A grammar of Hewramî

A grammar of Hewramî Masoud Mohammadirad lan guage sc ie n ce pr ess Comprehensive Grammar Library 1 1 Comprehensive Grammar Library Editor: Martin Haspelmath In this series: 1. Jacques, Guillaume. A grammar of Japhug. 2. Grimm, Nadine. A grammar of Gyeli. 3. Maurer-Cecchini, Philippe. A grammar of T uatschin: A Sursilvan Romansh dialect. 4. V isser , Eline. A grammar of Kalamang. 5. Caballero, Gabriela. A grammar of Choguita Rarámuri: In collaboration with Luz Elena León Ramírez, Sebastián Fuentes Holguín, Bertha Fuentes Loya and other Choguita Rarámuri language experts. 6. Barlow , Russell. A grammar of Ulwa (Papua New Guinea). 7. T erhart, Lena. A grammar of Paunaka. 8. Pebley , Carol J. & Thomas E. Payne. A grammar of Kagayanen. 9. Rohleder , Jean. A grammar of V amale. 10. Kasak, Ryan. A grammar of Mandan. 1 1. Mohammadirad, Masoud. A grammar of Hewramî. ISSN (print): 2748-971X ISSN (electronic): 2749-7798 A grammar of Hewramî Masoud Mohammadirad lan guage sc ie n ce pr ess Masoud Mohammadirad. 2026. A grammar of Hewramî (Compr ehensive Grammar Library 11). Berlin: Language Science Press. This title can be downloaded at: http://langsci- press.org/catalog/book/517 © 2026, Masoud Mohammadirad Published under the Creative Commons Attribution 4. 0 Licence (CC BY 4.0): http://creativ e commons.org/licenses/by/4. 0/ ISBN: 978-3-96110-543-4 (Digital) 978-3-98554-163-8 (Hardcov er) ISSN (print): 2748-971X ISSN ( ele ctr onic): 2749-7798 DOI: 10.5281/zenodo.17140766 Source code available from ww w .github .com/langsci/517 Errata: paperhive.org/documents/remote?type=langsci&id=517 Cover and concept of design: Ulrike Harbort T ypesetting: Masoud Mohammadirad Proofreading: Carlos Fasola, Aidan Winberr y , Elliott Pearl, Gracious T emsen, Jeroen van de W eijer , Lachlan Mackenzie, Mary Ann W alter , Matthew K orte, Nicoletta Romeo, Patricia Cabredo, Ste ven Kay e Fonts: Libertinus, Arimo, DejaV u Sans Mono T ypesetting software: X Ǝ L A T E X Language Science Press Scharnweberstraße 10 10247 Berlin, Germany http://langsci- press.org [email protected] Storage and cataloguing done by F U Berlin Contents Ackno wledgments xiii Abbreviations xv 1 Introduction 1 1.1 Hewramî and its varieties . . . . . . . . . . . . . . . . . . . . . . 1 1.2 The place of Hewramî within Iranian diale ctology . . . . . . . . 4 1 . 3 H e w r a m a n T e k h t .......................... 7 1.4 The affiliation of T ekht Hewramî and its status within He wramî d i a l e c t o l o g y ............................. 8 1 . 5 E a r l i e r r e s e a r c h ........................... 1 0 1.6 Fieldwork b ehind this study . . . . . . . . . . . . . . . . . . . . 11 1 . 7 M a i n t e x t c o r p u s ........................... 1 3 1 . 8 F o l k t a l e c o r p u s ............................ 1 6 2 T ypological overview 19 2 . 1 P h o n o l o g y .............................. 1 9 2 . 2 M o r p h o l o g y ............................. 2 1 2 . 3 S y n t a x ................................ 2 7 2 . 3 . 1 W o r d o r d e r ......................... 2 8 2 . 3 . 2 A l i g n m e n t .......................... 2 9 2.3.3 Grammatical r elations . . . . . . . . . . . . . . . . . . . 31 2.3.3.1 Differential A indexing . . . . . . . . . . . . . 31 2.3.3.2 Differential P indexing . . . . . . . . . . . . . 33 2.3.3.3 Differential A flagging . . . . . . . . . . . . . 34 2.3.3.4 Differential P flagging . . . . . . . . . . . . . . 36 2.3.3.5 Differential case marking on non-core a r g u m e n t s .................... 3 8 3 P honetics, phonology , and morphophonology 41 3 . 1 P h o n e m e i n v e n t o r y ......................... 4 1 3.1.1 V owel phonemes . . . . . . . . . . . . . . . . . . . . . . 41 3.1.1.1 Description of vowels . . . . . . . . . . . . . . 41 Contents 3.1.1.2 Phonetic realisation of vow els . . . . . . . . . 43 3 . 1 . 1 . 2 . 1 / i / ................... 4 4 3 . 1 . 1 . 2 . 2 / e / ................... 4 5 3 . 1 . 1 . 2 . 3 / æ / ................... 4 5 3 . 1 . 1 . 2 . 4 / ɛ / ................... 4 7 3 . 1 . 1 . 2 . 5 / ɑ / ................... 4 9 3 . 1 . 1 . 2 . 6 / o / ................... 5 0 3 . 1 . 1 . 2 . 7 / u / ................... 5 1 3 . 1 . 1 . 2 . 8 / ʊ / ................... 5 2 3 . 1 . 1 . 2 . 9 / ɨ / ................... 5 2 3.1.2 Consonant phonemes . . . . . . . . . . . . . . . . . . . 53 3.1.2.1 Description of consonants . . . . . . . . . . . 56 3 . 1 . 2 . 1 . 1 / p / ................... 5 6 3 . 1 . 2 . 1 . 2 / b / ................... 5 6 3 . 1 . 2 . 1 . 3 / t / ................... 5 7 3 . 1 . 2 . 1 . 4 / d / ................... 5 7 3 . 1 . 2 . 1 . 5 / k / ................... 5 8 3 . 1 . 2 . 1 . 6 / g / ................... 5 8 3 . 1 . 2 . 1 . 7 / q / ................... 5 9 3 . 1 . 2 . 1 . 8 / ʔ / ................... 5 9 3 . 1 . 2 . 1 . 9 / t  ʃ / ................... 5 9 3.1.2.1.10 /dʒ/ . . . . . . . . . . . . . . . . . . 59 3 . 1 . 2 . 1 . 1 1 / m / ................... 6 0 3 . 1 . 2 . 1 . 1 2 / n / ................... 6 0 3 . 1 . 2 . 1 . 1 3 / ŋ / ................... 6 1 3 . 1 . 2 . 1 . 1 4 / f / ................... 6 1 3 . 1 . 2 . 1 . 1 5 / s / ................... 6 1 3 . 1 . 2 . 1 . 1 6 / z / ................... 6 2 3 . 1 . 2 . 1 . 1 7 / ʃ / ................... 6 2 3 . 1 . 2 . 1 . 1 8 / ʒ / ................... 6 2 3 . 1 . 2 . 1 . 1 9 / x / ................... 6 3 3 . 1 . 2 . 1 . 2 0 / ɣ / ................... 6 3 3 . 1 . 2 . 1 . 2 1 / ħ / ................... 6 3 3 . 1 . 2 . 1 . 2 2 / ʕ / ................... 6 4 3 . 1 . 2 . 1 . 2 3 / h / ................... 6 4 3 . 1 . 2 . 1 . 2 4 / ɾ / ................... 6 4 3 . 1 . 2 . 1 . 2 5 / l / ................... 6 6 3 . 1 . 2 . 1 . 2 6 / ɫ / ................... 6 6 3 . 1 . 2 . 1 . 2 7 / j / ................... 6 6 ii Contents 3 . 1 . 2 . 1 . 2 8 / w / ................... 6 7 3.1.3 Phoneme-grapheme associations . . . . . . . . . . . . . 67 3 . 2 P h o n o t a c t i c s ............................. 6 7 3.2.1 Phoneme distribution . . . . . . . . . . . . . . . . . . . 68 3.2.1.1 Phonotactics of Consonants . . . . . . . . . . 68 3.2.1.2 Phonotactics of vo wels . . . . . . . . . . . . . 70 3.2.2 Syllable structur e . . . . . . . . . . . . . . . . . . . . . . 71 3 . 3 S t r e s s p o s i t i o n ............................ 7 2 3 . 3 . 1 N o u n s ............................ 7 3 3 . 3 . 2 A d j e c t i v e s .......................... 7 4 3 . 3 . 3 A d v e r b s ........................... 7 5 3 . 3 . 4 V e r b s ............................. 7 5 3 . 3 . 5 C o p u l a ............................ 7 8 3.4 Major morphophonemic processes . . . . . . . . . . . . . . . . . 78 3.4.1 h-initial insertion . . . . . . . . . . . . . . . . . . . . . . 78 3.4.2 Inversion in the voicing of phar yngeals . . . . . . . . . 79 3.4.3 Glide insertion . . . . . . . . . . . . . . . . . . . . . . . 79 3.4.4 Anaptyctic vowel insertion . . . . . . . . . . . . . . . . 79 3 . 4 . 5 M e t a t h e s i s .......................... 8 0 3 . 4 . 6 V o w e l h i a t u s ........................ 8 1 3 . 4 . 7 A s s i m i l a t i o n ......................... 8 2 4 Nouns and nominal morphology 83 4 . 1 N o m i n a l i n fl e c t i o n .......................... 8 3 4 . 1 . 1 G e n d e r ............................ 8 7 4.1.1.1 Gender assignment . . . . . . . . . . . . . . . 87 4.1.1.2 Gender of loanwords . . . . . . . . . . . . . . 93 4.1.1.3 Functions of the base-final feminine morphemes 95 4.1.1.3.1 Expressing the female counterpart of masculine nouns . . . . . . . . . . 95 4.1.1.3.2 Deriving the fruit names . . . . . . . 95 4.1.1.4 Gender agr e ement . . . . . . . . . . . . . . . . 96 4 . 1 . 2 N u m b e r ......................... .. 9 8 4.1.2.1 Inherently plural nouns . . . . . . . . . . . . . 99 4.1.2.2 Associative plural . . . . . . . . . . . . . . . . 100 4 . 1 . 3 C a s e ........................ ..... 1 0 1 4.1.4 Classifiers and measure nouns . . . . . . . . . . . . . . . 110 4 . 1 . 5 B a r e n o u n s ......................... 1 1 3 4 . 1 . 6 D e fi n i t e n e s s ......................... 1 1 5 iii Contents 4 . 1 . 7 I n d e fi n i t e n e s s ........................ 1 1 7 4.2 Nominal w ord formation . . . . . . . . . . . . . . . . . . . . . . 119 4 . 2 . 1 D e r i v a t i o n .......................... 1 1 9 4 . 2 . 1 . 1 - î ......................... 1 2 0 4 . 2 . 1 . 2 - d a r ........................ 1 2 0 4.2.1.3 -eke/ -ekê . . . . . . . . . . . . . . . . . . . . 120 4 . 2 . 1 . 4 - î n e ........................ 1 2 1 4 . 2 . 1 . 5 - l e / - ɫ e ...................... 1 2 1 4.2.1.6 -ga/ -ge , -gê . . . . . . . . . . . . . . . . . . . 122 4 . 2 . 1 . 7 - g e r î ....................... 1 2 2 4 . 2 . 1 . 8 - w a n ....................... 1 2 2 4 . 2 . 1 . 9 - y a n e ....................... 1 2 3 4 . 2 . 1 . 1 0 - g e l ........................ 1 2 3 4.2.2 Nominal compounding . . . . . . . . . . . . . . . . . . . 123 4.2.2.1 N-N compounds . . . . . . . . . . . . . . . . . 124 4.2.2.2 N- V compounds . . . . . . . . . . . . . . . . . 124 4.2.2.3 N- ADJ compounds . . . . . . . . . . . . . . . 125 4.2.2.4 Echo compounds . . . . . . . . . . . . . . . . 125 5 Noun phrase 127 5.1 Noun phrase properties . . . . . . . . . . . . . . . . . . . . . . . 127 5.2 Modifiers in the noun phrase . . . . . . . . . . . . . . . . . . . . 128 5.3 Linkers in the noun phrase structure . . . . . . . . . . . . . . . 129 5 . 3 . 1 G e n i t i v e e z a f e ........................ 1 2 9 5.3.2 Attributive ezafe . . . . . . . . . . . . . . . . . . . . . . 132 5.3.3 Compound ezafe . . . . . . . . . . . . . . . . . . . . . . 134 5.4 Interaction of nominal case marking and ezafe marking . . . . . 135 5.5 W ord order in the noun phrase . . . . . . . . . . . . . . . . . . . 136 5.6 Agr eement in the noun phrase . . . . . . . . . . . . . . . . . . . 140 5.7 Syntax of definiteness . . . . . . . . . . . . . . . . . . . . . . . . 142 5.8 Syntax of indefiniteness . . . . . . . . . . . . . . . . . . . . . . . 146 5 . 9 Q u a n t i fi e r s.............................. 1 4 8 5 . 9 . 1 g i r đ ‘ a l l ’ ........................... 1 4 8 5.9.2 her ‘ each, every’ . . . . . . . . . . . . . . . . . . . . . . 148 5 . 9 . 3 ç i n n ê ‘ s o m e ’ ......................... 1 5 0 5 . 9 . 4 b e ʕ z ê ‘ s o m e ’ ......................... 1 5 1 5.9.5 kuçê ‘a bit, few’ . . . . . . . . . . . . . . . . . . . . . . . 151 5.9.6 ter ‘ other , another’ . . . . . . . . . . . . . . . . . . . . . 152 iv Contents 6 Pronouns and demonstratives 155 6.1 Personal pr onouns . . . . . . . . . . . . . . . . . . . . . . . . . . 155 6.1.1 Spee ch act pronouns . . . . . . . . . . . . . . . . . . . . 155 6.1.2 Third person pronouns . . . . . . . . . . . . . . . . . . . 159 6 . 2 C l i t i c p r o n o u n s ........................... 1 6 5 6 . 2 . 1 F u n c t i o n s .......................... 1 6 6 6.2.2 Morphosyntactic behaviour . . . . . . . . . . . . . . . . 171 6.3 Nominal demonstrativ e pronouns . . . . . . . . . . . . . . . . . 177 6.3.1 Independent demonstrative pronouns . . . . . . . . . . 177 6.3.1.1 Exophoric use . . . . . . . . . . . . . . . . . . 179 6.3.1.2 Anaphoric use . . . . . . . . . . . . . . . . . . 180 6.3.1.3 Discourse pr esentative use . . . . . . . . . . . 181 6.3.1.4 Empathetic use . . . . . . . . . . . . . . . . . 182 6.3.1.5 Predicative use . . . . . . . . . . . . . . . . . . 182 6.3.2 Demonstrative determiners . . . . . . . . . . . . . . . . 183 6.3.3 Local adverbial demonstratives . . . . . . . . . . . . . . 186 6.3.4 Manner adv erbial demonstratives . . . . . . . . . . . . . 187 6 . 4 R e fl e x i v e p r o n o u n s ......................... 1 8 9 6.5 Reciprocal pronouns . . . . . . . . . . . . . . . . . . . . . . . . . 192 6.6 Indefinite pr onouns . . . . . . . . . . . . . . . . . . . . . . . . . 192 6.7 Interrogative pronouns . . . . . . . . . . . . . . . . . . . . . . . 195 6 . 7 . 1 ç ê ş ‘ w h a t ? ’ ......................... 1 9 6 6 . 7 . 2 k ê ‘ w h o ? ’ .......................... 1 9 7 6 . 7 . 3 ç î ‘ w h y ? ’ ........................... 1 9 8 6 . 7 . 4 k o ‘ w h e r e ? ’ ......................... 1 9 9 6 . 7 . 5 ç e n î ‘ h o w ? ’ ......................... 2 0 0 6 . 7 . 6 k a m ‘ w h i c h ? ’ ........................ 2 0 0 6.7.7 çinne ‘ho w many , how much’ . . . . . . . . . . . . . . . 201 6 . 7 . 8 k e y ‘ w h e n ’ ......................... 2 0 1 7 Adjectives and numerals 203 7 . 1 A d j e c t i v e s .............................. 2 0 3 7.1.1 Adjectival infle ction . . . . . . . . . . . . . . . . . . . . 