Re-Engaging Comparative Semitic and Arabic Studies 3447111186, 9783447111188

Citation preview

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

ABHANDLUNGEN FÜR DIE KUNDE DES MORGENLANDES Im Auftrag der Deutschen Morgenländischen Gesellschaft herausgegeben von Florian C. Reiter Band 115 Board of Advisers: Christian Bauer (Berlin) Desmond Durkin-Meisterernst (Berlin) Lutz Edzard (Erlangen/Oslo) Jürgen Hanneder (Marburg) Herrmann Jungraithmayr (Marburg) Karénina Kollmar-Paulenz (Bern) Jens Peter Laut (Göttingen) Joachim Friedrich Quack (Heidelberg) Florian C. Reiter (Berlin) Michael Streck (Leipzig)

2018

Harrassowitz Verlag . Wiesbaden

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Re-engaging Comparative Semitic and Arabic Studies Edited by Daniel Birnstiel and Na‘ama Pat-El

2018

Harrassowitz Verlag . Wiesbaden

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the internet at http://dnb.dnb.de.

For further information about our publishing program consult our website http://www.harrassowitz-verlag.de © Deutsche Morgenländische Gesellschaft 2018 This work, including all of its parts, is protected by copyright. Any use beyond the limits of copyright law without the permission of the publisher is forbidden and subject to penalty. This applies particularly to reproductions, translations, microfilms and storage and processing in electronic systems. Printed on permanent/durable paper. Printing and binding: Hubert & Co., Göttingen Printed in Germany ISSN 0567-4980 ISBN 978-3-447-11118-8 e-ISBN 978-3-447-19823-3

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Contents Editors’ preface ........................................................................................ VII Ahmad Al-Jallad What is Ancient North Arabian? ................................................................

1

Daniel Birnstiel Neither Clear Nor Clarifying – Yet Clearly Arabic .................................... 45 Lutz Edzard The Marked Nominative in Arabic, Semitic, and Afroasiatic .................... 105 Phillip W. Stokes The Plural Demonstratives and Relatives Based on *ʾVl in Arabic and the Origin of Dialectal illī .......................................................................... 127 Janet C.E. Watson and Abdullah Musallam al-Mahri A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar ............. 151

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Preface Already about fifty years ago, the late Edward Ullendorff lamented the fact that the interests of people working in the field of Semitic languages have often become diverted from comparative and historical linguistics to other spheres of interest. He noted that “Arabists with a concern for other Semitic languages gave way to islamisants who were bound to step outside the area of Semitics and seek contacts with other Islamic languages, such as Persian or Turkish, which had very disparate linguistic affiliations.”1 Even as Arabic linguistics, especially the growing interest in Arabic dialectology and internal subgrouping, grew to become a field in its own right, Arabists by and large avoided engaging with Semitic philology and linguistics. A sub-field that was once at the core of Semitic studies, with scholars such as Theodor Nöldeke, Carl Brockelmann, Edward Ullendorff, Joshua Blau and many others, veered away from engaging with the non-contemporary languages of the region, potentially loosing linguistic and cultural context. This has resulted in a situation, in which Semitic languages are studied in many academic institutions almost exclusively within the framework of individual fields, for example Arabic in institutions for Arabic and Islamic studies, Hebrew and Aramaic in divinity schools and institutes for Jewish studies, or Syriac and Ethiopic in departments studying Eastern Christianity. Even in cases, where several Semitic languages are studied in near vicinity to each other due to the co-existence of respective infrastructures, cross-departmental cooperation of a linguistic nature is rare. At the same time, institutes dedicated to comparative Semitic philology have become increasingly rare, and as a result fewer scholars are trained in the discipline. Recognizing the importance and need of rooting philological research on Arabic Islamic texts within a solid linguistic framework, the organizers of the congress “Horizons of Islamic Theology” which took place in September 1 – 5, 2014, at the Institute for the Study of Islamic Culture and Religion at the Goethe-University in Frankfurt, wished to take a stand against this trend and included a section on Comparative Semitic and Arabic Studies. This section consisted of three panels, “Arabic in Context”, “Arabic and Semitic: How Archaic Is the fuṣḥā?” and “New Insights in Arabic Syntax”.

1 Ullendorff, Edward. “Comparative Semitics”. In Linguistics in South West Asia and North Africa. (Current trends in linguistics 6). Edited by Thomas A. Sebeok. The Hague: Mouton, 1970, 263.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

VIII

Preface

This section attempted to highlight the benefits of a linguistic and comparative engagement of Arabic with other regional languages and societies for an improved understanding of Early Islam and its literary heritage. The articles collected in this volume largely emerged from the papers presented at these three panels. They are ordered alphabetically and cover an area reaching from general and comparative Afro-Asiatic linguists to South Arabian phonology to diachronic and synchronic Arabic linguistics. Despite the diversity of their topics, the contributions all illuminate various ways in which a re-engagement with comparative Semitic and Arabic linguistics can benefit research on Arabic Islamic texts. Ahmad Al-Jallad explores what linguistic features are diagnostic of Ancient North-Arabian. He discusses previous definitions of Arabic and the classification of languages commonly bundled as Ancient North-Arabian with regard to shared isoglosses. Al-Jallad contextualizes the various dialects both in relation to Arabic and to West Semitic and offers a novel understanding of Ancient Arabia and its linguistic borders. Daniel Birnstiel investigates the meaning of the Qur’anic term mubīn and additional related terms from a synchronic perspective. He shows how these terms have often been misinterpreted both by Muslim exegetes and modern scholars and suggests several emendations to the text of the Qur’an. Evidence from other Semitic languages assists in entangling this issue. The contribution by Lutz Edzard discusses various functions of the Arabic accusative from a comparative Semitic and Afro-Asiatic perspective. He demonstrates that many of the cases attested in Arabic grammar can be much better understood and made sense of when regarded as cases of a marked nominative within the larger Afro-Asiatic and Semitic framework. Phillip Stokes traces the history of the common Arabic dialectal relative marker illī and its variants. After rejecting previous explanations connecting illī and variants to the Classical Arabic relative allāðī on linguistic grounds, he adduces evidence suggesting a derivation from the definite article *al followed by a plural demonstrative *ʾulay. In the process, Stokes reconstructs the demonstrative in Arabic and contextualizes it within Semitic. Janet Watson and Abdullah al-Mahri take a close look at word stress in Mehri from the perspective of Stratal Optimality Theory. They show how the complexity of the rules determining Mehri word stress may convincingly be analyzed as the combination of lexical stress and the interaction of constraints functioning at different levels. Our thanks are due to the organizers of the congress “Horizons of Islamic Theology” and the Institute for the Study of Islamic Culture and Religion, especially Ömer Özsoy and Udo Simon, to Lutz Edzard, to the editor-in-chief

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

IX

Preface

of Harrassowitz Verlag, Barbara Krauß, and to the editor of the series Abhandlungen für die Kunde des Morgenlandes, Florian C. Reiter, who made the publication of this volume possible. Finally, the editors wish to thank all the young dragons in their life, real and imaginary. ’

Frankfurt and Austin, October 2018

Daniel Birnstiel and Na ama Pat-El

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?1 Ahmad Al-Jallad, The Ohio State University, Columbus

1 Introduction The biggest barrier to understanding the linguistic diversity of pre-Islamic Arabia has traditionally been the dearth of sources. Scholars had only the fragmentary accounts of medieval Muslim scholars to rely upon. These presented the Arabian Peninsula as more or less linguistically homogenous – Arabic was spoken across its vast territory, with the exception of the southwestern corner where Ḥimyaritic still held sway. And even the few surviving remarks about Ḥimyaritic made it seem like an exotic Arabic dialect rather than an independent language.2 While the Arabic grammarians recorded a significant degree of dialectal variation in the Arabic of their time, their accounts gave little reason to suspect that Arabic was not the language of Arabia’s tribespeople since time immemorial. The exploration of the Arabian Peninsula and adjacent parts of the southern Levant beginning in the 19th century remedied the problem of sources. Tens of thousands of inscriptions in varieties of the South Semitic script, spanning the entire Peninsula, provided a new vista from which to study Arabia’s once pre-history. Medieval references to Ḥimyaritic had accustomed scholars to viewing ancient Yemen as linguistically distinct from the rest of the Peninsula. The epigraphic evidenced corroborated this, but rather than attesting a single, ancient Ḥimyaritic language, Ancient South Arabia was home to at least four distinct languages, spanning from the early first millennium BCE until the 6th century CE.3 None of these languages could be considered Arabic in any sense, 1 Editorial signs for inscriptions are: [restored letter]; {damaged letter}; / word divider; // line break; the glottal stop (hamza) is represented as ʔ in phonological and phonetic transcription and as ʾ when transcribing the glyphs of an inscriptions; the same goes for the pharyngeal fricative ʕayn, ʕ and ʿ, respectively. 2 For an excellent discussion on the relationship between the medieval Ḥimyaritic and the epigraphic record of the pre-Islamic kingdom of Ḥimyar, see Stein (2008). The classic reference for the features of Medieval Ḥimyaritic is Rabin (1951); for a more recent discussion, see Robin (2007). 3 On the chronology of Ancient South Arabian, see Stein (2013), Drewes et al. (2013).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

2

Ahmad Al-Jallad

and none matched the “Ḥimyariyyah” of medieval Islamic sources. The exact relationship between the Ancient South Arabian languages continues to be debated by specialists as does their position, as a whole or individually, within the Semitic family.4 The problems of classifying Ancient South Arabian, however, are not the subject of the present essay; it is the more fragmentary inscriptions and shadowy languages of Central and North Arabian that concern us. Unlike Ancient South Arabia, where the epigraphic languages are preserved in monumental inscriptions, the inscriptions of Central and North Arabia are usually classified as graffiti; in fact, only those texts composed at the oasis of Dadān in the Higāz appear to have been commissioned by their nominal authors.5 The inscriptions of Central and North Arabia differ from South Arabia in another important respect: while a single script was used across Ancient South Arabia to write various languages, Central and North Arabia attest a stunning variety of alphabets, the exact number of which remains unknown. The informal and laconic nature of these texts has posed considerable problems for the classification of their scripts and has made linguistic diagnosis sometimes impossible. The first successful attempt at classification was that of Winnett (1937), which divided the scripts of Central and North Arabia into categories of “Thamudic”6 A through E, to which were added “Liḥyanite” (= Dadanitic) and Safaitic, forming seven script groupings. Winnett, however, did not embark on a linguistic classification and it is clear that he regarded the language behind all of these alphabets as more or less homogenous. In a seminal paper, M.C.A. Macdonald (2000) established the conventional terminology for the scripts and languages of the Arabian Peninsula before Islam used today. For North and Central Arabia, Macdonald distinguished the following categories of the South Semitic script, based on whether they were employed at an oasis or by nomads (see Map 1 on page 36), and two linguistic groupings, Ancient North Arabian and Arabic.7

4 For the classic discussion, see Avanzini (2009). For balanced discussions on the linguistic classification of Ancient South Arabian within Semitic, see Huehnergard 1995; Huehnergard and Rubin 2011; Rubin 200. 5 On this phenomenon, see Macdonald (2010; 2015). 6 The term “Thamudic” is entirely conventional; there is no demonstrable link between the historical tribe of Thamūd and all of the inscriptions placed under this label, although a few texts can be connected to members of this tribe. 7 The South Semitic script is a conventional term for the family of alphabets used in Arabia before the rise of Islam, ranging from 27 to 29 glyphs. See also Macdonald (2010) for a good introduction to their history and Sass (1991) on their possible development.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

Ancient North Arabian

Desert North Arabian

Oasis North Arabian

Taymanitic (formerly Thamudic A)

Oasis of Taymāʾ, North Arabia

Dadanitic (formerly Oasis of Dadān, northern Ḥigāz Dadanite and Lihyanite) Dumaitic

Oasis of Dūmah, North Arabia

Dispersed North Arabian Texts from Mesopotamia and other places not classifiable as one of the above scripts Safaitic

Syro-Jordanian basalt desert, northern Saudi Arabia, isolated finds elsewhere

Hismaic (formerly Thamudic E)

Southern Jordan, northern Saudi Arabia

Thamudic B

North and Central Arabia

Thamudic C

North and Central Arabia

Thamudic D

North and Northwest Arabia

Southern Thamudic

Southwest Arabia

Old Arabic

Arabic Pre-Islamic ʾl-dialects Attested in various places, but concentratattested in various scripts ed in North Arabia and southern Levant

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

3

4

Ahmad Al-Jallad

This classification was based primarily on script type8 and a single linguistic isogloss, the phonetic realization of the definite article. Since 2000, our understanding of the epigraphy of Arabia has increased dramatically and the discovery of tens of thousands of new inscriptions has helped fill in the gaps in the grammar and lexicon of these languages. We are therefore now in a better position to re-assess the linguistic landscape of pre-Islamic North and Central Arabia, especially in terms of genetic classification. The main goal of this essay is to determine if the scripts classified by Macdonald (2000) as Ancient North Arabian also form a linguistic unity, and whether they, in part or whole, descend from Proto-Arabic or are distinct from it.

2 What is Arabic? Since Beeston (1981), scholars have generally divided the ancient languages of Central and North Arabia into two categories: Arabic and Ancient North Arabian. Arabic was largely defined by the language of the Qurʾān and the norms of Classical Arabic while the linguistic category Ancient North Arabian hinged on a single isogloss – the definite article h-. Macdonald (2000) softened the distinction between the two groupings, referring instead to Arabic and Ancient North Arabian as two separate, but perhaps mutually intelligible, “dialect bundles”.9 In 2009, Macdonald added two more features to distinguish Arabic and Ancient North Arabian dialects:10 1) The feminine singular relative pronoun is ʾlt (= ʾallatī); 2) The consonantal realization of root medial and final w and y: Safaitic rʿy vs. Classical Arabic raʕā; Safaitic s1yr vs. Classical Arabic sāra.

8 Note, however, that the inter-relationships between the Ancient North Arabian scripts remain poorly understood and it has yet to be proven that they constitute a single unit against Ancient South Arabian. At present, Ancient North Arabian is a negative definition, meaning simply the non-Ancient South Arabian South Semitic scripts. 9 This point is made even clearer in Macdonald (2009: 318, n.198) where he argues against Gawlikowski (2006: 46), stating that ʾl and h(n)- dialects were in no sense separate languages. This point is rarely recognized in the secondary literature, and Ancient North Arabian is often treated as a single entity, entirely distinct from Arabic. 10 Macdonald (2009: 312–313) states clearly that these criteria do not form a solid basis for the identification of the language of any inscription as Arabic, but at the time, the interpretation of the texts permitted little more.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

5

From this definition, it is clear that the Arabic to which Macdonald refers is in fact “Classical Arabic”, as the features listed are by no means characteristic of all forms of Arabic or even reconstructible to Proto-Arabic. Two of these features, the relative pronoun ʔallatī and the definite article ʔal, would moreover exclude many modern dialects of Arabic from the definition of Arabic. E. A. Knauf (2010: 207) advanced an opposite opinion: ““Ancient North Arabian” genetically is Arabic… Diachronically and genetically, we are talking about two stages in the development of a single language. “Ancient North Arabian” is “Proto-Old Arabic”, just as (vulgar) Latin is proto-French and French (extremely corrupted) is Late Latin.” His argument for this classification is based on three features: 1) The merger of s1 (Proto-Semitic [s], Hebrew šin) and s3 (Proto-Semitic [ts], Hebrew sāmek); 2) A pre-positive, rather than post-positive, definite article; 3) Broken plurals. The fact that both classifications have summoned overlapping linguistic features to support opposite conclusions underscores an issue that has been intuited rather than scientifically defined in the epigraphic literature – the linguistic definition of Arabic.11 Linguistic classification is dependent upon the identification of shared morphological innovations, that is, developments in grammar that are shared by a group of languages against other members of the larger family to which they all belong. These features alone can imply descent from an exclusive shared ancestor, in our case, a Proto-Arabic.12 The modern dialects of Arabic, Classical Arabic, and the language of the Qur’anic Consonantal Text (QCT)13 share a large number of features to justify assuming that they descend from a common ancestor to the exclusion of other Semitic languages. Huehnergard (2017) established a list of innovations prov-

11 This question has recently been asked by Retsö in his learned contribution to the Oxford Handbook of Arabic Linguistics (2013). For a critique of his approach, see Huehnergard (2017, n.19, n.83). 12 For an excellent introduction to this methodology as applied to Arabic, see Huehnergard 2017: 1–14). 13 For a definition of the QCT as separate from Classical Arabic, see Van Putten (2017: 47–48).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

6

Ahmad Al-Jallad

ing this genetic relationship, a list that I have expanded in a recent publication (2018a). Turning back to Knauf’s list of features, I find his argument that Ancient North Arabian, as a whole, is the ancestor of Old Arabic difficult to justify on methodological grounds. The presence of broken plurals cannot be used for classification, as it is most likely a Proto-Semitic retention rather than a shared innovation. Moreover, this pluralization strategy is found in Arabic, Ethio-Semitic, Modern South Arabian, and Ancient South Arabian.14 While Knauf regards the phonetic realization of the definite article as trivial, he argues its pre-positive position is an indicator of genetic descent. Huehnergard (2017), whom I follow, does not admit the definite article into the list of innovations characterizing Arabic; the article is certainly an innovation from the Proto-Semitic perspective (Huehnergard and Rubin 2011: 269–70), and is likely a contact feature that spread throughout Central Semitic sometime in the late 2nd millennium BCE. As such, it is a most unsuitable feature for linguistic diagnosis.15 The single Proto-Arabic innovation on this list is the merger of s1 and s3 (Huehnergard 2017: 19), yet as a sound change, it is a very weak feature and certainly cannot be used to argue for the claim that all of the epigraphy of Ancient North Arabia is Proto-Arabic. As sound changes go, s1>s3 could easily be an areal feature of Central and Northern Arabia, just as p > f seems to have been as well.16 Moreover, as we shall see, this sound change does not encompass all of the epigraphic groups of Ancient North Arabia, as Taymanitic seems to have undergone a different development (Kootstra 2016: 74–79, and below).

14 While the system of broken plurals appears to be Proto-Semitic, Ratcliffe (1998) has argued that the system was expanded in the Semitic languages of Arabia and Ethiopia, and that this expansion was evidence for a South Semitic sub-grouping. Huehnergard and Rubin (2011: 272–73) make a convincing case that the broken plural system was in fact a retention from Proto-Semitic, while admitting possibility that the “expansion” of the system was an areal feature of Arabia. I personally do not find the “expansion” scenario convincing; even if one accepts it, it does not allow for the use of broken plurals for classification unless one isolates innovative patterns unique to Arabic and locates them in the epigraphic record. The great ambiguities in the scripts of Ancient Arabia, however, make such a possibly very unlikely, as one would not be able to prove that supposedly Arabic patterns are absent in the consonantal skeletons of the epigraphic languages. 15 The exception to this, I would argue, is the suffixed article which is integrated into the morphology of the language, with allomorphs based on number: Syriac malkā ‘the king’vs. malkayyā ‘the kings’; Sabaic ṣlmn ‘the statue’ vs. ṣlmnhn ‘the statues’. 16 Huehnergard and Rubin (2011: 272) explain this as an areal feature but note that its existence in the earliest stages of Arabic is purely hypothetical and, in fact, there is good evidence to think that the reflex of this phoneme was still [ph] (Al-Jallad 2015a: 41–42; Al-Jallad 2017: 125–26).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

7

Knauf’s essay takes for granted that Ancient North Arabian is a linguistic unit. It is discussed as a whole as the linguistic antecedent to “Old Arabic” and “Classical Arabic”. The approach of this essay will be different – we will begin with the linguistic isoglosses outlined by Huehnergard 2017 and modified by Al-Jallad (2018a) and ask: can the various corpora of Ancient North Arabian be classified linguistically and if so, do they belong to the same genetic category as Classical Arabic and the modern Arabic dialects? The innovations characterizing Arabic are (Huehnergard 2017):17 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11)

the merger of s1and s3; the loss of the 1st person singular pronoun ʔanāku;18 the replacement of mimation with nunation;19 the levelling of the -at allomorph of the feminine ending to nouns terminating in -t; relics survive in words like bint-‘daughter’ and ʔuḫt- ‘sister’; the levelling of the -na ending of the 3rd feminine plural prefix conjugation to the suffix conjugation, producing qatalna from earlier *qatalā;20 the mafʕūl pattern as a paradigmatic passive participle of the G-stem; the vowel melody u-i for the passive; the preposition fī ‘in’, grammaticalized from the word ‘mouth’; the replacement of the anaphoric demonstrative with demonstratives derived fro the proximal base; the presence of nunation on nominal heads of indefinite asyndetic relative clauses; feminine demonstratives with a t-onset.

17 I have excluded Huehnergard’s proposal that pharyngealization is Proto-Arabic as the ancient evidence is ambiguous and several modern dialects have glottalized reflexes of the emphatics. I also excluded the suggestion that the loss of the paradigmatic infinitive qatāl is a Proto-Arabic phenomenon, as the complex maṣdar system of Arabic seems to be original on the principle of archaic heterogeneity. Huehnergard suggests that qad faʿala is an Arabic innovation as well; I agree with this but I see it as part of a larger trend in pre-verbal aspect/tense marking. 18 Huehnergard mentions this as a minor feature, also shared with Aramaic. 19 The replacement of mimation with nunation may seem like a minor change at first, but it requires more than expansion or levelling. Mimation on singular nouns and broken plurals was a simple m following the case vowel, while nunation on duals and plurals was followed by a vowel, either na or ni. If the system was simply expanded to singulars, we would expect a replacement of m with na or ni. Rather, it is the consonant that was replaced, suggesting a more abstract change. 20 Huehnergard notes that this is a minor feature.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

8

Ahmad Al-Jallad

To these innovations, I would add: 12) the complex and asymmetrical system of negation, mā, lā, lam, and lan, with certain moods and conjugations of the verb; 13) pre-verbal tense and aspect marking, qad faʕala, sawfa yafʕalu, etc.;21 14) the use of ʔan(na) as a complementizer; 15) the independent object pronoun base *(ʔiy)yā; 16) the use of the a-marked prefix conjugation (yafʕala) as a subjunctive.22 The following section will examine the features traditionally identified as diagnostic of Arabic in the literature, beginning with Macdonald (2009), and then treating Mascitelli (2006) and Robin (2001). Before moving on, I wish to be clear about the term Proto-Arabic, which will be used with some frequency in this paper. Proto-Arabic has sometimes been used to refer to the Ancient North Arabian inscriptions (Rabin 1951); this is misleading. A proto-language is the ultimate ancestor of all later forms of that language, and so Proto-Arabic would be the ultimate ancestor of not only Classical Arabic but other ancient dialects and the modern spoken forms as well (Huehnergard 2017).23 The term Proto-Arabic only implies that its constituent members form a linguistic unity, and does not make any claims about what the speakers of these languages may have called their spoken language or how they may have self-identified. 2.1 Traditional isoglosses of Arabic reconsidered The relative pronoun ʾlt This category can be expanded to include the entire relative pronoun series consisting of an ʔalla + demonstrative element. Huehnergard (2017: 22–23) did not include this feature as an isogloss of Arabic, as several dialects of Arabic retain the older form comprising of simply ḏV. In the Classical Arabic sources, this is found in the Yemeni dialects and in the dialect of Ṭayyiʔ, and it is widespread in the modern dialects.24 I have therefore argued that the ʔalla-forms 21 While Modern Aramaic also employs pre-verbal TAM markers, I wonder if this is not due to contact with Arabic. 22 This was suggested by Huehnergard (2017: 14) but not placed on his primary list of innovations. 23 It is now well established that Classical Arabic, a literary language of Islamic civilization, is not the direct ancestor of any modern spoken forms of Arabic. 24 The d-relative pronoun is attested in the Maghreb and in Yemen (Behnstedt 2016: 74);

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

9

are an innovation of only a sub-section of the Arabic dialects (Al-Jallad 2015a: 13–14; forthcoming). If we take their first occurrence in the Dadanitic inscription JSLih 384 and in the QCT as indicative of their geographic origin, then it would seem that this feature emerged in the dialects of the Ḥigāz.25 As such, the ʔalla-relative pronouns are a feature for the sub-classification of Arabic, but were certainly not a Proto-Arabic feature. The reflex of weak roots The reflex of III-y/w verbs: To this class belong verbs that terminate in a long /ā/ in Classical Arabic and many, but not all, modern dialects, e.g. banā ‘he built’, daʕā ‘he invoked’, etc. The collapse of this sequence was therefore considered an Arabic isogloss while their preservation was regarded as characteristic of Ancient North Arabian. Before investigating whether this distribution holds true, it must be first emphasized that the original forms of such verbs contained a triphthong and not a long vowel. The second issue is that this feature is not suitable for classification as it constitutes a simple and common sound change, the monophthongization of triphthongs, that took place independently in several different branches of Semitic (Huehnergard and Rubin 2011: 268–269; Al-Jallad 2014). While Classical Arabic and most of the modern dialects of Arabic reduced the triphthongs of this verb class to a final /ā/, the situation in the QCT is more complicated. According to the recent and exhaustive study of Van Putten (2017), III-w and III-y roots had different reflexes, /ē/ and /ā/, respectively – this much is indicated in the orthography, ‫ بنى‬vs.‫دعا‬. The maintenance of this distinction implies that in the direct ancestor of the language of the QCT, the triphthongs /aya/ and /awa/ were kept apart, and collapsed independently.26 This stage is in fact witnessed in the Graeco-Arabic inscription A1 (Al-Jallad and al-Manaser 2015), which attests both the al-article and the verb αθαοα [ʔatawa] ‘he came’, contra Classical Arabic ʔatā. These facts combined show that the collapse of the triphthongs to /ā/ cannot be attributed to the common on the ḏū of Ṭayyiʾ, see Rabin (1951: 204). Of course, the original form of the pronoun is preserved in the Classical Arabic relative-determinative, ḏū ‘the one of’. 25 Today, the form ʔallaḏī is found primarily in Yemen (Behnstedt 2016: 74), but its ubiquity in Middle Arabic texts may suggest that it was more widespread in former times. The form ʔilli in the modern dialects has been argued to be a reduced form of ʔallaḏī, although there are other explanations. See Stokes’ contribution to this volume. 26 Note that this situation must have obtained in the early Islamic period as one the Muʿāwiyah dam inscriptions attests the form bny-h, spelled unambiguously with dots under the y, for ‘he built it’, pointing towards [banaya-hu] or [banayoh], or something of the like; see Al-Jallad (2014, n. 33) and Miles (1948) for the inscription.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

10

Ahmad Al-Jallad

ancestor of all varieties of Arabic (Proto-Arabic), and therefore cannot be considered characteristic of Arabic, but of only some varieties of the language. The maintenance of the triphthong in a North Arabian inscription does not disqualify its status as “Arabic” nor can it act as a classificatory feature of a separate language family. The reflex of II-w/y verbs: In Safaitic, medial-weak verbs are written with a glide, myt [mayeta] ‘he died’, ṭwf [ṭawopha] ‘he returned’. Safaitic orthography does not employ matres lectionis, so the medial glide must represent a true consonant.27 Since the hollow paradigm seems to have been tri-radical at the Proto-Semitic stage,28 these spellings certainly reflect the archaic situation, directly antecedent to the apophonic situation encountered in Classical Arabic and the modern dialects.29 As an archaism, they are not appropriate for classification. Moreover, it is unclear whether such forms are common to all Ancient North Arabian corpora, as we shall see below. The definite article The most widely cited difference between Arabic and Ancient North Arabian is the definite article ʔal vs. ha(n), respectively. And while the h-article is certainly not unique to Ancient North Arabian, as it is common to Canaanite as well, the ʔal-article is a feature claimed only to be found in Arabic. Yet, without considering other isoglosses, this reasoning is entirely circular. First, it should be clear that the basic division of ʔal and ha(n) into two neat linguistic groups does not reflect the actual attested situation. In Safaitic, five article forms have so far been identified (Al-Jallad 2015a: 74–79): -----

ʾl ʾ h hn

27 Knauf (1991: 94) has advanced some arguments that y and w could stand for long /ā/, but these fail to convince, especially in light of Greek transcriptions. 28 For good arguments for this, see Bauer 1912; Voigt 1988; Suchard 2016; Huehnergard, however, suggests that the sound changes responsible for the collapse of the triphthongs in medial-weak verbs occurred at the Proto-Semitic stage (1995: 177, n.75). 29 While all dialects and Classical Arabic exhibit the forms Cu/iCtV and CāCa, some ancient forms attributed to the Ḥigāz exhibit an ē vowel in medial position, e.g. ḫēfa, hēba, ṭēba (Rabin 1951: 111), etc., suggest a different resolution of the original triphthong rather than sporadic raising of long ā. Rabin’s explanation of these forms preserving a reflex of a fourth Proto-Semitic long vowel, ē, has not been accepted. On the phonology of Proto-Semitic, see Kogan 2011.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

11

-- Ø Now, while h-article is certainly the most common, the ʾ-article is not rare. The ʾl- and hn- articles, however, occur far less frequently, the latter only a handful of times. One must remember that the Safaitic inscriptions are highly formulaic and that the distribution of article forms in the inscriptions may not reflect the exact situation in the spoken dialects. M.C.A. Macdonald brought to my attention two inscriptions authored by the same individual. In these, the writer identified himself as ‘the bandit’ (= lṣṭ)30, but in one he used the h-article while in the other the ʾ-article.31 ShNGA 1:  l ḥmlt bn s¹lm ʾ-lṣṭ ḏ ʾl {ḍ}{f} ‘By Ḥmlt son of S¹lm, the bandit, of the lineage of Ḍf’ AMSI 50: l ḥmlt bn s¹lm h-lṣṭ ‘By Ḥmlt son of S¹lm, the bandit’ The most obvious explanation for this is that both article forms existed in the dialect of this writer and that they were perhaps in variation. Indeed, if we imagine a situation where several competing article forms were available, but the h-was simply part of the register of writing, we might then explain the disproportionately high usage of h-.32 In the Thamudic B inscriptions, the h-form is virtually the only form found, but these are far less numerous than Safaitic and in most cases the h can be construed as a demonstrative. There is one attestation of the h-article in the Thamudic C inscriptions (see the discussion of Thamudic C below), but in a divine name and so this may not reflect the language of the texts themselves. To my knowledge, the definite article has not yet been attested in Thamudic D. Thamudic F exhibits a variety of article forms, but these are restricted to the onomasticon. 30 Safaitic lṣṭ is most certainly a loanword from Greek, Aramaic lesṭā, Classical Arabic liṣṣun, ultimately from Greek ληστής. 31 While the ʾ and h can sometimes be close in shape in Safaitic, the ʾ distinguished by a single crossbar, these texts are written in the so-called square script, where both forms are rather distinct. This ʾ-article in this inscription could in theory be the ʾl-article, with the assimilation of the l to the following coronal. 32 While the Safaitic inscriptions certainly do not represent a formal writing tradition as such, with a regulated scribal language, their formulaic nature certainly allows for conventions of spelling and language to exist, even in this informal context. This much is suggested by the spelling of etymological *q as both q and ʾ in a single inscription (Al-Jallad 2015a: 53).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

12

Ahmad Al-Jallad

Of the Oasis North Arabian scripts, the h-article is found in Taymanitic while both h(n) and ʾ(l) are attested at Dadān, but the former is much more common.33 No article is attested in the three inscriptions carved in the Dumaitic script.34 In Hismaic, the h-element seems to be purely a demonstrative. The language appears to lack any morphological means of definition.35 This preserves the ancient Central,West, and Proto-Semitic situation, as both Ugaritic and Gəʕəz lack a definite article (Huehnergard and Rubin 2011: 269). The article-less situation is not restricted to Hismaic, as some Safaitic inscriptions lack the article as well. This is clearly the case in the inscription HshNSMI 5 (Al-Jallad 2015a: 251). The text contains two common expressions without the definite article, where in all other inscriptions the article is present, e.g. ḥl dr ‘he camped in this region’ compared to the normal expression ḥl(l) h-dr and even ḥll ʾl-dr (C 5137); and lm yḫbl sfr ‘may the inscription not be effaced’ compared to similar expressions such as ʿwr ḏ yḫbl h-sfr ‘blind him who would efface the writing’. In light of the foregone discussion, a better characterization of the Ancient North Arabian situation with regard to the article is this: the definite article still possesses a strong demonstrative force and several allomorphs are in use, but the h-article is the most common form. Archaic strands lacking a definite article are attested in the Hismaic and, rarely, in the Safaitic inscriptions. The h-article is found across Central and North Arabia, while theʾl- and ʾ-articles are concentrated in the north and northwest, in the Ḥigāz (Dadanitic), the southern Levant (Nabataean),36 and rarely in Safaitic (the Syro-Jordanian Ḥarrah), although scattered examples can be found in the south, such as at the oasis of Qaryat al-Fāw and in the Ḥimà region near Naǧrān.37 33 In Dadanitic, the form with an n regularly appears before laryngeals, and only in one case before qbr ‘grave’, but the two words are on separate lines (JSLih 81; Winnett and Reed 1970: 124). 34 These are WTI 21–23 (Winnett and Reed 1970: 80). 35 This is clearly seen in inscriptions associated with rock art. For example, AMJ 67 reads ʿm ḫṭṭ gml ‘ʿm carved (the) camel’, clearly associated with the camel on the same panel. In Safaitic and Thamudic B, the same construction occurs almost always with the article, while in Hismaic, the h is rare and King, correctly in my mind, interprets it as a demonstrative (King 1990, §C.6). The clearest example is found in a bilingual Hismaic-Nabataean inscription published by Hayajneh (2009), in which the author of the text gives his name as ʿbdʾlʾyb in Nabataean, with the ʾl-article, and ʿbdʾyb, in Hismaic, clearly showing that h was not the equivalent of Nabataean Arabic ʾl. 36 For example, the ʿĒn ʿAvdat inscription, JSNab 17 (Macdonald in Fiema et al. 2015), and personal names such as ʿbdʾlbʿly (Negev 1991, s.v.). 37 These examples will be discussed in more detail under §6.4 and §7.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

13

2.2 The relationship between the ʔal- and ha-articles Two main positions exist regarding the relationship between the ʔal-and ha-articles. The first sees them as deriving from two different morphemes. Rubin (2005: 75–76) argues that the ʔal-article is in fact the reflex of the plural demonstrative, which he reconstructs as *ʔulli. This requires several ad-hoc changes, including the reduction of the geminate l, the loss of the first syllable and the introduction of a “prothetic” ʔa-. Ad-hoc sound changes are possible in a situation of grammaticalization, but the number required here certainly prevents a straight-forward reconstruction. Moreover, the demonstrative element *ʔulli, or *ʔul, does not belong to the same paradigm as the han particle (Hasselbach 2007: 20). It would indeed be strange if the plural demonstrative grammaticalized as an article but not the singular members of the same paradigm, i.e., demonstratives containing the element ḏ-. Since no Central Semitic language exhibits inflection for number or gender in the definite article (Pat-El 2009: 23–24), this strongly suggests that inflection was never a feature of this prefix, in turn undermining a connection with the demonstrative ḏ-/ʔul-series. Finally, the existence of the hl-article in Thamudic F (see below) precludes a derivation from the base *ʔul, as the sound change ʔ > h is unknown in Arabic, while the reverse is common. The second position derives ʔal from han (Tropper 2001), through perhaps an irregular, but not inexplicable, sound change of n >l. The main challenge to this is that presentatives with an l are attested in other Semitic languages, Akkadian allû; Ugaritic hl, the interrogative particle hal in Classical Arabic, and indeed the Thamudic F definite article hl (see also the discussion in Pat-El 2009, especially 40–42). This fact may suggest that both forms *han and *hal existed at an earlier stage; the l element is after all an established deictic suffix (Hasselbach 2007: 22). Since Safaitic shows that multiple article forms co-existed and were used in the writing of a single author, it is possible that this situation extends into the distant past, and that *hal and *han were used interchangeably as definite articles, at least in Arabia, until one form eventually dominated the other. While it is impossible to prove which solution may have given rise to the hal-article, and its later form ʔal, we can, I think, with certainty rule out a source from the plural demonstrative base. Thus, Central Semitic article forms not only share their peculiar syntax but also derive collectively from a hā deictic particle with a suffixed n and l.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

14

Ahmad Al-Jallad

Explaining the article as a contact feature in Proto-Arabic The very existence of forms of Arabic written in the Hismaic and Safaitic scripts lacking the definite article strongly suggests that their common ancestor lacked a morphological means of marking definiteness. I would therefore suggest that the definite article entered these languages through contact with Northwest Semitic languages in the southern Levant. Evidence for such a contact situation is in fact attested in an Ancient North Arabian inscription published by Hayajneh, Ababneh, and Khraysheh in 2015. The text is in a difficult-to-classify North Arabian script but contains a prayer formula typical of Thamudic B and Safaitic. Following the edition, the prayer reads: h mlkm w kms1 w qws1 bkm ʿwḏn … ‘O Mlkm (= Malkom), Kms1 (= Kemōš), and Qws1 (= Qaws) we have placed under your protection …’.38 The short inscription attests to cultural contacts between speakers of North Arabian languages and the Canaanite languages east of the Jordan. In addition to this, the inscription may be a bilingual, as the same stone bears a Canaanite text, which remains undeciphered. The published photographs of this component, like those of the Arabian inscription, are too poor to be of use for re-editing the text. Nevertheless, this cultural nexus would have produced a bilingual environment in which the definite article, as a morphological feature and perhaps the morpheme itself, could have spread to Arabian languages lacking it. If contact was responsible for the introduction of the definite article in Arabic and other North Arabian languages, then the earliest form would have been *ha(n), with the assimilation of the n to the following consonant – and this is exactly what we encountered in the inscriptions. The introduction of the *h/ʔal- article could be the result of a calque, imperfect borrowing, or have arisen from the functional overlap of *han and *hal in other contexts Accordingly, it may be a later, more restricted innovation. Indeed, our first clear attestations of *h/ʔal come from the Nabataean period, centuries after the attestation of *ha(n).

38 This is the reading of the first line; the second line is read by the edition as h ʾsḥy m mḏwbt (oder mdwst) and translates the entire texts as “O Mlkm und Kms und Qws, wir suchen bei euch den Schutz (für) die/diese Brunnen vor Leck (oder: Zerstörung, Verderben)”. The spellings of Kms1 and Qws1 with the same sibilant, s1, indicate that this language has merged s1and s3. The second line is read from right to left, but I suspect that the text is boustrophedon. The published photographs are not of a high enough quality to undertake a re-editing of the text, however, and so I will withhold judgement of the reading and interpretation of the second line until new photographs are published. Fortunately, H. Hayajneh informs me that he is preparing the text for a new round of 3D photography and will publish a new edition soon.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

15

A simple sound change produces the greatest diversity in the shape of the definite article, h >ʔ #C_CC(v́) (Al-Jallad 2015b).39 From an original han one gets ʔan, the n of which would regularly assimilate producing the ʾ-article – both forms are found frequently in Safaitic.40 The ʔal-article derives in a straightforward way from hal. The fact that the sound change h > ʔ #C_CC(v́) operates normally across North Arabia, affecting for example the C-stem verb (ʔafʕala rather than hafʕala) further supports the idea that the *ha(n) article is a borrowed form, as it is not affected by this change. Now, the assimilatory patterns of the ʔal-article require an explanation. It is important to emphasize that in most of the Old Arabic data, the coda of ʔal rarely exhibits assimilation; thus, there is no reason to assume that the assimilatory patterns of Classical Arabic held true for the ʔal-article across space and time.41As I have argued above, some varieties may have used several allomorphs of the article in free variation. Classical Arabic may have inherited both the *ʔan and *ʔal article variants and redistributed them along phonological lines: ʔal before dorsals and ʔan before coronals. The assimilation of the n in this context is not problematic, as even Classical Arabic experiences n assimilation with particles, e.g. ʔan+ lā >ʔallā. Finally, contact between the languages of Arabia – not simply Northwest Semitic – and sociolinguistic factors arising from that, must have played an important role in the distribution of definite article forms in the pre-Islamic period. The ease at which definite article forms can be transferred across closely related languages undermines any attempt to use its particular phonetic shape – whether *ha(n) or *ʔal – to draw linguistic boundaries. An informative parallel is seen in Southwestern Arabian dialects of Arabic today. Traditionally, these dialects of Arabic used the ʔan- or ʔam-articles. Both forms are receding today at the expense of the spread of the ʔal-article.42 The latter is no doubt considered prestigious because of its presence in the major urban dialects. The restricted distribution of the ʔal-article in pre-Islamic times speaks to the absence of prestige attached to dialects using that form, and to the absence of a 39 That is, h shifts to ʔ in a closed word-initial syllable, possibly pre-tonic, while it remains h otherwise: so hafʿala> ʾafʿala but huwa remains huwa. 40 The article forms ʔan and ʔa with assimilation to all following consonants are attested in modern dialects of Southwest Arabia (Watson 2014: 150) and were known to the Classical Arabic Grammarians as well. The nasal article with m was said to be a characteristic of the Ṭayyiʾ dialect (Rabin 1951), and was widespread in South Arabia as well. 41 This is what Macdonald has called the “Northern Old Arabic isogloss” (2000: 51); See Al-Jallad (2017) for a full discussion of this feature in light of Greek transcriptions. 42 The ʔam-article, which for Classical Arabic grammarians was characteristic of “Ḥimyaritic” could derive from the ʔan-article. The /n/ could have assimilated to a following labial consonant, becoming /m/, and was thereafter generalized to all positions.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

16

Ahmad Al-Jallad

pan-Arabian linguistic order that would have caused the pre-Islamic dialects to converge in a single direction. 2.3 The isoglosses of Mascitelli and Robin Mascitelli (2006: 19–20) and Robin (2001) provide a few more features regarded as diagnostic of Arabic. I will focus here on those that do not overlap with Huehnergard (2017). 1) The ʔa-morpheme of the C-stem — ʔafʕal: This feature was also proposed by Robin (2001). While it is true that the causative morpheme of Arabic is a glottal stop in the suffix conjugation, the sound change of h >ʔ in this environment is rather common and can hardly be used, on its own, to diagnose a language. The same sound change occurred in Gəʕəz and in the attested history of Aramaic. As we shall see below, the Dadanitic corpus has both h- and ʔ-morphemes, suggesting a diachronic development from the former to the latter in that language’s history. 2) The onomasticon: Mascitelli considers certain personal names to be unique to Arabic, while acknowledging that these are also found in the North Arabian inscriptions. Personal names cannot be used to genetically classify languages.43 3) T-morphemes of the suffix conjugation: While Proto-Arabic certainly levelled the t-morpheme from the second person series to the first (qataltu rather than qatalku), this seems to have been an areal feature of all Central Semitic languages north of Yemen, common to Northwest Semitic as well (Huehnergard 2005: 168–169). The k-endings are an areal feature of southern Arabia, found not only in Ancient South Arabian, but all languages of this region regardless of their classification, e.g. Modern South Arabian, Gəʕəz, and even some modern Arabic dialects.44

43 For an excellent presentation of this point, see Macdonald (1999). 44 The fact that k-endings do not occur outside of Yemen strongly suggests that they spread to Arabic once it moved into the region; all attestations of Arabic, both ancient and modern, outside of Yemen have t-suffixes. On the distribution of the modern k-dialects, see Behnstedt (2016, verbs) and on the ancient dialects, see Rabin (1951, chapter 5).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

17

3. Why the Safaitic and Hismaic inscriptions should be classified as Arabic With the linguistic definition of Arabic in hand, let us now turn to the dialects/ languages attested in the epigraphic record and attempt to classify them genealogically. I have already argued in several places (2014, 2015a, 2017, 2018a) that the dialects attested in the Safaitic and Hismaic inscriptions should be considered Arabic, in genealogical terms, and constitute a continuum of dialects including the Arabic of the Nabataeans. The following discussion will synthesize my opinions on the matter and make the case for this as clear as possible. Putting aside the superficial shape of the definite article, Safaitic, and to a lesser degree Hismaic (on account of the laconic nature of inscriptions in that script), exhibit most of the innovations of Arabic. The system of negation Classical Arabic shows a unique system of negative adverbs tied to modality. The past tense is negated with lam and the short prefix conjugation (the jussive), the present tense with lā and the indicative (yafʕalu), and the future with lan and the subjunctive (yafʕala). In addition to this, the suffix conjugation is negated with mā.45 lam: Classical Arabic lam yafʕal ‘he did not do’ Safaitic: lm yʿd ‘he did not return’ lm ygd ‘he did not find’ mā Classical Arabic: mā faʕala ‘he did not do’ Safaitic: m mṭr ‘it did not rain’ m nm ‘he did not die’ m hnʾ ‘he was not happy ‘ lan Classical Arabic: lan yafʕala ‘he will not do’ Safaitic: lʾn yqtl ‘he will not be killed’

45 All examples come from Al-Jallad (2015a: 155–156).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

18

Ahmad Al-Jallad

Mafʿūl passive participle The productive passive participle of the G-stem (form I) in Classical Arabic is mafʕūl, although faʕīl and faʕūl patterns exist. The same is true of Safaitic, for example: mqtl [maqtūl] ‘killed’, mḥrb [maḥrūb] ‘plundered’, msby [masbeyy] ‘captured’.46 The subjunctive yafʿala Verbs terminating in an /a/ are found throughout Semitic, and the relationship between them has yet to be explained. Nevertheless, it is only in Arabic, and in Classical Arabic to be precise, that such verbs function as a marked subjunctive.47 The ability to detect such forms is challenged by the orthography of the ANA scripts, which does not indicate final vowels at all. Yet it is this orthographic practice that allows us to demonstrate that some verbs terminate in an /a/. Indicative prefix conjugation verbs of the III-y/w class end in a long vowel that is not indicated orthographically, ydʿ [yadʕū] ‘he reads’ (Al-Jallad and Jaworska, forthcoming, s.v). Thus, when the final radical is written, we can be sure that it is followed by a vowel, and in these cases, the verb is in a syntactic environment where Classical Arabic would require the subjunctive. Hismaic f ygzy nḏr-h w yzd ‘in order to fulfill his vow and do more’; ygzy = [yagzeya]48 Safaitic h lt qbll ʾhl slm f nngy ‘O Lt, [grant] a safe reunion with family that I may be saved’; nngy = [nangeya]49

46 See Al-Jallad (2015a: 117–123) for a discussion on the various reflexes of the participle. 47 While verbs terminating in an -a morpheme are attested in Akkadian, Ugaritic, Hebrew, and Amarna Canaanite, the relationship between these remains poorly understood and their syntax is quite different from Arabic. In Ugaritic and Hebrew, they function as cohortatives and there is no consensus on the function of this morpheme in Amarna Canaanite. 48 As Zwettler and Graf note, the indicative of this verb would have been yagzī and would have been spelled ygz in Hismaic orthography (Zwettler and Graf 2004: 64). 49 Again the significance of this spelling is brought into relief when we consider that the indicative prefix conjugation of dʿy ‘to read, invoke’ is spelled ydʿ (QZMJ 468) /yadʕū/ or /yadʕī/.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

19

The complementizer ʾn Arabic alone among the Semitic languages makes use of the particle ʔan(na) as a complementizer: Safaitic50 smʿ ʾn myt flfṣ ‘he heard that Philip had died’ srt smʿ ʾn ʿm-h… ‘he served in the military having heard that his paternal grandfather…’ Nunation and the vocative form with m The northern dialects of Old Arabic had lost final nasalization on singular nouns in the pre-historic period. Yet, forms with otiose n persist and these can be interpreted as vestiges of ancient nunation. A handful of examples are discussed in Al-Jallad (2015a: 69), but no system as such can be identified. In the QCT, and therefore in Classical Arabic, the divine name ʾallāh can take a special suffix when in the vocative: ʾallāhumma. The same peculiarity is found in Hismaic, where both the divine name lh (= (al)lāh) and lt (= (al)lāt) can take a final m in the vocative, e.g. h lhm = /hā-llāhumma/, h ltm = /hā-llātumma/ (King 1990: 63). 3rd feminine plural of the suffix conjugation The Proto-Semitic third person feminine plural of the suffix conjugation terminated in a long -ā, while Arabic levelled the prefix conjugation ending, -na, to the suffix, producing faʕalna (Huehnergard 2017). Feminine plurals are rare in the inscriptions, but one clear example has so far been attested in Safaitic, exhibiting the Arabic innovation: w ṭrdn h-ḫl (QZMJ 128) ‘and the horses were driven together’. Nevertheless, Huehnergard explains that this is a very obvious analogy and therefore the feature should probably not be assigned too much weight.

