Ahmad Ibn Hanbal’s Book of Renunciation 345

Ahmad Ibn Hanbal’s Book of Renunciation

by C h r i s t o p h e r M e l c h e r t
University of Oxford

Abstract
Ahmad ibn Hanbal’s book al-Zuhd (‘renunciation’) is one of the largest extant
collections of renunciant sayings from the first two Islamic centuries. It was
assembled by his son
Abd Allah, who contributed about half the sayings in it inde-
pendently of his father. The extant text is only half or a third of the version available
to Ibn Hajar in the Mamluk period. Some of what is missing can be recovered from
quotations in Abu Nu
aym, Hilyat al-awliya#. It is notably dominated by data from
Basra. Its contents are highly miscellaneous, but rejection of worldly goods appears
to be the theme that comes up most often.

Renunciation (zuhd) is a major feature of early Islamic piety. Its values,

especially fear of God and insistence on taking seriously the question of
one’s place in the Afterlife, apparently predominate in securely datable
Islamic inscriptions of the seventh century ce, to the point that little else
can be made out about the religion at that stage, such as the importance
of law and the Prophet.1 Modern scholarly consensus has for some time
agreed with medieval Islamic scholarship in locating the origins of Sufism,
which flourished from the later ninth century, in the early renunciant mo-
vement.2 Our principal sources for the history of renunciation are collecti-

1) V. Solange Ory, “Aspects religieux des textes épigraphiques du début de

l’Islam”, Les premières écritures islamiques, dir. Alfred-Louis de Prémare, Revue

due Monde Musulman et de la Méditerranée 58 (Aix-en-Provence 1990), 30–9; Fred
Donner, Narratives of Islamic Origins, Studies in Late Antiquity and Early
Islam 14 (Princeton 1998), chap. 2.
2) The classical account is Louis Massignon, Essay on the Origins of the Techni-

cal Language of Islamic Mysticism, trans. Benjamin Clark (Notre Dame 1997). Re-
cent historical overviews are Alexander Knysh, Islamic Mysticism: A Short History,
Themes in Islamic Studies 1 (Leiden 1999), and Ahmet Karamustafa, Sufism: The
Formative Period, The New Edinburgh Islamic Surveys (Edinburgh 2007). For the

Der Islam Bd. 85, S. 345–359 DOI 10.1515/ISLAM.2011.007

© Walter de Gruyter 2011
ISSN 0021-1818 Brought to you by | provisional account
Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
346 Christopher Melchert

ons of stories and sayings from the ninth to eleventh centuries, among
which the second largest is the Kitab al-Zuhd (“book of renunciation”) at-
tributed to Ahmad ibn Hanbal (d. 241/855). The intention of this study is
to sketch its extent and character.
As for Ahmad ibn Hanbal, modern biographies in Arabic have stressed
his involvement in the Inquisition and the formation of Islamic law.3 Nimrod
Hurvitz’s more recent biography in English rightly stresses Ahmad’s piety
as one basis of the regard in which he was held and of the Hanbali school
of law that formed after his death; however, it cites al-Zuhd very seldom, in
line with its general neglect of Ahmad
s activity as a collector of hadith.4
Two versions of al-Zuhd are in print, based on different manuscripts,
both without a critical apparatus. The first appeared in Mecca in the
mid-1930s with an introduction by one
Abd al-Rahman ibn Qasim, who
presumably also edited it on the basis of one Moroccan manuscript.5 Dar
al-Kutub al-
Ilmiya of Beirut published a photomechanical reprint in the
1970s, then a resetting of it with new pagination in the 1980s.6 The second

present writer’s understanding of the historical development of Sufism, v. Christopher

Melchert, “Basran Origins of Classical Sufism”, Der Islam 83 (2005), 221–40.
3) Oustandingly, Abu Zahra, Ibn Hanbal: hayatuhu wa-
asruhu wa-fiqhuh

(Cairo n.d.), Mustafa al-Sˇ ak
a, al-A#imma al-arba
a (Cairo and Beirut 1399/1979),
687–973, idem, al-Imam Ahmad ibn Hanbal (Beirut 1404/1984), and Fahmi Jad
an,
al-Mihna (Amman 1989).
4) Nimrod Hurvitz, The Formation of Hanbalism: Piety into Power, Culture

and Civilisation in the Middle East (London 2002). Other recent treatments of
Ahmad’s piety, both excellent, are Michael Cook, Commanding Right and Forbid-
ding Wrong in Islamic Thought (Cambridge 2000), chap. 5, and Michael Cooper-
son, Classical Arabic Biography: The Heirs of the Prophets in the Age of al-Ma#mun,
Cambridge Studies in Islamic Civilization (Cambridge 2000), chap. 4. V. also Chris-
topher Melchert, Ahmad ibn Hanbal, Makers of the Muslim World (Oxford: One-
world, 2006), chap. 5.
5) Ahmad ibn Hanbal, K. al-Zuhd (Mecca 1357), 400 pp. Sezgin identifies the

MS as Rabat, Kattani 292 (GAS 1:506, no. 3), 236 ff., confirmed by Roger Delad-
rière, Introduction to Bayhaqi, L’anthologie du renoncement, Collection “Islam
spiritual” (Lagrasse: Verdier, 1995), 9. According to Sezgin, the manuscript is
from the 12th century H. (i.e. approximately the 18th century CE). A facsimile of
the title page shows the year 1243 under the name of an owner, mostly crossed out,
but this seems to be the year someone acquired it, not when it was copied: Ahmad
ibn Hanbal, al-Zuhd, ed. Yahya ibn Muhammad Sus (n.p. n.d.), 29.
6) Ahmad ibn Hanbal, K. al-Zuhd (Beirut 1396/1976), 400 pp.; idem, al-Zuhd

