ON LOGICAL 'RELATIVISM' (1937) t

J6zej: M. Bochetiski

Objections raised against modern formal logic in the tars of the persons
who took part in the discussion cover almost all which is usually considered to
be a critique of that science m and for this we are grateful to all the participants
for giving us the occasion for an universal presentation of our point of view2.
The objections can be divided into three groups:

1. Conventionalization of a science by modern formal logic

2. Mechanization of the process of thinking
3. Metaphysical and theological analogy

In this answer we will limit ourselves to the first group. People who raise objec-
tions belonging to this group share with us the same intellectual and absolutistic
point of view - - the ground then is the same for both of us.
If we have understood well, the following objections appear:

A) Modern formal logic make science conventional because: (a) it introduces

the relativity of axioms, (b) does not search for truth, but only for connections
between particular theses,, and (c) approves of many-valued logics.
B) The consequence of (A) is: (a) a relativistic definition of truth, (b) positivism
and scientific pragmatism, according to which all that is not an experimental
thesis is a hypothesis examinable only by experience or by some criterion of
usefulness.
C) These mistakes result from the fact that modern formal logic is under the in-
fluence of mathematics and physics.

A further objection has been expressed, more or less as follows: given (A) and
(B), the watchword of modern formal logic for the unity of sciences is only a
pretence, because modern formal logic itself recognizes deep differences
between the scientific conceptions of reality. I do not understand this last
objection and, for this reason, I will not answer to it. Perhaps it concerns the so-
called neopositivistic 'Einheitswissenschaft' w but if this is true, we have here

I "Tradycja mygli katolickiej a .~cislo.~", in IC Michalski, My~l katolicka wobec logiki wsp61czesnej
[Catholic mind in relation to modem logic], Poznatt. [This paper was written more than half a century ago.
Several statements it contains are no longer upheld by the author. J.M.B.].
2 [Bochefiski is answering some objections raised in the meeting held in 1936 at the Cracow Catholic
Scientific Institute. Transl. note: TN].

Axiomathes, tt 2, settembre 1993, pp. 193-209.

194 JOZEF BOCHEIqSKI

a question that has no connections with modem formal logic. I will answer
these objections in turn.

A. The conventionalism of modern formal logic

We have to begin with the clear difference between what results from the
principles of modem formal logic and what results from the beliefs held by logi-
cians. It is obvious that someone who practises modem formal logic may have,
beyond the logical, many different philosophical opinions not connected with
modem formal logic - - he can be e.g. a conventionalist or a pragmatist, etc.
What will concern us, obviously, are not the views held by any logicians, but
the consequences beared by the very same modem formal logic3. Speaking in
the language of schoolmen, we discuss here the views of logicians as logicians,
not as philosophers: formaliter ut logistici sunt. But it will soon appear, as we
will see, that this distinction is valid also in relation to the objections of group
(B).
(a) First objection: the point of view which admits various logical axiomati-
zations is conventionalistic; modem formal logic admits various logical axioma-
tizations; modem formal logic is conventionalistic. We understand by the word
'conventionalistic' (in accordance with the intention of the persons who
advanced that objection) a viewpoint according to which the truth or falsity of
statements depends on conventions, not on material criteria.
The sense of this objection seems to be as follows: while the old logic esta-
blished a system approved by everyone and based on obvious axioms, modem
formal logic contains many entirely different systems based on different, arbi-
trarily chosen, axioms. Strictly speaking, the first part of this sentence is not
true, because even in 'classical' logic, at least in ancient logic, there were
different systems based on different principles. A very interesting example of
this multiplicity of systems and principles is the ancient (and, as I think, the
medieval) logic of modality: Theophrastus built a system of modal logic in
which at least one third of the theses is different from Aristotle's analogous
theses'*; certainly there were many systems in stoic logic too, because different
stoic thinkers accepted different definitions of implication5. If 'classic' logic in
general contains just one system, this is due to the limitation to a small part of

3 On account of some deeper objections not mentioned here but possible, we should distinguish modem
formal logic as a method from the scientific directives of the modern formal system of some particular
logician. E.g. between the 'ontology' of Le~niewski and the system of Principia there are very important and,
for philosophy of logic, significant differences.
4 See C. Prantl, Geschichte der Losik im Abedlande, Leipzig 1927, bd. I, 370 if; A. Becker, Die
Aristotelische Theorie der MOgfichkeitsschl~sse, Begin 1933, 66 ft. The undersigned has written an essay on
that question and is going to publish it soon [Z historii logiki zdmi modalnych [On the history of modal logic],
Lvov 1938. TH].
s See J. t~ukasiewicz, "Zur Geschichte der Aussagenlogik", Erkenntnis, 5 (1935), 116 ft.
O N LOGICAL'RELATIVISM' 195

the ancient and medieval logical knowledge m it is a decadent logic.