50 Al-Jallad (2015a: 165).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

20

Ahmad Al-Jallad

Prepositions, particles, and varia The value of prepositions is difficult to assess when it comes to classification. Mascitelli (2006: 19), followed by Huehnergard (2017: 22), conclude that the locative preposition fī ‘in’ is restricted to Arabic, and this is attested in Safaitic. A number of other prepositions and particles unique to Arabic are found in Safaitic as well, most notably ʿnd ‘at’,51 and possibly ḥt = (QCT ḥty, CAr ḥattā) ‘until’.52 Safaitic also attests the independent pronominal base y- (= Classical Arabic ʔiyyā) (Al-Jallad 2015a: 95).53 The form ʔanāku seems to have disappeared in Safaitic, but one must bear in mind that there are only a handful of cases requiring the 1st common singular independent pronoun (ibid.). In Hismaic, only the form ʾn is attested, 14 times in the OCIANA corpus. In addition to these, a unique Arabic development is attested in one Hismaic inscription that deserves remark. In many Arabic dialects, the imperative of the verb ‘to give’ has a quasi-suppletive form hāt, hātī, hātū, etc, a frozen form of the C-stem of √ʔtw ‘to come’.54 This feature is found in a rather clear context in a Hismaic inscription.55 The combination of irregular developments required to produce this forms speaks against a parallel development. T-demonstratives Huehnergard (2017) has suggested the t-demonstrative as Proto-Arabic innovation and I have followed him in previous works (e.g. 2015a: 12). If he is correct, then this is another example of a Proto-Arabic innovation attested in Safaitic, e.g. qbll m-t h-snt ‘may this year bring a reunion (with loved ones) (HAUI 198, Al-Jallad 2015a: 84); t h-ḫṭṭ ‘this writing’ (BS 870); t l-wṣn ‘this (writing) is for Wṣn’ (KRS 209). However, I am now inclined to follow Hueh-

51 AAEK 173: ʿrf sfr ʾb-h ʿnd h-rgm ‘he recognized the writing of his father at the cairn’. 52 This particle is attested once in an unformulaic context, so full confirmation of its existence must await future attestations; ṣyd ḥt wqf ‘he hunted until exhaustion (lit. stopping)’ (Al-Jallad 2015: 164). 53 AWS 218: s1ʿd-h rḍw nʿm w bḏ ḍr w yh brk ‘may Rḍw help him through divine favour because there is danger here and may he (Rḍw) bless him’. 54 See Al-Jallad (2017: 104, n.22) on this feature. The form exhibits two remarkable features, the preservation of the h- of the causative morpheme and the loss of the glottal stop: haʔti >hāt, instead of the expected ʾʾt in Hismaic. 55 The text was deciphered by King (KJC 46) with the first line re-edited by Al-Jallad (2017: 104, n.23) as: w m ḥll ḍyr-h / ht ʿśw w rsl / smʿt ḏśry w ktby ‘and whoever washes his wounds, give [an offering of] an evening meal and milk that Dusares and Kutbay may hear’.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

21

nergard’s other suggestion, namely, that the t-demonstratives are a “bizarre remnant from Afro-Asiatic”.56

4. The classification of Dadanitic Dadanitic is the term Macdonald gave to the script of the oasis of Dadān, in north-western Arabia, replacing earlier terms such as Dedanite and Lihyanite. The Dadanitic corpus is far smaller than the Safaitic one but it is the only corpus of Ancient North Arabian to have a true monumental tradition.57 The PhD thesis (in preparation) of F. Kootstra argues convincingly that the oasis of Dadān was not monolingual and that the diversity of spoken languages at the oasis may have had an effect on the linguistic profile of the inscriptions. The main language of the Dadanitic inscriptions appears to be distinct from Arabic and exhibits features that cannot be otherwise reconstructed for Proto-Arabic. These are: The causative In addition to the ʔ-causative common to most of the Arabian languages north of Yemen, Dadanitic has h-causatives like Sabaic. This feature cannot be understood as the result of influence from the Minaean colony at Dadan (Rossi 2014), as Minaic has an s-causative. Like Aramaic, it would seem that Dadanitic experienced the shift of h to ʔ in the causative morpheme in the historical period (Gzella 2015: 34). The co-occurrence of both forms in the same formulaic context, however, suggests that the progressive ʔ-form had entered the scribal school as well (Kootstra, forthcoming). While some remnants of the h-causative can be identified in Arabic, most of which are loans, no productive h-morpheme is reconstructible for Proto-Arabic.

56 A paper on the reconstruction of the Proto-Arabic demonstrative series is currently in preparation. The main reason for the shift of my opinion is that the t-onset cannot be plausibly derived from any other part of the grammar, all other feminine markers being suffixes. Moreover, the principle of archaic heterogeneity favors a demonstrative series ḏ (masculine), t (feminine), ʾul (plural), where subsequent ḏ feminine can be explained through leveling. 57 According to OCIANA, the number of Dadanitic texts is 1969 while Safaitic has 33164, and this does not include the tens of thousands of texts discovered since the publication of the database. For an overview of Dadanitic (nomenclature and grammatical features), see Macdonald (2000; 2004) and the forthcoming dissertation of F. Kootstra.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

22

Ahmad Al-Jallad

The anaphoric pronoun hʔ Dadanitic retains the use of the 3rd person pronouns as anaphoric demonstratives, meaning it did not participate in the Arabic innovation of replacing these forms with distal demonstratives derived from the proximal base (Huehnergard 2017: 22).58 Note also that the form of the pronoun is hʾ [huʔa], with a glottal stop, rather than hw [huwa]. Only the latter form is reconstructible for Arabic. The feminine ending Dadanitic does not seem to have levelled the -at allomorph of the feminine ending (Huehnergard 2017: 20), e.g. qrt ‘settlement, village’ /qarīt/ vs. Arabic qaryatun, which would appear in Dadanitic orthography as qryt.59 The form of the dual Verb forms originally terminating in -h have previously been regarded as duals (Sima 1999; Macdonald 2004: 526). These contrast with the dual ending on pronominal suffixes, which is -y, -hmy ‘both of them’. I suggest that this asymmetric situation should probably be reconstructed for Proto-Semitic, with languages like Sabaic levelling ending -ay to the verb while Classical Arabic levelled the -ā of the verb to the pronominal suffixes: Dadanitic

Classical Arabic

Sabaic

Verbal suffix

-h /ā/

/ā/

y /ay/

Pronominal suffix

-y /ay/

/ā/

y /ay/

In this case, Dadanitic preserves the Proto-Semitic situation and, as such, the feature cannot be used for classification.60 58 For example, JSLih 78: w ʿl-h hʾ ‘and the aforementioned is for him’. 59 This occurs in the inscription JSLih 64; a convincing interpretation of this difficult text is offered in Lundberg (2015: 134). 60 The Classical Arabic situation is not likely reconstructable for Proto-Arabic. Most of the Arabic dialects have lost this feature, and so it is unclear what forms they would have had in the verb or pronouns. Safaitic does not employ matres lectionis, and so it is unclear if the pronouns inflect for the dual. The verb, on the other hand, seems to attest a dual

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

23

The conjunction ʿdky A minor feature. In all forms of Arabic, the Proto-Central Semitic subordinating conjunction ʕad+kay has been replaced by a new particle ḥattay, which, I have argued, is a combination of the nominal form ‘border’ ḥadd and the subordinator tay/tā.61 Dadanitic retains the original conjunction.62 The numerals Huehnergard (1995) identified the form of the tens terminating with the masculine external plural -ūna, -īna as a Central Semitic innovation. These forms are found in Dadanitic, e.g. ʾrbʿn, ʿs2rn, etc. (Macdonald 2004: 522), distinguishing it from Ancient South Arabian, which has -y. These features, I think, suggest that Dadanitic did not descend from Proto-Arabic but is rather a sister language. Whether Proto-Arabic and Dadanitic constitute a separate sub-grouping or whether their common ancestor is Proto-Central Semitic will have to await the discovery of more texts. 4.1 Arabic at Dadān While Dadanitic proper may be excluded as a variety of Arabic, there was another Semitic language in the area that seems to have been a variety of Arabic, one which I have called Old Ḥigāzī (Al-Jallad 2015a: 13–14; forthcoming). The main isogloss of Old Ḥigāzī is the relative pronoun based on the portmanteau demonstrative (h)alla+DEM, e.g. CAr ʔallaḏī. JSLih 384 attests the feminine form ʾlt /ʔallatī/ (Müller 1982: 32–33; Macdonald 2008 and Macdonald in Fiema et al. 2015: 409). The feature, attested once, may suggest that speakers of Old Ḥigāzī used Dadanitic as a literary language, much like speakers of Nabataean Arabic commissioned their official inscriptions in Aramaic. The presence of speakers of Old Ḥigāzī may also explain the occasional use of the ʔal-article in Dadanitic, e.g. ʾ-ẓll ‘the ẓll-ceremony’ (AH 119, 138); ʾl-ʾsd ‘the lion’ (Al-Ḫuraymāt 4). Macdonald considered these examples of mixed Old Arabic texts (2000, 2008), and if JSLih 384 indicates that Old Ḥigāzī was spoken at the oasis at the same time as Dadanitic, then it is certainending y (Al-Jallad 2015a: 103), suggesting a similar innovation to Sabaic. If so, then the asymmetric paradigm of Dadanitic should also be reconstructed for Proto-Arabic. 61 See Al-Jallad (2014) and (2018b). 62 The particle is attested in JSLih 72: mn mʿn h-gbl hn-ʾʿly ʿdky mʿd h-gbl hn-ʾs{f}l ‘from the assembly place of the upper border to the sanctuary of the lower border’ (Lundberg 2015: 135).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

24

Ahmad Al-Jallad

ly possible that the ʔal is the result of substrate influence. However, considering the Safaitic situation and, indeed, that of Thamudic F (below), it is possible that variation in the shape of the definite article was something inherent to Dadanitic as well. Another feature typical of Classical Arabic and the QCT is the replacement of the infinitive as a verbal complement with a subordinated clause introduced by ʾn, cf. CAr ʾan yafʕala ‘that he do’. This feature is attested only once in the North Arabian epigraphy, in a fragmentary Dadanitic inscription from al-ʿUlā. AH 203 1: hm ---- [ḏ]– 2: ġbt/ʾ{n}/yk{n}---3: l-h/{w}ld/f rḍy [-h] ---4: w ʾḫrt-h {ḏ}---‘[PN made an offering to] Ḏġbt that he may have offspring so satisfy him and his progeny’ The text contains no other diagnostic features, so it is impossible to say if in the text was carved in Old Ḥigāzī, like JSLih 384, if the feature is simply substrate influence on Dadanitic, or if this is originally a Dadanitic feature that was taken over by the Arabic dialects of the Ḥigāz but nowhere else.

5. Taymanitic Taymanitic refers to the North Arabian script and its associated language used at the oasis of Taymāʾ and surrounding areas. The corpus and the classification of its language was the subject of a comprehensive study by Kootstra (2016), in which special attention was given to its linguistic classification. Kootstra convincingly argues that Taymanitic should be excluded from the Arabic classification as it merges *s3 and *ṯ rather than *s1. This merger belongs to a broader shift that collapses the interdentals and sibilants, merging *ḏ and *z and *ṯ ̣ and *ṣ as well (2016: 105). Taymanitic also does not participate in the Proto-Arabic sound change of iwa to iya, spelling the verb *raḍiwa as rḍw rather than rḍy (= /raḍiya/). A final development is the shift of w to y in word-initial position, attested clearly twice, e.g. yrḫ ‘month’ (ibid.),63 which may suggest a closer affinity to Northwest Semitic. Confirmation of this hypothesis must await the discovery of longer and clearer texts. 63 Proto-Semitic *warḫum, Safaitic wrḫ (Al-Jallad 2015: 353), but Proto-Northwest Semitic *yarḫu.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

25

6. Thamudic Before discussing the classification of the Thamudic inscriptions, we should make it absolutely clear that Thamudic does not refer to a single language or script, nor does it refer to a group of languages and scripts that have more in common with each other than the other groupings. Following Macdonald, it is simply a category for inscriptions that do not fall into one of the better understood classifications. The interrelationships between the different scripts remain poorly understood, as do their chronologies, and languages.64 Today, four categories of Thamudic are distinguished, B, C, D, and F. These groupings are certainly too broad and will be revised in upcoming years as the texts are subjected to closer and more systematic studies. B, C, and D are what remain from Winnett’s seminal 1937 study of the Thamudic inscriptions,65 and the Franco-Saudi surveys of the Ḥimà region near Naǧrān have added a new member, F or Himaitic, to the group (Robin and Gorea 2016). The following discussion will offer some preliminary impressions of the linguistic affiliation of these texts and will certainly be subjected to modification as our knowledge improves. 6.1 Thamudic B66 The classification of Thamudic B at this point is impossible. The corpus contains primarily personal names and signatures associated with rock art; moreover, the scripts classified under this rubric are quite diverse and certainly form better defined sub-groupings. The definite article, which is mostly used with a demonstrative force, is always h-. The first and second person suffix conjugations terminate in t, unlike the k’s of South Arabia.67 One peculiar feature is the use of nm as a dative preposition, perhaps reflecting an original form lm, as in Taymanitic (Kootstra 2016: 101), with regressive assimilation, *lima > nima. The imperative verb is often followed by an energic n, ʾtm-n ‘restore to heath!’ (Winnett 1987). Winnett has suggested the existence of the dative preposition k- but confirmation must await the discovery of clearer examples (Winnett, no date). Despite the fact that Thamudic B is the most widespread category, the limited linguistic 64 For a good overview of Thamudic, see Macdonald (2000). 65 Originally Winnett also distinguished a Thamudic A and E, but these are today called by most Taymanitic and Hismaic, respectively. 66 On the identification of the letter shapes of Thamudic B, see Winnett (1937: 28–34) and the script chart in Macdonald (2000). 67 For example, the phrase ʾn rfʾt ‘I am healed’ (Eskoubi-B 175 in Hayajneh 2011: 770).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

26

Ahmad Al-Jallad

facts provided by the inscriptions prevent any judgement regarding its linguistic classification. In terms of content, Thamudic B overlaps with Safaitic and Hismaic in a few ways, namely, prayers of the structure h + divine name + sʿd ‘help’ + PN and the signing of rock art, something rarely encountered in other corpora. 6.2 Thamudic C68 The Thamudic C inscriptions are characterized by an introductory formula with wdd. The exact meaning of this word is not agreed upon by scholars (Tsafrir 1996). These texts can be subdivided into several categories, with two main branches C1 and C2, based on geography and the phonemic values of the glyphs (Al-Jallad 2016). Considering the brief nature of these texts, there is only a few things to say about their linguistic character. The verb and pronouns The first person possessive suffix pronoun is consonantal, y, at least when attached to a noun in the accusative case and the personal suffix of the first person suffix conjugation is a t, e.g. ktmt s¹my = */katamtu simay(y)a/ ‘I have concealed my name’,69 in contrast with the k endings of Ancient and Modern South Arabian. In a single inscription with an invocation to dgn (= Dagan(?)), an s-based 3ms pronoun is attested: WRTH 8070 h dgn l- yd h-ʾlht mlt -s¹ ‘O Dagan, may his people be in the company of the goddess/ gods.’ 68 On the script type and contents, see Winnett (1937: 34–38). The corpus is spread from Northwest to Central Arabia and no chronological information has come to light, so any dating of this grouping is entirely conjectural. 69 There is no evidence for the use of matres lectionis in Thamudic C so the y in this inscription must be construed as consonantal. On this inscription, see Winnett and Reed (1970: 131–32, no. 74). 70 This inscription makes use of the bar with two circles on each end for the glyph g, the same value this glyph has in Hismaic. Most previous studies have regarded the divine name here as dṯn; however, a connection with dgn seems more plausible as the latter is a well-known deity. This issue will be taken up in a comprehensive article on the Thamudic C inscriptions, which is in preparation. My interpretation of the inscription differs entirely from the editio princeps and so I will provide a short commentary here. l-yd I take as /la-yad/ ‘at the hand’, cf. Aramaic lyd ‘next to’ (CAL) and perhaps related to the Arabic preposition ladā ‘at’< *la-yadā. The term mlt should be equated with Arabic

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

27

This inscription also provides the single attestation of the definite article in Thamudic C, h-ʾlht. It is also possible that hʾlht is a proper noun and so the definite article does not reflect the situation in the language of the Thamudic C inscriptions. The main prepositions of this corpus are l- ‘authorship, to, for, by’ and possibly k- ‘to, for’ (JS 50; NTQT 83). The conjunction f is attested perhaps with a comitative meaning,71 and w to connect two clauses. Several personal names are attested with mimation: rs¹qm */risqum/ (NTJT 82); ʾmnm */ʔamanum/ (NTJT 93); gmʿm */gumʿum/ (NTJT 105); ʾmtm */ ʔamatum/ (NTJT 115). If this is a feature of the language of these inscriptions rather than of their source, then this would disqualify Thamudic C as a form of Arabic. Finally, two proximal demonstratives are attested, zn and zt, indicating the loss of the interdental ḏ. Although short, the Thamudic C inscriptions exhibit an array of features that are completely alien to Arabic and likely reflect a distinct language. This fact is brought into relief when we consider texts such as the following, which has so far defied interpretation, suggesting that if longer texts were available, we would be dealing with a variety as distinct from Arabic as the non-Sabaic Ancient South Arabian languages. Esk 204: wdd f s¹w//tʾlʿs¹ s¹wʾ//wdd 6.3 Thamudic D Thamudic D furnishes us with one dated inscription: JSTham 1 accompanies the famous Nabataean inscription JSNab 17, dated to 267 CE. As such it is the latest securely dated Ancient North Arabian inscription. But when it comes to linguistic classification, we can say nothing in any detail about the affiliation of these inscriptions. As far as I know, not even the definite article has been attested in this corpus. Two introductory particles are known, zn and zt, which could reflect proximal demonstratives and attest to the merger of ḏ and z. These millatun ‘people, nation’ perhaps related to the root √mll ‘to speak’, referring to co-linguals originally. The sense of ‘to speak’ is entirely lost in Classical Arabic, but Safaitic attests the term mly ‘words’ once, in reference to a Greek inscription (!) (Macdonald, Al Muʾazzin, and Nehmé 1996, MISSI 2; Al-Jallad 2015a: 327). 71 The f frequently occurs after wdd, overlapping with b-, and so it is reasonable to assume a similar function. While some have connected the f to the Arabic preposition fī, it is probably better to regard it as closer in function to the comitative wāw (wāw al-māʿiyyah), lit. wdd f PN would render ‘greetings be with PN’.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

28

Ahmad Al-Jallad

alone suggest that the language is not ancestral to Arabic or closely related to Classical Arabic. The inscriptions sometimes contain expressions of love and longing, with the verb ʿs2q, meaning presumably ‘to love’, and a participle mḥb or mḥbb. These variant spellings suggest that it is a passive particle of the C-stem rather than a G-stem mafʿūl pattern; thus, muḥabb or muḥbab. A few longer texts exist in Thamudic D and these have so far defied interpretation.72 6.4 Thamudic F Thamudic F is one of the terms suggested by Robin and Gorea (2016) for the non-South Arabian inscriptions of the site of Ḥimà, near Naǧrān.73 These texts consist almost exclusively of names and the verb wqr ‘to carve’. In some of the theophoric names, however, the definite article is attested in many shapes, hl, ʾl, h, hn, h, and ʾ, and the Ancient South Arabian suffix -n. While this says nothing about classification, it does provide yet another example of the extreme diversity of article forms attested across Arabia. Some of the personal names are also marked with mimation. Prioletta (2018) suggests that there is evidence for the shift of w to y in word-initial position, citing the following by-forms wqr = yqr and wdd = ydd. The verb form tqr was identified as a Gt stem (ibid), equated with CAr ittaqara, suggesting the infixation of the t.

7. The question of the “Northern” dialect of Sabaic The Sabaic inscriptions from the Yemeni Jawf, called in the literature by various names, including Haramic, Amiritic, Sabaeo-North-Arabian, and simply the ‘Northern Dialect’ (Stein 2004: 228), are characterized by a number of

72 Consider, for example, WTI 63, which Winnett and Reed read and translate as follows: wznzʾ(g)wzfry|wḫmyztfʿlmrḥḫm ‘And Zʾg and Zfry have committed adultery. And this deed stinks worse than a stinking fart’. While this interpretation strains credulity, the syntax and etymological analysis are not convincing. 73 These texts have previously been labelled “southern Thamudic” (Macdonald 2000) and were preliminarily described by Ryckmans (1956). I have settled on the term Thamudic F rather than Himaitic because there remains a great deal of graphic variation in the script to sort out. Thamudic F, in the end, may encompass several scripts and so the label Thamudic seems appropriate for now.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

29

features that have suggested to scholars an “Arabic” admixture. These are, following Stein: 1) 2) 3) 4) 5) 6) 7)

The assimilation of n to a following consonant; The use of the s3 glyph for etymological *ṯ and s1 for etymological *s; The spelling of the preposition ʿdy as ʿd; Inconsistent use of mimation; The preposition mn rather than Sabaic bn; The conjunction b-hn ‘because’; The negative preterite formed with the use of lm and the short form of the prefix conjugation.

In addition to this list, we may add the person-number-gender suffixes on the 1st and 2nd person of the suffix conjugation in t (cf. Arabic katabtu, katabta, katabti) rather than k (Sabaic, fʿlk ‘I have done’). Macdonald (2000: 55) discusses these features in his usual thorough and balanced way. While such a list would suggest that these ‘intrusions’ are a regular part of the dialect, Macdonald emphasizes the rarity of these forms. From a corpus of twenty texts, the negative lm + prefix conjugation occurs in only four. The use of s1 for s3 in only two texts, and the use of s3 for ṯ in only one; these texts are in all other respects Sabaic. But more important than their rarity is the fact that none of these features, with the exception of lm yfʿl, points exclusively towards Arabic. All are characteristic of any non-South Arabian Semitic language. By assuming an Arabic source, one risks essentializing the languages of Arabia to two groups – Arabic and Ancient South Arabian. Macdonald (2000: 57) offers a compelling solution to the linguistic nature of these texts: “At Haram, the limited range and nature of the non-Sabaic elements in the Sabaeo-North-Arabian inscriptions suggest that they are clumsy attempts at writing correct Sabaic by people whose mother tongue was either a different language, or a dialect of Sabaic which contained elements from another language.” The final possibility should be emphasized again. We should not necessarily assume a strong linguistic border separating Sabaic from the Arabian languages spoken to the north of it. Instead, it is likely that Arabia constituted a continuum of Central Semitic languages, for which we are afforded only glimpses by the epigraphic record. The Yemeni Jawf could have been a frontier area, where features typical of North Arabian languages, broadly defined, and Saba-

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

30

Ahmad Al-Jallad

ic mixed, very similar to the dialects of the ʿAsīr today.74 In other words, the Haram inscriptions may be a point on a dialect continuum between a language like Dadanitic, which has h-causatives, dual pronouns in hmy, the anaphoric use of the 3rd person pronoun hʾ, but a prefixed definite article, and Sabaic. About 100km to the northeast at the oasis of Qaryat al-Fāw, we encounter another artefact attesting to this linguistic continuum.75 While perfectly good Sabaic and Minaic texts come from the oasis, an epitaph in a language so far unique has been discovered, the so-called rbbl bn hfʿm grave inscription. Unlike the seemingly ‘mixed’ character of some of the Haram inscriptions, this text was composed in an entirely different language. Most scholars have regarded the inscription as an example of “Old Arabic” on account of the definite article ʾl.76 In addition to all the arguments I have presented above for treating ʾl as the single isogloss defining whether a language is or is not Arabic, it is clear that ʾl was simply one of many article forms used in this region. Other features of this text mark it off from Arabic, including the productive use of mimation, yet according to a distribution distinct from Sabaic, a negative construction with bn, the conjunction ʿdky, and more.77 This epitaph could attest a dialect that sat on this continuum of Central Semitic languages, a close relative of Proto-Arabic, yet not participating in some of the important isoglosses that define that group of languages, such as spread of nunation to singular nouns.

8. East Arabia The linguistic situation of East Arabia is far less clear than that of the west. Like West Arabia, Aramaic was used widely as a literary language (Stein 2018). But in the Late Antique period, Syriac writers from the region refer to a non-Arabic 74 For example, the Arabic dialect of Riǧāl Almaʿ (Saudi Asir), the word for daughter is brat, similar to Modern South Arabian and Aramaic (Watson 2011a: 861)! Watson has produced a series of studies (Watson 2011b; Watson 2014; Watson et al. 2006) on the Arabic dialects of south-western Arabia. While these dialects exhibit a comparable mix of features, they do not represent a natural continuum from Arabic to the ancient languages of South Arabia. Instead, they likely resulted from a convergence between the later movement of Arabic into this region and early Semitic languages. Features like bar and brat, for example, are not typical of Sabaic, but can be found in Modern South Arabian, which is today spoken hundreds of miles away. 75 To date, the only resource on the excavations of this site and its inscriptions is al-Ansary (1982). 76 For example, Beeston 1981; Robin 2001; Macdonald 2008. 77 The genetic position of the language of this epitaph is discussed in detail in Al-Jallad (2014).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

31

and non-Aramaic colloquial called qaṭrayith.78 This language is only known from glosses in Syriac manuscripts – its lexicon appears to be distinct from Arabic (Kozah, forthcoming), as we know it, but little more can be said.79 The inscriptional evidence attests a local Arabian language in a script called Hasaitic by scholars. It is a close derivative of the Ancient South Arabian alphabet, but its language was considered a part of the North Arabian group because some of the personal names attested in the inscriptions contain the hn article, e.g. ʾmt-hnʾlt ‘maiden of the goddess’, ʾws-hnʾlt ‘gift of the goddess’ (Sima 2002: 189) and certain formulaic expressions.80 As Macdonald has so carefully argued in the past, theophoric elements in personal names cannot serve for the diagnosis of the language of the text within which they are couched. The relatively small corpus of formulaic grave inscriptions, however, offered little else to go on. A single feature, discussed by A. Sima (2002: 193–194), in a bilingual Hasaitic-Aramaic inscription from Thaj suggested that the article in Hasaitic was a post-positive glottal stop, mlkʾ, as in Aramaic.81 With only one attestation, and the presence of an Aramaic component in the same inscription, this identification could be little more than a suggestion. The excavations at Mleiha in the Arabian Gulf, however, have brought to light several new texts in Hasaitic. A minuscule inscription carved on a silver plaque confirms that the article in Hasaitic was a suffixed ʾ, as the following phrase is entirely Hasaitic and no appeal to Aramaic can be made: ṣlḥft-ʾ zʾt ‘this metal plaque’ (Stein 2017). The same inscription also gives the verb hqnt as the causative of qny ‘to offer’. If this is a native feature, and not a borrowing from Sabaic, then it would

78 Note that in Syriac literature from the 4th to 9th centuries, Qatar refers to the entire Northeast Arabia, including present-day Qatar, Bahrain, and parts of the UAE. 79 For a good overview of Qaṭrayith, see Contini (2003); Kozah (2014). 80 Macdonald (2008: 492–493) characterizes the Hasaitic script as basically the Ancient South Arabian one with some modifications. The language was included in the Ancient North Arabian category on account of expressions like ḏ ʾl ‘of the lineage’. This of course is a non-linguistic criterion and is in any case absent so far in Thamudic C and D, and is extremely rare in Dadanitic and Thamudic B. 81 The inscription sits at a bottom of a well and is badly weathered. Lines four and five are partially legible and read following Sima l(h)/snt/ʾḥdy/ʾtbl/mlkʾ/wdʾb ‘im Jahre eins von ʾtbl, dem König. Wadd-ʾAb.’ Robin and Prioletta (2013) very cleverly argue that ʾtbl is the Characene king Attembelos, but do not explain the missing final s and must appeal to the ad-hoc writing of mb as simply b. The inscription has been re-discovered during the first season of the Saudi-Dutch-French Thaj Archaeological Project and it seems that Sima’s reading of the first two letters as ʾ and t is rather difficult and certainly impossible to substantiate on the stone itself.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

32

Ahmad Al-Jallad

suggest that Hasaitic had h-causatives as well. The spellings hqnt instead of hqnyt further suggest that triphthongs had collapsed in certain environments.82 The inscription Thaj.Has 1 spells the sibilant of the common Arabian name zayd with the hashtag glyph #, which is the South Arabian alphabet signifies /ḏ/. This can only indicate that the voiced interdental and sibilant had merged to /z/, similar to Ḥaḍramitic, Thamudic C and D. The most famous of these is the bilingual tomb inscription of ʿmd son of Gr, the bqr83 of the king of ʿmn (= ʕumān) (Overlaet, Macdonald, and Stein 2016). The inscription furnishes a single isogloss that has been the subject of some debate. In addition to the word bn in the Hasaitic text, the author uses br. While the edition takes this as an Aramaicism, it is possible in light of the use of a similar form in the Modern South Arabian languages to consider this the result of substrate influence from a language similar to Modern South Arabian. The text also attests a dual relative pronoun zy, perhaps /zay/,84 giving us the following paradigm for Hasaitic (ibid. 137):

82 Stein (2017: 115) suggests that hqnt should be vocalized as haqnēt, reflecting the collapse of the triphthong to ē. While certainly possible, one cannot rule out haqnat or even haqnit/haqnet, as the vowel would likely have been reduced in a closed syllable. According to the most plausible reconstruction of the language of the QCT, we would have the pair bny /banē/ and bnt /banat/ (Van Putten 2017). 83 The meaning of this term is unclear; see the edition for a discussion. 84 The edition considers the possibility that zy is a plene spelling of the singular, which usually appears as z. However, given that zy occurs in a second inscription with a dual antecedent and never with inscriptions with a singular antecedent, this explanation seems most unlikely, as they concede (Overlaet, Macdonald, and Stein 2016: 137, n. 22).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

Masculine Singular

z /zV/

Feminine Singular

zʾt /zāʔat/ or /zaʔt/

Dual

zy /zay/

Plural

ʾlwt /ʔVlawāt/

33

The numerals The number twenty is attested once as ʿs2rn (Robin 1994, no. 1), exhibiting Huehnergard’s Central Semitic innovation in the tens (1995). The combination of these features makes Hasaitic quite unlikely to be descendent of Proto-Arabic and certainly rules out any genetic relationship with the dialects of Arabic spoken in the region today. It is impossible at this point to assess the relationship between Hasaitic and Qaṭrayith. 8.1 The Dhofār Inscriptions The Dhofār region of Omān is home to a small, yet ever-growing, corpus of painted inscriptions in a South Semitic script, not identical to any of the better known Ancient North Arabian alphabets or the Ancient South Arabian alphabet of Yemen. These short texts, found almost exclusively in caves, have remained undeciphered.85 It is possible, although impossible to prove at the moment, that the language of these texts is ancestral to the Modern South Arabian languages of the region, such as Mehri, which have so far not appeared in the epigraphic record.

85 The texts discovered as of 1991 are catalogued in Al-Shahri and King (1991), but this edition contains only hand drawings. New technologies will hopefully improve their study (Le Quellec et al. 2018).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

34

Ahmad Al-Jallad

9. Conclusion This essay has hoped to demonstrate three arguments: 1) Ancient North Arabian is not a well-defined linguistic family to the exclusion of Arabic. 2) Arabic cannot be defined by a single isogloss, the shape of the definite article. A more robust linguistic definition includes Safaitic, Hismaic, and some inscriptions in the Dadanitic script, at least, into the group of languages we must regard as Arabic. 3) Ancient Arabia comprises a linguistic continuum of Central Semitic languages, stretching from the southern Levant to northern Yemen, possibly including Sabaic. As argued by Stein and Kottsieper (2014), the Non-Sabaic Ancient South Arabian languages and the Modern South Arabian languages could reflect a more archaic stratum, pre-dating the southern expansion of Central Semitic.86 The place of the languages attested in Thamudic awaits better documentation, but one can speculate, based on their impenetrable language, that they too reflect a pre-Central Semitic stratum. From a geographic perspective, the isoglosses typical of Arabic are clustered at the northern end of this continuum, in the north-western Ḥigāz and southern Levant. While this may be in part due to the better documentation of the languages in this area, inscriptions from North Central Arabia, South Arabia, and East Arabia all reveal languages considerably distinct form Arabic and do not participate in several Proto-Arabic innovations. Thus, it would seem that the language which we would identify as Arabic emerged sometime in the second millennium in this region and then eventually spread across the Peninsula, replacing its sister languages on this continuum. The reasons for this are at present unclear. A preliminary classification of the languages of pre-Islamic North and Central Arabia is as follows (Figure 1). I have omitted Thamudic B because of the lack of information and the emerging picture that the scripts classified under this rubric are quite diverse and require a complete rethinking. The languages in the dotted circle were argued to be on a continuum between Sabaic, Dadanitic, and Arabic. The question marks next to Thamudic C and D indicate that their linguistic features do not allow for a more precise classification. 86 I cannot, however, follow the suggestion of a special relationship between Sabaic and Aramaic.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

35

Figure 1: Classification of the languages of pre-Islamic North and Central Arabia

Proto-West Semitic

Ethio-Semitic Central Semitic Thamudic C (?) Thamudic D (?) Modern South Arabian

Northwest Arabic Dadanitic Ancient South Hasaitic Semitic Arabian Haramaic Sabaic inscriptions and rbbl bn hfʿm

Northern Old Old Ḥiǧāzī Arabic

epitaph of Qaryat al.-Fāw

Safaitic Hismaic Nabataean Arabic

A linguistic definition of the languages of the Arabian Peninsula also allows us to re-consider a popular question in the literature: “why did Arabic not appear in the epigraphic record until so late”? This impression was in part the result of the very narrow definition of Arabic as Classical Arabic. The isoglosses associated with Arabic do in fact appear in the inscriptional record, and from this perspective Arabic is widely attested in the pre-Islamic period, across many scripts. It would therefore seem that the isoglosses of Classical Arabic, as defined by Macdonald at the beginning of this paper, perhaps have a shallow chronological origin and were extremely geographically restricted. Thus, the question would be akin to asking why English is not attested until the 18th century, while defining English as a language that collapsed plural distinction in the second person pronoun. Despite the tens of thousands of inscriptions from across the Arabian Peninsula, Classical Arabic as such has not yet been attested. No inscription attests a fully functioning system of nunation, for example, and even when Classical Arabic forms such as ʔallatī are found, they co-occur with features that are rather distinct from the literary register of the Islamic period.87 87 In JSLih 384, the 3fs verb is bnh /banah/, similar to some modern Yemeni dialects and

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

36

Ahmad Al-Jallad

Map 1: Distribution of the languages and South Semitic scripts of pre-Islamic Arabia

Sigla AAEK: Safaitic inscriptions in Al-Manaser 2008 AMSI: Safaitic inscriptions in Al-Manaser 2016 AH Dadanitic inscriptions in Abū l-Ḥasan 1997 Al-Ḫuraymāt Dadanitic inscriptions on OCIANA AMJ Hismaic inscriptions in King 1990 BS Safaitic inscriptions published on OCIANA from the 2015 Badia Survey C Safaitic inscriptions in Ryckmans 1950 HAUI Unpublished inscriptions from the Mafraq Museum JS Thamudic inscriptions in Jaussen and Savignac 1909 JSLih Dadanitic inscriptions in Jaussen and Savignac 1909 distinct from the language of the QCT and Classical Arabic (Behnstedt 2016: 277).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

37

KJA Hismaic inscription in King 1990 KRS Safaitic inscriptions published on OCIANA NTJT Thamudic inscriptions in Al-Theeb 2000a NTQT Thamudic inscriptions in Al-Theeb 2000b NTST Thamudic inscriptions in Al-Theeb 2000c QZMJ Unpublished Safaitic inscriptions to appear on OCIANA ShNGA Safaitic inscription in Shdeifat 2003 Thaj.Has Unpublished Hasaitic inscription from Thaj, to appear in first season’s report. WRTH Thamudic inscriptions in Winnett and Reed 1973 WTI Thamudic inscriptions in Winnett and Reed 1970

Bibliography Abū l-Ḥasan, Ḥusayn. 1997. Qirāʾah li-Kitābāt Liḥyāniyyah min ǧabal ʿAkmah biminṭaqat al-ʿUlā. Al-Riyāḍ: Maktabat al-Malik Fahd al-Waṭaniyyah. Al-Jallad, Ahmad. 2014. “On the Genetic Background of the Rbbl Bn Hfʿm Grave Inscription at Qaryat Al-Fāw.” Bulletin of the School of Oriental and African Studies 77: 445–465. ———. 2015a. An Outline of the Grammar of the Safaitic Inscriptions. Studies in Semitic Languages and Linguistics 80. Leiden: Brill. ———. 2015b. “Yusapʿil or Yuhapʿil, that Is the Question: Two Solutions to the Sound Change *s1 >h in West Semitic.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 165: 27–39. ———. 2016. “The Thamudic C (Wdd F) Inscriptions.” Presentation Handout, North American Conference on Afro-Asiatic Linguistics 44, UT Austin. ———. 2017. “Graeco-Arabica I: The Southern Levant.” In Arabic in Context: Celebrating 400 Years of Arabic at Leiden University, edited by Ahmad Al-Jallad, 99–186. Studies in Semitic Languages and Linguistics 89. Leiden: Brill. Al-Jallad, Ahmad. 2018a. “The Earliest Stages of Arabic and Its Linguistic Classification.” In Routledge Handbook of Arabic Linguistics, edited by Elabbas Benmamoun and Reem Bassiouney, 315–331. New York: Routledge. ———. 2018b. “The Etymology of Ḥattā.” In To the Madbar and Back Again: Studies in the Languages, Archaeology, and Cultures of Ancient Arabia dedicated to Michael C.A. Macdonald, edited by Laïla Nehmé and Ahmad Al-Jallad, 338–345. Leiden: Brill.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

38

Ahmad Al-Jallad

Al-Jallad, Ahmad. Forthcoming. “The Damascus Psalm Fragment: Middle Arabic and the Legacy of Old Ḥigāzī”. LAMINE 2. Chicago: Oriental Institute. Al-Jallad, Ahmad, and Ali al-Manaser. 2015. “New Epigraphica from Jordan I: A Pre-Islamic Arabic Inscription in Greek Letters and a Greek Inscription from North-Eastern Jordan.” Arabian Epigraphic Notes 1: 51–70. Al-Jallad, Ahmad and Karolina Jaworska. Forthcoming. A Grammar and Dictionary of Safaitic. Leiden: Brill. Al-Manaser, Ali. 2008. Ein Korpus Neuer Safaitischer Inschriften Aus Jordanien. Semitica et Semitohamitica Berolinensia 10. Aachen: Shaker Verlag. ———. 2016. “Einige Altnordarabische Inschriften Aus Jordanien.” Orientalia Academiae Scientiarum Hungaricae 69: 457–69. Al-Ansary, Abdul Rahman. 1982. Qaryat Al-Fau: A Portrait of Pre-Islamic Civilisation in Saudi Arabia. Riyadh; London: University of Riyadh; Croom Helm. Al-Shaḥrī, Ali A.M. & King, Geraldine M.H. 1991. The Dhofar Epigraphic Project. A Description of the Inscriptions Recorded in 1991 and 1992. Unpublished manuscript. Al-Theeb, Sulayman. 2000a. Dirāsah li-nuqūš Ṯamūdiyyah min Ǧubbah bi-Ḥāʾil. Ar-Riyāḍ: Maktabat al-malik Fahd al-Waṭaniyyah. ———. 2000b. Nuqūš qārā aṯ-Ṯamūdiyyah bi-minṭaqat al-Ǧawf bi-l-Mamlakah al-ʿArabiyyah as-Saʿūdiyyah. Ar-Riyāḍ: Muʾassasat ʿAbd ar-Raḥmān as-Sudayrī al-Ḫayriyyah. ———. 2000c. Nuqūš Ṯamūdiyyah min Sakākā (Qāʿ Furayḫah wa-ṭ-Ṭuwayr wa-l-Qudayr) al-Mamlakah Al-ʿArabiyyah as-Saʿūdiyyah. Ar-Riyāḍ: Maktabat al-Malik Fahd al-Waṭaniyyah. Avanzini, Alessandra. 2009. “Origin and Classification of the Ancient South Arabian Languages.” Journal of Semitic Studies 54: 205–20. Bauer, H. 1912. “Mitteilungen zur Semitischen Grammatik.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 66: 103–14. Beeston, Alfred F.L. 1981. “Languages of Pre-Islamic Arabia.” Arabica 28: 178–86. Behnstedt, Peter. 2016. Dialect Atlas of North Yemen and Adjacent Areas. Handbuch des Orients 114. Leiden: Brill. Contini, Ricardo. 2003. “La Lingua Del Bēt Qaṭrāyē.” In Mélanges David Cohen: études sur le langage, les langues, les dialectes, les littératures, offertes par ses élèves, ses collègues, ses amis présentées à l’occasion de son quatre-vingtième anniversaire, edited by Jérôme Lentin and Antoine Lonnet, 173–81. Paris: Maisonneuve etLarose. Drewes, Abraham J., Thomas F.G. Higham, Michael C.A. Macdonald, and Christopher Bronk Ramsey. 2013. “Some Absolute Dates for the Development of the Ancient South Arabian Minuscule Script.” Arabian Archaeology and Epigraphy 24: 196– 207. Fiema, Zbigniew T., Ahmad Al-Jallad, Michael C.A. Macdonald, and Laïla Nehmé.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

39

2015. “Provincia Arabia: Nabataea, the Emergence of Arabic as a Written Language, and Graeco-Arabica.” In Arabs and Empires before Islam, edited by Greg Fisher, 373–433. Oxford: Oxford University Press. Gawlikowski, Michał. 2006. “Arabes et Arabies Dans L’antiquité.” Topoi 14: 41–46. Gzella, Holger. 2015. A Cultural History of Aramaic: From the Beginnings to the Advent of Islam. Hanbuch der Orientalistik 111. Leiden: Brill. Hasselbach, Rebecca. 2007. “Demonstratives in Semitic.” Journal of the American Oriental Society 127: 1–27. Hayajneh, Hani. 2009. “Ancient North Arabian-Nabataean Bilingual Inscriptions from Southern Jordan.” Proceedings of the Seminar for Arabian Studies 39: 203–22. ———. 2011. “Ancient North Arabian.” In The Semitic Languages: An International Handbook, edited by Stephan Weninger, Geoffrey Khan, Michael P. Streck, and Janet C.E. Watson, 756–781. Handbücher zur Sprach- und Kommunikationswissenschaft 36. Boston, Berlin: De Gruyter Mouton. Hayajneh, Hani, Mohammad I. Ababneh, and Fawzi Khraysheh. 2015. “Die Götter von Ammon, Moab und Edom in einer neuen frühnordarabischen Inschrift aus Südost-Jordanien.” In Fünftes Treffen der Arbeitsgemeinschaft Semitistik in der Deutschen Morgenländischen Gesellschaft vom 15.-17. Februar 2012 an der Universität Basel, edited by Viktor Golinets, Hanna Jenni, Hans-Peter Mathys and Samuel Sarasin, 79–105. Alter Orient und Altes Testament 425. Münster: Ugarit-Verlag. 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 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 P. Streck, and Janet C.E. Watson, 259–78. Handbücher zur Sprach- und Kommunikationswissenschaft 36. Boston, Berlin: De Gruyter Mouton. Jaussen, A., and M. Savignac. 1909. Mission Archéologique En Arabie. 6 vols. Paris: Leroux/Ernest Paul Geuthner. King, Geraldine M. H. 1990. Early North Arabian Hismaic. PhD Dissertation. London: University of London, School of Oriental and African Studies. Knauf, Ernst Axel. 1991. “More Notes on Ǧabal Qurma, Minaeans and Safaites.” Zeitschrift des Deutschen Palästina-Vereins 107: 92–101. ———. 2010. “Arabo-Aramaic and ʿArabiyya: From Ancient Arabic to Early Standard Arabic, 200 CE-600 CE.” In The Qurʾān in Context: Historical and Literary Investigations into the Qurʾānic Milieu, edited by Angelika Neuwirth, Nicolai Sinai, and

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

40

Ahmad Al-Jallad

Michael Marx, 197–254. Texts and Studies on the Qurʾān 6. Leiden-Boston: Brill. Kogan, Leonid. 2011. “Proto-Semitic Phonetics and Phonology.” In The Semitic Languages: An International Handbook, edited by Stephan Weninger, Geoffrey Khan, Michael P. Streck, and Janet C.E. Watson, 54–151. Handbücher zur Sprach- und Kommunikationswissenschaft 36. Boston, Berlin: De Gruyter Mouton. Kootstra, Fokelien. 2016. “The Language of the Taymanitic Inscriptions and Its Classification.” Arabian Epigraphic Notes 2: 67–140. ———. Forthcoming. Language Variation at Dadan. PhD Disserstation, Leiden University. Leiden: Leiden University. Kozah, Mario. 2014. “Introduction.” In The Syriac Writers of Qatar in the Seventh Century, edited by Mario Kozah, Abdulrahim Abu-Husayn, Saif S. Al-Murikhi, and Haya Al-Thani, 1–22. Gorgias Eastern Christian Studies 38. Piscataway, NJ: Gorgias Press. ———. Forthcoming. “The Arabian Oryx or Unicorn in Qaṭrayith: The Ancient East-Arabian Language of Beth Qaṭraye.” Arabian Epigraphic Notes. Le Quellec, Jean-Luc, F. Duquesnoy, V. Charpentier, and A. al-Mashani. 2018. “The Ẓufār painted inscriptions in Oman: Epigraphy and New Technologies”. Arabian Epigraphic Notes 4: 53–68. Lundberg, Johan. 2015. “Prepositional Phrases in the Dadanitic Inscriptions.” Arabian Epigraphic Notes 1: 123–38. Macdonald, Michael C.A. 1999. “Personal Names in the Nabataean Realm: A Review Article.” Journal of Semitic Studies 44: 251–89. ———. 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 Roger D. Woodard, 488–533. Cambridge: Cambridge University Press. ———.2008. “Old Arabic (Epigraphic).” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, Mushira Eid, Alaa Elgibali, Manfred Woidich and Andrzej Zaborski.Vol. 3, 464–477. Leiden: Brill. ———. 2009. “Arabs, Arabias, and Arabic before Late Antiquity.” Topoi 16: 277–332. ———. 2010. “Ancient Arabia and the Written Word.” In The Development of Arabic as a Written Language, edited by Michael C.A Macdonald, 5–27. Supplement to the Proceedings of the Seminar for Arabian Studies 40. Oxford: Archaeopress. ———. 2015. “On the Uses of Writing in Ancient Arabia and the Role of Palaeography in Studying Them.” Arabian Epigraphic Notes 1: 1–50. Macdonald, MichaelC.A, M. Al Muʾazzin, and L. Nehmé. 1996. “Les inscriptions safaïtic de Syrie, cent quarante ans après leur découverte.” Comptes rendus de séances de l’Académie des Inscriptions et Belles-Lettres, 435–494. Macdonald, Michael C.A. and Geraldine M.H. King. 2000. “Thamudic”, in: Encyclo-

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

41

paedia of Islam, second edition, edited by: P. Bearman, Th. Bianquis, C.E. Bosworth, E. van Donzel, W. P. Heinrichs, 436–438. Leiden: Brill. Mascitelli, Daniele. 2006. L’Arabo in epoca preislamica: formazione di una lingua. Roma: L’Erma di Bretschneider. Miles, G.C. 1948. “Early Islamic Inscriptions near Ṭā’if in the Ḥijāz.” Journal of Near Eastern Studies 7: 236–42. Müller, Walter.W. 1982. “Das Altarabische der Inschriften aus vorislamischer Zeit..” In Grundriß der Arabischen Philologie. Band I: Sprachwissenschaft, edited by Wolfdietrich Fischer, 30–36. Wiesbaden: Reichert. Negev, Avraham. 1991. Personal Names in the Nabatean Realm. Qedem 32. Jerusalem: Institute of Archaeology, Hebrew University of Jerusalem. Overlaet, Bruno, Michael C.A. Macdonald, and Peter Stein. 2016. “An Aramaic−Hasaitic Bilingual Inscription from a Monumental Tomb at Mleiha, Sharjah, UAE.” Arabian Archaeology and Epigraphy 27: 127–42. Pat-El, Na’ama. 2009. “The Development of the Semitic Definite Article: A Syntactic Approach.” Journal of Semitic Studies 54: 19–50. 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 Michael C. A Macdonald, 53–70. Supplement to the Proceedings of the Seminar for Arabian Studies 48. Oxford: Archaeopress. Rabin, Chaim. 1951. Ancient West-Arabian. London: Taylor’s Foreign Press. Rabinowitz, Isaac. 1956. “Aramaic Inscriptions of the Fifth Century B.C.E. from a North-Arab Shrine in Egypt.” Journal of Near Eastern Studies 15: 1–9. Ratcliffe, Robert R. 1998. The “Broken” Plural Problem in Arabic and Comparative Semitic: Allomorphy and Analogy in Non-Concatenative Morphology. Current Issues in Linguistic Theory 168. Amsterdam; Philadelphia: John Benjamins. Retsö, Jan. 2013. “What Is Arabic?” In The Oxford Handbook of Arabic Linguistics, edited by Jonathan Owens, 433–50. Oxford: Oxford University Press. Robin, Christian J. 1994. “Documents de l’Arabie Antique, III.” Raydān 6: 60–90. ———. 2001. “Les Inscriptions de L’arabie Antique et Les Études Arabes.” Arabica 48: 509–77. ———. 2007. “Ḥimyaritic.” inEncyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, Mushira Eid, Alaa Elgibali, Manfred Woidich and Andrzej Zaborski.Vol. 2, 256–261.. Leiden: Brill. Robin, Christian J., and Maria Gorea. 2016. “L’alphabet de Ḥimà (Arabie Séoudite).” In Alphabets, Texts and Artifacts in the Ancient Near East: Studies Presented to Benjamin Sass, edited by Israel Finkelstein, Christian J. Robin, and Thomas Römer, 310–75. Paris: Van Dieren Éditeur. Robin, Christian J., and Alessia Prioletta. 2013. “Nouveaux arguments en faveur d’une identification de la cité de Gerrha avec le royaume de Hagar (Arabie Orientale).”

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

42

Ahmad Al-Jallad

Semitica et Classica 6: 131–85. Rossi, Irene. 2014. “Minaeans beyond Maʿīn.” In Languages of Southern Arabia, edited by Orhan Elmaz and Janet C.E. Watson, 111–123. Supplement to the Proceedings of the Seminar for Arabian Studies 44. Oxford: Archaeopress. Rubin, Aaron D. 2005. Studies in Semitic Grammaticalization. Winona Lake, Ind.: Eisenbrauns. ———. 2008. “The Subgrouping of the Semitic Languages.” Language and Linguistics Compass 2: 61–84. Ryckmans, Gonzague. 1950. Corpus Inscriptionum Semiticarum: Pars Quinta, Inscriptiones Saracenicae Continens: Tomus I, Fasciculus I, Inscriptiones Safaiticae. Paris: E Reipublicae Typographeo. ———. 1956. “Aspects nouveaux du problème Thamoudéen.” Studia Islamica 5: 5–17. Sass, Benjamin. 1991. Studia Alphabetica: On the Origin and Early History of the Northwest Semitic, South Semitic, and Greek Alphabets. Frieburg, Schweiz; Göttingen: Universitätsverlag ; Vandenhoeck & Ruprecht. Shdeifat, Younis M. 2003. “Naqš Ṣafawiyy min ǧabal ʿAnāzah fī šamāl šarq Al-Urdunn: ‘iʿādah,qirāʾah wa-taḥlīl.’” Maǧallat Muʾtah li-l-Buḥūṯ wa-l-Dirāsāt 18: 213–29. Sima, Alexander. 1999. Die lihyanischen Inschriften von al-ʻUd̲ayb (Saudi-Arabien). Rahden, Westfalen: Leidorf. ———. 2002. “Die Hasaitischen Inschriften.” In Neue Beiträge zur Semitistik. Erstes Arbeitstreffen der Arbeitsgemeinschaft Semitistik in der Deutschen Morgenländischen Gesellschaft vom 11. bis 13. September 2000 an Der Friedrich-Schiller-Universität Jena, edited by Norbert Nebes, 167–200. Jenaer Beiträge Zum Vorderen Orient 5. Wiesbaden: Harrassowitz. Stein, Peter. 2004. “Zur Dialektgeographie des Sabäischen.” Journal of Semitic Studies 49: 225–45. ———. 2008. “The ‘Himyaritic’ Language in Pre-Islamic Yemen.” Semitica et Classica 1: 203–12. ———. 2013. “Palaeography of the Ancient South Arabian Script. New Evidence for an Absolute Chronology.” Arabian Archaeology and Epigraphy 24: 186–95. ———. 2017. “South Arabian Zabūr Script in the Gulf: Some Recent Discoveries from Mleiha (Sharjah, UAE).” Arabian Archaeology and Epigraphy 28: 110–24. ———. 2018. “The role of Aramaic on the Arabian Peninsula in the second half of the first millennium BC”. In Languages, scripts and their uses in ancient North Arabia, edited by Michael C.A. Macdonald, 39-49. Supplement to the Proceedings of the Seminar for Arabian Studies 48. Oxford: Archaeopress. Stein, Peter and Ingo Kottsieper. 2014. “Sabaic and Aramaic - a common origin?”. In Languages of Southern Arabia, edited by Orhan Elmaz and Janet C.E. Watson, 81-87. Supplement to the Proceedings of the Seminar for Arabian Studies 44. Oxford: Archaeopress.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

What is Ancient North Arabian?