205 7.1.1.1 Attributive adje ctiv es . . . . . . . . . . . . . . 205 7.1.1.2 Predicative adjectives . . . . . . . . . . . . . . 206 7.1.1.3 Substantivised adjectives . . . . . . . . . . . . 208 7.1.1.4 Adjectives in the NP . . . . . . . . . . . . . . 209 7.1.2 Adjectival derivation . . . . . . . . . . . . . . . . . . . . 209 7 . 1 . 2 . 1 - î ......................... 2 0 9 v A cknowledgments This bo ok is the product of sev eral rounds of fieldwork in He wraman b etween 2016 and 2023. It would not hav e b een possible without the support of many individuals. It is my great pleasure to acknowledge their contributions here . I am deeply grateful to my informants and stor ytellers in Hewraman T ekht for welcoming me into their community and offering me unique access to their lan- guage and culture . I owe special thanks to my primar y informant, Amir Qasemi, who patiently answered my many questions about his native language and kindly introduced me to some memb ers of the local community . My friends K amran Za- mani and Heyder Fatemi provided valuable assistance with the logistical aspects of the fieldwork. I was warmly r e ceiv ed during each visit to Hewraman, and I dedicate this b ook to its people in recognition of their ongoing effort to preser ve their language in a rapidly changing world. The idea of writing a grammar of Hewramî emerged during my collaboration with Geoffrey Khan on our joint study , Language contact in Sanandaj: A study of the impact of Iranian on Neo- A ramaic . Through that collaboration, I came to appreciate the importance of documenting Hewramî as an endangered language. Geoffrey’s support and encouragement have been invaluable over the years. I am especially indebte d to my friend and colleague Shuan Osman K arim, who went above and bey ond in assisting with the te chnical aspects of the grammar , including typesetting in L A T E X and developing Python scripts to export the text corpus from FLEx into the requir e d L A T E X template for linguistic examples, as well as to extract glossaries fr om the text corpus. I am grateful to Martin Haspelmath for accepting this manuscript into the se- ries ‘Comprehensive Grammar Librar y’ , for arranging the review pr ocess with multiple revie wers, and for sharing his own constructive comments on se veral occasions. I am equally grateful to anonymous revie wers whose insightful feed- back on earlier versions of the book significantly improved the w ork. My thanks extend to all the individuals who contributed to proofreading and offer ed sug- gestions on the revised version. I thank Sebastian Nor dhoff for his consistent assistance and professionalism throughout the publication process. Finally , I extend my heartfelt thanks to my wife , Slava, for her unflagging support throughout the years. Ackno wledgments The resear ch on the b ook from September 2021 onwards was made possible by funding from the European Resear ch Council (ERC, ALHOME, 101021183), for which I am grateful. xiv Abbre viations General abbreviations Ar . Arabic A v . A vestan CK. Central Kurdish H. Hewramî NK. Northern Kurdish P . Persian SK. Southern Kurdish T r . T urkish Grammatical abbreviations 1 first person 2 second person 3 third person A agent-like argument of a transitive clause add additive adj adjective adjzr adjectivizer adp adposition adv adv erbial a ug augment a ux auxiliar y clf classifier cmpr comparative compl completive conj conjunctive cop copula def definite deic deictic dem demonstrative dir direct disc.pt cl discourse particle dist distal emph emphatic ep epenthesis exist existential particle ez ezafe linker ez.a ttr attributive ezafe ez.cmpd ezafe compound ez.gen genitive ezafe f feminine hort hortative imp imperative ind indicative indf indefinite inf infinitive intj interjection intr intransitive loc locational m masculine NC non-canonical subject n.f feminine noun n.m masculine noun neg negative nmlz nominalizer O direct object of a transitive verb obl oblique Abbreviations ord ordinal P patient-like argument of a transitive clause pl plural pn proper noun post postposition povb post-verb pro pronoun proh prohibitive pro x proximal prs present prsv presentative PSR possessor pst past pt cl particle pt cp participle pvb prev erb q.pt cl question particle R recipient-like argument in ditransitive clauses R non-core arguments of intransitive and transitive clauses recp reciprocal base refl reflexiv e S single argument of an intransitive verb sbjv subjunctive sbrd subordinate particle sg singular supr superlative T theme-like argument in ditransitive clauses t am tense-aspe ct-mood tr transitive v oc vocative Codes of example sources Main text corpus (Mohammadirad 2025c) BP Oral history DG Local anecdote/myth DP Oral history HB Lo cal anecdote/myth JE Process narrative JH Folktale JM A utobiography JP Oral history KŞ Folktale PM Local anecdote/myth RE Process narrative ŞC Lo cal anecdote, recent history ZB Local anecdote/myth ZP Oral histor y ZQ Local anecdote/myth xvi Codes of examples from the folktale corpus (Mohammadirad in prep[b]) BB Folktale BM Folktale ÇH Folktale ÇK Folktale DB Folktale ED Folktale HJ Folktale HM Local anecdote HS Folktale H W Folktale JC Folktale JF Folktale JL Folktale KK Folktale KT Folktale ME Folktale MF Folktale MM Folktale MP Folktale MR Folktale PK Folktale PP Folktale P W Folktale SH Folktale SK Folktale ŞE Folktale ŞŞ Folktale WL Folktale XŞ Lo cal anecdote XX Folktale YX Lo cal anecdote xvii 1 Introduction The mountainous Zagros regions of northwestern Iran and northern Iraq are characterised by rich linguistic diversity . V arieties of Kurdish ar e sp oken acr oss much of the area, along with remaining varieties of Gorani and Neo- Aramaic. Over the last two decades, there has been a growing interest in documenting peripheral Gorani varieties, leading to the publication of a few sketch grammars of these endangered languages (see Mahmoudve ysi et al.’s 2012 description of the Gewrecû variety , Mahmoudveysi & Bailey’s 2013 description of the Zer de variety ). Despite these welcome documentation efforts, the more conser vative Hewramî varieties of Gorani have lacked a decent monograph-length descrip- tion since MacK enzie’s (1966) seminal sketch grammar of the Luhon variety of Hewramî. This bo ok aims to fill this gap by pro viding a comprehensive grammar of the T ekht variety of Hewramî, one of the most conser vative Hewramî varieties spo- ken in the high mountainous region of Hewraman T ekht straddling Iranian and Iraqi Kurdistan. The grammar is accompanied by a Hewramî-English glossary , an English-Hewramî glossary , and a verb list. This chapter is structured as follows: §1.1 pro vides a general description of Hewramî and its varieties. §1.2 discusses the place of Hewramî within Iranian lan- guages. Then, I move on to giv e a brief description of Hewraman T ekht (§1.3), fol- lowed by the affiliation of the T ekht variety within Hewramî dialectology (§1.4). In §1.5, I give an ov er vie w of the literature on Gorani varieties. §1.6 summarises the fieldwork behind producing this grammar . The chapter ends with informa- tion about the main corpus of narrative texts (§1.7) behind this study , followed by an additional corpus of folktales used to back up the description of morphosyn- tactic features (§1.8). 1.1 Hewramî and its varieties Hewramî is an Iranian language spoken in the remote mountainous r egion at the heart of the Kurdish-speaking region along the western border of Iran and neigh- bouring areas in Iraqi Kurdistan. Hewramî is a name used by Hewramî people 1 Introduction to refer to themselves and their language . Published references to the language appear in the following forms: A wromānie (Christensen & Benedictsen 1921); A uramânî (Mann & Hadank 1930); Hawrāmī (MacK enzie 1966, Mahmoudveysi & Bailey 2018); Hawrami (Holmberg & Odden 2008, Haig 2008, Stilo 2019). The ex- onyms maço ‘he/she says’ and maço zuwan ‘maço language ’ are sometimes used by neighbouring Kurdish-speaking pe ople to refer to the language. In addition, in some linguistic studies, e.g., Khan & Mohammadirad (2024a, 2024b), the cov er term Gorani has be en used interchangeably with He wramî. MacK enzie (1987) classifies the Hewraman region into four main divisions: Luhon (in the south), T ekht (in the centre), Dizlî (in the north), and Razaw (ar ound Sar v Abad). The last one could include the Jawer o and Gawero sub- regions (Mahmoudv eysi & Bailey 2018). 1 Hewramî varieties ar e traditionally di- vided into three major groupings: T ekht, Luhon, and Jawer o. Geographically speaking, these varieties are spoken in the centre, south, and east of the gr eater Hewraman region, r espe ctiv ely . The T ekht region is linked to Jawero thr ough a stretch of valleys, while the Luhon r egion is locate d in the western valle y . Within Iran, the most significant concentration of Hewramî speakers is in the cities of Mariwan, Pawa, and Sarv Abad, though Kurdish is also spoken in these three cities. In Iraqi Kurdistan, He wramî sp eakers can be found in cities such as Khur- mal and Halabja. Figure 1.1 shows the r egion of Hewraman according to the men- tioned divisions. 2 The list of localities was partly up dated based on a recent atlas of language distribution in K ordestan province ( see Anonby et al. 2019). As can 1 It has yet to be determined whether these geographical divisions match linguistic sub divisions, especially in the Dizlî and Razaw regions. 2 The villages in each part are as follows. The local name has appeared in the Latin Kurdish al- phabet for each locality , follow ed by the official name in the transcription common in Iranian philology . The localities in Iraqi Kurdistan have been transcrib ed only in the Kurdish alpha- bet. Dizlî region : (1) Dizɫî (Dezli); (2) Qeɫacê (Qal’ eh Ji); (3) Baramawa (Bahrām Ābād); (4) Tifɫî (T efli); (5) T azawa (Tāzeh Ābād); ( 6) Bindoɫ (Bendowl); (7) Deymeyo (Demay o); (8) Derokî (Daraki); (9) Gorgeyî (Ebrāhim Ābād); (10) Zelke (Zalkeh); (11) Dere (Darr eh); (12) Zeɫm; (13) Ehmewawa; (14) Hanew Qulî. T ekht region : (1) Hewraman T ekht (Owrāmān T akht); (2) W eysîyan (V eysiān); (3) Biɫbeř (Bel- bar); (4) Serû Pîrî (Sar Pir); (5) Jîwar (Zhivār ); (6) Silên (Selin); (7) Espeřêz ( Aspeh Riz); (8) Del (Dal); (9) Nwên (Nevin); (10) K elcî (Kalji); (11) Naw (Nav ); (12) Daɫemerz (Daleh marz); (13) Zom (Zom); (14) K emaɫe (Kamāleh); (15) Bennen (Bannan); (16) W ergewîyare (V argeh Vir); (17) Rûwar (Ruvār); (18) He wasawa (Abbās Ābād). Luhon region : (1) Nođşe (Nodsheh); (2) Newsû (Nowsud); (3) Pawe (Pāv eh); (4) Keymne (K eym- neh); (5) Zawer (Zāvar ); (6) Hane Geřmeɫe (Hāneh Garmaleh); (7) Hecîc (Hajij); (8) Ner we (Narveh); (9) Şoşme (Shushmeh); (10) Şêxan (Sheykhān); (11) Xaneqa (Khaneqā); (12) Deremûr (Darreh Mur ); (13) Giraɫ (Gerāl); (14) Neysane (Naysāneh); (15) Sosekan; (16) T ewêɫe; (17) Bîyare; (18) Baxe K on; (19) Serget; (20) Dere Mar; (21) Gulp; (22) Dega Şêxan. 2 1.1 He wramî and its varieties be se en, most He wramî-sp eaking localities are mainly located on the Iranian side of the border . Figure 1.1: The region of He wraman and its divisions There ar e no precise figures on the number of Hewramî speakers. MacKen- zie (1987) puts the number of sp eakers at ar ound 10,000. More recently , Mo- Razaw region : (1) Řazaw (Raz Āb); (2) K eřawa (Karr Ābād); (3) Degaga (Degāgāh); (4) Dêw ez- naw (Divaz Nāv); (5) Coɫandê ( Jowlān Deh); (6) Naw e (Nāveh); (7) Xaneqay Řazaw (Khāneqāhe Raz Āb); (8) Dořû (Dorud); (9) Mehmûawa (Mahmud Ābād). Jawero region : (1) Serûmaɫ (Sar o Māl); (2) Dêw er (Divar); (3) Çemşîyer ( Chashmidar); (4) Nese- nar (Nasanār); (5) Hersîn (Harsin); ( 6) X waşt (Kho vāsht); (7) Sipîbin (Sefid Ben); (8) Nîce (Nijeh); (9) Saɫîyan (Sāliān); (10) Ewêheng ( A vihang); (11) Sê Pîran (Seh Pirān); (12) Hoye (Howyeh); (13) Ser Hoye (Sar Ho w yeh); (14) Bêsaran (Bisārān); (15) Jan (Zhān); (16) Ar yan ( Āryān); (17) Sûretifî (Sureh tefi); (18) Jinên (Zhenin); (19) Paygelan (Pāygalān); (20) Tifên (T efin); (21) Paɫin- gan (Palangān); (22) Gowaz (Go vāz); (23) Dejin (Dazhen); (24) Ser Řêz (Sar Riz); (25) Borîyer (Boridar); (26) Jirîje (Zherizheh); (27) Mazîbin (Māzi Ben); (28) T engîwer (T angi V ar); (29) Kanî Hoseyn Beg (K ani Hossein Bag); (30) Kêɫane (Kilāneh). Gawero region : (1) Heşemêz (Hashemiz); (2) Gelên ( Galin); (3) Xaneqay Gelên (Khāneqāhe Galin); (4) W esê Jûro (V asi-ye Olyā); (5) W esê X waro (V asi-ye Soflā); (6) Doɫaw (Dowlāb); (7) Bizɫane (Bezlāneh); (8) Heɫwan (Halvān); (9) T exte (T akhteh); (10) Suw er yan (Sovāriān); (11) T a (Tāy ); (12) Farsawa (Fāres Ābād); (13) Nîyer (Nier ); (14) Şîyan (Shiān); (15) Derwêşan (Darvishān); (16) Dêr Moɫî (Dir Mowli). 