THE MUSANNAF OF .

ABD AL-RAZZ^Q AL-SAI{'ANT

AS A SOURCE OF AUTHENTIC AHÄDITN OF
THE FIRST CENTURY A.H.*
HARALD MOTZKI, (Jniversitöt Homburg

I
THr question of when and where afuädTth-especially those of the Prophet-
arose is nearly as old as the kadTth itself. Muslim scholars tried generally, but not
exclusively, to check the path of transmission of the traditions (isndd) and the trans-
mitters (rijdl) mentioned in each isndd. Western studies of Islam since the second half
of the nineteenth century have pointed out that this method of ltadlth criticism is
unreliable and have concentrated on the content of the text when judging the authen-
ticity of a l.tadTth.Ignaz Goldziher's thesis that the traditions ascribed to the Prophet
and the Companions (;afuAba) contained in the classical collections of afudd1th are not
authentic reports of these persons but rather reflect the doctrinal and political develop-
ments of the first two cinturies after Mullammad's deathr is based primarily on
analysis of the content of the l.tadlth (matn) and not the transmitters.
Joseph Schacht, too, when trying to date ahädTth, first studied their contents and
classified them within the framework of the development of the issue to which they
refer.2 He considered criteria from the asantd only secondarily and only if they were
consistent with the chronology first arrived at after consulting the contents (mutün).
Otherwise, he rejected the information of the asänTd as false or fabricated. Like Gold-
ziher, Schacht proposed general statements concerning the time when certain groups
of traditions and types of transmissions originated. He regarded these general conclu-
sions on the development of the fuadlth not as heuristic assumptions, but as historical
facts, and he did not limit his conclusions to the legal al.ttidTtft on which he had based
his theories.3
The low esteem in which Goldziher and Schacht held the isndd and Muslim r'snal
criticism in tackling the problem of dating al.tädTth was challenged by a research
approach which may be called "tradition-historical" ("überlieferungsgeschichtlich").
This approach, familiar in Western Islamic studies since the work of Julius Wellhausen,
tries to extract earlier sources from the compilations we have at hand, which are not
preserved as separate works, and it focuses on the materials of certain transmitters-

* The German version of this paper was pre- Fred Donner and L. Paula Woods for their help in
sented at the Colloquium on HadTth and Histori- revising my English translation.
ography held in Oxford in September 1988. I thank I See I. Goldziher, Muhammedanische Studien
(Halle, 1888-90), vol. 2, p.5 and passim.
2 See J. Schacht, "A Revaluation of Islamic Tradi-
tion," Journal of the Royal Asiatic Society 49
UNES 50 no. I (1991)l Q9a\:143-54, esp. 147; idem, The Origins of
o l99l by The University of Chicago. Muhammadan Jurisprudence (Oxford, 1950), p. I
All rights reserved. and passim.
0022-2968 I 9l / s00I -0001$ I .00. 3 See n. 2 above.
 JounNnr oF Nnln EnsrenN Sruorns
rather than on hadlth clusters dealing with specific topics. The source-analytical works
of Heribert Horst, Georg Stauth, Fuad Sezgin, and othersa suggest that Goldziher and
especially Schacht, viewed the isnqd too skeptically and that they generalized too
quickly from single observations. The tradition-historical method, however, runs the
risk of overestimating the historical value of the isndd, as Schacht rightly emphasized
in connection with the work of Leone Caetani.s
In this article, I will once again address the source-analytical and tradition-historical
approaches and try to show how we can ascertain whether, or to what degree, the
chains of transmission of al.tddTth are reliable. A few valuable source-analytical studies
exist in the field of tafslr,6 but I will show in what follows that the issue can also be
examined successfully in the realm of legal traditions, those on which the hadTth
theories of Schacht depend. As did Schacht,T I maintain that the methods used and the
results obtained in this special area of traditions apply, in principle, in other areas as
well, for example, in the realm of historical traditions.s

II

Among the many existing l.tadlth compilations,the Mu;annaf of the Yemenite'Abd

al-Razzdq al-San'änT (d. 2ll1826) is, for reasons which will be explained below, espe-
cially well suited for a source-analytical approach. This work, the eleven-volume
edition of which is based on the rare manuscripts of it which survive,e admittedly
raises some questions regarding completeness and original composition because it is
compiled from different riwdyät (transmissions). Ninety percent of it, however, goes
back to one single transmitter, Isheq b. Ibrählm al-DabarT (d. 285/898). He probably
received it in written form from his father, a pupil of 'Abd al-Razzdq, but skipped his
father in the riwäya because he had, or claimed to have had, an ijdza (permission to
transmit) for the Mr4annaf from 'Abd al-Razzdq himself, having attended his lectures
as a child together with his father. Islräq was six or seven years old when nbd al-
Razzaq died.lo The great age difference between lAbd al-Razzdq and Ishäq al-DabarT
does not seem to effect the validity of his transmission, at least for a historian. There is
no hint that Ishäq fabricated the texts in toto or even partially and ascribed them to
'Abd al-Razzäq" Other than a few rare notes of transmitters, 'Abd al-Razzdqmust be
considered the real author of the Musannaf.
Even a cursory reading of the work reveals that most of its books (kutub\ contain
materials said to come mainly from three persons: Ma'mar, Ibn Jurayj, and ath-

4 H. Horst, "Zur Überlieferung im Korankom- Ihre Entwicklung in Mekka bis zur Mitte des 2.15.
mentar at-Tabarfs," Zeitschrtft der Deutschen Mor- Jahrhunderts, Abhandlungen für die Kunde des
genltindischen Gesellschaft 103 (1953): 290-307; Morgenlandes, vol. 50, 2 (Wiesbaden, 1990).
G. Stauth, Die Überlieferung des Koronkommentors 9'Abd al-Razzaq b. Hammäm al-Sancänr, Al-
MugAhid b. Gabrs (Gießen, 1969); F. Sezgin, BufiI- Musannaf, ed. HabTb al-Rahmän al-A'zamT, I I
rthin kaynaklan hakkmdo aro$tumorlar (Istanbul, vols. (Simlak, l39l I 1972).
le56). r0 Adh-DhahabT, MTzan al-ftidal
fT naqd al-rijdl,
5 Schacht, "Revaluation," p. 148. ed. M. B. al-Nacsänr (Cairo 1325 I 1907), vol. l,
6 See n. 4 above. p. 58; Ibn Hajar al-cAsqalänT, Lisdn al-mTzan (Hyder-
7 Schacht, "Revaluation," pp. 148, 150 f. abad, 1329-3t), vol. l, pp. 349 f .; al-gafadl, Al-
8 The arguments covered in this article are more Wdfl bi-l-wafayat: Das biographische Lexikon des
fully developed and documented in my forthcoming $alal.taddTn b. Aibak (Wiesbaden, 1972), vol. 6,
book Die Adönge der islamischen Jurisprudenz- pp. 394 f.
 T*rn MustNt'ttr or lAr-Rtzz^TQ Ar,-SIN'ANI 3

Thawrr. Exceptions are the books al-MaghdzT and, al-JAmi', which are overwhelmingly
composed of texts from Ma'mar, and the Kitdb al-Buyü', where transmissions from
Ibn Jurayj occur only rarely. On the basis of a representative sample of 3,810_single
traditions-comprising 21 fercent of the relevant sections of the entire workrl-the
supposed origins of the texts transmitted by i{bd al-Razzaq can be more exactly
defined: 32 percent of the material comes from Ma'mar,29 petcent from Ibn Jurayj,
and 22 percent from ath-Thawrf. Transmissions from Ibn 'Uyayna (4 percent) follow.
The remaining 13 percent ofthe texts are said to stem from about 90 different persons
(from each only I percent or less), among them famous legal scholars of the second
century .t.n. such as Abü Fan-rfa (0.7 percent) and Mälik (0.6 percent)
If the particulars lAbd al-Razzdq gives about the origin of his material are correct,
then the work is compiled from three large sources which are themselves made up of
several thousand traditions. The enormous size of the supposed sources suggests that
we may be dealing with either originally independent works-or at least parts of
them-or with the contents of the teachings of the three named authorities who could,
judging from their age, be teachers of lAbd al-Razzdq, or both. On the other hand, we
cannot rule out a priori the possibility that 'Abd al-Razzäq generally fabricated the
information on the origin of his material and attributed it fictitiously to these people.
Which of these two hypotheses is the more probable could perhaps be decided with the
help of biographical and bibliographical reports about the persons in question. But
since the reliability of such reports is no more certain than that of the statements of
our author, we have to find a solution from the work of 'Abd al-Razzäq itself. The
clue to it can be found by analyzing the four largest clusters, or complexes, of trans-
missions in a bit greater depth.
Let us suppose that 'Abd al-Razzdqhad arbitrarily ascribed his material to the four
above-mentioned informants: Ma'mar, Ibn Jurayj, ath-ThawrT, and Ibn 'Uyayna. If
this were the case, we would expect that the transmission structure of these four
groups of texts would be similar because they were put together at random-a pro-
ieOuie that Schacht proposed for certain links in the asäntd.t2 As background, I have
summarized below the information on the origins attributed to the traditions contained
in the four groups oftexts.