(Beirut 1403/1983), 480 pp. Henceforth, citations of page numbers in the former will
appear in roman, in the latter in italic. From the same publisher is now available
Ahmad, al-Zuhd, ed. Muhammad
Abd al-Salam Sˇ ahin (Beirut 1420/1999), 327 pp.,

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 347

appeared in successive volumes from Alexandria and Beirut at the end of

the 1970s and beginning of the ’80s, editing attributed to a Muhammad
Ğalal Šaraf, based on one Libyan manuscript.7 Later in the 1980s, there ap-
peared an index to both editions from Yusuf
Abd al-Rahman al-
Mar
ašli.8 Roger Deladrière speaks of the Meccan edition as compri-
sing 2,379 items, which my own count confirms.9 A more recent reprinting
of the Meccan edition, edited by Yahya ibn Muhammad Sus, counts 2,418
items.10 Mar
ašli’s index indicates 893 items in the Meccan edition not
found in Sˇ araf ’s, 327 in Sˇ araf ’s edition not found in the Meccan, sugges-
ting an extant collection of about 2,700; however, his index is faulty on this
point and exaggerates the number of additional items in Sˇ araf ’s edition.
The two editions, which is to say the two manuscripts, follow the same plan
of beginning with the sayings of prophets, proceeding to the sayings of
Companions, then of persons who came after them, but the order of chap-
ters is slightly different and Sˇ araf ’s edition has fewer sayings in inappro-
priate chapters.
Both manuscripts come with an account of the book’s transmission up
to the year 708/1309–1011:
Nasir al-Din Abu
Abd Allah Yusuf ibn Muhammad ibn
Abd Allah al-Dimašqi
al-Šafi
i, known as Ibn al-Munhar <
Taqi al-Din Abu Muhammad
Abd al-Rahman ibn Abi l-Fahm ibn
Abd al-Rah-
man al-Buldani al-
Abbasi <

with items numbered but again without indices or cross-references. Another reprint
of the Meccan edition with items numbered and some cross-references is Ahmad,
al-Zuhd, ed. Muhammad al-Sa
id Basyuni Zaglul (Beirut 1423/2002), 568 pp. The
Meccan edition has also been reprinted by Dar Rayan and Dar
Umar ibn al-Äattab,
according to Sus, “Introduction”, Zuhd, 21. I have not myself seen Ahmad, al-Zuhd,
ed.
Isam Faris al-Harastani and Muhammad Ibrahim al-Zughli (Beirut 1994).
7) Ahmad ibn Hanbal, al-Zuhd, ed. Muhammad Ğalal S ˇ araf, 2 vols. (Alexan-
dria 1980, then Beirut 1981). Citations of page numbers that include a volume
number will be to this edition. MS identified as al-Ğami
a al-Libiya 3856, 358 ff., by
Sˇ araf, Zuhd, I, 6.
8) Yusuf
Abd al-Rahman al-Mar
ašli, Fihris ahadith K. al-Zuhd lil-imam

Ahmad ibn Muhammad ibn Hanbal, Silsilat faharis kutub al-sunna 6 (Beirut
1408/1988).
9) Deladrière, Anthologie, 9.
10) V. supra, n. 5. This edition, of 752 pages, includes marginal cross-references

to the Meccan edition and the Moroccan manuscript, also notes commenting on the
asanid. My guess is that it was published in Cairo in 2003.
11) Ahmad, al-Zuhd, 3 8 = ed. S ˇ araf, I, 23; cf. ed. Sus, 37.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
348 Christopher Melchert

Abu l-Qasim Muhammad ibn Yusuf ibn Yahya ibn Yunus al-Tağir <
Abu Talib
Abd al-Qadir ibn Muhammad ibn Yusuf al-Yusufi <
Abu
Ali al-Hasan ibn
Ali ibn Muhammad ibn
Ali ibn al-Muühib, by qira#a in
Rabi
I 443/July-August 1051, <
Abu Bakr Ahmad ibn Ğa
far ibn Hamdan ibn Malik al-Qati
i.

The earliest three names are known.12 Ibn Hağar once says of the ear-
liest two that they alone transmitted the Musnad of Ahmad ibn Hanbal
and al-Zuhd al-kabir.13 However, other sources concerning Ibn al-Muühib
(including other references by Ibn Hağar) call the latter book only al-
Zuhd, so we probably need not infer that there was ever a lost Kitab al-
Zuhd al-sagir. Al-Äatib al-Bagdadi disparages Ibn al-Muühib: “He also re-
lated from Ibn Malik K. al-Zuhd of Ahmad ibn Hanbal. He had no old copy
of it. His copy was in his own hand. He wrote it in his old age and it is not to
be argued by.”14 Fortunately, Ibn Hağar was wrong inasmuch as Abu
Nu
aym al-Isbahani also transmitted great parts, at least, of al-Zuhd.
Comparison with his transmission suggests that Ibn al-Muühib’s transmis-
sion was reasonably exact, on which more below. The Libyan manuscript
begins with almost the same list of transmitters, so the text traditions re-
presented by it and the Moroccan manuscript must have diverged subse-
quently to 708/1309–10. It seems fairly certain that neither manuscript
depends on the other, but they may go back to a common ancestor subse-
quent to 708/1309–10.
Saud Al-Sarhan, whose doctoral dissertation on the works of Ahmad
ibn Hanbal we eagerly await, has shown me photocopies from two other
manuscripts of al-Zuhd. Except for section headings, one seems to be
practically identical to the Moroccan text published by
Abd al-Rahman
ibn Qasim. The other includes the first nineteen folios from a recension at-
tributed to Salih ibn Ahmad (d. Isfahan, 266/880?) rather than
Abd Allah
(d. Baghdad, 290/903). It begins very similarly to the familiar Moroccan

12) On Abu Bakr al-Qati
i (d. 368/979), v. al-Üahabi, Siyar a
lam al-nubala#,

ed. Šu
ayb al-Arna#ut, &al., 25 vols. (Beirut 1981–8), XVI, 210–13, with further
references; on Ibn al-Muühib (d. 444/1052), v. ibid., XVII, 640–3; on Abu Talib al-
Yusufi (d. 516/1123), v. ibid., XIX, 386–7.
13) Ibn Hağar, Lisan al-Mizan, 7 vols. (Hyderabad 1329–31, repr. Beirut

1406/1986), I, 146, s.n. Ahmad ibn Ğa
far ibn Hamdan.

