Al-Jallad. 2019. The Classification of the Languages of North Arabia: Remarks on the Semitic Language Family Tree of the 2 nd edition of Routledge's The Semitic Languages

The Classification of the Languages of North Arabia: Remarks on the Semitic Language Family Tree of the 2nd edition of Routledge’s The Semitic Languages Ahmad Al-Jallad Ohio State University al-jallad.1@osu.edu ‫ﻣﻠﺨﺺ‬ ‫ اﻟﻠﻐﺎت‬:‫ﺗﻨﺎﻗﺶ ھﺬه اﻟﻤﻘﺎﻟﺔ ﺗﺼﻨﯿﻒ ﻟﻐﺎت ﺷﻤﺎل اﻟﺠﺰﯾﺮة اﻟﻌﺮﺑﯿﺔ اﻟﻘﺪﯾﻤﺔ واﻟﻌﺮﺑﯿﺔ ﻛﻤﺎ ھﻲ ﻣﻌﺮوﺿﺔ ﻓﻲ ﻣﻘﺪﻣﺔ ﻛﺘﺎب‬ .‫إﯾﻞ‬-‫اﻟﺴﺎﻣﯿﺔ ﻟﺼﺎﺣﺒﯿﮭﺎ ھﻮﯾﻨﺮﺟﺎرد وﺑﺎت‬ Abstract This paper discusses the classification of the languages of Ancient North Arabia and Arabic as presented in the introduction of Routledge’s The Semitic Language’s handbook. Keywords: Arabic, Historical Linguistics, Language Classification, Semitic language family, Arabian epigraphy 1- Introduction The second edition of Routledge’s The Semitic Languages provides a state-of-the-art introduction to the classification of the Semitic languages(Huehnergard and Pat-El 2019: 1-21). It includes a comprehensive cladistic model based on the principle of shared innovations. And unlike most trees representing the Semitic language family, this one includes the poorly documented epigraphic languages of North Arabia. When this model was compiled, there was not yet a comprehensive grammatical summary of the epigraphic languages of North Arabia and so the authors followed the general treatment of epigraphist, namely, to lump them all together in a ©The International Journal of Arabic Linguistics (IJAL) Vol. 5 Issue 2 (pp. 86-99) single sub-grouping – “North Arabian”.1 A welcome departure from the old model, however, is the creation of a new sub-grouping comprising Safaitic and Arabic: “Arabic-Safaitic”. In this short note, I will discuss the classification of the North Arabian “languages” based on evidence from newly discovered and deciphered texts, and push back on the idea of an Arabic-Safaitic subgrouping. It is hoped that my remarks will complement this outstanding presentation of the Semitic language family. Figure 1: the classification of the languages of North Arabia (from Huehnergard and PatEl 2019) Figure 2: textual description of the classification (from Huehnergard and Pat-El 2019) 2- Ancient North Arabian scripts or languages? Is there such a thing as a ‘North Arabian’ genetic sub-grouping? The poor state of documentation provides us with a very low resolution view of Ancient Arabia’s linguistic map. But the little evidence that we do have absolutely rules out treating the Ancient North Arabian (ANA) epigraphic corpora as closely related varieties of a single language, as had once been thought. Before moving on to the linguistic facts, it is perhaps important to highlight that the terms used in the classification under discussion refer primarily to ‘scripts’ and not languages as 1 See Beeston 1981; Macdonald 2000, 2004, but Macdonald correctly notes in 2000 that the classification is preliminary and it may turn out that the Ancient North Arabian ‘dialects’ are indeed more distinct once more evidence emerges. 87 such. While interrelationships between these scripts have not yet been sorted out, it is clear that they in no way constitute a single sub-grouping that derives from the South Arabian script, and there is moreover no evidence to suggest that they descend from a single Proto-North Arabian script.2Huehnergard and Pat-El correctly state that the chronology of Ancient North Arabian documentation remains vague, only adding to this complexity. The following approximate dates are possible: Dadanitic: Latter half of the first millennium BCE (See the Leiden University dissertation of F. Kootstra for a comprehensive discussion of the chronology of Dadanitic.). Hismaic: Definitely written during the Nabataean period (1 st c. BCE-1st c. CE); unclear how much earlier (On the chronology of Hismaic, see King 1990.). Taymanitic: Mid-first millennium BCE (but