1. In the group of texts which allegedly came from Ma'mar [b. Reshid], 28 percent of
the material is said to stem from al-Zuhr! 25 percent from Qatäda [b. Di'äma], ll
percent from Ayyüb [b. Abi TamTma], a little more than 6 percent from anonymous
persons, and 5 percent from Ibn Täwüs. Ma'mar's own statements amount to only I
percent. The rest (24 percent) is distributed among 77 names.13

2. In the group of transmissions ascribed to Ibn Jurayj, 39 percent is supposed to go

back to 'Atä' [b. AbI Raba[r], 8 percent to unnamed persons, 7 percent to 1A,mr b.
Drnär, 6 percent to Ibn Shihäb [al-Zuhrl], and 5 percent to Ibn Täwüs. Assertions by
Ibn Jurayj himself amount to only I percent, and the remaining 37 percent is spread
among 103 persons.

I I The three "atypical" books had been left out. 13 The calculations are based on the sample men-
12 See Schacht, Origins, pp. 163 ff. tioned on p. 2 of this article.
 JounNnr oF NBen EasrnnN Srunrrs

3. In the material said to have been received from ath-Thawri, his own legal opinions
dominate, representing 19 percent of the total, followed by texts from Mansür [b.
al-Mu'tamir] (7 percent) and Jäbir [b. Yaz-rd] (6 percent), and from anonymous per-
sons (3 percent). The remaining 65 percent is said to come from 161 different authori-
ties or informants.

4. The texts put under the name of Ibn 'Uyayna consist of up to 23 percent of
transmissions from 'A,mr b. DTnär; 9 percent are said to come from Ibn AbI Najlh, 8
percent from Yahyä b. Sa'Id [al-Ansär1], 6 percent from IsmäcTl b. AbT Khälid; 3 to 4
percent of the texts are anonymous, and the remainder (50 percent) is said to come
from 37 persons. There is no ra'y (opinion) of lbn 'Uyayna himself.

These profiles indicate that each of these four collections of texts has quite an
individual character. It seems very improbable that a forger arranging material in a
specific order and labeling them falsely would have produced such highly divergent
collections. Besides, we have to bear in mind that these profiles are no more than
coarse grids and that differences emerge the more we go into details and ask, for
instance, about the geographic origins of the authorities or informants, formal pecu-
liarities of the texts, etc. The investigation of the transmission structures of 'Abd al-
Razzäq's Mu;annaf leads, therefore, to the conclusion that the materials he places
under the name of his four main authorities are genuine sources, not the result of
fictitious attributions which he has invented himself.
There are several other formal features of 'Abd al-Razzäq's presentation of trans-
missions that indicate that they are authentic. One of these is the fact that he is
sometimes uncertain about the precise origin of a tradition and that he admits this
openly. In one case, for instance, a tradition is introduced by: "1A,bd al-Razzäq from
ath-ThawrT from Mughlra or someone else-Abü Bakr [i.e., i{bd al-Razzäq] was
uncertain about it-from Ibrählm, who said: An actual forger would surely
not express such doubts, since it would undermine his main purpose, that of forging a
definite and uninterrupted transmission from an acknowledged authority. Furthermore,
'Abd al-Razzäq gives the impression that he received thousands of texts directly from
Ibn Jurayj, ath-Thawrr, and Ma'mar. This could be untrue, but if so, we may ask why
we also find, asdnld such as *'Abd al-Razzdqfrom ath-Thawrr from Ibn Jurayj
or-more rarely-"1A.bd al-Razzdqfrom Ibn Jurayj from ath-Thawrl . . . ,"tu or *'Abd
al-Razzdqfrom ath-Thawrr from Ma'mar The fact that there are also indirect
transmissions from his main authorities supports my argument still further. The origin
of his material is not arbitrary; but he specifically labels the source the tradition comes
from.
Forgery seems still more unlikely because there are also anonymous transmissions
by 'Abd al-Razzaq from authorities for whom he cites, in most cases, one of his main
informants as a source. Two examples state *'Abd al-Razzäq from a Medinese shaykh
who said: I heard Ibn Shihäb report from . . ." or "'Abd al-Razzäq from a man (rajul)

14'Abd al-Razzäq, Musannaf, vol.6, no. 11825. l6 Ibid., vol. 6, no. 10984.
15Ibid., tro. 11682; vol. 7, nos. 12631, 13020, and l7 Ibid., oo. 10798.
t3607 .
 THn MuseNN.tr oF i{r-R nfrp Ar-SaN'ANt

from Hammäd from Such asdnld are strange because, in general, 'Abd al-
Razzaq receives Ibn Shiheb's traditions from Ibn Jurayj or Ma'mar and Hammäd's
material from ath-Thawrr or Ma'mar.
Let us turn to the biographical literature; as noted above, this material requires a
separate treatment for methodological reasons, since the authenticity of the biographi-
cal traditions is as controversial as that of t}ire akddTth and early legal texts. According
to the biographical literature, at the age of eighteen, 'Abd al-Razzäq attended the
lectures of the Meccan scholar Ibn Jurayj (d. 1501767) when the latter visited the
Yemen, probably in l44l76l-62.re Ma'mar b. Räshid (d. 1531770) is said to have been
tAbd al-Razzäq's most important teacher. Basran by origin, he later lived at $an'ä',
fhe birthplace of lAbd al-Razzdq. He studied seven to eight years with Ma'mar,
probably from 1451762-63 until his death in 1531770.20 The Kufan Sufyän ath-Thawrr
(d. 16l/ 778) was in Yemen in the year 1491766,2t and the Meccan scholar Sufyän b.
'Uyayna (d. l9S/814) was there in 1501767 and,1521769." It is quite likely that on
these occasions lAbd al-Razzdq received the bulk of the material transmitted from
these authorities. The statements in the biographical literature about 'Abd al-Razzäq's
teachers thus coincides broadly with our findings from the Musannaf itself, the main
sources of his work.
In addition, it is important to keep in mind that these four, fAbd al-Razzäq's most
important teachers, are numbered among the first authors of works of a similar type.
They are regarded as the pioneers of mu;annafliterature. Ibn Jurayj, probably one of
the first mu;annaf authors, is said to have compiled a book called Kitäb al-Sunan;
ath-Thawrl, the al-Jdmi' al-kabtr and al-Jami' al-;aghtr; and Ibn 'Uyayna, the Kitdb
al-Jawämi'f t l-sunan wa-l-abwdb." There are no titles of Ma'mar's books preserved
in biographical or bibliographical literature, as far as I know. All these works seem to
be lost, but it is obvious that they must have been the sources from which 'Abd
al-Razzdq compiled his Mu;annaf. The fact that the author of the Kitdb al-JAmi',
which is attached to tlrre Mu;annaf, is not lAbd al-Razzdq himself but without doubt
his teacher Ma'mar further supports my argument.
This evidence leads to the conclusion that the the bulk of the Mu;annaf of 'Abd
al-Razzaq is a compilation of texts from older works of varying size, which can be

18 Ibid., vol. 7, nos. 12795 and 13622. 22lbn Sa'd, Taboqdt, vol. 5, p. 365.
re Ibn AbT Fätim, Taqdimat al-ma'rifa li-kilab 23 See Ibn al-NadTm, Fihrist (Cairo, 1348), pp.
al-jarlt wa-t-tacdTl (Hyderabad, l37l I 1952), pp. 52 f .; 315-16. According to him, Ibn'Uyayna did not
Ibn Hajar al-cAsqalänT, TahdhTb al-tahdhrä (Hyder- have a book but, rather, that people could only hear
abad, 1325-27), vol.6, pp. 311-12. Adh-DhahabT, him speak. This does not necessarily mean that he
MTzan, vol. 2, p. 127 . did not write his transmissions down, but only that
20 See Ibn AbT
Fätim, Kilab al-Jarlt wo-t-tacdTl he did not use a book in his lectures andlor did not
(Hyderabad, l37l-7311952-53), vol. 3, p. 38; Adh- place a book at his pupil's disposal for copying.
Dhahabl , Tadhkirot al-l.tuffA7 (Hyderabad, 137 5), There are works ascribed to him which, therefore,
vol. 1 , p.364; idem, MTzan, vol.2, p. 126 (instead of must be records of his lectures made by his pupils: a
'Umar it has to be read Ma'mar). TafsTr (see Ibn al-NadTm, Fihrist, p. 316) and a
2l See Ibn Sa'd, Kilab al-labaqdt al-kobTr, ed. Kitab al-Jawamf fT l-sunan wq-l-abwfib (see Abü
E. Sachau et al. (Leiden, 1905-17), vol. 5, p. 365 Tälib al-Makkt, Qüt ol-qulüb [Cairo, 1961], vol. 1,
(biography of Ibn 'Uyayna; source of information is p. 324 and Sezgin, Bufuärthm kaynaklan, p. 42).
Ibn'Uyayna); Adh-Dhahabt, Tadhkira, vol. l, For Matmar's Jami', see Sezgin, "Hodis musonne-
p. 346 (biography of Hishäm b. Yüsuf; source: fotmtn mebdei ve Ma'mer b. RASid'in 'Chmi'i',"
Ibrählm b. Müsa); Ibn $ajar, Tahdhlb, vol.6, Türkiyat 12 (1955): 1t5-34.
pp. 3l I and 313.
 JounNnr oF Nrnn EesrBnN Sruores

reconstructed, at least partly, from the asänid of the texts. 'Abd al-Razzäq acquired
his four main sources between the years 144 I 671 and 153 I 770. They were composed in
the course of the first half of the second century ,q,.H. and are among the oldest known
compilation s of altddlth and legal texts of relatively large size.

