Lateral fricatives and lateral emphatics in southern Saudi Arabia and Mehri

Proceedings of the seminar for arabian studies Volume 41 2011 Papers from the forty-fourth meeting of the Seminar for Arabian Studies held at the British Museum, London, 22 to 24 July 2010 seminar for arabian studies archaeoPress oxford Orders for copies of this volume of the Proceedings and of all back numbers should be sent to Archaeopress, Gordon House, 276 Banbury Road, Oxford OX2 7ED, UK. Tel/Fax +44-(0)1865-311914. e-mail [email protected] http://www.archaeopress.com For the availability of back issues see the Seminar’s web site: www.arabianseminar.org.uk Seminar for Arabian Studies c/o the Department of the Middle East, The British Museum London, WC1B 3DG, United Kingdom e-mail [email protected] The Steering Committee of the Seminar for Arabian Studies is currently made up of 13 members. The Editorial Committee of the Proceedings of the Seminar for Arabian Studies includes 6 additional members as follows: STEERING COMMITTEE Dr Robert Carter (Chair) Dr Mark Beech Dr Nadia Durrani Dr Robert G. Hoyland Dr Derek Kennet Mr Michael C.A. Macdonald Dr Ardle MacMahon (Secretary) Dr Venetia Porter Dr St John Simpson Mrs Janet C.M. Starkey (Editor) Mr Andrew Thompson (Treasurer) Professor Janet Watson Dr Lloyd Weeks EDITORIAL COMMITTEE: ADDITIONAL MEMBERS Professor Alessandra Avanzini Dr Ricardo Eichmann Professor Clive Holes Professor Khaleel Al-Muaikel Professor Dan Potts Professor Christian Robin Opinions expressed in papers published in the Proceedings are those of the authors and are not necessarily shared by the Editorial Committee. The Proceedings is produced in the Times Semitic New font, which was designed by Paul Bibire for the Seminar for Arabian Studies. © 2011 Archaeopress, Oxford, UK. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of the publisher. ISSN 0308-8421 ISBN 978-1-905739-40-0 Proceedings of the Seminar for Arabian Studies 41 (2011): 425–432 Lateral fricatives and lateral emphatics in southern Saudi Arabia and Mehri Janet C.e. Watson & Munira al-azraqi Summary Arabic was traditionally described as lughat al-Ρād ‘the language of Ρād’ due to the perceived unusualness of the sound. From Sībawayhi’s description, early Arabic Ρād was clearly a lateral or lateralized emphatic. Lateral fricatives are assumed to have formed part of the phoneme inventory of Proto-Semitic, and are attested in Modern South Arabian languages (MSAL) today. In Arabic, a lateral realization of Ρād continues to be attested in some recitations of the QurΜān. For Arabic, the lateral Ρād described by Sībawayhi was believed to be conined to dialects spoken in ДaΡramawt. Recent ieldwork by Asiri and al-Azraqi, however, has identiied lateral and lateralized emphatics in dialects of southern ΚAsīr and the Saudi Tihāmah. These sounds differ across the varieties, both in their phonation (voicing) and manner of articulation — sonorants and voiced and voiceless fricatives — in their degree of laterality, and in their phonological behaviour: the lateralized Ρād in the southern Yemeni dialect of GhaylΉabbān, for example, has a non-lateralized allophone in the environment of /r/ or /l/. Recent phonetic work conducted by Watson on the Modern South Arabian language, Mehri, shows a similar range of cross-dialect variety in the realization of the lateral(ized) emphatic. In this paper, we discuss different relexes of lateral(ized) emphatics in four dialects of the Saudi Tihāmah; we show that some of these dialects contrast cognates of *Ρ and *·; and we show that lateral emphatics attested in dialects of the Modern South Arabian language, Mehri, spoken in areas considerably to the south of the Saudi Tihāmah, show a similar degree of variation to that of the Arabic dialects of the Saudi Tihāmah. Keywords: Arabic, Mehri, Ρād, lateral fricatives, lateral emphatics Introduction In this paper, we begin by discussing the presence of lateral fricatives and lateral emphatics in Semitic. We then examine examples of lateral fricatives and the lateral(ized) emphatic in a dialect of the Modern South Arabian language, Mehri. In section two, we consider the early Arab grammarians descriptions of Ρād and of the ‘incorrect’ Ρād. In section three, we discuss the presence of the lateral Ρād in Arabic dialects and show that its geographical area is