However, the second part of that statement is true: in modem formal logic
there are indeed many systems containing different axioms. E.g., speaking only
of propositional logic, Frege's system contains 6 axioms 6, the system of Princi-
pia 5 7, Hilbert's system has 4 s, Lukasiewicz's system has 3 9, as several other si-
milar systems 10 and finally a number of systems contain only one axiom as e.g.
dr. Sobocifiski's systems n.
But these systems are different only for the selection of undefined words,
axioms and, frequently, directives [dyrektyw] m however it is possible to
deduce from any of them the whole two-valued propositional logic and, on the
basis of one of them, it is obviously possible to obtain the principles of the
others and vice versa 12. Our systems then choose different starting-points, but
from these points we obtain always true assertions contained in the whole field
of theses of two-valued propositional logic. On account of this, we cannot talk
of conventionalism: it is as if we have here an enormous (infinite) net of mutual
connected logical theses. It does not matter from which place we start as long as
we obtain the same true theses.
One can observe that if not truth, then at least evidence is excluded from
these systems: modem formal logic does not claim evidence as a necessary
feature of its axioms. Practically it is the contrary: the majority of systems
depends on wholly evident axioms. Even if some systems - - as e.g. the one-
axiom systems m depend on very complex theses and, as a result, they are not
evident to non professionals, this is not yet a sufficient reason to regard them as
conventionalistic: the logicians such as e.g. Sobocifiski try to find the shortest
and unique axiom for propositional logic that tends to satisfy certain theoretical
and aesthetical necessity; even if the systems built in this way cannot claim a
didactic value for beginners, by no means does this stand in the way of
approving them as objectively true and not conventional systems; for in some
other system we can obtain such complex axioms from very simple theses
which are obvious to everyone.
It is also possible to observe that not all systems cover the whole field of
logical theses. E.g. Heyting's system does not contain tertium not datur and an
implication-negative system does not contain the theses in which the Sheffer
stroke [dysjunkcja] occurs. Nevertheless, even this objection does not prove the

6 See above, 126.

7 A.N. Whitehead, B. Russell, Principia Mathematica, Cambridge 1935, vol. I, 96 ft.
$ D. Hilbert, W. Ackermann, Grund~ge der theoretischenLogik, Berlin 1928, 22.
9 j. t.ukasiewicz, Elementy logiki matematycznej [Elements of mathematical logic], Warszawa 1929, 45.
lo See above, 83.
it B. Soboci~ski, "Z bada~ had teori~ dedukcji" [Investigations on the theory of deduction], Odbitka z
Ksi~gi Pamiotkowej Kota Fiiozoficznego Stuchacz6w Uniwersytetu Warszawskiego, Warszawa 1932 (see
also PrzegiqdFilozoficzny, 35 (1932).
12 See the article in note 5, in which l~ukasiewicz deduces from his assumptions both Frege's axiomatic
(80) and the axiomatic of Principia(87). It is equally possible to deduce on the basis of this system Hilbert's
axiomatic and the axioms of dr. Soboci6ski m the contrary process is obviously also possible.
196 J6ZEFBOCHEI~SKI

conventionalism of modem formal logic: these systems contain only part of the
field of logical theses, but by no means does it follow that they deny the truth of
other theses. E.g. tertium non darer is not denied in a three-valued logic, and
disjunctive theses are not denied in an implication-negative logic. The logicians
who construct these systems are interested only in some parts of our field m
and they can, as we believe, do it as much as a scientist can work only on a part
of his science and not on the whole of it, or as Aristotle could construct a
system of assertoric syllogistic from which it is impossible to obtain the modal
modes.
I believe that the heart of the objection discussed above lies in the lack of
distinction between logical systems as a tool for our human knowledge and the
way in which they exist in God's mind. We can come up to them in various
ways, we can not recognize them at all m the method of coming up and
recognizing has to be conventional by nature. But the use of such conventional
methods of modem formal logic by no means entails that only the theses in
agreement with our conventions are true. There are logicians asserting this
thesis, but the method itself of modem formal logic is not in charge of this.
We answer this objection in a scholastic form: the view which admits
various logical axiomatics is conventionalistic - - I distinguish the following:
the statement that there are axiomatics in contradiction one to the other, but
equally true; the statement that, on methodological accounts, there are various
axiomatics which do not deny the unity of the field of logical theses. Modem
formal logic admits various logical axiomatic.s, but while I deny the first
meaning, I admit the second.

(b) Second objection: The view according to which logic does not concern
truth but only the connection between theses is conventionalistic; modem
formal logic is the view according to which logic doesn't concern truth but only
the connection between theses; modem formal logic is conventionalistic.
If we rightly understand, according to that objection logicians should be
indifferent to the truth of conclusions: their only interest is in the conclusions
deduced from the given assumptions. That problem is quite complex and we can
distinguish three questions:

(i) Is modem formal logic interested in the truth of the extralogical conclusions
obtained by means of its formulas?
(ii) Is modem formal logic interested in the truth of the conclusions deduced
from its own thesis?
(iii) Is modem formal logic interested in the truth of the sentences by means of
which the conclusions are deduced, i.e. in the truth of the laws?

It seems that the objection admits that we should give a positive answer to
question (iii), because otherwise we cannot explain to ourselves what the above
 ON LOGICAL 'RElaTIVISM' 197

mentioned 'relation' between theses means. We should be concerned then only

with (i) and (ii). In the case of (i) the problem is quite easy: modem formal
logic, as any logic, cannot be interested in the truth of the extralogical
conclusions obtained by means of its formulas, because such truths belong to
other sciences: the task of logic is only to provide formulas which let us pass,
correctly, from sentences admitted as true (by other sciences) to other sentences,
which should be admitted, in this case, as also true. Logic is interested only in
this passage. There are assumptions, i.e. extralogical conclusions, that are true
per definitionem but do not belong to logic. On account of this, modern formal
logic is neither worse nor better than 'classic' logic.
We have then (ii): Is modem formal logic interested in the truth of its own
theses? The answer has to be positive. There are, obviously, logicians who
speak of the possibility to construct non-true logics [logik nieprawdziwych]13,
but none of us has ever seen these 'logics'; if one could ever arise, we should
say, in agreement with the prevailing opinion in logical circles, that it is not at
all a logic 14. We can obviously treat modem formal logic as a game, without
treating of truth, but once again we will then be concerned with some specific,
not scientific, interpretation of modem formal logic, for which the method itself
of modem formal logic is not responsible 15.
We answer then in a scholastic form: we admit the major premise, we deny
the minor premise and the conclusion. Precisely speaking, we can distinguish
here the three senses of the statement: 'no search for truth', but after the above
explanation this is rather unnecessary.