43

Suchard, Benjamin. 2016. “The Hebrew Verbal Paradigm of Hollow Roots: A Triconsonantal Account.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 166: 317–322. Tropper, Joseph. 2001. “Die Herausbildung des bestimmten Artikels im Semitischen.” Journal of Semitic Studies 46: 1–31. Tsafrir, Nurit. 1996. “New Thamudic Inscriptions from the Negev.” Le Muséon 109: 79–93. Van Putten, Marijn. 2017. “The Development of the Triphthongs in Quranic and Classical Arabic.” Arabian Epigraphic Notes 3: 47–74. Voigt, Rainer M. 1988. Die infirmen Verbaltypen des Arabischen und das Biradikalismus-Problem. Stuttgart: Steiner. Watson, Janet C.E. 2011a. “Arabic Dialects (General Article).” In The Semitic Languages: An International Handbook, edited by Stephan Weninger, Geoffrey Khan, Michael P. Streck, and Janet C.E. Watson, 851–96. Handbücher zur Sprach- und Kommunikationswissenschaft 36. Boston-Berlin: De Gruyter Mouton. ———. 2011b. “South Arabian and Yemeni Dialects.” Salford Working Papers in Linguistics and Applied Linguistics 1: 27–40. ———. 2014. “Southern Semitic and Arabic Dialects of the South-Western Arabian Peninsula.” In Languages of Southern Arabia, edited by Orhan Elmaz and Janet C.E. Watson, 147–153. Supplement to the Proceedings of the Seminar for Arabian Studies 44. Oxford: Archaeopress Watson, Janet C.E., Bonnie G. Stalls, Khalid al-Razihi, and Shelagh Weir. 2006. “The Language of Jabal Rāziḥ: Arabic or Something Else?” Proceedings of the Seminar for Arabian Studies 36: 35–41. Winnett, Fidelity V. 1987. “Studies in Ancient North Arabian.” Journal of the American Oriental Society 107: 239–44. Winnett, Frederick V. No date. Studies in Thamudic. Self-published. ———. 1937. A Study of the Liḥyanite and Thamudic Inscriptions. University of Toronto Oriental Series 3. Toronto: University of Toronto Press. Winnett, Frederick V., and William L.Reed. 1970. Ancient Records from North Arabia. Near and Middle East 6. Toronto: University of Toronto Press. ———. “An Archaeological Epigraphical Survey of the Haʾil Area of Northern Saudi Arabia.” Berytus 22: 53–100. 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.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic1 Daniel Birnstiel, Goethe-University Frankfurt

1. Introduction In this paper I analyse the semantics of of mubīn and etymologically related lexemes in the Qur’ān. The term mubīn, often translated as “clear” or “plain”, occurs as an attribute of lisān ʿarabī “Arabic tongue/language” in two Qur’ānic passages (16:103; 26:195) that have excited much comment from both Muslim scholars and Western academics. In addition, mubīn also occurs several times as an attribute of several apparently self-referential terms designating the Qur’ān or parts thereof, i.e. (al-)kitāb and qurʾān.2 Based on a thorough analysis of the occurrences of mubīn and related lexemes I conclude that the basic connotation of the root byn is “identifiability” with mubīn denoting “identifiably, identifiable”, and not “being or making clear”, as has been claimed. 1.1 Structure of the paper Following a break-down of the various forms of the root byn in the Qur’ān, I present a survey of common interpretations of mubīn in certain contexts and I raise objections against these interpretations, show why they are unlikely to be correct. Sections (3–8) discuss the meanings of these forms as they arise from a contextual analysis of their respective loci. The last section, which is followed by an appendix listing all attestations of mubīn, summarizes the findings. 1 This contribution is the extensive revision, refinement and completion of a paper originally commenced and partly drafted during my time as research fellow at the Cambridge Muslim College (2009–2011). Parts of its results were originally presented in September 2014 at a conference in Frankfurt. It has greatly benefited from discussions with and feedback by Judith Birnstiel, Michael Marx, George Warner, Timothy Winter and Maiko Yamaji. The insights of Kropp (2015) partially overlap with my own findings; however, he differs with regard to the interpretation of lisān ʿarabī mubīn; in addition, I attempt to determine the distinctions between the different verb stem derivations in a more refined way. 2 Other terms belonging to the related semantic field of prophets and prophecy such as balāġ (communication), naḏīr (warner), rasūl (messenger), etc. are likewise characterized as mubīn.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

46

Daniel Birnstiel

1.2 The root byn: attested forms The root byn is altogether attested 524 times.3 These attestations include 267 occurrences of the preposition bayna “between”, which is however irrelevant for this discussion. Forms of stem II (bayyana) occur 42 times (perfect 5 times, imperfect indicative 17 times, subjunctive 9 times, energetic twice, jussive 3 times, active participle 6 times), while stem IV (ʾabana) is attested 120 times (active participle, i.e. mubīn, 119 times, imperfect once); stem V (tabayyana) occurs 18 times (perfect 15 times, subjunctive 3 times); stem X (istabāna) is found twice (1 subjunctive, 1 active participle). Stem II Perfect

Stem IV Stem V

5

Imperfect

Indicative

16

12

Totals 17

1

Subjunctive 9

17 3

1

13

Energetic

2

2

Jussive

3

3

Imperative Act. ptcp.

Stem X

3 6

119

Totals

41

120

3 1

126

18

2

181

To this must be added 76 attestations of diverse nominal forms, namely bayyinah (19 times) and its plural bayyināt (52 times), bayān (thrice), tibyān (one attestation) and bayyin (one attestation).

3 These numbers and the following statistics referring to the different derivations of byn are based on the reading (qirāʾah) of Ḥafṣ.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

47

2. Common interpretations of qurʾān ʿarabī mubīn and similar expressions As mentioned above, mubīn occurs in the Qur’ān as an attribute of lisān ʿarabī “Arabic tongue/language”. Bobzin in the glossary of his translation of the Qur’ān (2010: 786) explains ʿarabī “Arabic” in the glossary to his translation of the Qur’ān as “the language (lisān), in which the Qur’ān is written, and which is labeled in the Qur’ān itself often as ‘plain’, ‘clear’ (mubīn)”; similarly, Reynolds (2004: 677) calls the the “Qur’ān’s own claim to be “clear” or “plain” (mubīn)” the “standard understanding”. According to two frequently encountered, mutually non-exclusive views, mubīn should be understood as “easily comprehensible” and/or “explaining, elucidating matters”: “The word mubīn means both “clear in itself” and “that which clarifies (something else).” (Mir 2006: 88). Thus, the Qur’ān claims both that it yields its meaning unambiguously and that it elucidates matters, dispelling doubt and eliminating error. One must keep both meanings of mubīn in mind for a correct understanding of the Qur’ān’s claim to be “clear”.” Carter (2006: 120) likewise believes that in the context of such passages “mubīn has to be understood as “making or being clear…”.” 2.1 mubīn denoting “clear, easily comprehensible” The idea of mubīn denoting “unambiguity in meaning” is frequently encountered in Muslim writings, e.g. in al-Kafāwī’s Kullīyāt (1419/1998: 723), where he explains the Qur’ān having been revealed in lisān ʿarabī mubīn to mean “ʾannahū ʾunzila bi-lisānin lā yaḫfā maʿnāhu ʿalā ʾaḥadin mina l-ʿarabi walam yustaʿmal fīhā luġatun lam yatakallami l-ʿarabu bihā” [“that it was revealed in a language the meaning of which was obvious to (literally: not hidden from) any Arab and in which no idiom was employed that the Arabs do not use in their speech”].4 Similar explanations are often found in Qur’ānic 4 Note that this and similar explanations assume that the pre-Islamic Arabs were “a people of poets, where everyone, from the most noble sayyid to the outcast brigand, was if not himself a poet then at least a connoisseur and aficionado of poetry” [Ambros (1998: 167)]. This assumption, as pointed out by Ambros, is however not corroborated by hard evidence outside the Islamic tradition; see Ambros (1998) for further details. Furthermore, the ignorance regarding certain Qur’ānic expressions ascribed to Muḥammad’s contemporaries in Islamic literature indicates that generalizing claims of “a language the meaning of which was obvious to any Arab and in which no idiom was employed that the Arabs did not use in their speech” [al-Kafawī (1419/1998: 723) are to be regarded as literary topoi and should be treated as such. Cf. also Wansbrough (2004: 79).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

48

Daniel Birnstiel

commentaries in the elucidation of the above-mentioned passages; thus ʾAbū Suʿūd (d. 1575) glosses in his commentary (n.d.: iv.232) bi-lisānin ʿarabīyin mubīnin in 26:195 as “wāḍiḥi l-maʿnā ẓāhiri l-madlūli li-ʾallā yabqā lahum ʿuḏrun mā” [“(in clear Arabic language), of apparent meaning and evident sense, so that no excuse whatsoever (to reject the message) would be left to them (= the Qurayš)”]. Unsurprisingly, the interpretation of mubīn as “comprehensible without difficulty” is a view readily adopted and employed by adherents of theological schools inclined towards literalism. Thus the Radd ʿalā l-Ğahmīyah wa-z-Zanādiqah ascribed to Aḥmad b. Ḥanbal (2002: 118) asserts after listing a number of Qur’ānic verses dealing with God’s divine speech (kalām) and his addressing human beings, especially his talking to Moses “directly” in 4:164 “fa-hāḏihī nuṣūṣun bi-lisānin ʿarabīyin mubīnin lā yaḥtāḡu ʾilā tafsīrin huwa mubīnun bi-ḥamdi llāhi” [“These are passages in clear Arabic which has no need for interpretation for it is clear by the praise of God.”]. Likewise, the son of the eighteenth century Nağdī reformer and founder of the Wahhabist movement ʿAbd Allāh b. Muḥammad b. ʿAbd al-Wahhāb addresses the problem of refraining from dealing with the modality of God’s establishment upon the throne (al-istiwāʾ ʿalā l-ʿarš) and similar issues by having recourse to the interpretation of mubīn as “unambiguous in meaning” (Muḥammad b. ʿAbd al-Wahhāb & ʿAbd Allāh b. Muḥammad b. ʿAbd al-Wahhāb 1981: 129): “ʾammā muḫālalafatu luġati l-ʿarabi fa-lā yağūzu laka ʾan tuḫālifahā wa-tufassira kitāba llāhi ʿazza wa-ʿalā bi-ġayrihā li-muḫālalafatika li-mā ʾanzala llāhu fīhi wa-qad qāla taʿālā “nazala bihi r-rūḥu l-ʾamīnu ʿalā qalbika li-takūna mina l-munḏirīna bi-lisānin ʿarabīyin mubīnin” wa-qāla taʿālā “ḥ’m wa-l-kitābi l-mubīni ʾinnā ğaʿalnāhu qurʾānan ʿarabīyan laʿallakum taʿqilūna” ʾilā ġayri ḏālika mina l-ʾāyāti fa-hal yağūzu laka ʾan taqūla stawā bi-lā kayfin baʿda ʾan qāla mubīn” [“As for contradicting the language of the Arabs, you are not permitted to contradict it and to interpret the book of God (high and mighty) by other than this language because of your contradicting that which God has revealed therein. For the sublime has said: ‘The trusted spirit brought it down, upon your heart so that you would be among the warners, in clear Arabic speech’ (26:193–195). And the sublime has said: ‘By the clear book. We have made it a Qur’an in Arabic so that you may understand’ (12:1–2, 43:2–3) and similar verses. So how can you be permitted to say ‘He established himself without how’ after (God) has said “clear”?”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

49

2.2 lisān ʿarabī mubīn and the nature of Qurʾānic Arabic Based mainly on the idea of “clarity of meaning”, the interpretation of lisān ʿarabī mubīn has also entered into the discussion surrounding the nature of this “clear Arabic language”. Vollers (1906), as is well known, opted for a vernacular, dialectical form of Arabic, whereas other scholars regarded the Qur’ān’s language as (nearly) identical with the supra tribal language of the poets. Indeed, Zwettler (1978: 186f., n. 151) goes so far as to claim that “mubīn and bayān are, possibly, quasi-technical allusions to the poetic koine.” [Zammit (2002: 37)], while Corriente suggests that mubīn means “that Qur’ānic Arabic falls ‘between’ … the Old Arabic koine and Eastern bedouin Arabic on the one hand and Middle Arabic on the other” (ibid. 37f.).5 A somewhat different approach to the question is adopted by Ambros (1998: 170, passim), who claims that at the time of the Qur’ān’s emergence Nabatean Aramaic, Thamudic, Old South Arabian as well as possibly Greek and Middle Persian in certain circumstances, but not Arabic (in the sense of al-ʿarabīya or classical Arabic) were the languages of prestige. He asserts (171) that being clear and comprehensible without difficulty does generally not have an awe-inspiring and fascinating effect and suggests that the expression lisān ʿarabī mubīn “clearly comprehensible Arabic language” should be understood as having the same emotional value as expressions such as “auf gut Deutsch” or “in plain English” and consequently stressing the functional and communicative merits of a language with low prestige value which should be used without feelings of inferiority (171f.). Zaehner (1975: 170) seems to follow a similar view when he states that the Qur’ān “repeatedly speaks of the ‘Book’ (one of the stock names for the Koran itself) as being mubīn, ‘clear, making’ clear, almost ‘self-explanatory’ because it was manifested in plain Arabic speech, in the local vernacular, that is, and not in a sacred language that could not be understood.” Native philologists and Muslim theologians and commentators, however, have regarded the “clear Arabic language” always as being “grammatically correct” and “literary” (Classical) Arabic and tended to treat passages such as 16:103 and 26:195 as proofs for both the dogma of the inimitability of the Qur’ān and the idea that its language was “pure Arabic”, i.e. free from any foreign words. The Javanese nineteenth century scholar Muḥammad Nawawī 5 Donner (2008: 36) lists several difficulties that arise from the assumption that the Qur’ān’s language is identical to the language of ancient Arabic poetry. Note also the recent literature regarding the nature of Qurʾānic Arabic which shows that it was not identical to Classical Arabic (Birnstiel forthcoming; Van Putten 2018; Van Putten & Stokes 2018).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

50

Daniel Birnstiel

e.g. explains 16:103 (kalāmu llaḏī yulḥidūna ʾilayhi ʾaʿǧamīyun wa-hāḏā lisānun ʿarabīyun mubīn “The speech of the one they allude to is barbaric,6 and this is clear Arabic language” as follows (1417/1996: i.607): “ʾay kalāmu llaḏī yansubūna ʾilayhi ʿibrānīyun lam yatakallam bi-l-ʿarabīyati wa-lam yaʾti bi-faṣīḥi l-kalāmi wa-hāḏā l-qurʾānu kalāmun ʿarabīyun ḏū bayānin wa-faṣāḥatin fa-kayfa yuʿallimu muḥammadan wa-huwa ğāʾakum bi-hāḏā l-qurʾāni l-faṣīḥi llaḏī ʿağaztum ʿanhu wa-ʾantum ʾahlu l-faṣāḥati fa-kayfa yaqduru man huwa ʾaʿğamīyun ʿalā miṯli hāḏā l-qurʾāni wa-ʾayna faṣāḥatu hāḏā l-qurʾāni min ʿuğmati hāḏā llaḏī tušīrūna ʾilayhi” [“This means the speech of the one they refer to is Hebrew; he does not speak Arabic nor does he bring clear, eloquent speech. However, this Qur’ān is Arabic speech, characterized by clarity of expression and eloquence. So how can he teach Muḥammad, when the latter brought you this clear, eloquent Qur’ān that you were unable to imitate, even though you are the people of eloquence? So how can someone who is a non-Arab produce something like this Qur’ān? And what a difference lies between the clear eloquence of this Qur’ān and the barbaric way of expression of this person that you allude to!”]. A very similar explanation also is given by al-Ḫāzin (1415/1994: iii.99), who moreover glosses lisān ʿarabī mubīn as bayyin(i) l-faṣāḥati wa-l-balāġati, i.e. characterized by manifest purity of speech and eloquence.7 As noted above, these loci, especially 16:103, were also taken by a large number of Muslim scholars as a Qur’ānic proof that it could not contain loanwords, even though this claim is patently wrong from a linguistic point of view.8 Thus ninth century grammarian ʾAbū ʿUbaydah (1981/1401: 17) claimed: “nazala l-qurʾānu bi-lisānin ʿarabīyin mubīnin fa-man zaʿama ʾanna fīhi ġayra l-ʿarabīyati fa-qad ʾaʿẓama l-qawla wa-man zaʿama ʾanna ṭā-hā bi-n-nabaṭīyati fa-qad ʾakbara wa-ʾin lam yaʿlam mā huwa fa-huwa ftitāḥu kalāmin wa-huwa smun li-s-sūrati wa-šiʿārun lahā wa-qad yuwāfiqu l-lafẓu l-lafẓa wa-yuqāribuhū wa-maʿnāhumā wāḥidun wa-ʾaḥaduhumā bi-l-ʿarabīyati wa-l-ʾāḫaru bi-l-fārisīyati wa-ġayrihā” [“The Qur’ān has come down in clear Arabic language. Consequently, whosoever claims that it includes non-Arabic (words) has made an arrogant presumption, and whosoever claims that ‘ṭā-hā’ [the beginning of surah 20] is a Nabatean expression, has made it even worse. If he does not know what this is – it is the opening of a 6 The exact meaning of ʾaʿǧamī “barbaric” is subject to controversy. 7 Cf. also Ibn ʿĀšur (1404/1984: xiii.95), Ibn al-ʿArabī (1424/2003: iv.21), as well as asSuyūṭī’s chapter on ʾiʿğāz in his ʾitqān (2005/1426: v.1873–1905). See also an-Nasafī (1419/1998: ii.554) on verse 26:2. 8 Cf. Carter (2006) with further literature, Jeffery (1938), Zammit (2002, especially 51– 61).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

51

speech and a name for the surah and a distinguishing mark for it. However, one expression may agree with another one and approximate it, their meanings being the same, while one of them is in Arabic and the other in Persian or another (language)”].9 The idea that the resemblance in sound and meaning between words in Arabic and other languages is merely coincidental is also supported by aṭṬabarī (1422/2001: i.15), while according to as-Suyūṭī (2005/1426: iii.936) other scholars suggested that “all these expression (i.e. the expressions that appear to be foreign) are unadulterated Arabic, but the Arabic language is very extensive, and it is not unlikely that even the venerable elders are unaware (of all its expressions)”, adducing Ibn ʿAbbās’ ignorance of the meanings of fāṭir (6:14) and fātiḥ (in reference to 7:89) and aš-Šāfiʿī’s assertion that only a prophet fully comprehends a language as proofs.10 2.3 mubīn denoting “clarifying, elucidating matters” Although, as noted above, some Western as well as Muslim scholars have advanced the idea that mubīn in said context also means “to make clear, to clarify”,11 this view is not usually found in the commentaries on 16:103 and 26:195. It is, however, quite common in elucidations of the phrase al-kitāb al-mubīn (“the clear book/scripture”). E.g. aṣ-Ṣanʿānī (1410/1989: i.317) on 12:1 explains “verses of the clear book” as “bayyana llāhu taʿālā rušdahū wa-hudāhu” [“God the Almighty made clear his right path and his guidance”], an opinion that is also mentioned by aṭ-Ṭabarī (1422/2001: xiii.5f.). Similarly, Tafsīr al-Ğalālayn [al-Maḥallī & as-Suyūṭī (n.d.: 235] glosses al-mubīn as “al-muẓhir(i) li-l-ḥaqqi mina l-bāṭili” [“which makes visible truth and distinguishes it from falsehood”]. Aṭ-Ṭabarī (1422/2001: xiii.5f.) similarly explain that the book “makes clear the matters which are permitted and prohibited”.12 A related, though somewhat different explanation is given by az-Zamaḫšarī (1418/1998: iii.250): “ʾay tilka l-ʾāyātu llatī ʾunzilat ʾilayka fī hāḏihi s-sūrati ʾāyātu s-sūrati ẓ-ẓāhiri ʾamruhā fī ʾiʿğāzi l-ʿarabi wa-tabkītihim ʾawi llatī 9 Cf. also Al-Ġazālī (1413/1993: 84). For a general overview of the different positions adopted by Muslim scholars, see Carter (2006: 122ff.), Jefferey (4–11), Kopf (1956: 40–45; 1961) and Reynolds (2004: 677f.). 10 Note the contradiction between linguistic ignorance ascribed to some Companions and the assertion that complete linguistic comprehension can only be ascribed to a prophet on the one hand and the characterization of lisān ʿarabī mubīn as “a language the meaning of which was obvious to any Arab”. 11 Cf. e.g. the meanings offered in the dictionaries of Ambros & Procházka (2004: 46f.) and Badawi & Abdel Haleem (2008: 123–126). 12 Cf. also Al-Māwardī (n.d.: iii.5) and an-Nasafī (1419/1998: ii.554, 627).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

52

Daniel Birnstiel

tubayyinu/tubīnu li-man tadabbarahā ʾannahā min ʿindi llāhi lā min ʿindi l-bašari ʾawi l-wāḍiḥatu llatī lā taštabihu ʿalā l-ʿarabi li-nuzūlihā bi-lisānihim ʾaw qad ʾubīna fīhā mā saʾalta ʿanhu l-yahūdu min qiṣṣati yūsufa” [“This means: These verses, which were revealed to you in this surah are verses of the surah whose concern regarding the incapacitation of the Arabs (to produce something like it) and their reproach is obvious; or which make it clear to anyone who contemplates them that they are from God, not from a human being; or the clear (verses), the meanings of which are not obscure to the Arabs due to their being revealed in their language; or in which the things you asked the Jews regarding the story of Joseph have been made clear”]. The sheer bulkiness of the information that needs to be mentally supplied by the audience in order to arrive at such an interpretation as well as the fact that mubīn is never attested with an object show that such a reading hardly corresponds to the originally intended meaning of the text. Furthermore, such an interpretation is only feasible in the context of a community which has already accepted the Qur’ān as a scripture that clarifies permitted and prohibited matters; this does not harmonize with the scenario depicted by the Qur’ān itself, namely that of a scripture directed at people that are either disbelieving the new, allegedly divine message or a community in statu nascendi, which therefore must continuously be told and reminded of the text’s divine origin. Additionally, such generalizing interpretations rarely fit the context of the respective passages. Contrary to most Muslim as well as Western Qur’ānic scholars, Gilliot (2008: 94f.), thinks that the understanding of lisān ʿarabī mubīn in 16:103; 26:195 as “pure Arabic” (in the sense of an unadulterated language) is incorrect and opts instead for an interpretation as “making [things] clear.” 2.4 Objections to these interpretations There are several problems arising from these interpretations. Firstly, even if the Qur’ān is a single coherent and unified corpus, its language is in many cases far from being easily comprehensible. Despite the obvious attempts of Islamic theologians to render certain dogmas unassailable by invoking the “clarity of the language”, the Qur’ān itself points out in surah 3:7 that it consists of clear and unambiguous verses (muḥkamāt) and ambiguous or parabolical verses (mutašābihāt). The fact that both the intended meaning of the term mutašābihāt as well as the identity of these verses have remained a controversial issue as evidenced by the various interpretations found in the exegetical

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

53

literature on 3:7 proves that understanding mubīn to signify “comprehensible without difficulty” is hardly compatible with the facts.13 Mir (2008: 89) attempts to counter such arguments by asserting that “the claim of any book to be clear does not necessarily mean that all its readers, regardless of their backgrounds – that is, their age, experience, mental acumen, level of knowledge, and linguistic ability – will understand it equally well or fully” and that “a document will be called “clear” if it treats its subject in language that is clear relative to that subject”. Given the far-spread disagreement regarding many issues among the Qur’ānic exegetes, whom one may consider possessing similar levels of experience, knowledge, mental acumen, etc. when compared with the laity, his objections do not seem convincing. Secondly, as pointed out above, the Qur’ān includes more than only a few words of ultimately non-Arabic origin. However, the earliest exegetes do not seem to have regarded their existence as problematic; see also Carter (2006: 120). Their understanding of “clear Arabic” can therefore not yet have been concerned with the purity and eloquence of language. Thirdly, mubīn occurs also as an attribute of nouns that are completely unrelated to the semantic fields of prophecy, revelation, scripture or language, yet these occurrences are usually left out completely in discussions regarding the meaning of mubīn. As will be argued below, it is impossible that the term could denote “easily comprehensible” and/or “explaining, elucidating matters” in these cases. It seems reasonable to assume that mubīn should have a relatively unified meaning regardless of what nouns it modifies. Fourthly, although mubīn is an active participle of stem IV, which is usually said to be causative, the form occurs nowhere in the Qur’ān with any object that would explicitly name the matter clarified or elucidated. As a matter of fact, Muqātil14, whose tafsīr is one of the earliest exegetical works, glosses mubīn in almost every passage simply with bayyin “clear”, including passages in which it is found as an attribute of “(the) book”, e.g. 12:1, 27:1 (1423/2002: ii.318, iii.296).

13 As pointed out by Carter (2006: 120), “the Qur’ān contains at least five kinds of acknowledged obscurity. There are the famous “mysterious letters” at the beginning of twenty-nine sūras, there are the mutashābihāt, the “ambiguous verses”, there are the numerous “strange, rare” (gharīb) expressions, mostly native words, which puzzled the lexicographers, there are the textual variants (qirāʾāt), and there are the words which look and sound “foreign” (ʿajamī, aʿjamī).” 14 Cf. eg. Muqātil (1423/2002: i.155, 180, 311, 517, 550, ii.245, 252, 272, 278, 318, ii.331, 400, iii.118, 131, 185, 190, 193, 261, 272, 298, 299, 300, 316, 577, 653, 710, 800, iv.131, 148, 325, passim).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

54

Daniel Birnstiel

3. The semantics of mubīn 3.1 mubīn in context As mentioned above, mubīn is attested 119 times. It occurs always in verse-final position, always qualifying a definite or indefinite noun or else an adjective, and only in its masculine singular form. The most frequently attested among these are ḍalāl “error” (19 occurrences), sulṭān “authority” (12 occurrences), naḏīr “warner” (11 occurrences), ʿadūw “enemy” (10 occurrences), ṣiḥr “magic” (10 occurrences), kitāb “scripture” (12 occurrences, 7 times definite, 5 times indefinite) and balāġ “transmission, communication, report” (7 occurrences).15 The occurrences with ʿadūw include several examples that seem well-suited to illustrate the function of mubīn. In example (1), which occurs at the peak of the well-known episode narrating the fall of Adam (7:19–25), Satan succeeds in tempting Adam and his wife, who have been warned by God not to approach the tree (7:19): (1) fa-dallāhumā bi-ġurūrin fa-lammā ḏāqā š-šağarata badat lahumā sawʾātuhumā wa-ṭafiqā yaḫṣifāni ʿalayhimā min waraqi l-ğannati wa-nādāhumā rabbuhumā ʾa-lam ʾanhakumā ʿan tilkumā š-šağarati wa-ʾaqul lakumā ʾinna š-šayṭāna lakumā ʿadūwun mubīnun (7:22) “ And (Satan) brought them to fall through deception and when they had tasted (the fruit of) the tree, their shame became evident to them and they began (to cover themselves) by sewing together leaves of the garden, and their Lord called them: “Is it not the case that I forbade you from that tree (known to you) and that I told you: Behold, the devil is unto you an obvious enemy?”16 The context of the passage leaves no doubt that what is meant by “an obvious/ clear enemy”, namely an enemy who manifests or has manifested his enmity, someone whose enmity is clear, i.e. evident, obvious, or perceptible without 15 An examination of all loci in Al-Ḫaṭīb (1422/2002) shows that there seem to be no variant readings attested for any of these instances. A breakdown of the different nouns and adjectives qualified by mubīn is given in the appendix. 16 For this understanding of the demonstrative pronoun as “recognitional” and the rhetorical usage of the negation of the predicative relationship, see Birnstiel (2010: 37–48, 149, 214). For ʾinna functioning as particle presenting highly assertive statements that do not permit contradiction on the discourse see Bloch (1986: 113–117).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

55

difficulty. The text asserts that God has warned Adam and his spouse against the devil. This is apparently also the understanding of aṭ-Ṭabarī (1422/2001: iii.37) in his commentary on 2:168: “yaʿnī ʾannahū qad ʾabāna lakum ʿadāwatahū bi-ʾibāʾihī ʿani s-suğūdi li-ʾabīkum wa-ġurūrihī ʾiyyāhu… “[“this means that he has shown you his enmity by his refusal to prostrate himself to your father and by his deceiving him…”]. Similarly, al-Bayḍāwī (1418/1998: 118) glosses the phrase in 2:168 as “ẓāhiru l-ʿadāwati ʿinda ḏawī l-baṣīrati” [“(someone) whose enmity is evident among those endowed with insight”]: (2) yā ʾayyuhā n-nāsu kulū mim-mā fī l-ʾarḍi ḥalālan ṭayyiban wa-lā tattabiʿū ḫuṭuwāti š-šayṭāni ʾinnahū lakum ʿadūwun mubīnun (2:168) “O people! Eat from what is found on the earth, for it is lawful and good!17 But do not follow the footsteps of Satan! Behold, he is to you an obvious enemy!”18 The meaning of an “enemy who has manifested himself as such” is also present in the following three examples: (3) ʾa-lam ʾaʿhad ʾilaykum yā banī ʾādama ʾan lā taʿbudū š-šayṭāna ʾinnahū lakum ʿadūwun mubīnun (36:60) “Is it not the case, o sons of Adam, that I charged you not to serve Satan? Behold, he is unto you an obvious enemy!” (4) wa-qul li-ʿibādī yaqūlū llatī hiya ʾaḥsanu ʾinna š-šayṭāna yanzaġu baynahum ʾinna š-šayṭāna kāna li-l-ʾinsāni ʿadūwan mubīnān (17:53) “And tell my slaves to speak in the most beautiful manner. Behold, Satan stirs discord up among them! Behold, Satan has always been unto mankind an obvious enemy.”

17 I understand ḥalālan ṭayyiban as circumstantial expression, referring predicatively to the “what is found on the earth” to highlight the assertion that the things which are found thereupon are indeed wholesome. 18 Similar examples occur in 2:208 and 6:142. See also aṭ-Ṭabarī (1422/2001: iii.602) and al-Bayḍāwī (1418/1998: i.134) on 2:208, where they both offer explanations similar to those mentioned above. See also al-Bayḍāwī (ibid.: v.94) on 43:62 “ṯābitu ʿadāwatihī bi-ʾanna ʾaḫrağakum ʿani l-ğannati wa-ʿaraḍakum li-l-baliyyati” [“whose enmity has become established by the fact that he expelled you from the Garden and subjected you to tribulation”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

56

Daniel Birnstiel

(5) wa-ʾiḏā ḍarabtum fī l-ʾarḍi fa-laysa ʿalaykum ğunāḥun ʾan taqṣurū mina ṣ-ṣalāti ʾin ḫiftum ʾan yaftinakumu llaḏīna kafarū ʾinna l-kāfirīna kānū lakum ʿadūwan mubīnān (4:101) “And when you travel the earth, shortening prayers is not held against you, if you fear that those who disbelieve might attack you. Behold, the disbelievers have been unto you an obvious enemy.” In (3), God rhetorically asks humanity, if they had not been warned against the devil. The phrase occurs here in a sentence presenting the assertion that the devil is so obviously an enemy of mankind that mankind should have been able to perceive and recognize his enmity.19 In (4), Muḥammad is commanded to instruct his followers regarding the manners of speech. Here the phrase occurs in an assertion highlighting the devil’s history as an enemy of mankind who has always proven himself as such and can therefore be recognized as what he is. In (5), the last sentence presents a fact that ever since the emergence of Muḥammad as a prophet, the disbelieving Meccans had revealed themselves as an enemy of the believers. In one additional case, the phrase is expanded by an additional participle:20 (6) wa-daḫala l-madīnata ʿalā ḥīni ġaflatin min ʾahlihā fa-wağada fīhā rağulayni yaqtatilāni hāḏā min šīʿatihī wa-hāḏā min ʿadūwihī fa-staġāṯahu llaḏī min šīʿatihī ʿalā llaḏī min ʿadūwihī fa-wakazahū mūsā fa-qaḍā ʿalayhi qāla hāḏā min ʿamali š-šayṭāni ʾinnahū ʿadūwun muḍillun mubīnun (28:15) And (Moses) entered the city at a time when its people were unaware and he found therein two men fighting each other, one of his own people, the other of his enemy. And the one who was of his own people called out to (Moses) to help him against the one who was of his enemy, and Moses struck him and put an end to him. He said: “This is from the work of the devil. Behold, he is recognizably an obvious enemy!” These examples21 permit to draw several conclusions. Firstly, all examples occur in highly salient discourse peaks, presenting a strong assertion regarding the enmity of a referent vis-à-vis the addressee. Secondly, as the context shows, the presented assertion calls attention to the fact that the referent’s en19 Note that Muqātil (1423/2002: iii.583) glosses here ʿadūw mubīn as “bayyinu l-ʿadāwati” [“whose enmity is obvious”]. 20 Al-Bayḍāwi (1418/1998: iv.174) glosses this simply as 28:15 “zāhiru l-ʿadāwati” [“whose enmity is manifest”]. 21 Additional examples are 12:5 and 43:62.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

57

mity has either been shown previously or is so blatantly visible that it can easily be recognized. Similarly, mubīn occurs also with several nouns used to refer to prophets, especially Muḥammad, e.g. (an-)naḏīr “(the) warner” and rasūl “messenger”. In all these cases, mubīn highlights and marks the referent as being identifiably such a person, as can easily be seen in the following cases:22 (7) wa-la-qad ʾarsalnā nūḥan ʾilā qawmihī ʾinnī lakum naḏīrun mubīnun (11:25) And We had indeed sent Noah to his people (, saying): “Behold, I am to you an apparent warner!” (8) qul ʾinnamā l-ʿilmu ʿinda llāhi wa-ʾinnamā ʾana naḏīrun mubīnun (67:26) “Say: It is with God that the knowledge lies, while it is an apparent warner that I am.” In (7), the verse introduces a short narrative regarding Noah and his mission; the phrase occurs here in the opening sentences, in which Noah asserts that his having been sent to warn the people is evident and clearly discernible. As such, the verse functions to establish his authority as messenger and warner within the discourse; it is precisely the obviousness of his messengerhood that is rejected by the leadership of his people in 11:27, when they claim him to be only a man like they themselves. In (8), the verse occurs in the context of a question asked by the people and is used to (re-)establish Muḥammad’s authority by asserting that he is manifestly someone who has been commissioned to warn people but has not been given knowledge of certain matters pertaining to the end of time. In (9) and (10), the text asserts that someone who can easily be identified as messenger has arrived, i.e. somebody whose prophet- and messengerhood are evident. The Qur’ān highlights the identifiablility as messenger in order to contrast it with the negative reaction of the people mentioned in the immediately following verse in both cases. This depicts the non-believers as rejecting the truth despite its recognizable nature and as regarding the messenger as a sorcerer or a possessed person; in neither case is an understanding as “a messenger who clarifies things” justifiable by the context.23 22 Additional examples involving naḏīr “warner” occur in 7:184, 15:89, 22:49, 26:115, 29:50, 38:70, 46:9, 51:50, 51:51, 71:2. See also Sprenger (1869: ii.352 n.2). 23 Some of the glosses offered by aṭ-Ṭabarī and al-Bayḍāwī may also be understood in light of these findings, as is indicated by the adverbial clauses and adjuncts of their explanations. Thus aṭ-Ṭabarī (1422/2001: x.602f.) explains the phrase naḏīr mubīn in

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

58

Daniel Birnstiel

(9) bal mattaʿtu hāʾulāʾi wa-ʾābāʾahum ḥattā ğāʾahumu l-ḥaqqu wa-rasūlun mubīnun | wa-lammā ğāʾahumu l-ḥaqqu qālū hāḏā siḥrun wa-ʾinnā bihī kāfirūna (43:29–30) Nay! I have granted these people here and their fathers an enjoyment until there came to them the truth and an apparent messenger. Yet when the truth came to them, they said: “This is sorcery and we certainly reject that!” (10) ʾannā lahumu ḏ-ḏikrā wa-qad ğāʾahum rasūlun mubīnun | ṯumma tawallaw ʿanhu wa-qālū muʿallamun mağnūnun (44:13–14) What would a reminder avail them? for an apparent messenger had come to them, but then they turned away from him and said: “A tutored and possessed man!” Mubīn is attested as attribute of several other nouns that indicate agenthood or an animate referent possessing volitional faculty; it is also found twice with a noun indicating a referent from the animal kingdom. In all these examples, the context shows that the only way to understand mubīn is as an assertion that the quality indicated by the noun is question is obvious, manifest or perceivable. Thus in (11), the Qur’ān highlights man’s quarrelsomeness and impertinence as someone who despite his inferior status as creature confronts his creator and reveals himself as a clearly bellicose person. This is also the understanding of Muqātil as shown by his gloss (1423/2002: ii.586) “bayyinu l-ḫuṣūmati” [“whose contentiousness is clear”]. (11) ḫalaqa l-ʾinsāna min nuṭfatin fa-ʾiḏā huwa ḫaṣīmun mubīnun (16:4; similarly, 36:77) He created man from a drop – and behold! suddenly he is clearly contenious!

7:184 “qad ʾabāna lakum ʾayyuhā n-nāsu ʾinḏārahū mā ʾanḏarakum bihī min baʾsi llāhi ʿalā kufrikum bihī” [“He has made evident to you the fact that he is a warner, o people, when he warned you against the harm coming from God due to your denial of him”]. Al-Bayḍāwī (1418/1998: iii.44) glosses the same locus as “muwaḍḍiḥu/mūḍiḥu ʾinḏārahū bi-ḥayṯu lā yaḫfā ʿalā nāẓirin” [“who makes the fact that he is a warner clear in such a manner that it cannot be hidden from a complating person”]. The latter also offers the following explanations of rasūl mubīn in 43:29: “ẓāhiru r-risālati bi-mā lahū mina l-muʿğizāti ʾaw mubīnun li-t-tawḥīdi bi-l-ḥuğaği wa-l-ʾāyāti” [“the messenger whose mission (i.e. messengerhood) is evident from the miracles that appertain to him or who makes clear God’s oneness by proofs and signs”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

59

(12) fa-ʾaṣbaḥa fī l-madīnati ḫāʾifan yataraqqabu fa-ʾiḏā llaḏī stanṣarahū bi-l-ʾamsi yastaṣriḫuhū qāla lahū mūsā ʾinnaka la-ġawīyun mubīnun (28:18) The next morning found (Moses) in the city, while he was in a state of fear, looking around him, and behold! suddenly the one, who had requested his help the day before was calling out to him. (Moses) said: “Behold, you are clearly in error!” (13) wa-bāraknā ʿalayhi wa-ʿalā ʾisḥāqa wa-min ḏurriyyatihimā muḥsinun wa-ẓālimun li-nafsihī mubīnun (37:113) And We blessed him and Isaac. Yet among their offspring are some that do good and some who manifestly transgress against themselves. (14) ... ʾinna l-ʾinsāna la-kafūrun mubīnun (43:15) … Behold, man is clearly an ingrate. (15) ... qāla l-kāfirūna ʾinna hāḏā la-sāḥirun mubīnun (10:2) The disbelievers say: “Behold, he is obviously a sorcerer”. (16) fa-ʾalqā ʿaṣāhu fa-ʾiḏā hiya ṯuʿbānun mubīnun (7:107; 26:32) Then (Moses) threw his rod and behold, there it was clearly a snake! In (12), Moses designates the man who had ask for his assistance and now requests his help again as someone who is qualified by this repeated behaviour as someone whose error and deviousness are perceivable. This understanding is also shared by al-Bayḍāwī (1418/1998: iv.174) who explains the phrase as “bayyinu l-ġawāyati li-ʾannaka tasabbabta li-qatli rağulin wa-tuqātilu ʾāḫara” [“whose error is evident, because you were the cause for the killing of a man and are fighting another one”]. The same interpretation of the phrase is also found in examples (13–16): “ẓālim li-nafsihī mubīn” = “someone whose transgression against himself is evident”; “kafūr mubīn” = “someone whose ingratitude is evident”;24 “sāḥir mubīn” = “someone whose sorcery is perceivable”; “ṯuʿbānun mubīnun” = “something the snake-ness of which is perceivable”. The explanations given by some of the exegetes regarding example (21) are grammatically very illuminating, as will be shown below: Aṭ-Ṭabarī (1422/2001: x.343) comments on 7:107 with “tubīnu li-man yarāhā ʾannahā ḥayyatun”, i.e. “(a snake) which shows to whoever sees it that it is a snake”. Al-Bayḍāwī (1418/1998: iii.68) explains the same locus with “ẓāhiru ʾamruhū lā yušakku fī ʾannahū ṯuʿbānun” [“one whose quality (literally: matter) is evident/perceptible, whose being a snake cannot be doubted”], while he glosses 24 Note the explanation given by al-Bayḍāwī (1418/1998: v.88): “ẓāhiru l-kufrāni” [“whose ingratitude is perceivable”]; similarly, al-Muqātil (1423/2002: iii.790): “bayyinu l-kufri” [“whose ingratitude/disbelief is clear”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

60

Daniel Birnstiel

the same phrase in 26:32 simply as “ẓāhiru ṯuʿbānīyatihī”, i.e. whose snakeness is evident” (ibid.: iv.137). The same semantic relation can also be shown to exist between mubīn and a nominal head that refers to an abstract notion or object. One of the abstract nouns that occurs rather frequently is ṣiḥr “sorcery, magic”:25 (17) fa-lammā ğāʾahumu l-ḥaqqu min ʿindinā qālū ʾinna hāḏā la-siḥrun mubīnun (10:76) Yet then, when the truth came to them from us, they said: “Behold, this is clearly a case of sorcery!” (18) fa-lammā ğāʾathum ʾāyātunā mubṣiratan qālū hāḏā siḥrun mubīnun (27:13) Yet then, when our signs came to them in noticeable form, they said: “This is clearly a case of sorcery!” In all these cases, it is evident from the context that “clear magic” or “open sorcery” can only be such sorcery that can be identified as such, i.e. magic the “magic-ness” of which is perceivable. The statement made by the disbelievers in all these examples is that the evidence they have been given leads them to identify the signs as sorcery and cheap conjurations, and thus necessarily as not real and true. This is also the understanding of aṭ-Ṭabarī (1422/2001: ix.115f.) on 5:110 where he explains siḥr mubīn as “yubīnu ʿammā ʾatā bihī li-man raʾāhu wa-naẓara ʾilayhi ʾannahū siḥrun lā ḥaqīqata lahū” [“concerning that which he brought, it/he shows those who see it and look at it that it is magic lacking reality”]. Moreover, al-Bayḍāwī (1418/1998: iv.156) glosses 27:13 as “wāḍiḥu siḥrīyatihī”, i.e. “(magic) whose magic-ness is evident”.26 Another abstract noun that is attested even more frequently is ḍalāl “error, straying from the way”. In these examples, the context shows that mubīn characterizes the fact that someone has gone astray, i.e. his being astray or his astray-ness as manifest and evident:27

25 The additional attestations are 5:110, 6:7, 11:7, 34:43, 37:15, 46:7, and 61:6. 26 Cf. also one of al-Bayḍāwī’s glosses on 10:76 (1418/1998: iii.120): “ẓāhirun ʾannahū siḥrun” [“evident that it is sorcery”] as well as al-Ḫafāğī’s explanation of 6:7 (1974: iv.23): “ʾay ẓāhirun kawnuhū siḥran”, i.e. “its being magic is evident”. Cf. also al-Ḫafāğī on 10:76 (1974: v.51), where he explains that the reference in al-Bayḍāwī’s gloss was to the entire species of sorcery as such. 27 The other attestations are 6:74, 12:8, 12:30, 19:38, 21:54, 26:97, 28:85, 31:11, 34:24, 36:24, 36:47, 39:22, 43:40, 46:32, 62:2, and 67:29. Notice al-Bayḍāwī’s glossing mubīn in 3:164 with ẓāhir “evident, visible, perceivable” (1418/1998: ii.46).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

61

(19) la-qad manna llāhu ʿalā l-muʾminīna ʾiḏ baʿaṯa fīhim rasūlan min ʾanfusihim yatlū ʿalayhim ʾāyātihī wa-yuzakkīhum wa-yuʿallimuhumu l-kitāba wa-l-ḥikmata wa-ʾin kānū min qablu la-fī ḍalālin mubīnin (3:164) God did indeed confer a favour upon the believers when he sent amidst them a messenger from among themselves to recite to them his signs, to purify them and to teach them scripture and wisdom, for they had been before that evidently in error. (20) qāla l-malaʾu min qawmihī ʾinnā la-narāka fī ḍalālin mubīnin (7:60) The leaders of his people said: “Behold, we are aware that you are evidently in error!” (21) ... wa-man yaʿṣi llāha wa-rasūlahū fa-qad ḍalla ḍalālan mubīnān (33:36) … and whoever opposes God and his messenger has thereby clearly gone astray. The verse in (19) asserts that Muḥammad’s mission constitutes a divine favour, because previously they had been in a state of error which is characterized as perceivable, i.e. a state “the strayed-ness of which is perceivable”. Similarly, in (20) Noah is assured by the leaders of his people that he is in a perceivable state of error. Examples (21) is to be understood in the same way, as shown by the context. Similarly, people’s mistaken behaviour leads to their act being recognizable as a sin in the following example:28 (22) wa-man yaksib ḫaṭīʾatan ʾaw ʾiṯman ṯumma yarmi bihī barīʾan fa-qadi ḥtamala buhtānan wa-ʾiṯman mubīnān (4:112) And whoever commits a mistake or a sin, but then throws it upon someone innocent has thereby burdened himself with calumny and an evident sin. Another frequently attested noun is sulṭān “authority, authorization”, which occurs often in conjunction with Moses, e.g. in (23) and (24). The context of these two examples is clear: Moses has been sent with signs as well as an evidently dinive authorisation:29 28 The phrase ʾiṯm mubīn also occurs in 4:20, 4:50 and 33:5. 29 Further attestations of this expression occur in 4:91, 4:144, 4:153, 14:10, 27:21, 37:156, 44:19, 40:23 and 51:38. Note that al-Bayḍāwī glosses sulṭan mubīn in 4:153 as tasalluṭ ẓāhir “evident authorization” (1418/1998: ii.107).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

62

Daniel Birnstiel

(23) wa-la-qad ʾarsalnā mūsā bi-ʾāyātinā wa-sulṭānin mubīnin (11:96) And indeed we did send Moses with our signs and clear authority. (24) ṯumma ʾarsalnā mūsā wa-ʾaḫāhu hārūna bi-ʾāyātinā wa-sulṭānin mubīnin (23:45) Thereafter we send Moses and his brother Aaron with our signs and clear authority. (25) ʾam lahum sullamun yastamiʿūna fīhi fa-l-yaʾti mustamiʿuhum bi-sulṭānin mubīnin (52:38) Or do they have a ladder on which they would listen? Then let the one of them who has listened bring clear authority! (25) presents a challenge to the unbelievers by hypothetically asking them if they had a ladder leading up to heaven whereby they could listen to God’s councils. The Qur’ān then mocks them and demands them to show something that would make the authoritativeness of their claims perceivable, i.e. something that would prove their being authorized to utter such claims. A similar reading is advanced by al-Bayḍāwī (1418/1998: v.155), who explains the phrase as “bi-ḥuğatin wāḍiḥin tuṣaddiqu stimāʿahū” [“with a clear proof that certifies his listening”]. Mubīn occurs also as attribute with several additional nouns, abstracts as well as concrete objects. In (26), the text asserts that a certain behaviour results in a type of loss that can be identified as such; in other words, the perceptibility and manifestness of the outcome as loss renders its denial impossible. Similarly, the context shows that in (28) God’s actions on the day of judgement will make it obvious that he is the actual true reality; i.e. God is a reality whose actuality and realness has become undeniably evident and manifest to the people on that day. This is also the understanding of al-Bayḍāwī (1418/1998: iv.103) who explains the phrase as “aṯ-ṯābitu bi-dātihi ẓ-ẓāhiru ʾilāhiyyatuhū lā yušārikuhū fī ḏālika ġayruhū wa-lā yaqdiru ʿalā ṯ-ṯawābi wa-l-ʿiqābi siwāhu ʾaw ḏū l-ḥaqqi l-bayyini ʾayi l-ʿādilu ẓ-ẓāhiru ʿadluhū” [“the one who is established by his own self, the one whose divinity is evident, wherein none other than He participates and apart from whom nobody has the power to reward and to punish, or the owner of the clear turth, the just one whose justice is evident”]. Likewise, the trial and tribulation in (28) is qualified by mubīn as the type of trial that can be perceived as such or that manifests itself as such:30

30 Additional examples with ḫusrān mubīn occur in 22:11 and 39:15; ḥaqq mubīn occurs also in 27:79, al-balāʾ al-mubīn in 37:106. The phrase al-fawz al-mubīn is also attested in 6:16. Further occurrences are 24:12 (ʾifk mubīn) and 44:10 (duḫān mubīn).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