3 1 Introduction (5) T ekht H. Luhon H. prs pst prs pst wen- wena- wan- wana- ‘read’ piseř- piseřa- seř- seřa- ‘wipe ’ cen- cena- incen- incena- ‘mince ’ biřfan- biřfana- řfan- řfana- ‘ab duct’ Both varieties employ pre verbal T AM prefixes (i.e., indicativ e and subjunctive prefixes) to a limited extent. Phonological factors condition the occurrence of these prefixes, e .g., b efore v owel-initial v erbs in b oth varieties. How ever , Luhon H. tends to use these prefixes with more v erbs; see (6). For example, w-initial verbs tend to take the indicative prefix in Luhon, but not in T ekht (See Moham- madirad & Karim 2025, for e xplanation). (6) T ekht H. Luhon H. wan-o mi-wan-o ‘he/she reads’ wer-o (mi)-wer-o ‘he/she eats’ taw-o mi-taw-o ‘he/she can’ zan-o mi-zan-o ‘he/she kno ws’ řem-o mi-řem-o ‘he/she runs’ 1.5 Earlier research Gorani languages, among them the Hewramî varieties especially , have long in- trigued philologists and linguists alike. A first major – and often unnoticed – monograph-length study on a variety of Hewramî was a descriptiv e grammar of T ekht Hewramî and the Pawe dialect (spoken in the northeast of Kermanshah in the city of Pawe) by Christensen & Benedictsen (1921). This grammar resulted from a field trip to Sanandaj and Hewraman by Å ge Meyer Benedictsen in 1901. The Hewramî description comes from a collection of seven te xts, four of which were collecte d fr om a young speaker of Ruwar dialect (P. Rūbār ), whom Bene- dictsen met in Sanandaj. The remaining three texts w ere collected in Hewraman, in the village of Nawe Sûte, which I have been unable to locate on the map. The Pawe material features one text and four poems. A second major work on Gorani varieties was carried out by Mann & Hadank (1930). The bo ok contains chapter-long descriptions of eight Gorani varieties, including Kanduali, A uramani, Bajalani, Biwaniji, Gahwarai, Rijabi, Say yidi, and Zardai. Among these , only Kandulai has been describ ed in detail. Little grammat- ical description is offered for the rest of the varieties, and the r espe ctive sections 10 1.6 Fieldwork behind this study consist mainly of lexicon and a few texts. The A uramani sketch deals with the linguistic analysis of proper Hewramî, though, as MacK enzie (1966) notes, it is not evident which dialect of Hewramî is investigated here . The best-known grammar of Hewramî is MacK enzie’s (1966) description of the Luhon dialect of Hewramî. The monograph is based on the spe ech of a single male Kurdish-Hewramî bilingual whom MacKenzie met in London. The gram- mar is detailed and remains the only reliable description of a variety of Hewramî ever since . Nonetheless, it is brief and economical. Inde ed, the bo ok mainly co v- ers morphology , perhaps in line with the tradition of grammar writing at the time. Less co verage has be en giv en to phonology and esp ecially syntax. In recent years, sev eral scholars have dev oted themselves to describing the most endangered varieties of Gorani, situated outside Hewraman and considered peripheral Gorani varieties. This has resulted in the publication of two sketch grammars of the peripheral Gorani varieties of Gewrecû (Mahmoudveysi et al. 2012) and Zerde (Mahmoudve ysi & Bailey 2013), spoken in western Iran. Mah- moudveysi & Bailey (2018) offer a short grammatical description of the Hewramî dialect of Pawe. The authors have now embarked on preparing a grammatical description of the Shabaki dialect of Gorani spoken in the Mosul P lains. Despite the recent rising interest in Gorani varieties, MacK enzie’s (1966) de- scriptive grammar is the only standar d description of a Hewramî variety to date. Other varieties of Hewramî remain largely under-inv estigated. This b ook aims to provide a detaile d grammatical description of the T ekht variety of Hewramî, which, as seen in §1.4, differs in many respe cts fr om the more southern Luhon variety . The grammar is based on field recordings I collected over the course of several field trips to the Hewraman r egion over the past sev en years. It provides a detailed account of the phonetics, phonology , morphology , syntax and lexicon of Hewramî, grounded in current linguistic methods. Despite what is often as- sumed, there is a high degree of dialectal variation within Hewramî varieties. Indeed, it is unclear whether the vernaculars of Dizlî and Şamyan belong to any known three varieties of Hewramî. Like wise, the morphosyntactic features of the Jawaro variety remain largely unknown to scholars w orking on Hewramî dialectology . It is hope d that the current monograph will encourage scholars to produce comprehensiv e descriptions of the remaining Hewramî varieties. 1.6 Fieldwork behind this study The material for this bo ok was mainly gather ed during various rounds of field- work conducted in Hewraman T ekht b etw een March 2016 and A ugust 2023. I 11 1 Introduction visited the region for the first time in March 2016 and conducted a pilot linguis- tic fieldwork. On this first field trip , my goal was mainly to get familiar with the spee ch community and get an idea of patterns of language use on a daily basis. I recorded spontaneous sp eech and dialogues, and conducte d a few elici- tation tasks, primarily focusing on verb conjugation. A recording I collected on this trip appears as a glossed text in Khan & Mohammadirad’s (2024a: 557 –564) study of language contact in Sanandaj. The second field trip took place in June and July 2017 . This trip was con- ducted as part of my P hD dissertation on pronominal clitics in W estern Iranian languages. During this trip, I conducted elicitation tasks using visual stimuli and a questionnaire with nativ e speakers from Hewraman T ekht. In addition, I recorded some spontaneous sp oken data and recor de d a narration of the Pear story . The elicite d and spontaneous data collecte d in this trip formed the basis for the study of argument indexing and syntax of pr onominal clitics in T ekht Hewramî ( see Mohammadirad 2020b: 365–372), within the context of W estern Iranian languages. The material from these two trips, additional elicited grammar surveys carried out with Hewramî speakers, and further recordings served as the basis for the description of linguistic features of T ekht Hewramî in Khan & Mohammadirad’s (2024a) detailed study of language contact in Sanandaj entitle d: Language contact in Sanandaj: A study of the impact of Iranian on Neo- A ramaic . The documentation of Hewramî began to take shape during the work with Geoffrey Khan on the mentioned b ook between 2020 and 2023. The linguistic material for this bo ok comes principally from a field trip to the region in A ugust 2022. During that field trip, which lasted three weeks, I con- tacted the locals in Hewramî, which greatly facilitated fieldwork and contact with the inhabitants. I made recordings of 15 narrativ es. The recordings wer e made using a Zoom H5 Handy Recorder , which produced audio files in W A V for- mat. I conducted my fieldwork mostly in Hewraman T ekht but also visite d Serû Pîrî (a village north of He wraman) and Benen (the summer habitat for the in- habitants of Hewraman T ekht, locate d in the highlands). During this fieldwork, I transcribe d most of the r e cor dings and double-che cked my interpretation of recordings with my nativ e assistant, Amir , to ensure that the corr e ct interpr eta- tion had be en achie ved. In addition, I carried out many elicite d grammar surveys (including a questionnaire) with Amir and, occasionally , with a few other pe ople in Hewraman T ekht. The questionnaire was developed within the framework of the project “ A Linguistic Histor y of Minorities in the Near East” at the Univ ersity of Cambridge for studying phonological, morphosyntactic, and lexical variation 12 1.7 Main te xt corpus within Kurdic. 6 The 15 narratives collecte d in this trip w ere transcribe d and trans- lated and later fully glosse d as a te xt corpus volume representing He wramî oral narratives ( see Mohammadirad 2025c for details). The fourth field trip took place in July 2023. On this trip, I mainly conducted elicitation tasks on issues I encountered while preparing a first draft of the cur- rent grammar . Additionally , I compiled the verb list pr esente d in Appendix C and collected more recordings in Hewraman T ekht. During this trip, I collecte d further recordings of spontaneous spe ech in Hewraman T ekht, Serû Pîrî, and Bennen. This trip allowed me to trav el to the nearby villages and get a first im- pression of linguistic div ersity within the lo calities in which T ekht Hewramî is spoken. I was lucky to meet two competent stor ytellers in Nwên ( one of whom was originally from Silên), who narrated some 25 spoken narratives characteris- tic of the folkloristic tradition in Hewraman. These narratives, as w ell as a few additional tales that I gathered in Hewraman T ekht in July 2023, constitute a col- lection of tales, a comprehensive pr o cessing of which is planned as a projected publication (Mohammadirad in prep[b]). In discussing the linguistic features of T ekht Hewramî, I sometimes draw on the folkloristic material just discussed (see §1.8) to back up the description of Hewramî featur es. Still, the main grammatical analysis is based on the glossed text corpus in Mohammadirad (2025c). 1.7 Main text corpus The 15 narratives collected in A ugust 2022 (see the pre vious section) comprise the main corpus behind the current bo ok. These narratives yield a total of 96 minutes of running spee ch. The recordings hav e be en time-aligned with transla- tion using the annotation software ELAN. The r ecordings, along with the ELAN files, and time-aligne d texts, have b een archived on the op en-access platform Zen- odo (see Mohammadirad 2025a), and can b e found at https://zenodo.org/records/ 15419952. The recorded texts belong to different supposed genres: folktales, local anec- dotes, myths, oral histor y of the r egion, and autobiographies. The resulting texts form the T ekht Hewramî database and ar e a touchstone for the grammatical de- scription and lexicon. The text corpus behind this study has been entirely glossed in FLEx, from which the glossed texts wer e converted into the spe cific format- ting for linguistic examples required by Language Science Pr ess in L A T E X, using 6 https://www.ames.cam.ac.uk/research/project/echoes- vanishing- voices- mountains- linguistic- history- minorities- near- east 13 1 Introduction a Python script. The glossaries at the end of the book were drawn from the T eX file using additional Python scripts. The narrators who recounted the narratives were all o ver 60 years old at the time of recording. In many ways, then, this descriptive grammar is indicativ e of the language as spoken by the older generation. Nonetheless, in a few cases, the grammar highlights the generational difference in language use . As for the linguistic profile of the narrators, except for the narrator of the te xt JE, who is a monolingual, the rest exhibit some bilingualism in Kur dish. These speakers have a w eak bilingualism pattern, with Hewramî as their dominant language and Kurdish as less dominant. In addition, some of these narrators show ed weak competence in Persian. The 15 narratives consist of appro ximately 10,000 wor ds. T able 1.1 lists the ti- tles of the texts and their identifier codes in Mohammadirad’s (2025c) volume. The examples taken from the latter v olume are cited in the grammar using the initials of the texts along with an identifier number corresponding to the num- bered annotation units–typically a sentence or a clause. For instance, ZB.20 cor- responds to sentence number 20 in the ZB text. This guides the readers to the larger linguistic context from which the examples ar e taken. There are some glossing and translation conventions worthy of noice . Object language examples are pr esented in two versions, an orthographic one that cor- responds to actual surface realisations, and the morphologically segmented ver- sion, which contains repr esentations of the morphemes that are closer to their assumed underlying forms. Example (7) illustrates how the surface form awekê ‘the water’ in the orthographic