IBN MADA'S CRITICISM OF ARABIC

GRAMMARIANS

KOJIRO NAKAMURA
The University of Tokyo

The origin of the Arabic grammar is lost in obscurity. Traditionally

its foundation is traced back to Abu al-Aswad al-Du'ali (and further back to
the fourth Caliph, 'Ali).(1) We cannot, however, be sure of this tradition, since
none of Abu al-Aswad's works are extant (not to mention 'Ali).(2) It is certain,
however, that in the course of the second century A. H. there appeared two
groups of grammarians in Basra and Kufa, and that from the former group
came out two eminent scholars: al-Khalil b. Ahmad (d. ca. 175) who was the
first to compile an Arabic dictionary, entitled Kitab al-'Ain and to lay down the
principle of the Arabic prosody ('arud), and Sibawaihi (d. ca. 177), one of his
disciples, who composed a "Book," al-Kitab, which laid the foundation of the
later theoretical development of the Arabic grammar.
Although the grammar in this early period was already based on the theory
of regent ('amil) as an explanatory principle for the change of the case-ending
of word,(3) it was essentially no more than a list of the rules induced from the
materials, or "une sorte d'inventaire des faits grammaticaux dont us s'efforcent
d'expliquer chacun d'eux en lui-meme, pour un classement, une mise en
ordre."(4) In due course, however, the grammarians became more concerned
with systematizing the grammatical explanations, rationalizing the grammatical
facts, and elaborating the logical construction-in short, philosophieren on the
Arabic grammar in the process of commenting on Sibawaihi's al-Kitab.(5) G.
Weil summarizes the philosophical premise of this process of rationalization as
follows:

Ebenso wie Allah selbst, so ist auch der Qur'an als Gottes Wort die Vernunft schle-
chthin. Da aber die Sprache des Qur'an Reprasentant und Muster des Ausdrucksder
arabischen Sprache ist, so muss die gottliche Vernunft and Vollkommenheit auch im
Bau der arabischen Sprache allenthalben zum Ausdruck kommen. Identifizierten die
Griechen bewusst Sprachen und Denken, Sprachgesetze und Denkgesetze, so schufen
die Moslemeunbewusstdie Gleichung von arabischer Sprache und absoluter Vernunft.(6)

89
Every single word or phrase has its own rationales for its form and place so that
the whole is kept in harmony and balance.(7) There should not be any excep-
tions to the rules (qawa'id). Seeming exceptions can be explained by finding
out their rationales by means of qiyas (analogy) and taqdir (virtual assumption).(8)
Generally speaking, once the grammatical rules, principles and systems are
established, they are likely to be regarded as absolute and unalterable. And
this is particularly true with the Arabic grammar, since it emerged and develop-
ed, inextricably bound up with the interpretation of the direct Word of God
(the Qur'an) and the theological presuppositions.
There were some who raised an objection against such a development,
but their efforts bore little fruit. Ibn Mada' was one of those few opponents.(9)
It is our intention in this article to make an attempt to understand Ibn Mada's
criticism of this "orthodox" grammatical tradition in relation to the development
of the Islamic legal theory (usul al-fiqh).

Life of Ibn Mada'

The Almohad (al-Muwahhid) dynasty (515-668 A.H./1121-1270 A.D.)

had its inception, like its predecessor, the Almoravids, in a religious reform
movement in North Africa started by a Berber called Muhammad Ibn Tumart
(d. ca. 524).(10) He was born in southern Morocco. Still in his youth he made
a pilgrimage to Mecca, and stayed in the Muslim East for nearly ten years.
During this period he studied the Ash'arite theology, particularly that of al-
Ghazali (d. 505). He was also influenced at the same time by the writings
of the great Spanish Zahirite, Ibn Hazm (d. 456).(11) This combination pro-
duced a complicated character in his religious system (and also in that of the
Almohad dynasty)(12)-that is to say, he was legally a Zahirite and theologic-
ally an Ash'arite.
Concerning his legal position, I. Goldziher describes as follows:

Il ressort de sa dispute d'Aghmat avec les fouqaha, que l'element essentiel de sa

doctrine sur les fondements du droit peut se resumer en cet axiome: al-'aql leisalahou
fi al-char' madjal,c'est a dire "qu'on ne peut pas accorder au raisonnement la moindre
place dans les lois de la religion." Ce sont les sources objectives, materielles,du droit,
que l'on doit prendre pour base de la legislation,c'est a dire le Coran, la tradition tran-
smise de facon authentique, er le consensus de la oumma,fonde sur des traditions qui, a
travers toutes les generations, out ete soutenues par des autorites nombreusesformant

90 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

une chaine ininterrompue (tawatour). Ainsi se trouve absolument exclu l'element sub-
jectif, personnel, ce qu'il appelle dhann,hypothese, opinion, et qui, devons-nousajouter,
a, sous la forme du consensus de la oumma,trouve sa place necessaire parmi les sources
regulieres du droit, des le debut des speculations juridiques dans l'Islam.(13)

Just as he discredited ra'y (in the above quotation, zann or 'aql), or qiyas, as one
of the usul (legal "roots"), so he rejected the furu' (positive laws) of the establish-
ed legal schools, which were deduced from the other usul and formulated by ra'y,
and he attacked the taqlid (blind imitation) of those legal systems.
Dependent on the Ash'arite theology,(14)he stressed the tawhid (unity) of God,
but he went too far to the point that he denied the divine attributes.(15)-Thence
came the appellation ("al-Muwahhid") of the dynasty. On his return to
Maghrib with the reforming zeal, he set out to attack, from this standpoint, the
"anthropomorphism" which was dominant among the Malikite Almoravids.
He believed himself as sent by God to reform the corrupt religious situation
and took the title of the Mahdi ("Messiah"). He declared jihad against the
"unbelievers" of the Almoravids.
The actual foundation of this dynasty, however, came with his successor,
friend and general, 'Abd al-Mu'min b. 'Ali (d. 558), who carried out the master's
enterprise and put his doctrine into practice. He successfully conquered Mar-
rakesh and put an end to the Almoravid dynasty. He even sent the army to
Muslim Spain and placed it under his control. And the whole coast from Egypt
to the Atlantic Ocean and Spain was united under a single creed and sceptre for
the first time in the Islamic history. He took the title of Amir al-Mu'minin, and
Friday prayers were delivered in the name of the Mahdi or his Caliph (instead of
the 'Abbasid Caliph) in the whole empire.
The second Caliph, Yusuf b. 'Abd al-Mu'min (reg. 558-580) was himself
a well-educated eminent scholar, while he was in Spain as governor on be-
half of his father, he met many scholars in the Arabic language, grammar and
the Qur'an.(16) He also showed an extraordinary interest in learning and patro-
nizing philosophy (falsafah). He ordered to collect books and invited scholars.
Suffice it to mention the names of two philosophers: Ibn Tufail and Ibn Rushd
(Averroes). During his reign the capital was transferred to Cordova in Spain.
The third Caliph, Abu Yusuf Ya'qub (reg. 580-595) was also an eminent
scholar like his father, and he himself wrote a book on hadith. He used to convene
a debate by scholars and philosophers.(17) Thus Ibn al-Athir reports, "Er
bekannte sich offentlich...zur Zahirijja und wandte sich von der malikitischen

