On the Term "Khādim" in the Sense of " eunuch " in the Early Muslim Sources

Author(s): David Ayalon

Source: Arabica, T. 32, Fasc. 3 (Nov., 1985), pp. 289-308
Published by: BRILL
Stable URL: http://www.jstor.org/stable/4056896 .
Accessed: 20/01/2011 23:09

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .
http://www.jstor.org/action/showPublisher?publisherCode=bap. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

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
 ON THE TERM KHADIM IN THE SENSE OF <<EUNUCH?)
IN THE EARLY MUSLIM SOURCES

BY

DAVID AYALON

IN a very recent article A. Cheikh Moussa studied meticulously the

main passages of al-Jahiz dealing with the eunuchs in his Kitab
al-Hayawan and Kitdb Mufakharat al-Jawiari wal-Ghilmnn' (later
Hajawan and Jawdri). This is a welcome and timely contribution. In
spite of the obvious centrality of the eunuch institution in Muslim history
and civilization the subject was very much neglected (with the possible
partial exception of its study regarding the Ottoman empire). Any
systematic research into that subject will quickly show that its impor-
tance far surpasses all previous expectations. The passages of al-Jahiz are
essential for the study of certain aspects of the eunuch phenomenon in
Islam, and their thorough analysis is long overdue. Their detailed
scrutiny by A. Cheikh Moussa (later M.) is certainly superior to
anything written about them earlier. This does not imply the absence of
certain flaws in that scrutiny, which are outside the scope of the present
lines.
In an addendum to the above cited article2, the same author expresses
disagreement with a statement and a conclusion of mine included in the
first installment of my study <<Onthe Eunuchs in Islam?>3.
The statement relates to the attitude towards the black eunuchs as
compared to the attitude towards unemasculated blacks in major
Muslim centers. The conclusion is about the term khadim in the sense of
?eunuch>>in the Muslim Medieval (especially historical and related)
sources.
Since he deals with the first of the two very briefly, and since I myself

I A. Cheikh Moussa, Gahiz et les Eunuques ou la confusion du meme et de l'autre,

Arabica, XXIX, 1982, pp. 184-214.
2 Ibid., pp. 212-214.

3 Jerusalem Studies in Arabic and Islam (JSAI), Jerusalem, vol. I (1979), pp. 67-124, and
especially pp. 64, 72, 83 and 85.

Arabica, Tome XXXII, 1985

290 D. AYALON [2]

am discussing that subject in much greater detail in other works of mine,

some of which are not yet published, I shall concentrate at this stage on
replying to his arguments concerning the term khadim.
In a nutshell: M. disagrees with my attempt (a most tentative one, as
far as the very early part of the Muslim era is concerned, a fact which I
stress in my study), to establish the date of the introduction of khadim
in the above mentioned nmeaning.In particular he contradicts my
attributing the use of that term in that sense to al-Jahiz. From his
conviction that that author never employed khadim as equivalent to
khasi he draws much wider conclusions.
My comments on his criticism are these: From his presentation the
reader cannot form a proper idea about the range and variety of the
source material with which I back my conclusion on this subject, or
about the reasoning and the line of argumentation which I use in order to
reach it. Even if he is right about al-Jahiz and khadim, which I believe he
is not, what he infers from it is not correct. Finally, what he attributes to
me as stating is not always accurate.
These comments on M.'s criticism had been written only wih reference
to two instances which I bring in my above mentioned study, where al-
Jahiz uses, in my view, the term khaidimin the sense of <<eunuch>>4. M., in
opposing that identification, deals only with one instance, and overlooks
the other. In both cases individual persons are involved.
Those two instances were the ones I knew (or, more precisely,
remembered) when I wrote my article on the eunuchs, and in an earlier
reply to M. (which I shall call <<thefirst version>>)I had to confine myself
only to them. It goes, of course, without saying, that in following a term
and its development throughout a number of centuries, exhausting the
whole data pertaining to that term is impossible. If, however, one is on
the right track, there are very good chances that with further reading he
will find not only additional or better proofs to support his thesis, but
even such proofs which will overshadow the earlier ones. Within the
context of the present difference of opinion between M. and myself, I
found, after having written the first version, a proof of that category in
the writings of the selfsame al-Jahiz. Furthermore, it is included in one of
the two works which M. used in his research on that author and the
eunuchs.

4 JSAI, I, p. 85.
[3] KHADIM 291

That piece of evidence is not only decisive in connection with the

present difference of opinion, but is of the highest importance for the
reconstruction of the history and development of our term in the early
Muslim period. I shall, therefore, deal first of all with it, and in some
detail. The discussion regarding the two original instances will be only
somewhat abbreviated, because it contains points which are both
pertinent to our subject and can be better clarified here than elsewhere.
The work of al-Jahiz which contains the evidence in question is the
epistle Jawa-ri5, mentioned in the opening lines of the present study. But
before presenting and analyzing the evidence, a few remarks about the
epistle have to be made, which will explain the background of the
evidence.
Al-Jahiz contrasts there the qualities of the boys and the girls (mainly
the slave-boys and the slave-girls), shifting very often to men and women
in general, and putting special emphasis on the sexual aspect (including
adultery and pederasty)6. He does it by means of a debate or dispute
between a protagonist of the girls (siihib al-/awairi)and a protagonist of
the boys (sahib al-ghilman). A repeated argument of this second prota-
gonist is that the lovers of the females are primitive and rude, whereas the
lovers of the males are refined and sophisticated. He includes the early
Arabs (and particularly their poets) in the first category and the later
ones in the second.
The passage on the eunuchs occupies two full pages7 out of a debate
covering about thirty sparsely printed pages (including the scholarly
apparatus)8, and it comes at the very end of that debate. It is represented
as the pronouncement of sahib al-,jawdr, and it opens with these words.

I. <<You[i.e. sahib al-ghilmnan] mentioned the khisyan and the beauti, of their,fgures
and the smoothness of their complexions, and the [practice of] having [carnal] pleasure
with them. And [you also said] that that is something with which thefirst ones were not
acquainted. You thus drove rmeagainst my will to describe the characteristics of the
khisicanin spite of the fact that that is senseless in our [present] book, for we confined
ourselves to speaking only about the jawari and the ghilman s>(wa-dhakartaal-khisiJin
wa-husn qudiidihim wa-na mat absharihim wal-taladhdhudhbihim wa-anna dhtilika Ia

It is included in 'Abd al-Salam Muhammad Hariun'sRasd'il al-Jdhiz, vol. 11 (1965),

pp. 87-137 (it was published earlier in the form of a booklet by C. Pellat, Beirut, 1957,
94 pp.).
6 <<Weliked to mention what went on between the pederasts and the adulterers)>-

ahbabnd an nadhkur mdaard bayna al-ldta wal-zundt (ibid., p. 95, 11.6-7).This statement is
included in the introductory part of the epistle.
7 Ibid., pp. 123, 1.4 -125, 1.4.

