You are on page 1of 16

Oral vs.

Written Transmission: The Case of abar and Ibn Sad


Author(s): Ghada Osman
Source: Arabica, T. 48, Fasc. 1 (2001), pp. 66-80
Published by: BRILL
Stable URL: http://www.jstor.org/stable/4057591 .
Accessed: 27/09/2013 13:01
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

BRILL is collaborating with JSTOR to digitize, preserve and extend access to Arabica.

http://www.jstor.org

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITTEN TRANSMISSION:


THE CASE OF TABARI AND IBN SA'D*
BY

GHADA OSMAN
HarvardUniversity

al-Rusulwa-l-Muliik,The History of Prophets and Kings, by


Muhammad b. Oarir al-Tabari (224-310 A.H./839-923 C.E.)
marked the beginning of a type of history-writing unprecedented among
Muslims. For the first time, there existed a universal history that spanned
from the beginning of creation until the historian's own time. Its author,
a Tabaristan-born hadtt scholar who spent many of the years until his
death in Baghdad, aimed at writing a history that asserted "on the one
hand, the essential oneness of prophetic mission and on the other, the
importance of the experiences of the umma and the continuity of these
experiences through time."'
In order to undertake his project, Tabari relied on a melange of
works. Contrary to earlier writers such as Ibn Ishaq, who used a variety of sources including the Quran, Arab genealogies and pre-Islamic
events, the magazf materials, poetry, a little documentary material
such as the Constitution of Medina, and general anecdotes, third and
fourth century A.H. authors more or less relied on the writings of their
predecessors for information. A perusal of Tabari shows that in fact
he relied on a variety of historians and other authors such as Abui
Mihinaf, Sayf b. cUmar, Ibn al-Kalbi, 'Awana b. al-HIakam, Nasr b.
Muzahim, al-Mada'ini, 'Urwa b. al-Zubayr, al-ZuhrT, Ibn Ishaq,
Waqidi, Wahb b. Munabbih, Ka'b al-Ahbar, Ibn al-MatnT, al-Haggag
'T"'ARI1H

* My thanks go to Professor Roy P. Mottahedeh for his encouragement and comments on the initial draft of this piece.
I 'Abd al-'AzIz Duri, The Rise of Historical WritingAmong the Arabs. Translated by
Lawrence I. Conrad. Princeton: Princeton University Press, 1983, p. 70.
? Koninklijke Brill NV, Leiden, 2001

Arabica,tome XLVIII

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRIT-TEN TRANSMISSION

67

b. al-Minha, Hisham b. 'Urwa, al-Zubayr b. Bakkar and so forth, in


addition to oral accounts that were circulating at the time.2
In recounting his history, Tabari used numerous turuq(channels) to
give accounts. These are both turuqthat are given by the same author
in a work, such as for example three different accounts that start with
the isnddal-Harita-Muhammad b. Sa'd-Muhammad b. 'Umar then
branch off in different directions, all with a matn suggesting that the
first man to accept Islam was Zayd b. Harita, as well as by different
authors, such as this same example coupled with an account with similar content and an isnadof 'Abd al-Rahman b. 'Abdallah b. al-Hakam'Abd al-Malik b. Maslama-Ibn Lahi'a-Abu al-Aswad-'Urwah. In
addition to these, Tabarr sometimes interjects and inserts general information that was presumably circulating at the time, sometimes introducing it with the words, "qala Abiu Gafar," thus creating "a basic
chronological framework and an outline of the main events... that
were matters of widespread and generally accepted knowledge."3
Tabarr relies overwhelmingly on the accounts of Ibn Ishaq. He frequently quotes them verbatim, so that we find accounts with the exact
wording in Tabari and Ibn Ishaq, as well as his later redaction in Ibn
Hisam. Yet although Ibn Ishaq figures prominently in Tabari's writings, to ignore Tabari's other sources would be shortsighted. This would
particularly be the case in overlooking his transmission of information
from Muhammad b. Sa'd.4
Born in Basra in 168 A.H./784 C.E. as a mawla, Ibn Sacd lived
there for the first part of his life, then moved to Medina and later to
Baghdad. Although he had studied under other scholars, including
Hisham ibn al-Kalbi (d. 207 A.H./822 C.E.), his name became especially linked with that of Waqidi (d. 207 A.H./823 C.E.). Virtually his

2 W. Montgomery Watt, "The Materials Used by Ibn Ishaq" in Historiansof theMiddle


East, ed. Lewis & Holt. London: Oxford University Press, 1962, pp. 24-31. See also
Watt's introduction in The Histoy of al-Tabari. Vol. VI: Muhammadat Mecca. Translated
by M.V. McDonald & M.W. Watt. Albany: State University of New York Press, 1988,
pp. xvi-xvii; Duri, p. 7, Jawad 'Ali, "Mawarid Tarih al-Tabarl," in Magm' al-'llmTal'Iraqi. Baghdad, vol. 2, 1952 & vol. 4, 1954.
I Watt, 1988, pp. xix-xx. For more detail on this see xix-xxii. Examples of occurrences of "qdla Aba (6afar" are in Muhammad b. Oarfr al-Tabari, Td'rih al-RusulwaI-Muliuk.Lugh. Bat.: EJ. Brill, 1964, 3:1181, 1183, 1188, 1222.
4 See Gordon Darnell Newby's examination of Tabari's reliance on Ibn Ishaq and
his quote from Yaqiut in 7he Making of the Last Prophet:A Reconstruction
of the Earliest
Biographyof Mu4ammad(Columbia, SC: University of South Carolina Press, 1989): "Tabari
erected his universal history on the foundations laid by Ibn Ishaq" (p. 16).