(c) Third objection: The view which approves of many-valued logics is con-
ventionalistic; modem formal logic is the view which approves many-valued
logics; modem formal logic is conventionalistic.
In the presence of many misunderstandings concerning the meaning of the
term 'many-valued', we should shortly mention here what is understood by that
term in contemporary logic 16. In stoic logic and in the logic of Frege (the
versions of which are all 'classic' systems, among the others the system of

,3 See the excellent article EJ. Nelson, "Deductive systems and the absolumessof logic", Mind, 92
(1933), 30-42. The author discusses with the conventionalist C.J. Lewis proving that his 'various' logics are
different ways of writing the same true logic. CJ. Lewis is, moreover, an 'outsider' of classic logic also on
other points.
14 In classic treatises of modem formal logic it is difficult to find a clear declaration of this point - - facts
seem clearly obvious to the authors. It is enough to read e.g. the preface to Principia or the articles of b
ukasiewicz, "O znaczeniu i potrzebach logiki matematycznej" [On the meaning and needs of mathematical
logic], Naulm Polslm, 10 (1929), 604-20 and "Logistyka a fdozofia" lLogistic and philosophy], Przeglqd
Filozoflc2ny, 39 (1936), 115-31 in order to convince oneself of bow these two prominent modem logicians
see our problem. G.H. Luquet (Logique [ormelle, Paris 1925, V) says clearly: "La logiqne a pour r61e de
d~terminet les moyens d'atteiodre la v~rit6".
15 For this question, see A. Tarski, "O logice matematyc2nei i metodzie dedukcyjnej" [On mathematical
logic and deductive methodl, Warszawa (bez daty) [Warsaw, no date], 88 ff. The author - - not suspicious of
absolutism - - warns us of regarding modern formal logic as a game.
• 16 See 7.. Zawirski, 0ber das Veth~ltnis der mehrwertigen Logik zur Wahrscheinlichkeitsrechnung',
Studio Philosophica, 1 (1935), 40742.
198 J6ZEF 8OCHENSKI

Principia and of t~ukasiewicz) we learn that every propositional variable has

only one of two values: 1 or 0, eventually: T or F. When we construct a logic in
a purely formal fashion, we are concerned only with signs whose sense is
defined by the laws of substitutions and by the functional tables: The
interpretation usually holds that '1' signifies 'true' and '0' signifies 'false', and
our assumption can be expressed in the incorrect, but easy to understand, form:
'every sentence can be only true or false'. It should be emphasized that that
interpretation, which is very convenient, has no connection with a system
understood in a formal way.
A logic which accepts the above mentioned assumption is called two-valued
logic. Many years ago Lukasiewicz and, after him, E. Post struck upon the idea
to construct systems in which a propositional variable could have values other
than 1 and 0. hukasiewicz worked out, at first, a system of three values in
which, beyond 1 and 0 appears the third value ½, then he developed n-valued
and infinitely-valued logics. From a formal standpoint these systems are correct
(non-contradictory, complete) and can be axiomatized as has been proved for
the three-valued logic and some others. Until we consider many-valued logics
as a system of signs, in abstraction from their interpretation, we cannot take up
against them any objection m because they fulfil as the two-valued systems all
the conditions that today we assign to these kind of formal systems.
Since logic is not a game of signs but a tool of knowledge, there arises the
problem of the interpretation of those systems. We have some interpretations of
this kind 17. Even if, at the moment, none of them is free from conventionalism,
a non-conventionalist could even say that, as he does not find a correct logical
sense for many-valued systems, he regards them as purely formal systems with-
out any logical meaning, at least until a correct interpretation will be found. But
if one interpretation of such a system would be free from conventionalism, we
could approve many-valued logics as an useful tool of science.
We have some interpretations which, by no means, do bear conventionalism.

(1) At first, many-valued systems can be interpreted as logic of probability

[logiki prawdopodobiefistwa]. Reichenbach treats them in such a wayl8; accor-
ding to this interpretation the value 0 is ascribed to the sentence for which the
probability of its truth amounts to O, the value 1 to the sentence for which this
probability amounts to 1, and various degrees of probability are ascribed to the
fraction values between 0 and 1. Being understood in that way, many-valued lo-
gic does not deny any of the laws of two-valued logic, particularly the so-called
law of excluded middle: it does not assert that between true and false sentences
there are 'indeterminate' [nijakie] sentences, but it just presumes that between
the sentences with probabilities 0 and 1, there is a series of degrees of

17 [The Author is referring here to the ideas of C.I. Lewis. See note 13. TN].
18 H. Reichenbach. "Die Iogischen Grundlagen des Wahrscheinliehkeitsbegriffs", Erkemunis. 3 (1933).
401 ft.
ON LOGICAL'RELATIVISM' 199

probability. If someone observes that many-valued logics, interpreted in such a

way, bear the negation of the law of excluded middle, this proves that he is not
able to understand any other interpretation of logic beyond the one which he
knows from two-valued logic, i.e. in which '1' means 'true' and '0' means
'false'. In our interpretation, signs have got the same shape as in the mentioned
logic, but we ascribe to them entirely different meanings. As a consequence,
when a many-valued system, interpreted as logic of probability, rejects the law
of the excluded middle, it is quite clear that we are not referring here to the
classical thesis called in that way, but to a false statement according to which
there could not be any intermediate degree between the probability of 1 and the
probability of 0.
(2) The first interpretation is subjective: here it can be formulated in a
simpler form, in which ~ e.g. for a five-valued logic ~ we can accept the
following 19 meanings of numbers:

1 - obviously true;
¾ - probably true;
½ - unknown: or true or false;
¼ - probably false;
0 - obviously false.

