On the placing of ṣ in the Maghribi abjad and the Khirbet al-Samrāʾ ABC

journal of Semitic Studies XXXVII/2 Autumn

ON THE PLACING OF S IN THEMAGHRIBI ABJAD AND THE

KHIRBET AL-SAMRA' ABC

M. C. A. MACDONALD

UNIVERSITY OF OXFORD

In 1986 I published a study of all the known letter-ordersof Arabic and the ancient North Arabian scripts.1 In it Ishowed that the earliest letter-order of Arabic is the so-called'Maghribi' abjad which shares several features with the muchearlier Thamudic E2 abecedary (KnA) found at Khirbet al-Samra'. Both are based on the Aramaic letter-order, withcertain significant differences, and both place at the end of thealphabet the rawddif, or 'extra' letters representing phonemeswhich occur in North Arabian and Arabic, but not in Aramaic.

A curious feature in both these abecedaries is the position ofs, which is placed after «, in the position of Aramaic semkat. Inmy previous article I argued that the Maghribi abjad repre-sented the conflating of two traditional orders stemming fromthe falling together of \z\ and /s/ in Phoenician (see Fig. 1).

Ugaritic n % s p s qAramaic n s p s qThamudic E (KnA) n s [ ]3 ' / qMaghribi abjad n s f d q

Fig. 1

In one, the letter / (representing \z\ + /s/) could have been

1 See Macdonald 1986, particularly 1 j 7 ff. Since I shall be discussingphonemes which, though etymologically identical, were realised as differentsounds at different periods, I have used // for phonemes and [ ] for theirphonetic realisation. For the alphabetic signs which expressed these phone-mes I have used letters in italics (e.g. /) or letter-names (e.g. sad).

2 In Macdonald 1986 I called this script 'South Safaitic' following E. A.Knauf. For my reasons for returning to Winnett's original label 'ThamudicE', rather than his later term 'Tabuki Thamudic' or KnauPs 'SouthSafaitic', see Macdonald 1991: n. 10.

3 Inexplicably, r is placed between / and ' in this ABC.

M5

at Bodleian L

ibrary on Decem

ber 9, 2014http://jss.oxfordjournals.org/

Dow

nloaded from

http://jss.oxfordjournals.org/

S IN THE MAGHRIBI ABJAD AND THE KHIRBET AL-SAMRA' ABC

placed after n, i.e. in the position of Ugaritic £, while in theother, it occupied the place of Ugaritic / (i.e. between p and q).The first tradition would be represented by KnA and thesecond would be that of the well-known Phoenico-Aramaicorder.4 In the Maghribi abjad, therefore, sad was given itsposition according to the first tradition (after nun), and its'formal derivative'5, dad, was placed in the position of saccording to the second tradition (i.e. after pjf). In this it isunique among the formal derivatives, the remainder of whichmake up the rawddif added on at the end of the abjad.

However, as I recognized in that paper, the evidence for my'first tradition' is extremely slight, there being no trace of itbetween Ugaritic and Thamudic E. While this lack is not quiteas damning as it may seem, and there are a certain number ofarguments in favour of this tradition,6 its existence canobviously be no more than an hypothesis. Given this uncer-tainty, other explanations for this arrangement are to bewelcomed.

It is well-known that in a minority of the Latin, Greek andAramaic loan-words in Arabic, s, sigma and semkat are trans-cribed by sad. While some of these borrowings are clearly early,it is by no means possible to date all of them, even roughly,just as it is equaliy impossible to date the introduction of most