considerably larger than previously assumed. Looking at our data, we then show that the realizations of lateral emphatics in dialects of the Saudi Tihāmah vary considerably within a very small geographical area and that some dialects distinguish between cognates of *Ρ and *·; inally we examine crossdialectal realizations of the lateral emphatic in three dialects of Mehri. We conclude with suggestions for future research. Lateral fricatives in Semitic Semitic had three contrasting non-emphatic s-sounds, one of which, based on working with cognates, was most probably a lateral fricative (Steiner 1977). The Modern South Arabian languages are the only extant Semitic languages to maintain a three-way series of plain sibilants: a dental sibilant /s/, an alveopalatal sibilant /š/, and a lateral sibilant /ś/; this latter is similar in articulation to the Welsh double l in words such as Llanelli, llid and Lloyd, and represented in the International Phonetic Alphabet as ‘l’ with a belt: [τ]. Consider the following spectrograms of inal geminate /s/, /š/, and /ś/ in the dialect of Mehri spoken in Gabgabat, near Salālah in Dhofar (Zufār), Oman. The sound iles were all produced by a 45-year-old male speaker during a single recording session. The onset of the sibilants is to the immediate right of the inserted dotted line on the spectrograms; we can see that the /s/ in Ήiss has the highest pitch (shown by the high frequency of the start of the darkest part of the spectrogram corresponding to the consonant); and /š/ has the lowest pitch (shown as the relative low frequency of the start of the darkest part of the spectrogram corresponding to the consonant). The Modern South Arabian languages also have an emphatic counterpart to ś, cognate with Arabic *Ρ. This is realized as a pharyngealized lateralized fricative in some dialects, including the dialect of the male speaker from 426 Janet C.E. Watson & Munira Al-Azraqi Salālah, and as an emphatic lateral sonorant in others. The words in 1–4 are differentiated principally by the initial consonant: 1. sibb ‘to insult’ 2. šubb ‘be off [dog]’ 3. śibb ‘he climbed’ 4. ṣ́abb ‘it became warm [water]’ Arabic as lughat al-Ρād and descriptions from the early Arab grammarians figure 1. Spectogram of Ήiss ‘he felt’ figure 2. Spectogram of nišš ‘he fell precipitously’ Arabic Ρād is generally pronounced either as a voiced pharyngealized dental stop, or a voiced pharyngealized interdental fricative. The fact that Arabic was traditionally described as lughat al-Ρād, however, indicates that the grammarians considered this sound unique. There are several reasons for assuming that early Arabic Ρād was realized as a lateral or lateralized velarized or pharyngealized consonant: the Arab grammarians were very careful in their description of the articulation of consonants, and various descriptions of the articulation of Ρād show that the edge of the tongue was by the molars, giving a lateral articulation. Thus, Sībawayhi (1982) describes the correct articulation of Ρād in eighth-century Arabic as: ‘the beginning of the edge of the tongue and the adjoining molars is the place of articulation of the Ρād. Ρād is majhūr (unbreathed) and it is also rakhw (lax) and muΓbaq (lidded, emphatic)’ (trans. Dickins 1990: 237). Cantineau (1960) interprets this as describing a lateral or lateralized velarized voiced interdental fricative [·l]. Furthermore, ‘incorrect’ pronunciations of Ρād include alongside al-dāl al-mufakhkhamah (emphatic dāl) and al-zāy al-mufakhkhamah (emphatic zāy), al-lām almufakhkhamah (emphatic lām). This sound was said to have been used by people known as al-Zayālaρ (Ibn alJazari 1405 AH). Further evidence for the laterality of Ρād is in the pronunciation of Arabic loans involving Ρād in various languages. In Spanish, *Ρ is realized as [ld], as in: qāΡī ‘judge’ > alcalde, aΡ-Ρayρa ‘estate, hamlet’ > aldea; Akkadian ruldayu goes back to Arabic ruΡāy (Steiner 1977); and in some African languages *Ρ is realized as [l] in Arabic loans, as in Fulfulde: farilla < farΡ ‘obligation’, and Hausa: háyla < ΉayΡ (a) ‘menstruation’ (Versteegh 2006: 545). < Lateral Ρād within Arabic dialects figure 3. Spectogram of ρiśś ‘he got up’ For Arabic, the lateral or lateralized Ρād described by the early Arab grammarians was believed to be conined to dialects spoken in ДaΡramawt (Landberg 1901; El-Jindi 1983; Habtour 1988). As late as 2006, Versteegh writes: Lateral fricatives and lateral emphatics in southern Saudi Arabia and Mehri 427 figure 4. Saudi Tihāmah dialects investigated. ‘It is not surprising that this special sound disappeared in the New Arabic dialects, which all merged the relexes of Classical Arabic Ρ and ·’ and ‘The only alleged exception to the general merger is the dialect of DaΕīna, which according to Landberg (1901; 1905–1913) has /ά/ as relex of Classical Arabic /Ρ/ and /·/ as relex of /·/’ (Versteegh 2006: 544). The sound has, however, recently been discovered in parts of Asir and the Saudi Tihāmah (Al-Azraqi 2007; 2010; Asiri 2009). Al-Azraqi, a native of Abhā in ΚAsir, irst became aware of the lateral articulation of Ρād around her region when listening to her husband talking to a friend from the Saudi Tihāmah on the phone. She heard him pronouncing the cognate of Ρād as a distinctly lateralized fricative, and when she asked him about it was told that he was imitating the person he was speaking to. A short time later she arranged for recordings to be made in three villages in the Saudi Tihāmah. As a Saudi woman she was unable to travel and record men on her own — and women were reluctant to be recorded. She therefore engaged ive male ieldworkers from the area and supplied them with video cameras to record monologues and conversations. She asked them to focus where possible on words involving standard Arabic * Ρ and *·. Twenty hours of audio-visual data based on interviews and word lists were produced. The informants ranged in age from 20 to 80, and in educational levels from illiterate to university level. All informants, irrespective of educational level, exhibited lateral or lateralized relexes of * Ρ in their speech. AlAzraqi used Pinnacle Studio 11 to illustrate articulation of the sounds in video clips and presented her initial indings at the Association de Dialectologie Arabe conferences in Vienna (2006) and Colchester, Essex (2008). In April 2009, Al-Azraqi came over to the UK for a weekend meeting with Janet Watson and Samia Naïm. During this time, we transcribed one relatively short passage of speech individually and later compared our results. We then carried out initial impressionistic and acoustic phonetic analysis of *· cognates with the help of an instrumental phonetician from the University of Leeds, Barry Heselwood, and using the computer programme Praat for acoustic analysis (www.fon.hum.uva.nl/ 428 Janet C.E. Watson & Munira Al-Azraqi praat/). From Al-Azraqi’s initial research it was already known that dialects of the Saudi Tihāmah exhibited lateral relexes of Arabic *Ρ, however, until 2009 the degree of variation, the degree of laterality, and the interaction of voicing and laterality had not yet been determined. It was also unclear whether any of the dialects exhibited distinct relexes of *Ρ and *·, something held to be impossible by Versteegh (2006, see above) and Al-Wer (2004). Our initial work in 2009 showed three main realizations of the lateral(ized) Ρād in three different dialects, and two dialects in which the cognates of *Ρ and *· were distinct. At the same time, Yahya Asiri was completing his PhD thesis at the University of Salford on the dialect of Rijal AlmaΜ, another dialect in the Saudi Tihāmah, but some distance from the area in which Al-Azraqi’s ieldworkers collected their data. His work showed distinct lateralized relexes of *Ρ and *·, with the relex of *Ρ being a voiced lateralized pharyngealized interdental fricative, and the relex of *· a voiceless lateralized pharyngealized interdental fricative. The lateral emphatics we discovered in this area are: a pharyngealized lateral sonorant [τ], a voiced lateralized pharyngealized dental fricative, transcribed here as [·l], a voiceless lateralized pharyngealized dental fricative, transcribed here as [εl ], and a voiceless dental fricative with median friction, a sound which gives the impression of strong sibilance, tentatively transcribed here as [sl]. In most cases the lateral(ized) sounds correspond to *Ρ, but in some cases (see al-ΚIzzi below) they correspond to *·, and in other cases the cognate of *Ρ in some words is lateral and in others non-lateral (see al-RubūΚah older generation below). The dialect of al-Ghayil exhibited a pharyngealized lateral sonorant, which is most probably the sound described by Ibn al-Jazarī (1405 AH) as al-lām almufakhkhamah (emphatic lām). The following exchange took place between the interviewer and the language consultant: 5. Interviewer — antū ·īfāπn̥ ‘you [were] guests’ Consultant — Ήan łīfā̃(n̥) ‘we [were] guests’ ał-łuh 6. Interviewer — šūru ·-·uhur awlā m-xalf am-·uhur ‘before noon or after noon?’ Consultant — lā billāh / tġaddaynā baρd ał-łuh ‘no by God, we have lunch after noon’ In the dialect of al-ΚIzzi, the cognate of *· is lateral, while the cognate of *Ρ is not: here a voiceless lateral dental fricative with median friction contrasts with a voiced pharyngealized dental fricative cognate of *Ρ as in: antū 7. slāmī ‘thirsty’ (*·āmī) In al-RubūΚah, there appear to be two age-conditioned varieties. Among the older generation a pharyngealized lateral sonorant contrasts with a voiced pharyngealized dental fricative, as in: 8. fałil ‘blessing (*faΡil) 9. a·-·amān ‘social welfare’ (*al-Ρamān) Among the younger generation two lateralized pharyngealized consonants contrast: a voiced lateralized pharyngealized dental fricative relex of *Ρ with a voiceless lateralized pharyngealized dental fricative relex of *·, as in the following minimal pair provided by one of Al-Azraqi’s informants: 9. ·laym ‘pain’ (*Ρaym) 10. εlaym ‘has been affected’ These are very similar to the cognates of *Ρ and *· in Asiri’s dialect of Rijal AlmaΜ, for which he provides the following (in some cases) near-minimal pairs (Asiri 2009). Note that while ·l usually corresponds to *Ρ, as in: yi·lārib ‘he ights’ and ρa·l·lan ‘she bit’, and εl to *·, there is not always a clean correspondence: 11. ·larf ‘circumstance’ (*·arf) versus εlāmī ‘thirsty’ (*·āmī) 12. fā·l ‘it m. overilled’ (*fāΡ) versus māεl ‘he churned’ 13. tanū·l ‘she reminds’ versus tamūεl ‘she churns’ Of particular interest to us is the fact that three neighbouring dialects exhibit three main lateral or lateralized relexes of *Ρ within a very small geographical area, and that Rijal AlmaΜ, considerably to the north-west of Al-Azraqi’s area, also exhibits lateralized emphatics. Our future work involves discovering the geographical spread and density of dialects which exhibit lateral(ized) emphatics within the area, and the degree of phonetic variation. The lateral emphatic in dialects of Mehri We initially examined lateral fricatives in Mehri as part of the discussion of lateral fricatives in Semitic. Recent impressionistic and acoustic phonetic work, however, showed that dialects of Mehri exhibit a similar range of variation in realization of the emphatic lateral to the Arabic Saudi Tihāmah dialects. The Mehri dialects examined here are the dialect of Jōdab in Hawf in the easternmost province of Yemen, the dialect of Gabgabat Lateral fricatives and lateral emphatics in southern Saudi Arabia and Mehri 429 figure 5. Mehri dialects investigated. near Salālah in Dhofar, Oman, and the dialect of Ndēd on the border of Yemen in Dhofar. Male and female informants ranged in age from 20 to 45, and in education levels from basic education to university level. Audio data was gathered by Watson from Yemen in 2008 and from Dhofar between January and April 2010. The data included semi-structured interviews and elicited sentences and word lists. Data was recorded onto an Olympus LS10 recorder or directly onto a Sony laptop using Adobe Audition 1.5 software, and was saved as raw data in WAV format. Phonetic analysis was done using PRAAT for acoustic analysis (www.fon.hum.uva.nl/praat/). The dialects investigated showed three main variants: the Mahriyōt dialect in Hawf exhibits a voiced lateralized pharyngealized fricative, but with considerably less friction than the voiced lateralized pharyngealized fricative attested in Rijal AlmaΜ and in the speech of the younger generation of al-Rubūρah. This can be seen in the spectrogram of ·lōfar ‘plait’ in igure 6. Women speakers of this dialect tend to produce an affricate which is sometimes realized as voiceless, and strongly resembles the [τ] of Gabgabat. In the following three spectrograms, arrows indicate the mid-point of the articulation of the sound in question. figure 6. Spectogram of Ψlōfar ‘plait’ 430 Janet C.E. Watson & Munira Al-Azraqi figure 7. Spectogram of aτaΉ ūk min hāśan ‘what is the laughing for?’ figure 8. Spectogram