This five-valued logic could not obviously include the law of the excluded
middle, but it also obviously could not deny that law: according to it we could
affirm that not every sentence is obviously true or obviously false, but that there
are other possibilities - - and I doubt that anyone can contradict us about it. It is
worth adding that that interpretation seems extremely useful because in
scientific practice (except in mathematics and metaphysics) we have nearly
always to use sentences on account of their certainty m and it is very desirable
to have a tool to operate with them. E.g. not so many people can offhand give
an answer to the question about the value of the conclusion of some syllogism
in Camestres whose major premise is probably true and the minor probably
false.
(3) Finally, there is also another interpretation, the objective one, of some
many-valued logics. We ground it on some text of John of Saint Thomas, one of
the most famous scholasticians, who ascribes to the Divine Doctor (giving an
inaccurate footnote on the source) the thesis: the false has degrees 2°. That point
of view authorises the use of an objectively interpreted many-valued logic in
which '1' amounts to the truth and all fractions together with '0' to various de-
grees of falsity. In such a conception the assertion 'every European is a Mus-
sulman' is falser than the assertion 'every African is a Mussulman' (on account
of quantity) - - and the assertion 'a tree is a stone' is less false than the assertion

19 See C.I. Lewis, C.H. Langford, SymbolicLogic, New York - London 1932, 229 ft.
20 See Joannis a S. Thoma, Ars Logica, ed. B. Reiser, Taurini 1930, 777a.
200 J6ZEF BOCH~SKI

'a soul is a stone' (on account of quality). And this interpretation does not bear
any conventionalism; rejecting the law of the excluded middle we have in our
mind only the assertion that between the true and the 'border' false do not exist
(infmite) mediate degrees; what matters here is that between undifferentiated
truths and undifferentiated falsities there is no middle point.
As it has been seen, we have at our disposal interpretations of many-valued
logics quite well founded in the tradition of 'classic' logic, and also very useful.
It is true that there is also an interpretation of three-valued logics which bears,
indeed, consequences which cannot be accepted from our philosophical stand-
point. This is the objective interpretation done by the constructor of that system
gukasiewicz. According to this excellent Polish logician, there are really
sentences which are neither true nor false but indeterminate [nijakie] - - what in
his thought is connected with the indeterministic theory of nature 21. But no one,
even Lukasiewicz himself, says that such interpretation is necessarily connected
with the system discovered by him, a system that can be interpreted as anyone
wants. People who burden many-valued logics with responsibility for the
philosophical theses of logicians make always the same error of confusing
science with its representatives.
At the end of the discussion of this objection, we want to pay attention to the
circumstance that many-valued logics can be I as it seems to us ~ very useful
for the axiomatisation of theology, and this is what we should do. In theology
we make continual use of sentences qualified otherwise than true or false: we
talk of sentences 'de fide', 'fidei proximae' etc. Here it can be said that these
qualifications can be regarded as a private problem of theologicians, and
theological sentences can be treated as usual true assertions. Unfortunately that
method is entirely disappointing where we have to establish scientifically, i.e.
on the ground of some logic, which qualification will be right for our
conclusions. Theologicians know the heroic fights that thomists had more than a
dozen years ago concerning these qualifications. Perhaps many-valued logics
could render to that question a large service 22. The problem should be regarded
more closely, but after the rejection of the fear of conventionalism it should be
not too difficult to do it.
After the explanation of the objection, we answer in a scholastic form. The
view which approves many-valued logics is conventionalistic; I admit many-
valued logics as systems that bear conventionalism; I deny them as systems that
make abstraction from conventionalism and for which there are non-conventio-
nalistic interpretations. Modern formal logic is the view that approves many-
valued systems D but: I deny those of the first kind; I admit those of the second
kind. The conclusion is based on a paralogism.

21 A similar logical apprehension of indeterminism can be found in Aristotle (De lnterpretatione, 9) and,
under Aristotle's influence, it has been discussed also by Middle-age philosophers (ed. note).
For instance, the Marin-Soli theory, which distinguishes two kinds of 'conjuctinn' [l~jcanik]between
subject and predicate, cannot be sensibly expressed without the help of many-valued logic.
ON LOGICAL"RELATIVISM' 201