4 See Macdonald 1986: 108-noarid 117-119.5 In my 1986 article I classified those letters in the Arabic script which

are identical in form into 'homomorphs' and 'formal derivatives' (Mac-donald 1986: 148, n. 119). Homomorphs are those letters whose forms hadbecome identical in the late Nabataean prototype of the Arabic script, thusI_J and O, r and r-, j and j , (see Macdonald loc. cit. for a fuller discussion).Formal derivatives, on the other hand, were created to give separateexpression to pairs of Arabic phonemes which in the Nabataean script hadbeen represented by a single letter. Thus i from J, r- from r-, J» from i>, j-from £, ^ from, ^ô, J^ from ,_,-, J J from O. Where no diacritical pointsare employed these formal derivatives are identical to the base-forms fromwhich they were created. However, the fact that, with the exception of ^y),the derived forms were collected together at the end of the abjad shows thatthey were considered to be distinct letters, albeit of identical shape to theirbase-forms within the abjad itself. This is a quite different concept, there-fore, from that of the Nabataean script where a single letter represented twophonemes. Compare the English letter c, which represents a number ofphonemes, with the letter /, which represented both a vowel and aconsonant until the 17th. century when a developed 'formal derivative', j ,came to be used to distinguish the consonant, ; being left to represent thevowel.

6 See Macdonald 1986: 109-10, 117-19.

156

at Bodleian L

ibrary on Decem


Dow

nloaded from



of the much larger number of words in which these letters aretranscribed by sin.

However , a small number of Greek and Latin names haverecently been found in the Safaitic and Thamudic E inscrip-tions, and these t h row some light on this problem. In themajority of them, Greek sigma and Latin s are represented by /in Safaitic and Thamudic E , a l though in Nabataean and Palmy-rene the same names appear with semkat. The same applies tothe title KocTaap.7 T h u s : -

Greek Nabataean Safaitic/Thamudic E

KocTaap -10'p8 qsr9

'HpcpSr)? [NTTin]10 hrds"[ND-)in]12 grfsn

7 It seems probable that the Semitic title came into Aramaic, Safaitic andArabic, via the Greek, rather than direct from Latin Caesar. See Angeli1970: 60-63, 88, 103-105.

8 See Cantineau 1930-32: ii, 143; Angeli 1970: 103-105; to which addYadin 196a: 241-42, 246. The spelling qysr probably occurs in CIS ii 170/2.CIS reads qysr, but de Voyiie's reading (1868-77: 100), qysr, supported byNoldeke (1865: 639), is surely confirmed by the photograph. The samespelling, qysr, has now been found, in the phrase b snt ts' I qldys qysr, in aNabataean inscription from Sur in the Leja, and here the reading qysr isobsolutely clear, see Starcky 1985: 180. In Palmyrene the form is qsr, seeAngeli, loc. cit.

9 See SIJ 88, WH 1698, 1725b, Ms 44, 48-54, and an inscription foundby the Basalt Desert Rescue Survey. See Macdonald 1993 on this last textMs 44. When the word was first found Milik suggested that it representedcastra rather than xaurap (i960: 179). However, the recent discovery of theexpression bn qsr, 'the son of Caesar', in Ms 44, as well as the otherexamples of Safaitic s for sigma\semkat, renders this suggestion improbable.

10 Milik suggests the Aramaised form, Sinn, as the patronymic in CIS ii406, though with reservations (1976: 148, 150). It is difficult to see howsuch a reading is achieved from either of the copies published in theCorpus, but Milik may well have worked from the original.

11 In a text from the Basalt Desert Rescue Survey. For a discussion ofthe text see Macdonald 1993.

12 LP Nab 102. Littmann reads 0D11H, but Starcky corrects this to XDlin(1985: 175).

13 See King 1990a: 62 and PI. Ila, and Macdonald 1993.14 LP Nab 101 and Milik 1976: 149. It also occurs in an unpublished

Nabataean inscription from the Wadi Feiran area in Sinai.15 Ms 44 and an inscription found by the Basalt Desert Rescue Survey.

See Macdonald 1993.

157

at Bodleian L

ibrary on Decem


Dow

nloaded from



nqtsxl

*Aa(iCTai/'Aa(jia(Ti18 '<0»"T1 9 : • dmsj™ •

In addition, the Latin name Germanicus appears in Palmyreneas grmnqws21 but.in Safaitic as grmnqs.22 • .

