire, induces an enquiry for this absence of fire. But
enquiry "is nothing but a desire for knowledge.
When a man sets
about to infer the presence of fire it is natural for him to entertain a
desire for the knowledge of fire and not of the opposite.
Doubt
again cannot be the cause of desire for knowledge. Desire for knowledge or for anything else is always induced by the knowledge that the
object of desire is a means to the realization of a desired end. Now
the person, who is interested in the discovery of fire, cannot be
supposed to think the absence of fire to be conducive to the realization of his objective. On the contrary he should entertain a desire
for the ascertainment of fire in which he is interested on account of
its instrumentality in the realization of his objective. Thus, there
is no logical or psychological justification for Udayana's contention
that doubt induces an enquiry after the opposite alternative, and
reasoning comes to eliminate such desire which precludes the
knowledge of the original object of inference1.
Rucidatta observes in defence of Udayana's position that the causal
relation between any two phenomena can be ascertained only by
observation of unconditional concomitance in agreement and in
difference. It does not behove a man to call in question the relation
of casuality on a priori grounds.
It is true that desire of knowledge
is found to be induced by the assured conviction of the utility of its
object in which he is interested. But that does not prove the impossibility of any other fact being the cause of it. The charge of the plurality
of causes can be averted by the hypothesis of the qualitative difference
of the effects, namely, of the desire for knowledge, as induced by
knowledge of the utility of a thing from that induced by doubt. That
doubt produces a desire for knowledge about its object is known by the
application of the Joint Method.
Accordingly one should not raise
a difficulty about the causal efficiency of doubt regarding desire for
1 nanu samsayo na jijnshetuh, tasy icchtvena jnnescasdhanatjnnasdhyatvt. sdhyrthinas ca sdhybhvajnnam nestasdhanam iti na tatra
jijns. tath ca katham tannvarttakat tarkasye' ti cct?
NKuPM, eh. Ill, pp. 4-5.
28
INDUCTIVE REASONING
[CHAP.
knowledge about the opposite alternative, which is one of the alternative objects of doubt.
It is admitted that the emergence of a countermanding reason results in the emergence of desire for knowledge
of the opposite alternative.
The case of doubt is exactly on a par
with the countermanding reason.
As regards the contention that
when the person is interested in the knowledge of fire he ought not
to have a desire for the contrary, it should suffice to say that solicitude
for a thing not infrequently induces a doubt about the undesirable
possibility1.
Thus has Klidsa, the greatest poet of India, observed
that affection Is apt to be apprehensive of the evil about the beloved
person. According to the line of argument adopted by the opponent,
Klidsa should have said that affection is sanguine of the welfare of
the object of love2.
Udayana's position that doubt is the cause of
diversion of enquiry towards the opposite alternative thus stands
unassailed.
Vacaspati Misra thinks that reasoning is the cause of the elimination of doubt about the opposite;? whereas Udayana regards it as the
cause of the elimination of the resultant enquiry about the opposite.
But both these views seem to be illogical, as reasoning is not opposed
to doubt or enquiry in respect of the relevant subject-matter. The
opposition of cognitions always takes place when one posits an object
and the other negates it. But we have already shown in our exposition of the position of Udayana that there is no opposition in respect
of subject-matter between reasoning and doubt. And the same lack of
opposition is found in the case of enquiry and reasoning also. Rucidatta observes that though the opposition may not be logical, there
can be no doubt that the opposition is factual.
It is true that the
application of reasoning results in the elimination of doubt, which
is the position of Vacaspati Misra, and in the elimination of
opposite enquiry, which is the position of Udayana. That the
elimination of doubt or of enquiry takes place after the operation of
i
anvayavyatirekbhyam satpratipaksasy eva samsayasy'pi jijsjanakat-
snehah ppasanki
Sakuntal, Act IV., p. 183.
vt.
NKuPM, ch. Ill, p. 5.
ii]
29
reasoning is a matter of experience.

So there is no reason to doubt
the causal relation between the two.
We have already shown that
LJdayana's affirmation of opposition between reasoning and opposite
enquiry is not on the same footing with the position of Vcaspati
Misra.
The latter opposition has been shown to be an instance of
a universal law, whereas Vcaspati's position can be explained on the
postulation of a specific causal relation determined by a specific observation of concomitance in agreement and difference1.
The modus operandi of reasoning as shown by Rucidatta differs
from what is assumed by Bhaglratha. Rucidatta explains the procedure as follows.
First a man observes smoke in a hill and then
comes to entertain doubt whether there is fire in it or not. This
doubt about the possibility of absence of fire induces an enquiry about
it. The enquiry consists in the search for the conditions of lack of
fire. So long as this diversion endures, the knowledge of the
universal concomitance between fire and smoke already ascertained in
the kitchen does not occur and consequently the synthetic judgment,
which could be produced by the knowledge'of the two premises, does
not materialize.
But reasoning serves to eliminate this opposite
enquiry, and after this the knowledge of the universal concomitance
of smoke with fire, and of the occurrence of smoke in the hill, results
in the synthetic judgment (paramara), which immediately culminates
in the knowledge of the conclusion. Thus reasoning is regarded by
Rucidatta as an instance of the first type set forth above. It is argued
that the diversion of enquiry caused by doubt about the opposite
possibility prevented the realization of the synthetic judgment, which
could end in the knowledge of the conclusion. The subject-matter
of the synthetic judgment is the probans-as-existent-in-the-subject.
The elimination of opposite enquiry which obstructed the realization
of the synthetic judgment thus amounts to an elimination of an
obstruction of the subject-matter. It should not be thought that the
elimination of the obstacle set forth above is equivalent to the removal
1 nanu virodhivisayatvbhvt tarkah katham samsayajijsayor nivarttaka
iti cet? na,...icchym aniscapratisanclhnatven'pi virodhitvc ca, madhuvisasamprktnnabubhuksdau tath darsant.
NKuPlV, ch: III, p. 5.
3o
INDUCTIVE REASONING
[CHAP.
of doubt about the universal concomitance, which is the province of the

second type of reasoning. What Rucidatta maintains is this. The
doubt or enquiry about the opposite possibility is not the same thing
as doubt about the universal concomitance. The certitude of the
universal concomitance was already in the mind of the person concerned. But this certitude was rendered ineffectual by the diversion
of enquiry and so the synthetic judgment which is the immediate
antecedent of inference according to the Natyyika could not take
place. The elimination of the obstacle by reasoning thus sets free
the conditions of inference already present, which thereafter produce
their natural and expected effect, namely, the knowledge of the
conclusion1.
The difference of Rucidatta from Bhagiratha is clear. The latter
makes the emergence of diverted enquiry a subsequent event to the
synthetic judgment produced by the knowledge of the premises,
whereas the former makes it out as an antecedent occurrence to the
same. Secondly, the elimination of the obstacle is explained by
Bhagiratha to relate to "the knowledge of the probandum, whereas
Rucidatta makes it centre round the probans as the subject-matter of
the synthetic judgment. The result is that Rucidatta regards reasoning; as an antecedent sub-org-an of the instrument of inference whereas
Bhagtratha's explanation gives it the status of a subsequent sub-organ.
Rucidatta in support of his position appeals to the authority of Vardhamana Upadhyya, who, as reported by him, is said to assert that reasoning is always an antecedent sub-organ in the work called, the Prameyatattvabodha. It deserves to be noted that according to Rucidatta
reasoning is always an antecedent sub-organ and has two different
functions, namely, (i) one conducive to the knowledge of the universal concomitance and (2) the other conducive to the realization of
synthetic judgment (paramara)2.
NKuPM, ch. Ill, p. 5.
2 ... tad ha tatr aiva Vardhamfmah tftarko nyyasya prvngam nyyavisayaparisodhakatvt, vyptigrhakatvc ca1' ity huh Ibid.
CHAPTER III
GAGEA'S SPECULATIONS ON REASONING
Gangesa Updhyaya is an epoch-maker in the domain of Nyya
speculations. His work the Nyyatattvacintamani is an extraordinary
work which created a revolution by a thorough revaluation
o: ail the problems of realistic philosophy
that had been
discussed by generations of philosophers for about a thousand years
and a half. The special attraction of GangesVs work in spite of its
studied brevity and consequent difficulty lies in its judicial crystallisation of the substantial results achieved by previous speculations and
the avoidance of unessential details. It is perhaps the most popular
work in the sense that generations of scholars and thinkers have given
their minutest attention and most anxious consideration to every word
of the work. There are a very few works which seem to have
attracted so many commentaries and that again from scholars of all
parts of India, whose labour is spread out over at least six centuries.
The most illuminating commentaries and the best expository treatises
have, however, been composed by scholars of Mithili and Navadvipa
of Bengal. Latterly, however, the Bengal School came almost to
monopolize the entire attention of the academic world in India.
Gagesa's acute analysis of the problems and assessment of the value
of his predecessors' views, coupled with his wonderful marshalling of
the data furnished by the previous writers, have made his work and
his name immortal. No study of the problems of Nyaya philosophy
can pretend to be thorough, unless it makes a special and elaborate use
of this magnum opus which has created a history in the domain of
Indian philosophy. Its influence on the later speculations m the
other branches of philosophy in India is no less pronounced than on
the future development of Nyya epistemology.
These prefatory
remarks are made by us to help the reader of this humble work to
form an idea of the gravity and the onerous responsibility of the task
we propose to undertake.
We have not avoided the difficulties and
we shall endeavour to make our representation as loyal as it will be
32
INDUCTIVE REASONING
[CHAP.
made thorough.
We approach our task with full consciousness of
the abstruseness of GagesYs treatment accentuated by the fact that
we shall not have the benefit of the study made by any previous
writer in English or any of the modern Indian languages. Gagea's
work is a sealed book to the English-knowing public interested m
philosophy.
It will be presumptuous on our part to suppose that our
exposition will be free from the drawbacks incidental to pioneer
attempts.
In the chapter on induction {vyptigraha) Gangesa institutes a
thorough enquiry into the means of arriving at the knowledge of
universal concomitance. Gagesa observes in the beginning that the
ascertainment of universal concomitance of the probans with the probandum, in one word, induction, is made possible by the knowledge
of co-association of the probans and the probandum coupled with the
absence of the knowledge of the failure of their concomitance. There
can be no conviction of the necessity and universality of concomitance
in question if there be a doubt or conviction of the fallibility of the
said concomitance. So the condition of induction can be propounded
in brief, firstly, to be the knowledge of co-assocration of the two
phenomena, and secondly, the absence of the knowledge that the
said co-association is not necessary i.e., accidental. The term 'knowledge* in the second condition is used in a comprehensive sense
inclusive o doubt.
Now, doubt about the contingency of the concomitance in respect of the probans may be due to the cognition of
common attributes coupled with the lack of the knowledge of specific
attributes, which makes the determination of one of the alternative
objects of doubt impossible. Secondly, the doubt of the necessity of
concomitance may be due to the doubt about the presence of a condition (ttpdhi).
The condition is that, which has by its nature necessary concomitance with the probandum, but not with the probans.
So the concomitance of the probans with the probandum 1$ really
due to the presence of a condition with which the probans may be
accidentally associated. Thus, for instance, the concomitance of fire
with smoke (as opposed to that of smoke with fire) though observed
in a large number of cases cannot be regarded as necessarily universal,
as fire is associated with smoke only when the former is associated
i]
GAGESA'S SPECULATIONS ON REASONING
33
with a carbohydrate. If fire is in contact with a carbohydrate, smoke

invariably occurs.
But the occurrence of smoke is due to the carbohydrate which has no necessary connexion with fire. The concomitance of fire with smoke is thus subject to the condition that carbohydrate should be associated with fire. This concomitance is thus
accidental and conditional.
But the concomitance, which is the
condition of inference, must be unconditional, that is to say, it must
be known to be due to nothing extraneous but to the intrinsic
character of the probans and the probandum1.
Now, the first kind of doubt can be removed by the definite
knowledge of the specific characteristics of one of the terms. An
example will make the position clear. Sound shares in common with
space the characteristic of invisibility. A doubt may naturally arise
whether sound will also participate in the other characteristic of space,
namely, its eternity.
Now, this doubt of the concomitance of sound
with eternity can be set at rest by the definite knowledge of sound
being a product.
The latter characteristic is the differentia of sound
and removes the doubrthat it could be eternal. About the second
kind of doubt of the concomitance of two terms, the doubt in question
can be eliminated only by the demonstration of the impossibility of
an extraneous condition, i.e., if it can be shown that the concomitance
of the terms is due to the essential nature of the terms under consideration, and not to an external condition associated with the probans
by way of accident.
But how can the doubt about a condition be
eliminated?
In other words, how can the concomitance be proved
to be necessary and integral to the essential nature of the terms? The
instrument of the elimination of doubt about such a condition is
reasoning, which operates as a reductio ad absurdam of the opposite
possibility.
There may be cases of concomitance, which are not
assailed by doubt.
The non-emergence of doubt may be due
to the absence of defective knowledge, which is the universal
1 vyabhicrajanavirahasahakrcam sahacradarsanam vyptigrhakam. jnanam
niscayah sak ca. s ca kvacid updhisandeht kvacid visesdarsanasahitasdhranadharmadarsant. tadvirahas ca kvacid vipaksabdhakatarkt kvacit svatah
siddha eva.
TCM, p. 662.
34
INDUCTIVE REASONING
[CHAP.
condition of doubt.
Secondly, it may be due to the sheer failure of
the opposite alternative to present itself to the mind. The absence
of doubt, however, is the universal condition of the assured knowledge
of the concomitance being necessary and universal. If there be the
slightest tinge of doubt, inference becomes impossible. The most
essential condition of inference is the unwavering conviction of the
infallibility of concomitance1.
The elimination of doubt is not equivalent to the cessation of
doubt. Now doubt, being a psychical event, is made to cease by
the appearance of another psychical phenomenon. The cessation of
doubt as noted here is a natural occurrence and the service of reasoning is not required for the same. But the elimination of doubt
should be understood to mean the elimination of the possibility of
recurrence of the doubt. Such total elimination of doubt, initial or
recurring, is secured only by means of reasoning. It may be objected
that reasoning is founded upon similar absolute conviction of necessary
concomitance between the ground (apdaka) and the consequent
(pdya). Reasoning is a hypothetical argument, in which the consequent is shown to follow necessarily from the ground assumed on the
assertion of the opponent. Suppose, a sceptic doubts that smoke is a
necessary concomitant of fire (vahnivyafyd). Now this doubt is
sought to be removed by a redttctio ad absurd um as follows. "If
smoke were not a necessary concomitant of fire, it would not be a
product of fire, which it undeniably is". The effectiveness of this
hypothetical argument, in which reasoning invariably expresses itself,
lies in the belief in the necessary relation of the ground, namely, the
absence of smoke's concomitance with fire, with the consequent,
the denial of smoke being a product of fire. Now this necessary
relation, on which reasoning is based, being itself a case of universal
concomitance, is exactly on the same footing with the other cases of
concomitance. Consequently the foundational concomitance is open
to doubt. And if recourse be had to another reasoning for the elimination of the second doubt, the second reasoning would again presuppose another necessary concomitance, which will in its turn be
i
TCM, p. 662.
m]
35
liable to be called in question. So the removal of one doubt would

be the starting point of the emergence of an endless series of doubts.
The efficacy of reasoning as an instrument for elimination of doubt
about universal concomitance thus turns out to be a will-o '-the-wisp or
a dangerous quagmire 1 .
Let us now consider the solution of the problem as propounded
by Gagesa. Gagesa does not think that the admission of reasoning
would make a regressus ad infinitem unavoidable. Reasoning is
required to eliminate doubt of the contingency of a case of concomitance. Though reasoning is at bottom itself based on concomitance,
it is not inevitable that the series of doubts will proceed without limit.
The limit of doubt is self-contradiction or contradiction of a selfevident truth. Reasoning is required only for the elimination of an
actual doubt and will not be necessary if there be no doubt. When
doubt is rendered impossible by the emergence of self-contradiction,
reasoning will have no raison d'etre. So neither the chain of doubts
nor that of reasonings will be unlimited. The charge of infinite
regression as a necessary consequence of reasoning is thus unfounded2.
It has been said that reasoning serves to eliminate doubt about the
contingency of the universal concomitance. It requires to be explained how reasoning comes,to remove such doubt. In other words, its
modus operand! should be made clear. Gagesa explains the modas
operand* of reasoning in the matter of elimination of doubt by a
concrete example. Now smoke and fire are known to be concomitant. But there may be a doubt about the necessity and universality
of such concomitance. In other words, it is possible to doubt whether
smoke should occur in a place, e.g., a lake, which is devoid of fire,
just as it is observed to occur in association with fire in a kitchen. If
smoke could exist in a fireless place like a lake, its concomitance with
fire would only be contingent. If, on the other hand, the occurrence
tarkasya vyptierrahamlakatvena anavasth iti cet?

T C M , p. 675.
na, yvadsankani
tarknusarant.
yatra ca
vyghtena
sakai va
nvatarati tatra tarkam vinai va vyptigrahahIbid.

Vide also Nku, chapter III, p. 25.
36
INDUCTIVE REASONING
[CHAP.
of smoke be impossible in a fireless place, then the concomitance of

smoke with fire will not be contingent or a matter of accidental coincidence. When such a doubt occurs reasoning comes in to exclude
the possibility of such doubt in the following manner. Smoke is a
product and as such must have a cause of its own. A product must
be caused either by fire or something other than fire. This holds
good of all products. It cannot be the case that a product will be
the effect of neither fire nor not-fire, i.e., something other than fire,
because this would only mean that the product under consideration is
not a product, which is a contradiction in terms. Now smoke is a
product and as such cannot be considered to be the effect of neither
fire nor not-fire. So the doubt about smoke cannot be of the form
stated above. If smoke be not the product of either fire or not-fire it
cannot be regarded as a product at all. But smoke, having a definite
time of occurrence, cannot be other than a product. The doubt about
its being a product or not is thus precluded by contradiction of an
undeniable truth 1 .
Let it be accepted that doubt about smoke being a product is
impossible. But it may possibly be doubted in the first place, that,
though a product, smoke is the effect of a cause which is different
from fire; in the second place, smoke may be thought to be occasionally the product of fire and occasionally to be the product of something
else; in the third place, it may be thought to be a product without
any cause. The doubt, as set forth just now, is thus seen to alternate
among three possibilities. The second alternative is based on belief
in the possibility of a plurality of causes and the third assumes the
possibility of uncaused production. Both these alternatives are shown
to be impossible on account of contradiction. The first alternative in
which smoke is believed to be the product of something else than fire
is equally liable to be dismissed on the ground of self-contradiction2.
1 tath hi dhmo yadi vahnyasamavahitjanyatve sati vahnisatnavahitjanyah syn no tpamiah syt.
TCM, p. 675.
2 ... kirn dhmo Vahnher eva bhavisyati? kvacid vahnim vinpi bhavisyati?
ahctuka eva v utpadyate? iti anlt syt, sarvatra svakriyvyghtah syt.
Ibid,, p. 676.
ni]
37
The usual organ for ascertaining causality is the Joint Method,

if the causal relation between two events is ascertained by the observation of the unconditional concomitance in agreement and in
difference, a subsequent doubt would involve contradiction of an
established and well-founded belief. Let us consider the case of
smoke and fire. That fire Is the cause of smoke is established by the
observation of the concomitance of smoke with fire in presence and
in absence. Smoke has been repeatedly observed to be associated
with fire and never where fire is absent. It is invariably seen to be
the case that when a man wants to produce smoke for the expulsion of
mosquitoes, he fetches fire and puts combustible substances in it. It
is preposterous to suppose that such a person can entertain doubt that
smoke is not the effect of fire. The regular adoption of fire for the
production of smoke is incompatible with such doubt. If doubt were
possible in such a case, the unfaltering movement towards fire would
be impossible. A man's activity is caused by a volitional urge, which
in its turn is generated by an unwavering belief. It is unthinkable
that a man would doubt that A is the cause of B and at the same
time unhesitatingly resort to A for the production of B. The contradiction of the behaviour and doubt is quite obvious. So doubt
cannot be entertained in such a case. This proves the untenability
of the charge that if necessary concomitance were only ascertained
by reasoning, there would crop up an infinite series of doubts and
reasonings, the latter being based upon a belief in necessary concomitance in each case. Reasoning, it has been observed before,
is necessary only when there is doubt. But when doubt is finally
set at rest by self-contradiction, reasoning would have no necessity
of its own. Thus doubt is not infinitely possible and consequently
the appeal to reasoning for rebutting doubt is not necessary ad
infinitum1.
Gagesa's view as recorded here is substantially a reproduction
1 yadi hi grhitnvayavyatiiekam hetum vin'pi kryotpattim saketa, tad
svayani eva dhmitham vahneh, trptyartham bhojanasya, parapratipattyartham
sabdasya co 'pdnam niyamatah katham kuryt? ,..cad idam uktam tad eva hi
sakyate yasmin nsaiikyamne svakriyvyghto na bhavat 'ti.
TCM, p. 676.
38
INDUCTIVE REASONING
[CHAP.
of the arguments adduced by Udayana in the Nyydkusumnjdli1

in refutation of the scepticism of the Crvka Materialists. But
Udayana's defence of the validity of inference and of the possibility
of induction has been subjected to destructive criticism by rlharsa
in his work, the Khandanakhandakhdya. Gagesa's contribution
to logic consists in the attempt to rebut the criticism of Srlharsa.
We shall now consider the objections of Sriharsa to Udayana's
position and the defensive arguments propounded by Gagesa to
rebut the contention of Srharsa.
Sriharsa raises a query about the cogency of practical behaviour
serving as an impediment to the emergence of doubt regarding the
necessity of concomitance.
Udayana thinks that the sceptic's
behaviour is a contradiction of the aforesaid doubt.
But,
what is the nature of contradiction constituted by behaviour?
Sriharsa asserts
that this contradiction is nothing but the
mutual opposition of doubt and behaviour. Opposition, again, is
only a case of universal necessity of the incongruence of two terms.
In other words, opposition involves that the terms under consideration must not occur in the same substratum as a matter of universal
necessity. But this universal absence of concomitance, which is the
condition or the essence of mutual opposition, is itself a case of universal concomitance. The difference of the latter from the usual
cases of concomitance, e.g. between smoke and fire, is this that the
latter holds between two negative terms, whereas the former is one
between two positives.
The difference of the nature of terms,
however, is irrelevant to the determination of the question of universal
concomitance. The fact remains unchallenged that when Udayana
seeks to avoid the charge of infinite regress by appeal to contradiction, he only assumes another case of universal concomitance as the
solvent of the problem. But the solution only culminates in the
emergence of another problem. Let us revert to the old example of
smoke and fire and apply the criticism of Snharsa. Udayana and his
adherent Gagesa maintain that doubt about the necessity of the concomitance of smoke and fire is ultimately resolved by the contradiction
i
NKu, chapter III, pp. 23-25.
i]
GACEA'S SPECULATIONS ON REASONING
39
furnished by the volitional activity of the person regarding smoke

and fire. Sriharsa thinks that this contradiction lies in the supposed
Incompatibility of the activity with doubt of the necessity of the
concomitance of smoke and fire. The implication of the mutual
incompatibility of doubt and activity is that there is a necessary concomitance between activity and absence of such doubt. So the
so-called contradiction transpires to be a belief in another necessary
concomitance.
And a philosopher, who wants to be satisfied about
the possibility of the knowledge of universal concomitance, is only
sought to be silenced by a show of necessity.
It is conveniently
ignored that the doubt raised is not confined to a particular case of
concomitance, but extends to concomitance as a universal and
necessary proposition. If the doubt about a particular concomitance
be sought to be satisfied by the show of another concomitance, it
would only result in the shifting of the centre of dispute.
The
second concomitance, which is the foundation of the first concomitance, would equally be liable to doubt. And if another concomitance in the shape of contradiction be put forward in defence,
chat would also fail to effect an improvement, being equally subject
to doubt.
Udayana sought to avoid infinite regression entailed by
reasoning by the evidence of practical behaviour pointing to a contrary
belief. But this contradiction by practical behaviour has turned out
on enquiry to be itself an assertion of another necessary concomitance.
So this is not a solution of the problem. Sriharsa's contention is that
it is not humanly possible to be convinced of the universal and infinite
repetition of concomitance of two phenomena, observed in a limited
number of cases, in future time and remote places. Udayana trots
out the plea of contradiction as a means of the knowledge of universal
concomitance.
But contradiction itself being a case of universal concomitance is equally open to doubt. The fact, that one is a case of
concomitance of positives and another is a case of concomitance of
negatives, does not make an appreciable difference so far as the epistemological problem is concerned.1
We have considered at length the contention of Srharsa and we
1 KhKh, pp. 677-93.
4o
INDUCTIVE REASONING
[CHAP.
now propose to examine and elucidate the answer given by Gangesa

to Sriharsa's doubt.
Gagesa asserts that the interpretation of Udayana's position by Sriharsa is unfair and illegitimate. Udayana does
not mean thac contradiction involves only a case of concomitance of
negations of two incongruous elements. Udayana's statement is crystal
clear.
He simply affirms thac practical activity per se constitutes
contradiction of the doubt about the necessity of concomitance and
this makes such doubt impossible. Udayana's intention can be elucidated as follows.
It is absurd that a man should doubt whether
smoke be the product of fire and at the same time should seek for
fire whenever he feels a necessity for smoke. The inconsistency of
the activity with the doubt is obvious and does not require an elaborate logical apparatus to make itself felt.
Udayana only lays stress
upon this inconsistency which is self-evident. He is quite sure that
the exposure of the absurdity involved in plain self-inconsistency constitutes a refutation of the alleged doubt. He does not take pains to
prove this absurdity as he is convinced of the truth that absurdity
cannot be refuted like truth. It is worthy of remark that the doubt
about the causal relation of smoke and fire is ultimately resolvable into
doubt about necessary concomitance, since causality is only a species
of necessary concomitance. Although Udayana states that the doubt
relates to the causality of smoke and fire, he cannot but intend that
the doubt is of the necessary concomitance.
Contradiction by volitional activity of the doubc of causal relation, as shown by Udayana,
should thus be understood to relate to the doubt of the necessary concomitance. The position, which Udayana and Gagesa are interested
to establish, is that doubt about necessary concomitance is resolved
by reasoning; and a further doubc about the foundation of reasoning,
which is itself a case of necessary concomitance, is precluded by the
awareness of absurdity involved in such doubt 1 .
i etena vyghto virodhah sa ca sahnavasthnaniyama iti tatr'py anavasthe' ti nirastam. svakriyy eva sankpratibandhakatvt. ata eva
"vyghto yadi sank' sti na cec chaka tatas tarm.
vyghtvadhir sak tarkah sankvadhih kutah'*.
iti Khandanakramatam apy apstam. na hi vyghtah sanksritah, kintu
svakriy aiva sankpratibhandhike' ti.
TCM, pp. 719-22; vide also KhKh, p. 693.
i I
41
Gagesa puts forward a further consideration against Srlharsa's

contencin abouc the impossibility of inductive knowledge, if universal concomitance cannot be ascertained on account of the inevitability
of infinite regression as alleged by Sriharsa, the removal of doubt by
observation of specific characteristics would become impossible. The
elimination of doubt in such cases is a matter of universal experience.
To return to the classical example of doubt about the man- ortree, we have already found that the doubt is resolved by the knowledge of hands and feet or branches and leaves, the specic characteristics of man and the tree respectively. But what is the logical necessity that the doubting person ceases to have doubt when he observes
either the hands and feet or the branches and leaves in the object of
doubt lying ahead? It is nothing but the unmistakable knowledge of
the necessary concomitance between a human being and hands and
feet, or between a tree and its branches and leaves, which are called
specific characteristics. Knowledge of a specific characteristic removes
the doubt because it Is in essence the knowledge of a necessary concomitant (vypyadharma). What is not a necessary concomitant is
not a specific characteristic, e.g., the tapering height, which is common
to both tree and man, the terms of doubt. Thus not only metaphysical
doubt about the necessity and universality of causality, but also such
ordinary doubts of every day experience as illustrated by the case of
man and tree, will turn out to be impossible of elimination, if the
knowledge of universal concomitance remains unrealizable by us.
This is certainly scepticism in excelsis. But such scepticism, which
makes even ordinary activity impossible, cannot be regarded as a
serious obstacle to the pursuit of philosophical speculations1.
1 We cannot help adverting to a point of epistemological interest which

arises in this connexion. Later Naiyyikas, for example Mathurntha Tarkavagisa and others, hold that the resultant knowledge, either of the man or of the
tree as the case may be, is perceptual and not inferential. It is true that the
conditions of perception and of inference are equally present in such a situation.
But mere operation of the sense-organ is incompetent to deliver the decisive
knowledge that it is a tree or a man, unless and until the knowledge of the
specific characteristics as the necessary concomitant o either of them supervenes.
That such knowledge is of the nature of the synthetic judgment {-parRmarsa)
42
INDUCTIVE REASONING
[CHAP.
We cannot conclude our discourse here without adverting to one

or two sage observations made by Gagesa in this connexion.
efficiency
The
of reasoning as an organ of induction (vyptigrbakatarka)
comes in evidence prominently when the concomitance of the terms is

repeatedly observed and ascertained.
It cannot be realized that smoke
cannot be the product of" anything
else than fire, if the repeated
knowledge of their concomitance is not at the back of our mind.
solitary observation of concomitance of smoke and fire is not adequate

to the realization of its universality and necessity.
be supposed
that repeated
observation
But it should not
by itself has got any special
efficiency
with regard
Repeated
experience only prepares the ground for the application o
reasoning.
The Buddhist
causality by itself
possible.
to the knowledge of
logician
thinks
universal concomitance.
that the knowledge of
makes the knowledge of universal concomitance
But universal concomitance,
as has been observed by the
Buddhist himself, is not exhausted by causal relation.
Observation
of the necessary concomitance and of the minor premise does not admit of
doubt. It is also universally admitted that the immediate consequence of such
synthetic judgment is inference. It seems incongruous on the part of the
later Naiyayikas that they should believe the resultant knowledge after doubt to
be an exceptional case, which it must be if it be perceptual. It is worthy of
remark that the older Naiyayikas regard such cognitions as inferential in
character and in this they are more consistent than their successors. It is
remarkable that Vardhamna Updhyya endorses the view of the older
Naiyayikas in his Parisuddbiprakasa {vide p. 115). The older view has the
further merit of effecting an economy in logical thought. It ensures the uniformity of causal relation in that it does not make a reservation to the causality
of the combined knowledge of premises in regard to inference. This reservation,
has, however, been made by the later Naiyayikas who think that the conditions
of perception prevail over those of inference, when the object of cognition is
self-identical. But this seems to be an unwarranted innovation. Moreover
the later Naiyayikas cannot place implicit faith in the infallibility of the Law
of the Uniformity of Nature,, if an accredited cause be admitted to produce an
unwonted effect. The Naiyayika at any rate cannot be a party to such a
confession of failure of the causal law. The Vedntist makes a verbal proposition an instrument of perceptual intuition in exceptional cases and the Naiyayika
has accused the former of disloyalty to the Law of Uniformity. This accusation
can come with ill grace from the mouth of a neo-logician who pays scant respect
to the law in the case of the knowledge emerging after doubt.
,]
GAGEA'S SPECULATIONS ON REASONING
43
o the concomitance in agreement or in difference or in both where

possible is no doubt the organ of the knowledge of causality. But,
as we have seen, an initial or a subsequent doubt may frustrate the
knowledge of causality and for the elimination of doubt we have to
call in aid the resources of reasoning which works as a reductio ad
absurdum. Observation of concomitance is the accredited organ of
induction. But it is rendered inoperative and its findings are neutralized by the emergence of doubt about its universality and necessity.
Reasoning eliminates the doubt and thus removes all obstacles from
che path of the aforesaid organ. Repetition of instances is only
necessary for the certitude of brute concomitance and by itself has
no particular bearing upon the knowledge of the concomitance being
necessary and universal1.
Gagesa further admits the possibility of reasoning being correct
or incorrect, right or wrong. An incorrect reasoning is only a simulation of true reasoning. An incorrect reasoning cannot give us
authentic knowledge of universal concomitance which can be had
with the aid of correct reasoning.2 The case of reasoning is exactly
on a par with the case of the knowledge of universal concomitance.
If the universal concomitance as stated in the major premise be based
on a wrong or imperfect finding, the resultant inference cannot but
be wrong, it is necessary to exercise sufficient caution to determine
that the reasoning applied is correct or otherwise. The possibility of
reasoning going astray does not however give a warrant for scepticism.
It is also given to us to discover our mistakes and to correct them,
just as it is natural to make mistakes. We reserve the consideration
of correct and incorrect reasonings and their conditions and results for
a subsequent chapter3.
VIEWS OF THE OLDER NAIYYIKAS
We have already considered the solution offered by Gagesa
Updhyya in the Tattvacintamani of the problem of infinite regress
1
2
TCM, p. 722.
sattarkd vyptiprani tadbhst tadapram.
See chapter x.
Ibid., p. 724.
44
INDUCTIVE REASONING
[CHAP.
urged by Sriharsa as the inevitable consequence o the application of

reasoning.
Let us now consider the views of older Naiyyikas as
recorded by Gagesa. it was admitted by the predecessors of Gangesa
that reasoning is instrumental in the ascertainment of universal concomitance.
It was also acknowledged by them that reasoning itself
is based upon the knowledge of necessary concomitance between the
ground (padaka) and the consequent (apadya). There may be a
necessity for resorting to another reasoning, if the necessary concomitance between the ground and the consequent requires to be ascertained for the first time. In plain language the position may be stated
as follows: "If the universal concomitance between two terms be
subject to doubt, resort to reasoning is necessitated for the elimination
of the doubt; and if the credentials of the reasoning again be called
in question, resort to another reasoning will be inevitable for the
removal of doubt about the necessary concomitance, on which it is
founded.**
But this does not mean that the process will go on indefinitely. Doubt about the necessary concomitance of the ground and
the consequent can occur only when such concomitance is explicitly
felt. But if the said concomitance be implicitly present in the mind
or only recalled from the previous knowledge of it, then there can be
no doubt.
In such circumstances one reasoning alone will suffice
for the ascertainment of the necessary concomitance between the probans and the probandum, since the foundational concomitance of
reasoning is not subjected to doubt.
The position maintained is
based upon the postulate that doubt is possible only with regard to
an original belief and not when the same is recalled by memory. It
seems to be psychologically correct. In support of their position the
older Naiyyikas cite the example of the original volitions and activities
o an infant. The infant's impulse to suck the mother's breast or the
tendency to avoid a fall from the mother's lap are cases in point. The
activities in question are possible because of the knowledge that food
causes satisfaction of hunger and that fall causes pain. Certainly the
knowledge of these truths, which are cases of necessary concomitance
of cause and effect, cannot be supposed to be ad hoc discoveries, as
the infant had no occasion for such experience in its present life. Such
knowledge, therefore, should be regarded as the legacy of past experi-
iuj
45
ence o past lives.

And as these truths ire implicitly felt or recalled,
doubt cannot arise1.
The cogency of the argument will not be admitted by modern
psychologists who would explain these biological activities as automatic and instinctive, which take place without being premediated
by definite knowledge.
But Indian philosophers have all along
insisted that instinctive activities are not blind impulses of nature
which have no psychological backing. It is not open to dispute
that what was once a matter of elaborate process of discovery becomes
by repeated habit a natural possession and the chain of psychical
processes, which were necessary to bring it into being, is rendered
superfluous. Even instinctive activity must have an original belief
for its cause, though habit tends to make us forget or ignore the
nature of its genesis.
The logical validity of the position is not, however, liable to
question on the ground of the dubious origin of instinctive activities.
Other instances of habitual inferences and customery beliefs in the
necessary concomitance, on which such inferences are based, may be
utilised for establishing the contention of the older logicians. We
shall find reasons in the course of our enquiry to dispense with this
ancient theory.
But we have thought that a consideration of this
respectable theory, however inadequate ic may be, is necessary for the
understanding of the process of evolution of logical thought in India.
Gagesa has given his reasons for rejecting the old theories and we
shall explain them later on. We now propose to discuss the views of
some other schools of thinkers which have been recorded by Gagesa
in the Tattvacintamani.
Gagesa has again referred to the views of an older school of
logicians who have sought to solve the problem of infinite recurrence
of doubt by appeal to psychological evidence. They assert that it is
not universally true that the necessary concomitance underlying reasoning will be subjected to doubt as a matter of necessity. Now there
are causal relations, which are accepted to be true on the evidence of
beliefs current from time immemorial.
When a reasoning is
1 TCM, p. 731.
46
INDUCTIVE REASONING
[CHAP.
founded upon such a belief there is no possibility of its being assailed

by doubt. The view recorded here seems to be an anticipation of
Descartes' advocacy of innate ideas. It is stressed here that there arc
certain convictions ingrained in our very mental constitution and psychologically speaking, no doubt can be entertained about their truth.
When a reasoning is immediately based upon such beliefs or can be
logically shown to be derived from such beliefs in ultimate analysis,
the necessary concomitance as the foundation of such reasoning cannot
be made an object of doubt.
Thus when an accepted consequence
is shown by reasoning to necessitate the admission of a condition
which invalidates an established position, that operates as a final
reductio ad absurdum\ and a further emergence of doubt of the
validity of such a reductio ad absurdum is rendered impossible because
of its being directly or indirectly based upon such an irresistible conviction.
What is true of inherent beliefs in causality is also true of
the relations of opposition, the truth of which is equally accepted as a
matter of innate conviction. The fact, that such beliefs are ingrained
in and integral to the very make-up of the human mind, makes such
beliefs the universal property of all men.
So the very statement of
such a proposition makes it acceptable to all1.
But Gagesa does not accept this psychological solution as satisfactory. He admits that there are universally accepted beliefs, but
he does not agree that their universal acceptance constitutes an
evidence of their ultimate truth. There may be innate beliefs regarding a specific causal concomitance or a specific opposition being
universally valid; psychologically considered, there may not be any
spontaneous doubt about their validity. But that cannot satisfy a
philosopher who wants to know the logical foundation of such beliefs.
The beliefs in question can be accepted as logically valid if they can
be shown to be proved by an accredited cognitive organ. The sheer
length of time, or the numerical strength of the adherents of such
beliefs, cannot be accepted as equivalent to logical proofs. There
must be a valid cognition of the proposition at bottom. Now what is
i
yat tu andisiddhakryakranabhvavirodhdimlh kecit tark iti.

TCM, p. 731.
ni]
GANGESA'S SPECULATIONS ON REASONING
47
the nature of such cognicin? The consideration of probability makes

the admission almosc inescapable chat ic cannot be anything but
inference. But inference is based upon necessary concomitance and
necessary concomitance is always liable to doubt, for the elimination
of which recourse must be had to reasoning. And that will only
open the flood-gate of doubt, since reasoning depends for its validity
upon a belief in a necessary concomitance, which will never be exempt
from doubt, unless ic is removed by something other than reasoning1.
Gagesa following Udayana has shown that it is self-contradiction,
which is different from reasoning, chac can finally remove all doubts.
The view which has just been considered seems to be very old.
It has also been recorded by Srharsa. Gagesa only dittoes the
criticism thac has been propounded by Sriharsa. Buc the criticism of
rlharsa is more elaborate. Sriharsa observes thac even a craditional
belief handed down from undated past and shared by generations of
men does noc give it a privileged position. The question is whether
the belief is valid or noc. If ic is valid, che magnitude of the time
of ics currency and duracin or che superiority of number of the
persons encercaining ic has no logical relevancy to the question of
its truth. Even the valid cognition of one man will prove the validity
of a proposition. If however the belief be universally current, still ic
cannoc be accepced to be crue, unless an independenc proof of ics
validity can be produced. It is noc a face chac because a belief
is currenc from time immemorial or shared universally by che largesc
majoricy of men, ic must be true merely for that reason. In logic at
any rate che rule of majoricy cannoc command allegiance. People are
accuscomed co chink cheir bodies as cheir real selves. So when a
person is assaulted by a ruffian, he lodges a complaint in a law court
to the effect that he himself has been assaulted, akarcrya has
proved that even philosophers are not immune from the superstition,
so far as their activity and conduct are concerned, that the self is
identical with the body. The belief is ingrained in the animal mind
1 tatra pramnnuyoge 'numna eva paryavasnt. na ca vyptigrahnyathnupapattyai va tarkasya andisiddhavyptikatvajnnam ici vcyam. anupapatter apy anumnatvt.
TCM, P . 731.
48
INDUCTIVE REASONING
and is really universal.
[CHAP.
But the universality of the belief or the
infinitude
of the time of its currency does not vest it with the status
of truth.
So the contention
of the older Naiyayikas that doubt is
not possible with regard to such beliefs cannot be accepted as necessarily and universally
valid.
The question of doubt in philosophy is
not a question of psychology or practical
behaviour.
It is always
logical in character and as such can be refuted by logical considerations

alone.
Gagesa
recognised the justice and cogency of Srlharsa's
criticism1 and he has reproduced it faithfully.

Gagesa
has referred
to a class of Mlmmsists who held that
observation of concomitance alone was sufficient to prove the necessity

and universality of the relation between two terms.
Accordingly
they did not acknowledge the necessity of reasoning for the elimination of doubt regarding the necessity and the universality of a concomitance.
They did not think such doubt or the condition of doubt
to be an obstacle to the realization of necessary relation by means

of observation of concomitance.
Doubt
of the necessity of the
relation was thus dismissed as an ineffectual and innocuous event and

the employment
of
reasoning for the removal of such doubt was

2
deemed a superfluity .
Gagesa observes that the view smacks of a dangerous sophistry.
If the view is followed
with consistency and logical exactitude,
result will be the impossibility of belief

itself.
in the validity of
If the unfailing observation of concomitance
were regarded as
the proof of its universality and necessity, inference would

raison d'etre.
the
inference
have no
In every case of contemplated inference the observation
of the unfailing
concomitance of the probandum with the difference
of the subject (paksabheda) outside the subject (minor term) would

operate as a bar.
Suppose a man sets about to prove the existence of
fire in a hill on the observation of smoke.
But the inference will be
precluded by the knowledge of another concomitance.
Fire has been
1 andisiddhavyptiks te tark iti cet? na, tadbuddheh pramitatvsiddheh. sarire svtmapratyayasya tdrsasy'py apramtvopagatnt. anditvasiddhes co 'bhayatr' visest.
KhKh, p. 69 x.
2 TCM, pp. 733-34.
in]
GANGERS SPECULATIONS ON REASONING
49
observed as not only associated with smoke, but also with the difference
of the hill, in which the existence of fire is sought to be proved. It
is undeniable that fire has been observed in places different from the
hill, the subject of the contemplated inference. The presence of fire
in the hill concerned is certainly unproved and unknown, as
the previous knowledge of it would render recourse to inference
unnecessary. Now, the concomitance of fire with difference of hill
has been observed without any break or exception. In the case under
consideration, the difference of hill cannot exist in a hill, simply
because the hill is only a hill and not different from it. This knowledge of the absence of the difference, which has been shown to be a
necessary co-associate of fire, should make the absence of fire a necessary consequence. The absurdity shown to be inevitable in the case
of the hill will be inevitable in all other cases of inference. Whatever
be the subject and the predicate (sadbya), the fact will be true that
the concomitance of the predicate has been observed in whatever is
different from the subject and never in the subject. If the observation
of concomitance of the predicate with the difference from the subject
were the condition precedent of the said concomitance being universal
and necessary, then the subject would never be the locus of the
predicate, since it could not be the locus of difference from itself,
which is to be taken as the determinant concomitant (yyapaka) of
the predicate according to the view propounded by the Mimmsists1.
But the alleged absurdity cannot arise, if reasoning be regarded
as a necessary condition for the elimination of doubt, and doubt be
looked upon as an obstacle to the ascertainment of necessary concomitance. In spite of the observation of the concomitance of the
probandum with difference from the subject, the necessity of concomitance cannot be accepted, unless the last vestige of doubt about it be
removed. But one can legitimately doubt that the element of difference is irrelevant. Why should not fire, to take a concrete example,
exist in the hill, though it be different from the previously observed
1 ye ca anuklatarkam vinai va sahaerdidarsanamtrena vyptigraham
vadanti, tesm paksetaratvasya sdhyavypakatvagrahe anumnamtratn
ucchidyeta.
TCM, p. 734.
5o
INDUCTIVE REASONING
[CHAP.
places of its occurrence ? This doubt and query cannot be set at

rest unless the necessity of the concomitance of the probandum and
difference be placed on a footing of indubitable certitude. There is
no possible means other than reasoning which can remove this
doubt. The Naiyyika therefore would not admit the necessity of
such concomitance, because doubt of the possibility of its being
otherwise is not ruled out by a reductio ad absttrdttm.
Moreover, the denial of reasoning as a condition of the realization
of the universal concomitance would lead the Mimmsist to prove
more than he intends. Prabhkara seeks to prove the existence of
energy as the property of a cause. But whatever probans may be
adduced, the same can be assumed to prove something else in addition
to the energy. One may argue that fire is the cause of combustion,
because it is possessed of an energy for that. But the same probans
may be shown to prove that fire is possessed of an unknown property
also. Why should not fire possess some other unperceived attribute,
if it can possess an unperceived attribute like energy? The Mimmsist can prove his contention, only if he can adduce a reductib ad
abstfrdttm of the opposite possibility. But the reductio ad absurdum
of the opposite is what is called reasoning by the Naiyyika. The
Naiyyika insists that mere observation of concomitance is not adequate to the proof of its necessity, unless the opposite possibility can
be ruled ouc by a reductio ad absurdum. It may be hoped that the
Naiyyika has succeeded in making out a strong case for the necessity
of the admission of reasoning in the ascertainment of the necessity of
concomitance.
Let us conclude the dissertation on the utility of reasoning by a
resume of the views as given by Sakara Misra in his Vdivinoda.
According to Srvallabha and Gagesa the result of the application of
reasoning is the removal of doubt about the necessity of the relation
between the probans and the probandum. As regards the doubt about
the necessary concomitance of the ground and consequent in a reasoning,
that also can be removed by another reasoning. The employment of
a series of reasonings may be necessary, if a series of doubts emerge in
the mind. But that is not possible and so there is no necessity of
assuming contradiction by practical activity as the ultimate eliminator
ni]
GANGELA'S SPECULATIONS ON REASONING
51.
of doubt.
This seems to be the position of the author of the
JSIyyaltlvatJ. But we have seen that Udayana thinks contradiction
to be the ultimate solvent of doubt and its repudiation makes infinite
regress inevitable. The position has been endorsed by Gagesa in his
Tattvacintmani. There was a class of Naiyyikas, who preceded
Gagesa, who held that doubt is not possible with regard to universally
accepted beliefs handed down from generation to generation from an
undated past. We have seen that this position has been subjected to
smashing criticism by Gagesa. There was another class of Haiyyikas, who were of the opinion that the infinite recurrence of doubt was
psychologically impossible inasmuch as the mind cannot be infinitely
preoccupied with a particular mental event. The mind must move
towards some other objects and thus the chance of doubt cropping up
in an endless succession would be cut off. This position does not
seem to be logically sound. The charge of infinite regress is in every
case based upon logical possibility, which is never actualized. It is the
logical necessity and the actual unreality that make the charge of
infinite regress operate as an insuperable difficulty. In fact it is the
element of absurdity that is sought to be brought home by the show
of the logical necessity of the infinite regress. The absurdity would
not be real, if the alleged infinite regress were an actual fact. Another
drawback in this theory is that it misunderstands altogether the nature
and purport of a logical objection. Though it is not possible as a
matter of psychological truth that the doubts should go on occurring
indefinitely, the psychological evidence does not give an advantage.
Unless the doubt be shown to be logically absurd, the concomitance on
which the inference is based will not be established. It is not the
psychological cessation of doubt, but the logical impossibility of it,
that alone can guarantee the truth of inference or of the knowledge of
the concomitance which is its foundation1.
A CRITIQUE OF GAGESA'S SOLUTION
We feel called upon to attempt an evaluation of the solution
offered by Gagesa of the problem of infinite regress urged by Sriharsa.
1 y v , p. 58.
52
INDUCTIVE REASONING
[CHAP.
The role of an exponent, that has been assumed by us, should not
make it imperative that we must hold brief for any particular school;
and in the interest of intellectual honesty and unbiased quest of truth,
we should endeavour to convince ourselves whether a particular exposition is satisfactory or not. It will-be "obvious to a dispassionate enquirer
of truth, who will study the exposition of the views recorded above,
that the Naiyyika has laid exclusive stress upon the logical value of
empirical knowledge and practical activity.
Gangesa's solution, that
the contradiction constituted by activity by itself renders doubt impossible without presupposing another concomitance, does not seem to
be based upon a correct appraisal of Sriharsa's spirit.
It is hard to
think that practical activity removes doubt by dint of a natural causal

efficiency,
just as fire removes cold.
We have every reason to look
upon the opposition in question as logical and not causal.

activity removes doubt,
belief with activity.
The
because it involves the concomitance of true
It is assumed that if the belief were absent,
activity could not take place.
the
But this is tantamount to the admission
of a necessary concomitance between absence of activity and absence

of belief, which manifests
contradiction
afforded
itself in the form of a doubt.
Thus the
by activity derives its force from the belief in
another wider concomitance.
The latter belief is also equally open to
doubt, and so the charge of infinite regress cannot be avoided.
It
should further be taken into consideration that our practical activities

are not dependent upon a philosophical conviction of the indubitable
necessity of concomitance of facts.
It is not possible to doubt that
belief in the presence of fire, on the evidence of smoke, is a condition

of activity towards fire. But there is no necessity to suppose that this
belief should be unchallengeable on philosophical grounds.
In fact,
the degree and kind of certitude, that is sought to be achieved by

philosophical speculations, can neither be presupposed nor proved by
our work-a-day beliefs and activities.
Even our firmest convictions
are not founded
upon a profound
analysis of the conditions that give
them currency.
So the contention of Udayana and the interpretation
of Gagesa do not touch the fringe of the problem that has been raised
by Sriharâ. The pragmatic consequences of beliefs have by themselves
little or no metaphysical validity.
The utmost that can be asserted in
ml
53
favour of the Naiylyikas is that they have succeeded in giving an

explanation, which is valid for the practical purposes of life, the claims
of which should also be duly considered so that scepticism too cannot
be pushed without limit. When our doubts come in conflict'"with the
unchallengeable and self-evident convictions, than which nothing can
be surer, then certainly it is the doubts which are pushed to the wall.
This has been the criticism of Raghuntha Siromani in his commentary
on the Khandanakhandakhdya,
quoted by Vamsidhara Misra in the
Tattvavibhkara, the commentary on the Smkhyatattvakaamudt. We
could not refrain from reproducing the criticism though it gives a
hard shaking to the defence of Udayana and Gagesa1.
STV, p/78.
CHAPTER IV
SECTION I
We have considered the nature of the service that is rendered by

reasoning and have also shown the necessity of reasoning as an instrument for the elimination of doubt. We now propose to consider the
claims of reasoning as an organ of knowledge. The tradition is
unbroken among the Naiyyikas that reasoning is not an instrument
of knowledge, though it is a necessary modus operand* of the same.
In the course of our work we will show that the Mdhvas, the Rmnujyas, and the Jainas have accorded to reasoning the status of a
cognitive instrument. Vatsyayana, Uddyotakara, and Vacaspati regard
reasoning rather as a helpful condition and an auxiliary to an independent instrument1. Udayana seeks to establish the subordinate character
and role of reasoning from a different line of approach. He
asserts that reasoning is not knowledge, but an assumption and an
imaginary imposition (hryaropa). The assumption of an untrue
ground is meant to make the assumption of an untrue consequent
unavoidable2. But the question arises, why should the ground and
the consequent and the relation between them be untrue? Taken by
themselves each one of these factors of the judgment in question is
true. Thus for instance, smoke is a concomitant o fire and their
relation is also necessarily true. If the relation in question were not
true, the reasoning would be only a counterfeit show. It is admitted
even by those who do not acknowledge the objective truth of the
judgment set forth by reasoning that there are a true reasoning and a
false simulation of reasoning (tarkbhas). It stands to reason to conclude that reasoning should not be regarded as an assumption (arofa)3.
1 tarko na pramnasamgrhlto, na pramnntaram pramnnm anugrliakas
tattvajnnsya kalpate.
NB, p. 53; vide also NV, p. 53; NVTT, pp. 53-54.
2 NVTTP, p. 325.
3 nanu nyam ropah, pdypdakasvarpe tasya ttvikatvt, n'pi tayh
samsarge, tasy'py abdhant
na tarkasya hryaroparpatvam.
p, pp. 325-26.
iv]
THE LOGICAL STATUS OF REASONIiNG
55
It is answered by the Naiyyika that though the ground and the

consequent are real by themselves and the relation between them is
also necessarily true, still there is no real relation between the ground
and consequent on the one hand and the locus on the other in which
they are sought co be affirmed.
The form of such assertion is therefore
necessarily hypothetical, which implies that the assertion is only an

attempt at affirmation and not a statement of categorical truth.
In a
hypothetical judgment it is not the truth, of the ground or of the

consequent or of the relation between them, that is the matter of
assumption; but the affirmation
of the unreal relation of the ground
and consequently of the consequent with a subject is sought to be

effected
by means of the same.
In fact, the ground does not belong
to the subject and consequently the consequent cannot exist in it.

Thus, a hypothetical judgment is not a real judgment at all, since it
does not correspond to a true and actual objective situation.
It is
the unreality of the relation of the ground with the subject that
makes the judgment unreal and false.
A false judgment, asserted
in full consciousness of its falsity, can be characterised as only a

deliberate and voluntary
it an assumption.
proposition.
imposition
and so the
Naiyyika
Let us consider the following hypothetical judgment,
which is an illustration of what is called reasoning (tarka).

the lake were possessed of smoke,
It is a fact
calls
An example will drive home the truth of the
that
the
possess smoke or fire.
subject,
it would be possessed of
namely
the
lake,
does
"If
fire".
not
But still the formal hypothetical assertion
makes us understand as if a relation were actually present between the

lake and the smoke.
An
assumption is the understanding of a rela-
tion, which does not actually exist.
Thus, smoke and fire both are
only ascribed to the lake, though they do not belong to it.
The
hypothetical assumption thus differs in a material respect from inference.

subject.
In inference the probans and the probandum must exist in the

The lack of the former condition makes the probans non-
existent in the subject (svar-pasiddha-hetu) and thus makes the minor

premise impossible.
The non-existence of the probandum in che
subject constitutes the fallacy of contradicted reason (badha).
But
these very conditions are essential to a hypothetical judgment of the
56
INDUCTIVE REASONING
[CHAP.
nature of assumption (ropa). The assumption of a false probans

makes the admission of a false probandum inescapable,
it is useful
as a reductio ad absurdurn, and the unreality of the probans and the

probandum is the very condition of the redttctio ad absurdum
becom-
ing effective .
The problem, which we have set out to determine, is the logical
status of reasoning,
in the course of our enquiry we have found that
reasoning gives assumption and not knowledge.
It seems necessary
that we should determine the epistemological nature of assumption,

as that will make the determination of its logical status an easy natural
procedure.
Uddyotakara has observed that the nature of reasoning is
difficult to determine.
The difficulty is apparent from the fact that
there is such wide divergence of opinion regarding the nature of

reasoning.
The Naiyyikas regard it as a separate category.
The
author of the Vaisesikastra as well as Prasastapda do not admit

the reality of such a category.
The Mlmamsists and the Buddhists
have sought to include it within the fold of inference.

regard
it
as
an
independent
instrument
The Jainas
of knowledge.
The
difference of opinion among philosophers is proof of the intractability of the problem.
Certainly, it is difficult to ascertain whether
there can be a species of cognition, which is neither doubt nor certitude.

The difficulty
is further accentuated by the consideration that the
existence of a cognition is known by introspection in the form 'I

know'.
Similarly, desire and aversion are known by introspection
as 'I desire' and 'I dislike it'.
But in the case of assumption, the
verdict of introspection is not so definite and that is the cause of

its difficulty and consequently of the difference of the views among
philosophers.
A class of philosophers,
though they regard it as
different from doubt and certitude, still think it to be a species of

cognition.
But a question arises whether assumption, which is the
natural form in which reasoning manifests itself, is felt by introspection in the form, 'I know'.
gives itself out as follows.
To cite one or two instances, reasoning
'If the hill were devoid of fire, it would
i mai vam. tarko hi nirvahnir ayam nirdhma iti nirvahnkvdin she

'dantvasya smndhikaranyanii ullikhati. etan-mlakam abhidhanam api tatsmndhikaranyavisayam, na ca tad asti vstavam.
NVTTTPP, p. 326.
iv]
be devoid of smoke'.
be possessed of
fire'.
57
'If the lake were possessed of smoke, it wouid

But is it definite that the character of the
hypothetical judgment is felt as 'I know the absence of smoke m

the hill or the presence of fire in the lake' ?
Certainly the cogni-
tive character is not at all definite, as in that case there would be

no reason for divergence of views.
Some neo-logicians expound the
introspection as 'I assume (tarkaydmi) absence of smoke or presence

of fire'. But this seems to be an attempt to avoid the real problem.
Is assumption felt as a species of cognition ?
There is no doubt that when a man sees a jar, he feels by introspection that he sees it and also knows it as jar.
It never happens
that the man thinks that though he perceives the jar, he does not know
it.
So with regard to inference.
A person*, who infers the presence
of fire in a hill, is made sure by introspection that he not only

fire, but also knows it.
be necessary concomitants of cognition.

to be so ?
infers
Perception and inference are thus seen to

But is assumption known
The case seems to be quite otherwise.
A person actu-
ally sees smoke and then reasons that, 'if the hill be devoid of fire,
it must be devoid of smoke'.
But is this
reasoning competent
to give knowledge of absence of smoke in the hill ?
It seems
preposterous to suppose that the same person, who sees smoke, would
be made to posit his knowledge of the absence of smoke ; but this
is the result, in which the reasoning is shown to culminate.
it assumption or anything you like, but it is difficult
Call
to regard it as
a species of cognition.
There is another consideration, which makes it next to impossible
to classify it under cognition.
judgment, but the adjective
difference
Reasoning is called a hypothetical

'hypothetical'
from a judgment proper.
covers a multitude of
It is a truism thac a judgment is
cognisant of a relation between two terms, adjective and substantive,

subject and predicate.
If there be no knowledge of a relation, or if
there be the knowledge that relation does noc exist, no judgment

will be possible.
The definite knowledge, that a relation does noc
exist, is called by the Naiyyikas knowledge of concradiction, which

is regarded by chem as an impediment to the realization of judgment.
But the very reverse is seen to be the case with a hypothetical
$8
INDUCTIVE REASONING
judgment in which reasoning manifests itself.
[CHAP.
The knowledge of
contradiction is rather the essential condition of

judgment.
hypothetical
So what is an impediment to judgment in general is
the very condition of reasoning, hypothetical judgment and assumption, which mean one and the same thing.
Now, the predicate e.g.
the absence of smoke, the assumption of which is necessitated by

reasoning, is known to be absent in the hill.
Thus, the definite
knowledge of the absence of the predicate in the subject is the essential condition of assumption.
It would
not be an assumption,
if the posited predicate were not known to be contrary to what is

true.
The upshot is that the condition of hypothetical judgment
is the very reverse of what makes judgment possible.
Viewed from
the point of view of the conditions of their genesis, the hypothetical

judgment is thus diametrically opposed to judgment proper.
It is
difficult to suppose that though produced by the exact opposites,

the effects should come to have a common character.
The commu-
nity or divergence of nature of effects is determined by the corresponding character of the conditions.
they
Linen is not a jar, because
are possessed of contradictorily opposed character.
To the
question why they should have such opposite character, the answer
will be found in the difference of the causal conditions of the effects.
Reasoning or hypothetical judgment has been shown to be derived
from causal conditions, which are the diametrical opposites of the
conditions of categorical judgments.
As we have indicated before,
the designation of hypothetical judgment as judgment is only an

extension of linguistic usage on the ground of formal analogy and
has no material validity and justification.
"The internal structure
of an assumption is the same as that of a judgment.
The only
difference is that a judgment is an assertion, while an assumption is

not...
And it is true of an assumption, as it is of a judgment,
that it has meaning". 1
Barring this similarity, the two have no
community so far as its logical value is considered.

There is another fundamental difference, which puts hypothetical
judgments in a class apart from categorical judgments.
i
The Nature of Existence, vol. u, p. 106.
A judgment,
iv]
59
which refers to a desirable objective or to a means of its realization,

produces a conation1 for the same.
Thus conation is always the
consequence of a judgment and not vice versa.
But a hypothetical
judgment or an assumption, which are the same thing, cannot produce

a conation about its object.
conation.
On the contrary it is itself the result of a
An assumption is the consequence of a conation of the form
"Let there be a cognition of an attribute in a subject, which we

know to be devoid of the same".
In fact assumption is an attempted
judgment for which none of the necessary conditions is in existence.

It is difficult
therefore to suppose that assumption should be a species
of cognition.
There is another characteristic, which distinguishes assumption
from knowledge or judgment.
A judgment is either invalidated by an
opposite knowledge or it itself invalidates the latter.
But assumption
can neither invalidate, nor can be invalidated by any judgment, inasmuch as the knowledge of its invalidity and falsity is its very condition.
Assumption makes an attempt to think of an attribute as exis-
tent in a substratum, in which it is known to be non-existent.

knowledge of contradiction were to invalidate
If the
an assumption, it would
not come into being at all, since the very condition of its origination
is the knowledge of the aforesaid contradiction.
It follows from the consideration of the nature of assumption, of its
object, and of its conditions, that assumption or hypothetical judgment
cannot be placed under the same category with knowledge.
The
Naiyayikas were cognisant of this fundamental difference of character

and so they have called it hypothetical knowledge or assumption. The
difference of terminology only serves to stress its difference as a psychical activity.
It is refreshing to note that McTaggart also recognises
assumption as a psychical activity, which is placed by him in a different category from other psychical activities.
He of course places
assumption together with judgments and volitions under cogitation,

which is wide enough to include what is an accredited cognition and
what is not so.
r 'Conation' is used here to mean desire for the attainment of an object
together with a volitional activitv.
6o
INDUCTIVE REASONING
[CHAP.
A class of Naiyyikas has regarded assumption as a species of

mental perception. But this seems to be unjustifiable. A perception,
mental or sensuous, is the generator of a latent mental trace, which
produces recollection. But assumption cannot be supposed to generate
either a mental trace or a memory, and thus it is misleading to call it
a case of perceptual cognition. Some modern Naiyyikas have sought
to deny that reasoning as hypothetical judgment or assumption is the
product of a conation. But this seems to be due to a confusion. It
would not be an assumption, if it were not the effect of a conation.
There are some other Naiyyikas who have made an attempt to subsume assumption under erroneous knowledge. But this seems to be
also an inaccurate appraisal of the nature of assumption. If assumption
were a kind of error, it must be shown, in the first place, to be produced by some defect ; and, in the second place, it must be shown
liable to be contradicted by a subsequent judgment. But it is neither
the effect of a defect nor is it liable to be contradicted. In the circumstances, the assertion that assumption, though not superseded by a
subsequent experience, is an error, is tantamount to admission of an
error, which does not admit of contradiction. But this is obviously a
case of contradiction in terms. If the Naiyyika seeks to justify his
classification of assumption under error on the ground of its recognised
difference from truth, then there would be no excuse for not placing
assumption under the category of true knowledge. Assumption is
found to share these two fundamental characteristics of true knowledge,
namely, that it is not the effect of a defect in any condition of knowledge, and secondly, it is not contradicted by a subsequent experience.
We are constrained to observe that Udayana in his work Kiranvati
has also sought to place it under the head of erroneous knowledge1
out of loyalty to Prasastapda, who does not admit a third variety of
knowledge apart from error and truth. But our criticism will also
apply to Udayana. To sum up, assumption has none of the characteristics of knowledge proper.
It does not produce memory or any
volitional activity, which are invariably the consequences of knowledge.
This shows that assumption cannot claim to have the logical character
l viparyayabhedatvt prthan no 'padistah.
KV, p. 260.
iv]
61
of truth or falsity ^ as it lacks the fundamental psychical characteristics

of knowledge, which alone can be true or false.
Assumption
again is to
be distinguished from
error.
Error is also the attribution of a false predicate to a subject
and in assumption also there is the same false attribution. But
whereas in error the erring man is not conscious of the falsity,
in assumption we are aware that the attribute, which is ascribed
to the subject, does not actually belong to it. In Sanskrit terminology both these cases are called cases of ascription (aropa). But
error is an ascription, which is not conscious of itself and assumption
is on the contrary conscious of the falsity of the ascription
(abryropa). For the sake of brevity and simplicity we propose to
call error an unconscious ascription and assumption a conscious
ascription. We shall use the term assumption as equivalent to
conscious ascription henceforward.
We have found that both error and assumption are cases of
ascription. Ascription can again be divided into (i), ascription
of identity and (2), ascription of relation. The ascription of
identity makes the substratum and object of ascription appear as
identical. The ascription of relation serves to make them understood as related. If the numerical difference or the unrelatedness
of the terms is not felt owing to some defect, the ascription of
identity or of relation will be a case of error. But if the knowledge
of numerical difference and of the lack of relation be present in
the mind and still identity or relation be superimposed by a fiat
of will it would be a case of assumption. In poetical figures,
particularly metaphor (r-paka), one thing is felt to be identical
with another. This conscious identification of different objects,
(for example, when the face of a lovely woman is imagined to be
identical with the moon or lotus), is a case of assumption of identity1.
The poet's imaging of the face as the moon is inspired by their
close similarity.
But whatever may be the psychological reason for
the attempted assertion of identity or of relation in assumption, the
fact is indubitable that the attempted identification or relation takes
1
Sdhityadtirpana. pp. 440-41.
6z
INDUCTIVE REASONING
[CHAP.
place in full consciousness of their unreality. Thus the hypothetical

assertion, viz. "If the lake were possessed of smoke, it would be possessed of fire," is an instance of assumption or an imaging, as called
by McTaggart1. In McTaggart's terminology the psychical process
involved is called imagino;.
The separate existence of imaging or conscious ascription as
psychical reality cannot be denied or ignored. It is Udayana who
seems to have drawn our attention to the reality of this psychical existent. In fact, the reality of this psychical process was implicitly
recognised even in Vedic times. The processes of meditation recommended in the Upanisads and the later Paurnic literature are cases of
conscious ascription. Thus it is recommended that one should meditate
the mind as the Absolute Brahman or a male person as fire or a woman
as mother2. The fact that the object of meditation is not what it is
recommended to be meditated is perceived throughout the duration
of the religious exercise. This also explains the etiology of the worship
of images as Gods and Goddesses. The identification is always deliberate, whether inspired by similarity or otherwise.
Reasoning has been shown to be possessed of the character of
assumption. But reasoning is a hypothetical judgment and as such
consists of two members. The first member s the assumption of the
ground (vyapyhryaropa) and the second is the assumption of the
consequent (yyapakbryropa). The first is the condition and the
second is the consequence. Now, assumption (abryajnna) is defined
by the neo-logicians as "Judgment having for its predicate an attribute the opposite of which is known to be a real determination of the
subject3". Assumption as defined here cannot be confounded with
authentic knowledge, the characteristic of which is the actual presence
of the predicated attribute in the subject. Thus, for instance, the
judgments, "The hill is possessed of smoke", or "The lake abounds
in weeds and fish", are instances of authentic knowledge. The reason
1 The Nature of Existence, vol. ii, p. 106.
2 mano brahme'ty upsta. Chnd. Up. III. 18.
puruso vava gautamgnih. Ibid., V. 7.
3 svavirudclhadharniadharmiivacchcclakakam svaprakrakam janam hryam.
iv 1
of the authenticity lies in the face that smoke, winch is predicated of

the hill, is actually present in it and the weeds and fish predicated of
the lake actually exist in the subject.
brror is also a judgment, but
the predicate docs not in point of reality belong to the subject.
The
main point of its difference from assumption consists in the fact that
though in error the predicate does not belong to the subject as a
matter of ontological fact, yet it is not realized that the subject does
not actually possess the predicate, but on the contrary has an attribute, which is the contradictory of what is predicated.
In fact, the
knowledge of the absence of the predicate and of the presence of its

opposite in the subject would make an error an impossibility.
In
assumption, however, the person who makes this assumption is quite

alive to the unreality of the predicate as an attribute of the subject and
is fully conscious that the subject on the contrary possesses an attribute,
which is contradictorily opposed to the
predicate.
assumption is a false knowledge known to be false.
In one word,
This knowledge
of falsity of the attribute, which is predicated notwithstanding the

consciousness of its falsity marks it out from error, which is false
knowledge honestly believed to be true.
It is for this reason we have
taken exception to the attempt of a later section of Naiyayikas to

include it under error.
Reasoning
assumptions.
is a complex judgment or rather a compound of

In it one assumption makes another assumption an
irresistible logical consequence.
Thus, for example, when an arguer
seeks to rebut the denial of fire by an opponent in a place, which

is known to possess smoke, he pretends to accept the opponent's
assertion of the non-existence of fire as a truth.
But he makes the
assertion of the opponent the ground for enforced admission of the

non-existence of smoke, which flatly contradicts the accepted datum
of experience.
The argument is put in the form, "If the hill did
not possess fire as you maintain, it must not possess smoke".
That
absence of fire necessarily involves absence of smoke is not liable to

doubt even by the opponent.
So the demonstration of the necessity
of denial of smoke amounts to the demonstration of an absurdity,

which is bound to be admitted as a necessary consequence of the
denial of
fire.
The particle 'if, which is prefixed to the preliminary
6*4
INDUCTIVE REASONING
[CHAP.
assertion, indicates that the assertor of the hypothetical

does not believe in its truth.
proposition
Both the clauses, the ground and the
consequent, are known to be false and are also proved in the end
to be absurd.
This rather elaborate consideration of the nature of the hypothetical judgment, in which reasoning invariably delivers itself, will
make the determination of the logical problem an easy procedure.
The foregoing discussion has resulted in showing that reasoning is
not knowledge, but a compound of assumptions.
So the logical
problem as to whether reasoning should be regarded as an organ of

valid knowledge should not arise.
But philosophers are not prepared
to let down the matter on such rather easy terms.
It has been urged
that, whatever be the psychological character of reasoning, its contribution to the determination of truth by means of the elimination of
doubt about the opposite possibility is undeniable.
The intention
of the opponent seems to be that reasoning, though it may not be

itself a species of valid knowledge, may be an instrument of valid
knowledge.
The opponent is also not prepared to accept the position
that Aksapda is definite that reasoning is not a cognitive instrument.