III
The question now is whether we can prove the authenticity of the materials con-
tained in lAbd al-Razzäq's principal sources. This has to be decided for each source
separately. For a trial solution, however, I have chosen the transmission of the Meccan
scholar Ibn Jurayj which covers about one third of the entire Mu;annaf.2a From these
more than 5,000 traditions, a representative selection of about 20 percent was ex-
amined.25 The following conclusions are based on this sample.
According to the information Ibn Jurayj gives about the origin of his material, it is
distributed among different authorities in a rather striking way. The largest part, about
39 percent, is said to come from 'Ata' b. AbI Rabäh. The following five scholars taken
together compose 25 percent: 'Amr b. DTnär (7 percent), Ibn Shihäb (6 percent), Ibn
fäwus (5 percent), Abü l-Zubayr (4 percent), and lAbd al-Karrm (3 percent). The next
five together constitute only 8 percent: Hishäm b. 'Urwa and Yahyä b. Sa'r-d (2per-
cent each), Ibn AbT Mulayka, Müsä b. 'Uqba, and 'Amr b. Shu'ayb (between 1.5 and
I percent each). Another group of ten names compose 7 percent, each only between I
and 0.5 percent. The remaining 20 percent.come from 86 persons, each with very few
texts. Ibn Jurayj's own legal opinions are rare (l percent).
This strange distribution of authorities in the work of Ibn Jurayj argues-in my
view-against the suggestion, which cannot be ruled out a priori, that he is a forger
who projects his own ra'y and the accepted legal opinions and practices in Mecca
during his lifetime backwards onto the preceding generation of scholars. Why would
he have made the work of forgery so difficult for himself ? Is it not more plausible to
expect that a forger would mention only one, or at most a few, of the most famous
older fuqahd', and these more or less with the same frequency? Why would he have
run the risk of having the entire forgery detected by introducing a host of additional
minor informants?
There may be still another reasonable interpretation for the varied distribution of
Ibn Jurayj's authorities: the Meccanfaqth'Atä' b. AbT Rabäh (d. 115/733) could have
been the teacher of Ibn Jurayj over a longer period of time. Since he was the oldest of
his more important authorities-this conclusion arrived at by their dates of death-I
believe that he probably was his first teacher. After !Atä"s death, or even during his
lifetime, Ibn Jurayj may have heard the lectures of other-somewhat younger-schol-
ars of Mecca, such as 'Amr b. DTnär and Abü l-Zubayr. He may also have sought'ilm
from scholars who lived not permanently in his city, such as the Medinese Ibn Shihab
and others, or with whom he may have come in contact while they were in Mecca for
the l.tajj. He may have traveled to them or have acquired copies of their lectures from
their pupils. In my opinion, the high number of sporadic authorities and informants

2a See p. 2 above. 10243 -14053 (kilab al-nikdl.t, kilab al- laldq).

25 'Abd al-Razzäq, Mu;annaf, vols. 6-7, nos.
 TnE Musnwx.tp oF 'Ar-R tzzT.Q AI-QaN'ANI

can be explained by his living in Mecca, which as the place of the pilgrimage gave him
many opportunities to meet scholars from all over the Islamic world, and this picture
coincides with what we read in the biographical literature.
If Ibn Jurayj had been a forger who ascribed his texts more or less arbitrarily to
certain older authorities, we would expect that the materials summarized under the
different narnes woüld not be essentially different from each other, at least formally.
But if one studies the transmissions from his 2l most frequently cited authorities and
informants-these amount to 79 percent of the entire source-it becomes clear that
the difrerences are so significant that we have to regard them as coming from distinct
and different sources. The divergences in these groups of transmission ascribed by Ibn
Jurayj to different individuals can be observed on several levels.

L The proportion of ra'y to traditions in the said sources or in the texts of their
principal authority varies substantially. The ratio of ra'y is, for example, 80 percent in
the material of !Ata' b. AbT Rabah, 85 percent in Ibn Täwüs from Täwüs,42 percent
in Ibn Shihäb, 42 percent in 'Amr b. DTnär, 40 percent in Ibn 'Urwa from 'Urwa b.
tl-Ztbayr,30 percent in Yahyä b. Sa'Id from Ibn al-Musayyab, and 3l percent in
i{bd al-Kartm. Others such as lAmr b. Shu'.ayb, Sulayman b. Müsä, Ibn AbT Mulayka,
and Müsä b. 'Uqba, rarely or never cite their own legal opinions.

2. There are remarkable variations, too, if we look at the relationship between Ibn
Jurayj's informant and the main authority of that informant and the number of
accounts transmitted from him. In some cases, the relationship was that of a pupil to
his teacher, as in the cases of 'Atä' b. AbI Rabäh and Ibn 'Abbäs, 'Amr b. DTnär and
Abü l-Sha'thä', Abü l-Ztbayr and Jäbir b. :Abd Alläh, Yal.ryä b. Sa'Id and Ibn
al-Musayyab, and Müse b. 'Uqba and Näfi'. But there are also other relationships,
such as the transmission by a son from his father, as in the cases of Ibn Täwüs and
!äwüs, Hishäm b. 'Urwa and 'Urwa b. al-Zubayr, and Ja'far b. Muhammad and
Muharnmad b. 'AlI. There are traditions of a mawla which come from his patron, as
in the case of Näfi' and Ibn 'Umar. Some of these pairs are almost exclusive, that is,
the younger informant transmits only material from the respective master or father
and from nobody else; this is the case with Ibn Täwüs and Täwüs, Ibn 'Urwa and
cUrpa, Müsä b.
'Uqba and Nafi', and Ja'far b. Mulrammad and Muhammad b. 'AlI.
Others rely more or less heavily, but not exclusively, on their most important teacher,
for example, 1{!A' b. AbI RabäI.r, 'Amr b. Dtnär, Abü l-Zubayr, Yabyä b. Sa'Td, and
Ayyüb b. AbT Tamima. In addition, there are sources where such relationships of
pupil/teacher or son/father do not dominate the transmission but in which we find
either many different authorities-as in the case of Ibn Shihäb, Sulaymän b. Müsä,
ahd others-or a choice which focuses on a certain region or on a certain group of
authorities, a phenomenon that can be obserwd, for instance, in the cases of 'Abd
al-Karrm, 'Ate' al-Khuräsän! lAmr b. Shu'ayb, and Ibn AbT Mulayka.

3. Ibn Jurayj's sources vary considerably in their proportions of traditions from the
Prophet, the ;aftdba, and the tdbi'ün. Only one transmission, that of 'Amr b. Shu'ayb,
contains mainly prophetic aftAdfth.In other collections, the proportion of this type of
transmission oscillates between 20 and 30 percent, as, for example, in that of :A.ta' b.
 JounNlr oF Nnan EasrnnN Sruorrs
AbI Rabälr, Abü l-Zubayr, Ibn AbI Mulayka, Ibn ShihAb, Hishäm b. .Urwa, and
1{tä' al-Khuräsänl Some have only a few or no prophetic traditions at all, such as
1A,mr b. Drnar, Ibn Täwüs, Yal.rya b. Sa'-rd, Müsä b. 'Uqba, 1A,bd al-KarTm, and
Näfi'. High proportions of sal.tdba traditions can be found in the works of !Atä'b.
AbT Rabäh, Abü l-Zubayr, Ibn AbI Mulayka, Müsä b. 'Uqba, Näfi., Yahyä b. Sa.rd,
'Abd al-KarIm, and 1\tär al-Khuräsänl; the proportions are only between 35 and 45
percent with 'A,mr b. DTnär, Ibn Shihab, and Hishäm b. .Urwa, and there is a
remarkably low percentage in 'Amr b. Shu'ayb's and Ibn fäwüs's work, the trans-
mission of the latter containing mainly tdbi'ün material. Regarding Ibn Jurayj's other
authorities, the texts from tdbi'ün reach only a ratio of 30 to 40 percent, as in the case
of 'Amr b. DTnär, Hishäm b. 'Urwa, Yalryä b. Sa'-rd, and lAbd al-Karr-m. Many fewer
tabicün texts are found in the collections of Ibn ShihAb, Abü l-Zubayr, 'A1ä, b. AbI
Rabäh, Ibn AbI Mulayka, and 'Amr b. Shu'ayb, and none are found in the Müsä b.
'Uqba, Nefi', and !Atä' al-KhurAsAnl collections.