On the other hand, Greek. Tzwpyioc, (Nabataean DW)appears in. Safaitic23 as_grgj-;, while the Latin names Claudius andTitus are found as tts1 and 'qlds1 respectively.24 In Thamudic E,wrqnsx could represent 'Ypxocv6<; or Oupyevioi;.25 Other Greeknames are found either without the nominative ending,26 or inan Aramaised form- with the ending — \2 7

In some of the earliest Greek, Latin and Aramaic loan-wordsin Arabic, s, sigma and semkat are transcribed by sad. Thusqaysar, from Kafaap or via Aramaic ICp,2 8 is found not only in

16 CIS ii 303 (JS N a b 106) which , if the reading is correct , probablyrepresents Nixr]-nr)i; o r NIXTJTIOC;.

17 K J C 28 (in K i n g 1990b). Pace K i n g , this name is surely more likely torepresent NixT]T/)t; than NOOJTIOI;.

18 Nei ther of these forms is at tested and bo th are reconstruct ions fromthe Nabataean name. Milik and Starcky interpreted the name as a hypoco-ristic of Aa[xa(Ti7C7to? (1970: 142). Winne t t assumes the form *Aapux<Ti (1973),bu t Milik suggests *Aa(xa'ai (1976: 149): Jaussen and Savignac (1909-22: i,224) took it to represent Damasius. •

•.' J S N a b 84, and see Milik and Starcky 1970: 142.20 SIJ .287 , 823, (on which see Winne t t I973) , .S IAM.36 .21 CIS ii 3913, Palmyrene line 103.22 L P 653 in which the dat ing formula mus t surely read snt lgy[n] grmnqs

b nq't ' the year the t roops of German icus were at N q ' t ' . See Macdonald1993 for the justification of this re- interpretat ion. '

23 K i n g 1990a: 65, n. 4. .24 Tts1 in C 2308, 2309, ISB 176, N S R 47, 'qlds' in W H 837 and K i n g

1990: 58. L i t tmann suggests that the extremely c o m m o n Safaitic name slivrcould represent Latin Severus (1940: 104). H o w e v e r , as G. Ryckmans(commentary to C 87) and Sartre (1985: 237) poin t ou t it is more likely torepresent Arabic sawwdr (misprinted as fawwdr in Sartre).

25 Wrqns1 is in T I J 208. F o r t he G r e e k n a m e s see Pre is igke 1922: 451and 249 respect ively.

26 T h u s tlmy (C 567 [?], 5298, SIJ 889, a n d several unpub l i shed texts , cf.J S L ih 45 , 54, etc.) which cou ld rep resen t e i the r T7ToXe(xixroi; o r Semit ictalmay (2 Sam. 3 : 3 , etc.) , and hrkld ( W A M T 60) w h i c h cou ld represen t'HpaxXei8T)<; (Pre is igke 1922: 122-23).

27 T h u s Jtny' ( W H 2833a) f r o m 'A(p66vtoi;, see Mii l ler 1980: 73 .28 The case of Arabic qasr is more complicated. It has usually been

assumed-that it is ultimately derived from Latin castra, via Aramaic K1SJ?,rather than the phonetically more correct form KlBOp.See Jeffery 1938:240. However Conrad argues convincingly for an Arabic etymology (1981:8-11) .

158

at Bodleian L

ibrary on Decem


Dow

nloaded from


S. IN THE MAGHRIBI ABJAD AND THE KHIRBET AL-SAMRA' ABC

ancient North Arabian but in the pre-Islamic poetry.29 Arabicqamts 'a shirt, tunic' from Latin camisia, probably via GreekxocfAicnov, is also clearly an early borrowing since it occurs in theQur'an and quite. frequently in the pre-Islamic poetry.30 Tothese can be added sir at which, derives, ultimately from Latinstrata, almost certainly via Greek. arpara. and Aramaic s'ratjd','istrdtd', etc.31 There is less certainty, about, the date of theborrowing of the rare word dims 'a course, of stone in a wall'(.<• AramOiaH / OinVT < Sofô^32 and oi fan 'a1 stone in aring' ( < Aram. ODOD. < iji?jcpo<;),33 / qiss or jissM 'gypsum,plaster' (ultimately < yuô*;), mastaki 'mastic' ( < Aram-. "ODDO< (locaTix?)),35 and fursah 'an opportunity' ( < Syr. pursd\emphatic state oip'rus < 7ropo<;).36 . , . . . .