possibly as early as the late 2 nd millennium BCE) (The chronology of Taymanitic is discussed in detail by Kootstra (2016)). Thamudic B: Early first millennium BCE based on the contents of some inscriptions (See Macdonald (2000), and for the other Thamudics as well.). Thamudic C: No dated texts. It seems to be much earlier than Hismaic as there are no interactions as such between Thamudic C or Hismaic in areas where they geographically overlap. Thamudic D: One inscription connected to a Nabataean text dated to 267 CE from the Ḥigāz. No other datable inscriptions are known. Thamudic F: This category is missing in the classification under discussion. There are no dated texts (On the tentative dating of these texts, see Prioletta (2018)). 3- Linguistic Classification The attempt to classify Hismaic, Dadanitic, Taymanitic and the various Thamudics together into a ‘North Arabian’ family is dependent upon shared morphological innovations. Early arguments in support of a ‘North Arabian’ subgrouping were dependent entirely upon the shape of the definite article, a prefixed h-. Huehnergard and Rubin (2011) have already argued that the definite article is not a suitable feature for sub-classification,3 and in the case of North 2 The developmental trajectories of the North Arabian scripts were the subject of a presentation I gave at the 2017 Seminar for Arabian Studies in London; the paper is currently in preparation. 3 This argument was made broadly for Central Semitic by Huehnergard (1995) and in particular for Arabic in (2017). 88 Arabian it would appear to be even less significant as we are dealing not with the position and morphology of the article but simply its phonetic realization. Even so, the h-article alone does not support the existence of a North Arabian subgrouping. Various article forms are attested across and within the North Arabian corpora (For the data and citations for individual corpora, see AlJallad (2018)): Dadanitic: h-, hn (before laryngeals), but rarely ʾ(l) Hismaic: no definite article Thamudic B: h-, ʾl (See Inscription 1 in the appendix.) Thamudic C: hThamudic D: no definite article attested Thamudic F: h, hl, ʾ, ʾl In Safaitic, we have a similar diversity of article forms: h, ʾ, ʾl and no article. Another feature of the North Arabian subgrouping is the merger of *s1 and *s3, something witnessed in Classical Arabic and all modern dialects as well. But even this sound change does not hold for the entire corpus.4 Hismaic, Dadanitic, Thamudic B: merger of s1 and s3 Taymanitic: merger of s3 and ṯ Thamudic C: s3 possibly kept distinct Thamudic D: Unknown 4- Key features in individual linguistic corpora Taymanitic did not participate in the merger of s3 and s1. In fact, it attests a sound change that is otherwise typical of Northwest Semitic: w > y /_#. This change is occasionally encountered in Thamudic F in a variant of the verb wqr ‘to engrave’, written sometimes as yqr. But as Prioletta (2018) has pointed out, this could very well be the prefix conjugation. For a comprehensive study of the features of Taymanitic and its classification, see Kootstra (2016). Thamudic B is found across the Peninsula and was most certainly used to write more than one linguistic variety. Most of the texts contain short prayers and signatures, and do not provide 4 Again, for the data of the individual corpora, see Al-Jallad (2018). At the time of that publication, the existence of s3 in Thamudic C was not yet known; see below for the evidence. 