The mere omission of inclusion of reasoning in the list of the instruments of knowledge docs not prove that reasoning cannot be an
additional instrument of knowledge.
The argumentum
ex silencio
has no force even from the standpoint of the author of the
Nyya-
stra, who does ^not care to be exhaustive in his statement.
Thus,
for instance, Aksapda does not include doubt and the like in the
list of twelve knowables formulated
by him.
But that does not
prevent the recognition of doubt etc. as additional knowables.
So
it muse be regarded as an open question that though reasoning may

have been omitted from the list of cognitive instruments, it may be
regarded as an additional instrument of knowledge1.
But this contention of the opponent cannot be accepted as the
authentic
representation
of
the
intention
of
the
author of the
i astu tarhi pramnapadasamgrhitebhyah pramnebhyah pramnntaram

asamgrhtam pramnapadena, prariieyapacen 'evasamgrhith
samsaydayah
prameyh.
NVTT, p. 53-
iv]
THE LOGICAL STATUS, OF REASONING
65
Nyyastra.
We shall produce reasons to show that the position
maintained by the opponent is incorrect, both psychologically and logically.
But even from the point of view of historical and traditional
interpretation
the position will be found to be indefensible.
It is a
fact that Vtsyyana has definitely asserted that reasoning is not included under any one of the four recognised cognitive instruments nor it
is an additional organ of knowledge1.
Uddyotakara observes that the
reason for rejecting the claim of reasoning to be a valid instrument of

knowledge lies in the fact that reasoning does not produce certitude of
belief in an object of knowledge2.
Udayana explicates the meaning
of Uddyotakara's statement by pointing out that reasoning cannot

independently and directly produce certitude.
It is in fact dependent
upon the service of a recognised cognitive instrument and thus has not
independent efficacy as productive of knowledge.
A cognitive instru-
ment is what is a direct condition of certain knowledge by dint of its

own causal efficiency.
But reasoning does not and cannot produce
certitude independently of an organ of knowledge and so it is not

regarded as an independent cognitive
dinate and subservient to an
instrument.
It is always subor-
independent cognitive instrument and so
can never appropriate the status of the latter 3 .

Udayana observes that if reasoning were to be an independent
instrument and were to be accepted as such, it should be shown to
produce valid knowledge in either of these two ways.
A cognitive
instrument may be either a valid cognition by itself, or something

different from cognition.
Thus, the instrument of inference is a valid
knowledge of necessary concomitance.
But the instrument of percep-
tual knowledge is a sense-organ, which is a material entity and so nonpsychical.
The difference
between the two kinds of cognitive instru-
ments is that the first type produces its effect only when it is known;
the other type is competent to produce knowledge by remaining un1 tarko na pramnairngrhto na pramnntaram.
2 pramnarn paiicchedakam na tarkah.
3 NVTTP, p. 324.
NB, P . 53tasmn na pramanam.

NV, p. 53.
66
INDUCTIVE REASONING
known.
[CHAP.
About the first type ic should be recognised that it can be an
instrument of knowledge, if it is itself a valid cognition.
Now, reason-
ing being a psychical reality cannot operate like sense-organ as unknown and unperceived.
produce credentials
If it is to be a cognitive instrument it must
of its being itself a valid cognition.
But we have
seen that reasoning is only an assumption and not a cognition at all.

And so the question of its being a valid cognition does not arise.
But
a cognitive instrument is necessarily found to be either an unperceived

sense-organ or a valid cognition.
As reasoning is neither the one nor
the other it cannot be regarded as an independent instrument of

knowledge 1 .
Again, reasoning can be an instrument of valid cognition, only
if it can be shown that there is a valid cognition, which is its result.
In inference the resultant knowledge is the conclusion.
Let it be
supposed that reasoning as a cognitive instrument has this conclusion

as its effect.
But reasoning has been found to be the assumption of a
consequent necessitated by the assumption of a ground.
If reasoning
as the assumption of the consequent is to be the organ of inference, it

must be shown that the object of the said assumption stands in a
definite and necessary relation with the object of inference.
Thus,
for instance, knowledge of smoke produces knowledge of fire, because

smoke is known to be necessarily related to fire. But the object of
assumption is not seen to stand in a necessary relation to an object of
inference.
The reasoning is of the form, 'If it were devoid of fire,
it would be devoid of smoke".
Here the assumption of the negation
of smoke has for its object the said negation of smoke, which has no
direct relation with fire as the object of inference.
It can at most be
said that the negation of smoke s only concomitant with the negation
of fire and the object of inference is only the opposite of the said
negation, namely,
fire.
But this relation of the object of inference
with the opposite of the object of assumption is not a necessary and

essential relation.
Thus, assumption cannot be supposed to lead to
i ctad uktam bhavatitarko hi pramriipam pramnam syt ligadarsanavat,

apramipam v indiiyclivat. na tvad dyali. tasya ahryatroparpatvt. n 'pi
dvityah
NVTTP, p. 326.
iv]
67
the knowledge of the object of inference. There is a necessary relation, no doubt, between the negation of negation of smoke, that is to
say, between smoke, and fire, which is the object of inference. The
knowledge of smoke is the necessary cause of the knowledge of fire,
since smoke has a necessary relation with fire. The advocate of the
validity of reasoning can at most assert that the object of inference is
necessarily related with the negation of the object of reasoning. But
this is not equivalent to the proof that the object of reasoning has a
necessary relation with the object of inference. Thus reasoning cannot
be regarded as an instrument of the inferential knowledge. It can at
most be regarded as a contributory to the inferential knowledge and
this has been done by the Naiyyika, who accepts reasoning as an
auxiliary to a cognitive instrument1.
Furthermore, the advocate of the independent validity of reasoning
cannot point to any independent valid cognition as its result. If it is
intended that reasoning is the instrument of the cognition, which is
produced by a sense-organ or an instrument of inference, then it must
be admitted that reasoning can produce such cognition only when it is
associated with those accredited instruments. But this amounts to a
confession that reasoning only serves as an auxiliary to those instruments. The Naiyyika also admits that reasoning has a bearing upon
a valid cognition that is produced by an instrument of knowledge, and
this bearing is due to the fact thac reasoning is only a function of an
instrument of knowledge and has no independent status of its own2.
It is true that reasoning is the assumption of the consequent necessitated by the assumption of the ground. In other words, reasoning
is a reductio ad absurdum. Though it is essentially an assumption, it
1 ......svavisayavypyaviparyaye hy asya prmanyam sakkam, tena ca n
'sya niymakah sambandhah, asambaddhasya gamakatve 'tiprasagah, tadvisayaviparyayasya tu tena saha svbhavikah sambandho' s. tena tatsamvalitasya pravrttes
taditikartavyattvam eva.
NVTITP, p. 327.
2 W e have already shown that function is the product of a cause, which
is the necessary condition of* die production of the effect for which the cause
is responsible. In other words, function is the operation (vypra) of the cause
by means of which the cause produces the effect . Thus function is the intermediate stage between the cause and effect, being related to the cause as its
effect and to the effect as its cause, (karanajanyah karanajanyajan&kQ. v
68
INDUCTIVE REASONING
[CHAP.
serves to demonstrate the falsity of the opposite conclusion and thus by

the implication of opposition, becomes an instrument of proving the
truth of a given argument.
An example will make the position clear.
Suppose, one seeks to prove the existence of fire on the ground of the
existence of smoke.
Suppose, again, that an opponent questions the
validity of the inference of fire. The opposition is rebutted by the

assertion that if there were no fire, there would be no smoke. . The
negation of smoke is however unreal and the unreality of the negation
of smoke proves the unreality of the negation of fire. The proof of
the unreality of the negation of fire is tantamount to the proof of the
reality of fire. Thus, though an assumption, reasoning really culminates in the proof of the probandum.
The plausibility of the oppo-
nents argument is undeniable, but it will not stand scrutiny.
It is
true that reasoning shows that the denial of fire necessitates the denial
of smoke; and the denial of smoke is contradicted by the unmistakable
perception of smoke. The result is that reasoning with the aid oEperceptual knowledge serves to establish the proposition that the subject is
possessed of the negation of negation of smoke, which is the equivalent
of smoke. If smoke or negation of negation of smoke be the ground of
the proof: o the negation of negation of fire, it becomes only a case of.
inference of fire by means of smoke.
The only point of difference
is that in the: positive inference, fire is proved by smoke and in the

negative inference the negation of negation of fire is proved by the
negation of negation o smoke.
The difference is only formal.
The
truth of the matter is that reasoning can be regarded as an independent

organ of knowledge only whea it surrenders itself and gives rise to an
inference^
It is for this reason asserted by the Naiyiyika that reason-
ing cannot have the status of an independent instrument of knowledge.

There is absolutely; n room for doubt that reasoning is a necessary
step, but its necessity and utility lie in the removal o an impediment,
which facilitates the functioning
of an independent organ of know-
ledge, held so long in abeyance owing to the presence of an obstacle1.

There is another consideration, which detracts from the claim of.
reasoning to be an independent cognitive instrument.
? NVTTP, p. 327; vide also NVTTTPP, p. 327.
Though it is a
iv]
69
useful auxiliary to a cognitive instrument, it is not a universal and

necessary factor even in its subordinate role. In perceptual cognition
reasoning has no part to play.
limited.
Thus its scope as an auxiliary is only
It would not, however, be a sound objection if it is argued
that when reasoning is not a necessary condition of knowledge even as

an auxiliary, it should be totally dispensed wich.
Because the lack of
its universal necessity does not prove that it is not necessary n any
case.
la fact reasoning is called into requisition where there is doubt
of the opposite possibility.
And this doubt prevents the successful
operation of a cognitive instrument.
Reasoning thus renders a useful
and necessary service by eliminating the doubt, which

cognitive instrument in question in suspension.
held the
It miglit be urged
against this claim of reasoning that though it be a necessary instrument for the elimination of doubt, reasoning cannot be regarded even
as an auxiliary to a cognitive instrument.
So far as the knowledge of
an object is concerned it can. be achieved only By an accredited cognitive instrument, and reasoning has. no direct or indirect contribution to
the realization of such knowledge.
It has been shown that there are
cases of knowledge which are not preceded by reasoning.
Thus,
however necessary be the services of reasoning, its bearing upon the

knowledge of an object achieved by means of a cognitive instrument
is found to be nil.
It has been claimed by some Naiyyikas that
reasoning is a function of a cognitive instrument and we liave already

explained this position.
But it is now contended that reasoning cannot
even have the status of such a function.
The fact, that there are
cases of knowledge impreceded by the service of reasoning and the

further consideration, that reasoning has no bearing upon the achievement of knowledge, prove to demonstration that the claim of reasoning
being a necessary function of a cognitive instrument is unfounded and
indefensible1'.
The contention of the opponent serves to show that reasoning
has no necessary connexion with a cognitive instrument
resultant knowledge produced by it.
or
the
But the contention cannot be
accepted as a true evaluation of the character of the service of reasont
NVTTPP, p. 328.
70
INDUCTIVE REASONING
[CHAP.
ing. It is not denied that there are cases of knowledge, which are independent of the service of reasoning. But there are also cases of knowledge, which are produced by cognitive instruments only when they
are aided by reasoning. The qualitative difference of the resultant
cognitions is proof of the efficacy and bearing of reasoning upon the
former. The knowledge that is secured by a cognitive instrument
independently of reasoning is only assertory in character, whereas
the cognition produced by a cognitive instrument reinforced by
reasoning is definitive and apodeictic in nature1. Thus it is universally true - that definitive knowledge has always reasoning for its
antecedent condition. Though reasoning cannot be an independent
instrument of knowledge, its necessity and utility cannot be denied.
The only explanation of this is that reasoning by elimination of doubt
serves to put the resultant knowledge on a footing of definitive
certitude and it can achieve this if it can serve as the necessary
function of an instrument3.
Our exposition of the Naiyyika's speculations on the logical
status of reasoning will remain incomplete if we do not advert to a.
point of striking importance, which we had occasion to mention in
connexion with our exposition of Uddyotakara's standpoint.
It
has been observed that reasoning serves to create a presumption in
favour of one of the competing alternatives by setting forth the
logical grounds on its behalf. But the presumption or a sense of
heightened probability should not be construed as a kind of belief
falling short of the rank of the certitude. Udayana interprets the
concept of probability here as equivalent to the proof of non-contradiction3. Presumption is at most an attenuated form of doubt in
which the degree of probability in favour of one of the alternatives
is enormously greater than in the other. But this carmot be the
full nature of the service rendered by reasoning. So long as the
other alternative is not ruled out as an impossibility, the truth of
a particular position cannot be accepted as established. Udayana
1 NVTTPP, p. 328.
2 Ibid.
3 sambhvan ce' h'virodhamatram.
NVTTP, p. 329.
tv]
71
therefore interprets this concept of probability as nothing short of

the demonstration that the given position is not contradicted by the
testimony of any accredited organ of knowledge. This explanation
of Udayana again furnishes the explanation of the assertion,
which is frequently made by the Naiyayikas, that reasoning
lends support to a cognitive instrument at work. Udayana further
observes that this is also in agreement with the view of Vacaspati
Misra who asserts that reasoning consists in the demonstration of
logical tenabiity or untenability of the competing alternatives.
Vacaspati states that reasoning lends support to a cognitive instrument by setting the seal of approval upon its finding1. Udayana
observes that this process of approval consists in the demonstration
that a particular proposition is not in opposition to the deliverances
of accredited sources of knowledge2. Let us explain the position
by means of a concrete illustration. The inference of fire on the
ground of smoke is said to be approved and supported by reasoning.
The approval and support consist in showing that the absence of
fire is incompatible with the presence of smoke and so by opposition
the presence of fire is true. The existence of fire is in concord
with the evidence of all the sources of knowledge that are relevant
to it. And the concord and harmony of fire with smoke is deduced
from the absurdity of the opposite possibility.
Reasoning qua
reductio ad absurdum shows that the absence of fire is incompatible
with the presence of smoke. This is the interpretation of Vardhamna and we think that Vardhamna is more correct and precise
in his statement than even Udayana and Uddyottkara3. Uddyotakara
characterized the service rendered by reasoning as the creation of
presumption4. This was interpreted by Vacaspati as the demonstration
of logical validity and invalidity of the competing alternatives.
Udayana clarified the conception further by showing it to be the
demonstration of absence of conflict and opposition with established
truths. Vardhamna has taken pains to clarify this demonstration
1 tarkas ca.......
pramnam anugrhnti.
2 NVTTP, p. 329.
pramnam yukte tattve pravaifcamanam anujaruri

NVTT, p. 54.
3 Ibid.
4 NV, p. 326.
7*
INDUCTIVE REASONING
[CHA>.
of non-contradiction by showing that it is only an indirect result

achieved by proving the absurdity of the contradictory possibility.
To round off the discourse we may observe that the proof of the existence of fire is afforded by inference alone, and reasoning only supports
the deliverance of inference by showing that the non-existence of fire
is logically incongruent with the presence of smoke.
SECTION
II
THE LOGICAL STATUS OF REASONING. AS ELUCIDATED IN THE
NYYAKAKDALI AND IN THE VYOMAVATl

It is necessary in order to make our treatment of reasoning complete and thorough that we should take into account the exposition of
reasoning as found
in the Nyayakandati
Srldhara in the Nyayakandati
and in the Vyomavati. .
alludes to the Naiyyika according to
whom reasoning has . been supposed to be a species of knowledge,

which cannot be subsumed either under doubt or certitude 1 .
It has
been thought to be of the nature of presumption which however serves

to eliminate the doubt of the opposite alternative and thus paves the
way for the effective application of an organ of knowledge.
Thus,
for instance-, whe*n one philosopher asserts that the self is subject to
birth and death and another asserts that it is unborn and immortal,
these two conflicting assertions necessarily give rise to doubt.
Now
the considerationthat if the soul were subject to birth and death, then
rebirth and emancipation would be impossible, and if it were, on the
other hand, an eternal principle, then its association with the successive physicalorganisms called samsara, and it* ultimate emancipation
from this cycle of birth and death, which are the inevitable fate of
physical organisms, would be possiblecreates a presumption in favouf
of the immortality of the soul, which is however proved by inference.
Reasoning thus helps the determination of truth by inference by means
of removing an obstacle caused by doubt.
It can thus be regarded
rather as a function of inference. It cannot however be considered as an

i
anubhyate
hy an tara samsaynm nirnayam ca tarkah.

'NK, p. 103.
iv]
73
organ of knowledge by itself, inasmuch as the prbans employed in

reasoning is hypothetically assumed and as such the minor premise is
a false assumption1.
This is only a succinct representation of the position of the older
Naiyyikas by Sridhara and we have elaborately dealt with this theory
before.
What is of importance in this connexion is that Srldhara does
not think this to be a sound view.
He asks what is the nature of
reasoning ? Is it knowledge of the impossibility of the opposite alter

native endorsed by the opponent ? Or, is it a presumption in favour
of one's own position3 ? On the first alternative reasoning must be
held to be an organ of knowledge.
If the knowledgethat the cycle
of births and rebirths and final emancipation from this course of events
would be impossible if the soul were a historical eventwere invalid,
the opposite alternative advocated by the opponent would not be ruled
out as an impossibility.
And consequently the proof of the immor-
tality of the soul by inference would not be possible3.
If, however, it
is held that the opponent's position is validly rebutted by reasoning,

and the denial of the course of rebirth and death and of the eventual
emancipation from this undesirable course of events, which is the inevitable consequence of the opponent's position, is an absurdity,, then
reasoning must be regarded as the instrument which establishes a valid
conclusion.
If such be the case, it must be subsumed under a recog-
nised organ of knowledge, be it perception or inference4.
In point of
fact, reasoning operates as a reductio ad absurdum and it succeeds in

proving the absurdity of the opponent's position by adducing a stronger reason in support of the contrary position.
It is thus an instru-
ment for proving the impossibility of the position advocated by the

opponent and as such has all the characteristics of inference.
1 NK, p. 103.
2 kirn parapaksbhvapradtis
vj svapaksasambhvan?
Ibid.
3 dye pakse pramnam eve 'dam. j^tur anityatve samsrpavargayor asambhava iti jnnam yady apramnam n 'stnd vipaksbhvasiddhih, apramajiena
kasyaeid arthasya siddher ayogd ity anumnasy 'pravrttir eva visayavivekbhvt.
Ibid.
4 atlia siddhyaty asmd vipaksbhvas tad pramnam idam pratyaksdisu kasmimseid antrbhavisyati.
Ibid.
10
tarkah?
kirn
74
INDUCTIVE-REASONING
[CHAP.
It has been contended that reasoning does not record a definitive

verdict, but only operates as a presumption in favour of a particular
position.
But the question may legitimately be posed, what makes
this presumption possible?
In other words, what is the cause of it1?
Is it the organ supposed to establish the conclusion in view? N o , that

is not possible.
If an organ has established the conclusion, then there
will be no necessity for presumption at all and so reasoning would be

entirely superfluous2.
As a matter of fact the argument employed to
establish the immortality of the soul would never succeed in establishing the conclusion, unless the position, advocated by the opponent,
namely, that the soul is subject to birth and death, is proved to be a
logical impossibility.
If, however, it is supposed that reasoning serves
to prove the absurdity of the opponent's position and inference thereafter succeeds in proving the immortality of the soul, then it must be
held that reasoning furnishes the logical ground, which makes the
inference of immortality, a logical possibility.
Reasoning shows that
the soul cannot be mortal and the assertion of the impossibility of

mortality is tantamount to the assertion of the immortality of the
soul.
It is thus nothing but inference3.
Even if it be regarded that
the knowledge of the immortality of the soul follows as a consequence

directly from che proBans employed by the arguer, viz.
'that the
soul is not subject to birth and death', still the conclusion will not
follow so long as the contention of the opponent that the soul is subject to birth is not demolished.
The demolition of the opponent's
position is however secured by reasoning acting as a rednctio ad absurdam.
Reasoning proves that the soul cannot be mortal, as it involves
the absurd consequence of the impossibility of emancipation from this

cycle of births and deaths.
And even it this proof of the absurdity
of the assertion of mortality be not regarded as tantamount to the

1 atha svapaksasambhvantmakah pratyayas tarkah, asyo 'tpattau kim
kranam?
:
-:,,
NK, p. 104.
2 na tvat svapaksasdhakam pfamnam, tasy' pravrtteh.
Ibid.
3 vipaksbhve pratite svapaksasambhvano' pajyate iti vipaksbhvapratitir asya kranam iti cct? tarhi vipaksbhvalingakam anumnam evai tat.
Ibid.
iv]
75
proof of its immortality, but rather as a logical ground of this proof,

yet the validity of reasoning as an independent organ cannot be impeached.
Reasoning proves the absurdity of the opponent's position
by means of a logical probans and by reason of the. necessary concomitance between the probans and the probandum.
but inference.
It is thus nothing
And if the conclusion recorded by reasoning e.g. the
impossibility of the mortality of the soul be regarded to operate as the

logical ground for inferring its immortality and thus it be regarded as
a helpful condition of inference,
the Naiyyika must admit that
whether as an independent instrument for establishing the logical

ground, or as an auxiliary condition, reasoning has got an intrinsic
validity of its own 1 .
If, on the contrary, reasoning be denied all
validity, it cannot be supposed to function

proof or of disproof.
as the condition either of
When the Naiyyika asserts that reasoning only
serves to eliminate doubt which Isolds inference in check, he must

admit that it can succeed in its mission only if it possesses a validity
of its own, since an invalid cognition cannot be of any avail in the
determination of a truth 2 .
We need not give a survey of the arguments recorded in the
Vyomavatt,
since they are substantially in agreement with the argu-
ments of rdhara.
Vyomasivacrya, the author of the
Vyomavatt,
clearly and emphatically maintains the view that the finding of reasoning is necessarily valid and the attempt to deny validity to reasoning
is suicidal 3 .
If reasoning were invalid it could not make any contri-
bution to the determination of truth. Reasoning has been held to eliminate doubt or diversion of enquiry after the opposite alternative. But it
can rebut doubt and preclude opposite enquiry only if it can succeed
in giving a valid assurance that the opposite alternative is logically
untenable. And this assurance could not be expected to emanate from
reasoning if it were an invalid assumption.
Vyomasivacrya accor-
dingly concludes that reasoning muse be an organ of knowledge, and
1 NK, p. 104.
2 Ibid.
3 sa tu tasmims tad itiriipatvn
niscaya eva.
Wr, p. 533.
76
INDUCTIVE REASONING
[CHAP.
as such must be subsumed either under perception or inference1.

Sridhara, we have found, maintains that reasoning is nothing but a
case of inference, pure and simple. Barring this difference between
the two commentators, both of them are agreed that reasoning is an
organ of knowledge and its finding s valid. Udayana, as the author
of the Kiranvati which is another commentary on Prasastapada's
work, has sought to include reasoning under erroneous knowledge.
We have already criticised this venture of Udayana and so we abstain
from an elaborate treatment of this novel position3 in the present
context.
Wr, P- 533-
z Vide chapter iv,vp. 87.
CHAPTER
THE STATUS OF REASONING IN MADHVA'S SCHOOL AS

ELUCIDATED IN THE PRAMNAPADDHAT1
BY JAYATRTHA
We indicated at the outset of chapter iv that reasoning has been
given the status of an independent cognitive instrument in
Madhva school of thought.
the
The followers of Madhva, the redoubt-
able advocate of absolute dualism, have shown a remarkable tendency

for original thinking not only in the field of metaphysics and exegesis
on ancient texts, but also in the realm of epistemology and logic.
The Mdhvas are uncompromising realists and as such are expected
to be in agreement with the representative thinkers of the other
realistic schools of India, viz. Nyaya, Vaisesika, Mimamsa and Jaina.
But though points of agreement are considerable between the Mdhvas
and the other realistic philosophers so far as the ultimate conclusions
are taken into account, the approach and the mode of argumentations
adopted by the writers of the school of Madhva do indeed present
a striking contrast to those of their predecessors in the cognate schools
of thought.
The Mdhvas are not the persons to say ditto, to the
findings of their predecessors, whether opponents or friends.
It is
no doubt true that they have utilised and benefited by the speculations of the previous thinkers, but there is found in a remarkable
measure a freshness of approach and originality in the consideration
of every important problem.
It is almost certain that a student of
Indian thought can expect to learn something new in the writings

of the philosophers of this school.
We have shown in our exposition
of the speculation of the Nyaya-Vaisesika school that the representative

exponents of this school of philosophy have given a half-hearted recognition to reasoning as a logical organ, although it cannot be denied
that they have bestowed most critical and minute thinking on the
problem and have handsomely recognised the necessity and utility of
reasoning as a % helpful condition for the realisation of truth and
certitude in belief.
Reasoning has been made rather a parasite and
the tradition in the Nyya school s absolutely uniform that it cannot
78
INDUCTIVE REASONING
[CHAP.
be given a status of an independent instrument of knowledge,
though
its services cannot be dispensed with without jeopardy

Jayatirtha is an outstanding writer of the Mdhva school.
In
learning, scholarship and dialectics, he may be favourably compared

with some of the greatest intellectual stalwarts that every school
of Indian
greatest
thought has produced.
He is assuredly one
protagonists of Madhva's school and
of
the
it can be asserted
without any mental reservation and without the slightest exaggeration that the prestige
in a remarkable
enjoyed by this
degree
school of philosophy is
due to the contributions of this
man.
Jayatirtha stands in a class apart whether he writes a commentary

or an independent treatise.
He is a bold thinker* and a man of
unshakable conviction and has the courage to set forth his views in
a challenging manner bolstered up by convincing logic.
Jayatirtha
gave the lead, which was followed up by Vysatrtha and others, in

the enterprise of re-examination and re-valuation of the contributions
of previous thinkers.
In the Nyymrta,
Vysatlrtha delivered a
terrible onslaught on the monistic philosophy of akara's Vednta,

And but for the masterful defence of monistic philosophy by Madhusdana Sarasvat in the Advaitasiddbi
the chances are that the Sakarite
Vednta would have suffered an irrecoverable set-back, if not total

eclipse.
The attacks of these writers on the speculations of the
Nyya-Vaisesika school even in their latest orientation given by the

logicians of Mithil and NavadvTpa are of a formidable character.
It is a singular phenomenon that the criticisms of Nyya speculation
did not evoke a counter-reply from the Naiyyikas.
The reason for
this silence on the part of the Naiyyikas is perhaps to be found in

the comparative lack of publicity of these writings, particularly in
Mithila and Navadvlpa.
Thanks to the boon of the modern press
a student of Indian philosophy finds himself in a position of advantage

in that he has the inestimable privilege of access to these works of
Madhva's school, which were formerly jealously guarded from the
irreverent gaze of intruders.
The jealousies of medieval philosophers
have died out for good and for evil.
For good, because a modern
student starts without a preliminary bias and is not subject to any

extra-academic influence and consideration in his evaluation of these
vj
STATUS OF REASONING IN MADHVA'S SCHOOL
79
treasures of Indian thought, which impose upon him an equal claim

for study and examination.
For evil, because a modern student
cannot be expected to approach these classics with the ardour and

the enthusiasm of an orthodox zealot.
nation is irrevocably gone by.
The old age of epical fasci-
But a student of philosophy who
possesses a legitimate passion for philosophical speculation can be

expected to do justice to the philosophers of the past age, though he
may somewhat lack in the intensity of burning faith, which characterised his predecessors who could swear by their convictions.
Jayatirtha has written a standard work on logic and epistemology
from the. standpoint of Madhva's school.
He has given due consi-
deration to reasoning and the results of his study of it are entirely

at variance with the conclusions reached in the whole literature of
the Nyya and partly of the Vaisesika schools, which had been expounded by us in the previous chapters.
Jayatirtha regards reasoning
as a species of inference1 and as such an independent organ of knowledge, whereas the Naiyyikas of the old school have considered it to
be only a helping condition and an auxiliary to an accredited organ of
knowledge.
We have seen that Gagesa Upadhyaya, the founder of
the modern school of logic, has regarded it as a helpful condition for

the intuition of universal concomitance, and as such he makes it subordinate and subservient to perceptual cognition 2 .
Jayatirtha challenges
all these writers on logic and unhesitatingly declares reasoning to be a

valid cognitive instrument in its own right.
In the chapter on inference of the Pramanapaddbati,
propounds several ways of classification of inference
Jayatirtha
In one of these
classifications: he places inference under two broad heads, namely, (i)

inference for establishing-a conclusion (sadbantwmana) and (2) inference for refutation of the opponent's position (dsannumana).
He
divides the latter variety into two sub-kinds, namely, (i) for proof of a
defect in the argument employed by the opponent (dustipramitisdhana)
and (2) reasoning
{tarkaf.
The first consists in proving the incom-
1 tarko' nmnam eva. PP, p. 14.

2 Cf. sn-pra, c. iii.
*
3 sdhannumnarp dsannumnam ce' ti. dsandnumnam api dvedh.
dustipramitisdhanaqi tarkas ce' th
PP, p. 36.
So
INDUCTIVE REASONING
[CHAP.
petency of the probans employed for establishing the probandum. In

other words, it consists mainly in the exposure of a fallacy in the opponent's argument. The second variety is mainly concerned with the
rebuttal of the objection, advanced by the opponent, by a reductio ad
ahsurdum. Jayatlrtha accordingly considers reasoning to be a variety
o inference for refutation of a position (dsananumanavi'sesa). He
has defined reasoning as the admission of an undesirable contingency
necessitated by the admission of a false issue1. The commentator
Janrdana Bhatta explicates it as the enforced admission of a contingency as the condition and presupposition by the admission of a fact
necessarily conditioned by the former by reason of the necessary concomitance between them. When the occurrence of A is necessarily
determined by the occurrence of B, that is to say, when A cannot
possibly occur if B is not present, A is said to be a necessary concomitant (vypya) and B as the determinant of the concomitance (vypaka).
The determinant of concomitance (yyapaka) is inferred on the ground
of the determinate in concomitance (vypya). We will call the determinant in concomitance for the sake of brevity determinant concomitant {yyapaka) and the determinate in concomitance (yypya) simply
determinate concomitant. The meaning of Janardana's comment can
thus be expressed: the admission of a determinate concomitant makes
the admission of a determinant concomitant inevitable2. Janrdana
Bhatta regards such reasoning to be a case of inference and he seeks to
prove his contention by means of inference, namely, "Reasoning is a
species of inference, because it generates mediate cognition by virtue
of necessary concomitance, like the well-known case of the inference
of fire from smoke by reason of necessary concomitance of the latter
with the former"3. Just as the probans is the instrument of proof of
the probandum in virtue of necessary concomitance between them, so
also the probans .hypothetically assumed, that is the ground according
to the terminology we have adopted, is the instrument of the proof of
i kasyacid dharmasy 'gikre 'rthantarasy pdanam tarkah.
pp. pp- 36-37i PPT, pp. 36-373 tatko' numnam vyptibalena paroksajanajanakacvc.
Ibid., p. 38.
y]
STATUS OF REASONING N MADHVA'S SCHOOL
81
the consequent by virtue of necessary concomitance. It is argued that

when the ground of reasoning works on the basis of necessary concomitance, just as the probans does in the accredited cases o inference, the
former (reasoning) should also be regarded as a case of inference and
nothing else.
It is necessary concomitance that is the basis and foun-
dation of inference and reasoning has also this basis alone.
So there
is no reason to demarcate reasoning from inference and to give the

latter a status superior to the former.
It may be contended against this that in inference the major premise and the minor premise are the necessary condition.
Or to put it
the other way about, in accredited cases of inference the probans is

shown to stand in necessary concomitance with the probandum in the
major premise (yypti) and is again shown to be actually present in
the subject in the minor premise (paksadbarmat).
These two are the
necessary conditions of inference and the lack of one or the other

makes inference impossible.
In
reasoning,
how/ever, there is the
necessary concomitance between the ground and the consequent no

doubt, but the ground is never actually present in the subject.
Thus,
there is the major premise, but the minor premise is conspicuous by

its absence.
The ground is only hypothetically superimposed upon
the subject and the arguer is all the while conscious of the unreality
of the relation of the ground with the subject.
will make the position clear.
A concrete illustration
"If the lake were possessed of smoke,
it must be possessed of fire", is a case of reasoning.
The arguer is
definitely aware that smoke is not and cannot be actually present in

the lake and so the minor premise is not possible and thus the conclusion
cannot follow.
The lack of the minor premise, that is to say, the
lack of the actual presence of the probans or the ground in the subject,
is called the fallacy of the non-existent probans.
The awareness of
the fallacy precludes the knowledge of the conclusion, that is, of the
presence of the probandum in the subject.
a case of inference.
So reasoning cannot be
Were it to be a case of inference, the knowledge
of the ground would have to be regarded as the condition and the

knowledge of the consequent would have to be regarded as the
resultant.
But there is no such thing as knowledge of the ground
or the knowledge of the consequent, but only an assumption in either

11
02
INDUCTIVE REASONING
case.
[CHAI.
Assumption can never be a case of knowledge and as reasoning
is nothing but assumption in respect of both the minor premise and

the conclusion, it ought not to be confounded with inference proper.
Inference is knowledge definite, authentic and categorical,
whereas
assumption is only a hypothetical supposition, and an attempt to

reach the status of knowledge which is never successful.
It will be
nothing but a flagrant violation of the logical fabric of human

knowledge to attempt a confusion of the two, which is necessarily
involved in placing them under one category1.
Jayatlrtha observes that the objection of the opponent proceeds
from lack of correct appraisal of the logical value of the fallacy called
the 'non-existent probans1.
It is no doubt the convention in logic
that the probans should be accepted by both the parties in a debate,

the proponent (ydin)
and the opponent (frativadin),
property of the subject.
to be a true
Though the actual presence of the probans
in the subject is the general rule, it ought not therefore to be regarded

as the essential condition of inference.
What is essential is the belief
in the presence of the probans in the subject, and this belief is the
condition of the belief in the presence of the probandum, whether
spontaneously admitted or necessarily enforced.
If the opponent
believes that the probans in question is present in the subject and

this probans actually stands in the relation of necessary concomitance
to the probandum, then the admission of the presence of the probandum in the subject on the part of the opponent is a matter of unavoidable logical necessity.
The probans in question may be actually
absent in the subject, but so far as the opponent is concerned, he

cannot make a grievance of it against the proponent and deny that
the conclusion does not follow for the lack of the minor premise.
The proponent is not to blame because he only accepts the opponent's
version and shows that the latter's belief in the presence of the
probans in the subject makes the belief in the presence of the
probandum a matter of inescapable necessity.
No conclusion can
follow from a probans definitely known to be absent.
The presence
i PP, pp. 38-39.

Vide also PPT, pp. 38-39: nanu tarkasy'numnatve
apdakam nirvahnitvam nirdhmatvam prati ligam iti vcyatn. tiirvahnitvasya
vahnimati parvate pakse'siddhatvena katham tarkasy'numnatvam.
y]
83
of the probans makes the presence of the probandum inevitable, and

if the opponent believes the former he cannot consistently refuse to
believe the latter. The presence of the probans is a matter of belief
with the opponent, and though this belief is mistaken and the proponent is definitely aware of this mistake, the opponent cannot
make a complaint that the argument is vitiated by the fallacy of
the non-existent probans, because the opponent is persuaded that the
probans is actually existent in the subject1.
Reasoning is a case of hypothetical inference and acts as a reductio
ad absurdum. The probans of such an inference is necessarily
assumed on the assertion of the opponent, and so is not accepted by
both the parties, as is the case with ordinary inference. Though
not accepted by both the parties it is accepted by one of them, .
namely, the opponent, for whose edification and correction reasoning
is applied as a reductio ad absardum.
The probans in such cases
of inference is certainly not believed to be present in the subject by
the arguer himself. But his want of belief does not affect the valid- "
ity of the inference so far as the opponent is concerned, because
the latter believes in the presence of the probans. And the presence
of the probandum as the necessary determinant concomitant and *
condition is only brought home by the arguer as a matter of unimpeachable logical necessity. So the lack of admission of the '
presence of the probans by both the parties cannot be made the
ground of repudiating the cogency of reasoning. The admission of
the presence of the probans in the subject by the arguer and the
opponent alike would on the contrary take away the raison d'etre of
the employment of reasoning as a dialectical instrument. Reasoning
is a case of reductio ad absurdum and its purpose is to reduce the
opponent to a position of absurdity by proving the untenability of the
argument sponsored by him and by contrary implication ending in
the vindication of the truth of the opposite conclusion.
This
characteristic function of reasoning, namely, its culmination in the
proof of the opposite of the conclusion advocated by the opponent
1
anumnatve'pi tarkasy padaksiddhir adsanam. parbhyupagamamtrasya tatra siddhipadarthatvc.
PP, pp. 38-39. Vide also PPT, p. 39.
84
INDUCTIVE REASONING
cannot be
[CHAP.
accomplished unless the presence of the probans in the
subject be shown to be wrongly assumed by the opponent1.
Let
us make it clear by a concrete example.

"IE the lake were possessed of smoke, it would be possessed of fire.
But as a matter of fact it is not possessed of fire.
It cannot be possessed of smoke/'
If smoke were actually present in the lake and also believed to be
so by the arguer he could not argue into the conclusion that the
lake is not possessed of smoke.
The culmination of reasoning In
demonstration of the truth of the opposite conclusion can be secured

only by proving the absence of the probans and the probandum both
in the subject.
The statement, "If the lake were possessed of smoke,
it must be possessed of fire", shows the concomitance of the ground

and the consequent in agreement.
And j:he negative conclusion,
"The lake is not possessed of fire and so it cannot be possessed qf

smoke" shows concomitance in difference between them.
It is the
general custom for logicians to stress in reasoning the negative conclusion alone, e.g.
"The lake is not possessed of fire". This does
not mean that the arguer is not interested in proving the absence of
smoke also in the lake.
absence of smoke.
The absence of fire necessarily entails the
And so the express statement of the absence of
smoke in the lake is deemed superfluous.
But whether expressly
stated or not, the absence of smoke is a necessary concomitant of

the absence of
fire.
be
to
compelled
If under the circumstance the arguer were to

believe in
the
presence
of
smoke
in the
lake as a preliminary condition as the opponent does, he would

have no difference with the opponent.
stoii could not be brought home.
conclusion presupposes
that the
And so the negative conciu-^

The possibility of the negative
belief in the
probans or the ground in the subject
existence of the
must be a pretence and
make-believe and not spontaneous on the part of the. arguer.
So
i vastutah pakse vidyamnatve pdakatva-vyghtah. katham anyath na

c'yam nirdhmah tasmn na niragnika id viparyaye paryavasnam. tadabhve ca
tarka cv' bhsah syt. PPT, p. 39.
y]
85
the belief in the non-existence of the probans in the lake on the

part of the arguer does not make inference impossible 1 .
To sum up.
The minor premise is- undoubtedly a necessary
condition of inference. And it shows that the probans is in connexion

with the subject.
This connexion is actual in inference and is also
accepted to be such by both the parties in a debate.
In reasoning
the ground does not actually exist in the subject, but assumed to be
present by the arguer on the admission of the opponent.
There is
however no logical difference between the fact that the probans is

actually present in the subject and the fact that it is admitted to be
present.
And we have shown that the conclusion follows to the satis-
faction of both the parties when the probans is admitted to be present

by both of them, and to the satisfaction of the opponent when it is
admitted by him alone.
So the actual absence of the probans
in the subject does not/preclude the deduction of the conclusion and

as such reasoning cannot be denied the status of inference on this
ground.
Another objection against reasoning being given the status of inference is that the conclusion established by reasoning is necessarily false,
since the probandum does not actually exist in the subject.
The ab
sence of the probandum in the subject is called the fallacy of the contradicted probans (badbdf.
But the Mdhvas do not regard it as a
fallacy or a defect in reasoning.
Undoubtedly such cases are truly ob-
jectionable in categorical inference.
But the absence of the consequent
in the subject is not at all a defect in reasoning, which is regarded as

inference of the type called reductio ad absurdum*.
The arguer does
not believe in the existence of fire in a lake, nor is it an accepted conclusion.
He is aware of the absence of fire in the lake as the esta-
blished conclusion and the admission of the presence of fire on the

1 PPT, p. 39.
2 nanu tath'pi na tarkasy'numnatvain sagacchate. tarkasy'numnatve
hy pdyam nirdhmatvam Hngi'ti vcyam. tasya dhmavaty abhvd badhah.
Ibid.
3 sdhyarupasy aiva iigino bdho dosah. na tv pdyarpasya. pdyatvavyghtat. yad pdyam tat pramnabdhitam eva. anistam hy pdyam, prctmnabdha ev' nistatvatn, na c'tra nirdhmatvam sclhyam..
Ibid.
86
INDUCTIVE REASONING
[CHAP.
contrary would constitute the fault called acceptance of a position

opposed to his own accepted conclusion (apasiddhnta).
This is cer-
tainly a fault in categorical inference, which seeks to establish an

independent conclusion (sadhaknnmana).
But in a reductio ad absur-
dum the enforced emergence of the conclusion, showing the necessity

of the admission of the absence of smoke, does not involve any such
fallacy or defect1.
What the arguer aims at proving is this that the
admission of the false minor premise makes the admission of the false
conclusion unavoidable, and thus by negative implication he proves
that the truth lies in the contradictory of the conclusion. What are the
defects of categorical inference, that is, an inference which seeks to
establish a conclusion by means of a probans and a probandum actually
present in the subject, are not the defects of a hypothetical inference,
which is called a reductio ad absurdum.
On the contrary the actual
absence of the probans and the probandum in the subject, which resr
pectively involves the fallacy of non-existent probans and contradicted
probans in categorical inference, which aims at a true conclusion, are
really the conditions of the inference known as reductio ad absurdum2.
If however we look far and close, we shall find that both the types of
inference, which we have called categorical and hypothetical, conventional inference and reductio ad absurdum, will be found to culminate
in the finding of a truth.
In categorical inference neither the major
premise alone nor the minor premise by itself can establish a conclusion
but the combination of the two does it. So also the reductio ad absurdum apparently records a false conclusion when taken by itself.
taken by itself it is only an incomplete argument.
But
It is only when
it is taken in combination with the negative conclusion, which is the

logical culmination, that a reductio ad absurdum can function as a
completed logical organ..
Understood in this light the reductio ad
absurdum can be free from all the objections that have been advanced
against it by the Naiyyikas who took it apart from its negative im-
1 ata eva n'pasiddhnto'pi. sa hi siddhntaviruddhasya sadhane syt, na

tv pdane.
PPT, p. 39.
2 tasmt sdhannumna eva bdhpasiddhntayor dsanatvam na prasaga- .
numnarpe tarke, pratyuta tayor bhsanatvam cva. bid. Vide also PP, p. 39.
v]
plication.
87
Jayatirtha quite legitimately emphasises the necessity of
taking a full and comprehensive view of reasoning as a logical organ.

Reasoning consists of reductio ad absurdum
as only one half and the
negative implication as the other half; the two together make a completed whole, and it is only as such a whole that it functions as a
logical organ1.
The hostile criticism of the Naiyyikas has been in-
spired by their acceptance of reasoning in the first half and their failure to take account of the other.
That the two together end in the
proof of a true proposition does not admit of doubt.
Jayatirtha observ-
es that there is not the slightest justification for the repudiation of reasoning as an organ of knowledge.
It is not only an accredited organ
of knowledge, but falls under inference.
The essential character of
inference lies in the necessary concomitance of the probans and the

probandum and it is this fact alone which makes the deduction of a
conclusion a possibility.
It is admitted on all hands that reasoning also
draws its sole probative force from the actual presence of necessary concomitance between the ground and the consequent.
This should be
accepted as legitimate ground for placing reasoning under the head of

inference.
Jayatirtha accordingly concludes that reasoning is an acc-
redited organ of knowledge and is subsumed under inference.

It is remarkable that European logicians also regard reductio ad
absurdum as a species of inference and this may be regarded as lending
further countenance to the position of Jayatirtha.
difference between them.
But there is a vital
Whereas European logicians are satisfied
with formal consistency alone and as such even a false major premise
may be regarded as warranting a logically consistent conclusion, Jayatirtha in common with the whole class of orthodox Indian logicians
will not admit the validity of such inference.
All men are immortal.
''-"'
Socrates is a man.
.v
Socrates is immortal.
i yath sdhannumane na vyptimtram, n' pi paksadhatmatamtram

sdhyapramisdhanam kintu militam cva. evam anistapdaiiam viparyaye
paryavasnain co' bhayam militatn eva 'numitisadhanam bhavati' t tarkasya
prmnyam upapadyate.
PP, p. 40.
88
INDUCTIVE REASONING
is a perfectly valid syllogism in formal logic.

not give any countenance to it.
[CHAP.
But Jayattrtha would
The European logicians, therefore,
have no special difficulty with the reductio ad absurdurn being regarded

as a case of inference.
But Indian logicians insist upon the material
truth of the premises as well as the conclusion as the fundamental

characteristics of inference.
Inference with Indian logicians and so
with Jayattrtha is not an exercise in formal consistency, but an

accredited organ of knowledge.
the cause of a real difficulty.
And so the false minor premise is

Jayatlrtha points out in defence that
though the minor premise is false, the major premise is true and the
negative conclusion which is the ultimate outcome of reasoning is a
true finding. So Jayatrtha overcomes the difficulty caused by the
lack of truth of the minor premise by his arguments, which we have
produced beforehand he lays stress quite legitimately upon the material
truth of the major premise and the negative conclusion, and on this
basis declares reasoning to be an accredited organ of knowledge.
Reasoning is regarded by the Naiyyikas as only an auxiliary to
an organ of knowledge and as what only helps the activity of a cognitive organ by removing an impediment in the way.
This removal of
impediment is asserted by Vacaspati Misra to consist in the elimination of doubt and by Udayana to consist in the removal of enquiry
after the opposite possibility1.
Jayatlrtha partially agrees with Vacas-
pati Misra when he admits that reasoning on some occasions serves to

eliminate doubt of the opposite alternative and in this way helps the
realization of truth by means of an accredited organ3.
But he differs
from the Naiyayikas by regarding reasoning as an independent organ

irrespective of its bearing upon another organ.
Reasoning no doubt
acts as a helpful condition by serving to eliminate doubt, which creates

a deadlock; but this is not the only role played by reasoning.
plays an independent part as an instrument of inference.
It also
Thus, for
instance, when a person is deterred from inferring fire in a hill on the

ground of smoke by a doubt of the necessity of fire's occurrence,
1 Cf. sufra, c. i.
.
2 kvacid viparitasaknirasanadvrena pramannm anugrhako'pi bhavat
PP, p. 40.
y]
89
reasoning comes to his aid in the removal of the doubt. Here reasoning aids the realization of inference. JayatTrtha is thus in agreement
with the Naiyyikas in so far as reasoning is regarded as a helping
factor. But as has been pointed out by Janrdana Bhatta, reasoning
can also serve to establish a conclusion independently of a categorical
inference. For instance, suppose that the subject of debate is propounded by the umpire on the basis of the divergence of views in the
form "The hill is either possessed of fire or not". That the hill is
possessed of fire can be established by a reductio ad absnrdum independently of the aid of a categorical inference is shown as follows. "If
the hill were devoid of fire, it would be devoid of smoke. But it is
not devoid of smoke and so it cannot be devoid of fire". Reasoning
as a hypothetical argument in the present case establishes the conclusion that the hill cannot be destitute of fire and so by the negation of
the absence of fire establishes by/implication that the hill is possessed
of fire1. It cannot be objected that the argument in question is only
negative in character being based upon the concomitance in difference,
which is only an auxiliary of the concomitance in agreement and as
such has no independent probative value of its own. But there is not
the slightest reason for denying the cogency of the negative concomitance. On the contrary the observed concomitance in agreement
cannot be established on a footing of universal and necessary truth
unless the opposite alternative is ruled out as an impossibility, which
is effected by negative concomitance. The cogency and independent
validity of reasoning as a type of inference based on negative concomitance cannot, therefore, be repudiated without weakening the foundation of inference, that is to say, without leaving the knowledge of
universal concomitance open to doubt2.
Jayatlrtha definitely maintains that reasoning is an independent
organ and is subsumed under inference. Reasoning is a case of hypothetical inference and the latter is as much possessed of cogency as
1 adrir agnimn na ve'ti vipratipattau sdhannumnam vinaiVa yadi

niragnikah syt tarhi nirdhmah syt na c'yam nirdhma itd tarkarpanumancn
aivl 'gnisiddheh.
PPT, p. 40.
2
12
Ibid.
INDUCTIVE REASONING
[CHAP.
categorical inference is accepted to be. The Naiyyikas have consistently ignored the claims of hypothetical inference and think that it is
only of service as a pendant to categorical inference. We have shown
how reasoning as a hypothetical argument serves to establish a conclusion, primarily a negative and secondarily a positive one. And even
when it serves to confirm and reinforce the finding of a categorical
inference, it does not forfeit its status as an independent organ. That
one cognitive organ may serve to confirm the finding of another
cognitive organ without forfeiting its status as an independent cognitive organ is admitted by the Naiyyikas when they declare the
possibility of the convergence of several cognitive organs on a self-same
object of knowledge (pramnasamplava)1. To take a concrete example.
A man is told by a reliable person that there is a lake in the neighbourhood. On moving forward he observes the flight of aquatic birds
and thereby infers the presence of water. ' On moving further on he
finally reaches the lake and observes the water there. In this case the
knowledge of water is first derived from verbal testimony; in the second
place, the same is confirmed by inference and in the third place, it is
secured by perceptual observation. All the three species of knowledge, verbal, inferential, and perceptual have for their object the water
of the lake and nothing else. The first knowledge is confirmed by
the second and the second by the third. Though the second cognition thus reinforces and confirms the first, and the second in its turn
is corroborated by the third, nobody would commit the blunder of
repudiating the independent status of the second and third cognitions
as independent cognitive instruments. The case of reasoning is exactly
on a par. Though it confirms and reinforces the finding of another
cognitive instrument, it does not thereby assume the status of a subordinate appendix. Like the second and third cognitions of the aforesaid example reasoning never forfeits its independent character. The
findings of Jayatirtha and his followers are in complete agreement with
those of the Jaina logicians, with this exception that whereas the latter
i tasmd yath pramanasamplave dvitiyam pramnabhtam eva, prathamadardhyahetutvena-pram^nugrhakam tath" tarko'pi pramanam eva pramnnu*
grhakah.
PPT, p? 40.
y]
91
regard it as an extra kind o organ, the former have subsumed it under

inference. It is remarkable that the Buddhist logicians call reasoning
a species of inference (prasagnumnd) and the agreement of the
Mdhvas is more pronounced with the Buddhists than the Jainas. It
is a pity that the Buddhist's speculations on reasoning as frasaganumana are not available in a systematic form. The contributions of
the Mdhva logicians may however be regarded as a rational elaboration of the Buddhist position.
CHAPTER VI
REASONING AS ELUCIDATED IN THE NYYASUDHd
BY JAYATTRTHA
Jayatrtha has written a learned commentary on Madhvcrya's
bhasya on the Brahmastra. The commentary is called the Nyaya*
sudha and is an authoritative standard work of the Mdhva school.
Students of Indian philosophy are undoubtedly aware of the fact
that commentaries are not merely parasitical works, but independent
contributions necessitated by the criticisms of the rival philosophers.
It has been the fashion in Indian academies that scholars preferred
the style of commentaries to pretentious original works, although
the commentary is to all intents and purposes an independent work,
setting forth views not necessarily sponsored by the original authors,
and being chiefly preoccupied with the refutal of the criticisms that
have been levelled against the original works by hostile critics. The
Nyayasudh is an extremely elaborate work and whenever any
problem is broached in it, it has received a thorough-going treatment
and elaborate consideration in which the views of the rival thinkers
have been expounded with meticulous care and the criticism has been
as elaborate as it has been relentless. In his comment on the Brahmastra III. ii. i, Jayatrtha has dwelt at length on reasoning and has
given a survey of the views of the writers of the Nyya school
beginning with the author of the Nyayasutra and coming down to
Udayanacarya1. We have already dealt with the views of these
writers and so we do not propose to give an account of the treatment
that has been accorded to the speculations of these % writers by
Jayatirtha. The special point of interest lies in the position adopted
by Jayatrtha with regard to the logical status of reasoning, which
he takes considerable pains to prove to be a variety of inference.
>
y Nsu, pp. 477-78.
yi]
REASONING AS ELUCIDATED IN THE NYYASUDH
93
The speculations recorded in the Pramnapaddhati by Jayatirtha are

rather a succinct summary of the findings recorded by him in the
Nyayasudh. We do not therefore hesitate to go to the main source
in the Nyyasudh even at the risk of some amount of reduplication
in order to give a fuller account of Jayatlrtha's disquisition on reasoning,
since we are conscious that our representation as confined to the
Pramanapaddbati is bound to give the appearance of a scrappy and
fragmentary treatment, unless supplemented by the findings of
the Nyayasudha.
Jayatirtha observes that a debater whether in the role of a proponent or an opponent is mainly concerned with a two-fold task,
namely, the establishment of his own tnesis, and the refutation
of the thesis of the other party1. In order/to accomplish the first
objective, that is to establish his own position, it s necessary that
all the factors of syllogistic reasoning, namely, the subject, the
probans, necessary concomitance and the example cited by him
should be acceptable to both the parties3. Thus, for example, in
the syllogism "The hill is possessed of fire, because it is possessed
of smokev All that is possessed of smoke is possessed of fire, as for
example, the kitchen", the hill as the subject, smoke as the probans,
the universal proposition, and the kitchen as the example, are accepted
and approved by both the parties. This is an instance of the syllogism
employed for the establishment of one's own thesis. It is essential
to bear in mind that whenever a syllogistic inference is employed
for the proof of one's own position, all the factors involved in it
must receive the certificate of approval from the arguer himself and
the opponent as well. It is not possible that such syllogisms can
pass muster if any of the factors concerned be a matter of doubt.
It is possible to contend that insistence on the unanimous acceptance
of all the factors of inference is rather a counsel of perfection, and
if this be regarded as the minimum condition, the scope of inference
1 dvivdham hi kryam kathakasya svapaksasadhanam parapaksadsanam
ce'ti.
Nsu, p. 478.
2 tatra sadhainam sviaparasammatair eva nyyaih kryam nS'nyataramtra. siddhaih.
Ibid.-
94
INDUCTIVE REASONING
[CHAP.
will be necessarily narrowed down to a field of trite cases and it will

have very little value as a logical organ in metaphysical speculations,
where the differences between the rival schools are much too fundamental. Let us elucidate the position by a concrete example. The
Naiyyika is definitely of the opinion that words are perishable,
whereas the Mimamsists and the Grammarians are equally strongly
persuaded that words are eternal facts. The Naiyyika would
seek to prove his position by the following syllogism: "Words
are perishable, because they are products. All that are products
are perishable, for instance, a jar" 1 .
The Mimamsists would
not accept the probans to be true.
According to them words
are eternal entities, which are revealed by the sounds produced. The Naiyayikas on the other hand are definitely convinced that there is no difference between word and sound.
So according to them a word is as much a product as sound is.
However firm may be the conviction of the philosophers concerned,
neither can hope to convince the other party of the truth of his position,
since the opponent will always disagree about the truth of the probans
employed. The Mmamsist will naturally refuse to believe that word
is a product and hence the use of the syllogism in question by the
Naiyyika will be entirely abortive. The Mmamsist will plead that
the probans is not acceptable to him and hence the condition of inference is lacking. What is then the remedy in such a situation ? It
seems that the situation is hopeless if unanimity on the truth, of the
subject, the probans, the universal proposition and the example be
made the necessary condition of the validity of the syllogistic argument.
But the situation is not so desperate as it seems at first sight. If the
Mmamsist takes exception to the validity of the probans, the remedy
lies in the proof of the probans to the satisfaction of an impartial judge,
if the opponent pretends to be unconvinced. In the cas under consideration the Naiyyika will be called upon to prove that word is a
product, and if he succeeds in establishing the proposition that word
is a product, the Mmamsist will be obliged to accept the consequence
involved by the truth of this proposition. The Naiyyika may argue
i yath sabdo'nityah kryatvt
Nsu, p. 478.
yi]
95
that there is no difference in kind between the effort required for the
production of a jar and that required for the production of a sound1.
There is no evidence that word is a pre-existent fact manifested by an
effort like the water in a well. Besides, there is no convincing argument for belief in a metaphysical entity, which is not cognisable by a
sense-organ, but which is to be regarded as the real word-essence.
The Naiyyika would contend that word and sound are identical, since
our knowledge of word is always secured by means of the auditory
organ. And so sound and word being proved to be identical, there
is no justification on the part of the Mimamsist for refusing to regard
it as a product, which he admits sound to be.
The main point at issue is this. If one has to prove one's position
in the teeth of the opposition of a rival, one can succeed in the task
if the probans employed is accepted to be true by both. If however
there is a difference about the truth of a particular probans, the centre
of debate is shifted to the probans itself, and the successful culmination
of the initial argument depends upon such proof.
The inference for refutation can assume two forms:
(1) this is
not possible; (2) the admission of it leads to an undesirable consequence. So far as the first type is concerned, the apparatus of inference
must be acceptable to both the parties2. To take a concrete example.
Kumrila Bhatta holds that word is a substance and the Vaisesika
maintains that it is a quality. The Vaisesika argues that the position of
Kumrila is untenable and in support advances the following syllogism :
"Word is not a substance, since it is perceptible by the organ of hearing. All that is so perceptible is not a substance, just as the universal
is" 3 . In this form of inference the probans is acceptable to both the
parties. It is in the same position with the inference for establishing
one's own position, so far as the unanimity on the apparatus of inference is the ruling condition. In case there is a difference of opinion
1 karyai ca 'sv avyanjakaprayamnantaropalabdher ity adi.
Nsu, p. 478.
2 dsanam tu dvividham. idam ittham na bhavati v anistopadar&mena
v. tatra'dyam ubhayasiddha-nyayair evai kryam.
Ibid.
3 iabdo na dravyam srotragrhyatvat smanyavad iti.
Ibid.
9^
INDUCTIVE REASONING
[CHA>.
on the validity of the probans or the universal proposition the inference is brought to a standstill, and it can be set in motion only by
meeting the objection of the opponent. In the syllogism cited here
the Mmamsist may contend that the probans is inconclusive, since
there is no bar to a substance being audible. Time is perceptible by
all the sense-organs and so is perceptible by the organ of hearing, but
time is a substance. The Vaisesika can prove his contention only by
proving that the probans is conclusive. In actual fact, the Vaisesika
has argued that time is a supersensible entity and its existence is capable of being proved only by inference. The Vaisesika's argument
will be effective and valid only if he can succeed in refuting the
Mmmsist's contention and proving that the probans and the
universal proposition are to be accepted by the Mmamsist. The
truth is this. A categorical inference whether employed for establishing one's own position or for refutation of the opponent's thesis must
be based on an apparatus, which is accepted by both. In refutation,
the probans and the . universal proposition must be acceptable to the
opponent otherwise it will fail of its purpose 1 . In the second type of
inference, stress is laid on the agreement of the opponent in respect
of the probans and the universal proposition, and the arguer's belief or
disbelief is considered irrelevant. This type of inference is called
reasoning and assumes the form of - a hypothetical proposition 2 . The
example of it is the following argument. . "If the hill were devoid of
fire, it would be devoid of smoke"... The first clause states the ground
or the probans and the second the consequent or the probandum. The
ground is not believed to be proved by the arguer and is stated by
him hypothetically in order to disarm the opposition of the opponent,
who would dispute the presence of fire in spite of the presence of
smoke. But it is not only not necessary that the probans should be
accepted by the arguer to be true, but on the contrary he must be
aware of its falsity. If it were not so, the hypothetical inference would
i Nsu, p. 478.
t dvityam tu parasiddhair eva nySyaih (karyatn).-.
iti vyavabriyate.
Ibid.
.tad eva ca .tarka
vi]
REASONING AS ELUCIDATED IN THE NYYASDH
not operate as reductio ad absurdum.
97
What is necessary in this type
of inference is that the probans or the ground should be acceptable to

the opponent.
The universal concomitance between the ground and
the consequent need not also be objectively true.

accepted to be true by the opponent.
It will suffice if it is
The type of inference, in
which the major premise and the minor premise are
necessarily
accepted to be true by the opponent and only assumed by the arguer

for bringing out an absurd consequence, which necessarily follows from
the assumed ground, is called reasoning (tarka)1.
reductio ad absmdum
in European logic.
It is known as
The falsity of the subject
or the minor term, and of the probans, which is a real defect in categorical inference, is not only not a defect in the reductio ad absurdum,
but rather is the condition of it.
The absence of fire in the hill,
which is the ground in the hypothetical inference, is not established by

an accredited organ of knowledge, but rather assumed by the arguer,
tvho actually believes in the presence of fire, for bringing out the
absurd consequence.
It may not be out of place to make an observation on the conception of universal concomitance as advocated by the Naiyyikas on the
one hand, and the Jainas and the followers of Madhva on the other in
order to refute the charge of the lack of co-existence of the probans
and the probandum that has been advanced against the claim of
reasoning as an accredited type of inference.
It has been urged by the
Naiyyikas that the ground and the consequent (for instance, absence
of fire and the absence of smoke) are not co-present in the subject (the
hill), and so there can be no objective universal concomitance between
them.
Thus, the lack of universal concomitance makes inference
impossible.
Apart from the consideration that the lack of objective
concomitance between the ground and the consequent is no impediment in reductio ad absurdum,
which may also be a matter of
assumption for the arguer, it is worthy of remark that the Mdhvas

and the Jainas do not believe that the co-presence of the two terms is
a necessary factor of universal concomitance at all.
Nsu, p. 478.
According to
98
INDUCTIVE REASONING
[CHAP..
them the universal concomitance can be established where the one

term cannot be conceived to be possible without the presence of the
other 1 .
There is a necessary relation between them no doubt, but the
relation is not necessarily one of co-existence in the same substratum.

So there is no difficulty when the presence of one term in one place
and time involves the presence of the other term in another place and
time.
Spatial or temporal co-existence is not a necessary factor of
universal concomitance.
Thus, there is no difficulty in the case where
the occurrence of the flood in the lower region of the river is the
ground of the inference of a heavy downpour in the upper region.
There is neither temporal nor spatial co-existence between them, but
still there is no difficulty
in inference, because the concomitance
between them is necessary and universal 2 .
W e have seen that tempo-
ral or spatial co-existence, is not an element in universal
concomitance
and so the lack of it in the universal proposition in redactio ad absurdum cannot be made grievance against those who think reasoning to
be a type of inference nonetheless.
W e now propose to take into consideration a set of objections that
have been_advanced by the Naiyayikas against the claim of reasoning
to the status of inference.
The first objection is that the conditions
of accredited inference and reasoning are fundamentally different
and
this difference of conditions must entail a fundamental difference in

the character of the result.
Accredited inferences consist of categorical
assertions and the probans is admitted to be materially true by both

the parties, whereas reasoning consists of a hypothetical minor premise,
and the probans affirmed of the subject is materially untrue.
It is
preposterous that the two cases should be he|d to be similar in spite of

this material difference
in the conditions.
there is also a material difference
And as regards the result
to be observed*
In inference, the
result is . a true conclusion, while in reasoning, the conclusion is a

i yaddesakalasambaddhasya yasya yaddesaklasambaddhena yena vina anupapattih tasya tena s vyptih.
XT. Ms. pp. 3-4 (chapter on inference).
,2 ctena shacaryaniyamo v, anaupadhikah sainbandhq v, sdhytyantabhvavadavrttitvam v vyptir iti nirastam. vyadhikaranayoh pravrstyoh desatah
klato v shacaryasya sambandhasya v asambhavt nadiprasya vrstirpasdhya-!
bhvavadadhodesavrttitvc ce'ti.
.
Ibid.
yi]
false issue.
99
Jayatirtha in reply maintains that this difference in con-
ditions is not fundamental.
We shall show that the essential condi-
tion of inference is the same in categorical and hypothetical inference

known as reasoning.
The difference of minor detail does not involve
a difference in the fundamental character of the result.
The Naiyayi-
kas admit the validity of three types of inference, namely, exclusively

affirmative (kevalanvay), exclusively negative {kevalavyatireki) and affirmative-cum-negative (anvaya-vyatireki).
different in each type.
The conditions of inference are
For instance, in the
affirmative-cum-negative
inference, the probans is said to possess five characteristics, namely, (1)

presence in the subject (paksasattva);
{sapaksasattva)\
(2) presence in the homologue
(3) absence in the heterologue (yipakssattva); (4)
immunity from opposition by a countermanding probans (asatpratipaksitatva); (5) immunity from contradiction (abdhitatva).
In the first
type of inference, the heterologue is an impossibility and so the third

characteristic, namely, absence in the heterologue is out of the question.
Similarly in the second type of inference there cannot possibly be a
homologue and so the second characteristic is conspicuous by its
absence.
Thus, it is found that the character of the probans is
different in each type. In the first, the third characteristic is impossible

and in the second, the second characteristic is found to be wanting1.
It might be contended that the first and the second types are not
true cases of inference, since the conditions are found to be deficient
in one or the other characteristic of the probans2.
The Naiyayikas
may maintain that though these specific characteristics may differ,

there is no difference in the fundamental conditions, which
are
nothing more or less than the major premise and the minor premise;
in other words, the presence of the probans in the subject and the
necessary concomitance of the probans with the probandum.
The
fivefold characteristic of the probans is only symptomatic of. this

necessary concomitance with the probandum and the lack of one or
the other characteristic does not disqualify
1 Nsu, p. 478..
2 anyath kevalnvayidharmasya
mnatvpattch.
a probans, provided its
kevalavyatirekiny abhvt
IM*
tasynanu-
roo
INDUCTIVE REASONING
[CHAP.
necessary concomitance with the probandum is certified.
The lack
of presence in the homologue or absence in the heteroiogue where

the homologue or the heteroiogue are not possible, does not detract
from the validity of the necessary concomitance between the probans
and the probandum.
What is essential in inference is the presence
of the probans in the subject and necessary concomitance of the

probans with the probandum1.
Jayatlrtha in answer to this defence
of the Naiyayikas maintains that the Naiyyikas have hit upon the
truth undoubtedly, but have failed to gauge the relative value and
cogency of the conditions asserted by them.
The presence in the
subject is only a matter of empirical truth and has no bearing upon

the logical necessity, which makes the deduction of the conclusion
irresistible.
The logical necessity lies in the necessary concomitance
of the probans with the probandum and the existence of the probans
in the subject is more or less an irrelevant accident, so far as the
.validity of inference is concerned.
The miner premise only serves
to.establish the actual incidence of the probandum shown in the

conclusion through the incidence of the probans, but the deduction
of the probandum is made possible by the necessary concomitance
alone.
The presence of the necessary
concomitance alone is the
.sufficient guarantee and warranty of inference and the presence of

the said probans in the subject is as irrelevant as the presence in the
homologue or the absence in the heteroiogue.
And this necessary
concomitance is equally operative in ( i ) reasoning, (2) categorical

inference for proof (sdhannumna), (j) categorical inference for disr
proof >
. (dsannumana)
and
(4)
implication
(arthapatti)2.
The
Mdhvasv regard, all these four types of arguments as cases of m>

ference.
The difference in details is only responsible for their variation
as subspecies of.the same genus, namely, inference.