Early borrowings from Aramaic include him 'fortress'1, whichJeffery37 argues is to be derived not from Arabic hasuna 'to beinaccessible', but from Aramaic |0n 'to be strong', and moreparticularly-from Syriac hesnd' 'a fortress', and possibly sayqal 'aweapons polisher' ( < Aram. MVÔ, cf.Syr. s'qal 'to polish'),38

both of which are found in the pre-Islamic poetry,39 andpossibly the very rare sariqah 'thin cakes'.40

By contrast, there is a fairly large number of Greek andAramaic loan words in early Arabic in which sigma and semkathave been transcribed by sin. For instance, the names 'I/yds andYiinus are found in the Qur'an, as are the words sigill (<(TtyiXXov < sigillum), simd ( < ayj[i.a), qirtds ( < âp-nrji;), all ofwhich come via Aramaic, and many words from Aramaic itself,such as nushah ( < Nabataean nnoi), masjid (Nabataean

), kursiyy (Aramaic ICOTO), etc.41

29 Fo r instance, Imru', a l-Qays, Ra'iyjah, 34.30 See Jeffery 1938:. 243, and Lewin : 1978: 361. .31 Jeffery 1938: 195-96. • .32 Fraenkel 1886: 12-13.33 Fraenkel 1886: 59-60. . .34 A l so qass, jass, o r juss. A c c o r d i n g t o Fraenke l the form jibs is 'eine

ziemlich spate E n t l e h n u n g und y\)t\ioc,' (1886: 9,, 11.-3).35 F r a e n k e l 1886: 151. .36 F r a e n k e l 1886: 243. •, •37 Jeffery 1938 : 109-10.38 Fraenkel 1886: 254. . - •39 On,him see Fraenkel 1886: 236 and Lewin. 1978: 86; on sayqal see

Fraenkel 254 and Lewin 242. . . .40 See Fraenkel 1886: 186, n. 2,41 On all of these see Jeffery 1938.

M9

at Bodleian L

ibrary on Decem


Dow

nloaded from



If we accept McDonald's adaptation of Beeston's analysis,42

ancient North Arabian43 and Arabic, up to and including theperiod of Slbawayhi, possessed no [s], since North Arabian s1

(Arabic sin) represented 'a kind of post-dental apical sibilant'44

for which McDonald uses the symbol [s], while North Arabians2 (Arabic sin) represented 'an approximation to the German"Ich-Laut"', [c].45 There would therefore have been no pho-neme in Arabic which would have clearly corresponded toLatin s, Greek sigma or Aramaic semkat. According to Murto-nen, Slbawayhi's description of the point of articulation of sinas 'between the point of the tongue and a little above theincisors'46 corresponds closely to that of Finnish /s/ 'which canbe described as standing in the middle between English s and /Qshl) phonemes'.47

As McDonald suggests,48 there was probably 'a wide mea-sure of free variation in the realization' of the sound repre-sented by s1 and sin. This would explain the inconsistencies inthe transcription of [s] in foreign names and loan-words bothin ancient North Arabian and in pre-Slbawayhi Arabic. In somedialects, or idiolects, the sound represented by s1 / sin may havesounded close enough to [s] to seem the natural transcriptionfor sigma and semkat, in others the sound represented by sadmust have appeared more appropriate.