B. Relativity of truth, positivism, pragmatism

After the answers to the objections of the first class it should be clear that
also objections of the class (B) are based on paralogisms; a logical error is even
more visible here than in the objections of the class (A). People who make these
objections against the logicians' do argument in the form of the following
syllogism.
The followers of modern formal logic should approve relativistic-
positivistic-pragmatistic theses; you are a follower of modern formal logic; you
should consequently approve relativistic-positivistic-pragmatistic theses.
In this syllogism the words 'modern formal logic' have received a different
meaning in the major premise and in the minor one. Nowadays it has become
consolidated to call The Vienna Circle the school of modern formal logic, and
from this it is easy to deduce that modern formal logic affirms the same that the
Viennese affirm. The substance is entirely different: The Vienna Circle is a
philosophical school which uses with great skill modern formal logic with its
own assumptions and on account of that can be called a 'school of modern
formal logic'. But The Vienna Circle itself neither constructed a modern formal
logic, nor (in our opinion) can be compared with the work of the Polish school;
therefore there is no reason to regard modern formal logic as an exclusive
property of that Circle23. Modern formal logic, like any logic, is a purely formal
tool by means of which anyone can deduce all that he wants from any
assumption he wants, with just one restriction: consistency. Then the Viennese,
assuming their positivistic axiomatic, can deduce their positivistic theses by
means of modern formal logic, but thomists e.g. could do the same and we hope
that they will do it soon. When it will happen, it will be impossible to assert that
modern formal logic bears thomistic theses - - because modern formal logic is
both for thomists and for the Viennese only a tool, not responsible for the
philosophical theses of the philosophers that use it.
All the objections of class (B) depend, we think, on such a confusion
between the Vienna Circle and modern formal logic. It is true that the
assumptions of that philosophical school affirmed a relativistic definition of
truth, pragmatism and positivism; and it is not true that these consequences
result from the assumptions of modern formal logic itself. It will be, perhaps,
clearer if here we describe shortly the method used by modern formal logic. As

23 hukasiewicz perfectly explained this point it in his article "Logistyka a filozofia" [Logistic and philo-
sophy] (see note 14). Also H. Scholz, one of the most famous German logicians, forms an entirely clear opi-
nion in that question, it is worth citing some sentences from his work: "In keinen Falle steht es so, das ein
itherzeugter Logistiker nicht zugieich Metaphysiker sein kann in dem streng determinierten Leibnizischen
Sinne eines denkenden Menschen, fOr welchen sogar die Gottesfrage als ein dutch keinen noch so
charaktervollen Positivismus totzumachendes philosophisches Problem mit dem ganzen Gewicht eines
solchen existiert. Man hfite sich also die yon uns mit Leibniz bahauptete zentralphilosophische
Leistungsf~ihigkeit der Logistik dadurch zu weleher die nene Logik allerdings in einiger ihrer stitrksten
Vertreter mit einem extremen Posifivismus gegenwartig exisfiert" (Geschichteder Logik, Berlin 1931, 65).
202 ~6ZEFBOCHENSKJ

is generally known, we talk here about two methods of modern formal logic, at
least in propositional logic: the matrix method and the axiomatic one. Since the
matrix method can be correctly founded only on some axiomatic system, which
has been highly clarified, we will be interested only in the axiomatic method,
which moreover is the more popular.
If we construct a system of modern formal logic by means of that method we
move as follows24:
1. We choose some signs (words, letters etc.), which are not defined in any
given system, but we assume that these signs have got some clear sense for
everyone. In the system of Lukasiewicz we have the following words of that
kind: (a) The letters 'p', 'q', 'r'... called 'propositional variables' i.e. letters
standing for any sentence; (b) The inscription 'Cpq', that should be read: 'if p,
then q'; the inscription 'Np', that is read: 'not true that p', or 'not-p'. As we can
see this concerns signs whose meaning is quite simple.
2. By means of these signs, some others can be defined. E.g. in the
mentioned system the sign 'Apq' is defined by means of the expression 'CNpq',
writing: "Apq = CNpq'. The sense of 'Apq' is univocally [jednoznacznie]
established by the definition - - and it is easy to assert that 'Apq' means
precisely the same as 'p, or q, or both'.
3. We choose some axioms, which are not proved; as we said, the point here
concerns entirely simple and obvious axioms and always true theses. In e. g.
Hilbert's axioms they are some expressions which can be translated into every-
day language as follows: (a) if p or p, then p; (b) if p, then p or q; (c) if p or q,
then q or p; (d) if (if p then q), then if r or p, then r or q.
4. We choose some regulas, called the rules. A rule of this kind is, e.g., the
'rule of detachment' [regu|a odrywania], which permits the following
operation: If we have in the system a thesis of the following shape, 'if p, then
q', and if we approve a thesis of the shape p, then it is allowable to approve a
thesis of the shape q. As we can see, the point here is a very simple principle,
which can be expressed also (though not precisely) by the words: if A implies
B, and A is approved, then also B should be approved. All the rules used in
practice have got the same features: simplicity and obviousness.
5. When the list of the 'elementary' [pierwotnych] elements, mentioned in
the points 1-4, has been established we can start with deduction: we apply the
rules to the axioms and gain from them new theses, and theses from the latter by
means of the same rules, etc.
This is the fundamental method of modern formal logic. It is amazing how it
could be possible to see in such a method a relativistic concept of truth, positi-
vism, pragmatism and other philosophical errors, with whom it has nothing in
common. It only shows us how we should proceed starting from some given
assumptions; if we consider some logical assumptions we are interested always
in true and obvious theses - - up to now we have not heard anything important