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

68

GHADA OSMAN

only extant work is the chronologically arranged multi-volume "Book


of Classes," I?it&bTabaqatal-Kubra.Ibn Sa'd died in Baghdad in 230
A.H./845 C.E.5
Just as Tabari's work is the earliest universal history to have reached
us, Ibn Sa'd's compilation is the first example of tabaqdtaccessible to
us. Along with Ibn Ishaq's sTra,Ibn Sa'd's Tabaqdtis also generally one
of the earliest Islamic works that we have. Yet whereas we have only
a fragment of the original Ibn Ishaq manuscript, leaving us to rely on
the altered recension of Ibn HiSam, the entire Tabaqdtof Ibn Sa'd is
believed to have reached us through the redaction of one of his students. His work is therefore also the earliest source used by Tabari
that survives in full. Furthermore, as Donner points out, whereas Ibn
Ishaq's Sra, WaqidY's Ktab al-Ma4rdz and Baladuri's Futih aI-Buld&n
focus on particular aspects, the Tabaqdt,on the other hand, provides
a wider selection of material, including contradictory reports about the
same event, in much the same way that Tabari does.6
Tabarl's transmissions from Ibn Sa'd are rather complex. While he
was using Ibn Ishaq as his foremost source, he appears to have been
referring to Ibn Sa'd as a source for supplementary material to augment and/or contrast Ibn Ishaq's account. Thus he omits a great many
of Ibn Sa'd's accounts, as opposed to the thorough inclusion of those
from Ibn Ishaq. This is not to imply, however, that Tabar- does not ever
quote at length from Ibn Sa'd; on the contrary, he quotes both long
and numerous passages from him throughout his history of Muhammad.
Perhaps in part due to the importance of accurate transmission of
data, numerous studies have been undertaken to assess the degree to
which oral versus written transmission was used in the classical Islamic
era. While studies by Nabia Abbott and Fuat Sezgin illuminated the
role of early written records, Rudolf Sellheim, based on an analysis of
amtal, concluded that despite the existence of written works by this
time, "eine freie muendliche ueberlieferung bis ins 4./10. Jahrhundert
hinein gang und gaebe war." Independently, Khoury explained that
On est tout A fait en droit de mlaintenir une attitude d'aspect double: d'une part
la reconnaissance d'une activite scriptuaire de plus en plus intense, des la deuxieme moitie du deuxieme siecle, d'autre part l'admission que la voie orale dans
la transmission est restee longtemps apr6s, et cela jusqu'au IV'W/X siecle, un moyen

Watt, 1988, p. xv.


Fred Donner, The Beginningsof IslamicHistoricalWriting.Studies in Late Antiquity &
Early Islam, vol. 14. Princeton: Darwin Press, 1998, p. 132.
5
6

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITTEN TRANSMISSION

69

important auquel on a recouru pour finir definitivement par ecrit beaucoup de


livres du IIIP/IXC siecle (et bien sur des siecles anterieurs).7

Yet "the degree of change imposed by such a transfer from oral to literary remains tantalizingly hard to grasp," as explained by Cameron
and Conrad. The two media should not be viewed as separate overlapping spheres, but rather as one homogenous entity. In his articles
discussing the commonalities of oral transmission,Gregor Schoeler refers
to the practice of reading written works aloud, combining orality and
written transmission. It is almost as if written works were in some cases
part of oral transmission in the early years of Islam. One such case
was that of Tabari with regard to Ibn Sa'd. While Tabarr may have
had access to some parts of Ibn Sa'd's Tabaq&t,I contend that he used
an oral account to transmit at least certain portions of the work. Since
nearly two volumes of Ibn Sa'd are dedicated to the life of the Prophet,
and it is here that the bulk of Tabari's transmission from Ibn Sa'd lies,
I shall use this section as the. focus of this article.8
In the centuries of early Islam, students were used to learning their
teachers' work by heart. As Holyland points out,
a Muslim scholar... might disseminate his knowledge by lecturing (samra'a)or
having his students recite to him from memory (qira'a)... This is important for
understanding the fluidity and mobility of texts, which would be to some degree
inevitable while oral and written tradition coexisted, priority beinggiven to oral
communzcation.