14) Al-Äatib al-Bagdadi, Ta#riä Bagdad, 14 vols. (Cairo 1349/1931, repr. n.d.),

VII, 391 = Ta#riä madinat al-salam, ed. Baššar
Awwad Ma
ruf, 17 vols. (Beirut
1422/2001), VIII, 394. A similar charge from al-Silafi apud Ibn Hağar, Lisan, II,
237, s.n. al-Hasan ibn
Ali ibn Muhammad.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 349

text but with occasional slight differences in the order of reports, addi-
tions, and omissions. It seems to regularly omit items in the Moroccan text
that come through
Abd Allah from someone else than Ahmad (more on
these below). Unfortunately, it comes with no account of its transmission
from Salih, nor does Salih’s name reappear after the first line. Having come
across no literary source that attributes any recension to Salih, I suspect
that it was originally someone’s selection from an earlier version of the
Moroccan text, omitting
Abd Allah’s name and items through him from
someone else than Ahmad. Someone else, then, noticing the omissions, as-
cribed the whole at the beginning to Salih.
The largest extant collection of renunciant sayings is Abu Nu
aym al-
Isbahani (d. 430/1038), Hilyat al-awliya#, which comprises about 15,600
items altogether.15 Even if we exclude its approximately 4,000 prophetic
hadith reports and 1,000 items from ninth- and tenth-century Sufis, this
remains our most abundant source by far. To my knowledge, the next-lar-
gest extant collection, after al-Zuhd of Ahmad ibn Hanbal, is K. al-Zuhd
wa-l-raqa#iq attributed to Ibn al-Mubarak (d. 181/797), which comprises
altogether about 2,050 items. After this come the kitab al-zuhd included
in Ibn Abi Šayba (d. 235/849), al-Musannaf, which comprises about 1,500
items, and the Kitab al-Zuhd of Hannad ibn al-Sari (d. 243/857), which
comprises almost as many.16
However, al-Zuhd of Ahmad ibn Hanbal was originally much longer
than the extant text. Ibn Hağar al-
Asqalani (d. 852/1449) states that Ah-
mad
s Musnad is three times as long as al-Zuhd.17 Ibn Hağar worked from a

15) Abu Nu
aym al-Isbahani, Hilyat al-awliya#, 10 vols. (Cairo 1352–7/1932–8).

A more recent edition is attributed to Mustafa
Abd al-Qadir
Ata#, 12 vols. (Beirut
1418/1997), but it represents no more than a retyping of the first edition with added
mistakes – scholars should avoid it as long as photomechanical reproductions of the
Cairo edition remain in print.
16) Ibn al-Mubarak, Kitab al-Zuhd wa-l-raqa#iq, ed. Habib al-Rahman al-

A
z. ami (Malegaon 1386). Indices by Yusuf
Abd al-Rahman al-Mar
ašli, Fihris
ahadi© Kitab al-Zuhd, Silsilat faharis kutub al-sunna 5 (Beirut 1408/1987). The
standard edition of Ibn Abi Šayba, al-Musannaf, is now that edited by Hamd ibn

Abd Allah al-G ˘ um
a and Muhammad ibn Ibrahim al-Luhaydan, 16 vols. (Riyadh
1425/2004). K. al-zuhd appears at XII, 133–468. To be sure, other renunciant say-
ings appear elsewhere in the larger work; e.g. k. al-du
a#, about 800 items at X,
5–204. Hannad ibn al-Sari, K. al-Zuhd, ed.
Abd al-Rahman ibn
Abd al-Ğabbar al-
Faryawa#i, 2 vols. (Kuweit 1406/1985).
17) Ibn Hağar, Ta
ğil al-manfa
a bi-zawa#id riğal al-a#imma al-arba
a (Hydera-

bad 1324), 8 = ed. Ikram Allah Imdad al-Haqq, 2 vols. (Beirut 1416/1996), I, 243.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
350 Christopher Melchert

Musnad of about 28,000 items, very like the one in print today.18 This im-
plies that he knew a Zuhd of more than 9,000 items, over three times as
long as either version of al-Zuhd in print today. It is impossible to say how
long after Ibn Hağar’s day the full version was lost.
Ibn Hağar’s estimate is confirmed by quotations in Abu Nu
aym, Hi-
lyat al-awliya#. Abu Nu
aym seldom mentions books by name; rather, like
al-Äatib al-Bagdadi and other traditionists, he prefers to cite everything
by isnad going up to the speaker of the item at hand. Yet many of Abu
Nu
aym’s reports can be identified as coming from particular books, just
as many of al-Äatib al-Bagdadi’s can.19 I have counted 480 certain quota-
tions of al-Zuhd in the Hilyah; for example, from the entry for the Yemeni
Successor Tawus (d. 106/724–5?),20
< Abu Bakr ibn Malik <
Abd Allah ibn Ahmad ibn Hanbal < my father <
Abd
al-Razzaq < Ma
mar that Tawus occupied himself with an ill comrade of his
until he had missed the pilgrimage.