8 Ibid., pp. 95, 1.12-125, 1.7.

292 D. AYALON [4]

ta'rifuhual-awd'il /a-alja'tand ila9 an nasif ma fl al-khisydanwa-in lam vakun li-dhalika

mnananft kitdbina idh kunna innandanaqul fi al-jawiri wal-ghilmdn)t?.

These opening lines are followed by the enumeration of those charac-

teristics with the clear major aim of demonstrating that the eunuchs are
neither men nor women " and, therefore, should be excluded from the
debate.
What the protagonist of the Jawiari affirms in the clearest possible
terms is that the whole discussion about the eunuchs had been forced
upon him by his fellow-disputant, because it was that disputant who had
been the first to refer to them. This affirmation has a sole, single and
inevitable meaning: the protagonist of the boys must have already
mentioned the eunuchs. One has only to examine the few preceding
pages which contain that protagonist's statements, and he will surely
discover that reference. It is most regrettable that M. did not do that
obvious thing.
In looking for that reference in those preceding few pages one will not
find the word khisyadn(or khasi). The inescapable conclusion to be drawn
from this fact is that the eunuchs must appear under a different name.
And, indeed, they do appear under such name in the following circum-
stances.
II. According to the assertion of sdhib al-jawdrfit never happened that
the love of a ghuldm caused the death of any lover (lam nasma' bi-'ashiq
qatalaha hubb ghuldm). As by contrast, he names seven poets who died
because of their love to a woman. Amongst them were Kuthayyir, Jamal
and 'Urwa, whose respective love to 'Azza, Buthayna and 'Afra' had
been the cause of their death'2.
The answer of sahib al-ghilidin to this assertion is the following one.
III. ?Had Kuthayyir, Jamal and 'Urwa, as well as their likes whom you named, seen
some of the k h a da ni of the people of our lime -[I mean] those who had been bought
for huge sums of money -[and perceived]how good looking, clean coloured, and well
balanced [they are], and how beautiful are their figures [literally: the beautj of their
figures], they would have cast off Buthayna, 'Azza and 'Afra' ... and discarded them

9 Lane in his dictionary translates alja'ahu ild shay'in thus: ?He constrained, compelled,
forced, drove, or necessitated, him to have recourse to or betake himself to, or to repair to,
or to do a thing; he impelled him, or drove him against his will, to it, or to do it>>.I preferred
the underlined.
10 Jawdri, etc. p. 123, 11.4-6.To help the reader follow my line of argument I gave
Roman numerals to the five major citations from al-Jahiz's epistle. For convenience sake I
call all of them ?passages>?even when that author's words are fully or partly paraphrased.
" Ibid., pp. 123, 1.7-125, 1.4.
12 Ibid., pp. 104, 1.11-105, 1.2.
[5] KHADIM 293

as if they [i.e. these women] had been [mere] dogs (law nazara Kuthayvir wa-Jumayjil
wa-'Urwa wa-man sammajvtanuinnuzara'ihim ild ba'd k hadam ahl 'asrina mimman
qad ushturivabil-mil al-'azimfardhatan wa-shatdtanw a - n a q a' la w n wa-husni'tiddl
w a -j a u,d a t q a dd wa-qawdmla-nabadhuiButhayna wa-'Azza wa-'Afrd' min hdliq13
wa-tarakkhunnabi-mazjar al-kildb). But your line of argument was [to bring as an
evidence] against me [the view of] rude and uncivil Beduins (wa-lakinnaka ihtajajta
ala'nadbi- A 'rdb ajlIf jufdt), who were nurtured in misery and wretchedness and
grew up in them. They know nothing about the luxuries of life and the worldly
pleasures. They live in the deserts and shy away from [civilized] people like wild
animals. They eat hedgehogs and lizards and cut open the colocynth [in order to eat its
kernel]. The maximum that any one of them can reach is to cry over the remnants of
the [Beduin]encampment, and liken the woman to a cow or a gazelle, when, in fact, the
woman is more beautiful than both. Moreover, he [i.e. the Beduin poet] would liken
her to a serpent and call her the disfigured and the scabby, alleging that he does it for
fear of the evil eye> 14.

That sdhib al-jawarl decided to speak about the k h is y d n only in

reaction to what sahib al-ghilman said about the k ha d am, there cannot
be the slightest doubt. This is because another alternative simply does not
exist in those few pages which contain the statements of the boys'
protagonist. This alone settles the whole matter. Since, however, I have
already decided to perform an ?overkill>>in connection with the term
under discussion long before I came across M.'s article and comments 15,
I shall do the same again.
First of all, there is complete agreement between what the protagonist
of the boys says about the khadam and what his antagonist attributes to
him as saying about the khisyain:it is their being distinguished by their
beautiful features and their serving as an object of carnal pleasure. Both
antagonists use also very close expressions in their respective passages
(III and I): compare jawdat qadd with husn qudi7d,and naqa' al-lawn with
safi' al-lawn16.
Secondly, the opening lines about the khisyjn, which we have reprod-
uced and translated (passage I), contain the word al-awJ'il'7 (?the
first ones>>), which is a key term in our context. The same word appears
two additional times in the epistle (passages IV, V below), and once in
the form of al-awwah7n(passage IV).

13 Pellat (op. cit., p. 27, 1.1 and note 1) corrects JL, to _JL-. I am not sure about the

translation, but this has no bearing on the meaning of our passage.