63

(26) ... wa-man yuttaḫiḏi š-šayṭāna walīyan min dūni llāhi fa-qad ḫasira ḫusrānan mubīnān (4:119) … and whoever takes the devil as a friend against God has thus suffered a perceivable case of loss. (27) yawamʾiḏin yuwaffīhimu llāhu dīnahumu l-ḥaqqa wa-yaʿlamūna ʾanna llāha huwa l-ḥaqqu l-mubīnu (24:25) On that day God will pay them their proper (literally: true, real) due and they will know that it is God that is evidently the truth. (28) wa-ʾātaynāhum mina l-ʾāyāti mā fīhi balāʾun mubīnun (44:33) And we gave them the type of signs in which there was an evident trial. (29) fa-ʾammā llaḏīna ʾāmanū wa-ʿamilū ṣ-ṣāliḥāti fa-yudḫiluhum rabbuhum fī raḥmatihī ḏālika huwa l-fawzu l-mubīna (45:30) Yet as for those who believe and do righteous deeds, their lord will enter them into his mercy. Suchlike constitutes the evident triumph. (30) yā ʾayyuhā n-nāsu qad ğāʾakum burhānun min rabbikum wa-ʾanzalnā ʾilaykum nūran mubīnān (4:174) O people. A proof has come to you from your lord and we have sent down to you that which can is evidently a light. (31) … fa-ʾatbaʿahū šihābun mubīnun (15:18) … and then a noticeable flame followed him. (32) qāla ʾa-wa-law ğiʾtuka bi-šayʾin mubīnin (26:30) (Moses) said: “Even if I brought you something of a perceivably (divine) nature?” In (29), the mercy bestowed by God upon the believers is qualified as a type of triumph the nature of which is obvious to the onlookers. Similarly, the verse in (30) asserts to the audience that a light has been sent by God whose nature is such that they can recognize it as light. The perceptibility and recognizability of a flame as such is also marked as salient in (31); this also seems to be the interpretation of al-Bayḍāwī (1418/1998: iii.208), who glosses mubīn as “ẓāhirun li-l-mubṣirīna”, i.e. “visible/evident to the observers”. In (32), Pharaoh had threatened to throw Moses into prison if he should worship a god other than Pharaoh. Moses asks whether that threat was valid even if he were to give Pharaoh something the nature of which would be perceptible to Pharaoh, i.e. something that would manifestly be a sign or miracle of divine origin and evidence of divine authorization of Moses’ action. Similarly, in (33) the text asserts that the bounties that have been bestowed upon Solomon constitute the kind of favour the being-a-favour of which is perceptible; interestingly, al-Bayḍāwī (1418/1998: iv.157) glosses the expression as “llaḏī lā yaḫfā ʿalā ʾaḥadin” [“that cannot be concealed to anyone”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

64

Daniel Birnstiel

Similarly, (34) must be understood in the context of the peace treaty of Ḥudaybīyah of 628, which was at first perceived by Muḥammad’s followers as an unsuccessful conclusion of their pilgrimage. The text asserts that nevertheless a victory has been achieved by Muḥammad whose victory-ness will manifest itself and thus be perceivable: (33) wa-wuriṯa sulaymānu dāwūda wa-qāla yā ʾayyuhā n-nāsu ʿullimnā manṭiqa ṭ-ṭayri wa-ʾūtīnā min kulli šayʾin ʾinna hāḏā la-huwa l-faḍlu l-mubīnu (27:16) And Solomon inherited from David and said: “O people! We have been taught the speech of the birds and we have been given of everything. Behold, this is what constitutes an evident kind of favour. (34) ʾinnā fataḥnā laka fatḥan mubīnān (48:1) Behold, we have given you a perceivable victory. (35) wa-la-qad raʾāhu bi-l-ʾufuqi l-mubīni (81:23) And he did indeed see him on the perceivable horizon. In the light of the previous interpretations, (35) must also be understood in a similar sense: here mubīn qualifies the horizon as perceptible, i.e. the angel’s appearance manifests itself undeniably on the horizon. Al-Qaysī’s explanation (1429/2008: xii.8093) is quite noteworthy: “[ʾay] wa-la-qad raʾā [muḥammadun] ğibrīla ʿalayhi s-salāmu fī ṣūratihī fī n-nāḥiyati [llatī] tatabayyanu fīha l-ʾašyāʾu fa-yurā man qablahā wa-ḏālika nāḥiyatu maṭlaʿi š-šamsi min qabli l-mašriq” [“[I.e.] [Muḥammad] did indeed see Gabriel (on him be peace) in his (real) shape in that direction in [which] things become clearly distinguishable and someone (standing) in front of it becomes seen, and that is the direction from which the sun comes out at the moment before sunrise”]. The observation made by ʾIsmāʿīl Ḥaqqī (1330/1912– 1347/1928: x.352) is similarly noteworthy, namely “wa-l-mubīnu min ʾabāna l-lāzimi bi-maʿnā z-zāhiri”, i.e. “and mubīn is derived from the intransitive ʾabāna (stem IV of bāna) in the meaning of ẓāhir “evident, perceivable, manifest””. Since the adjective “clear” is used also in English (and other languages) in meteorological contexts referring e.g. to a bright, cloudless, sky or horizon, or clear weather, it is important to note that mubīn is not used in this way. Rather, as shown by the preceding examples, it indicates that its referent is easily identifiable as the very thing it has just been called. Consequently, “seeing him on al-ʾufuq al-mubīn” means “seeing him on a place whose identity as horizon is evident”, i.e. a place or direction that is easily identified as the horizon. The focus lies on the assertion of the angelic vision as physical and not spiritual.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

65

In the light of the above mentioned observation made by Ḥaqqī (1330/1912– 1347/1928: x.352), i.e. that stem IV of bāna (√byn) is – at least in the mentioned example – intransitive, it seems worthwhile to discuss at this point the semantics of verb, the etymology of the root and the function of the attested structure, before continuing to analyse the instances, in which mubīn occurs as attribute of a noun from the semantic field of “scripture, message, transmission”. 3.2 Stem IV in Arabic grammar It is generally thought that stem IV of the Arabic verb system (as well as its cognate forms in other Semitic languages) is causative and/or factitive. This view is evident in Gilliot’s assertion (2008: 94f.): “Mubīn is the active participle of the causative-factitive abāna, which can be understood as “making [things] clear”.” It is also found underlying the above-mentioned interpretations of mubīn, as evidenced e.g. in explaining “verses of the clear book” (12:1) as “bayyana llāhu taʿālā rušdahū wa-hudāhu”, i.e. “God the almighty made clear his right path and his guidance” [aṣ-Ṣanʿānī (1410/1989: i.317)] or as “al-muẓhiru li-l-ḥaqqi mina l-bāṭili”, i.e. “which makes visible truth and distinguishes it from falsehood” [al-Maḥalli & as-Suyūṭī (n.d.: 235]. However, in many cases, verbs of stem IV can be both, transitive and intransitive, or only intransitive. This fact is also repeatedly observed by several Muslim commentators. Thus al-Ḫafāğī (1974: v.51) on 10:76 explains: “yašīru ilā ʾanna mubīn min ʾabāna bi-maʿnā ẓahara wa-ttaḍaḥa lā bi-maʾnā ʾaẓhara wa-ʾawḍaḥa ka-mā huwa ʾaḥadu maʿnayayhi”, i.e. “it indicates that mubīn is derived from ʾabāna (stem IV) in the sense of ẓahara “to be come evident” and ittaḍaḥa “to become clear, manifest, apparent”, not in the sense of ʾaẓhara “to make evident” and ʾawḍaḥa “to make clear, manifest, apparent”, which is one of its two senses”. Similarly, al-Ḫafāğī (1974: ii.296) comments on Bayḍāwī’s gloss ẓāhiru l-ʿadāwati “whose enmity is visible/perceptible” (2:208) with: “ʾišāratun bi-ʾanna ʾabāna lāzimun bi-maʿnā ẓahara”, i.e. “(it is) an indication that ʾabāna is intransitive, having the sense of ẓahara “to become evident, perceivable””.31 31 Al-Ḥafāğī (1974: ii.266) expresses himself similarly regarding al-Bayḍāwī’s interpretation of 2:168: “yaʿnī ʾannahū min ʾabāna bi-maʿnā bāna wa-ẓahara” [“That means that it is derived from ʾabāna (stem IV) in the sense of bāna (stem I, “to become evident”) and ẓahara “to become evident, perceivable””]. Elsewhere, he offers similar comments (ibid.: iii.196, 226; v.288; vii.449).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

66

Daniel Birnstiel

The alleged double sense of ʾabāna is also mentioned by an-Nasafī (1419/1998: ii.627) in his commentary on 28:2: “yuqālu bāna š-šayʾu wa-ʾabāna bi-maʿnan wāḥidin wa-yuqālu ʾabantuhū fa-ʾabāna lāzimun wa-mutaʿaddin”, i.e. “one says bāna š-šayʾu (stem I) and ʾabāna (stem IV) “the thing has become evident” with a single meaning, and one says ʾabantuhū (IV) “I have made it evident”, thus ʾabāna is intransitive and transitive”. AnNīsabūrī (1416/1996: iv.64) expresses himself similarly: “madāru hāḏihi t-tafāsīri ʿalā ʾanna ʾabāna lāzimun mutaʿaddin yuqālu ʾabāna š-šayʾa wa-ʾabāna huwa bi-nafsihī” [“These interpretations rotate around the fact that ʾabāna is intransitive as well as transitive. One says: ‘he made something evident’ and ‘it was in itself evident’”]. These views are also corroborated by the findings of the Arab lexicographers. Az-Zabīdī (1385/1965–1422/2001: xxxiv.297) states that stem I, II, IV, V and X are all “lāzimatun mutaʿaddiyatun” [“intransitive-transitive]. Ibn Manẓūr (1415/1994: xiii.67) likewise mentions that these five stems bimaʿnan wāḥidin, i.e. they have a single meaning. In addition, the following important observation made by al-ʿAkbarī (n.d.: i.155) seems to indicate that the intransitive meaning is the primary one: “yuqālu tabayyana š-šayʾu wa-bāna wa-ʾabāna wa-stabāna kulluhū lāzimun wa-qad yastaʿmilu ʾabāna wa-stabāna wa-tabayyana mutaʿaddiyatan”, i.e. “one says tabayyana š-šayʾu (stem V) and bāna (stem I) as well as ʾabāna (stem IV) and istabāna (stem X) “the thing became evident”; all these forms are intransitive, however one may use ʾabāna (stem IV), istabāna (stem X) and tabayyana (stem V) transitively”.32 Nöldeke (1904: 37 n. 1), where he mentions that bāna as well as ʾabāna, tabayyana and istabāna mean “to be obvious/clear/manifest” (“deutlich sein”). The fact that stem IV is not necessarily causative has also been observed by Western Arabists and Semitists alike. Chouémi (1966: 128) shows that of the 301 Qurʾānic verbs attested in stem IV, 159, i.e. more than half, cannot be characterized as causative.33 Zaborski (2007: 31) similarly mentions the findings by Chekayri (1994) which likewise shows the frequency with which stem IV does not function as causative of stem I. Kropfitsch (1991:65) makes the important observation that in Modern Literary Arabic forms of stem IV may semantically be analysed as causatives, 32 Similarly, al-ʿAkbarī (n.d.: i.341) on 4:19, where he includes stem II (bayyana) among the intransitives and mentions mubīnah as variant reading of mubayyanah; see further below the discussion of these examples. Also an-Nīsabūrī (1416/1996: ii.16) mentions the intransitive use of stem II. Regarding the intransitivity of bayyana and the interpretation of 4:19, see also Ibn ʿĀdil (1419/1998: vi.260). On the issue of intransitive stems of this root, se also ar-Rāzī (1401/1981: vii.16) and aṯ-Ṯaʿlabī (1422/2002: iv.152). 33 Some of these are nevertheless transitive.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

67

factitives, denominals, intransitives and ingressives. Vollers (1906: 110) and Zaborski (2007: 31) also mention that the characterization of stem IV as causative was already questioned by several Arab grammarians in the middle-ages, including the well-known az-Zağğāğ. Moreover, as Zaborski (ibid.) points out, there is a larger number of verbs attested in both stem IV and stem II (or III) with identical causative or factitive meaning (despite the attempts of grammarians to find shades of meaning distinguishing them), while there is also “a considerable number of forms with the same non-causative meaning in classes no. I and no. II.” This phenomenon is also known from other Semitic languages. Gesenius (1910: 145f.) goes into considerable length in his attempt to explain how various verbs are causative “according to the Hebrew point of view” (ibid.: 145), even though these verbs would generally hardly be thought of as causatives and include verbs “which express the obtaining or receiving of a concrete or abstract quality”, verbs “which express … the entering into a certain condition and, further, the being in the same”, and “stems which express action in some particular direction” (ibid.). Non-causative verbs that formally belong to the causative š-stem occur also in Akkadian [von Soden (1995: 145f.)], while also the Syriac causative stem (afel) features intransitive verbs, e.g. ʾaṷreq “to turn green”, ʾaʿteq “to grow old”, ʾarḥeq “to depart, distance onself”, or ʾaskel “to act foolishly” [Brockelmann (1908: 527)]. All of this indicates that non-causative transitive and intransitive meanings of stem IV verbs are not unusual and even to be expected from a diachronic point of view.34 3.3 The semantics of byn: comparative evidence It seems likely that the idea originally expressed by the root byn is that of “seperation”. This meaning is attested in Qatabian byn “to keep apart” in the sense of “to keep one’s distance”, e.g. ybnwn “let him keep his distance” [Ricks (1989:24)]. It is also evident in Sabaic byn, imperfect ybnn “to remove (punishment), to intervene, separate (a boundary)” [Beeston, Ghul, Müller & Ryckmans (1982: 34)] and “to be dislocated; to cease < to be cut off < to be separate” [Biella (1982: 40)]. Lane (1863–1893: i.285) also records similar meanings for Arabic bāna (stem I): “it became seperated, severed, disunited, or cut off”.35 To this one must add the noun bayn “separation, interval”,36 the 34 For a recent treatment of the issue of semantic transitivity of Semitic verb stems see Bjøru (2014). 35 Cf. also Leslau (1991: 773). 36 Cf. Leslau (1991: 115) for Ge’ez bayn “interval, distance”. Cf. Lane (1863–1893: i.287f.)

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

68

Daniel Birnstiel

existence of which is attested for numerous branches of Semitic (Arabic, South Arabian, Ethiopic, Hebrew, Phoenician, Ugaritic, Aramaic) by its development into the preposition “between”.37 How the semantic notion of “separation, separateness” may develop connotations of “apparentness, visibility, recognizability” is easily comprehensible. Thus Lane (1863–1893: i.286) records the following meaning for bāna given by the Arab lexicographers “it … was, or became, [distincs, as thought separate from others; and thus,] apparent, manifest, evident, clear, plain, or perspicuous: … and it was, or became, known”. Against this background, it becomes understandable, why Sabaic bynn (noun, used as adverb) should come to mean “openly, publicly” [Biella (1982: 40)]: something that is separate stands out from the mass; it is thus recognizable. This semantic development is especially evident in Ge’ez bayyana “to discern, distinguish, remark, pay attention, notice, recognize, consider, demonstrate, judge, pass judgement, expound”; abayyana “to decide, make clear, make evident, make see clearly, state clearly, demonstrate, prove, manifest, make understand, make a difference” [Leslau (1991: 115)]. The meanings of “identifiability” and “visibility” are also present in the modern South Arabian languages, e.g. in Mehri bǝyōn / yǝbyōn / yǝbyēn “to appear”, abyēn “to make something clear, appear clearly”, hǝbyōn / yǝhǝbyōn / yǝhábyǝn “to reveal, clarify” [Johnstone (1987: 60f.)]; in Ḥarsūsi bǝyōn “to appear”, abyōn / yabyōn / yǝhábyen “to distinguish something from something, to reveal all of something” [Johnstone (1977: 22)]; in Jibbāli (Śḥeri) bɛ̄n / ybɛ́n “to appear”, ebyín / yǝbyínǝn / yɛ́byǝn “to reveal; to make something out from a distance”, ǝbtīn “to become visible”, s̃ǝbyín “to become visible” [Johnstone (1981: 32)]. In North-West Semitic languages, on the other hand, the semantic development is focused more on “inward discernment of a thing”38 rather than its outward identification, i.e. it is more concerned with the understanding of a matter, and consequently with explaining it. This is borne out by Ugaritic b-n “to understand, to take note (?)” [del Olmo Lete & Sanmartín (2004: i.224)]; Jewish Palestinian Aramaic bayyen “to explain D”, iṯbōnēn, “to investigate, for bayn, bīn. Cf. also Gesenius (1954: 94f.). 37 Cf. Brockelmann (1908: 498). For Arabic bayna see Lane (1863–1893: i.287f.), for Sabaic byn, bn, bynn “between; among; in both of two things”, see Biella (1982: 40), Beeston, Ghul, Müller & Ryckmans (1982: 34); for Qatabanian byn “between”, see Ricks (1989: 25); for Ugaritic bn “between, among”, cf. del Olmo Lete & Sanmartín (2004: i.227f.); for Syriac baynay, baynāṯ “between, among, within, in the midst”, cf. Payne Smith (1998: 42); for Hebrew bên “between”, see Gesenius (1954: 94f.). Cf. also Schulthess (1903: 25) for Christian Palestinian Aramaic, Leslau (1991: 116) for Ge’ez and other Ethiopic languages, and Johnstone (1977: 22) for modern South Arabian. 38 Cf. also Nöldeke (1904: 40).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

69

understand; to pay attention Dt” [Sokoloff (1992: 96)];39 and Syriac bayyen “to show, point out, to give instruction, understanding, to discern D”, eṯbayyan “to notice, take notice of, regard, consider, understand, have discretion Dt” [Payne Smith (1998: 38)]. The situation in Hebrew is very instructive: The qal-stem bīn means “to notice, to discern/perceive; to take care/heed, pay attention to; to be familiar with/skilled in”, while the hip̄ ʿīl-stem heb̠īn means “to pay attention to; to understand, to know about; to have/gain introspection”, but also “to give insight/introspection, to teach”. This shows that, although the semantic field of “understanding, explaining” is better attested, meanings related to the field of “identification” are by no means absent. All of this seems to indicate strongly that in Arabic, the meanings may be expected to connect primarily with the semantic field of “identity/identification” and “perception/ perceivability”.40 In this context, it is noteworthy that certain dialects of Yemeni Arabic use stem II “to make clear” specifically in the context of questioning an interlocutor whether a man is a tribesman or a Jew, while stem VIII is attested in the meaning “to appear, to be visible, be/become clear, to be recognized” [Piamenta (1990: i.47)]. 3.4 The expression [noun + mubīn]: internal syntactic and semantic structure Translating ʿadūw mubīn as “an enemy who is recognizable as such” or naḏīr mubīn as “a warner who is evidently one” show that the attribute does not refer to the noun (nucleus) itself, but rather only to a part thereof. More precisely, mubīn refers to the quality inherent in the noun. This is also clear from some of the glosses given by the commentators, who utilize what is known as “improper annexation”(al-ʾiḍāfah ġayru l-muḥaqqiqah),41 e.g. Muqātil (1423/2002: iii.583) “bayyinu l-ʿadāwati” [“whose enmity is obvious”], al-Bayḍāwi (1418/1998: iv.174) “zāhiru l-ʿadāwati” [“whose enmity is manifest”], (ibid.: iv.137) “ẓāhiru ṯuʿbānīyatihī” [“whose snake-ness is evident”] and (ibid.: iv.156) “wāḍiḥu siḥrīyatihī” [“(that) 39 See Schulthess (1903: 25). 40 All this is not to say that e.g. Arabic forms can never imply “understanding” and “explanation” of the subject matter or only rarely do so, rather, it simply appears that the southern Semitic languages and Arabic used this root to primarily indicate the identification and perception of a thing and only secondarily its understanding and elucidation. 41 For this construction, see Fischer (1987: 79f., 178) and Wright (1898: 64f., 198f.). Note, however, that some of the glosses and explanation utilize other constructions, such as the indirect attribute (an-naʿt an-sababī; see Beeston (1970: 94f.), Fischer (1987: 194), Wright (1898: 283f.), Polotsky (1978).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

70

Daniel Birnstiel

whose magic-ness is evident”]. This means, however, that there is an internal predication of some sort. Expressions like ʿadūw mubīn, etc. may be compared to expressions such as mechanic engineer, nuclear scientist, marine biologist, cardiovascular biologist, and oceanic carthographer.42 In all these examples, the adjective in question does not refer to the occupational title, i.e. the noun, but rather to the underlying occupation from which it is derived. This is to say, cardiovascular biologist does not refer to a biologist who is cardiovascular, but to someone whose profession is cardiovascular biology. Such expressions may also be compared to German compounds with an attribute that refers to the first element of the phrase, but not the basis of the compound (i.e. the second element), e.g. deutsche Literaturwissenschaft “German literature studies”, where the literature is characterized as German, but not the studies (even though structurally the attribute appears to refer to it) or protestantische Kirchengeschichte “protestant church history”, i.e. the history of the protestant church, not the protestant history of the church. In German, such compounds are said to by grammatically incorrect and acceptable in the standard language only for several set expressions [cf. Duden (1995: 257f.)].43 However, Jespersen in his discussion of adjuncts (1949: ii.283–292) mentions expressions such as heavy sleeper and early riser. While nuclear scientist may be paraphrased as a scientist who deals with nuclear science, such a paraphrase is impossible with heavy sleeper and the like. According to Jespersen (ibid.: 283), “Here the adjunct is a shifted subjunct of the verb contained in the substantive riser : he rises (vb) early (adv) = he is an early (adj) riser (sb).” Another way to paraphrase this using Jespersen’s notation would be “riser : his rising (verb noun) is early (adjective) = he is an early (adj) riser (sb)”; in other words, “someone who rises early” is “someone whose rising is/happens early”, i.e. it is structurally comparable to “someone whose hair is black”, “a black-haired person”. In this context, it is important to note the fact emphasized by Goldenberg (1995: 6ff.) especially with relation to the Semitic languages, where the dis42 I owe a number of these expressions to Maiko Yamaji. 43 Duden (1995: 257f.), however, seems to indicate that such expressions are not completely absent from the spoken language. See also Bergmann (1980) and Sandberg (1984). For the different semantic relations between attribute (satellite) and noun (nucleus), see also Bär (2002). The English equivalent of these German phrases are expressions such as public school-boy or dirty laundry box; cf. Jespersen (1949: ii.284). Note also that in English, mechanical engineering is not a type of engineering that is mechanical, but a type of engineering that is related to mechanics; i.e. mechanical, cardiovascular, oceanic are here relational adjectives.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

71

tinction between noun and adjective is difficult to draw, that adjectives “do not name qualities but possessors of qualities”. Thus, “strong” denotes “someone/ something possessing strength”, or in Arabic qawīyun = ḏū quwwatin. Similarly, ʿadūwun “enemy” could be paraphrased as ḏū ʿadāwatin “possessor of enmity”, or rasūl “messenger” as ḏū risālah “owner of a mission”. Since a structure like ʾaswadu š-šaʿri “black-haired” is equivalent to ḏū š-šaʿri l-ʾaswadi “the possessor of the black hair”, it becomes clear that even substantives like “sorcery, magic” may on a more abstrect level be converted into a phrase like “that which possesses the quality of magic-ness”, as is implicitly indicated also by al-Bayḍāwī’s gloss (1418/1998: iv.156) “wāḍiḥu siḥrīyatihī” [“(that) whose magic-ness is evident”]. One may thus summarize this part of the discussion by noting (a) that [noun + mubīn] in the Qur’ān may be compared to English expressions such as early riser or heavy sleeper and must be regarded as related to the improper annexion (al-ʾiḍāfah ġayru l-muḥaqqiqah) and the indirect attribute (an-naʿt as-sababī), (b) that stem IV of bāna is clearly attested with non-causative, intransitive meaning, and (c) that bāna is also attested with meanings indicative of the semantic field of “recognition, identification, identity” and the like; (b) and (c) are also corroborated by other Semitic languages. The exact semantic differences expressed by the different stems as attested in the corpus remain to be analysed and will be dealt with further below. The following section treats those instances of mubīn, in which it co-occurs with a noun related to scripture, revelation or language. 3.5 mubīn following a noun referring to revelation, scripture or language Among the examples following into this category are a number of occurrences of the phrase al-balāġ al-mubīn. The context of (36) shows that the findings made above also hold true here: the expressions does not mean “a conveyance that makes things clear”, but rather indicates that Muḥammad’s duty to transmit the message entrusted to him in a way that enables the audience to ascertain that a transmission of a specific nature has indeed taken place. This also holds true for the examples (37) – (39):44

44 The additional occurrences of al-balāġ al-mubīn are 16:35, 24:54, and 64:12.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

72

Daniel Birnstiel

(36) wa-ʾaṭīʿū llāha wa-ʾaṭīʿū r-rasūla wa-ḥḏarū fa-ʾin tawallaytum fa-ʿlamū ʾannamā ʿalā rasūlinā l-balāġu l-mubīnu (5:92) And obey God and obey the messenger and beware! And if you turn away, so know that what is incumbent upon the messenger is a conveyance which can be identified as such. (37) fa-ʾin tawallaw fa-ʾinnamā ʿalayka l-balāġu l-mubīnu (16:82) And if they turn away, then what is incumbent upon you is a conveyance which can be identified as such. (38) wa-ʾin tukaḏḏibū fa-qad kaḏḏaba ʾumamun min qablikum wa-mā ʿalā r-rasūli ʾillā l-balāġu l-mubīnu (29:18) And if they give the lie – so nations before you have given the lie, and only giving a conveyance which can be identified as such is incumbent upon the messenger. (39) wa-mā ʿalaynā ʾillā l-balāġu l-mubīna (36:17) Only giving a conveyance which can be identified as such that is incumbent upon us. In these cases, the context indicates that Muḥammad will not be taken to task for failing to make people understand or accept the message; what is demanded of him is to transmit it in a way that permits the addressees to recognize the fact that a message of divine origin is being conveyed. In examples (40) – (41), the text asserts the audience that the scripture they are presented with is one the very nature of which permits the conclusion that it is indeed of divine origin. This seems also to be the view of an-Nasafī (1419/1998: ii.554), who glosses 26:2 as “aẓ-ẓāhiri ʾiʿğāzihī wa-ṣiḥḥati ʾannahū min ʿindi llāhi”, i.e. “the one whose mirculous nature and the correctness of its being from God are evident”. A similar idea is expressed by al-Bayḍāwī (1418/1998: iii.154), who explains 12:1 with “ʾay tilka l-ʾāyātu ʾāyātu s-sūrati ẓ-ẓāhiri ʾamruhā fī l-ʾiʿğāzi” [“That means: These verses are verses of the surah whose issue (i.e. nature/essence) regarding the miraculous nature is evident”]:45 45 Sprenger (1869: ii. 352 n. 2) expresses a similar view: “In Arabic: mubīn “evident”, “distinguishable”. Muḥammad often says that he is an evident messenger of God, i.e. one who can easily be identified as such, concerning whose vocation there can be no doubt. Here mubīn has a similar meaning: “There is no doubt concerning the existence of the book” (cf. Qur’ān 2:1), but it does not mean: The plain book.” The German reads as follows: “Im Arabischen: mobyn „offenbar“, „unterscheidbar“. Moḥammad sagt oft, daß er ein offenbarer Bote Gottes sei, d.h. einer, den man leicht als solchen erkennen kann, über dessen Beruf kein Zweifel obwaltet. Hier hat mobyn eine ähnliche Bedeutung: „Es ist kein Zweifel über die Existenz des Buches vorhanden“ (vergl. Ḳ 2, 1), es heißt aber nicht: Das deutliche Buch.”

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

73

(40) ʾ“l”r tilka ʾāyātu l-kitābi l-mubīni (12:1; similarly 26:2, 28:2) ʾalif – lām - rāʾ These are signs of the syripture that can be identified as such.. (41) wa-l-kitābi l-mubīni (43:2; 44:2) And by the scripture that can be identified as such (42) ... qad ğāʾakum mina llāhi nūrun wa-kitābun mubīnun (5:15) ... Indeed, a light has come to you from God, and a scripture that can be identified as such. (43) wa-mā min dābbatin fī l-ʾarḍi ʾillā ʿalā llāhi rizquhā wa-yaʿlamu mustaqarrahā wa-mustawdaʿahā kullun fī kitābin mubīnin (11:6) There is no beast upon the earth except that its sustenance is incumbent upon God, and He knows its time-span and its resting-place; everything is (kept) in a book that can be identified as such. (44) wa-mā min ġāʾibatin fī s-samāʾi wa-l-ʾarḍi ʾillā fī kitābin mubīnin (27:75; similarly, 6:59, 10:61, 34:3) And the there not a hidden thing in the heavens and on the earth except that it is (kept) in a record that is identifiable as such. In examples (43) and (44), kitāb does not refer to a revealed scripture, but rather to the heavenly record that is kept of everything happenining in the universe. Although one might think at first glance, that a record that is mubīn would imply a record that clarifies things or the like, applying the established meaning of mubīn shows that the text tells the audience that all things are kept in a record which is such that it can be identified as record; i.e., all things recorded in a document that the audience could be identified and recognized as a record, if they were to be presented with it. A somewhat similar interpretation is also given by aṭ-Ṭabarī (1422/2001: xii.328) on 11:6 who explains “kullun fī kitābin mubīnin” with “yubayyinu li-man qaraʾahū ʾanna ḏālika muṯbatun maktūbun qabla ʾan yaḫluqahā wa-yūğidahā”, i.e. “it makes evident to whoever reads it that this has been established and written down before he created them and brought them into being.” Also in the following examples mubīn in conjunction with a scriptural term must be understood as “making its nature perceivable”:46

46 The expression qurʾān mubīn occurs also in 36:69, the phrase ʾimām mubīn is also found in 15:79.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

74

Daniel Birnstiel

(45) ṭ”s tilka ʾāyātu l-qurʾāni wa-kitābin mubīnin (27:1) ṭāʾ - sīn These are signs of the recital and of a scripture that is perceivably one. (46) ʾ”l”r tilka ʾāyātu l-kitābi wa-qurʾānin mubīnin (15:1) ʾalif – lām - rāʾ These are verse of the scripture and of recital that is perceivably one. (47) ʾinnā naḥnu nuḥyī l-mawtā wa-naktubu mā qaddamū wa-ʾāṯārahum wa-kulla šayʾin ʾaḥṣaynāhu fī ʾimāmin mubīnin (36:12) Behold: It is we who resurrect the dead and write down whatever they have send ahead as well as their remains, and everything – we preserved it in an account that is clearly identifiable as record. When considered in context, understanding [X + mubīn] as “that which is recognizably X” also leads to a better understanding of 16:103 and 26:195 than interpreting mubīn as “making things clear” or “easily comprehensible”:47 (48) wa-la-qad naʿlamu ʾannahum yaqūlūna ʾinnamā yuʿallimuhū bašarun lisānu llaḏī yulḥidūna ʾilayhi ʾaʿğamiyyun wa-hāḏā lisānun ʿarabiyyun mubīnun (16:103) And we do indeed know that they are saying “It is a human being that teaches him. The languages of the one they point at is foreign, while this is perceivably Arabic language. (49) bi-lisānin ʿarabiyyin mubīnin (26:195) “in a language that is recognizably Arabic” The context shows that to counter the accusation that Muḥammad has been taught by someone, the text asserts that the language spoken by the alleged informant is non-Arabic, while the language of the message is one that is perceivable to the addressees as Arabic. In other words, the language of the message is such that its rootedness in Arabic language, culture and mentality can be recognized by the audience, which can deduct from it that the source cannot be the alleged informant, whose steepedness in non-Arabic culture and language precludes the possibility of borrowing. One may possibly go so far as to state that the point made by the Qur’ān is that even though it (re-)tells stories and events of Judaeo-Christian provenance, these narratives are presented in 47 Translating lisān ʿarabī with “Arabic language” should not be taken as similar with “Classical Arabic”. The question of what exactly is meant by this expression and what Arabic culture and customes mean to the Qurʾāns original audience still demands further research, especially in light of recent findings regarding the nature of Qurʾānic Arabic and the history of Classical Arabic. See section 2.2. above, especially note 3.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

75

such a thoroughly Arabic manner that the possibilty of their being borrowed should be excluded by the addressees upon consideration. It, however, does not imply per se that this language form must be devoid of loan words or that the message must necessarily be composed in supra tribal poetic language. All it implies is its Arab-ness and Arabic-ness, whatever that meant to the audience. This was already recognized by Sprenger (1869: i.482), who translates 26:195 as “es ist augenscheinlich in arabischer Sprache abgefaßt” (, i.e. “it is ostensibly/evidently composed in Arabic language”) and notes (ibid: i.482 n. 1): “Literally: in unmistakable, i.e. pure, idiomatic Arabic language, the like of which no stranger can write; consequently, it is an original revelation and not copied from the earlier scriptures. Cf. 16:105.”48 More recently Reynolds (2008: 15) has touched upon this notion of Arab-ness and Arabic-ness that is asserted by the Qur’ān: “Perhaps what the Qur’ān means when it describes itself as ʿarabiyyun mubīn (16:103, 26:195) “clear [or, better, ‘demonstrative’] Arabic” is not that it is in pure, classical Arabic, but that it is in the language of its people. Perhaps this phrase reflects the text’s pride at being the first effort at an Arabic scripture.”49 As shown, all attestations of mubīn can credibly be analysed as an intransitive attribute qualifying its nucleus as something the nature of which is evident or identifiable. There remain only two instances of forms derives from stem IV of bāna to be discussed, namely 43:18 and 43:52. 3.6 mubīn as predicate In (50), mubīn does not occur as an attribute of another noun. It seems that the personal pronoun huwa “he” refers back to the indefinite pronoun man “someone” at the beginning of the verse. The phrase ġayru mubīn implies reference to some expression of agent-hood, animateness and person-ness, and one may regard the expression as equivalent to an instance of [“someone” + mubīn]. The entire verse can thus be understood as referring to someone, who is reared in luxury, yet cannot articulate himself in a debate, i.e. someone who 48 The German original reads: “Wörtlich: in unverkennbarer d.h. reiner, idiomatisch arabischer Sprache, dergleichen kein Fremder schreiben kann, folglich ist es eine Originaloffenbarung und nicht aus den frühern Schriften copirt. Vergl. 16, 105.” 49 See Kropp (2015) for a different understanding. Even though he comes to a similar conclusion regarding the meaning of [X mubīn] as “identifiably a case of X”, he read lisān ʿarabī mubīn as Ethiopian calque where “clearly Arabic language” is synonymous with “clearly in the language of revelation”. To me it seems that it is not necessary to invoke an Ethiopian calque to realize that in the context of these examples, the obviously Arabic character – whatever that meant – of the language serves as a guarantee that the message is indeed revealed and not some sort of plagiarism.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

76

Daniel Birnstiel

is unable to make himself (or what he says) identifiable. It does not refer to understanding the content of his arguments, but rather to the ability to perceive his arguments: (50) ʾa-wa-man yunaššaʾu fī l-ḥilyati wa-huwa fī l-ḫiṣāmi ġayru mubīnin (43:18) And someone who is brought up in finery, while being indiscernible in a contention? 3.7 Other forms of IV Example (51) is the only instance of a finite verb form of stem IV. Pharaoh is attempting to assert his superiority viz-à-viz Moses and claims that Moses is of low status and moreover almost impossible to understand. Although the verb form here relates to language, it does not seem to imply a lack of clarity of the actual words regarding their content or form; rather, Pharaoh seems to point towards the modality of Moses’ speaking, which due to his speech empediments – as indicated by the Qur’ān (20:27) and Exodus (4:10) – is almost not perceivable as speech. In other words, Pharaoh is mocking Moses for speaking in a manner that can only with difficulty be perceived as speech. (51) ʾam ʾana ḫayrun min hāḏā llaḏī huwa muhīnun wa-lā yakādu yubīnu (43:52) (Pharaoh says): Am I not better than this one here who is despicable and hardly discernible (in speech)? This sense is maintained in the variant reading attested for this passage with stem I yabīnu instead of stem IV (Al-Ḫaṭīb 1422/2002: 8.385).

4. bayyin- compared with mubīn Although no verbal or participle forms of stem I are attested in the corpus, a large number of derivates of the adjective form bayyin-, namely bayyinah and bayyināt, occur.50 This form, however, may be regarded as a quasi-participle 50 The variant readings attested in al-Ḫaṭīb’s encyclopedia for loci where the Cairo edition reads bayyinah or bayyināt following the recitation of Ḥafṣ from ʿĀṣim concern only cases of reading the singular bayyinah instead of the plural bayyināt in 3:97 and 29:49 (i.548; vii.117), reading bayyinatu mā (genitive construct) as bayyinatum-mā
bi-l-bayyinaṯ ṯumma in 2:92 and 5:32 (i.153; ii:265). 51 Additional occurrences of bayyināt are 2:87, 2:99, 2:159, 2:185, 2:209, 2:213, 2:253 (2x), 3:86, 3:97, 3:105, 3:184, 4:153, 5:32, 5:110, 7:101, 9:70, 10:13, 10:15, 10:74, 14:9, 16:44, 19:73, 20:72, 22:16, 22:72, 24:1, 29:39, 29:49, 30:9, 30:47, 34:43, 35:25, 40:22, 40:28, 40:50, 40:66, 40:83, 43:63, 45:25, 46:7, 57:25, 58:5, 61:6, and 64:6.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

78

Daniel Birnstiel

(52) wa-la-qad ʾātaynā mūsā tisʿa ʾāyātin bayyinātin fa-sʾal banī ʾisrāʾīla ʾiḏ ğāʾahum fa-qāla lahū firʿawnu ʾinnī la-ʾaẓunnuka yā mūsā masḥūrān (17:101) And indeed we gave Moses ten signs of self-evident nature! So ask the children of Israel! Behold, we came to them and Pharaoh said to him: “O Moses, indeed, I think that you are equipped with sorcery!” (53) fa-lammā ğāʾahum mūsā bi-ʾāyātinā bayyinātin qālū mā hāḏā ʾillā siḥrun muftaran wa-mā samiʿnā bi-hāḏā fī ʾābāʾinā l-ʾawwalīna (28:36) And when Moses came to them with signs of self-evident nature, they said: “This is invented sorcery and we have heard heard this thing from our forefathers of old.” (54) huwa llaḏī yunazzilu ʿalā ʿabdihī ʾāyātin bayyinātin li-yuḫriğakum mina ẓ-ẓulumāti ʾilā n-nūri ... (57:9) He is the one who sends on his slave signs of self-evident nature, in order to take you from from the darness into the light… In (54), the text asserts that the signs given to Muḥammad are of a kind the nature of which is self-evident; in other words, Muḥammad has been given signs whose divine origin is obvious, i.e. not demanding any mental activity on part of the audience in order to recognize this fact. The following examples show the independent usage of bayyināt: (55) wa-la-qad ğāʾakum yūsufu min qablu bi-l-bayyināti fa-mā ziltum fī šakkin mim-mā ğāʾakum bihī... (40:34) And indeed Joseph had come to them previously with the evident things, yet you did not cease to be in doubt about that which he brought … (56) wa-la-qad ğāʾakum mūsā bi-l-bayyināti ṯumma ttaḫaḏtumu l-ʿiğla min baʿdihi ... (2:92) And indeed Moses did come to you with the evident things. (But) thereafter you took the calf … In (55) the children of Israel are critisized for doubting the message brought to them by Joseph despite its self-evident and self-exlanatory nature. Example (56) is again an instance of incongruence between the obious and plain nature of the evidence and the subsequent behaviour of the people: although Moses has been given evidence that manifests the oneness of God and leaves no doubts regarding it, the children of Israel begin to worship an idol and disregard the self-evident existence of the one God. In (57), the verse draws a connection between the self-evident indications and the fact that after the knowledge had come, people began to disagree. The

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

79

text identifies bayyināt with knowledge. i.e. the things indicated are self-evident and of an immediately known, manifest nature: (57) wa-ʾātaynāhum bayyinātin mina l-ʾamri fa-mā ḫtalafū ʾillā min baʿdi mā ğāʾahumu l-ʿilmu baġyan baynahum … (45:17) And we gave them evident things regarding the issue, yet after the knowledge had come to them, they became divided out of jealousy among them… (58) llaḏīna qālū ʾinna llāha ʿahida ʾilaynā ʾallā nuʾmina li-rasūlin ḥattā yaʾtiyanā bi-qurbānin taʾkuluhu n-nāru qul qad ğāʾakum rusulun min qablī bi-l-bayyināti wa-bi-llaḏī qultum fa-li-ma qataltumūhum ʾin kuntum ṣādiqīna (3:183) As for those, who say: “Behold, God has made a covenant with us that we should not believe any messenger until he comes to us with a offering that is consumed by the fire, say: “A messenger has already come to you before me with evident things and with the very same thing you had demanded. Why then did you slay them, if you speak the truth?!” Example (58) counters the excuse given by the rejecters of the message as to why they cannot follow it by pointing out that a previous messenger not only fulfilled their demand, but also brought signs of a nature that speaks for itself; the text asserts that despite the fact that the demand had been met and despite the fact that they had received proofs the divine origin of which was evident and which therefore established the messenger as true, they still acted in an inacceptable way. 4.2 bayyinah “self-evident (thing)” Similarly, the singular bayyinah is used to refer to “evidence the nature of which is clearly established”, i.e. evidence the divine origin of which is undoubtable. Like its plural, it occurs either independently or else as attribut of ʾāyah “sign”. In (59), the opponents of the prophet Hud defend themselves by claiming that he did not bring a sign of an evident nature, i.e. of divine origin. Likewise, in (60), the disbelievers’ rejection of a sign is countered by the rhetorical question if they could continue to reject the message even if a proof of the kind known from the previous scriptures, whose heavenly origin is self-evident were given to them:52 52 Additional examples of bayyinah occur in 2:211, 6:57, 6:157, 7:73, 7:85, 7:105, 8:42

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

80

Daniel Birnstiel

(59) qālū yā hūdu mā ğiʾtanā bi-bayyinatin wa-mā naḥnu bi-tārikī ʾālihatinā ʿan qawlika wa-mā naḥnu laka bi-muʾminīna (11:53) They said: “O Hud! You did not bring us something self-evident, and we are not deserters of our gods on account of your word and we do not believe in you. (60) wa-qālū law-lā yaʾtīnā bi-ʾāyatin min rabbihī ʾa-wa-lam taʾtihim bayyinatu mā fī ṣ-ṣuḥufi l-ʾūlā (20:133) And they said: “Even if he came to us with a sign from his Lord!” Is it not the case that something evident of the kind that is recorded in the former scriptures has come to them? (61) qul ʾa-raʾaytum šurakāʾakumu llaḏīna tadʿūna min dūni llāhi ʾarūnī māḏā ḫalaqū mina l-ʾarḍi ʾam lahum širkun fī s-samāwāti ʾam ʾātaynāhum kitāban fa-hum ʿalā bayyinatin minhu … (35:40) Say: “Have you seen your partners whom you invoke apart from God? Show me what part of the earth they have created! Or do they have a share in the heavens?” Or have we given them a scripture and now they base themselves on some evident thing thereof? (62) ʾa-fa-man kāna ʿalā bayyinatin min rabbihī ka-man zuyyina lahū sūʾu ʿamalihī ... (47:14) Or is someone who bases himself some evident thing from his lord like someone to whom the evil of his action has become beautiful … In (61) and (62), bayyinah refers explicitly to the divine origin of the evidence. In the former case, the polytheists are rhetorically asked if they had been given a scripture furnishing them with an undoubtable piece of evidence for their behaviour. The latter example compares someone who bases himself on evidence the evidentiality of which cannot be doubted with someone whose evil actions have become beautiful to him, i.e. the text compares someone whose proof is self-evident and therefore must be accepted by all as proof and evidence with someone who is self-deluded. 4.3 bayyin There is also one occurrence of bayyin in the singular attested in the corpus (without any attested variant readings). Here the sleepers of the cave express their reservation against their people’s idol worship with their lacking an authorisation whose nature as divine authorisation was evident.

(2x), 11:17, 11:28, 11:63, 11:88, 29:35, 98:1, and 98:4.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

81

(63) hāʾulāʾi qawmunā ttaḫaḏū min dūnihī ʾālihatan law-lā yaʾtūna ʿalayhim bi-sulṭānin bayyinin fa-man ʾaẓlamu mim-mani ftarā ʿalā llāhi kaḏibān (18:15) Behold, our people have taken gods apart from him. Why do they not bring against them an evident authorisation? Who then is more in the wrong than someone who invents a lie against God? This last example can easily be contrasted with the attestations of sulṭān mubīn: while in (64) and (65) the authorization is characterized as being identifiable to the audience as such, the sleepers of the cave in (63) criticize their people for lacking an authorization which is self-evident. (64) wa-la-qad ʾarsalnā mūsā bi-ʾāyātinā wa-sulṭānin mubīnin (11:96; 40:23) And indeed, we did send Moses with our signs and an authority of identifiable nature. (65) ṯumma ʾarsalnā mūsā wa-ʾaḫāhu hārūna bi-ʾāyātinā wa-sulṭānin mubīnin (23:45) Thereafter we send Moses and his brother Aaron with our signs and an authorization of identifiable nature. These near minimal pairs corroborate the theory that mubīn (and consequenly stem IV of byn) indicate identifiability, i.e. the audience can work out on their own that the claimed authorization is indeed one, since its nature is such that it permits this identification, even if this is not clear at first sight. Stem I and bayyin with derivaties, on the other hand, indicate that something is self-evident, already identified, clear or known.