They nearly all came to Abu Nu
aym by the links < Abu Bakr ibn Malik
<
Abd Allah ibn Ahmad ibn Hanbal or < Ahmad ibn Ğa
far ibn Hamdan <

Abd Allah ibn Ahmad.21 Abu Bakr ibn Malik and Ahmad ibn Ğa
far ibn
Hamdan are the same person, more usually known as Abu Bakr al-Qati
i,
from whom Abu Nu
aym collected hadith in Basra in about 360/970–1. Be-
sides these, however, I have also counted 737 apparent quotations not
found in the published texts of al-Zuhd; for example, the next after the one
just quoted22:
< Ahmad ibn Ğa
far ibn Hamdan <
Abd Allah ibn Ahmad < his father < Mahdi
ibn Ğa
far < Damra < Bilal ibn Ka
b: Tawus, when he went out of Yemen,
would drink only from ancient, Ğahili waters.

18) V. Christopher Melchert, “The Musnad of Ahmad ibn Hanbal”, Der Islam

82 (2005), 32–51, at 37–8, for these calculations.

19) Pedersen identifies quotations by Abu Nu
aym of Sulami’s Ta#riä al-su-

fiya in his introduction to Sulami, Kitab Tabaqat al-sufiyya, ed. Johannes Peder-
sen (Leiden 1960), 51–3, 57–9. On al-Äatib al-Bagdadi, v. Akram Diya# al-
Umari,
Mawarid al-Äatib al-Bagdadi fi Ta#riä Bagdad (n.p. 1395/1975).
20) Abu Nu
aym, Hilya, IV, 10 = Ahmad, Zuhd, 376 450.
21) For one exception among a handful noticed by me, v. Abu Nu
aym, Hilya, I,

70–1, which quotes a story on the authority of Muhammad ibn Ahmad ibn al-Hasan
that appears in Ahmad, Zuhd, II, 71.
22) Abu Nu
aym, Hilya, IV, 10.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 351

(The point was to avoid wells dug by tyrannical rulers.)

Identifying quotations of lost works is a difficult operation.23 In this
case, on the one hand, we are often able to compare our texts of al-Zuhd
with Abu Nu
aym’s quotations, and they are reassuringly close. Sometimes
one appears to give us a better text, sometimes the other, but discrepancies
are always within the usual range between variant manuscripts. An exam-
ple of Abu Nu
aym’s giving us the better text is where he quotes Ahmad
as relating an item (on al-Hasan al-Basri
s longstanding sadness) from

Ali ibn Hafs, whereas the extant Zuhd presents Ahmad as relating the
same item from
Ali ibn Ğa
far.24
Ali ibn Hafs of Baghdad is an historical
personage, Ahmad’s source for 38 hadith reports in the Musnad, whereas
no
Ali ibn Ğa
far is mentioned among his shaykhs.25 Ğa
far is easily expli-
cable as a scribal error. An example of a slightly better text in the printed
version is where it has Bilal ibn Sa
d enjoin us, la takun wali Allah fi
l-
alaniya wa-
aduwahu fi l-sirr (“Be not God’s friend in public but his
enemy in private”), whereas Abu Nu
aym has rather waliyan lillah, which
spoils the grammatical parallelism and is easily explicable as another scri-
bal error.26
On the other hand, many of Abu Nu
aym’s quotations from al-Qati
i
from
Abd Allah are manifestly not from al-Zuhd but rather from
Abd Al-
lah’s version of al-
Ilal wa-ma
rifat al-riğal or from the Musnad. Some quo-
tations of the
Ilal or Musnad are easy to spot, as they concern questions of
hadith transmission or law, but others concern matters of piety that one
might expect to find in al-Zuhd. For example, this item from Abu Nu
aym’s

23) For recent treatments of the difficulty, v. inter alia Lawrence I. Conrad,
“Recovering Lost Texts: Some Methodological Issues”, Journal of the American
Oriental Society 113 (1993), 258–63, Stefan Leder, “Grenzen der Rekonstruktion
alten Schrifttums nach den Angaben im Fihrist”, Ibn al-Nadim und die mittelalter-
liche arabische Literatur (Wiesbaden 1996), 21–31, and Ella Landau-Tasseron,
“The Reconstruction of Lost Sources”, al-Qantara 25 (2004), 45–91. The difficul-
ties appear to be greatest for works from before around the middle of the ninth cen-
tury CE, which is related to the very fluidity of texts before then, on which v. the
works of Gregor Schoeler, esp. “Die Frage der schriftlichen oder mündlichen
Überlieferung der Wissenschaften im frühen Islam”, Der Islam 62 (1985), 201–30.
24) Abu Nu
aym, Hilya, III, 19; Ahmad, Zuhd, 266, 326.
25) V.
Amir Hasan Sabri, Mu
ğam shuyuä al-imam Ahmad ibn Hanbal fi l-Mus-

nad (Beirut 1413/1993), 268–70; Ibn al-Ğawzi, Manaqib al-imam Ahmad ibn Han-
bal, ed. Muhammad Amin al-Äan ği al-Kutubi (Cairo 1349), 45, chap. 5, fi tasmiyat
man laqiya.
26) Ahmad, Zuhd, 385 461; Abu Nu
aym, Hilya, V, 228.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
352 Christopher Melchert

chapter on ®abit al-Bunani (d. 720s/738–48) is found also in the
Ilal but
not in al-Zuhd27:
< Abu Bakr ibn Malik <
Abd Allah ibn Ahmad ibn Hanbal < his father: I have
heard that Anas said to ®abit, “How your eyes resemble those of the Messenger
of God …”, whereupon he ceased not to weep until he had damaged his eyes.

