Supplement to the proceedingS of the Seminar for arabian StudieS volume 40 the development of arabic aS a Written language Papers from the Special Session of the Seminar for Arabian Studies held on 24 July, 2009 edited by M.C.A. Macdonald Seminar for arabian StudieS archaeopreSS oxford 2010 This Supplement is available either as a set with volume 40 of the Proceedings or separately from Archaeopress, Gordon House, 276 Banbury Road, Oxford OX2 7ED, UK. Tel/Fax +44-(0)1865-311914. e-mail bar@archaeopress.com http://www.archaeopress.com For more information about the Proceedings, see the Seminar’s website: http://www.arabianseminar.org.uk Steering Committee of the Seminar Dr. R. Carter (Chairman) Prof. A. Avanzini Dr. M. Beech Dr. N. Durrani Dr. R. Eichmann Prof. C. Holes Dr. R.G. Hoyland Dr. D. Kennet Mr. M.C.A. Macdonald Dr. A. MacMahon (Secretary) Prof. K. Al-Muaikel Dr. V. Porter Prof. D. Potts Prof. C. Robin Dr. St.J. Simpson Mrs. J. Starkey (Editor) Mr. A. Thompson (Treasurer) Prof. J. Watson Dr. L. Weeks Seminar for Arabian Studies c/o the Department of the Middle East, The British Museum London, WC1B 3DG, United Kingdom e-mail seminar.arab@durham.ac.uk Opinions expressed in papers published in this book are those of the authors and are not necessarily shared by its editor or by the Steering Committee of the Seminar. Typesetting, Layout and Production: Dr. David Milson This Supplement is produced in the Times Semitic font, which was designed by Paul Bibire for the Seminar for Arabian Studies. © 2010 Archaeopress, Oxford, UK. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of the publisher. ISSN 0308-8421 ISBN 978-1-905739-34-9 M.C.A. Macdonald (ed.), The development of Arabic as a written language. (Supplement to the Proceedings of the Seminar for Arabian Studies 40). Oxford: Archaeopress, 2010, pp. 47–88. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material LaïLa Nehmé Summary This contribution aims at presenting a corpus of epigraphic texts in a script that is “transitional” between Nabataean and Arabic. In order to establish this corpus, the author irst collected all the texts which are dated to between the third and ifth centuries AD, whatever their origin. Secondly, “evolved” forms of characters, which occur in these dated texts, were sought for in undated ones. When identiied, the texts in which these characters are found were included in the corpus. In Appendix 2, a sample of thirty-four texts is presented, together with photographs and facsimiles, of which ifteen are previously unpublished. The forms of the letters are analysed and those which can be termed “evolved” are identiied and described. Keywords: Arabia, Nabataean, Arabic, script The last ten years have witnessed the discovery of increasing numbers of inscriptions of a type that was previously known only from a few texts recorded by earlier surveys in north-west Saudi Arabia.1 These texts have usually been considered to be Nabataean. However, they show features which make them distinct from the Nabataean monumental and non-monumental inscriptions of the irst century AD (called here, conventionally, “classical” Nabataean) as they appear mainly in Дegrā, Petra, and other sites in Jordan. Considering that these particular features appear to indicate a certain degree of development of the script, we have labelled these new texts, faute de mieux, “transitional”, i.e. transitional between the Nabataean and Arabic scripts. This label is not entirely satisfactory because it can be understood only in the context of Nabataean and early Arabic epigraphy, but it is useful, at least provisionally, in order to identify, isolate, and describe these texts. It has been suggested that they should be called “Late Nabataean”2 but this terminology would imply that they are closer to Nabataean than they are to the earliest examples of the Arabic script. This is probably true of some, but not all, of them. The following contribution3 has a very practical For instance during the survey undertaken in 1962 by F.V. Winnett and W.L. Reed (on which see Winnett & Reed 1970). 2 “Tardo-nabatéen,” Robin 2008: 174. 3 This contribution owes a lot to Michael Macdonald, who has not only read and corrected it as the editor of the Supplement would do, but has also made a lot of corrections, suggestions, and comments, including on the reading of some of the 1 purpose and will not deal with the issue which is usually discussed when considering questions related to scripts in this ield: the debate on the origin of the Arabic script. I consider, indeed, that at least in its early stages, the Arabic script did develop from Nabataean, not from Syriac, and that the corpus I shall be presenting is suficiently convincing in itself to obviate presenting the historiography of the debate, the arguments for and against each theory, etc. Instead, I shall concentrate on what seems to me the most important element in the present state of our knowledge, i.e. the texts which have recently been found and the identiication of the “transitional” ones, trying to answer the following question: which texts can be labelled as such and why? A few major contributions have already dealt with the development of the Nabataean script into Arabic and I am very much indebted to them.4 What follows is based on the work they have already done, which has shown me the way. inscriptions. If I were to acknowledge each one of them, his name would appear under each paragraph. I would therefore like to express my warmest thanks to him for reviewing this text so carefully and making it ready for publication. I am however responsible for any mistakes which have remained in the text. 4 Grohmann 1971; Gruendler 1993; Healey 1990–1991; Yardeni 2000, vol. B: 219–263. More recently, Macdonald 2009a; al-Muraykhī, in press. It should also be noted that this contribution deals only with epigraphic material written on stone and will therefore not consider the script on the papyri, which shows “evolved” forms of letters at an earlier date than in the inscriptions. In the analysis of the letters, I shall simply indicate when a similar form of a particular letter exists in the papyri. 48 Laïla Nehmé The deinition of “transitional” texts The establishment of a new corpus of texts requires, at least at an initial stage, the use of objective criteria. Considering that I am interested in the development of the script from Nabataean to Arabic, I irst decided to include in the initial corpus all the texts which can be dated to between the third and ifth centuries AD (I have labelled them “late Nabataean” texts), even if they do not necessarily show evolved features (see Appendix 1 for the list of texts dated to this period). The reason for choosing the ifth century is obvious: it is the last century before the appearance of the pre-Islamic texts from Syria in what is considered to be the early Arabic script, Zebed (AD 512), Jabal Says (AD 528), and Дarrān (AD 568).5 The reason for choosing the third century is less obvious. One may ask, indeed, why not put the starting point of the corpus in the second century, which marks the end of the political independence of the Nabataean kingdom. However, this political event would be irrelevant to the use and development of the script. The main reason is the following: the third century is the period during which there are epigraphic texts which can be regarded as still being written in the “classical” Nabataean script as well as texts which show signs of a development towards something different.6 This can best be illustrated by two texts, which are dated respectively to the beginning and the end of this century. The irst one is CIS ii 963, from Wādī Mukattab in southern Sinai, dated to AD 206 (Fig. 23),7 while the second one is UJadh 309, dated AD 295 (Fig. 48). One can easily see, on the facsimiles, that there are many more “evolved” characters in the text of AD 295 than there are in that of AD 206. Thus, in the text of AD 206, only inal y and the h of mΜh have “evolved” forms, whereas in that of AD 295, this is true of — in order of their appearance in the text — the y, d, š, r, Μ, Ή, inal h, inal t, m, and Κ. All the letters that show “evolved” characteristics having been identiied in the dated texts, the criteria used to include undated texts in the corpus of transitional texts were based on the fact that they contained such letter forms. Several hundred texts have been examined and a selection of those that have been included in the corpus, On these texts, see Robin 2006: 330–338 with previous bibliography in notes. 6 This does not mean that all the later texts are written in transitional characters. 7 The igure numbers in the text do not start at 1 because priority has been given to the sequence of igures in Appendix 2, where the inscriptions are presented by their number in ascending order. 5 all of which are illustrated by photographs and facsimiles, is given in Appendix 2. The following methodology has been used: all the letter forms which appear in texts dated after AD 200 have been drawn and numbered individually, starting usually with the “classical” Nabataean form of the letter. The number of different letter forms does not exceed nine (for the m, for example) but there are usually no more than four or ive. Each letter of each text which appears to contain evolved forms of letters or which is dated after AD 200 was then described in a database, using the numbers attributed to each of the various letter forms. It thus became possible to search for any form of any letter in initial/medial or inal form (when relevant). Examining all the forms of all the letters led to the conclusion that only some of the letters had forms which could be considered as diagnostic and were therefore useful for the classiication. The letters whose forms are not relevant within this corpus are the following: — b: not only because the variations are small but also because these variations do not seem to be systematic; — z: because the letter does not vary signiicantly; — Γ: there are three or four forms of the letter which can be reduced to two, deined basically by how the letter is traced: 1) like a Latin “S”, i.e. a wavy line starting from the top, the letter remaining open on each side (as in UJadh 178, Fig. 40, or UJadh 219, Fig. 41); or open at the top and almost closed at the bottom (as in S 1, Fig. 29, and UJadh 375, Fig. 52); or 2) as a diagonal stroke terminated by a loop — either completely closed or open on the left — at the bottom of the letter (as in JSNab 18 line 2, Fig. 25). The irst form is much more widespread than the second, of which there are only four examples in our corpus. Form 1, traced like an “S”, is the one usually found in the cursive script of the papyri of the early second century AD, but there the bottom part of the letter forms a loop which is closed (Yardeni 2000: 247); — l: because the looped form of the letter, which is the normal form in the “classical” Nabataean texts, where it coexists with the unlooped — one could say angular — form, does not occur in the texts dated after AD 200. The only attested form in the latter is the unlooped/angular form, which consists of the right and bottom sides of a rectangle and which does not vary significantly. This is also the form in the papyri (Yardeni 2000: 252); — medial n: because the form of the letter is very consistent throughout the Nabataean and post-Nabataean period with, however, a general tendency to diminish the height of the vertical stem and give it the same height as A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 49 that of the b, with which it becomes easily confused (a good example is in QN 2, Fig. 28); — s: because there are not enough examples of the letter in the texts. Where it does occur, however, the form of the letter is different from the “classical” Nabataean form. It is either open on the left, as in CIS ii 963 (Fig. 23), or closed but with a circular shape on the right and an angular shape on the left, as in UJadh 219 (= ThNUJ 84, Fig. 41). The form of the letter in the papyri is mostly open on the left but tends to become closed for ease of writing (Yardeni 2000: 256); — Κ: because the letter has so many variations that it is almost impossible to trace a clear development of its form. Note, however, that the letter becomes more and more horizontal and tends to sit on the line, as in Κnmw in UJadh 219 (Fig. 41). Having excluded these seven letters from our criteria, we are left with the other ifteen, which have forms which can clearly be termed “evolved” and which, in combination with each other, provide diagnostic criteria for the inclusion of a text in the corpus of transitional inscriptions. — Μ (Fig. 1).8 There are two forms of the letter, which can be considered as evolved forms, numbered 2 and 3 on Figure 1, while 1 is the most ordinary form of the letter in “classical” Nabataean. Form 3 can safely be considered as more evolved than 2. The best example of form 2 is ARNA.Nab 17 (Fig. 21), in Μy, mΜh, and khnΜ, while very good examples of form 3 appear in UJadh 309 (Fig. 48) in Μwšw, ΜlΉ{b/n}h, ktbΜ, and mΜt. For lām-alif, see the end of this section. — g (Fig. 1). The texts show two forms of the letter. The “classical” form is 1 (as in UJadh 360, Fig. 50), and it should be noted that this form is very similar to the evolved form of Ή (form 3). However, there is another form of the letter (form 2), in which the diagonal stroke gets shorter and shorter and inally does not go down beyond the horizontal stroke (UJadh 3, Fig. 32) or only very little (Ar 19, Fig. 20). The last stage of the development of the letter, where the diagonal stroke stops at the horizontal, does not seem to be attested in the papyri. — d (Fig. 1). The texts show three forms of the letter but only form 3 may be considered as “evolved“ because it seems to appear only in texts from the third century onwards whereas 1 and 2 appear throughout Nabataean epigraphy, including in late texts. It should be noted, as Note that the letter shapes given in Figures 1,2, and 4 are a synthesis of all the letter forms, which were found in the texts and are not necessarily exactly how they appear in any particular text. 8 2 3 1 2 ʾ 1 g or 3 1 2 d 2 1 medial h 4 3 1 2 final h 3 2 1 w or 3 2 1 or ḥ figure 1. The forms of Μ, g, d, h, w, and Ή. M.C.A. Macdonald has pointed out, that form 3 is a far more “archaic” form, in terms of Aramaic epigraphy, than forms 1 and 2, which shows that an “evolved form” can develop the appearance of an “ancestral” form. It should also be emphasized that in all the transitional texts, d is clearly distinguished from r. Despite this, in a number of cases, a diacritical dot is placed above it (see the section “Dots on letters” below). A good example of d of form 3, with a