14 Jawdri, etc., p. 105, 11.3-12.
15 In the articles ?From Ayyfubids to Mamlufks>> (REI, in the press), and <<Harunal-
Rashid and his Eunuchs>>(in an advanced stage of preparation).
16 This expression is used by sdaib al-jawari in connection with the khiiyyn a few lines
after his opening words on the eunuchs, which were reproduced above (Jawdri, p. 123, 1.10).
17 Ibid., p. 123, 1.5.
294 D.AYALON [6]

IV.Sahibal-jaHwdri- quotes verses of the earlier poets who wrote love

songs about the women (shabbabuibil-nisd') and states that those verses
are far superior to the ones composed by the modern poets (about the
boys) 8. He considers Imru'u al-Qays as sha'ir al-shu'ara' min al-awwalun
wal-dkhirin 9, and calls the rest of the earlier poets al-qudamd'fi al-
Jdhilivya wal-IshAi20 and al-aivawij21, He names the modern poets
al-muhdathun22
The retort of sdhib al-ghilman to his antagonist's praise of the earlier
poets is this.
V. ?<Youhave been wrong in the debate and unjust in your argumentation. For we
have not denied the merit of the a u a 'iI among the poets. The only thing we said was
that they had been rule and uncivil Be(uins ( a r aab aj ldafj u.f,a t ). They do not know
the delicacies of life and the worldly pleasures. For even if any one of them would exert
himself he would [not go beyond] comparing the woman to a cow or a gazelle or a
snake. And if he wanted to say that she had been well balanced and straight he would
compare her to a stick and her thigh to a reed. For they grew up with wild animals and
snakes, and were acquainted with nothing else. But we know that the very good
looking girl is more beautiful than the cow or the gazelle, or anything else with which
she had been compared>>23.

The passage just quoted (V) proves most decisively that the a w a 'iI of
the protagonist of the jawJrl(passage I) and the A 'r ab a]jI f j ufa t
are identical from the point of view of the protagonist of the ghilman. It
goes also without saying that the a 'r a b aj a/J'fj
I uf'at of this passage (V)
are exactly the same as their namesakes of the k h ada m passage (III).
Note also the identity of the detailed characterization of the early Arab
poets (awa'il, A'rab aj/lf jufat) as primitive people by the boys'
protagonist in passages III and V.
This leads to the following inescapable conclusion: When sdhib at-
jawar[ quotes sahibal-ghilman as saying that the a w a '1i were ignorant
of the [carnal]pleasure experienced with the khiy Jun(passage I), he could
refer to nothing else but to that selfsame person's statement about the
ignorance of the a'rJb ajlJf jufaJt (whom he identifies with the
a wa 'i I-passage V) of the [carnal] pleasure experienced with the khadam
(passage III). The first line and a half of the khisyJn passage (I) is purely
and simply a brief summary of the khadam passage (III).

18 Ibid., pp. 1 4, 1.1-116, 1.2.

19 Ibid., p. 114, 11.1-2.
20
Ibid., p. 115, 1.10.
21 Ibid., p. 115, 1.13.
22 Ibid., p. 115, 1.12.
23 Ibid., p. 116, 11.3-10.
[7] KHADIM 295

The synonymity of khadam and khisy,jn in al-Jahiz's Jawari has thus

been determined in the most definitive way24, and this fact has several
implications. First of all, it alone demolishes the whole structure which
M. built in order to disprove my argumentation about the early use of
khadim as equivalent to ?eunuch?) in the Muslim historical and related
sources25. By far his major point is that, contrary to my claim, al-Jahiz
never employs that term in that sense in any of his writings which he (i.e.
M.) or I quoted. This is how he formulates his final verdict in the very
last sentence of his criticism:
<Rien dans les passages consider&s [de Oahiz] ne permet de relever la moindre
ambiguit&:hadlimii
signifie bien serviteur mais jamais eunuque?>(my italics -D.A.)26.

Anything else he says becomes meaningless with the elimination of this

point. Secondly, it completely vindicates my claim in my article <<Onthe
Eunuchs in Islam?), that in the two instances I brought there al-Jahiz
meant ?eunuch?> when he mentioned khadim. The proofs that I
mobilized in the first version of my reply to M. thus lose their primary
place. However, because of their much wider inferences I shall repeat
most of them here, leaving the main implication of the evidence from
Jawiri for a later stage in the present paper.
What I said in my article <<Onthe Eunuchs, etc.>>about al-Jahiz's
employment of khddim as the synonym of khas[ is this: <<Jahizalso uses
it, and al-Mas'uidialso attributes its use to him>>27.What M. attempted
to refute is the first part of this short passage, overlooking the second.
We shall treat both of them in the same order.
Al-Jahiz, in his well known passage (or chapter) on the eunuchs
in Hayawvun,mentions Hadij (or Khadij) al-khasi khaidim al-
Muthanna b. al-Zubai'r28. In the light of the evidence from Jawari,
khddimn here can have only one meaning, especially when it is attached to
khasi. This is, however, what I said in the first version of my reply.
Grammatically M. is right, of course, in separating between the two,
but practically he is not so. The combination of khddimand khasf is most
frequent in the sources, even long after the first as equivalent to the

24 Needless to say that Pellat was mistaken when he stated in connection with the
opening words of the khisyDn passage that al-Jahiz did not mention the eunuchs in the
previous pages of our epistle (lam A'adhkurhumfi ma sabaqa min nass al-kitdb, p. 52, note 4
of his Beirut edition).
25 See also below.
26 Op. cit., p. 214.
27 Op. cit., p. 85.

28 Havawdn, I, P. 118.
296 D. AYALON [8]

second had already become not only well established, but absolutely
dominant. The same authors, who usually use only khadim, revert from
time to time to the combination, when the addition of khasFis utterly
unneccesry29. Quite often one will find in one source khadim (or khadam)
and in another khadim khasi (or khadam khisydn) in connection with the
same person or group of persons. Of particular significance is the
following instance: Hilal al-Sab1, speaking of the court of Caliph al-
Muqtadir, states that there were in it 11,000 khadim, whereas al-Khatib
al-Baghdadi mentions them as 11,000 khiidimkhasi. For him every single
khddim of that huge group was a khasi30, and with full justification.
Incidentally, this practice of combining the two terms also helps very
much in establishing their identity.
In contradiction to M.'s view, al-Jahiz's illuminating remarks about
the khasi's aptitude for <<service?)(khidma)31, only strengthens the
connection between the two terms. They give us an insight into why the
euphemism khaddimwas chosen to designate khasi32.
Thus, in each of the two works which served as the main sources for
M.'s study on al-Jahiz's eunuchs, that author employed khadim as the
equivalent of ?eunuch?) (once in the singular and once in the plural).
We shall now turn to the second part of my above cited statement,
which M. overlooked. Since al-Jahiz had been proved to have used the
term in question in the meaning under discussion, there is no wonder that
other Muslim writers would attribute its use to him. That is how one of
them did it.
Al-Mas'iudi brings several anecdotes related by al-Jahiz about a
Romeo and Juliet type of love affairs between slave-girls and slave-boys
in the Muslim courts. The last of them starts like this: <<Al-Jahizsaid:

29 Even in my very limited selection that combination appears quite frequently (<(On

the Eunuchs, etc.>>.See e.g. passages XII (p. 78), XXXI, XXXII (p. 82)). In No. XII I put
al-khasi khddim on an equal footing with khddim khasi, and this is the correct thing to do
from a practical point of view. The closeness of the two is evident. After all, the
development of a language and its expressions is not necessarily subject to grammatical
pedantry.
30 See <<Onthe Eunuchs, etc.>?,passage VII, p. 77.
31 Moussa, op. cit., p. 214.
32 I have already alluded briefly to the connection between khddim (= khasi) and
khidma on the basis of certain data included in the published part of my study (<<Onthe
Eunuchs, etc.>?,p. 83, note 60). In my -Harrun al-Rashid and his Eunuchs>>I am discussing
in some detail al-Jahiz's evidence on the eunuchs and the khidma, and my conclusions are
totally different from those of M. (incidentally, this had been done long before I came
across his Jahiz article). Similarly to khadim becoming the equivalent of khasi the reason
for calling the eunuch ustadh was that he used to teach small children (ibid., p. 90, note 91).
[9] KHADIM 297

'I told this story to Abii 'Abdallh Muhammad b. Ja'far al-Anbari in al-Basra'. He
[al-Anbari] said: 'I shall tell you a story similar to the one you told me. Fa'iq
al-khddirm,who was the mawll of Muhammad b. Humayd al-Tisi33 told me:
'Muhammad b. Humayd was sitting once with his boon companions. A slave-girl
sang behind the curtain (sitira)"' .

What should be emphasized about this quotation is that al-Jahiz

speaks and cites here in the first person. This increases considerably the
chances that our author preserved al-Jahiz's own words. Fa'iq is a quite
frequent name among eunuchs35.
A central aspect of M.'s interpretation of al-Jahiz's khiidim is that he
does not confine himself to that author's terminology. He has a much
wider goal. By means of it he wants to question my whole attempt at
tracing the earliest occurrences of that term in the sense under discussion
(an attempt which he does not present with absolute accuracy). Since
that interpretation had been proved to be without foundation, anything
he wants to base on it belongs perforce to the same category. However,
the examination of his line of argument and his conclusions contribute to
the clarification of the issue of the present debate.
In the last paragraph of his Addenda M. states that I am completely
right (a entierementraison) to point out the uncertainty which dominates
the employment of the term khadim by the authors of the second half
of the third/ninth century, like al-Tabari, or those of the fourth/tenth
century like al-Mastudi36. To say the least, this is not exactly my view.
Here is what I say.
(<Itis very difficult to establish, with any degree of certainty, when the term khadim
started to be used in the sense of eunuch. Some scholars believe it to have been in the
beginning of the 4th/IOth century (see below). It can be proved, however, that this had
happened considerably earlier. The chronicle of al-Tabari is replete with that term,
and includes even instances pertaining to the very beginning of the 'Abbasid period.
[Jahiz (160-250/776-868) also used it, and al-Masuldi also attributes its use to him.i
?Of great interest is the episode of Fakhita, Mueawiya's wife (passage I). The term
khaditn, mentioned in the account of that episode, might belong to the vocabulary of
al-Masiidi, or al-Mada'ini, or an earlier traditionalist who had been al-Mada'ini's
source (even the possibility that it had been used in the reign of Mu'awiya himself

' 33 The commander whom al-Ma'mu-n sent to fight BTbak, and whom that rebel

defeated and killed in 214/929 (Tabari, II, pp. 1099, lines 3-7, 1109, lines 9-1 1; al-Safad1,al-
Waft bil-Wafiqdt, Wiesbaden, 1961, vol. III, p. 29, Ziriqll, A'kim, VI, p. 343).
4 Murij, VII, p. 227 (Pellat's edition, V, p. 19).
3 The famous Fa'iq, the Samanid commander (died 389/999) was a eunuch (kana
khasiyyan minimawdli Ni7hb. Asad Ibn al-Athir, al-Kdmil, Beirut, 1966, vol. IX, p. 149,
line 2). There are, of course, non-eunuch Fa'iqs, but they are not called khddim.
36 Op. cit., p. 214.
298 D.AYALON [10]

cannot be excluded). The same uncertainty prevails about the khaidimsmentioned by

al-Tabarl, and who lived in a period preceding his own time. This ambiguity, however, is
not confined to the term under discussion, but extends practicallv to all, or most, oJ the
terminology relating to the earlYperiod of Islam, which came down to us, in most cases,
through considerably later sources. In fact, this is the basic difficulty in dealing with
early Muslim history in general since we possess for it mainly this kind of compara-
tively late sources. Under these circum.stances,the on/v thing that can be said about our
term is that it appears in the sources in connection wvithevents and occurrenceswhichtook
place from the beginning of the Umavvad period (or even earlier -see the testimony
attributed to Husayn b. Ali, passage XVIII)>>37.

M.'s interpretation of what I say about the account regarding

Mu'awiya and his wife is included in another passage of his criticism,
which will also be cited in this paper, and will be answered there. At the
present moment I shall limit myself to my view of al-Tabarl and later
historians as expressed in the just quoted passage of mine, stating
beforehand that the data contained in al-Tabarl's chronicle about our
term constitute the backbone of the study of the eunuchs in the lands of
Islam during the period which it covers. What I state in that passage is
precisely the contrary to what M. ascribes to me38. The uncertainty as I
see it is not about al-Tabari's own time, but about the period preceding
it39. In the second half of the 3rd/9th century the term khadim
(<<eunuch>>) becomes not only very frequent, but is the predominant
one40. As for the beginning of the 4th/lOth century and later on, there
are no uncertainties whatsoever about the use of khadimby the historians
of that period, including al-Mas'iudi, as M. believes me to think.
The greatest part of M.'s comment on my attempt at dating the earliest
references to kha7dimin the sense under discussion is dedicated to the
Fakhita-Mu'awiya incident, which I mentioned in the passage quoted
above in full. For that purpose he reproduces al-Mas'uidi'sevidence on