Sama'a was especially crucial for the study of religion, and "the tabaqat
division" was, after all, as Rosenthal points out, "more practical for the
purposes of the religious sciences."9

7 Nabia Abbott, Studiesin Arabic Literay Papyri. Chicago: Chicago University Press,
1957-72. Fuat Sezgin, Geschichte
des arabischenSchriftums.Leiden: EJ. Brill, 1967ff. Rudolf
Seilheim, "Abu 'Ali al-Qali-Zum Problem muendlicher und schriftlicher Ueberlieferung
am Beispiel von Sprichwoertersammlungen" in Studienzur Geschichte
undKulturdes Vorderen
Orients.Leiden, EJ. Brill, 1991, p. 374. Raif-Georges Khoury. "Pour une nouvelle comprehension de la transmission des textes dans les trois premiers siecles islamiques" in
Arabica34 (1987), p. 191.
8 Averil Cameron & Lawrence I. Conrad, The Byzantineand Early IslamicNear East
Problemsin the Literary SourceMaterials.Studies in Late Antiquity and Early Islam, vol. 1.
Princeton, NJ: Darwin Press, 1992, p. 17. Gregor Schoeler, "Die Frage der schriftlichen
oder muendlichen Ueberlieferung der Wissenschaften im fruehen Islam," in Der Islam
62 (1985), p. 204, "Weiteres zur Frage der schriftlichen oder muendlichen Ueberlieferung
der Wissenschaften im Islam," in Der Islam 66 (1989), p. 38.
9 Robert G. Holyland, SeeingIslam as OthersSaw It: A Survoyand Evaluationof Christian,
jewish and ZoroastrianWritingson Early Islam. Studies in Late Antiquity and Early Islam,

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

70

GHADA

OSMAN

Tabari almost always uses the word haddia with regard to Ibn Sa'd's
reports. This term can be taken to definitely or subtly suggest orality,
or can be viewed as a neutral term; it cannot by any means be construed as inevitably indicating written transmission. In fact, none of the
terms used by Tabari when transmitting from Ibn Sa'd about the life
of the Prophet, qala, dakara,or rawa 'an, can be viewed as such.
What appears through a comparison of accounts common to both
Tabari and Ibn Sa'd is an overwhelming similarity in wording, where
the former quotes the latter verbatim. A presentation of the complete
list of these accounts would be rather cumbersome here; I believe mentioning a few of these instances will suffice. Among them are the accounts
of enmity between Hasim and Umayyah, the judging between 'Abd alMuttalib b. Hasim and Harb b. Umayyah, the reports of predictions
of Muhammad's Prophethood, the Emigration to Abyssinia, and the
naming of Qurays, where despite the extra length of Ibn Sa'd's account,
the part that overlaps in the two is identical in wording.'0
An examination of the works of the two compilers for the period in
question reveals eighteen instances where their accounts are identical.
Yet there are also numerous instances of divergent accounts. Accounts
common in the two sometimes contain discrepancies in the wording,
structure and details of the matn,as well as in the asnaidsthemselves.'
It could be argued that the differences in accounts are due to
differences in recensions. While the author of Ibn Sad wa Tabaqtulhu,
'Izz al-Din cUmar Misa, alledges that in fact it is the version of alHarit that has reached us today, according to Fueck, the edition of Ibn

vol. 13. Princeton, NJ: Darwin Press, Inc, 1997, p. 41, emphasis mine. Schoeler, 1985,
Leiden: EJ. Brill, 1968, p. 95.
p. 227. Franz Rosenthal, A HiWtoty
ofMuslim Historiography.
zum KtItdbal-cIqdal-fand des AndalusiersIbn
Cf. Walter Werkmeister, Quellenuntersuchungen
Berlin: Klaus Schwarz Verlag, 1983, esp. pp. 463-9.
'Abdrabbih.
'0 Tabari, 3:1090-1, 1104-5, 1144, 1181-2; Mu1;ammad Ibn Sa'd, Al-Tabaqatal-KAbra.
Beirut, Lebanon: Dar al-Kutub al-'Ilmiyya, 1990, 1:62, 70-71, 58-59, 128, 159. Ibn
Sa'd cites Ibn al-Kalb! as his source for the first two accounts, and Waqidi for the last
three.
11 These similarities and differences are beyond the type brought up by Marsden
Jones in "Ibn Ishaq and al-Waqidi," Bulletin of the School of Orientaland Afiican Stdies
XXII, I, 1959, pp. 41-51, from which he surmises that "the greater part of the sira
was already formalized by the second century A.H., and that later wrnters showed a
common corpus of qass and traditional material, which they arranged according to their
own concepts and to which they added their own researches" (p. 51). Furthermore,
whereas Jones is examining texts not attributed to a previous author, I am looking at
instances where Tabari does cite Ibn Sa'd.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRFITEN TRANSMISSION