24 See R. Camap, Abriss der Logistik, Wicn 1929, 70 ft.

ON LOGICAL' RELATIVISM' 203

about any relativistic logic: as far as we know a system of that kind does not
exist. Here also we cannot talk about any form of positivism. In the present-day
phase of the progress of modern formal logic, when systems are exactly forma-
lised, we can abstract them from any philosophical assumptions. And e.g. we
have the thesis of modern formal logic which often appears as the axiom 'if p,
then (if p then c0, then q' and whatever we wish to put under 'p' and 'q' in that
thesis m they can be sentences empirically examined by the criterion of
usefulness, dogmas of faith, metaphysical assertions - - modern formal logic,
absolutely, is not and will not be interested in it. It teaches us, we repeat, how
we deduce from some sentences some other sentences, it does not teach us how
to examine them extra-logically. At last, to accuse modern formal logic of some
'specific pragmatism' seems to us a major misunderstanding: modern formal
logic stresses very much the point that theses should be rationally proved by
means of the most exact methods, and it is rather questionable that a real
pragmatist could approve modern formal logic as the most prolific scientific
method at all. At least in the International Philosophical Congress in Prague a
leader of pragmatists, F.C.S. Schiller, expressed himself against modern formal
logic and in favour of Aristotle's logic because, in his opinion, both the first and
the second form of logic are just not necessary tools, but Aristotle's logic can be
(though with difficulty) understood by students, while modern formal logic
cannot 25.
We do not deny, considering the progress of modern formal logic due to no-
minalists, that here and there in their systems arises some assumptions which
we cannot agree with. And e.g. in Russell's system necessarily arises the dogma
that the construction of all logic is based on some individual assertion - - a
dogma today rejected even by R. Carnap 26. But, once again, the point here is
not about the individual views of logicians, but about the method of modern
formal logic. While Russell constructed his logic on a nominalistic basis, G.H.
Luquet constructed a radically idealistic system, in which an individuum is
defined by means of some universal 27. That all by no means burdens in any way
the method itself; in our opinion that method should take into its hands rather
than reject people interested in the correct philosophical foundation, or rather in
the correct philosophical interpretation, of contemporary logic.
We answer that objection in a scholastic form: The followers of modern for-
mal logic should, consequently, approve relativistic-positivistic-pragmatistic

25 AS tO the pragmatistic views of formal logic see A. Reymond, Les principes de la Iogique et la critique
contemporaine, Paris 1932, 49 ft. We should beware that authors like L. Rougier (La structure des theories
d~ductives, Paris 1921) do not know sufficiently well the thomism against which they oppose modem formal
logic. They should be warned of the acceptance of their conclusion concerning the contradiction between
these two doctrines. In fact no contradiction on account of the, here mentioned, facts arises - - moreover w e
could easily show that a thomistic philosophy of logic can serve to a philosophical establishment of modern
formal logic easier than to a logic called 'classic' (the point concerns especially the analysis of the sentence of
the form: for every x: if x is A, then x is 13).
26 See R. Carnap,AbrissderLogistik, Wien 1929, 107, n. 47, 3 A.
27 G.H. Luquet, Logique formelle, Paris 1925.
204 J6zEFeOCHm~S~

theses. By contrast I admit it for the followers of 'modem formal logic' i.e. the
followers of the views of the Vienna Circle and of similar groups, but I deny it
for the followers of modem formal logic, i.e. of pure logic in the contemporary
fashion. We are followers of modem formal logic m I distinguish: I deny it in
the first meaning; I admit it in the second one. The paralogism is simply
evident.

C. Modern formal logic, mathematics and physics

One of the most popular objections against modem formal logic is the asser-
tion that modem formal logic introduces a 'mathematization of philosophy' and
that its method cannot be used beyond the range of physical-mathematical
sciences. This objection seems right for the two following circumstances: from
one side the books of modem formal logic are full of signs similar to the mathe-
matical ones and there is used such an exact deductive method known to laymen
only from mathematics; from the second one - - modern formal logic is often
directly called 'mathematical logic' and to some extent it was constructed for
the foundation of mathematics. Here we can also add a third objection based on
some assertions of Aristotle and S. Thomas: not in all sciences is an identical,
mathematical exactness possible m and the fact that modem formal logic wants
to be the logic of all sciences forces upon sciences an equal degree of exactness.
We will answer in succession these three objections.
First objection: a discipline that uses the methods of mathematics is not a
logic useful for other sciences; modem formal logic is a discipline that uses ma-
thematical methods; modem formal logic is not a logic useful for other sciences.
It could be undoubtedly true that in the first phase of its existence modem
formal logic was extremely similar to mathematics; and e.g. Boole, who con-
structed 'logical equations" [r6wnania logiczne], used mathematical signs,
talked about multiplication and addition etc28. But Boole lived in the mid 19th
century, and since then modem formal logic has entirely changed in this
respect. Today no one seriously asserts that it is possible to transfer the methods
of mathematics to modem formal logic - - except for notation and ways of
deduction. But these two features can be called 'mathematical' only with some
very important restrictions.
If we first regard the use of notation (which, moreover, in modem formal
logic is entirely different from that in mathematics) we should say here that its
inventor is neither Leibniz, nor Boole, nor Frege, nor any other logician, but
Aristotle, in whose name modem formal logic is very often deprecated. In the
first works of that great thinker, 'mathematical' symbols, i.e. variables, do not
appear at all. But in some moment, which is not clearly known to us, Aristotle

2S See J. hukasiewicz, Elementy logiki matematycznej [Elements of mathematical logic], War~awa 1929,
1-7.
O N LOGICAL'RELATIVISM' 205