- We need not elaborately deal with the objection of the Naiyayika
on the score of the actual absence of the probans in the subject.
.1 smnyalaksanasadbhvt visesalaksanbhve'pi na dosa i
cet?
Nsu, p. 478.
3
tt....,..vyptyaj)eksay
gamakatvasya tarke'pi vidyamnatvat.
Ibid,
It is
yi]
101
a truism that the probans does not exist in the subject as a matter of
fact in reasoning.
In categorical inference the truth of the minor
premise, which shows the incidence of the probans in the subject, is

regarded as a necessary factor by all schools of logicians in view of the
consideration that the arguer is interested in establishing the necessary
incidence of the probandum in the subject.
The demonstration of the
incidence of the probandum in the subject is inspired by a practical

consideration and so the incidence of the probans in the subject is to
be shown as a matter of practical necessity.' We have shown that the
minor premise has very little bearing on the theoretical and logical
necessity and so its material truth is not regarded as a material condition of inference in general, although it is not denied that it is a necessary factor in categorical inference in view of the practical interest of
the arguer.
In reasoning, the minor premise is hypothetically assumed
in conformity with the assertion of the opponent for the purpose of

showing the necessity of admission of the prdbandum in the subject,
which does not as a matter of fact possess the same.
The lack of the
material truth of the minor premise is thus no valid objection against

the status of reasoning as a variety of inference1.
There is another objection, which we have touched upon in our
treatment of reasoning in the Pramanapaddhati.
discussion in the Nyyasudb
But the elaborate
throws an additional light upon the prob-
lem and this is our apology for embarking upon it over again.
The
objection is that the probandum that is necessarily deduced in reasoning

does not actually belong to the subject;
The deduction of a proban-
dum in a subject in which its absence is proved by another source of

knowledge is called the fallacy of contradicted probans.
The infer-
ence of the absence of heat in fire is an instance of this fallacy.
Rea-:
soning, if given the status of inference, will be vitiated by this fallacy"

as the probandum is invariably unreal.
Another defect is that the
inference of an unreal probandum in opposition to one's previous commitment regarding the" absence of the probandum is what is technically called the admission of a contrary conclusion (apasiddhnta) and
the perpetration of this defect deprives the arguer of victory in a dei
Nsu, p. 478.
IO2
INDUCTIVE REASONING
bate.
[CHAP.
But these charges are out of place in reasoning, because it is
not a case of inference, and so not an organ of knowledge, which are

alone liable to be assailed by such charges.
But the advocate of the
inferential status of reasoning will have no room for escape from these
strictures1,
Jayatlrtha seeks to clarify the ground of the objection of the
Naiyyika by posing a dilemma.
Does the Naiyyika mean to say
that the absence of smoke cannot be inferred in the hill, because it is

contradicted by the knowledge of the presence of the smoke in it?
Or does he mean to show that the inference of absence of smoke
involves contradiction of his previous commitment that the smoke is
present in the hill? 2
But neither of these contingencies arises and so
the apprehension of the Naiyyika, which leads him to put reasoning

in a separate class from inference, is groundless.
In reasoning the pro-
bandum is not sought to be proved on its own account, but rather

what is intended is that the admission of the probans makes the
admission of the probandum inevitable.
When the opponent denies
the presence of fire in spite of the presence of smoke in the hill, the
denial of fire, is shown to involve the denial of smoke, the presence of
which is indisputably attested, and so it does not lie in the mouth of
the opponent to acquiesce in the absence of smoke, which is shown to
follow as a necessary consequence of his refusal to accept the true
conclusion, namely, the presence of fire in the hill 3 .
The charges of
contradiction would be inescapable if the arguer intended to establish

the conclusion independently.
But this is not the case.
That the
arguer does not independently seek to establish the absence of smoke

in the hill in defiance of undisputed knowledge of the presence of;
smoke is apparent from the fact that the minor premise is given in a
hypothetical form, namely,
1
"If the hill be devoid of fire, it will be
bdhpasiddhntasadbhvn na tarkasy 'numnatvam iti ce$?

Nsu, p. 478.
2 kirn pdyatopetasya nirdhmatvder bdhyatvam tatproktau c'pasiddhantah.

Ibid., p. 479.
3 n'dyah. yata pdanam nm'ngikartavyatjnpanam, na punah sadbhavapratipdanam.
yadi visam bhaksayisyasi tarhi marisyasi'tt yath.
'
Ibid.
vil
103
devoid of smoke". The case is exactly on a par with such hypothetical proposition as "If you swallow poison you will die". The hypothetical proposition only states the necessary connexion between two
events, namely, swallowing of the poison and death and not the actual
historical truth of the events. Similarly in the reasoning "If the hill
be devoid of fire, it will be devoid of smoke", only the logical connexion between the ground and the consequent is demonstrated. And
the implication of it is that the admission of the ground makes the
admission of the consequent inescapable and not that the ground and
the consequent are actual historical truth. The fact that the arguer is
not concerned in proving an independent conclusion is obvious not
only from the hypothetical form of assertion, but from the fact that the
logical culmination of reasoning is the demonstration of the truth of
the contradictory. The employment of reasoning as reductio ad
absurdum aims at proving that the hill is not devoidôf fire as it is not
devoid of smoke. The charges of-contradicted reason (badha) and of
the contradiction of the accepted position (pasiddhnta) are therefore
found to be creatures of a muddled logical sense1*. As regards the
charge of unreality of probans (asiddhi) we have shown that it is
groundless and our conclusion is further reinforced by the fact that the.
probans _ is stated in a hypothetical way. The meaning of this
hypothetical form of statement "If the hill be devoid of fire" is that,
"If you assert it to be devoid of fire" and this assertion of the opponent is a real fact. The accusation of the fallacy of the unreal probans.
is thus shown to be as unreal as the charge of contradiction (badba)
and the charge of self-contradiction (apasiddhanta) have been found to
be2..
There is another objection of the Naiyyika, which seems to
be really formidable. Inference has been broadly classified under two
heads, namely, one for proof of a position and the other for refutation
of the position of the opponent. The Mdhvas have endeavoured
to place reasoning under the second class of inference.' But reasoning
cannot even be regarded as a species of inference for refutation,
1 Nsu, p. 479.
2 Ibid.
t>4
INDUCTIVE REASONING
[CHAP.
because inference for refutation is valid only if it serves to generate the

valid cognition of a real defect. Reasoning employs a logical apparatus which is not objectively true, but accepted only in consonance with
the assertion of the opponent. Inference to be valid must be based
upon a probans and a probandum, which are necessarily true and
accepted as such by both the parties. Reasoning, on the contrary, is
based on a probans and a probandum which are unreal and so cannot
but be a false pretender to the status of inference. The defect that is
sought to be proved by reasoning cannot but be an apparent defect
and not a real one 1 .
In reply to this charge Madhva has observed in his commentary
that reasoning is in the same position as casuistry (chala) and sophism
(jatf.
Casuistry and sophism are shown to be vicious arguments
only by the help of the probans and the probandum, that is, the
logical apparatus, employed by the opponent and admitted to be true
by him alone. If the logical apparatus is to be objectively -valid-and'
as such to be accepted by both the parties, then casuistry and sophism
can never be proved to be forms of sophistry, which they actually are.
The Naiyyika may contend that the employment of the same vicious
logical apparatus by him for the refutation of the sophistical arguments
of the opponent is not made by way of assertion of objectively valid
truth. The adoption by the Naiyyika of the same logical apparatus
employed by the sophist is made for the purpose of demonstrating
that the logical apparatus employed by the latter for picking holes in
the argument of the proponent acts as a boomerang against the
sophist, as it serves to invalidate his own position also. In other words,
the sophist is shown that the logical apparatus employed by him is
suicidal, since it contradicts his own position. The Naiyayika never
gives it to understand that the logical apparatus is admitted by him to
: i- dvividham ev'numnam, sdhanam dsanam ca. tatra tarkp na sdhanam
iti tvat bhavatm api sammatam,.n'pi dsanam, dustipramjanakam hi dsanam
nma. na ca tarkasya tat sambhavati
.....asdhun ca jnpyamn dusrir
bhlsabhtai ve'ti katham tarko'numnatn?
l
Nsu, p. 479.
2 chalajtyor dustattamlavyutpdaiienai vai tat samnayogaksemam. ft
Ibid.
vi]
105
be valid but is all along conscious that it is unsound and vicious and so
he does not expose himself
to the charges that arc levelled against
reasoning. Madhvacarya observes in this connexion that he has compelled the Naiyyika to make a momentous admission that the employment of a false logical apparatus in seeming deference to the opponent
for the purpose of exposure of contradiction in the opponent's argument
is a legitimate logical device in casuistry and sophism.
also in the same position as casuistry and sophism.
Reasoning is
The assertion of
the false pro bans in the subject, in one word, of the false minor premise, is hypothetically made by the arguer out of pretended deference
to the opponent and not as a valid proposition.
He means to assert
that the assertion of the false probans by the opponent makes the
assertion of the false probandum inescapable.
The admission of the
absence of fire in the hill makes necessary the admission of the absence
of smoke, which is opposed to the evidence of every organ of knowledge.
The admission, which makes necessary the admission of an
issue contradicted by the testimony of all accredited organs of knowledge,
is worthy of being dismissed as an absurd statement.
The
difference between sophism and reasoning, though both are cases of

inference alike, is that sophism is vitiated by self-contradiction, whereas
reasoning culminates in a conclusion, which is contradicted by an accredited organ of knowledge.
But this difference is only apparent, since
the admission of a proposition contradicted by the testimony of an

accredited organ is ultimately a case of self-contradiction.
To make an
assertion opposed by the testimony of all accredited organs is to make

an assertion subversive of a truth accepted universally and so by the
assertor himself also.
Viewed in this light reasoning will be seen to
serve the same purpose as sophism does.
In other words, both types
of argument end in convicting the opponent of self-contradiction.

The truth of this finding was anticipated by Udayancrya, who
asserted 'reasoning to operate as the limit of doubt'1.
'By limit' he
meant reasoning to be the instrument by which the doubt of the

sceptic was shown to be contradicted and this contradiction of the
1 tarkah sankvadhir matah.

NKu, III.
14
* INDUCTIVE REASONING
[CHAP.
doubt of the doubter is nothing but a case of self-contradiction.

Madhva concludes that reasoning should be legitimately held to be a
case of inference inasmuch as it is conditioned by the same logical
apparatus, which makes an accredited case o inference a valid organ
of knowledge and inasmuch as the objections levelled by the Naiyyikas against reasoning being accorded the rank of inference have been
found to be inspired by inadequate appraisal of the essential condition
of inference.
We may refer en passant to Vekatantha, the celebrated exponent of Rmnuja's school and the elaborator of the Rmnuja school
of logic, who substantially agrees with the philosophers of Madhva's
school in regarding reasoning to be a species of inference. Venk'atantha has devoted considerable attention to this problem, but
the logical result achieved by him is rather meagre and may be regarded as a faint ec&o of Jayatirtha's conclusion. The interest of
Venkatantha is rather textual than logical, and he has strained every
nerve to demonstrate that his conclusion that reasoning is a valid instrument and a species of inference1 at that, is not in conflict with the
views of Rmnuja and the traditional exponents of Rmnuja's philosophy. We refrain from reproducing this exegetical enterprise, since
it has very little speculative interest of its own and is calculated only
to disarm the suspicion and hesitation of the loyal adherents of the
school.
I tarkasy'aumnarsyaprthagblivat
NP, p. 324.
CHAPTER VII
REASONING AS ELUCIDATED IN THE TARKATNDAVA
BY VYASATIRTHA
Vyasatirtha flourished in the early part of the 16th century A . D .
and is a representative writer of the Mdhva school, whose contributions have served to raise the status and prestige of the Madhva school
of philosophy to the highest pinnacle of glory.
His mastery of the
dialectics of the Navya-Nyya school of Mithila" and Navadvlpa, his

meticulous knowledge of the writings of the rival schools of thought
and the writings of the previous philosophers of his own school, and
his extraordinary critical outlook and penetrating logical insight are
abundantly proved in his writings.
The Nyayamrta
and the Tarka-
tandava are the two works of Vyasatirtha which deserve special mention.
The first, as has been observed before, created a perilous situa-
tion for monistic Vedlnta and the second is a smashing onslaught on

the fabric of logic built by the Naiyyikas. The Tarkatndava is beinopublished by the Mysore University, and the publication of the part
which deals with inference is still awaited with considerable eagerness
by scholars interested in logic.
I had the singular good fortune of
having access to this unpublished part in a manuscript copy, which was

in the possession of the late lamented Professor S. N . Das Gupta.
Professor Das Gupta with his characteristic patronage of young scholars
allowed me to keep this manuscript for a considerable time, and the
present chapter on reasoning is a result of my study of it in the
manuscript.
We have already dealt with the dissertation on reasoning by Jayatlrtha in his two works.
Jayatrtha naturally exercised a profound in-
fluence on Vyasatirtha and the speculations of the latter on reasoning

have naturally been influenced by Jayatrtha Y arguments.
Vyasatirtha
endorses the conclusion of Jayatrtha and follows up the arguments

produced by him.
There would have been no justification for us to
embark upon a fresh dissertation on Vysatirtha's writing, if it had

been only a slavish reiteration of the old arguments.
But fortunately
che treatment of reasoning by Vyasatirtha, though for the most part
io8
INDUCTIVE REASONING
[CHAP.
directed into the channels carved out by Jayatrtha, abounds in original

reflections not only in respect of manner, but also matter.
There will
be some amount of restatement of the same topics, but we trust that it

will not be unjustified, since this will be the means of introduction of
new matters contributed by Vysatlrtha.
Accordingly I appeal to the
good reader who will favour this humble attempt of mine with a perusal,
to make allowance for unavoidable reformulation of the same problems.
Vysatlrtha regards reasoning to be a case of inference in conformity with the traditional finding of his school.
In order to establish
this conclusion he first of all seeks to show the inaccuracy of the conception of reasoning as a case of conscious ascription, as has been set
forth by the Naiyayikas in their traditional definitions.
We have
shown in the concluding portion of chapter v how the conception of

service rendered by reasoning has undergone a radical change in the
hands of the Naiyayikas themselves, from being only a presumptive
evidence to a categorical
finding.
Vardhamna has taken pains to
establish that reasoning primarily shows the absurdity of the contradictory possibility, and secondarily
shows the absence of conflict and
opposition of the finding of an organ of knowledge with established

truth.
Thus, there has been a veering round from the hypothetical
assumption to a categorical finding as the service rendered by reasoning.
Vysatlrtha lays stress upon the categorical nature of the finding
registered by reasoning and gives an altogether new orientation to the

conception of the nature of reasoning and its result.
If it were a
mere case of ascription, even deliberate (hryropa), it would be impossible not to regard a case of deliberate ascription, asset forth in
the instance we are going to state, as a case of reasoning.
Suppose,
for instance, the object lying ahead is ascertained to be a tree and

not a human being.
Suppose further, that a person still deliberately
insists on regarding it asa human being, and for that purpose

ascribes to it hands and feet, and concludes that he sets a, human
being.
This would be an act of deliberate perception of the tree1
i ... hryo 'yam ropah sa ca bdhito'pi bhavati'ti cet, na, purusatvbhavaniscaynantaram pratyaksdV hrye vypyakaracaraniopj janye pratyaksc hryapurusrope ativyptch.
JT. Ms. p. 192.
yii]
109
as a human being made plausible by the wilful ascription of the

characteristic features of a man to it.
Here the ascription or assump-
tion of a determinate concomitant (yyapyadharma)

determinant concomitant (yypakadharma).
leads to that of a
But it is not regarded as a
case of reasoning even by the Naiyyika.
He would contend that
it is a case of perception and Vyasatirtha would endorse it as the correct finding. But why should it not be
a case of reasoning?
answer is given by Vyasatirtha that the knowledge is
The
not necessitated
by necessary concomitance, but by the help of the visual organ or the

mind as aided by it Reasoning becomes effective only because it
operates through the knowledge of the necessary concomitance alone.
And this is exactly the determinate characteristic of inference. And so
there is no reason to repudiate the inferential nature of reasoning1.
Vyasatirtha thus disposes of the definition of reasoning as a case of
ascription and defines it in terms of categorical knowledge as follows :
"Reasoning is the valid cognition of the negation of probans or ground
as the determinant concomitant of the negation of the probandum or
the consequent as admitted by the opponent". It may be also defined
as 'a valid cognition of the unavoidable admissibility of the negation
of the probans entailed by the admission of the negation of the probandum. 12 The denial of fire in a smoking hill in volcanic eruption is
tantamount to the admission of the negation of smoke.
Reasoning
serves to show that the negation of fire stands in the relation of

necessary concomitance with the negation of smokethe former being
the determinate concomitant (vypya) and the latter being the determinant concomitant (yypaka).
The relation between negation of fire
and negation of smoke is really one of necessary concomitance, and

reasoning only brings home the knowledge of this truth.
There is
thus no reason to deny it the status of an organ of knowledge.

If reasoning be an organ of knowledge and a subvariety of infer1
tasmt tarkah anumnavisesa eva, vyptibalena gamakatvt.

TT. Ms. p. 194.
2 manmate tv agikrtena sdhybhvena saha anagkrtasya sdhanabhvasya vypakatvapram v. sdhyabhvgikranimittik sdhanbhvasy

f
gikarttavyatvapram v".
'Ibid, pp. 194-95.
i io
INDUCTIVE REASONING
[CHAP.
ence at that, as made out by the Mdhvas, it must be shown that

reasoning has the same logical apparatus as inference. There must be a
probans and a probandum and a necessary concomitance between the
two. To revert to our old example. "If the hill be devoid of fire,
it must be devoid of smoke," the probans here is the negation of fire,
and the probandum is the negation of smoke. Though there is
necessary concomitance between the two, the incidence of the probans
in the subject, as indicated by the predication of it in the first clause of
the hypothetical proposition, s unreal. Similarly also the existence
of the probandum, namely, the negation of smoke as asserted in the
second clause is a myth. It has been shown that these are serious
defects, which make inference illegitimate and inadmissible. We have
shown in the preceding chapter how Jayatirtha admits the truth of the
assertions of these facts by the Naiyyikas and how he takes away the
sting from their allegations by asserting them to be the necessary
conditions of reasoning as a hypothetical inference. r In other words,
Jayatirtha saves himself from the strictures of the Naiyyikas on the
score of the false minor premise and the false conclusion by the plea
that he does not make these assertions as objective truths, but only as
hypothetical assumptions. It has been argued that the falsity of the
minor premise and the absurdity of the conclusion, though they are
real defects in categorical inference, are not only not defects in reasoning as reductio ad absurdum, but are the necessary conditions of the
same. Reasoning is only a means, which though objectively false,
serves to correct the opponent of his vagary. Here the end justifies
the means. We have already dealt at length with the defence of
Jayatirtha in the preceding two chapters and so we need not reproduce
the version of the same as given by VyasatTrtha. We only
propose to deal with the novel vindication of reasoning as inference
given by Vysatirtha and to show how he boldly equates reasoning
with categorical inference, and still avoids the charge of the false minor
premise and the false conclusion without taking cover under the
privilege of hypothetical inference, which has been claimed by his
predecessors. It becomes his modesty and reflects credit on his loyalty
to the traditional exponents of the Mdhva school of thought that
Vysatlrtha seeks to emphasise that all the findings made by him are
VII]
itt
nothing but the logical deductions from the implications thrown out
by Jayatirtha.
The arguments of Jayatirtha have sought to prove that though there
is contradiction involved by the falsity of the minor premise and of the
conclusion, still the status of reasoning as inference is not affected by it.
But the problem arises. How can reasoning be accorded the status of
an organ of valid cognition, when the minor premise and the conclusion
are in conflict with universally accepted truths? Vysatrtha does not
believe that the minor premise or the conclusion of reasoning is false
and so he thinks the problem to be only a false creation of the
Naiyyika, arising from his misconception of the' nature of reasoning.
Reasoning is requisitioned when there is doubt. Now, the doubt that
a smoking hill may be destitute of fire is possible only in either of the
following ways: (i) Firstly, negation of smoke may not be the determinant (yypaka) of the negation of fire. Secondly, though one may
be the necessary concomitant of the other and negation of smoke be the
determinant of the negation of fire, negation of smoke may yet be
possible in the subject.1 Now, the first case of doubt is dismissed by
showing that negation of fire is a determinate of the negation of smoke,
and this is accomplished by the hypothetical proposition, "If it be devoid of fire, it must be devoid of smoke." The hypothetical form does
not imply that the proposition does not assert a categorical truth. As
a matter of fact necessary concomitance between two sets of facts is
usually asserted in a hypothetical form. Thus, for instance, such propositions as, "If the gayal be similar to the cow, the cow also must be
similar to the former," or "If a person alive is not at home, he must
be present outside," or "If there be smoke there must be fire," do not
assert a tentative possibility, but a necessary and universal truth. The
hypothetical form is the usual medium of expression of a necessary
relation between the antecedent and the consequent clauses.2 The second
t sadhmo'pi niragmko'sti iti sank hi dvedh "sambhavati. niragnikatvam
prati nirdhmatve vypakatvasy aiv' bhvena v, saty api vypakatve pakse
nrdtimatvasy* pi sattvena v.
TT. Ms. p. 2 0 1 .
2 tatr'dyah nirdhmatvasya vypakatvapradarsakena yadi niragnikah tarhi
nirdhmah ity anena nirasyate.
ibid.
INDUCTIVE REASONING
[CHAP.
clause states the conclusion which necessarily follows from the first
clause as the premise.
The second case of doubt is effectively
dis-
missed by showing the nadmissibility of the negation of smoke in the

subject, which is directly perceived to be issuing smoke.
Now, the reasoning, "If the hill be devoid of fire, it will be devoid
of smoke," is nothing but a case of inference in which the negation
of smoke is shown to be the determinant concomitant of the negation
of fire, and the whole proposition is to be considered to be the
ment of
state-
the probans. The probandum oE this is the proposition that
the negation of the negation of smoke is the determinate concomitant

of the negation of the negation of fire, according to the terminology we
have adopted before. The cognition of the aforesaid probans is the instrument of the cognition of the aforesaid probandum. And both these
cognitions are true, since the negation of the determinant is the determinate concomitant of the negation of the determinate.1
To illustrate
it by a concrete example, smoke is the determinate concomitant (yyapya)

of fire and the fire is the determinant concomitant (vypaka) of smoke,
but the negation of fire is the determinate concomitant of the negation
of smoke.
We can infer the determinant concomitant from the deter-
minate concomitant and not vice versa, because the determinate cannot
exist independently of the determinant, though the latter can exist independently of the former.
Thus, smoke is the determinate of fire,
because it cannot exist independently of the latter.
But fire can exist
independently of smoke and so the negation of smoke does not entail

the negation of fire. But smoke is impossible without fire and so the
negation of fire necessarily entails the negation of smoke.
The relative
position of two terms qua determinant and determinate is reversed in

the case of their negatives.
When the positive term is determinant,
the negative of it becomes the determinate.2

1 atra ca niragnikat;vam prati nirdhmatvasya vypakatvam Ungarn,
tajjnm karanam, niragnikatvbhvam prati nirdhmatvbhvasya vypyatvam
lg tajjnnam phalam dvayam api pramrpam eva
vypakbhvasya
vypybhvam prati vypyatvc ca.
TT. Ms. p. 203.
2 vypyavyapakabhvo hi bhvayor ydrgisyate. tayor abbvayos tasmd
viparitah pratiyate: $lokavrttika, sl. 121-122 (chapter on inference).
yn]
113
We have made this digression to show that the doubt that there
may be smoke without fire is set at rest by showing that fire is the
determinant concomitant of smoke and negatively by showing that the
negation of fire necessarily entails the negation of smoke, the latter
being the determinant concomitant of the former. Likewise the negation
of the negation of smoke (to put it positively, the smoke) is the determinate concomitant of the negation of negation of fire (that is to say,
fire).
Let us now apply the results we have obtained to the problem at
hand, viz. whether reasoning is a case of inference.
The proposition
''If A were devoid of fire, it would be devoid of smoke" shows that

negation of fire is the determinate concomitant of the negation of smoke
and so the former necessarily entails the latter. There is nothing in theproposition which can be taken exception to.
It is a true assertion and
serves as the probans of the assertion, "The negation of negation of

smoke is the determinate concomitant of the negation of negation of
fire," which is to be considered as the probandum by virtue of the fact
that it follows from the previous assertion. Now both the propositions,
the probans and the probandum, are true and as reasoning consists of
these two propositions, there is no earthly reason to call in question its
validity.
The necessary concomitance between the two negations thus
proved is applied in the negative implication of reasoning, which is the

ultimate upshot. To be explicit, "If A were devoid of fire, it would be
devoid of smoke," is the first part of reasoning, which works as the probans. "But it is not devoid of smoke" and so "it cannot be devoid of
fire" is the final part of reasoning and states the conclusion or the probandum.
The conclusion only asserts that negation of negation of smoke
being the necessary determinate concomitant of the negation of negation of fire, the actual incidence of the former entails the incidence of
the latter.
The first proposition is true, because it states a necessary
universal truth and the second is equally valid, as it also states necessary truth which follows from the former.
These two are universal
propositions and assert the necessary universal concomitance between

two terms, which is not liable to be denied by any person.
The asser-
tion that the subject, say the hill, is not devoid of smoke is based on
i*4
IHDUCTIVE REASONING
[CHAP.
an objective fact, and the consequence that it cannot be devoid of fire

follows as a matter of necessity. Vyasatrtha observes that since both
the probans and the probandum are true and since the necessary concomitance between them is not liable to be repudiated as false, there
does not seem to be the slightest logic in the condemnation of reasoning. Reasoning has the same logical apparatus as inference and as such
should be classed with it.1
The objection of the Naiyjrika to the validity of reasoning is based
on the assumption that 'negation of fire' is made the probans and 'the
negation of smoke' is made the probandum in reasoning.2 Were it
really the case, the Naiyyika's objection would be unanswerable. But
Vysatlrtha shows that the case is entirely different and so the accusation of the Naiyyika is absolutely groundless. Jayatlrtha first accepted
the assumption of the Naiyyika to be true. In other words, he accepted the version of the Naiyyika that 'negation of fire' was the probans
and 'negation of smoke' was the probandum. But he maintained that
the assertion of the false probans and that of the false probandum did
hot entail the logical fallacies alleged by the Naiyyika, since they were
only assumed for argument's sake and not independently stated as objective facts. Such assumption is the necessary presupposition of a
hypothetical inference adduced as a reductio adabsttrdttm. Vyasatrtha
maintains that this defence of Jayatirtha only aims at showing that the
Naiyyika can be defeated even on his own terms. It does not, however, give out the real significance of reasoning and as such the defence
should be considered to be inspired rather by a sense of chivalry which
would find satisfaction in vanquishing the opponent with his own weapons freely chosen by him.
Vyasatrtha now throws off this chivalrous pose and asserts that the
charge of the falsity of the minor premise and the conclusion is groundless. Even if we assume that in the reasoning, "If the hill were devoid
of fire, it would be devoid of smoke,1' the probans is negation of fire
and the probandum the negation of smoke, as has been asserted by the
Naiyyik and accepted by Jayatirtha, yet there will be no difficulty on
1 TT. Ms. p. 204.
2 Ibid.
vii]
115
the score of the unreality of the minor premise and. the conclusion.
It
has been the complaint of the Naiyyika that the probans, viz. the
negation of fire and the probandum, viz. the negation of smoke, do not
actually belong to the subject, viz. the hill and so the minor premise
and the conclusion are false.
And this is the greatest stumbling-block
in the way of reasoning being given the status of inference.
Vysa-
tirtha observes in reply that it is not a fact that the probans and the
probandum do not stand in relation to the sujpject.
The relation f
smoke and fire to the hill is one of conjunction and in the inference of
colour from taste, (cf. the mango is possessed of colour, because it is possessed of taste) taste as the probans and colour as the probandum stand
to the subject in the relation of inherence.
In the inference, again,
*A must be knowable, because it is existent/
the probans 'existence* is
related by way of inherence to the subject, but the probandum

'knowability* stands to the subject in the peculiar relation which
subsists between a cognition and its object.
relation (yisayatasambandha).
We may call it cognitive
The Naiyyika has admitted two kinds
of relation: (1) One actual, that is to say, the relation by which the
actual incidence of one term in the other is determined
(vrttiniymaka-
sambandha); (2) another which makes us understand that one term is

related to another term, though there is no factual incidence of one in
the other (vrttyaniyamakasambandha).
Conjunction, inherence and
according to some, identity are examples of the former.
The jar
actually exists on the ground in the relation of conjunction.
The
examples of inherence between a quality and a substance and so on are

matters of common knowledge.
The proposition, "A is identical with
A " shows that A stands to itself in the relation of identity.

has been regarded as a factual relation by some logicians.
are illustrations of the first kind of relation.
Identity
All these
But the relation between
a subject and an object or between cognition and its content is an

example of the second kind.
Nobody can deny that they are related,
though the factual incidence of the one in the other is not possible.
The logical requirements that there must be a true minor premise iand
a true conclusion^ in other words, that the probans and the probandum
must stand in relation to the subject, are satisfied irrespectively by both
the kinds of relation. If actual relation, such as conjunction, inherence
1x6
INDUCTIVE REASONING
[CHAP.
and the like, were thought to be the prerequisite of inference, a thing

could not be inferred to be knowable on the ground of-its existence,
since the relation between a thing and the quality of being knowable
is not inherence.
Any kind of relation which makes us understand
that the probans and the probandum are related to the subject will
suffice to constitute a true minor premise and a true conclusion1.
Let us now apply the results of our finding to reasoning and see
whether a true minor premise and a true conclusion are possible in it.
Negation of fire is given as the probans and negation of smoke as the
probandum.
Do they stand in any relation to the hill?
Vyasatlrtha
maintains that there is a relation between the subject on the one hand
and the probans and the probandum on the other even in the case
under consideration.
Negation of fire is asserted by the opponent to
be true of the hill and this means that negation of fire stands in the
relation of being admitted to be present in the hill.
And as regards
the probandum, viz. negation of smoke, it is asserted by the arguer

as the necessary consequent of the assertion of the negation of fire by
the opponent.
In other words, negation of smoke is asserted by the
arguer as the consequence which is to be admitted by the opponent as

a matter of undeniable logical necessity.
Negation of smoke as the
probandum cari thus be said to stand to the hill in the relation of being
one to be necessarily admitted by the opponent.
Thus, the probans
and the probandum are seen to be related to the subject in the example
of reasoning cited above, and this is typical of all such cases2.
The
charge of the lack of minor premise and conclusion on the ground of

the lack of actual relationship is found to be derived from a hasty
generalisation that the relation must be one of actual incidence, which
we have shown to be only a small part of the field of relations.
Vyasatlrtha gives another alternative interpretation. "If the hill be
devoid of fire" means "If you admit that the hill is devoid of fire.*/
- t
TT. Ms. p. 206.
2 yadv pdakartipalingasya niragnikatvasya pakse parbhyupagatatvam

cva paksasambandha itt n'siddhih...lingino nirdhmatvasy'pi pakse abhyupagamaniyatvam eva hi paksasambandhah.
bicf.
Vii]
REASONING AS ELUCIDATED IN THE TARKATJNDAVA
The admission of the negation of fire is the probans

fact, and so there can be no question of the
probans in the subject.
117
and this is a true
non-existence of the
If the admission were to be
repudiated, there
would be no occasion for the application of reasoning as reductio ad

absurdnmL.
The minor premise which is stated in the antecedent
clause of the hypothetical proposition, "If the hill be devoid of fire,

then it will be devoid of smoke" is true and real and so the objection
based upon the unreality of the minor premise ig found to be baseless.
As regards the charge of the unreality of the probandum, viz. 'the
negation of smoke,* which is given in the consequent clause, it will be
equally found to be a baseless allegation.
The clause 'then it will be
devoid of smoke' does not mean to assert that the negation of smoke is
actually present in the hill. The meaning of the clause is the assertion
that in that case you cannot but admit that the negation of smoke is
true of the hill. The hypothetical form of assertion has this very logical
implication and c is exactly on a par with the proposition, "If you
take poison you will die." 2
What reasoning aims to establish is that
the admission of the negation of fire necessarily entails the admission

of the negation of smoke.
The opponent may decline to admit the
consequent, in other words,, he may refuse to believe that there is no

smoke in the hill.
But this refusal does not exempt him from the
charge of self-contradiction, since the consequent follows from the ground as a matter of unimpeachable logical necessity.
In the light of
this interpretation reasoning turns out to be a case of categorical inference.
It may be syllogistically propounded as follows:
"The hill
must be admitted to be devoid of smoke, since it is admitted to be

devoid of fire, as for example is the case with a lake."3-
The hypothe-
tical mode of assertion does not, as we have explained before, imply

x athav yadi niragnikah syd icy asya yadi niragnikatvgkra icy
arthah. tathgikras ca siddha eva. anyath tarkasy'py anutthnt.
*
TT. Ms. p. 208.
v
2 evam tarhi mrdhmah syd ity asya niirdhmatven 'gikartavya ity
arthah. pdanam hi agikartavyataya jpanam. na punah sadbhvaprattpdanam.
'Ibid.
3 parvato nirdhmatven'gikartavyah nragnikatven 'gikrtatvt hradavat
ity anumnam eva tarka iti sdhannumane tarkc'pi asiddhyadikam dosa eva ti
na tarkasya sdhannumnt vaisamyam api
lbidt
n8
INDUCTIVE REASONING
[CHAP.
that the meaning of the proposition is a tentative possibility. On the

contrary this is the usual form of assertion of a necessary and universal
relation between two facts.
The difference of the third interpretation from the second consists
in the fact that in the latter negation of fire is made the probans and
it stands to the hill as the subject in the relation of being asserted to
be true. In the third alternative the probans is the assertion of the
opponent and it has been shown that it is a true fact.1 In the first
interpretation the hypothetical proposition, "If the hill be devoid of
fire, it will be devoid of smoke" has been shown to be the statement
of necessary concomitance between 'negation of fire' and 'negation of
smoke.1 The negative proposition, "The hill is not however devoid
of smoke," which is the logical culmination oE reasoning, furnishes the
minor premise. In all these three interpretations the minor premise
asserts an objective fact and the conclusion is shown to be a necessary
logical consequence. So ail the allegations of the Naiyayika based on
the unrealness of the minor premise and the conclusion are proved to
be baseless and irresponsible utterances.
There arises one question in this connexion which cannot be ignored.
Reasoning has been placed under the head of inference for refutation,
which is a different type from categorical inference, which intends to
establish a positive conclusion. But, as interpreted here, reasoning has
been shown to be on the same footing with categorical inference, since
it ends in.registering a positive conclusion. This seems to be inconsistent with the classification proposed by the exponents of the Mdhva
school of logic. In reply it may be observed that reasoning serves to
show the untenability of the assertion of the opponent. There is thus
no inconsistency in the fact that reasoning has been classed under inference for refutation*. As regards the question of form, we can also
show that there is a difference between categorical inference and reasoning in that the probans employed in reasoning is fundamentally different from that, of categorical inference as a class. Neither admission is
made a probans, nor is it considered to be the relation between the
.:-: aira ca dvitiye pakse mragnikatvam eva lingam abhyupagamas tu tasya
paksena sambandhah. trtiye tu abhyupagama eva lingam iti bhedah.
.
.-;.;..
XX. Ms. p. 208.
VII]
REASONING AS ELUCIDATED IN THE TARKATJNDAVA
119
probans and the subject in categorical inference, as; is done in the

second and third alternatives.
The vindication of reasoning by Jayati-
rtha from the charges of the Naiyayika on the ground that the probans
is a matter of assumption, which makes the assumption of the probandum an inescapable consequence and so the charges have no raison
d'etre, has already been considered by us.
It goes without saying that
reasoning thus interpreted is to be placed in a different category from

categorical inference.1
The novelty of Vyasatirtha's exposition of the nature of reasoning
is undeniable.
Of course Vyasatrtha gives quotations from Jayati-
rtha's writings and these quotations have been shown to adumbrate

the alternative interpretations propounded by Vysatlrtha.
Vyasati-
rtha has no doubt followed up the implications, but the bold formulation of the nature of reasoning in a manner commensurate with the
nature of categorical inference in respect of its perfect immunity from
the logical fallacies involved in the falsity of the minor premise and
the conclusion, cannot but evoke spontaneous admiration for his ingenuity and logical acumen.
Jayatlrtha has further pleaded that in
assessing the logical validity of reasoning it must not be taken apart

from the negative implication, which gives out the true conclusion.
But this was not in point of fact an original discovery, since Udayana
has shown that reasoning, apart from its negative implication showing
the truth of the opposite conclusion, cannot be accepted to record a
true finding. Vyasatirtha maintains that reasoning as defined and'
interpreted by him is both formally and materially valid and that
again even independently of the negative implication2.
He, however,
defends Jayatirtha by asserting that Jayatirtha's insistence on the re^

cognition o the negative implication as a necessary factor of reasoning1
is due to his realization that reasoning can operate as a reductio ad ati~
st*rdnm, only if it culminates in the proof of the opposite conclusion*.
Vysatlrtha submits that reasoning is valid formally and materially in
1 JT. Ms. p. 208.
2 etena. viparyayaparyavasnarpnumnanirapeksasya tarkasya na prmnyam iti nibandhdau Udayanadyuktarn nirascam tannirapeksasy aiva prasangasya
prrnnyopapdant.
Ibid., p. 210.
i2o
INDUCTIVE REASONING
itself and apart from the negative conclusion.

defined by him to be,
firstly,
the valid
[CHAP.
Reasoning has been
cognition of necessary
concomitance of the negation of the probandum with the negation

of the probans and,
secondly, to be the valid cognition of the
necessity of admission of the negation of the probans entailed by

the admission of the negation of the probandum.
Though both
these cognitions are valid in their own right and independently of

the aid of the negative culmination, yet reasoning has to be admitted to be necessarily dependent upon the negative culmination in
order to bring home the truth that the hill is not destitute of fire,
which was the original thesis, the vindication of which is the ultimate
aim and purpose of the application of reasoning,
Vyasatrtha's exposition of reasoning leaves no room for doubt
that it is to be classed with inference, and as such must be admitted
to b a valid logical organ.
In the course of our exposition of the
speculation of the Nyaya-Vaisesika school on reasoning in the first four

chapters, we have demonstrated the difficulty felt by the advocates of
these systems, regarding the psychological and logical status of reason*
ing.
The Naiyyikas have recognised that reasoning could not be
placed under the head of doubt or of error, because reasoning is noc

a case of vacillation between two conflicting alternatives, nor again is
it liable to be sublated by subsequent cognition.
to place it under valid cognition.
They also hesitated
Gagesa thought that reasoning
constitutes by itself a species of cognition, which is different from

doubt, error, and valid knowledge.
It has been sought to be explain-
ed as a case of conscious ascription like the ascription of a deity on

an image.
The image is known to be different from the deity,
but still the former is deliberately identified
with the latter.
This
identification is not a case of error, since it is deliberate and the

belief in the identity is all along known to be untrue.
The Naiy
yika contends chat the assertion of the hill to be devoid of fire and
consequently of smoke in spite of the awareness of the presence of
smoke and fire in it, is the result of an act of conscious superimposiCion like that of the deity on the image.
The identification of a
smoking hill with a smokeless hill is certainly false.

error as it is deliberate.
But it is not an
Such an act of cognition may be called
vii]
ni
assumption or imaging or ascription or conscious superimposition.

Whatever may be the nomenclature, the Naiyyika is positive that
such an act of cognition must not be confounded with valid cognition,
which is occasioned by the accredited organs of knowledge, Vyasatirtha however maintains that it is nothing short of perversity to deny
it the status of valid cognition. The assertion,""*'If the hill be devoid
of fire, it will be devoid of smoke", has been shown by him to be a
legitimate form of assertion of necessary concomitance between two
facts. And the hypothetical form is not symptomatic of the hypothetical nature of the cognition. Moreover the attempt to equate it with
the act of deliberate identification, which is involved in the case of
images of deities, is due to the reluctance of facing an unwelcome
situation. Undoubtedly there is deliberate identification of an idol
with a deity, but it is occasioned by a profound religious necessity.
The contemplation of the deity is facilitated by this identification with the idol and such contemplation is productive of a result of
supreme spiritual importance. But no such motive is to be discerned
in the case of the smoking hill being identified with a smokeless one1.
There is again a vital difference. There is no logical connection
between the idol and the deity, whereas the necessary concomitance
of a fireless entity with a smokeless entity is all that makes possible
the realization of the absurdity of the opponent, who believes the
presence of smoke to be possible in the absence of fire.2 This is the
motive and objective for the application of reasoning and ic can be
fulfilled only if there is necessary concomitance between the facts
asserted in the hypothetical proposition. The hypothesis of conscious
ascription or deliberate identification is entirely irrelevant, since it
has no bearing upon the result aimed at, namely, the reduction of the
opponent to absurdity. Vyasatirtha maintains that reasoning is nothing but a case of inference, since it works by means of the same
1 ...tasya phalnusandhnaspeksatvt vyaptijnanirapeksatvc ca, tarkasya

ca phalnusatidhnanirapeksatvt vyptijnnaspeksatvc ca.
TT. Ms. p. 213.
2
pratimaym
'nanubhavac ca.
16
nryanadhynasy
lbids
eva
parvatc
mrdhmatvadhynaaya
Hi
INDUCTIVE REASONING
[CHAP.
logical apparatus, namely, the universal concomitance and the minor

premise, as accredited cases of inference.
Furthermore, reasoning is resorted to with a view to effective
rebuttal of doubt.
It will be shown when we shall consider the
utility and service of reasoning that reasoning succeeds in rebutting
a doubt, because there is opposition between them in respect of contents. This opposition of contents shows that it is a logical opposition
and reasoning as the vehicle of such logical opposition should be regarded
as a logical weapon. The contention of Gangesa that reasoning is a
species of cognition different from the other recognised species is
nothing short of sophistry1. If the opposition of reasoning to doubt
be derived from the class-character they possess, then it will be a case
of natural opposition. But in that case Gagesa will not be able to
account for the fact that reasoning rebuts one kind of doubt and not
every other. The opposition is, therefore, logical and is based upon
the necessary concomitance at bottom. The doubt of the negation of
fire in the hill is rebutted by the consideration that negation of fire
presupposes negation of smoke as its determinant concomitant, which
is however contradicted by the knowledge of the presence of smoke in
it. Thus, reasoning only shows that the doubt cannot be entertained,
because it has not the sanction of the necessary concomitance at its back
and is on the contrary directly opposed by the necessary concomitance
between the negation of fire, which he regards as a possibility, and
the negation of smoke, which he cannot admit on account of the
contradiction involved by the cognition of smoke.
The Utility Of Reasoning
Reasoning is called i request to remove the doubt of the inefficiency of a probans adduced to prove a probandum, in other words,
when there is doubt of the necessary concomitance of the probans
with the probandum. Now, the doubt about the fallibility of the
concomitance between a probans and a probandum may be universal,
that is to say, without reference to the particular logical subject under
i tarkatvasya jtitve tarkasya tvadangikrtasamsayatnvartakatvam na syat,
jnnasya virodhivisayakatvea aiva tannivartakatvt.
TT. Ms. p. 215.
yn]
123
consideration1. Secondly, the doubt may have specific reference to the

subject under consideration though the concomitance of the probans
and the probandum may be accepted outside the present context2. To
take a concrete example. The concomitance of smoke with fire may
be doubted as a case of contingent occurrence and so the presence of
smoke need not be an instrument of proof of fire as a matter of universal necessity. Now, doubt of the infallibility of necessary concomitance, which is expressed n the major premise, may arise antecedently
to the employment of inference, or immediately subsequent to the
knowledge of the major premise and minor premise on account of the
operation of a conflicting organ of knowledge3. The possibility of
the latter occurrence makes necessary the admission of the two other
fallacies, viz. contradicted probans and countermanded probans in
spite of the opposition of the Buddhist logicians. We have already
shown in chapter 11 in the course of the exposition of Bhagratha
Thakkura's view how reasoning facilitates the: knowledge of the probandum, that is to say, of the conclusion by removing, the obstacle
imposed by the operation of a conflicting organ of knowledge. We
are already familiar with the objections of Srfvallabha to the possibility
of doubt subsequent to the knowledge of the major premise and the
minor premise. We have seen how Bhagratha and Rucidatta defend
the possibility of such doubt and how reasoning is of service in the
elimination of the obstruction.
Vyasatirtha admits the possibility of the two kinds of doubt,
antecedent and consequent to the knowledge of the premises. He
holds that these two types of doubt are at bottom concerned with the
necessity of the universal concomitance. The doubt is to the effect
1 ...tarkanivarttyaprayojakatvasanka hi dveciha paksataditarasdhranyena
sadhiimo'pi kascin niragniko'stv iti v.
TT. Ms. p. 216.
2 paksd anyatra niyame'pi pakse samabalasya prabalasya v sdhybhvapramnasy'bhimnena paksa eva sadhmo'pi niragniko'stv it va.
Ibid., p. 217.
3 tatra cly vyabhicrasakrp ligaparmarst prcin. dvitiy tu
pratipramnasakrp, tasmd arvacln.
Ibid.
i24
INDUCTIVE REASONING
[CHAP.
that the concomitance of the two terms used as the probans and the
probandum is contingent and accidental. This doubt of the contingency of concomitance is removed by reasoning, which shows that the
concomitance is necessary and universal, because the probans in
question is impossible without the necessary accompaniment of the
probandum. Reasoning is classified by Vyasatlrtha under two heads
in conformity with the twofold division of doubt, antecedent and
consequent to the knowledge of the premises. Reasoning, which
eliminates the doubt of universal concomitance, is instrumental to the
formulation of the major premise, which is regarded as the organ of
inference (anumitikarananispadaka). The second type of reasoning,
which removes the doubt arising subsequently to the knowledge of the
premises, directly helps the realization of the conclusion1. Vysatrtha
thus admits the two conventional types of reasoning, viz. (i) one
conducive to the knowledge of the universal proposition (yyptigrahaka) and (2) the other conducive to the knowledge of the
conclusion (visayaparisodhaka) by removing obstruction caused by
doubt in each case. Reasoning in this twofold role acts as an
auxiliary to inference. But the role of reasoning as an independent
organ as shown before is also admitted. We have shown in our
exposition of Jayatirtha's discourse on reasoning, that reasoning,
whether it acts as an auxiliary or as an independent organ, never
forfeits its status as an independent organ, and Vysatrtha has
endorsed this position in his work2.
We have expounded the nature of reasoning and the service it
renders in the light of the exposition of Vyasatlrtha. We now propose
to consider Vysatlrtha's criticism of the exposition of the nature of
service rendered by reasoning according to the Naiyyikas. Vyasatlrtha maintains that if reasoning be regarded as a case of assumption
or false ascription, as has been done by the Naiyyikas, it cannot be
supposed to make any contribution either to the realization of the
1 nivarttyabhedt nivartakas tarko'pi dvedh tatr'dyo'numitikarananispdakah, dvitiyas tu nispannena karanena pratipramnasanknirsrtham apeksitatvena phalopakrakatvd anumannugrhakah.
v
XT. Ms. p. 218,
2 Und., p. 219.
vii]
REASONING AS ELUCIDATED IN THE TARKATNDVA 125
conclusion or to that of the necessary concomitance. Reasoning as we

have repeatedly shown works as a reductio ad dbsurdum, because it is
based upon necessary concomitance. The denial of fire in a smoking
hill is shown to involve absurdity, because smoke is concomitant with
fire and consequently the denial of fire would involve the denial of
smoke, which is directly contradicted by the perception of smoke in
the hill. That the denial of fire involves the denial of smoke is due
to the fact that negation of fire is necessarily concomitant with the
negation of smoke. If, however, the denial of fire were of the nature
of an assumption and a make-believe and so also were the denial of
smoke, it passes our understanding how the pretended denial of fire
would be incompatible with the presence of smoke.1 This is the
fundamental weakness of the Naiyyika's position and infects all their
speculations on the utility of reasoning with an air of unreality and
make-believe. Vyasatlrtha has taken up the different views of the
utility of reasoning, (1) that it serves to prove the untenability of the
opponents position, (2) or to rebut the doubt of the possibility of the
opposite conclusion, (3) or to preclude enquiry after the opposite
possibility, which have been already examined by us in all their details in the first four chapters dealing with the position of the NyayaVaisesika school of philosophy. The objections that have been levelled
by Vylsatlrtha against these views have already been anticipated by
these philosophers and they have given suitable replies. It will serve
no purpose to dwell upon these arguments, which are only summed up
by VySsatirtha. It must be recognised that Vyasatlrtha's definition of
reasoning and his exposition of its nature and service really register a
high water-mark in the logical speculations of India and they ought to
be accepted as a distinct improvement upon the theories of-the NyayaVaisesika school.
1 ... pratnnaviruddhanirdhtnatvapdakatay niragnikatvam eva 'nistam,

na tu tajjnam tasyroparpatvena tadanapdakatvt.
JT. Ms. p. 225.
CHAPTER VIII
The Jaina logician regards reasoning as the organ of cognition of
universal and necessary concomitance between two terms.
It is an
entirely different species of cognition which can be subsumed neither

under perception, nor inference, nor any other kind of knowledge.
The condition of such cognition is observation and non-observation of
the co-occurrence of two terms.
is as follows:
The form in which it delivers itself
"A occurs only and necessarily when B occurs, and is
not seen to occur as a matter of universal necessity when B is absent."

The content of this cognition is the universal and necessary relation
between two terms e.g. A and B. 1
It is not however to be contended
that what is called the resultant cognition is nothing different from

what has been laid down as the condition of it.
It has been asserted
that observation and non-observation of two terms as two co-occurrent

events is the condition of the knowledge of the universal concomitance
of two terms. But the supposed two cognitions, the so-called condition
and the result are identical both in respect of form and matter, since
the observation of two terms in association together with the nonobservation of one term in the absence of the other is equivalent to the
cognition of necessary relation between the two.
The Jaina however
maintains that mere observation of co-presence of terms or non-observation of them out of relation, cannot give insight into the universality
and necessity of their relation.
Observation and non-bservation are
nothing but cases of perceptual cognition.
And as such their jurisdic-
tion is limited to present individual data.
But reasoning is a mental
act, which though it arises from the perceptual cognition of the present
data, has jurisdiction over past, present and future, and as such it is a
different type of cognition both in respect of form as well as matter.
It is a'fact that inference is possible if there is knowledge of universal
i upalatnbh'nupalambhasambhavam trikllkalitasdhyasadhanasambandh.
dylambanam idam asmin saty eva" bhavatTty akram samvcdanam hparanm
tarka iti.
PNT, III. 7, p. 500.
yin]
THE JANA CONCEPTION OF REASONING
and necessary concomitance at its back.
127
Ic is also equally an undenia-
ble fact that perception is necessarily limited to present data, and as

such is incompetent to have commerce with past, future and remote
occurrences.
However much the sceptic may decry the validity of
inference as an organ, the very possibility of the theoretical and practical activity of our life presupposes the infallibility of inference.
It
must therefore be admitted that the knowledge of universal concomitance is possible.
And it is for the acquisition of such knowledge
that reasoning has been postulated by the Jainas as the organ for the
same.
A question has been raised that an organ of knowledge must bear
a definite relation to the object which it is supposed to cognise.
Now
it is a universal rule that in cognition other than perceptual, the cognition of the relation between it and. the object is a necessary
tion of the cognition of the object.
pre-condi-
Thus, in inference cognition of
smoke is the ground of the cognition of fire, only because it is preceded

or accompanied by the cognition of the necessary relacin between the
probans and the probandum, as for example smoke and fire. In verbal
cognition also where a word denotes a fact which is its meaning the
cognition of the relation between word and meaning is the necessary
antecedent condition of the same.
Reasoning is considered to be a
species of cognition, which is different

cognition.
and distinct from perceptual
Accordingly, in conformity with the rule formulated above
it can eventuate in the cognition of its object, namely, universal concomitance, provided the relation between reasoning on the one hand
and the universal concomitance on the other is cognised before.
But
what will be the organ of such cognition ? For aught we know it cannot be perception, since it is rigidly delimited to present daca and universal concomitance is ex hypothesi thought to extend beyond all limitations of time and space. 1
Nor can the knowledge of this relation be
secured through the good offices of inference, since the latter invariably
depends upon the knowledge of the universal concomitance, and if the
knowledge of the relation between universal concomitance and reasoning
1 .., kucas tatsambandhapratipattih? na tvat piratyakst, tasya tadavtsayatvt.

SVR, p. 503.
128
INDUCTIVE REASONING
[CHAP.
were to be secured by it, it would presuppose another reasoning, and the

latter again another inference for the cognition of the relation with universal concomitance, and this inference would again require another reasoning and there will be no end of the process/ Nor can it be supposed
that a second reasoning will be of help for the cognition of the relation
between the first reasoning and its object, namely, universal concomitance. The second reasoning will be in the same position as the first
and so it will be as ineffectual as the first being subject to the same
limitation. Nor again can any other organ of knowledge be appealed
to for the resolution of the impasse, since being different from
perceptual cognition, it too will presuppose the cognition of the
relation between itself and its object.2
The Jaina replies to this objection by denying the reality of the
problem based upon the universal rule that non-perceptual cognition
can operate only by being preceded by cognition of the relation between the said cognition and its object. The Jaina thinks that this
universal rule is based upon a hasty generalization and not upon
comprehensive knowledge. The opponent has made a reservation in
the case of perceptual cognition and what is the reason of it? It is
nothing but the recognition of the brute fact that perceptual cognition
takes place without the mediation of any other cognition. The laws
of nature,, whether physical, or psychological, or logical, can be derived only from our knowledge of things and their behaviour. And
this has been the reason why the Naiyyika has been constrained to
exclude perceptual cognition from the province of this law. The same
is the case with reasoning. That it is different from perceptual cognition does not impose a disability upon it, nor again perceptual
cognition because of its difference from non-perceptual cognition is
entitled to a prerogative, which cannot be shared with others 3 . Ic is
merely a question of fact and not of reason that a particular species
of cognition is preceded by cognition of the relation that obtains
Between itself and its object. Reasoning is certainly the organ of
1 npy anumnt, anavasthnusagat.
SVR, p. 503.
2 bid.
3 Ibid.
yin]
129
cognition of universal concomitance and it is undeniable that there

must be a relation between this organ and its object. But as has
been observed, reasoning, that is set into activity by observation and
lion-observation of two particular cases of concomitance, does not
stand in need of the cognition of the relation between itself and its
datum as che antecedent condition, and directly delivers the cognition
of the universal concomitance by its inherent and necessary power.
And so the problem does not arise at all.1
But a really formidable objection has to be encountered. Reasoning can be regarded as an organ of knowledge only if it can -establish
its claim to deliver the knowledge of an objectively existent fact.
It has been claimed that universal concomitance is the object of reasoning. But what can be the nature of universal concomitance?
Does it mean the necessary co-presence of two terms in an identical
area of space, or in a determinate period of time, or in both3? It
will transpire on examination that neither of these alternatives is a
tenable supposition. Firstly, spatial co-presence is not found between
smoke and fire. The volume of smoke is found to wander in the
firmament, whereas fire is" present in a particular pare of the;hill;
There is again supposed to be a necessary connexion between rainfall
and flood. But rainfall has occurred in the upper region and the flood
ts found in the lower region of the river, and so there \ is- no spatial
co-presence between these two events also. Yet it is supposed that
these sets of occurrences are necessarily connected3. As regards temporal co-presence, that also is not a necessary element of the supposed
universal concomitance. Thus there is no temporal concomitance
between flood and rainfall, since there was no flood when there was
rainfall, and there is no rainfall when there is flood. Again the rise of
1 tarkasy 'pi svayogyatvisesastnarthyad cva svavisayapratyayanasiddhir
bhavatu,
SVRp. 503.
2 tarkasya ca visayo vyptaksanih pratijnyatc.vyptis ca vicryamna
'nvasthm badhnti. yatah sdhyasdhanayoh sambandho vyptih. sambandhas
caikadelatvam ekaklatvam v tayor bhavet ?,
ibid., p. 504.
3 Ibid.
i3o
INDUCTIVE REASONING
the star called Krttika

the rise of Robini1.
[CHAP.
is supposed to be the ground of inference o
But they are not synchronous events as the latter
occurs after the former2.

It has been found that the so-called universal concomitance cannot
be determined either as spatial or as temporal co-presence, which are the
only possible ways in which necessary relation can be understood.
Let
us now consider the nature of the terms between which this relation can
be supposed to hold.
Suppose, in the first place, that the universal
concomitance subsists between one universal and another universal.

Or, in the second place, that it subsists between a universal and its
particulars.
Or, in the third place, that it subsists between the par-
ticulars and the universal.
Or, in the fourth place, that it subsists
between particulars and particulars3.
The first alternative is not a
tenable hypothesis, since the concomitance between the universals is

not conducive to the interests of inference.
The object of inference,
that is the major term is an individual and not a universal.
Thus, for
example, the fire which is sought to be inferred on the ground of

smoke is a concrete individual and not the fire-universal. Besides
univcrsals arc regarded as ubiquitous principles, and as such the cognition of the presence of a particular universal in a particular place will
be a superfluity, inasmuch as fire-universal, tor example, exists everywhere, both in the hill and outside the hill.
And so what will be
gained by proving the presence of fire-universal in a particular hill by

inference4?.
The second alternative also does not stand in a better position.
Is
the relation of the universal with the particulars to be supposed to

. i Ktttik. The . pleiades. The third of the 27 lunar mansions or asterisms (consisting of six stars). Rohini. Name of the fourth lunar mansion
(containing five stars) figured by a cart. PSED, pp, 371, 807.
.
2 SVR, p. 505.
....::
3 kim ca kasya ken/yam avinbhvah? kim snanyasya smnyena,
snranyasya visesaih, visesanm smanyena, visesnm visesair v?
"
; . -
"
na prathamapaksah. yato na smanyasya saianyena vyaptir anum^nn-
gam...tasya ca nityatvavibhutvbhym sakaladesaklasambandhitay

ddhasya sadhane siddhasdbyat syt.
Ibid.
suprasi-
yin J
131
obtain between the universal on the one hand and the -particulars on
the other, undetermined by space and time, or definite particulars determined by space and time 1 ? The former supposition does not give any
advantage or a new information, since it is much too well-known that
universals are embodied in their particulars. The inferable predicate,
that is the major term, is a definite individual and a concrete real, and
so the knowledge of the mere possibility of the presence of fire-universal embodied in some unknown individual, does not add to our knowledge, for everybody is aware that a universal is related to its particulars2. But the very fact that a universal can exist independently of
particulars in its disembodied form takes away all the point from this
hypothesis. What che arguer is interested to prove by inference is not
the presence of the pure universal or of the universal embodied in an
indefinite individual, but of a definite particular individual. If in the;
alternative it is supposed that the relation of the universal is with the
definite individuals existing in time and place, which are exclusively
their own, that will be in direct contradiction with facts. The relation
of a universal with a particular individual is as particularistic and con-?
tingent as the particulars are. Thus, there is no necessary concomitance between smoke-universal and the individual fire existing in the
hill, since it is found that the hill-fire is not present in the. kitchen*;
though smoke-universal is found therein3*
The third alternative is in the same position with the second
since the particulars are always contingent to the universal whereas:
the universal is equally present in different particulars. Thus i t is
found that the relation between a probans and a probandum isnot
one between two, universals, nor between a universal and particulars^
-.. 1 dvitiyapakse'pi desaklbhyam anavacchinnair visesaih sara^nyasy*
vinbhavoVacchinnair v?. . - '-.'**"': '
v
SVR, p. 505.
2 yady anavacchinnaih, cad siddhasdhyataiva.
Ibid.
3 atha deJakiVacchmnaih, cacl' nugambhavah. na hi dhmasmnyasya
parvat4ischair gnivtsesatr amiga mo''s ti, tasya tadantarena' pi plcpraideslday
upalambht.
.' * . ". . .
132
INDUCTIVE REASONING
[CHAP.
nor vice versa. The reason is, in the first place, that the object of
inference is always a definite concrete individual and not a universal,
which is ex bypotbesi known to be present at all times and at all
places, and so the proof of this fact will be proof of what is obvious and
undisputed. The second and third alternatives are scouted on the
ground that the relation between a universal and particulars, and
conversely of particulars with a universal, is always vitiated by a contingency, since no particular can either restrict or exhaust the incidence of the universal in it. Besides, the possibility of the existence
of universals, pure and simple, independently of embodiment in a
particular, robs these hypotheses of all their practical value1.
As regards the fourth hypothesis, that the universal concomitance
subsists between two sets f particulars, it will be found on examination to be tantamount to either a counsel of perfection, or a counsel
of despair. Certainly it is impossible to determine the necessary
relation between all individual smokesf and all individual fifes, since
their number is literally infinite and our resources arc hopelessly
unequal to the task2.
Again, the proof of the positive concomitance is supposed to b e
furnished by the truth of the negative concomitance between the
opposites. Thus, for instance, smoke and fire are supposed to benecessarily related as determinate and determinant concomitant, andthe corroboration of this belief is thought to be found in the necessary
concomitance of the negation of the detcrhiinaht concomitant With the
negation of the determinate. The absence of fire is supposed to
entail the absence of smoke. But does the absence of one particular
fire entail the absence f all individual smokes ? This is not the case.
The absence of the fire-in-thc-hill does not entail absence of the smoke
-ia-thc^kitchen, and the fire-in-the-kitchen that is present is numerically different from the fifc-in-thc-hill. So it must be admitted that**
i
A
SVR, p. 505.
yikskjam
rstoa nahtyat.
vi&sair iti tuciyapaks^pi
klhlakam avi
ym]
THE JAINA CONCEPTION OF REASONING ^
133
the negative concomitance between the absence of fire and the absence
of smoke can be held to be valid if the absence of all fires entails the
absence of all smokes. But even if this be possible it is never within:
the range of human knowledge which is circumscribed by definite
limitations. It is therefore a counsel of perfection which can never
be realized, and if it is offered in full consciousness of this impossibility, it turns out to be a counsel of despair1.
Furthermore, the negative concomitance is supposed to be founded
upon the inherent opposition between the absence of the proband urn
and the presence of the probans. Thus, for instance, the absence of
fire is supposed to entail the absence of smoke, because the absence o
fire is supposed to be incompatible with the presence of -smoke or
vice versa. But on scrutiny it will be found that the opposition of
smoke is with the absence of smoke and not with the absence of fire/
Thus, it is found that in the burning charcoal which is completely
dehydrated, there is found absence of smoke and this shows that the
opposition of smoke is with the absence of smoke and not with the
absence of fire, since the absence of absence of fire, that is, the presence of fire, is found to exist in harmony with the absence of smoke2.
Moreover though it is conceded that there is necessary concomitance
between the reddish colour of the fire and smoke,, yet smoke is never:
used as the ground of inference of the colour. Likewise smoke is
considered to be the probans of fire and not the darkness of the colour
of the former. So it is an arbitrary assertion that smoke is the concomitant of fire and not the colour of them- The concept of necessary
and universal concomitance is thus bound to be dismissed asan
arbitrary and capricious figment of the logician's imagination, and thedismissal of universal concomitance as a pious hdpe atid consummation
bound to remain ever unrealized, knocks down a fortiori the "claim of
reasoning as an organ of this chimera3.
The Jaina in r*rj>ty to this formidable array of objections observes
1 SVR, p. 505.
2 tad iha dhmbhva eva sati dhiimasya nivrttir drsyata ici clhmlbhvcnaiv' sya vifodho na- tv gnyabhaveta.
Ibfd,
3 Ibid., pp. 505-6,
i34
INDUCTIVE REASONING
[CHAP.
that tlic alleged difficulties proceed from a perverted misconstruction

of the data of experience, and the failure to distinguish between
essential characteristics and unessential accidents* The relation of
necessary concomitance holds between two terms in their essential
character. That one term is related to the other is due to the fact
that the essential nature of one necessarily involves or presupposes
that of the other1. Thus, for example, the essential nature of smoke
is what is called smokehood, in other words, what makes a smoke
smoke The essential nature of fire is likewise what is called firehood,
in other words, what makes a fire fire. Fire and smoke will not be
themselves minus what is called firehood and smokehood respectively
And the relation f necessary concomitance has its foundation and
basis upon this essential nature of things, and time and space or any
other phenomena that happen to be associated with them are absolutely; irrelevant accidents without the least bearing upon the relation
in question. So the objections based upon temporal and spatial
concomitance are incoherent prattle. The necessary concomitance holds
between fire as such and smoke as such without reference to time and
place, which, though they are the unavoidable setting in which the
individual fire and smoke manifest themselves, are nothing but external determinations2.
As regards the difficulties raised regarding the nature of the terms
as to whether they are particulars or un i versis, these are the creations
of abstractionistic tendency of logical thought. The Jaina does not
believe in the existence of pure universals or pure particulars, which
are the fictions of abstract logic, however necessary and convenient.
they may prove to be for logical thinking 3 .. The dichotomy between
universal, ancj particular is false. And so the objections based on
these abstractions of logic will not cut ice with philosophers, who will
_ j v svarpaprayuktasy 'vyabhicmrasya vyptitvapratijnac.
"K
SVR, p. 506.
2 dcsaklau parihrtya svarpamacrenai va dhmader agnydin sah
'vinbhvasya nirbdhabodhbhirdhatvlt.
/ -
' - -
- -
ibid.
3 na ca kcyalau smnyavi&sau gamyagamakarupatay'nubhyete.
.;
...
vin].
verify their speculations by observation
135
of objective data.
It may
however be asked between whom then the necessary concomitance

can hold good.
The answer is that it holds between two reals, which
are by their very constitution inseparable from each other1.
Reals
are neither universis nor particulars, but constitute a third type,

which is distinct from both of them severally and jointly.
And it is
between two such reals, which may be characterised as concrete and

individuated universals, that the relation of necessary concomitance
can hold good.
As regards the question, ''Between what two reals
does necessary concomitance hold good as a matter of intrinsic necessity ?"
it can be determined by reasoning, which is set into activity
by observation and non-observation of concomitance between two

individual facts, as we have elucidated at the outset of this chapter.
As regards the question whether
negative
concomitance holds
between a particular and a particular, or between one particular and all

particulars, our answer is that it subsists between two reals taken in
their universal reference.
Thus the absence of smoke is concomitant
with the absence of fire as such, that is to say, with the absence of
all individual fires. The relation is cognised between negation of the
whole class of fire and the negation of the whole class of smoke, that
is to say, between the absence of all smokes and the absence of all
fires3.
And in this apprehension of the negative concomitance no
reference to the spatio-temporal setting, in which they may actually

occur, enters as an element.
But it has been asked, how can the
absence of all fires and all smokes be apprehended without apprehension of the negata, that is, - all the individual fires and all the individual smokes?
And the apprehension of all individual smokes and all
individual fires s not possible to ordinary human beings destic&te of

omniscience.
1
The Jaina answers that the matter is not so recondite
tatrayasya yena Vyabhicaras tasya ten 'vinabhvah. ~

.
. .
" SVR, p. 5o6..
2 yad api kim ekasya kasyacid agncr abhvc dhmnupapactir grhyate
sarvasya ve* tyay upanyastam, tatra sarvasye' ti brmah.' yato dhmnupapattih
sarvkscpena prariyate y ay an kaicd dhiimah sa sarvah sarvasya 'gncr abhave'nu papanna iti.
ibid.
13.6
INDUCTIVE REASONING
as it has been made out to be. The difficulty proceeds from the
misconception of the nature of negation. Negation is not an abstract
non-entity, but a positive entity. Thus by the negation of fire we
should not understand an ethereal abstraction having no local habitation of its own. It is, on the contrary, nothing but the positive locus
different from the locus of fire1. Negation of one entity is always
another entity.. Thus the negation of fire is a locale in which all fires
are non est, such for instance as a lake. The apprehension of the lake,
together with the recognition of the fact that fire is not present in it,
is equivalent to the apprehension of the negation of fire. It is not
necessary that the knowledge of all the negata should precede the
knowledge of the negation. That this is the fact is proved in the
apprehension of individual entities, which is a commonplace occurrence.
Let us illustrate this truth by the examination of the apprehension of
a particular entity, say, a jar.. The jar is felt as jar and this means
that it is felt as distinct and different from all the infinite number of
things which are not jars. And this is in entire conformity with the
view propounded before that the reals are neither universals nor particulars in the state of divorce, but constitute a third type which embodies both the elements and at the same time is different froth either.
Thus to know jar is not only to know a positive existent, but also t
know the negative element in it, which is responsible for the distinctive individuality of it.
The contention, therefpre, that.to know a
thing negatively, that is to say, as distinct and different from all that
is other than it, presupposes the knowledge of the infinite number of
existents from which it is different, is an illustration of the absurdities
to which love of abstract logical speculations may lead. The admission
of the necessity of the condition alleged would on the contrary make
#11 our knowledge of individual facts as individuals impossible. W e
have shown how an individual may be known as different from the
infinite number of individuals without the knowledge of the specific
individuals standing in opposition. And the knowledge of the negation of fire is nothing but the knowledge of an entity which is bereft
i yato'guyabhvas taclanyadesdisvabhvo bhvntarasvabhavatvd abhvasya.
SVR, p. 507.
yin]
137
of fire and for this the knowledge of the infinite number of fires is
not necessary1.
As regards the contention that there is no opposition between the
absence of smoke and the presence of fire, but it holds between the
presence of smoke and the absence of smoke, it will suffice to say that
it (the contention) is based upon insufficient observation of data.
one thing is opposed to another
at any rate in all cases.
That
thing cannot be determined a priori,
Without entering into the epistemological
controversy that the opposition of smoke and not-smoke is known

a priori, the condition of the knowledge of opposition can be formulated
in conformity with the testimony of experience as follows.
B is
opposed tb A, if the presence of A invariably entails the absence of B.

Thus the relation of opposition is known from experience to subsist
between two terms when the presence of one is inevitably attended
with the absence of the other 2 .
It is of course observed that if there
be nt-smoke in a plaice, smoke will not be perceived to be present in it.