3'7 ?On the Eunuchs, etc.?), p. 85. All the italicized words or phrases in this passage are

added. The passage in the original is also accompanied by a considerable number of notes,
which are not repeated here.
38 The extraordinary thing is that M. reproduces in the note the opening lines of the
present passage (up to the word ?earlier>>)(op. cit., p. 214, note 105).
39 In order to remove any doubt, in spite of its being self-evident, I would like to

emphasize that khadims mentioned in al-Tabari's chronicle and in writings of other authors
during the periods preceding their own time, especially if they are prominent persons, are
also undoubtedly eunuchs. The uncertainty is solely about whether the early lost sources
call them already khddim (pl. khadam), or this represents only the terminology of later
extant sources, which copied from them and changed their original wording (see also
below). The insurmountable difficulty of having to reconstruct early Islamic history on the
basis of later sources, which I raise so emphatically a propos the term khadim, will again be
referred to below.
40 This by no means implies that the term khasi became extinct (see also below).
[I1] KHADIM 299

the incident, as well as the evidence of al-Jahiz, which is almost, but not
absolutely, identical.
Each of the two versions will be dealt with separately.
Al-Mas'iudi says:
Wa-dliakara(al-Maaddinianna Mu dwiva b. Abi Sufvdn dakhalaclhata vawmin 'ali
inmrca'atihi Fdllkhitawa-kadnatclhdt 'aql wa-hanmwa-ma'aihuk ha1si wa-kanat mnakshfidut
l-ra s fia-lanniira 'atIia 'ahu a / - k h d di in ghattat ra 'sahd fa-qaa Mu 'dwiva innahu
k h as i fa-qdlat va amir al-miuminin a-tartdal-muthla bihi ahall/atlahu mad/iarramahu
Allaih 'alavhifr,-istar/le'a Mu'dwiva wa- 'alinia anna al-haqq nmaqalathu fri-lam vudk/iil
b/-datl/iii-a 'ai-dharamihi k/iddiman illi kahbi-rain f_ni.van41.

Khadimnis mentioned twice in this excerpt. According to M.'s

reasoning it is more likely that in the first time it means <<servant>,
for otherwise why should Mu'-awiyatell his wife <<heis a eunuch>>,if she
had already known it. On the other hand, M. argues, the lines from the
word istarja'a onwards, where khadim appears for the second time, are
not repeated in any parallel form in the earlier version of al-Jahiz. From
this he concludes that this second reference to khadim seems to reflect
only the language and connotations of the time of the later al-Mas'iidi.
These considerations are unacceptable. First of all, because M.'s view
about the meaning of that term in the vocabulary of al-Jahiz, has already
been proved to be erroneous. Secondly, because they are unreasonable in
themselves (we shall return to this point), and the absoluteness of that
unreasonableness comes into bold relief when one considers the context
in which al-Mas'fidi brings these lines. They are part of a passage in
which he comments on the murder in Damascus of Abui al-Jaysh
Khumarawayhi, the son and successor of Ahmad b. Tiilfun, by his
eunuchs (Dhui al-Qa'da 282/December 895). He says that those who
killed him were khadam min khaclarnihim(<<some khadam of their
khadam>>,i.e. of himself and of Tughj), and he continues: <<Wehave
already exhausted the subject of khadamnbelonging to the [ethnical
groups of] the Suidan, the Saqaliba, the Rum and the Sin in our book
Kitdb al-Zamdn. There we mentioned that the people of al-Sin castrate
(yakhsu-na)their children like the Rum [= Byzantines] [who] do [the
same] to their children. [We also spoke there] about the contradictory
traits of the khisydn, resulting from the cutting of that [sexual] member.
[We have also spoken about] the changes which nature brought upon
them as a result of this [operation] in accordance with what people
told about them and about their characteristics>>.The lines about

41 Murk-j,VIII, pp. 148-149.

300 D.AYALON [12]

al-MadW'ini'saccount on Mu'awiya and his wife follow here uninter-

ruptedly, and immediately after them al-Mas'iudi states that people
distinguish between two kinds [of eunuchs]: the ?deprived one>>[of
his virility] (al-mashlb), and the ?one whose sexual organ is cut
off>>(al-majbuib).He then expresses his support to the standpoint of
Mu'awiya's wife, namely, that the fact that the eunuchs are beardless and
without their male reproductive organ does not turn them into women.
They remain males (rijal, dhukiur).Al-Mas'iudi concludes his expose in
referring to the fact that the armpits of the khadam do not exude bad
odor (repeating the word khadam twice in this connection), and declaring
that this is one of the virtues of the khadam42.
Now how on earth can khadim in the sense of ?servant>>infiltrate into
such a context? Whereas M. contends that nothing indicates in the
Mas'tdi-Mada'ini account that khadim is equivalent there to khasi
(< Rien n'indique,etc...>>.? the fourth paragraph on p. 213, 11.2-3 of that
paragraph), the truth is that everi'thing indicates it. Furthermore, this
passage alone refutes M.'s assertion about uncertainty in the use of
khddimby al-Mas'adi (to say nothing about numerous other instances in
his Muru7j)43.The recurrence of khadim as equivalent to khasi in the
historical and related literature of the 4th Hijra century and the centuries
which follow is overwhelming in its frequency and its clarity.
Even when the account attributed to al-Mada'ini stands alone, the
claim that in a passage of about four lines, Khadim, which is mentioned
there twice in the same connexion, means ?servant> in the first time and
khasi in the second, does not seem to have any foundation. Furthermore,
it is utterly inconceivable that Mu'awiya would be so inconsiderate and
heedless that his wife would suspect him of bringing an adult unemascu-
lated male to her private quarters, and at a moment when she had been
barefaced. If that had been the case, he would have entered first and
warned her. Even with such a forewarning it is extremely doubtful that
under ordinary circumstances a Muslim would allow a man to enter the