71

Sa'd we now have gives the recension of Ibn Sa'd's student Ibn
Hayyawiyah (d. 381 A.H./991 C.E.), while Tabari, on the other hand,
quoted Ibn Sa'd's Tabaq&t
from his teacher al-Harit b. Abi Usama (186282 A.H./802-895 C.E.). If the differences are due to the various recensions, then one would expect there to be no differences among authors
who used the same recension. A comparison between Tabari and Ibn
Sayyid al-Nas (1273-1334) is fruitful here. Although reported by Fueck
to have used Ibn Hayyawiyah's version, Ibn Sayyid al-Nas himself
acknowledges that he saw portions of it, but that he used that of alHarit for his work. The wording of Ibn Sayyid al-NMs'saccounts from
Ibn Sa'd is exactly the same as that of the Tabaq&t,with only the occasional word change or insert. Furthermore, we find that while accounts
are exactly the same in Ibn Sa'd and Ibn Sayyid al-NMs,they are presented with both a different isndd and matn in Tabari. Thus despite
both relying on al-Harit's version for their work, Ibn Sayyid al-NMsand
Tabari end up with divergent accounts, whereas in spite of relying on
different recensions, Ibn Sa'd and Ibn Sayyid al-NMsdo end up with
nearly identical ones.'2
in the Two Accounts
Divergences
A look at the accounts common to Ibn Sa'd and Tabari reveals that
there are many instances in which Tabari, for no apparent reason,
rewords and reshapes Ibn Sa'd's reports. These changes do not serve
the sake of brevity or clarity, nor do they. appear to be "in the interests of sundry literary,juristic and socio-politicalconcerns."'3Furthermore,
such reworded accounts tend to also differ to some extent from the
original in their isnads.
One example of this is the intertwined account of the marriage and
the death of 'Abdallah b. 'Abd al-Muttalib, the Prophet's father. In
this account, Tabari's only mention of 'Abdallah's marriage and death,

12 'Izz al-Din 'Umar Muasa, Ibn Sa'd wa-Tabaqatuhu.


Beirut, Lebanon: Dar al-Garb
al-Islami, 1987, p. 7. J. Fueck, "Ibn Sa'd," Engyclopedia
of Islam, 2nd Edition, vol. 8,
p. 922. For examples of changed words and/or inserts, cf. Ibn Sayyid al-Nas, 'Uyuin
al-Atarfi Funu2nal-Ma4zi wa-al-Sama'ilwa-al-Sgar. Beirut: Dar al-Affaqal-4adida, 1977,
1:63, 2:55, 2:67, 2:81, and Ibn Sa'd, 1:104-5, 2:39, 2:44, 2:52. For a comparison of
the three works cf. Ibn Sayyid al-Nas, 1:352, Ibn Sa'd, 2:21-2, Tabari, 3:1360, and
Ibn Sayyid al-Nas, 1:269-70, Ibn Sa'd, 2:3, Tabari 4:1757, 1763.
'3

Holyland,p. 38.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

72

GHADA OSMAN

and the sole instance in which he uses Ibn Sa'd as a source of information about 'Abdallah, he writes: "'Abdallah's father sent him to
al-Madinah for provisions, where he died. When he was late in returning, 'Abd al-Muttalib sent his son al-Harit to look for him but alHarit found that he had died." Tabari then inserts a segment that is
introduced by the words "Al-Waqidf: This is an error,"' saying that
'Abd al-Muttalib married at the same time as his son. Tabari then
returns to the death account, prefacing it with the isndd al-HaritIbn Sa'd-al-Waqidi, and writes a few more lines about 'Abdallah's
death and burial place, ending with the words, "There is no dispute
about this among our fellow scholars."'14
Ibn Said's report is written in a completely different manner. First
of all, the accounts of the marriage and the death are not interwoven,
but are rather completely separate. Furthermore, both accounts are presented in much more detail. The marriage account, although traced
back to Waqidi, does not contain the words "This is an error." The
details present before and after the WaqidT interjection in Tabari's
death account are interwoven into one paragraph in Ibn Sa'd."5
The isn&dsin the Ibn Sa'd and the Tabari accounts also diverge.
Both Ibn Sa'd and Tabari trace the marriage account to al-ZuhrY.
However, whereas the latter includes "Ma'mar and others" in his isndd
before reaching Zuhrf, who is his last link in the chain, the former
omits this link but includes a transmitter after Zuhri-, as well as an
alternative chain. The chain for the death account in Ibn Sa'd is also
longer than that of Tabari. If TabarY copied these accounts from Ibn
Sa'd's written account, why did he not follow the same order? And
why did he insert Waqidi's objection, not present in Ibn Sa'd's written account in the middle of the account?
In this case, we find that although the accounts are presented in a
different way, their contents are similar. Yet there are also cases where
there are discrepancies in the accounts themselves. In Tabari and Ibn
Sa'd we find that, for example, TabarT writes that Halid b. Sa'id was
the fifth person and the fourth man to accept Islam, and presents his

14
Tabari, 3:1081-2. All translations of TabarI are from the State University of New
York Press series "The History of al-Tabari," ed. Said Amir Arjomand. This quotation
is on 6:8-9.
15 The marriage is in Ibn Sa'd 1:76, the death 1:79-80.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITTEN TRANSMISSION