4. The use of lhe isnad, or the mentioning of informants for traditions, varies in the
several sources of Ibn Jurayj. Asdnld from !Atä' b. AbI Rabäh and Ibn fäwüs occur
very rarely, the occurrence of asdnTd in the transmissions from Ibn AbI Mulayka,
'Amr b. Shu'ayb, "Abd al-Karlm, and !A1a' al-Khuräsänl is under 50 percent. They
are frequent, however, in the materials of the Medinese, such as Ibn Shiheb, Hishäm
b. 'Urwa, YalryA b. Sa'rd, and Müsä b. 'Uqba, but also in that of the Meccans lAmr
b. DTnär and Abü l-Znbayr; one of these generally exhibits some Medinan influences,
and the other is known to have Medinan origin.

5. Considerable differences are to be observed when checking the terminology of

transmission, that is, how Ibn Jurayj quotes his sources. The use of the word ",an,"
for instance, varies between never in the case of Ibn AbI Mulayka and 60 to 80 percent
in the transmissions of Yahyä b. Sa'-rd, Müsä b. 'Uqba, and 1A,mr b. Shu'ayb.
Between these extremes lie transmissions with relatively few 'an traditions, such as
those of Abü l-Zubayr and 'Amr b. Drnär, and others that show a frequency between
30 and 45 percent, such as those of Hishem b. 'Urwa, Ibn Shihäb, Ibn Tawüs, :A,tä'
b. AbI Rabäh, and 'Abd al-KarTm. There are similar fluctuations in the use of the
formula "sami'ttt." With some of his authorities, Ibn Jurayj does not use it at all; with
others, he uses it sporadically. Sometimes, however, it appears frequently, as in the
traditions of Ibn AbI Mulayka. Similar preferences for certain terms of transmission
are to be found in the work of some of Ibn Jurayj's authorities too, for example, the
nearly exclusive use of sami'tu by Abü l-Zubay.r. The heterogenous structure of trans-
mission speaks, in the final analysis, against the suggestion that it might be possible to
decide, on the basis of the transmission terminology, the question of whether the
material was received in oral or written form. In the case of Ibn Jurayj's transmission,
such conclusions are generally not safe, except in the odd case, such as that of Mujahid.
The above are the five most important levels by which the differences among the
several chains of transmissions can be formally described. They indicate that each
source has an individual character. This clearly contradicts the assumption that Ibn
Jurayj produced all the characteristic variations, fabricated the texts himself, projected
them back on older authorities, or forged the transmission chains or parts of them.
Such a diversity can hardly be the result of systematic forgery, but, rather, must have
 Tsr MuseNtt,ap oF 'Ar-R azÄp Ar-S.qN'ANt

developed over the course of time. We must, therefore-until the contrary is proven-
start from the assumption that the traditions for which Ibn Jurayj expressly states a
person as his source really came from that informant, and thus Ibn Jurayj's trans-
mission, in my opinion, should be regarded as authentic.
There are common explanations adopted by critics to reject the authenticity of a
transmission in such a case. For example, it could be maintained that the respective
transmitter-in our case Ibn Jurayj-was not the actual forger, or only the partial
forger, but that the forgery was the work of others, his contemporaries perhaps, those
from whom he actually took the material and then called it his own. Another explana-
tion might be that a later author arbitrarily used his name. These are the sort of
arguments Schacht proposed: "The bulk of the traditions which go under his [Näfi"s]
name must be credited to anonymous traditionists in the first half of the second
century A.H."26 But suggesting as an explanation for the contradictions in a transmis-
sion that instead of a single transmitter named in the text that many anonymous
forgers must have been at work cannot be accepted as a scholarly sound argument
because it shifts the problem from the level of facts, which can be checked, to the
sphere of speculation. I do not argue against the possibility that there were forgers of
aftddTth and asanTd in the first and second centuries of Islam. It is indeed one of the
most important tasks of the historian to detect if in fact texts and transmission chains
were fabricated, and if so, where, how, and why it was done. Schacht himself pointed
to the fact, already well known to Muslim l.tadlth critics, that the asänld of later
collections are much better and more complete than those of the older ones. This is
one possible point of departure in the attempt to unmask forgeries of and improve-
ments on asdntd and their authors. But the mere fact that al.tddTth and asantd were
forged must not lead us to conclude that all of them are fictitious or that the genuine
and the spurious cannot be distinguished with some degree of certainty.
The study of one chain of transmission in an early collection of traditions, i.e., the
material of Ibn Jurayj in the Mu;annaf of 'Abd al-Razzäq, shows that it is indeed
possible to separate trustworthy from suspicious traditions or texts of uncertain relia-
bility. A comparison of this early state of transmission (first half of the second century
,r.H.) with that of the more recent collections of the second half of the third century
and later could certainly give insights into the extent of fabrication, the forgers, and
their motives. This is ceitainty a topic for future research.2T

IV

The reliability of Ibn Jurayj and the aut'henticity of *ris-sources can be further
examined. To demonstrate, I have chosen the largest of his sources, i.e., his trans-
mission from lA.tä' b. AbT Rabäh. At first, it is striking that this material consists of

26 Schacht, Origins, p. 179 and passim. bridge, l9S3). His research has produced many
27 G. H. A. Juynboll has recently tackled the valuable results, especially concerning the extension
question of the authenticity of the ltodTth anew. He and the techniques of isndd falsification, partly
has concentrated on the biographical material, espe- known even to the Muslim scholars themselves. But
cially that of Ibn $ajar, and the akAdfth of the he has treated the early mu$annaf works such as
Prophet as preserved in the classical and other that of 'Abd al-Razzäq and Ibn AbI Shayba quite
collections of the third century A.H. and later. See harshly. In my opinion, they offer us many interest-
his Muslim Tradition: Studies in Chronology, Pro- ing new ideas, as will be argued in this article.
venonce ond Authorship of Early Hadnh (Cam-
 r0 Jounxer oF NEan EasrnnN Sruorns
texts of two different genres which appear with nearly the same frequency. One half of
the texts may be classified ut ,"tpoäo, the other half as dicta. By responsa, I mean
answers of 1A.ta' on questions of Ibn Jurayj himself or anyone else, named or un-
named. Dicta are defined as statements of 'AtA' which are not preceded by a question.
Both genres contain opinions (ra'y) of iA.tä' himself or traditions from others (ahddith,
dthdr).
By far the largest number of the responsa consist of the answers of .Atä, to Ibn
Jurayj's own questions. Anonymous questions do not amount to l0 percent, and those
from other named persons are very rare. The responsa are dominated by cAtä"s own
ra'y, whereas the traditions have a frequency in this genre of only l0 percent. Among
the dicta, the difference is not so marked; here the proportion of ra'y to traditions ii
70 to 30 percent. comparing the ratio of these two principal genres in .Alä', 50:50, to
the material of other important authorities of Ibn Jurayj, a remarkable result emerges:
among the texts of 1A,mr b. Drnar, the proportion of responsa is only 9 percent
(exclusively on questions of Ibn Jurayj); among that of Ibn Shihab about 14 p".cent
(here only 1.5 percent on questions of Ibn Jurayj!); from Ibn Täwüs,5.5 percint are
reported; from lAbd al-Karrm, 8 percent responsa (all on questions of Ibn Jurayj); and
from Abü l-Zubayr, there is not even one responsum.
How does the study of genres contribute to the question of text authenticity? The
mere fact that the two main genres are distributed in the sources of Ibn Jurayj in such
a different way seems to contradict the possibility of a systematic projection back-
wards upon the preceding generation of scholars. If this were so, we would expect
more uniformity in the way they were forged. The same is true for the varied frequJncy
of types of questions among the responsa which Ibn Jurayj transmits from .Ata;. ca;
we explain the different kinds of questions-the direct, the indirect, the anonymous,
and those from named authorities other than 'A,tä' as mere stylistic devices which Ibn
Jurayj had used according to the principle of variatio delectat|
The pattern of question and answer implies a strong claim of authenticity, inasmuch
as the question is asked by the transmitter, or pupil himself, and is immediately
answered by the respective authority, or teacher. With his question, the questioner has
a share, in some way, in the answer (as the actual instigator of it). Tlie directness of
transmission can hardly be expressed more strongly. Formulas such as ,,sami.tuhü
yaqül," "qkhbarant," or "qälo /t" which indicate direct oral transmission as well
(which does not exclude their being written doqq), clearly suggest less reliability, not
to mention the totally uninvolved "'an x qala."28 If one supposes, on the strength of
the many direct questions of Ibn Jurayj to :Ata', that he intended to simulate the
highest degree of authenticity, how can the following two introductions be explained?
Ibn Jurayj said: "I charged someone to ask fAtä' about . . . , because I could not hear
him," or: "I sent someone to .AtAt with the question about. ..?"2e Why would he
have invented, in addition to the many direct questions, several anonymous ones,
questions which are generally taken to be less authentic because they identify the
transmitter only in a passive and not active role? why, then, does Ibn Jurayj transmit
28 In this context, I interpret these
words with were systematically followed in earlier times.
their normal meaning, and I do not rely on the rules zs %,bd al-Raziaq, Musannaf, vol. 6, no. 10g25;
connected with these terms by the later "science of vol. 7, no. 13g93.
ftadTth," since it cannot be assumed that these rules
 THp Mustxxtp oF lAr-RazzTq AI-SIN'ANT ll
in addition to the responsa so many dicta from'Atä', two-thirds of them with the
simple formula 'an A1a' qdla"? Those who propose this theory of projection or
forgery based on this type of transmission must examine the question further. I pro-
pose that the study of the genres argues against the hypothesis of outright forgery.
In addition to the two extrinsic formal criteria of authenticity, those of extension
and of genres, there are further indicators which suggest that Ibn Jurayj's transmission
from 1A.tä' was authentic. I call them intrinsic formal criteria of authenticity because
they are based on an investigation of how Ibn Jurayj presents 'AtA"s material. The
leading questions thereby were whether we can recognize a distinctive profile of Ibn
Jurayj differing from that of 'Atä' and whether there are critical remarks of his about
the opinions of his teacher, or other formal hints which are not in agreement with the
supposition of back-projection and fabrication.
As intrinsic formal criteria of authenticity, I propose six types of material: Ibn
Jurayj's own legal opinions, his commentaries on texts of 'A1ä', indirect transmissions
frorn 'Atä', expressions of uncertainty by Ibn Jurayj, the existence of variants, and the
reporting of 1{te"s deficiencies.