discovered variables 29. Since then we can talk of the existence of formal logic,
which is not thinkable without variables. The invention of variables (mainly
individual variables 'A', 'B', 'C'..., sometimes appears also propositional
variables) 30 seems so important to Aristotle that in his later works he misused
them: e.g. in Analytica posteriora these symbols appear almost in every page
both where they are needed and where they are not needed. Scholasticians in
this respect have been faithful to Aristotle - - it is enough e.g. to take into our
hands the works of Saint Albertus or the commentaries of S. Thomas Aquinas
in order to convince us that they also use everywhere the letters 'A', 'B', 'C'
and similar ones.
In this aspect, then, modem formal logic introduces nothing essentially new.
All the differences between it and the old logic depend on this, that scholastic
and 'classic' logicians used artificial symbols in a limited range (individual
variables), whereas modem logicians extend their use to all the elements of the
thesis: the point here is a purely quantitative difference. The result of this is the
appearance of our treatises in modem formal logic, which contain very few
words and are similar to the mathematical works. We can then, if we want, call
this symbolism a 'mathematical' one. But we have to remember two things: 1)
the constructor of logic is Aristotle, 2) symbolism of this kind has been
transplanted from logic to mathematics and not from mathematics to logic. Here
we should talk of the use in mathematics of the logical method rather than of the
contrary.
Here we should also add that while the mathematical symbols mean mathe-
matical objects, the logical ones are signs of objects not connected in any
special way with mathematics. And e.g. when we write the modus ponendo
ponens in the symbolism of Lukasiewicz, CpCCpqq, the letter 'C' means
implication (if then), and the letters 'p' and 'q' any sentence. As has been seen
-

with mathematics, it has also no connections to theology.

The question of using exact deductive methods arise in a similar way.
Nobody can prove that an exact formalised deduction is just a privilege of
mathematics - - because the same deductions really exist also beyond that
science. In this point logic goes also historically, as it seems, before
mathematics because the first handbook of deductive method - - Aristotle's
Analytica posteriora ~ under whose influence all m o d e m formal logic still
lays, is not a mathematical, but a logical work. In addition, how is it possible to
read in the following sentence in Lukasiewicz's handbook, nowadays logic
outdistance mathematics: "The theses we will prove can be formalised on the
ground of any axiomatic system, but it could be a very difficult work. We will

29 On the evolution of Aristotle's logic, see F. Solmsen, Die Entwicklung tier Aristotelischen Logik und
Rhetorik, Bedin 1929. This work, excellent for many reasons, is a proof of the necessity of some education in
modern formal logic for historians of philosophy: the author does not possess it and in consequence he simply
does not see many significant things.
30 E.g. Analytica priora A 15, 34a, 27-33.
206 J6zcF ~,oc~Et~sva

prove our thesis as is usually done in mathematics... "31. Then gukasiewicz

himself retreats here from logical exactness and turns himself to the
mathematical method, which is far less exact. Moreover, we repeat that if we
can prove that logical theses are entirely exact, why cannot we use that method.'?
We answer now the objection in a scholastic form: the discipline which uses
mathematical methods is not an useful logic for other sciences m I distinguish
here: the methods which can be used only in mathematics, I admit them; the
methods which are partially used in mathematics, but which are logical
methods, I deny them. Modem formal logic uses mathematical methods. I
distinguish here: I deny the methods which can be used only in mathematics; I
admit the methods which are partially used in mathematics, but which are
logical ones. That objection is based on paralogism.
Second objection: mathematical logic is not a general logic useful for other
sciences; m o d e m formal logic is a mathematical logic; m o d e m formal logic is
not a general logic useful for other sciences.
Also this objection depends on a misunderstanding. In the first place m o d e m
formal logic is, as it was correctly said in Cracow by Wilkosz, 'mathematically
n6e' i.e. it has been developed by mathematicians, but today it is developed by
many specialists who are not mathematicians; moreover many famous
mathematicians regard it as something that has got no connections with
mathematics - - e.g. in Polish modem formal logic Wilkosz talks about it.
Indeed, m o d e m formal logic today is logic without any addition, it is connected
with mathematics only by a common origin: it contains thousands of theses
which can be used in mathematics and which are really very often used there,
but which can also be used elsewhere. The objection denies this, but it is very
obvious that all Aristotle's syllogistic belongs to modem formal logic and the
same can be said also about such formulas as 'if p, then if p then q, then q',
which are everywhere used in science. After all, contra factum non valet
argumentum: and the fact is that we successfuly apply modern formal logic to
other sciences; e.g. Salamucha applied it to theodicy 32.
A central point of that objection is, I think, more or less, a presentment of the
following kind: modem formal logic uses mathematical signs, then it can be
related only to quantity and not to quality, then m o d e m formal logic is, in its es-
sence, the logic of mathematics. But this is a very large misunderstanding,
because at the very beginning of every work of m o d e m formal logic it is said
that the signs used do not refer to numbers or any other quantitative objects, but
to sentences or words. This objection seems to be a very sad certificate of the
limited power of abstraction of many people: they are accustomed to signs that

3t j. i~ukasiewicz, Elementy logiki maternatycznej [Elements of mathematical logic], Wasszawa 1929, 99

ft.
32 j. Salamuche, "Dow6d 'ex motu' na istnienie Boga", Collectanea Theologica, 15 (1934), fasc.l-2. [See
its English translation: J. Salamucha, "The Proof 'ex motu' for the Existence of God. Logical Analysis of SL
Thomas' Arguments", transl. T. Oierymski, M. Heitzmann, New Scholasticism, 32 (1958), fasc. 3, 334-72.
TNI.
ON LOGICAL'RELATIVISM' 207