5. The semantics of stem II Both finite and nominal forms are attested for stem II. Since it has been established that stem I signals the evidentiality of something, while stem IV signals its identifiability, it may be assumed that stem II would indicate that something is made identifiable or caused to be evident. This is clearly borne out by the attested examples.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

82

Daniel Birnstiel

5.1 Finite verb forms Altogether 35 finite verbforms of stem II of byn occur in the Qurʾān.53 In (66), the Israelites request from Moses to ask God to make the heifer they are to slaughter identifiable or evident for them; in (67), the verse asserts that God has caused the guidance to be identifiable or evident in the book: (66) qālū dʿu lanā rabbaka yubayyin lanā mā hiya qāla ʾinnahū yaqūlu ʾinnahā baqaratun lā fāriḍun wa-lā bukrun ʿawānun bayna ḏālika fa-fʿalū mā tuʾmarūna (2:68, similarly 2:69, 2:70) They said: “Implore for us your Lord to make evident to us, what she is!” He said: “Behold, He says that she is a cow who is neither old nor youngt, but of an age in-between. So do what you have been ordered!” (67) ʾinna llaḏīna yaktumūna mā ʾanzalnā mina l-bayyināti wa-l-hudā min baʿdi mā bayyannāhu li-n-nāsi fī l-kitābi ʾulāʾika yalʿanuhumu llāhu wa-yalʿanuhumu l-lāʿinūna (2:159) Behold, as for those who conceal whatever we have sent down, namely evident things and guidance, after we have made it evident to the people in the scripture – it is such people that God curses as well as those who curse. (68) ʾillā llaḏīna tābū wa-ʾaṣlahū wa-bayyanū fa-ʾulāʾika ʾatūbu ʿalayhim wa-ʾanā t-tawwābu r-raḥīmu (2:160) except those who repent and set right their affairs and make (this) evident – for it is such people that I return to and I am the merciful, oft-returning one. Example (68) introduces an exception to those cursed in the previous example: such people are characterized by repentence, mending of ways and making their change publicly noticeable. Also here, bayyana cannot mean “to explain” in the sense of “explain and clarify the content of something”, rather it means here that the people who have repented must show their repentance, i.e. their change of heart must be such that it is perceptibe to, and perceived by the rest of the community. 53 The perfect is also attested in 2:118, 3:118, and 57:17, and the imperfect in 2:187, 2:221, 2:242, 2:266, 3:103, 4:176, 5:15, 5:19, 5:75, 5:89, 24:18, 24:58, and 24:61. The subjunctive also occurs in 4:26, 6:105, 9:115, 14:4, 16:39, 16:44, and 22:5, and the energetic in 3:187. The variant readings collected by al-Ḫaṭīb concern only cases of assimilation of the final -n to a following l-, e.g. yubayyil lanā < yubayyin lanā, e.g. in 2:69 or its lack (i:122f.), reading the plural bayyannāhu as singular bayyanahū in 2:159 (i.221), reading li-nubayyina as li-yabayyina in 22:5 (6.80), and similar things.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

83

In the following two examples, the verb refers to the matter of disagreement. It seems, however, that what is meant here is not that the details of the disagreement are to be explained and expounded, but rather that the addressees will be enabled to identify the exact issue and to realize its nature. In (69), the verse asserts that the revelation of the scripture took place inter alia in order to facilitate this identificaiton and as a means of guidance; scripture, by revealing the ultimate truths, unmasks the issues of strive and lays open their identity, thereby showing that the matter was utterly unworthy of causing disagreement. (69) wa-mā ʾanzalnā ʿalayka l-kitāba ʾillā li-tubayyina lahumu llaḏī ḫtalafū fīhi wa-hudan wa-raḥmatan li-qawmin yuʾminūna (16:64) And we have revealed to you the scripture only in order that you may make evident to them the (matter) about which they came to differ and as guidance and mercy for a people who believe. (70) wa-lammā ğāʾa ʿīsā bi-l-bayyināti qāla qad ğiʾtukum bi-l-ḥikmati wa-li-ʾubayyina lakum baʿḍa llaḏī taḫtalifūna fīhi fa-ttaqū llāha wa-ʾaṭīʿūni (43:63) And when Jesus came to them with the evident things, he said: “I have come to you with the wisdom, and in order that I shall make evident for you some of that, concerning which you disagree. So fear God and obey me!” (71) ... wa-la-yubayyinanna lakum yawma l-qiyāmati mā kuntum fīhi taḫtalifūna (16:92) … and certainly he will make evident to you on the day of Judgement wherein you were used to disagree. Example (82) constitutes a divine assertion that God will make evident to mankind on the day of judgement wherein they disagreed. Also here, what is meant is not that the particulars of the disagreement will be explained, but rather that the issue will be made manifest and known. Examples (72) is clearly an instance of “making visible”: The verse claims that God makes his boundaries manifest to people who know, i.e. he manifests them in such a way that people of knowledge can perceive and identify them. Similarly, in (73) and (74), the verse indicates that God makes manifest his signs or verse not in the sense of explaining or expounding them, but in the sense that by stating rules of behaviour as detailed in (74), he makes his signs identifiable as such. In other words, the wording of the verses is such that God facilitates thereby their identification and recognition as divine signs:54 54 ʾāyāt appears almost as some sort of default object of yubayyinu in several loci, e.g. in

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

84

Daniel Birnstiel

(72) fa-ʾin ṭallaqahā fa-lā taḥillu lahū min baʿdu ḥattā tankiḥa zawğan ġayrahu fa-ʾin ṭallaqahā fa-lā ğunāḥa ʿalayhimā ʿan yatarāğaʿā ʾin ẓannā ʾan yuqīmā ḥudūda llāhi wa-tilka ḥudūdu llāhi yubayyinuhā li-qawmin yaʿlamūna (2:230) And if he divorces her, she is not permitted to him thereafter, until she marries a husband other than him, and if he divorces her, then there can be no blame on the two of them, if they think that they will establish the boundaries of God. And these are the boundaries of God, which he makes evident for people who know. (73) ... ka-ḏālika yubayyinu llāhu lakumu l-ʾāyāti laʿallakum tatafakkarūna (2:219) … in such manner God makes evident to you the signs, so that you might reflect. (74) wa-ʾiḏā balaġa l-ʾaṭfālu minkumu l-ḥuluma fa-l-yastaʾḏinū ka-mā staʾḏana llaḏīna min qablihim ka-ḏālika yubayyinu llāhu lakum ʾāyātihī wa-llāhu ʿalīmun ḥakīmun (24:59) And when the children among you reach puberty, then let them ask for permission, just as those before them ask for permission. In such manner God makes evident to you his signs, for God is knowing and wise. The analysis of the finite verb forms of stem II has revealed that the context indeed points towards understanding the verb as indicating “to make identifiable”, “to make evident”. However, it seems that the distinction made between stem I (identified-ness) and stem IV (identifiability) is neutralized in stem II. Examples (67) and (71) for example, must contextually be understood as God having made the issue so manifest that it is impossible to be ignorant of it, while in (68) and (72) the recognition and identification is apparently only made possible. In other cases, such as (66) the matter remains undecided. 5.2 Participles of stem II Apart from the finite forms discussed above, there are also six attestations of the active feminine participle of stem II, three times in the singular (mubayyinah) and three times in the plural (mubayyināt). However, there are some variant readings to be considered here. In all three instances of the plural participle,

2:221, 2:266, 24:18, and 24:51, i.e. “signs that are made evident” stand in opposition to “signs that are self evident”.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

85

the qirāʾāt tradition offers the passive form mubayyanāt as an alternative reading.55 This reading, i.e. mubayyanāt, is clearly the grammatically correct reading. When reading the active participle mubayyināt, the commentators generally tend to substitute “lacking objects”, e.g. as aṭ-Ṭabarī (1422/2001: xvii.294) who explains mubayyināt in 24:34 as “bi-maʿnā ʾanna l-ʾāyāta hunna tubayyinu l-ḥaqqa wa-ṣ-ṣawāba li-n-nāsi wa-tahdīhim ʾilā l-ḥaqqi” [“meaning that it is the signs themselves that make clear truth and rightness to the people and guide them to the truth”].56 The active form would have to be regarded either as synonymous with mubīn, which does not harmonize well with the meaning established for stem II, as lacking its object, or else as intransitive or reflexive form with some “lacking” self-referential expression such as nafs- to be understood, much like ʾaslama “to surrender oneself” is often constructed with an elliptical wağhahū or nafsahū.57 A “reflexive interpretation” is offered by an-Nasafī (1419/1998: ii.505): “ʾaw min bayyana bi-maʿnā tabayyana wa-minhu l-miṯlu qad bayyana ṣ-ṣubḥu li-ḏī ʿaynayni”, i.e. “or it is derived from bayyana (stem II) in the sense of tabayyana (stem V); to this category belongs for example ‘dawn has become visible/made itself clear (stem II) to someone endowed with two eyes’”. However, given that all finite attestations of stem II either have an explicit direct object or else are elliptical with the direct object being retrievable from the context, a reflexive-intransitive of an active participle seems impossible; the same holds true for a transitive active participle with elliptic direct object. More over, two of the attestations of the participle, namely in 24:34 and 24:46 are framed by attestations of finite verb forms of stem II with ʾāyāt “signs” as direct objects (in 24:18, 24:58. 24:59. 24:61). It seems unlikely that the text would first use stem II as transtive, then switch to an intransitive use for the participle, and then switch back to the transitive use for the remaining finite forms. The conclusion that suggests itself is that reading the consonant text as active participle is a later misreading. In the context of recent findings regarding the language of the Qurʾān58 I would carefully advance the theory that Qurʾānic Arabic was not only lacking final short vowels, but also short vowels in word-internal open syllables as syllable peaks were syncopated. In other words, short vowels may have occurred only in closed syllable or else on the phonetic level as anaptyctic vowels. 55 Cf. al-Ḫaṭīb (1422/2002: vi.263f., 288; ix.512). 56 Cf. also al-Bayḍāwī (1418/1998: iv.106, 111). 57 See also section 4.2. above. 58 See Van Putten (2018) and Van Putten & Stokes (2018).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

86

Daniel Birnstiel

This would lead naturally to a neutralization of the distinction between active and passive participles in the derived stems in all forms apart from the masculine singular: masc. sgl. act. m(ə)bayyin vs. masc. sgl. pass. m(ə)bayyan but: fem.sgl. act./pass. m(ə)bayy(ə)nah, etc. All three examples (75) – (77) speak about signs which have been made evident or identifiable by God. As such, they are to be contrasted (a) with signs the divine origin of which is self-evident (bayyin-) and signs the nature of which allows the addressee to recognize them as signs (mubīn), though the idenfication is not conclusive but only possible; mubayyan on the other hand indicates that the identified-ness or identifiability is not naturally present, but rather the outcome of intervention. It might be claimed that in (77) the recitation of the signs is what causes them to become identifiable or manifest as being of divine origin by the audience.59 (75) wa-la-qad ʾanzalnā ʾilaykum ʾāyātin mubayyanātin wa-maṯalan mina llaḏīna ḫalaw min qablikum wa-mawʿiẓatan li-l-muttaqīna (24:34) And we have indeed sent down upon you signs which have been made evident/identifiable and an example from those who passed away before you and an admonition for the God-fearing. (76) la-qad ʾanzalnā ʾāyātin mubayyanātin wa-llāhu yahdī man yašāʾu ʾilā ṣirāṭin mustaqīmin (24:46) And we have indeed sent down signs which have been made evident/ identifiable, and God guides whomever he wishes on a straight path. (77) rasūlan yatlū ʿalaykum ʾāyāti llāhi mubayyanātin li-yuḫriğa llaḏīna ʾāmanū wa-ʿamilū ṣ-ṣāliḥāti mina ẓ-ẓulumāti ʾilā n-nūri ... (65:11) a messenger who recited to you the signs of God, something made evident/identifiable in order to bring out those who believe and do righteous deeds from the darkness into the light … In this context, it is important to note that the grammarian and commentator ʾAbū Ḥayyān employs the expression “ʾāyāt mubayyanāt” with a meaning that is close to the interpretation given above in his commentary on 2:118 “… qad bayyannā l-ʾāyāti li-qawmin yūqinuna” [“… we have indeed made 59 The text has been changed from the reading of Ḥafṣ in the relevant cases, replacing mubayyināt with mubayyanāt.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

87

manifest the signs to people who are filled with certainty”] (1413/1993: i.): “ʾay ʾawḍaḥnā l-ʾāyāti … wa-hiya ʾāyātun mubayyanātun lā labsa fīhā wa-lā šubhata li-šiddati ʾiḍāḥihā lākin lā yaẓharu kawnuha ʾāyātun ʾillā li-man kāna mūqinan” [“This means: we have made evident/visible the signs, … and they are mubayyanāt signs, without confusion and uncertainty regarding them due to the strength of their manifestation, although their being signs is only apparent to someone who is filled with certainty.”] The feminine singular forms of the stem II participle mubayyinah appears in all three instances as attribute of fāḥišah “an act of sexual misconduct”. Also here variant readings are attested in the qirāʾāt literature. Thus al-Ḫaṭīb (1422/2002: 2.42) records the forms mubayyinah, mubayyanah, and mubīnah for 4:19; for 33:30 and 65:1, the variant readings mubayyinah and mubayyanah have likewise been attested (al-Ḫaṭīb 1422/2002: 7.277, 9.500). The variant reading mubīnah would obviously indicate an act of sexual misconduct that is perceivable. Ar-Rāzī (1401/1981: xxx.33) offers the following two variant based interpretations in his commentary on 65:1: “quriʾa bi-fāḥišatin mubayyinatin wa-mubayyanatin fa-man qaraʾa mubayyinatin bi-l-ḫafḍi fa-maʾnāhu ʾanna nafsa l-fāḥišati ʾiḏā tafakkara fīhā tabayyana ʾannahā fāḥišatun wa-man qaraʾa mubayyanatun bi-l-fatḥ fa-maʾnāhu ʾannahā mubarhanatun bi-l-barāhīni wa-mubayyanatun bi-l-ḫuğaği”, i.e. “one reads ‘bi-fāḥišatin mubayyinatin’ and ‘mubayyanatin’; if someone reads mubayyinatin (active form), then the meaning is that when the sexual misconduct is contemplated, it becomes evident by itself that it is sexual misconduct, and if someone reads mubayyanatun (passive form), then the meaning is that it (i.e. its occurrence) has been demonstrated by proofs and made manifest by evidence”. The intransitive, reflexive reading is also supported by an-Nasafī’s commentary (1419/1998: iii.28) of 33:30: “bayyinatin - ẓāhiri fuḥšihā”, i.e. “mubayyinatin – (a sexual misconduct) the indecency of which is evident, derived from bayyana (stem II) in the meaning tabayyana (stem V)”.60 The arguments mentioned with regard to ʾāyāt mubayyināt/mubayyanāt hold also true in these three cases, i.e. the lack of explicit direct objects following the participle and the exclusively transitivity of the finite verb forms. Consequently, the reading mubayyinah is to be regarded as a mistake and to be substituted by the variant reading mubayyanah as has been done below. A 60 See also al-Ḫafāğī (1974: iii.118) on 33:30 and Ḥaqqī (1330/1912–1347/1928: ii.181) on 4:19. Cf. also aš-Šawkānī’s gloss on 33:30 (1428/2007: 1166): “ʾay ẓāhirati l-qubḥi wāḍiḥati l-fuḥši” [“i.e. (a sexual misconduct) the ugly nature of which is apparent, the indecency of which is evident”].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

88

Daniel Birnstiel

fāḥišah mubayyanah is an act of sexual misconduct that has been made evident, i.e. it has become manifest and known:61 (78) ... wa-lā taʿḍulūhunna li-taḏhabū bi-baʿḍi mā ʾātaytumūhunna ʾillā ʾan yaʾtīna bi-fāḥišatin mubayyanatin ... (4:19) … and do not place constraints on them in order to take away something you have given them, except in the case that they commit an act of sexual misconduct that is made evident … (79) yā nisāʾa n-nabiyyi man yaʾti minkunna bi-fāḥišatin mubayyanatin yuḍāʿaf lahā l-ʿaḏābu ḍiʿfayni wa-kāna ḏālika ʿalā llāhi yasīrān (33:30) O wives of the Prophet! Whosever among you commits an act of sexual misconduct that is made evident, the punishment will be doubled for her … (80) ... lā tuḫriğūhunna min buyūtihinna wa-lā yaḫruğna ʾillā yaʾtīna bi-fāḥišatin mubayyanatin ... (65:1) … do not expell them from your houses and they shall not leave, unless they commit an act of sexual misconduct that is made evident …

6. The semantics of stem V Stem V is attested altogether 18 times: 12 times in the perfect, 3 times in the subjunctive and 3 times in the imperative. Since stem V usually functions as reflexive of stem II, the expected meaning of tabayyana is “to make oneself distinguishable/identifiable/evident; to become identifiable/evident”.62 At first glance, it seems problematic to apply this sense to all examples. A few examples feature the construction tabayyana X min Y “X has become/made itself distinguishable/distinct from” which clearly fits the established meaning of the root and the expected meaning of stem V:63 61 The occurrence of mubīnah as variant reading 4:19 is unlikely to be correct in the face of the better attestation of mubayyanah and the occurrence of mubayyanāt in a nearly identical contect. 62 A number of variants connected to assimilatory processes across word boundaries are mentioned, e.g. yatabayyallakum < yatabayyana lakum in 2:187 (similarly in 2:259, 9:34, 9:113, 9:114, 29:38, 41:53, 47:25, 47:32) or qat tabayyana < qad tabayyana in 2:256 (similarly in 29:38) are mentioned by al-Ḫaṭīb (1422/2002: i.261, 361; iii.394, 468, 470; vii.113; viii.300; ix.26, 32). 63 Similarly in 2:256.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

89

(81) ... wa-kulū wa-šrabū ḥattā yatabayyana lakumu l-ḫayṭu l-ʾabyaḍu mina l-ḫayṭi l-ʾaswadi mina l-fağri ... (2:187) … so eat and drink until the white thread of dawn makes itself/becomes distinguishable to you from the black thread … In the majority of the remaining cases, the verb is intransitive, featuring either an expliticly mentioned subject, an anaphorically retrievable subject or else referring to the preceding propositional content as subject. In these cases, the expected meaning “to make oneself distinguishable/identifiable/evident; to become identifiable/evident” is clearly recognizable:64 (82) ʿafā llāhu ʿanka lima ʾaḏinta lahum ḥattā yatabayyana laka llaḏīna ṣadaqū wa-taʿlama l-kāḏibīna (9:43) May God forgive you! Why did you grant them permission until those who told the truth had made themselves evident to you and you recognized the liars? (83) sa-nurīhim ʾāyātinā fī l-ʾafāqi wa-fī ʾanfusihim ḥattā yatabayyana lahum ʾannahu l-ḥaqqu … (41:53) We will show them our signs on the horizons and in themselves, until it becomes evident to them that it is the truth … (84) wa-ʿādan wa-ṯamūda wa-qad tabayyana lakum min masākinihim … (29:38) (God has destroyed a number of previous sinful nations) And also (the people of) ʿĀd and (the people of) Ṯamūd. And (the fact the God has destroyed them) has become clear to you from their (abandoned) dwellings ... In all these cases, the semantic element of “identification/identity” is obvious. As in stem II, the opposition between being evident (stem I) and being identifiable (stem IV) seems to be neutralized. In a few cases, a passive reading tubuyyina is attested instead of the active tabayyana:65 64 Additional instances are 4:115, 9:113, 9:114, 47:25, 47:32. 65 For 2:109, 2:259 and 8:6, al-Ḫaṭīb (1422/2002: i.176, 371f.; iii.260) also lists the reading buyyina, i.e. a passive form of stem II. This reading is contextually possible, but runs counter the consonant text tbyn and the underlying rasm (‫)ٮٮٮں‬. For 14:45, al-Ḫaṭīb (1422/2002: iv.512f.)) also lists the variants nubayyinu and nubayyin „we will make evident/clear“, which is contextually possible and in harmony with both the consonant text and the underlying rasm, but rather unlikely with regard to the tense-aspect structure of the passage.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

90

Daniel Birnstiel

(85) … min baʿdi mā tabayyana lahumu l-ḥaqqu … (2:109) … after the truth has become/made itself evident to them… (86) … fa-lammā tabayyana lahū qāla ʾaʿlamu ʾanna llāha ʿalā kulli šayʾin qadīrun (2:259) (God tests a person by making him die and reviving him for a hundred years. Then he lets him guess how long he was without consciousness and clarifies to him the actual amount of time that has passed, explaining that the resurrection of the dead will proceed likewise) … and when (this) had become clear to him (i.e. that man), he said: “I know that God is capable of anything!” (87) yuğādilūnaka fī l-ḥaqqi baʿda mā tabayyana ka-ʾannamā yusāqūna ʾilā l-mawti wa-hum yanẓurūna (8:6) arguing with you about the truth after it has become/made itself evident, just as if it they are being driven towards death, while they are looking on! (88) wa-sakantum fī masīkini llaḏīna ẓalamū ʾanfusahum wa-tabayyana lakum kayfa faʿalnā bihim … (14:45) And you dwelt in the dwelling places of those, who were injust against themselves. And it has become evident to you how we dealt with them … Although it is tempting to try to read the passive as “having been made evident/identifiable” in opposition to an active “having become evident/identifiable”, such a distinction could only be established after examinating active and passive verb forms of other verbs in stem V (and other reflexive stems). Example (89) is problematic at first sight, since it appears that tabayyana is constructed transitively with an explicit subject (al-ǧinn) and an explicit object (ʾan law kānū yaʿlamūna l-ġayba mā labiṯū fī l-ʿaḏābi l-muhīni), indicating something along the lines of “it had become evident to them that” or “they had realized that”. A closer examination reveals that stem V here has the regular meaning of “becoming/making itself evident”. The passage describes who Solomon made the Jinn subservient to his will and had them perform all sorts of works for him. Even though Jinn are supposed to be conginizant of hidden facts, they were unaware of Solomon’s death. Rather, his death became only known, when a worm had gnawed at his staff and his corpse had fallen to the ground. Since the Jinn had continued working at his projects in humiliation, they showed their true character, when

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

91

Solomon’s corpse fell to the ground, because everyone realized at this point that the claim of the Jinn’s knowledge of the unseen was untrue:66 (89) fa-lammā qaḍaynā ʿalayhi l-mawta mā dallahum ʿalā mawtihī ʾillā dābbatu l-ʾarḍi taʾkulu minsaʾatahū fa-lammā ḫarra tabayyanati l-ğinnu ʾan law kānū yaʿlamūna l-ġayba mā labiṯū fī l-ʿaḏābi l-muhīni (34:14) And when we decreed death for him, only the animal of the earth eating his staff showed them (that he had died). And when he fell, the Jinn made themselves evident, (meaning) that if they had known the hidden, they would not have lingered in (the state of) offensive punishment! As mentioned above, stem V is attested also three times in the imperative (masculine plural), in 4:94 (twice) and 49:6). These cases also seem to be instances of an underlying transitive construction with an eliptic object “the matter, the affair”. As shown be the context, the addressees are urged to examine the matter at hand before making any claims that may turn out to be wrong. It seems difficult to reconcile the notion of “making sure, assuring oneself” with the established meaning of stem V “to make oneself distinguishable/identifiable/ evident; to become identifiable/evident”. However, according to al-Ḫaṭīb (1422/2002: i.131; ix.79), the variant reading taṯabbatū instead of tabayyanū is well established for all three instances. This reading concurs with the underlying rasm (‫)اوٮٮٮٮ‬. The root ṯbt “(to be) firm, strong” is well established in the Qurʾān. It occurse e.g. in stem II in 2:250, 3:147, 8:22, 8:12, 11:120, 14:27, 16:102, 17:74 in the sense of “making strong, making firm, strengthening”. A semantic devellopment of stem V from “making oneself firm, strengthening oneself” to “assuring oneself, making sure, ascertaining” seems plausible. It seems therefore grammatically preferable to read taṯabbatū in all three instances, as follows: 66 Many reciters seem also to have regarded this as a difficult passage, several of which even change the consonant text, as evidenced by the variant readings given by al-Ḫaṭib (1414/2002: vii.349f.): tubuyyinati l-ǧinnu [the Jinn had made themselves eviden]; tabayyanati l-ǧinna [(the fact that, if they had known …) has made evident the Jinn (?)], tabāyanat al-ʾinsu [it has become clear to humans (that the Jinn do not have secret knowledge)], tabayyanati l-ʾinsu l-ǧinna [humans have recognized the Jinn as what they are (?)], tabayyanati l-ʾinsu ʾan law kāna l-ǧinnu yaʿlamūna l-ġayba [humans have realized that that if the Jinn know the hidden, …] and tabayyanati l-ʾinsu ʾanna l-ǧinna law kānū yaʿlamūna l-ġayba [humans have realized that that if the Jinn know the hidden, …]. The grammaticality of these readings is rather doubtful considering the other attestations of stem V.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

92

Daniel Birnstiel

(90) yā ʾayyuhā llaḏīna ʾāmanū ʾiḏā ḍarabtum fī sabīli llāhi fa-taṯabbatū wa-lā taqūlū li-man ʾalqā ʾilaykumu s-salāma lasta muʾminan tabtaġūna ʿaraḍa l-ḥayāti d-dunyā fa-ʿinda llāhi maġānimu kaṯīratun ka-ḏālika kuntum min qablu fa-manna llāhu ʿalaykum fa-taṯabbatū ʾinna llāha kāna bi-mā taʿmalūna ḫabīrān (4:94) O you who believe, when you set out (to war) in the path of God, assure yourselves and do not say to anyone who offers you (the greeting of) peace “You are not a believer”, while you desire the petty goods of the earthly life. For plentiful booties lie with God. You were suchlike before, but then God granted you favours. So assure yourselves. Behold, God is cognizant of what you do! (91) yā ʾayyuhā llaḏīna ʾāmanū ʾin ğāʾakum fāsiqun bi-nabaʾin fa-taṯabbatū ʾan tuṣībū qawman bi-ğahālatin fa-tuṣbiḥū ʿalā mā faʿaltum nādimīna (49:6) O you who believe! When a deliquent brings you a piece of news, then assure yourselves, lest you cause people harm out of ignorance and thus become filled with remorse for what you have done. Apart from these last cases, in which the variant reading taṯabbatū seems to be more likely and original, all attestations of stem V fit the expected meaning of tabayyana “to make oneself distinguishable/identifiable/evident; to become identifiable/evident”

7. The semantics of stem X It is very difficult to draw any distinction between stem IV and X or stem V and X because there are only two attestations of stem X in the Qur’ān.67 A reflexive verb derived from stem IV “to be identifiable” would somewhat conincide in meaning with a reflexive form of stem II “to make identifiable/ cause to be identified”, i.e. “to make oneself identifiable/cause oneself to be identified”. Moreover, although stem X was originally a reflexive counterpart to the Semitic causative stem (*safʿala : istafʿala), the development of the causative prefix (*sa- > *ha > ʾa-) in Arabic resulted largely in a dissolution of the relationship between these two stems. Absent a thorough investigation of stem X in the Qurʾān, this stem generally seems to indicate some personal involvement of the agent, often with desiderative, declarative, or conative function. It seems possible to understand 67 No variant readings are given by al-Ḫaṭīb.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

93

stem X in these two examples as expressing something like “to prove oneself as identifiable”: (92) wa-ʾātaynāhumā l-kitāba l-mustabīna (37:117) And we gave the two of them the scripture of a nature which show itself to be identifiable. (93) wa-ka-ḏālika nufaṣṣilu l-ʾāyāti wa-li-tastabīna sabīlu l-muğrimīna (6:55) And in such fashion we are wont to expound the signs, and in order that the path of the sinners might show itself recognizable. In example (92), the reference is to the Torah which is characterized as a scripture which proves itself by being recognizable as such, i.e. the “personal involvement” somehow makes the entire form “heavier” and “more salient”, expressing a notion of successful and proven recognizability. The same is true in example (93). Such an understanding seems to underly the explanation of Ibn ʿĀšūr (1404/1984: xxiii.164) on 37:117: “al-mustabīna l-qawīyu l-wuḍūḥi fa-s-sīnu wa-t-tāʾu li-l-mubāliġati yuqālu stabāna š-šayʾu ʾiḏā ẓahara ẓuhūran šadīdan” [“mustabīn means the something, distinctness of which is firm and the sīn and tā (i.e. the use of the stem X) serve as means of hyperbolism. One says istabāna š-šayʾu when it strongly evident/manifest].68

8. Remaining derivatives of byn The remaining derivations of the root byn that have yet to be discussed comprise three instances of bayān and one occurrence of tibyān. The former is formally a verb noun (maṣdar) of stem I,69 the latter a suppletive verbal abstract functioning as verb noun of stem II.70 However, it seems that in most of these cases, the meaning of the verb noun has become concretized71 and adopted a meaning similar to that of the participles (mubīn) or verbal adjectives (bayyin, bayyinah), although the scope of reference is larger, i.e. referring to the entire 68 Several variants are attested for 6:55 by al-Ḫaṭīb (1422/2002: ii.438f.), namely li-yastabīna s-sabīlu, which is due to constructing sabīl as masculine noun, and li-tastabīna/ yastabīna s-sabīla, which constructs the verb as transitive with sabīl as direct object; in this case, the subject pronoun inherent in the finite verb form refers to Muḥammad either as addressee or else in the third person. 69 Cf. Chouémi (1966: 151), Lane (1863–1893: i.288). 70 Cf. Brockelmann (1908: 385), Lane (1863–1893: i.289). 71 This can be compared to the development of Latin mansio “the act/fact of staying” > “the

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

94

Daniel Birnstiel

corpus of the revelation or a larger part thereof, etc. In examples (94) and (95), bayān is apparently equivalent to the revelation, i.e. the Qur’ān or a part thereof; its being a revelation, i.e. being of divine origin is placed in focus by calling it “something of self-evident nature”: (94) hāḏā bayānun li-n-nāsi wa-hudan wa-mawʿiẓatun … (3:138) Here is something of self-evident nature, namely a guidance and an admonition … (95) ʿallamahu l-bayāna (55:4) He taught him a self-evident thing. (96) ṯumma ʾinna ʿalaynā bayānahū (75:19) Moreover, (ensuring) its being a thing of a self-evident nature lies upon us. (97) wa-nazzalnā ʿalayka l-kitāba tibyānan li-kulli šayʾin wa-hudan wa-raḥmatan wa-bušrā … (16:89) … and we have revealed unto you the scripture as something that makes everything evident, a guidance, a mercy and glad tidings ... In (96), the parallelism with the preceding verse 75:17 makes it clear that it is God’s duty to ensure the collection and recitation of the Qurʾān as well as “its being of a self-evident nature”. In (97), tibyān is used as a concrete noun referring to the role of the scripture in its entirety as something which enables to recognition of everything, i.e. something that permits the audience and readers upon reflection to identify and recognize the nature of everything.

9. Conclusions Given the evidence presented throughout this article, it becomes possible to draw a comprehensive picture of the semantics of the root byn as attested in the Qur’ān, based on contextual analysis and comparative Semitics. The semantic field to which the root and its derivatives belong may be termed as “distinctness, identity, perceptibility”; the semantics of this root refer thus to the faculties of perception and identification of entities in the sense of establishing and identifying their nature. They do not refer to the elucidation and comprehension of contents, details or modalities.

place of staying” > French maison ‘house’, or Latin exercitus “the act of exercising” > “exercise” > “army”.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

95

Furthermore, stem I and IV are both intransitive, referring to factuality (of identity) and possibility (of identification) respectively. To this may be added stem X with an element of “inner participation”, expressing possibly a heightened salience of the manifest-ness. Stem II and V stand in opposition to these intransitive stems and express “causation” or “bringing about of perceived-ness and perceptibility”, with a distinction of non-reflexive vs. reflexive. In addition to showing that the commonly accepted readings mubayyinah/ mubayyināt as well as tabayyanū are likely wrong and should be exchanged for the attested variant reading mubayyanah/mubayyanāt and taṯabbatū respectively, the two most important findings of this contribution are (a) that bayyinah/bayyināt qualifies entities that are characterized as possessing a self-evident nature, i.e. their being Divine signs and indications is self-evident, and (b) that attributive mubīn qualifies its head (nucleus) as an entity whose nature is discernible. Thus, contrary to the common translations and the wide-spread understanding, the Qur’ān does not claim that it is a clear book, i.e. a book that clarifies matters or that portrays its message in a clear, i.e. easily comprehensible manner. Furthermore, when the Qur’ān asserts that it is in lisān ʿarabī mubīn, it is only claiming that the Arab-ness and Arabic-ness of its language is clearly manifest and discernible, whatever Arab-ness and Arabic-ness may have meant in the 7th century Hiǧāz. The Qur’ān neither states that it is clear nor that it is clarifying. It does, however, claim to be clearly Arabic.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

96

Daniel Birnstiel

Appendix: Attestations of mubīn with their loci

nucleus

translation

number

loci

ḍalāl

error, going astray 19

3:164, 6:74, 7:60, 12:8, 19:38, 21:54, 28:85, 34:24, 36:24, 36:47, 43:40, 46:32, 62:2, 26:97, 33:36

kitāb

book, scripture

12

indef: 5:15, 6:59, 10:61, 11:6, 27:1, 27:75, 34:3 (7 times); def: 12:1, 26:2, 28:2, 43:2, 44:2 (5 times)

naḏīr

warner

12

indef: 7:184, 11:25, 22:49, 26:115, 29:50, 38:70, 46:9, 51:50, 51:51, 67:26, 71:2 (11 times); def: 15:89 (once)

sulṭān

authority

12

4:91, 4:144, 4:153, 11:96, 14:10, 23:45, 27:21, 37:156, 40:23, 44:19, 51:38, 52:38

ʿadūw

enemy

10

2:168, 2:208, 4:101, 6:142, 7:22, 12:5, 17:53, 28:15, 36:60, 43:62

siḥrun

magic

9

5:110, 6:7, 10:76, 11:7, 27:13, 34:43, 37:15, 46:7, 61:6

al-balāġ

communication

7

5:92, 16:35, 16:82, 29:18, 36:17, 64:12

ʾiṯm

sin, guilt

4

4:20, 4:50, 4:112, 33:58

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

12:30, 31:11, 39:22, 67:29,

24:54,

Neither Clear Nor Clarifying – Yet Clearly Arabic

97

nucleus

translation

number

loci

al-balāʾ

trial, affliction

2

indef: 44:33 (once); def: 37:106 (once)

al-fawz

triumph

2

6:16, 45:30

al-ḥaqq

truth

2

27:79, 24:25

al-ḫusrān

loss

3

indef: 4:119 (once); def: 22:11, 39:15 (twice)

ḫaṣīm

quarrelsome

2

16:4, 36:77

ʾimām

faith

2

15:79, 36:12

lisān ʿarabī Arabic language

2

16:10, 26:195

qurʾān

reciation, Qurʾān

2

015:001, 036:069

rasūl

messenger

2

43:29, 44:13

ṯuʿbān

snake, serpent

2

7:107, 26:32

al-faḍl

favour

1

27:16

duḫān

smoke

1

44:10

fatḥ

victory

1

48:1

ġawī

(gone) astray

1

28:18

ġayr

other, un-

1

43:18

ʾifk

lie, slander

1

24:12

kafūr

ingrate, full

unthank- 1

43:15

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

98

Daniel Birnstiel

nucleus

translation

number

loci

nūr

light

1

4:174

sāḥir

magician

1

10:2

šayʾ

something

1

26:30

šihāb

flame

1

15:18

ʾufuq

horizon

1

81:23

ẓālim li-naf- someone wronging 1 sihī himself

37:113

Bibliography ʾAbū Ḥayyān, Muḥammad b. Yūsuf. 1413/1993. Tafsīr al-baḥr al-muḥīṭ. Edited by ʿĀdil ʾAḥmad ʿAbd al-Mawjūd and ʿAlī Muḥammad Muʿawwaḍ. Eight vols. Bayrūt: Dār al-Kutub al-ʿIlmiyyah. ʾAbū Suʿūd b. Muḥammad. n.d. ʾIršād al-ʿaql al-salīm ʾilā mazāyā al-kitāb al-karīm. Edited by ʿAbd al-Qāder ʾAḥmad ʿAṭāʾ. Five vols. Ar-Riyāḍ: Maktabat ar-Riyāḍ al-Ḥadīṯah. ʾAbū ʿUbaydah, Maʿmar b. al-Muṯannā. 1981/1401. Mağāz al-qurʾān. Edited by Muḥammad Fuʾād Sezgīn. Two vols. Second edition. Bayrūt: Muʾassasat ar-Risālah. al-ʿAkbarī, ʾAbū al-Baqāʾ ʿAbd Allāh b. al-Ḥusayn. n.d. At-tibyān fī ʾiʿrāb al-qurʾān. Edited by ʿAlī Muḥammad al-Buḫārī. Two vols. Al-Qāhirah: Dār ʾIḥyāʾ al-Kutub al-ʿArabiyyah. al-Bayḍāwī, ʾAbū al-Ḫayr ʿAbd Allāh b. ʿUmar. 1418/1998. ʾAnwār at-tanzīl wa-ʾasrār at-taʾwīl. Edited by Muḥammad ʿAbd ar-Raḥman al-Marʿašlī. Five vols. Bayrūt: Dār ʾIḥyāʾ at-Turāṯ al-ʿArabī. al-Ġazālī, ʾAbū Ḥāmid Muḥammad. 1413/1993. Al-Mustaṣfā. Edited by. Muḥammad ʿAbd as-Salām ʿAbd aš-Šāfī. Bayrūt: Dār al-Kutub al-ʿIlmīyah. al-Ḫafāğī, Šihāb al-Dīn ʾAḥmad b. Muḥammad. 1974. ʿInāyat al-qāḍī wa-kifāyat

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

99

ar-rāḍī = Ḥāšiyat aš-šihāb ʿalā tafšir al-Bayḍāwī. Eight vols. Bayrūt: Dār Ṣādir. al-Ḫaṭīb, ʿAbd al-Laṭīf Muḥammad. 1422/2002. Muʿğam al-qirāʾāt. Eleven Volumes. Al-Qāhirah: Dār Saʿd ad-Dīn. al-Ḫāzin, ʾAbū al-Ḥasan ʿAlāʾ ad-Dīn ʿAlī b. Muḥammad. 1415/1994. Lubāb at-taʾwīl fī maʿānī at-tanzīl. Edited by Muḥammad ʿAlī Šāhīn. Four vols. Bayrūt: Dār al-Kutub al-ʿIlmīyah. al-Kafāwī, ʾAbū al-BaqāʾʾAyyūb b. Mūsā. 1419/1998. Al-kullīyāt: muʿğam fī l-muṣṭalaḥāt wa-l-furūq al-luġawīyah. Edited by ʿAdnān Darwīš and Muḥammad al-Miṣrī. Bayrūt: Muʾassasat ar-Risālah. al-Maḥallī, Ğalāl ad-Dīn and as-Suyūṭī, Ğalāl ad-Dīn. N.d. Al-qurʾān al-karīm wa-bi-hāmišihi tafsīr al-Ğalālayn. Edited by ʿAbd al-Qādir al-ʾArnāʾūṭ. Dimašq: Dār Ibn Kaṯīr. al-Māwardī, ʾAbū l-Ḥasan ʿAlī b. Muḥammad b. Ḥabīb. n.d. An-nukat wa-l-ʿuyūn. Edited by as-Sayyid b. ʿAbd al-Maqṣūd b. ʿAbd ar-Raḥīm. Six vols. Bayrūt: Dār al-Kutub al-ʿIlmiyyah. al-Qaysī, ʾAbū Muḥammad Makkī b. ʾAbī Ṭālib. 1429/2008. Al-Hidāyah ʾilā bulūġ an-nihāyah. Thirteens vols. Aš-Šāriqah (Sharjah): Kulliyyat al-Dirāsāt al-ʿUlyā. an-Nasafī, ʾAbū al-Barakāt ʿAbdallāh b. ʾAḥmad. 1419/1998. Madārik at-tanzīl wa-ḥaqāʾiq at-taʾwīl. Edited by Yūsuf Bidīwī and Muḥyi ad-Dīn Dīb Mistū. Three vols. Bayrūt: Dār al-Kalam aṭ-Ṭayyib. an-Nīsabūrī, Niẓām ad-Dīn al-Ḥasan b. Muḥammad. 1416/1996. Ġarāʾib al-qurʾān wa-raġāʾib al-furqān. Edited by Zakarīyā ʿAmīrāt. Six vols. Bayrūt: Dār al-Kutub al-ʿIlmiyyah. ar-Rāzī, Muḥammad b. ʿUmar. 1401/1981. Mafātīḥ al-ġayb. Thirty-three vols. Bayrūt: Dār al-Fikr. as-Suyūṭī, Ğalāl ad-Dīn. 2005/1426. Al-ʾitqān fī ʿulūm al-qurʾān. Seven vols. Al-Madīnah al-Munawwarah: Markaz ad-Dirāsāt al-Qurʾāniyyah/Muğammaʿ al-Malik Fahd. aṣ-Ṣanʿānī, ʿAbd ar-Razzāq b. Hammām. 1410/1989. Tafsīr al-qurʾān. Edited by Muṣṭafā Muslim Muḥammad. Three vols. Ar-Riyāḍ: Dār ar-Rušd. aš-Šawkānī, Muḥammad b. ʿAlī. 1428/2007. Fatḥ al-qadīr al-ğāmiʿ bayna fannay ar-riwāyah wa-d-dirāyah min ʿilm at-tafsīr. Edited by Yūsuf al-Ġuš. Bayrūt: Dār al-Maʿrifah. aṭ-Ṭabarī, ʾAbū Ğaʿfar Muḥammad b. Ğarīr. 1422/2001. Tafsīr aṭ-Ṭabarī – ǧāmiʻ al-bayān ʻan taʼwīl ʾāy al-qurʼān. Edited by ʿAbd Allāh b. ʿAbd al-Muḥsin atTurkī. Twenty-six vols. Al-Ğīzah: Dār Ḥiğr. aṯ-Ṯaʿlabī, ʾAbū ʾIsḥāq ʾAḥmad b. ʾIbrāhīm. 1422/2002. Al-kašf wa-l-bayān ʿan tafsīr al-qurʾān. Edited by ʾAbū Muḥammad b. ʿĀšūr and Naẓīr al-Sāʿidī. Ten vols. Bayrūt: Dār ʾIḥyāʾ at-Turāṯ al-ʿArabī. az-Zabīdī, Muḥammad Muṣṭafā al-Ḥusaynī. 1385/1965–1422/2001. Tāğ al-ʿarūs.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

100

Daniel Birnstiel

Edited by ʿAbd as-Sattār ʾAḥmad Farāğ & al. Forty vols). Al-Kuwayt: Matbaʽat Ḥukūmat al-Kuwayt. az-Zamaḫšarī, Ğār Allāh ʾAbū al-Qāsim Maḥmūd b. ʿUmar. 1418/1998 Al-kaššāf ʿan ḥaqāʾiq ġawāmiḍ at-tanzīl wa-ʿuyūn al-ʿaqāwīl fī wuğūh at-taʾwīl. Edited by ʿĀdil Aḥmad ʿAbd al-Wuğūd and ʿAlī Muḥammad Muʿawwiḍ. Six vols. Ar-Riyāḍ: Maktabat al-ʿAbīkān. Ambros, Arne and Stephan Procházka . 2004. A Concise Dictionary of Koranic Arabic. Wiesbaden: Reichert. Ambros, Arne. 1998. “Arabisch kontra Nabatäisch: Überlegungen zur Prähistorie des Prestige der ʿArabīya.” In Annäherung an das Fremde (Vorträge XXVI. Deutscher Orientalistentag vom 25. bis 29.9.1995 in Leipzig), edited by Holger Preissler and Heidi Stein, 167–172. Stuttgart: Franz Steiner. Badawi, Elsaid M. and Muḥammad Abdel Haleem. 2008. Arabic-English Dictionary of Qur’anic Usage. Leiden: Brill. Bär, Jochen A. 2002. “Kürze als grammatisches Problem: determinative Verschränkungen. Phänomene der Ersparung im Übergangsbereich von Wortbildung und Syntax.” In Das Wort. Seine strukturelle und kulturelle Dimension. Festschrift für Oskar Reichmann zum 65. Geburtstag, edited by Vilmos Ágel, Andreas Gardt, Ulrike Haß-Zumkehr and Thorsten Roelcke, 310–338.Tübingen. Max Niemeyer. Beeston, Alfred F.L., M.A. Ghul, W.W. Müller and J Ryckmans. 1982. Sabaic Dictionary (English-French-Arabic). Louvain-la-Neuve: Édition Peeters. Beeston, Alfred F.L. 1970. The Arabic Language Today. London: Hutchinson and Co. Bergmann, Rolf. 1980. “Verregnete Feriengefahr und Deutsche Sprachwissenschaft. Zum Verhältnis von Substantivkompositum und Adjektivattribut.” Sprachwissenschaft 5: 234–265. Biella, Joan Copeland. 1982. Dictionary of Old South Arabian (Sabaean Dialect). Chicago: Scholars Press. Birnstiel, Daniel. 2010. Selected Features of Arabic Syntax in the Qur’ān. Unpublished PhD Dissertation). Cambridge: Cambridge University. Birnstiel, Daniel. Forthcoming. “Classical Arabic.” In: The Semitic Languages, editied by John Huehnergard and Na’ama Pat-El. Second edition. New York: Routledge.. Bjøru, Øyvind. 2014. “Transitivity and the binyanim.” In Proceedings of the Oslo-Austin Workshop in Semitic Linguistics Oslo, May 23 and 24, 2013. Edited by Lutz E. Edzard and John Huehnergard, Wiesbaden: Harrassowitz 2014, 48–63. Bloch, Ariel. 1986. Studies in Arabic Syntac and Semantics. Wiesbaden: Otto Harrassowitz. Bobzin, Hartmut. 2010. Der Koran (aus dem Arabischen neu übertragen). München: C.H. Beck. Brockelmann, Carl. 1908. Grundriss der vergleichenden Grammatik der semitischen Sprachen. I. Band: Laut- und Formenlehre. Berlin: Reuther and Reichard.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

101

Carter, Michael. 2006. “Foreign Vocabulary.” In The Blackwell Companion to the Qur’ān, edited by Andrew Rippin, 120–139. Oxford: Blackwell. Chekayri, Abdellah. 1994. “Racines et Formes.” In Langues & grammaire: actes du premier colloque, Paris 8, Juin. Paris, edited by Léa Nash and Georges Tsoulas, 47–68. Paris: Université de Paris 8 (Département des sciences du langage). Chouémi, Moustafa. 1966. Le Verbe dans le Coran: racines et formes. Paris: C. Klincksieck. del Olmo Lete, Gregorio and Joaquín Sanmartín. 2004. A Dictionary of the Ugaritic Language in the Alphabetic Tradition. Second Revised Edition. Leiden and Boston: Brill. Donner, Fred M. 2008. “The Qur’ān in recent scholarship: challenges and desiderata.” In The Qur’ān in its Historical Context, edited by Gabriel Said Reynolds, 29–50. London and New York: Routledge. Duden. 1995. Duden. Grammatik der deutschen Gegenwartssprache. Siebte, völlig neu bearbeitete und erweiterte Auflage. Mannheim, Leipzig, Wien und Zürich: Dudenverlag. Fischer, Wolfdietrich. 1987. Grammatik des klassischen Arabisch. Zweite, durchgesehene Auflage). Wiesbaden: Otto Harrassowitz. Fox, Joshua. 2003. Semitic Noun Patterns. Winona Lake, Indiana: Eisenbrauns. Gesenius, Wilhelm. 1910. Hebrew Grammar edited and enlarged by the Late E. Kautzsch Second English Edition Revised in Accordance with the Tweny-Eight German Edition. Oxford: Clarendon Press. Gesenius, Wilhelm. 1954. Hebräisches und Aramäisches Handwörterbuch über das Alte Testament. Reprint of the Seventeenth Edition 1915. Berlin, Göttingen und Heidelberg: Springer Verlag. Gilliot, Claude. 2008. “Reconsidering the authorship of the Qur’ān: is the Qur’ān partly the fruit of a progressive and collective work?” In The Qur’ān in its Historical Context, edited by Gabriel Said Reynolds, 88–109. London and New York: Routledge. Goldenberg, Gideon. 1995. “Attribution in Semitic Languages.” Langues Orientales Anciennes: Philologie et Linguistique 5–6: 1–20. Ḥaqqī, ʾIsmāʿīl b. Musṭafā. 1330/1912–1347/1928. Rūḥ al-bayān. Edited by ʿUṯmān Bek. Ten vols. ʾIstanbūl: Maṭbaʿah ʿUṯmānīyah. Ibn ʿĀdil, Abū Ḥafṣ Sirāǧ ad-Dīn ʿUmar b. ʿAlī. 1419/1998. Kitāb al-lubāb fī ʿulūm al-kitāb. Edited by ʿĀdil ʾAḥmad ʿAbd al-Mawjūd and ʿAlī Muḥammad Muʿawwiḍ. Thirty vols. Bayrūt: Dār al-Kutub al-ʿIlmīyah. Ibn al-ʿArabī, Muḥammad b. ʿAbd Allāh. 1424/2003. ʾAḥkām al-qurʾān. Edited by Muḥammad ʿAbd al-Qādir ʿAṭāʾ. Four volumes. Third edition. Bayrūt: Dār al-Kutub al-ʿIlmīyah. Ibn ʿĀšūr, Muḥammad aṭ-Ṭāhir b. Muḥammad aṭ-Ṭāhir. 1404/1984. At-taḥrīr wa-t-tanwīr. Thirty vols. Tūnis: Ad-Dār at-Tūnisiyyah li-n-Našr.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

102

Daniel Birnstiel

Ibn Ḥanbal, ʾAḥmad. 1424/2002. Ar-radd ʿalā l-ǧahmīyah wa-z-zanādiqah. Edited by Ṣabrī b. Salāmah Šāhīn. Ar-Riyāḍ: Dār aṯ-Ṯabāt li-n-Našr wa-t-Tawzīʿ. Ibn Manẓūr, ʾAbū al-Faḍl Ğalāl ad-Dīn Muḥammad b. Mukarram. 1415/1994. Lisān al-ʿarab :Fifteen vols. Bayrūt: Dār Ṣādir. Jeffery, Arthur. 1938. The Foreign Vocabulary of the Qur’ān. Baroda: The Oriental Institute. Jespersen, Otto. 1949. A Modern English Grammar on Historical Principles. Seven vols. London: Allen and Unwin. Johnstone, T. M. 1977. Ḥarsūsi Lexicon and English-Ḥarsūsi Word List. London and New York: Oxford University Press. Johnstone, T. M. 1981. Jibbāli Lexicon. Oxford: Oxford University Press. Johnstone, T. M. 1987. Mehri Lexicon and English-Mehri Word-List. London: School of Oriental and African Studies. Kopf, Lothar. 1956. “Religious Influences on Medieval Arabic Philology.” Studia Islamica 5, 33–59. Kropfitsch, Lorenz. 1991. “Zur semantisches Struktur der Verbalstämme im Neuhocharabischen.” In Festgabe für Hans-Rudolf Singer, edited by Martin Forstner, two vols., 55–65. Frankfurt am Main and New York: Peter Lang. Kropp, Manfred. 2015. “Lisān ʿarabiyy mubīn – „klares Arabisch”? oder: „offenbar Arabisch”, gar „geoffenbartes Arabisch”?” In Books and written culture of the Islamic world studies presented to Claude Gilliot on the occasion of his 75th birthday, edited by Andrew Rippin and Roberto Tottoli, 271–287. Leiden: Brill. Lane, Edward William. 1863–1893. An Arabic English Lexicon (eight parts). London: Williams and Norgate. Leslau, Wolf. 1991. Comparative Dictionary of Geʿez (Classical Ethiopic). Wiesbaden: Otto Harrassowith. Mir, Mustansir. 2006. “Language.” In The Blackwell Companion to the Qur’ān, edited by Andrew Rippin, 88–106. Oxford: Blackwell. Muḥammad b. ʿAbd al-Wahhāb and ʿAbd Allāh b. Muḥammad b. ʿAbd al-Wahhāb. 1401/1981. Fī ʿaqāʾid al-ʾislām: ǧawāb ʾahl as-sunnah an-nabawīyah fī naqḍ kalām aš-šīʿah wa-z-zaydīyah. Edited by Muḥammad Rašīd Ridā. Bayūt: Manšūrāt Dār al-ʾĀfāq al-Ǧadīdah. Muqātil b. Sulaymān al-Balḫī. 1423/2002. Tafsīr Muqātil b. Sulaymān. Edited by ʿAbd Allāh Maḥmūd Šiḥātah. Four vols. Bayrūt: Dār ʾIḥyāʾ at-Turāṯ. Nawawī, Muḥammad b. ʿUmar. 1417/1996. Marāḥ labīd li-kašf maʿnā al-qurʾān al-karīm. Edited by Muḥammad ʾAmīn aṣ-Ṣanāwī. Two vols. Bayrūt: Dār al-Kutub al-ʿIlmīyah. Nöldeke, Theodor. 1904. Beiträge zur Semitischen Sprachwissenschaft. Strassburg: Karl J. Trübner. Payne Smith, Jessie. 1998. A Compendous Syriac Dictionary founded upon the The-

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

Neither Clear Nor Clarifying – Yet Clearly Arabic

103

saurus Syriacus of R. Payne Smith. Reprint. Winona Lake, Indiana: Eisenbrauns. Piamenta, Moshe. 1990. Dictionary of post-Classical Yemeni Arabic. Two vols. Leiden: Brill. Polotsky, Hans Jakob. 1978. “A Point of Arabic Syntax: The Indirect Attribute.” Israel Oriental Studies 8, 159–173. Reynolds, Gabriel Said. 2004. “A Reflection on Two Qurʾānic Words (Iblīs and Jūdī), with Attention to the Theories of A. Mingana.” Journal of the American Oriental Society 124.4, 675–689. Reynolds, Gabriel Said. 2008. “Introduction: Qur’ānic studies and its controversies.” In The Qur’ān in its Historical Context, edited by Gabriel Said Reynolds, 1–26. London and New York: Routledge. Ricks, Stephen D. 1989. Lexicon of Inscriptional Qatabanian. Roma: Pontificio Instituto Biblico. Sandberg, Bengt. 1984. “Der Bezug des Adjektivattributs bei substantivischen Zusammensetzungen.” Beiträge zur Geschichte der deutschen Sprache und Literatur 106, 159–183. Schulthess, Friedrich. 1903. Lexicon Syropalaestinum. Berlin: Georg Reimer. Sokoloff, Michael. 1992. A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period. Ramat-Gan, Israel: Bar Ilan University Press. Sprenger, Aloys. 1869. Das Leben und die Lehre des Moḥammad nach bisher grösstentheils unbenutzten Quellen. Second edition. Three vols. Berlin: Nicolaische Verlagsbuchhandlung. Vollers, Karl. 1906. Volkssprache und Schriftsprache im alten Arabien. Straßburg: Trübner. Van Putten, Marijn. 2018. “Hamzah in the Quranic Consonantal Text.” Orientalia N.S. 87, 93–120. Van Putten, Marijn and Philip W. Stokes. 2018. “Case in the Qurˀānic Consonantal Text.” Wiener Zeitschrift für die Kunde des Morgenlandes 108: 143–179. Von Soden, Wolfram. 1995. Grundriss der akadischen Grammatik. Third Edition. Roma: Pontificio Istituto Biblico. Wansbrough, John. 2004. Quranic Studies. Sources and Methods of Scriptural Interpretation. Foreword, Translations and Expanded Notes by Andrew Rippin. Amherst: Prometheus Books. Wright, William. 1898. A Grammar of the Arabic Language. Volume 2. Reprint of third edition revised by W. Robertson Smith and M.J. De Goeje, 1967. Cambridge: Cambridge University Press. Yaḥyā b. Salām. 1425/2004. Tafsīr Yaḥyā b. Salām. Edited by Hind Šalabī. Two vols. Bayrūt: Dār al-Kutub al-ʿIlmīyah. Zaborski, Andrzej. 2007. “Non-Causative Verbs of the Causative ʾaqtala Class in Arabic and *yuqtilu Conjugation in Proto-Semitic.” In Studies in Semitic and General

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

104

Daniel Birnstiel

Linguistics in Honor of Gideon Goldenberg, edited by Tali Bar and Eran Cohen, 31–43. Münster: Ugarit-Verlag. Zaehner, R.C. 1975. “Why not Islam.” Religious Studies 11.2, 167–179. Zammit, Martin R. 2002. A Comparative Lexical Study of Qur’ānic Arabic. Leiden, Boston and Köln: Brill. Zwettler, Michael. 1978. The Oral Tradition of Classical Arabic Poetry. Columbus: Ohio State University Press.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic, and Afroasiatic Lutz Edzard, University of Erlangen-Nürnberg and University of Oslo

1. Introduction In standard descriptions of Semitic, the nominative has the role of the unmarked case, typically in the role of the subject, whereas the accusative constitutes the marked case, typically in the role of the direct object. This paper argues that the concept of a “marked nominative” language, which is highly relevant on the Afroasiatic level, also plays an important role in Semitic and especially Arabic, as it catches many roles of the “accusative” (dependent case) that have no connection to the role of the direct object. Generally speaking, the term “marked nominative” refers to a scenario, in which the nominative constitutes a longer or more complex, or in other words “marked”, form vis-à-vis the accusative. In Gothic and Old Norse, for instance, the nominative exhibits more complexity as compared with the accusative (cf. https://en.wikipedia.org/wiki/Nominative_absolutive_language): (1) Nominative vs. accusative in Gothic and Old Norse Nominative Accusative Gloss Gothic dag-s dag ‘day’ Old Norse arm-r arm ‘arm’ A comparable scenario is found in Harar Oromo (cf. Owens 1985: 101, 251) : (2) Nominative vs. accusative in Harar Oromo a. sárée-n adii-n ni iyyi-t-i dog-nom white-nom foc bark-f-ipf ‘The white dog is barking.’ b. haat-tii okkóttée goot-t-i mother-nom pot.acc make- f-ipf ‘Mother is cooking (lit. making the pot).’