This item from the chapter on the Basran Maymun ibn Siyah (fl. earlier
2nd/8th cent.) is found also in the Musnad but not in either printed version
of al-Zuhd28:
< Abu Bakr ibn Malik <
Abd Allah ibn Ahmad < his father < Muhammad ibn
Bakr < Maymun al-Muradi < Maymun ibn Siyah < Anas < Prophet: There is
no people who meet to recollect God, wanting by that only his face, without
there calling to them a caller in Heaven, saying “Go forgiven: your faults have
been replaced by virtues.”

A few prophetic hadith reports are found in both the Musnad and the
extant text of al-Zuhd; for example,29
< Abu Bakr ibn Malik <
Abd Allah ibn Ahmad ibn Hanbal < his father <
Abd
al-Rahman ibn Mahdi < Hammam < Qatada < Äulayd al-
Asari < Abu
l-Darda# < the Messenger of God …: The sun does not rise without there being
sent next to it two angels who cry out, “A little that suffices is better than
much that distracts.”

(The point is not to let possessions distract from one’s dependence on

God.) It is possible that the example just quoted, on people who meet to re-
collect God, originally was in al-Zuhd as well as the Musnad. Altogether,
I feel fairly sure that my count of Abu Nu
aym’s quotations of al-Zuhd is
below rather than above the true number.
It is impossible to tell for sure whether Abu Nu
aym or the anonymous
abridgers behind the printed versions of al-Zuhd present us more nearly
with a random selection of items from the original, long version. Among
Abu Nu
aym’s quotations, the ratio of those found in the printed version to
those not found is about 2:3, whereas it would be 1:2 or more if Ibn Hağar’s

27) Abu Nu
aym, Hilya, II, 323; Ahmad ibn Hanbal, al-
Ilal wa-ma
rifat al-

riğal, ed. Wasi Allah ibn Muhammad
Abbas, 4 vols. (Beirut 1988), II, 373 = idem,
al-Ğami
fi l-
ilal wa-ma
rifat al-riğal, ed. Muhammad Husam Baydun, 2 vols. (Bei-
rut 1410/1990), I, 332.
28) Abu Nu
aym, Hilya, III, 107–8; Ahmad, Musnad imam al-muhaddithin,

6 vols. (Cairo 1313), III, 142 = Musnad al-imam Ahmad ibn Hanbal, ed. Šu
ayb al-
Arna#ut et al., 50 vols. (Beirut 1413–21/1993–2001), XIX, 437.
29) Abu Nu
aym, Hilya, IX, 60; Ahmad, Musnad, V, 197 = ed. Arna#ut,

XXXVI, 52–3; Ahmad, Zuhd, 19 26.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 353

figure were certainly correct, if Abu Nu
aym’s selection were certainly ran-
dom, and if my undercounting of prophetic hadith quoted by Abu Nu
aym
from the lost part of al-Zuhd were certainly negligible. If we assume that
Abu Nu
aym’s selection from the original was random, then the original,
long Zuhd should have comprised about 6,800 items, not 9,000. It is also
possible that Abu Nu
aym was guided by some of the same principles of se-
lection as the anonymous abridgers, and I have somewhat undercounted
Abu Nu
aym’s quotations of prophetic hadith from al-Zuhd; therefore, an
original size of around 9,000 remains at least credible.
Just as Ahmad’s Musnad includes a significant number of additions
from his son
Abd Allah, meaning items he heard from other persons than his
father, so does al-Zuhd include many additions from him: a little more than
one-third of the printed versions, almost exactly one-half of the quotations
from Abu Nu
aym. Here, the parallel with the Musnad is weak evidence that
the proportions of material from Ahmad and
Abd Allah in the original, long
version of al-Zuhd were more like those in Abu Nu
aym’s sample; that is,
equal. Abu Nu
aym also quotes a substantial number of prophetic hadith
from
Abd Allah that he did not hear from his father. It is possible that these
were once transmitted with the Musnad, then excised with other hadith
from
Abd Allah. Such excision must have happened on a considerable scale
if we are to harmonize medieval reports that
Abd Allah’s additions compri-
sed about a quarter of the Musnad with the extant text, of which
Abd Allah’s
additions make up less than 5 percent.30 If there was a tendency over time to
drop material from the Musnad that did not come through Ahmad, the same
tendency might account for the diminished proportion of items from
Abd
Allah in al-Zuhd. (I have supposed the same tendency accounts for the ma-
nuscript of al-Zuhd attributed to Salih ibn Ahmad.)
The table of contents to al-Zuhd suggests stories of twelve qur’anic pro-
phets, not in chronological order (nor with all reports of particular ones
gathered together; e.g. sermons, wisdom, and the renunciation of
Isa are
distributed among three). Then come Companions, Successors, and others
of the eighth century CE. A comment from Ibn Taymiya confirms that the
original, long version of al-Zuhd was likewise arranged biographically, for
he reports preferring this arrangement to the topical one of Ibn al-Muba-
rak, al-Zuhd.31 The Musnad, likewise assembled by
Abd Allah from his fat-

30) V. Melchert, “Musnad”, 37, 47.