Vol. X 1974 91
Richtung ab; die Sache der Zahiriten nahm denn auch zu semen Zeiten einen
grossen Aufschwung."(18) He ordered the jurists (fuqaha') to reject the furu'
and not to follow (qalladu) any of the ancient Mujtahids of the orthodox schools,
but to make legal decisions (ahkam) by their own efforts (ijtihad) based on the
Book, the Sunnah of the Prophet, and Ijma'.(19)
This policy was pursued to such an extent that the Caliph finally ordered to
burn the fiqh-books of the orthodox schools. We see in this development a
revolt against the established authority of the four legal schools (madhahib).
And Ibn Mada' occupied the position of Chief Judge (qadi al-jama'ah) until he
died in 592 A.H. under the two Caliphs, Yusuf b. 'Abd al-Mu'min and his son,
Abu Yusuf Ya'qub.
Ibn Mada' (Ahmad b. 'Abd al-Rahman b. Muhammad b. Sa'd b. 'Asim
b. Mada' al-Lakhmi al-Qurtubi) was born in 513 (or 511) in Cordova. His
family was of noble origin, and "he grew up, devoted to searching for knowledge
and concerned solely with meeting the teachers and learning from them."(20)
He learned the Arabic grammar, lexicography, and adab-literature from
Ibn Mashkwal, Ibn Sahnun and others. His biographers particularly mention
that he studied Sibawaihi's al-Kitab under Ibn Rammak. He learned hadith
and fiqh under the Qadi. 'Iyad, a great hadith-expert and jurist at that time in
Maghrib. In addition he studied speculative theology (kalam), medicine,
arithmetics and geometry. It is also said that he was an eminent poet and a
writer (katib). He was, however, most distinguished in the field of the science
of the Arabic language ('ilm al-'arabiyah), in which he made most efforts and for-
med his own independent opinions.(21) For this very reason, he is said to have
been "isolated from the convention of scholars"(22) at that time.
He was sharp in mind, firm in belief, good in reputation and righteous in the
way of life. Yusuf b. 'Abd al-Mu'min appointed him Chief Judge and he
remained in this position in Marrakesh, Fez and others until he died in Seville
in 592 (or 593)(23) under the reign of his son, Abu Yusuf Ya'qub.
Al-Suyuti reports that Ibn Mada' wrote three books(24): Al-Mashriq fi al-
Nahw, Tanzih al-Qur'an 'amma la Yaliqu bi-al-Bayan, and Al-Radd 'ala al-Nuhat.
And concerning the second book, he also says, "In his book, Tanzih A'immah
al-Nahw 'amma Nusiba ilai-him min al-Khata' wa-al-Sahw, Ibn Khuruf opposed
him (Ibn Mada') for this book." From the title of Ibn Khuruf's book, we
may conclude that all these works by Ibn Mada' are more or less on the Arabic
grammar.

92 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

Since he did not compose any books on fiqh and the first two of the above-
mentioned works by him are not extant, there is no way of knowing his ideas on
the usul al-fiqh proper. However, judging from the high position he occupied
under his masters Yusuf and Ya'qub, and the contents of his only extant work,
Kitab al-Radd 'ala al-Nuhat, he was a Zahirite, and his criticism of the current
grammatical theories would be better understood from this point of view.(25)
This is what we shall undertake in this essay.

Legal Theory

Historical Background: In the four orthodox legal schools (madhahib al-fiqh),

the Qur'an, the Sunnah, Ijma' and Qiyas are the established basic sources of
the Islamic law (usul al-fiqh). Although no one doubted from the very outset
the authority of the Qur'an as the first source (asl), however, fervent disputa-
tions were waged over the other sources (usul) in the early history of the theoretical
development of fiqh: What does sunnah mean? Whose consensus is meant
by ijma'? Is qiyas to be approved? If so, to what extent? etc. etc.
Above all, the problem of qiyas, or ra'y (see infra, p. 94) was the object of
ardent controversies, so much so that we can even say that the differences among
the schools were mainly determined by the extent to which they approved the
application of ra'y in the legal procedure.(26) Between the two extremes-Abu
Hanifah (d. 150), who enthusiastically vindicated ra'y and Dawud b. 'Ali (d.
270), who denied it (in the sense currently used among the other Imams)-we
can arrange Malik b. Anas (d. 179), al-Shafi'i (d. 204), Ahmad b. Hanbal (d.
241) in this order.(27)
No doubt, the use of ra'y was motivated by the actual necessities. As Islam
expanded in the vast areas outside its cradle of the Arabian Peninsula, the Mus-
lims faced many problems urgently requiring instant solution, to which the
Qur'an and the sunnah(28)did not give any direct answers. Thus they were
obliged to make decisions and regulate their conducts in such cases according to
the results of their intellectual efforts based on the sacred sources.
The extent of the application of ra'y varied in accordance to how far it was
needed-this need in turn depended upon many factors such as time and
place-, and to how much the sunnah was known. Iraq, above all, went
furthest in making use of ra'y in the legal decisions. Some, like Hammad b.
Abi Sulaiman (d. 119 or 120), went so far that he did not give much importance

Vol. X 1974 93
 to the sunnah.(29)
At this stage of development there was no unified concept of ra'y in the
technical sense, nor unanimous acceptance of it yet. Its application was still
the object of severe criticism at that time. Abu Hanifah was the first to make
attempts to codify fiqh based on the principle of qiyds, following his predecessors
in Iraq. It is also true, however, that along with these attempts and efforts went
the systematic attack on this principle and its adaptation in furu', (30)
It was al-Shafi'i who adopted qiyas as one of the usul by giving it a theoreti-
cal refinement and systematization. At the same time he criticized the con-
cept of sunnah as the living traditions of "the ancient schools of law"(31) which
were traced back to the Companions, and established the principle of the hadith-
critique, and the sunnah as the Prophetic precedent transmitted uninterruptedly
in the form of hadith. Thus the theoretical foundations of fiqh were now laid by
al-Shafi'i. To use Goldziher's words:
... hat die Lehrmeinung al-Safi'i's zwei Seiten. Von der einen Seite macht er
den Ausgangspunkten des Abu Hanifa Concessionen; freilich geht er aber nicht so weit
wie dieser, und diese Beschrankung bildet die zweite Seite seines Systems: vor Allem,
Berucksichtigungder Tradition. Er giebt dem Abu Hanifa die Berechtigungdes Kiyas
als Rechtsquelle nur insofern zu, als dasselbe auf die geschriebenen und uberlieferten
Queue begrundet ist.(32)
From this eclectic character of al-Shafi'i, though its emphasis was more or less on
the Sunnah, there appeared among his followers the "people of ra'y" (ashab al-ra'y)
on the one hand, and the people who extremely stressed the traditional ele-
ment on the other. And the founder of the Zahirite school, Dawud b. 'Ali, came
from this latter group.
Qiyas: The word ra'y in its original sense, had a good connotation of "good,
right, discrete, rational opinion" and was used as an antonym of hawan (careless
decision, misleading passion).(33) At the same time it attained among the
conservative hadith-scholars and theologians a bad connotation which was nearly
equal to its antonym hawan, in spite of the fact that the "opinion" was used
with some reservation even among the Companions of the Prophet.
At this early stage of the development it meant vaguely an "individual
opinion or decision" and it was not yet methodologically defined and system-
atized. Later on, however, it was specified and narrowed down to the meaning
of qiyas, logical reasoning of analogy.
Sagte man fruher: Dort, wo kein geschriebenesoder uberliefertes positivesGesetz

94 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

vorfindlich ist, mag der vor emnim Gesetz nicht vorgesehenes Verhaltniss gestellte Ri-
chter seine eigene Einsicht anrufen, so sagte man von nun ab: Dieses Walten der eigenen
Ansicht hat sich im Rahmen der Analogie zu bewegen, welche der freien Willkur der
subjectiven Einsicht die Richtung vorzeichnet, in welcher sie zur Geltung kommen
darf.(34)