can refer only to quantities, and they cannot understand that similar (but not the
same) signs can be used also to refer to any other object.
We answer this objection in a scholastic form: mathematical logic is not a
general logic useful for other sciences n I distinguish here: I admit logic which
uses only quantities; I deny logic which gives the most general directions of
reasoning. Modern formal logic is mathematical logic n I distinguish here: I
deny mathematical logic in the first sense; I admit mathematical logic 'n~e',
which gives general directions of reasoning. Again, here we have to do with a
paralogism.
Third objection: the view which demands the use of the same mathematical
exactness in all fields of science is false and contradicts the tradition; modem
formal logic is the view which demands the use of the same mathematical
exactness in all the fields of science; modem formal logic is false and
contradicts the tradition.
Aristotle in his Metaphysics expressed the opinion that the same
mathematical exactness cannot be reached in all the fields ~ for example it
cannot be reached where we have to do with material objects 33. St. Thomas
Aquinas' comment approves the view of Aristotle 34 even if he asserts that there
are places, e.g. ethics, where we cannot apply the same rigour as that required
for metaphysics and mathematics. That view is then traditional; it seems also
true, e.g. it is difficult to imagine history of art with the same degree of
exactness as mathematics or logic. Modern formal logic, undoubtedly, requires
in all the sciences methods that are as much exact as possible. It could seem that
modern formal logic requires something impossible and stands in contradiction
with thomism.
But, once again, we have here a misunderstanding. Modern formal logic is a
theory of deduction and for deduction the exactness of which we are here
talking about is required; modern formal logic does not have do with other
scientific methods - - e.g. with those that establish experimental theses in
physics or in history of art. With regard to this, modern formal logic itself has
nothing to say. It is true that the followers of modern formal logic claim,
together with Aristotle, that the ideal of science is an axiomatic system, in
which from some clearly precise assumptions other assertions are deduced; but
neither Aristotle, nor any logician claims that the use of that ideal could be
contemporarily possible in all the sciences as they stand. Here only one logical
postulate remains: all the sciences, without any concern of their matter, when

33Met. II, 3, 995a 14-17. "t$n d'akribologhian t$n mathematik~n ouk en hdpasin apaitet~on, all'en tot s
m~ ~chousin h~len. Di6per ou physikDs ho tr6pos hdpasa gdr Sos he ph~sis ~chei h~len" (Mathematical accu-
racy is not to be demanded in everything, but only in things which do not contain matter. Hence this method
is not that of natural science, because presumably all nature is concerned with matter. - Translation of H.
Tredennick).
34 S. Thomae Aquinatis, In Metaphysicam Aristotelis commentaria, ed. M.R. Cathnla, Taurini 1926, lib.
II, lect. V, N. 336. It is worthwhile to be aware that SL Thomas does not speak o f exactness (exact/tudo), but
of certainty (certituto, certa ratio).
208 J6ZEFBOCHEI~SKI

they use deduction should use a correct, i.e. an exact, one.

A postulate formulated in such a way is, we think, right and agrees with tho-
mistic tradition. There is no reason here which could justify the use of inexact
deduction in ethics or history of culture rather than in mathematics; in this
moment we have to do with an entirely similar situation: some premises are
given and it is necessary to obtain a conclusion from them. If we have tools,
which lead us to conclusions with an as little as possible risk of error, excluding
all the hidden possibly false premises, why shouldn't we use these tools also in
ethics and history? Aristotle and St. Thomas thought in such a way; in spite of
the restriction we are talking about, St. Thomas uses in his ethics (If part of
Summa) the very same syllogisms used in other works of theology and
philosophy; Aristotle does the same in his Physics: when he discusses material
objects he does not hesitate in using syllogisms not worse than the ones used in
metaphysics and logic. These great thinkers distinguish perfectly well the
uncertainty connected with some non-deductive methods of justifying the
premisses from the unexactitudeness of deduction itself. Postulating the first as
a necessity imposed by the object, they did not conclude that in this field of
philosophy they can inexactly deduce.
It is worth adding that though modern formal logic itself has nothing to say
of non-deductive methods, a logical preparation, created by someone
accustomed with exact deduction, gives something more, namely the capacity to
work also in other fields of science, even if not with the exactness of modern
formal logic. To logicians it is forbidden to use incomprehensible words,
metaphors, vague style and similar things, which unfortunately reduce the level
of most of contemporary philosophical works. We do not think that such
postulate of correct expression could be in any way injurious for non-deductive
sciences: on the contrary, it can help them. This postulate is not against
tradition, because catholic tradition, so long ridiculed, is the tradition of
defining, distinguishing the meanings of words; in short it is the tradition of
purely scientific work. We do not understand why today, when European
thought returns to it, we Catholics should turn away from it and defend the
hopeless vagueness of the philosophy of XIX century.
It is understood that irrationalists do not like such traditions and its postulate
of correct scientific work: they are afraid of the clear meaning of the words used
and of the correct proofs. Here we can wonder if that irrationalistic philosophy
is a science at all u does that philosophy merit the name of philosophy? As for
us, we are inclined, in agreement with Catholic tradition, to answer this
question in the negative.
To this objection, we answer in a scholastic form: the view which demands
the use of the same mathematical exactness in all the sciences is false and con-
tradicts tradition u I distinguish here: I admit what demands the construction of
all sciences by means of the same, only deductive, methods; I deny what
demands the use of exactness for all the deductions, without prejudicing
ON LOGICAL* RELATIVISM' 209

anything about usefulness and exactness of non deductive methods. Modem

formal logic is the view demanding the use of the same mathematical exactness
in all the fields of science m I distinguish here: I deny what demands the
construction of all the sciences by means only of the same deductive methods; I
admit what demands the use of exactness for deduction, without prejudicing
anything about the usefulness and exactness of non-deductive methods. The
conclusion is based on a paralogism.
At the end we can say that the objections of the above discussed group are
all based on misunderstandings: whereas modem formal logic is the only
convenient and exact method of deduction, it is reproached with mistakes which
have nothing in common with it.

[Translated by Ryszard Puciato]