So smoke and not-smoke are to be understood as mutually opposed.
Bur this is not the only possible case of opposition.
There is J another
fact which is found to stand in opposition to smoke, for instance,

absence of fire. It is invariably observed that if there be absence of
fire, there is absence of smoke.
Smoke cannot possibly exist in a place
where fire is non-existent. So the opposition of smoke is found to hold

riot only with negation of smoke, but also-with negation of fire. Thus
we find that there is not only positive concomitance between smoke
and fire, but also the negative concomitance between negation of fire
and negation of smoke.
The opponent's objection is therefore fiihd
to be absolutely baseless.
:;
As regards the objection that there is necessary concomitance

between smoke and the colour of the fire and still one is not, made the
ground of inference of the other, c will be found to be based upon a
hasty and unreal assumption.
The relation of necessary concomitance
is understood between fire and smoke in respect of their essential

1 SVR,p. 5 p7.
2 ... yatsadbhave yasya niyamena nivrttis tena tad viruddhara cva.
agnyabhave ca sari dhumasya niyamena nivarttamanatvd dhmbhaven cva
ten' pi tasya virodhah. Ibid.
18
138
INDUCTIVE REASONING
[CHAP.
character and not of colour and the like, because the qualities in question are not the essential characteristics of them1. Thus the yellowish
colour of fire is found to be present in gold also; its brilliance in a
flash of lightning; the quality of existence is found in all entities; and
its burning capacity is also found in the sun; its upward motion is
found also in the wind; its capacity to melt ice is found in hot oil.
Thus all these qualities of fire are contingent and variable. Similarly
also the dark colour and other attributes of the smoke are not essential
characteristics. A buffalo may be as dark as the smoke; its irritating
effect on the eyes is also found to belong to oil; its upward rise in the
sky is found to be present in vapour2. So these qualities are not uncommon and peculiar characteristics of smoke or of fire. The necessary
and universal concomitance is determined only by those qualities,
which are known to constitute the essential nature of the things, e.g,
smoke and fire. There need be no difficulty to apprehend what is an
essential and what is an unessential attribute. The essential attribute
is the connotation of a term without the knowledge of which the
entity denoted by the term cannot be understood. The essential
attributes of smoke and fire are therefore those, the knowledge o
which is the pre-condition of the knowledge of smoke as smoke and
fire as fire. Such attributes are called, for want of better and "more
expressive terms, ,'smokehood and firehood. The qualities which as
specified above are found to belong to other substances, can by no
means enter into the constitution of the essential nature of things, and
so the objection based upon the supposition that these qualities are the
terms of necessary concomitance falls to the ground with the collapse
of the foundation on which it is based3.
1 ... yato vyptyanusren'numnam vidhiyate. vyptis ca agnitvadhmatvadvren

aiv'vasiyate. na
paingalydidharmadvrena.
tesm
nantyd
vyabhicrc ca.
'
SVR, p. 507.
1
i b i d .
'
.. '
-' ;
3 ato yen at va rpena visvavivaravarttinyo vahnivyaktayo

dhmavyaktayas
ca samgrhyante tad eva rpam vyaptim niyamena vyavasthpayati. tac c
'gnitvadhmatve muktv" na' nyad bhavicum arhati. na khalu yath vastvantarasdhrainah paingalyadayas tath 'gnitvadhtnatve.
Ibid., p. 508.
. yin]
THE JA-INA CONCEPTION OF REASONING
But it may be asked
139
that i there is necessary concomitance
between smoke and fire, why is it that smoke and fire were not understood to stand in this relation,
perceived for the first time.
not in operation.
when they were severally or jointly
The answer is "that the relevant organ was
The absence of the organ is due to the absence of
the conditions of it.
And we have already set forth the conditions as
observacin and non-observation of co-presence and these conditions were

not certainly present at the time of the first perception.
The failure
of the cognition does not imply the absence of the necessary concomitance, since it is due to the non-emergence of the relevant organ,
namely, reasoning1.
To sum up the Jaina position.
Reasoning is a separate organ of
cognition and its content is universal concomitance.
The conditions
of reasoning are observation and non-observation of the co-occurrence.

of two terms-in one word the Joint Method as called by Mill.
THE NAIYYIKA'S OBJECTION REFUTED

The Naiyayikas hold that reasoning cannot be an organ of knowledge, since it has no object of its own.
The object of reasoning is
held by the Jaina to be universal concomitance.
But it is the object
exclusively of perceptual cognition. .Even the first perceptual cognition^

is competent to take note of the universal and necessary concomitance
of smoke with fire. It is not possible, that there should arise the doubt
whether smoke is the effect of fire, or of any other entity, such as water.
Nor is there any ground for mistaking it to be the effect of anything
other than fire. The perceptual cognition of the co-presence of smoke
and fire is the sufficient warrant that smoke is the necessary concomitant of fire. The repeated observation of concomitance in presence
and in absence only serves to place the finding of first perception on a
footing of security.
The subsequent observation does not add an iota
i nanu yady anayor vyptir asti tarhi prathamadacsanakle kasmn no

'Uikhati 'ti cet, grhkbhvd iti brmah...vyptijnasya hi kranam upalambhanupalambhau. na ca prathatnadarsanakse tau stah.
SVR, p. 50$.
140
INDUCTIVE REASONING
[CHAP.
to the content of the perceptual cognition1. The Naiystyika also offers

an alternative explanation. The organ of cognition of universal concomitance is perception aided by the repeated observation of concomitance in presence and in absence3. It will therefore be no sound
objection to urge why this necessary concomitance is not cognised at
the time of the initial perception. The answer is that the full condition of the perception of necessary concomitance is the Joint Method,
and this cannot possibly come into being at the time of the initial
perception3.. The repeated observation of concomitance in presence
and in absence serves to rebut all doubt about the determinant of necessary concomitance. Even though smoke and fire are found to be copresent, there may legitimately arise the doubt whether the necessary
concomitance of fire is determined by smoke as such or by such qualities of smoke as dark colour, etc. This doubt is resolved by the consideration that the qualities of smoke, such as dark colour, etc., are
I invariably the accompaniment of smoke alone and have no reference to
fire4. That smoke as such, that is to say, smoke as determined by
smokehood, is the necessary concomitant of fire, is definitely ascertained
by the repeated observation of concomitance in presence and in absence.
The cognition of concomitance as necessary and universal' is felt
to be direct and intuitional in character, just like the perceptual
cognition of a man bearing a staff. . The cognition of necessary concomitance is no doubt a judgment, cognisant as it is. of two terms and
1 ... tarkah pram^nam na bhavati nirvisayatvt...tarkasya hi visayah
sdhyasdhanayor vy ptih paraih pratipdyate. s ca praty aksasy aiya visayah,
prathamapratyakse'py agnisambandhitvena dhmasya pratibhsanat. na c 'yam
dhmo vahner anyato'py anyasmd eva'y jlader bhavati *ti sandehayipryasau
:
pratyakse sambhavatah.
*:
SVR, p. 509. '
2 bhyo dars"andarfanVagatnvayavyatirekasahakrtena -v -pratyaksena
vyptih pratiyate.
.. . .
Ibid.
. 3 anvayavyatirekbhym hi bhyodarsandarsanvagatbhym vyptir
uHikhyate. na ca prathamapratyaksakle tau stah.
Ibid.
4 kitn dhmaprayukto'yam niyamah kirn v syamatvidiprayukta iti, tatra
symatvdayo dhmpeks na vahnispeks id dhmasya 'gnisambandhitvc
dhmatvam eva prayojakam na tu symatvdayah. dhmatve hi sati na kadcid
agnitvam vyabhicarati 'ti. Ibid*
yin]
14t
the relation holding between them. But the judgment is as perceptual

as the judgment of the man wich the staff*. It should therefore be
held that the organ of necessary concomitance is the instrument of
perception and not reasoning as advocated by the Jainas.
The Jaina re fly
As regards the objection that reasoning has no object of its own,
since universal concomitance, which is claimed to be the object of
reasoning, falls within the ken of perception, it will be found on enquiry to be devoid of substance.
Perception wtiether sensuous or
mental cannot be competent to get hold of universal concomitance,

which holds not only between present terms, but has reference to past,
future and even remote objects, which can never be supposed to come
within the range of perception2.
Perception is by the very nature of
it necessarily limited to things, which exist at the present moment and

which become related to the sense-organ.
that perception,
or the sense-organ
It is preposterous to suppose
in question,
can come into
relation with what has become defunct and what is yet not in
existence.
It is equally an idle' supposition that the sense-organ,
or its result
sense-perception comes . into a sort of
transcendent
relation.with the infinite number of individuals, once existent and

hot yet existent, through the medium of the universal which is
cognised in the individual object.
We have discussed the possibility
and the character of such perception at the very commencement

of chapter i.
It ought to . be remarked here that the. Naiyayika's
theory of transcendent perception of objects,
even
defunct and
unborn and situated. beyond the ken of sense operation, has been
repudiated by the Mlmmsist the Vedntist, the Buddhist and the
Jaina philosophers;
These philosophers would fain account for these
cognitions by means of inference.
The inference will be like this
1 bhyodrstanvayavyatirekasya pramtur aparokskratayo 'pajyamnatvad

visistadandydipratyayavac pratyaksam eve 'dam vyptijnnam iti.
SVR, p. 509.
2 ...na ca catre* ndriyasya grahanasmarthyam sambhavati. sakaivypyavy^pakavyakcibhih srdhm tasya sambandhsambhavt.
142
INDUCTIVE REASONING
[CHAP,
"The past, future and present unperceived coins are like the coin that
is under observation, because they belong to the self-same class, and
hence possess the same class-character as the present one".
And even
if it is admitted that rhe cognition of one individual carries with it the

cognition of all the individuals belonging to the same class, yet the
knowledge of relation that holds between two heterogeneous individuals
cannot be cognised as necessary and universal in reliance on the verdict
of perception, transcendent or non-transcendent.
Thus the mere per-
ception of a horse in the stable, even though it be repeated for a hundred times, cannot give us the assurance, even by virtue of its transcendent capacity credited to it by the Naiyyikas, that the connexion
of the horse with the stable is necessary and universal.
So something
more is needed for the acquisition of the knowledge called induction.

It must; therefore be admitted that perception, whether ordinary
or. extra-ordinary, is incompetent to give insight into necessary and
universal 'concomitance,
that holds between two classes of facts.
The contention of the Naiyyika cannot therefore be accepted as a

satisfactory explanation of inductive judgment1.
Let us
consider the alternative
explanation
offered by the
Naiyyika, namely, that though pure perception be incompetent to

realise necessary relation, the same aided and reinforced by repeated
observation of the concomitance of individuals of a class in presence and
in absence, is an efficient organ of the same.
But the hypothesis of
reinforcement, however plausible it may appear, cannot be supposed to

have any special bearing upon the problem.
However much an entity
may be assisted by a number of auxiliaries, it cannot be supposed that

the auxiliaries in question will confer upon it a capacity which it does
not possess independently and on its own account.
Thus the visual
organ, though aided by a powerful light, cannot come to acquire the

capacity ior perceiving taste or odour2.
The light can only facilitate
the perception of colour for which the visual organ is competent per se.
So the repeated observation of concomitance in agreement and in
SVR, pp. 510-11.
na khalu pradipasahakrtam locanam rasdau pravarttamnam pratiyate.

' ':-\
bid., p. 5 1 1 ,
yni]
THE JAlNA CONCEPTION OF REASONING
143
difference, if viewed as che ally and auxiliary of perception, can be of

help to it only in the realisation of what it is competent for.
But we
have seen that perception is incompetent with regard to things and

their relation, which stand outside its purview.
Nor can the cognition of universal concomitance be regarded as a
sort of mental intuition.
It is a truism that the mind taken by itself
is incompetent to perceive external objects1.
Its jurisdiction is entirely
limited to only such mental phenomena as feeling, willing or knowing.

It must be admitted that observation of mere concomitance is unable
to penecrate into its universality and necessity.
Such knowledge is
. secured only when the assurance that A cannot possibly occur without
B is acquired after elimination, of all doubt about the opposite possibility.
The organ of the elimination of such doubt and the conviction
that the relation between A and B is necessary is nothing but reasoning.
And even if it be supposed to be an aiding ally for the realisa-
tion of the conclusion, when thwarted by doubt, it must be admitted

that it can be of real help if it has validity of its own 2 .
The
, Naiyayika makes reasoning a subordinate ally of perception as the organ

of necessary concomitance.
dogmatism.
But this is only nothing short of perverse
He would deny reasoning the credit it is entitled to and
transfer it to perception,which we have seen to be absolutely incapable

of the task.
The Naiyayika may contend that reasoning is not
regarded by him as a valid instrument of knowledge 3nd this is the

reason why he would accord to it the status of a subordinate ally and
nothing more.
But there is no reason why reasoning should not have
validity of its own.
We have shown that it has a legitimate object
of its own in the shape of universal concomitance and also that its
finding is not contradicted by any other cognition.
The Madhyas have sought to equate it with inference and this is

also the position of a class of Buddhists, who call it prasanganitmna,
that is to say, inference employed for the demonstration of an absurdity.
The Jaina demurs to accept this position.
Inference is possible
only if there is knowledge of necessary concomitance behind it.
vt.
To
1 manaso bahyendriyanirapeksasya bahirarthasksatkarane smarthybha.

SVR, p. 5 n .
2 Ibid., p. 512.
'
.- >
144
INDUCTIVE REASONING
[CHAP.
make.this knowledge of universal concomitance a case of inference will

be a case of petitio principa or infinite regress. An infinite series of
inferences will be necessary for the establishment of one universal
concomitance, if inference be the source of knowledge of it. Each
inference will fall back upon another universal concomitance and thus
upon another inference. To get rid of chis vicious infinite series it is
necessary that there must be an independent source of knowledge of
universal concomitance which is the necessary condition of inference
in its turn. It seems that Vysatirtha in the Tarkatndava makes
reasoning the organ of necessary concomitance in the first instance,
which is the literal meaning of the definition propounded by him1.
Of course the finding of a reasoning can serve as the ground of inference, but this is only a secondary and derivative function of it. The
criticism of the Jaina will also hold against the position of the European logicians who seek to make out induction to be a case of inferer
ince. - This point will be presently elaborated in connexion with the
examination o the Buddhist position on reasoning.
-:
"
THE BUDDHIST POSITION EXAMINED
The content of reasoning is necessary concomitance and the modas

operand! is the conviction of the absurdity of the opposite possibility.
Both these findings are of the nature of judgments. Judgment is
necessarily concerned with two concepts, which are synthesised by it.
Concepts as such are tabooed in the Buddhist scheme of epistemology
in which simple intuition, absolutely untouched by concepts, is the
only organ competent to envisage reality as it is in itself. It will not
be to our purpose to enter in this context into an elaborate examination
of the soundness or otherwise of this extremist view. The Jafnas
along with the other realistic schools of philosophers, such as the
Naiyayikas, the Mlmmsists, particularly of the school of Prabhkara,
and others, do not see eye to eye with the Buddhist philosophers and
arc positively of the opinion that there is no ground a priori or
a posteriori for the condemnation of conceptual thought as inherently a
vicious construction pf human intellect.
;
i
Cf. supra p. 1 0 9 .
vm]
145
The Jinas would only observe in this connection that the Buddhist
has not been entirely consistent in his denunciation of conceptual
thought, since he has been constrained to accept the validity of inference as an organ of knowledge 1 .
The conditions of inference, that is
to say, the premises and the conclusion registered by them, are judgments and hence conceptual in character.
In spite of this fundamen-
tal weakness, in; the judgment of the. Buddhist, inference is still

regarded as an organ of knowledge, because of its verification and
agreement with reality.
If this be the criterion of validity,, then we
shall show that reasoning fully satisfies this criterion and as such is
fully entitled to the status, which is accorded, to inference by the
Buddhist. ;The Buddhist has however, contended that the validity of
inference is only vicarious and derivative, being ultimately traceable
to intuition : from which it emerges.
But this seems to be an attempt
at window-dressing inspired by the consciousness of weakness.
It
ought I to suffice for our purpose to observe, whether its yalidityjs

borrowed or 'simulated,' that it is conceptual thought;: which makes
-simple, indeterminate intuition, if possible at all, intelligible and
capable of being turned to account. Intuition uninterpreted by conceptual thought is as good as non-existent.
One cannot think, much less
.assert, that the intuition^ in question is of A, or B, or not of any .thing

.else,since.this involves conceptualthought,/ Besides- all philosophical
'discussion will b.e robbed of its valu, if*conceptual thought is damned
to be false on jtsôwn. ; account* because , philosophical; discussion is
capable of being conducted ^by means; of language, which, js: suspect
in the Buddhist eyes, dealing
.alone,. The Jaina would
as it does in concepts.' and concepts
however be:\. satisfied, if his arguments
an^compl the Buddhist to_give : reasoning the same ^status^ with

înference. <-.'; -..;\;:;": ^;.-;
,rv.'/.....-; :;,...
.'.:
. .: . ': .[.::
The Buddhist ..however has contended that even if reasoning be

regarded as competent to take stock of necessary concomitance, yet it
will not possess validity of its own, inasmuch as reasoning necessarily
brings up the rear of the cognition of the necessary concomitance,
1 savikalpakajlnarpatve'py anumnasy*eva tarkasya*pi pratnnatvnatiptat.
SVR, p. 513.
l
secured by some other organ, and hence is bound to repeat the finding
of the latter. Knowledge, to be valid must be not only in conformity
with reality, but also new and original (agrhUagthi). The Jaina
however does not accept the charge of the Buddhist that reasoning
makes known what is known by something else. It has been found in
the course of examination of the Naiyyika's position that no other
organ of knowledge, except reasoning is competent to take stock of
necessary concomitance between two classes of facts in their universal
reference1.
Let us however examine the Buddhist position critically and dispassionately. The Buddhist maintains that necessary concomitance
caii be conceived to subsist between a cause and an effect (tadutpatti)
and between facts, which are bound by identity of constitution and
-nature (tdtmya). As regards non-cognition it is nothing but a case of
constitutional identity and so it need not be considered separately
from it2. Causality is a ground of necessary concomitance and so the
determination of necessary concomitance can be achieved by the determination of causality. Let us consider die conditions of the knowledge of causality. It (causality) is determined by the observation of
concomitance in agreement and of concomitance in difference, which
has been analysed by the Buddhist into five cases of cognition, two
positive and three negative. In the first instance, there is non-apprehension of the effect, say smoke, in a particular place. This is called
non-cognition number one. In the second place, there is cognition of
fire and cognition of smoke. These are two positive cognitions. In
the third place, there occurs non-cognition of fire accompanied with
non-cognition of smoke. There are two non-cognitions in this case.
Thus in all there are five cognitions, viz. the initial non-cognition, the two subsequent cognitions, and finally the two non-
' "" 1iri atha grhitagrhitvam tad abhidhiyate. tad apy ayuktam. yatah sadhyas dhanayoh sihastyena vyptis tarkasya visayo, na caita tpratipattau pramnntaram prabhavatL
y
SVR, p. 513.
1 (
' ' 2 * riipalabdheh punah svabhvahetv ntarbhvn na tatra prthagavinbhavagrhakaprarnnacinte'ti.

Ibid., p. 514. -
Non-cognition is also a species of cognition 1 .
cognitions.
This
fivefold cognition at once gives us to understand that the two events

are causally related-the antecedent being called the cause and the
consequent the effect.
Now the knowledge of the causal relation
serves as the guarantor of the necessary concomitance between the

cause and the effect.
The effect is always determined by the cause.
If t were not so determined, then the effect could be independent of

the cause and hence could possibly occur without the antecedent
occurrence of the cause.
But this is simply absurd, since only what is
eternally existent or eternally non-existent can be thought to be independent of a cause.
The effect being a determinate event cannot be
supposed to be eternally existent or eternally non-existent on pain o

obvious self-contradiction2.
'-
As regards essential identity, it becomes the ground of necessary

concomitance, because the opposite possibility of two facts, bound by
essential identity of nature, occurring independently of-one another,
is contradicted by inference
(vipaksabdhaknumntJiTocltean
example from geometry, the- fact that the sum of three angles f #
triangle is together equal to two right angles is proved by necessary
inference and the possibility of the sum of the three angles being
more or less than two right angles is ruled out by inference alike.
The
Buddhist example of essential identity may be cited as the relation of

existence: with momentariness?.
All existents -arc inferred; tovjbe-
momentary on the ground of their being existent facts.
Existence
involves momentariness in the same way as the three sides of a- rian^

gle involve three angles.
That existence is necessarily concomitant
with momentariness is proved by the consideration that the conditions

of causal efficiency; which is the sole criterion of existence, are not
i krye hetau tvad yypteh pratipattih pratyaksnupalambhapacakj
jyate...anupalambho'pi pratyaksavisesa eva laksayitavyah...yad -ha (Dhrma:
y -''
: ^ / J
' " ' _ '
;f
\ -:' "::>
: i r . dhumdhirvahnivijanam dhmajnam adhis tayoh//

pratyaksnupalambhbhym iti pacavir anvayah//4
Quoted from the Pramanavarttika in die SVR, p. 514.
^;
'F . ': <S
3 svabhavhetau tv avinabhavaprattir vpak-se bdhakaniimnat. ytH

ksanikatvanumnam sattvahetau vypaknupalambhac. Ibid,
i48
INDUCTIVE REASONING
[CHAP.
possible- in the non-momentary. Now these conditions are succession

or non-succession. -Succession is the occurrence of an entity in two
different moments and the non-successon is the opposite.- These two
opposite characteristics cannot possibly exist in one and the same thing
without splitting up-its. identity into two. ;The non-momentary is
however supposed to be an identity in itself, and as such cannot be
susceptible of succession, which involves two opposite characteristics,,
determined by its relation to two different time divisions, one-past,
another, present, or one present another future. . The conditions of
exercise of causal efficiency, namely, succession or non-succession, being
proved, to be incompatible with the non-momentary by inference, it
follows that only what is momentary can.be existent. .The necessary
concomitance of existence with momentariness is thus proved by
inference of the impossibility of the non-momentary being existent1.
Thus .reasoning has no scope of its own, so far as the determination,
;?
pf necessary concoiitnce is concerned.
, ;...:.,.
v
: .\^
nr. :Theelina ; fin'reply observes'that the contention; of the Buddhistthar a series of cognitions and nn-cognitions gives out the . assurance
that the,observed data are causally related-is acceptable ia so far H as. it;
does hotbutstrip the, data in question. The cognitions and noneogriitioris are-nothing but perceptions, and as such are subject, to:
the.? limitations:_,.o,:,the present time and place. : There is always the!
questionioE here and now in case of perceptual cognition; Perceptual
cognition: can never give us any dependable information about things^
which are putside the context.of here and now,; in other words,, about!
things,- which are past, future and beyond the spatial context2. Thus
something more than perceptual cognition is required to give insight
into -the -causal relation^ 'which : by : the very ; nature of it hojds;
between two classes of events irrespective of limitations of space and
time..: In otherwords,; causality is a necessary and universal relation..
It has been contdec that: the very knowledge of causality affordedby the fivefold cognition carries .with it the unimpeachable^ assurance
1 Vide The Buddhist Philosophy Of Universal Flux, chapters i-iv.
2 . upalambhriupalambhasvabhvasya dvividhasy'pi pratyak<sasya sannihitamtravisayatay desdivyavahitasamastapadrthagqcaratvygt. * .
yin]
149
that it holds not only between the observed individuals,
but'extends
to all the individuals of the classes concerned. . It cannot be conceived

that smoke is an effect of fire and will yet exist without fire, since thi
will be equivalent to the repudiation of the causal relation, between
them1*
'.
: '
The Jaina replies, "True, but indeterminate sense-intuition, which

has been denied by the Buddhist of ratiocinative and discursive activity, cannot be supposed tobe competent for this elaborate process
of ratiocination." Nor again can it be supposed that the conceptual
thought,, which immediately
arises in . the trail of, indeterminate
sense-intuition and interprets the content, of the jatter in teems of

categories, can be equal to this undertaking.. Because, such conceptual
interpretative thought only clarifies the contents of intuition and so
cannot transcend its limitations2.
If,. however, some other, conceptual
thought is postulated.for the comprehension of causality injts universal

reference, then this . would hot repeat the finding Neither of -indeter*
minate sense-intuition, or of; the same; as interpreted! by 1 categorising
conceptual: thinking.: It has, for
its content; something, which is
beyond the capacity 1 of : the;: aforesaid., psychical o processes. ? i t -must

therefore be regarded;as an original ..organ of knowledge.
And this
organ is ho other than what' we call, reasoning, i > . - :
1 Âslot.
î I
tlie/ple;that in die case;of facts relatfed:by essential idem
tity the necessary concomitance between thni is understood by means

of inference of theimpossibility of ^ the j opposite - alternative, I it only
betrays slipshod fhinkingj and; hasty ^f'judgnientK.' .The fnfercnce;o
the absurdity of the opposite alternative, which js made by the Bud?
dhist the organ of necessary concomitance of facts related by essential
identity, is itself possible,
i
if there is the knowledge of another
tath ca Kirttin (Dharmakirttin) kirttitam Vrttike (Pramanavrttike)

kryam dhmo hutabhujah kryadharmnuvrttitah/^
s.a bhavam stadabhve'pi hetumattm vilaghayet/^
Quoted from the Pramanavrttika in the SVR, p. 515.
2 pratyaksaprsthabhvino vikalpasy'pi pratyaksagrhitamatrdhyavasayi-'
tvc sarvopasamhrena vyptigrhakatvbhvah.
Ibid,, pp. 514-15.
3 /w.,p.5151
. . : , ;_.,.__.v ,
i5o
INDUCTIVE REASONING
universal concomitance at its back.
[CHAP.
But what will be the organ of
this universal concomitance ? If you make inference the organ of it,

the result will be a vicious infinite series, as has already been shown
by us in connection with the examination of the Madhva position.
The Jaina therefore concludes that there is no way out possible for
the justification of inference without postulating an independent and
separate organ for the knowledge of universal
concomitance.
And
reasoning is this organ.

We have shown that reasoning possesses the criterion of originality.
It again has the further criterion of agreement and verification.
two are the tests of an organ of knowledge.
slightest
reasoning.
justification
These
Thus there is not the
for repudiating the independent validity of
If the finding of reasoning, which we have shown: t o b e
universal and necessary concomitance, be regarded as problematic and

subject to doubt, this will be laying the axe at the very root of inference.
In conclusion we may point out that reasoning possesses another
trait, which may be regarded as the third, criterion of truth; namely,

its capacity to rebut doubt and misconception.
We. have taken
elaborate pains to prove in this dissertation that reasoning

and successfully rebuts doubt of the opposite alternative1.
effectively
The Jaina
logician insists that this should be regarded-as the final clincher of the
issue, andi inv this? he. has* the support of rldhara, Vyornasivcrya* and
the whole school of Mdhva philosophers, who have unanimously
protested against the absurdity involved in the denial of validity to
reasoning;:and in having at the same times recourse to it as the instrument for the rebuttal of doubt.
,t
Yide chapter i.
CHAPTER IX
SECTION I
Udayana in his Atmatattvaviveka has classified reasoning under

five heads, viz. (i) self-dependence (atmasraya); (2) mutual dependence (anyonysraya); (3) vicious circle (cakraka)\ (4) the vicious
infinite (anavastb); (5) reduco ad absurdem {artistaprasangaliterally emergence of an undesirable issue)1. Varadarja has endorsed
this fivefold classification in the Tarkikaraksa and quoted Udayana's
ifissima verba in his support2. Visvanatha also in his gloss on the
Nyyastra (I. i. 40) has dittoed this classification in deference to
Udayana. He has however mentioned other varieties of reasoning as
propounded by older logicians, though he has repudiated them as
getiuine cases of reasoning. These cases are: (1) initial presentation
(prathamopasthitatva); (2) conventional or empirical induction(utsarga);
(3) absence of decisive proof i. e. of crucial evidence (yinigamamraha); (4) simplicity (laghava)\ (5) complexity (gaurava). Visvanatha
is not prepared to accept them as orthodox instances of reasoning on
the ground that they are not calculated to. enforce an undesirable
consequence which is the essential character of reasoning. They have
however been set forth s instances of reasoning on the ground of
their serving as auxiliaries to an accredited organ of knowledge.
Genuine orthodox cases of reasoning also operate as auxiliary factors
to organs of knowledge and it is this participation in this functional
characteristic in common with the accredited cases of reasoning,>which
is responsible for their subsumption under the genus of reasoning. It
is apparent that they can at most be regarded as cases of reasoning
1 sa c'tmasrayetaretatscayacakralunavasthnistaprasangabhedena pacavidho'pi...
' "
.
ATWp.863.
2 sa c 'tmsrayctaretar^rayacakrakasraynavasthnistaprasangabheden
pacavidha iti.
TR, p. .186.
152
INDUCTIVE REASONING
[CHAP.
only n a secondary sense, made possible by the usual device of extension of meaning on the basis of partial analogy.
But reasoning (tarka) has been used as the symbol of the number
six and Udayana too is said to have accepted this usage 1 , although
he has only spoken of the five species of reasoning.
The expression
sixfold reasoning (sattarki) which is widely current in philosophical

parlance is alsa an indication of the sixfold classification of tarka. The
-sixth variety has been given by Sriharsa in his
Khandanakhandakhadya
after enumeration of the aforesaid four instances and it is called

recrimination or tu quoqm (f ratibandl)2;
T h e fifth variety is however
called v.ydghata (contradiction),
. .
_.
W e do not meet with the slightest suggestion of classification of
"reasoning'in the Nyayastra,
Bhasya, Vdrttika, and the
Tdtparyattkd.
Udayan seems to speak of the fivefold classification for the first, time
in Uietmatattvaviveka*.
"
'r "
iz rnSrlharsa who is posterior to Udayana at least by a century or. tv/o

has referred to the sixfold
classification of reasoning in 'die'Kbandana-
4l
r:hbdHdkhdya .-^.nLliis is not an original formulation of Siiharsa, but

only a restatement of the position of an adherent of the Nyaya" school
who must remain anonymous for want of data available to us. It is
obvious from the enumeration and definition,of t h e six classes of reasoning that these have not been borrowed from Udayana's works. It is
further worthy of remark that reductio ad absurdum^ has been mentioned as the fifth variety of reasoning by Udayana 5 . j But if w e ' look
closely into the situation, it will transpire that the enumeratioiv of
reductio ad absurdttm as an additional variety of reasoning by Udyan'Crya is not logically justifiable.
1
W e have shown time and again ; in
the course of - our dissertation that

a*
[lili*
ark,2tnbar3nkaprmitesv :atitsu
emergence of ah undesirable
sakntatah/
' .-> ... v
vajrses Odayana cakre subodham Laksanavallmj /.

r ili-
" ''.* * > e e L a k s a n v a l i , p . 1 3 .
. ; :
api ctmsrayo 'nyonysrayas cakrakam vyghto'navasth pratibndi
ce?ty. apdyair bhidyamana sattark'syate. KhKh, p. 1291.
4 KhKh, p. 1291,
5 f A T V r p . 863.
ix]
153
consequence, which is known as redttctio ad absurdum% is rather the

general character of reasoning as such and ought not to be recognised,'
regard being had to logical consistency, as a specific variety of it.
All
cases of reasoning are cases of reductio adI absurd urn wichout exception.
Malllntha, the commentator on the Trkikaraks, has also expressed
the view that the recognition of redttctio ad absurdem as a separate
species of reasoning is unsupportable by logic 1 .
-Srlharsa has therefore'
stated the first four varieties enunciated by Udayana and completed the
list by adding two other varieties, namely, contradiction (vyghata) and
tu quo que argument (pratibandf)2.
Sriharsa's classification has been
accepted in toto by Vamsidhara Misra in his
Smkhyatattuavibhakara^
a commentary on the Smkhyatattvakaumttdl*.
* I
Both hagirath Thakkura and Raghuntha Siromani, the commentators on the tmatattvaviveka
have also remarked that reductio ad
absurdum is not a particular variety of reasoning, but the general

character of it, which appertains to all the: recognised varieties, ~ The
specific mention of it as a separate instance is inspired by the desire to
lay emphasis on its supreme importance.
It is a popular differentia-
tion, quite n a par with the differentiation of cattle and blcki

One generally speaks that a person has a hundred head of cattle and
twenty bullocks.
This is a loose
classification and can be justified
only for the sake of emphasis, which is ordinarily known

danyaya,
the maxim of cattle and bullock.
s gobativar*
This shows that the
classification of reasoning with reference to reductio d
absurdatii
rather loose and popular and is not logically necessary4.
vv
But such differentiation is also endorsed by convention even in theproceedings of legal and academic institutions.
Thus the expression
'the Vice-chancellor and the Syndicate', or 'the King and the country*
are approved forms of expression, although the Vice-chancellor is an
1 See Mallintha's commentary on TR, p. 187: ......katham anistaprasangasy'eha tarkvntarabhedoktih?

2
KhKh, p. 1291.
3 S T V . p . 73.
tmsrayadnm apy anistaprasagarpatay govrsanyyd vibhgah.
:_ 4
ATVP, p. 865, See also ATVD, p. 866.
20
154
INDUCTIVE REASONING
[CHAP.
essential part and parcel of the Syndicate, and the king is an essential
member of the nation, though the most important and prominent.
It
is for the sake of emphasising the paramount importance of a particular factor that the propriety of such differentiation is not called in
question.
The separate mention of the redttctio ad absurdum
is
equally a case of emphasis inspired by the recognition of the supreme

role played by it.
Sakara Misra in the Vadivinoda,
a work on practical logic, has
given almost the same classification of reasoning as Udayana.
But he
vhas very wisely and judiciously named the fifth variety emergence of
an absurdity other than that involved in the aforesaid cases (taditaranistaprasahga)1.
This saves the classification from the charge of loose
talk and makes this fivefold classification entirely self-sufficient and

self-consistent.
It is remarkable that Venkatantha, a celebrated exponent of
Rmnuja's system of thought, has followed Udayana in his classification of reasoning and has only sought to save the position by calling
it a pure reductio ad absurdum (kevalnistprasanga)2'.
But this emen-
dated nomenclature does not make an improvement and is exposed to

the same objection.
In the Prajnparitrna, a work of the Rmnuja
school, the pure reductio ad absurdum has beefi divided into two subvarieties, called opposition (y trod ha) and impossibility {asambhavf.
Venkatantha
varieties,
in
namely,
the
Nyayaparisaddhi
(1) tit
quoqtte
has
added three other
(pratibandt);
(2)
equalisation
(samavacana); and (3) double noose (ttbbayatahspsy.
Srnivsa, the
commentator of
two
the Nyayapariwddbi,
has
added
further
varieties, namely, complexity (gaurava) and simplicity (Jaghava)5.

.. 1 y v , p. 19.
2 NP, p. 347.
3 Prajnparitrne tu kevalnistaprasangam eva dvidhkrtya sodha tarka
uktah.
tmsrayanam anyonysrayanam cakrakam tath/
anavasth virodhas csambhavas ce'ty ami budhair iti//
quoted in the NP, p. 347.
4 etesm eva prakrabhedt pratibandisamavacanobhayatahspsdyah
prasangabheda drastavyh.
Ibid.
5 See ftinivsa's commentary on IMP, pp. 347-48: adtpadena gauravalghavatarkayor grahanaai....
ix]
Sriharsa
has
shown
five
other
addition
to the six noticed above.
decisive
proof
conventional
i.e. of
or
varieties
These
155
of
are:
reasoning in
( i ) absence of
crucial evidence (vinigamanaviraha) ;
empirical
induction
(utsarga);
(3)
(2)
complexity
(kalpanagatirava)\ (4) simplicity (kalpanalaghava); (5) impertinence

(anaudtya)
opinion
alias
that
impudence (vaiyatya).
these additional
He is definitely of
varieties should
the
be accorded co-
ordinate status with self-dependence and the rest, because they participate in the general character of reasoning viz. reductio ad
absurdum,
and also because they cannot be included under any of these recognised varieties by reason of difference of contents.
Thus Sriharsa states
eleven varieties of reasoning . Sankara Misra in his commentary on the

Kbandanakhandakbdya regards these additional cases as simulations
of reasoning and not as orthodox instances2.
But in the
Vdivinoda
he has definitely pronounced his opinion that though there is divergence of views regarding
these
varieties, as to whether they are
simulations or genuine cases of reasoning, it is proper that these

should be regarded as legitimate varieties of reasoning, because they
cannot be included under any one of the accepted varieties, and particularly because simplicity etc. have been employed in standard works
as cases of reasoning3.
Thus his difference from Sriharsa is more or
less a make-believe, since he dittoes the classification given by Sriharsa
almost totidem verbis.

We now propose to consider the nature of these varieties of reasoning by means of the definitions recorded by Sriharsa;
Histori-
cally speaking we are not aware of any work of the Nyya school in
between Udayana and Sriharsa, in which
ning and the definitions of the
the classification of reaso-
different varieties can be traced.
1 apare'pi visayabhedt tarkabhed tmsraydivan mantum ucith. tad

yath, avinig;amah, utsargah, kalpangauravalghave c'naucityam ce'ti.
KhKh, p. 1319.
2 See Sankara Misra's commentary on KhKh, p. 718 (ed. Mohanll Vedantcrya and R* Bhgavatcrya): tarkapratirpakatven* bhimatanm utsargdinm khandanam abhidhtum svarpam darsayati.
3...avinigamdayas tarkapratirpak ity anye. tark eve'ty apare, yukta
caitat, anyatrntarbhvayttum asakyatvt.
VV, p. 37.
156
INDUCTIVE REASONING
[CHAP.
Udayana however has barely specified the names, and though the
nature of these varieties of reasoning is constructible from his treatment, Udayana has not propounded formal definitions1.
,.:
'
SECTION II
DEFINITIONS OF THE DIFFERENT VARIETIES OF REASONING

"'. .
(0
Self-dependence (atmasraya)
It consists in the situation when A directly and immediately

presupposes its own self in respect of genesis, subsistence, and cognition 2 .
The adverbs directly' and 'immediately* are significant and
intended to preclude the extension of the definition of self-dependence

to mutual dependence and vicious circle.
In these two latter varieties
also,A presupposes A , that is, its own self, but only mediately and
indirectly.
The criterion of self-dependence therefore lies not in the
presupposition of its own self solely and wholly, but only when the
presupposition is direct and immediate.
dependence there is only one term,
In other words, in self-
whereas in mutual dependence
there are two, and in vicious circle more than two.

Let us now give an instance of self-dependence in respect of
genesis.
If, for instance, it were contended that A was the cause of
A, that s, its own self, then A would be split up into two entities
which are not identical, since causal relation cannot centre round one
term, but requires at least two terms.
The instance of self-depen-
dence in respect of subsistence may be propounded as follows:

suppose it were contended that the cup existed in its own self, in
other words, it were its own container and content; then the cup
"would be split up into two entities, since the container and the content cannot be one identical substance.
The case of self-dependence in respect of cognition would arise if
it were Contended that A for its cognition presupposed the cognition
of its own self, that is, the cognition o A itself.
The cognition
could never take place, because the same thing (that is to say,
x ATV,p. 863.
}.%. svasy'vyav.ahitasvpeksanamatmsrayah.
KhKh, p. 1291.
ix]
157
the cognition of A) cannot be the condition presupposed and the

result conditioned by it. The two must be numerically different;
Self-dependence in respect of cognition does not at bottom differ from
self-dependence in respect of genesis. If we substitute 'cognition*
for * the object* it will be found that cognitional self-dependence is
but a case of originational self-dependence and the two cases are
not materially different. To say that A depends for its cognition on
this very cognition of A, is tantamount to saying that the cognition
of A depends for its genesis upon its own self, that is, the self-same
cognition of A. This also holds good of mutual dependence, because
the dependence of the object for its cognition upon the self-same
cognition of it, is nothing but a case of the dependence of the. cognition for its genesis upon the self-identical cognition. So the cases of
self-dependence in respect of genesis and in respect of cognition are
at bottom a case of self-dependence in respect of genesis.
. I
(2) Mutual dependence (anyonysraya)
:;
This arises when between two terms, the first term presupposes the
second term and vice versa
immediately and directly,
its genesis, subsistence, or its cognition1.
either for
Thus if A depended upon
B and B in its turn depended upon A for its genesis or subsistence, and
A depended in order to be cognised upon the cognition of B, and B
again in order to be cognised were found to depend upon the cognition of A, the result would be a case of mutual dependence in respect
of genesis, subsistence and cognition respectively.
(3) Vicious circle (cakraka): .
.
,
When one self-identical term requires its ownself for genesis,

subsistence' and
cognition
through
or more terms, t gives rise to the

This
also
arises
when
between two
the
intervention
of
two
argument in vicious circle2*

terms, the
first
requires
the second, and the second requires the first through the intervention
1
anyonyasyVyavahitnyonyapeksitvatn anyonysrayah.
'''
"
KhKh,
p.
129T.
antaritasya tad eva dvayam tmasrayo'nyoyasrayas cakrakam.

Ibid., pp. 1291-92.
158
INDUCTIVE REASONING
of the third or the fourth.
The fundamental character of these three
types of reasoning is that the same term is required to serve as the

condition and the result,
and their difference lies in the number of
steps involved in between.
When the terms are two and each is
required by the other it is called mutual dependence, and when the

number of terms is three or more, it is called argument in circle.
It
must be borne in mind that the argument in circle arises only when
the number of terms is limited.
clusion comes back
to
the
In all these cases however the constarting-point.
When however the
number of terms is infinite, it will be a case of infinite regress as we

shall show immediately after treatment of this type.
these three different types as follows: when A
We can illustrate
requires A immediat-
ely and directly it becomes a case of self-dependence.
When A re-
quires B immediately and directly, and B requires A immediately and

directly, the argument in question will be a case of mutual dependence.
When A requires B and B requires C and C requires A
argument becomes a case of vicious circle.
The chaimof
the
intervening
terms may be extended indefinitely, that is to say, D , E, F, etc. and

all these cases will be instances of circular reasoning, provided we stop
at a limit.
In other words, in circular reasoning there must be a first
.term and also a last term.

Let us now give one concrete illustration.
The Vaisesika
pher defines the self as the substratum of cognition.
philoso-
This definition
if subjected to scrutiny transpires to be a*n example of an argument in

vicious circle.
In reply to the question, "What is a cognition?" the
Vaisesika would say "A cognition is that which possesses the attribute
cognitionhood".
But the answer is not obvious in its meaning and
requires further clarification.
'Cognitionhood' is not a self-evident
concept. Naturally the question arises, What is cognitionhood? The

Vaisesika gives this reply: it is a universal which subsists in a kind
of specific quality of the self, and does not subsist in pleasure, pain,
will and so on 1 .
But this also is not sufficiently clear.
The opponent
1 According to the Vaisesika the self is possessed of eight specific qualities,

namely, cognition, pleasure, pain, desire, volition, aversion> merit, and demerit.
Thus cognitionhood is the common attribute which exists in all cases of cognition and does not appertain to the remaining seven specific qualities.
ix]
159:
may ask, What is a universal? And the Vaisesika's answer is that the
universal is nothing but the uncommon condition of synthetic cognition, which makes possible the classification of a number of individuals under one genus.
Now, let us stop at this final answer and consider the logical value
it possesses. The Vaisesika started with the assertion that the self
was the substratum of cognition, and when .questioned about the
meaning of cognition, he was constrained to make a series of explanatory statements, the last of which involves synthetic cognition as an
element. But to a man who wants to be enlightened as to what cognition consists in, the answer, which makes cognition an element in it*
cannot be expected to throw any light on the problem. The Vaisesika's answer only seeks to explain cognition by means of cognition itself
only after making a series of intermediate statements, which would
rather serve the purpose of drawing'a red-herring across the track of
enquiry.
That the argument is a case of vicious circle will be apparent from
the following consideration. Well, cognition is asserted as the proof
of the universal; the universal is asserted as the proof of the particular
kind of universal, cognitionhood; the cognitionhood is asserted as proof
of cognition. Thus cognition is said to be cognition itself only through a number of intermediate assertions1.
This also may be cited as an instance of mutual dependence. If
cognition is proved, cognitionhood can be established, and cognitionhood if established can establish cognition. Here there are two terms
and so it is a case of mutual dependence. It may also illustrate seit
dependence, because the cognition is proved by cognition itself2.
(4)
The vicious infinite (anavastha)
The continuity of the chain of probans and probandum, of ground

and consequent without limit constitutes the vicious infinite3. If
: i jnnclhikaran^clilaksananirpanadvrena cakrakdy-patteh.
KhKh, p. 236.
See also nandaprna's commentary on KhKh, pp. 237-39.
2 See Ananda^purna's commentary on KhKh, pp. 238-39.
3 upapdyopapdakapravho 'navadhir anavastha.
KhKh, p. 1292.
i&>
INDUCTIVE REASONING
[CHAP.
this continuity be endorsed by recognised canons of proof, it will cease

to be vicious. In order to distinguish the latter from the infinite
series that is vicious, we propose to call it legitimate or valid infinite.
Srlharsa has spoken of two infinite series: one regressive and the other
progressive." The former is called regressus ad infinitttm and the
latter may be dubbed as progresses ad infinitum1.
Let us now elucidate it By rheans of the following illustrations. It is contended by
some philosophers that a thing can be known by means of a cognitive
instrument, which can be effective only if it is previously known.
But this contention is bound to; give rise to the vicious infinite
called regressus ad infinitum.
If it is insisted that the organ of
cognition must be cognised before it can make anything known,
the cognition of the cognitive organ will presuppose another cognitive
organ, because a thing, whether an organ or an object of cognition,
can possibly be known only 'by means of another organ. But the
second organ in the receding series would require a third organ to
make it known, and the third again a fourth and so on to infinity.
If we were to stop at some ultimate organ of cognition, that s to say,
if we were to posit that some organs remained uncognised the result
would be that no cognition would take place, since it was postulated
that, only a cognised organ could make anything known. The result
would be however not only failure of the cognition of the organs,- but
also the failure of the cognition of the matter in hand It.was in
order to explain the cognition of a particular object that the series o
organs and that of their cognitions were postulated. But the admission of an uncognised organ at some limit ends in proving that the;
cognition of the object under consideration cannot materialize. The
series of attempts has ended in smoke and amounts to the confession
of failure to explain how an object can be known. If, on the other
hand, the series of organs and their cognitions, which are postulated
for the justification of the cognition of the object in question, were
pushed back without a halt, the result would be the self-same
t ...tame tarn adhodhvantim anavasthm. caksa te... tame tarn rdhvam
dhvantim anavasthm eaksate.
KhKh, pp. 1294-95.
ix]
incongruity.
161
The series of cognitions required to justify the fact of
the felt cognition of an object, would never be completed and so the

required justification would not be feasible.
This is an instance of
the recessive infinite series designated by us as regressus ad infinitum.

The raison d'etre of infinite
series serving as a reductio ad
absurdum is that it fails to explain a datum, the necessity of which

has impelled its advocate to postulate a series of which there is no last
term, which could justify the original datum.
This is called che
infinite series which militates against the original datum, which is the
starting point of the series1.
Let us now consider an example of the second type of the infinite
series which we have called progressus ad infinitum.
The Naiyyika
believes that things are mutually discrete and distinct and their distinction is constituted by differentiating attributes. The statement of these
differentiating attributes is called definition.
The Naiyyika further
asserts that definition is essential to the understanding of the nature

of a thing.
But this overweening confidence in definition has been re-
garded as an uncritical attempt at cheap and easy evaluation of things

by Nagarjuna, the Vedntists, and by such western thinkers as Bradley.
Let us bring home this situation by means of a well-known
definition propounded
by the
Nyya-Vaisesika school.
defined as a substance possessed of smell.

characteristic of earth.
Earth, is
Smell is the differentiating
It is by means of this attribute that earth is
distinguished from all other substances.
But smell can serve as a
differentiating characteristic only if it can be proved in its turn to be a

differentiated
attribute.
But what can constitute its differentiation?
Certainly only those characteristics which are found only in smell and
not in others.
Thus, if we probe the question more closely, it trans-
pires that smell can operate as a differentiating attribute, only if it is

differentiated in its turn by some characteristics of its own.
So if we
arc not content to suspend our philosophical speculations owing to

intellectual indolence and inertia, which is the case with the average
mass of mankind, we must face the situation that the current definii mlaksatikarim hur anavasthm hi dsanam/;
~
Nyyamajari, p. ai (ed. Srya Nryana ukla, howkhamba)
ai
'
'
i6z
INDUCTIVE REASONING
[CHA>.
tious of standard text-books m philosophy are only devices to lull our

speculative impulse into the slumber of dogmatism.
Smell can act as
the differentiating characteristic of earth only if it is known to be

differentiated by its own characteristics. But these latter characteristics,
again, will fail to differentiate, if they are not again known to be
different.
Thus one definition will necessitate an unending series of
definitions with the result that the original datum with which we
started on our quest will fail to justify itself.
The justification of~a
prima facie definition which we accepted as the original datum, can

be achieved only if there is an ultimate definition, which is self-evident
and thus will not stand in need of another definition.
But the denial
of this possibility makes this an instance of progressive infinite series

which we have called progresses ad infinitem.
But all infinite series are not vicious.
If there be some datum,
primal or subsequent, which is established by independent logical

proofs, then the infinite series that may arise in connection with this
proved datum will not affect the validity of the original proposition,
and so will not be vicious.
For instance, the example .of seed and
sprout is a case of legitimate regresstts ad infinitum.
The seed is the
cause of the sprout and the sprout again is the cause of the seed.
But
the fact that the seed when planted in some favourable soil produces
the sprout, is established by the Joint Method which is the legitimate
canon for ascertaining causal relation.
The same also holds good
regarding the sprout, which eventually produces the seed.
So though
we are compelled to postulate that a seed is the cause of a sprout, and

a sprout again is the cause of a seed, and the chain of these links of
cause and effect is bound to be pursued infinitely in the past, yet the
causal relation between seed and sprout is not invalidated in the least,
because the causal relation between particular seed and a particular
">sprout s independently established by a legitimate organ of knowledge,
îrrespective of the consideration that the causal chain has had an
înfinite past history and will have an infinite future career. The raison
Wetre "of infinite series acting 'as a reductio ad hsardum lies in the fact
that there is mutual dependence between a supposed cause . and an
effect in respect of genesis, subsistence, or cognition.
But trie causal
relation between a seed and a sprout is ascertained independently by
ix]
163
observation of concomitance in agreement and difference without

reference to the past or the future series.
It would have been a vicious
series had the self-same individual seed and the self-same individual
sprout been required to serve as cause and effect alternately,,
But
though the seed is the cause of the sprout, and the sprout again is the
cause of the seed, the two individuals are self-contained, so far as their
causal relationship is concerned.
The individual seed which is the
cause of the individual sprout is not the effect of that sprout, but of a
different individual sprout that occurred in the past.
The pairs of
cause and effect are numerically different and hence there is no mutual
dependence.
The fact that the series is extended over the three divi-
sions of time, the past, present and future, is due to the fact that the
history of the physical world is an uninterrupted course of .events to
which we cannot set an arbitrary limit.
Let us now consider an instance of legitimate progressive infinite
series.
It is a fact that a quality generates a relation a n d a relation
generates a quality.
sweet".
Take for instance, the proposition "Sugar is
Here the quality Sweetness' by virtue of its belonging to
sugar generates a relation between the predicate and the subject.

Sugar can be sweet only if there is a relation between the sugar and
the sweet.
But this relation again necessarily generates a quality,
the quality of standing in a relation to sweetness.
Thus sugar has not
only sweetness as a quality, but also the quality of standing in a

relation to sweetness.
This derivative .quality must again generate a
relation between itself and sugar, because a quality can belong to the
subject, only if it stands in a relation to it.
This latter relation will
again give rise to a further relational quality, namely, the quality of

standing
in that relation.
In this way the chain of qualities and
relations will be generated infinitely.
But this infinite series does not
invalidate the original proposition "Sugar is sweet", because the

quality and relation involved is not in any way dependent in respect
of genesis, subsistence, and cognition upon the progressive series of
qualities and relations that will be generated.
(5)
Tu quoque argument (pratibandj)
It arises when the arguer docs not care to refute the charge of an
164
INDUCTIVE REASONING
[CHAP,
undesire .1 contingency brought home by the opponent against him,

but ascribes the same consequence to the opponent1.
For instance,
suppose that the Vedntic monist proves the unreality of the world,
and the dualist without refuting the Vedantist's argument affirms that
if the world were unreal, Brahman too would be unreal.
This is
nothing but a case of recrimination, because the dualist virtually

admits the validity of the argument of the Vedntist.
The legitimate
course of action would be to expose the fallacies of the opponent's

argument, and the mere allegation of an unacceptable emergency has
. no logical cogency.
A striking example of tu quoque
Nyyavarttika.
argument has been given in the
In order to refute the Buddhist doctrine of flux
Uddyotakara, the author of the Nyayavrttika,
asserts the proposition :
"A positive effect inheres in its material cause'1.
For instance, a piece
of rextile is produced from the yarns and again inheres in them.

It is obvious that the cause must be antecedent to the effect..
But
it is equally obvious that the effect should also inhere in the" causa
materialis.
Thus the material cause must exist in the antecedent
moment and also in the moment in which the effect is produced2.

It follows therefore that the material cause at any rate cannot be
fluxional in the sense of existing for one moment only.
it must exist for two consecutive moments.
In the least
The denial of this would
involve the consequence that a product will have no material as its

support.
An
unsupported- product
is
an
impossibility.
The
Buddhist in reply to this contends that an effect is without a support.

There is no logical necessicy that it must have a support to inhere in 3 .
Uddyotakara rejoins that this
support is an illogical assertion.
sweeping denial of a material
It could be accepted if an instance
of unsupported effect were produced by the Buddhist/ The truth
1 svbhyupagatadosatulyat pratibandi.
KhKh, p. 1292.
2 kranam ca krydhikaranam atas ten'pi kryaklo 'nubhavaniya t .
NV,p. 436.
3 andhram eye*ti cet? atha manyase pratitya kranam kryam utpadyate
ti kva kim varta^ta iti.
rx]
165
of a universal proposition can be brought home, if an example

able to
both
accept-
the proponent and the opponent is demonstrated.
no such example approved by both the parties is made

the Buddhist,
and
the
But
available
by
absence of such an approved example should
be the sufficient condemnation of the contention of the Buddhists 1 .

The
buddhist 'might
rejoin
example cannot be a fault.
that the
absence of
an
T h e Naiyyika also " cannot
approved
put
forward
an approved example of a product,having a material support.

ever might be produced as an instance,
accept it as an instance in point.
so far as the absence
of
the
Buddhist
will
What-
refuse
to
So the position is one of equality,
a commonly accepted
example
is
consi-
dered .
Uddyotakara
observes
that
this
is no argument at all
that the position is the same or similar
is devoid
of
logical
T o say
cogency
inasmuch as it amounts
to the admission of the truth of the charge
brought against him,
this admission
the opponent's
and
is sufficient
point
in
connection with the tu quoqtte
ment, which deserves consideration to make the

Srlharsa in the Khandanakbandakhdya
type of reasoning
is the
parties, regard being
complete.
upshot of
this
had to the fact that their positions are equally

kind of
Sriharsa quotes Kumrila's if se dixit:

open to objections of equal gravity,
tion" 4 .
treatment
thinks that the
argu-
demonstration of the equality of both the
same or similar
value, one of
condemn
position .
There is another
open to the
to
the parties
It is apparent from
should
objection.
In support of
it
" W h e r e both the positions arc
and
their solution is of
equal
not be singled out for condemna-
t h e position
adopted
by
Uddyotakara
that he does not regard it as a case of indecisive argument.
O n the
contrary he contends that the onus of proof lies on the opponent
and
1 tan na, drsdfntbhvt-na hi kascid ubhayapaksasampratipanno drstntah

kryam andhram asti't,
N V , p. 435.
2 kryam dhcavad ty atr'pi na drstnto 'sti'ti samnam.
Ibid.
3 samnam ity anuttaram abhyupagamt.
n'smat-pakse drstnto *sti.
4 samnam ity anuttaram...itt brvan
itydi vadato hattasya pratibhatikartavyah.
abhyupagatam tvad bhavat!

;
Ibid.
nUddyotakara "yatrobhayor*
.
. KhKh, p. 993.
i66
INDUCTIVE REASONING
[CHAP.
since he fails to establish his position by an independent line of

argument, the opponent alone should be convicted of failure.
The
position therefore is not one of equality on both sides, since Uddyotakara maintains that the charge of equalisation is capable of independent refutation, whereas Kumarila and Srlharsa are of the opinion that
both the proponent and the opponent in such a situation should be
held to be in the same predicament.
(6) Absence of decisive proof i.e. of crucial evidence
(vinigamanavirahd)
When more than one alternative is possible, and the balance of
logical evidence is equally distributed between them, the acceptance
of one and the consequential rejection of the other are open to the
charge of absence of decisive proof1.
For instance, it is a moot prob-
lem in the Vaisesika philosophy whether the two attributes, namely,

'being an element* and 'being possessed of limited magnitude* can be
regarded as class^characters.
We propose to call these attributes ele-
mentality and dimensionality for the sake of brevity and convenience.

The Vaisesika philosopher holds that two or more universals can
co-exist only if they are related as genus and species, that is to say,
when the extension of one universal is included in that of another
universal.
Thus, substance-universal (dravyatva) and earth-universal
(prthivltva) are found to co-exist in any earthy substance, say a jar.

The extension of substance-universal is wider and more comprehensive
than that of earth-universal in that the former occurs not only in
earthy substances but also in water, air, and. so on.
If two generic
attributes are not related as higher and lower in the sense of genus
and' species as noted above, they are not regarded as class-character
or universal in the strict sense of the term2.
The criterion of higher
and lower universals is that the latter in its extension is included

within the extension of the former, so that though the former may
occur independently of the latter, the latter is bound up with the
:
-t.i: vikalpen 'nvayvagamayogye ekasmin nabhyupagate tad ekadesnvayaniyamanirdhran'sak/attvam avinigamah.

KhKh, p. 1319.
a ..samvistayos ca parparabhYaniyamah, anynnatiriktavrttijtidvayakaipanym pramnbhve.
NKu, p. 79.
DC]
167
former as a matter of necessity. If the possibility of occasional independent occurrence and occasional co-existence of two universals
be admitted, then the co-existence of the cow-universal and the horseuniversal in the same substratum cannot be ruled out as a logical
impossibility 1 . O n e might contend that as there is no evidence
available of the co-existence of these two opposed universals, the confusion of the cow-universal and the horse-universal in the same substratum cannot be seriously entertained as an ontological possibility.
But this contention being based upon an empirical generalisation
cannot rule out the possibility, unless it be backed by an a priori
logical necessity. It has been found that empirical observation,
however extended and wide in its scope, is set aside b y experience of
an opposed coincidence. T h u s one m i g h t have contended that the
generalisation ''All crows are black" was a universal proposition. But
the discovery of white crows in Australia has proved the invalidity of
this universal proposition. So the mere testimony of hitherto available
experience that cows and horses are distinct and accordingly their respective universals have not been found up till n o w to co-exist in one
substratum, cannot be regarded.as an unimpeachable evidence of *thc
utter impossibility of their co-existence in future 3 .
T h e situation however becomes entirely different once the rule propounded by Udayana be accepted that two or more universals in order
to be co-existent must be related as higher and lower in the sense of
being includent and included in respect of extension.
Let us now apply the results attained by this metaphysical digression to the case of the co-existence of the attributes elementality and
dimensionality. These t w o attributes, though not related as includent and included, serve b y virtue of the fact that they occur in a
number of individuals as synthetic principles exactly as accredited
universals do. T h e attribute of elementality belongs to all the five
varieties of elements, namely, earth, water, fire, air, and ether. So
this synthetic attribute can prima facie be regarded as a universal.
1 parasparaparilwavatyos ca samVese gotvsvatvayor api tathbhavaprasangt.
- - >. :
NKu, p. 7 9 , .
2 NKu, pp. 79 et seq.
168
INDUCTIVE REASONING
[CHAP.
Likewise the attribute of dimensionality also belongs to a number of

substances including the first four elements, and the mind which is
held to be atomic in dimension by the Nyaya-Vaisesika school.
It has
every claim t o b e regarded as universal, since it satisfies the test of

synthesising a number of individuals.
But these two attributes in
spite of their synthetic function are not regarded as universals, because

they are not related as higher and lower.
The mind is not an element
and ether (aka) though an element is not possessed of limited

magnitude.
Thus the two attributes being found to exclude one
another, so far as their incidence in ether and mind respectively is

, concerned, are not accepted as universals, since they fail to satisfy the
requirement of the law of the co-existence of universals that they
should be related as higher and lower1.
. We have found that the attributes of elementality and dimensionality are not regarded as universals by the orthodox adherents of the
Nyaya-Vaisesika school. But it may be legitimately posed that though
both are not capable of being accepted as universals, we do not see
any logical incompatibility if either of them is accepted as an orthodox.instance of universal.
N o harm accrues if either elementality or
dimensionality be accepted as a universal, though we admit in deference to the law of Udayana that both cannot be regarded as universals.
The admission of one of them to the rank of universal does not
offend the law, and so there is no logical bar to this possibility.
But
here the type of reasoning we are. considering, namely, absence of

crucial evidence conies in.
Why should we accept one and reject
the other? As no reason for preferential treatment can be made out,

* both of them are on the same level, and so must sink or swim together.
When, both of them cannot be accepted as universals, neither
of them should singly be favoured with differential treatment2.
The
case under consideration is on a par with that of two rival arguments,

samkarai ca paraspartyantabhvasamndhikaranayor ekatra samvesah
H i
bhtatvder jtitve bdhakah.
Din, pp. 77-8,
},:.::<& See amkara Misra's commentary on KhKh, p. 718: ...kirn bhtatvam
mrtatvam v jtir atas tatrai 'katrasya jtitvagrhakam pramnam vinigamanvirah pratibadhnti.
ix]
which contradict each other 1 .
169
Here also the acceptance of the one
as a valid argument and the rejection of the other as fallacious might

avoid the contradiction involved.
But as there is no reason for prefe-
rence the two are rejected without any distinction being made between
them, because- of the lack of special reason in favour of one.
It is
concluded therefore that the two attributes, elementality and dimensionality
must be
rejected
as
false
claimants to the rank of
universal.
The ground of rejection of their claim is absence of crucial evidence in favour of one and for the exclusion of the other.
legitimacy of this type of reasoning has to be admitted.
So the
It might be
contended that absence of crucial evidence is nothing but absence of an

organ of cognition2.
It does not require an elaboration to establish
the proposition that nothing can be accepted as true, unless it is cognised by a legitimate cognitive organ.
One might therefore argue that
there is no case for acceptance of absence of crucial evidence as a novel

phenomenon and consequently as a type of reasoning.
But this contention
with scant regard.
in spite of its plausibility must be dismissed
True, absence of crucial evidence is symptomatic
of the absence of valid cognition by a legitimate cognitive instrument.

But they are not the same thing.
other is the result.
One is rather the condition and the
Absence of crucial evidence strikes our mind first,
and this leads to the conclusion" that there is no cognitive instrument

to certify the truth of the alternatives in question.
It is fit and proper
that we should accept absence of crucial evidence as the ground of

our rejection of the competing alternatives, and we ought not to confound it as a case of default of a cognitive organ, which is rather the
resultant discovery3.
1 satpratipaksatietvor iva nirddhrayitum asakynvayayoh parasparapratiksepa eva paryavasnt.

?
KhKh, p. 13.19.
2 nanv evam pramnbhva eva dosah syIn na'vinigama iti cct?
Ibid., p. 1320.
3 tasy'vinigamonncyattvcn'vinigamasyâiva
prathamotpaanasy'opanyf*
saucityau
Ibid.
22
i;o
INDUCTIVE REASONING
[CHAP.
(7) Empirical induction (utsarga)

When two contradictorily opposite predicates may be conceivable
with reference to an identical subject on a priori grounds and the
balance of logical proof is equally shared between them, it is the
general tendency of the human mind to give preference to one of
them, which is endorsed by experience as yet uncontradicted. This
may be called empirical induction1. Take for instance, the proposition
*'Fire is hot". It is a priori conceivable that fire might be cold, and
this possibility is not capable of being confirmed or rebutted by a priori
logical proof. Though the logical necessity of hotness as a predicate
.of fire, and the logical impossibility of the predicate coldness
cannot be demonstrated by means of pure logic and our only
ground for belief in. the proposition, "Fire is hot", is the testimony of
experience, collective and individual, yet the belief cannot be lightly
dismissed as an uncritical superstition on the basis of a priori grounds.
It is an innate tendency of our mind to accept the truth of a proposition which is continually confirmed by experience, unless and until an
empirical evidence contradicting the truth of it is encountered. These
empirical inductions are capable of being ultimately traced to our
belief in the Law of the Uniformity of Nature, though this belief is not
susceptible of logical justification. It maybe legitimately supposed
that whatever happens in the universe has got a sufficient reason at its
back. But this sufficient reason which can show its logical necessity
may remain a sealed book to us, regard being had to the limitations of
knowledge to which mankind is unfortunately compelled to remain
subject. Only an omniscient spirit can have an illuminative insight
into the ultimate secrets of Nature.
It is a pity chat there have been philosophers who have made
capital out of these human limitations, and have exploited them for
justification of scepticism. Whatever might be the philosophical
merits of scepticism as an attitude of mind, unqualified scepticism has
not commended itself as an explanation of the problems of the world.
The human mind has instinctively refused to be gagged into silence,
t fchulyadrstam apeksya bhulyadrstatay durbalasy 'nupagamrhato'tsargah.
KhKh, p. 1322,
ix]
171
or stifled into inactivity which is the necessary concomitant of intellectual death.
The sane and healthy attitude should be acceptance of
the testimony of experience as truth, provided we have convinced

ourselves that there is no ground for believing it to be vitiated by a
defect in the subjective and in the objective conditions of knowledge.
It is the recognition of this principle in British jurisprudence which
allows the accused to be regarded as innocent, until ancj unless his
guilt is brought home by indubitable evidence.
It is again in con-
formity with this principle that the Mimmsists of the Bhtta school
in India accept all judgments to be true, unless a contrary evidence
showing defect in the subjective apparatus or the objective conditions
or a subsequent
contradiction by
another
accredited
presents itself to set aside their truth-claim1.
experience
The sceptics of the
Buddhist school believe that all judgments should be suspect unless

their truth is established by independent evidence.
But this is scepti-
cism in excelsis, and as we have observed before the generality of

mankind looks askance at this attitude and believe it to be rather an
aberration.
Let us now turn our attention to another contention that these
empirical inductions are rather of the nature of doubt.
They differ
from pronounced cases of doubt in that the former beliefs are concentrated on one possible alternative, and not found to vacillate between
two alternatives of equal strength like the latter2.
But though the
element of vacillation is not pronouncedly felt, these empirical inductions are at their best of
the nature of doubt of an alternative of
superior strength showing only a high degree of probability.

this contention
seems more ingenious than convincing.
But
It is an
extreme view which puts u neon trad icted inductions on the same level
with doubt.
In the first place, such inductions serve as positive
evidence and doubt is rather in the nature of a disproof.
In the
1 yam tarkam etatn lamby huh :

tasmd bodhatmakatvena prpt biiddheh pramnat/
arthnyathttvahettthadosajnnd apodyate//,
5V, p. 61. Quoted in KhKh, p. 132a.
2 nanu balavad-ekakotikah samsaya evo' tsargas tat kathara tarkah syd
iti cet?,
KhKh, pp. 1322-23.
172
INDUCTIVE REASONING
[CHAP.
second place, doubt is an unsettled condition of the mind which

oscillates like a pendulum between two competing alternatives, whereas an empirical induction is of the nature of the unilateral cognition
of one of the extremes1.
Nor need we attach undue importance to
the modified scepticism of a section of philosophers who would rank

these inductions with presumption.
In presumption one of the alter-
natives is believed to be much more powerful than the other on

account of the preponderant balance of evidence in its favour.
But
though it stands higher in rank than plain doubt, the possibility of

the other extreme, however weak, is not entirely ruled out of court.
The difference between doubt and presumption is thus one of degree
and hot of kind 2 .
Empirical inductions on the otherhand are believ-
ed as .veridical -and the doubt created by philosophers is rather of the

nature of.after-thought.
If premium be put on such indulgence in
.metaphysical doubt, there will remain very few beliefs which can
Haim immunity from its* invasion.
If the opponent seeks to decide
the issue by, appeal to psychological evidence, and differentiates between assured conviction and presumption by the criterion of psychologically felt doubt in one case, and exemption from it in-the other,
then this will not afford any advantage to the sceptic, because empirical inductions of\ the kind, we have been considering,,will also lay
claim to this prerogative, as psychologically speaking such beliefs
are immune from the visitation of doubt in any manner or form3.
Before bringing the present discourse to a close, we propose to
advert to a trait which differentiates this type of reasoning from the
other types.
Empirical induction serves rather to prove a position
by appeal to non-contradiction, and the rebuttal of the opponent's

position is consequential.; Such .types of reasoning as infinite regress
etc. are primarily directed to the refutal of the opponent's position
and their cogency is derived: from the contradiction involved in them.
Non-contradiction is the main plank on which empirical induction
relies for its cogency, whereas the other types depend upon the con
'
'
'
"
"
'
..utsargasy aikakotinisthatvt samsayasya ca kotidvayvaghitvt.