This Qiyas includes technically two procedures. One is comparison. It

means, in order to make a decision over a case about which the Qur'an and the
Sunnah are silent, to compare it with the similar cases which are already
mentioned in the sacred sources. Secondly, the basis on which to compare is
the motive ('illah) or the spirit of law.(35) Suppose that the Qur'an prohibits
man to do four things, A, B, C, and D. Then you examine the motive of the
text (nass) or the intention of the Lawgiver, which is common to all these four,
and you find out the motive a. Now if you find the same motive a in other
cases E and F, about which no reference is made either in the Qur'an or the
Sunnah, then you may conclude that the same prohibition is applied to these
cases in like manner.
In addition to the method of Qiyas, each school developed other procedures
to rectify or supplement the conclusion drawn by Qiyas. The Hanafites have
istihsan. This principle is well explained by the following words of Abu Yusuf,
an early representative of this school: "According to the qiyas this and that
would be prescribed but I have decided according to my opinion (istahsantu)."(36)
This extremely subjective principle was limited by the Malikites by introducing
the concept of "common weal (maslahah)" (istislah). Al-Shafi'i rejected istihsan
and adopted instead much more limited technical procedure, istishab,(37) to
cope with the actual difficulties. Dawud, founder of the Zahirite school, rejected
all these procedures as "subjective." The theoretical attitude of his and his
school is succinctly described by Ibn Khaldun as follows:

Sie (die Anhanger Dawud's) reducirten die Quellen der Erfassung des Gesetzes
ausschliesslichauf das Ausdruckliche (in Koran und Tradition Bestimmte) und auf
den Consensus, als in welchen alles Gesetz enthalten sein soll. Sie fuhrten auch das
offenbare (nicht erst durch Speculation zu erschliessende)Kijas und die Ursachen des
Gesetzes, selbst da, wo die Ursache, als solche, in der Schrift ausdrucklich dargelegt
wird, auf die Schrift zuruck (d. h. erlaubten keine uber den in der Schrift erwahnten
Fall hinausgehende Nutzanwendung der Analogie und der Gesetzesursachen);denn-
so sagen sie-die schriftlich erwahnte Gesetzesursacheist, so oft wir ihr begegnen,
nichts anderes als die Bestimmungdes (concreten) Gesetzes (nicht aber die Bestimmung

Vol. X 1974 95
 eines Principes).(38)
This is technically what the Zahirites mean when they say that the text of the
Qur'an and the Sunnah must be interpreted and understood in its "external" or
"literal" (zahir) sense. It is strictly rejected to interpret a text by relating it
to, or comparing it with, others.
Two premises are presupposed in this Zahirite attitude. One is that if
the jurists had to, and could, deduce a more general proposition from the indi-
vidual materials as mentioned above, the Lawgiver would have expressed so;(39)
namely, "tout ce qui est signifie est ecrit."(40) In other words, the interpreta-
tion must be done exactly as the written words go. Nothing more nor less.
Furthermore (and this is the other premise), it is impossible (and even presump-
tuous for man) to ask about the motive of the Divine Law, as well as to ask about
the motive of God's creation.(41) Logically it is not difficult to draw from the
foregoing discussion another characteristic of them, namely, the rejection of
taqlid of the furu' which were systematized in the four schools on the basis of
Qiyas
Let us see an example(42) for a better understanding of the above-mentioned
Zahirite attitude. The Qur'an prohibits to take interest (riba) in several places
(2:275-9; 3:130; 30:39),(43) but no mention is made about the condition. At
this junction a hadith is quoted: "To take interest is forbidden; and furthermore:
gold, silver, wheat, barley, dates, raisins."
Concerning this hadith, the schools of analogy first try to find out the com-
mon motive because of which each of these is forbidden, namely, the genus
to which each of these species belongs, and draw the following conclusions.
According to Rabi'ah, teacher of Malik b. Anas, the prohibition of interest
has something to do with all the things that cannot be given as zakat. There-
fore, the same prohibition is applied to domestic and riding animals. The
Hanafites say that the first two items are the examples of the whole genus which
can be counted by weight. The Shafi'ites see in these items the representatives
of all valuable things and foods. Consequently the prohibition of usury is ap-
plied not only to the goods mentioned in the hadith, but also to all the goods
which are counted under the same category. In this way these schools extend
the meaning of the text and apply it to that which is not verbally mentioned
in the text.
On the other hand, the Zahirites cannot admit of this kind of "arbitrary,
subjective" interpretation of the text, based on syllogism. To do so is to submit

96 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

the Will of the Lawgiver to man's personal will. The text should be inter-
preted externally and literally as it is expressed by the words and should be
understood for its own sake. No comparison, nor analogy, nor motive-seeking
(ta'lil) is permitted. According to the Zahirites, therefore, the prohibition of
interest is simply applied to the seven items which are mentioned verbally in
the hadith, and no more.
Five Categories: The orthodox schools developed another technique to interpret
the texts and apply them to the actual situations. This is the so-called Five
Categories (al-ahkam al-khamsah). These are (1) the obligatory (wajib or fard),
(2) the recommended (sunnah, mandub or mustahabb), (3) the indifferent (mubah),
(4) the reprehensible or disapproved (makruh), which is further subdivided into
a) al-makruh karahata tanzih and b) al-makruh karahata tahrim,(44)(5) the forbidden
(haram or mahzur). It makes much difference according to which category
to interpret a particualr text. Let us see some examples.
The Qur'an says:(45) "O ye who believe! When ye rise up for prayer,
wash your faces, and your hands up to the elbow." (5:6) The orthodox scho-
ols are unanimous in regarding this as "recommended (mustahabb)," not "oblig-
atory (wajib)." One has only to make ablution once a day. Then the purified
state lasts throughout the whole day so long as it is not broken by some act which
requires another ablution. And they quote, in support of their assertion, a
hadith to the effect that the Prophet, on the day of the conquest of Mecca, per-
formed five ritual prayers (salat) with one ablution. (In the course of the first
century, it is said, it was so much neglected to perform ablution each time before
the five ritual prayers that the jurists were obliged to adopt such a compromise
with the changed actualities.)(46)
Whatever the actual situation may be, however, the Zahirites strictly reject
such an interpretation, stressing the literal meaning of the text, and claim that
the ablution is required before each ritual prayer since it is "obligatory."(47)
The Qur'an says:(48) "(And if ye fear that ye will not deal fairly by the
orphans,) marry (fa' nkihu) of the women, who seem good to you...."(4:3) The
generally accepted exegesis of the verse is that every Muslim is allowed to get
married, or at the best, God has recommened the married life to the Muslims,
but it is never "obligatory." On the other hand, the Zahirites claim from the
imperative form fa' nkihu that marriage is "obligatory" for all Muslims who
have filled the conditions required in the above-quoted verse. On the whole
the Zahirites tend to interpret the text either under the category (1) or (5) (or

Vol. X 1974 97
at most b) in (4)), while the orthodox schools are inclined to interpret the pro-
hibitions and commands under the intermediate categories.
To sum up, Zahirism started as a legal theory par excellence, which metho-
dologically denied Qiyas and interpreted Ijma' in the narrowest sense in con-
tradistinction to the other schools.(49) And its main characteristic lay in its
methodology of text-interpretation. It is Ibn Hazm who applied this method
to theology for the first time(50) and set out to attack Ash'arite and other theolo-
gies. Thus Ibn Tumart, being a Zahirite legally, could later on vindicate
ardently Ash'arism, and a famous Sufi, Ibn 'Arabi (d. 638) could also claim
to be a Zahirite.(51)
The methodology of the Zahirites lies in "literal" interpretation of the text
and rejection of Qiyas. Two premises underlie this: Anything necessary is
expressed in the text and no motive-seeking (ta'lil) is allowed in the divinely
inspired text. All this may be characterized as the spirit of respect for the text
(nass) in opposition to the "arbitrary" interpretation of it among the four schools.
It does not necessarily follow, however, that the Zahirite school respects the
text and the others do not.(52) The difference is that the former puts so much
stress on the "literal" meaning of the text, even without regard to the conclusion
or the actual situations, while the latter will not separate the text-interpretation
from the actual problems of the community.