73

source as Ibn Sa'd-Waqidi. Ibn Sa'd himself, on the other hand, is


not so certain, reporting from Waqidi that Halid could have been the
third or fourth."6
An example of a discrepancy in both the isnddand the matnis with
regard to the account of Israfil's role in the revelation. While both men
present the view that Israfil was the bringer of the revelation for three
years, there are clear discrepancies in the accounts. The Tabari account
reads:
Al-Harit-Ibn Sa'd-Muhammad b. 'Umar al-Waqid!--al-Tawr --Isma'il b. Abi
Halid-al-4a'bi: He (al-Waqidi) also said that he had this from al-Mansar by
dictation-al-A.g'at-al-ga'bi:
Israfl was associated with the Messenger of God's
prophethood for three years. The Messenger of God was aware of him, but could
not see his person. After that came Gabriel.
Al-Waqidi: I mentioned that to Muhammad b. Salih b. Dinar, who said, "By
God, nephew, I heard 'Abdallah b. AbT Bakr b. Hazm and 'Asim b. 'Umar b.
Qatadah conversing in the mosque. An Iraqi man was saying this to them and
they both denied it, saying, 'We have neither heard nor learnt anything other
than that it was Gabriel who was associated with him and used to bring him revelations from the day he became a prophet until he died.'"'7

The actual account in Ibn Sa'd is as follows:


Al-Mu'alla b. Asad al-'Ammi---Wahib b. Halid-Dawud b. Abi Hind-'AmirHalaf b. al-Walid al-Azdi; Ualid b. 'Abdallah-Dawuid b. Abi Hind-'Amir; Nasr
b. Sa'ib al-Hurasdani-Dawu1db. Abi Hind-'Amir: the Messenger of God received
the prophethood when he was 40 years old. Israfil was with him for three years,
then he was disassociated from him, and associated with Gabriel for ten years in
Mecca and ten years after the migration to Medina. The Prophet died when he
was sixty-three years old.
Ibn Sa'd said, "I mentioned this account to Muhammad b. 'Umar who said, "The
scholars in our land have never heard that Israfil was associated with the Prophet.
They and the people of the sira among them say that only Gabriel was associated with him from when he first received revelation until he died."'8

The accounts raise two points of comparison. The first-and more nebulous one-is the significance of the phrase "by dictation" in Tabari's
quote from Waqidi. Is the significance of the insert that Waqidi heard
the account rather than read it, or is it that the account was written
down and dictated rather than passed along solely by word of mouth?

16 Tabarl, 3:168-9; Ibn Sa'd, 4:71. This issue is hardly unique to Tabari and Ibn
Sa'd: in her "Processes of Redaction: the Case of the Tamimite Delegation to the
Prophet Muhammad," Bulletinof the Schoolof OrientalandAfricanStudies49 (1986), p. 264,
Ella Landau-Tasseron quotes an instance when the commentator Tabarsi claims he is
quoting Ibn Ishaq, whereas the actual passage in Ibn Ishaq is different.
17 Tabari, 3:1248, English quote, 6:155.
18 Ibn Sa'd, 1:149-50. The word "a/barana"is used for all the transmission in these
chains. All translations of Ibn Sa'd are mine.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

74

GHADA

OSMAN

The second point is the clear discrepancy between the accounts. The
only commonality in the two is the assertion of Israfil's role and its
subsequent rejection. The isnads,phrasing of the matns,and chief characters of the accounts are completely different.
Another characteristic of Tabari's writing that supports the use of
orality is his otherwise inexplicable tendency to take only the first couple of accounts presented by Ibn Sa'd about a particular topic. This
is most clear in Tabafi's reports of the Prophet's mules, camels, milch
camels, milch sheep, swords, bows and lances, and shields, all exclusively taken from Ibn Sa'd. Here Tabari quotes Ibn Sa'd's accounts
1, 2, & 4; 1 & 2; 1-4; 6, 2, 3 & 5 combined; 1 as the only option; 1
& 2; and 1 only, for each section respectively. If using a written source,
he seems to have picked and chosen rather indiscriminately and to
have, in one place, changed the order unnecessarily.19
There are a few other points that should be mentioned here which,
taken collectively, point in the direction of oral transmission. The first
is a discrepancy in the name of one of the milch camels, quoted by
"-this of course
Tabari to be "al-rayya"but by Ibn Said to be "al-dabba'
could simply be an orthographic error, due to the similarity of the letters in Arabic. The second is that apart from the account of the shields,
all the accounts are from Waqidi, who Tabari mentions frequently. Yet
Ibn Sa'd presents more 4abarsin his section on camels, milch camels
and milch sheep from Waqidi that Tabari does not use. If these had
been plentiful it could have been suggested that Tabari wanted to keep
his selections brief, but it is curious that he only omits a few accounts.
Furthermore, in his treatment of the swords, he collapses Ibn Sa'd's
second, third and fifth accounts, without starting a new isnJd each time
as he is wont to do in such cases. He also switches the order maintained by Ibn Sa'd, presenting the animals before the weapons.20
The differences between the two sources goes a step further. On
numerous occasions Tabari cites Ibn Sa'd as his source for an account,
yet a perusal of Ibn Sa'd fails to reveal this account. It is unlikely that
these traditions constitute a part of the Tabaqit that has not reached
us; had this been the case we would have also remarked fragments of
accounts, parts of sentences, and other indications of missing work.
Oral transmission can easily account for this lapse. These omissions

19 Tabarl, 4:1782-8; Ibn Sa'd, 1:376-383.


20

Ibid.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITTEN

TRANSMISSION

75

occur for about ten 4abarsin Tabari's selections on the Prophet.2'