dot, can be seen in UJadh 178 in the word dkyr (Fig. 40). In the table that contains the description of the characters for each text (Fig. 18), forms 1 and 2 are put together because they are sometimes very dificult to distinguish in the texts. Laïla Nehmé — medial h (Fig. 1). There are basically two forms of the letter but only form 2 appears in the late texts. Indeed, the “classical” form 1 disappears almost completely from the latter. The only examples are found in CIS ii 963 (Fig. 23), dated AD 206, and RES 528, which is unfortunately known only from a copy (Jaussen, Savignac & Vincent 1905: no. 2, copy p. 239). One incongruous form, in which the letter is closed at its bottom, can be found in ARNA.Nab 17 line 5 (Fig. 21), where the h in medial position has a normal medial form but, as noted by M.C.A. Macdonald, with the base line continued under it from the preceding k. The best examples of form 2 of medial h can be found in phmw in UJadh 375 (Fig. 52) and in yhwdΜ in ΚUlā 1 (Fig. 54). — inal h (Fig. 1). The “classical” form of the letter in Nabataean is 1 while form 3 is a variation of form 2, which is the form usually encountered in the late Nabataean and in the transitional texts.9 Form 4 is a variation of form 3. Form 3 appears for instance in UJadh 297, dated to AD 305–306, mentioned above (Fig. 45), as well as in the Namārah inscription and in UJadh 309 in ΜlΉ{b/n}h (l. 2) and dnh (l. 3) (Fig. 48). UJadh 266 (Fig. 44) is a very good example of form 2. Finally, form 4 is found much more rarely, for instance in UJadh 299 (Fig. 47). — w (Fig. 1). There is no need to give examples of form 1, which is the commonest form in “classical” Nabataean. Forms 2 and 3 are variations of the same, with a general tendency for the letter to be ligatured from the right halfway up the vertical stem of the letter. Note that the form where the ligature meets the stem at the back of the loop is the one that permits the transition to form 3. The two forms which appear under 2 (one where the ligature is attached more or less halfway up the vertical stem of the w, and one where it meets the stem at the back of the loop) should probably have been distinguished in the database. UJadh 219 (Fig. 41) has very good examples of form 2, the most widespread form of the letter in the transitional texts, while the patronym in UJadh 90 (Fig. 36) is the best example of form 3. — Ή (Fig. 1). Form 2 is only a variation of form 1, which is the “classical” Nabataean form. It is attested only in the 9 There are only four late or supposedly late texts which have form 1 whereas there are twenty-eight which have the other forms. In three of the former, several letters in the texts — not just the h — have “archaizing” forms. This is the case of JSNab 17, dated to AD 267 (Fig. 24), which has several other “archaizing” letter forms (Μ, š, t, etc.), of the Stiehl inscription from MadāΜin СāliΉ, dated to AD 356 (Fig. 31), and of the Fihrū text from Umm al-Jimāl, LPNab 41 (Fig. 26), dated to the third century. The fourth text is unpublished and comes from MadāΜin СāliΉ. It may either be earlier than expected or an exception. 3 2 1 medial y 2 1 final y 2 1 k or 6 5 3 4 2 1 m 9 8 7 1 2 final p p or 2 1 odd forms 50 ṣ or 2 1 q figure 2. The forms of medial and inal y, k, m, p, Β, and q. inscription from Wādī Mughārah in Sinai (NDGS 2) and in UJadh 19 (= ThNUJ 34, Fig. 35). One interesting form, which may show how the letter developed, can be seen, as suggested by M.C.A. Macdonald, in the word byrΉ in JSNab 17, line 6 (Fig. 24). Form 3 is very widely attested in the transitional texts, as illustrated by UJadh 309 (Fig. 48), in ΜlΉ{b/n}h (l. 2) and Ήd (l. 5), and in UJadh 298 (Fig. 46), in the name ΜlΉrt. — medial y (Fig. 2). The “classical” Nabataean form, 1, usually consists (from top to bottom) of an inclined stem followed by a loop — with a strong curve — open on the left. In late and transitional texts, the curve of the loop tends to diminish, to be less wide (form 2), and inally disappears, leaving only a more or less diagonal stem. It A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 51 should also be noted that the letter starts to be joined to both the preceding and the following letters. Very good examples of form 2 can be seen in UJadh 309 (Fig. 48), in dkyr (l. 1), ywm (l. 4), and tšΚyn (l. 6). As for form 3, which is probably a further step in the development of the letter, it can be best seen in UJadh 15 (= ThNUJ 30, Fig. 34), in dkyr, whereas in šlym{n}, in the same text, the y is between forms 2 and 3. Form 3 can also be seen in UJadh 178 (Fig. 40). — inal y (Fig. 2). There are basically two forms of inal y. The irst one is the same as medial form 1 but the transitional texts contain almost exclusively the inal y which is shown as form 2. Good examples can be seen again in UJadh 309 (Fig. 48), in bly and šly (l. 1), in btšry (l. 5), as well as in UJadh 405 (Fig. 53), in both names. See also UJadh 3 (Fig. 32). — k (Fig. 2). Form 1 and its variants are much less widespread than form 2,10 which is very similar to the developed form of the d, except for the horizontal line at the bottom of the letter, which does not occur in the d but is an essential part of the letter in the k. M.C.A. Macdonald notes that as with d, the top of form 2 is typical of Imperial Aramaic k, although there the “tail” is a straight diagonal line. Examples of form 2 of k can be found in UJadh 15 (Fig. 34), UJadh 90 (Fig. 36), etc. There are not enough examples of inal k to make any comment. — m (Fig. 2). This letter is complicated to analyse because of the variations in the developed form of the letter. The “classical” Nabataean form is represented by forms 1–3 while the developed form is represented by forms 4–9. Only sixteen texts in our group contain a letter m of forms 1, 2, or 3. In some of these texts, both the “classical” and developed forms of the letter occur in the same text. This is the case, for instance, in UJadh 10 (= ThNUJ 38, Fig. 33), where the m in šlymw and yΚmrw is of form 7 and is ligatured to the left, whereas the m in šmnw is of form 1 and is ligatured from the right. It seems that both forms were perfectly familiar to the writer. The other texts — those in which there are only “classical” forms of this letter — are the following: one is a grafito from MadāΜin СāliΉ, JSNab 18, (Fig. 25); four are formal texts, alNamārah, JSNab 17, (Fig. 24), the Stiehl text, (Fig. 31), LPNab 41, (Fig. 26); four others are grafiti from Sinai11 Twenty examples of form 1 against ifty of form 2. CIS ii 963 (Wādī Mukattab, AD 206), CIS ii 2666 (Jabal Munayjah, AD 218–219) and NDGS 2 (Wādī Maghārah, AD 266), CIS ii 1491 (Wādī Fayrān, AD 232). and are dated, apart from NDGS 2, to the irst half of the third century; the same is true of B 3, from BoΒra (Fig. 22), which is dated to AD 230–231. Finally, there are two texts with a possible transitional character, one from alΚUdhayb, north of al-ΚUlā, and one from Umm Jadhāyidh, which contain only “classical” forms of m. They are not dated but have been considered as “transitional” because of the other letter forms in them (medial h of form 2 for al-ΚUdhayb and possibly Κ for Umm Jadhāyidh). However, the Umm Jadhāyidh text at least (UJadh 360 = ThNUJ 62, Fig. 50), may have to be considered closer to “classical” Nabataean than we thought at irst glance.12 One last text, UJadh 172 (Fig. 39), is particularly interesting. If it is dated to AD 311–312 (see Appendix 2 for the reading of the date) and considering that it does not contain any transitional characters, it would show that the “classical” Nabataean script was still used in north-west Arabia in the fourth century AD. Fifty-six texts — thus many more than in the above category — contain forms of m in which the body of the letter is close to a circle (nos 4 to 9 in Fig. 2). The letter is sometimes ligatured on both sides (forms 4–6) and sometimes on one side only (7–9). Ligatures on both sides: when the letter is ligatured on both sides, the stem which makes the ligature is either at the bottom of the letter on each side (form 4), or in the middle of the letter on each side (form 5), or at the bottom of the letter on the right side and in the middle of the letter on the left side (form 6). Having examined all the examples recorded as form 4 (nineteen examples) and form 5 (sixteen examples) in the database, it appears that the position of the ligature depends partly on the form of the letter m: the latter it is at its base, the more the ligature starts from the bottom of the letter (as in Ar 19, in the m in Κmrw, Fig. 20, or in UJadh 266, in the m of Κmyyw, Fig. 44); the closer it is to a circle, the more the ligature starts from the middle of the letter (as in UJadh 90, in ynmw, Fig. 36, or in UJadh 375, in phmw, Fig. 52). Another reason for this difference may be the letter that occurs before the m. In Fig. 3, middle column, it appears that the m is always ligatured from the preceding letter to its middle part when the letter is a h, a Ή, or a y. As for the mixture of both (form 6, right column in Fig. 3), of which there are nine examples, it appears 10 11 UJadh 360 was considered as “transitional” because of the general form of the name mΚnw (ligatures, the form of the Κ) but in fact, none of the letters is really diagnostic. 12 52 Laïla Nehmé Stem at the bottom on each side (form 4) Letter before l Letter Letter Letter Letter no. no. after before after before {b/n} w 1 contains fourteen examples, in seven of which the letter that precedes the m is an Μ, a d, or a r, which are normally not ligatured to the left (see UJadh 3, Fig. 32). In four examples, the unligatured character of the m is shared by some of the other letters in the text, which would make the case of the m not signiicant (see, for example, the n-m in UJadh 219, Fig. 41, where, however, M.C.A. Macdonald notes that apart from the combinations of letters which quite often lack ligatures — such as k-y in dkyr — only the letters s-p in ywsp and b-Γ in bΓb are not ligatured). Finally, there are two examples in which we would expect the m to be ligatured from the right, after a y and a Ή, but where it is not. Note that there is only one example in which the m in medial position is not ligatured to the left, possibly Ulā-JSNab 386, but the text is known only from a squeeze (the word in question is šmΚwn, l. 3, but M.C.A. Macdonald notes that upright Κ often does not take a ligature from the right) (see Macdonald 2009a: 208 and n. 5). It is clear, therefore, that the letter m is normally ligatured to the left, whatever the letter that comes after it. As for the position of the ligature, there are three cases, represented by forms 7–9. In most examples, the stem of the ligature is halfway up the letter: there are twenty-three examples of form 7 (where m is followed by Μ, h, w, Ή, k, l, n, Κ, r, š, or t) against three of form 8 and only one of form 9. Stem in the middle Stem at the bottom on each side on the right and in (form 5) the middle on the left (form 6) h w 2 Ή y 1 Ή d 1* y w y Letter after {b/n} no 1 h w 1 1 y w 2 w 1 y n 1 {l/n} w 1 l w 1 l w 2 n w 2 n w 2 š Μ 1 l Κ ?** {b / n} w 1 n w 4 Κ Μ 1 Κ y 1 Κ y 1 Κ r 4 Κ r 2 š w 1 š Κ 2 š š 1 t n 1 š Κ 2 Note that “no.” in the headings refers to the number of examples. * This text (Дijr 1) is in fact early Arabic (see Fig. 39). ** There is some doubt about reading this letter as a m. It could also be a q. figure 3. Letters on each side of the m in forms 4, 5, and 6. that it is very much linked to the presence of a w after the m. Indeed, in six examples out of the nine, the m is followed by a w, as in ARNA.Nab 17 (Fig. 21). Ligature on one side only: for this category, one should distinguish the words which start with m or end with m from those in which the m is in medial position. When the word starts with m, it is naturally not ligatured from the right and when it ends with m, it is of course not ligatured to the left. For the words in which the m is in medial position, the presence or absence of a ligature from the right depends very much on the letter which comes before. The database Finally, it should be noted that all examples of inal m in this corpus of texts are derived from the “classical” Nabataean form of the letter, i.e. forms 1 or 2. In the existing examples, however, one should be careful to treat separately the inal m of šlm, a word that may have become, in late texts, an ideogram. In the following examples, the inal m occurs in words other than šlm: • • • Ar 19 (Fig. 20): the inal m in grΚm is “classical” whereas the medial m in Κmrw is “evolved”; UJadh 309 (Fig. 48): compare the inal m in ywm (l. 4) and the initial m in mΜt (l. 6); M 1 (Fig. 27): compare the inal m in ywm (l. 3) with the examples of initial and medial m in the text (rm{n}h, mn, mytt, mΜh, Ήmš). — p (Fig. 2). Only two different forms of the letter have been distinguished, the irst of which is the “classical” Nabataean form. In our corpus, only twenty texts contain the letter p. Among them, seventeen have form 1 and only three have form 2. The clearest A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 53 examples of form 2 are in the two names in UJadh 222 (Fig. 42). Note that the inal form of the p of ywsp in UJadh 219 (Fig. 41) is almost that found in early Arabic inscriptions. It is very similar to the inal f of the patronym of the author of the Jabal Says grafito, whose name is mΚrf, as suggested by M.C.A. Macdonald in his recent rereading of the text, as well as to the p of ywsp in the newly published inscription from TaymāΜ (Macdonald, 2009b; al-Najem & Macdonald 2009: 210). — Β (Fig. 2). There are basically two forms of the letter. The irst one is the “classical” form while the second one can be considered as the evolved form. Two other forms are attested in one text each and may be considered as oddities (they are indicated as such on Fig. 2). In one of them, S 3 (Fig. 30), the reading of the letter in the irst name, byΒw, is not certain. As for the second odd form, it is attested in S 1, which is due to be published by Kh. al-Muaikil (al-MuΚayqil). It occurs in line 1, in the word ΜΒΉbh, where it cannot be read as a š. Form 2 of the letter occurs in four texts only, one example of which is UJadh 248 (Fig. 43). This form of the letter is very close to the s of ywsp in UJadh 219 (see Fig. 41) but it cannot be read as a s because, as noted by M.C.A. Macdonald, it is ligatured to the left, whereas s is not. There is no example of inal Β. — q (Fig. 2). There are also two forms of this letter. Again, 1 is the form in “classical” Nabataean while form 2 has undergone the same development as the letter p, becoming a circle on the line in initial/medial position. The developed form of these two letters tends to look very much like the m (compare them with form 4 of the m). In all our texts except one, the q belongs to form 1 but the only example of form 2 is particularly interesting. It occurs in S 3 (Fig. 30), in the name mrΜlqyš, where there is no doubt about the reading. — r (Fig. 4). The letter r is dificult to analyse because of the great number of small variations. However, I have identiied four forms of the letter in the texts. The evolved forms are 2 and 3. Form 4 is an odd form of the r usually found in combination with a b, which has exactly the same shape, thus forming the word br with two inclined parallel lines joined at the base. It is found in only one text of our corpus (unfortunately not among the selected texts presented in Appendix 2) but it occurs in “classical” Nabataean texts from Umm Jadhāyidh, for example in ThNUJ 114, where the letters bdΜ in ΚbdΜlgΜ are also written as three diagonal strokes. The r is one of the letters for which we have the greatest number of examples. There are approximately 100 of them, and they are more less evenly distributed between forms 1, 2, and 3 figure 4. The forms of r, medial and inal š, medial and inal t and lām-alif. of the letter. Very good examples of form 2 can be seen in UJadh 3 (Fig. 32) in dkyr and grmw, as well as in UJadh 219 (Fig. 41). The variant of form 3, seen for example in S 3 (Fig. 30) and UJadh 222 (Fig. 42), in the word dkyr, may be considered as the most evolved form of the letter. — initial and medial š (Fig. 4). The development of medial š is very interesting because one can clearly see how the three stems of which the letter is composed move downwards until they become three small vertical strokes resting on a horizontal line, exactly like the Arabic letter. Forms 1 and 2 can be considered as variants of the same form and the same is true of forms 3 and 4 while form 5 is the inal outcome of the development. Fifty-two texts contain a medial š, out of which the examples in twentyive are of form 1, i.e. the “classical” Nabataean form.13 Form 2 occurs in a few texts such as M 1 (Fig. 27) and the Stiehl inscription (Fig. 31). 