Methodological Premises of Ibn Mada'

Retrospecting the history of the Arabic grammar, Ibn Mada' says:

Verily the grammarians (God's mercy be upon them!), as I see, laid down the
principle of grammar in order to protect the Arabic language from corruption and also
its principle from alteration, and thus they reached the goal and attained what they
seeked. However, they have adhered to that which is not required of them, and in the
process of systematization they have gone too far in their quest. Thus the method of
their grammar has become complicated, and its foundation has become weakened,
and its argument has become short of convincement.... In its beginning, however,
the approach was exempt from the superflous elements and free from the limitations
and the fantasies. So its proof was the most evident among the sciences and its rules
were the most acceptable on any trial among the sciences (ma'arif). It included only
certainty...." (pp. 80-81).(53)

This is what Ibn Mada' sees in the current grammatical systems and what he

98 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

is about to do. These systems have become fused and blended with superflous,
unnecessary, useless elements, and lost their original simplicity and conciseness,
and they are now complicated and difficult to understand. To get rid of these
"superflous" elements from the grammar and simplify it -this is the objective
of his book, Kitab al-Radd 'ala al-Nuhat ("Book of the Refutation of the Gram-
marians"). He says, "My intention in this book is to remove from the grammar
that which the grammarians do not need and to criticize the mistakes which are
committed unanimously by them." (p. 85)
Now, what is Ibn Mada's idea about the cause of the degeneration of the
grammatical theories? This brings us to a more fundamental question: What
is his basic attitude towards the grammar or his methodological premise for
it? He explains this by quoting a hadith which, as he confesses, motivated
him to compose the book in question:

To assert an extra meaning (ziyadah)in the word of the speakers without any
proof (dalil) which justifies it is an obvious mistake. This, however, does not result in
any punishment. On the other hand, to do this in the Book of God... and to assert
an extra meaning in it without any evidence or proof but to say that what is put in the
nasb (accusative)is so only by a governing word in the nasb and that the governingword
in the nasb is nothing but either a word which verbally conveysa meaning or a word
which is intentionally suppressed and whose meaning resides in the mind, is forbidden
(haram)to him who knows this. Indeed, the Apostle of God said, "He who says about
the Qur'an by ra'y and is even right has already committed a mistake." What is requir-
ed by this hadithis the prohibition (of it). That which is prohibited is inviolable except
when a proof shows it. The ra'y which is not related to a proof is (prohibited). He
also said, "As for him who says about the Qur'an without knowledge,let him take seat
in Hell-fire." This is a harsh threat. That which [the Apostle of] God prohibits with
threat is "forbidden." He who supposes an extra thing in the Qur'an by word or
by meaning following a wrong opinion (zann) is evidently misguided.... (p. 92)

This is the leit-motif throughout the book.(54) In this quotation the degeneration
of the grammar is ascribed to the grammarians' application of ray (technically
qiyas, or analogy)(55) to the grammatical theories and to their assertion of the
extra meaning in the text without proof (dalil), and this is said to have a grave
consequence with reference to Qur'an-exegesis.
According to Ibn Mada', the correct interpretation of the text or speech
can be attained only when we approach it wihout applying ra'y or analogy
and thus without adding any "extra meaning and word" to it. Though he

Vol. X 1974 99
does not mention positively and systematically his right method for interpreting
the text,(56) we can surmise what it is from his following passage negating the
"government" ('amal) of words:

It is wrong rationally as well as legally (shar'an)to assert that the utterances (alfaz)
bring one another into what they are. An intelligent person would never say so for
the reasons which it is tedious to mention in the work whose concern is brevity. But
one of them is that the agent (fa'il) exists, according to its condition, where its act is
done, and that the declension (i'rab) is produced where it is only after disappearance
of the regent ('amil). Zaid, therefore, is placed in the nasb (accusative) case after
inna in our speech: Inna Zaidan,only after disappearance of inns. (p. 87)
We see from this quotation that the words in a sentance are separated from one
another. Neither "government" nor relationship is there among them. Con-
sequently we must try to interpret the sentence word by word without making
comparison or analogy.

Problem of Regent and Taqdir

Meaning of Regent and Taqdir: When Ibn Mada' says that the Arabic grammar
has become deteriorated and complicated on account of the ra'y of the gram-
marians, what does he mean concretely by ra'y? Technically it is the theory
of regent (nazariyah al-'amil). For this reason, he asserts first of all the abolition
of this theory. Now, what is the meaning of the theory of regent?
As is well known, the Arabic grammar (nahw), in its narrow sense, is the
systematic theory of explanation for the declension of the case-ending of word.(57)
And the theory of regent is the core of this system-it explains declension
(or indeclension) by the concept of "government" ('amal) of word, expressed
explicitly or implicitly. Ibn Mada' exposes this as follows:
...the nasb-,khafd-, and jazm-declensionsare only due to a verbal regent ('amillafzi) and
the raf'-declension among them is due to a verbal or implicit regent ('amil ma'nawi).
They (grammarians) explain this with reference to our speech: Daraba Zaidun 'Amran
(Zaid struck 'Amr) and say that the raf'-declension in Zaidun and the nasb-declen-
sion in 'Amran are produced only by daraba. (p. 85)
Thence comes the rule, for example, that every word with nasb-declension must
have the word which governs it in that declension (kull mansubfa-la budda la-hu
min-nasib).
These governing words, however, are not always expressed in word (lafz).

100 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

They may be "suppressed" (mahdhub) or "implicit" (mudmar or mustatir). In

these cases the grammarians "assume a virtual meaning" (taqdir) and explain
away the "exceptions." Thus they keep up the rules of the grammar and the
harmony of the language. This is, according to Ibn Mada', the very cause of
degeneration and complication of the grammatical theories. He proves the
falsity of their theory extensively and intensively.
Taqdir of SuppressedRegents: (a) Suppressionof a Known Word. The Arabic
grammarians divide the suppressed words (mahdhufat) into three parts. One
is "the suppression of a word without which the speech is incomplete, but it
is suppressed because of the knowledge of the partners of speech (mukhatab).
(p. 88) For example, the word of God, "And it is said unto those who ward
off (evil): 'What hath your Lord revealed?' They say: 'Good' (khairan)."
(16: 30) In this case the words, "Our Lord hath revealed" (anzala Rabb-na) is
suppressed, because the addressed persons know it. And when they appear
(zahara), the speech becomes complete. Rhetorically, however, the suppression
is better.
(b) Suppressionof an UnnecessaryWord. In this suppression the speech is
complete without the suppressed word; nay, when it appears, the speech becomes
defective. The syntactical distraction (ishtighal) belongs to this category. For
example, A Zaidan darabta-hu (Is it Zaid whom you struck?). In an attempt
to explain the nasb (accusative) in Zaidan, the grammarians assume virtual sup-
pression (taqdir) of darabta between a and Zaidan as the governing word of Zaidan.
When the suppressed word appears, therefore, the speech becomes defective
like: A darabta Zaidan darabta-hu. This kind of awkwardness is produced only
by their above-mentioned rule that every word with nasb-declension must have
the word which governs it in that declension. And in addition, Ibn Mada'
shows untenability of the suppression by quoting the following example: A Zaidan
mararta bi-ghulami-hi. (p. 89)
(e) Suppressionof an Implicit Word. When the suppressed word appears,
it changes the modality of the speech. Accordingly it is always implicit (mudmar)
and never appears.
For example, Ya 'Abda Allahi (O 'Abd Allah!). The grammarians assume
an implicit verb, ad'u or unadi before 'Abda Allahi. However, if the suppressed
word, ad'u or unadi appears, then the vocative changes into an ordinary sentence.
The same applies to "the causative fa" (fa al-sababiyah). For example,
Ma ta'ti-na fa-tuhadditha-na. The grammarians imply (yuqaddiruna) before tuhad-