42 Murulj,VIII, pp. 147-8 (in Pellat's edition: V, pp. 151-2). Al-Tabari recounts the

murder of Khumarawayhi in these words: dhabahahu ba'd khadamihi min al-khdssa ...
wa-qutila min khadamihialladhina uttuhimu7 bi-qatlihinay if wa-'ishrunakhddiman(Tabari,
III, p. 2148, lines 15-18). Al-Tabari, like on numerous other occasions in his chronicle, does
not deem it necessary to explain here the meaning of our term, either in the singular or in
the plural. He was absolutely certain that his readers would know that eunuchs are meant.
" In the Khumarawayhi-Mu'awiya passage al-Mas'fudiemploys repeatedly khaddim(p1.
khadam) as equivalent to khasi (pl. khisydn) in connection with all the ethnical groups, even
before their entrance into Dar al-Isldm. See also note 21.
[13] KHADIM 301

secluded apartment of his wife. All this excludes the possibility that the
Caliph's wife could mistake her husband's companion for an uncastrated
servant. We have also to remember that we are not dealing with what
crossed the mind of Mu'awiya's wife when she saw her husband's
associate but with the terminology of the transmitters of the story. And
at least about the last of them, al-Mas'tudi,there cannot be the slightest
doubt44.
On top of all that, M.'s interpretation is contradictory to the very
purpose of the anecdote. The Caliph's retort to his wife was not at all to
dispel her fear about the masculinity of his companion. The only thing it
reflects is their contrasting views about eunuchs in the harim; and the
only reason that al-Mas'fudicalls that companion, within seven words,
once khddim and once khasi, is to avoid repetition.
As far as khadim vis-a-vis khasi is concerned, the incident just
discussed, together with the whole passage in which it is included, serves
as a good example for the employment of both words. Although khadim
became more and more dominant, the original khasi by no means
disappeared. It recurs in the sources quite frequently, and in passages
where ?eunuch>>has to be mentioned repeatedly (like in the Khumara-
wayhi-Mu'awiya one), they alternate the two words45. This greatly
facilitates the task of establishing their identity.
Yet another, and very interesting, aspect of the Mu'awiya incident is
revealed by M. As already stated, he quotes another version of it which
he found in al-Jahiz's writings, and which is also attributed to the same
al-Mada'ini. In that version khadim is not mentioned at all. Since al-
Jahiz was al-Mada'inl's contemporary he argues his citation should
be considered the reliable one, and not that of the later al-Mas'uid46.
This piece of evidence from al-Jahiz was unknown to me. However, it
still leaves insurmountable the difficulty caused by the lost early sources
which I have clearly mentioned, and to which I shall return in the present
paper. With the absence of the original works of al-Mada'ini, one can
argue, with the same degree of credibility, that these are not identical
accounts, but similar versions. Such versions of traditions grouped

4 The Mu'awiya incident reflects a certain revulsion from the introduction of eunuchs
into the harim. We know of short-lived attempts to mitigage that practice by letting into the
women's quarters only over age or under age eunuchs (i.e. below or above the age of the
virility of the unemasculated male).
4 See e.g. (<Onthe Eunuchs, etc.?), passage XXI, p. 80.
46 Moussa, op. cit., p. 213.
302 D.AYALON [14]

together in the same source, or scattered over numerous sources, and

with varying degrees of differences (including differences in wording) are
very common in Muslim writings. Al-Mada'ini was a prolific writer, and
both al-Jahiz and al-Mas'tudi could pick non-identical versions either
from the same or from different works of his. What supports that
possibility is: a) the fundamentally different language of the two
versions47; b) the fact that in al-Jahiz's rendition the wife in question is
Maysiun,the daughter of Bahdal, and in that of al-Mas'iudiit is Fakhita,
the daughter of Qaraza; c) the existence of an account about a very
similar incident connected with the sister of Husayn b. 'Ali48.
The introduction of emasculated males into the harim must have
caused hesitations and reservations, especially in the beginning; and
traditions of antagonism to it connected with this or that personality is
the natural thing to expect. A more thorough reading of the sources
might well bring an additional crop of similar traditions. Like the three
versions just referred to, they will, in all probability, differ both in
language and in details of varying degrees of importance. All this serves
only to diminish the chances that al-Jahiz and al-Mas'uidi copied from
al-Mada'ini exactly the same account49. Neither is it certain that the first
of them copied more accurately than the second.
We reach now the main implication to be drawn from the khadam of
al-Jahiz in Jawalri. The context of that term's appearance there shows
how widespread its euphemistic use had already become in that author's
time. Al-Jahiz employed khadam without any additional clarification,
because he had no doubt that his readers would know exactly what he
means. He then used khisyainwithout any addition. being again com-
pletely certain that those readers would know precisely to which earlier
passage in his epistle he refers". It is thus made absolutely certain that

The reader can easily verify that fundamental difference, because M. reproduced the
two versions.
48 See <<Onthe Eunuchs, etc.?), passage XVIII, on pp. 79-80.
49 As M. states (op. cit., p. 213, note 103), al-Jahiz repeats the Mu'awiya incident in
Kitah al-Jaw-ri, etc. (ed. Harun), II, p. 125. But there he omits the name of Mu'awiya, as
M. points out. However, he does much more than that: he replaces Mu'awiya by ?one of
the kings? (ba'd al-muluik).He omits also the name of al-Mada'ini. Only a small part of
al-Jahiz's two versions is identical in its wording. Now what is the guarantee that the
version of Kitdb al-Hayawdn is a verbatim copy from al-Mada'ini? The attribution of the
incident to an anonymous king corroborates my view that the same anecdote had been told
in connection with various personalities.
50 There is yet another instructive aspect of the use of the term in our epistle. The one

who employs the euphemism khadam in connection with the eunuchs is sahib al-ghilman,
[15] KHADIM 303

in our author's period khadam in the sense of <<eunuchs>> became so

common and so well rooted (and consequently long in use) that there was
no d-angerof its being confused with khadam in the original meaning of
?servants>>,even when it appeared in the most general and undefined
form. Thus the date of the frequent appearance of our term in its
euphemistic sense can be considerably advanced. In the first version of
my reply to M. I regarded the first half of the third/ninth century as a
safe estimate. In view of the fact that al-Jahiz was born in 160/775 that
estimate can be pushed well beyond that.
It is also worthwhile to link al-Jahiz's khadam with those of some
contemporary, near contemporary and later authors. For the purpose of
my argumentation I shall not deal with them in an exact chronological
order.
The earliest author I can quote is Ibn Khurdadhbih (205 [or 211]-280
[or 300]/820 [or 825]5'-893 [or 911]) who compiled his geography in
232/846 and made a number of additions to it in 272/885 52I shall
reproduce here two short passages.
They run as follows: a) <Jazirat al-Dhahab [= al-Rahib ?]. The
khadam used to be castrated there>>(wa-bihd kdna v ukhsa- al-kha-
dam)53. b) ?What is coming from the Western sea is the khadan1
belonging to the Slav, Byzantine and Frankish [ethnic groups] (wa-