31) Jean R. Michot, Musique et danse selon Ibn Taymiyya, Études musulmanes
33 (Paris 1991), 122–3. Similar quotation apud Katib Çelebi, Kašf al-zunun, ed.
Şerefettin Yaltkaya and Rifat Bilge, 2 vols. (Istanbul 1942–3), II, 1423, s.v. kitab
al-zuhd.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
354 Christopher Melchert

her’s dictation and notes, occasionally groups together hadith that Ahmad
heard from some particular shaykh. So does al-Zuhd; for example, a se-
quence of 53 items from, ultimately, a very miscellaneous collection of pro-
phets, Companions, and Successors, of which 45 came to
Abd Allah from
Abu Bakr ibn Abi Šayba.32
An oddity of al-Zuhd by comparison with other collections of renunci-
ant sayings is the high proportion of items from pre-Muhammadan pro-
phets. A little over a third of Ibn al-Mubarak’s K. al-Zuhd comes from the
Prophet, a tenth of Ibn Abi Šayba’s, fully 45 percent of Hannad ibn al-Sa-
ri’s. About one-fifth of the published versions of Ahmad, al-Zuhd are made
up of hadith from the Prophet, an unsurprising proportion. The surprising
portion is another fifth of al-Zuhd comprising items from prophets before
Muhammad: in descending order,
Isa, Luqman, Ayyub, Dawud, and ot-
hers. This is far more than in any other such collection of renunciant say-
ings. The proportion of sayings from pre-Muhammadan prophets in Abu
Nu
aym’s selection is a mere one in twenty, suggesting that the extant, ab-
ridged versions of al-Zuhd include most of the original, long version’s ma-
terial from pre-Muhammadan prophets and that the long original was less
anomalous in this regard. About one-eighth of Abu Nu
aym’s quotations go
back to Companions, who were probably also, then, less well-represented in
the original, long version than in the extant abridgements. By the way, al-
Zuhd includes 35 items transmitted by the Yemeni Wahb ibn Munabbih (d.
113/731–2?), almost all of them concerning biblical prophets but almost
half of them concerning
Isa, which seems to tell against the surmise that
he was a convert from Judaism.33
Conversely, it was probably sayings from Successors that predominated
in the original, long version of al-Zuhd, as among Abu Nu
aym’s quotations
and as in most other ninth-century collections of renunciant sayings. The
latest persons to be quoted in al-Zuhd at the other end (not just relating
earlier sayings but speaking in their own right) are apparently Muhammad
ibn al-Farağ (d. 236/850–1), Bišr al-Hafi (d. 227/841), and Fath al-Mawsili
(d. 220/835), all quoted by
Abd Allah,34 and Sufyan ibn
Uyayna (d.
198/814?) and Ibrahim ibn Adham (d. 163/779–80?), quoted by Ahmad.35

32) Ahmad, Zuhd, 210–17 259–66.

33) Contra Steven M. Wasserstrom, Between Muslim and Jew (Princeton
1996), 26, among others.
34) Muhammad ibn al-Farağ at Zuhd, 317 385; Bišr at Abu Nu
aym, Hilya, VIII,

337, 338–9, 345, 347; Fath at Zuhd, 206–7 254.

35) Sufyan at Abu Nu
aym, Hilya, VII, 288 and Ahmad, Zuhd, 148 185, to which

add Sufyan’s direct quotations of the prophets Luqman and
Isa at Zuhd, I, 154 and

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 355

Of Ahmad
s immediate informants in al-Zuhd, 95 percent also appear in

the Musnad, 92 percent somewhere in the Six Books. Prophetic hadith and
renunciant sayings were thus transmitted by much the same persons, at
least in the eighth and early ninth centuries. Of hadith reports in Ahmad’s
Musnad, 34 percent came to Ahmad from Baghdadi shaykhs, 28 percent
from Basran, 15 percent from Kufan. Altogether, the Musnad is about 86
percent Iraqi. Significant numbers of hadith reports there also came from
Syrian, Meccan, and Yemeni shaykhs. Al-Zuhd is likewise 86 percent Iraqi.
Proportions from different centres are similar except that Basra (36 %)
and Baghdad (23 %) exchange places and Syria is insignificant. Ahmad
settled in Baghdad from 204/819–20 and travelled from it only once, to Sy-
ria, around 211/826–7 (except for forced journeys to al-Raqqah and Sam-
arra in connection with the Inquisition and its aftermath).36 The sugges-
tion is that his interest in collecting renunciant sayings weakened over
time by comparison with his interest in prophetic hadith, so that when he
finally went to Syria, he concentrated on collecting prophetic hadith. Quo-
tations of Ahmad himself certainly suggest that he thought physical priva-
tion unsuited to the married man, such as he had now become. His disciple
al-Marruüi writes,37
I told Abu
Abd Allah [i.e. Ahmad ibn Hanbal] … that the self-deniers were
saying that there is nothing better than paucity and hunger, and that if a man
accustomed himself to not eating save every two or three days, he would be re-
warded the same as someone who fasted perpetually. He said, “This is possible
only for someone who is alone. As for one who has dependants, how can he be so
strong? I broke my fast yesterday, and today my lower self impelled me to
break it (again). There is nothing to equal poverty. I remember those young
men of prayer.”

In his youth, by contrast, when he was an impoverished student, the

strictest austerity had made more sense.
The circulation of items in al-Zuhd also testifies to the geography of in-
terest in renunciation over the eighth century. Here are comparisons with
the kitab al-zuhd of Ibn Abi Šayba, showing percentages more than 5.
Ahmad’s source in early 9th cent. (first name in isnad):

Abu Nu
aym, Hilya, VII, 273–4, 288, 300, VIII, 101; Muhammad ibn Nadr at
Ahmad, Zuhd, 86, 368 108 441.
36) Üahabi, Siyar, XI, 306.
37) Ahmad ibn Hanbal, K. al-Wara
, ed. Muhammad Sayyid Basyuni Zaghlul

(Beirut 1409/1988), 81–2 = K. al-Wara
, ed. Zaynab Ibrahim al-Qarut (Beirut

1403/1983), 100.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
356 Christopher Melchert

Basra 76 (38 %), Kufa 37 (18 %), Baghdad 27 (13 %), Syria 16 (8 %), Wasit 13
(6 %), Yemen 9 and Khurasan 8 (4 %), Mecca 7 (3 %), Mesopotamia 5 (2 %),
Medina 2 (1 %).