l. %.bd al-Razzaqtransmits from Ibn Jurayj not only legal opinions which the latter
ascribes to earlier authorities, but also his own ra'y. lf one suggests that Ibn Jurayj
was a forger who projected his own legal opinions back upon older authorities with
the intention of giving them more weight, one has to find a convincing explanation for
the fact that there are legal statements of Ibn Jurayj himself which are not attributed
to earlier scholars.

2. That the hypothesis of back-projection is untenable is further evident if we turn to

the commentaries which Ibn Jurayj provides for some of 'A1e"s transmissions. They
can be classified as additions, which explain or elaborate, or oppositions. Ibn Jurayj
obviously added both types of remarks to the texts later. It is not plausible to suggest
that Ibn Jurayj first invented the texts, then falsely attributed them to 'Atä', and at
the same time, or a later time, embellished them with commentaries and criticisms. But
it seems not too far-fetched to suppose that Ibn Jurayj, when he heard the teachings of
his master as a young pupil, did not have the competence and self-confidence to
complete or criticize them, a competence he surely acquired later.

3. Judging from the amount of transmitted material, :Ata' is clearly Ibn Jurayj's
principal authority. If the authorship of 'A1ä' were completely or partly fictitious, we
would not expect that Ibn Jurayj would also report opinions as having beel received
indirectly, i.e., through a third p"ttott. Such traditions do exist nevertheless'3o

4. Sometimes Ibn Jurayj points out that he is uncertain about the very wording or
meaning-of nta"s sayings.3t This confession of doubt must be seen as proof of his
truthfulnäss and as his intention to reproduce the teachings of his master as exactly AS

possible.

30 Ibid., vol. 6, nos. I 1080, I 1348, and I 1460; vol. 3l Ibid., vol. 7, nos. 13138 and 12835'
7, nos. 12553, 12571,13t21, and 14001.
 12 JounNer or Nr,tn Elsrnnu Sruorns
5. The attempt to relate precisely and word-for-word can also be observed in the cases
where Ibn Jurayj records variants of:A,tä' in a tradition he heard from him at a
different time or which he heard both fiom him and another informant. The dis-
:t"p"?ntl may be only slight but can also be accompanied by a real change of mean-
tng."' I hose cases where Ibn Jurayj preserves different versions of the same subject
from 1A.tä' himself are especially difficult to harmonize with the supposition thai he
also attributed texts falsely to 1A.ta'. If this were so, the contradictions in his own
fabrications would have been realized by Ibn Jurayj. Besides, he adds notes to several
of 'Atä"s legal opinions that this was also a view held by a certain "Companion of the
Prophet" or by a caliph. Normally, he states this clearly as his own note, without
qugtine any informant for it. A forger would hardly have resisted the temptation to
claim 1{tä"s authority for it.

6' The importance of 'Atä"s teachings for him notwithstanding, Ibn Jurayj does not
always let him appear as a legal scholar without fault as one might .*p""1 from a
forger who falsely attributes his own ideas or commonly heard traditions to a great
master of the past. Lacking a better term, I call this "hints at deficiencies of :Atä'.-
Among them are ignorance, uncertainty, change of opinion, and contradictions.33

All the extrinsic and intrinsic formal criteria mentioned argue in favor of the authen-
ticity of the 'Atä"s transmission as preserved by his pupil Ibn Jurayj and contained in
the Mu;annaf of 'Abd al-Razzdq. This material genuinely appears to belong to .Atä',
who must have been one of Ibn Jurayj's most important teachers, a conclusion which
is also substantiated in the biographical literature. Ibn Jurayj usually differentiates
precisely between %.ta"s statements, those of other authorities, and his own opinion,
and he does not hesitate to deviate from the teachings of his master. In this trans-
mission, we are surely not dealing with conscious back-projections or spurious attribu-
tions. In my opinion, his work can be considered a historically reliable source for the
state of legal development at Mecca in the first decade of the second century A.H.

141ä', who died 1151733, belongs mainly to the generation of the last quarter of the
first Islamic century called at-tabi'ün. He is therefore a connecting link between Ibn
Jurayj and the generation of the ;al.tdba, i.e., the companionr of th. prophet. The
question is now whether in 'A1a"s materials genuine traditions of the first ."ntu.y a.H.
are found and how we can ascertain that fact.
First we must note that the proportion of traditions (aftadfth, dthdr) in the body of
texts from 'Ata' is not very high. Only 20 percent of it contain traditions, as against
80 percent of his pure ra'y.ln the genre of the responsa, this imbalance is even gÄater:
8 to 92 percent. This can be taken as an indicator ofthe fact that traditions coiveying

32Ibid., vol. 6, no. 10532; vol. 7, nos. 13650-51, 13655, t4001, and 14030; vol. 6, no. 10780; vol. 7,
13107, 13108, l3l 10, 13217 , and 13220; vol. 6, nos. nos. 11954,11966,11680, 12974, and 13391; vol. 6,
10962 (109 t9, 10951), 10969. nos. 11620 (l 1603, I l6l8), t 1627 (1 l6l0); and I 1648.
33 Ibid., vol. 6, no. 11522; vol. 7, nos. 12658,
 THB Mus,txxnr oF iA'r-R d,zzfie AI-SaN'ANT 13

opinions and practices of others played only a minor role in his legal teaching. The
conclusion that there were in his time only a small number of traditions or that he did
not know more than that, would, however be incorrect and can be easily disproven by
the texts. Even if ra'y dominates in his teaching (indeed, precisely because of it), the
fact that he sometimes relies explicitly upon earlier authorities must not be over-
looked. The hierarchy of his authorities according to the frequency of their mention is:
(l) Companions of the Prophet (15 percent), (2) the Qur'än (10 percent), (3) the
Prophet (5 percent), (4) anonymous traditions (3 percent), and (5) contemporaries of
'Atä' (1.5 percent).34 In the following, I confine myself to examining in somewhat
more detail two of these "legal sources" which 1{tä' sometimes quotes: the Com-
panions and the Prophet.

THE COMPANIONS OF THE PROPHET

First of all, a formal matter springs to mind: 'A,ta"s citing of the Companions in his
responsa have, as a rule, no isndd and are extremely short. Partly, they are mere
references which assume either personal contact with the person mentioned or the
knowledge of a more detailed tradition about him. In the genre of the dicta, on the
other hand, longer traditions and sometimes even asdnTd also appear.
Among the ;al.tdba,'Ate' most frequently quotes Ibn 'Abbäs. Sometimes he says
expressly that he had heard a statement from him, sometimes not. Regarding the
authenticity of lAta"s transmissions from Ibn 'Abbäs, note the following points:

1. In'Atä"s responsa, references to Ibn'Abbäs are very rare (a little more than2
percent), and they are, in this genre, mainly of supplementary value, serving merely as
confirmation of !Atä"s opinion but without value of their own. Obviously 'AtA' did
not, as a rule, attempt to give his own legal opinions more weight by referring to the
authority of an Ibn iA,bbäs or any other Companion.

2. Although, in most cases, 'Ata' quotes Ibn 'Abbäs directly, sometimes even with
sami'tu, there are also indirect references."

3. In some texts, he refers to him not to confirm something, but rather to contradict
35
nlm.

None of these are the usual methods that would be chosen by a forger who claims to
have heard a great master and who ascribes his own opinions to him.

4. Whereas most of :Atä"s transmissions from Ibn 'Abbäs contain simple legal dicta,
there are a few texts of quite another style and content. I would call them stories
(qi;a;).In them, 'Atä' presents himself as a pupil of Ibn 'Abbäs.3? Criteria of content
point to ihe authenticity of these reports.

3a The basis of calculation is the quantity of the 36 Ibid., no. I1747. For the opinion of Ibn 'Abbäs,
texts studied. seenos. 11767-69.
35 See 'Abd al-Razzaq, Mu;onnaf, vol. 6, no. 37 lbid., vol. 7, nos. 14021-22.
t t07 6.
 l4 JounNer oF Nnan E^qsrnnN Sruolns

5. In view of the great many al.tadlth of the Prophet that Ibn 'Abbäs is alleged to have
transmitted in the biographical literature (the usual number is 1,660),38 it is striking
that 'Atä', as a rule, does not quote such al.tddlth. In the sample of texts I studied, not
a single one was to be found.