There are several examples of. the confusion of sdde* andsemkat in later Aramaic49 and Syriac,50 which suggests that thepronunciation of the two in the main source languages forthese transcriptions was not always clearly differentiated, atleast at certain times and in certain areas.51

42 Beeston 1962; McDonald 1974, particularly pp. 42-45; and Murtonen1966: 137-38, 142.

43 That is the h- dialects represented in the Dedanite, Lihyanite, Hasaitic,the various types of Thamudic, and the Safaitic inscriptions; and the al-dialects of the Arabic 'behind' Nabataean and Palmyrene, and the pre-Islamic Arabic inscriptions.

44 McDonald 1974: 44.45 Beeston 1962: 223-24; McDonald 1974: 42-43.46 The translation is McDonald's (1974: 42).47 Murtonen 1966: 138.48 McDonald 1974: 44.49 See, for instance, Dalman 1905: 104; Kutscher 1976: 17; Levias 1900:

10; Schulthess 1924: 29.50 See, for instance, Payne Smith 1879-1901: 2489, 3349.51 In Mishnaic Hebrew also there is variation in the use of 0 and X both

160

at Bodleian L

ibrary on Decem


Dow

nloaded from



It is therefore probable that ;, which may well have been de-emphasized,52 was considered, by some at least, to be thenearest sound available to transcribe [s].53 If this was so,/ would have been associated with semkat in the minds ofsome of those who wrote the ancient North Arabian dialects,including the author of the Thamudic E abecedary. The samewould be true of some speakers of early Arabic and inparticular those who created the Maghribi abjad.

The writer of the Thamudic E alphabet (KnA) thus had achoice of two positions in which to place s. It is likely that hechose to place it after « simply because this was the first of thetwo positions to occur. It is clear that his grasp of the Aramaicletter-order was not very strong,54 and he may well not havefelt the consequent 'gap' between / a n d q.

On the other hand, those who devised the Maghribi abjadwere clearly very familiar with the Aramaic letter-order andwould have been conscious both of the lack of a true equiva-lent to semkat and that the proper position of sad was betweenfa' and qdf. They therefore chose a compromise, which was notof course available to the writer of the Thamudic E abecedarywho was using a quite different script. They placed sad in theposition of semkat, possibly as representing the nearest equiva-lent to it in sound and one of the traditional transcriptions ofit, and placed sad's formal derivative, dad, between fd' and qdf,rather than in its proper place among the rawddif.

The impetus for this must surely have been a desire tomaintain parity with the numerical values of the letters in the

in native words and in the transcription of sigma and / in Greek and Latinwords, see Segal 1958: 32-33.

52 The sporadic convergence of [s] and [s] can be found in severalArabic dialects. For instance, Dussaud and Macler, noting the variations inthe spelling of the place-name Salhad / Salhat / Salkah, etc., remark that'dans la region du Hauran la confusion du sin et du sad, cf. Busr pour Busr,est d'ailleurs frequente' (1903: 22-23). Cantineau found it not only in theHawran but at Aleppo (Cantineau 1946: 103); while according to Blau [s] /[s] variation is also common in Christian Arabic (1966-67: 109-13), etc.

53 On the basis of the / / semkat correspondence in KnA, Knauf suggeststhat while 'aram. /s/ noch glottalisiert war, siidsaf. [i.e. Thamudic E] /s/hingegen schon velarisiert' (1984: 121). This seems to me to press theevidence too far. The most that this correspondence can be assumed tosuggest is that the author of KnA considered Thamudic E /s/ to be closer issound to Aramaic /s/ than any other phoneme in his dialect.

54 See Macdonald 1986: 112.

161

at Bodleian L

ibrary on Decem


Dow

nloaded from



Hebrew/Aramaic/Syriac order. The lack of a true equivalent tosemkat threatened to disrupt the numerical sequence and it wastherefore necessary to fill its place with the letter most closelyassociated with semkat. The two letters with which it wascommonly transcribed,. sad and sin, already had places of theirown in the sequence, as equivalents to other Aramaic letters (Xand B*). To put either of them in semkat's place would be toleave a gap later in'the sequence.