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

106

Lutz Edzard

Next to Cushitic and Omotic, comparable systems also obtain in Nilo-Saharan languages, thus pointing to an areal feature (cf. Dimmendaal n.d.: 1). In the latter case, the nominative is marked by a different tonal pattern: (3) Nominative vs. accusative in Eastern Sudanic (Nilo-Saharan) a. έ-dɔ́l έmbártá 3sg-see horse.acc (two rising accents) ‘he sees the horse’ b. έ-dɔ́l έmbartá 3sg-see horse.nom (one rising accent) ‘the horse sees him’ The proper terminology to adequately describe the form and function of the “accusative”, a term that only captures a small subset of the variety of functions of the Arabic naṣb, remains a problem (cf., e.g., Haspelmath (2009)). While the term “nominative” appears to be relatively unproblematic, one can point out that the term “accusative” represents a semantically narrowed translation of the more appropriate Greek term αἰτιατική by the Roman polymath Varro. As we will see, marking the direct object is but one function of “accusative” or, better: the “dependent case”; other functions include the marking of predicates, focalized subjects, and even the citation form and the vocative. In short: the functions of the case traditionally called “accusative” by far transcend the marking of the direct object, and the term “absolutive” is therefore preferable, as this term in linguistics refers to the unmarked citation form in “nominative-absolutive languages”. We will not touch upon ergative systems in this paper, as found, for instance, in modern Neo-Aramaic, even though these could be considered a further development of a “marked nominative” system. For the record, here are some standard definitions: Ergative-absolutive (or simply ergative): the argument (subject) of an intransitive verb is in the same case as the patient (direct object) of a transitive verb; this case is then called the absolutive case, with the agent (subject) of a transitive verb being in the ergative case. Ergative-accusative (or tripartite): the argument (subject) of an intransitive verb is in its own case (the intransitive case), separate from that of the agent (subject) or patient (direct object) of a transitive verb (which is in the ergative case or accusative case, respectively).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

107

2. The Afroasiatic frame Before investigating the situation in Semitic, especially Arabic and Ethio-Semitic, it is instructive to consider the scenario in a wider Afroasiatic perspective. One has argued that Afroasiatic originally had a subject (or “agent”) case associated with an u-ending and a predicative/absolutive (“nonagent”) case, considered the citation form, which also served for marking the object, associated with an a-ending (for an in-depth description of the scenario in Cushitic, cf. Appleyard 2012: 205–206 as well as Mous 2012: 369–373.). In the East-Cushitic language Borana, for instance, one finds an opposition between a subject case terminating in -í, vs. an absolutive case terminating in -a (which also functions as citation case), e.g., nam-í ‘(a) man’ (subject) vs. nam-a ‘(a) man’ (predicate), as in kunin nam-a ‘this is a man’ (cf. Sasse 1984: 112): (4) Subject vs. predicate case in Borana (East-Cushitic) a. nam-í ‘(a) man’ (subject) vs. b. nam-a ‘(a) man’ (predicate) A comparable functional opposition can be observed in the Berber “state” (elsewhere corresponding to case) prefix system, where the “independent form” characterized by an a-vowel designates the direct object, the nominal predicate, and a number of adverbial cases, whereas the “dependent form” characterized by an u-vowel designates the non-focalized subject, the adnominal genitive, and the complement of prepositions (see Sasse 1984: 120f.). Here is an illustration: (5) Dependent vs. independent “state” (case) in Berber (cf. Sasse 1984) a. aγiul iuγa-t urgaz def.donkey.indep buy.3m.sg.pret-it def.man.dep ‘The man bought the donkey.’ b. argaz iuγa aγiul def.man.indep buy.3m.sg.pret def.donkey.indep ‘It was the man, who bought the donkey.’ As we will see, the comparative Afroasiatic perspective lends support to the functional status of the Arabic dependent case (naṣb), i.e. it reflects its multifunctionality, e.g., marking the predicate of existential verbs or expressing adverbial function. Thereby, one has to be careful not to mix up the concepts

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

108

Lutz Edzard

of “dependent” and “independent” in Berber with the scenario in Arabic grammatical theory, where the term also captures dependency in subordinated sentences (subjunctive). Indeed, many Berber and Cushitic languages have to be considered marked nominative languages in a synchronic perspective.

3. Typological parallels in Indoeuropean The wide distribution of the dependent case in Arabic (and other Semitic languages) can also be viewed in connection with the ascendence of the oblique case in Semitic in general, notably the replacement of the sound masculine plural ending -ūn by -īn in the history of Arabic (for an overview of issues associated with case in Semitic, cf. Hasselbach 2013). Other Semitic languages, notably Hebrew, have lost case long ago, or rather only feature minimal residues of case. In Hebrew, the masculine pural ending -īm also reflects the historical oblique case. In other words: the oblique case tends to “win” over the nominative in the course of the leveling of a case system. The Romance languages in their ascendency from Vulgar Latin languages perfectly illustrate this development, e.g., Italian luce ‘light’ as compared with Latin lux (nom) ~ luce (abl). Still, one finds residues of the nominative, e.g., the final s in French fils ‘son’ as compared with Latin filius. Of course, it is not claimed here that the merger of independent and dependent (oblique) case is to be considered the same as the scenario in a “marked nominative” language. The development of the independent pronouns since late Middle English exhibits a comparable kind of convergence towards the oblique form, notably in the second person singular and plural (cf. Lutz 1996): (6) The development of pronouns since Late Middle English (Lutz 1996) Chaucer Shakespeare Modern (Late Middle English) Subject Object Subject Object Subject Object 1sg ich/I me I me I me 2sg thou thee thou/ye/you thee/you you you 3sgm he him he him he him 3sgf she hire she her she her 3sgnt (h)it (h)it it it it it 1pl we us we us we us 2pl ye you ye/you you you you 3pl they them they them they them

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

109

In this context it is interesting to compare the normative vs. the factual use of the personal pronouns in spoken English. Whereas normative English requires the predicate in a nominal sentence in the independent form (it is I), most people would actually use the dependent form for the predicate and say it is me. The same opposition holds for normative you and I vs. more common you and me. The latter version has an obvious parallel in Arabic ʾana wa-ʾiyyā-k ‘I (indep) and you (dep)’: (7) Normative vs. factual use of personal pronouns it is I vs. it is me you and I vs. you and me (cf. Arabic ʾana wa-ʾiyyā-k) Similar phenomena can also be observed in the development of the pronominal system in the Romance languages (this circumstance was already pointed out by A. Fischer (1922: 145f.); for details, cf. Buridant 2000: 431).

4. The situation in Semitic Both Biblical and (layers of) modern Hebrew can be adduced to illustrate the central point of this paper, namely that the position of the subject in existential or even possessive constructions can be marked by a nota accusativi. Alternatively, one could argue that the slot of the subject is synchronically reanalyzed as the slot of the (direct) object: (8) Existentials in modern colloquial Hebrew (cf. Glinert 1994: 149) yeš … ’there is …’ > yeš ʾet ... ‘there is ...’ exist > exist acc gam kan yeš ʾet ha-beʿayot ha-ʾele also here exist acc def-problem.pl def-dem.pl ‘Here too there are these problems.’ Comparable constructions are already attested in Biblical Hebrew, albeit only in the present and future tense:

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

110

Lutz Edzard

(9) Existentials in Biblical Hebrew a. ʾεṯ-šənē hag-gōyīm wə-ʾεṯ-šətē hå̄-ʾărå̄ṣōṯ acc-two.m.cs def-people.pl conj-acc-two.f.cs def-land.pl l-ī ṯihyεnå̄ for-I.dep be.nonpast.3f.pl w-īrašnū-hå̄ conj-inherit.nonpast_consec.3pl-she.dep ‘The two nations and the two lands shall be mine and we shall possess them.’ (Ezek. 35:10) b. ha-məʿaṭ lå̄-nū ʾεṯ-ʿăwōn pəʿōr int-little for-we-dep acc-sin.cs PN ‘Is the sin of Peor such a small thing to us?’ (Josh. 22:17) c. u-ḇnē ḥεṣrōn ăšεr nōlad-l-ō conj-son.pl.cs PN rel be.born.3m.sg.pf-for-he.dep ʾεṯ-yəraḥməʾēl wə-ʾεṯ-rå̄m wə-ʾεṯ-kəlūḇå̄y acc-PN conj-acc-PN conj-acc-PN ‘The sons of Hetzron that were born to him: Jerahmeel, Ram, and Chelubai.’ (1 Chron. 2:9) The situation in Ethio-Semitic is also illuminating. While Gəʿəz features existential and possessive constructions governed by a preposition “in”, modern Amharic makes use of an existential verb allä, which takes the possessor as an object suffix (“is to him”). When comparable experiencer verbs are preceded by an independent pronoun in topicalizing function, a syntactic ambiguity arises as this pronoun can be marked as an object (which it effectively is) or not: (10) Verboids in Gəʿəz and experiencer verbs in modern Amharic a. b-o / b-ottu ‘he has‘ (“in him [is]”) in-he.obj b. allä-w ‘he has’ (“there is for him”) exist.3m.sg.pf-he.obj c. əssu-n rabä-w ’he is hungry’ (“him it hungers him“) he.indep-acc be.hungry.3m.sg.pf-he.obj d. əssu rabä-w ’he is hungry’ (“he it hungers him“) he.indep be.hungry.3m.sg.pf-he.obj

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

111

Classical Ethiopic, Gəʿəz, also features the dependent case in predicate position, in addition to the object-marking function, as shown by the examples in (11) (cf., e.g., Leslau 1987: 299 and Weninger 1993: 27): (11) Object marking of the predicate in Gəʿəz konä nəgus-ä be.3m.sg.pf king-acc ’He became king.’ rəgəmt-ä təkun mədr cursed.f.acc be.3f.sg.juss earth ’The earth shall be cursed.’ A still puzzling case is the coccurrence of the dependent case in subject position, even though a focus-marking function cannot be discarded in this case (cf. Waltisberg 2002: 50): (12) Possible object marking of the subject in Gəʿəz yəkun fäqäd-ä-kä bäkämä bä-sämay be.3m.sg.ipf will-acc-you.2.m.sg.dep like in-heaven wä-bä-mədr-ni conj-in-earth-too ‘Your will be done on earth as in heaven.’ (Mt. 6:10) Modern Southern Ethio-Semitic, here represented by Amharic, also exhibits a variety of functions of the dependent case that are reminiscent of the scenario in Classical Arabic. Instances (13a) – (13d) illustrate the functions of marking the direct object, the predicate, the adverb, and focus, respectively (cf. Leslau 1995: 189 and 892ff.; cf. also Appleyard 2004): (13) Functions of the -n suffix in Amharic a. direct object (“accusative): mäskot-u-n zəga window-def-acc close.m.sg.imp ‘Close the window!’ b. predicative: əssu-n b-əhon al-adärg-äw näbbär he-acc in-be.1sg.ipf neg-1sg.ipf-it.dep be.3m.sg.pf ‘If I were him, I wouldn’t have done it’

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

112

Lutz Edzard

c. adverbial use: ləǧ-u əǧǧ-e-n yazä-ññ child-def hand-my.dep-acc take.3m.sg.pf-me.dep ‘The child took me by the hand (“handwise”).’ d. focus: awnät-wa-n näw truth-her.dep-acc be.3m.sg.ipf (cop) ‘She is right.’ A survey of the multitude of functions of the dependent case (naṣb) in Arabic is useful at this point. Starting with the function of marking the direct object (14a), (14) provides an overview of the various functions in Arabic (the following list is oriented at the survey provided in aš-Širbīnī’s treatise Nūr as-saǧīya fī ḥall ʾalfāẓ al-ʾĀǧurrūmīya, as edited, commented, and translated in Carter 1981): (14) Functions of the -a(n)-suffix in Arabic (according to aš-Širbīnī) a. the direct object (mafʿūl bi-hī): ḍarabtu Zayd-an strike.1sg.pf Zayd-acc ‘I struck Zayd.’ As the Arabic maṣdar typically occurs within the mafʿūl muṭlaq construction (inner object or “figura etymologica“), the dependent case is assumed to be the default case by the Arab grammarians. Hence, the maṣdar counts as another typical function of the dependent case (14b): b. the verbal noun (maṣdar): ḍarabtu ḍarb-an strike.1sg.pf strike.maṣdar-acc ‘I struck hard.’ The following instances (14c) – (14f), but also instance (14j), the object of reason (mafʿūl minʾaǧli-hī), have an adverbial function, whereby the Arabic grammatical terminology is more specific than the European counterparts:

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

113

c. the time-qualifier (ẓarf az-zamān): ṣumtu l-yawm-a fast. 1sg.pf def-day-acc ‘I fasted today.’ d. the space-qualifier (ẓarf al-makān): ǧalastu ʾamām-a-ka sit. 1sg.pf front-acc-you.dep ‘I sat in front of you.’ e. the circumstantial qualifier (ḥāl): ǧāʾa Zayd-un rākib-an come.3m.sg.pf Zayd-nom ride.ptc-acc ‘Zayd came riding.’ f. the specifying element (tamyīz): ṭāba Muḥammad-un nafs-an be_content.3m.sg.pf Muḥammad-nom soul-acc ‘Muḥammad was content of soul.’ The tamyīz also affects the morpho-syntax of numbers between 11 and 99, e.g., the example chosen by Sībawayhi ʿišrūna dirham-an ‘twenty dirham-acc’ (cf. Carter 1972). Examples (14g) – (14i), as well as (14j) and (14m), illustrate the focus-marking function of the dependent case, notably the focalized subject (cf. also Reckendorf 1921: 108–113 for various cases of the “exclamatory” accusative). The dependent case in (14g) presupposes that the set to which the excepted element belongs has already been mentioned: g. the excepted element (mustaṯnā): qāma l-qawm-u ʾillā Zayd-an get_up.3m.sg.pf def-people-nom except Zayd-acc ‘The people stood up except Zayd.’ h. the noun negated by lā (ism lā) (“absolute negation”): lā ʾilāh-a ʾillā llāh-u neg god-acc except God-nom ‘There is no god except God.’

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

114

Lutz Edzard

i. the first term of an annexation (ʾiḍāfa) in the vocative (munādā): yā ʿabd-a llāh-i voc servant-acc.cs God-gen ‘O ʿAbdallāh.’ j. the object of reason (mafʿūl minʾaǧli-hī): qāma Zayd-un ʾiǧlāl-an li-Bakr-in get_up.3m.sg.pf Zayd-nom honor.maṣdar-acc for-Bakr-gen ‘Zayd stood in honor of Bakr.’ k. the object of accompaniment (mafʿūl maʿa-hū): sirtu wa-n-nīl-a travel.1sg.pf conj-def-Nile-acc ‘I traveled with the Nile.’ Other examples of the pattern involving the wāw al-māʿīya include constructions such as ʾana wa-ʾiyyā-ka ‘I and acc.you.dep’ (cf. above section 3). Instance (14l) reflects the wider distribution of the dependent case in predicative position in Afroasiatic: l. the predicate of kāna and itsrelated verbs (“sisters”) (ḫabar kāna wa-ʾaḫawāti-hā): kāna Zayd-un qāʾim-an be.3m.sg.pf Zayd-nom stand.ptc-acc ‘Zayd was standing.’ m. the subject noun of ʾinna ‘indeed’ and its related particles (ism ʾinna wa-ʾaḫawāti-hā): ʾinna Zayd-an qāʾim-un foc Zayd-acc stand.ptc-nom ‘(Indeed,) Zayd is standing.’ Instance (14n) represents an entirely expected ditransitive construction, i.e. a verb taking two direct objects: n. the two objects of ẓanantu ‘I thought’ and its related verbs (mafʿūlā ẓanantu wa-ʾaḫawāti-hā): ẓanantu Zayd-an qāʾim-an think.1sg.pf Zayd-acc stand.ptc-acc ‘I thought Zayd was standing.’

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

115

Type (14o) features regular agreement, and would not, in a “Western” perspective be counted as a type of its own: o. the concordant of a dependent element, which comprises four things: the adjective, the coordinated element, the corroborative, and the substitute (at-tābiʿ li-l-manṣūb wa-huwa ʾarbaʿat ʾašyāʾ: an-naʿt wa-l-ʿaṭf wat-tawkīd wa-l-badal): raʾaytu Zayd-an il-ʿāqil-a see.1sg.pf Zayd-acc def-intelligent-acc ‘I saw Zayd, the intelligent (one).’ raʾaytu Zayd-an wa-ʿAmr-an see.1sg.pf Zayd-acc conj-ʿAmr-acc ‘I saw Zayd and ʿAmr.’ raʾaytu l-qawm-a kull-a-hum see.1sg.pf def-people-acc all-acc-them.dep ‘I saw the people, all of them.’ raʾaytu Zayd-an ʾaḫā-ka see.1sg.pf Zayd-acc brother.acc- your.m.sg.dep ‘I saw Zayd your brother.’ An a-ending is also characteristic of the predicate (or predicative state, cf. Lipiński, cf. 2001: 273f.) in Palaeo-Syrian, Old Akkadian, and Amurrite, as well as in some constructions with ʾammā in the oldest forms of Classical Arabic: (cf. Lipiński 2001: 495.): (15) Nominal clauses with predicate in the dependent case (naṣb) ʾammā ʾanta barr-an fa-qtarib if you.m.indep believing-acc conj-approach.m.sg.imp ‘If you are believing then come!’ To conclude this (non-exhaustive) list of instances of the dependent case in Arabic, let us cite the erroneously so-called ʾafʿāl taʿaǧǧub ‘verbs of wonder/ surprise’ (per the position of the Baṣran grammarians), which in reality constitute frozen elative constructions (per the correct position of the Kūfan grammarians), in which the original subject was reanalyzed as an object due to the

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

116

Lutz Edzard

misinterpretation of the elative form as a causative (form IV) (cf. Brockelmann 1913: 12–13 and Diem 2013): (16) Frozen elatives in Classical Arabic mā ʾakram-a zayd-an what noble.elative-acc Zayd-acc ‘How noble is Zayd!’ The dependent case in subject position also occurs in modern Arabic pseudo-verb constructions (“verboids”) such as the following: (17) Verboids in Arabic dialects (cf. Shlonsky 1997, Brustad 2000, Comrie 2008; for Hebrew cf. Kuzar 2013) ʿind-ī… ‘with me …’ (“chez moi …”) at-I.dep > ʿand-i yyā-ha ‘I have her/it (fem.)’ at-I.dep acc.she.dep bi-wudd-ī… ‘in my wish …’ in-wish-I.dep > bidd-i yyā-ha ‘I want her/it (fem.)’ want-I.dep acc.her.dep ʾabū-k bidd-o yyā-k father.cs-you.m.sg.dep want-he.dep acc.you.m.sg.dep ‘Your father requests you(r presence).’ (cf. Elihay 2005: 87) fī-nī ... ’in me ...’ in-I.dep > mā fī-nī rūḥ maʿ-ik neg in-I.dep go.1sg.ipf with-you.m.sg ‘I’m not able to go with you’ Synchronically, ʿand-, bidd-, and fī- can be characterized as pseudo-verbs (“verboids”) that govern the dependent case or, in the case of the third example, subordinate clauses. ʿand- and bidd- take the possessive, not the object form of the pronominal suffix, indicating the possessor and the author of the wish, respectively. In the case of fī-, the use as a mere preposition would entail the use of the possessive (independent) -yi, fī-yi ’in me’ (cf. Cowell 1964: 479; apud Comrie 2008: 739).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

117

Further examples, which also govern the dependent case or, in the case of verboids expressing modality, subordinate clauses (which amounts to the same in Arabic grammatical theory), are the following: (18) Further verboids in Arabic dialects (Cowell 1964; Brustad 2000) baʿd-nī b-əl-bēt still-I.dep in-def house ’I’m still at home’ (Syrian Arabic) ləssaʿt-nī b-əl-bēt not.yet-I.dep in-def-house ’I’m not yet at home’ (Syrian Arabic) ləssā-nī mā ʿaraḍṭ-ha b-əl-wāžha not.yet-I.dep neg put_on_view.1sg.pf in-def-showcase ’I haven’t yet put them on view in the showcase’ (Cowell 1964: 546) xəss-na nṭəyyebu l-ʿša necessary-us.dep cook.1pl.ipf def-dinner ’We must cook dinner’ (Moroccan Arabic) The original semantics of the particles that constitute the core of the verboids are not always transparent any longer. baʿd originally means “after“, and ləssa / ləssaʿ / ləssaʿt derives from ʾilā s-sāʿa “to this hour“. Impersonal passive constructions can likewise attract the dependent case. Retsö (1982–1983: 86) provides the following examples: (19) Impersonal passive constructions (Retsö 1982–1983) mā yinḥārab wiyyāh neg fight.pass.ipf.3m.sg acc.3m.sg ‘he cannot be fought with’ mā yinsōlaf wiyyāh neg speak.pass.ipf.3m.sg acc.3m.sg ‘he cannot be spoken with’ An alternative interpretation may have recourse to the use of wa- as a preposition (wāw al-māʿīya), as, e.g., in the expression bi-mā yattafiqu wa-hāḏā l-iltizām ’with what is in congruence with this commitment’. In the context of impersonal verbs, Retsö (1982–1983) also points to the fact that the object in Arabic passive phrases can remain in the dependent case,

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

118

Lutz Edzard

e.g., ʾuḫriǧa kitāban (bring_out.pst.pass.3m.sg book.acc) ‘a book was brought out (to him)’. In ditransitive verbal phrases set in the passive voice, the second object remains in the accusative under any circumstances: (20) Case transfer under passivization sammā-hu Muḥammadan call.3m.sg.pf-he.dep Muḥammad.acc ‘he called him Muḥammad’ summiya Muḥammadan was_called.3m.sg.pf Muḥammad.acc ‘he was called Muḥammad’

5. Case study: the masʾala az-zunbūrīya In the following, we will consider the famous masʾala az-zunbūrīya (the “hornet question”) (cf. A. Fischer 1922 and Blau 1963) in the light of the previous deliberations. The masʾala involves the grammarians Sībawayhi (d. 180/796?), representing the Baṣran grammatical school, and al-Kisāʾī (d. 189/805), representing the Kūfan grammatical school. This masʾala is a major case in point in terms of the multifunctionality of the dependent case (naṣb) in Arabic. The version of the masʾala reported in Ibn al-ʾAnbārī’s (d. 577/1181) ʾInṣāf runs as follows (p. 292) (as above, the abbreviation/interlinear glossing “acc” for ʾiyyā- is used here without prejudice to the grammatical function (deictic or object-marking) of this particle):

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

119

(21) Ibn al-ʾAnbārī’s version of the masʾala az-zunbūrīya kuntu ʾaẓunnu ʾanna l-ʿaqrab-a be.1sg.pf think.1sg.ipf comp def-scorpion-acc ʾašadd-u lasʿat-an mina z-zunbūr-i strong.elativ-nom sting.maṣdar-acc from def-hornet-gen fa-ʾiḏā huwa ʾiyyā-hā conj-interj he.indep acc-she.dep (al-Kisāʾī = position of the Kūfan grammarians) vs. fa-ʾiḏā huwa hiya conj-interj he.indep she.indep (Sībawayhi = position of the Baṣran grammarians ‘I always thought that the scorpion has a sharper sting than the hornet, (however,) the former is (like) the latter (in terms of sting).’ While the grammatical point is clear, the actual interpretation differs, a point which also has to do with slightly divergent versions of the masʾala. Ibn Ḫallikān (d. 681/1282), Wafāyāt III, 463–465, in a popularized and anecdotal version of the surrounding story, has the following version: kuntu ʾaẓunnu z-zunbūra ʾašadda lasʿan mina n-naḥla ... ‘I always thought that the hornet has a sharper sting than the bee …’. The translation above is based on the possible feminine gender of zunbūr. Slane ,the translator of Ibn Ḫallikān (1842–1871: II, 397) translates fa-ʾiḏā huwa ʾiyyā-hā / hiya as ‘and behold! It was so’. Carter (2004: 13) also adopts this interpretation, thus relating ʾiyyā-hā / hiya to the sting (lasʿa), an interpretation which only works with a feminine maṣdar. Versteegh (2014: 72) translates ’but is was the other way round’, an interpretation which matches comparable phrases in ancient Mesopotamian contexts (personal communication by Manfred Krebernik). Incidentally, the phrase fa-ʾiḏā huwa ʾiyyā-hā is also found in exactly this wording in the 35th Maqāma of al-Ḥarīrī (Maqāmāt II, 449). In this context, the Baṣran position essentially pointed to the status of fa-ʾiḏā huwa hiya as a nominal sentence, with the predicate standing by default in the independent case (rafʿ). In contrast, the Kūfan position resorted to the concept of taqdīr, i.e. the presupposition of an underlying governing element such a ditransitive verb, which by analogy would trigger the dependent case (naṣb). According to Ibn al-ʾAnbārī, al-Kisāʾī adduced the following (according to him) analogous case:

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

120

Lutz Edzard

(22) Analogous masʾala cited by al-Kisāʾī (according to Ibn al-ʾAnbārī) ḫaraǧtu fa-ʾiḏā ʿAbd-u llāh-i go_out.1sg.pf and-interj servant-nom.cs God-gen l-qāʾim-u def-stand.ptc-nom vs. l-qāʾim-a def-stand.ptc-acc ‘I went out, and (lo and behold) there stood ʿAbdallāh.’ Here, the version with the dependent case (naṣb) can be justified by the presupposition of the ditransitive verb waǧada ‘find’: (23) Purported analogy between (22) and ditransitive constructions fa-ʾiḏā ʿAbd-u llāh-i l-qāʾim-a and-interj servant-nom.cs God-gen def-stand.ptc-acc ‘And (lo and behold) there stood ʿAbdallāh.’ waǧadtu ʿAbd-a llāh-i l-qāʾim-a find.1sg.pf servant-acc.cs God-gen def-stand.ptc-acc ‘I found ʿAbdallāh standing.’ Contrary to the account by Ibn Ḫallikān, al-Kisāʾī thus did not go as far as to rule out the independent case (rafʿ) as an admissible option in the masʾala az-zunbūrīya. Rather, he allowed for both possibilities: al-ʿarabu tarfaʿu ḏālika kulla-hū wa-tanṣubu-hū ’The Arabs set all this in the independent case (rafʿ “nominative“) and they [equally can] set it in the dependent case (naṣb “accusative“)’ (cf. Ibn al-ʾAnbārī, ʾInṣāf, 293).

6. The distribution and functional breadth of the particle ʾiyyāWithin a modern non-normative linguistic approach to the issue at hand, a viable option is to look at the distribution of ʾiyyā- in connection with personal pronouns and at the whole spectrum of functions that this particle can adopt. As we will see, ʾiyyā- also occurs in other context than marking the direct object. (For an overview of suggested etymologies cf. Rubin 2005, chapter 5, and Wilmsen 2013b.) At first glance, the prototypical function appears indeed to be the one of the direct object marker as in (24):

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

121

(24) Verbs with two pronominal objects ʾaʿṭā-hu ʾiyyā-hā give.3m.sg-he.dep acc-she.dep ’He gave her/it to him.’ In this context, the typology of ditransitive verbal phrases in Arabic and Semitic in general established by Jan Retsö (1987: 224) is useful: (25) Typology of ditransitive verbal phrases in Semitic (Retsö 1987) construction type A: Verb + O2 (IO) + O1 (DO) A1: Verb + -PS + -PS A2: Verb + -PS + ʾiyyā-PS A3: Verb +- PS + IP A4: Verb + l-PS + ʾiyyā-PS A5: Verb + l-PS + IP construction type B: Verb + O1 (DO) + O2 (IO) B1: Verb + -PS + l-PS B2: Verb + -PS + -PS B3: Verb + -∅ + l-PS Abbreviations: O1 = DO: direct object (“accusative“); O2 = IO: indirect object (“dative“); PS: pronominal suffix; IP: “independent“ personal pronoun; l = a preposition indicating “to/for”.

Construction types A2 and A4 on the one hand and A3 and A5 on the other illustrate that the slot of the direct object can be filled either by ʾiyyā- followed by a pronominal suffix or by a (bi-)form of the independent pronoun, comparable to the scenario found in the masʾala az-zunbūrīya. Dialectal Arabic examples of the former type (A2/A4) and the latter (A3) are found in (26a) and (26b), respectively. For the record, Retsö notes that the independent pronouns can also appear in this position in copula function (26c). The examples are originally found in Sabuni 1980, Sasse 1971, and, Jastrow 1979, later compiled in Retsö 1987 (221f.):

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

122

Lutz Edzard

(26) Comparable distribution of subject pronouns, object pronouns, and copula in Arabic dialects a. ʿaṭētū-k-iya (type A2) give.1sg.pf-you.m.sg.dep-acc.she.dep ‘I gave her/it to you (m.sg.).’ (Mossul) b. ʿaṭaytu-hū-wē (type A3) give.1sg.pf-he.dep-he.indep ‘I gave it (m.) to him.’ (Mḥallamiye) ǧǟbū-l-o-hūwe (type A5) bring.3pl.pf-dat-he.dep-he.indep ’They brought it (m.) to him.’ (Aleppo) c. fə-l-bayt-wē in-def-house-he.indep ‘He is in the house.’ (Mḥallamiye) hāyi hiya dem.f she.indep ’That’s it!’ (Mossul) The overlap between focus and object markers is even more intriguing. An interesting case in point is an attested deviating reading (qirāʾa) of ʾāya 5 in sūra 1 (cf. A. Fischer 1922: 154. Again, “acc” is used in this paper as a technical gloss without prejudice to the function of ʾiyyā- in the given context): (27) Deviating reading (qirāʾa) of ʾāya 5 in sūra 1 ʾiyyā-ka tuʿbadu acc-you.m.dep venerate.2m.sg.impf.pass ’You [God] shall be venerated.’ instead of standard ʾiyyā-ka naʿbudu acc-you.m.dep venerate.1c.pl.impf.act ’We venerate you [God].’ Clearly, the first case has to be interpretated with a focus marker (which, after all, is one of the functions of the dependent case). Accordingly, one can argue that corroborative appositions (tawkīd) in Arabic – both Classical and Modern Standard – rather reflect a deictic function of the particle ʾiyyā- (cf. Wilmsen 2011: 316ff. and Wilmsen 2013a/b) (28):

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

123

(28) Standard Arabic appositive constructions with ʾiyyā- (Wilmsen 2011) ḏālika ṣ-ṣawtu ʾiyyā-hu dem.m.sg def-voice acc-he.dep ‘this voice itself’ ʾahlu d-dunyā ʾiyyā-hum people def-world acc-they.m.pl.dep ‘the people themselves’ In modern Hebrew, the analogous construction (which only developed in post-Biblical Hebrew) is entirely regular, with ʾot- being an allomorph (except before the second person plural) of ʾet, whether or not one considers ʾet to be a cognate of ʾiyyā- (for skepticism in this regard, cf. Wilson-Wright (2016) (29): (29) Analogous case in modern Hebrew ʾot-o ha-davar acc-he.dep def-matter ‘the same matter’

7. Conclusion Returning to the problem initially exposed, the Semitic and especially the Arabic scenario clearly shows the broad functional range of the dependent (or “absolutive”) case. Thus, the variety of functions associated with the dependent case (naṣb) in Arabic and elsewhere does not constitute anything idiosyncratic. Rather, this variety reflects many of the functions of this case in “marked nominative” languages, in which the dependent case (naṣb) would be called “absolutive”. In the Semitic and Afroasiatic languages surveyed, the independent case (“nominative”) is not morphologically more complex, rather the vice versa, as the dependent case may be marked by a nota accusativi. Still, in terms of frequency, the concept of markednes makes sense in this context.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

124

Lutz Edzard

Bibliography 1. Primary sources al-Ḥarīrī, Maqāmāt = Antoine Isac Silvestre de Sacy, Les séances de Hariri publiées en arabe avec un commentaire choisi. Two volumes. Paris, 1847–1853. Ibn ʾĀǧurrūm, ʾĀǧurrūmīya = Ibn ʾĀǧurrūm Muḥammad ibn Dāʾūd al-Ṣinhāǧī, Kitāb al-ʾĀǧurrūmīya. Edited and commented in Edzard and Bjørsnøs (eds.) 2008: Arabic pages 171–183. Ibn al-ʾAnbārī, ʾInṣāf = Kamāl ad-Dīn ʾAbū l-Barakāt ʿAbd ar-Raḥmān ibn ʾAbī Saʿīd al-ʾAnbārī, Kitāb al-ʾinṣāf fī masāʾil al-ḫilāf bayn an-naḥwīyīn al-baṣrīyīn wa-lkūfīyīn. Die grammatischen Streitfragen der Basrer und Kufer. Edited by Gotthold Weil. Leiden: Brill. 1913. Ibn Ḫallikān, Wafāyāt = Ibn Ḫallikān, Wafāyāt al-ʾaʿyān wa-ʾanbāʾ ʾabnāʾ az-zamān. Edited by ʾIḥsān ʿAbbās. 8 vols. Beirut: Dār Sādir, 1968–1977. Ibn Khallikan’s Bibliographical Dictionary. Translated from the Arabic by Bn Mac Guckin de Slane. 4 vols. New York and London: Johnson Reprint Corporation, 1842–1871. aš-Širbīnī, Nūr al-saǧīya = Muḥammad ibn ʾAḥmad aš-Širbīnī, Nūr as-saǧīya fī ḥall ʾalfāẓ al-ʾĀǧurrūmīya. Edited, commented, and translated in Carter 1981.

2. Secondary sources Appleyard, David. 2004. “Some thoughts on the origin of the Amharic object marker -ን, -[ə]n.” In Studia Aethiopica in Honour of Siegbert Uhlig on the Occasion of his 65th Birthday, edited by Verena Böll, Denis Nosnitsin, Thomas Rave, Wolbert Smidt, and Evgenia Sokolinskaia, 291–301. Wiesbaden: Harrassowitz. Appleyard, David. 2012. “Cushitic.” In Semitic and Afroasiatic. Challenges and Opportunities, edited by Lutz Edzard, 199–295. Wiesbaden: Harrassowitz. Blau, Joshua. 1963. “The role of the Bedouins as arbiters in linguistic questions and the masʾala az-zunbūriyya.” Journal of Semitic Studies 8: 42–51. Brockelmann, Carl. 1913. Grundriß der vergleichenden Grammatik der semitischen Sprachen. II. Band: Syntax. Berlin: Verlag von Reuther & Reichard. Brustad, K. E. 2000 The Syntax of Spoken Arabic: A comparative study of Moroccan, Egyptian, Syrian and Kuwaiti dialects. Washington D.C.: Georgetown University Press. Buridant, Claude. 2010. Nouvelle grammaire de l’ancien français. Paris: Sedes. Carter, Michael G. 1972. “‘Twenty dirhams’ in the Kitāb of Sībawayhi.” Bulletin of the School of Oriental and African Languages 35: 485–496.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Marked Nominative in Arabic, Semitic and Afroasiatic

125

Carter, Michael G. 1981. Arab Linguistics. An Introductory Classical Text with Translation and Notes. Amsterdam: John Benjamins. ———. 2004. Sībawayhi. London: I.B. Tauris. Comrie, Bernard. 2008. “Pseudo-verbs.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 739–740. Leiden: Brill. Cowell, Mark W. 1964. A Reference Grammar of Syrian Arabic. Washington, D.C.: Georgetown University Press. Diem, Werner. 2013. “Die arabischen Mirative in historischer Perspektive.” In Nicht nur mit Engelszungen. Beiträge zur semitischen Dialektologie. Festschrift für Werner Arnold zum Geburtstag, edited by Renaud Kuty, Ulrich Seeger and Shabo Talay, 73–86. Wiesbaden: Harrassowitz. Dimmendaal, Gerrit J. n.d. “Marked nominative systems in Eastern Sudanic and their historical origin.” Edzard, Lutz and Amund Bjørsnøs. 2008. New edition of Rudolf-Ernst Brünnow and August Fischer (7th ed. 1988). Arabische Chrestomathie aus Prosaschriftstellern / Chrestomathy of Classical Arabic Prose Literature. Wiesbaden: Harrassowitz. Elihay, Jean. 2005. The Olive Tree Dictionary. A Transliterated Dictionary of Conversational Eastern Arabic (Palestinian). Jerusalem: Minerva Publishing House. Fischer, August. 1922. “Die masʾala zunbūrīja.” In A Volume of Oriental Studies Presented to Edward G. Browne on the Occasion of His 60th Birthday (7 February 1922), edited by Thomas W. Arnold and Reynold A. Nicholson, 150–156. Cambridge: Cambridge University Press. Glinert, Lewis. 1989. The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. Haspelmath, Martin. 2009. “Terminology of case.” In The Oxford Handbook of Case, edited by Andrej Malchukov and Andrew Spencer, 505–517. Oxford: Oxford University Press. Hasselbach, Rebecca. 2013. Case in Semitic. Roles, Relations, and Reconstruction. Oxford: Oxford University Press. Jastrow, Otto. 1979. “Zur arabischen Mundart von Mossul.” Zeitschrift für arabische Linguistik 2: 36–74. Kuzar, Ron. 2013. “Verboids.” In Encyclopedia of Hebrew Language and Linguistics, edited by Geoffrey Khan, vol. 3, 933–934. Leiden: Brill. Leslau, Wolf. 1987. Comparative Dictionary of Geʿez (Classical Ethiopic). Wiesbaden: Harrassowitz. ———. 1995. Reference Grammar of Amharic. Wiesbaden: Harrassowitz. Lipiński, Edward. 2001. Semitic Languages. Outline of a Comparative Grammar, 2nd edition. Louvain: Peeters. Lutz, Angelika. 1996. “The interplay of external and internal factors in morphological restructuring: The case of you.” In Advances in English Historical Linguistics, ed-

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

126

Lutz Edzard

ited by Jacek Fisiak and Marcin Krygier, 189–210. Berlin and New York: Mouton de Gruyter. Mous, Maarten. 2012.” Cushitic.” In The Afroasiatic Languages, edited by Zygmunt Frajzyngier and Erin Shay, 342–422. Cambridge: Cambridge University Press. Owens, Jonathan. 1985. A Grammar of Harar Oromo (Northeastern Ethiopia). Hamburg: Buske. Reckendorf, Herrmann. 1921. Arabische Syntax. Heidelberg: Carl Winter’s Universitätsbuchhandlung. Retsö, Jan. 1982–1983. “Subjectless sentences in Arabic dialects.” Orientalia Suecana 31–32: 71–91. ———. 1987. “Copula and double pronominal objects in some Semitic languages.“ Zeitschrift der Deutschen Morgenländischen Gesellschaft 137: 219–245. Rubin, Aaron. 2005. Studies in Semitic Grammaticalization. Winona Lake: Eisenbrauns. Sabuni, Abdulghafur. 1980. Laut- und Formenlehre des arabischen Dialekts von Aleppo. Frankfurt am Main: Peter Lang. Sasse, Hans-Jürgen. 1971. Linguistische Analyse des arabischen Dialekts der Mḥallamiye in der Provinz Mardin (Südosttürkei). Dissertation, University of Munich. ———. 1984. “Case in Cushitic, Semitic, and Berber.” In Current Progress in Afro-Asiatic Linguistics, edited by John Bynon, 111–126. Amsterdam: John Benjamins. Shlonsky, Ur. 1997. Clause Structure and Word Order in Hebrew and Arabic: An Essay in Comparative Semitic Syntax. New York and Oxford: Oxford University Press. Versteegh. Kees. 2014 . The Arabic Language, 2nd edition. Edinburgh: Edinburgh University Press. Waltisberg, Michael. 2002. “Zur Ergativitätshypothese im Semitischen.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 152: 11–62. Weninger, Stefan. 1993. Gəʿəz (Classical Ethiopic). Munich: LINCOM. Wilmsen, David. 2011. “Dialects of the dative shift: a re-examination of Sībawayhi’s dispute with the naḥwiyyūn over ditransitive verbs with two object pronouns.” In In the Shadow of Arabic. The Centrality of Language to Arabic Culture. Studies presented to Ramzi Baalbaki on the occasion of his sixtieth birthday, edited by Bilal Orfali, 299–321. Leiden: Brill. ———. 2013a. “The demonstrative iyyā-: a little-considered aspect of Arabic deixis.” Arabica 60: 332–358. ———. 2013b. “More on the Arabic object marker iyyā: implications for the origin of the Semitic notae accusativi.” Folia Orientalia 50: 67–82. Wilson-Wright, Aren M. 2016. “A reevaluation of the Semitic direct objet markers.” Hebrew Studies 57: 7–15.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives Based on *ʾVl in Arabic and the Origin of Dialectal illī Phillip W. Stokes, The University of Tennessee, Knoxville

1. Introduction In this paper I attempt to explain the etymology and origin of the dialectal Arabic relative particle illī1 and its variants. Previous explanations suggested one of two etyma: 1) a reduced form of the Classical Arabic (henceforth CAr) relative allaḏī, via an ad hoc loss of the interdental ḏ intervocalically (Brockelmann 1908; Barth 1913; Fischer 1959; Blau 1966–7); 2) a derivation from a Vl-based demonstrative (Rabin 1951; Grand’Henry 1978). The present explanation will provide further support for the latter. While this paper will argue that Rabin’s (1951) proposal, deriving illī from a demonstrative *ʾullay, is the most likely solution, a number of details remain unexplained in the original suggestion. In order to accurately establish the etymology of illī, I will review the demonstrative and relative forms based on ʾVl in CAr and the Qurʾān, and argue that the morphological variants of these pronouns are derivable from one proto-form via a series of regular sound changes that are well-attested in numerous Arabic varieties. The remainder of section 1 of the paper contains an overview of the variants of illī in dialectal Arabic. Section 2 will present the previous explanations and any shortcomings that justify further comment on the topic. Section 3 provides a detailed discussion of the morphology of demonstratives and relatives based on *ʾVl in Arabic. Following the discussion of the forms from section 3, Section 4 will present a diachronic reconstruction, focusing on the quantity of the initial vowel, the length of the consonant l, and the quantity of the final vowel. Section 5 presents a reconstruction of illī and its variants. The paper concludes with Section 6, which presents a summary and conclusions based on the foregoing discussion. 1 The transcription of this form is often illi, with a short final vowel, especially in synchronic descriptions of a particular dialect. I have decided to transcribe it as illī because this paper is focused on the origin and diachronic development of the particle and, as I will argue below, the final vowel most likely originated in a historically long vowel.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

128

Phillip W. Stokes

1.1 Relative pronouns in the modern Arabic dialects Several relative pronouns are attested in early Islamic Arabic, the most common of which are: m.s. allaḏī, f.s. allatī, mpl. allaḏīna and fpl. allāti or allawāti. Most contemporary dialects attest one of a few forms. A few marginal dialects attest allazi (NW Yemen; Behnstedt 2016: 74), ḏī / ḏā (NW Yemen; Behnstedt 2016: 74), and (ə)ddi (N. Africa; Retsö 2004: 269). The most common form cross-dialectally by far is illī (see Vicente 2009 for a good overview of the attested forms). The morphology of this form varies across the dialects; the following represent most of the attested variation2: Table 1: Selection of Relative Pronouns in Modern Dialects Cherchell, Algeria (Grand’Henry 1972: 142) əlli / lli / li Mardin, Turkey (Jastrow 1978: 123–124)

la

Most of Mardin area (ibid.)

lə

Diyarbakir, Siirt, and Kozluk (ibid.)

lē

Andalusi (Corriente 2006: 106)

alli (alongside allaḏī and addi / aḏḏī)

Jewish Antiochene (Arnold 2006: 116)

əlli

Baghdad (Abu-Haidar 2006: 226)

l with əlli “for emphasis”

Jewish Baghdad (Mansour 2006: 239)

əlli / lli / l (the latter of which assimilates)

Bahraini (Holes 2006: 247)

illi

Beirut (Naïm 2006: 279)

əlli / əlle / əl / halli

2 I include in this list a few examples of lV-based relatives, as well as several that are composed of only ʾVl. I will briefly discuss these variants, including the question of whether they are indeed etymologically related, below (section 4). Since I am focusing on forms related, potentially or clearly, to illī, I have not included variants from other etymological origins, such as *allaḏī etc.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

Bʿēri (Woidich 2006: 303)

illi

Damascus (Lentin 2006: 549)

əlli / lli / yəlli / halli

Ḥassaniyyah (Taine-Cheikh 2007: 243)

lli / əl

129

Horn of Africa (Simeone-Senelle 2007: 271) elli Palestinian (Shahin 2008: 531)

illi / elli / alli / halli

Jordanian (Ammani; author’s fieldwork)

illi / yalli / elli

Kuwait (Holes 2007: 614)

illi

Khartoum (Dickins 2007: 569)

al

Maltese (Mifsud 2008: 151)

li / illi

Mecca (Abu-Mansour 2008: 182)

illi

Moroccan (Heath 2002: §7.5.2)

lli

Negev (Henkin 2008: 365)

allī / hallī

Omani (Holes 2008: 484)

illi / il

NW Yemeni (Behnstedt 2016: 74)

ʾallī / ʾalli / (ʾ)illī / (ʾ)illi / ʾil / llī / ʾalī / lī / li

Tunisian Ṣāḥel (Grand’Henry: 1972: 142)

alla

While the etymologies of the other main dialectal relatives, allazi (< *allaḏī), ḏī / dī (< *ḏī), ḏā ( allaʾī > allay > allī.3 More recently, Grand’Henry (1972: 142) speculated that it might best be viewed as an independent demonstrative form composed of the definite article ʾal + asseverative la. The main difference between Grand’Henry and the traditional interpretation offered by Barth and Blau is that the former argues that the dialect form, composed of only *ʾal + *la, was distinct from allaḏī and not a reduced form of it.4 While the process Barth and Blau suggest is of course possible, it is completely ad hoc and, thus, less convincing. Grand’Henry’s argument does not explain the diachronic relationships between the attested forms. He simply notes that there are variants that are similar and composed of similar material. More critically, he remains silent on why final -a of the asseverative particle would be realized in most cases as -i or -ī or how it was retained, given that final short -a was in most cases lost in the dialects, cf. *darasa “he studied,” but daras “idem” in most dialects. Chaim Rabin (1951: 153–155) rejected the traditional explanation proposed by Brockelmann and Barth, and championed by Blau, and suggested 3 Another possible parallel to the ad hoc loss of interdental ḏ, is the deictic particle hāy “this (fsg.),” attested in a number of modern Arabic dialects. If we suppose an etymology of hāḏihi “this (fsg.),” it would appear that the interdental was lost, resulting in the same hiatus to diphthong: hāḏihi > hāḏī > hāʾī > hāy. While that is possible, it is not certain that the origin of dialectal hāy was hāḏihi (or hāḏī); rather, it is possible that the fsg. suffix ī was added to the deictic hā (Magidow 2016: 99). 4 Corriente (2013: 16, n. 46) suggested a third proposal, similar to that of Brockelmann and Blau, arguing that the Maghrebi forms could be the result of a change of *ḏ > l / V_V, attested in e.g., ila “if” (< *ʾiḏā). In that case, əlli in Maghrebi dialects would be the result of the following processes: *ʾallaḏī > ʾallalī > əlli (syncope of vowel in open syllable, common in many dialects). While this is certainly possible, the change of intervocalic *ḏ > l is to my knowledge not attested consistently outside of some Maghrebi varieties. It is possible that the various Vllī forms in the dialects are ultimately derived from two originally separate etyma, however, one should prefer an explanation that relies as little as possible on multiple origins and ad hoc or otherwise unattested sound changes.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

131

instead that dialectal illī and related forms are derived from a demonstrative *ʾullay. It is worth considering Rabin’s arguments in some detail, since many of the issues he deals with are of importance for the present proposal. Rabin (1955: 153–154) follows Barth (1913: §48–49) in reconstructing the proto-Semitic plural demonstrative as *ʾullay, which he also reconstructs as the proto-Arabic plural demonstrative. Much of his discussion deals with the differences between this reconstruction and the attested forms, including ʾūlāʾi, ʾūlay, and ʾūlā. Rabin reconstructs these forms, which were typically spelled with a mater lectionis representing the first -u vowel, i.e., ‫اوال \ اولى‬, as long. For Rabin, this length is an “ersatzdehnung,” which contrasts with the geminated ll in the form he reconstructs: *ʾūlay // *ʾullay.5 On the differences in final vowel length, Rabin argues that the common CAr form ending in -āʾi developed from *ai in the same way that the Qurʾānic reading of Q2.16 ištaraʾū ḍ-ḍalālata “they acquired error” (< *ištaraw ḍ-ḍalālata) did (Rabin 1951: 153). He suggests first that a “double-peak stress” developed, followed by the insertion of a glottal stop in order to break up the diphthong (Ar. hamza bayna bayna). Rabin argues (ibid.) that the impetus for breaking the diphthong was to avoid the doubly closed unstressed syllable that would occur when the proximal demonstrative was followed by a noun: ʾulā + l-C. Finally, he proposes that the proto-Arabic demonstrative *ʾullay was used as a relative for both singular and plural heads. The relative *ʾullay was paralleled by the innovative form allaḏī, which, he believed, was made up of the demonstrative ʾullay + demonstrative ḏī (1955: 154–155). Proof of this is found in the quality of the second -a vowel of allaḏī, which he argues was reduced from -ā (< *ay). Ultimately, allaḏī replaced *ʾullay because the latter, perhaps by contamination with the article, was reduced to llay, which was “felt to have too little body” (ibid., 155). Thus per his proposal, the modern dialectal form continues the pre-modern usage of the plural demonstrative *ʾullay as a general relative. As I will argue below, Rabin’s proposed etymology for dialectal illī is essentially correct. However, his discussion does not address some of the significant morphological differences between the attested pre-modern and modern forms. For example, it is not entirely clear from Rabin’s discussion why the initial vowel in, e.g., CAr forms such as allāʾi changed from *u > a. More critically, he does not distinguish between forms that have only single -l versus 5 There is evidence from Arabic dialects for such a contrast in, e.g., the third person singular subject pronouns: cf. 3ms/3fs hū(wa) / hī(ya) alongside huwwa / hiyya (cf. Fischer & Jastrow 1980: 80). This latter contrast, however, is of a quite different nature. There is little if any phonological difference between /ū/ and /uw/, but there is between /ʾūl/ and /ʾull/.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

132

Phillip W. Stokes

those with geminate -ll, and how either patterns with the initial vowels. I will argue that these are crucial points in any attempt at identifying the etymology of illī, as well as understanding the relationships between the attested forms. Rabin rejects (ibid., 155) the traditional explanation, namely that it is a portmanteau relative composed of three elements: the definite article al + asseverative particle la + demonstrative ḏī (Brockelmann 1908: 324; Barth 1913: §68, p. 157; Huehnergard 2017). The reasons for his rejection of this etymology are not entirely clear, but he seems to find it problematic because allaḏī is equivalent to the Hebrew demonstrative hallāze, which apparently rules out the possibility for Rabin that al is the same as the Arabic definite article. As Huehnergard (2017: 17) notes, however, its presence in Hebrew as a demonstrative means that its morphology predates proto-Arabic. It is also possibly paralleled by Ugaritic hnd (Pat-El 2009: 41). It is therefore probable that Hebrew hallāze and Arabic allaḏī reflect a common inheritance from Central Semitic, in which it functioned as a demonstrative. The innovation in Arabic is its use as a relative (Huehnergard 2017: 17). Therefore, none of the issues that Rabin seems to find objectionable argue against the etymology of al + la + ḏī. In order to properly evaluate the morphology of the variants, as well as the relationships between the various pre-modern and modern *ʾVl-based forms, however, we will be served by a review and discussion of the attested *ʾVl demonstrative and relative forms in Arabic.