More light can be shed on the question of oral transmission if we
examine how Ibn Sa'd uses his main sources, Ibn al-Kalbi and Waqidi,
and how Tabari subsequently uses them. Unfortunately, Ibn al-Kalbi's
6amharatal-Nasab has not reached us in full, while Waqidi's Tabaqdt,
which will be demonstrated to have been the text used by Tabari, has
not reached us at all except through Ibn Sa'd's version.
The information presented by Tabari to be from Ibn al-Kalbi is
mainly genealogical; it also gives the names and kunyasof different individuals. It should be mentioned here that Tabari does not quote an
isndd to explain the source from which he obtained Ibn al-Kalbi; he
just writes "I was told by Hisam b. Muhammad" (huddittu'an Hisdm b.
Muhammad),unless he is obtaining information via Ibn Sa'd. Although
huddittucould be interpreted to accommodate for one or more transmitters between Ibn al-Kalbi and Tabari here, this seems to be an
unlikely option since first of all, Tabari is fairly meticulous in presenting inads, and second, it is too much of a coincidence that it is always
the chains going back to al-Kalbi that are never quoted.22
This could then lead us to the assumption that Tabari obtained the
information from a written source of Ibn al-Kalbi. But an examination
of the overlapping accounts reveals that Tabari's citations were gathered from different places in Ibn al-Kalbi, and, more significantdy,sometimes only some of the information is present, either indicating the loss
of the rest of Ibn al-Kalbi's details, or the interweaving of these details
with the popular lore that had been circulating during Tabari's time.
This exists for example in accounts of the Prophet's ancestors Qusayy
and Ma'add.23

2 See for example Tabari accounts 3:1129, 1145-6, 1167-9, 1263-4, 4:1764, 1766-7.
Werkmeister dealt with this issue with regard to the discrepancy between Ibn 'Abd
Rabbihi's citations from books and the citations in the books as we have them.
It should be pointed out here, however, that, as noted by Fueck, the frequent mention of Ibn Sa'd in the is7td "haddatantMuhammadb. Sad 'an ab2hi.. ." should not be
taken to refer to the author of the Tabaqdt, but rather to his namesake, Muhammad
b. Sa'd al-'Awfi (d. 276/888) (EI, Ibn Sa'd, p. 922). This observation, confirmed by the
absence of all accounts that begin with this isnad from the Tabaqdt, makes Heribert
Horst's "Zur Ueberlieferung im Korankommentar at-Tabaris" (Zeitschiiftder Deutschen
Morgenlaendischen
Gesellschaft103 (1953), pp. 290-307), where he confuses Muhammad b.
Sa'd al-'Awfi with the author of the Tabaqdt, irrelevant for the purposes of this article.
22 Tabari, 3:1008; Hisam b. Muhammad al-Kalbi, Oamharat
Kuwait: Kuwait
al-NAasab.
Government Press, 1983, p. 96.
23 Tabari, 3:1091-2, 1111-2; Ibn al-Kalbi, 90-91; 67. Note that this method of

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

76

GHADA OSMAN

In the few instances where the wording of the two writers' accounts
is similar, it is still not identical. In describing the genealogy of Hasim,
TabarTexplains that Hasim, whose name was actually 'Amr, was named
HIasim because "he used to break up (hasama)bread for tarFd[a kind
of broth into which bread was crumbled]." Tabari then gives two lines
of poetry that are also given by Ibn al-Kalbi regarding this. Then in
talking about Mudrika, both Ibn al-KalbTand Tabari explain how he,
his brothers Tabiha and Qama'a, and their mother Ijindif all acquired
their names, in accounts that use similar, but not identical, wording.24
Furthermore, there is at least one marked discrepancy between the
words written by Ibn al-Kalbi and the ones reported by Tabarf from
him through Ibn Sa'd. Ibn al-Kalbi writes, according to the chain
Hisam b. Muhammad-his father-Abiu Salih-Ibn 'Abbds, that the
Prophet would trace his ancestry back to Ma'add and then stop, saying that the genealogists were wrong, and that God had said regarding this, "And many centuries between this" [Quran 25:38]. Ibn cAbbas
then commented that had the Prophet wanted to teach this genealogy
he would have taught it, but that instead he just said, "Between Ma'add
b. cAdnan and Isma'rl there are thirty forefathers." Ibn al-KalbTtherefore starts his genealogy with the sons of 'Adnan and traces it down
to Muhammad. Tabari, on the other hand, mentions a tradition from
Ibn al-Kalbi through Ibn Sacd that says, "Someone told me on the
authority of my father, Muhammad b. al-SaVibal-Kalbi, although I did
not hear this from him myself, that he traced the descent as follows"
Forty names are then mentioned between Macadd and Isma'il. The
presence of the little disclaimer in the chain is interesting, especially
since the first account that includes the words of the Prophet also
includes Muhammad b. al-Kalbl.25
As for Ibn Sacd, he quotes both Ibn al-Kalbi's position and the
ancestry quoted by Tabari on the same page, then two other pieces of
information that he traces back to Ibn al-Kalbi that are also in Tabar,
before exploring other sources, including Ibn Ishaq, who is here known
to Ibn Sacd through Ru'yam b. Yazid al-Muqri-IHarun b. Abi (Isa
takingvariousaccountsand weaving them into one is directlystated in such accounts
as one that Tabari mentionsto be from Ibn Ishaq, Ibn al-Kalbi,and Waqidi through
Ibn Sa'd, where he says, "Some of them narratea versionwhich overlapsthat of others and some of them add to the version of others;Tabari, 3:1082-4, English quote,
9:9; Ibn al-Kalbi, 96.
24 Tabarl, 3:1088-9, 1107, Englishquote, 6:16; Ibn al-Kalbl, 92, 76-77.
25