However, one solution to this was to fill the resulting gapwith the formal derivative of the. displaced letter, which wouldotherwise have been placed among the rawadif. This is exactlywhat was done. Sad was put in the position of semkat and itsformal derivative, dad, was put in its place between fa' and qif.The letter dad had developed simply because Arabic /d/ hadbeen transcribed by side' in the Nabataean ancestor of theArabic script, and before the development of diacritical pointsit would have been identical in form to sad. Thus the letterrepresenting /d/ would naturally have been associated with thenormal position of/ in the alphabet, i.e. between p/f and q. Dadis the only one of the rawadif which could have been movedinto the main body of the abjad in this way since it was the onlyletter for which a suitable space, with which it was alreadyassociated, had opened up.

The only other letter which could have taken the place ofsemkat was sin. However, this would have meant removing itfrom its position as, the etymological and, more or less,phonetic equivalent of Aramaic V, and bringing sin fromamong the rawadif to stand in its place. Since by this periodAramaic /§/ (the cognate of the phoneme represented by sin)had fallen under /s/ and the pronunciation of Arabic /in was [c]not [f], the sounds represented by sin and V would have beenquite different. and this (etymologically incorrect) arrangementwould probably not have occurred to the creators of this abjad.It was only after the time of SIbawayhi, when the values of sinand sin had shifted to those in modern Arabic, that they wereplaced in these positions in the latest of the Arabic letter-orders, the Eastern abjad}1 , • . ...

55 See Macdonald 1986: 123-24. I am grateful to Professor A. F. L.Beeston and Dr. E. A. Knauf for their helpful,comments on an early draftof this paper. They are not, of course, responsible for my conclusions.

162

at Bodleian L

ibrary on Decem


Dow

nloaded from


S IN T H E MAGHRIBI ABJAD A N D T H E KHIRBET AL-SAMRA' ABC

S I G L A : . . . . . . .

C Safaitic inscriptions in the Corpus lnscriptionum S'emiticariimParsV. •

CIS Inscriptions in other parts' of the Corpus- lnscriptionum Semiti-.. c a r u m . • . ••. • i . . . . :

ISB . Safaitic inscriptions in Oxtoby 1968.JS Lih Lihyanite inscriptions in Jaussen and.Savignac 1909,-22.JS Nab Nabataean inscriptions in Jausseri and Savignac 1909-22.KJC Thamudic E inscriptions from site C in WadI Judayid,

edited in King 1990b. • ' • ' " •KnA The. Thamudic :E alphabet published in Knauf 1985 and

. discussed at length in Macdonald 1986. . . .LP Nab .Nabataean inscriptions in Littmann 1914. .Ms . Safaitic inscriptions in Macdonald 1991b.NSR Safaitic inscriptions in Abdallahr97O.SIAM ' Safaitic inscriptions in Macdonald 1979.SIJ Safaitic Inscriptions in Winnett 1957:TIJ Thamudic inscriptions in Harding and Littmann 1952.WAMT Thamudic inscriptions in Winnett 1971.WH Safaitic inscriptions in Winnett arid Harding 1978.

REFERENCES:

Abdallah, Y. M.1970 Al-Nuqii! al-safawiyyah ft majmu'at jami'at al-Riyad 'am 1966.

. Thesis presented for the degree, of M. A., American University ofBeirut [unpublished].

Angeli, M. G. B.1970 Nomendatura pubblica"e sacra di Roma nelle epigrafi semitiche. Pubbli-

cazioni dell'Istituto di Storia Antica e Scienze Ausiliarie dell'Uni-versita di Genoya, 7. Genoa. .

Beeston, A..F. L.1962Blau, J.1966-67

'Arabian Sibilants'. JSS 7: 222-233.