3. Morphology of *Vl-based demonstratives and relatives In this section, I will review the attested forms from CAr and the Qurʾān (both the QCT and the reading traditions), as well as the modern dialects, and discuss Rabin’s proposal in light of these forms. After a discussion of these issues, I will propose a proto-Arabic form of the plural demonstrative, discuss how the various forms relate to each diachronically, and show how we can in fact derive dialectal illī and the attested variants from the plural demonstrative.

6 Oddly, the nominative form of the determinative pronoun is spelled ‫ اوال‬in a surprisingly large number of early QCT manuscripts. It is not clear what, if any, difference in pronunciation this represents. I thank Marijn van Putten for bringing this spelling variant to my attention.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

133

D i s t a l ProxPlural imal Plural

Table 2: CAr Demonstratives (Wright I: §340; Fischer 2002: §274–275) ʾulā ʾulāʾi ʾulāka ʾulāʾika ʾulālika

‫ألى \ أولى \ أوال‬ ‫أالء \ أوالء‬ ‫أالك \ أوالك‬ ‫أالئك \ أوالئك‬ ‫أاللك \ أواللك‬

Table 3: CAr Determinative Pronouns (Wright I: §340.c) Plural

(nom)ʾulū / ʾūlū (obl) ʾulī / ʾūlī

‫ألوا \ أولوا‬ ‫ألي \ أولي‬

6

Table 4: CAr Relative Pronouns (Wright I: §347; Rabin 1951: 154) Plural

ʾallāyi ʾallāʾi ʾallāʾī al-ʾulā (nom) allāʾūn / (obl) allāʾīn

‫ألى‬ ‫ألى‬ ‫أالءي‬ ‫األلى \ األولى‬ ‫االؤون \ االئين‬

As noted above, Wright (I: §340.a) notes that the initial -u vowel is always realized short, and proposes that the spelling with the mater lectionis [w] is scriptio plena in order to distinguish demonstratives from e.g., the preposition ʾilā. Variants with long ū are reported by the grammarians, as Rabin himself points out (Rabin 1951: 156). It is difficult to know what to make of the forms with long ū, however. As is well known, most CAr spelling conventions follow the conventions used to write down the Qurʾān. Demonstratives in the QCT are as follows:

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

134

Phillip W. Stokes

Table 5: QCT Demonstratives Plural Proximal Plural Distal

‫اوال‬ ‫اولىك‬

In the reading traditions, these are read in line with the CAr forms /ʾulāʾi/ and /ʾulāʾika/ respectively. Based on the QCT spellings, several interpretations are possible, including /ʾulā/ or /ʾulāʾ(i)/, depending on one’s position regarding the hamza in the Qurʾān. The distal could be interpreted as /ʾulāʾika/ if long ā was present but not written. Alternatively, it could represent /ʾulayka/ or /ʾulēka/ without a long ā. The reading traditions transmit only short u realizations of the initial syllable. Rabin (1951: 156) argues that this might have been the case at the time of the original composition of the Qurʾān, but that the spelling hints at an originally long ū realization. There are, however, other examples in the QCT of the sequence /alif + glide to spell short vowels, e.g., Q21.37; Q7:145 ‫ساورىكم‬ /saʾurīkum/ “I will show you.” Other combinations of alif + waw/yaa to spell short vowel combinations are attested in the QCT as well: cf. Q6.34 ‫نباى‬ /nabaʾi/ “(of) the news (gen);” Q8.65 et passim ‫ ماىه‬/miʾah/ “one hundred.”7 Until we have a fuller picture of the early development of the Arabic script and scribal practices in the Ḥijāz, we cannot put much weight on the orthography alone. Even if we want to argue from the orthography, it is just as possible that those who relied on the QCT itself mistook the plene spelling with w to signify underlying ū, and that forms with long ū subsequently became accepted, canonical variants in CAr. Finally, in addition to the masculine plural relative ‫ الدىن‬/allaḏīn/, the QCT also attests a feminine plural relative spelled ‫( الى‬cf. Q33:4). There is some variation in the way the reading traditions transmit this form. It is read variously as: allāyi (Ibn Kathīr variant), allāʾi (Ibn Kathīr and Warš) and allāʾī (al-Kisāʾī) (Ibn Mujāhid: 518–519). The orthography admits of several 7 It is, of course, quite possible, perhaps even probable, that the glottal stop /ʾ/ had been lost in the dialect behind the QCT (cf. Vollers 1906; Versteegh 1997: 42). In this case, the spelling ‫ ساورىكم‬could represent underlying /sāwrīkum/ or /sawrīkum/ or even /sawurīkum/; the sequence ‫ نباى‬would likewise reflect underlying /nabāy/ or /nabay/ or even /nabayi/. Diem (1979) argues that these represent a mixed historical/phonetic spelling, where the alif is a historical spelling and the glide reflects the underlying realization. In that case, the spelling ‫ او‬in the demonstrative pronouns could be a historical spelling or, perhaps more probably, reflect /awlā/ or perhaps even /awulā/.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

135

different readings. Since it is common in the QCT to leave non-final long ā unwritten, the orthography could represent either /allāy(i)/ or /allāyī/. Alternatively the orthography is compatible with short a, namely /allay/ or /allē/. It seems to me unlikely, however, that the orthography can represent /allāʾ(i)/, for which we would expect ‫اال‬, the spelling that is consistently used to write the plural demonstrative (see above). We can, therefore, rule out the forms with long vowels, namely allāyi, allāʾi and allāʾī as well. The form allāyi can also be ruled out as it appears that etymological sequences āy/wv shifted to āʾv in the language of the Qurʾān (see below for discussion). Thus we find, e.g., *samāyv “heavens” spelled ‫سما‬ /samāʾ/ or /samā/ (Q41.12 et passim). Regarding allāyī, while the Cairo Edition avoids spelling the sequence yi with ‫ىى‬, older Qurʾānic manuscripts regularly spell them thus (Van Putten, p.c.). We would expect the combination yī to be written ‫ىى‬, but it is only spelled with one ‫ى‬. Finally, as word-final i preceded by a hamza is typically spelled with ‫اى‬, e.g., Q6.34 nabaʾi “the news (of the messengers),” it is unlikely as well. This leaves three possibilities: allay, allē, or allāy.8 Several aspects of the morphology of these forms should be noted. First, the quality of the initial vowel of the demonstratives is ubiquitously u, whereas in the relative forms it is always a in the pre- and early Islamic forms Rabin cites, as well as in a number of the modern dialectal forms (e.g., Sudanese allī; Reichmuth 1983: 113). The other modern relative forms have either initial i or ə; no dialect of which I am aware has initial u.9 Second, the only attested alternation is between short and long ū, with the majority of the evidence short u; there is no attested contrasts between **ʾul and ʾūl forms in Arabic to support Rabin’s proposal. Significantly, only relative pronouns have geminated ll, e.g., allāʾi. All of the demonstratives, conversely, have only a single l. Finally, Rabin’s explanation of the shift from *-ay > -āʾi is not particularly convincing. Regarding the example he cites, ištaraʾū ḍ-ḍalālata “they acquired error,” he argues that the reason for breaking up the diphthong aw was to resolve what would have been a heavy syllable in a closed, unstressed position. Resolving the diphthong in this way, however, still results in a super heavy syllable, namely ʾūḍ, in a closed, unstressed position. Nor, it should be noted, are the results of the two examples the same: the verbal process he cites reflects *aw > 8 I thank Marijn van Putten (p.c.) for pointing out the Qurʾānic evidence and its importance for this topic. 9 The one exception is in the dialect of Rijal Almaʿ (NW Yemen), which attests inflection for masculine singular (ḏā), feminine singular (tā) and plural (wulā) (Asiri 2008: 73). The plural form attests the common shift of ʾ > w/ #_. This dialect provides supporting evidence for the reconstruction of a demonstrative with an initial short *-u.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

136

Phillip W. Stokes

aʾū but the process attested for the relative forms is *ay > āʾi (or even āʾī). In short, Rabin’s etymology for illī is attractive, but the evidence in support of it is weak. In what follows, I will review each component of the demonstratives, contrasting them with the relative forms, and propose paths of development from the proto form to the attested forms.

4. Diachronic analysis of ʾVl-based demonstratives and relatives Scholars have long recognized the existence of ʾVl-based demonstratives, reconstructible with a high vowel, either *ʾil or *ʾul (Barth 1913: 118; Hasselbach 2007: 19). The primary points of contention regarding the Arabic forms are 1) the quality and length of the initial vowel, 2) the length of the consonant, and 3) the quantity of the final vowel sequence. 4.1 Quality and length of initial vowel While certain traditions within the CAr canon allowed long vowel realizations of the initial u in demonstratives, they are in the minority (Rabin 1951: 156).10 As noted above, the spelling of syllable-initial /ʾu/ with ‫ او‬is unusual, but not unique in the QCT. Rabin argues that the short u represents a shortening of long ū, which occurred when the second a was lengthened, i.e., from *ʾūlay > ʾulāʾi. This would have, he argues, resulted in a stress shift *ʾū́ lay > ʾullā́ ʾi. There are several problems with this proposal. First, it does not explain CAr al-ʾulā (< *al-ʾulay), which attests a short u that, per Rabin’s explanation should have retained the stress. More significant is the absence of attested alternation between ʾull and ʾūl. Pre-modern data strongly suggests a short u in the proto form. Evaluating the possible contribution of modern dialectal evidence in this discussion is difficult given that most plural demonstrative forms seem to reflect the combination of *hā + ʾul or *ḏā + ʾul (Magidow 2016).11 The NW Yemeni dialect of Rijāl Almaʿ, which is one of the few dialects with a *Vl10 As posited above, these forms may have been abagadical. Whether this is true or not, it is not clear what may be concluded from extremely marginal forms reported by the grammarians but rarely, if at all, attested. 11 This combination obscures the original length of the u because both *hā + ʾul and hā + ʾūl would manifest as either hawl or hōl in the modern dialects (see the discussion in Magidow 2016).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

137

based relative without a prefixed hā, supports a short u reconstruction as well. There, the relative form wulā probably reflects *ʾulā via the regular shift of word-initial *ʾ > w (cf. wēn “where?” < *ʾayn(a)), Asiri 2008: 73).12 Given the complete unanimity of the attested forms, a short initial syllable *ʾu for the proto-Arabic plural demonstrative seems safe.13 4.2 Single and Geminate ll in demonstratives and relatives The demonstrative forms reviewed here are without exception composed of only a single l. The relative forms, on the other hand, each have a geminate ll. This distribution is crucial to understanding the relationship between the demonstrative and relative forms. I follow Brockelmann, Barth and Huehnergard in arguing that the most likely origin of the initial morpheme in allaḏī, al, is the definite article, and the second morphene, la, is the asseverative marker. This is supported by variants such as * ʾaḏḏī, attested in a number of modern dialects, e.g., North African variants əddi and ddi (Retsö 2004: 267), both of which are composed of the definite article and demonstrative without the asseverative la, i.e., < *al-ḏī. An explanation for the difference between demonstratives based on ʾul, and relatives based on all is that the latter is composed of the definite article al + the plural demonstrative ʾul. This combination is transparently attested in the CAr variant form al-ʾulā (as recognized already by Barth 1913: 157–158). To explain why we get ʾall in most cases instead of al-ʾul, I suggest that short initial u of the demonstrative was syncopated in open syllables. Syncope of a short high vowel in open, unaccented syllables is well-known across both ancient and modern data (Teeple 2009; Al-Jallad 2017: §4.2.5). And, as David Testen (1998: 215) has pointed out, the same process proposed here is attested in equivalent phonetic contexts elsewhere, cf. allāh14 < *al-ʾilāh “God,” and ʾan-nās < *al-ʾunās “people.” There are, in fact, two possible paths that can account for this development. One is simply the loss of a high vowel in an open unaccented syllable: *al-ʾunā́ s > al-ʾnā́ s > an-nā́ s. As Marijn van Putten (2018a) shows, there is evidence in the QCT that post-consonantal hamza 12 It should be noted, however, that this shift is not well understood, nor is its conditioning clear. 13 The Semitic evidence points toward a high vowel, but whether it was *u or *i is unclear. Hebrew and Aramaic suggest *i, but Akkadian and Arabic both have *u. The evidence from Gəʿəz and Modern South Arabian is ambiguous. 14 As Marijn van Putten reminds me, the spelling rather suggests a pronunciation /alilāh/ in the dialect behind the orthography. Also inexplicable is the unexpected backed realization of the ā, typically conditioned by the presence of an emphatic consonant.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

138

Phillip W. Stokes

is lost. Further, pretonic short high vowels are often syncopated (cf. kāfirū́ n > kāfrū́ n > kafrū́ n “non-believers”; ibid.). The same process is clearer yet in Arabic in Greek transcription, cf. Αβδοαρθας /ʿabdo-ḥārṯa/ < /*ʿabdo-ḥāriṯa/ (Al-Jallad 2017: 151–152). The process was likely as follows: al-ʾunās > al-unās > al-nās > an-nās. Stress is the primary difference between these examples of syncope and the relative al-ʾulā. Indeed, it seems a priori likely that there was a great deal of variation in the way stress was realized in CAr. Plausibly, the stress pattern of the speaker’s dialect would play a determining role, as it does for users of Modern Standard Arabic today (McCarus 2008: 240). The variants of the relative pronoun that attest syncopated short vowel suggest that the stress did not fall on the syllable ʾu, otherwise it is hard to imagine its loss. It is possible, as Rabin suggested, that the stress fell on the ultimate syllable. It is also possible that stress shifted to the initial syllable of the relative forms when the (etymological) definite article al was prefixed. This stress pattern, in which the article is treated as part of the word with respect to stress, is attested in a number of dialects today, in, e.g., Negev Arabic: ál-walad “the boy;” ál-bil “the camels” (Henkin 2008: 362–363). The initial syllable of the plural relative forms beginning (ʾ)all- has its origin in the following process: *ʾal-ʾul > ʾalul > ʾall. The latter can possibly be considered a Ḥijāzī development, especially given the evidence for syncope in the QCT, while the unsyncopated forms such as al-ʾulā are precisely what we would expect in CAr. Despite the overwhelming evidence that suggests *ʾul as the correct reconstruction for proto-Arabic, it is nevertheless surprising that we cannot reconstruct *ʾull for proto-Arabic given the comparative data. The following are the attested reflexes of the ʾVl-based plural demonstrative in the Semitic languages for which we have vocalic evidence: Table 6: *Vl-based Demonstratives in Semitic Languages (Hasselbach 2007) far demonstrative m Old Babylonian sg. ullûm pl. ullûtum Biblical Heb.

near demonstrative

f sg. ullītum pl. ulliātum cp. ʾellē

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

139

The Plural Demonstratives and Relatives

Biblical Aram.

cp. ʾillēk

Syriac

cp. hālēn

cp. ʾellē

Jud. Pal. Aram. cp. ʾillên / hāllên / haylên Gəʿəz

nom.pl. ʾəlləku

Mehri

cp. ʾəlyēk

nom. pl. ʾəlləku

nom. pl. ʾəllu / ʾəllontu

nom. pl. ʾ əllā / ʾllāntu

cp. ʾəlyōməh

Each of these traditions attests a geminate -ll, with two exceptions, Syriac and Mehri. In Syriac, historically geminate consonant combinations were simplified to a singleton (Healey 2011: 645). The Mehri form attests a single l, thus patterning with Arabic. Historic gemination is typically simplified in Modern South Arabian (Simeone-Senelle 2011: §4.3.2), so nothing can be determined from the synchronic form of the relative in those languages. The single l here could, alternatively, be the result of contact with Arabic. Thus Arabic is essentially alone in attesting a single -l. Rabin assumes that a shift from *ʾull > ʾūl took place in some dialects, while others, mainly with forms like ʾulāʾi that underwent the lengthening of ā reflect a subsequent shortening of ʾū > ʾu. There is in fact no evidence for ʾull in Arabic, nor is ʾūl positively attested outside of some references in the works of the grammarians. However, Rabin’s proposed change of *ʾull > ʾūl > ʾul is still valid if we assume that the change occurred already in proto-Arabic. This would explain 1) the lack of geminate -ll in any attested variety of Arabic, and 2) the ubiquity of short u. 4.3 Quality and quantity of final vowel sequence The morphology of the second syllable of the demonstratives and relatives in Arabic is slightly more complicated. Rabin, following Barth (1913: 49, p. 120), reconstructs a final *-ay. This is reflected in the orthography of the QCT, where the form is written with the mater lectionis y. A final glide is corroborated by two unpublished inscriptions written in the Safaitic script discovered in NorthEast Jordan, in which the plural relative pronoun is spelled ʾly.15 In Safaitic, 15 These inscriptions, discovered in 2015 at Marabb aš-Šurafāʾ, will be published in a forthcoming work by the present author.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

140

Phillip W. Stokes

the letter y can only represent a diphthong (/ay/ or /āy/) or triphthong (/aya/ or /āyv/) in Safaitic (Al-Jallad 2015: §3.2.3). It is likely that the proto-form originally ended with *ay, and that the lengthened forms developed subsequently. There are two reasons for reconstructing a final *ay rather than a longer version *āy(i). First, evidence from other Semitic languages, such as Hebrew and Aramaic, is compatible with the former, but not the latter. The Hebrew demonstrative ʾēlle is derivable from *ʾillay but not **ʾillāy;16 the latter would have undergone the Canaanite shift of *ā > ō, and resulted in **ʾillōy. The proto-form *ʾillay is the likely origin of the various Aramaic plural demonstrative forms, such as Imperial Aramaic ʾlh (likely /ʾellē/; Gzella 2011: 577), as well as Syriac hālēn (< *hā +ʾillē + n). Second, as will be detailed below, the longer forms can be hypothetically derived from the shorter, but this seems far less likely. Before I address the origin of the lengthening, however, the issue of the change from final *ay > ā deserves a few comments. Final *-ay shifted to -ā in most CAr traditions (e.g., *ʿalay > ʿalā). Alternatively, evidence from the pre-Islamic dialects of the Southern Levant (Al-Jallad 2017: §5.1.1, p. 154), as well as the QCT (Van Putten 2017: §7), unlike CAr, attest the shift *-ay# to ē. While the shift of *-ay# to ē is also attested in the modern Yemeni dialect of Jabal Rāziḥ (Behnstedt 1987: 133), most modern dialects pattern with CAr in attesting the shift *-ay > ā / _#. This sound change, however, did not operate consistently on every instance of final *ay: for example, it is not reflected in non-indicative imperfectives of some II-weak verbs (e.g., 2fsg. tarḍay, indicative tarḍayna, “that you be satisfied”). If a regular phonetic change took place that shifted final *ay > ā, we would expect for **tarḍā < tarḍay. There are two possible explanations for the lack of operation of this sound change on, e.g., *tarḍay > tarḍay. The first is the possibility that the prepositions mentioned, e.g., ʿalā “upon; over” and ʾilā “to, toward,” along with the interrogative adverb matā “when?” and the conjunction ḥattā “until; so that” were all originally triphthongs, ending in *aya. In that case, then these prepositions and adverbs do not provide evidence for *ay# one way or the other. However, the triphthong *aya in CAr also shifted to ā.17 This is demonstrated by, e.g., III-y roots in the 3ms, where we find *banaya > banā “he built” and *banayat > banāt > banat “she built.” Unlike CAr, etymological *aya reduced 16 This pace Hasselbach (2007: 19), who argues for a long proto variant *ʾulli. This is not supported by the Arabic evidence, nor is it plausible for Hebrew or Aramaic, where final short vowels were lost. 17 That the triphthong was still realized in proto-Arabic is confirmed by the consistent spelling of these verbs with a final glide in Safaitic, where it can only be a consonant (see Al-Jallad 2015: §2.3).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

141

to ē in the southern Levant and Ḥijāz (Al-Jallad 2017: §4.6.1; Van Putten 2017: §6). Thus if the proto-form of CAr ʿalā were *ʿalaya, we would expect, when pronominal suffixes were added, that it would result in **ʿalāka “upon you” (< *ʿalayaka). That we find ʿalayka “idem” instead suggests that the prepositions ended in a diphthong ay.18 Stronger evidence against the originality of the triphthong comes from the Warš reading tradition of the Qurʾān. In that tradition, forms that ended in *aya are realized ē, which, as noted above, corresponds to evidence from the southern Levant and the QCT itself. However, ʿalā, ʾilā and ḥattā are realized with final ā. It therefore seems unlikely that these forms ended in triphthongs, at least not universally. A second possibility is that these forms ended in *ay, and those forms that seem to provide counterevidence to the sound change (*ay > ā / C_#) are best explained as analogical restorations of the diphthong. Or alternatively, we might assume that the sound change was blocked due to paradigm pressure. The primary source of counterexamples is the subjunctive and jussive forms of the 2fsg. imperfect of final glide roots. Since the indicative forms of these verbs end in na, e.g., tarḍayna “you are satisfied,” the diphthong ay in this form is not word-final, and thus would have been protected from loss. The change tarḍay > tarḍā was blocked in the non-indicative forms due to paradigmatic pressure. As we saw above, a number of variants with a long -ā are attested in CAr, as well as in the Qurʾānic reading traditions. Forms in CAr that end in -ā#, e.g., ʾulā and al-ʾulā, underwent the shift of *-ay > -ā that characterizes the prepositions ʿalā “upon, on; against” (< *ʿalay) and ʾilā “to, toward” (< *ʾilay). The more common demonstrative and relative forms in CAr end in -āʾi or, more rarely, -āʾī. Fischer (1959: 103) argued that the proto form was ʾulā, and that the glottal stop originated in a pausal variant, i.e., pre-pausal ʾulā // pausal ʾulāʾ. The final short i was then inserted in order to avoid a super heavy syllable. He further suggested that the pausal form was so common that it was eventually levelled to pre-pausal positions as well. While pausal glottalization is attested in some modern Arabic dialects (cf. NW Yemeni varieties, on which see Behnstedt 2016: 42), it is not regular in the dialects behind CAr. More significantly, his suggestion that the pausal form of the demonstratives were so common that they were levelled is, given the unmarked pre-nominal position of demonstratives, difficult to imagine. 18 It is admittedly possible that, when no pronoun was suffixed, the form ended in a tripthohng *aya, but when a pronoun was suffixed it was added to a ditphong: *ʿalaya + N, but *ʿalay + PN. Given the evidence, however, this seems unlikely.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

142

Phillip W. Stokes

I propose rather that the CAr āʾi forms developed from *āyi, which is in fact attested as a variant in a number of the reading traditions (cf. Warš allāyi). Proto-Arabic *āy/wv sequences shifted in the dialects behind CAr to āʾv (Brockelmann 1908: 138; Al-Jallad 2014: 11–12): *samāyv > samāʾv “heaven; sky.”19 It is likely that a proto form ending in *-ay is behind all of the attested variants, and that each development can be explained by regular phonetic and analogical processes. More specifically, the variants with final -ā and -āyi / āʾi represent three attested sound changes. The form ʾulā can be explained as result of the shift *-ay > ā_#. The more common variant ʾulāʾi is the result of lengthening of a in some forms and occurred before the shift *āyi > āʾi, resulting in *ulay > ʾulāy. Finally, the long form ʾulāʾi involves a further shift, namely the aforementioned *āy/wv > āʾv. In this reconstruction, the final i is an epenthetic vowel inserted in order to break up a heavy CVVCC syllable: *ʾulay n-nās > ʾulāy n-nās > ʾulāyi n-nās > ʾulāʾi n-nās. This same epenthesis to avoid the creation of super heavy syllables is attested elsewhere in the varieties behind CAr, e.g., ʿan + l-bayt > ʿani l-bayt “from the house.” The impetus for lengthening *-ay > *-āy remains to be explained. Rabin (1955: 153) appeals to secondary stress to explain the length of -ā, as well as a the insertion of a hiatus -ay > -aʾi to break up the diphthong and avoid a heavy syllable. I argue instead that the lengthening is due to analogy. The suffix *-ay occurs regularly in Arabic, mostly as a feminine singular marker (cf. kubrā < *kubrayv “bigger/biggest (fsg.)), but also to mark certain abstract nouns (Fischer 2002: §119; Van Putten 2018b). While there is no apparent connection between this suffix and plurality, it is nonetheless part of several plural patterns in Arabic, e.g., ʾafʿilāʾv < *ʾafʿilāyv, fuʿalāʾv < *fuʿalāyv, etc. I suggest that speakers formed the longer ending on the demonstrative and relative forms by analogy with this suffix, which was more transparently indicative of plurality than was *-ay. We see a similar, though not identical, process at work in the plural relative allaḏīna, which was analogically created by adding the adjectival plural morpheme -īna to the inherited portmanteau form allaḏī (Barth 1913: 156–157). Thus each of the variants found in CAr, the reading traditions, and the orthography of the QCT can be derived from the posited proto-form *ʾulay via regular processes, phonetic and/or analogical. In the final section of the paper I will show how the common dialectal relative illī and its variants can be plausibly derived from the relative form, attested in the QCT, *allay.

19 The proto form *samāy is likely reflected in the Safaitic form s¹my (Al-Jallad 2015: 38).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

143

5. Reconstruction of illī and related forms Previous scholars, who assumed illī originated from a demonstrative, did not address the difference between the relative with geminated -ll and the demonstratives with a single -l. Furthermore, the different vocalic endings attested: -i, -ī, -e, -ē and -a (see again above, section 1) were ignored. In this section I will argue in detail that the origin of illī is *allay, which is a combination of the definite article *al + demonstrative *ʾulay. The onset of the relative particle is typically il, but many dialects exhibit əl or al. I suggest that the initial Vl- of these forms is indeed etymologically identical to the definite article. Like the realization of the definite article in most dialects, the vowel is raised to i or ə. There are a few exceptions, e.g., Sudanese allī, where the definite article remained a low central vowel a (cf. Yemeni alli // alī). In dialects with al instead of il/əl, the waṣl vowel is a in general, cf. Yemeni aftakar “he thought” instead of iftakar “idem;” angatal “he was killed,” instead of ingatal “idem” (Behnstedt 2016: 241). These should probably be considered word-initial CC clusters, i.e., l(l)i, ftaham etc., where different dialects resolve the clusters with different vowels. The most common realization of the final vowel of the relative particle illī is i (< *ī). The widespread realization of the final vowel as i or ī might at first seem difficult to reconcile with an original *-ay since after all, outside of North Africa the typical realization of *ay is ē, with some dialects preserving ay (Fischer & Jastrow 1980: 54–55). However, in certain frequently occurring lexemes, *ay > ī is the norm across the dialectal continuum. For example, the dialectal form of *šayʾ “thing,” is typically realized šī or išī in dialects that normally attest *ay > ē (e.g., Syrian (i)ši “thing; something”). The interrogative *kayfa “how?” is commonly realized as kīf in those same dialects (for this change in, e.g., Egypt and Sudan, see Dickins 2011: 936–940). These words, like the relative illī, often occur relatively frequently and, prosodically speaking, without a primary accent. This realization of *-ay as -ī likely underlies the relative particle illī as well. As mentioned above, several variants that appear to be related to illī require comment. A minority of dialects attests relatives with either final -ē (e.g., Siirt lē) and -a (e.g., Tunisian Sāḥil alla; Mardin la; Jastrow 1978: 123–124). To contextualize the latter it should be recalled that various CAr and Qurʾānic forms show *-ay# > ā (e.g., ʾulā “these”), as well as āʾi (e.g., ʾulāʾi and allāʾi). Both of these sequences are realized as final -a in most dialects: CAr samāʾ “heavens,” dialectal sama “idem.” Therefore, it is plausible that dialectal forms with -a# in this form originated from a form that had either undergone *-ay > -ā or āʾi > āʾ > ā.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

144

Phillip W. Stokes

The forms with -e and ē are remarkable given that the majority of the modern dialects tend to pattern with CAr in the realization of original final -*ay sequences: CAr ʿalā “upon, on” // dialectal ʿala “idem;” CAr banā “he built” // dialectal bana “idem.” The unusual reflex of *ay and the widespread nature of the relative particle illī and its variants, however, strongly suggests the possibility that this particle spread via contact.20 The source dialect of this form, then, reflects a different set of changes than occurred in CAr. We have evidence of *-ay realized as /ē/ in the southern Levant (Al-Jallad 2017: 153–156), as well as in the Ḥijāz (Van Putten 2017: §7, 57–59), and thus we might imagine that this form originated in one of those regions as well. Further support for such an idea might be found in the epigraphic record. In an Arabic inscription JSLih 384, written at Dadān with the Dadanitic (an Ancient North Arabian) script, a singular feminine relative form spelled ʾlt is attested (Müller 1982: 32–33; Macdonald in Fiema et al 2015: 409). The relative ʾlt in JSLih 384 is identical to Qurʾānic ‫ التي‬and CAr allatī forms. Al-Jallad has suggested, based on the absence of such forms in available pre-Islamic Arabic corpora, that these portmanteau demonstratives as relatives might be innovations of Ḥijāzī dialects (Al-Jallad 2015: §1.2; see further Al-Jallad forthcoming for a more developed argument in favor of the Ḥijāzī provenance of these features). If the present proposal, that illī and its variants are derived from *allay (< *al-ʾulay), is correct, then the Ḥijāz, where we find *ay > -ay or -ē, can be considered a plausible place of origin.21 Such a hypothesis is strengthened by the fact that it is precisely in the southern Levant and Ḥijāz where we also find attestations of final *-ay realized as -ay or -ē. Indeed, as I suggested above, the spelling of the relative in Q33.4, written ‫الى‬, can be naturally read as /(ʾ)allay/ or /(ʾ)allē/. Several of the variants discussed so far have the shape lV or Vl, thus lacking the characteristic geminate -ll of illī. Examples of Vl include (Retsö 2004: 265): -- Iraqi il -- Lebanese əl -- Antiochene il 20 It is, however, not attested in the early papyri studied by Hopkins (1984), with one possible exception (ibid., §292). It is likewise very rare in Judaeo-Arabic texts (Blau 1999: 53). This rarity of pre-modern attestation can be explained in a few ways. It is possible that it was felt to be too vulgar, and thus tended not to be written. Another possibility is that it was simply still relatively restricted geographically, and began to replace previously more widespread forms recently. 21 Pace Rabin 1951: 154–155, who assumes that allaḏī and allay were competing forms, the former the native Ḥijāzī form

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

145

The Vl forms are generally assumed to originate in the definite article, an opinion I share. Dialects in which these forms occur typically exhibit relative clauses in which the head of the clause is in construct with the following clause, which is headed by Vl: sant il-fātat “last year (lit. “the year that passed”); zlimt as-sōlafta minu “the man with whom you spoke, who is he?” Examples of lV relatives include (Retsö 2004: 265; see above, section 1): -----

Diyarbakir lə Siirt lē Mardin la Maltese li

These lV variants appear to me to be reduced forms of illī ( allē > llē > lē. The occurrence of the head noun in construct with the lV forms is probably best considered an extension of an original Ncst + Relative Clause, which is un-controversially what happened in the dialect behind the Baḥrainī example cited above.22 While it is impossible to make conclusive statements based on the present evidence, the argument advanced here can account for all surface forms attested in the modern dialects.

22 Huehnergard & Pat-El (2018: 11) note that the same phenomenon, namely a head noun in construct with the relative pronoun, is attested in Biblical Hebrew as well. They plausibly suggest that this blurring of the distinction between relatives with construct heads and appositional heads is due, among other things, to the loss of inflectional case marking.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

146

Phillip W. Stokes

6. Summary and conclusions I have argued that the relative particle illī and its various morphological variants are likely derived from a combination of the definite article *al + the plural demonstrative *ʾulay. I further argued that, while the previous explanations reviewed above are possible, the present proposal accounts for the attested morphology via processes attested elsewhere in the language, without appealing to ad hoc sound changes. The following table is a review of the changes: Plural Demonstratives a) *ʾulay > (lengthening of a) ʾulāy > (epenthesis) ʾulāyi > (shift āyi > āʾi) ʾulāʾi; b) *ʾulay > (shift *-ay > -ā) ʾulā. Relative Forms a) *ʾulay > (addition of ʾal) *al-ʾulay > (shift *-ay > -ā) al-ʾulā; b) *ʾulay > (addition of ʾal) *al-ʾulay > (syncope of u) *allay > (*-ay > ē)ʾallē; c) *ʾulay > (addition of ʾal) *al-ʾulay > (syncope of u) *allay > (*-ay > ē, ī) allī / allē > (fronting/raising of *a > i) illī / illē > (syncope of initial i) lli / lle > (simplification of gemination -ll?) li / le (?). If accurate, then all modern dialects attest a relative form that functioned originally to mark either singular (those derived from *ḏī, *al-ḏī and *allaḏī) or plural. In other words, one form was levelled to mark all heads, regardless of number.23 This might have been helped by the phonological bleaching that occurred in the processes described above. It also entails that the plural relative that spread was based on the demonstrative *ʾulay, and that the plural forms that extended the adjectival endings to the relative, such as allaḏīna and allawāti, did not spread and are thus not attested in later dialects.

23 A similar case of the use of a plural pronoun referring to both singular and plural is the English colloquial use of the pronoun “they,” which can be used to refer to either a singular or plural referent, e.g., “Somebody left their book in my office.”

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

147

Bibliography Abu-Haidar, Farida. 2006. “Baghdad Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 222–231. Leiden: Brill. Abu-Mansour, Mahasen Hasan. 2008. “Meccan Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 179–187. Leiden: Brill. Al-Jallad, Ahmad. 2014. “On the genetic background of the Rbbl bn Hfʿm grave inscription at Qaryat al-Fāw.” Bulletin of the School of Oriental and African Studies 77: 1–21. . 2015. An Outline of the Grammar of the Safaitic Inscriptions. Leiden: Brill. ———. 2017. “Graeco-Arabica I: The Southern Levant.” In Arabic in Context: Celebrating 400 Years of Arabic at Leiden University, edited by Ahmad Al-Jallad, 99–188. Leiden: Brill. ———. Forthcoming. “The Damascus Psalm Fragment: Middle Arabic and the Legacy of Old Hijāzī”. LAMINE 2. Chicago: Oriental Institute. Arnold, Werner. 2006. “Antiochia Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 111–119. Leiden: Brill. Asiri, Yahya. 2008. “Relative clauses in the dialect of Rijal Alma’ (south-west Saudi Arabia)”. Proceedings of the Seminary for Arabian Studies 38: 71–74. Barth, Jakob. 1913. Die Pronominalbildung in den semitischen Sprachen. Leipzig: J. C. Hinrichs’sche Buchhandlung. Behnstedt, Peter. 1987. Die Dialekte der Gegend von Ṣaʿdah (Nord-Jemen). Vol. I. Wiesbaden: Harrassowitz. Behnstedt, Peter. 2016. Dialect Atlas of North Yemen and Adjacent Areas. Translated by Gwendolin Goldbloom. Leiden: Brill. Blau, Joshua. 1960. Syntax des palästinenischen Bauerndialekts von Bîr-Zêt. Walldorf: Verlag für Orientkunde. ———. 1999. The Emergence and Linguistic Background of Judaeo-Arabic: A Study of the Origins of Neo-Arabic and Middle Arabic. 3rd Edition. Jerusalem: Ben-Zvi Institute. Brockelmann, Carl. 1908–1913. Grundriss der vergleichenden Grammatik der semitischen Sprachen. Berlin: Reuther und Reichard. Corriente, Federico. 2006. “Andalusi Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 101–111. Leiden: Brill. ———. 2013. A Descriptive and Comparative Grammar of Andalusi Arabic. Leiden: Brill. Dickins, James. 2007. “Khartoum Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. II, 559–571. Leiden: Brill. Dickins, James. 2011. “Dialects of Egypt and Sudan.” In The Semitic Languages: An

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

148

Phillip W. Stokes

International Handbook, edited by Stefan Weninger, Geoffrey Khan, Michael Streck and Janet C.E. Watson, 935–953. Berlin and Boston: Mouton de Gruyter. Diem, Werner. 1979. “Untersuchungen zur frühen Geschichte der arabische Orthographie I. Die Schreibung der Vokale.” Orientalia N.S. 48: 207–257. Fiema, Zbigniew T., Ahmad Al-Jallad, Michael C.A. Macdonald and Laïla Nehmé. 2015. “Provincia Arabbia: Nabataea, the Emergence of Arabic as a Written Language, and Graeco-Arabica.” In Arabs and Empires before Islam, edited by Greg Fisher, 373–433. Oxford: Oxford University Press. Fischer, Wolfdietrich. 1959. Die demonstrativen Bildungen der neuarabischen Dialekte: Ein Beitrag zur historischen Grammatik des Arabischen. The Hague: Mouton & Co. ———. 2002. A Grammar of Classical Arabic. Translated by J. H. Rodgers. New Haven: Yale University Press. Fischer, Wolfdietrich and Otto Jastrow. 1980. Handbuch der arabischen Dialekte. Wiesbaden: Harrassowitz. Grand’Henry, Jacques. 1972. Le Parler Arabe de Cherchell (Algérie). Louvain: Institut Orientaliste Louvain-La-Neuve. Gzella, Holger. 2011. “Imperial Aramaic,” In The Semitic Languages: An International Handbook, edited by Stefan Weninger, Geoffrey Khan, Michael Streck and Janet C.E. Watson, 574–586. Berlin and Boston: Mouton de Gruyter. Hasselbach, Rebecca. 2007. “Demonstratives in Semitic.” Journal of the American Oriental Society 127: 1–27. Healey, John F. 2011. “Syriac.” n The Semitic Languages: An International Handbook, edited by Stefan Weninger, Geoffrey Khan, Michael Streck and Janet C.E. Watson, 637–652. Berlin and Boston: Mouton de Gruyter. Heath, Jeffrey. 2002. Jewish and Muslim Dialects of Moroccan Arabic. New York: Routledge Curzon. Henkin, Roni. 2008. “Negev Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 360–369. Leiden: Brill. Holes, Clive. 2006. “Bahraini Arabic,” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 241–255. Leiden: Brill. ———. 2007. “Kuwait Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. II, 608–620. Leiden: Brill. ———. 2008. “Omani Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 478–491. Leiden: Brill. Hopkins, Simon. 1984. Studies in the Grammar of Early Arabic Based on papyri datable to before 300 A.H./912 A.D. Oxford: Oxford University Press. Huehnergard, John. 2017. “Arabic in its Semitic Context.” In Arabic in Context: Celebrating 400 Years of Arabic at Leiden University, edited by Ahmad Al-Jallad, 3–34. Leiden: Brill.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

The Plural Demonstratives and Relatives

149

Huehnergard, John and Na’ama Pat-El. 2018. “The Origin of the Semitic Relative Marker.” Bulletin of the School of Oriental and African Studies 81: 1–14. Jastrow, Otto. 1978. Die mesopotamisch-arabischen Qəltu-Dialekte, Band I: Phonologie und Morphologie. Wiesbaden: Steiner. Lentin, Jérôme. 2006. “Damascus Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 546–555. Leiden: Brill. Magidow, Alexander. 2016. “Diachronic Dialect Classification with Demonstratives.” Al-ʿArabiyya 49: 91–115. Mansour, Jacob. 2006. “Baghdad Arabic Jewish.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 231–241. Leiden: Brill. McCarus, Ernest. 2008. “Modern Standard Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 238–262. Leiden: Brill. Mifsud, Manwel. 2008. “Maltese.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. III, 146–159. Leiden: Brill. Ibn Mujāhid. Kitāb al-sabʿah fī l-qirāʾāt. Edited by Shawqī Ḍayf. Cairo: Dār al-maʿārif. Müller, W.W. 1982. “Das Altarabische der Inschriften aus vorislamischer Zeit.” In Grundriß der arabischen Philologie, Band I Sprachwissenschaft, edited by Wolfdietrich Fischer, 30–36. Wiesbaden: Reichert. Naïm, Samia. 2006. “Beirut Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 274–286. Leiden: Brill. Pat-El, Na’ama. 2009. “The Development of the Semitic Definite Article: A Syntactic Approach.” Journal of Semitic Studies 54: 19–50. Rabin, Chaim. 1951. Ancient West-Arabian. London: Taylor’s Foreign Press. Reichmuth, Stefan. 1983. Der arabische Dialekte der Šukriyya im Ostsudan. Zürich und New York: Georg Olms Verlag. Retsö, Jan. 2004. “Relative-Clause Marking in Arabic Dialects: A Preliminary Survey.” In Approaches to Arabic Dialects: A Collection of Articles presented to Manfred Woidich on the Occasion of his Sixtieth Birthday, edited by Martine Haak, Rudolph de Jong, and Kees Versteegh, 263–273. Leiden: Brill. Simeone-Senelle, Marie-Claude. 2007. “Horn of Africa.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. II, 268–275. Leiden: Brill. ———. 2011. “Modern South Arabian.” In The Semitic Languages: An International Handbook, edited by Stefan Weninger, Geoffrey Khan, Michael Streck and Janet C.E. Watson, 1073–1113. Berlin and Boston: Mouton de Gruyter. Taine-Cheikh, Catherine. 2007. “Ḥassāniyya Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. II, 240–250. Leiden: Brill. Teeple, David. 2009. “Syncope,” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. IV, 389–391. Leiden: Brill. Testen, David. 1998. “Literary Arabic and Early Hijazi Contrasts in the Marking of

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

150

Phillip W. Stokes

Definiteness.” Perspectives on Arabic Linguistics XI, edited by Elabbas Benmamoun, Mushira Eid and Niloofar Haeri, 207–225. Philadelphia: John Benjamins. Van Putten, Marijn. 2017. “The development of the tripthongs in Quranic and Classical Arabic.” Arabian Epigraphic Notes 3. 47–74. ———. 2018a. “Hamzah in the Quranic Consonantal Text.” Orientalia N.S. 87: 93– 120. ———. 2018b. “The feminine endings *-ay and *-āy in Semitic and Berber.” Bulletin of the School of Oriental and African Studies 81: 205–225. Vicente, Ángeles. 2009. “Relative Pronoun (Arabic Dialects).” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. IV, 70–72. Leiden: Brill. Vollers, Karl .1906. Volkssprache und Schriftsprache im alten Arabien. Strassburg: K.J. Trübner. Woidich, Manfred. 2006. “Bʿēri Arabic.” In Encyclopedia of Arabic Language and Linguistics, edited by Kees Versteegh, vol. I, 299–308. Leiden: Brill. Wright, William. 1874–1896/2005. A Grammar of the Arabic Language, 2 volumes, 3rd Edition. Mineola, NY: Dover.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar1,2 Janet C.E. Watson, University of Leeds, and Abdullah Musallam al-Mahri, Salalah, Dhofar In this paper, we provide a synchronic account of word stress in the Modern South Arabian language, Mehri, as spoken by members of the Bit Thuwar tribe. The data is taken from the first author’s own fieldwork working in Central Dhofar with members of the Bit Thuwar sub-tribes–Bit Iqhōr in Rabkut and parts of the mountains that receive the monsoon rains, and Bit Āmawsh in Dhahbun–with reference, where appropriate, to Johnstone (1987).3 This paper is a significant expansion and a partial revision of the short discussion on word stress in Watson (2012: 34–35). We begin with a brief background to Mehri within Modern South Arabian. We then discuss word stress patterns in Mehri, following Hayes’ (1995) metrical stress theory; here we show that Mehri is a head-first or trochaic language, namely that in (Cv)CvCv(C) forms, stress is placed on the left-most syllable. We show that in contrast to Arabic trochaic dialects, the domain of stress is the stem and the stem with stem-level suffixes rather than the entire prosodic word. The implications for this is that Mehri exhibits an opacity with regard to stress as seen in Arabic dialects in which three consonant clusters receive vowel epenthesis after the left-most consonant, as in: šuft-ha > šufitha ‘she said it f.’. (These are termed vC-dialects in Kiparsky 2003, so called due to the position of the epenthetic vowel in relation to the medial consonant in a sequence of three). By opacity, we mean that word stress is not assigned as would be predicted by the stress algorithm. In contrast to Arabic vC-dialects, however, opacity is due not to the interaction of epenthesis and syncope (cf. Kiparsky 2003), but rather to the lack of 1 We thank the Leverhulme Trust for project grant: RPG-2012-599 (2013–2016), during which time the research for this paper was produced. We also thank Saeed al-Mahri and Ali al-Mahri for comments during the writing of the paper, and Musallam Hazmay al-Mahri for providing some of the data. 2 The abbreviations used in this paper are: cond conditional; dim diminutive; du. dual; f. feminine; ind indicative; m. masculine; pl. plural; s. singular; subj subjunctive. 3 Contra descriptions of an ‘Omani Mehri’ made by Bendjaballah (2017), Bendjaballah & Ségéral (2014, 2017), Rubin (2010, 2018, etc.) and others, the Mahrah themselves do not distinguish varieties of Mehri on national grounds, but rather identify language varieties according to tribal affiliation and geographical location.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

152

Janet C.E. Watson and Abdullah Musallam al-Mahri

visibility of word-level suffixes to stress. In line with van Oostendorp’s (2002) analysis of unstressable suffixes in Dutch, we argue that word-level suffixes are invisible to stress because they are not incorporated into the prosodic word, but rather adjoined to it. The adoption of Kiparsky’s Stratal Optimality Theory approach enables us to capture Mehri stress assignment succinctly: stress is assigned at the stem-level according to weight and position, and suffixation of word-level suffixes can no longer affect stress assignment due to the high ranking of StressIdent (Collie 2007), which requires stress to remain on the stressed syllable of the stem, and the low ranking of *Adjoin, which mitigates against the adjunction of affixes to the prosodic word. We also show that Mehri exhibits limited lexical stress, and suggest that attempts to account for stress in these cases in terms of a quantity model on the part of many researchers, including the main author of the current paper, has led to the incorrect transcription and interpretation of these elements.

1. Background to Mehri Mehri is one of six Modern South Arabian languages. It is the most widespread of the language group, and is spoken in the southern parts of the Arabian Peninsula in eastern Yemen, western Oman, and parts of central southern and eastern Saudi Arabia. Mehri is considered moderately endangered, with around 180,000–200,000 speakers, although the number is impossible to determine with any precision since it is spoken over three state borders, separate census figures within the states concerned are not provided for ethnic communities, and many younger generation Mahrah have limited, or in some cases, no command of the language. The language is further endangered by having no traditional script; thus, written communications and education are conducted in Arabic rather than Mehri, and Mehri, together with the other Modern South Arabian languages, is not taught or recognised at any formal educational level. The paper draws primarily on data collected since 2009, much of it as part of a documentation project funded by the Leverhulme Trust, RPG2012-599,4 which began in January 2013. All data was checked with Abdullah Musallam al-Mahri, the second author and a native speaker of the Bit Iqhōr variety of Mehri, and Saeed al-Mahri and Ali al-Mahri, native speakers of the Bit Ṣamōdah variety of Mehri, as spoken in Central Dhofar.