version can be segmented at the underlying repre- sentation, while the second line in example (8) illustrates the underlying analysis for winû ‘blood of’ and kîseɫê ‘tortoise (f.sg.obl)’ in the orthographic version. (7) awekê biřo . aw(î)-ekê water.f-def.f.sg biř-o cut.prs.ind -3sg:A ‘He cut off the water supply . ’ [DP .34] (8) winû kîseɫê sawî be... win(î)-û blood.f-ez.gen kîseɫ(î)-ê tortoise.f- obl.f sáw-î rub.prs.sbjv -2sg:A be to ‘Y ou may rub tortoise ’s blo od on ... ’ [DG.47] A feature of Hewramî narratives is the frequent use of the present tense as the narrative tense to recount past ev ents (see §9.3.1.1). Consequently , all the tales 14 1.7 Main te xt corpus T able 1.1: The main text corpus (Mohammadirad 2025c) title id t opic zaroɫe û bizê ‘The baby and the goat’ ZB Local anecdote/myth ab out an abandoned baby zaroɫe û qiřolû darî ‘The baby and the tree hollow’ ZQ Local anecdote/myth ab out an abandoned baby herbene ‘The donkey keeper’ HB Lo cal anecdote/myth about a talking donkey peɫê merekuř ‘ A swarm of grasshoppers’ PM Local ane cdote/myth, grasshoppers derde gulî ‘Leprosy’ DG Local anecdote/myth ab out a man suffering from leprosy Şêx ʕumer û Cafir san ‘Sheikh Omar and Jafir San’ ŞC Local anecdote ab out recent history duwê padşε ‘T w o kings’ DP Oral history , two kings claiming Hewraman jîwayû Pîr Şelîyarî ‘Pir Shaliyar’s life’ JP Oral histor y , hagiography zemawinew Pîr Şalîyarî ‘Pir Shaliyar’s wedding’ ZP Oral history , hagiography babaw Pîr Şalîyarî ‘Pir Shaliyar’s grandfather’ BP Oral histor y , hagiography kuřû şuwaney ‘The shepherd’s son’ KŞ Folktale jîwayû Heyasî ‘Hayas’s life ’ JH Folktale řisûmatû ewsayma ‘Our past traditions’ RE Recollections of traditional life jîwayû ewsayma ‘Our past life ’ JE Recollections of traditional life, autobiography jîwayû min ‘My life ’ JM autobiography and narratives in Mohammadirad (2025c, in prep[b]) — which are the sour ce for linguistic examples in the current bo ok — hav e b een translated into the past tense, including in cases wher e the narrative present is used to describe past time events ( see 7 for an example). Readers are encouraged to keep this in mind when interpreting the examples. In the translation of linguistic examples, square brackets indicate w ords and meanings that are implicit or not stated in the text, while r ound brackets clarify the reference of participants in the text, see (9). 15 1 Introduction (9) maço, ‘nezanam. ’ m-aç-o ind -say.prs-3sg:A ne-zana=m neg-know.pst=1sg:A ‘He (the man) said, ‘I didn’t understand [his p oint]. ’ [ JH.26] 1.8 Folktale corpus As briefly discussed in §1.7, the morphosyntactic description of Hewramî is fur- ther backed up by additional folktales which I gathered at Nwên and Hewraman T ekht in A ugust 2023. These folktales provide a rich source for studying the folk- loristic tradition of Hewraman. Once fully processed, they will be published as a collection of folktales from Hewraman. For the current grammar , I have occa- sionally included example sentences from these tales in the main te xt. All the examples in the book from the folktale collection come with the initials of the tales and an identifier number , which refers to the place where the e xamples have been cited wit hin the mentioned tale. The folktale collection consists of 31 tales, totalling around 25,000 w ords. Here , it suffices to name each narrative with its initials, with a detailed description awaiting a projected publication of these tales (Mohammadirad in prep[b]). 1. mûsa û řuwase (MR) ‘Moses and the fox’ 2. mama mama (MM) ‘Mama mama’ 3. pîrejenî û kitê (PK) ‘The old woman and the cat’ 4. dêđê û çwar kinaçê (ÇK) ‘The stepmother and four girls’ 5. jenê fêlebaze ( JF) ‘ A cunning wife ’ 6. jenî û lîrewir eş ( JL) ‘The woman and the lira-seller’ 7 . wiɫkɫe (WL) ‘Wilkle’ 8. hacî û jenekêş ( JC) ‘Haji and his wife’ 9. şaw mara û şaw melekuřa (ŞŞ) ‘The king of snakes and the king of grasshop- pers’ 10. jenî û ħewt we y wê (H W) ‘ A woman and seven daughters-in-law’ 11. patşa û pîrî (PP) ‘The king and the old man’ 12. kuře keçeɫe (KK) ‘The bald b oy’ 13. hêyasî jîr (HJ) ‘He yas the Wise’ 14. siɫtan mehmûđ û heyasî jîr (SH) ‘Sultan Mahmoud and He yas the Wise’ 15. meɫik ehmeđ (ME) ‘Malek Ahmad’ 16. Kinaçêw T aɫî meẍrêbî (KT) ‘The daughter of T al from Maghreb ’ 17 . mar û p eyxwmer (MP) ‘The snake and the Prophet’ 16 1.8 Folktale corpus 18. meɫa yoso û meɫa xiđir (YX) ‘Mullah Y oso and Mullah Khidr’ 19. şê bayzîdî boystamî (BB) ‘Sheikh Bayzid Bostami’ 20. çil paɫewanê û hezretû ʕelî ( ÇH) ‘Forty warriors and His Highness Ali’ 21. pađşa û wezîr (PW) ‘The king and the vizier’ 22. duwê birayê (DB) ‘T w o brothers’ 23. meɫa xiđir û şûwane (XŞ) ‘Mullah Khidr and the shepherd’ 24. xiđre xuɫekêş (XX) ‘Khidr the Soil Carrier’ 25. ehmeđe dize (ED) ‘ Ahmad the Thief’ 26. şa ʕebas (ŞE) ‘Shah Abbas’ 27 . sîyawehş û keyxesrew (SK) ‘Siyawahsh and Keykhosr ow’ 28. hezretû mûsay û fêrʕon (MF) ‘Moses and Pharaoh’ 29. hacî mehmûđ û řozgarya (HM) ‘Haji Mahmoud and the Rozgaris’ 30. bêjen û menîje (BM) ‘Bezhan and Manizha’ 31. heyas û siɫtan mehmûyî (HS) ‘Heyas and Sultan Mahmoud’ 17 2 T yp ological o ver vie w This chapter lays out the major typological features of T ekht Hewramî. This should provide r eaders with a first-hand touch on the language, though the cov- erage of features r emains concise and e clectic. Defining grammatical properties of Hewramî includes a primary phonological gender assignment system, split- ergative alignment, two-term case system, differ ential argument flagging, differ- ential argument indexing, disharmonic SO V order , phonemic stress placement, and a complex deictic system. 2.1 P honolog y The consonant inventory includes 29 consonant phonemes, four of which are pe- ripheral phonemes occurring in a few loanwords or limited in their distribution within the syllable or wor d. These are represented in parentheses in T able 2.1. T able 2.1: Consonant phonemes Labial Lab .-dent. Alv . Postalv . Pal. V el. Uv . P har . Glott. Stop p b t d k ɡ q (ʔ) Affricate tʃ dʒ Nasal m n (ŋ) Fricative f s z ʃ ʒ x (γ) ħ ʕ h T ap ɾ T rill r Lateral l (ɫ) Approx. w j The vow el inventory consists of nine phonemic vowels: four fr ont vow els <î> /i/, <ê> /e/, <ɛ> /ɛ/, <e> [ɛ ∼ æ]; three back vo wels <û> /u/, <o> /o/, <a> /ɑ/; and two central v owels <i> /ɨ/ and <u> /ʊ/. Of these, /i, e, ɛ, u, ɑ, o/ and [ɛ ∼ æ] are long vow els, while /ɨ/ and /ʊ/ are short. The vo wel phonemes are distinguished by height and backness, as well as by their phonetic realisation (see §3.1.1.2). 2 T ypological overview T able 2.6: The inflection of witey ‘sleep’ in 1sg acr oss different T AM categories Present indicative m-ûs-û [ind-sleep.prs-1sg:S] Present subjunctive b-ûs-û [sbjv-sle ep .prs-1sg:S] Present progr essive m-ûs-ay m-ûs-û [ind-sleep.prs-nmlz ind-go.prs-1sg:S] Past progressiv e wis-ay wis-ên-a [sleep.prs-nmlz go.prs- a ug-1sg:S] Habitual past wis-ên-a [sleep.prs- a ug-1sg:S] Irrealis past wis-ên-a [sleep.prs- a ug-1sg:S] Past perfective wit-a [sleep.pst -1sg:S] Past conditional wit-εn-ê [sleep.pst -cond.a ug-1sg:S] Perfect wite=na [sleep.pst.pt cp.m=cop.1sg:S] Perfect progressive wit-î wite=na [sleep.pst -nmlz sleep.pst.pt cp.m=cop.1sg:S] Irrealis perfect wite=b-û [sleep.pst.pt cp.m=be.prs-1sg:S] Conditional perfect wite=bî-ɛn-ê [sleep.pst.pt cp=be.pst - cond.a ug-1sg:S] Past perfect wite=b-ên-ê [sleep.pst.pt cp.m=be- a ug-1sg:S] Perfect pluperfe ct wite=bîye=na [ sleep.pst.pt cp.m=b e .pst.ptcp.m=cop.1sg:S] The verb forms with present-time r eference fall broadly into four classes. In all verb classes, the negation of the indicative is identical to the pr ohibitive, as opposed to the negation of the subjunctive. Class 1 features the majority of verbs, as exemplified by the verb berđey ‘take ’ . Here, indicativ e, subjunctive , and imp er- ative verb forms ar e prefix-less. The verbs beginning with m in this class excep- tionally have the prohibitiv e prefix ne- . Class 2 is specific to verbs with a C(V) structure , with the exception of bîye y ‘b e , become ’ (prs b- ; pst bî- ), which belongs to class 1. The verb forms in this class r egularly take the T AM pr efixes, except the imperative prefix is occasionally dropped. Class 3 is limite d to lo w-vowel-initial verbs, with the negativ e prefixes for both the indicative and prohibitive being nime- , unlike the verbs in classes 1 and 2. Class 4 is limited to high back-vowel and mid-vow el-initial verbs. Like the verbs in class 3, the v erb forms in this class feature vo wel coalescence of the T AM prefixes with the stem. Howe ver , unlike in class 3, the verb forms in class 4 use the negation forms mé- . An e xception is the verb êşay ‘to hurt’ , for which the negative of the indicative can be expr esse d by either mé- or nimé- (see §9.1). Hewramî has a mixed adpositional typology , which reflects its structure being affected by b oth O V languages like T urkish, and V O languages like Arabic, and Aramaic (Stilo 2009). Additionally , adp ositions e xhibit applicative-like properties when taking pronominal arguments ( see §10.1.4). Example: 26 2.3 Syntax T able 2.7: V erb classes with present time reference , infle cted in 2sg ind sbjv imp/proh 1 ber- ‘take’ aff ber-î  bér-î bér-e neg mé-ber-î né-b er-î mé-ber-e 2 de- ‘give ’ aff mi-đe-î  bi-đé-î (bi)-đ(e)-é neg mé-đe-î né-đe-î mé-đ( e)-e 3 az- ‘let’ aff m-az-î  b-áz-î b-áz-e neg nim( e)-áz-î n-áz-î nim(e)-áz-e 4 ûs- ‘sleep ’ aff m-ûs-î  b-û  s-î b-û  s-e neg mé-ws-î né-ws-î mé-ws-e (9) xway ketê pey kîyasen. xwa-î God.m-sg.obl ket-ê be d-indf pey to kîyase= n send.pst.pt cp.m=cop.3sg.m:R ‘[ As if] God had sent him a bed. ’ [ JP .69] 2.3 Syntax The basic order of modifiers within the NP is DEM N UM N ADJ POSS, e xem- plified in (10). Hewramî uses two different head linkers in the structure of the NP , dubbe d ezafe/ezafeh in Iranian linguistics, depending on the category of the modifier: -î marks attributive ezafe, whereas -û marks genitive ezafe. A ddition- ally , an ezafe compound -e is used in the language with tightly-knit comp ound NPS, e.g. nan-e taz( e)-êwe [ bread.m-ez.cmpd fr esh-indf] ‘a fresh [loaf of] bread’ . (10) a duwe kuře ʕale w emîrî a dem.dist duwe two kuř-e son.ez.cmpd ʕal-e-û good-def-ez.gen emîr-î pn-m.sg.obl ‘those two good sons of Emir’ Case marking and ezafe marking interact in the structure of the noun phrase. When two possessors follow the head noun, only one formative is r etained. In the following example , the expecte d construction would b e qefesû sîne-y-û minne [ cage.m-ez.gen chest.m-sg.obl-ez.gen 1sg=post]. Howev er , in competition for the slot on the first possessor , only the oblique suffix remains, and the ezafe gets deleted. 27 2 T ypological overview (11) qefesû sîney minne qefes-û cage.m-ez.gen sîne- y chest.m-sg.obl min=ne 1sg=post ‘in my chest [lit. in the cage of my chest]’ [DP .38] 2.3.1 W ord order Hewramî has a default SO V order . This ordering is characterised by the subje ct not carrying the nuclear stress. The immediate pre-verbal slot is associated with the basic place of the focus in the clause, illustrated in (12). Occasionally , the sub- ject constituent comes b etw e en the v erb and its direct object (13). This typically occurs when the subject constituent is in focus. The fo cality of the A argument in past constructions can trigger the absence of indexing of the A argument on the verb ( see §2.3.3.1). (12) çêrhur zaroɫeke şot wer o. çêr=hur under=post zaroɫe-( e)ke child-def.m.sg.dir şot milk.m wer-o eat.prs.ind -3sg:A ‘The baby drank [its] milk from below . ’ [ZB.45] (13) heywane awê b er de. heywane animal.f.sg.dir awê  waterf.sg.obl berd-e take.pst -3sg.f:O ‘The flood [ lit. water ] to ok away the animals. ’ [ZB.21] Post-verbal objects are rare . If they occur at all, they are limited to nominals with definite refer ence evoked in the previous discourse . They seem to be limite d to certain clause types, e.g., interrogatives (14) and imperatives (15). (14) maça, ‘şanat tomeke?’ m-aç-a ind -say.prs-3pl:A şana=t scatter.pst.3sg:O=2sg:A tom-eke seed-def.m.sg.dir ‘They would say , ‘Did you plant the see ds?’’ [ JP .39] (15) mekojdê a kabray! me-koj-dê proh-kill.prs-2pl:A a dem.dist kabra-î fellow-m.sg.obl ‘Do not kill that man!’ [SH.268] 28 2.3 Syntax Despite having O V order , Hewramî exhibits se veral head-initial con- figurations, including Noun- A djective, Possessed-Possessor , Matrix clauses- complement clause, V erb-Goal, and V erb-Re cipient, running against the predic- tions of head-directionality hypothesis (Dr yer 1992). As for non-cor e arguments, goals of verbs of motion (16), recipients (17), and addressees (18) are ov er whelmingly realised in the post-verbal position, repre- senting Hawkins’s (2008) SO VX typ e , wher e X is the non-core argument. Though note that Hawkins (2008) uses the notation ‘X’ to refer to all kinds of non-core arguments, including also comitatives, instrumentals, place , etc. These latter ar- guments tend to be realised pre-predicatively in Hewramî. (16) êtir milo law ađî. êtir disc.pt cl mi-l-o ind -go.prs-3sg:S la-û to-ez.gen ađî 3sg.obl.m ‘ Anyhow , he went to him. ’ [ JP .13] (17) nanîç miđa to. nan=îç bread.m=add mi-đ( e)-a ind -give.prs-3pl:A to 2sg ‘They will give y ou a meal. ’ [HB.40] (18) maço be xanî. m-aç-o ind -say.prs-3sg:A be to xan-î chief.m-sg.obl ‘He said to the chief. ’ [KŞ.97] 2.3.2 Alignment Hewramî features a tense-based split ergative system both in verbal argument indexing and argument case marking. The alignment system is nominative- accusative for verbs deriv ed from the present tense and ergative-absolutiv e for verbs derived from the past stem, though note that ergative alignment is most consistently evident in the pattern of argument indexing. This system of tense- sensitive alignment depends not on the transitivity of the clause in a semantic sense but on the lexical transitivity of individual verbs. That is, a v erb is lexically specifie d as either transitive or intransitive, and that will determine ho w it in- flects in T AM constructions base d on the past stem of the verb . Whether a v erb has an overt object or not is irrele vant. In terms of verbal argument indexing, Hewramî uses verbal affix person mark- ers to index A and S in verbs deriv ed from the present stem, while O is expressed 29 2 T ypological overview via person clitics. The indexing of S and A is obligator y , whereas the indexing of O is contingent on the absence of the coreferent nominal. (19) milo. mi-l- o ind -go.prs- 3sg:S ‘It (a tortoise) w ent. ’ [DG.61] (20) beroş. ber- o=ş take.prs.ind - 3sg:A =3sg:O ‘He took him. ’ [DG.17] In verbs derived from the past stem, O and S ar e indexed by verbal affixes ( or copula p erson endings in perfect tenses), whereas clitic pr onouns index the A argument. Only indexing of S is obligatory . O-indexing (see §11.1.2.2) and A - indexing ( see §11.1.2.1) are nearly obliqgator y . (21) amaymê. ama- îmê come.pst - 1pl:S ‘W e came. ’ (22) berđîmêşa. berđ- îmê=şa take.pst - 1pl:O=3pl:A ‘They took us. ’ In terms of case marking, in clauses where the verb is deriv ed from the present stem, S (23) and A (24) are marked in the direct case, whereas O (25) is marked in the oblique case. (23) seʕbê wiɫaxdarê mila. seʕbê morning.f.sg.obl wiɫaxdar -ê stableman-pl.dir mi-l-a ind-go.prs-3pl:S ‘In the morning, the horse grooms went. ’ [ŞC.66] (24) îse dêwê xeber zana îse now dêw -ê ogre-pl.dir xeber news zán-a know.prs.sbjv -3pl:A ‘If the ogres find out [ ab out y ou] now’ [SK.100] 30 2.3 Syntax (25) yeway ç bero. yew(e)-a =îç barley.f-pl.obl=add ber-o take.prs.ind -3sg:A ‘He took the barley seeds, to o. ’ [ JP .29] In clauses with verbs deriv e d fr om the past stem, S (26) and O (27) are marked in the direct case, whereas A (28) is marked in the oblique case. (26) karewanîyê amê yanê serere . karewanî-ê caravan_people-pl.dir amêya=nê come.pst.pt cp.pl=cop.3sg.pl:S sere=re . top=postp ‘Some passers-by had stayed there . ’ [DB.15] (27) hewarêşa wişkinɛnê . hewar -ê =şa summer_habitat.m-pl.dir=3pl:A wişkinɛ=nê scour.pst.pt cp.pl=cop.3pl:O ‘They scoured the summer habitats [ searching for food etc.]. ’ [ JE.3] (28) ênne paɫewana zorşa kerđen ênne so_much paɫewan-a warrior-pl.obl zor=şa pressure=3pl:A kerđe=n do.pst.pt cp.m=cop.3sg:O ‘The warriors put much pressure [ on the skin]’ [SK.132] 2.3.3 Grammatical relations 2.3.3.1 Differential A indexing Hewramî features tense-sensitiv e indexing of the transitive subject A argument: in clauses based on the present stem of the verb , verbal person/number suffixes index the A argument, wher eas in clauses derived from the past stem of the verb , the historical clitic pronouns inde x the A argument. Unlike the verbal affixes, the clitic pronouns indexing the A -past argument are mobile. Furthermor e, while in- dexation of A -prs through inflectional suffixes is obligator y , the p erson clitics in- dexing A -past arguments are sometimes missing. They may be, at least partially , in complementary distribution with an overt oblique case-marked A argument. In (29a), the oblique-marked A padşa is the sole way to express the A argument. In (29b), the 3sg clitic =iş , resumes the corefer ent absent A argument. (29) a. min taze padşay kerdena wekêɫ. min 1sg taze anyway padşa-î king.m-sg.obl kerde=na do.pst.pt cp.m=cop.1sg:O wekêɫ advocate.m ‘I—the king has put me in charge. ’ [ZP .107] 31 2 T ypological overview b . watenîçiş, mişyo neberûşo. wate=n=îç= iş say.pst.pt cp.m=cop.3sg.m:O=add=3sg:A mişyo a ux ne-ber-û=ş=o neg.sbjv -take.prs-1sg:A =3sg:O=compl ‘He (the king) has said [to me], “Y ou shall not take her back. ’” [ZP .108] This optional indexing of A -past arguments is triggered by the A argument displaying properties related to focus (Mohammadirad & Haig forthcoming). A focused A argument in Hewramî can b e in non-contrastiv e or constrastive focus. The non-contrastive focus is further divided into wh-fo cus and completiv e fo cus (see §11.1.2). By way of example , in (30), the A argument has nuclear focus, and the A -inde xing p erson clitic =ş is missing: (30) î zeře çermeme kê berden eçêge? î dem.pro x zeře money-ez.cmpd çerme=m=e white=1sg:PSR=deic kê  who berde=n take.pst.pt cp.m=cop.3sg.m:O eçêge here ‘ Who has taken my white money [that is now ] here?’ [PK.29] When the A argument is in completive focus, it is not indexed via mobile person clitics. In (31), in response to the wh-question, the focused 2sg A argument is not indexed. (31) maço ‘miɫk ehmeđ! î dijmenême kê kuştênê? to kuştênê!?’ m-aç-o ind-say .prs-3sg:A miɫk pn ehmeđ pn î dem.pro x dijmen-ê-m=e enemy-pl.dir=1sg=deic kê  who kuştê=nê kill.pst.pt cp.pl=cop.3pl:O tò 2sg kuştê=nê kill.pst.pt cp.pl=cop.3pl:O ‘He (the king) said, Oh Milk Ahmad! Who has killed my enemies? Y ou have killed them!?’ [ME.150] By contrast, when the oblique-marked A NP is in the topic position and not fo- cused, the person clitic resumes it. This explains the co-occurrence of the subject indexing clitic and the oblique-marked A argument in (32). Similarly , oblique- marked A argument in the post-verbal position co-o ccurs with the co-indexing person clitic (33). The clitic indexing here follows fr om the fact that the topical A argument is placed in the non-focal p ost-verbal position as an afterthought. 32 2.3 Syntax (32) î pîyay tawaş î kinaçêşe kerde be qerarê weşeş kerde we... î dem.pro x pîya-î man.m-sg.obl tawà =ş can.pst=3sg:A î dem.pro x kinaçê=ş=e girl.f.sg=3sg:A =dem kerd-e do.pst -3sg.f:O be to qerar-ê settlement-indf weş-e=ş well-f=3sg:A kerd-e=we do.pst -3sg.f:O=compl ‘[ And if] the man has be en able to cure the girl... ’ [ZP .45] (33) qotê aman. asawekeş bînan qotekey . qot( e)-ê box-indf ama=n come.pst.pt cp.m=cop.3sg.m:S asaw-eke= ş mill.m-def=3sg bîna=n block.pst.pt cp.m=cop.3sg.m:O qot-ekey box-def.obl.m ‘ A box came [floating on the water]. The box blocked the (water ) mill. ’ [MF .75]–[MF .7 6] 2.3.3.2 Differential P indexing As seen in §2.3.2, P arguments are inde xed by clitic pronouns in present tense constructions but via inflectional p erson/number suffixes in past transitive con- structions. Here , differential P indexing means a deviation fr om the canonical ergative construction wher eby the obje ct is not inde xed on the verb. V erbal af- fixes are obligatory indexes of direct objects in canonical ergative constructions, illustrated by the following example . (34) to minit quɫ kerđaw . to 2sg min =it 1sg=2sg:A quɫ pierced kerđ- a =û do.pst -1sg:O=and ‘Y ou disabled me. ’ [P W .30] In clauses with O A V order , the verb tends to agree with the topical O . (35) werêsekê min warđêne . werês( e)-ekê rope-def.f.sg min 1sg warđê=ne eat.pst.pt cp.f=cop.3sg.f:O ‘[The lion said], ‘I have eaten the rope . ’’ [ÇH.85] The expected P indexing is sometimes absent with inanimate Ps that are plural. In the following examples, the v erb has a default 3sg.m inflection and do es not agree with the plural object. 33 2 T ypological overview (36) penc çemçêşa nîyanre. penc five çemç( e)-ê=şa spoon.m-pl.dir=3pl:A nîya=n=re put.pst.pt cp.m=cop.3sg.m:O=povb ‘They (my family ) had set five spoons [on the tablecloth]. ’ [ JE.46] (37) ewêş nîya biraw wêş. ewê=ş 3pl.dir=3sg:A nîya put.pst.3sg.m bira-û brother.gen.ez wê=ş reflx=3sg:PSR ‘He made them (the ogres) his brothers. ’ [ME.99] O-past indexing may also be absent due to affix co-optation by a higher-ranked argument in terms of animacy . In the following e xample, the expected 3pl O- agreement suffix is absent on the verb because its slot has b een taken over by the 3sg suffix indexing the oblique argument. (38) kîyast sênze danê heserêşa da pene. kîyast send.pst sênze thirteen danê clf.pl heser( e)-ê=şa mule.f-pl.dir=3pl:A da-Ø give .pst -3sg:R pene to ‘He [the king] sent [his men]. They [ his men] gav e him (Imam Ali) thirteen mules. ’ [ÇH. 69] 2.3.3.3 Differential A flagging The alignment system licenses case marking for A arguments. In T AM construc- tions derived from the present stem of the v erb, A is marked in the direct case, realised as a zero suffix in the singular and - ê in the plural. By contrast, in v erbal categories derived from the past stem of the verb , the A argument should be, by default, accompanied by the oblique case suffixes. The split alignment is only relevant for thir d-person nouns and pronouns. Spe ech act pronouns have lost the case distinction (see §2.3). The following example illustrates the oblique case marking on the A -past argument. (39) cafir sanî fermawan, ‘lodê!’ cafir pn san-î pn-m.sg.obl fermawa=n say.pst.pt cp.m=cop.3sg.m:O lo-dê go.prs.imp-2pl:S ‘Jafir San said, ‘Go [and bring him]!’’ [ŞC.36] In reality , not all A -past arguments are oblique-marked. A token frequency count of overt A s rev eals that a quarter of the third person As are not oblique- marked (see §11.2.1). The data suggest that information prominence triggers 34 2.3 Syntax oblique case marking on A -past arguments. The latter operates at two lev els: “local” and “global” (Chapp ell & V erstraete 2019). In Hewramî, the former is gen- erally associated with oblique marking and the latter with direct case marking. “Local” prominence is associated with the A argument b eing in narro w focus and contrastive focus. In (40), the case marking on her ‘ donkey’ is triggered by its contrast with min . (40) î her-î zûwaniş zana min hîçim nezanan. î dem.pro x her -î donkey-sg.obl.m zûwan=iş language.m.sg.dir=3sg:A zana-Ø know.pst -3sg.m:O min 1sg hîç=im nothing=1sg:A ne-zana=n neg-know.pst.pt cp.m=cop.3sg.m:O ‘The donkey knew the [Sheikh’s] language; I didn’t kno w a thing!’ [HB.71] Case marking on the A argument can also be triggered by global prominence. Accor ding to McGregor’s (2006) “ expecte d actor principle ” , in episodes of dis- course with an expected actor , the actor can b e left unmarked after its intro- duction. Any deviation from the e xpe cted actor is marked in the ergative case. One of the factors conditioning differential subject marking seems to be topic continuity . In the following e xcerpt, the establishe d dir e ct-marked topic of the intransitive clause in (41a), is repeated with the transitive clause in (41b), ev en though the oblique form ađîşa is expected with the latter clause. (41) a. tenya ađê luwɛnê. tenya only ađê i 3pl.dir luwɛ=nê go.pst.pt cp.pl=cop.3pl:S ‘Only they i (Baba Khwada, Hama the Invisible, and Little Hama) went [to Iraq]. ’ b . êtir ađê watenşa, ‘ême diruwê meyeymê . êtir disc.pt cl ađê i 3pl.dir wate=n=şa say.pst.pt cp.m=cop.3sg.m:O=3pl:A ême 1pl diruwê lie.f me-de-îmê neg.ind -give.prs-1pl:A ‘They i said, ‘W e are not going to lie. ” [BP .116]–[BP .117] In terms of information structure, non-oblique marked overt A -past arguments tend not to carry the nuclear stress. In other words, they behave like topics and contain given information ( see §11.2.1 for discussion). 