13 54 Laïla Nehmé Form 3, which is best illustrated in UJadh 309 (Fig. 48) in the words šly, Μwšw, šlm, and šnt, occurs in eight texts. Note that in UJadh 309 the š of tšΚyn is clearly on its way to form 5. Other examples of form 3 can be seen in JSNab 18 (in šlm, Fig. 25), UJadh 266 (in šlm, Fig. 44), and M 1 (in šnt, Fig. 27). Finally, form 5 can be best seen in UJadh 266 (in Κšylh, Fig. 44) and in UJadh 299 (Fig. 47). — inal š (Fig. 4), of which there are only eight examples, appears in two forms, the second of which can be considered as the “evolved” one. The best example is the name ΚbdΜyš in UJadh 105 (Fig. 37). The evolution of the inal form 2 from form 1 is understandable only if we assume that there was an intermediate form equivalent to form 3 of the medial/initial sequence. However, this form does not occur in our texts. — initial and medial t (Fig. 4). All the texts that have form 1 of medial t, which is the normal form in “classical” Nabataean, are either third-century texts (CIS ii 963, 1491, NDGS 2, RES 528, B 3, JSNab 17) or formal ones such as Stiehl and al-Namārah. UJadh 309 (Fig. 48), which is dated to AD 295, also has this type of t. Finally, CIS ii 333 has it also but the text is not clearly dated. The t which is closed at the bottom (form 2) is a variant of form 1 and occurs in two texts only: Stiehl (Fig. 31) and ARNA.Nab 17 (Fig. 21). M.C.A. Macdonald notes that it is not simply the base line being continued under the t, but that even isolated examples have a closed base in the Stiehl inscription. It is not clear how forms 1 and 2 evolved to forms 3 and 4 because intermediate stages are missing. According to A. Yardeni (2000: 263), however, an intermediate phase in the evolution of the cursive t may be reconstructed between the looped form of the letter (which appears in the papyri not only in inal but also in medial positions) and the form resembling the early Nabataean y. This reconstruction is described as follows: in the intermediate phase, the right stroke would become longer than the left one, the loop would gradually close until a mere angle remains between the strokes. This angle would then gradually open up, leaving only a wavy stroke. UJadh 297 (Fig. 45) has a very good example of form 3. Note that form 5 is attested only in LPNab 41 (Fig. 26), in the word tnwΉ. — inal t (Fig. 4) can have ive forms. The irst is the “classical” Nabataean form and is identical to the irst initial/medial form. Form 2 (and 3, which is a variant of 2) is a particular form of inal t which is known also in “classical” Nabataean texts and which is best seen, in our corpus, in JSNab 17 (Ήrtt, brt, hlkt, šnt, Fig. 24), in Stiehl (brt, mytt, šnt, Fig. 31), CIS ii 963 (šnt, tltt, Fig. 23), etc. As was the case for the medial form, we lack the intermediate forms between forms 1–3 and forms 4–5 (but see A. Yardeni’s comment quoted above). Form 4 is well represented in our corpus, as in šnt in UJadh 109 (= ThNUJ 132–133, Fig. 38) and UJadh 297 (Fig. 45). Form 5 can be seen only in UJadh 298 (in ΜlΉrt, Fig. 46), where it may however, as suggested by M.C.A. Macdonald, be simply a lazy version of form 4 rather than a separate form (see the possible slight curve to the right near the top of the letter). In UJadh 309, one can see that the ligatured form of the letter (in šnt) is form 4 while the unligatured form (in mΜt) is form 3 (Fig. 48). — lām-Μalif (Fig. 4). In the texts in transitional script, this combination of letters usually receives special treatment, as already in the Namārah inscription. They are combined in such a way that they seem to form one single letter. UJadh 367 (Fig. 51) gives extremely clear examples of lām-alif preceded by Μ in the name ΚbdΜlΜšΉn, repeated twice. The lām-alif there clearly has the form found in the Namārah inscription as well as in the Arabic inscriptions of the irst century AH. The corpus of transitional texts The corpus of transitional texts I have collected so far contains 116 texts, one of which may have to be removed because it could be considered as “classical” Nabataean (UJadh 360, Fig. 50, see above). One other text, Дijr 1 (Fig. 5), which reads Κly mΉmd, was initially thought to be transitional but is in fact Arabic. M.C.A. Macdonald notes that the forms of all the letters, including the d, are perfect early Kuic.14 However, with the exception of the d and the y, the form of the letters would also it relatively well with the last stages of the development of the Nabataean script.15 Some of these 116 texts are published and have been known to the academic community for a long time. This is the case of the inscriptions from Sinai published in the CIS ii and by A. Negev (see the table below), as well as the Namārah inscription, JSNab 17 and JSNab 18, Stiehl, etc. Some others have been published more recently, mostly in Arabic, by S. al-Theeb (al-Dhīyīb), 14 See for instance the inscription on plates of copper at the entrance to the Dome of the Rock (Grohmann 1971: pl. 12/a), where the inal y of al-Ήusnā in line 5 and nunajjī in line 6 has this form (as opposed to the inal y in other words, e.g. Κalā), and see also the d in MuΉammad in line 7. 15 As noted by R. Hoyland, the y comes more directly down than in any of the Nabataean texts (where it tends to sweep to the left side irst before descending). R. Hoyland also points to an effortlessness and smoothness in the text, which does not exist in the other inscriptions. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 55 figure 5. Дijr 1. (Photograph MadāΜin СāliΉ archaeological project, facsimile L. Nehmé). Kh. al-Muaikil (al-MuΚayqil), M. al-Muraykhī, and ΚA. al-Ghabbān or in English, by M.C.A. Macdonald (see the bibliography under each of these names). Some are miscellaneous texts collected by scholars who were given photographs taken either during surveys or haphazardly. This is the case, for instance, of the photographs taken by F.V. Winnett and W.L. Reed in 1962, some of which show texts that are relevant to our corpus. Finally, the majority are texts that were recorded during the survey of the Darb al-Bakrah, which is the name given to the ancient caravan road between Дegrā and Petra, named after a pass in the mountains south of Tabūk. This survey was directed by ΚA. al-Ghabbān, and the present author is in charge of the publication of the texts in the Nabataean script, which were photographed in 2004. It should also be noted that ifty of these 116 texts were examined during a workshop organized by the present author in Paris in 2005 as part of a project on the development of the Nabataean script. During this workshop, seventy-one texts dating from the beginning of the third century AD to the end of the reign of the Umayyad caliph MuΚāwiyah, in AD 680, were read by the participants who were, apart from the author, ΚAlī al-Ghabbān, Robert Hoyland, Michael Macdonald, Khalīl al-MuΚayqil, MshallaΉ al-Muraykhī, Christian Robin, and Moulay Janīf. At the end of the workshop, the intention was to publish the whole corpus of texts dated to this time span. However, the publication was delayed and, in the meantime, the collection of texts examined during the workshop has been rendered incomplete by the discovery — mainly during the Darb al-Bakrah Survey — of dozens of new texts, most of which were previously unknown. Two regions have provided the greatest number of texts either dated to between the third and ifth centuries AD regardless of the degree of development their script shows, or written in transitional characters: Sinai and north-west Arabia, the latter being the region which, in the last few years, has been found to be the richest in late Nabataean, transitional, and early Arabic texts (Fig. 6). Dots on letters The only letter that receives a dot in the transitional texts is the d, and this is the case in twenty-eight texts in this corpus (within our selection, see JSNab 17, QN 2, Stiehl, UJadh 3, 105, 109, 178, 248, 375, 405, ΚUlā 1). J. Healey (in Healey & Smith 1989: 78), followed by C. Robin (2006: 364, and ig. 16) has suggested that, in JSNab 17, diacritical dots were placed on two other letters: on the r of Ήrtt and on the š of rqwš. In both cases, however, an examination of the original shows that these are chips in the stone and not dots marked intentionally. Outside our corpus of 116 texts, the use of dots is also not rare and can be found in several inscriptions, where it does not seem to be used exclusively for d:16 • JSNab 65 (Fig. 8), from the Jabal Ithlib area in MadāΜin СāliΉ. The text reads dkyr lwqys Μ---- Κdrw bΓb and there is a dot on the d of Κdrw which was not recorded by Jaussen and Savignac; • JSNab 123 (Fig. 9), also from the Jabal Ithlib area. The dot on the d of dkyr, at the beginning of the text, was not recorded by Jaussen and Savignac; • JSNab 181 = CIS ii 320D (Fig. 10) from Mabrak an-Nāqah north of MadāΜin СāliΉ. The text reads Κylw br šw{d/r}{d/r}mΜ šlm, and the fourth letter of the patronym has a dot over it. It was read Šūdrūmā by Jaussen and Savignac, thus suggesting that the 16 Some of these references had already been collected by M.C.A. Macdonald, to whom I am very grateful for making them available to me. On the use of diacritics in Nabataean, see Healey 1990–1991: 45. 56 Laïla Nehmé Site name Region No. of texts References al-AqraΚ North-west Arabia 1 – al-Muraykhī & al-Ghabbān 2001 al-ΚArniyyāt North-west Arabia 4 – four unpublished texts, one of which is presented in Appendix 2 under “Ar 19” ΚAvdat Negev 1 – RES 528 BoΒrā Дawrān 1 – B 3 (Unpublished. Note that this text is written in what M.C.A. Macdonald [2003: 44–46, 54–56] calls “Дawrān Aramaic”) Dūmat al-Jandal North-west Arabia 1 – ARNA.Nab 17 (= Macdonald 2009a) MadāΜin СāliΉ and vicinity North-west Arabia 9 – JSNab 17–18 (see Appendix 2, s.v.), – Stiehl (= Stiehl 1970), and – six unpublished texts, including Дijr 1 (Fig. 5) which is in fact Arabic Jabal Munayjah Sinai 1 – CIS ii 2666 Jubbah North-west Arabia 1 – CIS ii 345 Mābiyāt North-west Arabia 1 – M 1 (= al-Muraykhī, in press) al-Namārah Дawrān 1 – see Calvet & Robin 1997: 265–269, no. 205 QāΚ al-Nuqayb North-west Arabia 1 – QN 2 (Unpublished) Сadr ДawΞāΜ North-west Arabia 1 – One unpublished text Sakākā (QalΚah) North-west Arabia 10 –Two texts, one of which is S 3 (= al-Muaikil 2002 [=1993] unpublished in Winnett & Reed 1970 but it appears in one of their photographs, reproduced in Macdonald 2009a: pl. 4, no. 13a) – al-Muaikil & al-Theeb 1996: nos 30 (= ARNA.Nab 10) and 31 (= ARNA.Nab 13) – ARNA.Nab 2, 6–9, 11–12 (for these texts, see also Macdonald 2009a: pl. 4. [no. 7 = al-Theeb 1994a: 190, no. 6]) Sakākā (vicinity) North-west Arabia 4 – al-Theeb 2005: no. 3 – S 1 (= al-Muaikil, forthcoming) – Two unpublished texts TaymāΜ North-west Arabia 1 – al-Najem & Macdonald 2009 al-ΚUdhayb North-west Arabia 2 – Two unpublished texts al-ΚUlā and vicinity North-west Arabia 3 – JSNab 386, – CIS ii 333 – ΚUlā 1 (Unpublished) Umm al-Jimāl Southern Дawrān 2 – LPNab 41 – LPArab 1 Umm Jadhāyidh North-west Arabia 61 – al-Theeb 200217 + 46 unpublished texts, twelve of which are presented in Appendix 2, under the siglum UJadh Wādī Fayrān Sinai 1 – CIS ii 1491 Wādī Дajjāj Sinai 1 – Negev 1981: no. 9 Wādī Maghārah Sinai 2 – NDGS 1–2 Wādī Mukattab Sinai 1 – CIS ii 963 Wādī Ramm Дismā 6 – Grimme 1936: 90–95 (= Gruendler 1993: 13, A1)18 – Four unpublished texts 17 18 See inscriptions nos 14, 30, 34, 38, 41, 45, 48, 62, 84, 122, 128, 132–133, 145, 159, 203. See Hoyland (this volume), who identiies two texts. figure 6. Table showing the provenances of the inscriptions with transitional letter forms and texts dated to the period AD 200–500. Those for which readings are given in Appendix 2 are in bold script. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 57 figure 7. The geographical distribution of the texts in the transitional script. dot was carved above a r, but it may well be read Šūrdūmā, with a dot on the d. Neither of the names is attested in the Nabataean corpus but цrΨm occurs once in Safaitic, in CIS v 1955; • JSNab 212, from around the railway station of alΚUlā. The text is known only from a copy in which a dot appears above the r of br. Its presence, should, however, be checked on the original; • JSNab 321 from Sheqeiq edh-Dhib (Shuqayq alDhiΜb), c.20 km north-west of MadāΜin СāliΉ. The text is known only from a copy. It reads šlm bgrt br brdw bšnt 36 lrbΜl19 and there is a dot above the d of brdw; • CIS ii 344, from the area of TaymāΜ. The text is known only from J. Euting’s copy. It reads Κ{b}dt br t{y}mw br prk and the dot is above the d of Κbdt; Contra Jaussen and Savignac, who read the patronym as Nadrū and suggest therefore that the dot is above a r rather than a d. This rereading of the patronym was suggested by Macdonald (2009a: n. 51). 