89 enough linguistic information for classification. For a description of its grammatical features, see Al-Jallad (2018:25-26). Thamudic C, in some respects, seems to record the most divergent language(s) in the ANA corpora, and it also comprises considerable internal diversity. While it has a prefixed definite article h-, some inscriptions also exhibit mimation, as in ASA. It has merged the sibilant z and interdental ḏ to z, hence the demonstratives zn (ms.) and zt (fs.) But perhaps more remarkable is the fact that the onset of the 3 rd person pronouns remainsa sibilant (See Al-Jallad (2018: 26-27) for more details). For example: mlt-s1 ‘his people’5 f-s1w ‘with him’; and possibly an independent form s1wʾ ‘that’6 What is more, the 3rd feminine singular seems to have been realized as an s3. An unpublished inscription from the area of Medina reads:7 ḥwyʾ wdd-s3 ‘long life/greetings to her love’ or ḥwyʾ, her love’ The s3 shape is , which is attested here for the first time in Thamudic C. If my interpretation is correct, it would suggest that an important Modern South Arabian and Ḥaḍramitic innovation was attested in pre-Arabic times as far north as the central Ḥigāz! Full confirmation of this feature, however, must await the discovery of more texts. The Thamudic D texts, which span from the area of Taymāʾ to the Ḥigāz, give us no information for classification – even the definite article is not securely attested. In fact, when these texts contain anything more than personal names they tend to be uninterpretable, suggesting that their language was quite remote from Arabic. Consider for example: Thamudic Inscription #63 in Winnett and Reed (1970) ʾnʾs1gwʾfry wḫmyʾtfʿlmr{ḥ/ṭ}ḫm 5 This comes from the interpretation of WRTH 80 in footnote 70 of Al-Jallad (2018). These attestations come from Esk 204: wdd f-s¹w//tʾlʿs¹ s¹wʾ//wdd. I would now translate the first line as ‘greetings be with him’, according to the standard Thamudic C formula. The second line may contain the independent 3 rd person pronoun s¹wʾ, but confirmation is dependent upon the identification of tʾlʿs¹. It appears to be a verb form but the root, lʿs, does not provide a suitable meaning in any Semitic language. 7 The text can be found on the Twitter account @mohammed93athar; see inscription 2. 6 90 Thamudic F is still a nebulous category, comprising texts spanning from the ʿAsīr to south of Riyāḍ. Most of the published texts come from the region of Nagrān and contain only personal names. However, a few new texts from the region of Abha (unpublished; see inscription 3) are known and these contain a short sentence. Next to a rock drawing of a camel and lion, our author carves: gʿnt wbl w bʿd ḥbm ḏt ‘(1) Gʿntwbl(2) and this (drawing) is for (his) love’ The short text provides us with the feminine singular demonstrative ḏt, a form not attested in Arabic, and mimation on the noun ḥb-m ‘love’. The meaning of the verb wbl is not clear in this context – it could be a personal name. This brings us to Hismaic. As our discussion has shown, there are no isoglosses to support the placement of the aforementioned epigraphic categories into a single linguistic subgrouping. Hismaic, however, is the second best documented corpus of ANA scripts and it must unequivocally be placed with Classical Arabic and Safaitic as it exhibits several innovations characteristic of those groupings. In fact, the longest Hismaic texts are the closest examples to Classical Arabic prose we have in the pre-Islamic period (Zwettler and Graf 2004). Among the most interesting innovations connecting it with Classical Arabic and the modern dialects is the optional use of a suppletive imperative form of the verb ‘to give’, hāt, a frozen h-causative of the verb ʾatawa ‘to come’, and the vocative -m on divine names: Hismaic h-ltm/hā-llātumma/ ‘O Allāt’ vs. Classical Arabic ʾallāhumma ‘O Allāh’ (Al-Jallad 2018, §3). It is possible that Dadanitic and Arabic are rather closely related. But Dadanitic did not participate in a few Arabic innovations – it, for example, retains the deictic use of the 3 rd person pronouns, e.g. hʾ ‘that’ and, at least in some varieties, used an h-causative (Al-Jallad 2018, §4). The corpus is varied and it seems that Arabic was at least one language used at the ancient oasis and may have been an important substrate or adstrate language. So with this, it is clear that there is no justification for a