35 3 Phonetics, phonology , and morphophonology (2) toyç /tʰo .itʃ/ > [tojtʃ] ‘y ou to o ’ /e/ <ê> is a close-mid front unrounded v owel: (3) êş [ʔeʃ] ‘pain’ şêr [ʃeɾ] ‘lion’ kê [kʰe] ‘who’ /ɨ/ <i> is a close-central unrounded vow el. This vow el has restricted distri- bution. W ord-initially , it appears to b e occurring with disyllabic or trisyllabic words, e .g., iti ‘any more. ’ W ord-finally , it only app ears with clitic particles, that is, elements that cannot stand by themselves as a free w ord, e.g., gi ‘ each. ’ (4) iti [ʔɨ.ˈtɨ] ‘any more ’ çil [tʃɨl] ‘forty’ gi [ɡɨ] ‘ each’ /i/ is used as an epenthetic vowel to br eak up some consonant clusters syllable- initially , especially in careful spe ech (see § 3.4.4). (5) bira ∼ bra ‘brother’ /i/ may also be use d to break up comple x consonant clusters in the syllable coda, e.g.,: (6) gerim < germ ‘warm’ /ɛ/ <ɛ> is an open-mid front unrounded vow el. /ɛ/ is the result of the contrac- tion of /a/ and /ê/. When realised in open syllables, it is very close in phonetic realisation to a near-open front unr ounde d v owel /æ/, the difference being that /ɛ/ is more fr ont and has a shorter length than /æ/ (see §3.1.1.2.4). (7) amɛ [ˈʔɑmɛ] ‘they came ’ watɛt [ˈwɑtɛt] ‘[if] you had said’ /æ/ <e> is a near-open front unrounded vow el. Its realisation approaches /ɛ/ in open syllables (see §3.1.1.2.3). (8) esb [ʔæsb] ‘horse ’ ber [ bæɾ] ‘ door’ ce [ dʒæ] ‘in, from’ /ɑ/ <a> is an open back unrounded vow el: (9) ađ [ʔɑɹˠ] ‘he ’ par [pʰɑɾ] ‘last year’ ça [tʃɑ] ‘there ’ 42 3.1 Phoneme inventory /o/ <o> is an open-mid back rounded vowel. It do es not occur word-initially in monosyllabic words. (10) oge [ʔo.ˈgɛ] ‘ther e’ kor [kʰoɾ] ‘blind’ to [tʰo] ‘you ( sg.)’ /ʊ/ <u> is a near-close near back rounded vow el. It is limited to word-medial position in monosyllabic words in the text corpus. (11) gul [ɡʊl] ‘bad’ kul [kʊl] ‘blunt (knife)’ /u/ <û> is a close back rounded vow el. Like other back-rounded vowels, /û/ is excluded word-initially , except occasionally in disyllabic words. (12) ûne [ʔu.ˈnæ] ‘that’ dûr [ duɾ] ‘far’ şû [ʃu] ‘husband’ 3.1.1.2 P honetic realisation of vow els Figure 3.1: Phonetic realisation of vow els This section describ es the phonetic r ealisation of individual vowels plotted in vow el charts. The quality of the allophonic realisation of vo wels was mea- sured using the acoustic analysis software Praat (Boersma & W eenink 2015). The 43 3 Phonetics, phonology , and morphophonology F1 and F2 frequencies for each vo wel wer e plotte d on the v owel chart. F1 has an inverse relationship with v owel height: the higher the v owel, the low er the F1. On the other hand, F2 has an inverse r elationship with the backness of the vow el, such that back vow els are characterised by lower F2 than front v owels. Diagram 3.1 maps the average realisation of v owels onto the cardinal v owel di- agram. Square brackets repr esent cardinal vo wels. The realisation of vo wels is based on analysing the acoustic properties of vow els averaged out for at least ten words per vo wel. A 20-year-old male speaker of T ekht Hewramî pr o duced the words. 3.1.1.2.1 /i/ As seen in Figure 3.2, the allophones of the close fr ont unrounded vowel /i/ ar e generally scattered in the area between cardinal vo wels [i] and [ e]. The allo- 1. metî  ye ‘aunt’ 8. péştî ‘back’ 2. hêzî  ‘yesterday’ 9. tirazû  yî ‘balance’ 3. gicî  ‘shirt’ 10. tifî  ‘mulberr y’ 4. hévîcî ‘ carrot’ 11. avî  r ‘fire ’ 5. tûtî  ‘parrot’ 12. jîr ‘sage ’ 6. góşî ‘ ear’ 13. tarî  k ‘dark’ 7. mékî ‘salt’ 14. keɫeşî  r ‘rooster’ 15. qeyî  m ‘old’ Figure 3.2: Phonetic realisation of /i/ <î> 44 3.1 Phoneme inventory phones have a tendency to become more back, espe cially in the environment of an unvoiced obstruent, e.g., mékî [ˈmɛ.ki] ‘salt’ , tarî  k [tʰɑ.ˈɾik] ‘dark’ . 3.1.1.2.2 /e/ The positional variants of the close-mid front unrounded vo wel /e/ show a rel- atively narro w scatter; see Figure 3.3. The sample consists of stressed /e/ (1–7) and their non-stressed counterparts (8–10). The allophones of /e/ are primarily positioned close to the cardinal vow el [e] marked by the square . An exception is the unstressed /e/ in namê ‘name ’ , which has a more central realisation, possibly due to the effect of the preceding nasal consonant. 1. qijê  ‘hair’ 6. hê  zmê ‘firewood (pl.dir)’ 2. birê  ‘ eyebr ow’ 7. yagê  ‘place ’ 3. kinaçê  ‘ daughter , girl’ 8. pérê ‘the day before yesterday’ 4. sêtê  ‘husband’s sister’ 9. námê ‘name’ 5. dêyê  ‘stepmother’ 10. eyálê ‘mother ( vocative)’ Figure 3.3: Phonetic realisation of /e/ <ê> 3.1.1.2.3 /æ/ Figure 3.4 shows the allophonic r ealisation of /æ/ in closed syllables. It can be seen that /æ/ is scattered in an area corresponding to the realisation of the car- dinal vow el [ɛ] in most cases, e.g., tejné ‘thirsty’ . In addition, some allophonic 45 3 Phonetics, phonology , and morphophonology stressed /æ/ in closed syllables unstressed /æ/ in closed syllables 1. kérge ‘hen’ 11. texté ‘wood’ 2. péştî ‘back’ 12. tersáy ‘fear’ (inf) 3. řénge ‘ colour’ 13. heştal û  î ‘plum’ 4. çérme ‘white ’ 14. kerđéy ‘ do (inf)’ 5. kéşkî ‘ dried whey’ 15. berđéy ‘take ’ (inf) 6. yehér ‘liv er’ 16. merđéy ‘ die ’ (inf) 7. pél ‘leaf’ 17. pencéw pay ‘heel’ 8. çép ‘left’ 18. jenbirá ‘brother of wife ’ 9. hévr ‘ cloud’ 19. tejné ‘thirsty’ 10. héşt ‘ eight’ 20. nemdár ‘known’ Figure 3.4: Phonetic realisation of /æ/ <e> in closed syllables variants of /æ/ come close to the realisation of [ æ], e.g., pél ‘leaf’ . Figure 3.5 demonstrates the allophonic variation of /æ/ in open syllables. As can be seen, allophones of /æ/ in open syllables are predominantly scattered in an area between the cardinal vo wels [ɛ] and [e], regardless of whether they are str ess-b earing or not. Comparing Figure 3.4 to Figure 3.5, it can be said that allophones of /æ/ occupy a considerable phonetic space. Notably , the allophonic realisation of /æ/ shows mor e raising in open syllables than in close d ones. 46 3.1 Phoneme inventory stressed /æ/ in open syllables unstressed /æ/ in open syllables 21. taté ‘father’ 31. lû  te ‘nose’ 22. çénî ‘needle’ 32. vérve ‘snow’ 23. mékî ‘salt’ 33. tî  re ‘arrow’ 24. jénî ‘woman’ 34. táce ‘ crown’ 25. îsé ‘now’ 35. bíze ‘goat’ 26. tazé ‘new’ 36. çóge ‘knee ’ 27. tûté ‘ dog’ 37. şéve ‘night’ 28. kité ‘ cat’ 38. séye ‘shade ’ 29. pêré ‘the day after tomorrow’ 39. léme ‘b elly’ 30. m[é]y e ‘sheep’ 40. tevérg[ e] ‘hail’ Figure 3.5: Phonetic realisation of /æ/ <e> in open syllables 3.1.1.2.4 /ɛ/ Figure 3. 6 represents the phonetic realisation of /ɛ/. A s seen in T able 3.1, the average realisation of /ɛ/ is phonetically v er y close to the average r ealisation of /æ/ in open syllables. One difference is that /ɛ/ is more fr ont, i.e., it has a relativ ely higher F2. The other difference is that /æ/ is longer than /ɛ/. Inv estigating the average length of these vo wels per two words yields the follo wing figures. /ɛ/ has a length of 90 ms, whereas the figure is 17 0 ms for unstressed /æ/ and 150 ms for stressed /æ/. 1 1 Our phonetic analysis of /ɛ/ and /æ/ in Luhon H. shows that the same tendency holds for the difference between /ɛ/ and /æ/, running contrary to MacKenzie’s (1966: 8) description of the vow el system in Luhon, which presents /ɛ/ as a long vow el and /æ/ as a short vowel. 47 3 Phonetics, phonology , and morphophonology 1. lúwɛ ‘they went’ 6. zanábîyɛm ‘(if) I had known’ 2. nélɛ ‘they didn’t go ’ 7. yáwɛ ‘they arrived’ 3. luwɛ  nî ‘you (f) have gone ’ 8. yáwɛ ‘if they arriv e d’ 4. luwɛ  ne ‘she has gone ’ 9. wátɛt ‘[if] you had said’ 5. luwɛ  nê ‘they hav e gone’ 10. wátɛş ‘if he/she had said’ 11. pî  yɛ ‘men’ Figure 3. 6: P honetic realisation of /ɛ/ <ɛ> T able 3.1: P honetic realisation differ ence b etw een /ɛ/ and /æ/ vow el F1 F2 /ɛ/ 490 2061 stressed /æ/ in open syllables 519 1927 unstressed /æ/ in open syllables 506 1824 The distinction between /ɛ/ and /e/ <ê> is phonemically distinctive, as shown by the following pairs. (13) bî  yê ‘they were ’ bî  yɛ ‘if they had b een’ luwê  nê ‘they were going’ luwɛ  nê ‘they have gone ’ 48 3.1 Phoneme inventory 3.1.1.2.5 /ɑ/ As sho wn in Figure 3.7, the allophonic realisation of /ɑ/ sho ws a wide scatter between cardinal vow els [a] and [ɑ]. It can be seen that /ɑ/ has a back realisation in the environment of back consonants, e .g., qur waqî ‘frog’ , and velarised con- sonants ( cause d by the spr ead of phar yngealisation), e.g., gemaɫ ‘male dog’ . The great div ersity in the p ositioning of /ɑ/ seems to have trigger ed the raising of [æ] to [ɛ] (see the preceding sections) as an instance of a push chain, meaning that the diversity in the phonetic positioning of /æ/ could have potentially made the distinction between [æ] and [ɑ] difficult and this, in turn, resulted in the raising of [æ] to [ɛ]. 1. babá ‘grandfather’ 8. waɫê  ‘sister’ 2. xuyá ‘God’ 9. yagê  ‘place ’ 3. qurwáqî ‘frog’ 10. eçá ‘there ’ 4. gáve ‘ cow’ 11. hamsá ‘neighbour’ 5. degá ‘village ’ 12. gîváv ‘plant’ 6. laló ‘maternal uncle ’ 13. xás ‘goo d’ 7. tatê  ‘father’ 14. varán ‘rain’ 15. gemáɫ ‘male dog’ Figure 3.7: Phonetic realisation of /ɑ/ <a> 49 3 Phonetics, phonology , and morphophonology 3.1.1.2.6 /o/ The allophonic realisation of /o/ is mainly scatter e d between cardinal v owels [ o] and [ɔ]; se e Figur e 3.8. /o/ tends to have a more back r ealisation in the envi- ronment of the v elarise d consonant /ɫ/, e .g., zoɫfê ‘hair (collective)’ , the voiceless velar plosive /k/, e .g., çiko ‘from where ’ , and in closed syllables, e.g., sot ‘it burnt’; qoɫ ‘ deep’ . 1. góşê ‘ ear’ 8. sawró ‘ cow’s dung’ 2. laló ‘maternal uncle ’ 9. séro ‘ on top ’ 3. nîmeřó ‘midday’ 10. aró ‘today’ 4. çikó ‘from where ’ 11. sot ‘it burnt’ 5. dóga ‘village ’ 12. pos ‘skin’ 6. cûcóɫê ‘ chicken’ 13. çoɫ ‘vacant, empty’ 7. zóɫfê ‘hair ( colle ctiv e)’ 14. kor ‘blind’ 15. qoɫ ‘ deep’ Figure 3.8: Phonetic realisation of /o/ <o> 50 3.1 Phoneme inventory 3.1.1.2.7 /u/ Like other vow els, the allophones of the close back rounded vow el /û/ exhibit a wide scatter; see Figure 3.9. It is particularly interesting that the r ealisation of /û/ is close to the cardinal vo wel [ o], espe cially in the environment of v elar consonants, e.g., hangû  rî ‘grap e ’ , but also in closed syllables, e.g., sûr ‘red’ . 1. lû  te ‘nose’ 6. hengû  rî ‘grap e ’ 2. kîlû  ‘kilo ’ 7. dûr ‘far’ 3. tûté ‘ dog’ 8. sûk ‘light’ 4. zû ‘ early’ 9. sûr ‘red’ 5. şû ‘husband’ 10. berû  ş ‘I will take it’ Figure 3.9: Phonetic realisation of /u/ <û> 51 3 Phonetics, phonology , and morphophonology /d/-deletion also occurs in post-consonantal slots, where the rele vant conso- nant is a rhotic, though not deleted in merđ . (35) kursan [kʊɾ.ˈsɑn] ‘Kurdistan’ <* Kur d isan zer [zæɾ] ‘yellow’ < cf. CK. zerd merđ [mæɾɹˠ] ‘he died’ Similarly , /d/ often undergoes lenition to a velarised alveolar appro ximant [ɹˠ] in post-vocalic position. 2 (36) ađ [ʔɑɹˠ] ‘he ’ dađ [ dɑɹˠ] ‘shout’ 3.1.2.1.5 /k/ /k/ → { [kʰ]/# [k  ]/ # /k/ is a voiceless velar stop . It is usually strongly aspirated syllable-initially , except when follow ed by central vowels or v oiceless fricative consonants. In word-final p osition, it is unr eleased. Therefore , /k/ forms a natural class with the voiceless stops /p/ and /t/. (37) keř [kʰær] ‘ deaf’ kinaçê [kɨ.nɑ.ˈtʃe] ‘girl’ kul [kʊl] ‘blunt’ bêşkê [b eʃ.ˈkɛ] ‘ cot’ sûk [suk  ] ‘light’ 3.1.2.1.6 /g/ /g/ → [ɡ] /ɡ/ is a voiced velar stop which occurs both syllable-initially and syllable- finally . (38) gave [ˈɡɑ.vɛ] ‘ cow’ beg [ bæɡ] ‘ chieftain’ 2 See Mohammadirad & Öpengin (2024) for an overview of the lenition of voiced stops within Kurdic, including in Gorani dialects. 58 3.1 Phoneme inventory 3.1.2.1.7 /q/ /q/ → [q] /q/ is an unvoiced uvular stop. It occurs both syllable-initially and syllable- finally . Due to its back articulation, /q/ is not generally aspirated in the syllable- initial position. (39) qesab [ qɛ.ˈsɑb] ‘butcher’ qaqez [ qɑ.ˈqæz] ‘pap er’ teq [tʰæq] ‘knocking’ 3.1.2.1.8 /ʔ/ /ʔ/ → [ʔ]/# V /ʔ/ is a voiceless glottal stop. It does not have contrastiv e phonemic status in Hewramî and only occurs in syllable-initial position. Its occurrence is condi- tioned by the avoidance of empty onsets in wor d-initial p osition. (40) asaw [ʔɑ.ˈsɑw ] ‘mill’ avî [ˈʔɑ.vi] ‘water’ 3.1.2.1.9 /tʃ/ /tʃ/ → [tʃʰ]/# /tʃ/ is a voiceless post-alveolar affricate, r epresented as <ç>. It occurs b oth in syllable-initial and syllable-final position. In the former p osition, it is generally aspirated. (41) çêş [tʃʰeʃ] ‘what’ /tʃ/ undergoes assimilation in voicing when preceded by a vowel and follow ed by alveolar /d/. Here , the stop part of the affricate is deleted and b ecomes voiced. (42) wajdê [ wɑʒ.