Vol. X 1974 101

ditha-na a particle an which has the same governing force as the verb. Thence
fa-an tuhadditha-na. Furthermore, this an is the an al-masdariyah. Thus, what
is implied in that example is: Ma yakunu min-ka itydn fa-hadith. And this has
two aspects: One is Ma ta'ti-na fa-kaifa tuhaddithu-na (You do not come, therefore
how do you talk with us?), and the other is Ma ta'ti-na muhaddithan(You do not
come to us to talk). (pp. 89-91)
All these suppressed words in (b) and (c) are always implicit and never
expressed, but their meanings (ma'ani) are in the mind of the speaker. Ibn
Mada' criticizes this assumption as follows:

These implicit words (mudmarat)whose appearance is not allowed must belong

to either of the two cases, namely, either they are not existent in utterance (lafz), but
their meanings exist in the mind of the speaker, or they do not exist in the mind, like
the words whose verbal expressions are non-existent. If these implicit words do not
exist in the mind, nor their verbal expressionsin the speech,then what is it that produces
the nasb-declension? What is it that givesthe implicit meaning? It is absolutelyimpos-
sible to ascribe the "government" ('amal)to a non-existent word (ma'dum). If you say:
The meanings of these suppressedwords exist in the mind of the speaker, and the speech
becomes complete with those meanings, and it is part of the speech which resides in
the mind and is to be expressedby the words, but whose verbal expressionsare suppres-
sed for brevity (ijazan) as the words which are allowed to appear are suppressedfor
brevity, then it necessarilyfollowsthat the speech is defective (naqis), and that it becomes
complete only with the implicit words since they are part of the speech. Thus we
add to the speech of the speakers that which is not expressedverbally, without any proof
(dalil) but their assertion that every word in the nasb-declensionmust have the verbally
expressed governing word in that declension. (p. 91)

According to Ibn Mada', if the suppressed words whose appearance is not allowed
do not have the meaning in the mind of the speaker, it is a sheer nonsense to
assume a "governent" in them. On the other hand, if those words have the
meaning in the mind of the speaker and the speech must be understood with the
help of this meaning which is never expressed verbally, there is something wrong
with the speech, namely, it is defective. In reality, however, the speech never
needs such an assumption. It is perfectly understandable by itself. Those
suppressed words, therefore, are nothing but an unnecessary, extra element.
Do we not see in this argumentation the same attitude that the Zahirites
show in the exegesis of the sacred text; that is, anything necessary is expressed in
the text?

102 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

Die Zahirschule kann dieser auf speculativer Willkur beruhenden Erweiterung des
geschriebenenGesetzes ihre Zustimmung nicht geben; waren jene Gattungen gemeint,
so hatte der Prophet sicherlich dem kurzeren Ausdruck den Vorzug gebend, statt einzel-
ne Arten zu nennen, bloss den einen Gattungsnamen gebraucht.(58)

Taqdir of "Suspenders" (Muta'allaqat): To the same foregoing category belong

the implicit words which the grammarians assume in relation to the prepositional
phrase, which are either an enunciative (khabar), or a relative clause (silah), or
a qualificative clause (sifah), or a denotative state (hal).
For example, Zaidun fi al-dar (Zaid [is] in the house). The grammarians
assume an implicit word, mustaqirr (remaining) or qa'im (staying) upon which
fi al-dar is suspended. This assumption was simply needed by their rule (qa'idah)
that if the prepositions in the prepositional phrases are not redundant (za'idah),
they must have regents, either explicit or implicit. (p. 99).
According to Ibn Mada', there is no need of this kind of implication. We
can dispense with it since "all this is a complete speech composed of two nouns
which show two meanings and between which is there a relationship. And
this relationship is shown by fi (in). So we do not need anything other than
this." (p. 99)

Taqdir of Implicit Pronouns: (a) Implicit Pronouns in the Derivatives. The

grammarians assume an implicit pronoun in the derivatives (mushtaqqat)such as
the present participle (ism al-fa'il), the adjective assimilated to it (al-mushab-
bahah bi-hi) and the past participle (ism al-maf'ul). For example, in the sentence:
Zaidun daribun 'Amran (Zaidis a striker of 'Amr), they assume an implicit pronoun
(in this case, huwa) in daribun, which indicates the agent (fa'il). They quote, in
support of their claim, the following examples where the implicit pronoun appears
manifest or its existence is apparent:
Zaidundaribunhuwa wa-Bakrun 'Amran. (Zaid and Bakr are strikers of 'Amr.)
Marartu bi-qawmin 'arabin ajma'una. (I passed by a group of Bedouins
all of them.)
Ibn Mada' criticizes this assumption (taqdir) as follows: There are two ele-
ments implied in the present participle, namely, the action and its agent (but
without the agent's name specifically indicated). Thus in the example: Zaidun
daribun 'Amran, the participle daribun indicates the agent of darb (striking), whose
name is not shown by darib itself, but Zaidun. Why then, such an additional as
an implicit pronoun?

Vol. X 1974 103

 As for the case of conjunction (Zaidun daribunhuwa wa-Bakrun 'Amran),it is
certainly admitted, as the grammarians assume, that the manifest pronoun
(bariz) is a corroboration of the implicit pronoun which is not expressed verbally,
but how is it possible to extend this specific case of conjunction to all other cases
where the pronoun is not implied by the speaker himself? The same is true with
the other example. That is to say, the expression ajma'una is exceptional. It is
certainly admitted that the word ajma'una is a corroboration of the implicit pro-
noun in 'arabin, but why is it necessary to generalize these exceptional few cases
and assume an implicit pronoun in all cases?
(b) Implicit Pronounsin the Verbs. In like manner the grammarians suppose
an implicit pronoun in the verb as its agent, say, in qama, when we say, Zaidun
qama (Zaid stood up). Because of their rule that "the agent does not come
before the verb and that the verb must have the agent" (p. 103),(59)they cannot
take this Zaidun as the agent of qama. Thus they assume a pronoun in the verb.
According to Ibn Mada', however, we do not need this sort of assump-
tion of a pronoun (damir), since "the verb itself shows the pronoun as well as
the tense by its form (bi-lafz-hi)." (p. 105) For example, from ya in ya'lamu we
know it is the third pers. masc. sing.; from ta in ta'lamu it is known to be the
second pers. masc. sing. or the third pers. femi. sing.; from 'alima it is known to
be the third pers. masc. sing. in the finished form, and so on. As for the rules
for the feminization (ta'nith) and the pluralization (jam') of the verb when it
comes before the agent, he agrees with the grammarians.

"Conflict" and "Syntactical Distraction"

"Conflict": The "conflict" (tanazu') is the section (bab) concerning the

two agents and the two objects. (pp. 107-17) According to the grammarians,
two regents cannot govern one single regiment (ma'mul). Therefore, one of
the two regents must be chosen and the agent is assumed in the other regent.
Ibn Mada' does not oppose this rule basically (in that case he uses the word
ta'liq rather than 'amal). However, when the grammairans apply the theory
of regent here and coin the artificial expressions by analogy (qiyas) which are
not used by the Arabs in reality and go even so far as to reject the speech of the
Arabs which does not fit in their rules, he raises an objection and criticizes them.
He shows this in two examples: the verb zanna, which takes two objects and the
verb a'lama, which takes three objects.