North Arabian subgrouping comprising all of the aforementioned epigraphic corpora. Thamudic C and Taymanitic most certainly cannot be classified in the same group as Thamudic B and Dadanitic (both of which merge s1 and s3). In fact, it is entirely unclear if Thamudic C and D should even be considered Central Semitic – the evidence is simply not there and their otherwise exotic nature cautions 91 against simplifying assumptions about their linguistic affiliation. I would therefore suggest that we keep these groups out of our classifications or perhaps put them simply under West Semitic (based on the use of the suffix conjugation) until more evidence comes forth. 5- Is there any validity to the Arabic-Safaitic subgrouping? As I have argued in my 2018 article (and elsewhere), the Safaitic inscriptions exhibit the same linguistic innovations as Classical Arabic and therefore treating these inscriptions as “not Arabic” is entirely arbitrary and based on non-linguistic arguments, such as script, geography, and/or chronology. But the subgrouping Arabic-Safaitic implies that there was a Proto-ArabicSafaitic, and that Arabic (all of the modern dialects and Classical Arabic), on the one hand, and Safaitic, on the other, constitute a main split in the family. In order to justify this, we would have to prove that “Arabic proper”,8 which it is argued comprises Classical Arabic and the modern dialects, shares innovations that distinguish all of its varieties from Safaitic or the opposite. I can think of no innovations to support this division. The most obvious difference is the article form, ʾal for “Arabic” and ha for “Safaitic”, but even here both suggested branches of “Arabic-Safaitic” show internal variation and one form cannot be proposed for the ancestor of either group. - Articles attested in “Arabic proper”: am, an, aC-, ʾal (with various patterns of assimilation); emphatic article hal9 - Articles attested in “Safaitic”: h(n), ʾ, ʾl We should also ask what evidence exists to assume that all modern dialects and Classical Arabic (an already vague and to a large degree artificial amalgamation of ancient dialects) constitute a single grouping to the exclusion of Safaitic (and Hismaic and Nabataean, for that matter!). I can think of no linguistic innovations to group all of these forms together. In fact, the major differences between Safaitic and later forms of Arabic are archaisms. And if we look closely at the totality of the modern Arabic dialects as well as Classical Arabic beyond the 8 A term used to refer to Classical Arabic and the modern dialects by Huehnergard and Pat-El (2019: 11). While the ʾam article is attested mainly in Yemen today, frozen forms in the modern dialects suggest that it was much more widespread in former times (Levantine imbēreḥ ‘yesterday’ < *am+bāriḥ). I have further identified this article form in an unpublished Arabic-script inscription (5 th-6th c. CE) from the Tabūk area (northern Ḥigāz) in the dating formula ʾm-ʿm/ʾam-ʿām/ ‘the year’. It is customary to refer to this article form as influence from ‘South Arabian’ or ‘Himyaritic’, but no such prefixed article is known in the languages of Ancient South Arabia – definiteness there is marked by a final -n /ān/. 9 92 grammatical manuals, we find that the archaic forms of Safaitic are indeed present in “Arabic proper”.10 Final diphthongs and triphthongs: Safaitic realizes the reflex of the alif maqṣūrah as /ay/ and /aya/ while Classical Arabic and most modern dialects realize it as /ā/. However several Qur’anic reading traditions (most prominently that of the Kufans Ḥamzah and al-Kisāʾī) pronounce the alif maqṣūrah as /ē/ < *aya. And this realization is found in some Middle Arabic documents as well as in several modern dialects of South Arabia (e.g. in Rāziḥīt). Medial triphthongs in verbs: Some varieties written in Safaitic keep the glide of medialweak verbs intact in the 3 rd person forms: myt‘ he died’, kyn ‘he