Ibn al-Kalbi, 65; Tabari 3:1114-5, English quote, 6:38-39.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITITEN TRANSMISSION

77

al-9am!. None of the pieces that he quotes from Ibn al-Kalbi actually
contain a line of descent from Isma'il through 'Adnan.26
Another point of significance is that when one examines the accounts
that Tabarl obtains directly from Ibn al-Kalbi as opposed to those
taken from Ibn al-Kalbi through Ibn Sa'd, one discovers that whereas
some of the former group can still be found in Ibn al-Kalbi, the latter cannot. Although this latter group may have been included in the
lost portions of Ibn al-Kalbi, it seems too much of a coincidence that
they have all been lost. It is therefore possible that Ibn Sa'd had heard
orally the accounts that Tabarl quotes from Ibn al-Kalbi through him,
thus explaining why Tabari chose to quote them from him and not
from Ibn al-Kalbi. But the accounts that Tabari presents from Ibn alKalb! also do not point to his possession of a full manuscript of Gamharat
al-fasab. The phrase "huddittu'an Hisdm b. Muhammad"could be used
to indicate oral transmission of a well-known text that was known to
have been written down at some point, and so did not require an isnad
for legitimation.
I have already commented on the dissimilarity between accounts
given by Ibn al-Kalbi and those quoted from him by Ibn Sa'd, indicating a strong likelihood that Ibn Sa'd received these accounts orally
from his teacher. This can be expected more in the case of Waqidi.
Ibn Sa'd was Waqidi's secretary while he was just Ibn al-Kalbi's student, and so he was probably more familiar with the former's material. There is a greater chance that he himself would have transmitted
this material orally to al-Harit, who could have transmitted it to Tabari.
In his examination of the writings of Waqidi, Marsden Jones notes
that whereas Ibn Sa'd occasionally takes from Waqidi accounts that
have to do with the time before the mission, Tabari relies on him for
such information as fighting the Abyssinian invasion of Yemen and
the death of 'Abdallah b. 'Abd al-Muttalib. Jones uses this piece of
information to deduce that the book mentioned in the lists by Ibn
al-Nadim, Yaqiut and al-Safadi of Waqidi's works as Kit&bal-Tdrih
wa-l-MagazTwa-l-Mab'atwas in fact three books, that Tabari only had
the Ma4dzf volume in his possession, and that this was his source of
infomation. He does not consider at all that the information that comes
through Ibn Sa'd is actually from Waqidi's Tabaq&t,not his Mqgazf, although he takes the first step in acknowledging that we can reconstruct

26

Tabari, 3:1115-6; Ibn Sa'd, 1:46-47.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

78

GHADA OSMAN

Waqidi's Tabaqdtin light of Ibn Sa'd's work, which is edited in the


latter's style and copied from it.27
An examination of the three works reveals a pattern in transmission.
Let us take the accounts of the first expedition undertaken by the
Muslims, that of the Prophet's uncle Hamza. Tabari writes:
Al-Waqidi asserts that in this year, in Ramadan, seven months after the (ira
(about March 623), the Messenger of God entrusted a white banner to Hamza b.
'Abd al-Muttalib with the command of thirty men of the Emigrants. Their aim
was to intercept the caravans of Qurays. Hamza met Abtu (ahl at the head of
three hundred men. Madi b. 'Amr al-Guhani intervened between them, and they
separated without a battle. The banner of Hamza was carried by Abu Marad.28

Waqidl's account includes:


The expedition of Hamza b. 'Abd al-Muttalib took place in Ramadan, exactly
seven months after the Prophet's migration.
They say: the first banner given by the Prophet after his advent to Medina was
to Hamza b. 'Abd al-Mumtalib.He sent thirty riders in two groups, fifteen Emigrants
and fifteen Supporters. They reached the sea with him (IIamza), and obstructed
the way of Qurays's caravan that were coming from Syria and heading towards
Mecca. With (the caravan) was Abu Gahl with 300 Meccan riders. They met and
lined up for battle. Then Madi b. 'Amr walked (i.e. mediated) between them, continuing to go to this group and that group until they left: Hamza depared to
Medina with his companions, and Abiu C;ahl headed with his caravan and companions to Mecca. There was no fighting between them.29