104 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

Example 1:
In the singular,
Zanantu wa-zanna-ni Zaidun shakhisan. (I thought Zaid, and he thought
me, going out.)
Zanantu wa-zanna-ni-hi Zaidan shakhisan.
In the dual,
Zanantu wa-zanna-ni shakhisan al-Zaidain shakhisain.
In the plural,
Zanantu wa-zannu-ni shakhisanal-Zaidina shakhisina.
Example 2:
In the singular,
A'lamtu wa-a'lama-ni Zaidun 'Amran muntalaqan. (Zaid told me, and I
told him, that 'Amr was set free.)
A'lamtu wa-a'lama-ni-hi iya-hu Zaidan 'Amran muntalaqan.
In the dual,
A'lamtu wa-a'lama-ni-himd iya-huma al-Zaidain al-'Amrain muntalaqain.
In the plural,
A'lamtu wa-a'lamu-ni-him iya-hum al-Zaidina al-'Amrina muntalaqina.
Virtually,
A'lamtu al-Zaidina al-'Amrina muntalaqina wa-a'lamu-ni-him iya-hum.
Ibn Mada' asks: Do the Arabs speak such complicated sentences at all? In
fact, we find this kind of expressions only in the textbooks of the grammarians.
We must, therefore, strictly refrain from applying the rules induced from the verbs
which take only one object to the verbs requiring two or more objects, by analogy
"until it is heard from the Arabs." (p. 121) This is his fundamental attitude.
"Syntactical Distraction" (Ishtighal); This is the section concerning "the dis-
traction of the verb from the object by its pronoun" (p. 118) or the noun annexed
to its pronoun. For example, Zaidan darabtu-hu. This section is very difficult
and complicated to understand, since it is deeply related to the theory of regent.
The grammarians fervently debated with reference to many complicated ex-
amples.
We have already mentioned Ibn Mada's criticism of one aspect of this
theory in reference to "suppression." Here he argues another aspect of it
quoting three famous examples:
A Zaidan lam yadrib-hu illa-huwa. (quoted by Akhfash).
Akhawd-ka zanna-huma muntalaqain. (qouted by Akhfash).

Vol. X 1974 105

 A anta 'Abdu Allahi darabta-hu. (quoted by Sibawaihi).
And he criticizes as follows:
To occupy oneselfwith these problems-which are imagined and not actually used,
nor needed-is uselessfor him who intends to investigate that of which necessity is
felt. To remove these and the like from the science of grammar is to investigate and
simplify it. Thus it is more deserving to dive into the problems like these which are
useful to speech than to occupy oneselfwith that which is of no use to speech like their
argument: By what is the object placed in the nasb (accusative), the agent or the verb,
or both? (p. 127)
On the other hand, Ibn Mada' proposes his own simple rule: If the noun placed
first is referred by the pronoun annexed to the verb in the raf' (nominative), the
noun is placed in the nominative since its pronoun is in the position of the nomi-
native. (Ex. A Zaidun qama.) Otherwise, it is placed in the nasb (accusative),
when it is in the position of the accusative. As for the rest, we have only to
"follow the speech of the Arabs." (p. 121).

Abolition of the Theory of Regent: Ibn Mada' attacks the theory of regent
further in relation to the more fundamental grammatical problem. He critcizes
the idea of regent itself. Quoting the words of Ibn Jinni, he says as follows:
Concluding the discussionin his al-Khas a'is, on the verbal regents (al-'awamilal-
lafziyah) and the implicit regents (al-'awamil al-ma'nawiyah), Abu al-Fath [Ibn Jinni]
says: As for the reality and the conclusionof the discourse, it is that the "government"
in the raf'-, nasb-, jarr-, and jazm-declensions belongs to the speaker (mutakallim) himself,
not to any other thing. He emphasized "the speaker" by the word "himself" (nafs)in
order to elevate the probability. Then he adds a stress further with his words "not to
any other thing" (la li-shayghair-hi). This is the doctrine of the Mu'tazilites. As for
the tenet of the People of Truth, these sounds (aswat) are only from the Act of God.
They are related to man exactly in the same way as the rest of his volitional acts
(af'al-hu al-ikhtiyariyah) are. (p. 87)
He is not arguing out of spite. He is simply stressing the inviolability of the
grammatical materials or the reality of speech. Its phonetic and syntactical
modality is a given thing (by God, according to Ibn Mada'). It is never to be
arbitrarily ascribed to the speaker kimself, or more specifically the "government"
assumed by the grammarians. According to Ibn Mada', as we have seen before
(supra, pp. 99-100), it is impossible from the very beginning to imagine the "gov-
ernment" between two words, since one word comes only after the disappearance
of the other.

106 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

Furthermore, the grammarians are self-contradictory in their assertion:

"Every word in the accusative must have the verbal governing word. On
the other hand, these suppressed words whose appearance is not allowed must
not exist in utterance and will, and nevertheless the speech is perfect without
them." (p. 98) If one argues against Ibn Mada' by saying that the theory of
regent is an artificial device for understanding, like, say, the point and line sup-
posed by the geometricians, which are actually bodies, he answers, "The geome-
trical point and line are a help to understanding of the reality for the student.
There is nothing like this, however, in the supposition of these regents except
conjecture and fantasy." (p. 98) In short, this theory is unnecessary. "This
is concerning the speech of man. As for the Speech of God, however, it is for-
bidden (haram)." (p. 104, 141) If the grammar requires universality as a sci-
ence, naturally it is also applied to the text of the Qur'an, the direct Speech of
God. At this point, the grammar is involved inseparably in theology.(60) Here
we see the necessity of Ibn Mada's criticism of the grammatical theories as a
Zahirite.

Problem of Qiyas

Meaning of Qiyas: When Ibn Mada' asserts the abolition of qiyas, what does
he mean by qiyas? To be sure, the grammar in general as a science is based on
qiyas in its broadest sense of "analogy" or "comparison," not to speak of the
theory of regent and the implicit assumption (taqdir). (The theory of regent
is ultimately based on qiyas.) And certainly he does not deny qiyas in general.
But he simply intends to reject its abuses committed by the grammarians.
The qiyas, as he sees it, consists in attributing the motivation to something
else (ta'lil). And in case of attribution, the same motivation ('illah) in the prin-
cipal judgement (asl) must be in the subsidiary judgement (far'). (p. 157) The
grammarians, however, are wrong in applying qiyas. They compare the noun
to the verb in "government," and the particle inna and its cognates to the transi-
tive verb (al-af'al al-muta'addiyah) in "government" and so on. (p. 157)(61) Ibn
Mada' takes up two topics: the second and the third motivations (al-'ilal al-thaw-
ani wa-al-thawalith) and the (grammatical) exercise (tamrin).
The Second and the Third Motivations: Example: Qama Zaidun. Concerning
Zaidun, if one is asked, "Why is it put in the nominative?" he will answer, "Be-
cause Zaidun is the agent. All the agents are put in the nominative." (p. 151)