Ahmad’s sources in mid- to late 8th cent. (second name): Basra 89

(46 %), Kufa 54 (28 %), Syria 16 (8 %), Yemen 8 (4 %), Baghdad 5 (2 %)
Medina, Mecca, & Khurasan 5 each (2 %), Egypt 4 (2 %), Wasit 2 (1 %).
Transmitter from Successor (some overlap with previous category):
Basra 84 (53 %), Kufa 31 (20 %), Syria 9 (6 %), Yemen 7 (4 %), Medina 6 (4 %),
Wasit and Khurasan 4 each (2 %), Baghdad 3 (2 %), Mecca and Egypt, 2 each
(1 %).

Successors:
Basra 80 (51 %), Kufa 35 (22 %), Medina 13 (8 %), Yemen 11 (7 %), Syria 8
(5 %), Mecca and Egypt 3 each (2 %), Wasit 2 (1 %).

Ibn Abi Šayba’s source in early 9th cent. (first name in isnad), sample
of 138:
Baghdad 6 (4 %), Basra 23 (17 %), Khurasan 3 (2 %), Kufa 93 (67 %), Mecca 1
(1 %), Wasit 12 (9 %).

Ibn Abi Šayba’s sources in mid- to late 8th cent. (second name), sample
of 145, of whom 137 identified:
Baghdad 1 (1 %), Basra 44 (32 %), Egypt 1 (1 %), Khurasan 1 (1 %), Kufa 73
(53 %), Mecca 3 (2 %), Medina 4 (3 %), Mesopotamia 1 (1 %), Syria 6 (4 %),
unknown 8 (6 %), Wasit 3 (2 %).

Successors, sample of 146, of whom 137 identified.

Basra 46 (34 %), Hijaz 1 (1 %), Khurasan 1 (1 %), Kufa 52 (38 %), Mecca 12
(9 %), Medina 17 (12 %), Mesopotamia 1 (1 %), Syria 3 (2 %), Yemen 4 (3 %).

The Zuhd is notably dominated by data from Basra, a domination that

becomes stronger the further back in time one goes. Ibn Abi Šayba travel-
led much less than Ahmad and gathered most of his hadith in Kufa (two-
thirds of k. al-zuhd, likewise of al-Musannaf as a whole), and Kufan items
outnumber Basran at all points in the eighth century; however, his col-
lection shows the same pattern of increasingly more data from Basra the
further back in the century one goes.
A considerable number of hadith reports in al-Zuhd are duplicated in
the Musnad. A few items are likewise repeated in al-Zuhd. Most of this re-
petition was probably deliberate, mainly similar or even identical sayings
supported by alternative asanid; e.g. Abu l-
Aliya (d. 93/711–12?) against
learning the Qur’an, then not reciting it, at Zuhd, 303 368. A few are exact

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 357

duplicates and probably unintentional; e.g. al-Hasan al-Basri (d. 110/728)

on persons who feared to accept even wealth that was rightfully theirs at
Zuhd, 37, 262–3 48, 221.
The most interesting questions about al-Zuhd naturally have to do
with its doctrine. There are three main literary traditions that supply us
with reports of renunciants. The most voluminous is that of hadith, under
which category falls the bulk of Abu Nu
aym’s collection and the collecti-
ons of Ibn al-Mubarak and Ibn Abi Šayba. The outstanding characteristic
of the hadith tradition is its insistence on full asanid to document the pro-
venance of every saying. Less voluminous but significant are the literatu-
res of adab and Sufism. The outstanding ninth-century adab collections
are the kitab al-zuhd included in al-Ğahiz (d. 255/868–9), al-Bayan wa-l-tab-
yin, the k. al-zuhd included in Ibn Qutayba (d. 276/889?), al-
Uyun wa-l-
aäbar, and numerous works by Ibn Abi l-Dunya (d. 281/884).38 The adab
tradition is distinguished from the hadith by its attraction to elegant locu-
tions, also, more subtly, to humorous material and often to miracle stories.
The Sufi tradition crystallized only in the later ninth century, and only
fragments and quotations remain of the earliest Sufi biographical dictio-
naries: Ibn al-A
rabi (d. 340/952?), Tabaqat al-nussak, and a large col-
lection not referred to by title from Ğa
far al-Äuldi (d. 348/959).39 It is dis-
tinguished from the hadith tradition in sometimes projecting later,
mystical values back onto the early renunciants, more regularly in making
out renunciation as an early stage in the formation of a mystic and the his-
torical formation of Sufism. There is, of course, considerable overlap
among the three traditions: individual books would be most conveniently
graphed on a triangle whose points would be zuhd, adab, and tasawwuf.

38) Al-Ğahiz, al-Bayan wa-l-tabyin, ed.
Abd al-Salam Muhammad Harun,

4 vols., Maktabat al-Ğahiz (Cairo 1367–9/1948–50), III, 125–92; Ibn Qutayba, al-

Uyun wa-l-aäbar, 4 vols. (Cairo 1343–9/1925–30), II, 61–375. For the works of Ibn
Abi l-Dunya, v. Reinhard Weipert and Stefan Weninger, “Die erhaltenen Werke des
Ibn Abi d-Dunya. Eine vorläufige Bestandsaufnahme”, ZDMG 146 (1996): 415–55.
My attention was first drawn to the distinction between significant practitioners
and littérateurs by Jacqueline Chabbi, “Remarques sur le développement histo-
rique des mouvements ascétiques et mystiques au Khurasan”, Studia Islamica, no.
46 (1977), 5–72, at 24.
39) On Ibn al-A
rabi, v. GAS 1:660–1; on al-Äuldi, v. GAS 1:661. Äuldi is said to

have assembled a book concerning 6,000 persons from the time of Adam until his
own, all of whom espoused the doctrine of the Sufis: al-Äatib al-Bagdadi, Ta#riä,
VII, 228 = ed. Ma
ruf, VIII, 147–8. An apparent extract (short) has been pub-
lished: al-Äuldi, al-Fawa#id wa-l-zuhd wa-l-raqa#iq wa-l-mara©i, ed. Muhammad
Fathi al-Sayyid (Tanta 1413/1993).