A Grammar of Christian Arabic based mainly on Southern-PalestinianTexts from the First Millennium. Corpus Scriptorum Christiano-rum Orientalium 267, 276, 279 Subsidia 27-29. Louvain: CorpusSCO. . ..

Cantineau, J.1930-32 Le Nabateen. (2 volumes) Paris: Leroux.1946 Lesparlers arabes du Hdran. Collection Linguistique publiee par la

Societe de Linguistique de Paris, 52. Paris: Klincksieck.Conrad, L. I.1981 'The qusur of mediaeval Islam: some implications for the social

history of the Near East'. Al-Abhath 29: 7-23.Dalman, G.1905 Grammatik des jiidisch-paldstiniscben Aramdiscb. ' (2nd. edition).

Leipzig: Hinrichs.

163

at Bodleian L

ibrary on Decem


Dow

nloaded from



Dussaud, R. and Macler, F.1903 Mission dans les regions desertiques de la Syrie moyenne. Paris: Impri-

merie Nationale / Leroux.Fraenkel, S.1886 Die arama'ischen Fremdwdrter im Arabischen. Leiden: Brill.Harding, G. Lankester and Littmann, E.1952 Some Thamudic Inscriptions from the Hashemite Kingdom of Jordan,

Leiden: Brill.Jaussen, A. and Savignac, R.1909-22 Mission arche'ologique en Arabie. (6 volumes). Paris: Leroux/Geuth-

ner.Jeffery, A.1938 The Foreign Vocabulary of the Quran. Gaekwad's Oriental Series,

79. Baroda: Oriental Institute.King, G. M. H.1990a 'The Basalt Desert Rescue Survey and Some Preliminary

Remarks on the Safaitic Inscriptions and Rock Drawings'.PSAS 20: 55-78.

1990b 'Early North Arabian Thamudic E: a preliminary descriptionbased on a new corpus of inscriptions from the Hisma desert ofsouthern Jordan and published material'. Ph.D. thesis, (Schoolof Oriental and African Studies, London, 1990).

Knauf, E. A.1984 'Zum Vordringen des Arabischen im Libanon vor dem Islam'.

WO 15: 119-22.1985 'A South Safaitic Alphabet from Khirbet es-Samra''. Levant 17:

204-206.Kutscher, E. Y.1976 Studies in Galilean Aramaic. Bar-Ilan Studies in Near Eastern

Languages and Culture. Remat-Gan: Bar-Ilan University.Levias, C.1900 A Grammar of the Aramaic Idiom contained in the Babylonian Talmud

with constant reference to Gaonic literature. Cincinnati: Bloch.Lewin, B.1978 A Vocabulary of the Hudailian Poems. Acta Regiae Societatis

Scientiarum et Litterarum Gothoburgensis, Humaniora, 13. G6-teborg: Kungl. Vetenskaps- och Vitterhets-Samhallet.

Littmann, E.1914 Nabataean Inscriptions from the Southern Haurdn. Publications of

the Princeton University Archaeological Expeditions to Syria in1904-05 and 1909, Division IV, Section A, Leiden: Brill.

1940 Thamud und Safa. Studien %ur altnordarabischen lnscriftenkunde. Ab-handlungen fur die Kunde des Morgenlandes, 35/1. Leipzig:Brockhaus.

Macdonald, M. C. A.1979 'Safaitic Inscriptions in the Amman Museum and other Collec-

tions I'. ADA] 23: 101-19.1986 'ABCs and Letter Order in Ancient North Arabian'. PSAS 16:

101-68.1991a 'Hu 501 and the Use of s3 in Taymanite'. JSS 36: 11-36.

164

at Bodleian L

ibrary on Decem


Dow

nloaded from


J" IN THE MAGHRIBI ABJAD AND THE KHIRBET AL-SAMRA' ABC

1993 'Herodian Echoes in the Syrian Desert'. In S. Bourke, J.-P.Descoeudres, A. Walmsley (Eds.). Trade, Contact and the Move-ment of peoples in the Eastern Mediterranean: Papers in Honour of]. B. Hennessy. Mediterranean Archaeology Supplement Series [inthe press].