Vol. X 1974 107

 One may be further asked, "Why is the agent put in the nominative?" This
is the second motivation. To this question one can, and should, simpy answer,
"So speak the Arabs. That is established by induction from the speech which
is spoken customarily (al-kalam al-mutawatir)." (p. 151) It is a reality. There
is no need, nor use of further investigation. No answer is required for this kind
of question just as the jurists are not required to answer the question about the
motivation of the text (nass).
The grammarians do, however. They answer: In order to distinguish the
agent from the object. And they go further and ask: Why is the judgement
not reversed in favor of the accusative for the agent and the nominative for the
object?- This is the third motivation. And they give the answer to this ques-
tion!: Because the agents are few for the reason that the verb has only one
agent, while the objects are many. Therefore, the heaviest, namely, the raf'-
declension, is given to the agent, while the lightest, namely, the nasb-declension,
is given to the object. Consequently, the few, but heavy agents and the many,
but light objects are kept in balance. We see in this way of thinking the
grammarians' philosophy of harmony (supra, pp. 89-90).
According to Ibn Mada', the philosophical inquiry such as this is of no
use and does not help us understand the text itself. On the contrary, it simply
makes the grammar more and more difficult.
It does not increase our knowledgeabout the fact that the agent is put in the nomi-
native. Even if we did not know it, that ignorance would not do us any harm, since
it is an established rule that the agent is put in the nominative. And this is what we
attain by investigating the authentic material which comes to our knowledge. (p. 152)
The grammarians should be content with the first motivation and be concerned
with accumulating the rules such as the nominative case for the agent (raf' al-
fa'il), instead of occupying themselves with philosophical speculation.
"Exercise": Ibn Mada' discusses another example of the abuses of qiyas,
namely, grammatical exercise (tamrin). Example: Make the pattern fu'l out
of bai'. (p. 161) To this question, one answers bu'. And he explains: Its
original form is buy', but has substituted the waw for this ya on account of the
dammah of the previous letter, since it is difficult (thaqil) to pronounce buy' as it
is. In this case, he is based on the analogy (qiyas) of the speech of the Arabs,
namely, muqin and musir (instead of muyqin and muysir, the present participles of
the fourth derived form, ayqana and aysara respectively). When the motive
('illah) disappears, however, they take the regular formation. Thus, the plural

108 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

of musir is mayasir, and its diminuative (tasghir) is muyaisir.

Another answer to the above question is bi' deduced on the analogy of
bid, 'in and ghid in the speech of the Arabs (p. 162) (the plural forms of baida',
'aina' and ghaida' respectively). Their original forms are buyd, 'uyn and ghuyd
respectively, following the general pattern fu'l as the plural form of fa'la. The
substitution of the waws for the yas, however, is abandoned in favor of the kasrah
on the first letters of these words.
And both sides argue in support of their own assertions, based further on
the analogy of some other words. For example, the former vindicate their
position by saying "that bu' is singular and to follow the pattern of musir and the
like is better than to follow the pattern of the plural." (p. 162) The latter, on
the other hand, support their views by saying that "to substitute the kasrah for
the dammah in favor of the ya is lighter (akhaff) and it is better than to change
the ya for the waw, since the ya is lighter and it is more frequently used than
the waw." (p. 163)
What does this sort of argumentation mean to Ibn Mada'? It is already
obvious to us now. He says:
This is only one topic. How about more of the same kind? The argument about
it is lengthened and the table of speech concerning it is extended with no profit from it
and no need of it. People cannot learn the genuine, pure language. How much
less this unnecessary, speculated argument! (p. 164)
New Grammar: As we have seen in the foregoing, Ibn Mada' himself has never
constructed a new grammatical system. For it to be done, the old edifice must,
first of all, be smashed and destroyed. And Ibn Mada' undertook this task.
He dealt a heavy blow on the current grammatical theories. From his attack
and argumentation, however, we can readily know the course which the system-
atization of a new grammar would take in the future.
First of all, the new grammar must be simple and plain. Secondly, in
close relation to the first, it must be such that it will respect the reality of speech
and convey the meaning "faithful" to the text. The speech or the text comes
first, and then the rules (not vice versa!). The reality of speech must be res-
pected to the highest possible extent. For this purpose, anything which impairs
the linguistic reality by introducing an extra meaning in it must be strictly
shunned.
In this perspective, the theory of regent (nazariyah al-'amil) and "virtual
assumption" (taqdir) must be discarded. In fact, it is not the "government"

Vol. X 1974 109

('amal) nor the "regent" ('amil) that changes the case-ending of word. The
linguistic modalities are given things to be used by the speaker to express his
intention. Efforts must be made to understand this intention of the speaker,
not the "government." We must be strict in applying qiyas so that we may not
draw a conclusion which is not used at all. We should refrain from indulging
in the second and the third motive-seekings. We must stop assimilating, for
example, the verb to the noun in declension by wrong analogy. Both have
different realities with different modes (suwar) and, therefore, under different
rules. What we have to do is to collect as many modes of each section (bab)
as possible and to induce rules from them, without indulging in idle speculation.
And in this case, the actual speech or text must be, first of all, taken into con-
sideration. By accumulating the rules gained in this way we can construct a
new system of grammar.

Epilogue

From the above discussions, as we see now, there are many parallels bet-
ween the legal theory of the Zahirites and the grammatical theory of Ibn Mada'.
The main one is the fundamental attitude towards the textual reality or the
speech-respect of the text as a given inviolable reality. Thence comes their
premise that the text should be understood and interpreted "as it is," namely,
in its "external" (zahir) sense. One must not add any extra meaning or expres-
sion in the text on no evidence (dalil) for it, nor seek any motive in it. Thus
Qiyas is denied by the Zahirites. In case of Ibn Mada', too, qiyas and the
theory of regent based on it are denied. To both of them, the text or the speech
is perfect by itself. If something were needed, it would have been expressed
verbally. If we approve the parallels between the attacks of the Zahirites against
the orthodox legal schools and those of Ibn Mada' against the "orthodox"
grammatical system, then we might be able to speak in the same way about the
relationship between the four legal schools and the "orthodox" grammatical
schools. This is another interesting topic.
Between Zahirism as the legal theory and that of Ibn Mada's grammatical
theory, however, there is a large difference concerning qiyas. In the legal
theory of the Zahiristes, there is no room for Qiyas. It is absolutely denied,
at least theoretically. Ibn Mada', on the other hand, does not deny qiyas en-
tirely. He admits of it in so far as it helps us understand the reality. Thus he

110 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

approves the first motivation (al-'ilal al-uwal), without which the science of the
Arabic language would not exist.
There still remains, of course, a question concerning the adequateness
of the Zahirite approach to text-interpretation, namely, the premise that the
right interpretation comes from the "external" meaning. Leaving aside this
complicated problem, however, it cannot be denied that Ibn Mada' has opened
up a new vista for simplifying the Arabic grammar.