KhKh,p. 1323,
J2:.; Ibtdi .,,.
3 bid.
ix]
tradiction arising from their application.
173
The fundamental characteris-
tic of reasoning operating as reductio ad absurdum is also present in

it, since by proving its position it serves as an instrument of disproof
of the opponent's plea indirectly1.
We cannot help adverting to the enumeration of the types of
reasoning as given in the Sarvadarsanasam graba.
It is a matter of
regret that Mdhavacrya, the author of the work, has not explained
the significance of any of these different varieties of reasoning.
We
find in the list the inclusion of apavada (exception) as a special

variety2.
We do not know of any recognised authority of any school
of Indian logic who has mentioned it as a variety of reasoning.
We
have ransacked all available sources and have incorporated our findings
in this dissertation.
It is imperative that we should not ignore the
classification of the author of the Sarvadarsanasam grab a n view of

its wide popularity.
In this work there are given eleven types of
reasoning, namely, (1) contradiction (vyagbata);
(2) self-dependence
(atmasraya); (3) mutual dependence (itaretafasraya);

(cakraksraya);
(pratibandt,
ghavd);
(4) vicious circle
(5) the vicious infinite (anavastha); (6) tu
wrongly printed pratibandbt);
(8) complexity (kalpanagaurava);
quoque
(7) simplicity (kalpanali(9) empirical
(utsarga); (10) exception (apavda); (11) impudence
induction
(vaiyatya)3.
We have thrown ample light on ten varieties of reasoning, but

the inclusion of a pavada has been source of good deal of difficulty
to us.
We have been painfully constrained to conclude that the
classification given in this work is rather slipshod and hasty.
W e do
not see even a slight warrant for the recognition of apavda as a

variety of reasoning.
The
recent commentator
MM.
hstri Abhyakara has .rather touched, and let it go.
Vasudeva
He pimply
explains utsarga (empirical induction) as a general rule and apavda

as a special injunction which is to be taken as an exception to this
general rule.
It is well-known to students of grammar and Mimms
that these two terms go together as a pair of opposites.
1 KhKh, p. 1323.
2 SDS, pp. 230-39,
3 Und.
MM.
i74
INDUCTIVE REASONING
[CHAP.
Abhyakara has only given the meanings which are current in these
two systems of discipline.
It is obvious from our exposition that
mtsargdhas got a wider significance.
Apavda cannot be regarded as
a variety of reasoning even by the widest stretch of imagination.
It
is most probably a slip of an unlearned scribe or a superficial student's

emendation that is responsible for its inclusion in the enumeration
of the types of reasoning.
Apavda
or special injunction embodies
an exception to the general rule and this exception is warranted by

undisputed authority of valid cognition.
It only shows that the
jurisdiction of the general rule is valid only outside the sphere of

exception1.
It does not function as a reductio ad absurdum
which
is the essential characteristic of reasoning.

We have taken elaborate pains to demonstrate that even such
recognised types of reasoning as simplicity, complexity etc.
are not
regarded as orthodox cases of reasoning by the later logicians on the

ground of their failure to serve directly as cases of reductio ad
absurd urn.
But as they server to re-inforce
the finding of an
accredited cognitive organ, they have been regarded as species of

reasoning in a secondary sense.
We have also tried to show that
they are not entirely devoid of the character of an invalidating agent,

since they serve to weaken the presumption in favour of the opposite
propositions by negative implication.
But neither of these characte-
ristics is found to be present in apavda.
It is for this reason that
SriKarsa who has been very liberal in his recognition of the types of
reasoning of varying degrees of strength has not cared to make room
for apavda, although its close association with utsarga is almost a
truism.
It is impossible to suppose that Srlharsa would not have
cared to include it, had he not been impressed with the absurdity
involved in it.
Its omission in the inflated list given by Visvanatha
is also a proof of the futility of its claim to the status of reasoning.

We feel compelled to reiterate our suspicion that the inclusion of
Apavda in the list of the types of reasoning in the Sdrvadarsanasamgraba is the outcome of an error of judgment, either of an uninformed
i
apavdavisayam parityajya utsargo 'bhinivisate.
A paribhZsa of MB.
IX
175
scribe, or of a hasty student who felt a natural temptation to make

good what he considered to be an ellipsis.
(8)
Complexity (gaurava) and Simplicity (lghava)
These two types are complementary to each other and so will be

treated together.
When between two alternative hypotheses, one
involves complexity and as such is difficult to understand and the

other is transparent in simplicity and as such is easily intelligible, the
former is liable to condemnation, and the latter is entitled to preference.
The ground of preferance lies in its simplicity and that of
condemnation in complexity 1 .
It must be borne in mind in this
connection that these two considerations are of value only when. the
balance of logical support is equally present.
A theory ought not to
be condemned merely on the ground of its complexity, if

complexity is entailed by logical necessity.
this
Likewise simplicity: by
itself is not a virtue, unless it is shown to be adequate and justifiable

by independent evidence.
Complexity may be due to qualitative i
and quantitative consideration.
When a fewer number of terms can
account for a situation, the postulation of a larger number will constitute quantitative complexity 2 .
Again, a hypothesis, which is hard
to comprehend is liable to the charge of qualitative complexity, if a

simpler hypothesis is available and found to be equally adequate.
We find in the processes of the solution of the mathematical problems
examples of these types of reasoning in abundant number.
sophy the application of these types is quite frequent-
In philo-
The cosmo-
logical argument of the Naiyyikas, proving the existence of one God

as the creator of the world on the ground of its emergence as the
product of a process, is supported by this type of reasoning.
It has
been contended by the opponent that if world be a product and as

such should have an intelligent author, why should not a number of
agents be postulated on the analogy of building of the
houses etc?
The Naiyiyika meets this charge of the plurality of agents by appeal

to this quantitative difference between complexity and simplicity.
sugamsugamayor asugamaducbalattvam
tadvyatirekena kalpanlghavam sdhannuklam.

2
Ibid.
kalpangauravam
,,
KhKh, p. 1325.
176
'' INDUCTIVE REASONING
[CHAP.
The postulation of a plurality of agents is more complex than that

of one. So the law of simplicity requires that monotheism should
have preference over polytheism as a metaphysical theory1.
The law of simplicity is wider and more comprehensive than the
law of parsimony or Occam's razor, because the latter does not
exhaust its scope.
As we have observed that it has got both quan-
titative and qualitative forms and the law of parsimony s only an

example of the quantitative form.
On the qualitative side it differs
from the law of parsimony .^ In pursuance of this law of simplicity

we are compelled to discount supernatural factors and give preference
to a natural explanation.
Thus the explanation of a human malady
by the agency of unseen evil spirits is not favoured, when a natural

explanation of it by means of physiological disorder is found available.
It should be always borne in mind that simplicity is used as a consideration in support of one's position and complexity is urged as a
fault against the opponent.
r
The Naiyyikas have distinguished three varities of simplicity from
the corresponding cases of complexity.
Simplicity and complexity
may be (1) in respect of form (sarlrakrtam), (2) in respect of presentation (upasthitikrtam),
and (3) in respect o relation, (sambandhakrtam)2.
Extensity (mahattva)
is propounded as the condition of perceivablity.
Others put. forward that the attribute of being inherent in more than
one substance (anekadravyasamavetattva)
ceivablity)*
is the condition of it (per-
Extensity is lighter in form than 'inherence in more
than one substance* and hence is preferred to the latter3.

: are the cases of formal simplicity and complexity.
These two
-, -
It is a truism that an- effect must have been non-existent before its
production.
Thus pre-negation (prgabhva)
tion precedent of its coming into being.
of the effect is a condi-
The production of odour is
concomitant with the production of a colour as a matter of fact.
So
the production of odour must be preceded by the pre-negation of

* n & & > *
1 ? .
' : : . : . <
I: KhKh, p. 1325.
2 laghutvam ca sarirakrtamupasthitikrtam sambandhakrtam ca.
.,,?/,.;;;,.,,,>m-v:;^;.:-..
. , _ : . . . . . . D i n , p . 1 2 1 .
3 tatra prathamam anekadravyasamavetatvpeksay mahattve.
..
.;
ix)
t77
colour and the pre-negaton of odour both, as a matter of necessity.

Now, which of the two pre-negations should be held to be the
condition of odour?
It is maintained that the pre-uegation of odour
should be the condition of odour, because it presents itself readily to

the mind, as in the enquiry of the cause of odour, the pre-negation
of odour is more relevant than the pre-negation of colour.
Here the
simplicity lies in the promptness of the cognition of the pre-negation

of the odour and
complexity in the belated
occurrence of the
cognition of the other alternative (pre-negation of colour)1.

Again a pen is the antecedent of writing.
But the colour of the
pen and the pen-universal are equally antecedent to it.
Now which
of the three, the pen, the colour, and the pen-universal should be
regarded as the causal condition of writing?
The answer is that
the pen should be held as the condition of writing, and neither its
colour nor the pen-universal, because the latter would make the causal
relationship more cumbrous than the former.. It is an accepted pos^
tuiate of Inductive Logic that the cause and the effect must be intimately related and the effect should occur only where the cause is in
operation.
It is absurd that one throws a bomb in New York and a
man is killed in London.
The reason for this lies in the fact; that
the effect can occur in the place where the cause has occurred
Now,
writing takes place in the piece of paper with which the pen is in
contact.
The co-existence of the pen at work and writing is easily
intelligible and there is no difficulty in regarding the pen as the causal

condition of writing, because of their co-relationship with the paper.
If the colour of the pen or the pen-universal be sought to be made
out as the causal condition, the relationship will be cumbrous, because
the colour can be related to the effect through its inherence in the
pen.
The colour inheres in the pen1 and the pen is in conjunction
with the .paper in which the writing is done.
So the relation* will be
inherence-cum-conjunction in the case of the pen-universal and the

colour, whereas if we make the pen the cause of writing then the
relation being simply conjunction is found to be less cumbrous.
The
i dvitiyam gandham prati rpaprgabhvpeksay, gandhapragabhve.

gandhtmakapratiyjogijnnasattvena sighram tadupasthiteh. Din, pp. 121-22.
i78
INDUCTIVE REASONING
causality of pen is therefore admitted on the ground of this simplicity

of relation. The case under consideration illustrates relational simplicity
and complexity
We have not said anything about the cogency of 'initial presentation' (pratbamopastbitatva),
which has been regarded as a variety of
reasoning by some older logicians as it appears from the statement of

Visvantha in the Nyyastravrtti.
This has not attracted the atten-
tion of later logicians and thus goes unchallenged.
It is an accepted
convention that we should not reject what presents itself first unless
there is strong evidence against it.
It is based on the truth that
nothing should be condemned unless there is strong reason for it.

Accordingly, what appears first in the field as a claimant for consideration should not be brushed aside, but on the contrary ought to be
accepted unless subsequent evidence invalidating
presents itselfr
its
authenticity
It has not been regarded as a variety of reasoning by
later logicians and their attitude is perfectly justified in view of the

fact t;hat it naturally reduces itself to a case of simplicity in respect
of presentation which we have just treated of.
(9)
Contradiction (yyaghata)
It consists in the admission of co-existence of two mutually repellent attributes in a self-identical subject1.
This admits of classi-
fication under three heads, namely, (1) contradiction in language

(svavacanavyaghata),
(2) contradiction in action
(svakriyavyagbata),
and (3) contradiction in thought (svajnanavyagbata)2.
The proposi-
tion "My mother is barren" is an illustration of the first variety.

Motherhood connotes the attribute of being productive of a child
and barrenness on the contrary connotes the lack of such productivity.
The; opposition involved in the predication of productivity and nonproductivity in respect of the same subject and in the same reference,
is toa obvious to require elucidation.
contradiction in action.
The second variety is called
When a man actually asserts "I am dumb",
he is* guilty of self-contradiction
by his own act.
The very act of his
1 viruddha-samuccdyo vyghcah. KhKh, p. 1292.

Z' svavacanasvakriysvajnnavyghtdidosarahitam...
ATV, p. 533; vide also ATVD, p. 537.
ix]
179
assertion proves the falsity of the predicate "dumbness". It should

not be confounded with contradiction in language in which the connotation of the subject and the predicate are mutually incompatible and
the very assertion is suicidal and self-repellent. In the second variety
the opposition of the subject and the predicate is not inherent in the
proposition itself, but it becomes obvious when the proposition is
actually articulated by the person. The articulation qua act brings
out the contradiction. There is no contradiction involved in the
proposition in the written form. So it is articulation which furnishes
the contradiction. It is thus a case of extrinsic contradiction effected
by an external act, whereas the opposition in the first variety is intrinsic. The third variety is called contradiction in thought. The following may be given as an example of it. When a man asserts and seems
to think "I do not know this pen"*, he becomes guilty of contradiction
in thought. The very contemplation or assertion of the object (this
fen) presupposes the knowledge of it, and so the assertion of the lack
of knowledge of the pen contradicts the assertor's knowledge of the
same which is the presupposition of the assertion.
It should.be clearly borne in mind that the pronoun or the pronominal adjective 'this' indicates direct intuition of it\ and the assertion
of ignorance of it is clearly a case of self-contradiction in thought*
But for the pronoun 'this* the assertion would be perfectly innocuous.
There has been an unbridgable gulf of difference between the monists'
of akara's school and the philosophers of the dualistic persuasion
regarding the, meaning of the proposition "I do not know" or "I am
ignorant of X ' V The former interpret it as proof of a positive fact,
and the latter attempt to equate it with the assertion of lack of khow^
ledge* In spite of this difference there is perfect unanimity on the
point that the ; proposition is significant and not nonsensical or
self-discrepant ? like the proposition "I do not know this'*..
The self-discrepancy of _ the latter proposition lies in the employment of the pronoun 'this' which indicates direct perceptual *
cognition or in the language of Russell 'knowledge by acquaintance*;
So the meaning, of the proposition "I dp not know this" turns
out on analysis to be "I do not know what I directly perceive". This
is clearly a case of what has been called self-contradiction in thought.
iSo
INDUCTIVE REASONING
[CHAP.
(IO) Impertinence (anaucitya) alias Impudence (yaiytya)

' Impertinence is also called impudence. It consists in the employment of an issue which is condemned by philosophical convention, and
is not capable of refutation or justification on account of its absurdity1.
An honest error, though absurd is susceptible of refutation or correction, and as such it is not a case of impertinence. We can give some
examples which will throw light upon its nature. Suppose a man
multiplies question upon question interminably. If this process be
not brought to an end, no discussion can be fruitful. It is therefore a
recognised rule of debate that questioning should be confined to relevant issues. Irrelevant digression is a fault.
But the fault of impertinence is not constituted by irrelevant
digression {arthantara) alone. If one seeks to silence an obstinate
arger by the charge of irrelevancy, the latter may ask, "What is
irrelevancy?" and each answer may again be made the subject-matter
of a question. This interminable prolongation of the chain of debate
can be arrested only if closure be applied to it. Impertinence is the
ground of this application of closure3.
Again, the question, "Is non-entity real or unreal?1', is not fit for
an answer. The assertion of reality is self-contradictory, and that of
unreality, again, will involve the same issue, when it is considered that
the subject or predicate of a proposition cannot be a fiction. The
%
proper attitude with regard to such a question is contemptuous silence3,
Again, if any one ask "Is the son of a barren woman in question
beautiful or not?" it will be an instance of impertinence.
It might be argued that impertinence is not a distinct fault and it
can be subsumed under one or other of the recognised types of logical
defects. For instance, the interminable series of questions and answers
is either subsumable under irrelevant digression (arthantara) or infinite
regress. And as regards the impertinence involved in the questions
1 prmnikpvyavahryatvam asamdheyajatiyam anaucityam yaiytyanmakanu^^.-^.
.
KhKh, p. 1326.
2 Ibid., 2' X327*
3, yathWastuni vidhinisedhayoh kim icchasi'ti prcchati/..atra sahrdaynm mkatai'vQ*cite'ti.
:
*: : '
bid., p. 1326.
ix]
181
regarding non-entity or a barren woman's son, these are instances of

self-contradiction.
So where is the scope of impertinence as a separate
fault?
But the necessity of recognition of impertinence as a separate
category of fault will be apparent on reflection.
It is no doubt a fact
that impertinent questions of the types we have considered involve

the faults shown by the critic.
But it is not exhausted by this or
that recognised type of fault.
Suppose the proponent or the umpire
convicts the arguer of self-contradiction.
But the arguer may further,
ask the question, "What, sir, is self-contradiction ?"
Any answer
that will be offered may again be made the peg of a further question,
and so on to any length.
If you urge that the attitude involves in-
finite regress, that also will not suffice to silence the man.
So what'
is the remedy save and except the conviction of the arguer of impertinence?
And thus can the final closure be applied to bring the
debate to a termination.
It must however be borne in mind that the occasion for thev
application of this charge arises only when the questions are inspired
by perversity or vitiated by absurdity.
It has been observed by
Paulsen that absurdity shares with truth the advantage that both cannot be refuted.
Impertinence arises only when a question does not
deserve an answer or even a refutation.
The proper attitude towards
it, as we have said, is one of refusal to answer it.
But as silence is
liable to be misunderstood as incapacity, the pronouncement of the

verdict of defeat on the ground of impertinence is imperative on the
part of the umpire who presides over the debate,
Srlharsa has taken
elaborate pains to establish the necessity of recognition of impertinence

as a distinct type of reductio ad ab surd urn and he concludes his argument by citing the authority of Mahimabhatta who gives it a* place
of honour in the formulation of the types of literary defects in the
Vyaktiviveka,
an extraordinary work on literary criticism1.
dosam Vyaktiviveker'mum kavilokavilocane/, .

kvyamimmsisu prptamahim Mahim" drta//
KlKh p. 1327.
182
INDUCTIVE REASONING
(II)
[CHAP.
Equalisation (samavacana)
Equalisation has been urged by Vekatantha and his commentator Srinivscrya as a distinct type of reductio ad absurdum1.
They
maintain that it cannot be subsumed under ttt qnoqtte argument.

The tu qttoque argument consists in the allegation of the same or
similar defect in the opponent's position, which has been urged by
the latter against the proponent.
Equalisation, on the other hand,
does not rest on the allegation of a similar defect, but on the claim
of a similar advantage.
Thus it is urged that if the Absolute
Brahman is to be accepted on the ground of its being supported by

logical proofs, the reality of the phenomenal world too ought to be
accepted on the similar ground of its being supported by incontestable
logical evidence2.
It acts as a reductio ad absttrdum, because it estab-
lishes the untenability of the monist's contention that the world is a

false appearance.
.
(12)
Double noose (ttbhayatahps)
Double noose is the same thing as dilemma*
When an argument
is shown to involve the opposite consequence by impaling it (argument) on the horns of a dilemma, the result is achieved by the
application of this type of reductio ad absurdttm.
An example will
illustrate the issue.
The Vedntic monist has established the un-
reality of the world.
But the opponent asks "Is this unreality true or
false ?"
If this unreality be false, the reality of the world will be
proved.
If again it be true the truth of unreality will prove its reality
and so monism will not be established3.
Thus the argument is prove
to be false whatever alternative of the two extremes of contradictory

opposition be conceded.
The double noose argument is the equivalent
1 NP, p. 347. See also Srinivscrya's commentary on NP, p. 347.

2 See Srinivasadirya's commentary on NP, p. 347: yadi prmnikatvena
Brahmanah satyatvam syt tulyam prapancasy'pl'ti smypdanam
samavacanam.
3 See Srinivascrya's commentary on NP, p. 347: mithytvam mithya
va na v. yadi mithya tarhi praparkasya satyatvan na BrahmatHktanisedhasiddhih. yadi na mithy tarhi tasyaiva satyatvena na Brahmatiriktanisedhasiddhir
iti sambhavitakotidvaye 'py anistpdanam ubhayatahsps.
TABLES OF CLASSIFICATION OF REASONING

Udayana's fivefold classification
(1)
Reasoning (tarka)
Self-dependence (I)
(tmeraya)
Mutual dependence (II)

(anyonysraya)
Table I
Vicious circle (III)

{cakraka)
The vicious infinite (IV)

(anavasth)
Reductio ad absurd urn (V)

{anistaprasanga )
(2) Vivanatha's reference to the five additional varieties of reasoning which can be tabulated as follows:
Table II
Table I plus
Initial presentation (VI)

{prathamopasthitatva)
Empirical induction
indu
(VII)
(utsarga)
Absence of crucial evidence (VIII)

(v inigamanviraha)
Simplicity (IX)
lyhava)
Complexity (X)
(gaurava)
Srlharsa's elevenfold classification
(3)
Selfdependence
Mutual
dependence
Vicious
circle
The
Tu quoque
(pratibandi)
Contradiction.
(vyghta)
VICIOUS
infinite
Absence
of
crucial
evidence
Table III
Empirical
induction
Simplicity
Complexity
Impertinence
(anaucitya)
alias
Impudence
(vaiytya)
Tenfold classification in the Vdivinoda
(4)
Selfdependence
Mutual
dependence
Vicious
circle
Beductio adabsurdum
as distinct from the
first four varieties
(t adit am nistaprasanga )
vicious
infinite
I
Absence
of
crucial
evidence
Table IV
Empirical
induction
Complexity
Varadarja's and Sankara Mira's classification as set forth in his commentary on the Khandanakhandakhdya
Simplicity
Impertinence
or
Impudence
is the same with Udayana
Classification of reasoning in Rmnuja's school

Venkatantha's eightfold classification
(5)
Selfdependence
Mutual
dependence
Vicious
circle
The vicious
infinite
Pure reductio ad absurdum

(kevalnistaprasanga)
Sub-division of fche fifth variety in the
Table V
Tu quoque
Equalisation
(samavacana)
Double noose
{ubhayatahpe)
Prajfiparitrana
I
Impossibility
(asambhava)
Opposition
(virodha)
Table:
Srinivasa's tenfold classification
(6)
The same as Table V plus the two other varietiesviz.
(7)
Contradiction
Complexity (ix) and Simplicity (x)

(gaurava)
(lghava)
Elevenfold classification of reasoning as found in the
Selfdependence
Mutual
dependence
Vicious
circle
The vicio us
infinite
Tu quoque
Table Wi
Sarvadaranasamgraha
Simplicity
Complexity
Empirical
induction
Exception
(apavda)
Im I
ix]
183
of the dilemma o European logic with this terminological difference

that whereas the two opposite extremes, between which the opponent
is called upon to exercise his choice and the admission of either lands
him in absurdity, are called nooses in Indian logic, they are called
horns of the dilemma.
This finishes the dissertation on classification of reasoning and we
now propose to deal with the conditions and fallacies of reasoning in
the next chapter.
CHAPTER X
SECTION I
We have dealt with the classification of reasoning in the preceding

chapter, and we now propose to deal with the essential conditions and
factors of reasoning, the infringement of which severally and jointly
gives rise to different fallacies of reasoning.
These conditions and
fallacies have been enunciated by Udayana in the
dtmatattvaviveka1,
and they have been codified by Varadarja in the
Trkikaraks2,
and arlnivscrya in his commentary on the Nyayaparisuddhi

3
Vekatantha .
of
This shows that the importance of the matter has
received continuous recognition since the time of Udayana,. if not

earlier. * These conditions have been set forth under five heads:
The absence of rebuttal by a contradictory reasoning.
it must be devoid of a rival.
(i)
In other Words,
(2) There must be concomitance in
agreement between the ground and the consequent.
(3) The conse-
quence entailed by reasoning must be an undesirable contingency.

(4) The reasoning employed must not be conducive to the proof oi
the opponent's thesis.
(5) There must exist concomitance in differ-
ence, that is to say, a necessary logical relation between the opposites

of the ground and the consequent4.
This is the most important factor
inasmuch as it shows the fulfilment of the ultimate purpose of reasoning.
Reasoning as we have observed, time and again, is at bottom
a reductio ad absttrdum.
The exposure of the absurdity of the
opponent's position inevitably entails the consequence that the truth

lies in the opposite proposition.
The failure of reasoning to culminate
in the truth of the contradictory of the opponent's thesis stultifies the

very purpose of reasoning.
1 ATV, p. 5 5 3 i
2 vyptis tarkpratihatir avasnam viparyaye/j
anistnanuklatve iti tarkagapacakam//,
TR, p. 187.
3 See Srinivscrya's commentary on NP, pp.
4 TR, pp. 187-88.
x]
185
The default of any one of these five factors gives rise to five
fallacies of reasoning which are named as follows: (1) mutual contradiction (mithovirodha);
(2) the lack of logical sanction (mlasaithilya);
(3) the establishment of an issue endorsed by the opponent
\istapa-
dana); (4) to be helpful (in the proof of the opponent's thesis)

(anuklatva);
(5) failure to culminate in the proof of the opposite
(viparyayparyavasana)1'.
The first fallacy arises when a reasoning is set off by another
reasoning, and the opposition of one reasoning by another is at bottom
a case of mutual opposition.
The second arises when the concomi-
tance of the ground and the consequent is found to be lacking.
As
has been elucidated by Raghuntha Siromani, the pivotal condition of

reasoning is the definite cognition of the necessary concomitance
between the ground and the
the ground in a subject.
the lack of logical sanction2.
inference.
consequent
and the admission of
The lack of any one of these entails

The case is exactly on a par with
The essential conditions of inference are necessary con-
comitance of the probans and the probandum and the subsistence of

such a probans in the subject.
In other words, the major premise
and the minor premise are the essential conditions of inference.
In
reasoning also the lack of either of these two premises which serve to
demonstrate the necessary concomitance of the ground and the consequent and the assumed existence of the ground in the subject makes
the employment of reasoning ineffectual, just as the lack of either of
die premises in inference fails to establish the desired conclusion.^
The third fallacy of reasoning consists in logical emergence of an
issue which is not incompatible with the philosophical standpoint of
the opponent.
The reason why it is regarded as -fallacious is the
lack of the fundamental condition of reductio *ad absurdum,
namely,
the proof of an issue which is detrimental to the opponent's position.
'. - ? . \
. U V & f ?
:
t
'
;"
'
'-
-":-
1 michovirodhamlasaichilyescpdannuklatvaviparyayparyavasnais tarkbhsatvts 7f.

ATV, p. 553; see also TR, pp. 188-90; NP, pp. 348-50.
2 mla&uthilyam, tarkasya mlam pdypdakavyptiniscaypdakbhyupagamau tacchaithilyajn tad anyataravyatirekah.
:
ATVD, p. 557.
24
186
INDUCTIVE REASONING
[CHAP.
It is a case of labour wasted and serves to prove the arguer's logical

ineptitude, since he has failed to discriminate between what is endorsed by the opponent and what is repudiated by him.
The fourth variety consists in yielding a point which tends to
establish the position explicitly advocated by the opponent.
The
difference between the third and the fourth variety is a subtle one
and is apt to be ignored.
In the third variety, the arguer is found
to lay stress on the emergence of an issue which is regarded by him

as an unwelcome contingency from the opponent's point of view,
whereas in reality it is not offensive to the latter.
In the fourth
variety also the issue shown to be entailed by the opponent's argument is also equally unoffensive.
are not different.
So in this respect, the two varieties
But the difference lies in the fact that in the third
variety the issue shown to be logically entailed is without a bearing

upon the thesis advocated by the opponent, though it is approved by
the latter, as it does reot run counter to his general philosophical
standpoint.
In the fourth variety, on the other hand, the issue shown
to emerge is not only inoffensive to the opponent, but directly leads

to the establishment of the conclusion which he contemplates to prove.
So the difference is at bottom found to revolve round the question of
what is relevant and what is not to the conclusion sought to be
established.
This distinction has got to be admitted though there
is no express formulation of it in standard works, otherwise the

two varieties would be lumped into one.
The fifth variety consists in the absence of concomitance in
difference and the second variety arises from the lack of concomitance
in agreement between the ground and the consequent.
The
two
cases are similar in so far as they are derived from the lack of necessary concomitance.
But the difference is fundamental, not only be-
cause of the difference between the positive and negative concomitance,

but also because of the difference in the result.
The lack of positive
concomitance only serves to show that the reasoning has no raison

d'etre, whereas the lack of negative concomitance makes the reasoning
Inoffensive as a damp squib.
It shows that the reasoning fails to
serve its purpose which is to show that the truth lies in the
sition opposite to what is sought to be established.
propo-
xj
187
nature of these fallacies will be best elucidated by concrete

examples.
We therefore propose to illustrate these fallacies seriatim.

(1)
Mutual contradiction (mitbovirodha)
Let us take up the Smkhya argument on causation.
It is conten-
ded that a non-entity cannot be brought into existence by causal

operation.
A rabbit's horn is an example in point.
If the effect in
question were non-existent before the causal operation, it could never

be called into existence, just as a rabbit's horn which is a non-entity
is never found to be produced by any amount of exertion.
The
Smkhya concludes that the effect was existent even before the application of causal activity.
The Nyya-Vaisesika school of philosophers
oppose this argument by means of the following argument: Whatever

is in existence is not capable or in need of being produced, just as
spirit (purusa) and primal matter (prakrti) according to the Smkhya
philosopher.
If the effect in question were as existent as spirit or
primal matter even before causal operation, then the production of

the effect in question would not be either possible or needful.
Pro-
duction is nothing but the bringing into existence of what was not in
existence before.
But if the effect were in existence,' what would
causal operation avail?

Now these two arguments directly contradict each other and as
such both of them equally fail to establish the contention of each.
The case is closely analogous to countermanded probans {satpratipaksa). If we probe the question deeper, it would transpire that it is the
self-same fallacy as countermanded probans, since the contradiction of
one reason by another is the raison d'etre of its being regarded as a
fallacy.
This is also the considered view of SriharsaV

( 2 ) - - T h e lack of logical sanction (mlasaithilya)
This fallacy arises, as has been observed before, if there is no necessary

concomitance between the ground and the consequent.
It is necessary
concomitance^which makes it possible for the ground or the probans

1 mithovirodhas ca satpratipaksatai'va.
KhKh, p. 1337.
i58
INDUCTIVE REASONING
[CHAP.
to be determined by the consequent or the probandum. The former

is the determinate and the latter is the determinant. So the relation
of necessary concomitance is at bottom one of determinant and determinate. A can be the determinant of B, if A is necessarily concomitant with and so cannot occur without B. But if it is found on
examination that the ground employed is not the determinant of the
consequent and conversely the consequent is not the determinate of
the ground, the reasoning which employs such a ground is condemned
as fallacious on the ground of its being devoid of logical sanction. An
example will make this obvious. "If the lake were possessed of
smoke, it would be possessed oa jar". This is a typical case of the
lack of logical sanction inasmuch as there is no necessary concomitance
between smoke and jar and so the former is not the determinant and
the latter is not the determinate. It is a case of plain non sequitur.
(3)
The establishment of an issue endorsed by the opponent

(istpdana)
The Buddhists do not believe in the existence of a whole apart

from and in addition to the constituent elements.
The apparent
solids such as the chair and the table are nothing but a conglomeration
of atoms.
In support of this position they employ the following
reasoning:
There can be no whole as a unitary principle.
Because
the admission of such whole makes it the substratum of co-existence

of the contradictory qualities.
Take for instance, a tree.
sheer plurality and cannot be a unity.
It is a
A bird is found to perch on a
branch of it, and the other parts of the tree are unoccupied.
If the
tree were one unitary whole it would be in conjunction and out of

conjunction with the bird. Again, a table may be red in one part and
not-red in another,
ft may be covered in one part and uncovered in
other parts, 'r The parts are different and distinct from one another.
And so there is no contradiction involved in the situation.
But the
so-called whole being numerically distinct and different from the parts,
and again being one simple unitary entity, is at once riddled with
contradiction if you make it both conjoined and non-conjoined, covered
and uncovered, coloured and uncoloured.
The criterion of numerical
difference is the possession of the contradictory attributes, and as this
x]
189
criterion is found to be present in the so-called whole, it follows that

it is not one but many.
The whole is therefore non est.
. The Buddhist further draws forth this implication by observing

that, the so-called wholes, chair, table etc.
are not only not wholes,
but also neither perceptible bodies nor imperceptible entities,
it has
been proved that the whole pervading the parts is logically impossible.
It cannot, again, be maintained that the whole is an imperceptible
entity, because that would contradict the plain verdict of experience.
If it were imperceptible, it would not be perceived as such.
But
experience does not endorse any such intangible principle, and the
advocate of the whole cannot point to any other independent source of
knowledge outside perceptual experience. Experience, on the contrary,
attests that the tree and the table etc. are perceptible facts.
The second contention of the Buddhist that the so-called wholes,
the tree, the table and so on are not imperceptible entities, is endorsed by the Nyya-Vaisesika school which advocates the existence of
wholes.
The contention therefore yields an issue which is not only
not inoffensive to the position advocated by the opponent, but is

welcomed by him as a true estimate of the nature of the whole.
The
Buddhist therefore hurls a garland of flowers mistaking it to be

a weapon of offence.
Apart from rhetoric the reasoning' employed
by the Buddhists fails of its purpose to refute the thesis of the

opponent.
(4)
To be helpful (in the
(anuklatva)
proof of the
.,
opponent's
thesis)
Suppose the Buddhist, who does not believe in the existence of

God, contends that if God were to be a real existent, then the
Upanisads would have to be accorded the status of an organ of
authentic knowledge.
N o w this contention would be an instance of
the fallacy under consideration.
It would not be strictly accurate
to maintain^that, it is an instance of the third fallacy of reasoning,

since the alleged consequence is unreservedly accepted
by the theist.
There is a subtle difference between the two cases.
As has been
observed before, the third variety consists
in the establishment of a
consequence which is directly endorsed by the proponent; whereas

the fourth variety entails a consequence which is conducive to the
i9o
INDUCTIVE REASONING
[CHAP.
establishment of a fact which is advocated by the arguer.
In the
example given above, the proponent seeks to establish the existence

of God by means of logic, and he is not directly concerned to prove
the infallibility
of the Upanisads.
But the admission of the infalli-
bility of the Upanisads would directly lead to the establishment of

God's existence, which is the thesis of the proponent.
It is logically
sound and necessary that this variety should be distinguished from

the third variety, since there is a qualitative difference
between the
results achieved by each of them.

(5)
The fifth fallacy of reasoning is failure to culminate in the
proof of the opposite
(yiparyayparyavasna).
For
example the
reasoning "If the lake were possessed of smoke, it would abound in

fish and weeds" is a case of this variety.
absurdnm.
It fails to be a reductio ad
The consequence alleged, namely, the abundance of fish
and weeds is not calculated to invalidate the arguer's thesis, since it

cannot be shown that the opposite proposition denying the incidence
of fish etc.
cannot be alleged to be true.
In fact the most important
and fundamental characteristic of reasoning is that it acts as reductio

ad absurdum, that is to say, it ends in proving that truth lies in the
contradictory of the proposition upheld by the arguer.
In every case
of reasoning and in each and every variety of it, this is the fundamental key-note, and the features which have been made the basis
of this classification are but necessary concomitants of this fundamental character.
It is a truism that in the third and fourth varieties
of fallacious reasoning there is necessarily present the failure of

culmination in the proof of the opposite proposition.
But this does
not mean that the third and fourth varieties are not independent
instances of fallacious reasoning.
There may be a coincidence of
several fallacies in a particular argument, whether it be an independent inference or a parasitical reasoning.

fallacious character of each.
But this cannot annul the
We may, for instance, show
in the
. example given that the second and the third fallacy are also present
and there is nothing repugnant in it.
For the presence of one is not
opposed to that of ocher fallacies.

Srlharsa in the Khandanakbandakhadya
has enunciated another
varieey of the fallacy of reasoning, which he calls the unreality of the
x]
subject (sraysiddhi)1.
I9i
We have seen that the logical ground (pro-
bans) in reasoning is invariably fictitious, but it is assumed on the

admission of the arguer to show the absurd consequences, which
logically follow from it.
The present variety cannot be confound-
ed with the unreality of the logical ground, because the subject

is unreal in addition.
For example, "If the sky-lotus were fragrant,
it would be coveted by all", is an example of the unreal subject.

Not only the ground, fragrance, is unreal but also the subject,
the sky-lotus.
SECTION
II
We have hitherto dealt with the different conditions and factors of

reasoning, and have also shown how the infringement of any one of
these conditions gives rise to several fallacies of reasoning.
In order
to make our treatment of the problem complete, we feel that it is

imperative to consider again the different species of reasoning in order
to show that the absence of the fundamental character of each makes
them fallacious simulations of reasoning.
(i)
Simulation of self-dependence (tmsrayabhasa)
We have already considered the nature of self-dependence, and

how it detracts from the probative force of the argument.
But where
it fails to invalidate the argument employed by the opponent, it has

to be regarded as a false simulation.
Self-dependence acts as an
invalidating weapon when the same thing is found directly to depend

upon its own-self in respect of genesis, subsistence and cognition.
But where a thing is found not to depend upon its own-self, but
upon another individual of the same class, it becomes a case of simulated self-dependence. :
^ -
Let us consider an example of it.
;,
Knowability is a universal
subsistent (kevalnvayi), and is bound to subsist in every knowable

fact, and as such it subsists in its own-self also*
a knowable, just as a table or a chair is.
Knowability is itself
So the attribute of know-
ability subsists -in. knowability itself exactly as the attribute of know''r.1 'ye ca paris: tarkadosh sat svikriyante sraysiddhir anuklatvam %et seq
!
KhKh, p. 1336.
192
INDUCTIVE REASONING
ability subsists in the chair.
[CHAP.
There is no difference between know-
ability and chair, so far as they are knowable and thus possessed of
the attribute knowability.
Now, Is this not a case of self-dependence
in respect of subsistence?
self?
How can knowability subsist in its own-
Apart from a -priori logical consideration, the fact is undeniable
that knowabiiity is also knowable, and as such the attribute of being

knowable belongs to it.
What is then the solution?
In reply it, is
asserted that there is no question of self-dependence here, because the

knowability which is the substantive and the knowability which is
the adjective are not the self-same individual.
In the proposition
knowability is knowable', the predicate can be paraphrased as 'possessed of knowability' and hence it may appear as tautologous.
But
there is no tautology in reality, because the subject is a specific individual determined by its specific reference, whereas the predicate is a
universal attribute, which we shall show, is undetermined by any
external objective reference. *That a chair is knowable means that a
chair is the object of knowledge, and that knowability is knowable
means that it is also an object of knowledge.
But knowledge in the
two cases is not the identical individual.

Knowledge in spite of the common character of being knowledge
subsisting in every instance of it differs according as it has reference
to different objects.
Thus the knowledge of a jar is different from
the knowledge of a chair.
In the case under consideration, namely,
knowability, it is knowable because it is the object of knowledge

having reference to knowability as its object.
The knowability of the
chair is different numerically from the knowability of knowability

itself, because the terms of reference (i.e. the objects concerned, namely, chair and knowability) are numerically different.
Thus there is
no incompatibility in the knowability of the jar being itself knowable,

because , a knowledge of jar and a knowledge of knowability are two
different individuals, and so the knowability of the chair is numerically t different from the knowability of knowability.
Knowability is
nothing but the attribute of being a possible object of knowledge.

And because the knowledge of one object, say chair, is different
from, the knowledge of another object, say table, the knowability
of the chair is also numerically different from the knowability of the
xf
table.
193
When we say that knowability is the universal attribute of all
knowables, we do not mean that the self-same knowability of the

chair is an attribute of its own-self.
The knowability of the chair is
no doubt knowable and hence possessed of the attribute of knowability.
But the knowability which is the attribute of the chair h
not numerically the same knowability which is the attribute of the

knowability
o the chair.
They are as different from each other as
one man is from another man.
The same man cannot mount on his
own shoulder, but another man can.
Though the mounter and the
mounted are both men and belong to the same genus, there is no
incompatibility in it, because the locus and the content are two
different individuals.
- knowability.
The situation is exactly similar tit the case of
When we say that knowability is an attribute of chair
and* knowability is an attribute of knowability we da not offend

against either logic or factuality, just as we do not whet* we say 'man
sits upon man'
There is no repugnance in the latter observation
because the sitting man and the man sitted upon are not the sanie
individualv
Exactly also no exception can be taken to the assertion
'knowability is an attribute of chair and knowability is also an- attribute of knowability'.
Because the two knowabilitiesr one substratum
and the other attribute, are not the self-same individual;, but two
distinct individuals.
If it were meant that the knowabititjr of; the
chair is numerically the same attribute as the knowability of knowability itself, the charge of self-dependence would be unanswerable,
because the same thing cannot be the locus and content of itself.
But as they have been found on examination to be different individuals
the fallacy does not arise.
We have found that knowabifity of an
object differs from the knowability of another object, because of the

difference of the knowledge-act, and one knowledge-act qua individual
differs from another
knowledge-act qua individual owing* to the
difference of the term of reference.

We cannot resist the temptation of adverting to a point which
has become a bone of contention between a section of modern exponents of the Ny.aya-Vaisesika school and Madhusdana SarasvatI and
his followers in connection with the problem of self-dependence.
We
have already alluded to the fact that knowability is a universal subsis-
t$4
INDUCTIVE REASONING
tent (kevalanvayi).
[CHA*.
A universal subsistent is an entity which does
not admic of negation in any context.
Knowability, speakability etc.
are regarded as universal subsistents by the Nyya-Vaisesika school

and also by the Vedntists subject to a qualification1.
is nothing which can be said to be* not knowable.
Now, there
Whatever might
be posited as an instance would become by virtue of the Very fact of

its being affirmed, an object of knowledge.
So knowability is a
universal subsistent being present everywhere and because its negation

is impossible in whatever reference.
In order to make the definition
of a universal subsistent comprehensive the Naiyyika has added a

qualifying adjective to 'negation' [of which the universal subsistent
is affirmed to be the not counter-term (apratiyogi)].
The prima facie
definition of a universal subsistent may be formulated as follows:

" A fact which is not the counter-term of absolute negation is a universal subsistent'V \The term 'absolute negation'is necessary, because
even a universal subsistent is the object of reciprocal negation (in the
current terminology, numerical difference).
Thus though 'know-
ability' is different from 'speakability', and s such is the counterterm of reciprocal negation yet it does not affect its character as a
universal subsistent, because by its definition a universal subsistent
is affirmed to be one which is not the counter-term of absolute negation (and not of reciprocal negation)2.
i
The Vedntist
knowledge.
does not admit that -pure consciousness is an object of
It is the eternal subject and can never be an object.- To be precise
it is pure consciousness undetermined as subject or object.
Even in final realiza-
tion of the Absolute Brahman, Brahman qua object is qualified however detachedly by the act of realization
{vrttyupahita
caitanya).
Unqualified
Brahman is
self-realized and becomes automatically qualified when there is anything
exter-
nal to it including the final act of realization.
*.
:.:.> a
The Nyya-Vaisesika
absolute negation
(anyonybhva),
school affirms four varieties
(atyantZbbava),
pre-negation
of negation, viz.
reciprocal negation or numerical
(pragabbava),
and post-negation
difference
(pradbvamsab-
bva). The negation of colour in air is an instance of absolute negation, because

its negation is not relative to any place or time.
thus absolute.
The negation in question is
The numerical difference of A from B is its reciprocal
It is the negation of the idejntity of two differents.

before its production
is pre-negation
negation.
The negation of the effect
It has no prior time-limit.
It is there
x]
195
We have so far shown the logical necessity and significance of

each of the terms in the definition.
But yet the definition is incom-
plete, because it fails to include many other universal subsistents.

According to the Nyya-Vaisesika school eternal entities, such as space,
time, ether (akasa), atom etc. are self-subsistent, because they do not
stand in need of a locus for their existence.
non-residential reals (avrttipadartha).
locus in respect of
They are therefore styled
Since they do not exist in any
any relation such as conjunction, inherence etc.
their non-existence is universally present.
Therefore the negations
of these eternal entities are to be regarded as universal subsistents.

But the definition, so far as it is worded, fails to include the negations
of these eternal entities within its purview, because they cannot be
regarded as the non-counter terms of absolute negation.
Now the
negation-of-space may be conceived as the counter-term of the negatioa
of negation-of-space.
This negation of negation-of-space is
nothing but. space itself, and so it cannot be contended that this

double negation is a
fiction.
Negation of space therefore being
the counter-term of a real negation cannot be included under the

head of universal subsistents,
though it being
universally present
cannot be denied this status of universal subsistent in spite of the

definition.
A definition must conform to the nature of reality, and it
is preposterous to suppose that reality is to be' sacrificed in deference

to a definition, however authoritative the quarter from which it may
have emanated.
The Nyaya-Yaisesika logician realized this situation
and therefore hastened to qualify the prima facie definition by the

proviso
'that the negation must be resident in a locus' (vrttimat).
The negation of space thus, though it is the counter-term of negation,

from the very beginning of time, if we can say so. But it Is not timeless like
absolute negation, because it ceases to be as soon as the effect comes* into existence. The genesis; of the effect entails the cessation of pre-negation* * The
cessation of the I effect is its post-non-existence. The prefixes'preV am} post*
have reference to the being of the effect. Post-negation has no posterior timelimit because once it occurs there is no time subsequent in which it ceases to be.
If post-negation were to cease at any subsequent time, the result would be the
re-emergence of the effect.- But a defunct effect is never revived. There is no
resurrection of a thing which has perished. Only an eternal entity which docs
not admit of origin or death can subsist for all time,
ig6
INDUCTIVE REASONING
[CHAP.
does not fall outside the scope of the definition, because it is not the
counter-term of a negation resident in a locus.
The negation of
negation-of-space, for instance, is not resident in a locus, because it

is nothing but space itself which as we have shown does not exist in
a locus.
We have already found that
speakafbility etc.
such
attributes as knowabiiity,
are universal subsistents and the subsistence
of
these attributes in themselves does not involve the fallacy of selfdependence.
Now there seems to have been a group of Nyaya-
Vaisesika philosophers who did not admit that self-dependence in

respect of subsistence was a fallacy.
ence etc.
According to them self-depend-
is a fault only in respect of genesis and cognition.
But this
is a clear case of prevarication and is nothing but a tactics of logical

escapism*
The sense of repugnance involved in the same thing being
the locus and content with reference to its own-self is as pronounced

as in the cases of self-dependence in respect of genesis and cognition.
It is inconceivable that an acrobat however skilful can mount on Jbis
own shoulder.
So self-dependence in respect of subsistence is to be
condemned equally with that in respect of genesis and cognition as an

absurdity.
But we have shown that universal subsistents must be
subsistent in every entity including their own selves according to the

terms of die definition.
If knowabiiity etc,
are to be universal
siafcsistentSi they must be shown to be subsistent in themselves-also.

Otherwise they must forfeit their claim to universal subsistence.
We
have shown that knowabiiity etc.
are subsistent in themselves and
also feow this
self-subsistence
does
incompatibility.
not
involve
any
logical
-..
Let us consider the conception of * universal subsistent' of the

Nyya-Vaisesika school and examine the definition formulated by
them in order to find out/ how far it meets the requirements of *thc
situation. This definition In spite of its emended form fails to account
for the charge of self-dependence of knowabiiity and the like in
respect: of subsistence.
The Naiyyika's assertion that self-dependence
in respect of subsistence is not a defect is only an attempt to avoid a

difficult issue
Madhusdana SarasvatI saw the situation squarely
in the face, and has declared that self-dependence in respect of sub-
x]
197
sistence is equally repugnant to logical sense, with self-dependence in

respect of genesis and cognition1.
We have already shown that knowability etc. are universal subsistents without any reservation, and their subsistence in themselves
does not involve any incompatibility.
The Naiyayika's definition
therefore requires a further emendation.
The definition "A universal
subsistent is one which is not the counter-term of absolute negation

resident in a locus"
succeeds in showing that negations of eternal
entities are universal subsistents in spite of their being counter-terms

of negation.
But it fails to account for such recognised universal
subsistents as knowability etc,

-themselves.
hurdle.
in so far as they are subsistent in
The following emendation
will help us to cross the
This emended definition is formulated as follows:]
"The
fact of being a universal subsistent is nothing but the fact of its being
possessed of an attribute, which is not the determinant of the essential
character of a counter-term of an absolute negation resident in a
locus 2 ."
We need not be apologetic for the apparent cumbrousness and
complexity of the definition, though it may appear uncouth to the
uninitiated.
For the situation requires this complex definition, and
close scrutiny makes t necessary.
Let us now apply this definition
to the disputed instances, knowability and the like.
It is undeniable
that knowability is bound to differ with the numerical difference of

the Icnowable object.
The knowable chair is numerically different
from the knowable table and the 'knowability of the chair* is an

attribute numerically different from the knowability of the table', so
far as they are different individual facts.
It is equally undeniable
that 'the knowability of the chair' cannot subsist in itself in the relation of a content and locus, because the same thing cannot be its locus
: and
content at the same time and in the sanie reference.
Thus the
: knowability o the chair' is the counter:term of negation in its own1 yat tu prameyatvadau pratnitatvd anyonyavrttir adosa iti, tan na;
tmsrayditattva-dosena tatr'pi pramitatvsiddheh.
;
"
AS, p. 805. f
2 atyatitbhvapratiyogitnavacchedakameyatvatvavattvt.
See Laghticandrik on AS, p. 805.
198
INDUCTIVE REASONING
[CHAP.
self in respect of subsistence and hence cannot be a universal subsistent.

But this need not cause any difficulty, since the individual attributes
'knowability of chair1 etc.
are not regarded as universal subsistents.
But though the individual attributes are not universal subsistents,

this does not affect 'knowability as such', to be precise, knowability
so far as it is possessed of the essential character of the attribute of
being knowability without reference to the numerical difference of the
individual cases of knowability due to the numerical difference of
terms of reference.
There may be an indefinite number of knowabili-
ties qua the individual characteristics of individual knowables.
But
all these knowabilities do possess the universal common attribute

which makes 'knowability' a knowability.
Knowability in its uni-
versal reference is a universal subsistent, since it is not identified with

the individual attributes, but is an attribute which typifies the universal
-character of knowability that belongs to every instance of it.
Now this universal knowability, that is to say, khowability which
is understood as one embodying this universal character and which is
not identical with this or that instance, which differs owing to the
difference of terms of reference, exists in all cases of knowability.
So
knowability as such is a universal subsistent and subsists in its own self

as well as the rest of knowables.
This does not involve the fallacy of
self-dependence in respect of subsistence as we have shown that such

knowability is not confined to any individual instance. The difference
of the terms of reference of the individual knowabilities only serves to
make these individual instances numerically different, but does not
affect the unity of the universal attribute of being knowability which
is found to be present in all cases.
We have shown that the
'knowability of knpwability' is different from the * knowability of the

chair and the like', and so there is no question of the same thing
being the locus and the content of itself in the same reference.
But
the consideration which makes knowability a universal subsistent is

the fact that however much one may succeed in showing that the
knowability of the chair and the knowability of the table are two
different individual entities and the knowability of the knowability
is a third individual, it is impossible not to take cognizance of the
fact that all these different instances of knowability being themselves
x]
1.99
objects of knowledge do share the character of knowability. The

latter knowability is thus the universal characteristic of all. This
universal knowability is an unindividualized concept, because it
transcends the limitations entailed by the terms of reference. It is
not this or that knowability, but is the essence of all knowabilities
(jeyatvatva).
Hence this universal character being one identical
principle, comprising all the individual instances of knowability within
its range,-and yet transcending them, must be held to be a universal
subsistent. There is no question of self-dependence in regard to this
universal knowability even in respect of subsistence, because it is
different numerically and functionally from the different knowabilities
in which it is immanent.
Thus knowability qua knowability
(jeyatvatvena jeyatvam) is both a transcendent and an immanent
character, because it transcends the individual instances of knowability,
in so far as they are circumscribed by their terms of reference, and yet
it is immanent in all such cases, because its presence in each.and all
of them is undeniable.
To sum up. The knowability as determined by the terms of
reference as the knowability of chair or the knowability of the table
is not a universal subsistent. These are determinate characteristics of
determinate objects and as such belong only to their respective terms
of reference. The knowability which is a universal subsistent is a
universal characteristic of all knowables and as such is undetermined
by terms o reference, though it belongs to each of them. It can be
determined only as an attribute which has the characteristic of being
knowability ( jeyatvatvena jeyatva ).
It is this attribute which
belongs to all knowables including their specific knowabilities, which
vary with the individual knowables. And because this universal
knowability is different and distinct from the specific knowabilities, its
subsistence in each and all of them does not involve the fallacy of selfdependence. It exists in all that exists-all things and their attributes
including the attribute of knowability, which qua knowable is on the
same footing with the substantives. It is numerically a self-identical
characteristic, and because it is the qualifying adjective of all entities
it is held to be a universal subsistent. It is immanent in all because
it transcends each. The specific knowabilities of specific knowables
lo
INDUCTIVE REASONING
[CHAP.
are however particular attributes subsisting only in particular instances.

The knowability
as the universal characteristic of all knowables is
definable not in terms of their
objects, but by its intrinsic character of
being possessed of the attribute of being knowability.
The knowability
which has the character of being knowability (jneyatvatva) is obviously

distinct from an individual knowability which has the character of
bsing the knowability of a particular
knowable, say a chair or a table.
It must be borne in mind that it is this knowability
undetermined by
its objective reference which is a universal subsistent.

We regret we have had to make this lengthy digression while
examining the nature of self-dependence as a fallacy. But this was
necessitated, by the reluctance of a section
of Naiyyikas to regard
self-dependence in respect of subsistence as a fallacy.
We have found
that this reluctance is unjustified and that self-dependence in respect

of subsistence is as much a logical offence as it is in respect of genesis
and cognition.
We therefore assert that the position of the author of
the Nyymrta denying self-dependence in respect of subsistence as a

case of fallacy is untenable1. It is refreshing to observe that Rmarudra,
the author of the Tarangint,
the commentary on the Dinakari, has
recognised this truth and has adopted the same line of defence as
Madhusdana Sarasvat2
and Gauda Brahmnanda, the author of the
Lagbucandrik.
(2)
~
Simulation of mutual dependence (anyonyasraybbasa)
The charge of mutual dependence becomes false when there is
difference of individuals.
It is true that a cognition when it becomes
defunct leaves a trace or impression (samskra) in the unconscious

mind; which is revived in the form of a memory-act in the presence
of a stimulant.
latter cognition
This
memory-act is a species of cognition.
The
depends on another cognition for its genesis and
thus the case might be urged as a case of mutual dependence.
But
the charge would be false, because the generator cognition and the
1 ynt tu prameyatvdau pramitatvd anyonyavrttir adosa iti.
'
* AS, p. 805.
2 vrttimadabhvpratiyogitvarn na kevalnvayitvam...kintu vrttimadabhtvapratiyogitnavacchedakadharmavattvam eva tath.
Rarnarudr't, p. 101.
x]
generated
cognition are two different individuals.
01
Were it the
case that the cognition which generated the memory-trace, and the
cognition
which is generated by the latter were a self-identical
individual, the charge would be real.
But as this is not the case
the accusation of mutual dependence in respect of the instance cited*

falls to the ground.
This fallacy of reasoning is called simulated
mutual dependence.
(3)
Simulation of vicious circle (cakrakabbasd)
The case is analogous with the previous one.
A circular reason-
ing is vicious when the argument moves in a circle and culminates

in showing that the cause and the effect, the locus and the content,
the subject and object are identical.
But if the individual cause
and the individual effect are different, then this charge will not
hold good.
For instance, the seed produces sprout, the sprout
produces the plant, and the plant produces the seed.
Though the
circle of cause and effect starts with the seed as the cause, and
terminates in the seed as the effect, and thus is a closed circle, yet
the charge of vicious circle cannot be brought home, because the
seed which is the ultimate product, and the seed which is the initial
cause, are numerically different.
The charge of vicious circle against
this statement of fact thus transpires to be a fallacious simulation.

(4)
Simulation of the vicious infinite (anavasthbbasa) -
When the infinite series is legitimate and valid, the charge that
it is vicious becomes an example of simulation of it*
We have
already discussed at length the nature o the infinite series and have
found that there are two varieties, one legitimate and valid and
the other illegitimate and invalid.
For instance, when the relation
of causality is firmly established between two terms independently

of the past and future course of events, the mere fact that it is a
link in the infinite chain of the cosmic process, does not make the
causal relation between the terms under consideration invalid.
We
have already shown that the fallacies of reasoning beginning with

self-dependence and ending in the infinite series, arise only when
there is dependence of one situation upon others in respect of genesis,
cognition, and subsistence.
fallacies become unreal.
26
If there be no such dependence the
To take any instances of causal relation,
2O2
INDUCTIVE REASONING
[CHA>5
say the seed and sprout, it is not open to doubt that the seed is the
cause of the sprout, and the sprout is the effect of the seed. But on
further consideration it is found that the seed is also an event in
time and as such must have a cause of its own, and that cause
also cannot but be an event, because an eternally existent cause
would produce an eternally existent effect. W e know from experience that the seed which produces a sprout was produced in its turn
by a sprout in the past, and the sprout again was produced by a seed*And if we continue our enquiry we shall find that the causal relationship is extended over the infinite past history of the cosmic process.
Furthermore, if we sedulously pursue our enquiry into the
future course of the seed's history, we must find that the causal
relationship of the seed and the sprout is bound to go on uninterrupted as in the past. Here history is found continually to repeat
itself. This infinite series, both recessive and progressive cannot be
regarded as an invalidating charge, because it does not make the
causal relationship between the seed and the sprout unreal. It
is .true that the seed has an infinite history in the past, and will
have an infinite career in the future, unless the vegetable world is
to become extinct. But however exacting the attempt may be to
understand this unceasing chain of causal relationship, it does not
invalidate the relationship between a particular cause and a particular effect, because the particular cause does not depend upon the
particular effect either for its genesis, or subsistence, or cognition.
The fact that it depends upon another cause which is analogous to
the effect does not make them mutually dependent, because each
term, in the series is numerically different from the other. But if it
is found that a particular term cannot come into being unless a series
of terms can be brought into existence, and this series is bound to
be infinite and unending, then the charge of vicious infinite series
-will be legitimate.
: (5) Simulation of reductio ad absurdum (bdhitrtbaprasan gab basa)
:- .Reasoning, as we have defined it, must be a reductio act
absurdum*
In other words, it must show that the opponent's
position is absurd and the truth lies in the opposite proposition.
This is the universal characteristic of reasoning. If it is found thac
x]
203
the consequence alleged by a reasoning to follow is not opposed to

the position of the opponent, it will be a false simulation of reasoning.
We need not regard it as a distinct species as- it is the common
characteristic of all varieties of reasoning.
(6)
Simulation of contradiction (vyaghtbhsa)
The proposition "The mother is barren" is an instance of contradiction.
It is the most decisive weapon of refutation.
contradiction may be relative.
But this
The woman who is barren may be-
come a mother by virtue of medical treatment and so the contradiction

between motherhood and barrenness need not be absolute.
Thus
the charge of contradiction, when one of the opposite characteristics

ceases to exist, becomes a simulation of contradiction.
(7)
Simulation of recrimination or 'tu qttoque
(fratibandybhasa)
The tu quoque argument becomes false, when the reason assigned*

by the opponent is shown to be untenable.
Suppose for instance, the
Vedntist contends that the phenomenal world is unreal, because it is

perceptible, and the. dualist alleges "that the concomitance of perceptibility with unreality is due to aa unknown condition and so the
unreality of the world does not follow.: If the Vedntist contends
in reply that if an unknown condition makes the concomitance contingent, then no inference will be possible.
Even the concomitance of
smoke with fire may be urged as due to an undetected condition.

This is a tu quo que argument and will remain valid if the opponent
fails to establish that the suspicion of an undetected condition in the
concomitance of. smoke" and fire is unfounded.
If the opponent suc-
ceeds in showing the impossibility of the undetected condition, his

refutation will remain valid, unless and until the Vedntist shows the
hollowness of the charge of undetected condition in his argument.
;>
W e have shown how reasoning in its different varieties fails

to act as an instrument of refutation owing to the lack of one or the
other of the: essential elements and thus becomes fallacious.
The
fallacies of the recognised types of reasoning have been demonstrated

with more or less elaboration.
As regards the additional varieties of
reasoning which have been recorded, by Sriharsa and the . logicians of

Rmnuja school, we have given, an elaborate exposition of. their
nature.
We do not propose to undertake an elaborate treatment of
204
INDUCTIVE REASONING
[CHAP.
the fallacies of these varieties, as in our exposition of the character of

each of them we have given sufficient indication of the conditions which
make them effective instances of refutation and it is obvious that the
infringement of these severally and jointly will make them fallacious
and ineffectual.
Thus, Absence of Crucial Evidence will transpire to
be fallacious if such evidence can be shown to be available.
Empirical
induction will be set aside by the discovery of a contrary instance.

Complexity will not be a defect if it can be shown to be entailed by
unimpeachable
logical
necessity and similarly simplicity will not
operate as a helpful consideration and as a deterrent against the

opposite theory if it is found on enquiry to be inadequate
upon insufficient data.
and based
The charge of Impertinence or Impudence
will fail to be brought home if the series of queries

honest speculative demand.
is inspired by an
The charge of Equalisation will be inef-
fective if the claim of equal advantage can be shown to be false.

Vedntist does not agree that the reality
The
of the phenomenal world is
established by incontestable logical proof and the claim of equal advantage is set aside by a powerful array of arguments.
The fallacy
of the Double Noose becomes apparent when it can be shown that

the alternatives are not contradictorily opposed or that the predicates
of the alternative propositions are repugnant to our logical sense Thus
the dilemma "Is unreality of the world true or false?" is vitiated by
contradiction and hence cannot be entertained as a legitimate case of
opposition.
It is vitiated by a transparent sophism*
Where the
predicate is repugnant by its very nature to the connotation of the

subject the dilemma is to be dismissed on the ground of absurdity..
- The raison d'etre of reasoning operating as a deterrent consideration against the opponent's position is that it operates as a redactio ad
absurdum.
It. shows that the.,admission of a position entails the
admission of an absurdity.
If the consequence alleged to follow is
foind to be sanctioned by logical evidence and is not offensive to the

opponent s standpoint and general philosophical persuasion, the reasoning: becomes innocuous and this constitutes its fallacious character.
T h e fallacy o a reasoning lies in the fact that it does not serve as a*
confutation which it claims to be.
claim is the proof of its fallacy.
The proof of the falsity of Its

,
CHAPTER XL . .
In syllogistic argument the proponent first asserts a thesis and next
assigns a reason in support of it. But this assignation of reason is
liable to be vitiated by fourfold defect1. In the first place, the reason
itself may be fallacious. For instance, the reason may be inconclusive
or non-existent in the subject or contradictory. In such cases the
argument fails in its purpose2. In the second place, the argument
may be defective in respect of example in which the concomitance of
the probans and the probandum is usually demonstrated. If the
example is found to be destitute of the probans or the probandum or
both, the concomitance of the probans with the probandum is not
capable of being established, and hence the reason assigned will fail
to establish the desired conclusion3. In the third place, the, reason
may be vitiated by a defective statement of the example. For
instance, it the concomitance in,.agreement is not stated at all or stated
in the contrary order in the example, the reason becomes defective
because of the impossibility of the determination of the necessary
concomitance4. In the fourth place, the argument may be assailed
by a reductio ad absurdum .such as self-dependence, mutual dependence
and the like 5 .
i
If we look closely into the nature of these defects, it will be found
that all the four varieties of defect are nothing but cases of fallacious
reason {betvabhasd) and thus the fourfold classification of defects of
* i " sdhaiiapryogasya caturdha dustatv^m sambhavati.
'
NPa, p. 125. See also VV, p. 19.
2 hetuto yath anaikntikam itydi.
.
;-.;
;.,.,.;,.
i : : ^ . \ r > . i \ .
'
5
:..-
Ibid,
>
, . : - [
'. '.. . ,._
.;
". -
drstutato yath sdhanavikalam itydi,
"' ,''_.,,"
; . . k
.-
.V ; - , "
..'...
Ibid,
. '_ [
\ \
'
'î.^..'-[
uktito yath anupadarlitnvayaviparyayopadarsitnvaydi.
'
'
'
'
'
'
'
tarkato yathf tmasrayetaretarsraydi.

..
r y l . -: ;.:.;..'.
Ibid.
\.
..
; .
\r.
20$
INDUCTIVE REASONING
[CHAP.
argument possesses only procedural and methodological importance and

cannot be justified by reason of logical difference.
Their ultimate
logical value rests on the fact that they are all capable of being
dissolved into one or other types of fallacies of reason.
So far as the first variety is concerned it is openly a case of logical
fallacy.
In the second type the defect of the example serves to show
that the probans is not necessarily concomitant with the probandum.

The demonstration of this defect by the opponent is tantamount to the
assertion of the fallacy called the non-concomitant reason (vypyatvsiddhi).
The third type of defect consists in the non-statement or
misstatement of necessary concomitance and since it fails to convince

the opponent of the necessary concomitance of the probans and the
probandum, It becomes equivalent to the assertion that the probans
employed is lacking in necessary concomitance, and this is nothing but
a case of the fallacy of non-concomitant reason.
The third variety is
thus on a par with the second.

Let us examine the logical value of the fourth defect.
argument is rebutted by a reductio ad absardum
When an
and the rebuttal
stands unanswered, it only serves to show that the argument under

review is guilty either of the rejection of an attested truth (prmnikaparityga) or the admission of an absurd position
(apramanikasvtkara).
The rejection of an attested truth involves contradiction of logical

proofs and the admission of an absurdity amounts to the confession of
lack of logical sanction.
These two defects will be found on examina-
tion to be nothing but cases of logical fallacy.
The contradiction of
logical proofs only shows that the probandum is non-existent in the

subject and this is called the fallacy of contradicted reason (badha).
The lack of logical sanction implies that the probans is invalid either
for the lack of relation with the subject, or lack of necessary concomitance with the probandum (asiddhd).
Thus the re duetto ad absutdum
only ends in showing the presence of either of the logical fallacies just
mentioned.
W e may legitimately surmise that the omission of the
demonstration of the classification of reasoning on the part of the

author of the Nyyastra and the scholiast Vatsyayana was perhaps
due to the recognition of the truth that the employment of a redactio
ad absmdum
is nothing but a way of demonstration of a logical
xi]
07
fallacy, A reductio ad absurdem serves to function as a weapon of confutation only because it derives its logical cogency from that of a logical
fallacy and for this we need not postulate an independent basis.
The
logical value that a reductio ad absurdum possesses is borrowed from that

of a logical fallacy of which it is only an indirect way of demonstration.
Sakara Misra in his work Vdivinoda has not recognised contradiction (vygbata) as a variety of reductio
ad absurdum
and this is
in conformity with the views of Udayana and Gagesa1.
Reductio
ad absurdum is founded upon the necessary concomitance of the

ground and the consequent and this concomitance if called in question
would make resort to another reductio
ad absurdum
inevitable and
the process has got to be repeated until all possibility of doubt of the
necessity of the universal concomitance which is the basis of reductio
ad absurdum
is eliminated with final certitude. -The final solvent
of doubt regarding necessary concomitance cannot be a reductio ad

absurdumwhich
rather presupposes some other ultimate necessary
concomitance as its condition.
The charge of infinite regress is
avoided by Udayana and Gagesa by the plea that the ultimate

solvent of all doubt is contradiction (yyghta).
Contradiction is a
self-evident ultimate fact and does not depend on any other ulterior
consideration for its validity. The contention of Sriharsa that contradiction depends for its validity upon another necessary concomitance
is dismissed as a sophism on the ground that there is no psychological
evidence of its ever being assailed by a doubt of its validity.
Con-
tradiction is thus the final arbiter than which nothing more ultimate
is logically and psychologically conceivable2.
Sakara Misra
has
regarded contradiction as the third variety of censure (nigrahasthana)

called contradiction of the thesis
(fratijvirodha)3.
vyghto 'pi a tarkavisesah kintu virodhanigrahasthnntarbhtah.
..:,
. -
--..I ; - , . * *
> _ - ^ - *
'
i . ' . . : . " . ,
y V , p . 19,:
2" Cf. supra, chapter iii.