All these and a few additional observations indicate that the transmission of 'Atä'
from Ibn 'Abbäs, preserved by Ibn Jurayj and passed on to 'Ahd al-Razzäq, is a
generally trustworthy one.
In addition to Ibn 'Abbäs, 'Afä' emphasizes only from a very few of his other
;al.taba a samä', from Abü Hurayra and Jäbir b. 'Abd Alläh for instance3e (cases
which are very rare and archaic in content). Others, on the other hand, he quotes
without samä'or indirectly, i.e., through an informant, although direct contact with
them was possible or even probable. From these examples, we can conclude that those
traditions of iA,tä' from sqhdba which he explicitly claims to have heard must-until
the contrary is proven-be taken as genuine.
References to 'Umar b. al-Khattäb are the second largest group of 'A,tär's trans-
missions from sahäba. But as a group, they do not amount to even 3 percent of Ibn
Jurayj's materials from 'Atä'. If one classifies 'Umar's traditions, for example, it
appears that they belong mostly to genres which are connegted with his role as caliph:
legal judgments (aqf,iya),4o ordinances (interdictions, positive orders),ar legal answers
Qfatdwd) in which the authority of the caliph may have been sought,n' dicta which may
partly be the results of legal sentences or fatdwda3 and, rarely, acta, i.e., practices of 4
more private character.aa In this regard, 'IJmar's traditions differ greatly from those of
Ibn !A'bbäs, and this gives them an appearance of historical value. That cAtä) invented
them can surely be ruled out because they are so marginal in his legal teaching and are
not always accepted by him as legally binding. They were clearly already current
knowledge at his time, but where did 'Atä' get them from?
For most of his transmissions from 'Ilmar, 'Atä' does not quote any source.
Occasionally, he introduces them with the word "dhakarü" ("it was reported [to
me]').nt In a few cases, however, he names the informant from whom he "heard" the
tradition or an isndd reaching back to a witness who lived during 'Umar's time. 'Atä'
himself was born after 'Umar's death.a6 There are clues in the texts that 'Atä' actually
did acquire the traditions from the informants whom he mentions. This does not mean
that they are necessarily genuine, i.e., that they report the truth about 'Umar, but we
can at least be sure that they were in circulation during the lifetime of 'A,tä"s infor-
mant. Some of 'Ata"s traditions about 'Umar can, therefore, be dated with certainty
to before 80 or 70 l.n. We come to similar conclusions when studying cAtä"s refer-
ences to and traditions from 'A'isha and 'AlT.

38 See Ibn Hazm, "Asma' al-sahaba al-ruwAt wa- 4r Ibid., nos. 13508 and 13541.
md li kull wohid min al-'odad"; idem, Jawami' al- 42 Ibid., no. 13612.
sTro, ed. Ihsän cAbbäs and N. al-Asad (Cairo, n.d.), 43 lbid., vol. 6, no. 10726; vol. 7, nos. 12877 and
p.276. 12885.
39 See 'Abd al-Razzäq, Musannaf, vol. 7, nos. 44 Ibid., vol. 6, no. I t 140.
12566 and 13680. 45 lbid., vol. '7, no. 12877.
40 Ibid., nos. 12401
, 12858, 12884, I 365 I , and 46 Ibid., nos. 14022, 13541, and 13612.
14021.
 Tnn Mu;eNxtr on l{r-RszzÄQ Ar-SeN'Alü 15

Frequently, especially in the responsa,'Atä' quotes only fragments of traditions

known in more detail from other sources. Usually, in these cases, we can safely assume
that he knew their complete versions. There is no hint, however, that the fuller ver-
sions are secondary and werp later exppnded from 1{ta"s short references. This can be
helpful for the dating of traditions in that if
in Ibn Jurayj's material from
there is
.A!ä' a referencg to or a short version of a tradition on $ahaba, 'Atä"s date of death
(1151733) is the terminus ante quem for the existence of the tradition in question.
One example which may demonstrate that this method is helpful for the study of
fuadTth is an unusually long tradition about the nursing of adults contained in both of
the two most important oithe several preservgd versiöns of Malik's Muwalla'.41 Itis
composed of sever4l oingle traditions: one of the Prophet with some additional infor-
mation, another about {'isha, and a third which conperns the other wives of the
Prophet. Because of its artificial cor4position, it does not $9em to fit into the frame-
wort of the usual traditions of Mälik. One is thus tempted to consider it a relatively
recent addition. But M4lik's isnddnames 'Urwa b. al-Zubayr (d. between 92l7ll ar1:d
l0l1720) as its author and Ibn Shihäb (d. 1241742) as the transmitter and his own
informant. According to Schacht's view of the legal developmerlt of the question, the
origin of the different parts of the story cannot be attributed to Ibn shihäb or some-
one from this period, and the appeal to 'Urwa must, in any case, be regarded as
spurious. Rather, Schacht sges here counter-traditions by the circle of "traditionists"
whose aim it was to change established doctrine.l8 If, on the other hand, we refer to
one of the responsa of 'A!ä' concerning the same topic, quite another picture of the
history of the question emerges. According to it, 'AtA'-who is certainly not to be
counted among the traditionists-already held the opinion that the suckling of adults
was legally valid, and, in this context, he refers to a practice of 'A'isha: '"kdnat A'isha
ta'muiu bi-dhAlikq bandti gkhlhd."4e This fs connecte{, without a doubt, to the more
detailed tradition of 'Urwa 4s preserved \nthe Muwq!!q'.It reads: "{'isha used this
[method] in cases of men she wished to let visit her. She used to command her sister
Umm Kulthüm bint AbT Bakr [. . .] and the daughters of her brother Qfa-kanu tdmunt
ukhtuhd (lmm Kulthüm l. . .) wa-bandti qkhthd) to nurse the men she wished to let
enter with her."4e"
The tradition concerning 'A'isha was, therefore, already known to 'Ata'. He and
Ibn Shihdb drew from the same source, since it is unlikely that 'Ate' attended lectures
of tfg younger Ibn Shihab. dccording to the latter, 'Llrwa b. al-Zubayr is the author
of the story. He was a slightly older contemporary of 'Ata' and was his informant for
other traditions. It thus seerns highly probable that 'Urwa is 'Atä"s source as well. If
so, then the tradition about 'A'isha as contained in the Muwa!1d must be considered

47 Maük b. Anas, Al-Muwa!!a', riwdyat Yahyä b. matik" (Der Islam 68 [1991]). The Arabic word
Yahyä, ed. M. F. 'Abd al-Bäql, 2 vols. (Cairo, used in these texts, arSa'a, really means "suckle," in
1370 I 195 I ), chap. 30, no. 12; idem, riwayat Mu- this special case probably by mixing breast milk with
[rammad ash-Shayb4nl, ed. 'Abd al-Wahhäb 'Abd drinks or food. In this way, a "milk relationship"
al-Lat-rf (Cairo, 1387 11967), no. 627. was created which had the same legal status as a
a8 See Schacht , Origins, pp. 48,246. blood relationship: "milk relatives" were forbidden
49 'Abd al-Razzdq, Musonnaf, vol. 7, no. 13883. to marry one another or to have sexual intercourse
49a$ss n. 47 above. For a more detailed discussion and could, therefore, associate with each other
of this tradition and its variants, see my forthcoming without restrictions.
article "Der fiqh des az-ZuhrT: Die Quellenproble-
t6 JounNar oF Nnnn Elsrnnx Sruuss

a genuine transmission from 'Urwa dating to the second half of the first century A.H.
and not to the middle of the following century.

THE PROPHET

In his responsa, 'A7-a' very rarely refers to the Prophet. Among the 200 responsa
studied, only three even hint at him. In addition, there are a few statements about the
Prophet that arise through Ibn Jurayj's questions. None of these texts have an isnad;
sometimes we find the introduction"balaghana anna n-nablf rasüla ltdh. . . . "50 The
proportion of transrnissions from the Prophet among 'Atä"s dicta is somewhat higher
(6 percent). Whereas in the responsa, only references to or fragments of altddrth are
found, most of the prophetic traditions among the dicta are complete and quite
detailed. Only a quafter of them have an isnäd although one which is not always
complete.
?tä' knew many more traditions about the Prophet than he actually used in his
legal argumentS. This is revealed in texts in which Ibn Jurayj-partly after an answer
containing only 'Atä"s opinion about a problem-expressly asks him about the
Prophet and then receives an answer which indicates that 'Atä' was well acquainted
wittr a prophetic tradition.sr 'A1ä' also cites legal principles, which he clearly recog-
nized as traditions from the Prophet but in which he does not refer to the Prophet
directly. One example of this is the legal statement "al-walad li-l-firdsh ws-li-l-'ähir
al-Ltajar" ("the child belongs the [marriage] bed and he who had illegal sexual inter-
course gets nothing")." 'Ala' makes use of this maxim on two occasions without
saying that it was considered a legal judgment of the Prophet.53 One responsum of his,
however, reveals that he knew that it was:
Ibn Jurayj said: I said to 'AtA': "what is your opinion [in the case] when he [the man] rejects
[the paternity of] it [the child] after she [the woman] has borne it?" ['Atä'] said: "[in that case]
he has to anathematize her (yula'inaha) and the child belongs to her." I said: "did not the
Prophet say: Al-walad li-l-firAsh wa-li-l-'ahir al-l.tajar"'? ['A,tA'] said: "Yes! But this was be-
cause the people in [the beginnings of] Islam claimed children born in the beds of [other] men as
theirs saying: 'They are ours'! [That is why] the Prophet said: Al-walad li-l-firdsh wa-li-l-'dhir
al-llajar'."54