19.. 'North Arabian Inscriptions and Drawings I'. Arabian Archae-ology and Epigraphy [forthcoming].

McDonald, M. V.1974 'The Order and Phonetic Value of Arabic Sibilants in the

"Abjad"'./5T 19: 36-46.Milik, J.T.i960 Review of Winnett 1957. Syria 37: 178-81.1976 'Une inscription bilingue nabateenne et grecque a Petra'. ADA]

zi: 143-52.Milik, J.T. and Starcky, J.1970 'Inscriptions nabateennes'. Pp. 141-60 in Winnett, F.V. and

Reed, W.L. Ancient Records from North Arabia. University ofToronto Near and Middle East Series, 6. Toronto: University ofToronto Press.

Muller, W. W.1980 'Some Remarks on the Safaitic Inscriptions', PSAS 10: 67-74.Murtonen, A.1966 'The Semitic Sibilants'. JSS 11: 135-50.Noldeke, T.1865 'Bemerkungen zu den von de Vogiie herausgegebnen Nabatais-

chen und Hebraischen Inschriften'. ZDMG 19: 637-41.Oxtoby, W G.1968 Some Inscriptions of the Safaitic Bedouin. American Oriental Series,

50. New Haven, CT: American Oriental Society.Payne Smith, R.1879-1901 Thesaurus Syriacus. Oxford: Clarendon.Preisigke, F.1922 Namenbuch en thai tend alle griechischen, lateinischen, dgyptischen,

hebraischen, arabischen und sonstigen semitischen und nichtsemitischenMenschennamen, soweit sie in griechischen Urkunden (Papyri, Ostraka,Inschriften, Mumienschildern usw) Agyptens sich vorfinden. Heidel-berg: privately published.

Sartre, M.1985 Bostra: des origines a I'Islam. Bibliotheque Archeologique et His-

torique, 117. Paris: Geuthner.Schulthess, F.1924 Grammatik des christlich-paldstinischen Aramdisch. Tubingen:

Mohr.Segal, M. H.1958 A Grammar of Mishnaic Hebrew. (Corrected reprint of the 1927

edition). Oxford: Clarendon.Starcky, J.1985 'Les inscriptions nabateennes et I'histoire de la Syrie meridionale

et du nord de la Jordanie'. Pp. 167-181 in Dentzer J-M, ed.,Hauran I Kecherches archeologiques sur la Syrie du sud a I'epo-

165

at Bodleian L

ibrary on Decem


Dow

nloaded from


S IN THE MAGHRIBI Ah]AD AND THE KHIRBET AL-SAMRA' ABC

que hellenistique et romaine, i. Paris, Bibliotheque Archeologiqueet Historique de PInstitut Francois d'Archeologie du ProcheOrient, 124, Paris: Geuthner.

Vogue, M. de • : ' . , " ' . , .1868-77 Syrie centrale. Inscriptionsse'mitiques. Paris: Baudry.W i n n e t t , F . V . . ' " • • . • ' ' . , . . / '1 9 5 7 Safaitic Inscriptions from Jordan. University of Toronto Near and

Middle East Series, 2. Toronto: University of Toronto Press.1971 'An Arabian Miscellany'. AION 31: 443-54.1973 'The Revolt of DamasI: Safaitic and Nabataean Evidence'.

BASOR 211: 54-57,Winnett, F. V. and Harding, G. Lankester1978 Inscriptions from Fifty Safaitic Cairns. University of Toronto Near

and Middle East Series, 9. Toronto: University of TorontoPress.

Yadin, Y.1962 'Expediton D - The Cave of Letters'. IE] 12: 227-57.

166

at Bodleian L

ibrary on Decem


Dow

nloaded from


On the placing of ṣ in the Maghribi abjad and the Khirbet al-Samrāʾ ABC

Documents