3 There are twenty two varieties of censure, recognised in the Nyayasittra
of which* the logical fallacies (hetvbbsa) are the last variety. The difference
between a logical fallacy and other types of censure lies in the fact that the
latter indicate personal aberration and the former points to a logical defect.
Personal aberration rather shows that the arguer is not endowed with the proper
psychological discipline necessary for the successful tackling of logical problems*
2o8
INDUCTIVE REASONING
(CHA*.
The elaboration of the different varieties of reasoning and of the

defects of the example is a later development in the logical speculations.
There is not the slightest indication of these defects either in
the Nyyastra or in the B hasy a of Vatsyayana*
We have already
observed that these defects are ultimately traceable to one or the other
logical fallacy recognised by the author of the Nyyastra
innovations
v a l u e .
sponsored
by
later logicians
have only a
and the
procedural
Now, a question has been raised.
In view of the fact that the
defects of an example and the defects entailed by a reductio ad absurdum are reducible to one or the other recognised type of logical
fallacy, what would be the legitimate procedure in the demonstration
of defect in the opponent's argument ?
The question is whether we
should demonstrate the defect in the argument by specifying the

logical fallacy involved, or by naming the type of reductio ad absurdumy or the specific defect of example.
Should we for instance, point
out that the example adduced by the opponent is defective in this

or that respect, or that it involves a case of self-dependence and the
like ?
Or should we name the fallacy involved by these defects
directly and straightaway ?
The question is an intriguing one.
If
the defect is to be demonstrated by naming the logical fallacy involved,

the elaborate discourse on reasoning and defects of example becomes
pointless.
Besides the convention of philosophers shows that the
defects- should rather be demonstrated by the specific enumeration'of

the type of reductio ad absurdum or of the defect of example.
We
find that this latter procedure has been followed in such an ancient
, and respectable work as the Mabbhsya of Patajali.
Udayana expressly declares in his Nyayaparisista
that though
reductio ad absurdum and defects of example are in their essence cases

of. logical fallacies, the defects should rather be shown by the specific
mention of the type of reductio ad absurdum etc. 1
exponents of the Nyyastra
There were other
who held that the defects in question
are not defects of reason* but should rather be regarded as additional

i See NPa, pp. 125-26; see also VV, p. 37: ete dosh svasabdenaiVodbhvyh, na tu hetvbhsataye'ti.,.ity Udayancryh.
xi]
instances of censure1.
209
The admission that there are other varieties
of censure over and above the twentytwo types expressly stated in the
Nyyastra1,
does not involve disloyalty to the author of the Nyya-
stra, inasmuch as the aphorism may be legitimately regarded as

illustrative and not an exhaustive statement of fact.
We feel inclined
to accept the interpretation of Udayana as the correct appraisal of

the value of the reductio ad absurdum and of the defects of example.
It does not seem to be fair to regard these defects as non-logical
aberrations and we shall show immediately the close affinity and
kinship of reductio ad absurdurn with accredited inference.
It seems
to be a sound attitude to regard these-defects as logical in character

and Udayana has shown how these defects derive their capacity as
instruments of confutation from the logical fallacies involved by them.
When however the reductio ad absurdum or the alleged defects, of ex*
ample transpire to be false simulations, they should be included under
the head of sophism (jtyuttara).
To sum up :sAn authentic reductio ad
absurdum is a case of logical fallacy and a spurious imitation of it is

a case of sophism.
If we brush aside all considerations of allegiance to the tradition of
logicians and set ourselves to examine dispassionately the logical value
of reasoning as an instrument of confirmation and refutation, we
cannot but admit, the d o s e affinity of reasoning with inference.'
The
logical apparatus which makes inference possible is closely analogous

if not identical with- that presupposed by reasoning. * The universal
proposition and the minor premise are the conditions of inference and
these very conditions are the presupposition of reasoning also.
essential difference between inference and reasoning is this.
The
Inference
presupposes a veridical cognition of the universal proposition and the .

minor premise, whereas
reasoning
proceeds from the opponent's
assertion of the universal proposition and the minor premise as the data.
The assertion of the opponent is not based upon veridical cognition
and reasoning as a reductio
ad absurdum
seeks to prove that the
1 ete svatantr eva dosh...nigrahasthnatvena boddhavy ity anye. ~
yy, P . 3 7 .
2
a
Vide NS, V. ii, x.
2i
INDUCTIVE REASONING
[CHAP.
admission of a false major or a false minor premise by the opponent

compels him to endorse an absurd conclusion.
There is also a difference between inference and reasoning in the
result achieved. Inference seeks to establish a conclusion and redttctio
ad absurdum seeks to prove the absurdity involved in the opponent's
thesis. But this difference in the result achieved is irrelevant to the
determination of the logical status of reasoning. As has been shown
by the Mdhva logicians reasoning and inference stand on the same
footing, because both derive their probative force from the major and
the minor premises. The consideration that the minor premise is
assumed in reasoning and is believed to be true in inference is rather
the proof of a psychological difference and thus should be regarded
as entirely irrelevant to the determination of the logical status of
reasoning.
The truth of the conclusion in inference and the falsity of the
conclusion in reasoning should only be regarded as the distinctive
criterion of them. But this distinction can at most prove that they
are two varieties of the same logical function and it should not be
made the warrant for relegating reasoning to a different category.
Whatever may be the difference of views between the different
schools of logicians regarding the ultimarle status of reasoning, it is
undeniable that it is exactly on a par with inference, so far as the
conditions are taken into consideration. The infringement of the
conditions in inference gives rise to different fallacies.and the position
is entirely analogous in reasoning inasmuch as the infringement of the
conditions of reasoning also gives rise to similar defects which have
been called, false simulations of reasoning* But apart from the difference of nomenclature the intrinsic difference of the fallacies of inference from the fallacies of reasoning is not discernible. It would
be rather fair and legitimate to maintain that they are essentially the
same. Thus, for instance, the fallacies of reasoning, namely, the
unreal subject, the lack of necessary concomitance and the establishment of a conclusion endorsed by the opponent can all be subsumed
under the logical fallacy called the nonrexistence of the probans
(astddbi).
The. Naiyayikas propound three different varieties of the non-
xi]
211
existent probans, namely, (1) non-existent in respect of the subject

(sraysiddha),
(2) non-existent in respect of necessary concomitance
(yyapyatvasiddba),
and (3) non-existent per se (svarpsiddba).
The
fallacy of reasoning called the non-existence of the subject is clearly

a case of the first fallacy.
The lack of logical sanction
is not any-
thing different from the fallacy of reason called 'non-existent in respect

of necessary concomitance* (vypyatvasiddh).
When there is lack
of necessary concomitance between the probans and the probandum,

the probans fails to prove the probandum because of the lack of;
necessary concomitance between the two and when there is the same
lack of necessary concomitance between the ground and the consequent the ground fails to entail the consequent for the self-same
reason.
The case is exactly similar but for the terminological differ-
ence, which is a creation of the logicians, that one is called by the

name of 'the non-concomitant reason* and another by the name of 'lack
pi4 logical sanction* {mlasaithilya)*
As regards the fallacy of reason-
ing called 'the* establishment of an issue endorsed by the opponent*

(jstapdana) it is not different in kind and import from 'the proof of
an admitted truth* (siddhasadhana).
When an argument ends in proving a conclusion which was not
disputed by the opponent, it is said to be a waste of labour.
no use to. seek to prove the obvious.
It is hi
This fallacy is not given a.
separate status but is regarded as an instance of subjectless reason

(sraysiddba hetu).
The logical subject or the minor term must be
one which is not known to be possessed of the probandum antecedently to the employment of the syllogism.
This is called the
essential character of the subject (paksat) and is the universal condition of inference,, subjective or syllogistic.
If the opponent were
already aware oi the subject being possessed of the probandum and

did not entertaia, any doubt about it, then the argument would end
in proving a proved fact.
The raison d'etre of the fallacy lies in the
fact that the subject of inference lacks in the essential condition that
it must not be antecedently known to be possessed of the probandum.
And this means that the logical subject is wanting and so this would
be a case of subjectless reason or what is called *a reason non-existent
in respect of the subject*.
2i2
INDUCTIVE REASONING
[CHAP.
The fallacy of reasoning called 'the establishment of an issue endorsed by the subject' is exactly on a par with the fallacy called 'the
proof of an admitted truth'.
It should therefore be classed under 'the
subjectless reason*.
As regards the fallacy of reasoning called 'mutual contradiction1,
it is not anything materially different from what is called 'the countermanded reason [satpratipaksita hetu).
When one reason is opposed by
another reason and both are possessed of co-equal strength, both the
arguments become, ineffective.
The result is the same when one
reasoning is opposed by another reasoning of equal cogency.
Thus
this fallacy of reasoning is found to be the same as the fallacy of the

countermanded probans because the principle of mutual opposition is
present in them in the same fashion.
The" Naiyyikas have recognised the following three fallacies of
reason, viz.
(i)
reason (yiraddha),
contradicted
reason {badbita),
(2)
and (J) non-concomitant reason
contradictory
(savyabhicara).
All these three fallacies are also possible in reasoning.

consider the question of contradicted reason last.
We shall
As regards the
contradictory reason it arises when the reason employed ends in proving the contradictory of the intended probandum.
The fallacy of
reasoning called 'the failure to culminate in the proof of the opposite'

is exactly analogous to this.
Reasoning seeks to prove that the
opponent's position is wrong and by implication that the truth lies

in the opposite proposition.
When the reasoning fails to prove that
the opposite proposition is true, it implies that the opponent's thesis

is correct.
The failure is due to the employment of a ground
(padaka), which is concomitant with the opposite of the consequent.

This is also the case with the fallacy of reason called the contradictory
reason, which is so called because it proves the contradictory of the
intended probandum, being concomitant with the opposite of the
probandum.
So they are not two fallacies,, but one at bottom.
As
regards 'the non-concomitant reason' it arises from the lack of necessary concomitance between the probans and the probandum.
The
fallacy of reasoning called 'lack of logical sanction', which we have

just * considered, is exactly a case of this lack of necessary concomitance.
xi]
213
As regards 'the contradicted reason* it arises when the probandum

is found to be absent in the subject by another uncontradicted organ
of knowledge.
The conclusion "Fire is cold*' is an instance of it.
The contradiction of the probandum by an independent source of

knowledge proves that the probans employed is fallacious.
In reason-
ing, the consequent entailed by the ground must be absent in the

subject, otherwise it would not operate as a rednctio ad absurdnm.
So the fallacy of the contradicted probans arising from the contradiction of the probandum cannot be a fallacy of reasoning.
The ortho-
dox Naiyayika has drawn pointed attention to this difference between

inference and reasoning which he regards as fundamental.
And it
is because of this difference reasoning in their opinion assumes an

unreal character.
We have shown in the exposition of the Mdhva
standpoint how the charge of unreality is repudiated by the exponents

of the, Madhva school of logic.
The difference of views on the logical
status of reasoning either as inference or as an independent organ has

been elaborately considered by us.
But we do not bank upon this
disputable basis in our attempt to show the kinship of reasoning with

inference.
Whatever may be the differences of philosophers regarding
the ultimate logical status of reasoning it. is indisputable that the

kinship of reasoning with inference is intimate and profound.
As regards the fallacy of the contradicted reason it is not the case
that' there is no corresponding fallacy of reasoning.
It must be ad-
mitted that all cases of reasoning are not cases of reductio ad absur*
dum in the sense of directly aiming at negative results.
Empirical
induction is a variety of reasoning and we have shown that its finding

is positive.
And those who think that reasoning culminates in the
creation of presumption are bound to admit that reasoning acts as a

positive organ also.. Here contradiction will be fatal to reasoning.
So the three bases of logical fallacies are found to be present in fallacious reasoning also.
There are some logicians who allege another fallacy of reasoning
called 'equality of consequences', when mutually opposed consequences
can be alleged to follow from a position advocated by an opponent.
It will show that the objection of the opponent is met by a counterobjection and the absurdity alleged is common to both the parties.
2i4
INDUCTIVE REASONING
[CHAP.
We may regard c without offence to logical propriety as a case cf

mutual opposition, which is a recognised fallacy of reason {satpratipa*
ksitatva.)
The close kinship of the fallacies of reasoning with those of
inference, which we have shown, should create a presumption in favour
of the position that reasoning is rather a case of inference.
The
Madhvas have proved their identity and barring the differences we have
shown,
which are more psychological than logical, there is no very
convincing reason why reasoning should be denied the same or similar

logical status with inference.
. CHAPTER XII
The development of logical speculations in the school of Navyanyya is indebted to a prodigious extent to the hostile criticisms of
Srlharsa of the Nyya concepts in that bewildering work known as the
Khandanakhandakhdya.
The predominant interest of this work lies
in destructive criticism of the realistic categories.
It is the postulate
of the philosophers of the Nylya school that reals are knowable and
the knowledge of these reals'can be definitely set forth in definitions^
From the very beginning of Nyya speculations as recorded in the
Nyyastra and Vtsyyana's Bhsya, not to speak of the supplementary
literature which gathered round these twa basal works in the form of
exegesis, we find that tremendous importance has been attached taf
definition as the initial focus of logical elaboration.
Sriharsa who flourished in the 12th century A . D .
took upon
himself the task of exposing the hollowness of the pretensions of the

logicians of the Nyya school that the phenomenal objects have got
well-defined logical character, each distinct and different from one
another and that these characters can be exhaustively analysed and
intelligibly-formulated in verbal propositions.
Sriharsa's critical evaluation of the problem of reasoning a s a
logical weapon deserves serious consideration and we address ourselves
to this task, although we are painfully conscious of the fact that it may
result in upsetting the whole fabric that we have built in the preceding
chapters.
Our apology is this.
W e have set ourselves the task of
evaluating the logical speculations of the different schools of philosophy on reasoning as a logical category and we might run the risk
of being accused of partiality or prejudice or inadequacy at any rate,
if we omitted the consideration of Srlharsa's views on this subject.
This omission might also be interpreted as the tactics of a shirker who
deliberately omitted a difficult task.
Whatever may be the verdict of
scholars who would honour my work with a perusal, I felt that this
omission of the evaluation of Sriharsa's
speculations on reasoning
2i6
INDUCTIVE REASONING
[CHAP.
would be a drawback, so far as the thoroughness of the exposition is

concerned.
I am aware of the difficulty of the task and the intrct^
ability of Sriharsa's language and thought.
But I am making an
honest attempt in explaining SrTharsa in a language which I shall try

my utmost to make intelligible to the modern mind.
It must be
recognised that however one may try to make understandable the treatment of an abstruse problem, one cannot hope to be successful, unless
critical reflection is brought to bear upon it by the student concerned.
Reasoning is defined to be the logical entailment of the determinant
concomitant (major term or the consequent) with reference to a person
who endorses the determinate concomitant (middle term or the logical
ground).
Logical entailment is interpreted to be the logical necessity

But this definition is too narrow.1
of admitting the consequent.
The
logicians of the Nyya-Vaisesika school admit two types of reasoning,

viz. ( i ) consisting
following
from
to refutation:
in demonstrating
the position
in
an
one word it is the
reductio- ad absurdum,
illogical issue, necessarily
admitted by the opponent with a view
(2)
type
of
reasoning called
There is another type which consists
in stating a hypothetical possibility which is however based upon

necessary concomitance of two facts.
For instance, a hungry and
thirsty traveller is found to make -such assertions, **P could satisfy

my hunger, if there was
"1 could
allay
my
a confectioner's
thirst, if
shop at hand",
drink". -These assertions are of a hypothetical

regarded by the logicians
or
I could get a glass of ice-cream to

nature
as a species of reasoning.
and are
Obviously
these assertions do not aim at the refutation of a position and so the

definition of; reasoning under consideration which applies to reductio
ad absurdum does not cover these hypothetical assertions in spite of
the fact that they are regarded as authentic Ilustracin of a species
of reasonings
This is the first charge against the definition given
that it fails to meet the requirements of the two types.;

The lack of a common * definition
The definition as worded is also too wide.
is - not the only
defect.
It applies to the case
t abhyupagatavypyam prati vypakaprasanjanam sah tatprasajanatn ca

svikrrhatbodhanam iti cet? na, avypteh
KhKh, pp. 1276-77.
xiij
of fallacious reasoning called the entailment

by the opponent (istpdana)1.
To
the
of
217 ,
an
issue endorsed
opponent
who admits
the truth of both the determinant and the determinate concomitants

such entailment will be a case of wasted labour.
illustration.
To take a concrete
Suppose the opponent does not dispute that the smoking
hill is possessed of fire, yet the proponent argues, "If you think the
hill to be devoid of fire, you must admit it to be devoid of smoke
also".
The opponent does not deny the existence of fire and so the
entailment of the absence of smoke on the supposed denial of fire

does not affect the position of the opponent and thus fails to be a
reductio ad absurdum.
In order to avoid this contingency the
logicians may propose to add a qualifying proviso, namely, 'who does

not endors the
determinant concomitant*.
The definition would
therefore be 'the entailment of a determinant concomitant with reference to a person who endorses the determinate concomitant, but does
not endorse the determinant concomitant*.
will not be free from difficulty2.
But even this amendment'
Suppose the opponent admits the
determinate concomitant, the presence of smoke, but does not admit

the presence of fireyyet the entailment of fire oh the ground of the
presence of fuel, which is not necessarily conconiitant with fire, would
be. a case of valid reasoning in terms of the definition* although it is
not entitled to it.
In order;to escape the predicament the definition
may be further proposed to be amended by the insertion of the clause

'by means of a'determinate concomitant*.
scrutiny, because of the
dilemma 3 :.
But this also will not stand

(1) "Does
it
mean
that
the determinate and the determinant, concomitants are respectively

admitted and declined per se without reference to the relation of
necessary ' concomitance ; which . factually - subsists
between -them?,
(2) Or does it mean that the acceptance and denial have reference
to their ; character as determinate and determinant in concomitance4?."
1 istpdane'pi gatatvc ca.
KhKh^ p. 1277.
i:
- 2 anabhyupagtavypakana ity api'ti cet? na/
Ibid.^
:
"^ 3'S vypyeaetjr, api klryam iti cet? na. vikalpsahatvr. *
- '
Ibid.
4 kim paramrthato vypyavypakabhvavyavastliitayoh svarpenesta-nista^
tvam uta vypyavypakaycc bhvena tat ?,
Ibid.
2i
INDUCTIVE RASONNG
|>IAI>-
The first alternative is not tenable. Suppose the opponent admits

the presence of smoke, but declines the presence of fire, because of
his ignorance of the necessary concomitance of smoke with fire,
though it is a fact that the concomitance is true. In this situation
the entailment of fire on the presence of smoke with reference to
the opponent would pass for a genuine reductio ad absurdttm,
although it ought not to do so, because the opponent will remain
unconvinced and so will not be compelled to admit the consequent.
Smoke is known by the opponent as smoke and nothing more and
fire is known only as fire and not qua determinant concomitant.
The entailment of fire by ; means of smoke will be as ineffective as
the entailment of any other circumstance, say a pen, because the
compelling consideration in the shape of necessary concomitance is
absent with regard to him. The result will also be the same if the
terms were known reversely as determinate and determinant. For
instance, if smoke were known as determinant and firas determinate,
the entailment of fire on the ground of smoke would fail to produce
conviction. These cases cannot be regarded as reductio ad absttrdum,
because the necessary concomitance is not endorsed by the opponent
foe. vahse, conviction. the jeductio ad absurdam is applied. The
factuality of the concomitance will not make it a case of reasoning,
unless it is driven home to the opponent by an independent argument.
The - success of the reductio ad absurdem invariably hinges on the
production of conviction in the opponent of the necessity of admission
of the consequent following upon the admission of the. ground. But
this.can.be achieved only if necessary concomitance can.be proved
to be factually present. It is on the proof and conviction and not
the mere presence of necessary concomitance that the success of
reductio ad absurd um depends 1 .
The second alternative will also be found to be in the same
predicament. The definition as amended stands as follows: "Reasoning
is the entailment of the determinant by means of a determinate known
as such with reference to a person who endorses the determinate but
does not endorse the determinant". But it overlaps the fallacious
i
KhKh, p. 1277.
XII]
reasoning called suicidal (pratiklatva)
219
which arises when the conse-
quent entailed is hostile to the position of the proponent himself.

Take for instance the following argument in which the proponent and
the opponent both endorse the determinate qua determinate, but
equally repudiate the determinant qua determinant.
Suppose
for
instance, that the Naiyayika
against
the
Mimamsist:;
"All that is
employs the syllogism

believed
or
known or
asserted
existent is necessarily possessed of the universal 'existence',

tence* is also believed or known or asserted as existent.
tence
must also be possessed of the
universal
as
'Exis.^.Exis-
'existence'."
In
this argument the ground, 'the assertion of existence', is admitted

t be true by both the proponent and the opponent, but the consequent is equally denied by them.
It thus satisfies the requirements
of the definition of the rednctio ad absmdm in full.

be a case of reductio ad absurdum,because
But it cannot
the consequent entailed
cannot be regarded as true by the Naiyayika himself, because he does

not endorse the presence of the universal in 'existence*.
It thus trans- -
pires to be a suicidal argument and a false re duetto
ad absurdum,
although the definition applies to it *The opponent may argue:
;5
"Well, the situation can be averted
by qualifying the phrase 'by means of a: determinate* by the adjective

'not endorsed as such by the arguer himself'2.
In the argument just
considered the determinate is accepted to be true by both the parties.

The definition now amended does not apply to" it arid hence there
will be no difficulty."
But the definition will be too narrow, though
it succeeds in avoiding the charge of undue* extension.

instance, the following argument:
Take for
.
"All objects of knowledge are objects of perception.

Merit and demerit are objects of knowledge.
..Merit and demerit are objects of perception."
It is a case of
valid reasoning, because the middle term AS accepred and the major
KhKhlp. 1 2 8 0 /
svaym vyapyataya'nistcne 'ty api visesanyam id cet?

Ibid.
"
22Q
INDUCTIVE REASONING
[CHAP.
term is denied by the opponent. The Naiyyika believes that even

such humanly imperceptible entities as merit and dement are objects
of intuition to God and mystics. The proposed amendment of the
determinate 'as not endorsed by the arguer himself would make the
definition inapplicable to the instance just considered, since the determinate is endorsed by the arguer also. It is certainly a case of valid
reasoning and so the failure of the definition proposed to apply to
it is the fault of the definition.
The opponent may contend that the definition should be stated as
follows: "The entailment of the determinant by means of the determinate admitted by the opponent but not endorsed by the arguer alone
is a reductio ad absurdum with reference to the opponent who does
not admit the determinant, provided the determinant is not repudiated
by the arguer . himself'* 1 . This definition will not overlap suicidal
reasoning, since the proviso is absent in the latter, in which the determinant entailed is necessarily repudiated by the arguer also. It will
not fail to apply to the reasoning entailing the perceptibility of merit
and demerit on the ground of the determinate being not endorsed by
the arguer as such by himself. The meaning of the clause 'not endorsed by the arguer himself is 'not endorsed by the arguer alone*, that
is to say, the determinate in reasoning should necessarily t e endorsed
by the opponent.
;
. But this., amendment too will not save the definition from the
charge of narrowness, since it will fail to include the following case of
valid reasoning: "Had there been a jar in existence here, it would
have been perceived". The determinant, perception of the jar, is not
endorsed and the determinate 'the existence of the jar* is endorsed by
the rguer. The absence of the proviso stated in the definition in the
case places it beyond the range of the definition. It may be suggested:
"Let the determinate be qualified as 'one not admitted by the arguer
to be actually existent*. In the argument under review the jar is not
believed to be actually existent by the arguer and so the definition
i atha svayam anistavypake svayam vypyatay'estetia yan na bhavati
tatrnabhyupagatavypakam param prati par"bhyupagatena vypyena vypakaprasajanam tarkah. .
.
KhKh, p. .1280,
XII]
would be applicable".
221
But even this amendment will not succeed in
saving it from exposure to the charge of over extension to a case of

reasoning which fails to culminate in the proof of the opposite proposition 1 .
To take a concrete example.
redactio
ad absurdum
The Naiyayika propounds this
against the Vedntists:
"If the Absolute
Brahman were to be known from the Vedic text, it would not be

self-revealing like the sacrifices known from it".
The definition as
amended would apply to it, although the Naiyayika does not endorse
the universal proposition implied, namely, that which is self-revealing
is not known from the Vedas alone.
Here the determinate * to be
known from the Vedic text alone' is not endorsed by the Naiyayika
as actually existent.
But it cannot be an instance of valid reasoning,
since it does not end in proof. of the opposite inasmuch as the

Naiyayika does not endorse the universal proposition, "What is selfrevealing cannot be known from Vedic text alone*', the concept of
self-revealation, being entirely u repugnant to the Naiyayika.
The
Naiyayika might propose to qualify the determinate by the clause

* which leads to the proof of the opposite'- in order to escape the
p r e d i c a m e n t . -*
But this too will not make the definition free from defect.
It
would fail to cover those cases of reasoning which are employed

entirely for the i purpose of the confutation1 of the opponent on the
basis of a determinate admitted by the opponent alone.
The arguer
in this case -does;-not believe in the necessary concomitance of the

determinate and the determinant concerned, and so the proof of the
opposite does not follow as a consequence.
cited.
An example may be
The Naiyayika seeks to repudiate the conception o Absolute
Brahman by the following redactio ad absurdum:
''If the Absolute
Brahman were the material cause of the universe, it would be Hable

to transformation'' 5-:> It is regarded as a. valid reductio ad absurdum
by the Naiyayika, though it cannot lead to the proof of the opposite,
since the: Naiyayika does not believe in the opposite concomitance,
1 evambhte eva viparyayparyavasyini gatatay 'tivyapakatvt.
KhKh, p. 1280.
2 viparyayaparyavasyine'ty api praksepyam ti cet?
Ibid., pp. 1280-8ig
222
INDUCTIVE REASONING
[CHAP.
namely, 'that which is not transformable is not a material cause1.

The Naiyyika believes that aka'sa is the material cause of the sound,
although it is not liable to transformation1.
It may be urged by the Naiyyika that an argument which is not
capable of culminating in the proof of the opposite as truth cannot be
an instance of genuine rednctio ad absurdum. It is at most a case of
opposition2. But this defence is a device of escapism. The essential
character of reductio ad absurdum consisting in the entailment of a
consequence unacceptable to the opponent on the basis of a determinate admitted as such by him is present intact. What is essential
to re duetto ad absurdum is the refutation of the opponent's position
and the culmination in the proof of the opposita or otherwise is a
consequence of no logical importance. The proof of the opposite
conclusion possesses logical value, because it is a case of opposition.
It is illogical to think that culmination in the opposite is an essential
condition and so the repudiation o reductio ad 'absurdum on the
ground of this failure is entirely unjustifiable.
The difference of the
types o reasoning, one leading to the proof of the opposite and the
other standing on opposition alone can at most be regarded as a
variation in the theme of opposition. The former type only shows
that the admission of a false determinate leads to the contradiction of
an established truth and the Matter shows that the admission of the
determinate makes the admission of the unacceptable determinant
inescapable. In both cases the admission of a consequence denied by
:j
the opponent is shown to be inevitable.
The denial of the status of reasoning to hypothetical assertion
(sambhvan) on the ground of its failure to lead to the opposite conclusion is equally unjustifiable. - What is essential in reasoning s that
the admission of the determinate is made the ground of the necessity
of the admission of the determinant. It is inessential to insist whether the admission of the determinate on the part of the opponent is
based on the actual presence of the determinate, or on the actual
concomitance of the determinate and the determinant, for the
1 KhKh, p . 8 i .
2 sa prasanga eva na bhavati virodhamtram tad iti cet?.
Ibid,, p 1283.
*n]
223
purpose of the reductio ad absurdum. The objective of such a

restriction is to show that the reductio ad absurdum must lead to the
establishment of the opposite conclusion. But this is logically unnecessary. The reductio ad absurdum s possible even on the basis of
the admission of the necessary concomitance by the opponent1.
The
admission may be based on objective truth or a false belief.
What
the reductio ad absurdum seeks to emphasise is that the admission of
the determinate makes the admission of the determinant logically
irresistible, in spite of the fact that the determinate is not objectively
present or the concomitance between the determinate and the determinant is a matter of subjective belief. The reductio ad absurdum
which consists in showing the necessity of entailment of the determinant inadmissible to the opponent on the ground of the opponent's
admission of the determinate may not be based on objective truth.
The objective truth of the concomitance of the: determinate and the
determinant is not essential to reductio: ad absurdum, but the belief
of the, opponents It is evient.frpm' the consideration that the
objectivity of the necessary concomitance unknown and unbelieved by
the opponent makes ineffectual the employment of reasoning.
It is
necessary that the niecessity of the concomitance should be realised by.
the opponent, otherwise it will fall as a damp squib upon him2.
It shows slipshod thinking when the logician insists upon botb
the objective truth of the necessary concomitance and the admission
of the same byi the opponent asithe essential condition of the reductio
ad absurdum. Wha we have shown is that what is necessary is the
admission alone^ and the objective truth is irrelevant, because the
objective truth* so Jong as it remains unrealized gives.no advantage
either-iway^, v The-culmination /in the opposite; conclusion is only
symptomatic of the absurdity of the opponent's position, which has
already; been-driven home by the entailment of the determinant. Now
the sceptics i such as ? the Madhyamikas have relied on reductio ad
1 paramrthato vyaptyabhve 'pi parbhyupagamam dya prasagapravrtter upapatteh.
.
KhKh, p, 1283,
2 Ibid, pp. 1283-84.
24
INDUCTIVE REASONING
[CHAP.
absurdum as the weapon of confuting the positions of other philosophers.
The Mdhyamika
philosopher does not believe in the
reality of anything, objective and subjective both.
He has adopted
a peculiar position in logic, namely, to accept the arguments
of the
opponents at their face value, and to apply to them the logical tests
which are accepted by the latter.
He shows that all the doctrines of
philosophers are found on examination to be vitiated by self-contradiction by means of the logical apparatus adopted by them.
But the
refutation of a particular position does not commit him to the admission of the opposite alternative, because in pursuance of his negativistic
philosophy he does not believe in the truth of
any position.
And
certainly logically speaking there is no reason to take exception to the

position of Ngrjuna.
He adopts reductio ad absurdum
instrument in his criticism.
as the only
But he does not think that the denial
of one position will lead to the affirmation of the opposite, if it is

found to be riddled with self-contradiction like the others It must be
admitted that the reductio ad absurdum so far as it is employed for
the refutation of the opponent's position does not involve commitment
to the opposite of what is refuted;
. But reductio ad absurdum when employed for the establishment
of a particular position necessarily leads to the acceptance of the opposite truth.
In such a case the failure to culminate in the opposite
conclusion is a" real drawback1.
The
Buddhist fluxist seeks to establish the position that all existents are
momentary, because existence means causal efficiency.
efficiency is exercisable in
And causal
succession or non-succession which are not
possible of a non-momentary reaL . He certainly believes in the objective concomitance of existence with momentariness.
And he can
establish this concomitance by showing the impossibility of the concomitance of causal efficiency with the . non- momentary.
And this,
can be effected if the Buddhist can show that: the non-momentary is

non-existent, because it cannot be possessed of causal efficiency.
Thus
the Buddhist fluxist has to prove the concomitance of existence with
t tasmd yah prasagah svapaksasiddhyaagam tasya viparyayparyavasyita dosyai'va syt.

KhKh, p. 1285,
XII]
225
momentariness on the ground of the negative concomitance of norimomentariness with non-existence. Here the belief in the objective
concomitance of two positives necessarily involves the belief in the
concomitance of the corresponding negatives. The failure to lead to
the proof of the negative concomitance is tantamount to the failure f
the proof of the original affirmation. But with regard to those critics
who are interested only in the destructive criticism of the opponent's
position, and who do not propose to establish any conclusion of their
own, positive or negative, the only thing that matters is the provisional
acceptance of the opponent's assertion for criticism. He shows that
the assertion of the opponent is contradicted by the canons of logic
accepted by the opponent himself. It does not lie in the mouth of
the opponent to bring the charge against him that he does not believe
in the objective truth of the necessary concomitance or its oppositeV So
far-astihe critic is concerned, his belief or unbelief in the necessary
concomitance of any two terms and of their opposites, is matter
entirely personal, which has no bearing upon the success of a redact to
ad absttrdum1.
The Naiyyika may contend that the charge of incomprehensivenes$ is based upon a wrong: conception of the nature of i reasoning. - A
true reasoning is /one in which a consequence unacceptable to the
opponent: is driven home as the determinant on the basis o a: ^determinate endorsed by the proponent as well as the opponent'21;
Thus
there is no? possibility of. a purely . destructive reduetto ad absurdum
being regarded as a case of reasoning and hence the inapplication of the
definition to it does not give rise to the fault of incomprehensiveness.
But tips shifting of the ground does not improve upon the situation,
since.in-the aptempq to avoid one defect, it becomes exposed; to ranother. ^ The definition becomes too narrow, because it fails to embrace
the species of reasoning which consists in the logical entailment of
desirable consequences Thus the assertion, "If I were to have a glass
of water, I could allay my thirst" is a case of valid reasoning, because
:
1 y as tu; prasagah parapaksabdhanjigam tatra parbhyupagamamtram

prayojakatn....
KhKh, p. 1286.
a atha vypybhyupagamen'nistasya vyapakasya pratitis tarka itt cet?
Ibid., p. 1287,,
226
INDUCTIVE REASONING
[CHAP.
of the logical relation between the two as determinate and determinant1,
. . . . . .
The Naiyyika may propose to delete the word * unacceptable'

from the definition.
The definition so amended would include the
case just mentioned.
But it becomes too wide inasmuch as it overlaps
the fallacious reasoning called 'the entailment of a consequence endorsed by the opponent' 2 .
opponent*
But let the adjective 'unacceptable to the
in the first definition be replaced by the expression *un-
cognised by a previous cognition*.
The definition will thus stand as
follows: " A reasoning is one in which a consequence uncognised by a

previous cognition is driven home as the determinant on the basis of
the determinate endorsed by the proponent as well as the opponent*'.
This will not apply to the fallacious reasoning in which the consequence is endorsed by the opponent, because the consequence is not
previously uncpgnised.-
But even this amended definition will be too
wide, since it will overlap an act of primal inference in which the

determinant was previously uncognised3.
The Naiyyika may contend that such inference is not a case of
reasoning,, because the determinant is known by means of a determinate which is truly xognised and not merely endorsed.
defence cannot stand scrutiny.
But the
The- difference of true cognition from
endorsement of :nthe determinate is only apparent in the first stage;

though there:is a veridical cognition of the determinate, yet the
determinant cannot be made acceptable to the opponent, unless he is
made to endorse the determinate.
So even in inference employed for
the conviction of the opponent the endorsement of the determinate by

the opponent is . essential and thus the difference based on true cognition and endorsement is bound to disappear in the ultimate resort. It
is a truism of logic that the determinant cannot be proved to the
satisfaction o the opponent by means of the determinate, though
cognised and hence accepted by the proponent,
unless the opponent
i na, istrthasambhvanym avypteh.

KhKh, p. 12872
;.-.-nhi crl*.vy.Spakasya pratitih sa iti cet? na. istpdane'pi gatatvt.
Ibid.
3 apramitasya tathe'ti cec? na. prathamnumane'pi gatatvt,
"" 7 ^ - W : , , . V.
'
Ibid.
:
>-
xii]
is persuaded to endorse it.
227
Thus the definition extends to primal
inference.
The Naiyyika may contend: "Well,.the difficulty can be avoided
if the definition be amended as follows.
'Reasoning is one which
leads not to the veridical cognition (aprama), but to admission of the

determinant on the basis of the admission of the determinate1.
Admis-
sion of the determinant means not veridical cognition but the enforced
admissibility of it".
The substitution of admission for
veridical
cognition in the definition of reasoning prevents it from overlapping

inference no doubt, but it makes it exposed to another serious defect.
The amended definition will overlap the fallacious reasoning called
mutual contradiction (mtthovirodba)1.
In. the erroneous perception of the shell as silver, the ontological status
of silver is a matter of dispute.
The realists have found a crux in
it and the idealists have not refrained from exploiting the situation
in the interest of their philosophical position- The realists contend
that silver is real on the ground that an unreal cannot appear to a
subject.
The Naiyyika would employ this reasoning in his support:
"Were silver unreal, it would not appear.

of fact.
.*. It must be real".
ing by the following:

contradicted.
"If the silver were genuine, it would not be
But it is contradicted as a matter of fact by the subse-
quent perception of shell as shell.

Now,
But it appears as a matter
The idealist would contest this reason-
in both these instances
/ . T h e silver cannot be genuine'*.

there is enforced
admissibility of
the determinant by means of the determinate endorsed-by both the

proponent and the opponent.
And so in terms of the definition they
should pass muster as genuine cases of reasoning. But these reasonings

are fallacious owing to mutual contradiction and so -the application of
the definition proves its untenability.
The definition would also
apply to the fallacious reasoning called the unreality of the subject

(asrayasiddh),namely,
a variety of lotus".
"The sky-lotus must be fragrant, because it is

Here also the admissibility of the determinant
on the admission of the determinate is indisputable and so it should

pass for a legitimate case of reasoning, which it cannot be on account
of the unreality of the subject.
1
...mithoviruddhdau tarkbhse'pi gatatvt.
KhKh, pp. 1287-88.
228
INDUCTIVE REASONING
[CHAP.
!'' -But the Naiyayika may contend that he can easily avoid this
charge by adding a proviso, provided that it is not vitiated by any
fallacy, such as mutual contradiction or the unreality of the subject/ 1 .
But it shows that the Naiyayika has got a very weak case, since he has
to employ a series of defences, and each defence by a fresh amendment
makes it liable to a fresh onslaught. This amendment too is in the
same position. The Naiyayika lays stress upon 'the enforced admission of the determinant as opposed to the veridical cognition of it* as
an essential element in reasoning. The definition yet will fail to
include, a genuine instance of reasoning in the following situation.
Suppose that a person is in doubt whether smoke is actually present
in a situation, though as a matter of fact the smoke is veridical. But
it is legitimate on his part to argue: "If there were smoke* there must
be fire". Here the tentative assertion of smoke and fire transpires to
be an assertion of a veridical situation. So there is veridical cognition
of the determinant and thus the reasoning becomes true, The definition amended would not however be applicable to it, because the
Naiyayika seeks to exclude veridical cognition of the determinant in
order to prevent its extension.to primal inference. But the Naiyayika
might seek to avoid this unpleasent situation by putting, a new
interpretation on the clause / t h e , admissibility of the determinant as
opposed to the veridical cognition Q i t \ The Naiyayika may. contend
that the meaning of the clause is not that . there may not be an actual
veridical cognition, but that the veridical cognition though present
as a matter of fact, is not realised as such at that tim. The definition
thus interpreted will not apply to.the doubtful situation cited above,
because the cognition of fire though it happens to be veridical,:;is not
realised as such at that time, but as a tentative possibility2*
. ,^^
Even
this
amendment
cannot
be
successful.
The
first
objection
t
to it is that it employs the pronoun /that' (at that time). Now these
pronouns always stand for specific cases and so cannot convey; a generic
meaning which is applicable to; several situations. The logician only
seeks; to; create an; iilusion; of /uniformity by means of the' uniform
x * sraysiddhydivyatireke s'ti cet ? *
KhKh,p. 1288. :
2 tatklam pramtven'pramiyamn ity api*ti cet ?
Ibid.
,/
- --' ii;
' :
x]
229
pronoun, but as their meaning is bound to vary in each case, the

uniformity can at most be formal. Apart from this formal difficulty
the definition will extend to the fallacious instance called suicidal
reasoning, which as has been shown before occurs when the determinate
is accepted by both the parties and the determinant is unacceptable
to the opponent as well as to the proponent himself.
In this case
there is forced admissibility as opposed to the veridical cognition of

the determinant.
It has been proposed to avoid this contingency by
the proviso that the word 'determinate* in the clause 'on the basis of
the admission of the determinate' should be read as falsely assumed
determinate'1.
This amendment precludes its extension to suicidal
reasoning, since the determinate is accepted as true by both the parties.

But this amended definition will fail to include a case of reasoning
based upon the determinate accepted as veridical by both the parties.
I n a dispute between the Mlmanisist
and the Naiyayika regarding the cause of an effect, such as a 'sprout;

the Mlmamsist would explain ? the production of the sprout at*
determinate time and place by appeal to unseen destiny as its operative
cause.
The Naiyayika on the other hand would explain it by appeal
t an agent (God).
The sprout must be produced by a Conscious
agent as it is a product like a jar;
This is a case of legitimate reason-
ing according to the Naiyayika, since the determinant is entailed fey

a necessary concomitance which the Mimamsist cannot afford to deny;
But the definition fails to apply to i t / because of the insistence on the
determinate being falsely assumed.
The determinate in this cas,
/the sprot being a product,' is really present and recognised as*such

by both'.the parties.
We have examined all the possible definitions of reasoning arid
found them to be vitiated by one defect or another. This shows 'dit
reasoning cannot be ah acceptable logical category.' The result
achieved,by the "examination of the general definition will be confirmed
by its failure "to* apply to the specific instances which we are going to
discuss:
'
' " --^---.<-'- --,.- ,.
:
tadvyvcchedrthatn ropitasya
siJdhena vypytena prasangasy'vypant.
vyapyasy-abhyupgamene'tkarane
:
ca
; KhKh; 1288; \\-
CHAPTER XIII
SECTION I
We have already dealt with the various divisions of reasoning in

chapter ix and shown in chapter x how these species of reasoning
turn out to be fallacious instances on account of the lack of one or
other essential condition.
Sdharsa contends that the definition of
reasoning finally propounded as 'the entailment of the admissibility of

the determinant on the basis of a falsely assumed determinate', is
susceptible of application to the false simulations of self-dependence,
mutual dependence, vicious circle, the vicious infinite, contradiction,
and tu qptoque, which are the recognised six species of authentic
reasoning.
If the definition of authentic reasoning can be shown to
extend to fallacious simulations, the definition must be condemned on

the charge of undue extension.
And so long as a logically sound
definition cannot be formulated,

that
we must accept the conclusion
reasoning is not capable of being
defined
and hence not
capable of being conceived as a distinct logical category and as

such must be'dismissed as an uncritical andan irrational figment of
the imagination.
-, We have already shown that knowabiltty is a universal subsistent
and being again knowable in its turn, the attribute
becomes an instance of self-dependence.
knowability
But the difference of the
terms of reference and of relations is supposed to make the charge of

self-dependence in this
case a
enduring for a considerable
false
period
simulation.
of
time
Again, a jar
must change every
moment of its existence and to account for this transition of states in

terms of causality, the previous jar must be recognised as the causal
condition of the subsequent one.
Here the self-identical jar being
the cause and effect of itself involves the charge of self-dependence

which is however a false appearance.
due to the difference of time-relations.
The unreality of the fallacy is

We have shown that mutual
dependence and the vicious circle turn out to be innocuous if there is
xiii]
difference of individuals.
sive and regressive,
As regards the vicious infinite, both progres-
we have
independence of the
231
shown
how
in
terms in respect of genesis,
the case of the

cognition, and
subsistence the infinite series becomes legitimate and valid.
Regarding
tu quoque it has been shown how it ceases to be vicious if there is a

special consideration in support of one alternative.
Now all these
cases of reasoning fail to act as invalidating weapons in spite of their

seeming analogy with their compeers which are really vicious.
We have made this digression in order to drive home the fact that
the definition of reasoning as 'the entailment of the admissibility of
the determinant on the basis, of a falsely assumed determinate' applies
to these cases, though they are not legitimate cases of reasoning as an
instrument of refutation.
It is quite arguable that what is real cannot
be its own locus and content, that is to say, cannot depend upon
itself for its existence and things which are real cannot be dependent
upon' one another in respect of genesis, cognition and subsistence.
This is the reason why self-dependence etc. are regarded as logical
fallacies.
But in spite of appearances we have shown how exceptions
can occur.
The definition of reasoning under consideration applies
to these exceptions and so becomes exposed to the charge of undue

extension.
-/A-
The logician may contend in defence that the definition can be

made perfect if a proviso is added :v< 'subject to the consideration that
the validating conditions are absent'1. We have found that the validating-conditions ar the difference o the terms of reference, relations^
numerical difference, and so on.
When these conditions are absent
reasoning becomes a legitimate weapon of refutation
Sriharsa in
answer to this defence of the logician asserts that it is a stratagem of

escapism.
Now it is not possible that all the validating
conditions
should be absent in each case, and so it must be conceded that theabsence of the conditions one by one makes these species of rasoning;
invalidating' weapons.
But the difference of terms of reference arid
consequential relation which is supposed to be the validating condition

1 prasangajsthane tavatm visesnm abhven'pi lalcsanm visesnlyam
i:
itt cet ?
; n
KhKh, p. 1297* ^
*3*-
INDUCTIVE REASONING
[CHAP.
of self-dependence is found to be present in mutual dependence and,

vicious circle, because though they terminate in self-dependence,
the
difference of the intermediate links in the chain has every right to be

regarded as a. difference in the terms of reference.
But this does riot
vest mutual dependence and vicious circle with validity.
As regards
numerical difference which is supposed to legitimatize mutual dependence, it is.found to be present in the vicious infinite also.
does not confer validity upon the vicious infinite.
But that
It cannot be denied
that in the infinite series of terms there is numerical difference of the

cause and effect and so the charge, of self-dependence cannot be
brought home.
But in spite of this validating condition it cannot be
regarded as a valid case,
because the charge of self-dependence is
sought to be resolved by the indefinite extension of the chain through

the addition of successive links which cannot be successful, because the
chain cannot be completed.
Here there is mutual dependence coupled
with numerical .difference of the terms.
But the numerical difference
does not legitimatize the mutual dependence involved in it.
So these
validating conditions resorted to by the logicians to account for the

difference of legitimate self-dependence from illegitimate self-dependence and so, on transpire to be so many dodges calculated to throw
dust in the eyes of the superficial observer.
It cannot be denied that
the so-called fallacies of self-dependence, ^mutual dependence, vicious

circle and, the vicious infinite are ultimately resolvable inpo the identical fallacy of self-dependence, and the extension o the chain is only
a device;p> fly in the face of it by a show of numerical difference.
Sriharsa is too astute a thinker to be deceived by these tactics and
accordingly; lays his finger on the weak spot.
He asserts that the
difference of .reference is found in mutual dependence and vicious circle^

bu* the. fallacy of self-dependence which is involved in them is not
d thereby.
And numerical difference which is supposed t be
of mutual dependence, and vicious circle is present in the ,

infinite series, but it fails to overcome the fallacy of mutual dependence
which, if undoubtedly involved in it. So the incorporation pf the absence
of the validating conditions in the definition of reasoning to prevent its
extension to fallacious semblances of self-dependence etc. only results, in
making it top narrow and incomplete, since it will not apply to cases
an]
SPECIES OF REASONING CRITICISE
of self-dependence etc. involved in mutual dependence and the vicious

infinite, because the validating conditions are not absent in the latter1.
Furthermore, the definition of reasoning must overlap such cases
of censure as the admission of a proposition contrary to the established
conclusion (apasiddhnta ) and the contradiction of the thesis
(pratijnvirodha)2. Thus, for instance, it will be possible to argue:
"If the Buddhists admit pre-disposition to be permanent, it will be a
case of admission of a proposition contrary to their established conclusion of universal flux". Here 'the admission of permanence' is the
determinate which is shown to entail the admission of a contrary
position as the determinant and so it is a case of reasoning. Similarly
it may be argued : "If you assert a proposition to be false which you
have previously admitted to be true, you will be guilty of contradicting
an established truth and your previous admission of it". Here the
assertion of falsity is the determinate and contradiction is the determinant. The definition of reasoning applies to it. But the logician cannoc
accept these cases as instances of reasoning, since he has recognised
them to be cases of censure. He cannot plead that there is no difference
between reasoning and these censures, as in that case these censures
should be subsumed under a head in which reasoning will be included.
The different enumeration of them in the list of censures would become
absolutely indefensible if these censures were also cases of reasoning.
;. Up till now the general nature of reasoning in terms of the
definitions propounded by the logicians has been subjected to scrutiny
and the result has been disappointing in that no definite conception of
reasoning as a logical category either in the way of refutation or
confirmation could be spotted out. Sriharsa now proposes to examine
the different species of reasoning, namely, self-dependence and others
with a view to the determination of their logical validity. The result
will be found to be equally; disastrous.
\ : ::J:w A-'ip r.rn >;> : () Self-dependence
Let us take up the case of self-dependence. It is a species of
reasoning and as such must consist of a logical determinate and determ:
i KhKh, ppi 1297-98.
.x, -"* -d#*-.-.
2 api c'pasiddhncavirodhdisv api tarkalabananl gacchat kathakram
vraniyatn?.
Ibid., p. 1298.
30
2J4
INDCTVE REASONING
[CHAIS.
mant governed by the relation of necessary concomitance.

question pertinently arises:
Now the
"Is self-dependence sanctioned by neces-
sary concomitance or not?" If it is endorsed by the valid cognition of

necessary concomitance, self-dependence must be regarded as legitimate
and valid, far less a logical offence.
If on the contrary, there be no
necessary concomitance between the determinate and the determinant,

it will be a false simulation1.
But the logician may contend that this
sweeping condemnation on abstract consideration is inspired by a

carping spirit and not by a desire to understand the real issue.
What
is meant by the charge of self-dependence is 'that which is the locus

of anything cannot be the content of it1.
A thing cannot be its own
locus and content, its cause and effect, and its cognisant and cognised.
It ^preposterous that the same thing should be the cause and effect,
locus and content, subject and object. Sriharsa in reply asserts that such
formulation of reasoning is wide of the mark.
that a thing cannot be a content and a locus.
One cannot contend

It is seen that the jar
is the content of its parts, but is the locus of colour.
So the opposi-
tion of locus and content is neither generic nor absolute, but specific
and relative.
One can at best contend that A cannot be the locus
or content of A.
But it is a matter of specific observation and has no
jurisdiction beyond the specific data.
The relation holds between two
individual facts and so the question of universal and necessary concomitance is: ultra vires.
It will not be legitimate to deduce from a
single observation a general rule.
Each case will have to be attested
by individuar observation and so the relation can at most be empirical

in character and thus incapable of laying claim to apodeictic certainty2.
" Even granting the validity of the proposition 'A cannot be the
locus and content of A', one cannot jump to the conclusion that B,
or C9 or D cannot be related as locus and content of themselves.
The
matter is entirely empirical and because A is not B and B is not A one

cannot legitimately infer that what is observed to be true of A will be
true of B. In fact,-all the propositions which allege the charge of selfl atmasraydes ca mlavyaptau pramjiopa gamas cec tarhi prmnikatvan
na dosatvam, na cen mlasathilyam ity ubhayacah-psabandhah katham
mocanyah?,
KhKh, p. 1299.
3 Ibid,
xin]
235
dependence and the like are singular propositions, and in spite of the
formal character of universality asserted by the logicians their validity
is entirely confined to the observed cases, and does not extend to other
facts.
It is refreshing to note that Kant has placed singular proposi-
tions in a different category from the general propositions, having a

real universal reference.
In Formal Logic, the singular proposition is
regarded as an A proposition together with general propositions.

this is purely a formal affair and has no material validity.
But
The logician has contended that he does not bank upon the
universal reference of the proposition concerned.
Even admitting that
it is a singular proposition, the logician will show that self-dependence

becomes an effective instrument of refutation.
When it is asserted 'A
cannot be the locus or content of A', it is implied that the identity of

A would be imperilled, if the relation of content and locus were to be
asserted1.
So the charge of the lack of extra-individual reference on
which Srlharsa bases his contention' only illustrates the sceptic's

eagerness to pick a hole in the opponent's position withour caringa to
understand its meaning.
But Sriharsa is not satisfied by this defence.
He says that the very assertion of the reductio ad absurdam is vitiated

by self-contradiction.
The logician would have to assert ."A cannot
be A, if A is to be the locus and content of A". > The proposition 'A

cannot be A' " is self contradictory and as such cannot be asserted.*: So
the reductio ad absurd am becomes itself absurd.
It may be contend-
ed in defence by the logician that the assertion of the absurdity affects

only the opponent.
In fact, the opponent can be proved to be fn the
wrong only if it is shown that his position involves absurdity. Sriharsa

criticises this defence asacase of sophism.
It is asserted by the
logician himself that the validity of the reductto ad^bsurdum

ly consists in the proof of the opposite proposition.
ultimate-
Now the reductio
ad absrdptm "A cannot be A, if ft is to be its own locus* and content"

cannot lead to the proof of the opposite proposition.'^ The opposite of
the proposition' c A cannot be A* is A must be A \
is not a proposition.at all,
But the latter
A logical. proposition is one- in which the
1 etadasrayatvad etadsritatvd v nai'katvam syad iri vacanabhangy

'padyam iti cet ?
KhKh, p. 1300.
236
INDUCTIVE REASONING
subject and the predicate have different connotation.
[CHAP.
But in CA is A',
the same thing is the subject and the predicate, and as such it cannot
be a proposition for the proof of which che employment of t-redactio
ad absurdum can have justification1
The logician may contend that self-dependence and the like are
only the instances of the type of reasoning called * purely destructive'.
In destructive reductio ad absardum
the culmination in the proof of
the opposite proposition is entirely out of the question and so the

criticism of the sceptic does not affect the validity of self-dependence
etc. as types of reasoning2.
a device of escapism.
Sriharsa in reply observes that this is only
It is not a fact that self-dependence etc. are
used only for destruction of the opposite thesis and not for confirmation of one's position3.
x KhKh, pp. 1300-01.
2 na ca prasangamatram etadbdhyai'va astu, krtam asya viparyayaparyayasnene'ti.yuktam. .
.
. .
Ibid., p. 130!,.,...,
3 Sriharsa has contended that self-dependence and the like are not necessarily used for purely. destructive purpose and are therefore capable of being
used as confirmatory considerations. He has not however cited a concrete
illustration in support of this assertion. It can be shown that Sriharsa's argument is supported by facts. The definition of cause as the invariable antecedent,
provided it is not otherwise realised (anyathsiddha) will on analysis show that
self-dependence is used as an argument in. support of the logician's position.
Xhe word anyathsiddha literally means what is otherwise realised. The cause is
defined to be one which is not otherwise realised. Now the adverb 'otherwise'
is a relative term and cannot be understood without reference to a particular
datum. It can only, mean therefore that a cause is one which is not realised as
other than a cause. In other words, a fact which can be realised as other than
a cause is not a cause. : But if this be the meaning of the term, the definition
becomes, a case of self ^dependence. A cause is defined to be one which is not
realised as other than a cause. 'Not other than a cause' means the cause itself.
Jhe Naiyyika realised this drawback and took care to assert that the expression
'otherwise realised* (anyathsiddha) is not to be taken in its grammatical sense,
but'rather as technical expression which means three or five kinds of facts,
namel v, the invariable characteristic of a cause or its attributes or . the cause of
a cause and so on. Thus though the jar is the cause of drawing water, jarhood
which is an essential accompanying attribute of it, without which the jar cannot
be known, is not its cause. So also the colour of the jar is not the cause. SQ
xni]
37
Moreover the logician does not admit the validity of an argument

in which the necessary concomitance is not endorsed by both the
parties. A reductio ad absurdum
derives its cogency ultimately from
the positive concomitance of the determinate and the determinant and

if it happens to be the case that the logician does not believe either
in the positive or in the negative concomitance, he is precluded by the
logical code accepted by himself to propound a proposition, which
presupposes such concomitance.
The logician may rejoin:
"You are entirely mistaken.
I do not
assert the reductio ad absurdum in the form 'A.will not be A' # I mean
to say that 'A will be other than A, if A is to be its locus and
content'.
Sriharsa in reply asserts that the difference between the
present and the previous reductio ad absurdum is merely verbal.

;
content of the latter is the same with that of the former.
The
To say that
'A will be other than A' is the same thing as to say 'A will not be A \
Thus the charge of self-contradiction, remains as it is 1 .
The charge of
failure to lead to the opposite is also unescapable even in the amended

form. The opposite of the reductio ad absurdum A will be other than
A' is * A will be not other than A \ which can be further expressed
as 'A will be other than other than A*.
The propositions A is hot
other than A* or 'A is other than other than A' mean that *A is
A\2
It has been observed more than once that a tautologous proposition is

no proposition at all, because the predicate and the subject are one;
But the Naiyyika may observe that in the reductio ad absurdum *A
will be other than A \
v'i;
~.!v";?f
'.:-:fL-"<-
. v'v.,;-;.,;.
the predicate is simply Otherness* and A is

;.
iLi4i.-.
- ^
j w - -
.'::.:..%':
'
\ ; r . : r /.
:.*::;}..':[-.
again the potter who has produced the jar is not the cause of drawing water.
Each of these cases is regarded as anyathsiddha. Now the reason for departure from the etymological meaning of the term is nothing but the desire to
avoid self-dependence. Thus the necessity of the avoidance of self-dependence
is the reason for the avoidance of the etymological meaning of the expression
anyathsiddha and the interpretation of it as one or the other variety! stated
above.-*This shows that self-dependence is not necessarily a purely destructive
consideration, but is used as an argument in support of one's position also.
i
]
>l
ata evaitadanyat syld ity api na sakyaprasajanam...vyghtd cva.
' ^ ^ b : ; ^ : , - - - . v
-:-',-
>,:.
-.
>KhKh, p. IJ03.
j
- ? ^...etadananyatvasy' aitadanynyatvasy" aitatvd eva^
*
.
;
Ibid* , .
stated to specify the 'otherness* and should not be construed adjectivally as an integral part of the predicate.
A here serves as a pointer and
index (upalaksana) standing aloof and so the proposition does not

become tautologous, because A does not enter as an element into the
predicate1.
The ingenuity of the logician is admirable no doubt.
But
i Indian logicians have recognised two different types of adjectives,

namely, internal (to the substantive) and external. In the proposition 'the
blue lotus is waving in the air', blue is an adjective internal to the lotus and
so is associated with the action predicated of it. The external adjective is of
two types, namely, ( i ) one in actual association with the substantive and (2) the
other not in such association. The first is called updbi and the second is called
upalaksana. The internal adjective is called vi'sesana. The external adjective
stands aloof and apart from the substantive and is not integral to it. So it
cannot come into relation with the predicated attribute. The upad hi, though
external, s connected with the substantive by an external relation. It may
be called an associate adjunct. The Vaisesika maintains that auditive organ
is nothing but ether (kasa) enclosed in the ear-drum. The ear-drum is the
associate adjunct of ether. The difference between ..an updhi (associate adjunct)
ana upalaksana (a pointer or index) is that the former can serve to distinguish
the substantive so long as it is actually present, whereas the latter need not
be present in order to distinguish the substantive. It distinguishes the substantive by virtue of its association with it at some past time. Thus the deceased
author of a book serves to distinguish the book, though he is not present at
the time of distinction. The notice-board of a shop may be cited as another
instance. Even when the notice-board-is removed, we can tell the shop, because of the past association of the notice-board with it. The classical example
of upalaksana is -'the crow on the roof of DevadattaV house*. The crow
helped us to identify Devadatta's house when it actually perched upon it.
But the crow may have flown away and yet we do not find any difficulty in
distinguishing Devadatta's house by. reference to the crow. The crow, now
absent, serves to distinguish it.
To sum up. .Though all adjectives are distinguishers, thc-Vtsesana being
integral to substantive is adjective par excellence. Regarding the external
adjectives, the upadbi, the associate adjunct as we have called it, must needs
be actually present in order to distinguish the substantive. Thus there will
be no organ of hearing if the ear-drum is destroyed in spite of the fact that
kasa is ubiquitous and eternally present. As regards upalaksana which we
have named. pointer or index, its association with the substantive is not coeval
with it and as such, can serve to distinguish it in absence. - We may also call
it once-associate adjunct, but dissociated from it during its subsequent career.
it seeks to escape the issue.

is otherness'.
In self-dependence the consequent alleged
This 'otherness' without reference to A will be a
pointless objection.
The absurdity involved in making the same thing
the locus and content of itself is shown by the consequent 'A will be
other than A*.
If mere 'otherness' without reference to A is to be
predicated, it will not involve any absurdity, since A's otherness from
B, or C is an obvious fact.
The re duetto ad absurdum can be effective
only if it is shown that A will be other than A,

locus and content.
if A is to be its own
So A cannot be detached from 'otherness' and be
regarded as a mere pointer, if the absurdity is to be driven home 1 .

But the logician contends that the numerical difference of things
is undeniable.
In fact, the identity of one thing is not the identity
of another thing, and this difference of identity is integral to the

individuality of things fer
se.
So A's otherness is a self-subsistent
fact, and the characterisation of it by A as A's otherness, is only a

methodological device which enables us to prove its otherness to a
disputant.
Such being the case A's otherness stands intact even if
we do not -characterise it by reference to A . And it is this otherness

which is the characteristic of things per se and? is different in each case
that is alleged to be the consequence of self-dependence, and as A
does not enter as a qualifying element into otherness, the charge of
proof of the obvious cannot be brought home.
The alleged 'otherness*
is not understood as otherness than B's or G's, but as A's otherness

which is the concomitant of A's individuality.
And because A stands
apart and aloof as a pointer to this otherness, the charge of selfcntradictibn arid the
consequential
charge of tautology
acceptance'of the opponent's position do not arise*.
in the
>
Srlhars observes that the defence shows the desperate straits to

which the logician is reduced.
The logician posits that the phenom-
In the present context the logician seeks to wriggle out of the difficulty
by regarding A, in the predicate 'other than A', as an index dissociated from
the subject at the time of the affirmation of the predicate.
i anyatvvadher tmana upalaksanatve...paryavasyet.
' ' '" '
KhKh, p. 13033 *
2 svarupata eva vilaksartam anyatvavisesam avadhir tmo' palaksayati...
Ibid.
M&
DUCTIVE REASONING
[CHAP.
enal objects are each possessed of numerical difference per se. What is
the ground of this postulation? It is nothing but the cognition o their
difference.
But the cognition of difference does not necessarily
presuppose factual difference of identity of different things. It can be
accounted for by the difference of associate adjuncts, though the
'otherness* which is the object of differentiation be an undifferentiated
unity. Even the logician admits that space is one, but we differentiate
the space of one room from that of another and also from the spaces
enclosed within a jar or a ship. The differentiation of space is due to
the difference of associate -adjuncts and not to objective difference in
space. The numerical differences of the phenomenal objects can
likewise be explained by reference to the difference of the associate
adjuncts, though the things may not actually differ from one another.
T h e defence based on objective numerical difference of things thus
topples down 1 . .
r ; Even if the numerical difference of things be admitted as factual
and objective this will not give any advantage to the logician in the
present context. The otherness of A may be different from that of
B and also be self-contained. But when the reductio ad ab surd urn in
self-dependence* is to be demonstrated it can be done only with reference to the objects concerned. Let us examine the example * A will
be different from A , if A is to be its own locus and content'* N o w
if difference alleged* s the consequent be not specified asÂ's
difference, but the factual difference as pertains to it as an objective
characteristic without reference to A , it will prove at most that A has
a different identity of its own. But nobody is interested in disputing
this obvious fact. That A has an identity different from that of B or
C is not the object of dispute and the logician will not gain, anything
f he thinks that it is this difference that is entailed, in the reductio
ad absurdum.
Thus interpreted the proposition will become absolutely innocuous. The absurdity car* be driven home when it is shown
that the admission o A being the content and locus of A will entail
A *s numerical difference from A . Mere difference will not prove
anything damaging and so the difference must be qualified by A a .
i KhKh, p. 1303.
a Ibidft pp. 1303-4.
xmj
241
We have stated the ground and the consequent as individual facts. '
But the necessary concomitance between them must be stated in:
general form, namely, 'if a thing is to be its own locus and content,
it must be numerically different*. But we have shown that the general
concomitance is only formal and each case of self-dependence as a
redtictio
ad absurdum
reduces itself to a singular proposition. The
reductio ad absurdum can be effective only if it is couched in the form

A will be different from A, if A is to be its own locus and content' ,
We have already shown that the absurdity can be brought home only
if the difference alleged is to be : qualified by A.
difference proves nothing.
Mere unqualified
So whether the - terms of the redudio ad
absurdum be accepted either as specific individuals or general terms,

the difference alleged must be qualified by the terms themselves.
does not make any difference whether A .is interpreted
as an individual.
It
as a class or
The logician cannot avoid the absurdity involved
in making-A different from A, because his attempt7at eliminating A

from the difference as a pointer in detachment*has-proved to beTa*
bluff1,.
'.-:...
- - V
- - ' "
' T - : -'-"
.We have considered self-dependence in respect of subsistence and

not in respect of genesis and cognition.
But the sam line ^farg-J -
ment is adoptable in these two respects also.
Thus if A is to: be
regarded as the cause or product of A, A will be split up into two

different identities.
other than A*.
In other words, * A will not be A ' or 'Arwilt be
Similarly in respect of cognition 'if A is to be the
cogniserand the object of cognition of A, A will not be A or A will

be other than A*.
These are the typical forms in 'which self-depen-*-
dence can be formulated as a reductio ad absurdum.

raised by Srlharsa that the reductio
ad absurdum
The objections
as stated involves
self-contradiction and tautology in its negative implication apply as

effectively ras it has been shown to do against it i n j respect of
subsistence^; C^.
,''/
" r:--.\.*i
(2)
::.?:;:" ; ; ' r ' : ^ ^
.-- . - - < *
^Â}.
Mutual dependence
As regards mutual dependence or logical see-saw Srlharsa holds

that it is exposed to equally fatal objections.
1 Khfih, p, 1304.
3*
The charge of mutual
*4*
INDUCTIVE REASONING
[CHAP.
dependence arises if the cognition of the difference of A from B

requires as its condition the cognition of B's difference from A.
Now
what is the possible form of its assertion ? If A is to be known by

reference to B and B is again to be known by reference to A, A will
not be known at all.
If this be the form of assertion
ic becomes
downright nonsense.
If there be no possibility of knowledge of A ,
then how can it be asserted ? If the cognition of A is conceded to be

an observed fact, the assertion of its impossibility will involve contradiction.
If there be no such cognition then the necessary concomitance
between the cognition of A's difference from B as dependent on the

cognition of B's difference from A, and the impossibility of the
cognition ok A will never be realised.
So the re duetto ad absurd urn
will be exposed to the charge of lack of logical sanction
(mlasaithilya),
since the concomitance of the ground and the consequent
remains
'unknown1,.? It has however been maintained that if the necessary

concomitance is stated in a general form as: follows, "An entity cannot
depend on another entity for its cognition, if the cognition of the
latter presupposes the cognition of the former", then the alleged
difficulties will not arise.
The reductio ad absurd urn may be stated
in the form "If A and B are real, they cannot depend on each other
$n respect of cognition"*
But rharsa does not think that this formulation is free from
difficulty, though it succeeds in avoiding the charge of the impossibility
of the knowledge of necessary concomitance and self-contradiction
The reductio ad absurdum as stated in mutual dependence presupposes
the necessary concomitance between reality and the impossibility of
mutual dependence in respect of cognition..... But. this concomitance
ean be regarded as necessarily universal if the opposite possibility is
ruled out. , There is no" argument available to prove the impossibility
of the contention, "Let things be real, though they may be found to
depend on one another in respect of cognition".
So long as this
contention is not ruled out as an absurdity-, mutual dependence cannot

be adduced as an invalidating argument, because of the lack of
f na tVad yady etad . etadbodhdhinabodham syt tad na buddbyete'ti,
tath sati vyptyasiddheh...
KbKh, p. 1306.
XIII]
243
necessary concomitance which is the foundation of all cases of redactio

ad absurdum1.
(3)
Vicious circle
As regards the vicious circle it is nothing but a case of self-dependence or mutual dependence only protracted by the interposition of
additional links.
We have already noticed how vicious circle and
mutual dependence can ultimately be reduced: to self-dependence.

The consideration of an accredited instance of circular argument which
has received the imprimatur of approval will place the matter in a
clear perspective,
"Salvation is achieved by the elimination of each
immediately preceding link induced by the elimination of each succeeding link in the chain
of suffering, rebirth,
voluntary - activity,
defects of attachment, aversion, and wrong cognition 2 ".

of this aphorism is this.
The meaning
True knowledge of the nature of the self
and not-self is the cause of salvation.
But it does not immediately
lead to this final consummation,-but only in and through successive

stages.
The elimination of the immediate successor entails the! eli-
mination of the immediate predecessor in the chain.
True knowledge
eliminates wrong cognition and because .wrong cognition is- the cause
of attachment and aversion, its elimination directly entails che elimination of these defects and the elimination of the latter J entails the
elimination of voluntary activity, which is productive- of religious
merit and dement.
This again entails the elimination of rebirth which,
is nothing but the embodiment of the soul in* a distinctive psychophysical organism.
The elimination of the latter involves the elimina-
tion of suffering which is incidental to physical incarnation of this self.

Now the truth of the aphorism presupposes a circle of cause and
effect.
Wrong cognition is the cause of attachment and aversion
and the latter is the cause of merit and demerit induced by voluntary
activity.
And this in its turn results In incarnation which is again
the cause of suffering, which is nothing but the false cognition of

identity of the self with the body.
.
This false cognition again pro-
1 KhKh.p. 1306.
,
i :
...
. .'.,;., ,.:..,-.<. 2 duhkhajanmapravrctidosamichyjannm uttaroetarpye tadanantarapyd apavargah.
NS, I. i. 2. See also KhKh, p. 1313.
244
INDUCTIVE REASONING
[CHAR
duces- aversion and.attachment which again produce various activities,

which again produce a fresh birth and this again produces a fresh
series of suffering.
So the circle of cause and effect goes on uninterrup-
tedly until the saving knowledge takes place.
This is a case of argu-
ment in circle involving reciprocal causality.