19 • ThNUJ 81, from Umm Jadhāyidh (Fig. 11). The text reads dkyr wbrw br Κdrw bΓb and there is a dot on both examples of d; • an unpublished text from MadāΜin СāliΉ which is carved 1 m to the left of the eye-betyl which is reproduced in Jaussen and Savignac20 (Fig. 12). This text is best read šlm ddn. There is a dot on each example of d; • an unpublished text from the Darb al-Bakrah Survey, Ir 3, from the site of ΚĪrīn (Fig. 13), which reads šlm rbybw br mšlmw and in which both examples of r have a dot above them; Jaussen & Savignac 1909–1914, i: 426 and ig. 217 (it bears the number Ith 55 in the new catalogue of the monuments of MadāΜin СāliΉ). This text was photographed by J. Bowsher and identiied by M.C.A. Macdonald as having diacritical dots. It is part of the epigraphic point no. 83, associated with the eye-betyl, which contains the inscriptions JSNab 111–118 as well as twenty-six unpublished texts. 20 58 Laïla Nehmé figure 8. JSNab 65. (Photograph MadāΜin СāliΉ archaeological project). figure 11. ThNUJ 81. (Photograph Darb al-Bakrah project). figure 9. JSNab 123. (Photograph MadāΜin СāliΉ archaeological project). figure 12. Unpublished text from MadāΜin СāliΉ (epigraphic point no. 83). (Photograph MadāΜin СāliΉ archaeological project). figure 10. JSNab 181. (Photograph MadāΜin СāliΉ archaeological project). figure 13. Ir 3. (Photograph Darb al-Bakrah project). • UJadh 398, also unpublished (Fig. 14), which reads mΚnw br bg{d/r}t, with a dot above the third letter of bg{d/r}t. It is not certain whether this is a d or a r but note that there is no dot over the r of br, which may indicate that the name should be read bgdt rather than bgrt. The name bgdt is relatively common in Safaitic; • Savignac 1932: 591, no. 1. The dot over the d in ydh is not shown on the photograph and does not appear on the squeeze but it was noticed by G.M.H. King on the stone and appears on the photograph she took; • al-Theeb 1994b, inscription B: 36–38. There is a dot over d in Ήdh but not on the d of ΜΊdw, which should perhaps be read ΜΊrw. Thus in most cases, the dot is used in Nabataean to distinguish the d from the r and is generally put on the d. There are only two texts (JSNab 212 and Ir 3) in which it A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 59 figure 14. UJadh 398. (Photograph Darb al-Bakrah project). figure 15. UJadh 118–119. (Photograph Darb al-Bakrah project). is clear that the dot is above the r. In Arabic, the earliest examples of texts using diacritical marks are the Ahnās papyrus (Grohmann 1932: 32–34 [not seen, quoted in Robin 2006: 342]; see also Larcher, this volume: ig. 6) and the Zuhayr inscription (al-Ghabbān 2008), dated respectively to 22 and 24 AH, to which should perhaps be added the inscription written on a piece of wood found in Petra among the Greek papyri, published by O. al-Ghul in 2004.21 As shown by C. Robin, who recently re-examined the use of diacritics (2006: 343–345), eleven letters out of the ifteen which, from the third century AH onwards, bear diacritics, already appear with them in the early documents dated up to the reign of MuΚāwiyah. However, neither the d nor the r are among these letters. This is to be expected in Arabic because these two letters are no longer homomorphs.22 It is interesting to note that in the transitional texts, which are supposed to be “on their way to Arabic”, none of the letters which will later receive diacritics receive any and, surprisingly, the d is still almost the only letter which receives a dot, despite the fact that its form is clearly distinct from that of the r. Thus, among the twenty-eight texts of our corpus which bear a dot above the d, twenty-four contain also a r which has a form which is very distinct from that of the d.23 In these twenty-four texts, therefore, there was no need to put a dot above the letter d to distinguish it from the r. All in all, the habit of adding a dot on the d in “classical” Nabataean in order to distinguish it from the r, which had a very similar shape, was rare but nevertheless existed and there is a possibility that this use in later texts was inherited from “classical” Nabataean. Strangely enough, the use of the dot to distinguish the letters which were starting to become homomorphs (such as the d and the k, the n and the b, etc.) is attested neither in the corpus of texts we collected nor in any of the pre-Islamic Arabic texts of the sixth century AD. M.C.A Macdonald (2008: 467a) questions the authenticity of this piece for the following reasons: “the relatively late form of the inal letter, the arrangement of the diacritical dots under the yāΜ, and the dificulty of interpreting the word as anything other than the modern name Nāyif”. 22 For this term see Macdonald 1986: 148, n. 119. 21 The relationship between the “classical” Nabataean and transitional script My intention, in this section, is not to show similarities or dissimilarities between the scripts in these two categories of texts but to present a few examples of inscriptions in the two scripts whose spatial relationship — including superposition — allows us to establish a relative chronology between them. The irst example is UJadh 118 (= ThNUJ 122) and UJadh 119 (= ThNUJ 123) (see Fig. 15). UJadh 118 is a transitional text while UJadh 119 is a very “classical” Nabataean one and it is clear that the tail of the inal letter Ar 34, Дijr 2, 3, 4, 5, Stiehl, JSNab 17, QN 2, UJadh 3, 4, 11, 67, 69, 105, 109, 118, 122, 178, 220, 248, 287, 300, 301, 312, 343, 375, 386, 405, 467, ΚUlā 1. The numbers in bold are presented in Appendix 2 and those in italics contain d and r with distinct forms. 23 60 Laïla Nehmé figure 16. UJadh 27 and 31. (Photograph Darb al-Bakrah project). 344, which must have been carved after the b of zbΜbrh of UJadh 343;24 and secondly at the top of the r of yΚmr which, again, must have been carved after the b of bΓb. This order in the superposition of the characters shows either that the letter forms of the “classical” Nabataean script continued to be used along with the transitional forms of the same letters or that the transitional forms of the letters started to be used early. Concerning the use of the “classical” Nabataean script at a late period, note should be made of the inscription UJadh 172 (Fig. 39) which, if it is indeed dated to AD 311–312 rather than AD 151–152, may be a late attestation of this script. The letter forms attested in the selection of texts presented in Appendix 2 Commentary figure 17. UJadh 343–344. (Photograph Darb al-Bakrah project). of UJadh 118 runs across and over line 1 of UJadh 119. The Nabataean text reads šlm Ήny br krys bΓb / w šlmw ΜΉwhy; the transitional one reads dkyr z{b/n}y{b/n}w / bΓy, with a dot on the d of dkyr. The second example is UJadh 27, written in “classical” Nabataean, and UJadh 31, written in transitional characters (Fig. 16). UJadh 27 reads ΚbdΚbdt br Μbw while UJadh 31 reads dkyr ΜpΓy / br Μwšw bΓb / w šlm. The bottom of the stem of the second w of Μwšw of UJadh 31 clearly runs over the top of the d of UJadh 27 and must therefore have been written after it. The third example, represented by UJadh 343 and 344 (Fig. 17), is particularly interesting. The text in “classical” Nabataean characters (UJadh 344) reads zpr br yΚmr / w Κbydw šlm, and the text in transitional characters (UJadh 343), reads dkyr lΚmrw / br zbΜbrh bΓb / w šlm. Parts of two of the letters of the “classical” Nabataean text run over two letters of the text in transitional characters. This can be seen irstly at the top of the y of yΚmr of UJadh Only two letters appear in both their “classical” and transitional forms in the same inscription: the Μ and the m. In JSNab 18, the Μ in ktbΜ, line 2 (Fig. 25) cannot be form 1 because there has to be room for the b before it. One has therefore to assume that the sign which follows the t is a combination of the b and the Μ (a similar combination can also be found in UJadh 343, Fig. 17), the latter being an evolved form of the letter (form 2 on Fig. 1). All the other examples of Μ in this text, however, belong to form 1. In M 1 (Fig. 27), the Μ in bΜyr and mΜt in line 4 belongs to the straight form of the Μ, in contrast to the looped form in Μntth in line 1 and qryΜ in line 2. In UJadh 10 (Fig. 33), there are two examples of m which have a transitional form, in šlymw and Κmrw, and one which has a “classical” form, in šmnw, in none of which is m in inal position (see the discussion of the letter m above). The evolved form of the Μ does not appear in inscriptions before the last quarter of the third century, in combination with a b, in ktbΜ and bΜyr, respectively in JSNab 18 (Fig. 6) and M 1 (Fig. 27), and isolated in ARNA.Nab 17 (Fig. 21). It is then much more regularly and often used in the texts than the “classical” form. However, forms 2 and 3 already appear in the Nabataean papyri along with form 1 (Yardeni 2000: 237). The form of the g (no. 1 on Fig. 1), which looks like the evolved Ή (no. 3 on Fig. 1), is less widespread in this corpus of texts than the “evolved” form of the g. In the The letter that follows the z looks like an early Arabic Μ but if we compare the shape of the letter with the signs for bΜ in ktbΜ as they appear in JSNab 18 (see below), it is almost certain that the letter after z should be read bΜ. 24 A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 61 whole corpus of transitional texts, there are only four examples of the former and ten examples of the latter, out of which four seem to have more or less completely lost the lower part of the diagonal stroke. M.C.A. Macdonald has pointed out to me that this is interesting in view of the fact that this bottom part of the diagonal is characteristic of early Arabic g, and that g and Ή are homomorphs from the earliest examples of the Arabic script. The d and the k can be treated together. The form of the d in the transitional texts is characteristic and appears at the end of the third century in ARNA.Nab 17 (Fig. 21), but both forms 1 and 2 continue to be used in a relatively large number of texts, both dated and undated, whereas in these texts the k has form 2 much more regularly than form 1. A. Yardeni has pointed out that the form of the d we have in the grafiti from the third and fourth centuries AD is not a late development of the letter because it resembles almost exactly that in earlier Aramaic texts (2000: 241). M.C.A. Macdonald notes however that this does not necessarily mean that it is an “inheritance” from earlier Aramaic since it may have developed this form from the “classical” Nabataean form. Medial h, along with Ή, is one of the letters which evolves early in the development of the script. Form 2 of this letter appears in JSNab 17 (Fig. 24)25 and is used throughout the texts. It can thus be considered as a diagnostic letter. Form 2 of the letter is not widespread in the papyri. Evolved forms of inal h (forms 2 to 4) appear for the irst time in an early third-century text from Sinai, CIS ii 963 (Fig. 23), dated (with some uncertainty) to AD 205– 206. From AD 280 onwards (in M 1 [Fig. 27] and others), they are almost always used in the texts, except in JSNab 17 (Fig. 24), LPNab 41 (Fig. 26), and Stiehl (Fig. 31), which are formal texts. In the papyri, both the long inal h (in what A. Yardeni calls “calligraphic” cursive script) and evolved forms (in what she calls “extreme” cursive script, i.e. the result of rapid writing) are used (Yardeni 2000: 242). Within form 1 of w in Fig. 1, I have distinguished, on the one hand those cases where the letter is ligatured from the preceding letter (in which case the ligature is clearly made to the bottom of the w), and on the other, those in which the w is not ligatured from the preceding letter. It appears that form 2, in which the ligature is made to the middle part of the w, more or less at the level of the loop, is much more widespread in the texts in this corpus than the ligature to the base, as in form 1 (twentyone against ive). In the papyri, the w keeps its loop in Apart from the texts which are listed in Fig. 18, it also occurs in NDGS 1 from Wādī Maghārah, dated to AD 265–266, and in the Namārah inscription of AD 328. 25 the “calligraphic” cursive script whereas in the “extreme” cursive script, the loop is rendered by a mere thickening at the top of the letter (Yardeni 2000: 244). The Ή is, like the h, a letter that starts evolving early and is more diagnostic than others. The transitional form is used almost exclusively in all the texts of our corpus, except in LPNab 41 (Fig. 26), in JSNab 17 (Fig. 24, in byrΉ only, where an odd version of the “classical” form is used, while all the other examples of Ή in the text are evolved), and in UJadh 19 (Fig. 35) which, again, has a corrupted “classical” form. In the “calligraphic” cursive script of the papyri, the Ή is very close to form 2 whereas in the “extreme” cursive script, it is like form 3 (Yardeni 2000: 246). Both the medial and inal forms of y are also very distinctive. The “classical” medial form (no. 1), which was also used in inal positions in the irst century AD, is no longer used at all. As for medial y, the only exceptions to the use of form 2 are an early text (CIS ii 963, AD 206–206, Fig. 23), and UJadh 172 (Fig. 39), the date of which is doubtful (either AD 151–152 or AD 311–312). In the papyri, the medial y has mostly the wavy form 2, with small variations, and the inal y has form 2 (Yardeni 2000: 248–249). I have already mentioned that only one text mixes the “classical” and the evolved forms of the letter m in non-inal position, UJadh 10 (Fig. 33). One interesting feature about the m is that the inal form always retains the “classical” Nabataean shape. If we ignore the row “m 1 f” in Fig.18, we realise that a large majority of the examples of m in this selection of texts (twenty-ive against eight) have the evolved forms (4–9), but that the “classical” form goes on being used much more than, for instance, those of h, Ή, and y. Circular forms of m are already very well attested in the papyri (Yardeni 2000: 253). There is only one clear example26 of the evolved form of medial p, a letter which otherwise has a form which remains very stable throughout the corpus. Evolved form 2 of p does not seem to be attested in the papyri. The same is true of the q, of which there is only one evolved medial form in our grafiti, in mrΜlqyš of S 3 (Fig. 30). The r is one of the letters which varies considerably and it is sometimes dificult to attribute a particular example to one of the forms that have been distinguished. There is, however, a general tendency for the letter to lose its upper bar or lourish and — mainly when combined with b in the word br — to become smaller and to be ligatured from the preceding letter to its middle part. It 26 There are two possible other examples in the rest of the corpus, including in the inscription in al-Muraykhī & al-Ghabbān 2001. Laïla Nehmé Μ1 Μ2 Μ3 g1 g2 d 1/2 d3 h1 h2 hf1 hf2 hf3 hf4 w 1 ligatured w 1 unligatured w2 w3 Ή1 Ή2 Ή3 y1 y2 y3 yf1 yf2 k1 k1f k2 m1 m1f m2 m3 m4 m5 m6 m7 m8 m9 x 5th c. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 9 6 9 2 4 14 16 2 6 3 4 5 2 5 10 20 1 1 2 11 2 24 13 0 15 7 1 17 6 14 0 2 4 6 4 9 2 0 TOTAL OF x white/grey TOTAL OF x ΚUlā 1 Undated texts Ar 19 QN 2 S3 UJadh 3 UJadh 10 UJadh 15 UJadh 19 UJadh 90 UJadh 105 UJadh 178 UJadh 219 UJadh 222 UJadh 248 UJadh 266 UJadh 298 UJadh 299 UJadh 360 UJadh 367 UJadh 375 LPNab 41. 3rd c. UJadh 297. 305–306 UJadh 172. 311–312? Stiehl. 356 S 1. 428 CIS 963. 205–206? B 3. 230–231 JSNab 17. 267 JSNab 18. 267? ARNA.Nab 17. 275–276 M 1. 280 UJadh 309. 295 4th c. UJadh 109. 455–456 3rd century UJadh 405 62 9 15 2 4 14 16 2 6 3 11 5 10 21 3 11 2 37 0 15 8 17 22 25 Note: the letters followed by a number in the irst column of this table refer to the synthesized letter forms on Figs 1, 2, and 4. The rows with a grey background are the transitional letter forms, and those with a white background are the “classical” Nabataean forms. A letter followed by “f” indicates a inal form. figure 18. Description of characters. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 63 p1 p1f p2 Β1 Β2 q1 q2 r1 r2 r3 r4 š1 š2 š3 š4 š5 šf1 šf2 t1 t2 t3 t4 t5 tf1 tf2 tf3 tf4 tf5 x 5th c. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 7 1 1 1 1 5 1 12 14 16 0 12 2 2 5 8 2 2 6 2 3 0 1 1 3 2 5 1 TOTAL OF x white/grey TOTAL OF x ΚUlā 1 UJadh 405 Undated texts Ar 19 QN 2 S3 UJadh 3 UJadh 10 UJadh 15 UJadh 19 UJadh 90 UJadh 105 UJadh 178 UJadh 219 UJadh 222 UJadh 248 UJadh 266 UJadh 298 UJadh 299 UJadh 360 UJadh 367 UJadh 375 LPNab 41. 3rd c. UJadh 297. 305–306 UJadh 172. 311–312? Stiehl. 356 S 1. 428 CIS 963. 205–206? B 3. 230–231 JSNab 17. 267 JSNab 18. 267? ARNA.Nab 17. 275–276 M 1. 280 UJadh 309. 295 4th c. UJadh 109. 455–456 3rd century 8 1 1 1 5 1 12 30 14 15 2 2 8 4 6 6 figure 18 (continued). Description of characters. should, however, be noted that form 2, a vertical line with the ligature to the bottom of the letter, is also very well represented in the texts. The š is an interesting letter because it is the one in which the evolution of the letter from form 1 to form 5 can be traced through many examples. It should be noted, however, that the “classical” form continues to be used throughout the corpus in more or less the same proportion as the evolved forms (3–5): i.e. fourteen against ifteen. The presence of a “classical” š in a text does not, therefore, mean that this text is early. As for the appearance of form 3, which is the real transitional form, it seems to be not earlier than the second half of the third century, in JSNab 18 (Fig. 25), in the word šlm at the end of line 2. In the papyri, forms 2–5 of medial š are attested, although form 5 is found only in the “extremely” cursive script (Yardeni 2000: 262). The t, like the š, is a very interesting letter because neither the medial nor the inal transitional forms appear before the end of the third century, in M 1 (Fig. 27), dated to AD 64 Laïla Nehmé 280. After that date, the transitional forms become the rule, except in Stiehl (Fig. 31) of AD 356 and in UJadh 309 (Fig. 48) of AD 295, which are formal texts.27 In the latter, only the inal form, in šnt, line 5, is transitional. Note that the “classical” and transitional forms of the letter in initial/medial position are exclusive of each other. In the papyri, where the open form of the t (no. 4) appears earlier than in the inscriptions, one can trace the evolution of the letter from forms 1–3 to form 4. There is indeed, in the cursive t, a form that is not attested in the inscriptions (see Yardeni 2000: 263). All in all, if one was to draw the ideal alphabet of the evolved forms that appear in transitional texts, most of which, of course, already appear in the papyri, one would probably get something close to the “idealized” alphabet, which is given in Figure 19. The logic would be that the more a particular inscription contains characters that belong to this idealized alphabet, the later it is. However, the process does not exactly work like this because some letters show an early development and are used in a rather stable way throughout the corpus (the h, the Ή, and the y) while others show hesitations between “classical” and evolved forms, sometimes within one single text (Μ, m). It is therefore dificult, from the inscriptions, to trace a continuous development in the use of the letter forms. Indeed, very often there are, as one would expect, mixtures of “classical” and evolved forms in the texts. The reason for this, as suggested by M.C.A. Macdonald, may be that the calligraphic version of the script used in formal inscriptions — and often successfully attempted in grafiti — co-existed with a day-to-day scribal version of the script used for documents on soft materials. He considers rightly that the “changes in the letter forms and the increasing use of ligatures seen in the formal script only make sense as the transference to stone of features developed through writing swiftly with pen and ink. There would have been no reason for them to have developed independently within the process of carving on stone”.28 The result is a “growing, but haphazard, intrusion of day-to-day scribal forms (with occasional attempts to ‘monumentalize’ them) into the calligraphic version of the script used for monumental inscriptions, which is imitated in grafiti”. I have tried, in this modest contribution, to gather all the examples of these “intrusions”, both published Note that it appears in the Namārah inscription, which is also formal. See Macdonald 2003: 52–54; this volume; and forthcoming, the section entitled “The chisel and the pen”; and (this volume). 27 28 and unpublished, not only to present an up-to-date assessment of the material available for the study of the development of the Nabataean script but also to help identify the intrusions, determine which are the most recurrent, and which characters offer more “resistance” and in which contexts. The material I have collected is a perfect illustration of the fact that here is no continuous development of the script, and this should make us even more careful regarding the use of palaeographic studies for dating. hf h k yf ‘ d g ṭ y ḥ s š r z w n q ʾ b m l ṣ p l-ʾ t figure 19. The idealized forms of the evolved characters in the transitional texts. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 65 Appendix 1. Texts dated to the period AD 200–500 (third–ifth centuries), by date Readings of the texts which appear in bold script are given in Appendix 2. Date (AD) Site Region Reference(s) Third century 203 TaymāΜ North-west Arabia al-Najem & Macdonald 2009 204 ΚAvdat Negev RES 528 205–206 Wādī Mukattab Sinai CIS ii 963 218–219 Jabal Munayjah Sinai CIS ii 2666 222–223 Wādī Дajjāj Sinai Negev 1981: no. 9 230–231 BoΒra Дawrān B 3 (unpublished) 231–232 Wādī Fayrān Sinai CIS ii 1491 265–266 Wādī Maghārah Sinai NDGS 1 266–267 Wādī Maghārah Sinai NDGS 2 267 MadāΜin СāliΉ North-west Arabia JSNab 17 267? MadāΜin СāliΉ North-west Arabia JSNab 18 Third century Umm al-Jimāl Southern Дawrān LPNab 41. Dūmat al-Jandal North-west Arabia ARNA.Nab 17 (= Macdonald 2009a) 280 Mābiyāt North-west Arabia M 1 (= al-Muraykhī, in press) 295 Umm Jadhāyidh North-west Arabia UJadh 309 (unpublished) Umm Jadhāyidh North-west Arabia UJadh 297 (unpublished) 305+ Al-ΚUlā North-west Arabia CIS ii 333 306 Al-ΚUlā North-west Arabia JSNab 386 Umm Jadhāyidh North-west Arabia UJadh 172 (unpublished) Дawrān most recently Calvet & Robin 1997: 265–269, no. 205 North-west Arabia Stiehl (= Stiehl 1970) Sakākā North-west Arabia S 1 (= al-Muraykhī, in press) Umm Jadhāyidh North-west Arabia UJadh 109 (= ThNUJ 132+133, republished in Nehmé 2009: 49–52, ig. 3). 275–276 Fourth century 305–306 311–312 (or 151–152) 328 al-Namārah 356 MadāΜin СāliΉ Fifth century 428 455–456 Laïla Nehmé 66 Appendix 2. The selection of inscriptions used as examples for letter forms mentioned in this paper The corpus of inscriptions dated to between the third and ifth centuries AD (so-called “late Nabataean”), some of which are clearly written in “transitional” characters and some less so, at present contains 116 texts. The selection of thirty-four inscriptions presented below represents therefore almost one third of the total number. They are used in this paper for examples of letter forms. In this Appendix, the readings and translations are given but commentaries are kept to a minimum. A photograph and a facsimile are provided for each text. The complete corpus will be published elsewhere, in a special volume devoted to texts from late Nabataean to early Arabic, up to AD 680. Note that ifty texts out of the 116 were examined during a workshop, which was organized in Paris in 2005 by the present author as part of a project on the development of the Nabataean script. During this workshop, a total of seventy-one texts was examined, dating from the beginning of the third century AD to the end of the reign of the Umayyad caliph MuΚāwiyah in AD 680. Editorial sigla: { } enclose doubtful letters {.} represents an illegible letter [ ] enclose letters which are restored ---- represents a passage in which one or more letters are completely destroyed / between two letters indicates an alternative reading * indicates one of ifteen previously unpublished texts figure 20. Ar 19. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). * Ar 19 (Fig. 20) ARNA.Nab 17 (Fig. 21) This text comes from the site of al-ΚArniyyāt (see Fig. 7), which was visited during the Darb al-Bakrah Survey Project in 2004 during which 166 Nabataean texts were recorded. These will be published as part of the Darb alBakrah corpus of inscriptions. dkyr grΚm br Κmrw Note the different shapes of m: the inal m in grΚm has the “classical” Nabataean form while the medial m in Κmrw is much more evolved and is ligatured from the preceding and to the following letters at its base. The d and the k are representative of what these letters become in the transitional texts. The r in dkyr is a simple vertical stroke. All the letters are clear. The name grΚm is new in the Nabataean onomasticon whereas Κmrw is common in the Nabataean inscriptions. This text was photographed by F.V. Winnett and W.L. Reed in 1962, approximately 15 km north-north-west of Dūmat al-Jandal. It was published — but misread — by J.T. Milik and J. Starcky, from a very poor photograph, as ARNA.Nab 17, and republished in 1996 by Kh. al-Muaikil and S.A. al-Theeb who read it as two inscriptions, their nos 63 and 64. Again, no photograph was made available. The irst exact reading and full commentary from an excellent photograph of the text by the Department of Antiquities and Museums of Saudi Arabia is due to M.C.A. Macdonald (2009a). The reading and translation given below are taken from the latter publication. The text was examined during the 2005 Paris workshop. Μy dkyr Κwydw (l. 3) br šlymw (l. 4) khnΜ (l. 5) A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 67 figure 21. ARNA.Nab 17. (Photograph Macdonald 2009a: 237, pl. 1, facsimile L. Nehmé). dnh šnt mΜh (l. 2) w šbΚyn (l. 1) “Yea, may Κwydw be remembered son of Šlymw the oracle priest. This is the year one hundred and seventy” The text is dated to AD 275–276. Note the form of the Μ in Μy, mΜh, and khnΜ, which still has a small vertical stroke. Note also the ligature between the m and the w in šlymw. The medial h of khnΜ is odd because it is closed at its bottom, as a result of the continuation of the base line from the k. The t in šnt is also closed at its bottom but this feature is part of the letter itself, it is not the continuation of the base line. Κwydw is a common name in the Nabataean inscriptions. It should be noted that Ίlymw is attested elsewhere in Nabataean only in North Arabia, in ARNA. Nab 16 as reread by al-Muaikil and al-Theeb (1996: no. 35),29 and in al-Theeb 1993: no. 21, as well as in UJadh 10, see below. Milik and Starcky’s original reading of the name in ARNA.Nab.16, šlytw, has been corrected to šlymw. The examination of the original photograph taken by Winnett and Reed seems to conirm this reading (the letter is closed at the bottom), but see Macdonald 2009a: n. 31. 29 * B 3 (Fig. 22) This text is part of the corpus of Nabataean inscriptions from BoΒra. It was found in 1956 in the ruins of the “Nabataean gate” and was kept in the Museum of the Citadel of BoΒra (where, however, it was not found by the author in 2003). It was photographed by M. Dunand in 1961 and recorded by J.T. Milik whose reading differs slightly from the one given below. The text is dificult to read because the letters are inely and lightly incised in the basalt. M.C.A. Macdonald considers it to be written in Дawrān Aramaic rather than in Nabataean. It seems to me, however, that the incision in the hard basalt may be responsible for the lack of some of the ligatures and most of the letters are perfectly understood and read as Nabataean. The text was examined during the 2005 Paris workshop. Κbdw br {m}{.}{d/r}{d/r}{pw} šlm št{.} mΜ{t} w Κšryn w Ήmš l{h}{.}{.}y “Κbdw son of {m}{.}{d/r}{d/r}{pw}, peace ! Year one hundred and twenty of the eparchy” 68 Laïla Nehmé figure 22. B 3. (Photograph J. Milik’s archive, facsimile L. Nehmé). The text is dated to AD 230–231. Note the classical form of the Μ, the m, the š, the t. Of the very few letters which have transitional forms, note that inal y in {h}{.}{.