4 See https://www.leeds.ac.uk/arts/info/125219/modern_south_arabian_languages.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

153

2. A note on terminology In this section, we define the terms nonconcatenative morphology, prosodic word, and stem. Mehri exhibits a great deal of nonconcatenative morphology, where much derivational and inflectional morphology is expressed not through the linear concatenation of morphemes but rather through the association of elements of the consonantal root to a different prosodic template, and vowel umlaut. The degree of nonconcatenative morphology is greater than in Arabic and most other Semitic languages. For example, the adjective rḥaym5 ‘beautiful m.s.’ with the prosodic template CCīC, has the masculine plural form rīḥōm, with the root consonants inserted into the prosodic template CīCōC, and the elative form arḥām, with the prosodic template aCCāC. Similarly, the perfect verb syūr ‘he/they f. went’ is inflected for masculine plural through vowel umlaut to give syawr ‘they m. went’, and the distinction between 2ms and 2fs in the imperfect is expressed by vowel umlaut in the stem, as in the difference between tḥōm ‘you m.s. want’ and tḥaym ‘you f.s. want’. Following Peperkamp (1999) and others, the prosodic word is defined as the domain of word stress, phonotactics and segmental word-level rules. In Mehri, the raw stem is the vocalised core of a content word–a noun, verb, adjective, participle, pronoun or adverb(ial)–which lacks inflectional prefixes or suffixes, and may be realised independently of affixes or may form part of an affixed word. The results of nonconcatenative derivational or inflectional processes constitute a stem, thus both tēṯ ‘woman’ and īnēṯ ‘women’ constitute stems. The stem is minimally monosyllabic and bimoraic, and maximally, but rarely, trisyllabic. Thus, šarḥawm ‘he/they f. were blessed with rain’ is a bisyllabic stem, which comprises the triconsonantal root {r-ḥ-m}, the derivational prefix ša- to derive a Š-form verb (Rubin 2018: 139), and the vocalic melody /a-ū/.6 When it stands alone, it also functions as a prosodic word. The prosodic word šarḥawm-an ‘we were blessed with rain’ comprises the vocalised stem šarḥawm and the inflectional suffix -an. Similarly, śxawlūl ‘he/they f. sat, stayed’ is both a prosodic word and a stem, and the prosodic word śxawlūl-an ‘we sat’ comprises the stem śxawlūl plus the inflectional suffix -an; and rīḥōm ‘beautiful m.pl.’ is a prosodic word and a stem, while rīḥam-tan ‘beautiful f.pl.’ comprises the stem rīḥōm plus the feminine plural suffix -tan.7

5 < rḥīm with diphthongisation of /ī/ following pharyngeal /ḥ/. 6 With /ū/ > [aw] after /ḥ/. See below. 7 /ō/ and /ē/, when shortened, are realised as [a].

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

154

Janet C.E. Watson and Abdullah Musallam al-Mahri

3. Word stress patterns in Mehri Mehri generally assigns word stress according to syllable weight in basic stems, and, as seen below (3.2), exhibits lexical stress in a few cases. The syllables in Mehri can be divided into light syllables, ((C)C)v; heavy syllables, ((C)C)vv and ((C)C)vC; and superheavy syllables, ((C)C)vvC and (C)vCC.8 The final consonant in a word does not count for weight; thus, while ((C)C)vC counts as heavy in word-medial position, it counts as light in word-final position. Note that the diphthongs /ay/ and /aw/ pattern as long vowels in final position, rather than as vC sequences, and that initial consonant clusters, including initial geminates, and vowel-initial syllables are permissible word initially. 3.1 Mehri stress algorithm In prosodic words comprising unaffixed stems, the stress algorithm for Mehri reads as follows: (1) a. Monosyllables: Stress a monosyllable of the shape CvCC, (C)CvvC or Cvv: -- ˈśī ‘thing’, ˈlā ‘no; not’, ˈbū ‘people’, ˈbark ‘knee’, ˈḏēk ‘that m.s.’, ˈḏayk ‘that f.s.’, ˈʕiśś ‘he got up’, ˈśxōf ‘milk’, ˈbʕayr ‘adult male camel’, ˈmray ‘camel with milk but no young’. b. Multisyllabic words: Stress final Cvv or CvvC, CvCC: -- Nouns: bīˈrōk ‘knees’, śīˈwōṭ ‘fire’, ṭuwāˈdēd ‘one [dim]’, agˈzē ‘sunset’, mūˈsē ‘rain’, ḥāˈbū ‘people’, ḥabˈrē ‘son [of]’, ḥāˈmē ‘mother [of]’; -- Pronouns: aˈtēm ‘you m.pl.’, aˈkay ‘we [dual]’, aˈtay ‘you [dual]’ -- Verbs, perfect: haśˈkūr ‘to get lots of milk while milking’, šūˈṯūḳ ‘to be fixed’, ūˈbūd ‘to hit; to be’; -- Verbs, imperfect: yˈḥaym ‘they m. want’, yāˈṣōṣ ‘he is afraid’, yāˈmēr ‘he says [subj]’; -- Adjectives: īˈdīn ‘new m.s.’, ṭīˈwōl ‘long m.pl.’, ṭīˈḳōl ‘heavy m.pl.’; -- Adverbs: ḥaˈlawk ‘there’, arˈḥāḳ ‘further’, rwāˈḥāk ‘further [dim]’.

8 As is conventional in works on syllable structure, C stands for any consonant and v stands for any vowel.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

155

c. In the absence of final Cvv or CvvC or CvCC, stress the rightmost non-final heavy syllable (Cvv or CvC), as in: -- Nouns: ˈśēḥaz ‘frankincense’, ˈtaywi ‘meat; flesh’, ˈtayyah ‘billy goat’, ˈxlōwaḳ ‘dresses’; -- Verbs, perfect: (a)sˈsōfar ‘to travel’, ˈsatbaṭ ‘to hit o.s.’, ˈfaś ̣ś ̣aḥ ‘to be embarrassed’, ˈnūka ‘to come’, ˈṯībar ‘to break [intrs]’; -- Verbs, imperfect: yˈhaśkar ‘he gets lots of milk while milking [subj]’,ˈśxawwal ‘sit m.s.!’, yiˈġōrab ‘he knows’; -- Adjectives: ˈḥōwar ‘black m.s.’, ˈlēban ‘white m.pl.’,ˈḳāṣam ‘cold’; -- Adverbs: ˈrēḥaḳ ‘far’, ˈbawmah ‘here’, bīˈyawmah ‘here [dim]’, ṣaˈrōmah ‘now’, ṣarīˈyōmah ‘now [dim]’. d. In the absence of a final superheavy or non-final heavy syllable, stress the left-most Cv syllable of the stem, as in: -- yiˈgawar ‘he falls [ind]’, yiˈbawar ‘he goes at night [ind]’,9 ˈbihi ‘with them [dual]’, ˈšiki ‘with you/us [dual]’, ˈšihi ‘with them [dual]’, ˈhina ‘which’, ˈlahina ‘but’.10 3.2 Lexical stress Final -ah, -eh, -oh, -uh, -ɛʔ11 are stressed in the following lexeme types: in the perfect and subjunctive of final weak verbs, in adverbs ending in -oh or -ih, in a closed class of non-deictic nouns, in familial diminutives, and in dual verbal subject markers. In all but ḥruh, final /h/ is a non-root consonantal phoneme. Rubin (2010: 90–91, 2018: 48), Bendjaballah (2017) and Bendjaballah & Ségéral (2017), following a strictly quantity-based framework, give these endings as (underlyingly) -ū, -ō and -ē. Johnstone (1987) gives bō, with boh as an alternative, for the shortened allomorph of bawmah ‘here’. Simeone-Senelle (1997: 394, 2011: 1087) gives yeˈmoh alongside yeˈmō ‘today’ and boh ‘here’. Watson (2012: 85, 90, etc.) gives adverbs ending in -ōh, and the dual and final-weak suffixes as -ōh and -ah, with a note that -ōh (and the feminine dual suffix -tōh) is often realised as -oh (and the feminine dual suffix -toh) (e.g. Watson 2012: 23). Dufour (2016) generally gives the vowels as short with final 9 The indicative of these verbs has the form yigawr among some members of sub-tribes of the Bit Thuwar tribe (cf. Johnstone 1987; Rubin 2018: 207), and yigɛ̄ran among members of the Bit Yishōl and Bit Ḳamṣayt tribes in western Dhofar. 10 Incorrectly given in Watson (2012) as lahinna. 11 The short front vowel in the dual endings of the subjunctive of some verb forms is mid /e/ rather than the /a/ posited by Watson (2012: 90). The phonological re-analysis of these endings leads to an addition to the vocalic phoneme inventory of marginal /o/, /e/ and /ɛ/.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

156

Janet C.E. Watson and Abdullah Musallam al-Mahri

/h/, although this transcription alternates with a final long vowel. Rubin (2018: 47–48) describes these endings as taking a final long vowel, which in the case of -ō always has final -h, which is then shortened. He claims there appears to be a general rule whereby a long vowel is shortened before final -h. Bendjaballah & Ségéral (2017: 175, note 25) acknowledge the presence of -vh# and -vʔ# in dual forms and final-weak verbs, but dismiss them as pre-pausal variants resulting from devoicing, which obtains at the phonetic level and not at the phonological level. There are three reasons we contest the vowel-shortening claim and the claim that -vh# is simply a pausal variant of a final long vowel: Close phonetic and phonological work with Mehri consultants shows that the vowels are phonetically and phonologically short. Contra Bendjaballah & Ségéral (2017), our field work has shown that final -h is realised not only in pause but also in strings, as in: bduh wa-bkuh wa-bkuh ‘he lied and he cried and he cried’, īmoh bkuh ‘today he cried’. In the subjunctive of transitive final-weak verbs ending in -vh#, /h/ remains on suffixation: the verb twuh ‘to eat’ has the subjunctive form yitah. When this takes a pronoun suffix, /h/ remains, as in: yiḥōm yitahs ‘he wants to eat it f.’ (cf. also examples in Rubin 2018: 54). When a vowel-initial suffix is added, the vowel is similarly realised as short and /h/ is present, as in: yiḥōm yitahan ‘he wants to eat us’. The verb śīni ‘to see’ in the variety spoken by Bit Afāri in Wadi Habrūt towards the Yemeni border takes the subjunctive form yiśnah, which when suffixed gives yiḥōm yiśnahs ‘he wants to see it f.’. This contrasts with the subjunctive of śīni ‘to see’ in the Bit Thuwar variety, yiśnεʔ, which, when suffixed, lengthens the final vowel: yiḥōm yiśnɛ̄s ‘he wants to see it f.’ (cf. also Rubin 2018: 48), and the subjunctive of kūsa ‘to find’, yiksē, maintains a long vowel in pausal and word-final position which, when suffixed, is maintained: yiḥōm yiksēs ‘he wants to find it f.’. Finally, -vh# and -vʔ# contrast with final stressed long vowels in verbal and nominal/adverbial forms, as in: yimˈlē ‘he fills [subj]’, yikˈsē ‘he finds [subj]’ v. yibˈdah ‘he lies [subj]’, yiśˈnεʔ ‘he sees [subj]’; ḥabˈrē ‘son of’ v. ḥabˈrih ‘his son’, ḥāˈmē ‘mother’ v. ḥāˈmah ‘his mother’; imˈšeh12 ‘yesterday’ v. mūˈsē ‘rain’; bˈkuh ‘he cried’ v. ḥāˈbū ‘people’. Examples of stress in these cases are given below:

12 Contra Rubin (2018: 48), imšeh is realised with a short vowel in the final syllable.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

157

(2) Lexical Stress: -- Final weak verbs: bˈkuh ‘to cry’, bˈduh ‘to lie’, šiḳˈroh ‘he/they f. hid’, habˈnuh ‘he/they f. built’, yiˈtah ‘he eats [subj]’, yibˈkah ‘he cries [subj]’, yiśˈnɛʔ ‘he sees [subj]’; -- Adverbs: īˈmoh ‘today’, imˈšeh ‘yesterday’, ilˈloh ‘last night’, buh ~ boh ‘here’, ūˈṭoh ‘like that’, ūṭīˈyoh ‘like that [dim]’; -- Non-deictic nouns: ṯruh ‘two’, ḥruh ‘head’; -- Familial diminutives: ḥadˈduh ‘[paternal] uncle’, xalˈluh ‘[maternal] uncle’, ʔamˈmuh ‘grandfather’, ʔammaˈtuh ~ ʔamˈtuh ‘grandmother’; -- Dual verbal subject markers: ghēm ‘he went’ v. gihˈmoh ‘they m.du went’, yigˈhōm ‘he goes’ v. yighaˈmoh ‘they m.dual go’, yibˈdēd ‘he separates [subj]’ v. yibadˈdoh ‘they m.dual separate [subj]’, tˈhaśkar ‘she/you m.s. get lots of milk when milking [subj]’ v. thaśkaˈreh ‘they f./you [du] get lots of milk when milking [subj]’. 3.3 Stem-level suffixes Stem-level suffixes attract final or penultimate stress according to quantity and position, or shift (in some cases vacuously) stress predictably due to quantity and position. We list stem-level suffixes in (3), and provide examples of these suffixes affixed to stems in (4): (3) Stem-level suffixes: -- Derivational adjectival suffixes: -ay and -ōy; -- Future participle suffixes: -ōna, -īta, -ōni, -ayti, -ēta; -- Nominal number/gender markers, e.g. noun and adjectival feminine suffixes: -ūt, -īt, -ēt, -ōt; noun and adjectival plural suffixes: -ūtan, -awtan, -tan, -īn; -- Diminutive suffixes: -yēn, -yōt; -- Verbal 3 person suffixes: perfect feminine singular -ōt, -ūt, -ēt; -- The ending -ayyan in the dual endings in the indicative or conditional of several derived verbal stems (analysed correctly as such by Bendjaballah & Ségéral 2017: 177–179). Stem-level suffixes with long vowels attract stress onto the long vowel of the suffix, and final -tan shifts stress predictably to the heavy penultimate syllable (CvC). Examples of these stem-level suffixes affixed to stems are given below.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

158

Janet C.E. Watson and Abdullah Musallam al-Mahri

(4) Stem-level suffixes -- Derivational adjectival suffixes: ḥāˈwēl ‘first’ > ḥāwaˈlay ‘first m.s.’, ḥāwaˈlōy ‘first m.pl.’, ʔamḳ ‘middle’ > ʔamˈḳay ‘middling, medium m.s.’, ʔamˈḳōy ‘middling, medium m.pl.’, ˈmahrah ‘Mahrah’ > mahˈray ‘Mehri’; -- Future participle suffixes: sidˈdōna ‘will suffice m.s.’, nˈkayta ‘will come f.s.’, kasˈyawti ‘will find f.dual’, mhaxalˈfēta ‘will change s.th. f.s.’; -- Nouns: –– Noun and adjectival feminine suffixes: baḳˈrēt ‘cow’, bahˈlīt ‘word’, āfəˈrūt ‘red f.s.’, ūbəˈnīt ‘white f.s.’, raḥbɛ̄ˈnōt ‘small town’, šinḥaˈrūt ‘complaining’; –– Noun and adjective plural suffixes: ˈḥayd ‘hand’ > ḥāˈdūtan ‘hands’, ḥambaˈrawtan ‘boys’, ġigˈgīt ‘girl’ > ġagˈgūtan ‘girls’, ˈnōhar ‘eagle’ > nahˈrīn ‘eagles’, ḥayˈḏēn ‘ear’ > ḥayˈḏantan ‘ears’, rīˈḥōm ‘beautiful m.pl.’ > rīˈḥamtan ‘beautiful f.pl.’; -- Diminutive suffixes: wōz ‘female goat’ > wuzɛ̄ˈyōt ‘female goat [dim]’, digrīt ‘bean’ > digrɛ̄ˈnōt ‘bean [dim]’, ˈwōram ‘road; way’ > wurɛ̄ˈmōt ‘road; way [dim]’, frayś ̣ ‘female camel foal’ > farś ̣āˈyēn ‘female camel foals [dim]’; -- Verbal 3 person subject suffixes: –– Perfect fs suffixes: śiˈnūt ~ šinˈyūt ‘she saw’, nˈkōt ‘she came’, harˈḳāt ‘it f. was stolen’, amaḳṣaˈdēt ~ amaḳeṣˈdēt ‘she took a short cut’; –– Dual imperfect subject suffixes: yāˈzēman ‘he gives [subj]’ v. yūzaˈmayyan ‘they m.dual give [cond]’. Object and possessive pronouns are stressed in a very few cases: the penultimate or final syllable is assigned stress according to quantity and position on object suffixation to the 3ms/3fpl perfect verb, on suffixation to the prepositions hāl ‘at, with’ bərk ‘in, among’ bīn ~ mān ‘between’, fanwē ‘in front of’ sār ‘behind’ including the sub-minimal prepositions la- ‘to, against’, bi- ‘with, in’, ka- ‘with’, which takes the pre-pronominal suffix for š-, and the accusative prefix tē-; on suffixation to the nouns ḥabˈrē ‘son of’ and ḥāˈmē ‘mother of’; and on suffixation to the adverbials waḥˈśē ‘on own’, ṭādid- ‘each other’, and to ḥnōf ‘self’ before plural pronouns. The object/possessive pronouns take a different allomorph in suffixation to these lexemes (Rubin 2010: 36, 2018: 58; Watson 2012: 70). The suffixed allomorphs of the 3ms/3fpl verb stem and the prepositions hāl, bīn ~ mān, sār (glossed above) take a different prosodic pattern from the non-suffixed allomorphs:

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

159

(5) Object and Possessive Pronouns: -- Verbs: wzūm ‘he/they f. gave’ > wuzˈmīs ‘he/they f. gave her’, šxaˈbūr ‘he/they f. asked’ > šxabˈrūk ‘he asked you m.s.’, ġrūb ‘he knew’ > ġarˈbih ‘he knew him’; -- Prepositions: fanˈwē ḏa- ‘before’, fanˈwīs ‘before, in front of her’, fanˈwihi ‘before them [dual]’; sār ‘behind’ > sˈrūk ‘behind you m.s.’, bərk ‘in’ > bərˈkīham ‘among them m.’, mān ‘between’ > manˈwīn ‘between us’, hāl ‘at’ > hˈnay ‘with me’, ka- ‘with’ > ˈšūk ‘with you m.s.’, ˈšīkam ‘with you m.pl.’, la- ‘to’ > ˈlūk ‘to you m.s.’, ˈlīham ‘to them m.’, ˈlihi ‘to them [dual]’, tē- >ˈtēham ‘them m.’; -- Nouns: ḥabˈrayš ‘your f.s. son’, ḥāˈmēs ‘her mother’, ḥāˈmah ‘his mother’; -- Adverbials: waḥˈśay ‘on my own’, waḥˈśūk ‘on your m.s. own’, waḥˈśīham ‘on their m. own’, waḥˈśih ‘on his own’, ḥanˈfayham ‘yourselves m.pl.’, ḥanˈfayyan ‘ourselves’, ṭaytīˈdayki ‘each other f.dual’. With all but two of these pronoun suffixes, stress appears to be assigned according to quantity and position. Where the 3ms possessive suffix, -ih ~ -ah, is concerned, however, stress is assigned to this syllable irrespective of whether the word is disyllabic or monosyllabic or whether the allomorph is the less common long form -ēh, as in šēh (~ ših) ‘with him’, or the short form -ih ~ -ah (Watson 2012: 69), as in: hˈnih ‘with him’, ˈbih ‘with him, in him’, ˈlih ‘to him, against him’, ˈtah ‘him’ [direct object], fanˈwih ‘before him’, sˈrih ‘behind him’, wuzˈmih ‘he/they f. gave him’. We assume the pre-suffixal allomorph in the disyllabic cases takes the pattern (Cv)CˈCē (e.g. Rubin 2018: 57); this analysis is supported by the defined allomorphs of bər ‘son’ > ḥabrē ‘the son of’ and ḥāmē ‘the mother of’: ḥabrē ḏa-saʕīd ‘Sa’id’s son’, ḥāmē ḏa-ṭfōl ‘Tfol’s mother’, fnōhan ‘before’ > fanwē ḏa- ‘before s.th.’. On suffixation of the vowel-initial allomorphs of the object/ possessive pronouns, hiatus is resolved by left-most vowel deletion with association of the original stress to the preserved vowel: wuzˈmē-ih > wuzˈmih ‘he gave him’; fanˈwē-ayš > fanˈwayš ‘before you f.s.’, fanˈwē-ihi > fanˈwihi ‘before them [dual]’; ḥabˈrē-i > ḥabˈray ‘my son’. 3.4 Stress preservation and opacity In the case of suffixation of word-level suffixes, stress remains on the original tonic syllable, irrespective of quantity and position. When a singular noun takes a possessive suffix, stress appears to be assigned according to syllable

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

160

Janet C.E. Watson and Abdullah Musallam al-Mahri

quantity and position; however, since stress is maintained on the same syllable as on the unsuffixed noun, it is difficult to determine, from this evidence alone, whether this is a question of stress preservation or of stress assignment according to syllable quantity and position: (6) bark ‘knee’ > aˈbarki ‘my knee’; bayt ‘house’ > aˈbitk ‘your m.s. house’, aˈbayti ‘my house’; ˈśxōf ‘milk’ > aˈśxafs ‘her milk’; īˈbīt ‘camel f.’ > ḥayˈbitk ‘your m.s. camel’; ḳlīˈfūt ‘bark’ > aḳlīˈfats ‘its f. bark’; ʔāfəˈrēt ‘redness’ > aʔāfəˈrats ‘its f. redness’; raḥˈbēt ‘town’ > arḥaˈbēti ‘my town’; š ̣āˈrīr > aš ̣āˈrīrək ‘the back of your m.s. neck’; ˈwō.ram ‘road’ > ˈḥar.mək ‘your m.s. road’; ˈgawf ‘chest’ > aˈgawfk ‘your m.s. chest’; ˈṣayḥ ‘voice’ > aˈṣayḥk ‘your m.s. voice’.13 3.4.1 Object suffixes With the exception of the 3ms/3fpl perfect verb discussed above, stress is preserved on the stressed syllable of the unsuffixed stem in object suffixation. Even in case of the 3ms/3fpl perfect verb and the prepositions and adverbials examined above, assuming a pre-suffixal allomorph (Cv)CˈCē, stress here is also preserved on the tonic vowel of the suffixed stem. 3.4.1.1 Imperfect verb When the imperfect verb takes object suffixes directly, stress remains on the original tonic syllable of the stem. In most cases, stress also appears to adhere to syllable quantity and position: (7) aˈġōrab ‘I know’ > aˈġarbəs ‘I know her’, aˈwūzam ‘I give’ > aˈwizmək ‘I give you m.s.’, yiḳˈbūb ‘he roasts’ > yiḳˈbibham ‘he roasts them m.’, yharˈdūd ‘he takes back’ > yharˈdidham ‘he takes them m. back’.

13 In the case of CvvC, the long vowel is shortened before consonant-initial suffixes where noun-final C is aspirated or ‘breathed’ (first shown in Bendjaballah & S̍égéral 2014, who describe these consonants as ‘idle glottis’). With the exception of bayt > abitk ‘your m.s. house’, diphthongs are not shortened in this position, as in agawfk ‘your m.s. chest’. Johnstone (1987: xiv) assumes the diphthong /ay/ in bayt, which shortens to [i] on suffixation of C-initial possessive pronouns, has developed from bīt, as attested in Western varieties of Mehri.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

161

However, evidence in favour of stress preservation in object suffixation is found when we consider affixation of an object suffix to the singular imperfect inflections of verbs ending in vvC. Here we note that the long tonic vowel of the imperfect is subject to pre-suffix shortening (Watson 2012: 41), even where a vowel-initial suffix is affixed, but that stress remains on the originally long syllable, as in: (8) yiśˈlūl ‘he takes’ > yiśˈlili ‘he takes me’, tˈḥaym ‘you f.s. want’ > tˈḥiman ‘you f.s. want us’, yiˈśōm ‘he sells’ > yiˈśami ‘he sells me’, tˈkōb ‘you m.s. think’ > tˈkaban ‘you m.s. think us’.14 Three facts demonstrate that long-vowel shortening here is motivated by suffixation rather than by considerations of syllable repair: First, shortening occurs irrespective of whether the suffix is consonant-, or vowel-initial. Secondly, shortening fails to occur where the long vowel contains crucial morphological information (plural number): we see this in case of object suffixation to masculine plural forms of the perfect and imperfect verb:15 hadlīl ‘they m. guided’ > hadlīli ‘they m. guided me’, yikīb ‘they m. think’ > yikībham ‘they m. think them m.’ and yiśīmsan ‘they m. sell them f.’ compare with yhadlūl ‘he guides’ > yhadlilan ‘he guides us’, yikōb ‘he thinks’ > yikabi ‘he thinks me’, and yiśōm ‘he sells’ > yiśamsan ‘he sells them f.’. Thirdly, shortening is not restricted to suffixation of object pronouns to the imperfect verb: it also occurs where the perfect 3fs takes an object pronoun, and where possessive pronouns are suffixed to plural nouns and to some prepositions. In these cases, pre-suffix shortening leads to the opaque assignment of stress to a penultimate or antepenultimate light syllable (Cv). Below we examine suffixation of possessive pronouns to plural nouns, affixation of subject suffixes to perfect verbs, and affixation of object pronouns to perfect verbs. 3.4.1.2 Plural nouns with possessive pronouns When plural nouns such as ḥaˈbūn ‘the children’, bīˈrōk ‘knees’, ḥāˈdūtan ‘hands’ take a possessive pronoun suffix, the long vowel of the original stressed 14 Shortening does not affect underlying diphthongs, as we see in yiśaynkam ‘he sees you m.pl.’. 15 Rubin (2018: 61–62, 162) claims that the masculine plural verbal forms have underlying final -ə (< *u) ‘that affects syllable structure when suffixes are added’ (ibid: 61). We believe rather that lack of vowel shortening is originally due to allomorphy: masculine plural forms have the allomorphs /aw/ and /ī/. Since /aw/, as a diphthong, may not shorten in pre-suffix position, the other allomorph, /ī/, can also not shorten.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

162

Janet C.E. Watson and Abdullah Musallam al-Mahri

syllable is subject to pre-suffix shortening; and as for the imperfect verb, this syllable preserves stress, despite, in the case of a dual or plural possessive suffix, the following long vowel of the suffix,16 as in: ḥaˈbinīham ‘their m. children’, abīˈrakīham ‘their m. knees’, ḥāˈditīkan ‘your f.pl. hands’.17 Similarly the prepositions nˈxāli ‘under’ and ˈtwōli ‘towards’, which take the allomorphs of pronoun suffixes that typically attach to plural nouns (Rubin 2010: 36), are realised on suffixation with pre-suffix vowel shortening, as nˈxalīsan ‘under them f.’, nˈxalīham ‘under them m.’, and ˈtwalīham ‘towards them m.’, ˈtwalīsan ‘towards them f.’ respectively, with stress preserved on the original stressed syllable of the stem. Thus, both syllable shortening and stress appear to be unmotivated: shortening is not required to reduce a word-medial CvvC syllable to CvC, and stress is assigned to the antepenultimate light (Cv) syllable rather than to the penultimate heavy (Cvv) syllable. 3.4.1.3 Inflected perfect verbs Perfect verbs inflected for the marked first or second person take stress on the same syllable as the unmarked 3ms/3fpl inflection, i.e. as the perfect stem. In several cases, whether we are witnessing stress assignment according to weight or through preservation of stem accent cannot be determined, since stress on the original stem syllable is identical to stress on the right-most heavy syllable, as in: ˈktūb ‘he/they f. wrote’ > ˈktibk ‘I/you m.s. wrote’; (a)sˈsōfar ‘he/they f. travelled’ > (a)sˈsafrak ‘I/you m.s. travelled’. The behaviour of T-stem verbs of the structure CvTCvC and of quadriliteral verbs, however, demonstrates that stress is indeed preserved on the original stressed stem syllable, as in: ˈwatxaf ‘he/they f. came in the evening’ > ˈwatxafk ‘I/you m.s. came in the evening’, ˈwatxafkam ‘you m.pl. came in the evening’; ˈśatūḳ (< *ˈśatwaḳ ‘he/they f. longed’) > ˈśatūḳak ‘I/you m.s. longed’, ˈśatūḳkam ‘you m.pl. longed’; ˈġatyaṯ ̣ ‘he/they f. became angry’ > ˈġatīṯkan ‘you f.pl. became angry’ (cf. Watson ̣ 2012: 34; Rubin 2018: 50). Similarly, basic quadriliteral verbs such as aˈš ̣anʔaf ~ aˈš ̣anaf ‘to turn a corner’ and aˈdaġdaġ ‘to tickle’ preserve stress on the penultimate stem syllable even where consonant-initial inflectional suffixes are added, and thus where considerations of quantity and position would demand stress assignment to the penultimate syllable, as in the unattested *amaḳˈṣadkam. Examples include: aˈš ̣anʔafkam ~ aˈš ̣anafkam ‘you m.pl. turned a corner’ and aˈdaġdaġkam ‘you m.pl. tickled’. 16 Rubin (2018: 58) gives the plural possessive pronouns as -ihəm, etc., with short -i-. According to our acoustic and phonological analysis the vowel is long. 17 On suffixation of a possessive suffix to nouns ending in the -an of nunation, -an is deleted: ḥādūtan > ḥāditya ‘hands’ > ‘my hands’ (e.g. Rubin 2018: 93).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

163

3.4.1.4 Perfect verbs with object suffixes Only perfect verbs inflected for the unmarked singular and feminine plural 3 person take object pronouns directly in the varieties of Mehri spoken by tribes based in and around Dhofar.18 In these cases, the 3ms/3fpl exhibits a different template for the pre-suffixal stem, as discussed above (3.3); in verbs in which the 3mpl is expressed nonconcatenatively rather than through suffixation,19 the verb preserves the unsuffixed stem form and stressed syllable, and adds the suffix directly, with no long-vowel shortening, as we see for imperfect masculine plural verbs above. Examples include: śiˈnīw ‘they m. saw’ > śiˈnīwham ‘they m. saw them’, (a)xxˈṣayb ‘they m. sent’ > (a)xxˈṣaybham ‘they m. sent them m.’. The 3fs perfect verb invariably takes the object suffix directly, shortens the suffix vowel due to pre-suffix vowel shortening, as we have seen for the imperfect (3.4.1.1), and preserves stress on the shortened syllable of the 3fs subject suffix, as in: śiˈnūt ~ śinˈyūt ‘she saw’ > śiˈnitham ~ śinˈyitham ‘she saw them’, haˈmawt ‘she heard’ > haˈmitan ‘she heard us’, taˈbawt ‘she followed’ > taˈbati ‘she followed me’20 (compare ˈhimaki ‘we/ you [dual] heard’ and ˈtabaki ‘we/you [dual] followed’, which exhibit the same syllable structure).

4. Metrical Theory 4.1 The mora Following Hayes’ (1995) metrical model for word stress assignment, short vowels are assigned a single mora, long vowels are assigned two moras, geminate consonants are assigned a single mora, and non-geminate consonants in coda position are assigned a mora through a process known as Weight-by-Posi18 In contrast to Mahriyōt, the Mehri varieties spoken principally in far eastern Yemen (Watson 2012: 201), where all perfect verbs may take object pronouns directly. 19 E.g. syawr ‘they m. went’ as opposed to amaḳṣad-am ~ amaḳəṣd-am ‘they m. took a short cut’, where the 3mpl is expressed through the suffix -am. 20 Bendjaballah & Ségéral (2017: 183, note 34; also Rubin 2018: 61) argue that “[t]he case of the 1s -i# is particular. When it is suffixed to the pf 3fs, the preceding short stressed vowel is short: wəzməti she gave me, *wəzmūti. This can be understood only if the morpheme is underlyingly /y#/ and not /əy#/.” Rubin (2018: 163) further assumes an underlying 1pl pronoun suffix -n. Recognition of long-vowel shortening before word-level suffixes enables us to account for shortening without the positing of unmotivated underlying forms.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

164

Janet C.E. Watson and Abdullah Musallam al-Mahri

tion, whereby a consonant following a vowel receives a mora due to occupying the coda of a syllable (Hayes 1989: 90). Weight-by-Position is diagrammed below through a broken line. Metrical structure creation is non-exhaustive, which means that it need not exhaust the string of syllables in the word, and some syllables may be left unaccounted for by metrical structure creation (Hayes 1995: 2–3). μ μ μ μ μ

C V C V V C V C 4.2 The minimal word Mehri generally respects the bimoraicity principle for minimal words: in the unmarked case, stressable words have a minimum of two moras. Thus, sub-minimal non-suffixed prepositions and clitics such as min ‘from’, bi- ‘in, with’, ka- ‘with’, li- ‘to, against’ and wa- ‘and’ are not assigned stress. However, final stress is assigned to -vh in the lexical cases reviewed in 3.2 above; this suggests that degenerate syllables may be permitted in strong position. The rules for Mehri stress assignment are laid out below: Mehri stress assignment 1. Consonant Extrametricality:21 C > /__]word; 2. Foot Construction: Assign moraic trochees from left-to-right; degenerate feet are permitted in strong position; 3. Word Layer Construction: End Rule Right–stress the head of the rightmost foot.

21 Consonant extrametricality assures that the final consonant in a prosodic word is not assigned moraic status; thus final CvC counts as a light syllable (3.1 above).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

165

According to these principles, the adjective ūbənīt ‘white f.s.’ is stressed as follows: 1. The final consonant is rendered extrametrical; 2. Moraic trochaic feet are assigned from left to right (foot layer); 3. The head of the right-most trochaic foot is assigned stress according to End Rule Right (ERR). ( x ) word construction ERR ( x x ) foot layer σ σ σ μ μ μ μ μ ū b ə n ī 4.3 Secondary stress and stress fluctuation Mehri exhibits secondary stress, assigned rhythmically to the head of pre-tonic trochaic feet. Thus in ˌūbəˈnīt ‘white f.s.’, the final superheavy syllable receives word stress and the initial heavy syllable is assigned secondary stress; in ˌxadˈmēt ‘work’, one trochaic foot is constructed over the initial heavy syllable, and one over the final superheavy syllable; the final superheavy syllable receives word stress and the head of the initial foot, the initial syllable, is assigned secondary stress. Where a word comprises three feet, stress is assigned rhythmically,22 with secondary stress assigned to the initial foot, and primary stress to the right-most foot, as in: ˌraḥbɛ̄ˈnōt ‘little town’, ˌraḥbɛ̄ˈnawtan ‘little towns’ and ˌnāfīˈrūr ‘he became red’. In a metrical model, assignment of secondary stress means that foot layer of the metrical grid is not subject to conflation (Watson 2002: 121). In words comprising two heavy syllables, stress shift typically occurs before a word with initial stress to avoid stress clash: thus, bahlīt ‘word’ exhibits ultimate stress in ˈḏīma ˌbahˈlīt ‘this word’, but initial stress in aˈbahˌlīt ˈḏīmah ‘this word’. 22 Probably due to the constraint *Clash, which mitigates against two adjacent stressed syllables (Alber 2004).

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

166

Janet C.E. Watson and Abdullah Musallam al-Mahri

5 Stratal analysis To sum up the discussion so far, stress in Mehri is assigned to a syllable in the stem according to its quantity and position, or to a stressable stem-level suffix, including the cases of lexical stress discussed above (3.2). Where an unstressable suffix is attached, the long tonic vowel is reduced, unless that long vowel contains crucial morphological information–i.e. plural number information. A metrical model can account for stress assignment on stems and stem-level suffixes, but it cannot alone account for stress preservation. The apparent opacity in stress in Mehri can be explained within a model that makes reference to the levels within the phonology. The fact that suffixation frequently fails to shift stress demonstrates that stress is cyclical in Mehri–in order for stress to be preserved it has to have been assigned at a lower level in the first place (Collie 2007). The phonological cycle predicts that stress is inherited only from the immediately embedded word (Collie 2007: 167). Here we propose that the stratal approach to Optimality Theory, which draws on the cyclical observations of Lexical Phonology, can provide a neat account of stress assignment and stress opacity in Mehri. Optimality Theory approaches phonological processes through a set of universal violable constraints on the output rather than phonological rules. These constraints interact and are ranked relative to each other. Languages and dialects differ not in the nature of the constraints, but in their ranking. There are two types of constraints within Optimality Theory: markedness constraints which require the output to be structurally well-formed, and faithfulness constraints which require the output to be faithful to the input: thus, the prosthesis of a glottal stop to an onsetless syllable is motivated by a markedness constraint on onsetless syllables; and weak preservation of stress in the second syllable of English orìginálity is motivated by a faithfulness constraint preserving stress on the immediately embedded input oríginal. The goal of Stratal Optimality Theory (henceforth Stratal OT) as opposed to classical Optimality Theory is to acknowledge phonological levels and to reduce cyclicity to faithfulness of the output (in this case the embedding word) to the input (in this case the embedded word), and opacity to the masking of constraints between levels (Kiparsky 2000, 2003). Constraint ranking in Stratal OT differs from that in classical OT in that the constraints are re-ranked across morphological levels within a language. We follow Kiparsky’s (2000, 2003) and Bermúdez-Otero’s 2018) Stratal OT model in assuming two morphological levels: the stem and the word. Stem-level suffixes are affixed to the stem, and word-level suffixes to the word. In Mehri, as we have seen above in 3.3, stem-level suffixes are derivational

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

167

(adjectival) suffixes, number/gender subject suffixes for the unmarked third person, gender/number suffixes in nouns and adjectives, and singular possessive pronouns; word-level suffixes are the marked (first and second) subject pronouns in verbs, object pronouns, and plural possessive pronouns. The following faithfulness and markedness constraints account for the data in Mehri.23 Faithfulness constraints: StressIdent: An accent in the input must have an accent in the output (from Collie 2007) MorphIdent: Stem-internal morphological information in the input must be maintained in the output Foot-Head: Preserve the head of the foot Markedness constraints: Parse-σ: Parse syllables into feet FtBinμ: Parse syllables into binary moraic trochees Weight-StressPrinciple: Only heavy syllables are stressed; unstressed heavy syllables are reduced AllFtL: Align the foot with the left edge of the domain RightmostAlign: Align the strongest foot with the right edge of the domain *Adjoin: Do not adjoin structures to the prosodic word *HeavyAdjoin: Adjoined structures are not heavy The faithfulness constraints account for stress preservation and secondary stress at the word level: StressIdent requires preservation of lexical stress, and requires stress in the embedding word to be identical to stress in the embedded word; and Foot-Head requires foot heads to receive (primary or secondary) stress. MorphIdent prohibits vowel shortening where shortening would result in a loss of crucial morphological information. The markedness constraints account for assignment of stress at stem level according to quantity and position, and preservation of stem-level stress at the word level: Parse-σ requires all syllables to be parsed into feet; FtBinμ requires the creation of binary left-headed feet; AllFtL accounts for the direc23 As is convention in Optimality Theory, a constraint preceded by an asterisk means ‘does not have’: thus, *HeavyAdjoin below means Adjoined structures are NOT heavy. A constraint with no preceding asterisk means ‘does have’: thus, MorphIdent below means stem-internal morphological information in the input MUST BE maintained in the output.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

168

Janet C.E. Watson and Abdullah Musallam al-Mahri

tionality of footing, aligning the foot with the left edge; RightmostAlign requires assignment of stress to the head of the right-most foot; Weight-StressPrinciple (WSP) requires all and only heavy syllables to receive primary or secondary stress, and requires the shortening of unstressed long vowels; *Adjoin states that structures adjoined to the prosodic word are disfavoured; and *HeavyAdjoin requires long vowels to be shortened before an adjunct. We claim that stress opacity in Mehri results from restriction of footing and stress assignment to the stem level; syllables in word-level suffixes are not parsed into feet but rather adjoined to the prosodic word through Adjunction-to-Word, expressed here through the markedness constraint *Adjoin: Do not adjoin. This draws on van Oostendorp’s (2002) account of the invisibility to stress of certain suffixes in Dutch resulting from adjunction of the suffix to the prosodic word. Opaque shortening of stressed long vowels at the word level is motivated by a requirement for adjoined structures to be lighter than non-adjoined structures.24 In the tableaux below, an asterisk * signals that an output violates a particular constraint on one occasion. Where two or more * are given, this means that the output violates the constraint that number of times; the exclamation mark ! to the right of * signals the highest-ranking constraint disfavouring the loser relative to the winner (McCarthy 2008: 44). The → symbol highlights the favoured candidate. At stem level 1, RightmostAlign and StressIdent outrank WSP and FtBinμ. Stem level 1

RightmostAlign

StressIdent

WSP

FtBinμ

*

*

Input: šiḳróh →šiḳ.ˈroh ˈšiḳ.roh

*!

*

24 In Adjunction-to-Mora advocated initially by Broselow (1992), the long vowel that shares a mora with the following consonant in a non-final CVVC syllable has been analysed phonetically in Arabic and Mehri as shorter than an open long vowel.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

169

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

Stem level

RightmostAlign

StressIdent

FtBinμ

WSP

Input: śīwōṭ śī.ˈwōṭ

*!

→ˌśī.ˈwōṭ ˈśī.wōṭ

*!

*

At stem-level 2, on suffixation of stem-level suffixes, constraints undergo reranking. RightmostAlign and WSP outrank FtBinμ and StressIdent, assuring assignment of stress to the head of the right-most foot, shortening of the penultimate long vowel of the stem and secondary stress where appropriate. Stem level 2

RightmostAlign

WSP

FtBinμ

StressIdent

Input: ūbōn, -īt, Base: [ū.ˈbōn] ū.bō.ˈnīt

**!

→ˌū.bə.ˈnīt

* *

ū.ˈbə.nīt

*!

***

ū.ˈbō.nīt

*!

**

*

**

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

170

Janet C.E. Watson and Abdullah Musallam al-Mahri

Stem level 2

RightmostAlign

WSP

FtBinμ

StressIdent

Input: ḥāwēl, -ōy, Base: [ḥā.ˈwēl] ḥā.wē.ˈlōy

**!

→ˌḥā.wa.ˈlōy

* *

*

ḥā.ˈwa.lōy

*!

***

**

ḥā.ˈwē.lōy

*!

**

Stem level 2

RightmostAlign

WSP

FtBinμ

StressIdent

*!

*

*

Input: ḥād-ūtan, Base: [ˈḥayd-ūtan] ḥā.ˈdū.tan →ˌḥā.ˈdū.tan ˈḥā.dū.tan

* *!

*

*

At the word level, constraints undergo reranking (Kiparsky 2000, 2003). The faithfulness constraint IdentStress and the markedness constraint *HeavyAdjoin now outrank all other constraints ensuring stress preservation: by *HeavyAdjoin outranking WSP, the tonic syllable of the adjoined stem is shortened; outranking of *Adjoin permits adjunction; the footing and parsing constraints, AllFtL, Parse-σ and FtBinμ, are outranked due to footing not taking place at the word level. The interaction of these constraints is exemplified in the following tableaux of a suffixed 3 feminine singular perfect verb and a suffixed plural noun (distinguishing constraints only given):

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

Word level

IdentStress

*HeavyAdj

*Adjoin

WSP

FtBinμ

*

*

171

Input: /tabawt, -an/ Base: [ta.ˈbawt] ta.ˈbaw.tan

*!

→ta.ˈba.tan ˈta.ba.tan

*!

Word level

IdentStress

*

*HeavyAdj

*

*Adjoin

WSP

FtBinμ

*

**

*

*

*

Input: /ḥādūt, -īham/ Base: [ˌḥā.ˈdūt] ḥā.ˈdū.tī.ham

*!

→ˌḥā.ˈdi.tī.ham ˈḥā.di.tī.ham

*!

*

*

ḥā.di.ˈtī.ham

*!

*

*

There remains a question as to why the stressed syllable in internally inflected masculine plural verbs fails to shorten on suffixation of object pronouns (cf. above). Here masculine plural is expressed, depending on verb type, by the long high front vowel /ī/ ([ay] following an emphatic or pharyngeal) or by the diphthong /aw/ (Johnstone 1987: xx). Examples include: yiśtīm ‘they m. buy’, hātīm ‘they m. spent the night’, yidlawl ‘they m. know’, syawr ‘they m. went’, wzawm ‘they m. gave’. Pre-suffix shortening does not occur in case of diphthongs, as we know in the verbal system from forms such as yiśayn-kam ‘he sees you m.pl.’, and in the nominal system when possessive pronouns are attached to nouns ending in -aw/ayC, as in: agawf-k ‘your m.s. chest’, aṣayḥ-k ‘your m.s. voice’ (see 3.1); thus, we do not expect it in the suffixation of object pronouns to yidlawl or wzawm; however, pre-suffix shortening should occur when a verb ends in -īC. Originally lack of vowel shortening of internally inflected masculine plural verbs on object suffixation is due to allomorphy: since the diphthong allomorph

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

172

Janet C.E. Watson and Abdullah Musallam al-Mahri

/aw/ cannot shorten as shortening is restricted to long vowels, the long vowel allomorph can also not shorten. Today object suffixation to internally inflected masculine plural verbs fails to result in shortening because shortening would erase crucial morphological information. This gives, for example: tḥaym ‘you m.pl. want’ > tḥaymsan ‘you m.pl. want them f.’, thadlīl ‘you m.pl. guide’ > thadlīli ‘you m.pl. guide me’, yidlawl ‘they m. know’ > yidlawlsan ‘they m. know them f.’. We maintain here that the internal long vowel is maintained on suffixation of object pronouns through high-ranking MorphIdent. In the following tableaux, only distinguishing constraints are given: Word level

MorphIdent IdentStress

*HeavyAdj

WSP

FtBin

Input: /tḥaym, -an/ Base: [tˈḥaym] →tˈḥayman

*

tˈḥiman

*!

tḥiˈman

*!

*

* *

*

*

*

6 Conclusion To sum up: the complexity of word stress assignment in Mehri is attributed to a combination of lexical stress and the interaction of violable constraints that are re-rankable between stem and word levels. This model enables us to dispense with unmotivated underlying forms. At the stem level, stress in Mehri is assigned according to quantity and position or to a morpheme requiring lexical stress; thus at stem level 1, RightmostAlign and IdentStress outrank other constraints assuring primary stress to the right-most foot and lexical stress. At stem level 2, on the affixation of stem-level suffixes, RightmostAlign and WSP outrank all other constraints, ensuring primary stress to the right-most foot and secondary stress to pre-tonic heavy syllables. The relative high-ranking of FtBinμ assures assignment of secondary stress to the head of pre-tonic feet. At the word level, on the affixation of word-level suffixes, IdentStress outranks *HeavyAdjoin, WSP and FtBinμ, ensuring stress preservation. *HeavyAdjoin outranks WSP and FtBinμ to ensure shortening of stressed pre-adjunct syllables. Finally, MorphIdent outranks *HeavyAdjoin to mitigate against vowel

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

A Stratal OT Account of Word Stress in the Mehri of Bit Thuwar

173

shortening in case the potential target of shortening expresses crucial morphological information.

Bibliography Alber, Birgit. 2004. Clash, lapse and directionality. http://ruccs.rutgers.edu/~prince/papers/alber_clash_lapse.pdf - accessed 21st March 2018. Bendjaballah, Sabrina. 2017. “Gutturals and glides, and their effects on the Mehri verb.” In Linguistic Studies in the Arabian Gulf, edited by Simone Bettega and Fabio Gasparini, 13–32. Torino: Dipartimento di Lingue e Letterature straniere e Culture moderne – Università di Torino. Bendjaballah, Sabrina and Philippe Ségéral. 2014. “The phonology of ‘idle glottis’ consonants in the Mehri of Oman (Modern South Arabian).” Journal of Semitic Studies 59: 161–204. ———. 2017. “The vocalic system of the Mehri of Oman: Stress, length and syllable structure”. Brill’s Journal of Afroasiatic Languages and Linguistics 9: 160–190. Bermúdez-Otero, Ricardo. 2018. “Stratal Phonology”. In The Routledge Handbook of Phonological Theory, edited by S.J. Hannahs and Anna R. K. Bosch, 100–134. Abingdon: Routledge. Broselow, Ellen. 1992. “Parametric variation in Arabic dialect phonology”. In Perspectives in Arabic Linguistics IV, edited by Ellen Broselow, Mushira Eid and John McCarthy, 7–45. Amsterdam and Philadelphia: John Benjamins. Collie, Sarah. 2007. English stress preservation and Stratal Optimality Theory. PhD thesis: University of Edinburgh. Dufour, Julien. 2016. Recherches sur le verbe sudarabique moderne. Memoire inedit presente dans le cadre l’habilitation a diriger les recherches sous la direction de M. Gilles AUTHIER. Hayes, Bruce. 1989. “Compensatory lengthening in moraic phonology.” Linguistic Inquiry 20: 253–306. ———. 1995. Metrical Stress Theory: Principles and case studies. Chicago: University of Chicago Press. Johnstone, Thomas M. (compiled by G. Rex Smith) 1987. Mehri Lexicon and English– Mehri Word-List. Routledge: London. Kiparsky, Paul. 2000. “Opacity and cyclity.” The Linguistic Review 17: 351–367 ———. 2003. “Syllables and moras in Arabic.” In The Syllable in Optimality Theory,

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233

174

Janet C.E. Watson and Abdullah Musallam al-Mahri

edited by Caroline Fery and Ruben van de Vijver, 147–182. Cambridge: Cambridge University Press. Lonnet, Antoine and Marie-Claude Simeone-Senelle. 1997. “La phonologie des langues sudarabiques modernes.” In Phonologies of Asia and Africa, volume 1, edited by Alan S. Kaye and Peter T. Daniels, 337–371. Winona Lake: Eisenbrauns. Peperkamp. Sharon. 1999. “Prosodic words”. GLOT International 4.4: 15–16. Rubin, Aaron. 2010. The Mehri Language of Oman. Brill: Leiden. ———. 2018. Omani Mehri: A new grammar with texts. Brill: Leiden. Simeone-Senelle, Marie-Claude 1997. Modern South Arabian. In The Semitic Languages, edited by Robert Hetzron, 378–423. London: Routledge. ———. 2011. Modern South Arabian. In The Semitic Languages: An International Handbook, edited by Stefan Weninger, Geoffrey Khan, Michael Streck and Janet C.E. Watson, 1073–1113. Berlin and Boston: Mouton de Gruyter. Van Oostendorp, Marc. 2002. “The phonological and morphological status of the prosodic word adjunct”. In: The phonological and morphological status of the prosodic word adjunct. Linguistische Berichte Sonderheft 11: 209–236. Watson, Janet C.E. 2002. The Phonology and Morphology of Arabic. Oxford: Oxford University Press. ———. 2012. The Structure of Mehri. Harrassowitz: Wiesbaden.

© 2018, Otto Harrassowitz GmbH & Co. KG, Wiesbaden ISBN Print: 9783447111188 — ISBN E-Book: 9783447198233