Notes

(1) See, for example, Jarmanus Farhat, Kitab bahth al-matalib fi 'ilm al-'arabiyah (Beirut, 1882)
p. 134.
(2) Ahmad Amin explains this traditional ascription of the Arabic syntax (nahw) to Abu
al-Aswad as follows: He is supposed to be the first to use the vowel signs. He told his
scribes to put different signs above or below the letters according as they were "opened"
(maftuh) or "closed" (madmum) or "broken" (maksur) when he read them. This is how
the declension (i'rab) and its rules (qawa'id) became the focus of attention. Thus "he
who says that Abu al-Aswad invented the syntax (nahw) means something like this,"
and later on when the word nahw received the technical meaning, he became to be
regarded as the founder of nahw in its technical sense (Duha al-Islam [3 vols., Cairo,
6th imp, 1961-62], II, pp. 286-87).
(3) Ahmad Amin, Zuhr al-Islam (4 vols., Cairo, 3rd imp., 1962-64), II, p. 118.
(4) H. Fleisch, Traite de philologie arabe (Vol. I, Beirut, 1961-), p. 13.
(5) For example, al-Mubarrad (d. 285), who wrote Kitab al-madkhal ila Sibawaihi, Kitab
ma'na Kitab Sibawaihi, Sharh shawahid Kitab Sibawaihi and Kitab al-radd 'ala Sibawaihi, and
al-Sirafi (d. 368), one of whose books was the Commentary (Sharh) on al-Kitab.
(6) G. Weil, "Zum Verstandnis der Methode der moslemischen Grammatiker," Festschrift
Edward Sachau (Berlin, 1915), p. 386.
(7) H. Fleisch, ibid., pp. 2-3.
(8) This tendency is particularly strong in the Basran school in the early history of the Arabic
grammar.
(9) Al-Jahiz (d. 255) is claimed to have perceived the necessity of constructing a new gram-
matical system, but he did not go further (Shawqi Daif, Madkhal to Ibn Mada', Kitab
al-radd 'ala al-nuhat [Cairo, 1947], p. 48), and Ibn Jinni tried to lay down another syn-
tactical theory (Ahmad Amin, Zuhr al-Islam, II, p. 118).
(10) For its general history, see R. Le Tourneau, The Almohad Movement in North Africa in the
Twelfth and Thirteenth Centuries, Princeton, 1969.
(11) I. Goldziher, Le livre de Mohammed Ibn Toumert Mahdi des Almohades (Alger, 1903), p. 54.
(12) See infra, p. 98.
(13) I. Goldziher, ibid., p. 44.
(14) I. Goldziher, ibid., pp. 54-56.
(15) According to one of his biographers, 'Abd al-Wahid, "in most question, he followed the
system of Abu al-Hasan al-Ash'ari, but he agreed with the Mu'tazilites in their denial of
the Divine Attributes and in a few matters besides...." (quoted from R. A. Nicholson,
A Literary History of the Arabs [London, rep. 1956], p. 431).
(16) Shawqi Daif, Madkhal, p. 5.
(17) Shawqi Daif, ibid., p. 6.
(18) I. Goldziher, Die Zahiriten, ihr Lehr-system und ihre Geschichte (Leipzig, 1884), p. 174.

Vol. X 1974 111

(19) Shawqi Daif, ibid., p. 6. The Zahirites strictly confine Ijma' to the Companions of the
Prophet (ijma' al-sahabah). This is another main difference of the Zahirite school from
the four orthodox schools. But we do not discuss this problem in this article, since it is
not directly connected with our main topic.
(20) Ibn Falhun, al-Dibaj al-mudhahhabfi ma'rifah a'yan 'ulama' al-madhhab (Cairo, 1351 A. H.),
p. 48.
(21) Suyuti, Bughyah al-wu'a fi tabaqat al-lughawiyin wa-al-nuhat (Cairo, 1326 A. H.), p. 138.
(22) Ibn Falhun, ibid., p. 48.
(23) This latter date is given by Hajji Khalifah, Kashf al-zunun (ed. by G. Flugel. 7 vols.
London, 1835-58), III, p. 355.
(24) Suyuti, ibid., p. 139.
(25) This is against the fact that Ibn Falhun mentions his name among the Malikites in his
biographical dictionary of this school (al-Dibaj). And one of his teachers, Qadi 'Iyad
(supra, p. 92) was apparently a Malikite (See C. Brockelmann, Geschichte der arabischen
Litteratur [2 vols., Leiden, rev. ed. 1943-49], II, pp. 455-56). R. Arnaldez takes it for
granted that Ibn Mada' was a Zahirite (Grammaire et theologiechez Ibn Hazm de Cordoue
[Paris, 1956], pp. 89-97).
(26) I. Goldziher, Die Zahiriten, p. 3.
(27) I. Goldziher, ibid., pp. 3-4.
(28) The Sunnah as the Prophetic precedent transmitted in the form of hadith is the concept
established later in the Islamic history (chiefly by the efforts of al-Shafi'i). And there
was no unanimous agreement about the concrete content of the sunnah in this early period.
(29) I. Goldziher, ibid., p. 13.
(30) I. Goldziher, ibid., p. 13.
(31) See J. Schacht, An Introduction to Islamic Law (London, rep. 1966) and his "Usul," Shorter
Encyclopaedia of Islam (Leiden, 1953).
(32) I. Goldziher, ibid., p. 25.
(33) I. Goldziher, ibid., p. 10.
(34) I. Goldziher, ibid., p. 11.
(35) I. Goldziher, ibid., p. 11.
(36) Quoted from R. Paret, "Istihsan and Istislah," SEI.
(37) That is, "seeking for a link (to something which is known and certain). It means: when
it is not ascertainable that the conditions under which a certain fiqh rule is applicable have
changed, the rule remains valid as if the conditions had not changed" (Th. W. Juynboll,
"Istishab," SEI).
(38) Quoted from I. Goldziher, ibid., p. 30. Cf. F. Rosenthal's translation of this passage,
The Muqaddimah (3 vols., Princeton, rev. ed. 1967), III, pp. 4-5, and his footnote 171.
(39) I. Goldziher, ibid., p. 42.
(40) R. Arnaldez, Grammaire et theologie,p. 90.
(41) I. Goldziher, ibid., p. 43.
(42) I. Goldziher, ibid., pp. 41-42.
(43) As for the verse-numbering of the Qur'an, we follow that of the Cairo edition, and the
English translation is from M. M. Pickthall, The Meaning of the Glorious Koran (Mentor
Books, 1953).
(44) The former means the act which is "reprehensible," but whose non-committal is "recom-
mended" only to those who lead a pious life, although whose committal is not to be blamed.
The latter means the act which is so "reprehensible" that it almost belongs to the next
category.
(45) I. Goldziher, Die Zahiriten, pp. 48-50.
(46) I. Goldziher, ibid., p. 49.

112 ORIENT
 IBN MADA'S CRITICISM OF ARABIC GRAMMARIANS

(47) I. Goldziher, ibid., p. 50.

(48) I. Goldziher, ibid., pp. 74-75.
(49) I. Goldziher, ibid., p. 131.
(50) I. Goldziher, ibid., p. 119.
(51) I. Goldziher, ibid., p. 185.
(52) For example, Dawud himself is said to have admitted of Qiyas only in the most evident
case of necessity. This concession was later repealed by a fanatic Zahirite, Ibn Hazm
(I. Goldziher, ibid., p. 37). However, Ibn Hazm himself, who relentlessly attacked the
figurative interpretation of the Qur'anic verses, was obliged to "assume the implicit mean-
ing" (taqdir) with reference to the text whose external meaning shows the corporality
of God (I. Goldziher, ibid., p. 167).
(53) The pagination in the bracket shows that of Ibn Mada's Kitab al-radd 'ala al-nuhat.
(54) He repeats similar expressions throughout the book (Cf. p. 80, 82, 106, 141, etc.).
(55) See infra, pp. 107-108.
(56) As far as I know, interestingly enough, Ibn Mada' has never used the word "zahir" in his
book in question.
(57) See the definition given by Farhat in his Bahth al-matalib, pp. 134-35.
(58) I. Goldziher, Die Zahiriten, p. 42.
(59) According to the footnote by the editor (Shawqi Daif), this rule is asserted only by the
Basran school.
(60) For this interesting topic of the relationship between the Zahirite philosophy of language
and theology, see R. Arnaldez, Grammaire et theologie.
(61) Ibn Mada' does not dwell upon this point. Al-Sirafi mentions five "similarities": (1)
Both by nature have the general meaning and are specified by adding letters. Example:
sa-yaqumu and al-rajl. (2) Both can be suffixed by lam al-ibtida'. (3) As the unfinished
verb has two tenses, i, e., the present and the future, so the noun has two meanings. Exam-
ple: 'ain, meaning "eye" and "fountain." (4) Both can be attached to a noun as the
concomitant state. Example: Marartu bi-Zaidin yadribu or daribun. (5) The unfinished
verb and the noun in the form of fa'il have the same order of vocalization a-i in yadribu
and darib (Quoted from G. Weil, "Zum Verstandnis," p. 390).

Vol. X 1974 113