}y. Κbdw is a well-known Nabataean name but the reading of his father’s name is too doubtful to make any suggestion. In the word šnt, the n is missing. The letters št are followed by the remains of a letter, possibly an Μ. No satisfactory explanation can be given for this sign, which may be accidental. CIS ii 963 (Fig. 23) This text (= RES 128) is known only from a squeeze published in CIS ii. It was examined during the 2005 Paris workshop. dkyr tymΜlhy br yΚ{l/n}y šnt mΜh Κl dmyn Κl ---- tltt qysryn “May be remembered TymΜlhy son of YΚny year one hundred which equals ---- the three Caesars” It has been suggested (CIS ii and Negev 1967: 253) that the signs Κ l after mΜh were in fact the igures 5+1 but this is very unlikely. The date should therefore be understood as only “one hundred”, i.e. AD 205–206, and not 100+5+1, i.e. 106 = AD 211. The three Caesars were Septimius Severus, Caracalla, and Geta, who reigned together from 198 to 211. The translation of the expression Κl dmyn as “equals” in Negev is based on his interpretation of the meaning of the verb dmy in Aramaic, “to resemble, be like” (Sokoloff 1990: 151/a). However, Negev does not explain how a substantive dmyn meaning “equals” can be derived from dmy and the syntax, preposition + dmyn + preposition, is odd. The whole phrase requires reconsideration but the date of the text is clear from what precedes it. JSNab 17 (Fig. 24) This well-known inscription has been widely discussed figure 23. A facsimile of CIS ii 963 based on the squeeze published in CIS (L. Nehmé). since it was irst discovered by C. Huber and published by Jaussen and Savignac. I do not intend to propose a full reexamination here (for which see Healey & Smith 1989; Healey 2002; Robin 2001: 547; 2006: 324–326). It has been included in this selection because a new photograph and a new facsimile have recently been made and are given here, along with the reading and translation, for reference. This text was examined during the 2005 Paris workshop. dnh qbrw ΒnΚh kΚbw br Ήrtt lrqwš brt Κbdmnwtw Μmh w hy hlkt py Μl Ήgrw šnt mΜh w štyn w tryn byrΉ tmwz w lΚn mry ΚlmΜ mn yšnΜ Μl qbrw d[Μ] w mn yptΉh Ήšy w wldh w lΚn mn yqbr w {y}Κly mnh “This is the tomb which was built by KΚbw son of Дrtt for Rqwš daughter of Κbdmnwtw his mother. And she died in al-Дijr in the year one hundred and sixtytwo in the month of Tammūz. And may the Lord of eternity curse anyone who alters this tomb or opens it except his children and may he curse anyone who buries and removes [a body] from it.” It should be noted here that the demonstrative pronoun before qbrw, in line 1, is not th, and is therefore not A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 69 figure 24. JSNab 17. (Photograph MadāΜin СāliΉ archaeological project, facsimile L. Nehmé). feminine (contra Healey & Smith 1989: 80). There is absolutely no doubt, after a minute examination of the stone, that we have here the normal masculine Nabataean demonstrative pronoun dnh. The text is dated to AD 267. Note the forms of some of the letters, such as the Μ, the m, both medial and inal š, the t, etc., which are closer to the “classical” Nabataean form than to the transitional one. The Ή and medial h are more evolved. Note the use of diacritical dots on the d (ll. 3, 8, 9). JSNab 18 (Fig. 25) ARNA.Nab 89–90, RES 1106.A, Healey 2002: 84–85. This text was carved below and to the right of JSNab 17 and has been given much less attention by scholars. It mentions the builders of the tomb of Raqūš and is therefore either contemporary with, or a little later than, JSNab 17. It was examined during the 2005 Paris workshop. w dkyr Κdmn hwΜ ktb ktbΜ dΜ bΓb w šlm dkyr bnyΜ hnΜw w ΜΉbr{w}h d{y} bn{w} qbrw Μm kΚb{w} “And may Κdmn who wrote this text be remembered for good and may he be secure. May the builders HnΜw and his companions, who built the tomb of the mother of KΚbw, be remembered” In Healey 2002, the end of the second line is read w bšlm but there is no b in front of šlm. The stroke which was interpreted as the letter b belongs in fact to the š (form 3). Note, as in JSNab 17, the “classical” form of some of the letters, the Μ (except in ktbΜ and in ΜΉbrwh), the w, the m, the t, etc. The letters which are more evolved are the h, the Ή, and the y. Note also the š in šlm at the end of the second line, which is clearly on its way to the evolved form of the letter. The name Κdmn, which was read gzmn by Jaussen and Savignac and Κd mn in ARNA.Nab 90, is clear on the photograph. It is new in the Nabataean onomasticon. LPNab 41 (Fig. 26) CIS ii 192, RES 1097, Cantineau 1930–1932: 25. This text has a Greek counterpart, which was discovered in a different place in Umm al-Jimāl. It is dated to the third century by the mention of a Jadhīmah king of Tanūkh. It was examined during the 2005 Paris workshop. dnh npšw phrw br šly rbw gdymt mlk tnwΉ “This is the nefesh of Phrw son of Зly the tutor of Gdymt the king of TnwΉ” Note, as in JSNab 17 and 18, the “classical” form of inal h, of the w, the m, the p, the š, etc. Note also the peculiar form of the t in tnwΉ. M 1 (Fig. 27) This inscription was found in a reused position during the second season of excavations at the site of al-Mābiyāt, ancient QuraΉ, which is about 40 km south of al-ΚUlā. 70 Laïla Nehmé figure 25. JSNab 18. (Photograph MadāΜin СāliΉ archaeological project, facsimile L. Nehmé). figure 26. LPNab 41. (Photograph courtesy of M.C.A. Macdonald, facsimile L. Nehmé). figure 27. M 1. (Photograph courtesy of M. al-Muraykhī, facsimile L. Nehmé). It has been published by M. al-Muraykhī (in press). The reading given below differs slightly from the reading proposed in the editio princeps. ------- šlm Κl q[b]r r{mn}h Μntth brt ywsp br Κrr dy mn qryΜ dy mytt ywm Κšryn w šth bΜyr šnt mΜh w šbΚyn w Ήmš “---- Šlm on the tomb of R{mn}h his wife daughter of Ywsp son of Κrr, who is from QryΜ (?), who died on day twenty-six of Iyār year one hundred and seventy-ive” The names r{mn}h and Κrr, if the reading is correct, are new in the Nabataean onomasticon. Ywsp is a well-known Jewish name (Ilan 2002: 150–168), which occurs also in the recently published inscription from TaymāΜ (al-Najem & Macdonald 2009, see their commentary on the name on p. 210) as well as in UJadh 219 = ThNUJ 84 (see below) and JSNab 262. It is possible that the toponym qryΜ is to be equated with modern Qurayyā, in the land of Madian. The woman of whose monument this stone was part would have died in Qurayyā and was either buried or commemorated (if the stone belongs to a nefesh) in A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 71 figure 28. QN 2. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 29. S 1. (Photograph courtesy of Kh. al-Muaikil, facsimile L. Nehmé). the region of Mābiyāt. The exact provenance of the stone is not known but if it comes from Mābiyāt itself and not from its surroundings, it means that this site was occupied before the Umayyad period. The text is dated to AD 280. Most of the letters in the text have an evolved form, especially the Μ (except perhaps in Μntth), inal h, Ή, medial y and m, and medial and inal t. Note, however, the form of the š in šnt, which is on its way to a more evolved form. * QN 2 (Fig. 28) This text comes from the site of QāΚ an-Nuqayb (see Fig. 7), which was visited during the Darb al-Bakrah Survey Project in 2004 and at which forty-nine Nabataean texts were recorded. dkyr ΜbwqΓ{b/n}h Note that the possible b and n have very similar heights. Almost all the letters, except inal h, have a “classical” Nabataean form. The name ΜbwqΓ{b/n}h is new in the Nabataean onomasticon. S 1 (Fig. 29) This inscription, which is due to be published by Kh. al-Muaikil in Adumatu along with other texts from the Sakākā region, was found on an outcrop not far from Sakākā. The reading is given here with the permission of Khalīl al-Muaikil, to whom I am very grateful. This text was examined during the 2005 Paris workshop. dkyrw mΉrbw w ΜΒΉbh Μl Κšrh w Κnymw w [w]Μlw w Ήrtw w {Γ/k}Ήšw bΓbw mΉrbw br ΚwydΜlt ktb ydh ywm Κšrh w tmnh bΜyr šnt 2 × 100 +100 +20 +3 {Μ}{d}{.}{Ήg}--- Μl Ήyrh “May MΉrbw and his ten companions and Κnymw and WΜlw and Дrtw and {К/K}Ήšw be remembered for good. MΉrbw son of ΚwydΜlt wrote [with] his hand day eighteen of Iyyār the year 323{Μ}{d}{.}{Ήg}---- alДīra[?]” This text is dated to April AD 428. For the commentary on this very interesting text, see al-Muaikil, forthcoming. Most of the letters in this inscription have evolved forms: Μ, d, inal h, w, Ή, y, k, m, r, š, t. The Β in ΜΒΉbh looks very much like a “classical” Nabataean š but it is different from the other š in the text, especially from the only initial š, in šnt. The line crossing the ligature between the b and the h, visible on the photograph, is accidental and was carved before the text. The name mΉrbw is attested in a recently published inscription from Umm al-Jimāl, the editors of which interpreted it as the Arabic name MuΉārib (Said & al-Hamad 72 Laïla Nehmé figure 30. S 3. (Photograph courtesy of Kh. al-Muaikil, facsimile L. Nehmé). 2004: no. 2). Κnymw is probably a variant of Κnmw, well attested in Nabataean, as is wΜlw. The name Ήrtw, which is a variant of Ήrt, is attested only in one text from Umm al-Jimāl, LPNab 55, in which the reading of the name is doubtful. Neither kΉšw nor ΓΉšw are known in Nabataean and ΚwydΜlt is a theophoric name which occurs here for the irst time, although Κwydw is a well-known Nabataean name. S 3 (Fig. 30) This inscription was found in 1991 by S. al-Theeb on a small hill known as al-QalΚah, about 5 km north of Sakākā, and was published by Kh. al-Muaikil (1993: no. 2, in Arabic, and 2002, no. 2, in English). Note that part of it also appears on a photograph taken in 1962 by F.V. Winnett and W.L. Reed and reproduced as plate 4 in Macdonald 2009a, where it was numbered ARNA.Nab 13a. This photograph was not included in Winnett and Reed 1970 and the inscription was not mentioned in the publication. This text was examined during the 2005 Paris workshop. by{Β}w dkyr mrΜlqyš br mlkw The editio princeps reads bΚΒw for the irst name but y should be preferred to Κ. On the second line, dkyr and mrΜlqyš should be preferred to br Κbd mrΜlqyš of the editio princeps. It is therefore better to consider the three lines, not as a single text, but as belonging to two different inscriptions: on the one hand, the name byΒw which is new to the Nabataean onomasticon, and on the other, a commemorative text starting with dkyr. The name mrΜlqyš is new in the Nabataean onomasticon. Note the peculiar form of the Β in byΒw, the shape of which was reconstructed from the Winnett and Reed photograph. This text shows considerably evolved characters, especially the Μ, the y, the m, the q, the r, and the inal š. The d and the k have the typical form they have in the transitional texts. This text has been dated to the ifth century by Kh. al-Muaikil (2002: 165) but this dating is not secure at all and is no more than a hypothesis. There are other texts with evolved characters on the same rock face, which can be seen in the Winnett and Reed photograph and which are discussed in Macdonald 2009a. Stiehl (Fig. 31) This inscription was found in Jedda but is said to come from MadāΜin СāliΉ. It was irst read by F. Altheim and R. Stiehl (1968: 305–309) and was republished by R. Stiehl in 1970. Al-Najem and M.C.A. Macdonald (2009) have recently republished it, with Stiehl’s photograph and a reading which differs slightly from the earlier ones. It is this reading and translation which are given below. This text was examined during the 2005 Paris workshop. dnh ----{š}----b{rt}Μ dy----30 Κdy---- br Ήny br šmwΜl ry{š} ΉgrΜ Κl mwyh Μtth brt31 Κmr{w} br Κdywn br šmwΜl ryš tymΜ dy mytt byrΉ Μb šnt mΜtyn w Ήmšyn w ΜΉdy brt šnyn tltyn w tmny Stiehl (1970) has dnh [npšΜ] [w] qbrtΜ dy [Κbd lh] for line 1. The readings Κdywn (instead of Κdnwn) suggested by Al-Najem and Macdonald (2009: 213–214 and n. 35) in lines 2 and 4, and mwyh (instead of mwnh) in line 3 (irst suggested by J. Starcky [1978: 47], followed by al-Najem and Macdonald [loc. cit.]) are very convincing. 30 31 A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 73 figure 31. Stiehl inscription. (Photograph R. Stiehl, facsimile L. Nehmé). figure 32. UJadh 3. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). “ ... Κdy[wn] son of Дny son of ЗmwΜl chief man of ДgrΜ for Mwyh his wife, daughter of Κmrw son of Κdywn son of ЗmwΜl chief man of TymΜ, who died in the month of Ab in the year two hundred and ifty-one at the age of thirty-eight” Note the “classical” form of most of the letters: Μ, inal h, w, m, r, š, t. All the examples of d in this text are dotted despite the fact that this was not necessary because r and d are not identical in shape. The letters which are more evolved are the Ή, as well as the medial and inal y. Note the peculiar form of the medial and inal t, almost all the examples of which are closed at their bases. Note that none of the examples of w is ligatured from the preceding letter. For the commentary on the names, see al-Najem & Macdonald 2009. UJadh 3 = ThNUJ 48 (Fig. 32) This text and the following ones were discovered during the Darb al-Bakrah Survey Project, in 2004, in Umm Jadhāyidh, 150 km north-west of MadāΜin СāliΉ. This site contains 488 Nabataean or transitional texts, written on rock faces or boulders, among texts written in Ancient South Arabian, Arabic, Hismaic, “Thamudic”, and Greek. Among the Nabataean texts, 230 had already been photographed and published by S. al-Theeb in 2002 (= ThNUJ). There is a spring not far from the site and archaeological structures have been identiied in the wadi that runs at the foot of the rocky outcrops on which the inscriptions are written. The number, variety, and sometimes very sophisticated character of the inscriptions are an indication that in antiquity this site was probably much more than a simple stop on the caravan road between Дegrā and Petra. Was it a sanctuary? Only a more thorough survey, or excavation, of the archaeological structures will tell. This text was examined during the 2005 Paris workshop. bly dkyr grmw br Κwnyw The editio princeps has gzmw instead of grmw but the latter is preferred here. The vertical r is very similar to 74 Laïla Nehmé figure 33. UJadh 10. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 34. UJadh 15. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). the one in dkyr, except that it is not ligatured from the g on the right. Note the forms of the medial and inal y, the triangular form of the g, and the form of the r, which is a simple vertical stroke. The d is dotted. The name grmw is very well known in Nabataean. The name Κwnyw is new in the Nabataean onomasticon but may be compared to Κwnw, attested in two inscriptions, one from al-ΚUlā, JSNab 202, in which the name could also be read Κwyw, and one from Дegrā, JSNab 285, the reading of which was checked on the original, Κwnw being the best one. UJadh 10 = ThNUJ 38 (Fig. 33) This text was examined during the 2005 Paris workshop. dkyr šlymw br yΚmrw bΓb šmnw This text and the next share the same characteristics. However, they cannot have been written by the same person because their authors do not have the same patronym and the name in UJadh 15 is better read as Šlym{n}. Note the occurrence in this text of both the “classical” and evolved forms of m. The š has a “classical” form and the script has a very upright aspect. The name šlymw is attested elsewhere in Nabataean only in ARNA.Nab 17 and other texts from North Arabia (on which see the commentary under ARNA.Nab 17). The name yΚmrw occurs only once elsewhere, in CIS ii 195, from Umm al-RaΒāΒ, but it is attested in the form yΚmr in an inscription from Jabal СarbūΓ Thulaythah, south-west of Tabūk (al-Theeb 1993: no. 49). The name šmnw is new in the Nabataean onomasticon. UJadh 15 = ThNUJ 30 (Fig. 34) This text was examined during the 2005 Paris workshop. dkyr šlym{n} br mΚnw bΓb w š{l}m dky The sign above dkyr was read as [bl]{y} in the editio princeps but the letter visible on the photograph can hardly be a inal y. The last letter of the irst name is A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 75 figure 35. UJadh 19. (Photograph al-Theeb 2002: 289, facsimile L. Nehmé). figure 36. UJadh 90. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). doubtful because it is a peculiar n, but it cannot really be read as a w because there is no loop and because there is a right angle at the base. Other letters have a peculiar form: the Γ, the l hooked at the top, and the inal m of šlm. Note the very vertical character of the y. The m, except the inal one in š{l}m, is very evolved. On the fourth line, there is probably an uninished dkyr. The name šlym{n} is new in the Nabataean onomasticon whereas the name mΚnw is very well attested in it. the name lΉm occurs in Safaitic and laΉīm “plump” or luΉaym “little plump one” are conceivable names. Note the form of the Ή, which is between the “classical” and evolved forms of the letter. Only the y and the m are clearly evolved. The names {l/r}Ήymw and lwyΜ have not been found before in Nabataean. The latter may be a Jewish name (see Ilan 2002: 182–185). UJadh 19 = ThNUJ 34 (Fig. 35) dkyr ynmw br Ή{b/n}y{b/n}w The d and the k are typical of the transitional texts and most of the characters have evolved forms, particularly the last w, which is very close to Arabic. The patronym is either Ήnynw or Ήbybw, both well attested in Nabataean. The name ynmw has been found before only in JSNab 285, mentioned above, which was checked on the original. This text was examined during the 2005 Paris workshop. dkyr {l/r}Ήymw br lwyΜ bΓb The k of dkyr does not have a horizontal line at its base. The irst letter of the second line is more like a l because it is joined to the following letter (this would not have been the case if it was a r) and it has no bar at the top, unlike the other examples of r. The name is therefore either lΉymw or rΉymw, both of which are new in the Nabataean onomasticon.32 As suggested by M.C.A. Macdonald, For rΉymw, see rΉym, in al-Theeb 1993: no. 60; rΉmh in JSNab 304, the reading of which was checked on the original; rΉmy in JSNab 355; inally rΉymbl in RES 1427D from Petra (this name was listed under rΉymbΚl by A. Negev [1991: no. 1066] but it is deinitely rΉymbl). This inscription was photographed by the author in 2003 and the reading was checked. 32 * UJadh 90 (Fig. 36) UJadh 105 = ThNUJ 128 (Fig. 37) This text was examined during the 2005 Paris workshop. dkyr šΚdw br ΚbdΜyš bšlm All examples of d in this text are dotted. Note the evolved forms of the Μ, the y, and the š. The inal m in šlm has retained its classical form. 76 Laïla Nehmé figure 37. UJadh 105. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 38. UJadh 109. (Photograph al-Theeb 2002: 311, facsimile L. Nehmé). The name šΚdw is very well attested in Nabataean. As for ΚbdΜyš, it may be considered as a defective form of ΚbdΜysy, a theophoric name meaning “the servant of Isis”, which is well attested in Petra (with a samekh), in RES 1431B as reread by J.T. Milik and by the present author on the original, in RES 1382 and 1435 (checked by the author on the original), in Milik & Starcky 1975: 128–129, pl. 47/2 (also checked on the original), and inally, perhaps, in CIS ii 481, as reread by J.T. Milik, an inscription for which, however, no photograph is available. UJadh 109 = ThNUJ 132–133 (Fig. 38) This text was examined during the 2005 Paris workshop. It was republished as part of an article on the Roman period in north-west Arabia in Nehmé 2009: 49–52, ig. 3. bly dkyr phmw br Κbydw šlm šnt 2 ×100 +100 +20 +20 +10 ΜdΉlw Κmrw Μlmlk “Yea, may Phmw son of Κbydw be remembered [and] may he be secure, year 350 [when] they introduced Κmrw [ΚAmrū] the king” This text is dated to AD 455–456. See the commentary in Nehmé 2009. Two of the three examples of d are clearly dotted and it is just possible that the d of Κbydw has a very small and faint dot above it. Note the evolved form of the Μ, the h, the w when ligatured from the right, the Ή, the y, the m, and the r. However, the p, the š, and the A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 77 figure 39. UJadh 172. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). final m in šlm, are “classical” Nabataean. The inal k in Μlmlk is different from any of the shapes which are shown in the tables. It should be noted that it has kept a more or less horizontal line at its bottom, which normally disappears in the classical Nabataean inal k but is present in the Kuic form. Note the use of the Arabic deinite article in Μlmlk. Κbydw is very well attested in Nabataean. The name phmw recurs only in UJadh 375 (see below) and in two unpublished inscriptions from al-ΚUdhayb, north of al-ΚUlā. ΜdΉlw is of course the third person plural perfect of the ΜafΚal form of the Arabic verb dakhala, “to cause to enter, to introduce”. * UJadh 172 (Fig. 39) bly dkyr šΚydw br ΚbdΚdnwn mΜbyΜ šnt 100 +100 +5 +1 +1 (or, less probably, 20 +20 +5 + 1 +1)33 There is doubt about the date because the irst two igures could be read either as two 20s or as two 100s. However, if each of these signs were to be read as 20, one would expect the loop to have remained opened on the left (for examples, see Milik & Seyrig 1958: ig. 2). If the signs are 20s, the date is year 47 of the eparchy, i.e. AD 152–153. If the signs are 100s, the date is year 207 of the eparchy, i.e. AD 312–313. A date in the middle of the second century would it more with the script of the text, in which none of the letters is evolved. The characters are “classical” Nabataean. Finding them at the beginning of the fourth century is surprising, especially if we compare this text with the script of UJadh 297, which comes from the same 33 M.C.A. Macdonald has suggested to me that there are two strokes on top of the ive, not just one, as I initially thought. site (see below and Fig. 45), dated to AD 305–306. ŠΚydw is a common name in the Nabataean inscriptions and ΚbdΚdnwn is attested in JSNab 38, carved on the monumental tomb IGN 100 in Дegrā. MΜbyΜ is either a nisba form, like ΉgryΜ in JSNab 150, or a professional name. The most obvious explanation is that it is the nisba form derived from the toponym mΜb, thus perhaps, as suggested by M.C.A. Macdonald, “the Moabite”. In Hebrew, it is spelt with a w (môΜabî), and mwbyΜ in JSNab 157 is supposed to be the Nabataean form, but one can envisage the shortening of the initial vowel to produce mΜbyΜ. * UJadh 178 (Fig. 40) dkyr Κnmw br zk{yw} bΓb w šlm The last two letters of the patronym may safely be read as zk{yw} because the same name and patronym appear in another text of Umm Jadhāyidh, probably written by the same man. Note the evolved forms of the y, the medial m, and the š. The d and the k have the forms they normally have in the transitional texts. The only d in the text is dotted although it does not have the same shape as the r. Κnmw is a common name in the Nabataean inscriptions. UJadh 219 = ThNUJ 84 (Fig. 41) This text was examined during the 2005 Paris workshop. dkyr ywsp br Κnmw bΓb w šlm The letters in this text are not as ligatured as they normally would be (between the n and the m and between the b and the Γ). It seems that the k had a dot over it but it is also possible that it is a crack in the stone. Note the form of 78 Laïla Nehmé figure 40. UJadh 178. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 41. UJadh 219. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 42. UJadh 222. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). the s, which is different from that in classical Nabataean. Note inally the evolved forms of w, y, and medial m. The Κ clearly sits on the line. On ywsp, see M 1 above. On Κnmw, see the previous text. * UJadh 222 (Fig. 42) dkyr {b/n}pnw br Μbw ypny The medial k in dkyr does not have a horizontal bottom line and looks therefore like a transitional d. Note the A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 79 figure 43. UJadh 248. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 44. UJadh 266. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). form of the r in dkyr, which is almost exactly that of early Arabic. Note also the form of the medial p, which is very close to Arabic as well as the evolved forms of Μ and y. Neither of the names is attested in Nabataean. Names formed with Μbw are rare in the Nabataean onomasticon. * UJadh 248 (Fig. 43) dkyr {Μ}š{p}r br ΚbdΒdpw Note the evolved form of Β as well as of the Μ, the y, and especially the š. All three examples of d have dots above them. The names, if correctly read, are new to the Nabataean onomasticon. They may be derived from the roots З-F-R and С-D-F, both of which exist in Arabic. * UJadh 266 (Fig. 44) Κšylh br Κmyyw šlm Note the very evolved form of the š, especially the irst one, as well as the form of the inal h, the y, and the medial m. The inal m in šlm has kept its classical Nabataean form. The name Κšylh is new in the Nabataean onomasticon and may be derived from the Arabic root Κ-S-L, Κusaylah meaning, among other things, a drop of honey. This interpretation of the name would mean that inal -h here represents a tāΜ marbūΓah and this would possibly be the earliest example of its use, the other earliest example being the Jabal Says grafito (see the contribution of P. Larcher, this volume). According to A. Negev (1991: no. 903), the name Κmyyw is attested 289 times in Sinai.34 * UJadh 297 (Fig. 45) This text was examined during the 2005 Paris workshop. Κwpw br wΜ{y}lw ktb ydh šnt 2x100 “Κwpw son of WΜylw wrote with his hand, year 200” The date is written in the form of two units attached to the symbol of the hundred, thus 200, and the text is dated to AD 305–306. It is very clear and there is no ambiguity in the reading. 34 Note that, contra Negev, the name is not attested in ARNA.Nab 20, where it is Μmyw. 80 Laïla Nehmé figure 45. UJadh 297. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 46. UJadh 298. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). Most of the letters of this text are very evolved: the Μ, the inal h, the y, and the medial and inal t. The name Κwpw is new in the Nabataean onomasticon, but ΚAwf is common in Arabic. The name wΜylw has not been found before, though wΜylt occurs once and wΜlw is very common. It is very probable that it represents a diminutive form *wuΜayl. šlm šlm Note the very evolved forms of the Μ, the inal h, the y, the m, the r, and the š. The two examples of inal m in šlm have retained their classical form. Κmrw is a common name in the Nabataean inscriptions and it is preceded here, for the irst time, by Μbw. On Κšylh, see above, UJadh 266. * UJadh 298 (Fig. 46) * UJadh 309 (Fig. 48) This text was examined during the 2005 Paris workshop. ΓΚlbh br ΜlΉrt šlm Note the very evolved forms of the Μ, the inal h, the Ή, the Γ, the r, the š, and the inal t. The inal m has retained its classical form. The name ΓΚlbh is new in the Nabataean onomasticon and its origin is dificult to trace. The name ΜlΉrt occurs twice in JSNab 382 and is very well known in early Arabic. bly dkyr šly br Μwšw br ΜlΉ{b/n}h bΓb w šlm w ktbΜ dnh ktb ywm Ήd btšry šnt mΜt w tšΚyn “Yea! May Šly son of Μwšw son of ΜlΉnΜ be remembered for good and may he be secure. And this writing he wrote the irst day of Tišrī, year one hundred and ninety” The text, which is a commemorative text of the dkyr + šlm type, is dated to the month of Tišrī year 190 of the Province, i.e. AD 295. Part of the irst letter of UJadh 309 is carved over part of the last letter of UJadh 317, a Nabataean text written * UJadh 299 (Fig. 47) This text was examined during the 2005 Paris workshop. Μbw Κmrw br Κšylh A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 81 figure 47. UJadh 299. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 48. UJadh 309. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 49. UJadh 317 and UJadh 309. (Photograph Darb al-Bakrah project). in more “classical” characters, which reads zbdΚdnwn br ΜnΚm (Fig. 49). The size of the text is not known. The reading is perfectly clear and there is no ambiguity apart from the possible confusion between b and n in the name ΜlΉ{b/n}h. This name does not occur in other Nabataean texts, although Ήnh is attested once, in JSNab 31, on tomb IGN 64 in MadāΜin СāliΉ, possibly as a woman’s name, Дannah. The Arabic roots from 82 Laïla Nehmé figure 50. UJadh 360. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 51. UJadh 367. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). which the name could be derived are either Д-N-N, “to be affected by an intense emotion”, or Д-B-B, “to be in love”, or L-Д-N, “to be intelligent”. The alternatives, Е-B-B, “to amble, to trot”, Е-N-N, “to speak nasally”, and L-Е-N, “stench”, are less appropriate roots for a name. The names Дunn, Дabbah, and Дabbā (but not *Al-Дabbah), can be found in W. Caskel’s Вamharat an-Nasab (1966), but it does not give any name derived from the root L-Д-N. Theoretically, the letters Μ-l at the beginning of the name could represent either the article or the ΜafΚal form of a root beginning with l, but the latter hypothesis is less likely despite the fact that the name ΜlΉn, “more intelligent” (?), is attested in Safaitic in WH 1322 and 1328 while ΜbΜlΉn is attested in Dadanitic in JSLih 291. The inal h in the name is probably a tāΜ marbūΓah. ΜlΉnh is the name of the grandfather of the author. His son and grandson, Μwšw and Ίly, bear names which are common in the Nabataean inscriptions. Note that šnt is written with an open form of the inal t (no. 4 on Fig. 4) whereas the inal t in mΜt (a word which is normally spelt mΜh in Nabataean in the absolute state) is written with a looped t. Note the forms of the Μ, the inal h, the Ή, the medial and inal y, the m, and the š. The d and the k have the typical form of these letters in the transitional script. The medial t in ktbΜ and ktb has a classical Nabataean form as has the inal m. UJadh 360 = ThNUJ 62 (Fig. 50) This text was examined during the 2005 Paris workshop. mΚnw br grmw MΚnw and grmw are well-known Nabataean names. * UJadh 367 (Fig. 51) This text was examined during the 2005 Paris workshop. Ήbšh br ΚbdΜlΜšΉn br ΚbdΜlΜšΉn The father and grandfather of the author have the same name. This text offers two very nice examples of lām-alif combined into one grapheme. The letters Μ, Ή, r, and š have very evolved forms. The name Ήbšh is new in the Nabataean onomasticon. It may be derived from the Arabic root Д-B-З, the basic meaning of which is “to collect”. The name of the father and of the grandfather is also new. The name consists of the Arabic article and the Arabic root З-Д-N, which has several meanings in the ΜafΚal form. The fact that it is preceded by Κbd suggests that ΜšΉn is a divine name but I have found no parallel for this.35 35 The closest is the name šhr, which is an epithet of the moon god (Cross 1986: 391), but the inal n is clear. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 83 figure 52. UJadh 375. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 53. UJadh 405. (Photograph Darb al-Bakrah project, facsimile L. Nehmé). figure 54. ΚUlā 1. (Photograph MadāΜin СāliΉ archaeological project, facsimile L. Nehmé). UJadh 375 = ThNUJT 38 (Fig. 52) This text was examined during the 2005 Paris workshop. bly dkyr Κbydw br phmw bΓb w šlm Almost all the letters of this text are evolved, except the inal m in šlm. The two examples of d are dotted. Κbydw is a common name in the Nabataean inscriptions. On phmw, see UJadh 109 above. UJadh 405 = ThNUJ 145 (Fig. 53) dkyr Ή{b/n}y br nΉmy The d is dotted. The letters of this text all have an evolved form. The name Ήny is attested only once in Nabataean, in the Stiehl inscription, but it is common in Safaitic. A name Ήby was thought to be attested in JSNab 14 but it was reread as Ήpy by J. Healey (1993: 148). NΉmy may be the Hebrew name nΉmyΜ/nΉmyh (Ilan 2002: 197). The alternative reading bΉmy is less satisfactory and the second letter is not likely to be a g, which allows us also to exclude ngmy. * ΚUlā 1 (Fig. 54) dkyr {l/n}Ήmw br yhwdΜ bΓb The two examples of d are dotted. The letters in this text are almost all evolved. The name lΉmw (see UJadh 19 for lΉm) is attested in JSNab 136 from MadāΜin СāliΉ but if we consider that the irst letter of the name in ΚUlā 1 is rather short for a l and in view of the Jewish patronym, the Jewish name nΉmw is more likely (see nΉmy in the previous text). YhwdΜ is of course the Hebrew name Judah (Ilan 2002: 112–125), which is attested in the form yhwdh in the signatures of the Starcky Papyrus, line 39 (Yardeni 2001: 129, 133). 84 Laïla Nehmé Appendix 3. Names contained in the inscriptions listed in Appendix 2 Previously unpublished inscriptions are followed by *. Μbw ypny Μbw Κmrw Μbw qΓ{b/n}h ΜwΊw ΜlΉ{b/n}h ΜlΉrt {Μ}Ί{p}r by{Β}w {b/n}pnw gdymt grmw grΚm hnΜw wΜ{y}lw [w]Μlw zk{yw} Ή{b/n}y Ή{b/n}y{b/n}w ΉbΊh Ήny Ήrtw Ήrtt {Γ/k}ΉΊw ΓΚlbh yhwdΜ ywsp ynmw yΚmrw yΚny {k/Γ}ΉΊw kΚbw lwyΜ {l/n}Ήmw {l/r}Ήymw mwyh mΉrbw mlkw UJadh 222* UJadh 299* QN 2* UJadh 309* UJadh 309* UJadh 298* UJadh 248* S3 UJadh 222* (see also under {n/b}pnw) LPNab 41 UJadh 3 (= ThNUJ 48), UJadh 360 (= ThNUJ 62) Ar 19* JSNab 18 UJadh 297* S1 UJadh 178* UJadh 405 (= ThNUJ 145) UJadh 90* UJadh 367* Stiehl S1 JSNab 17 S 1 (see also under {k/Γ}ΉΊw) UJadh 298* ΚUlā 1* M 1, UJadh 219 (= ThNUJ 84) UJadh 90* UJadh 10 (= ThNUJ 38) CIS ii 963 S 1 (see also under {Γ/k}ΉΊw) JSNab 17, 18 UJadh 19 (= ThNUJ 34) ΚUlā 1* (see also under {n/l}Ήmw) UJadh 19 (= ThNUJ 34) (see also under {r/l}Ήymw Stiehl S1 S3 mΚnw r{mn}h rqwΊ Ίly Ίlymw Ίlym{n} ΊmwΜl Ίmnw ΊΚdw UJadh 15 (= ThNUJ 30), UJadh 360 = (ThNUJ 62) S3 UJadh 405 (= ThNUJ 145) UJadh 222* (see also under {b/n}pnw) ΚUlā 1* (see also under {l/n}Ήmw) UJadh 105 (= ThNUJ 128) UJadh 367* B 3* JSNab 17 UJadh 172* UJadh 248* UJadh 109 (= ThNUJ 132–133), UJadh 375 (= ThNUJT 38) Stiehl JSNab 18 S1 ARNA.Nab 17 UJadh 3 (= ThNUJ 48) UJadh 297* UJadh 266* Stiehl, Ar 19*, UJadh 109 (= ThNUJ 132–133) S1 UJadh 178*, UJadh 219 (= ThNUJ 84) M1 UJadh 266*, UJadh 299* UJadh 109 (= ThNUJ 132–133), UJadh 375 (= ThNUJT 38) LPNab 41 UJadh 19 (= ThNUJ 34) (see also under {l/r}Ήymw) M1 JSNab 17 LPNab 41, UJadh 309* ARNA.Nab 17, UJadh 10 (= ThNUJ 38) UJadh 15 (= ThNUJ 30) Stiehl UJadh 10 (= ThNUJ 38) UJadh 105 (= ThNUJ 128) ΊΚydw tymΜlhy UJadh 172* CIS ii 963 mrΜlqyΊ nΉmy {n/b}pnw {n/l}Ήmw ΚbdΜyΊ ΚbdΜlΜΊΉn Κbdw Κbdmnwtw ΚbdΚdnwn ΚbdΒdpw Κbydw Κdywn Κdmn ΚwydΜlt Κwydw Κwnyw Κwpw Κmyyw Κmrw Κnymw Κnmw Κrr ΚΊylh phmw phrw {r/l}Ήymw A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 85 Sigla and abbreviations Ar Site name: al-ΚArniyyāt (see Fig. 7). ARNA.Nab Nabataean inscriptions published or mentioned in Winnett & Reed 1970. B Site name: BoΒra (see Fig. 7). CIS ii Corpus Inscriptionum Semiticarum. Pars II. Inscriptiones Aramaicas continens. Paris: Imprimerie nationale, 1889–1954. CIS v Corpus Inscriptionum Semiticarum. Pars V. Inscriptiones Saracenicas continens. Tomus 1. Inscriptiones Safaiticae. Paris: Imprimerie nationale, 1950–1951. Ir Site name: ΚĪrīn. JSLih “LiΉyanite” (Dadanitic) inscriptions published in Jaussen & Savignac 1909– 1914. JSNab Nabataean inscriptions published in Jaussen & Savignac 1909–1914. LPArab Arabic inscriptions published in Littmann 1949. LPNab M MS NDGS QN RES S Stiehl ThNUJ ThNUJT UJadh ΚUlā WH Nabataean inscriptions published in Littmann 1914. Site name: Mābiyāt (see Fig. 7). Site name: MadāΜin СāliΉ (see Fig. 7). Nabataean inscriptions published in Negev 1967. Site name: QāΚ an-Nqayb (see Fig. 7). Répertoire d’épigraphie sémitique. Paris, 1900–1968. Site name: Sakākā (see Fig. 7). Nabatean inscription published in Stiehl 1970. Nabataean inscriptions published in alTheeb 2002. Nabataean inscriptions published in alTheeb 2005. Site name: Umm Jadhāyidh (see Fig. 7). Site name: al-ΚUlā (see Fig. 7). Safaitic inscriptions published in Winnett & Harding 1978. References Altheim F. & Stiehl R. 1968. Die Araber in der alten Welt. v/1. Weitere Neufunde — Nordafrika bis zur Einwanderung der Wandalen — Аū Nuwās. Berlin: de Gruyter. Calvet Y. & Robin R. (eds) 1997. Arabie heureuse, Arabie déserte. Les antiquités arabiques du musée du Louvre. (Notes et documents des musées de France, 31). Paris: Éditions de la Réunion des musées nationaux. Cantineau J. 1930–1932. Le Nabatéen. (2 volumes). Paris: Leroux. Caskel W. 1966. Вamharat an-Nasab. Das genealogische Werk des HiΊām ibn MuΉammad al-Kalbī. (2 volumes). Leiden: Brill. Cross F.M. 1986. A New Aramaic Stele from TaymāΜ. Catholic Biblical Quaterly 48: 387–394. al-Ghabbān ΚA.I. 2008. The inscription of Zuhayr, the oldest Islamic inscription (24 AH/AD 644–645), the rise of the Arabic script and the nature of the early Islamic state. (Translation and concluding remarks by R.G. Hoyland). Arabian Archaeology and Epigraphy 19: 210–237. al-Ghul O. 2004. An Early Arabic Inscription from Petra Carrying Diacritic Marks. Syria 81: 105–118. Grimme H. 1936. À propos de quelques grafites du temple de Ramm. Revue Biblique 45: 90–95. Grohmann A. 1932. Aperçu de papyrologie arabe. Études de papyrologie 1: 23–95, pls 1–9. 1971. Arabische Paläographie. ii. Das Schriftwesen. Die Lapidarschrift. (Österreichische Akademie der Wissenschaften, Wien. Philosophisch-historische Klasse. Denkschriften, 94/2). Vienna: Böhlaus. 86 Laïla Nehmé Gruendler B. 1993. The Development of the Arabic Scripts. From the Nabatean Era to the First Islamic Century According to Dated Texts. (Harvard Semitic Studies, 43). Atlanta, GA: Scholars Press. Healey J.F. 1990–1991. Nabataean to Arabic: Calligraphy and Script Development Among the Pre-Islamic Arabs. In J. Bartlett, D. Wasserstein & D. James (eds), The Role of the Book in the civilisations of the Near East. Proceedings of the Conference held at the Royal Irish Academy and the Chester Beatty Library, Dublin, 29 June – 1 July 1988. Manuscripts of the Middle East 5: 41–52. 1993. The Nabataean Tomb Inscriptions of MadaΜin Salih. Edited with Introduction, Translation and Commentary. With an Arabic section translated by Dr. Solaiman al-Theeb, Department of Archaeology, King Saud University, Riyadh. (Journal of Semitic Studies Supplement, 1). Oxford: Oxford University Press. 2002. Nabataeo-Arabic: Jaussen-Savignac nab. 17 and 18. Pages 81–90 in J.F. Healey & V. Porter (eds), Studies on Arabia in Honour of Professor G. Rex Smith. (Journal of Semitic Studies Supplement, 14). Oxford: Oxford University Press. Healey J.F. & Smith G.R. 1989. Jaussen-Savignac 17 — The Earliest Dated Arabic Document (A.D. 267). Atlal 12: 77–84, pl. 46. [Arabic version pp. 101–110]. Hoyland R. (this volume). Mount Nebo, Jabal Ramm, and the status of Christian Palestinian Aramaic and Old Arabic in Late Roman Palestine and Arabia. In M.C.A. Macdonald (ed.), The development of Arabic as a written language. (Supplement to the Proceedings of the Seminar for Arabian Studies 40). Oxford: Archaeopress. Ilan T. 2002. Lexicon of Jewish Names in Late Antiquity. Part I. Palestine 330 BCE – 200 CE. (Texts and Studies in Ancient Judaism, 91). Tübingen: Mohr Siebeck. Jaussen A. & Savignac R. 1909–1914. Mission archéologique en Arabie. i. De Jérusalem au Hedjaz, Médain Saleh. ii. El-ΚEla, d’Hégra à Teima, Harrah de Tebouk. (2 volumes). Paris: Leroux/Geuthner. Jaussen A., Savignac R. & Vincent H. 1905. ΚAbdeh (4–9-février 1904) (suite). Revue Biblique Internationale [N.S.] 2: 74–89, 235–257, pls 6–10. Littmann E. 1914. Nabataean Inscriptions from the Southern Дaurân. (Publications of the Princeton University Archaeological Expeditions to Syria in 1904–1905 and 1909. Division IV. Section A). Leiden: Brill. 1949. Arabic inscriptions. (Publications of the Princeton University Archaeological Expeditions to Syria in 1904–1905 and 1909, Division IV. Section D). Leiden: Brill. Macdonald M.C.A. 1986. ABCs and Letter Order in Ancient North Arabian. Proceedings of the Seminar for Arabian Studies 16: 101–168. 2003. Languages, Scripts, and the Uses of Writing Among the Nabataeans. Pages 36–56, 264–266 (endnotes), 274–282 (references) in G. Markoe (ed.), Petra Rediscovered: Lost City of the Nabataeans. New York: Abrams/Cincinnati, OH: Cincinnati Art Museum. 2008. Old Arabic (Epigraphic). Pages 464–477 in K. Versteeg (ed.). Encyclopedia of Arabic Language and Linguistics, iii. Leiden: Brill. 2009a. ARNA Nab 17 and the transition from the Nabataean to the Arabic script. Pages 207–240 in W. Arnold, M. Jursa, W.W. Müller & S. Procházka (eds), Philologisches und Historisches zwischen Anatolien und Sokotra. Analecta Semitica In Memoriam Alexander Sima. Wiesbaden: Harrassowitz. 2009b. A note on new readings in line 1 of the Old Arabic grafito at Jabal Says. Semitica et Classica 2: 223. (forthcoming). On the uses of writing in ancient Arabia and the role of palaeography in studying them. A glimpse of the development of the Nabataean script into Arabic based on old and new epigraphic material 87 Milik J.T. & Seyrig H. 1958. Trésor monétaire de MurabbaΚât. Revue Numismatique, 6e série, 1: 11–26, pls 1–3. Milik J.T. & Starcky J. 1975. Inscriptions récemment découvertes à Pétra. Annual of the Department of Antiquities of Jordan 20: 111–130, pls. 37–47. al-Muaikil Kh.I. [= al-MuΚayqil Kh.I.] 1993. Naqshān Κarabiyyān mubakkarān min sakākā. Addarah 19/3: 112–131. 2002. Pre-Islamic Inscriptions from Sakākā, Saudi Arabia. Pages 157–169 in J.F. Healey & V. Porter (eds), Studies on Arabia in Honour of Professor G. Rex Smith. (Journal of Semitic Studies Supplement, 14). Oxford: Oxford University Press. (forthcoming). [Inscriptions from the region of Sakākā]. Adumatu. al-Muaikil Kh.I. & al-Theeb S.ΚA. [= al-MuΚayqil Kh.I. & al-Dhīyīb S.ΚA.] 1996. Al-āthār wa-Μl-kitābāt al-nabaΓiyyah fī minΓaqat al- jawf. Riyadh: JāmiΚat al-Malik SaΚūd. al-Muraykhī M. (in press). КarΉ jadīd Ήawla manshaΜ al-Ήarf al-Κarabī wa mawΓinihi al-aΒlī fī ΡawΜi muktashafāt athariyyah jadīdah. Dirāsāt fī Μl-āthār 2. al-Muraykhī M. & al-Ghabbān ΚA.I. 2001. Naqsh wāΜil bin al-jazzāz al-tidhkārī al-muΜarrikh bi-Κām 410 M. Silsilat mudāwalāt al-liqāΜ al-Κilmī al-sanawī li-l-jamΚiyyah 3: 127–153. al-Najem M. & Macdonald M.C.A. 2009. A new Nabataean inscription from TaymāΜ. Arabian Archaeology and Epigraphy 20: 208–217. Negev A. 1967. New Dated Nabatean Grafiti from the Sinai. Israel Exploration Journal 17: 250–255, pl. 48. 1981. Nabatean, Greek and Thamudic Inscriptions from the Wadi Haggag–Jebel Musa Road. Israel Exploration Journal 31: 66–71, pls. 7B–10. 1991. Personal Names in the Nabataean Realm. (Qedem, 32). Jerusalem: Hebrew University of Jerusalem. Nehmé L. 2009. Quelques éléments de rélexion sur Hégra et sa région à partir du IIe siècle après J.-C. Pages 37–58 in J. Schiettecatte & Chr.J. Robin (eds), L’Arabie à la veille de l’Islam. Bilan clinique. (Orient & Méditerrannée, 3). Paris: De Boccard. Robin C.J. 2001. Les inscriptions de l’Arabie antique et les études arabes. Arabica 48: 509–577. 2006. La réforme de l’écriture arabe à l’époque du califat médinois. Mélanges de l’Université Saint-Joseph 59: 319–364. 2008. Les Arabes de Дimyar, des « Romains » et des Perses (IIIe–VIe siècles de l’ère chrétienne). Semitica et Classica 1: 167–202. Said S. & al-Hamad M. 2004. Three Short Nabataean Inscriptions from Umm al-Jimāl. Proceedings of the Seminar for Arabian Studies 34: 313–318. Savignac R. 1932. Chronique. Notes de voyage. — Le sanctuaire d’Allat à Iram. Revue Biblique 41: 581–597, pls. 16–19. Sokoloff M. 1990. A Dictionary of Jewish Palestinian Aramaic. (Dictionaries of Talmud, Midrash and Targum, 2). RamatGan: Bar Ilan University Press. Starcky J. 1978. Langue, écriture et inscriptions. Pages 47–52 in F. Baratte (ed.), Un royaume aux conins du désert: Pétra et la Nabatène. Catalogue de l’exposition du Muséum de Lyon tenue du 18 novembre 1978 au 28 février 1979. Lyon: Muséum de Lyon. 88 Laïla Nehmé Stiehl R. 1970. A New Nabatean Inscription. Pages 87–90 in R. Stiehl & H.E. Stier (eds), Beiträge zur alten Geschichte und deren Nachleben. Festschrift für Franz Altheim zum 6.10.1968. ii. Berlin: de Gruyter. al-Theeb S.ΚA. [=al-Dhīyīb S.ΚA.] 1993. Aramaic and Nabataean Inscriptions from North-West Saudi Arabia. Riyadh: Maktabat al-Malik Fahd al-WaΓaniyyah. 1994a. Dirāsah taΉlīliyyah jadīdah li-nuqūsh nabaΓiyyah min mawqiΚ al-qalΚah bi-Μl-jawf bi-Μl-mamlakah alΚarabiyyah al-saΚūdiyyah. Majallat JāmiΚat al-Malik SaΚūd 6 (Arts 1): 151–194. 1994b. Two New Dated Nabataean Inscriptions from al-Jawf. Journal of Semitic Studies 39: 33–40. 2002. Nuqūsh jabal umm jadhāyidh al-nabaΓiyyah. Riyadh: Maktabat al-Malik Fahd al-WaΓaniyyah. 2005. Nuqūsh nabaΓiyyah fī Μl-jawf, al-Κulā, taymāΜ, al-mamlakah al-Κarabiyyah al-saΚūdiyyah. Riyadh: Maktabat al-Malik Fahd al-WaΓaniyyah. Winnett F.V. & Harding G.L. 1978. Inscriptions from Fifty Safaitic Cairns. (Near and Middle East Series, 9). Toronto: University of Toronto Press. Winnett F.V. & Reed W.L. 1970. Ancient Records from North Arabia. (Near and Middle East Series, 6). Toronto: University of Toronto Press. Yardeni A. 2000. Textbook of Aramaic, Hebrew and Nabataean Documentary Texts from the Judaean Desert and Related Material. A. The Documents. B. Translation, Palaeography, Concordance. (2 volumes). Jerusalem: Ben-Zion Dinur Center for Research in Jewish History. 2001. The Decipherment and Restoration of Legal Texts from the Judaean Desert: A Reexamination of Papyrus Starcky (P. Yadin 36). Scripta Classica Israelitica 20: 121–137. Author’s address Laïla Nehmé, CNRS UMR 8167, Orient & Méditerranée, 27, rue Paul Bert, 94204 IVRY CEDEX, France. e-mail laila.nehme@ivry.cnrs.fr