was’, etc. While all modern dialects and normative Classical Arabic collapse this triphthong to /ā/, some Qur’anic reading traditions(among the canonical readers this is most regularly applied by Ḥamzah) have an /ē/ vowel here as well: ḫēfa, mēta, etc. This reflects a similar collapse of the triphthong as in final position, and in any case, cannot reflect the conditioned raising of an original /ā/ vowel to /ē/. The preservation of archaic features like this in Safaitic is certainly not surprising. These texts date at least half a millennium before the Arabic grammatical tradition and its sources, and many inscriptions were carved more than twenty centuries before the modern dialects. In some cases, Safaitic seems to share some innovations in common with Classical Arabic, like the smarker for the explicit future tense, which no modern dialect, as far as I know, exhibits: Safaitic: s1-yʿwr ‘he will efface’ Classical Arabic: sa-yafʿalu ‘he will do’ In fact, the first inscription popularly regarded as ‘Arabic proper’, the Namārah inscription (328 CE)11, differs from normative classical Arabic and the Qur’an in many ways but agrees with the features found in the Safaitic inscriptions. There is clearly no linguistic basis to begin Arabic’s written history with this text. 10 The following two points are based on the excellent article of Van Putten (2017) on the fate of triphthongs in Arabic, and why they must be reconstructed for Proto-Arabic. 11 It is common to begin Arabic’s written history in the 4 th c. CE, usually based on the Namārah inscription. Note, however, that this is not the earliest Arabic-language inscription in the Nabataean script – The ʿEnʿAvdat inscription dates to 150 CE, at the latest, and JSNab 17 is dated to 267 CE. 93 Namārah Safaitic Normative Classical Qurʾan Arabic ‘this’ fs. ty t hāḏihī hdh Relative pronoun dw ḏ ʾallaḏī ʾldy -h [-ah] -h [-ah] -hā -hʾ 3fs suffixed pronoun In a recent PCA (Principal Component Analysis) study of lexical distance in the core vocabulary in the Semitic languages by H. Sidky,12Safaitic fits well within the range of Arabic dialects in terms of distance from Classical Arabic. There is clearly no basis in the lexicon to posit a general split between Safaitic and the rest of Arabic. Figure 3: Lexical distance of Arabic varieties from Classical Arabic (courtesy of HythemSidky) 12 This study is still in preparation by L. Souag, H. Sidky, S. Čeplö, and M. Van Putten. 94 The linguistic facts, I think, support an Arabic sub-grouping of Central Semitic defined by the isoglosses identified by Huehnergard (2017) and Al-Jallad (2018). This sub-grouping includes Safaitic, not as an independent branch but simply another, perhaps archaic, variety of Arabic on a continuum with Hismaic and Nabataean Arabic. Based on the epigraphic record, the only clear, innovative group in Old Arabic is what I have termed “Old Ḥigāzī”, characterized by the relative pronoun ʾallaḏī, ʾallatī, distal demonstratives with l ḏālika, tilka, etc., and the ʾan yafʿala construction replacing the infinitive. The placement of the modern dialects is unclear at this point, and perhaps their precise genetic classification is impossible considering the centuries of contact and leveling. What is clear, though, is that they cannot all derive from Old Ḥigāzī. For example, reflexes of the basic relative pronoun *ḏV- are attested across the Arabic-speaking world, from Yemen to the Maghreb. Dadanitic is certainly a Central Semitic variety closely related to, yet distinct from, Arabic, as it did not participate in important Arabic innovations. Taymanitic (not on tree) is very difficult to classify. If the sound change w > y /#_ is considered, then it could be classified as a Northwest Semitic language. But without any clear attestations of the plurals, this classification remains very tentative. As for the Thamudics, we simply cannot classify them – they should remain under West Semitic until positive evidence is discovered to support a particular subgrouping. I reproduce here the classification of the Arabian languages from my 2018 article (Fig. 1, pg. 35). Thamudic B and Thamudic F do not contain enough linguistic material for classification. 