of āτamkī ‘we/you [dual] used to’ In the Mehreyyet of Gabgabat, near Salālah, the lateral emphatic is realized as a voiceless lateral dental fricative, with slight affrication, as in: aτaΉūk min hāśan ‘what’s the laughing for?’: In the Mehreyyet of Ndēd, near Mazyūna on the border with Yemen, the lateral emphatic is realized mainly as a pharyngealized lateral sonorant, as in: āłamkī ‘we/you [dual] used to’: assumed, and initial indings from the Mahriyōt dialect of Hawf suggest that pronunciation differences may be linked to gender. Our future research intends to establish the geographical distribution of lateral(ized) emphatics in the Saudi Tihāmah, and the geographical distribution of dialects that exhibit phonemic distinctions between *Ρ and *·; to establish the extent to which the lateral(ized) emphatics in these dialects correlate with cognates of *Ρ; to establish through instrumental phonetic analysis the different phonetic correlates of lateral(ized) emphatics in these dialects; and to determine the extent to which these correlate with the phonetic correlates of the lateral(ized) emphatic in dialects of Mehri. Conclusion In this paper, we have presented the initial indings of a project for which we have considerably more questions than answers. What we have shown is that lateral and lateralized emphatics are far more prevalent in dialects of Peninsular Arabic than previously thought; that some dialects do (continue to) make ancient distinctions between *Ρ and *·; and that the degree of variation in the realization of lateral emphatics is not peculiar to the Arabic dialects but is also found in at least one Modern South Arabian language. For Mehri, research shows greater variation in the pronunciation of the emphatic lateral than previously Acknowledgements We are very grateful to Barry Heselwood, Yahya Asiri, Alex Bellem, and Samia Naïm for help and advice at various times, and to our funders, the Leverhulme Trust for a research fellowship for Watson (2007–2008), and the King Faisal Research Centre. Thanks also to James Dickins for comments on an earlier draft of the paper. References Asiri Y.M. 2009. Aspects of the phonology and morphology of Rijal Alma’ dialect (south-west Saudi Arabia). PhD thesis, University of Salford. [Unpublished]. Lateral fricatives and lateral emphatics in southern Saudi Arabia and Mehri 431 Al-Azraqi M. 2007. AΡ-Ρād in south-west Saudi Arabia as described by the old grammarians. Pages 43–50 in Proceedings of the 7th AIDA Conference, held in Vienna 5–9 September 2006. New Brunswick, NJ & London: Transaction Publishers 2010. The Ancient Ρad in Southwest Saudi Arabia, Arabica 57: 57–67. Cantineau J. 1960. Études de linguistique arabe (Mémorial Jean Cantineau). Paris: Klincksieck. Dickins J. 1990. Extended Axiomatic Functionalism: A contrastive assessment with application to aspects of Arabic. PhD thesis, Heriot-Watt University, Edinburgh. [Unpublished]. Habtour M. 1988. L’arabe parlé à Ġaylħabbān: Phonologie et morphologie. PhD thesis, University of the Sorbonne, Paris. [Unpublished]. El-Jindi A. 1983. Al-lahajāt al-Κarabiyyah fī al-turāth. Tripoli: al-Dār al-Κarabiyyah li-l-kitāb. Landberg C. de 1901. Études sur les dialectes de l’Arabie Méridionale. i. HaΡramaût. Leiden: E.J. Brill. 1905 –1913. Études sur les dialectes de l’Arabie Méridionale. ii. Dathina. Leiden: E.J. Brill. Sībawayhi Abu Bishr Ibn Qanbar/ed. ΚAbd al-Salām ΚUthmān ibn MuΉammad Hārūn. 1982. Kitāb Sībawayhi. iv. Cairo: Maktabat al-Khanji. Steiner R.C. 1977. The Case for Fricative-Laterals in Proto-Semitic. New Haven, CT: American Oriental Society. Versteegh K. 1984. Pidginisation and Creolisation: The case of Arabic. Amsterdam: Benjamins. 2006. Ḍād. Pages 544–545 in K. Versteegh et al. (eds), Encyclopedia of Arabic Language and Linguistics. i. Leiden: Brill. Al-Wer E. 2004. Variability reproduced: A variationist view of the [·]/[Ρ] opposition in modern Arabic dialects. Pages 21–32 in M. Haak, R. de Jong & K. Versteegh (eds), Approaches to Arabic Dialects: A collection of articles presented to Manfred Woidich on the occasion of his sixtieth birthday. Leiden: Brill. Authors’ addresses Munira Al-Azraqi, University of Dammam, Dammam, Saudi Arabia. e-mail [email protected] Janet C.E. Watson, University of Salford, School of Languages, Salford M5 4WT, UK. e-mail [email protected]