who approves of pederasty with boys and eunuchs. By contrast, sdhib al-jaw ri, who
disapproves of it and is particularly opposed to the inclusion of the eunuchs in the debate,
calls them by the unveiled designation khisadn with all its pejorative connotations.
Pederasty and sexual relations with eunuchs are discussed in considerable detail in M.'s
study (op. cit., pp. 206-209). One reason for al-Jahiz's rare use of khidirn or khadam in
his discussion of the eunuchs is that when one deals with castration and the castrated,
describing the operation and the results ensuing thereof, there is little need for euphemisms.
Another reason is that a major aim of that author was to show that the eunuch is the very
contrary of the perfect model (<<modeleparfait>>)of the male, as M. so aptly puts it (op. cit.,
pp. 210-21 1).
51El2, , p. 839a; Ziriqll, A'Idm, IV, p. 343a.
52 Miquel, op. cit., 1, pp. XXI, 9a. According to Brockelmann Ibn
Khurdadhbih
<oschriebzwischen 230-4/844-8>>(GAL, I, p. 225 (p. 258 of the new edition). I can express no
opinion about the unresolved controversy in regard to the date of the composition of the
said source, and to the authenticity of the version which survived (M. Hadj-Sadok, El2, II,
p. 839b). However, in view of the fact that we all treat it as a major geographical and
historical source, and use it abundantly (with full justification, in the present state of our
knowledge), there is no reason whatsoever to have any reservation towards these two
passages (especially as long as there is no very specific and weighty cause to doubt the
earliness of each one of them).
53 Kilub al-Masalik wal-Manidlik,p. 112, line 8, and also <<On the Eunuchs, etc.>>,p. 76,
passage IVa.
304 D.AYALON [16]

alladhi X'ai min al-Ba/zr al-Gharbi al-khadam al-Saqaliba (or: al-

Saqlabijy uin) wal-Ifranjiyji7n)5
Thus the khadam, who are <(thecastrated ones?>(khisyan) according to
al-Jahiz, are the khadam who are castrated (yukhsd) according to Ibn
Khurdadhbih. Equally there cannot be the slightest doubt that the
Saqaliba and Frankish khadam of passage b from Ibn Khurdadhbih are
as thoroughly castrated as the khadam of passage a from the same
author.
Other comparatively early authors are al-Baladhuri, al-Ya'qiibi and
even al-Tabari. We shall skip them for the moment and go straight to the
evidence of al-Mas'iudi and al-Muqaddasi of the 4th/lOth century.
Al-Mas'uidi, in the passage we have already cited, speaks about
the khadam belonging to the following ethnical groups: Sudan,
Saqaliba, Rum and Sin. All of them undergo the process of castration
(Yakhsufnahum)and all of them are eunuchs (khisydn)55.
Al-Muqaddasi speaks about the various kinds of khadam brought over
to the lands of Islam (Barbar, Habash, Saqaliba, Rum). All of them are
castrated (the verb khasd in the relevant passage is mentioned four times,
and salla once)56.
I cannot see any difference between the khadam of al-Jahiz and those
of the three other authors. All of them demonstrate the perseverance of
that term in that meaning, and these are only very few instances out of
many. Note also the all-embracing character of the term, not limited to
any social group or class, and including the various races.
The evidence just quoted from Ibn Khurdadhbih and al-Muqaddasi
(and corroborated by numerous other instances) is most revealing in yet
another way. The eunuchs mentioned there are called khadam even in the
process of being imported to the lands of Islam (or being the target of
such importation), and, therefore, even before being assigned to any job
or task. Other names for slaves represent the state in which they actually
are: either their state of slavery or their youthfulness (raqiq, 'abid,
mamaclik,ghilmadn,jadari, etc.). This can have only one meaning: these
khadam are not ?servants?>,because at that stage they are jobless. They
must be something else. And that <<somethingelse>>can be nothing but
eunuchs, as proved from the evidence of Ibn Khurdadhbih and al-

54 Kitab al-Masilik wal-Mandilik, Leiden, 1889, p. 92, lines 4-6. Cp. also <<Onthe
Eunuchs, etc.)), p. 76, passage V, no. e.
55 Murtuj, VIII, p. 148 (V, p. 152 in Pellat's edition).
56
Al-Muqaddasi, pp. 242, 1.2-243, 1.3 (see also JSAI, I, pp. 75-6).
[17] KHADIM 305

Muqaddasi and many others, as well as from the whole history of the
term in the Muslim sources57.
We shall return now to the authors whom we skipped.
Ahmad b. Yahya al-Baladhur1, who died well inside the third Hijra
century (279/892 at the latest 58), mentions Faraj al-khddim twice as the
builder of Tars-us and Adhana in the years 171/787 and 194/81059.
Khalifa b. Khayyat calls him Faraj al-khasi60. The instructive thing
about this piece of evidence is that al-Baladhuri did not deem it necessary
to explain khadim in the sense under discussion to his readers. Which is a
decisive proof that in his time that sense had become a matter of
common knowledge61.
The historian-geographer al-Ya'qiubi (died 284/897) wrote his geo-
graphical book quite late in life in Egypt, in 276/889-90 or 278/891. But
his chronicle, which ends with the year 259/872, is considered to have
been written well before his geography, while he still had been in the
East62. Each one of the important eunuchs of Samarra he mentions in
his geographical book is called khddim, and the whole group al-khadam
al-kibir63. In al-Ya'qiubi's earlier chronicle that term is repeatedly
mentioned.
What certainly constitutes the backbone of the whole study of the
eunuch institution in the early centuries of Islam, is the combination of
the chronicles of al-Ya'qiubi and al-Tabari. Al-Tabari's chronicle is, of
course, much richer. But there is more than sufficient data in the first of
the two, to establish that a good number of the individuals who are called
khadim by al-Tabari, are already so designated by the earlier source of
al-Ya'q-ubi.
That all these very numerous khadims are eunuchs can be established
by an argument which had not been mentioned in the published part of
my study, to which M. refers. The overwhelming majority of these
individuals are prominent people. In spite of the fact that most of them
are repeatedly mentioned in the sources, there is no trace (or almost so),