95 Figure 4: Classification of the languages of pre-Islamic North and Central Arabia (reproduced from Al-Jallad 2018) 96 Appendix: Unpublished Thamudic texts Inscription 1: Thamudic B text with the ʾal article. Photo: https://twitter.com/falah1alnze/status/1115645400084549632 nmbʾtr hnhyʾlhʾl-kbr s1mʿ l{y/n} ‘By Bʾtr’ ‘O Nuhay, god of greatness, give ear to me/us’ Inscription 2: Thamudic C with possible sunburst s3 Photo: https://twitter.com/mohammed93athar/status/1086335902933229568 ḥywʾ wdd-s3 ‘Ḥywʾ his beloved’ or ‘may his beloved have long life’ Inscription 3: Thamudic F text from Abhā region Photo: https://twitter.com/mmonseuor/status/1099692922868514816?s=20 gʿnt wbl Gʿntwbl w bʿd ḥbm ḏt and this (drawing) is for (his/a) love(r)’ The text is near a drawing of a lion and camel 97 Bibliography Al-Jallad, Ahmad. 2018. “What Is Ancient North Arabian?” In Re-Engaging Comparative Semitic and Arabic Studies, edited by Na ’ama Pat-El and Daniel Birnstiel, 1–43. Wiesbaden: Harrassowitz. Beeston, Alfred F. L. 1981. “Languages of Pre-Islamic Arabia.” Arabica 28 (Fasc. 2/3, Numéro Spécial Double: Études de Linguistique Arabe): 178–86. Huehnergard, John. 1995. “Features of Central Semitic.” In Biblical and Oriental Essays in Memory of William L. Moran, edited by A. Gianto, 155–203. Rome: Pontificio Istituto Biblico. ———. 2017. “Arabic in Its Semitic Context.” In Arabic in Context: Celebrating 400 Years of Arabic at Leiden, edited by Ahmad Al-Jallad, 3–34. Studies in Semitic Languages and Linguistics 89. Leiden: Brill. Huehnergard, John, and Na ’ama Pat-El. 2019. “Introduction to the Semitic Languages and Their History.” In The Semitic Languages: Second Edition, 1–22. Routledge Language Families Series. New York: Routledge. Huehnergard, John, and Aaron D. Rubin. 2011. “Phyla and Waves: Models of Classification of the Semitic Languages.” In The Semitic Languages: An International Handbook, edited by Stephan Weninger, Geoffrey Khan, Michael Streck, and Janet Watson, 259–78. Handbücher Zur Sprach- Und Kommunikationswissenschaft 36. Boston-Berlin: De Gruyter Mouton. ———. 2017. "Arabic in its Semitic Context" In Arabic in Context: Celebrating 400 years of Arabic at Leiden, edited by Ahmad Al-Jallad, 3–34. Leiden: Brill. King, Geraldine M. H. 1990. Early North Arabian Hismaic. PhD Dissertation. London: University of London, School of Oriental and African Studies. Kootstra, Fokelien. 2016. “The Language of the Taymanitic Inscriptions and Its Classification.” Arabian Epigraphic Notes 2: 67–140. Macdonald, M.C.A. 2000. “Reflections on the Linguistic Map of Pre-Islamic Arabia.” Arabian Archaeology and Epigraphy 11: 28–79. ———. 2004. “Ancient North Arabian.” In The Cambridge Encyclopedia of the World’s Ancient Languages, edited by R.D. Woodard, 488–533. Cambridge: Cambridge University Press. 98 Macdonald, Michael C. A. 2000. “Thamudic.” The Encyclopaedia of Islamic New Edition. Leiden: Brill. Prioletta, Alessia. 2018. “New Research on the Thamudic Graffiti from the Region of Ḥimā (Najrān, Saudi Arabia).” In Languages, Scripts and Their Uses in Ancient North Arabia, edited by M.C.A. Macdonald, 53–70. Supplement to the Proceedings of the Seminar for Arabian Studies 48. Oxford: Archaeopress. Van Putten, Marijn. 2017. “The Development of the Triphthongs in Quranic and Classical Arabic.” Arabian Epigraphic Notes 3: 47–74. Winnett, Frederick Victor, and William LaForest Reed. 1970. Ancient Records from North Arabia. Near and Middle East 6. Toronto: University of Toronto Press. Zwettler, Michael J., and David F. Graf. 2004. “The North Arabian ‘Thamudic E’ Inscription from Uraynibah West.” Bulletin of the American Schools of Oriental Research 335: 53– 89. 99