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
358 Christopher Melchert

With its ubiquitous asanid and few lines of poetry, al-Zuhd is plainly part
of the hadith tradition.
Major emphases of al-Zuhd are difficult to make out, stories and quo-
tations being so miscellaneous. Abu Nu
aym’s selection apparently inclu-
des more hostile sayings about rulers and fewer qur’anic glosses than the
extant abridgements. Otherwise, I have noticed no recurring differences in
the content of what they preserve of the original, long version. In a content
analysis of a random sample of 117 quotations from Abu Nu
aym, the ca-
tegory best represented is rejection of worldly goods; for example, that Mu-
jahid (Meccan, d. 104/722–3?) glossed Q. 102:8 (Jones translation: “Then,
on that day, you will be asked about bliss”), “About everything of the plea-
sures of the world.”40 This is followed by items praising particular indivi-
duals in fairly general terms; for example, that on the Day of Siffin (the
great battle between Mu
awiya and
Ali), a Syrian related of the Prophet
that Uways al-Qarani (said to have died in this very battle) was the best of
the Successors at doing well.41 Of ritual activities, prayer (salah) is the sin-
gle one most often commended; of austerities, restricted eating and drin-
king. Naturally, however, there is some overlapping of categories; for exam-
ple, when al-Hasan al-Basri says that the believer is sad morning and
evening, so that just a little food and water suffice him – is this to be clas-
sified as commending sadness or restricting one’s food and drink?42 In fact,
I did classify it as commending sadness, along with three other items in the
sample. To classify it as a commendation of restricting one’s food and drink
would suggest that it is about techniques to produce moral states, whereas
the actual sayings stress rather that physical austerities are the natural
outcome of a desirable moral state. Like other renunciant literature, al-
Zuhd is much more concerned with moral states than with teachable tech-
nique.
The contents of al-Zuhd are evidence first of all for what items of the
early renunciant tradition Ahmad and especially
Abd Allah ibn Ahmad
thought admirable. How far one takes them to be direct evidence of that
tradition depends first on how reliable one thinks hadith in general. Ah-
mad is quoted as calling for a lower standard of reliability concerning al-
targib wa-l-tarhib (“making to aspire and making to dread”): “When we re-
late (hadith) from the Messenger of God … concerning the licit and illicit,

40) Abu Nu
aym, Hilya, III, 281, quoting a lost section of Ahmad, Zuhd. Al-Ta-
bari quotes Mujahid the same way apropos of Q. 102:8, also by a completely differ-
ent isnad as glossing it “security and health (amn, sihha)”.
41) Abu Nu
aym, Hilya, II, 86, quoting from a lost section of
Abd Allah, Zuhd.
42) Abu Nu
aym, Hilya 2:132–3;
Abd Allah, Zuhd, 258 316.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM
 Ahmad Ibn Hanbal’s Book of Renunciation 359

the precedents and ordinances, we are strict about asanid; but when we re-
late (hadith) from the Prophet concerning the virtues of works and what
neither lays down nor suspends any ordinance, then we are easygoing
about asanid.”43 However, it does not appear that hadith reports in the
Musnad pertaining to al-targib wa-l-tarhib are any more liable to be weak
than hadith pertaining to ordinances.44 Neither have we reason to suppose
he filled up al-Zuhd (or instructed
Abd Allah to fill up al-Zuhd) with items
he considered weak. And, of course, the interest of al-Zuhd is primarily in
what it tells us of the piety of the eighth century, about which scholars no-
wadays tend to be markedly less sceptical than about the seventh. The ex-
tent to which its picture of eighth-century piety contradicts what is repor-
ted of Ahmad ibn Hanbal’s own – for example, the extremes of self-denial it
extols, as compared with the more moderate self-denial we are told that he
practised – raises our confidence that this material goes back well into the
century from which it purports to come.
The main findings of this study may be briefly summarized. The book
al-Zuhd attributed to Ahmad ibn Hanbal is the largest extant collection of
early renunciant sayings from its century, exceeded for all centuries only
by Abu Nu
aym, Hilyat al-awliya#. It was assembled by Ahmad’s son
Abd
Allah, who added to what he had heard from his father half or even as many
items again that he had heard from other sources.
Abd Allah’s text was two
or three times as long as the extant version, to judge by quotations in Abu
Nu
aym, Hilyat al-awliya# and a description by Ibn Hağar. The usefulness
of Hilyat al-awliya# is incidentally confirmed, both inasmuch as the Hilya
gives us a better idea of the original version of Ahmad’s al-Zuhd and inas-
much as we see that it accurately transmits the knowledge of the ninth
century. Ahmad’s al-Zuhd, finally, is an important source for the recon-
struction of his own piety, that of the early Sunni circles around him, and
more generally of Muslims in the eighth century, possibly also to some ex-
tent in the seventh. It is unusually rich in quotations of prophets before
Muhammad. Scholars are still at an early stage of figuring out how to in-
terpret it.

43) Al-Äatib al-Bagdadi, al-Kifaya fi
ilm al-riwaya, ed. Ahmad
Umar Hashim
(Beirut 1406/1986), 163 = ed. Muhammad al-Hafiz al-Ti ğani (Cairo 1972), 213.
44) Melchert, “Musnad”, 46–7.

Brought to you by | provisional account

Unauthenticated | 123.231.127.144
Download Date | 4/29/14 7:04 AM