5 See also the examples in passage V, p. 76 of (<Onthe Eunuchs in Islam)), which can
be multiplied.
58 A. Miquel, La geographie humainedu monde musulman,Paris-La Haye vol. 1 (1967),
p. XX. See also C. H. Becker and F. Rosenthal in El2, vol. I, p. 971b.
5 Futi7hal-Bulddn, Beirut, 1957, pp. 231, lines 1-5, 232, lines 4-12 (in the Leiden, 1866,
edition: pp. 168, line 19, 169, line 17, 170, line 2).
60 Ta'rikh Khalifa b. Khayydt, al-Najaf, 1967, p. 481,
line 9.
61 The same is true of other contemporaries of
al-Baladhuri.
62 Miquel, op. cit., p. XXI; Brockelmann, EI1, IV, pp. 1152b-1153a.
63 See e.g. Kitab al-Bulddn, Leiden, 1892, p.
261, lines 10-13.
306 D.AYALON [18]

of any offspring of them64. Suffice it to take any small number of

unemasculated dignitaries and follow their careers in the sources, and the
existence of their offspring would emerge quite quickly.
The implication of this formidable fact is the following one. It is very
regrettable that for the period which precedes the time of the historians
whose works came down to us, we cannot establish for sure whether the
term khadim (= khasi) had already been then in use65. However, any
prominent person of that early period who is called khadimnin later
sources can be considered as eunuch with a very high degree (perhaps
almost absolute) certainty66. This gives us the key for a sure and
unhesitant reconstruction of the eunuch institution in the crucial Muslim
period for which we possess almost no contemporary sources. This fully
applies to the important reign of H-ar-unal-Rashid, on whose eunuchs I
am now completing a study (see notes 15 and 32). All the prominent
khadimnsof that famous Caliph (as of other Muslim rulers) are never
mentioned as having children.
A no mean factor which helped the newer meaning of khadim to
overshadow the original one is this. With its constant spread at the
expense of other synonymous terms, more and more anonymous indi-
vidual, as well as groups of, eunuchs were designated by it. By repeated
use, however, khddim lost its euphemistic character (or, at least, most of
it), thus defeating the purpose for which it came into being. In spite of
that it had not been replaced by another euphemism. That is what quite
usually happens to euphemisms of similar character. The implication of
this phenomenon is that there would be a very great reluctance to use
such terms in their original (usually literal) sense, lest people would
confuse them with the prevailing (or at least very common) basically
offensive meaning. Consequently, there is every chance that an un-
emasculated man servant would be believed to have been a eunuch, if he

64 There are both khddimsand khasis, who are called Abfuso and so. But these are non-
existent sons. Some of the khddims or khasis have even wives and concubines. All this was
given them in order to mitigate their mutilation.
65 There is a reasonable possibility that it had been. It is hard to believe that later
sources had been so systematic as to replace every single khasi of earlier sources by khadim.
Thus the existence of a residue of khddimsfrom the earlier sources should not be ruled out.
This supposition is considerably strengthened by al-Jahiz's use of khddim and khadam in
that sense.
66 The number of those earlier khddims,who are called also khasi is not negligible. And
this can serve as an excellent Stichprobefor verifying the identity of the two terms. The only
reason for designating a prominent person with the not too flattering title of khddim is to
avoid using a much worse designation.
[19] KHADIM 307

had been called khddim. The same goes for the term ?eunuch>>,the
original meaning of which is ((bed chamber attendant>>(the Greek
eunoukhos). A virile bed chamber attendant would not be called ?eu-
nuch>>.For the same reason ?rest room>>,?water closet>>,bajt al-rdha
and the numerous other euphemistic appellations of the ?privy>>in the
various languages, would be hardly used in their literal sense.
One should not infer from the above series of arguments that every
single khadim encountered in the historical and related sources is a
eunuch beyond any shadow of doubt, especially if he is not a prominent
person. Such an absolute certainty does not exist with regard to any term
which underwent a similar process of development. I have already
stressed this point several times in my study <<Onthe Eunuchs in Islam>>,
and I deal with a number of specific instances in ?Hariun al-Rashid and
his Eunuchs>>.My sole aim was to demonstrate how overwhelming are
the chances that that term means ?eunuchs>>in this kind of sources. The
absolute assertion found in a work of Ghars al-Ni'ma Muhammad, the
son of Hilal al-Sab! (died 480/1087), that k h d dim does not have a child
(khddim lI yakun lahu walad)67 in the sense that it is impossible for a
khddim to have a child, should be interpreted in this way. That is how I
actually interpreted it:
<(It[khddimas equivalent to khasi] had become dominant to such a degree ... that it
had relegated the original meaning of this common word to a very secondary place>>68.

Thus, even as late as the eleventh century, where the predominance of

our return in the sense under discussion is generally accepted, there is no
100 percent certainty concerning each single case.
In concluding this reply we shall return to al-Jahiz. His employment of
khadam as the synonym of khisydn in Jawari, and the circumstances and
context in which he did it, necessitated a thorough reevaluation on my
part of the contribution of that author to the proper understanding of
that term and its historical development69. The reevaluation is, however,
in the very opposite direction to M.'s conclusions.

67 JSAI, I, pp. 84-5. Al-Hafawdt al-Nddira, Damascus, 1967, p. 47, 11.9-110.

68 Ibid., pp. 84-5. See also p. 84 above.
69 On the basis of the data available then to M. and myself the evaluation in the first

version of my reply was completely justified. See also the detailed remarks in note 50. The
importance of al-Jahiz's contribution is mainly qualitative. That of al-Tabari is quantitative
as well. Yet al-Jahiz as a source for the study of the place and functioning of the eunuchs
within the socio-military institution is of a quite marginal importance. An important aspect
of the evidence of al-Jahiz, Ibn Khurdadhbih, al-Baladhuri, al-Ya'qub! and al-Tabarl is
308 D.AYALON [20]

This is as far as the reply is concerned. A remark of a much more

general character has to be made. The clarification of the meaning of
the term Khadim forms the pivot of the whole study of the eunuch
phenomenon in Islam during the Middle Ages. Establishing its un-
70
equivocal sense is, therefore, of the highest possible importance .
D. Ayalon

that the lives of all these authors are overlapping. Thus we have an unbroken continuity in
the study of our term.
70 Remarks made in the present paper without the mention of sources are based partly
on data scattered in the already published chapter of my study on the eunuchs and used
there for a different purpose, and partly on forthcoming chapters. In those chapters the
important contribution of C. Pellat to the study of the subject in his article Khasi in El2 is
acknowledged.