But though a circle, the Naiyyika does not regard it as vicious.
It is true that wrong cognition presupposes wrong cognition as its
condition
through the interposition of certain intermediary links.
But the two wrong cognitions are not numerically identical.
The
wrong cognition which has led to the emergence of the present birth
. and the wrong cognition which is again produced by the latter are
numerically different individuals.
The difference of the individuals
- as cause and effect shows that the circle is not vicious.
The circle
becomes vicious only when the initial link and the final link in the
chain are one and the same.
.^ This defence : is ^ apparently satisfactory. - But a difficulty is confronted when we try to understand the problem a bit- more closely.
It is held that suffering
is the effect of birth; birth is the effect of
activity; activity is the effect of defect; and defect is the effect of

.wrong cognition.
It is because each preceding link is the effect
of each- succeeding link that the elimination of the succeeding,link

:> entails the elimination of the preceding one r because the disappearance of the cause induces the non-emergence of the effect. Now this
-. relation of causality is ascertained by the application of., the. Joint
Method.
It is plain that the causal relation as understood between
two facts has necessary reference to two classes of facts and not individuals. . If the causal relation were apprehended between the ob ser v, ed facts per nobody could understand: another individual oL, the
- same class as the cause.
Causal, relation must_ be admitted to be
, possessed of a general reference.
A man observes, that fire produces
^smoke. -The observation is strictly speaking of two individual entities,

.fire and smoke. But the man,i who apprehends fire as the* cause and
smoke as the effect by observing the presence of fire attended with the
presence of smoke and the absence of fire attended with the absence
of smoke, invariably produces fire in order to produce smoke.
But
the latter fire and smoke are different numerically from the previously
MIX]
245
observed fire and smoke. So the causal relation is always understood

not between two individual phenomena qua individuals, but between
two facts, so far as they typify and embody the class-character which
belongs to all members of the class. The causal relation that obtains
between suffering and birth and so on as stated n the aphorism should
in the light of these findings be understood to subsist between two
classes and not individuals. Individually suffering and birth are
different in each case. But as the causal relation does not obtain between them as individuals, but so far as they represent the classes,
the difference of individuals has, no bearing upon it. The numerical
difference of individuals which is entirely irrelevant to the causal
relation cannot therefore be regarded as a solvent of the mutual dependence involved in it. The defence by the Naiyyika of the validity of the circular argument by reference to the numerical difference of individuals is thus found to bc~a dodge, which may succeed
with, unreflecting persons1. The failure of numerical difference to
serve* as : a; validating condition:-* does away with the defence of the
logician of circular argument as valid in one case and invalid in others.
If circular argument be valid at all, it must be valid in all cases, and
if invalid it must be invalid without exception.
,-..- It may however be argued by the logician that the charge of selfcontradiction: and tautology alleged by Sriham to arise necessarily in
self-dependence :etc. is rather formal. It is asserted by the logician
no doubt that, the admission of the same thing being the. cogniser and
cognised, locus and content,! cause and effect; will involve the admission of the absurdity *A is not A', while in reality 'A is always A \
To assert 'A is A 1 is : to assert a truism. But the assertion of this
truism is necessary to bring home the absurdity involved in the denial
of identity which is the implication of self-dependence in respect of
genesis,/ subsistence, and cognition. True,; that 'A is A* is not a
proposition andso also *A is not A \ The logician does not categori-cally^ assert these - propositions, but is impelled by circumstances to
-allege them as consequences. It may be asked how can the logician
r s;. j'Mfcryakaranabhvasya'tajj tiyatay niyatattvena vyaktibheda^y cakra
knantarbhutatvt; "
- KhKh, p. 1313.
246
INDUCTIVE REASONING
[CHAP.
entertain - these propositions ? even as hypothetical possibilities, regard

being had to the position o radical empiricism which s advocated by
him?.
The logician of the Nyya school cannot assert that in the
proposition-'A is-not A' the predicate 'not A' i$ an ideal construct, as

the Buddhist would have it.
The logician thinks that all our know-
ledge and assertion are derived element by element from veridical

experience.
Such being the case the logician cannot seek to escape
the contradiction involved by regarding it as ideal as .opposed to actual.

The Nyya logician particularly in so far as his latest interpretation of
perceptual error goes cannot save his realism by making the relation
an unreal fiction, which is the position of Vcaspati Misra.
Even in
error the logician is not prepared to admit any subjective element*

In the erroneous perceptual judgment '.this is silver' the subject and
the predicate are both objective facts.
Though Vacaspati Misra
admits that the relation between the subject and the predicate is
subjective and hence unreal in the context, Gagesa differs from him
and asserts that the relation is also objective, which is as* much veridical, as the predicate. .. Such being the case the logician o the Nyya
school cannot escape the chirge of contradiction and tautology asserted
by Sriharsa.
i The raison d'etre of self-dependence, being a logical absurdity
lies in the ^recognition of a- necessary universal concomitance viz. a
self-identical
entity
cannot? be. its own locus and content.
The
relation of locus and content, as also the relation of cause and effect,
and of subject and object, is a necessary concomitant of numerical
differences But,whatever the frequency with which it may be observed,
concomitance < of two iacts
must .be regarded as contingent, and
precarious, if even a single exception can be found.
The source of
these inductions is ultimately experience and if a contrary experience

is found to record an opposite finding, the .necessity and the universality
of the concomitance .cannot bel pressed with confidence.
Let;. us ? now consider whether the concomitance on which selfdependence is J. based is necessary, and
universal.
We find that
the concomitance of the relation of locus and content with numerical

difference breaks down in the case of knowability, speakability, and
the like.
Knowability is a universal
subsistcnt and as such is a
MU]
SPECES OF REASONING CRITICISED
147
universal predicable. Knowability is also an object of knowledge and

knowability is nothing but the quality of being the object of knowledge, which as much belongs to itself as it does to other objects.
Thus knowability is subsistent in its own-self also. In other .words,
it becomes its own locus and content, though it is numerically
identical. Likewise speakability being itself expressed by a speech is
found to be speakable. The attribute of speakability therefore subsists
in itself in spite of the lack of numerical difference between locus and
content. So also the concept of usability, being itself usable is found
to behave as its own locus and content. This is also found'to be the
case with inherence which is admitted to serve as a relation and the
term of a relation. A quality is related toa substance by way of
inherence. But this inherence being different from the terms, quality
and substance both, ought to require another relation to belong to the
terms. In other words, a relation being a separate < entity is also a
term in addition to the two terms <: which are brought into relation by
it. But as the admission of another relation makes a regresses -ad
infinitum inevitable it is regarded as self-related. To be explicit.
Relation as a term is supposed to act as its own relation. ~It is thus an
object of relation as well as an instrument of it. Relatedness is~ a
quality of the terms induced by a relation and when a relation is
found to make itself related by itself it becomes- the -effect nd^ cause
of itself. To be related means to stand in a relation. To be-precise,
it means to be possessed of a relation.- If relation be supposed, to be
self-related it would mean that a relation is possessed of a relation and
as the second relation is the same as the first, the relation becomes its
possessor and possessed, locus and content. In this instance the relation of locus and content is found to centre round self-identical
entity and * so the alleged concomitance of numerical difference with
the relation of locus and content falls down1. " -- '
- Similarly with regard to the counter-term! of negation. I the
proposition: "The jar is not here", the jar is the counter-term and as
such it has the attribute of being the counter-term of negation. Now
i kin ca prameyatvabhidheyatvasryavah*ryatva,.,katham tmairayatkhandik vylptth savyabhicar* na syc?
KhKh, p. 131a
248
INDUCTIVE REASONING
this attribute belonging as it does to jar cannot belong to the negation?

of jar.
So the attribute is absent in the negation of jar and this
attribute then becomes the counter-term of the second negation and

must possess the attribute of being the counter-term.
Thus the
attribute of being the counter-term must be admitted to be possessed

again of the attribute of being the counter-term.
This means that a
self-same attribute is the possessor and possessed of itself, in other

words, as its own locus and content1.
The logician may contend that this charge of self-dependence
cannot occur, because there is either numerical difference of the terms
or difference of the terms of reference, or there is difference of relation.
In dealing with the charge of self-dependence regarding knowability,'.
we have shown that the difference of terms of reference serves to make
it legitimate and innocuous.
But Sriharsa observes that this difference :
of relation or, terms cannot be regarded as the solvent of contradiction

involved in self-dependence, because there are cases of self-dependence
etc which are accepted by the logician as legitimate, .although no
such difference can be trotted out as validating condition.
Thus
Prabhkara holds that a cognition illumines the subject, the object, and
its own-self at once.
Here the cognition cognises itself, though it is
numerically one and. the same.

numerical difference
Prabhkara cannot . lay hold of
as. the < solvent of the incompatibility which
becomes inevitable if the object of cognition and the act of cognition

are held to be possible <$nly in numerical difference.
numerically
Likewise A,, is *
different from B. The proposition means that A and B
stand in the relation of difference.
This difference being a third term
is again different from each of the terms.
And for this a second
difference qua relation is necessary which can make the first difference
a third term different from the original terms.
But the second relation
being in the same position with the first being a relation and a term
alternately,will require another difference to make it different from the
terms in question. But since difference qua relation must needs assume
the position of a term, a fourth difference will be necessary and the
result will be an inevitable regressus ad infinit urn.
diCh, p.-1310.
In order to avoid
xni]
SPECIES PF REASONING CRITICISED
149
this situation the logician holds that the relation of difference makes
not only the terms different from each other, but also its own-self
different from them. Here the same act of difference becomes the
agent and object of differentiation.
In plain words it is the
differentiator of and differentiated by itself. Obviously it. is a case
of self-dependence. But che logician has to pass it as legitimate.
Take again the case of self-cognition.
The Naiyyika holds that
the self is the cogniser of everything. When pressed with the question
"Is the self cognised and if so by whom?' 1 , he answers that the: self
is cognised by the self, because there is no other cogniser than the
self.
In making this admission the Naiyyika admits that the self is
the subject and object of its own cognition.
It is a clear case of self-
dependence, but has got to be approved by him.

proposition: "A word denotes a meaning"..
i;
Take again the
Now the subject word*
stands for the whole class of words and as such stands for its own-self
also.
A word is thus denpter f itself and so is the denoter and
denoted both.
Here I neither the subject, word, nor the act, denotar
tion, is different.
But the self-dependence
with the logician 1 .
involved passes muster

\
Again, the Buddhist fluxist holds that things perish automatically

without requiring a destroyer. *.A thing is thus destroyer of itself "'and
in admitting the same thing to be the subject and object of destruction, the Buddhist ought to bz guilty of committing the fallacy of selfcdependence.
But the Buddhist does not regard this self-dependence
as illegitimate.
The logician of the Nyya-Vaisesika school admits
that a relation ultimately relates itself to the terms without another

relation, : just in the same way as it relates the terms.
In admitting
self-relation of relation the logician* makes it the subject and object

of the.same act and yet does not feel an offence in it.
As regards the
laterjjlogician,: who:jholds that ultimately /it is the terms themselves

which, behave as; a relation, he makes the terms the relater of them1 dvravyaktibhedasy'pi vyabhicrivyarirekatvt. svapraksavdin svayam eva svajaaatvasya...itydi bahulam upagamd tmirayatadbhsavivekya
v
kim niymakam upeyam.
KhKh, p. 1310.
3*
250
INDUCTIVE REASONING
[CHAP.
selves. The terms are thus the agent and object of the act of relation.
In spite of self-dependence and in spite of the fact that there is no
numerical difference in the act or in the terms themselves, the selfdependence involved in the situation has met with the seal oE
approval1.
We- need not multiply instances. We have shown that though
there are cases of self-dependence in respect of subsistence, or of cognition, or of genesis, the philosophers of different schools have to pass
them off as entirely valid and legitimate. The logician has not been
able to furnish a criterion of legitimacy in one case and illegitimacy
in the other. The plea of difference of terms of reference, or of
relation cannot'be substantiated in the cases cited above.
'
As regards mutual dependence also the philosophers are found to
contradict themselves by admitting certain cases of mutual dependence
as valid, although they cannot find any special validating condition^
The Vaisesika admits that sound produces another sound and the
succeeding sound destroys the preceding one. The last sound which
is perceived in ' the ear-drum does not produce another sound, yet it is
destroyed. But what is the agent of destruction? The Vaisesika holds
that it is the penultimate sound, which produces the last sound,
destroys the latter and the latter again is supposed to destroy the
former. There is a plain case of mutual dependence in making A
the destroyer of B and B the destroyer of A 2 . ' ,
'...':: .;. ',
Again in equipollence as illustrated in the proposition '*An entity
which is produced is liable to destruction and whatever is liable to
destruction is an entity produced", the first determines the second and
the second determines the first. Here is a clear case of see-saw. Bue
it passes muster. Again, several conditions are combined to produce
an effect. Here the conditions mutually assist themselves and thus
they are mutually dependent. But the realist does not notice any
incompatibility in the situation3. Again, in asserting A is smaller
We 1 rKhKh,pvi3io.
'
. :: 2 'anyonyisraye c'ntyopntyasabdayor anyonyansakatyam..,katbam na
vyptibhagah.
Ibid.
xiii]
251
than B, it is implied that B i s bigger than A . The smallness of A

is determined by the bigness o B and the bigness of B is determined
by the smallness of A . N o w it is a clear case of mutual dependence,
which however does not cause worry to the philosopher. What makes
these cases valid ? The philosophers concerned maintain reticence on
the question, because no criterion can be found. The logician may
plead that there is no contradiction here and so these cases are passed
off as valid. But this is not an honest attitude. W h y should there
be contradiction n other cases and the absence of contradiction in the
cases under review? The basis of non-contradiction is propounded to
consist in the absence of observation of self-dependence or mutual
dependence in other cases and the same criterion should be applied
in these cases also. If self-dependence and mutual dependence be
logical fallacies in other cases> they should be regarded as logical
fallacies in the cases considered above. If you think that in the cases
under review self-dependence and mutual dependence are endorsed by
an accredited organ of knowledge and in the fallacious instances
there are no such validating conditions, it is proper to point out the
validating conditions or their absence as the criterion 1 . If these
conditions cannot be pointed out the assertions of self-dependence
and mutual dependence as fallacies in one case, and as. legitimate
instances in other cases, become absolutely confusing. ; It has
been shown that self-dependence and mutual dependence are not
fallacious universally and necessarily and the logician has failed, to
formulate a criterion by means of which one can distinguish between
a legitimate and an illegitimate instance. The conclusion seems
irresistible that self-dependence and mutual dependence cannot be
relied upon as instruments of refutation or of confirmation with
absolute certitude.
:
SECTION II
:i
'
'""
(4) Contradiction
Contradiction is regarded as a species of reasoning. It consists

in the admission of the co-existence of opposite attributes in the same
1
pramanbhva ey'opajivyo dsanam isyatm, krtam anyonysrayena vyabhicaritadosatvene *ti.
KhKh, pp. 1310-n.
252
INDUCTIVE REASONING
[CHAP.
substratum. But the question arises, "How can it be asserted and

in what form ?" As has been shown to be the case with selfdependence, contradiction also cannot be asserted in the way of a
proposition both in its positive form and in its negative implication.
It is nonsense to say that an existent does not exist. It is equally
nonsensical to assert, "Were it not existent, it would be non-existent",
or "Were it not non-existent, it would be existent". The ground and
the consequent in the propositions being identical, the propositions
are tautologous. 'Not existent* is equivalent to non-existent and not
non-existent is the same thing as existent. So how can they be
regarded as determinate and determinant, since they invariably
presuppose numerical difference as the condition precedent ?V
Besides contradiction is also based upon necessary concomitance.
Thus it is affirmed that "A cannot be both cause and not-cause of B".:
The impossibility of causality and non-causality with reference to
the sanie effect in one substratum presupposes that the two : should
occur in two different
substrata.
But
this supposedly
necessary
relation is found to be contradicted by their successive occurrence n

the same thing at different times.
Thus a seed is found tobe
productive of sprout at one time and non-productive of it at another.

The logician may reply that the necessary concomitance of incompatibility of productivity with non-productivity is valid with reference to
the same time.
It is never found that the seed, which produces
Sprout at a;time, does not produce it at that time.
So the successive
occurrence of opposite attributes in one substratum does not involve

contradiction at all.
Srlharsa observes that the difference of time is irrelevant to the

question of incompatibility. Now a things and its annihilation are
contradictorily opposed, though they occur at different times. It is
never possible that a thing should exist at the time of ânnihilation.
The .contradiction between them is therefore necessarily related to
different times. So it ,must be admitted that the co-occurrence of
^tne supposedly contradictory attributes, is not universally impossible
and the attempted solution ol the contradiction by reference to
.
i. See KhKh,p. 1308.
[nix
253
difference of time is unconvincing.

The logician has failed t)
demonstrate the basis of contradiction, which can be accepted as
universally valid.
The contradiction of supposedly incompatible
attributes in one substratum is ultimately reducible to the opposition
between existence and non-existence, being and non-being. But even
this fundamental basis turns out on examination to be illusory, as it
refuses to fit in with all accepted canons of logic.
;
Now what is the nature of opposition between being and nonbeing?
It cannot be understood as non-com presence or non-co-occurr-
ence, because the being of a pen is found at one place with its nonbeing at another place at one and the same time.
So contradiction
cannot be understood as non-compresence in one time.
Nor can it
mean non-compresence at one place, because the presence of a thing

and the absence of it can occur successively in the same place.
Nor
can it be supposed to consist in non-compresence in the same place

and the -same time, ; because : there are certain qualities, which are
admitted by the
Nyya-Vaisesika
their opposites in one substance.

disjunction,
of the soul.
sound, and
These
philosopher
to
coincide with
Such for instance are conjunction,'
cognition
can occur only
and
other
specific
qualities
in one part and therefore
coincide with their non-being in the remaining parts at one arid the
sametime.
So the compresence of the being of conjunction etc/ and
their non-being in the same substratum with reference to the same

time beiner veridical fact, the definition of contradiction ^would apply
to it and as such must be rejected as invalid.1
??
:I -
It may be argued that the objection does not affect the position of
those who do" not believe in the existence of partially extensive qualities
anc asTsucK are not bound to admit the compresence of an entity and
its negaibrori in the same substratum at the same time.
Accordingly
the definition of contradiction as the impossibility of the compresence

of being arid non-being in the same substratum at the^ ^sanie tune
stands unassailable.
-tv.
But what is the meaning of the impossibility of cc&presence of

1 bhvatadabhvayoh ko virodhah? sahnavasthnam iti cet? na......
KhKh, p. 1263,
254
INDUCTIVE REASONING
[CHAP.
being and non-being ? Does it mean the negation of compresence of

the two or the compresence of the negation of the two? 1 The first
alternative is not entertain able, because the compresence of the two
which is the counter-term of negation is not attested by an accredited
organ of knowledge. The Nyya-Vaisesika logician is emphatic in
his assertion that the negatum, that is, the counter-term of negation,
must be a veritable real and can never be an ideal fiction. Even in
the negation of such supposedly unreal fictions as a hare's horn, the
logician regards the negata as objective facts. When it is asserted
'the hare's horn does not exist', the assertor does not commit himself
to the admission of an unreal fiction. The meaning of the negative
proposition under consideration's that a horn does not exist on the
hare's head. The horn is an objective real being veridically perceived;
in a cow, and the hare and its head are also objective facts. It means
therefore nothing more than this; the horn is not perceived on the
hare's head.. So negation is on the same footing with affirmation, so
far as its reference to the objective datum is concerned. The counterterm of negation being universally and necessarily an objective fact,
contradiction cannot, be interpreted as the negation of compresence of
being and non-being, because such compresence is never an object of
cognition2.
; The secon alternative is also found on examination to be equally
afesurd. The compresence of the negation of being and non-being
means. the presence of the negation of being together with the
presence of the negation of non-being. But the negatioio being
means the affirmation of non-being and the negation of non-being
means the affirmation of being. The second alternative thus analysed
comes to mean the compresence of being and non-being wrrich is
absurd on the face of it. So the definition of contradiction as the
non-com presence of being and non-being in the same time and place
is found to be absurd whatever interpretation may be possibly put
upon it*.
1 tad - dhi tadubhayvasthnasahityanisedho
nisedhashityam v syt?
2 Ibid. ;
,
y
Ibid., p. 1265.
tadubhayvasthnaKhKh, p. 12635
xiii]
255
Let us try another definition: "Contradiction consists in mutual

exclusion1." But the meaning of the word 'exclusion1 is not clear.
Now the logician divides all conceivable reality into being and nonbeing.
And if exclusion be anything different from them, it
becomes entirely unintelligible. If it is held that being excludes
non-being and non-being excludes being, the meaning of exclusion
common to both is not apprehensible. On the other hand, the
meaning of exclusion is only found to shift from being to non-being,
and from non-being to being. Exclusion of being is not the same
as the exclusion of non-being. So the opposition constituted by
the exclusion of being is not the opposition constituted by the
exclusion of non-beings
So opposition common to both is, not
understandable.
(Contradiction has been alternatively defined to
consist in the noo-occurrence of one in the substratum of the other
and vicet versa** But then there is no contradiction, since being
and non-being occur i n different loci*, Even -if it ê interpreted^ that
occurrence means its relation and non-occurrence means absence of
relation, and , relation is different from the relata* i t des: not result
in an improvement. When ; you /assert that the occurrence of being
is invariably concomitant with the non-occurrence of non-being and
this is contradiction, you seek to keep pff the.opponent only by a
barrage of words, which do not convey an intelligible meaning. 1 You
all along i tacidy assume that the existence of being is the same
thing as the non<xistence of non-being. ^Such being the case,; to
assert that non-being doesnpt occurin a locus in which being occurs
is tantamount to the assertion / that being occurs in which being
occurs. But this is tautology plain and simple3. This will be made
dear by a concrete example. The jar does not exist in a place
where non-being of the jar occurs'\ Or, '-'The non-being of the
jar does not occur in the place, where the jac exists" These are
die forms of-assertion of .contradiction. ' What is the meaning of
1 parasparapratiksepakattvam virodha iti cet?.
KhKh, p. 265
2 yatr'aikasya sattvam tatr 'parasy 'sattvam niyamena yat sa virodhas
tayor iti cet?.
. Ibid*
3 .^...jambandhasya tadvyaririktattvc ekasattvasy* aiva......
paunaruktydy-ptt.
lbtd*9 p. 12655
*5$
INDUCTIVE REASONING
[CHA*.
the clause where non-being of the jar occurs?' Occurrence has

been interpreted as a relation. In the light of this interpretation
the clause would mean 'to which non-being is related as its
content*. The meaning of the predicative clause 'the jar does not
exist in a place* means to which the non-being of the jar is related
as its content*.
So the assertion of contradiction reduces itself
on analysis to a case of hopeless tautology l .
Let us examine the second proposition. The meaning of the

clause * where the jar exists* is to which the jar is related as a
content. And the meaning of the clause the non-being of the
jar does not occur in the place* is to which the negation of the
relation of non-being is related as the content, if occurrence is taken
tornean relation, distinct from the relata. But this interposition
of relation does not make the least improvement upon the situation,
because the relation of negation is not anything different from
negation, such as conjunction or inherence is. So the relation of
negation is negation itself. The negation of non-being of the jar
means the 'jar' rand so the second proposition only asserts that the
jar exists in the place in which it exists. : The double affirmation of
existence does not serve any purpose save and except that of deception
of those who are guilible 2 ; ;
M , : *'; ^ : 7v ;? ^ i <;;?
To take a retrospect, the logician has sought to define contradiction as the compresence of being and non-being. J But -being
and non-being by their very nature can occur only : in different
substrata and in this there is no contradiction, and the attempt at
putting them together has grieviously failed.
The logician has made another attempt to resolve the deadlock
necessitated by the assertion, "Between being and non-being, t otic
occurs in the place where the other - does not occur". > It is held
that"?the meaning of the clause 'where one occurs'means -where
one occurs alone and exclusively' and thus there is no tautology3.
pp. 1265-66.
2
lbid.Jf.1266.
3 y atr'aikasy' vasthnam tatr* aikasy' aive* ti niyambhipryena na patinar*
uktyadir iti cet?,
KhKh, p. 1268.
xiii]
25.7
But the meaning of exclusion implied by the adverbial phrase 'alone

and exclusively' is not intelligible. If it is interpreted to mean
'that nothing else exists in the place', it will constitute a flagrant
contradiction of fact. When it is asserted "The non-being of the jar
does not exist, where the jar exists", the meaning of the clause
'the jar exists' cannot be the non-existence of anything else than
the jar, because as a matter of fact the colour *of the jar exists as
an additional fact.
So the meaning of the implied exclusion
is intelligible only with reference to the . opposite of it. But the
opposite cannot be understood without knowing the meaning of
opposition1.
-;.
It has however been asserted that the position.is quite simple and
the difficulty is only the outcome of cussedhess. The meaning of the
proposition under consideration is that inthe alternative occurrence of
non-being, it excludes being and reversely being excludes non-being.
So when it is; affirmed that either being or non-being is alone possible
in locus, it means the exclusion of the other : alternative3* But this
attempt at oversimplification by the logician only indicates haste and
impatience. What is wanted t o b e understood is the meaning of
opposition and for this a:, definition of it has been asked for. In the
latest attempt, opposition is sought to be explained as exclusion of
being in one case and exclusion of non-being in the other, and there
is nothing common between the two and so opposition is not
understandable. Moreover the . exclusion of being means the assertion
of non-being and the exclusion of non-being means the assertion of
being. So it boils down to the assertion, that being occurs where
being occurs, and non-being occurs where non-being occurs. This is
an outraging ricase of tautology?. 3^ ?
c>. . ;; ...V; : :; :;i
:/ It may now;be/contended that the difficulty, is due to the consideration that opposition is a relation external to the terms being and
non-being;^But it is not anything different from the terms. Being
1 virodhivyavacchedakatvasya ca virodhnirvacane' nirvacant.
/
.\\\.\-%
KhKh, p. 1268.
j s a abhvapakse bhvavyavacchedo bhyapakse c'bh"vavyavacchedo niy amrtha iti cet?,
Ibid.
3 Ibid,
33
258
INDUCTIVE REASONING
[CHAP.
and non-being are opposed per se> and so opposition is nothing but
their intrinsic nature1.
It is exactly on a par with existence*.
'Existence' is the nature of an entity and its incidence in a thing
makes it felt as existent. Opposition likewise makes an entity felt as
opposite, in which it occurs. Being and non-being are thus felt as
opposite to each other, because opposition is the very nature of them.
One may ask, "If opposition be the very nature of the terms, being
and non-being, and thus be* identical with them, then what will
behave as opposite? The answer is, the opposite is that which has
opposition as an intrinsic element of its nature. Thus an entity,
which possesses existence and opposition both as elements in its
nature, behaves as an opposite existent. As regards a non-existent it
has non-being and opposition as its character. And so it is felt as a
non-existent, opposite of an existent. Opposition being the very
nature of being and also of non-being, being and non-being are felt as
opposites. Another important consequence of opposition is that
things which are possessed of existence and non-existence as part
of their nature behave as numerically different. Thus numerical
difference and opposition go together2.
This ingenious defence by the logician of opposition as the inalienable character of being, makes being and opposition identical.
Similarly non-being and opposition are also to be regarded as identical. Let it be so. But we may ask whether opposition is the
character of being and non-being severally and distributively, or of
both of them collectively taken together If opposition be the
character of being or non-being distributively, then each of them will
be split up into two opposites. Opposition cannot exist in one and
the same thing and so if opposition be the same thing as being,
being is bound to be split up into two opposites3. It has however
1 bhvbhvayoh svarpam eva virodhah...^
KhKh, p.* 1269.
kasy' aitau virodha id c'nuyoge svasrayasy* ety uttaram, kim tatra
hl
i i prasne bhedavyavasthanam ity abhidheyam.
virodhaphalam
iti
3 etayor virodhattvarn, pratyekam, v syan militayor va?. nadyah pratyekam
eva" srayaikatvabhangaprasanga't.
; .
lbid,t p. .1270.
XIII]
259
been contended that the consequence of the incidence of opposition

is not the bifurcation of a unit in which it occurs, but differentiation
or distinction.
Opposition makes an existent different from a non-
existent and difference is always understood to be the characteristic

of being and non-being in relation to one another.
A n existent is
that which is possessed of being and a non-existent is that which is

possessed of non-being and opposition being identical with being and
non-being, the loci of being and non-being are cognised as numerically different, since numerical difference is the consequence of opposition.
Opposition does not involve split of identity.
being opposite not to itself, but to non-being.
It only makes
So also non-being is
made opposite to being, because of its identity with opposition1.

Let us take it for granted that opposition is the very nature of
being and non-being, and also the contention that being and non-being
and their loci are numerically different by virtue of this inherent
opposition.
There is no possibility therefore for being or non-being
to be split up, because the opposition is always understood between

being and non-being and their loci in relation to one another.
It is
asserted that the locus of being will be opposite to and numerically

different from the locus of non-being.
But a thing which is possessed
of being and non-being successively maintains its numerical identity.

Thus there may be a pen on the table at one time and on the removal
of the pen, negation of the pen will be there.
Thus the table is
found to be possessed of being and non-being of the pen in succession,

although there is no opposition or numerical difference which is alleged
to be its consequence in the table per se.
It may be contended that
the opposition and numerical difference occur only at one time and not
at Jdifferent times,: But time being admitted to be a single real,
there can be no numerical difference of it taken by itself.
the: qualifying rphrase
*at one time'
So
is a pointless re-iteration2,.
If however time be regarded to vary with the difference of conditions

associated with it, then of course the qualification becomes intelligible.
;
2* klabhedeh* aikasya bhvbhvsrayatvbhyupagamt tadabhedati:cet?,

na. tadabhedasya svbhvikasya vivaksitatve visesanavaiyarthyt.
-^ Ibid., p. 127 ia
26o
INDUCTIVE REASONING
[CHAP.
But i a span o time qualified by one identical condition be regarded

as a temporal unit, then the same difficulty will recur.
For instance,
a day may be regarded as one temporal unit being qualified by the
association o the sun.
In that case the table may be possessed of the
pen and o its negation in the course of the day and the absence of
numerical difference of the table in spite of its being the locus of the
being and non-being of the pen at one time shows that there is no
opposition between being and non-being. But it may be argued that
though the day is one unit, it admits of variation, being associated
with different conditions occurring within it.
Thus the division of
the day into hours, minutes and seconds is possible. So the phrase
'at the same.time*, stands for the unit of time which is not associated
with diverse conditions.
The period or the time called the day is
however associated with different conditions, such as the movements
of the hands or die watch.
But this makes the determination of a
unit of timeJmpossible,.because however small the unit may be taken,

its simultaneous association with the movements of the different stars,
the moon, and the sun cannot be eliminated.
And if the association
of different conditions be responsible for the difference of time-unit,

then even a small moment will have to be regarded as diversified into
various units, as its association with diverse conditions mentioned
above cannot, be debarred.
It has however been contended that this
contingency cannot. arise, because the unit of time is determined by

conditions which do not coincide.
But coincidence is not intelligible
without reference to a unit of time*
And the unit of time again
exposes itself to the question of natural and conditional difference.;

W e have found that time ferse
is one and the difference of time is
only determined by the difference of associated conditions.
But as the,
conditions can overlap, the temporal determinations cannot be made,

exact and reliable distinctions.
So the determination of opposition.
and numerical difference of being and non-being with reference to.

time is doomed to futility 1 .
W e have seen that the logician has not succeeded in explaining
the nature of opposition by making it a characteristic of being and
i
KhKh, p. 1271.
XIII]
261
non-being severally.
Let us try the alternative which makes opposition the characteristic of both being and non-being collectively. But
what can be the meaning of collectivity? We may understand by it
the com presence of being and non-being at one place or at one time,
or their occurrence in the same place and in the same manner of
relation, or in some other way different from the mode of relation1.
The first alternative is unentertainable, because being and non-being
can never subsist in one place. The second too is equally absurd,
because being cannot coincide with non-being in the form of prior
and posterior negation. Thus, for example, the being of the jar did
not coincide with its non-being previous to its production or will not
coincide with its non-being entailed by the cessation of the jar. The
third alternative is found to be equally futile. The coincidence of
conjunction of a jar in a locus along with its absence may be cited as
an instance of coincidence of being and non-being. But the logician
does not regard this coincidence as a case of opposition and i{ we look
closely into the situation, we shall have to admit that the coincidence
of being and non-being is only apparent. Conjunction occurs in a
part of the space different from the part in which its non-being
occurs. So the identity of locus is found to be impossible. Andas
regards the manner of relation, being and non-being are found to
radically differ
Being is related by inherence and non-being by the
relation of substantive-adjectival type. So the compresence of being
and non-being is not intelligible with reference to the identity of time,
of place, or of co-relation, or of relation to the substratum.
As regards those who do not believe in the partial incidence of conjunctionthe question of the compresence of being and non-being
cannot; arise with them. It is held by them that conjunction is
possible only with space. But space is a unitary whole without parts.'
And so if anything, say, a jar is conjoined with it, it cannot be regarded as out of conjunction with space in respect of; another part,
because space has no parts. The coincidence of being and non-being
therefore in the same substratum is absolutely out of the question 2 .
1 Mbaiiitattvm ' ca'nayor ekadelatvam vl ^bhunatam, ekaklatvam: va,
ekaprakrena vrttic v, vrttiprakrnyaikopdhyavacchedo v? KhKh, p. 1272.
2 Ibid., pp. 1272-73.
262.
INDUCTIVE REASONING
[CHAP.
Even if we imagine that being and non-being may coincide in

some inscrutable way different from the modes considered above, yet
the impossibility of the coincidence of the being of a jar with its
non-being entailed by non-production or cessation, cannot make them
the substrata of opposition as they are never found to subsist together.
If being were to be supposed to coincide with non-being entailed by
non-production or cessation, the result would be that a thing would
not fail to come into being or cease to be and in that case a jar would
exist even before its production and cessation.
However one may try,
one cannot bring being and non-being, either generically or specifically,

together which may serve as the basis of opposition.
The truth of
the matter is that opposition is never an empirical fact. It cannot be

perceived to belong to an objective situation, because being and nonbeing can never coincide in any sense.
The occurrence of being and
non-being in separate locus is a fact no doubt, but that does not involve
opposition. And compresence of them is never found.
What Sriharsa
seeks to drive home is that the empiricist cannot give a concrete

instance of opposition.
It only makes itself felt when we ideally try
to bring being and non-being together. Opposition is thus a priori and

ideal and never actual.1
The realist who is also an empiricist is there-
fore bound to contradict himself whenever he would allege contradiction in a supposition, because supposition according to him must
correspond with an objective situation.
Thus contradiction cannot be
regarded as a species o re duetto ad absurdum just like self-dependence

etc. by the logician who pins his faith on realism*
(5)
The vicious infinite
Existence is an ultimate category.
Other things are made existent
by virtue of their participation in it.
But existence per se is existent
in its own right.; The postulation of another existence in existence

will give rise to an infinite series of existences.
and no rest at any stage.
There will be no end
But this logical entailment of 'lack of end
or rest' by the postulation of 'another existence* is possible if there

be a necessary
and universal concomitance between them.
1 KhKh, pp. 1273.74.
But there
XIII]
263
is no instance in which such concomitance can be cognised.
If it
were cognised, the fact of the veridical coincidence of the two would
make it veridical and so the question of entailment of a false issue
which is the very soul of reductio ad absurd urn would be absurd.
If
the concomitance were never cognised the entailment could not be

pressed as necessary and true1.
If it be argued that
there is no proof
of the coincidence, the safe procedure will be to show that we cannot

accept it as true in the absence of proof.
But the attempt to show
hypothetical entailment will in that case be bereft of all justification.

(6) Tu quoque or recrimination
We have seen that tit quoque consists in the counter-allegation
of the sam or similar defect against the opponent by the proponent.
This is regarded as a kind of reasoning and is employed to refute
the opponent when other forms of refutation do not occur to the
party
Sriharsi has elaborately refuted the concept of opposition and
we have just given an expositiqn of it.

pose a question to him:
"Well, you
The logician may however
have shown by your refutation
that the concept of opposition is not defensible.
But do you think
that such expressions as 'opposition', or 'contradiction* etc. have got

a significant meaning or jiot? If they are. significant, you cannot
deny that they stand for a conceivable fact, which your refutation
seeks to disprove.;,. If they are nonsensical expressions, then why have
you taken pains^to refute them?" T o put it the other way about.
Is the concept;of opposition refuted by you understood by j o u or not?
If not understood, how can you condemn it by showing defects in
the concept? And if you refute it after understanding its nature, we
would: ask a you whether your understanding is right or wrong. J
your understanding of the concept of opposition be correct, you' cannot refute it without self-contradiction.
If hqwever it be contended
that the understanding is erroneous, then also you cannot escape the
charge of self-contradiction.
Erroneous
cognition is possible, if there
is a veridical cognition behind its back.
The shell is erroneously
":;^ : ^ : i:f%^|l^5-::o-^ ; .
~: -. " -;* '
";-
r""""'
; ;'.y;; " /
anavasth tu yath sattym api sattntaram ity anavadhau sattpravhe
isyamne^dosagrastatvt.
KhKh, p. .1308.
264
INDUCTIVE REASONING
[CHAP.
perceived as silver only because silver was correctly perceived in the

past. The sceptic cannot show a single instance in which the datum
of error was not the datum of a veridical cognition in the past. So
the sceptic's refutation of opposition as illogical and unreal is bound
to presuppose the reality of it, whether he accepts opposition as based
upon the data of right or wrong cognition.
The logician's refutation of the sceptic takes the form of a tu
quoque argument, because he thereby only confronts the sceptic with
the same charge of self-contradiction, which was alleged by the latter
against him. But rharsa asks, "What is the purpose of the employment o tu quoqueV Does the logician mean to say that the difficulty alleged : by the opponent is equally present in the latter's
position? Or, secondly, does the assertor of the tu quoque mean that
the solution of the difficulty by the opponent will be the solution of
the 'alleged difficulty in the proponentls position1 ? The first alternative is not approved by the canons of legitimate debate. There are
three types of debate sanctioned in logic. (1) Legitimate discourse
aiming at the determination of truth; This type is confined to the
discourse between a teacher and pupil or between friends. (2)
Disputation, the objective of which is the defeat of the opponent by
refutation of his position and the establishment of one's own position
(3) Wrangling, the objective of which is the opponent's defeat alone
without regard to the proof of one's position. Now in the second
and the third type of argument- the refutation of the allegation of the
opponent is the expected course of action on the part of the proponent; The counter-allegation of the same or similar difficulty < in
the Opponent's ^position is riot quite relevant. The failure, to refute
the alleged defect is tantamount to the admission of its justice and
legitimacy" and: that will a t ' once secure his defeat and so the debate
cannot proceed further3; ''* ~ ,i v ; ; ; . . - : .
:
::
. 1 pratibandim grhnatah kim parapakse dospdanamtram vvaksitam,
uta y as tvaya tatra samdhir abhidheyah sa Giay Jpi 'ty abhprayt svapakse
samdhih?,
KhKh,p. 983.
2 na tvad dyaK aprastutatvat^w.^
lbit
.:;...
:.;:./.
[xtii

The second alternative is not also legitimate.
265
The logician
intends to wait for the refutation of the counter-allegation of the same

difficulty
from the opponent's side.
But what is expected is that
the logician should directly give the solution of the difficulty alleged
by the opponent.
The opponent, particularly in
the third type of
debate, is not interested in the establishment of any position.
In such
a case the refutation of the opponent's position will be secured by

showing that the allegation of the defect by the opponent is false.
But the failure to adduce a defensive argument on this line proves
the untenability of his position.
The logician may however contend
that the employment of the tu
quoque argument is tantamount to
the refutation of the opponent's position.
And the onus of proof of
the legitimacy of the allegation made by the opponent lies on him and
if he succeeds in offering a solution of the counter-charge which is
necessary for the defence of his position, the same solution will be
offered by the logician himself1.
But Sriharsa asserts that this is a false expectation3.
tion of the tu quoque
futation of the
The refuta-
need not be necessarily equivalent to the re-
original allegation.
The allegation and counter-
allegation need not be analogous and so the solution of one will not
necessarily be the solution of the other.
For instance, the followers
of Prabhkara maintain that all cognitions must
be veridicaL
If
however any cognition be admitted tobe false, that is to say, to be

contrary to fact, that would open the flood-gate of scepticism*
will be no ground for believing any cognition to be true.
There
The
logician may retort that this possibility of unbelief is-equally present

in the position of the aforesaid philosophers also, because the admission
of the possibility of error is inescapable.
So here the charge of
scepticism is balanced by the counter-allegation of the same charge

and thus it is a case of tu quoque.
In this case the defect alleged viz.
the inevitability of scepticism is the same.

by the followers of Prabhkara
But the solution offered
is entirely different
from that of the
x bhavms tvad abhidhattm kas tatra samdhth tato maya 'py abhidheya
iti cet?
KhKh, p. 984.
2 maya tadabhidhnasya smpratant aprastutatvc,,....
34
266
INDUCTIVE REASONING
[CHAP
logician. Prabhakara regards all cognitions to be true and the truth

of the cognition is believed by him to be self-certified. But the
logician does not believe in the self-validation of cognition. He holds
that the truth of a cognition is determined by verification. So though
the charge and counter-charge are similar, the solution is different.
The logician may think that there must be a solution of the charge
and the solution must be the same, the charge and the counter-charge
being identical or similar.1 The justification of the tu quoqae lies in
this belief. But it has been shown that the solution of the same
difficulty is not the same as propounded by the different schools of
philosophers. The logician may urge that the difference of solution is
due to the difference in the fundamental philosophical standpoint of
the different schools. But where such difference cannot be shown the
solution of the same charge is bound to be identical. So the employment of the tu quoqite argument is justified, because it only serves to
draw the attention of the opponent to the truth of the necessity of the
solution, the realisation of which will prompt the opponent to offer a
solution and this will clinch the difficulty.3
. /
But this also is a vain expectation on the logician's part. He has
admitted that the difference of philosophical standpoint is responsible
for the different solution of the same problem. If he asks for the elucidation of the difference of philosophical position adopted by the opponent,
the latter is not bound to oblige him, when he adopts the procedure
sanctioned in the third type of debate, because the purpose of such a
debate is fulfilled by the allegation of a fallacy in the proponent's
position.3 The onus of solution therefore lies on the proponent and
not the opponent. The failure of refutation on the proponent's part
of the difficulty alleged by the opponent * should bring the debate to
'termination by the adjudication of defeat of the proponent.
Apart from the difference of philosophical position the difference
n the nature of the object of dispute and consequently of solution also
1 srayd evam iti cet?. na. vaisamyasy 'vasy-bhyupeyatvt,
*:
KhKh, p. 985.
2 Ibid.
3 ko'sau tatra visesa iti cet?, na. may smpratam tadabhidhanasy
prastutattvam ity uktatvc.
lbidH pp. 985-86.
xm]
warrants different solution of the same problem.1
267
W e here propose to
give one or two concrete instances in support of our position that the
same problem admits of different, solutions.
The logician posits the
necessity of the admission of objective existence of the objects of

experience on the ground of their being cognised as existent.
Our
cognitions cannot be unfounded and baseless, as this will mean the

obliteration of the difference between truth and falsehood.
So the
cognition of diverse things as existent presupposes that they must participate in a common characteristic.
'existence'.
And this
characteristic
is
The ground of postulating objective existence as the com-
mon character of diverse objects of experience is this synthetic reference

to them as Existent*.
It is on this ground* that substance, quality,
and action in spite of their diversity inter sezte
believed to be possess-
ed of and united by the common bond of 'existence'. Now let us take

the case of existence itself, which is supposed to account for the identity
of reference regarding the categories mentioned above.
existent or not?
Is existence
The latter alternative must be dismissed as absurd,
since the idea of existence being non-existent is unentertainable on

account of contradiction.
accepted as true.
The first alternative is therefore to be
But it is not immune from a logical difficulty.
'Existence must be existent'.

tion?
(2)
(1)
But what is the meaning of the proposi-
Is it to be interpreted on the same line as the proposition

'The pen is existent*.
as the predcate.
The latter proposition affirms'existenc'
We have reiterated the necessity of numerical differ-
ence of the subject and the predicate in a proposition.

which is unimpeachable.
It is a law
There is no difficulty regarding the second
proposition, because of the numerical difference of the subject and the

predicate.; But in the case of the first proposition, "Existence is
existent" (i.e. possessed of existence) k is open to the charge of tautology if the predicated existence be regarded as numerically identical
with the subject.
If to avoid tautology i t is asserted that 'existence'
that is predicated is numerically different from the subject 'existence',

then it will invoLve a regressus ad infinitem.
i
The second existence
dsyagatavesabhvbhvavisesitatvdin 'pi tadvaisamyasambhav^t.

KhKh.p. 987.
268
INDUCTIVE REASONING
[CHAP.
being competent to be affirmed as existent will presuppose a third

existence.
The third also being in the same position will require a
fourth existence and there will be no end of it.
The logician escapes
this dilemma by asserting existence to be an ultimate fact, which is

existent in its own right and on its own account1.
In spite of the identity of the philosophical standpoint it is
apparent from the consideration of this problem that a self-identical
problem admits of different solutions.
The problem was how to
account for the synthetic reference of diverse objects of experience as

-existent in spite of their diversity.
In the case of substance,^quality,
and action the postulation of one identical attribute informing all these
different categories has been offered as the solution.
Regarding
existence itself the postulation of another numerically different existence has been found to lead to an infinite regress and so existence has
been believed to be self-sufficient.
It is clear that there may be
different solutions of a self-same problem.

the case regarding inherence.
This is also found to be
The Vaisesika philosopher believes that
substance, quality, action, universal, particularity (visesa) and inherence

are positive categories.
And among them the first five are capable of
being related by way of inherence.
A s regards inherence, though it
is a positive category it is denied that it can be related by way of

inherence to the terms.
A quality is related to a substance by
inherence and though inherence is neither quality nor substance and as

such is a third entity, it does not stand in need of an additional
inherence to make it related to the terms.
The postulation of another
inherence will make an infinite regress inevitable, because how many

times inherence may
be postulated each inherence being numerically
different from the relata will require another inherence to make it

related to the terms.
In order to avoid this predicament the Vaisesika
asserts that inherence is self-related.

term and a relation.
In other words, it is both a
Here also the problem is the same regarding the
first five categories and the sixth one.
But the solution is different*
It must be admitted by all lovers of justice, equity and fairplay in

logical discussion that the logician's reliance on ta quoque as necessitating
A KhKh, p. 987.
XHI]
269
an identical solution is absolutely unjustified and unwarranted by

facts. We now propose to drive home the untenability of tu quoque
as a form of logical argument What does the logician intend to
prove by advancing a tu quoque against the Vedntist without caring
to refute the fallacies urged by the latter? He may intend either of
these two things. In the first place, the fallacies urged by the
Vedntist may be regarded as inoffensive on the ground of their
applicability to the position of the monist himself. In the second
place, the logician may urge that though the fallacies may be actually
present, they ought not to be urged on the ground that a defect which
is common to both the parties and whose solution is bound to be the
same should be ignored and slurred.over in pursuance of the dictum of
Kumrila "In a case which is exposed to the same defect, that is
bound to be solved equally by both the parties, it is against the canons
f logic that one of the parties should be embarrassed by the demand
of its solution1",'.j
'
The first alternative is not capable of being pressed with success.
When the Vedntist urges the fallacies of non-existent or inconclusive
reason, it is open to the logician to refute the charge by showing that
the fallacies do not actually exist.
Does the definition of the fallacies
concerned apply to the fallacies urged by the Vedntist or not? If the
definitions do not apply then the logician can dismiss it immediately
on the ground that they are not fallacies, being devoid of their
characteristics. But then the logician's resort to to quoque is absolutely
unavailing and uncalled for. If, on the other hand, the fallacies are
"possessed of the characteristic features of fallacies propounded by the
logician himself, then the original proposition of the logician will be
vitiated by the fallacies in question. The logician may however contend that the fallacies urged by the Vedntist are not fallacies proper,
because they are admittedly present in the position of the Vedntist
himself, which is believed by him to be correct. Admitting the
kica parodbhvitam asiddhydikam aparhrtya pratibandy pratyava1.
tisthamnasyâ ko'bhipryah kim y ad idam dosatay pareno 'dbhvitam tad
adsanam..J...,"yatrobhayor itydi",
KhKh, pp. 987.88.
*7*
INDUCTIVE REASONING
[CHAP.
ingenuity of .the defence the Vedntst may retort: "Let it be so.

But then'the logician's definition of the fallacy will be wrong, because
if it applies fully to a correct reason, it will be guilty of over-extension."
The purport of the logician seems to be this. A fallacy which
is the subject-matter of tu quoque should not be regarded as a fallacy,

but as a legitimate reason.
But the logician conveniently ignores that
the definition of the fallacies applies to it, and apart from the tu
quoque there is nothing to discriminate between the present instance
and the other instances of the fallacies.
a tu quoque
And if the opponent employs
against him in the form "If the fallacy urged by me be
not a fallacy proper in spite of the presence of all the factors in it,
then the other cases of fallacy also should cease to be fallacies" and
how will the logician escape this tu quoque1 ?
It may however be contended that tu quoque is a case of opposition.
One cannot deny the. plausibility of the contention of the
Vedntist that the fallacies urged by him should be regarded as

fallacies proper, when all the conditions of fallacy are present in them;
But the logician's standpoint cannot be lightly brushed aside that a
position .which is admitted by both the parties in a debate cannot be
regarded as fallacious.
The result of these two rival claims, one
pitted against the other, is a case of doubt, whether the fallacies under
dispute are cases of fallacies
proper2.
The employment of tu quoque
is then justified in creating a deadlock.
But this defence admits
that the reason employed by the proponent is incompetent to prove

the position advocated by him.
So the question is not of equal
advantage as the- opponent's contention
that the thesis of the pro-
ponent is untenable stands unrefuted.

Le us now examine the second alternative that a fallacy which
equally affects the position of the proponent and the opponent should
not be exposed in a debate, because both the parties are under duress
to offer a solution of it.
But the Vedintist is not prepared to submit
to this convention, because it is
bound to give rise to complications
- I - KhKi, p. 98a.
;
2 niyamakbhve.,.sandehah paryavaslta iti cet?,
Ibid.
XIII]
which will make all debate impossible.
271
If the tu quoque be admitted
to be a legitimate argument, no debate will be possible.
Let us take
up the argument oi the Naiyyika; "Syllables are non-eternal, because

they are audible, like sound".
N o w it is admitted by both the
Naiyyika and the Mmamsist that sound is non-eternal and is also

audible.
The Mlmmsist however thinks the argument to. be. vitiated
by fallacies.
Audibility according to him is an inconclusive reason,
because even sound-universal (dbvanitva)
is audible,
admitted to be eternal by both the parties.
though it is
Furthermore, the Naiy-
yika's theory of sound-production makes all the sounds except the

final one not capable of being perceived.
The Naiyyika believes
sound to be a quality of ether which is produced by articulation or

impact of ether with an instrument.
Now sound being a quality
cannot move, because only a substance can move.
So the initial sound'
produced by a speaker perishes in the, very place in which it occurs

and accordingly it cannot be perceived by a person standing at a
distance.
It is therefore held by the Nyya-Vaisesika philosopher
that one sound produces another sound- in the proximate point of

space, and the series is completed only when the last sound is produced
in the ether enclosed within the ear-cavity of the hearer..
sound which is perceived by the hearer.
It is this
AH other sounds beginning
with the initial and ending in the penultimate sound, are bound to
remain unheard.
These sounds therefore are not audible and so
audibility employed by the Naiyyika as the ground of inferring the

perishability of sound cannot belong to them, and as such cannot prove
that they are perishable.
Thus the reason, viz. audibility, is vitiated
by a two-fold fallacy, namely, inconclusiveness and non-existence in

the subject.
Even, if the charge of inconclusiveness be waived, the
charge of non-existence in the subject, viz. all the sounds including the
penultimate which are also believed by the Naiyyika to be perishable,
cannot be overcome1.
The Naiyyika in defence
may employ a tu quoque.
rejoin i, "Your criticism proves your carping spirit.
He may
If this line of
argument were correct and legitimate that would make all inference
1 KhiCh, p. 990. See also nandaprna's commentary on KhKh, p. 990^
272
INDUCTIVE REASONING
[CHAP.
invalid including even an inference which is admitted to be valid by

both the parties to the dispute.
The hill is possessed of fire, because
it is possessed of smoke, is an instance of inference recognised as valid

by both the Mimmsist and the Naiyyika.
But a censorious debater
may easily pick holes in it and show that the argument is vitiated by
fallacies.
For instance, it may be contended that smoke is not neces-
sarily concomitant with fire, because smoke is spread over a large area
of space which is devoid of
fire.
This shows that the association of
smoke is not necessarily a case of association with
fire.
Moreover
smoke is present in a part of the hill possessed of a carbohydrate and

is non-existent in other parts.
So the probans is vitiated by the
fallacy of partial non-existence in the subject, in
addition to the fal-
lacy of non-concomitance urged before1*

It is obvious from the above consideration that the fallacies which
equally affect the position of the proponent and the opponent should
not be urged, because that would make an inference of acknowledged
legitimacy vicious and illegitimate.
Naiyyika.
This is the contention of the
But the Mimmsist avers that the employment of tu
quoque on the basis of a legitimate argument is fraught with fatal

consequences.
If a legitimate inference is made a bugbear by exploit-
ing it as a tu quoque argument with a view to gagging the opponent

into silence, the same procedure may be followed by the opponent for
the vindication of an apparently illegitimate argument.
it may be argued by the
For instance,
Mimmsist, "Sound is eternal, because it is
visible". "Now visibility is non-existent in the subject and is also

devoid of necessary concomitance with the proband urn.
But if the
Naiyyika urges this fallacy, the Mimmsist may take shelter under
tu quoque standing on the aforementioned argument of smoke and
fire.
He may contend that these two fallacies should not be urged,
as they are also present in the smoke-fire

be legitimate on all hands.
argument acknowledged to
And the plea of common difficulty put
V i tath&hy ubhayavdisammataclsyatvam dhtnnumandikatn yadi prattbandikaroci paras tad...pratibandy bhayd..*...

KhKh, p. 990.
XIII]
a73
forward by Kumrila can be advanced by the upholder of

fallacious argument1.
t
:
.
the
The Naiyyika may complain that he has been entirely misunderstood.
He does not assert that the fallacy of non-existence or non-
concomitance should be excluded in the examination of the opponent's

argument.
But what he intends to establish is this.
Only those
fallacies which are possessed of a trait susceptible to tu quoque
should
be excluded. Now the Vedntist seeks to clarify the implication of the

logician's contention by examining the nature of the trait.
It will
boil down on examination to the assertion that fallacies which are

subject to tu quoque
character.
should be regarded as devoid of fallacious
But even if we take it for granted that the fallacy is not
a fallacy - proper, the application of the terms of the definition would

prove that the definition should be revised.
If the definition be
defended by a qualifying proviso, 'except those cases which are liable

to tu quo queand hence proved to be legitimate,' then there will be no
reason for the application of the tu quo que, because the definition of
fallacies will not apply to visibility etc.
which may pass for legitimate
reason, because of this inapplicability of this definition2: "
"
The contention of the Vedntist is. that tu quoque is not a

legitimate form of argument.
The Naiyyika has been constrained
to admit that tu quoque serves to prove that the fallacies subject to

tu quoque are not fallacies or should not be exposed.
The Vedntist
. turns the table upon him by showing that the tu quoque argument
can be exploited for the vindication of an undisputed fallacy on trie
basis of an argument of commonly acknowledged legitimacy.
r; It has been contended by some logicians that the diflSculties have
all been alleged against the tu quoque argument based upon a common
problem. It has been hotly debated whether the defect which is
i. ~.esa maya 'pi sugrahai va tarn prati pratibandl atra'pi iakyata eva
pathitum...itydi.
;
KhKh, p. 990. See also nandaprna's commentary on KhKh, p.. 990.
a atha'yam vtsesam dya pratibandi syt tanmltrasy'uudbhvanam.j;
krtam pratibandy.
KhKh, pp, 990*91. ,
35
274
INDUCTIVE REASONING
[CHAP.
equally present in the positions of the opponent and the proponent

should be urged or not.
And if it is to be urged, and to be remedied,
on whom will the onus lie? It seems to be a question of priority of

right and liability.
quoque
Both the advocate of the legitimacy of the tu
and the disputant of it have striven utmost to claim the right -
of priority and to shift the onus of justification to the other side. Even
if it be admitted that the result has been indecisive regarding bilateral
tu quoque the logician may claim that it must be effective when it is
unilateral.
When a tu quoque confronts the opponent with contradic-
tion of his own position alone and exclusively, the proponent is in a

position of advantage, because the difficulty is to be met not by him
but by the opponent alone.
In such a situation the tu quoque is
unilateral and is bound to act as an effective weapon of refutation.

,The logician seeks to justify his contention by means of a concrete
illustration.
The Yedantists deny the possibility of a true relation between
consciousness and an external object.
The logician affirms this rela-
tion to be of a subject-object type (visaya-visayibbava).
He contends
that the denial of this relation will involve the Vedantist in selfcontradiction.
The Vedntist holds:
"The objective world is
unreal, because it is the object of cognition."
Now it is admitted in
this argument that there is a relation between the cognition and its
object, and this relation is made the ground for denying the reality
o the objective world.
So the denial of the cognitive relation on the
Vedntist's part will preclude him from proving his favourite thesis,
namely, the unreality of the world.
quoque
Here the logician employs a tu
to show that the Vedntist's denial will involve him in self-
contradiction by making it impossible for him to prove the unreality of

the world. This shows that the Vedntist should not seek to repudiate
the cognitive relation asserted by the logician. The Vedntist however
contends in reply that the failure to assert the impossibility of the
xoghitive - relation will be tantamount to the proof of the logician's
fM)stion.J.>iThus the Vedntist will employ tu quoque against the
logician^ by showing that the logician's argument* seeking to prevent
the Vedntist from denial of the cognitive relation, is confronted by
the argument showing the necessity of such denial, in order to maintain
XIII]
275
his thesis, viz. the unreality o the objective world1. It is thus a case
of double noosethe logician's argument showing the impossibility of
the denial of cognitive relation by the Vedantist and the latter's
argument showing the logical necessity of the same. The Vedntist
is put in a dilemma. If he admits the reality of the cognitive relation,
he will be constrained to admit the reality of the subject and the
object. If on the other hand he denies the reality of the cognitive
relation, he cannot majee it the ground of denying the reality of the
objective world2. In either case his denial of the reality of the objective world will become unjustifiable.
The Vedntist does not admit that the dilemma affects him alone:
The Vedntist shows that the cognitive relation cannot be ultimately
valid, because it is neither one of identity nor of otherness. If the
object of cognition were different from the cognition in the same
sense as an unknown %object is, there would be no relation between the
object and the cognition. If on the other hand the cognition and the
object were identical the object would be one with the cognition and
so the question of relation would not arise in the absence of the
cognitum. But the felt relation between a cognition and its object is
not capable of being repudiated as a psychological fact. What is
denied is the ultimate logical validity of such a relation. And when
the Vedintist makes this empirical relation the ground of the unreality
of the objective fact, he is not committed to the admission of its
; ultimate validity. The logician therefore cannot interpret the affirmation of this cognitive relation by the Vedntist as the affirmation of an
ultimate reality. Apart from the philosophical standpoint of the
Vedntist, he contends that he has shown a logical difficulty in the
assumption of the reality of the cognitive relation advocated by the
logician. The logician has made a counter-allegation that the denial
of the reality of the cognitive relation by the Vedntist will preclude
him from using this cognitive relation as the ground of the denial of
the objective reality. The Vedantist in defence has advanced a
1 See nandaprna's commentary on KhKlvpo 992. r:- ; - ''', ~
2 ...ubhayatah-ps pratibandirajjur bhavata eva durnivr syt. "m
KhKh, p. 992.
276
INDUCiiVH REASONING
[CHAP.
counter-plea that the failure of repudiation of the cognitive relation

will amount to the admission of the opponent's position.
Apart from
the ultimate justification of the position which we have indicated above

the Vedntist has only confronted the logician's counter-allegation with
another counter-allegation.
met by a tu quoque.
equalisation.
Sp the tu quoque of the logician has been
From the logical standpoint it is a question of
What is intended by the Vedntist is that tu quoque
cannot be a legitimate argument, because it is always possible to meet

one tu quoque by another tu quoque and there is nothing to choose
between them.
If you admit the validity of one, you must admit the
validity of the other.
But then the position becomes similar to what
is called mutual opposition.
When two parties seek to prove two
opposite conclusions by equally plausible arguments, employing reasons

of equal plausibility, the result is the failure of proof of either position.
The position is also the same when one tu quoque is assailed by
another tu quoque1.
As regards the dictum of Kumrila that the tu quoque serves to
act as a weapon of refutation by showing the nadmissibility of an
argument exposed to the same charge, the Vedntist categorically
affirms that he does not admit the validity of it.
argumentum
ment.
ad verecundiam.
It is at best an
Mere appeal to authority is no argu-
And if again authority is admitted to be a logical argument
by the logician, it can be countered by the authority of Uddyotakara.

Uddyotakara maintains
difficulty
that when the opponent
seeks
to avoid*
by means of a tu quoque, he admits that his position
is exposed to the defect
alleged
and the
counter-allegation of
the same defect in the opposite position does not constitute its
justification.
On the contrary it amounts to the admission of the
defect and this admission should be adjudged
as the ground of his
defeat2.
mai'vam. mithoviruddhyh pra,tibandys tarkbhsatvt...

KhKh, p. 992.
2 ...iti bruvann. Uddyotakaro "yatrobhayor>f itydi vadato Bhattasya
pratibhajclkartavyah.
Ibid.y p. 993.
CHAPTER XIV
In the preceding chapter we have dealt with the criticism of
reasoning as propounded by Sriharsa and the finding of Srlharsa has been
that reasoning as an instrument of refutation is entirely undependable.
Reasoning is based upon empirical evidence and is consequently subject to the limitations of human experience. The criticism of rharsa
has made it abundantly clear that the concept of opposition is not
capable of being derived from experience, because experience does not
record a single case of compresence of opposites and the observation
of their separate incidence cannot give a guarantee of the universality
of their mutual exclusion. Empirical evidence is vitiated by contingency. It must therefore be admitted that the fundamental concept
of opposition which we unhesitatingly believe to be true of experience
universally and necessarily, is derived from the inherent constitution
of our minds. In other words, it must be a priori and subjective and
ideal. The compresence of opposites, of being and non-being of the
same thing, is not endorsed by experience. Were it so there would
be no opposition. Again the separate incidence of the opposites is
not calculated to give rise to the idea of opposition, because there is
no repugnance felt if the pen be present in one place and absent in
the other. The opposition is felt when these two elements are ideally
brought together. So opposition is always ideal and never actual and
hence sense-experience can never be an organ of the discovery of
opposition. No guarantee can be held out that this concept will
hold good of the objective reality in a necessary and universal reference.
We have seen that the logicians of the Nyaya-Vaisesika school
have posited two types of reasoning, namely, (i) which serves to confirm the conclusion (yisayaparisodhaka), and (2) as a condition of the
knowledge of universal concomitance (vyptigrhaka) by means of the
elimination of the opposite possibility. We now propose to deal
with the criticism of the claims of reasoning as an organ of induction
by which the Jaina logicians have set enormous store. In the light
278
INDUCTIVE REASONING
[CHAP.
of Sriharsa's critique we shall have to admit that this claim is not

absolutely beyond dispute. The pivotal basis of inference is universal
and necessary concomitance between what is called a probans and
what is called a probandum. We have observed time and again that
human experience is unequal to this task. A term can be a probans
if it is indubitably known to be necessarily and universally concomitant
with the probandum. This necessary and universal relation can be
asserted only if it can be definitely known that the probans does not
occur in a place which is definitely known to be bereft of the
probandum. The possibility of doubt of the occurrence of the probans
in a locus known to be bereft of the probandum will make this
probans ineffectual as the proof of the probandum. It is required
that the probans should be above suspicion like Caesar's wife, so far
as its occurrence in a locus bereft of the probandum is concerned.
Only that can be regarded as a probans which is definitely known to
be impossible of occurrence in a place lacking in the probandum and
if subsequent evidence attests its occurrence in a place which is bereft
of the probandum or is calculated to create suspicion of the occurrence
of the probans in such a place, the conclusion will not follow.
What is then the guarantor of the probans being necessarily concomitant with the probandum and of the impossibility of the occurrence of the probans in a place devoid of the probandum? As a
matter of fact the multiplication of instances of the co-occurrence of the
two terms cannot confer an immunity upon the probans from its
possible deviation. In one word, the doubt of the contingency of the
concomitance cannot be overruled, unless and until the occurrence of
the probans without the probandum can be shown to be logically and
ontologically impossible. The logicians have banked upon reasoning
for the achievement of this certitude. But reasoning as we have seen
is effective only when the concomitance of the ground and the consequent is securedly established. In other words, the necessary concomitance of two terms can be established only if we are sure of the
necessary concomitance between their corresponding negatives. Thus
smoke can be universally and necessarily known to be concomitant
with fire only if it can be known that the negation of fire is
necessarily concomitant with the negation of smokewhich
1_V J
is the supposed deliverance of reasoning. The concomitance o the

probans and the probandum is supposed to be guaranteed by reasoning. But as reasoning can be effective only if it is based in its turn
upon the concomitance of the negatives, the question naturally arises,
what is the guarantor of this second concomitance? If we make
another reasoning the guarantor of it, the result will be a regressus
ad infinitum. If reasoning be supposed to lack in necessary concomitance, it will be open to the charge of lack of logical sanction. The
dilemma is inevitable whether reasoning is supposed to be based upon
necessary concomitance or proclaimed to be independent of it1.
The logician may retort that the criticism of Sriharsa is suicidal.
Sriharsa's argument itself is a case of reasoning. The argument that
the admission of necessary concomitance as the basis of reasoning will
lead to infinite regress presupposes the necessary concomitance between
such admission and infinite regress. So Srlharsa also will be.faced
with the dilemma. If he admits another reasoning as the guarantor
of the necessary concomitance in question, he will have to face the
music of infinite regress and if he denies necessary concomitance, the
lack of logical sanction will stare him in the face. The criticism may
be regarded as a case of tu quoque. But tu quoque or no tu
quoque the critic cannot evade the responsibility of vindicating the
justice of his argument when he elects to enter the arena of debate
which is governed by strict rules of the logical game. If the critic
refuses to play the role, his criticism will be dismissed as a case of
blatant sophistry2.
Srlharsa in defence does not rely upon tu quoque or its defects.
He avers that he believes like the logician in the empirical validity
of the logical instruments. But he asserts that these logical categories
are nothing more than postulates, which, however useful they may be
1 ...tarkasya vyptimlatvbhyupagame 'navasthprasangt. tadanabhyupagame mlasaithilyena tarkbhsatvpatt.