Only through Ibn Jurayj's question are we informed that this legal maxim was not a
creation of 'A,ta' himself, but one which already was well known and attributed to the
Prophet at the turn of the first century l.s. This allows us to test with traditions from
the Prophet the rule we formulated when discussing the dating of traditions from
;altdba: with the help of 'AtA"s references to certain traditions or with his fragments
of traditions otherwise known, their asdnTd can be checked and their time of origin
defined more exactly. This will be demonstrated further below.

s0 Ibid., vol. 6, no. 10969; vol. 7, no. 12632. Zao--rdT, TAj al-'arüs (Cairo, 1306 I 1888), vol. 3,
sl Ibid., vol. 6, no. 10651. p. 127; al-Qastallänr, Irshad al-sarT ild sharh Sahfh
52 The Arabic dictionaries and the, commentaries al-Bukhdrl(Büläq, 130411886), vol. 4, p. 10.
of l.tadltft compilations rightly prefer this meaning 53 'Abd al-Razzdq, Mu$annaf, vol. 7, nos. 12381
of ol-I.tajar to the also possible rojm, "stoning." I and 12862.
thank J. Burton for this idea. See Ibn Manzür, 54 Ibid., no. 12369.
Lisdn al-'arab (Beirut, n.d.), vol. 4, p. 166; al-
 Tgg Mus,a,Nv,np oF iA.r-R d^zzTe Ar-SnN'ANr l7

Early detailed transmissions about the Prophet using this maxim, "al-walad li-l-
firdsh wa-li-l-'dhir al-I.tajar," are to be found in Mälik's Muwattd and in the Mu;an-
naf of 'Abd al-Razzäq. Different versions can be distinguished:

l. There are several variants of a story relating the quarrel between Sa'd b. AbT
Waqqäs and 'Abd b. Zam'a concerning who the real father (nasab) of a boy was.
They had, it was reported, appealed to the Prophet as a judge, and he made a decision
uttering this maxim.tt This I call the qi{;a version.

2. There is a short tradition containing only the dictum itself.56 All the early variants
of the qrypa version have an isnad ending with "Ibn Shihäb al-Zuhrl from 'Urwa
b. al-Zubayr from {,'isha." The short version sometimes has this isnad, and some-
times the following one: "Al-ZuhrI from Ibn al-Musayyab and Abü Salama from Abü
Hurayra."IbnShihab (d.1241742) isthe"commonlink"inallof thesetexts,if wedo
not count tAtat's references to it for the moment.
According to the procedure of dating with the aid of "common links," as done by
Schacht, the time of Ibn Shihäb would be the earliest point at which this complex of
traditions came into being.s? But since Schacht was convinced that there were exten-
sive forgeries of the chains of transmission, he held al-Zuhrt "hardly responsible for
the greater part of these traditions" from the Prophet, the Co_mpanions, and the
Sucessors in the asqnld of which he appears as a "common link."58 He thus places the
origin of such traditions in the second quarter of the second century or later. Schacht
felt that the above-mentioned maxim has to be dated to the second quarter of the
second century based on the fact that in the Kititb al-Umm of al-Shäfi'I (d.20418?0)
Abü Hanr-fa (d. 150| 767) is reported to have known it as a dictum of the Prophet.5e In
addition, Schacht quotes a text from the Kifib al-Aghanl of Abü l-Faraj al-IqbahänT
(d. 356196'l), used earlier by both Wellhausen and Goldziher, where a dispute about
the paternity of a child is reported that allegedly occurred "in the middle Umayyad
period." Since in this case the rule of the Prophet is neither mentioned nor followed,
Schacht concludes that "it had not yet asserted itself in the time of the dispute re-
corded in Aghani."6o It wur therefore obvious to him that the first century .1,.n. cannot
be accepted as a possible time of origin of this maxim. Thus the reference to the
Prophet must be regarded as historically untenable and as a clear forgery.
If we turn to Ibn Jurayj's and 'AtA"s references to this maxim of the Prophet, it
becomes clear that Schacht's chronology is incorrect. Since 'AtA' quotes the rule
several times, it is my opinion that it must have been widely known by the first decade
of the second century a.H. (i.e., the middle of the Umayyad period) at the latest. 'Ate'
obviously knew the qlppa version.ut We have already seen that he did not transmit

5s Ibid., nos. 13818 (Ma'mar from al-Zuhfr.); 56 'Abd al-Razzdq, Musannof, vol. 7 , no. 13821.
138 19 and 13824 (Ibn Jurayj from Ibn Shihab); 57 See Schacht , Origins, pp. 177 ff.
"AhodTth AbT l-Yamdn," no. I (Shu'ayb from al- 58 lbid. , p. 246.

ZuhrT); M. M. Azami, Studies in Early Hadith Lit' 5e Ibid., p. 182.

erature (Indianapolis, 1978), pt. 2 (Arabic texts), 60 lbid., p. l8l. Cf. Goldziher, Muhammedanische
pp. l4l f.; MAük, Muwatta' (Yafya), chap. 36, Studien, vol. l, p. 188, n. 2 and Azami, Studies,
no. 20 (Mälik from Ibn Shihäb). In most of the p.266.
qß;a versions, the second part of the maxim is 6l See 'Abd al-Razzaq, Musannof, vol. 7, no.
lacking; see also Azam|, Studies, p. 16l. The text is 12369.
also to be found in the " SahThAn."
18 JounN.q,r oF Nnen EnsrsnN Sruntrs

from the younger Ibn Shihdb but sometimes directly from .Urwa b. al-Zubayr.62
'Urwa is Ibn Shihäb's informant of the qlssa variants according to their assrud, so
there is evidence for the assumption that he was !Atä"s source as well. If so, the story
must have been in circulation by the second half of the first century A.H., since 'Urwa
died about the end of it. But the possibility cannot be ruled out that the tradition was
widely known earlier, asthe qsantd with the alleged authors 14,,isha (d.571676) and,
Abü Hurayra (d,. 591678) claim, and it is possible that the story has a kernel of truth
and that Mutrammad really made such a judgment.63
Schacht considered it improbable that the Prophet had anything to do with this
legal rule also for other reasons. But in my opinion, he was wrong here, too, as I will
demonstrate below. In his short discussion of our legal maxim backed up by systematic
and historical arguments, he adopts Goldziher's hypothesis that the alleged prophetic
dictum may have been taken from Roman Law, which has a similar rule: pater est
quem iustse nuptiae demonstrant The pre-Islamic Arabs decided disputes of paternity
in another way, by calling in "professional physiognomists" (qafa).From this, he
concludes that the maxim cannot be of Arabic origin and was therefore not current in
Arabia in Muhammad's time. Furthermore, he claims that this legal rule was "strictly
speaking incompatible with the Koran" and that the problems it should solve "could
hardly arise under the Koranic rule regarding'idda."6a He thus seems to conclude-
without actually saying it directly-that this dictum cannot have been the Prophet's.
These arguments are not convincing, however. Let us begin with his premise that the
legal maxim "al-walad li-l-firdsh" is incompatible with the Qur'än: Schacht suggests
that the disputes about the paternity of a child arose in cases where the waiting period
after the separation from the legitimate sexual partner was not correctly observed. But
this is only one possibility which, as he rightly points out, should not arise in Islam.
There is another much more important context for disputes over paternity: in cases of
illegal sexual intercourse with a married woman or a slave. The Qur'än struggles with
this question by issuing clear legal norms concerning marriage, divorce, and concu-
binage and by proscribing illegal sexual intercourse with heavy punishments in this
world and the world to come. The early texts describing the use of the legal maxim
"al-wqlad li-l-firdsh," i.e., the ql'ssa version and !A.1ä,'s responsaut reveal, nevertheless,
that in the early Islamic community there were particular social contexts where the
Qur'änic norms had not yet gained a firm footing. One of these problematic areas was
the relation between a master and his female slave, which even in Muhammad's
lifetime was unclear.66
This is the background of the disputes in which our legal maxim first emerges in the
texts: the possibility that a man who committed fornication with another man's wife
or slave would benefit from the child resulting from this illegitimate union was elimi-
nated. Moreover, many cases of adultery were prevented from becoming public be-