None of the details about the thirty men all being Emigrants, the color
of the banner, or the role of Aba Martad are mentioned here, despite
Waqidr's lengthy accounts that contains may other details than those
mentioned here. We find in Ibn Sa'd, however:
The first banner made by the Prophet was for Hamza b. 'Abd al-Muttalib b.
Hasim in Ramadan exactly seven months after the migration. It was a white banner carried by Abui Martad Kunaz b. al-Husayn al-Ganawi, Hamza's confederate.
The Prophet sent him with thirty men of the Emigrants.
Some have said that they were two groups of the Emigrants and the Supporters,
but what is agreed upon is that they were all among the Emigrants. The Prophet
did not send any Supporters on expeditions until Badr....
Hamza went out and obstructed Qurays's caravan that had come from Syria and
was heading towards Mecca. With it was Abui 4ahl b. Hism with 300 men. They
reached the coast, from al-Is, and met and lined up for battle. Magdi b. cArnralOuhani, an ally of both groups, walked between the two, first to this group and
then to that one, until he separated them and they did not fight. Abu Oahl headed
toward Mecca with his companions and caravan, and Hamza left with his companions to Medina.30

27 Muhammad b. 'Umar al-Waqidi, Kitab al-Madzz2. Ed. Marsden Jones. London:


Oxford University Press, 1966, 1:13-15.
28
Tabari, 3:1264-5, English quote, 7:10.
29
Waqidi, 1:9-10.
30 Ibn Sa'd, 2:3-4.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

ORAL VS. WRITTEN TRANSMISSION

79

These two accounts are much closer; here we find Tabari's details
that were not in the Waqidi account. But again, the accounts are far
from identical. In this example, Tabari may have been concerned about
length, and so abridged the original account, but we can also note his
rearrangement of facts in other accounts such as that of B. Qaynuqa',
where Tabari's account is closer in both length and content to Ibn
Sa'd's than to Waqidf's Magazf, but still represents a significant departure from the Tabaqdt.We find a similar situation for Tabari's account
of al-Kudr, which does not even exist in Waqidi, but is still not worded
the same as that of Ibn Sa'd.3'
The dissimilarities in Tabari's accounts from Waqidi and the latter's
reports in al-Magazf have been taken as an indication of Tabari's cautious use of the work due to his conviction of Waqidi's unreliability.
Yet an examination of Tabarl, Waqidi's MagazT,and Ibn Sa'd reveals
that whereas Tabari's accounts are quite distant from those of al-Ma4zT,
they are quite close to Ibn Sa'd's accounts, which he mentions were
at least in part derived from Waqidi. Thus we can assume one of two
things. Either the accounts circulating were so divergent, that when
Tabari received reports supposedly from Waqidi, these were very far
from what he actually wrote. Yet we know from the examination above
that the changes obtained through orality do not tend to be so drastic; furthermore, even Ibn Sa'd, Waqidi's ka-tib,gives accounts that are
supposedly from him but that cannot be found in the Magazf. What
more likely happened is that Ibn Sa'd and Tabari obtained these from
accounts of Waqidi that are no longer extant, such as the Tabaqat,on
which we know Ibn Sa'd's work is based. Due to the loss of the original Waqidi text, however, we can point out but only guess about the
likelihood of oral rather than written transmission from Waqidi to Ibn
Sa'd and Tabari.32
Conclusion
In our day and age, although we are willing to admit the importance of orality to a large extent, it is difficult for us to fully conceive
31 Tabari, 3:1360, 1363, Ibn Sa'd, 2:22, 23, Waqldl, 1:177.
32
On the view that Tabari's accounts are dissimilar from those of Waqidi due to
the former's distrust of the latter see for example Michael Fishbein introduction The
Histogyof al Tabart. Vol. VIII: The Victotyof Islam. Albany: State University of New York
Press, 1997, pp. xx-xxi. For accounts that are in Ibn Sa'd but not in Waqidi, see for
example Ibn Sa'd, 2:6, 23.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

80

GHADA OSMAN

of the central role it played in the early centuries of Islam. The emergence of writing often leads us to believe that it competed and was on
a par with orality as a medium. In our eagerness to embrace written
sources, we often forget the practical problems associated with them:
the time and effort required to copy a manuscript, the slow pace of
travel to disseminate a written document from place to place, and the
resulting shortage of written works for students throughout the ninth
and tenth century Islamic world. All this served to maintain the crucial part played by orality and oral transmission.
A comparison between accounts common to Ibn Sa'd and Tabari,
although revealing parallels in many cases, also unearths numerous discrepancies in the wording, structure and details of the accounts, as well
as in their isnads, pointing to oral transmission. But our investigations
should not stop here; we should rather free ourselves from the confines
of our mindsets and surroundings, and ask further questions. How frequently did oral as opposed to written transmission take place? Can
we even take verbatim accounts to indicate a written transmission?How
did oral and written transmission actually affect reliability in the classical Islamic era? Regardless of the degree of faith that one places in
the early Islamic sources in general, if "oral tradition and oral teaching can continue to be of major importance long after the introduction of written records and literary accounts," then should orality be
viewed to be less reliable than writing? It is through the study of particular texts that we can grasp the role of orality and elaborate on its
effect on learning and transmission in the classical Islamic era.33

Cameron & Conrad, p. 17.

This content downloaded from 193.54.110.35 on Fri, 27 Sep 2013 13:01:33 PM


All use subject to JSTOR Terms and Conditions

You might also like