KhKh, p ; 677.
2 See nandaprna's commentary on KhKh, p. 677: nanv ayam api
kascit tarko 'sya vyptimlatve tatkhandanavyghtas tadamlatve prasithilatnlatvam...
28o
INDUCTIVE REASONING
[CHAP.
in the conduct of theoretical and practical transactions in life, cannot

be believed to be possessed of ultimate and final validity, because of
their self-contradictory nature. The Vedntists accept the logical
categories as postulates and believe in their empirical validity. What
they are interested to prove is that these postulates have no metempirical and metaphysical validity1. Our whole life and its activities
are subjected to move in the circle described by nescience and beyond
this circle the concepts of logic or ethics have neither relevancy nor
necessity. The Vedantist insists that our empirical sciences such as
logic, ethics, or whatever else it may be carry us to a great distance,
but not long enough to enable us to cut this circle of original ignorance. All our enterprises and adventures are doomed to futility, if
we depend entirely and unhesitatingly upon axioms and postulates
of these sciences as the wherewithal to solve the ultimate mysteries.
For the realisation of the ultimate truth we have to devise another
organ, which is not irrational but suprarational, which is not opposed
to human reason, though it transcends the latter.
The logician does not agree to make an unconditional surrender to
the Vedantist's terms. He admits that reasoning depends for its
effectiveness upon the necessary concomitance between the ground and
the consequent. But the admission of this basal concomitance does
not necessitate infinite regress, because there can be no possibility of
doubt regarding an ultimate fact, and the final concomitance from
which reasoning derives its probative force is such an ultimate fact.
Take for instance, the concept of causality. It is an ultimate fact
which makes doubt about its validity self-contradictory. When a
phenomenon is found to occur occasionally and contingently, the
human mind instinctively asks for an antecedent condition as its
explanation. And this antecedent condition is called the cause,
provided it satisfies the other tests, namely, invariability (niyama),
immediacy (avyavahitatva), and unconditionality (anattpadbikatva or
ananyathsiddhatva). By observation of concomitance in agreement
and in difference, fire is ascertained to be the invariable, immediate
and unconditional antecedent of smoke. Let us see whether a doubt
i See also nandaprna's commentary on KhKh, p. 677.
xiv]
281
about smoke being the effect of fire is possible. The doubt of smoke
being caused by an antecedent condition is not possible, because smoke,
if uncaused would be either an eternal entity or a fictitious non-entity.
But smoke is neither eternal, as it Is seen to occur at a determinate
place and time, nor a fiction, because in that case it would be nonexistent even at the time when it was affirmed to be existent by the
opponent himself1.
The opponent may contend that the concomitance of smoke with
fire is based upon the assumption of the impossibility of smoke in a
fireless placein other words, upon the concomitance of negation of
fire with the negation of smoke. But this basic concomitance may be
called in question, so far as its necessity is concerned. And if another
reasoning is pressed into requisition for the rebuttal of the doubt, the
result will be infinite regress, because the second reasoning also will
necessarily be in the same predicament. The logician would assert
that this doubt is impossible, because the necessity of the concomitance
based upon causality is final and ultimate, and as such is not open to
doubt. It is final and ultimate, because it does not stand in need of
corroboration by another reasoning, and also because the doubt of it
involves self-contradiction. The concomitance is regarded as ultimate
also because it is unavoidable. The denial of the validity of such
concomitance will make the knowledge of necessary concomitance
impossible in every reference, and this will again make in its turn the
knowledge of causality impossible, because causality is a case or
necessary concomitance. The denial of causality or of its knowledge
will make the opponent's participation in debate impossible. The
assertion of an argument for or against a position is made with a view
to the conviction of the opponent. This conviction is the effect, and
the assertion is the cause of it. The repudiation of causality will
make his own activity mpossible. In other words, the opponent will
be convicted of contradiction by his own action and by his own
speech. This will also be tantamount to self-contradiction in another
1 ...tath hy agnidhmavyabhicrasankym bdhakas tarko yam
abhidhiyate yadi dhmo'gnirn vyabhicared akranakah san nityah syt, na syd
eva v. sa c'yam anuttaras tarkah, tatra s'ankym vyghtpatteh.
KhKh, p. 578.
36
282
INDUCTIVE REASONING
[CHAP.
way, since to say that a thing is an effect and at the same time independent of a cause amounts to the assertion that it is not an eifect at
all.
To say that an effect is not an effect, which is involved in the
assertion o its independence of a cause, is equivalent to the repudiation

of the nature of a thing 1 .
It follows therefore that scepticism cannot
be pushed too far and must recoil upon the opponent when it confronts him with self-contradiction.
It is possible to argue:
immune from doubt.
"Let the inference based on causality be
Let it also be granted that causality is an
ultimate relation, the denial of which makes self-contradiction unavoidable".
It is possible to contend that after all the necessity of causa-
lity is only a subjective necessity.
Our thinking principle is such
that either by its natural constitution, or by acquired habit instilled

by the social environment that, the enquiry for the cause of an event
is an irresistible impulse.
Kant thinks this belief in causality to be
inherent in the human understanding and he accordingly calls it a

form of understanding, that is to say, the natural and inevitable way
in which the human mind interprets successive events.
that the belief in causality is rather customary.
Hume thinks
Whatever may be
the genesis of the belief in causality, whether it is derived from the

constitution of the human mind or fostered by custom, the necessity
of the causal relation need not be more than subjective.
There stand
formidable difficulties in the way of belief in the ontological necessity

o causation.
Ngrjuna has emphasised the difficulty in the under-
standing of the ontological relation between a so-called cause and a

so-called effect.
He asserts that the effect is not produced by itself.
In other words, the cause and the effect cannot be identical as the
Smkhya maintains, since that would make causal activity superfluous.
There is no meaning in supposing that the effect was existent before
its production, because production means the bringing into existence
of a fact which was non-existent before.
made existent is absurd.
To think that an existent is
Thus the Smkhya theory of pre-existence
i tad eva hy iankyate yasmin nsaqkyamne svakriyvyghatdayo dos

n 'vataranti'ti lokamaryd. evam sarvatr 'nuttaras tarko bdhako 'bhidheya
iti.
KhKh, p. 678,
xiv]
283
of cScct cannot commend itself. The Naiyyika's theory that the

cause is entirely different from the effect and the relation between
them is one of otherness is also unsatisfactory. We want to understand the intrinsic nature of causality, the belief in which seems to be
universal. To say that oil is produced from sesame seeds, though oil
was not pre-existent in them, and though there is no intrinsic relation
between the cause and the effect, which would compel us to affiliate
a particular kind of effect to a particular kind of cause, cannot amount
to a philosophical explanation. It is only the assertion of an empirical
fact that A follows B and we have not observed any deviation from
this rule. The question why should fire produce light and not darkness, the opposite of it, though both are on a footing of equality, so
far as the lack of intrinsic relation with fire is considered, the logician
only with iteration added to vehemence if not arrogance asserts that
it is light that follows fire and not darkness, and because the concomitance of fire as a class with smoke as a class has not been found to
admit of deviation, the causal relation cannot be impeached. But even
the most cussed sceptic does not doubt that there is regular
sequence between the two kinds of events and this sequence has been
uniformly observed. What he is sceptical about is the ontological
necessity of relation between them which will make doubt about its
variability impossible. The logician has failed to carry conviction to
the critic or to set at rest his honest scruple.
Ngrjuna contends that no intrinsic relation has been found between a cause and an effect. It is not identity nor pure otherness.
Nor can it be supposed that the relation is synthesis of both, because
apart from the logical incompatibility involved in this supposition, the
synthesis will be open to attack on both fronts. But Ngrjuna admits that the. repudiation of causality by the Crvka Materialists who
affirm that there is no need for postulating a causal relation, because
it is the nature of events to come into existence at a determinate time
and place without any compelling influence being brought to bear
upon them, is tantamount to confession of failure1. However much
1 na svato n 'pi parato na dvbhym n 'py ahetutah/, utpann jtu
vidyante bhvh kvacana kecana/^
Mdbyamikakrikj I. i.
284
INDUCTIVE REASONING
[CHAP.
che causal relation may be denounced it will never succeed in stifling

our natural curiosity and demand for an explanation, "Why should
an event occur at a determinate place and time, and not at another
time and place, or always or never, if there be no determining reason
for it?'* So absolute indeterminism has not commended itself to the
conscience of a thinking person as a satisfactory explanation. The
belief in causality is inevitable in spite of the demonstration of the
irrationality or rather lack of rational justification by the followers of
Ncrarjuna and the Vedntistic monists. Granted that the relation
of causality cannot be explained in conformity with the accepted
categories of logic, yet the philosophers have not been able to eradicate or dislodge this belief in the necessity of causal explanation.
Call it an instinct or superstition engendered by Maya it is there and
no amount of cajolery or logical coercion will succeed in making us
rid of ic
It may be urged that causality, and inference based upon it may
not be capable of being denied, but all inference is not reducible to
causality. What about other forms o inference? Take for instance,
the inference of perishability on the ground of being a product. It
is held that whatever is a product is liable to destruction. This relation between the fact of being a product and the fact of being
perishable is not causality. It is believed that destruction though a
product is not liable to further destruction. In other words, destruction though product is not perishable. What is then the guarantee
that a sound which is the product of an operation should not be
imperishable like destruction? So the argument of causality, though
ic succeeds in protecting inferences based upon it, is not the solvent
of the logical difficulty, so far as non-causal forms of inference are
concerned. The logician answers this objection by asserting that
causality is not the only ultimate fact which disarms doubt. That
also is not liable to doubt which makes self-contradiction inevitable. It is
this self-contradiction which is an ultimate fact and makes doubt impossible. The argument "Sound is perishable, because it is a product",
is absolutely valid though it is not based upon causality. If a product
were not perishable, ic would be an ecernal verity. To say that a thing
is a produce and is a.c the same time an eternal verity is to be guilty of
xiv]
contradiction in terms.
of
eternity.
285
The very concept of product repels the idea
Furthermore the assertion of
eternity of sound would
make its eternal perception inevitable, which is however, opposed to

the universal verdict of experience.
So self-contradiction here is the
re duetto ad absurdum of the doubt of the opposite possibility and is

the ultimate and final clincher.
relevant.
The analogy of destruction is not
Destruction is not believed to be liable to destruction,
because that would make the revival of the

ble.
object destroyed irresisti-
But no resuscitation of a thing once
experienced.
not been
destroyed has been
But so far as a positive entity is concerned it has

found to be existent
eternally
if it has been found
to be produced at a determinate place and time.
The logician
concludes that the validity of inference is determined by reasoning

which dispels all doubt by tracing the validity to an ultimate fact,
whose validity is self-evident
and
the denial of which involves
seif-contradiction.
The Vedntist would assert that the logician has tried to prove
the obvious and his is a case of wasted labour.
The Vedntist does
not deny that the belief in causality is an ultimate conviction which

no amount of logical argument can succeed in dislodging from the
human mind.
Ontologically the Vedntist holds that the belief in
causality is as much a superstition of ignorant minds,

the identity of the soul with the body.
as the belief in
But his logical difficulty is
not derived from his metaphysical theory.

does not call in aid the metaphysical argument.
In the arena of logic he

He takes for granted
rhat causality is an ultimate fact and is repugnant to doubt.
Though
the Vedntist does not believe in the absolute validity of causality as

a metaphysical fact, he accepts it as a valid empirical principle and
logic also does not presuppose more than this empirical validity.
The
Vedntist admits that smoke which is a product must have a cause.

It is not held that smoke is possible independently of a cause.
What
the Vedntist doubts is that there may be a plurality of causes

for smoke.
What is the guarantee that fire is the only possible cause
of smoke?
Smoke may be supposed to be the product of a cause
which is other than fire and this does not involve contradiction
2S6
INDUCTIVE REASONING
[CHAP.
of the belief in the ultimate validity and necessity of the causal

relation1.
It may be argued in defence that this doubt of a plurality of
causes does not occur normally to a person, who is acquainted with
the use of fire.
It may be so. But that does not make the doubt
of a philosopher impossible.
The possibility of doubt of fire being
the sole cause- of smoke makes the inference of fire from smoke
precarious.
This shows that the belief in the necessary concomitance
of smoke with fire is only possessed of empirical validity and not the
sort of validity which s presupposed even in logic.
The belief in the
abstract necessity of causality does not afford any advantage, so far

as its application to concrete instances is taken into consideration. So
long as the possibility of doubt of other causes is not eliminated, all
our inferences based upon causality will fail to achieve certitude. What
is true of inference is true of reasoning also, because both are dependent upon the validity of necessary concomitance.
Even if the
concomitance be reduced to causality which as an abstract principle

must be believed to be universally valid, the validity of the belief in
the necessary concomitance of smoke with fire cannot be securedly
established.
It is possible to argue that this doubt is not spontaneous
and rather bespeaks a hypercritical attitude, as no instance of smoke

being produced by anything other than fire has been observed by
anybody.
But non-observation is not the proof of its impossibility.
There is no assurance that such a case will not be observed
in future,
or is not observed in an unknown region.

The logician has argued that if smoke be capable of being produced from fire and what is not-fire, then smoke will be heterogeneous
owing to the heterogeneous
character of the causes.
A smoke which
is the product of fire will be different in kind from the smoke which
is produced from non-fire.
This will mean that smoke is not homo-
geneousan admission which is not warranted by experience2.

1 ...... kim ity evam sakitavyam y ad dhetuphalabhva eva na bhavisyati.
evan tu samkitavyam. agnim vihy 'nyasmd api hetor ayam udesyat'ti.
KhKh, p. 679.
2
na ca vcyam evam hi sati dhdmasy 'aikajatitvam na syd iti.

Ibid.
xiv]

The Vedntist is not convinced by this argument.
287
There may be
different species of smoke and in spite of the difference of one species

from another there is no reason why should the different
species not
possess a common generic character, just like the different human races.
The difference of the generic character of the causes of smoke can at
most account for the difference of characteristics of species belonging
to the same genus.
Furthermore the very assumption that the
difference in the generic character of the causes will entail the difference in the generic character of
by logic or by experience.
the effect is not guaranteed either
It is not a fact that different causes do
not produce a homogeneous effect.
Cognition is produced by various
causes. Thus the cognition of a jar is produced by sense-object contact.

The cognition of the sense-organ is generated by a concomitant probans and that of religious merit and demerit is engendered by the
language of scripture.
, produced by different
But all these instances of cognition, though

and
heterogeneous causal conditions are homo-
geneous, because they all share in the common generic character of

cognition which may be called the cognition-universal1.
So there is
nothing repugnant in the theory that various heterogeneous causes can

produce independently of one another the same kind of smoke possessing the same generic character.
The contention
that homogeneity of
effect can be accounted for only by the homogeneity of the cause

must be regarded as an approximate estimation of the causal relation.
It has been contended by the logician that it is preposterous to
suppose that various
heterogeneous causes can produce a homogeneous
effect. A jar is produced from clay and a linen textile is produced from
cotton yarns. It is never found that they are produced promiscuously
a linen from clay and a jar
from yarns.
There is no deviation from
the law that a homogeneous cause produces a homogeneous effect and

a heterogeneous effect is produced by a heterogeneous cause.
regards the exception adduced by the Vedantist
tion a little reflection will show
As
in the case of cogni-
that there is no breach of the law
1 kvacid indriyajanyatve kvacid anumndijanyatve 'pi janaikajtyavat

tadupapatteh.
KhKh, pp. 679-80.
2 8S
INDUCTIVE REASONING
even in this case.
CHAP.]
The cognition that is produced
by sense-object
contact is intuitional in character, the cognition produced by the

concomitant probans is inferential, and the cognition produced by
language is verbal in nature.
So there is no absolute homogeneity in
the cognition variously produced by various
causes.
Sense-object
contact is responsible for the intuitionality, probans for the inferentiality, and language for the verbality of the resultant cognition1
The Vedntist does not accept this explanation as satisfactory.
If
the different conditions of cognition are responsible only for the

specific character of cognition, such as intuitionality etc., then the
common condition of the generic character, viz. cognitionhood or
cognition-universal, which appertains to all cognitions in spite of their
variation as species, has to be formulated.
It will be immediately
shown that no such common condition is possible of assertion. And

in the absence of a common condition the generic character of all
cognitions, viz. their cognitionhood will remain unaccountable2.
If
again cognitionhood be undetermined in its incidence, there will be

no logical bar to the promiscuous occurrence of cognitionhood even
in inanimate objects, say a plate or a pen, and in such psychical entities as conation, aversion, and the like.
mining condition of the incidence
So there must be a deter-
of cognitionhood.
And as a
common condition is not apprehensible, it is logically inescapable that

we should assert a sense-organ, or a probans, or a linguistic symbol,
variously and independently of one another as the condition of it.
The law that the nature of effects is determined by the nature of
causes is more fundamental than the law of homogeneity.
In confor-
mity with this law we must admit that a plurality of heterogeneous

causes can produce a homogeneous and unvarious effect, particularly
when a homogeneous character in the causes is not comprehensible.
Let us waive for the present our objection regarding the lack o
a common ground in the plurality of causes of cognition and agree
1 KhKh, p. 68o, See also nandaprna's commentary on KhKh, pp.
680-81.
2
jnnatvasy'kasmikatvaparihrrtham tatkranasy' nugatasya
bhavat
'vasyam vaktavyatvat.
KhKh, p. 680.
xiyj
289
for the sake of argument that the different conditions only contribute
to the different characteristics in the effects as the Naiyyika insists.
Even then we can contend that smoke can be produced by fire as
well as by some other cause. We may presume in imitation of the
logician that fire is responsible for a particular difference in the nature
of smoke. In other words, there is no a priori difficulty against the
postulation of different species of smoke. Smoke produced by a fire
may have a different specific character which will distinguish it from
the smoke produced by some other cause. So the postulation of a
plurality of causes for smoke cannot be dismissed as a logical absurdity.
The result will be that the absence of fire will not be the ground for
the inference of the absence of smoke as such, though it may legitimately serve as the ground for the inference of the absence of a
particular species of smoke1.
The logician may contend that the contention of the Vedntist
could be accepted if the difference of character in smoke were discernible like the difference of nature in the various kinds of cognition
produced by various causes. In the absence of such variation how
can it be contended that smoke may be produced by fire as well as
what is not-fire? All our experiences of smoke attest the presence o
a uniform generic character in each instance. It is unthinkable that
a different species of smoke produced by a cause other than fire
would have escaped the observation of mankind up till now, if any
such species were possible2.
The Vedntist argues that the Naiyyika misses the point of his
contention and lays emphasis upon a fact which is not disputed. It
is not asserted that a different species of smoke is observed. But the
non-observation cannot a priori prove the impossibility of such a species.
As regards the contention that up till now no heterogeneous trait in
smoke has been observed, it is admitted as a face, but this may be
due to an accident that smoke which is produced by another cause
x dhme 'pi vahner vsese eva prayojakatvasya tad-vac chakitum
sakyatvt.
KhKh, p. 680. See also nandaprna's commentary on KhKh, p. 68o,
2 na drsyate tvad agniprayojyo dhme visesa iti......
KhKh, p. 680.
37
29o.
INDUCTIVE REASONING
[CHAP.
has not been observed. The condition of perception of a heterogeneous trait in smoke is the perception of different species and because
such a species has not yet been discovered all observed cases of smoke
which are invariably produced by fire are presumed to be possessed of
a homogeneous nature. One cannot deny the a priori possibility of
other kinds of smoke produced by causes other than fire simply on the
basis of our present knowledge. The savage people of an undiscovered
island may deny the possibility of aeroplane on the basis of his knowledge. But this only proves his ignorance and not the objective
impossibility of his object of ignorance. The same may be the case
with regard to the different species of smoke, which though logically
thinkable, may have escaped all observations. It is quite plausible
that such species of smoke may be discovered in future and then we
shall be aware of the difference of nature of the different species of
smoke inter sel.
The logicians have however attempted to account for the community of the nature of various cognitions by means of the co-operation of the mind with the soul which is alleged by them to be the
universal condition of all cognitions. The Vedntist's contention
that even heterogeneous causes can produce a homogenous effect is
based upon the supposed lack of a common condition in the case of
different species of cognitions. He has banked upon the impossibility
of the formulation of such a common cause in support of his
position. But the contact of the mind with the soul is regarded by
the logician as the universal condition of cognition and the different
conditions superadded to this, only account for the specific difference,
such as intuitionality and the like. The Vedantist however has demurred to accept this as the common cause of cognitions. The mind-soul
contact is also the condition of desire, volition, aversion, love, antipathy
etc., which are non-cognitional in character. So this cannot be regarded as the common condition of cognitions. If however the logicians
posit an unperceivable entity such as religious merit or demerit, or a
causal energy inherent in the different conditions of cognition, or the
I
......na vcyam. tadadarsanasy'ptato...iti sambhvanay durnivratvat.

KhKh, p. 680.
xiv]
'291
presence of cognition-universal, or the pre-negation of cognition, as the

common condition of the common cognitional character of the different
species of cognition, the Vedntist also may formulate some such unperceived fact as the common condition of smoke-character,
to the different
species of smoke.
pertaining
The differences of the specific
nature of the different species of smoke may then be accounted for by

the different conditions such as fire and not-fire superadded to this
common condition1.
The logician however does not regard the two cases as analogous.
The postulation of unseen moral destiny, or some such unperceived
fact, as the common condition of the different species of cognitions, is
necessitated by the consideration that the accepted
conditions of
perception or inference are seen to vary in different cases, and because

there is no observed common ground in the causal conditions concerned.
But the case of smoke stands entirely in a different position.
No
instance of smoke has been observed, to be independent of fire as its

condition.
And no instance of smoke again has been observed to
manifest a heterogeneous characteristic like perceptual and inferential

cognitions.
So the postulation of an unseen moral destiny as the
common causal condition of smoke is uncalled for.
Besides, this runs
counter to the accepted dictum of scientific and philosophical speculations that an unobserved condition should not be postulated when an
observed condition can account for an observed fact2.
however is not satisfied
by this explanation.
The Vedntist
He complains that the
logician has sought to minimise the difficulty in his eagerness for an

easy solution.
It is never disputed that no case of smoke unproduced
by fire has been observed.
But the possibility of such a species of
smoke is not ruled out by this empirical evidence and this makes the
inference of fire from smoke or the inference of the absence of smoke
from the absence of fire precarious3.
1 asty tmamanoyogo 'nugatam kranam
sakyata eva sakitum.
jnotpattav iticet? na
KhKh, p. 682.
2 drste vyabhicre yuktam adrstder aikajtypparikalpanam iti cet?
KihKh, p. 6S3.
3
Ibid.
2g2
INDUCTIVE REASONING
[CHAP.
The logician contends that the admission of the possibility of

plurality of causes will make all inference impossible. Whenever an
inference will be sought to be made from an effect the sceptic may
oppose it by pointing to the possibility of an unknown cause. The
result will be an unrelieved deadlock of all theoretical and practical
activities. If the Vedntist were sincere in his doubt, he would be
precluded from participation in a debate. The Vedntist argues for
the conviction of the opponent believing him to be an embodied spirit.
But the knowledge that the opponent is a spirit can be secured only
by an inference. And this inference may be precluded by the suspicion that the opponent may not be possessed of spirit like the proponent, since there may be a cause other than the spirit to account for
the opponent's vital activities on the basis of which the spirit is inferred. So the attempt to throw doubt on concomitance of an accepted
cause and an accepted effect by the Vedntist will entangle him in the
meshes of self-contradiction.1
The Vedntist asserts in reply that the logician only sees one side
of the issue which seems favourable to him. Just like Nelson he
conveniently puts the telescope on the blind eye to the other side of
the problem, and pretends to think that the difficulty is solved. The
Vedntist contends that the supposed necessary concomitance, which
is the basis of inference and reasoning alike, is susceptible to doubt.
Now this doubt of the fallibility of concomitance is actual. And
there are two courses open to the logician. He may either assert that
doubtful concomitance is not a fault, or that a probans of doubtful
concomitance though fallacious may yet be an organ of valid
knowledge. The first alternative is out of the question, since the
logician also regards it as a fallacy. The second alternative will make
a false inference also pass as valid. For instance, when a person is
aware of only two kinds of fire, one produced from faggots and another
produced from hay, he may naturally infer on seeing fire that it is
produced either from faggot or from hay. But this is certainly fallacious as there may be other causes of fire, say, friction of sticks, or
i evam sankamnasya bhavato na kvacid anumnam syt
sa eva
vyghta iti cet?,
KhKh, p. 683.
xiv]
coal and so on.
293
The argument Is fallacious, because fire's concomi-
tance with other causes Is possible of doubt.
The case is entirely
analogous with smoke, because its concomitance with a cause other

than fire is equally a matter of doubt1.
The logician cannot aver that the conditions of doubt are not
present in the case of smoke's concomitance with
fire.
It has been
shown that our ignorance of other causes of smoke is not proof of the
impossibility of the same, just as the ignorance of the person of other
causes of fire except faggot and hay does not make his inference of either
of these two alternative causes of fire immune from logical difficulty.
Now the conditions of doubt are the cognition of common attributes,
the non-cognition of special traits, and the cognizance of the two
alternatives.
Let us see whether these conditions are present in the
case of smoke.
cause.
Now smoke is a product and as such must have a
It is presumable that smoke may be the product of fire as
well as what is not-fire.
Now there is no known criterion by which
the observed smoke may be affiliated
to fire or to not-fire exclusively.
From our present knowledge we are not in a position to discriminate

between a smoke caused by fire from that produced by not-fire.
The
smoke-character or the attribute of being a product is common to both

varieties.
And this common character easily presents before us the
two alternatives, fire and not-fire as the possible causes.

are the conditions of doubt.
And these
But the logician who pleads for the
unquestionable validity of the inference of fire from smoke expects

us to believe that there may be a cause in all its fullness and yet the
effect will not materialise. The employment of smoke as ground for the
inference of fire is exposed to the same difficulty as the employment
of fire in the aforementioned
inference.
If smoke be accorded a
privileged position n spite of the presence of all the causal conditions

of doubt of the fallibility of the concomitance,
the natural deduction
from this will be that there is no necessary connection between cause

and effect.
But the acquiescence in this position which has turned
1 dhmavad vahnec api vahnikaranavisesnumnasy' aivam sati sadanumnatvaprasan gt.

KhKh.'p. 683.
INDUCTIVE REASONING
294
[CHAP.
out to be unavoidable for the logician will involve him in selfcontradiction1.
His employment of an argument in order to prove
or disprove a position is inspired by his belief in the infallibility of

the causal relation between argumentation and
tion.
production of convic-
The repudiation of the necessity of causal relation which is
implied by his denial of doubt in the presence of the full conditions

of it ill goes with his argumentation for the conviction of the opponent
which proceeds from his belief in the causal relation.
The logician however pleads in self-defence that the Vedntist's
attempt at putting accredited causal inference on the same level with
palpably false inference is bound to fail.
He does not agree that the
conditions of doubt are present in the case of inference of fire

from smoke.
The contradiction which such doubt involves serves
as the clincher of it.
The conditions of doubt
include the non-
cognition of special characteristics of a particular alternative, but the

contradiction resulting from doubt acts ipso facto as the cognition
of a special trait of fire.
So the charge of the Vedntist that all
causal inferences are affected by the doubt of a plurality of causes falls

to the ground, because the contradiction involved by such doubt shows
that the doubt is not entertainable and as such the conditions of doubt
are not present3.
The Vedntist does not accept this advocacy of
validity of
inference by appeal to contradiction as a satisfactory mode of argument.
He argues that contradiction may be regarded as a clincher
if the knowledge of it be secured by an accredited organ of knowledge.

If the contradiction alleged be inspired by a fancy it will not act as
a solvent of doubt.
If on the other hand, contradiction be a real
verity, then it must have a real condition.

contradiction is alleged to be doubt.
Now the condition of
Doubt is thus the support and
the abode of contradiction. If contradiction be proved by an accredited

1 sdhranadhartnadarsanaviiesdarsandau saty api
tulyah.
svakriyvyghtas
KhKh, pp. 683-84.

2 vyghtasy'aiva visesatvt taddarsanena sanksmagry eva n'sti matpakse
kuto vyghtasmyam iti cet ?
KhKh, p. 685.
xiv]
295
organ of knowledge, the latter will automatically and eo if so prove

the reality of doubt. Contradiction is not a floating attribute with
no local habitation. It is felt only when there is doubt. Unassociated
with and irrespective of doubt contradiction is an impossibility.
If contradiction be possible even without reference to doubt, then it
will be inescapable both for a person who entertains a doubt and for
a person who is not disturbed by such a visitation. Contradiction
arises from doubt and is thus an invariable concomitant of it. The
proof of one will be the proof of other. The logician may plead that
contradiction is opposed to doubt and though it may not destroy the
first doubt which is its condition, it will preclude emergence of further
doubt. But this does not afford any advantage, since the initial
doubt will make the necessary concomitance unreliable, and when
this doubt ceases, contradiction also will automatically cease. In the
absence of contradiction what will prevent the emergence of further
doubt1?
We have shown in our examination of contradiction in the
preceding chapter that contradiction is an ideal fact and is purely
subjective. But the logician of the. Nyya-Vaisesika school is an .
empiricist and a realist. He is constrained * by his metaphysical
commitments to postulate an objective ground for every cognition.
We have shown following Sriharsa that the admission of, an objective
basis for contradiction entangles the realist in a morass of self-contradiction. All the objections of Srlharsa are directed against this
realistic attitude. .An ordinary person may think that these arguments
are pure sophistry, because they only aim at demolishing our conception
of contradiction which is ineradicable. There is substance in the
logician's contention that the argument against contradiction can only
succeed in its objective by showing contradiction in the concept. And
Srlharsa has actually done this. So the argument is suicidal. But it
misses the fundamental point o Srlharsa's criticism. His criticism does
1
vyaghatarpasya viesasya darsanam sakpratipaksabhtam ucyate
tat kirn pramnt kutas cid upajyamnam vaktavyam...kah sakntarotpatter
vrayitc'ti vaktavyam.
KhKh, pp. 685-86.
296
INDUCTIVE REASONING
[CHAP.
not seek to prove that contradiction is an impossible or absurd

conception. In fact the absurdity of an argument can be proved only
by contradiction. What he seeks to establish is that the concept of
contradiction is an a priori one, and the attempt to derive it from
empirical evidence is bound to fail. Viewed in this light the criticism
of Srlharsa should not give a shock to our logical sense, however
paradoxical it may look at first sight.
We have seen that the logician insists that contradiction is the
final solvent of doubt, which (contradiction) according to him serves
as crucial evidence of the untruth of the opposite position and by
necessary implication of the truth of the remaining alternative. The
Vedntist has argued that doubt is the condition and support of
contradiction and as such is bound up with the latter and so there
can be no antagonism between the two. The relation between doubt
and contradiction is one of cause and effect and the effect cannot be
destructive of the cause at least at the time of its genesis. But the
logician points out that a subsequent psychical event is opposed to a
preceding psychical event, though the latter is the cause of the former.
So also with regard to sound. The subsequent sound is opposed to
the preceding sound", though the latter is the cause of the former.
It may therefore be argued that opposition is not incompatible with
causal relation. The logician would agree that there is no opposition
at the time of the genesis of the effect. But the effect may destroy
the cause after having come into existence. Besides, though contradiction may not be opposed to doubt as its cause, -the knowledge of
contradiction or the memory-impression left by it may be destructive
of doubt, because doubt and the knowledge or memory-impression of
contradiction are not causally related. Knowledge of contradiction or
the memory-impression of it is the effect of contradiction which is the
effect of doubt. But the effect of an effect is not the effect of the
original cause just as the cause of a cause is not the cause of the
latter's effect. The logician further insists that contradiction as an
existent is not opposed to doubt. The opposition is not between two
existents per se, but between the cognition o contradiction, as opposed
to the mere presence of it, and doubt. As has been observed before
contradiction serves as crucial evidence (yisesa^. Crucial evidence,
xiv]
297
which is rather the peculiar characteristic of one of the alternatives,

also does not io if so eliminate doubt, but only when it is cognised as
such.
Similarly the cognition or memory-impression of the cognition
of contradiction can destroy the doubt in spite of the fact that it is

the cause of contradiction.
There is a vital difference between contra-
diction and the cognition or the memory-impression of it and when

the relation of opposition holds between these two facts, causally
unrelated, the Vedantist's criticism of k on the basis of causal relationship does not affect the logician's position ia the least1.
The Vedntist would assert that the logician has entirely misunderstood the spirit of his contention.
In the first place, the relation of
opposition is not reducible to causal relation in its entirety.
There
is causal relation no doubt between doubt and contradiction, because

contradiction emerges immediately upon the emergence of doubt.
But doubt is integrally related to contradiction in that doubt is a
content of the latter.
along with it.
Contradiction is felt only when doubt is felt
Whether doubt be regarded as a qualifying adjective
or the content or the support of contradiction, it is undeniable that

doubt is a necessary concomitant of contradiction and is synchronous
with it.
So the analogy of causal opposition which obtains between
two successive psychical events or successive sounds is not relevant to

the problem under discussion.
is an integral part of it.
Contradiction is possible only if doubt
So doubt is bound to co-exist with contradic-
tion and the cessation of doubt will necessarily be concomitant with

the cessation of contradiction.
Thus it cannot be maintained that
there is opposition between doubt and contradiction in the sense of

non-compresence.
As regards the contention of the logician that
the cognition or memory-impression of contradiction will play the

role of crucial evidence and will be destructive of doubt, it will be
found on scrutiny to be as unsubstantial as the previous pleas.
In the
first place, the cognition of contradiction cannot be regarded as the

eliminater of doubt, because doubt is a part and parcel o contradiction
and as such will be an indispensable element of the content of such
cognition.
As for memory-impression it is only a replica of the
1 KJhKh, p. 688.
38
298
INDUCTIVE REASONING
[CHAP.
original cognition and as such is subject co the same limitation.

Moreover it is not warranted by experience that
memory-impression
of a special traic of one o the alternative terms of doubt serves to

eliminate doubt.
Contradiction has been put forward
as a special
trait and a crucial evidence, and its cognition or memory-impression

has been supposed to exercise destructive influence
this
supposition
is unwarranted.
upon doubt.
But
If memory-impression of an
opposite trait were to act as a deterrent or as a destructive agent of

the doubt of the contrary, the potter would not doubt that the black
unbaked jar would become red after its calcination in a potter's
furnace.
Black is opposed to white.
The knowledge or memory-
impression of the black colour of the unbaked jar should make such
doubt impossible. So mere memory-impression of one particular trait cannot be supposed to stand in opposition to the doubt of the emergence of
opposite, when the previously perceived trait has actually ceased to exist
therein.
The previous cognition of the memory-impression of contra-
diction cannot analogously be supposed to be deterrent or destructive

of doubt if contradiction has ceased to exist.
But contradiction is not
opposed to doubt even when it is present and so is not on the same

footing of advantage as the black colour which actually deters the
emergence of an opposite colour during its presence1.
And it has
been shown that contradiction is not only not opposed to doubt but
possesses it as a component part of its being.
It has been shown that contradiction is incapable of eliminating
doubt, as doubt is its very condition and support, whether it be regarded as a part of its content, or its determining condition.
If contra-
diction be supposed to have reference to fallibility of concomitance

between a probans and a probandum as its condition and as a part of
its content, the new content will have the same incidents as doubt has
been shown to have.
Contradiction will be dependent upon this new
substitute and so cannot be expected to act as a hostile agent against

the doubt of necessary concomitance2.
It has been shown that contradiction or the knowledge of it cannot
1 KhKh, pp. 688.89.
2 Ibid., p. 6&().
xiv]
be established by an accredited organ of knowledge.
299
If contradiction
were to be established by a cognitive organ as a veridical fact, the

reality of doubt, which contradiction is supposed to eliminate, would
also be established by it as an unquestionable factual verity.
It may
be supposed that the knowledge of contradiction Is secured, not by a

cognitive organ, but by a reductio ad absurdum in the form "If smoke
were doubted to be non-concomitant with fire, the result would be
contradiction".
But the question will pertinently arise whether or not
the reductio ad absurdum is based upon necessary concomitance between the ground, doubt of necessary concomitance and the consequent
contradiction.
If the necessary concomitance in question were lacking,
the reductio ad absurdum
would be a fallacy.
And even if such a
fallacious reasoning be supposed to prove its objective (sic. contradiction), it will itself be a case of self-contradiction, because to say that a
fallacy is proof is to assert the co-existence of two incompatible facts1.
If however the logician declares that a reasoning which is requisitioned
to prove necessary concomitance does not stand in need of a further
necessary concomitance as its basis, the result will be that any possible
reasoning can be adduced against it, as the question of necessary
concomitance is made irrelevant.
There will then be no criterion for
distinguishing a genuine reasoning from a spurious one, since the

only criterion of genuine reasoning, namely necessary concomitance, is
affirmed
to be irrelevant.
Certainly the logician should be the last
person to subscribe to this position.

If the logician admits that a reasoning is not possible without a
necessary concomitance as its basis, the result will be a vicious infinite
series, since the basal concomitance will be open to doubt like the
previous one.
According to the logician's standpoint reasoning is the
only proof of necessary concomitance by means of elimination of

doubt
So the doubt of the basal concomitance can be eliminated by
another reasoning, for the establishment of whose concomitance another

reasoning will be in request and the process will have no termination2.
1 n'pi yady atra vyabhicrah sakyeta tad vyghtah syd ity evamrpt
tarkd vyghtvagamah
vyghta upanetum.
KhKh, p, 692.
2 atha tarkasya vyptir mlabht 'bhyupagamyate tatr'pi vyabhicrasankym punar anavasthaiva
anavasthaiva.
Ibid., p. 693.
3 oo
INDUCTIVE REASONING
[CHAP.
Udayana asserts that the denial of necessary concomitance will

amount to denial of inference. But why should the sceptic think that
the knowledge of necessary and universal concomitance be impossible?
If it is argued that there is room for doubt whether the observed
concomitance will hold in unobserved places and times and this makes
inference precarious and undependable, Udayana retorts, how can doubt
itself of the concomitance breaking down in unobserved places be
possible? it presupposes that there are other places and times than
the observed ones and this knowledge of other places and times must
be secured by inference as perception is always concerned with the
datum that is here and now. So doubt itself proves the necessity of
inference. And if there be no doubt, then the concomitance will
remain unchallenged and inference based upon it will take place
without a hitch1.
But an honest enquirer may submit: "We admit the ingenuity
of the defence. Even if we take it at its face value, it will only
prove that inference in the abstract is not capable of being repudiated
without self-contradiction. We are however concerned with concrete
cases. It cannot be denied that there is no self-contradiction in
the denial or doubt of the necessary concomitance between smoke
and fire, unless the logician can vouchsafe us the knowledge of
the secret that makes the logician immune from the visitation of
doubt. We do not see how we can be sure of the infallibility
of a particular case of necessary concomitance and consequently of
inference. What is the guarantor of the necessity and infallibility
of concomitance? Our question is inspired by genuine and honest
doubt, and we appeal to the logician to give a straight and an honest
answer".
The logician affirms in reply that doubt is eliminated by a reductio
ad absurdum. The reductio ad absurdum shows that such doubt leads
to self-contradiction and the demonstration of this absurdity makes
the doubter desist from enquiry into the fallibility of concomitance.
i
sak ccd anum 'styeva na cec chaka tatastaram/

vyghtvadhir sak tarkah sankvadhir matah//
NKu, ch. III. 7.
xiv 1
301
With the cessation of enquiry about the opposite possibility, the

enquirer deduces the necessary conclusion from the probans.
ad absurdum
Reductio
served to cut off the diversion of enquiry which
was created by doubt.
W i t h the removal of this obstacle there is
nothing which stands in the way of inference.
But is not reductio
ad absurdum itself based upon necessary concomitance and is not this

basal concomitance open to doubt?
The answer is a yes and a no.
The necessary concomitance is no doubt the basis of the reductio

absurdum,
finitely.
ad
but the doubt of its infallibility cannot be pursued indeIt is possible up till self-contradiction is not confronted.
The final solvent of doubt is self-contradiction, which is the limit

of it.
The Vedntist would reply:
" W e have already evaluated this
defence and shown that it is only a superficial solution of the problem.

It has been shown
that the suspicion of a plurality of causes makes
all causal inferences precarious.
Furthermore, even the abstract state-
ment that doubt of the fallibility of concomitance is removed by

contradiction has been shown to be a hasty assessment, because contradiction has been shown to be unopposed to doubt.
Besides, contra-
diction itself is a case of reasoning and as such is dependent upon

necessary concomitance of the ground and consequent.
Self-contradic-
tion in respect of personal behaviour is not the ultimate solvent of

doubt, as we have shown in our critique of GagesVs position in
chapter iii that behaviour is not conditioned by philosophical conviction.
The concept of contradiction has been analysed by Sriharsa
and has been shown to be an unobjective idea.
Again, it has been
shown that contradiction presupposes doubt as its condition
and
continues so long as doubt exists.
Doubt therefore constitutes the
very life principle of contradiction".
Srlharsa facetiously observes:
" W e can refute Udayana using his own words mutatis
mutandis.
If you admit the reality of contradiction you must admit the reality
of doubt.
If there be no doubt there will be no contradiction a
fortiori because doubt is the condition and content of contradiction.

How can then contradiction be the terminus of doubt?
reasoning, how can it eliminate doubt?
And as for
Reasoning culminates in con-
tradiction and the latter has been shown to be bound up by an
3 o2
INDUCTIVE REASONING
[CHAP.
indissoluble tie of natural alliance with doubtit being the very life
and soul of contradiction" 1 .
But a difficulty may be urged : "Well, even if we admit for the
sake of argument that your analysis of contradiction is correct and
that we have failed to refute your argument yet it cannot be gainsaid
that nobody will be convinced by it.
we confront a contradiction.
We cease to doubt as soon as
This ought to satisfy the critic if he is
an honest enquirer of truth, that there is some sophistry in the argument, which has escaped detection."
The Vedantist will not deny the plausibility of the logician's
contention.
the case.
difficulties.
He will agree with the logician that such is no doubt

A feeling of contradiction appears to set at rest all our
But this is nothing but an appearance, though attested
by human psychology.
But our estimation of truth and reality cannot
be entirely based upon psychological evidence.

ment s at best an argument ad hominem.
Psychological argu-
We found
in chapter iii
that there were logicians who held that there were beliefs and convictions which were ingrained in our mental constitution and consequently universally accepted as truths.
But even Gagesa had to admit
that the wide-spread acceptance of these beliefs was not proof of

infallibility.
their
We had to remark that the question of majority of the
adherents of a belief was logically irrelevant and wide of the mark.

The number of wise men, who would not accept anything on trust
or on any other evidence than logical authentication, has been very
small in all ages and climes.
the salt of the earth.
But this microscopic minority has been
It is they who have been the friends,
philoso-
phers and guides of the academic bodies in this imperfect earth. To

revert to our problem, the attempt at resolution of the
doubt
by
appeal to innate
philosopher's
convictions or to the psychological
evidence has been futile with regard to the Vedantists, who have seen
the limitations of human reason.
i
But this scepticism of human
tasmd asmbhir apy asmin narthe na khalu duspath /

tvadgthai'v'nyathkram aksarni kiyanty api//
vyghto yadi sak'sti na cec chank tatastarm/
vyghtvadhir sank tarkah sankvadhih ktitah//
KhKh, p. 693.
xiv]
303
reason does not argue an anti-intellectual or anti-rational attitude of

mind. The Vedntists were not misologuesthey rather doted upon
reason but refused to make an idol of it. They insisted that logical
training was essential for the progress of humanity and was the
necessary propaedeutic to the realisation of truth. But though it goes
long long enough, it can make only an approximation to the truth*
Some other resource is necessary for the realisation of the ultimate
truth and this resource is supra-rational in character. It is not antirational, but is the efflorescence and fruition of rationality. The truth
is our ownself and self-realisation alone can cut off all our doubts and
problems and bursts all our ties and limitations. The devices of
logical thought are extremely useful in that they carry a sincere
searcher of truth progressively towards the realisation of truth. They
will lead us to the verge of the gulf that separates truth frjm untruth.
And this gulf can be crossed by self-realisation alone. But that 'does
not mean that we can dispense with logic. Logic is the only resource
on this side of perfection. The supreme service of logic consists in
training our minds to detect the flaw in our progressive findings and
to create divine unrest with everything that smacks of imperfection.
But we must be aware of the pitfalls and dangers lurking unsuspected
which even highly trained minds are apt to ignore, being lulled into a
sense of security by the apparent triumphs of logic. The philosopher
must not be afraid of doubt, but on the contrary, should welcome it
as the rude friend who is not afraid of causing pain to us by breaking
our self-complacence. The greatest enemy in the path of the aspirant
for truth is self-satisfaction in the midway. This self-satisfaction is
secured by our natural indolence or inertia to track the enemy to his
den. This enemy is the half-truth which masquerades as full-truth.
The Vedntist's love of logic is inspired by his love of truth and
if he criticises the logician's achievements, he does it with a view to
goad him to accelerate his step in the search of truth. He admits
that the logician's achievements are praiseworthy, but he does not
take them as the achievement of final triumph, which a flatterer and
a false friend would do. The ultimate achievement which can be
scored by logic is to demonstrate the futility of the weapons, which it
has created for the organisation of human thought, after it has been
3o4
INDUCTIVE REASONING
achieved.
[CHAP.
But thought however organised cannot give us direct
glimpse of the face of truth.
It must needs be supplanted by direct
and immediate vision of truth
for which logic is the necessary and
indispensable propaedeutic.
We thought that our treatment of this supreme logical problem,
which has led us to discuss the fundamental problems of logic, would
remain incomplete and imperfect, unless we embodied the results of
our study of the Vedntist's speculations on this subject.
This study
may appear to stultify the previous results of our * enquiries.
If this
be the impression left on the reader's mind, we shall be cordially and

sincerely sorry for it.
We have made a sincere attempt to give a
faithful and at the same time a critical exposition of the thoughts of

the different schools of philosophers on the fundamental logical problems
in order to acquaint a modern student of philosophy with the logical
speculations of ancient Indian thinkers.
We have studiously avoided
partisanship and deliberately refused to play the role of a yesman.

We have reached the end of our quest.
We have shown how the
problem has provoked the ardour and enthusiasm of the logicians of

India and the energetic discussion of the various issues bound up with
the problem and the threadbare analysis of its implications and bearings
as recorded by us in this book ought to convince a modern student of
logic that the logical speculations of Indian thinkers were as astute as
they were deep.
I. INDEX OF SUBJECTS
The numbers indicate pages
Absence of decisive proof i.e. of
crucial evidence, 151, 155, explained and illustrated, 166-69,
will be fallacious, 204.
Absolute Brahman, l94n, 221.
absolute negation, 194n.
adjective, two different types, 238n,
its elaborate consideration, 238n.
argument, its four types as cases of
inference, 100.
ascription, 61, divided into (I), ascription of identity and (2), ascription
of relation, 61, a case of error, 61,
and a case of assumption, 6i, 108,
109, conscious ascription recognised in the Upanisads and the later
Paurnic literature, 62.
assumption, elucidated 51-55, determination of its epistemolgica!
nature, 56, is the natural form in
which reasoning manifests itself,
56t cannot be regarded as a species
of cognition, 57-58, its difference
from knowledge or judgment elucidated, 58-59, Me Taggarfs view
referred to, 58, 59, cannot be a
species of mental perception, 60,
cannot be subsumed under erroneous knowledge, 60, is found to
share two fundamental characteristics of true knowledge, 60,
cannot claim to have the logical
character of truth or falsity, 61,
is a conscious ascription, 61, psychical process involved in it is
called imaging in Me Taggart's
terminology, 62, its definition
propounded by the neo-logicians,
62,distinguished from error,61,63.
Buddhist doctrine of flux, its refutation, 164.
casuistry, 104.
39
categorical inference, 96,99, for proof,

100, for disproof, 100, 117.
categorical judgment,58, its difference
from reasoning or hypothetical
judgment, 58-59.
cause, defined by the Naiyayikas,236n,
elucidated, 280.
causality, 146, its conditions of
knowledge as explained by the
Buddhists, 146, its elaborate criticism by Sriharsa, 280-94.
censure, 207.
cognition, its nature explained, 158-59.
cognitive relation, 115.
Complexity, 151, 154-55, 173, explained and illustrated, 175-76, its
three varieties, 176, will not be a
defect, 204.
condition, 32.
consequent, 34, 44.
-- --
Contradiction, 35-40, 51, enumerated
as the fifth variety of reasoning by
Sriharsa, 152-53, 173, elucidated,
178, its classification under three
heads, 178-79, not recognised as a
variety of reductio ad absurdum
by Sankara Misra, Udayana, and
Gangesa,207, regarded as the third
variety of censure by Sakara
Miara, 207, its elaborate criticism
by Sriharsa, 252-62, 294-302.
contradicted probans, 24-25, 85.
contradicted reason, 55, 103, 206, 213.
Conventional induction, 151, 155.
countermanded probans, 25, 187.
countermanding reason, '28, 212.
defects of argument, 205, its fourfold
classification, 205.
determinant concomitant, 49, 80, 109,
112.
determinant of the concomitance, 80.
determinate concomitant, 80,109,112.
306
INDUCTIVE REASONING
determinate in concomitance, 80.

Double noose, 154, elucidated, 182.
doubb, 3-4, its elimination secured by
tarka (reasoning), 4, 6,9-11, 20,
its nature as explained by Bhatta
Vdndra 10, by Udayana, 14, by
Srlvallabha, 19-23, 27, 35-36, 11112, its two kinds admitted by
Vyasatlrtha, 123-24, about the
probandum, 20, about the matter
to be known, 23.
earth, defined, 161.
Empirical induction, 151,155, explained and illustrated, 170, 173, 204.
Equalisation, 154, explained, 182.
error, 61, its difference from assumption, 61, 63.
Exception, included by Mdhavcrya
in the list of reasoning, 173, a
source of good deal of difficulty to
us, 173-75.
fallacies of reason, three different
varieties, 212-13.
fallacies of reasoning, set forth under
five heads, 185.
ground, 34, 44.
hypothetical assumption, 55, differs
in a material respect from inference, 55
hypothetical inference, not accepted
as a valid instrument by Indian
logicians, 8, 83, 90, 96-98.
hypothetical judgment, 55, its difference from judgment in general,
57-58, one and the same with reasoning and assumption, 57-58, its
nature elaborated, 63-64.
hypothetical proposition, 8, 96, 103,
111, 117, 121.
imaging, 62.
Impertinence alias Impudence, 155,
elucidated, 180-81.
imposition, 154,
impossibility, a sub-variety of Pure
reductio ad absurdum, 154.
Impudence, 173.
induction,32, is made possible according to Gangea by the knowledge
of co-association of the probans
and the probandum coupled with

the absence of the knowledge of
the failure of their concomitance,
32, 42.
Infinite series, 160, its two varieties
elucidated, 160.
inference, several ways of its classification by Jayatlrtha, 79, for establishing a conclusion, 79, for
refutation, 79, which assumes two
forms 95, its three types admitted
by the Naiyyikas, 99, it3 conditions, 99, employed for the
demonstration of an absurdity,
143, its validity is only vicarious
and. derivative according to the
Buddhist, 145.
Initial presentation, 151, 178.
introspection, 56-57.
Joint Method, 20, 22, 27, 37, 139, 140,
162.
lack of logical sanction, 185, explained
and illustrated, 187-88, is not
different from the fallacy of reason called 'non-existent in respect
of necessary concomitance', 211.
major premise, 81.
minor premise, 81.
mutual contradiction, 185, explained
and illustrated, 187, not materially different from 'the countermanded reason', 212, 227.
Mutual dependence, 151, explained
and illustrated, 157, 173, its elaborate criticism by Srlharsa, 24143.
necessary concomitance, 21, 100,
operative in (1) reasoning, (2)
categorical inference for proof,
(3) categorical inference for disproof, (4) implication, 100.
necessary concomitant, 80.
negation, 136, its four varieties, 194n.
non-concomitant reason, 206, 211.
non-existent probans, its three varieties, 210-11.
numerical difference, 194n.
opposition, a sub-variety of Pure
reductio ad absurdum, 154, elabo-
INDEX
rate refutation of the concept of
opposition by Srlharsa, 258-62.
petitio principiit 144.
plurality of causes, 258, 286, its elaborate consideration by Sriharsa,
286-94.
post-negation, 194n.
pre-negation, 176, 194n.
probans non-existent in the subject,
55.
progressus ad infinitum, explained
and illustrated, 160-62.
Pure reductio ad absurdum, 154, has
been divided into two sub-varieties, called opposition, and impossibility, JJ4.
Reasoning, its nature and utility,1-15,
is of two types,7, cannot be regarded as the necessary function of
inference according to Srivaliabha,
18, is classed under two heads, 19,
is regarded as a subsequent suborgan of the instrument of inference by Bhagiratha, 26, is always
an antecedent sub-organ and has
two different functions according
to Rueidatta, 30, is not an instrument of knowledge according to
the Naiyyikas, 5, 54-70, is accorded the status of a cognitive instrument by the Mdhvas, 77-125, the
Rmnujiyas,106, and the Jainas,
126-150, is not knowledge,
but an assumption and an
imaginary imposition according to
Udayana, 54, is a complex judgment or rather a compound of
assumptions, 63, is nothing but
inference according to Sridhara,
74, must be subsumed either
under perception or inference
according to Vyomasivcrya, 76,
is a variety of inference for refutation of a position according to
Jayatirtha, 80, is a case of hypothetical inference and acts as a
reductio ad absurdum according
to Jayatirtha, 83, is a species of
inference according to Venkata-
307
ntha, 106, is classified under two

heads by Vysatirtha, 124, is to
be classed with inference, and as
such must be admitted to be a
valid logical organ according to
Vysatirtha, 120, its classification
under five heads by Udayana,
which has been endorsed by Varadarja and Visvantha, 151,
and also by Sakara Misra,
154, its varieties as propounded by older logicians,
and repudiated by Visvantha,
151, its classification by Sriharsa,
152, which has been accepted by
Vamidhara,153, its classification
by Venkatantha,I54, and by Srinivsa, 154, and in the Prajparitrna, 154, its five other varieties
as shown by Sriharsa, 155, definitions of its different varieties, 156183, its conditions set forth under
five heads, 184, its fallacies set
forth under five heads, 185, qua
instrument of refutation examined, 205, its different species
criticised by rlharsa, 233-276,
qua organ of induction criticised
by Sriharsa, 277-302.
reciprocal negation, 194n.
Recrimination, enumerated as the
sixth variety of reasoning by
Sriharsa, 152.
Reductio ad absurdum, 6,7, 12, 22, 3334, 43, 50, 56, 67, 71, 73-74,
80, 83, 85-89, 97-98, 103,
110, 114, 117, 119. 125, 151-52,
(its enumeration as an additional
variety of reasoning by Udayancrya is not logically justifiable,
152-53), 154-55, 161-62, 173-74,
181-82.. 184-85, 190, 202, 204-10,
213, 216-25, 235-43, 263, 285,
299-301.
regressus ad infinitum, 35, explained
and illustrated, 160-64, 247-48,
267, 279.
relation, its two kinds admitted by
the Naiyyika, 115.
INDUCTIVE REASONING
3 o8
Smkhya theory of pre-existence of

effect, 282-83.
sattarkl (sixfold reasoning), 152.
self defined, 158.
Self-dependence, 151, explained and
illustrated, 156, 173, its elaborate
criticism by Srlharsa, 233-41.
Simplicity, 151, 154-55,173, explained
and illustrated, 175-76, its three
varieties, 176.
Simulation of reasoning, 54, 191, of
Self-dependence explained and
illustrated, 191, of Mutual dependence explained and illustrated,
200-1, of Vicious circle explained
and illustrated, 201, of The vicious
infinite explained and illustrated,
201-2, of Beductio ad absurdum
explained and illustrated, 202-3,
of Contradiction, 203, of Recrimi. nation or Tu quoque, 203.
sophism, 104, 209. :
synthetic judgment, 29-30, 4ln.
tarlca, in the Jaina epistemology an
independent source of knowledge,
1, serves to eliminate doubt and
thereby helps the determination
II.
of truth, 4, defined in the Ns, 4,

used as the symbol of the number
six, 152.
Tu quoque, its enumeration as the
sixth variety of reasoning by
Sriharsa, 152-54, explained and
illustrated, 163-C5, 173, 203,
230-31, its elaborate criticism by
Sriharsa, 263-76, 279.
universal, 167-68.
universal concomitance, 130,132.
universal proposition, 1, enquiry into
the possibility of its knowledge,
1-3.
universal subsistent, 191-93, defined,
194, its elaborate consideration,
194-200.
unreality of the subject, enunciated
by Srlharsa as a distinct variety
of the fallacy of reasoning, 190.
Vicious circle, 151, explained and
illustrated,157-59,173, its elaborate criticism by Sriharsa, 243-51.
Vicious infinite, 151, explained and
illustrated, 159-63, 173, its elaborate criticism by Sriharsa, 26263,
INDEX OF WORKS
Advaitasiddhi, $.
[tmatattvaviveka, 151-52, 184.
Bhsya (Vtsyyana's), 152, 208, 215.
Brahmastra, 92.
Dinakari, 200.
Khandanakhandakhdya, 38, 53, 152,
155, 165, 215.
Kiranvali, 60, 76.
Laghucandrik, 200.
Makaranddy 26.
Mahbhsya, 208.
Nyyakandali, 72.
Nyyakusumnjali, 26, 38.
Nyyalilvati, 15n, 16, 18, 51.
Nyayapariista, 20S.
Nyyaparisuddhi, 154.
Nyyasudh, 92-93, 101.
Nyyasutra, 4, 64-65, 151-52, 206,
207n, 208-9, 215.
Nyyastravrtti, 178.
Nyyatattvacintmani, 31, 51.
Nyyavrttika, 164.
Nydymrta, 78, 200.
ParUuddhipraksa, 42n.
Prajparitrna, 154.
Pramnapaddhati, 79, 93, 101.
Prameyatattvabodhat 30.
Sarvadarsanasamgraha, 173.
Smkhyatattvakaumudi, 53, 153.
Taragin, 200.
Tarkatdndava,107, 144.
Tattvavibhkara, 3, 153.
Trkikaraks, 15n, 151, 153, 184.
Ttpryailka, 152.
Vaisesikastra, 56.
Vdivinoda, 50, 154-55, 207.
Vrttika, 152.
Vyaktiviveka, 181.
Vyomavatl, 72, 75.
III.
INDEX OF AUTHORS
Aksapda, 64.
haglratha Thakkura, 23-24, 26, 2930, 123,153.
Bhatta Vdlndra, 10.
Bradley, 161.
Gagea, 20, 31-32, 35, 37-38, 40-48,
50-53, 79, 120,122, 207, 246, 301-2.
Hume, 282.
Janrdana Bhatta, 80, 89.
Jayanta Bhatta, 18.
Jayatlrtha, 78-80, 82, 87-90. 92-93, 99100. 102, 106-8, 110-11, 114. 119.
124.
Kiidsa, 28.
Kant, 235, 282.
Kumrila, 165-66, 269, 273, 276
Madhusdana Sarasvati, 78, 193, 196,
200.
Mdhavcrya, 173,
Mallintha, 15n, 153.
Mathurntha Tarkavgisa, 41n.
McTaggart, 59, 62.
Mill, 139,
MM Vsudeva Shastri Abhyakara,
173-74.
Ngrjuna, 161, 224, 282-84.
Paulsen, 181
Praaastapda, 56, 60, 76.
Prabhkara, 50,144, 248, 265-66.
Raghuntha Siromani, 53, 153.
IV.
INDEX OF SANSKRIT WORDS
abdhitatva (immunity from contradiction), 99.

agrhitagrhi (new and original), 146.
ananyathsiddhatva
(unconditionality), 280.
anaucitya alias vaiydtya (Impertinence alias Impudence), 155,180.
anaupdhikatva
(unconditionally),
280.
anivasth (Vicious infinite). 151,159,
173.
Rmarudra, 200.
Rucidatta, 26-30, 123.
Russell, 179,
Sankara Miara. 50.154-55, 207.
Sankarcrya, 47,179.
tfrivallabha, 15n, 16-24, 26, 50, 123.
Srldhara, 72-73, 75-76,150.
Srlharsa, 38-41, 44, 47-48, 51-52,152-53
155, 160, 165-66, 174, 181, 187, 190,
203, 207, 215-16, 230-36, 236o, 237,
239, 241-42. 245-46, 248, 252, 262-65,
277-79, 295-96, 301.
Srlnivsa, 154, 184.
Udayana, 7, 9-15n, 16-19, 23, 26-29,
38-40, 47, 51-54, 60, 62, 65, 70-1.
76, 88, 92, 105,119, 151-56. 167-68,
184, 207-9, 300-1.
Uddyotakara, 6, 54, 56, 65, 70-1,16466, 276.
Vcaspati Misra, 6-7, 9-10, 12-5n,
18-19, 28-29, 54, 71. 88, 246.
Vamsidhara Misra, 53, 153.
Varadarja, 15n, 151, 184.
Vardhamna, 26, 30, 42n, 71, 108.
Vtsyyana, 4, 6, 65, 206, 208
Venkatantha, 106,154,182.
Visrantha, 151, 178.
Vyomaaivcrya, 75, 150.
Vysatirtha, 78, 107-11, 114-16, 11921, 123-25,144.
anavasthbhsa
(simulation
Vicious infinite), 201.
of the
anekadravyasamavetattva
(the attribute of being inherent in more
than one substance), 176.
anistaprasanga
(reductio ad absurdum% literally emergence of an
undesirable issue), 151.
anuhlatva (to be helpful in the proof
of the opponent's thesis), 185,189.
io
INDUCTIVE REASONING
anumitikaranani$pdaka (instrumental to the formulation of the

major premise), 124.
anvayavyatireki (affirm ative-cum
negative), 99.
anyathsiddha (otherwise realised),
236x1.
anyonybhva (reciprocal negation or
numerical difference), I94n.
anyonyraya (Mutual dependence),
151,157,
anyonyeraybhsa (simulation of
Mutual dependence), 200.
apasiddhdnta (acceptance of a position opposed to his own accepted
conclusion), 8Q,
apavdda (exception), 173.
apratiyogi (not counter-term), 194.
apTtnnikasvikra (admission of an
absurd position), 206.
arthntara (irrelevant digression),
180.
asatpratipahsitatva (immunity from
opposition by a countermanding
probans), 99.
asiddhi (unreality of probans), 103,
206.
asambhava (impossibility), 154.
atyantbhva (absolute negation),
194a.
avrttipaddrtha (non-residential reals),
195.
avyavahitatva (immediacy), 280.
hryropa (imaginary imposition),
54.
ahryajftna (assumption), 62.
kea (ether), 168, 195, 222, 238n.
pdaka (ground), 34, 44, 212.
pdya (consequent), 34, 44.
aropa (assumption), 54, 56,
(ascription), 61.
sraysiddha (non-existent in respect
of the subject), 211.
eraysiddha hetu (subjectless reason),
211.
iraysiddhi (unreality of the subject),
19091, 227.
atmraya (Self-dependence), 151,156,
173.
tmsraybhsa (simulation of Selfdependence), 191.

bdha (contradicted probans), 24-25,
(contradiction), 103.
bdhita (contradicted reason), 212.
bdhitrthaprasangbhsa (simulation of reductio ad absurdum),
202.
bhsya, 92.
cakraka (Vicious circle), 151, 157.
cakrakbhsa (simulation of Vicious
circle), 201.
cakrakdraya (Vicious circle), 173.
chala (casuistry), 104.
dhvanitva (sound-universal), 271.
dravyatva (substance-universal), 166.
dsannumna (inference for refutation of the opponent's position),
79.
dsannumnavieesa (a variety of inference for refutation of a position), 80.
dustipramitisdhana (for proof of a
defect in the argument), 79.
gaurava (Complexity), 154,175.
gobalivardanyya (the maxim of
cattle and bullock), 153.
grdhyasamsaya (doubt regarding the
object of knowledge), 20.
hetvbhsa (fallacious reason), 205,
istpdana (establishment of an issue
endorsed by the opponent), 185,
188,211.
itaretardsraya (Mutual dependence),
173.
jdti (sophism), 104.
jdtyuttara (sophism), 209,
jeyatvatva (the essence of all knowabilities), 199,
jfieyatvatvena jfieyatvam (knowabiiity
qua knowabiiity), 199.
kdrikd 15n.
kalpandgaurava (Complexity), 155,
173.
kalpandldghava (Simplicity), 155, 173.
kevalavyatireki (exclusively negative), 99.
kevalnistaprasaga (pure reductio
ad absurdum), 154.
INDEX
kevalnvayi (exclusively affirmative)
99.
(universal subsistent), 191, 194.
mahattva (extensity), 176.
mithovirodha (mutual contradiction),
185, 187, 227.
mlasaithilya (lack of logical sanction), 185,187, 211.
nigrahasthdna (censure), 207.
niyama (invariability), 280.
paksabheda (difference of the subject),
48.
paksasattva (presence in the subject),
99.
paksatd (the essential character of the
subject), 211.
pardmara (synthetic judgment), 29,
24, 30, 41n.
pradhvamsdbhdva
(post-negation),
194n.
prakrti (primal matter), 187.
pramd (veridical cognition), 227.
pramdnasamplava
(convergence of
several cognitive organs on a selfsame object of knowledge), 90.
prasagdnumdtia (inference employed
for the demonstration of an absurdity), 143.
prathamopasthitatva
(initial presentation), 178.
.
pratibandl (recrimination or tu quo*
que), 152-54,173. "
pratibandybhsa (simulation of recrimination or tu quoque), 203.
pratijilvirodha (contradiction of the
thesis), 207, 233.
pratihulatva (the fallacious reasoning
called suicidal which arises when
the consequent entailed is hostile
to the position of the proponent
himself), 219.
prativddin (opponent), 82.
prgabhva (pre-negation), 176, 194n.
prdmdnikaparitydga (rejection of an
attested truth), 206.
prthivitva (earth-universal), 166.
purusa (spirit), 187.
rpaka (metaphor), 61.
samavacana (Equalisation), 154, 182.
311
sambandhakrtam (in respect of relation), 176.

sambhvan (hypothetical assertion),
222.
samskdra (trace or impression), 200.
sakdnivrtti (cessation of the doubt
about the opposite), 16.
sarirakrtam (in respect of form), 176.
sapaksasattva (presence in the homologue), 99.
satpratipaksa
(countermanded probans), 25,187.
satpratipaksita hetu (countermanded
reason), 212.
savyabhicdra (non-concomitant reason), 212.
sattarki (sixfold reasoning), 152.
sdhya (predicate), 49.
sdhaknumna (categorical inference, which seeks to establish an
independent conclusion), 86.
sddhandnumdn (inference for establishing a conclusion), 79.
siddhasdhana (the proof of an admitted truth), 211.
svajnnavyghta
(contradiction in
thought), 178.
svakriydvydghdta
(contradiction in
action), 178.
svarpsiddha (non-existent per se),
211.
svarpsiddha^hetu
(probans nonexistent in the subject), 55.
svavacanavyghta (contradiction in
language), 178.
taditarnistaprasanga
(emergence of
an absurdity other than that involved in the aforesaid cases),
154.
tannynit
(the weakening of the
opposite alternative), 16.
tarka (reasoning), 4, 55, 79, 152,
(symbol of the number six), 152.
tarkaydmi (1 assume), 57*
tarkdbhsa (simulation of reasoning),
54.
ubhayatahspd (Double noose), 154,
182.
upalak*ana 238, 238n.
yrr
INDUCTIVE REASONING
upasthitikrtam (in respect of presentation), 176.

updhi, (associate adjunct), 238n.
utsarga (Conventional or Empirical
induction), 155, 170, 173.
uttardiiga (subsequent sub-organ), 26.
vahnivypya (necessary concomitant
of fire), 34.
vaiydtya (Impudence), 173, 180.
vddin (proponent), 82.
vinigamandviraha (absence of decisive
proof i.e. of crucial evidence),
155, 166.
vipaksdsattva (absence in the heterologue), 99.
vipaksajijfldsd (diversion of enquiry
after the opposite), 19.
vipaksajijftsdnivrtti
(cessation of
desire for belief in the opposite),
16.
viparyayaakd (doubt of the opposite), 19.
viparyaydparyavasdna (failure to culminate in the proof of the opposite), 185, 190.
virodha (opposition), 154.
viruddha (contradictory reason), 212.
visayavirodha (opposition of contents)
10,
visayapariodhaka
tarka (reasoning
which removes an obstacle to the
, clear understanding of the sub*
ject-matter), 19
visaya-visayibhvd) 274.
visayatdsambandha
(cognitive relation), 115.
visesa (particularity), 268.
viesana, 238n.
vrttimat (resident in a locus), 195.

vrttiniydmakasambandha
(the relation by which the actual incidence of one term in the other is
determined), 115.
vrttyaniydmakasambandha
(the relation in which one term is related
to another term, though there is
no factual incidence of one in the
other), 115.
vrttyupahita caitanya (Brahman qua
object is qualified by the act of
realization), 194n.
vyghdta
(contradiction),
152-53.
173, 178, 207.
vydghdtdbhdsa (simulation of contradiction), 203.
vydpaka (determinant concomitant),
49, 80, ll)9, 112.
vydpdra (operation), 67n.
vydpya (determinate concomitant),
80,109, 112.
vydpyadharma {determinate concomitant), 109.
vydpydhdrydropa (assumption of the
ground), 62.
vydpti (necessary concomitance), 21,
vydptigraha (induction), 32.
vydptigrdhaka
tarka
(reasoning
which#is conducive to the knowledge of the universal concomitance between the probana and
the probandum), 19.
vydpyatvdsiddha (non-existent iu respect of necessary concomitance),
211.
vypyatvsiddhi
(fallacy called the
non-concomitant reason), 206.

A Study of Tarka and Its Role in Indian Logic

Uploaded by

Copyright:

Available Formats

A Study of Tarka and Its Role in Indian Logic

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Study of Tarka and Its Role in Indian Logic

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What are some of the logical concepts discussed in the book?

What are some of the logical concepts discussed in the book?

INDUCTIVE REASONING

A Study of Tarka and Its Role

LL.B., D.Litt. (CAL.)

Professor, Mithila Institute of Post-Graduate Studies and Research

thesis for the D. Litt. degr ee

unanimously approved by the

Board of Adjudicators in 1949'

to place materials at the hands

of the schol ar s of the E uro p e a n

American universities, w h i ch would enable them

Indian philosop hers.

of logical speculations recorded by

I thought that my cherished desire was bound

to remain unrealized if the volume was not speed ily published.

I sent my work to the press even at the risk of bei ng involved in

seen the light of t he day.

glad to see t h at this humble work

Before I finish, I must exp re ss my sen se of

p rofound gratitude to Prof. Anukul Chandra Mookerjee of the Allaha4

My heartiest tha nks are due

to Prof. JatiIcoomar Mookerjee, who, young in age but

origi nal writing in ma ny

as he is, he has compelled me to rev ise my

matter of pr o found regret to me that Sri Ksitindranath

in the land of the l i vi ng to see the result of my researches.

I must express my sense of ir redeem ab le debts which lowe to Maha4

Banerjee M.A., LL. B., Advocate,

to Sri: Binay aknath

Calcutta High Court,

Jyotirmoya Ba nerj ee M.B., D.P.H., D.T.M., for the keen appreciation

and constant encouragement which I have received from them in the

course of my scholastic pursuita debt which is too deep and personal

Calcutta, October 1953.

SriVallabha's tentative suggestion regarding the result of the

application of reasoning is the removal of doubt about the

Unbroken tradition among ttie Naiyyikas that reasoning is

view incidentally referred to, 58-59. Criticism of the different

THE LOGICAL STATUS OF REASONING AS ELUCIDATED IN

Sridhara's elucidation of the position of the older Naiyyika

Jayatrtha's elaborate dissertation on reasoning and his survey of

Vysatirtha's endorsement of the conclusion of Jayatirtha, 107.

images of deities is shown to be due to the reluctance of facing

THE JAINA CONCEPTION OF REASONING

Elucidation, of Jaina logician's contention that reasoning is a

134 H's position that the necessary concomitance holds

DEFINITIONS OF THE DIFFERENT VARIETIES OF REASONING 156-83

infinitum and the progresses ad infiniturn, 160. Their salient

Enunciation of conditions and fallacies of reasoning by Udayana,

another variety of the fallacy of reasoning, viz. the unreality of

FALLACIOUS SIMULATIONS OF REASONING

(i)( Simulation of self-dependence explained and illustrated,

Assignation of reason in a syllogistic argument liable to be

varieties of defect of argument are shown to be nothing but

Sriharsa's charges against the definition of reasoning elucidated,

(4) Contradiction, defined, 251-52, and subjected to scrutiny,