62 Ibid., ro. 13939. 6a Schacht, Origins, pp. 181 f.

63 Juynboll argues that the fact alone that this 65 'Abd al-Razzaq, Musannof, vol. 7, nos. 12369,
statement was ascribed to Ibn al-Musayyab has to 12381, 12529, and 12862.
be regarded as a proof that it cannot be older than 66 See my article " Wal-muhsanotu mina n-nisdi
that person; see Muslim Tradition, pp. l5 f. But this illa ma malokat aimanukum (Koran 4: 24) und die
is not convincing because he relies on limited data. koranische Sexualethik," Der Islam 63 (1986): 192-
His conclusion is purely hypothetical. 218, esp. 199 tr
 THE Muslxtt.tp oF !Ar-R nzTe Ar-SaN'ANt t9

cause both in the case of a man who claimed a child borne by the wife or the slave of
another man or in the case of a woman who affirmed that her child was not from her
husband or master, the claimant implicitly confessed to illegal sexual intercourse and
ran the risk of the punishment for it. 'Atä' limits the application of the legal maxim to
such cases where the paternity of a husband or female slave's master was not rejected
by the man himself but which was challenged by another party-thus presupposing
irregular or illegitimate sexual relations-and he justifies it by saying that the original
intention of the rule had been to stop such paternity disputes. 'Atä' dismisses the
pre-Islamic method of relying on the qdfa (physiognomists) who established paternity
by comparing the child with the contesting would-be fathers. He appears to maintain
that this method has been replaced by the "ql'walad li-l'firdsh" rule.
This maxim is, therefore, in congruity with the Qur'änic legislation concerning
marriage and family and with the mores of the early Islamic period even if it does not
fit so well with the Qur'änic tendency to insure the real paternity of a child. Is there a
legal system in the world wholly free from contradictions? We have to conclude that
even the Prophet may have used this legal maxim.
In locating the emergence of Islamic jurisprudence in Iraq at the beginning of the
second century A.H. at the earliest and the introduction of this legal maxim in the late
second half of the second century n.u., Schacht looked for Roman influences in
Islamic law, especially in Iraq, and he suggested late antique rhetoric as the channel.
But Patricia Crone has recently shown that this is quite improbable, especially as far
as this particular legal maxim is concerned.ut If my argument that the maxim was
known in .Iijdz by the first century and that it cannot be ruled out that the Prophet
himself had used it is sound, the hypothesis of a Roman (or better Roman provincial)
origin becomes even more dubious. It could only be possible if we could trace the
adoption of this legal rule to pre-Islamic times.
Such a supposition is not as extravagant as it may appear at first sight. The existence
of an old Arabian method of deciding paternity disputes by physiognomists does not
exclude that in some places, under the inffuence of other laws, the rule "al-walad
li-l-firdsh- may have been adopted. That this had indeed been the case is explicitly
stated in an aws'il tradition, usually regarded as anachronistic, reporting that the pre-
Islamic judge Aktham b. Sayfl decided according to that rule.us If this was not a new
idea in Arabia-which may be possible-which law could have been behind it? We
must consider Jewish, i.e., rabbinical law; there is indeed a parallel in the Babylonian
Talmud.6e Was the Jewish legal rule adopted from Ri:man law or was it originally
Jewish? Given the present state of our knowledge about pre-Islamic Arabia, too many
questions remain open and too much speculation is needed to push the origin of the
Islamic legal maxim to before the first century A.H., for neither Jewish nor Roman
origin can be proven.6e"

67 P. Crone, Roman, Provinciql and Islamic Law influences on Islamic law, see Crone, Romon, Pro'
(Cambridge, 1987), pp. l0 f. Cf. also Azami, Studies,_ vincial ond Islamic Law, esp. chap. l; my review of
pp. 265 f . Crone in Def Islam 65 (1988): 342-45; and W. B.
68 See Juynboll, Muslim Tradition, p. 16. Hallaq's article "The Use and Abuse of Evidence:
69 See Crone, Roman, p. I l. The Question of Provincial and Roman Influences
6ea Qn the general question of possible Roman on Early Islamic Law," JAOS I l0 (1990): 79-91.
20 JounNnr oF Nnnn EnsrEnN Sruuns

VI

I
have chosen and discussed in some detail the example "al-walad li-l-firäsh" be-
cause Schacht concentrated and relied on it in his monumental work on the origins of
Islamic jurisprudence. My thesis that by means of lAta"s references to prophetic
traditions this legal maxim can be dated back at least to the second half of the first
century 2..n., if not to the Prophet himself, undermines some of Schacht's fundamental
ideas, among them his well-known theories on the pattern of the development of
hadtth: Successors, Companions, and Prophet, that is, the traditions from the Prophet
concerning legal questions are the earliest link in the chain:

[...] Generally and broadly speaking, traditions from Companions and Successors are earlier
than those from the Prophet.To One of the main conclusions to be drawn t. . .] is that, generally
speaking, the "living tradition" of the ancient schools of law, based to a great extent on indi-
vidual reasoning, came first, that in the second stage it was put under the aegis of Companions,
that traditions from the Prophet himself, pat into circulation by traditionists toward the middle
of the second century,4.fL, disturbed and influenced this "living tradition", and that only
Shaf i secured to the traditions from the Prophet supreme authority.Tr . .. Every legal tradition
from the Prophet, until the coltrary is proved, must be taken not as an authentrc or essentially
authentic, even if slightly obscured, statement valid for his time or the time of the Companions,
but as the fictitiozs expression of a legal doctrine formulated at a later date.12 .. . We shall find
that the bulk of legal traditions from the Prophet known to Malik originated in the generation
preceding him, that is in the second quarter of the second century 4.H., and we shall not meet
any legal traditionfrom the Prophet which can be considered authentic (italics mine).73

The prophetic traditions connected with the legal maxim "al-walad li-l-firdsh wa-li-
l-'shir al-Ltajar" are made up of a"group of texts which clearly contradict Schacht's
theory on the time of the origin of those prophetic legal traditions. This is not an
isolated instance.Ta
We have seen that 'Atä' claims the Prophet only very rarely as an authority and
that he also gives his own opinion about legal problems for which he knows a tradi-
tion from the Prophet without referring to it. This is one argument against the assump-
tion that 'Atä' invented prophetic traditions himself. Those he quotes or hints at must
have already been in circulation during his lifetime, that is, that they can be placed in
the first century A.n. For the reasons already explainedTs and because of the general
lack of asantd, it also has to be ruled out that Ibn Jurayj falsely ascribed them to
'Ata'. His al.tddlth from the Prophet are-contrary to Schacht's generalization-not
earlier than his traditions from the Companions, they are not transmitted more care-
fully, and they obviously have no stronger probative force then the latter. In number,
!A.tä"s references to the Prophet are eclipsed by those to his teacher Ibn cAbbäs, but
the Prophet is mentioned more often than all the Companions, such as 'Umar, {,'isha,
or !AlI.
All this reflects the very subordinate role of the prophetic al.tddlth-and we can say
of the traditions in general-in the legal scholarship and teaching of 'Atä', and this

7o Schacht, Origins, p. 3. 73 Ibid.

7l Ibid., p. 138. 74 Cf . my forthcoming book Die Anfönge.
72 Ibid., p. 149. 75 See secs. IIIand IV of this article.
 Tnn MustxN.tp oF lAr-RnzzÄb AI-S,q,N'ANT 2l

state of affairs may be typical for Islamic jurisprudence of the first century ,q,.H. But we
have to emphasize that there were traditions from ;altdba and from the Prophet in the
first century, and they were sometimes used as sources or argumentsby thefuqahA'of
the late first and early second centuries to support their doctrines. We have to con-
clude that the last quarter of the first century of Islam was the beginning of a develop-
ment that made stormy progress in the course of the second century, reaching its peak
with the doctrines of ash-Sh4fi'r about a century later: the infiltration and incorpora-
tion of prophetic al.tödtth into Islamic jurisprudence.
The conclusion that the prophetic afuddtlh are marginal to the legal teaching of
'Atä' does not mean that they are worthless for us; on the contrary, they are excep-
tionally valuable. Since there is only one generation between 1A,ta' and Mul.rammad,
these texts are very close to the time and the people they report about, and their
authenticity cannot be ruled out a priori-as Schacht has done. qta"s transmissions
from the Prophet that have an isnad are especially precious in this respect. But his
transmissions without an isnad, too, can be successfully used to date traditions, if
variants from other sources are known.
While studying the Mu;annaf of 'Abd al-Razzdq,I came to the conclusion that the
theory championed by Goldziher, Schacht, and, in their footsteps, many others-
myself included-which, in general, rejects kadfth literature as a historically reliable
source for the first century LH., deprives the historical study of early Islam of an
important and useful type of source.
It goes without saying that this material cannot be regarded as completely truthful.
This even the Muslims themselves did not claim. Their method of sifting through the
material by means of the critical study of the transmitters was a quite workable
method of examination that may be of some use even for the modern historian, but it
was not entirely satisfactory and could not avoid misinterpretations. I think that with
the help of the newly available sources, such as lhe Mu;annaf of 'Abd al-Razzäq and
that of Ibn AbI Shayba, and the al.tddlth contained in early complexes of transmis-
sions-such as those of 'Atä' in 'Abd al-Razzäq's Musannaf, where the hadtth is not
the real object of the teaching but is only marginal-we are now able to raise the
question of the historical value of the kadilh texts anew.