ˈde] ‘Y ou (pl.) say!’ cf. waçdê 3.1.2.1.10 /dʒ/ /dʒ/ → [dʒ] /dʒ/ is a voiced post-alveolar affricate, repr esented as <c>. It o ccurs in syllable- initial and syllable-final position. 59 3 Phonetics, phonology , and morphophonology (43) cîya [ dʒi.ˈjɑ] ‘separate’ bêcge [ˈb edʒ.ɡɛ] ‘apart from’ Note that /dʒ/ has a low phonemic load in Hewramî. One of the known isoglosses within Iranian is that the Old Iranian palatal approximant /j/ is pr e- served syllable initially in Gorani while it has shifte d to /dʒ/ in Kurdish, e .g.: (44) Gorani Kurdish yew e co, ceh ‘barely’ cf. A v . yauua yeher cerg, ciger ‘liv er’ The low functional load of <c> in Hewramî may e xplain the realisation of [ dʒ] in the Arabic loan aciz ‘upset’ as /d/, hence adiz [ʔɑ.dˈɨz]. 3.1.2.1.11 /m/ /m/ → [m] [m] is a voiced bilabial nasal that occurs syllable-initially and syllable-finally . (45) mûso [mu.ˈso] ‘he/she is sle eping’ kam [kʰɑm] ‘which’ 3.1.2.1.12 /n/ /n/ → { [n] [ŋ]/ C [ +sibilant] /n/ is a voiced alveolar nasal that occurs syllable-initially and syllable-finally . (46) namê [ˈnɑ.me] ‘name ’ nan [nɑn] ‘food’ /n/ is weakened before the voiced sibilant consonants /z/ and /s/, and sounds like a nasal glide: (47) paŋze [pɑŋ.ˈzɛ] ‘fifteen’ sêŋze [seŋ.ˈzɛ] ‘thirteen’ paŋsew [pɑŋ.ˈsæw ] ‘five-hundred’ 60 3.1 Phoneme inventory 3.1.2.1.13 /ŋ/ /ŋ/ → [ŋ] The phoneme /ŋ/ is a voiced velar nasal, r epresented by the grapheme <ŋ> in words wher e it always app ears, as in paŋze ‘fifteen’ (abov e). In some cases, howe ver , the phoneme is not stable in the current state of the language , and some speakers pronounce it as <ng>. Notably , /ŋ/ only occurs syllable-finally . (48) deng [ dæŋ] ‘voice ’ ceng [ dʒæŋ] ‘war’ 3.1.2.1.14 /f/ /f/ → [f] /f/ is a voiceless labio-dental fricative . It occurs syllable-initially and syllable- finally . (49) fire [fɨ.ˈɾɛ] ‘much, very’ sêf [sef] ‘potato ’ 3.1.2.1.15 /s/ /s/ → ⎧ ⎪ ⎨ ⎪ ⎩ [ s ] [ ṣ ]/# 𝑉 [ + lo w + back − rounded ] ∼ 𝑉 [ + lo w + back − rounded ] /s/ is a voiceless alv e olar fricativ e. It occurs both word-initially and wor d- finally . (50) saɫe [sɑɫɛ] ‘y ear’ mas [mɑs] ‘yoghurt’ /s/ be comes pharyngealised in the environment of back vow els. This has be en shown with acoustic analysis for the following pairs (see Khan & Mohammadirad 2024a: 24). The greater F1 frequency in the pharyngealised words suggests the lowering of the v owel in the envir onment of the phar yngeal phoneme. Similarly , the lower F2 for the phar yngealised vo wels means that these vow els are realised further back compared to the vow els in non-phar yngealised wor ds. (51) şeṣ [ʃɑsˤ] ‘sixty’ /ɑṣ/ F1=623 F2=982 mes [mæs] ‘ drunk’ /æs/ F1=589 F2=1464 61 3 Phonetics, phonology , and morphophonology (52) ṣe [sˤɑ] ‘hundred’ /ṣɑ/ F1=604 F2=942 îse [ʔi:ˈsɛ] ‘now’ /sɛ/ F1=438 F2=17 08 3.1.2.1.16 /z/ /z/ → { [z] [ j] / V V ( optional) /z/ is a voiced alveolar fricative that occurs syllable-initially and syllable- finally . (53) zînan [zi.ˈnɑn] ‘prison’ payîz [pɑ.ˈiz] ‘autumn’ /z/ is sometimes lenited to a palatal approximant [ j] in inter vocalic position. (54) meyanû [mɛ.jɑ.ˈnu] ‘I don’t know’ cf. mezanû 3.1.2.1.17 /ʃ/ /ʃ/ → [ʃ] /ʃ/ is a voiceless post-alveolar fricative , represented as <ş>. It occurs both syllable-initially and syllable-finally . (55) şot [ʃot] ‘milk’ tiş [tɨʃ] ‘acid’ /ʃ/ undergoes assimilation in voicing when preceded by a vowel and follo wed by alveolar /d/: (56) kujdê [kʊʒ.ˈde] ‘Y ou (pl) kill!’ cf. kuşdê gojd [ɡoʒd] ‘meat’ cf. goşt, goşd 3.1.2.1.18 /ʒ/ /ʒ/ → [ʒ] /ʒ/ is a voiced post-alveolar fricative, r epresented as <j>. It occurs b oth syllable- initially and syllable-finally . (57) jenî [ˈʒɛ.ni] ‘woman’ řoj [roʒ] ‘ daylight’ 62 3.1 Phoneme inventory 3.1.2.1.19 /x/ /x/ → [x] /x/ is a voiceless velar fricativ e that occurs syllable-initially and syllable- finally . (58) xele [xɛ.ˈlɛ] ‘grain’ xas [xɑs] ‘good’ puxte [pʊx.ˈtɛ] ‘ clean’ bax [bɑx] ‘garden’ 3.1.2.1.20 /ɣ/ /ɣ/ → [ɣ] /ɣ/ is a voiced velar fricativ e, represented as <ẍ>. It only occurs syllable- initially . (59) aẍe [ʔɑ.ˈɣɛ] ‘lord’ ẍeyb [ɣæjb] ‘ disappeared’ 3.1.2.1.21 /ħ/ /ħ/ → [ħ] /# /ħ/ is a voiceless pharyngeal fricative. It occurs both in words of Iranian stock and loanwords from Semitic languages. In the nativ e lexicon, the non- etymological /ħ/ has developed through either pharyngealisation of the glottal fricative /h/ or in the environment of back vow els; see ʕine . (60) ħot [ħot] ‘sev en’ cf. P . haft ʕine [ˈʕɨnɛ] ‘buttock’ cf. CK. /NK. qûn ħîç [ħitʃ] ‘nothing’ cf. P . hîç Note that the phar yngealisation of /ħ/ in ħîç ‘nothing’ seems to be conditione d by information structure , such that when the word is in focus, /h/ tends to be pharyngealise d. This is likely due to the “perceptual magnet effect” associated with pharyngeals (Blevins 2017). /ħ/ is preserved in loanwords fr om Arabic, where it occurs mainly syllable- initially . The syllable-final position of /ħ/ results either from the r etention of /ħ/ from the source w ord, e.g., nîkaħ ( see b elow ) or from shifting the originally syllable-initial voiced pharyngeal [ʕ] to the syllable co da through metathesis and 63 3 Phonetics, phonology , and morphophonology then devoicing it to [ ħ], see cuħme . Note that variation exists between speakers for devoicing [ʕ] syllable-finally . Therefore , no categorical rule can be p osited here . (61) ħeywan [ħæj.ˈwɑn] ‘animal’ ħukm [ħʊkm] ‘rule’ nîkaħ [ni.ˈkɑħ] ‘marriage ’ cuħme [ dʒʊħ.ˈmɛ] ‘Friday’ cf. Ar . jomʕa ħegaɫ [ħɛˈɡɑɫ] ‘scarf’ cf. Ar . ʕiqal 3.1.2.1.22 /ʕ/ /ʕ/ → { [ʕ] [ħ] # (optional) /ʕ/ is a voiced phar yngeal fricativ e. It is loaned from Arabic and can occur only syllable-initially . [ʕ] sometimes undergoes devoicing to [ħ]. (62) ʕal [ʕɑl] ‘go od’ cf. Ar . ʕaliya temaʕe [tʰɛ.ˈmɑ.ʕɛ] ‘greed’ cf. Ar . ṭamaʕiyya cuħme [ dʒʊħ.ˈmɛ] ‘Friday’ cf. Ar . jomʕa deħfe [ˈdæħfɛ] ‘ exclusion’ cf. Ar . dafʕa 3.1.2.1.23 /h/ /h/ → [h] /h/ is a voiceless glottal fricative that occurs only syllable-initially . (63) her [hæɾ] ‘ each, every’ hamin [hɑ.ˈmɨn] ‘summer’ /h/ is regularly deleted from the syllable-final position in cognate wor ds. (64) mêman [me .ˈmɑn] ‘guest’ cf. P. mehmān şa [ʃɑ] ‘king’ cf. P . şāh 3.1.2.1.24 /ɾ/ /ɾ/ → { [ɾ] /σ , σ (not word-initial) [r ]/# [rˤ]/σ 64 3.1 Phoneme inventory There is a three-way distinction of rhotics in T ekht Hewramî. The rhotics can be realised as a tap /ɾ/, alveolar trill /r/, or an emphatic /rˤ/. /ɾ/ is a voiced alveolar tap, r epresented as <r>. It occurs syllable-initially and syllable-finally . (65) heser e [ hɛ.ˈsɛ.ɾɛ] ‘mule ’ here [hɛ.ˈɾɛ] ‘the donkey’ pîr [pʰiɾ] ‘ old’ şar [ʃɑɾ] ‘ city’ /ɾ/ is excluded word-initially , where voiced alveolar trill /r/, repr esented as <ř>, is used instead. In other words, the contrast between /r/ and /ɾ/ is neutralised wor d-initially . /r/ occurs syllable-initially and syllable-finally . In the latter posi- tion, it usually appears with low and back vow els. (66) řas [rɑs] ‘ correct, right’ kuř [kʊr] ‘boy’ keř [kʰær] ‘ deaf’ Hewramî also has the emphatic /rˤ/, which occurs word-internally but is lim- ited to the syllable initial position. (67) merˤe [ˈmɛ.rˤɛ] ‘ cave ’ herˤe [ˈhɛ.rˤɛ] ‘mud’ The distinction between these rhotic consonants has b een measured by in- strumental acoustic analysis in Khan & Mohammadirad (2024a: 30). The pha- ryngealise d /rˤ/ has low er F2 than /ɾ/ and /ř/, suggesting its back articulation in the vocal cavity . The adjacent vo wels also show low er F2 in the environment of pharyngealise d /rˤ/, suggesting flat resonance . Similarly , F1 is higher in the envi- ronment of pharyngealised /rˤ/, meaning that the tongue is closer to the b ottom of the oral cavity . It can also be seen that F2 is lower for the alveolar trill /ř/ in beř ‘product’ in comparison to the flap /ɾ/ in here ‘ donkey’ , reflecting flatter resonance . 3 (68) mer e [ˈmɛ.ɾɛ] ‘grassland’ /e/ F1=523 F2=1526 /ɾ/ F1=427 F2=1542 /e/ F1=490 F2=1562 merˤe [ˈmɑ.rˤɛ] ‘ cave ’ /e/ F1=725 F2=1119 /rˤ/ F1=637 F2=1036 /e/ F1=643 F2=1131 3 As the pharyngealised /rˤ/ has a very low functional load in Hewramî, and indeed there is considerable cross-speaker variation in its production, I use the trilled <ř> to represent it. 65 3 Phonetics, phonology , and morphophonology (69) her e [ hɛ.ˈɾɛ] ‘ donkey’ /e/ F1=559 F2=1526 /ɾ/ F1=486 F2=1525 /e/ F1=515 F2=1561 herˤe [ˈhɑ.rˤɛ] ‘mud’ /e/ F1=668 F2=1145 /rˤ/ F1=634 F2= 1049 /e/ F1=637 F2=1164 (7 0) beř [bær] ‘pr o duct’ /æ/ F1=528 F2=1403 /r/ F1=447 F2=1469 ber ˤ [ bærˤ] ‘ dried’ /æ/ F1=653 F2=1112 /rˤ/ F1=7 05 F2=1143 3.1.2.1.25 /l/ /l/ → [l] /l/ is a voiced alveolar lateral. It occurs b oth syllable-initially and syllable- finally . (71) lalo [lɑ.ˈlo] ‘maternal uncle ’ çil [tʃɨl] ‘forty’ 3.1.2.1.26 /ɫ/ /ɫ/ → { [l]/# [ɫ]/ # /ɫ/ is a velarised alveolar lateral. It occurs mostly syllable-finally . It can some- times occur syllable-initially; for instance, when the feminine suffix -e is added to masculine nouns and adjectives with /ɫ/ in their coda, resyllabification occurs and /ɫ/ ends up as the onset of the final syllable. Unlike its plain counterpart, /ɫ/ cannot occur word-initially . (72) laɫe [ˈlɑ.ɫɛ] ‘ deaf’ (f) saɫe [ˈsɑ.ɫɛ] ‘year’ 3.1.2.1.27 /j/ /j/ → [ j] /j/ is a palatal approximant r epresented as <y>. It occurs syllable-initially and syllable-finally . (73) yerê [ˈjɛ.re] ‘thr ee’ berzep e y [ bær .zɛ.ˈpæj] ‘standing’ 66 3.2 Phonotactics 3.1.2.1.28 /w/ /w/ → { [ v] ∼ [ w]/# [ w]/ # [ʋ]/ V V (optional) /w/ is a bilabial appro ximant. It is usually realised as a labio-dental fricative [ v] syllable-initially . (74) watiş [ˈvɑt.ɨʃ] ‘he/she said’ wînû [ vi.ˈnu] ‘I see’ In syllable-final position, realisation as /w/ is more common. (75) asaw [ʔɑ.ˈsɑw ] ‘mill’ masaw [mɑs.ˈɑw ] ‘fish’ The sound is sometimes lenited in inter vocalic position and realised as a labio- dental approximant [ʋ]. (7 6) aw eyanî [ɑ.ʋɛ.jɑ.ˈni] ‘prosperity’ miro kenê  we [mɨ.ˈɾo kɛ.ˈne.ʋɛ] ‘pear-picking’ 3.1.3 P honeme-grapheme associations The transcription system used in this b ook follows the ‘Hawar’ standar d Kurdish script ( Kurdish Roman (Hawar ) Orthography 2012). T able 3.3 exhibits ho w the graphemes in the Hawar script correspond to IP A symbols. The Standard Kurdish transcription system used in this book differs from MacK enzie’s (1966) transcription system used for Luhon Hewramî and, more broadly , the transcription systems in Iranian philology in the following aspects; see T able 3.4. 3.2 P honotactics The phonotactics of vow els and consonants have been describ ed separately for each phoneme in §3.1.1.1 and §3.1.2.1. Here , I present the distribution of vow els and consonants in separate tables. The most frequent syllable structures ar e CV , CV C, and CV CC. 67 3 Phonetics, phonology , and morphophonology (81) degakê  ‘the village ’ kuřî  ‘boy (obl.m)’ berđéy ‘to take ’ The derivational suffixes -î and -gerî are also stress-bearing. (82) wişkesaɫî  ‘ drought’ gewregerî  ‘grandeur’ Howe ver , the direct plural affix -ê and the indefinite suffixes -êw / -êwe / -ê are not stress-bearing. The stress thus remains penultimate. (83) hê  zmê ‘firewood’ řóê ‘ days’ kúřêw ‘a boy’ kinaçê  we ‘a girl’ Nouns in the vocative hav e the stress on the penultimate syllable: (84) tàte | ‘Father!’ èđa | ‘Mother!’ kinàçê | ‘Girl!’ Pronominal clitics are not stress-bearing and thus do not cause a change in the stress pattern of nouns they attach to . (85) yané꞊ta ‘your house ’ wê  ꞊ma ‘ourselv es’ yó꞊şa ‘ one of them’ Likewise , the additive clitic ꞊îç is not stress-bearing. (86) mín꞊îç ‘I too ’ 3.3.2 Adjectives The stress placement pattern in adjectives is similar to that in nouns. Thus, their stress pattern is not completely predictable . Masculine adje ctiv es follow the basic syllable-final stress pattern. On the other hand, the stress is on the penultimate syllable in feminine and plural adjectives. (87) ʕaqíɫ ‘wise (m)’ ʕaqíɫe ‘wise (f)’ ʕaqíɫê ‘wise (pl)’ 74 3.3 Str ess p osition The comparative and superlative suffixes -ter and -terîn receive w ord stress. (88) zil-tér ‘bigger’ zil-terî  n ‘biggest’ 3.3.3 Adv erbs Adv erbials, like masculine nouns, follow the syllable-final stress pattern. (89) êgé ‘here ’ îsé ‘now’ ití ‘any more , any way’ In some adverbials, the stress falls on the penultimate syllable: (90) wélê ‘but’ pérê ‘the day before yesterday’ bê  cge ‘except for’ péwkî ‘for the reason’ 3.3.4 V erbs Most verbs derived from the pr esent tense stem have no morphological distinc- tion between the subjunctive and imperfe ctiv e verb forms (see §9.1). Stress po- sitioning is the only way to distinguish between identical subjunctive and im- perfective moo ds for such verbs. In the imperfective , the stress is placed on the syllabic inflectional person suffix (91). In the subjunctive, the stress is placed on the first syllable of the verb (92). (91) yaw-û  ‘I arrive ’ yaw-î  ‘you arrive ’ yaw-ó ‘he/she arrives’ yaw-mê  ‘we arrive ’ yaw-dê  ‘you arrive ’ yaw-á ‘they arrive ’ (92) yáw-û ‘that I arrive ’ yáw-î ‘that you arrive ’ yáw-o ‘that he/she arrives’ yáw-mê ‘that we arrive ’ yáw-dê ‘that you arrive ’ yáw-a ‘that they arrive ’ 75 3 Phonetics, phonology , and morphophonology Figure 3.13a and 3.13b exhibit different stress patterns associated with the verbs beró ‘he takes’ [ JP .29] and béro ‘that he takes’ [HB.11]. For each verb , the higher intensity peak highlights the stressed vow el. (a) indicativ e ‘he takes’ (b) subjunctive ‘that he takes’ Figure 3.13: The stress position for the verb ‘he takes’ With disyllabic present stems in the subjunctive , the stress is on the first syl- lable. (93) gírîno ‘that he/she boils’ If the vow el of the inflectional p erson affix becomes a glide following a vow el- final verb stem, the stress is placed on the last syllable of the stem ( se e 94). Note additionally that ‘give ’ is one of the few verbs that take the indicativ e prefix (see §9.1.1). (94) mi-đé-y ‘you (sg) give ’ In verb forms deriv e d fr om the past stem, the stress is placed on the last sylla- ble of the verb stem. The bound person markers use d to inflect past stem verbs do not bear stress. These include inflectional p erson affixes, used to inflect past intransitive verbs, and clitic pronouns, used to inflect past transitive verbs ( see §9.1.7). Inflectional person affixes in past intransitive do not take stress because historically they originate from the enclitic copula, which under w ent univerba- tion with the past participle verb . (95) Past intransitive yawá-nê ‘I arrived’ yawá-y ‘you arrived’ yawá-∅ ‘he/she arrived’ yawá-ymê ‘we arrived’ yawá-ydê ‘you arriv ed’ yawɛ  ‘they arrived’ 7 6 3.3 Str ess p osition (96) Past transitive wát꞊im ‘I said’ wát꞊it ‘you said’ wát꞊iş ‘he/she said’ wát꞊ma ‘we said’ wát꞊ta ‘you said’ wát꞊şa ‘they said’ Similarly , in past imperfective forms, the stress falls on the penultimate sylla- ble. (97) yaw-ê  n-ê ‘I was arriving’ yaw-ê  n-mê ‘we were arriving’ The stress pattern of the imperative/subjective pr efixes dep ends on the syl- lable structure of the verb . The formativ e is stressed with stems starting with a consonant cluster or if the stem consists solely of a consonant. Howev er , the stress shifts to the stem in verb stems with CV C structure (see §9.1.2 for details). (98) Imperative/subjunctive bí-nvîs-e ‘W rite!’ bí-l-a ‘that they go ’ b-zán-mê ‘that we know’ p-sán-û ‘that I know’ The negative prefixes ar e stress-bearing. (99) Negative/prohibitiv e mé-don-e꞊m ‘ do not talk to me!’ né-ker-mê ‘let’s not do (it)’ né-yaw-ên-mê ‘we wer e not arriving’ As sho wn, stress retraction applies to the subjunctive form of the v erb, e ven when the subjunctive prefix is absent. Historically , the loss of the subjunctive prefix led to shift of stress to the stem (see K arim & Mohammadirad forthcoming, Mohammadirad & Karim 2025). (100) yáw-û ‘(if) I arrive ’ yáw-î ‘(if) you arrive ’ yáw-o ‘(if) he/she arrives’ yáw-mê ‘(if) we arrive ’ yáw-dê ‘(if) you arrive ’ yáw-a ‘(if) they arrive ’ 77 3 Phonetics, phonology , and morphophonology The completive particle ꞊we/꞊o is not str ess-b earing. (101) amá=we ‘he came back’ ker-û  ꞊we ‘I will open’ 3.3.5 Copula Present copulas are not str ess-bearing. The stress is thus realised on the predicate. Note that feminine and plural adjectives do not take syllable-final stress, as is the case with nominals (see §3.3.1). (102) xás꞊na ‘I (m) am well’ xáe꞊na ‘I (f) am well’ xás꞊a ‘he is well’ xásê꞊nmê ‘we are w ell’ Similarly , past copula suffixes do not receive w ord stress. (103) xas b-ê  n-ê ‘I (m) was well’ xase b-ê  n-ê ‘I (f) was well’ xasê b-ên-mê ‘we were well’ 3.4 Major morphophonemic processes 3.4.1 h-initial insertion As discussed in §3.2.2, vo wel-initial wor ds are preceded by a glottal stop [ʔ] due to the restriction against onsetless syllables. Sometimes, wor d-initial [ʔ] is realised as a glottal fricative /h/. A survey of the text corpus r eveals that /h/ is initially inserted primarily before the front vow els <e> and <ê>, both in the native lexicon and in loanw ords. (104) Before /e/ hetîm ‘ orphan’ cf. Tr . etim heɫbet ‘of course ’ cf. Ar . albatte hesere ‘mule ’ cf. NK. êstir hesare ‘star’ cf. CK. estêre henar ‘pomegranate ’ cf. P . anār heke ‘if, well’ cf. P. agar ‘if’ In some cases, /h/ alternates freely with /ʔ/: 78 3.4 Major morphophonemic pr ocesses (105) hesp ∼ esb ‘horse ’ cf. P . asb hême ∼ ême ‘we ’ cf. CK. ême Less commonly , /h/ is added before words starting with v owels other than <e>: (106) hêɫeke ‘fine sieve ’ cf. T r . elek harđî ‘flour’ cf. P. ār d hiɫoşe ‘sour plum’ cf. P . ālûçe 3.4.2 Inversion in the voicing of pharyngeals A pharyngeal phoneme is sometimes b orro wed, but its voicing is inverted in Hewramî. (107) seʕbe [ˈsæʕ.bɛ] ‘morning’ cf. Ar . sˤabāħ deħfe [ˈdæħ.fɛ] ‘ expulsion’ cf. Ar . dafʕ cuħme [ dʒʊħ.ˈmɛ] ‘Friday’ cf. Ar . jomʕa ħegaɫ [ħɛ.ˈɡɑɫ] ‘scarf’ cf. Ar . ʕiqāl 3.4.3 Glide insertion An epenthetic glide is inserte d between syllable-initial vow el sequences to avoid hiatus. /w/ is used following [u]; elsewhere /j/ is used. (108) /piɑ/ > [pi.ˈjɑ] ‘man’ /guinɛ/ > [ɡu.wi.ˈnɛ] ‘sickly’ /dɛ.gɑ.ɑ/ > [ dɛ.ɡɑ.ˈjɑ] ‘villages (pl.obl)’ 3.4.4 Anaptyctic vow el insertion In Hewramî and other regional languages, the vow el generally employed to break up illegal consonant clusters is <i> (IP A [ɨ]) or <u> (IP A [ʊ]) depending on the quality of adjacent consonants. These vo wels may break up consonant clusters in the onset of the syllable. (109) bira < bra ‘brother’ şuwane < şwane ‘shepherd’ The anaptyctic i and u may also break up consonant clusters in the coda. The clusters that allow breaking up by <i> include rm, ɫm, sk, rs, řk, wr . As can be seen, the first segment in the cluster is a sibilant or a liquid, whereas the second segment is usually an occlusive (nasals and obstruents) and less frequently a rhotic. 79 3 Phonetics, phonology , and morphophonology (110) zoɫim < zoɫm ‘tyranny’ gerim < garm ‘warm’ nasik < nask ‘thin (stick)’ biřik < biřk ‘thick (stick)’ quris < qurs ‘heavy’ hewir < hewr ‘ cloud’ The clusters that allow breaking up by u appear to be less common. The first segment in such clusters is often a back consonant, e.g., /k/, /q/. (111) şukur < şukr ‘praise ’ nuquɫ < nuqɫ ‘ candy’ 3.4.5 Metathesis Many disyllabic loans from Arabic containing a pharyngeal consonant in the onset of the second syllable or in the code of the se cond syllable ar e subje ct to metathesis in Hewramî. These loans are typically tr eated as feminine, with an unstressed -e added if the loan does not already end in a -e . Therefore , in the loan word the phar yngeal phoneme would occur in the onset ( or the coda of) the second syllable, but when metathesis is applied, it ends up as the coda of the first syllable. (112) seʕbe [ˈsæʕ.bɛ] cf. Ar . sˤabāħ ‘morning’ weʕze [ˈwæʕ.zɛ] cf. Ar . wazʕ ‘situation’ cuʕme [ dʒʊʕ.ˈmɛ] cf. Ar . jomʕa ‘Friday’ deħfe [ˈdæħ.fɛ] cf. Ar . dafʕ ‘ expulsion’ meʕsûd [mæʕ.ˈsud] cf. Ar . masʕûd ‘Masoud’ The metathesis occurs to avoid non-permitte d consonant clusters acr oss sylla- ble boundaries (see §3.2.1.1 for consonants phonotactics). The motivation b ehind the metathesis here is that following the sonority hierar chy , the coda of the first syllable should be more sonorant than the onset of the second syllable. In other words, sonority must not rise across syllable b oundaries (Gousko va 2001). The non-permitted patterns in T able 3.8 all r esult in a rise of sonority in the syllable boundar y . Therefor e, the metathesis occurs to resolve the issue . In the following pairs, onset metathesis is attested across multiple CV syllable structures; the onset of the second syllable is metathesised with the onset of the third syllable . (113) heteqî [ hɛ.tɛ.ˈqi] ‘truly’ vs. heqetî [ hɛ.qɛ.ˈti] kiřêɫe [kɨ.re.ˈɫæ] ‘key’ vs. kɫêře [kɨ.ɫe.ˈrɛ] 80 3.4 Major morphophonemic pr ocesses T able 3.8: Non-p ermitted consonant sequences across syllable bound- aries leading to metathesis permitted cluster non-permitted cluster Gloss across σ-boundaries across σ-boundaries seʕbe [ˈsæʕ.bɛ] *[ˈsæb .ʕɛ] ‘morning’ weʕze [ˈwæʕ.zɛ] *[ wæz.ʕɛ] ‘situation’ cuʕme [ dʒʊʕ.ˈmɛ] *[dʒʊm.ʕɛ] ‘Friday’ deħfe [ˈdæħ.fɛ] *[ dæf.ħɛ] ‘ expulsion’ meʕsûd [mæʕ.ˈsud] *[mæs.ˈʕud] ‘Masoud’ Coda metathesis is only attested with CV CV structures: (114) nûrî [nu.ˈri] ‘force ’ vs. nîrû [ni.ˈru] Less commonly , metathesis occurs within the consonant clusters in the coda. (115) tiɫf [tɨɫf] ‘ child’ vs. tifɫ [tɨfɫ] 3.4.6 V owel hiatus V owel hiatus is generally avoided across or within w ord boundaries. There are several strategies to r esolve hiatus. The first is that the second vow el is glided, yielding a rising diphthong, e.g., /o y/, /ay/: (116) toyç ‘y ou too’ < [ˈtʰo.itʃ] pîyay ‘man (m.sg.obl)’ < [pʰi.ˈɑi] The second strategy is to add an epenthetic glide /j/ or /w/ between vowels (see §3.4.3): (117) gûwînê /ɡu.i.nɛ/ > [ɡu.wiˈnɛ] ‘sickly’ degaya /dɛ.ɡɑ.ɑ/ > [ dɛ.ɡɑ.ˈjɑ] ‘villages (pl.obl)’ Another strategy is to omit the first vow el, which is generally unstressed: (118) kîseɫê ‘tortoise (f.sg.obl)’ < kîséɫî + -ê (f.sg.obl) miđo ‘he/she gives’ < mi-đe + -o jena ‘women (pl.obl)’ < jénî + -a jenû ‘wife of’ < jenî + -û Y et another strategy is to coalesce the two vo wels into one: (119) pîyɛ ‘men (pl.dir) < pîya + -ê mê ‘he/she comes’ < m-e-o 81 3 Phonetics, phonology , and morphophonology 3.4.7 Assimilation Assimilation is a pr ocess in which a pair of adjacent segments b ecome similar . There are tw o kinds of assimilation: progressiv e and regressive. These assimila- tion processes have the effect of creating geminate consonants which are other- wise not attested in the language. T otal progressive assimilation is seen in the following wor ds, where /d/ fully assimilates to the preceding nasal sound: (120) çinne ‘how much’ < * çinde cf. CK. çende inne ‘this much’ < * inde cf. CK. ewende Regressiv e assimilation takes place in the following example, wher e the rhotic /ř/ assimilates to the following liquid. (121) kulle ‘small boy’ < * kuřle 82 4 Nouns and nominal mor phology Nouns are marked for gender, number, and case, though gender is not expressed in the plural. These categories are repr esented on nouns via fusional suffixes. The inflection of a noun is predictable from the phonological shape of the base and from its gender . 4.1 Nominal infle ction The categories case, number, and gender can b e marked morphologically on nouns. Each categor y inv olves two distinctions, thus singular and plural for num- ber, masculine and feminine for gender (only in singular number), and direct and oblique for case. The gender assignment is phonologically based (see §4.1.1). For now , it suffices to say that the nouns ending in a consonant, stressed -é , -í , -û  , and -ó are mascu- line. Additionally , some nouns ending in stressed -á are masculine. On the other hand, all the nouns ending in -ê , whether stressed or not, and nouns ending in unstressed -e and -î are feminine. The class of feminine nouns also includes a subset in stressed -á . As r emarked, the infle ction of a noun is predictable from the phonological shape of the base and from its gender . Masculine and feminine nouns are in their base form when singular and in the direct case. Masculine nouns end in -î and feminine nouns in -ê , when singular and in the oblique case. In the plural, gender distinction is lost. The relevant inflectional suffixes are -ê in the dir ect case, and -a in the oblique case, see T able 4.1. T able 4.1: Nominal infle ctional suffixes–underlying forms sg.dir sg.obl pl.dir pl.obl m -Ø -î -ê -a f -Ø -ê 4 Nouns and nominal morphology (12) eđaw tatew jenê merđ eđa=w mother=and tate -û father-ez.gen jenê wife.f.sg.obl merđ- Ø die.pst -3sg.m ‘The wife ’s parents died. ’ [XX.110] Likewise , gender distinction is phonologically co ded for some animals. (13) Feminine Masculine gawe ‘ cow’ gaw ‘bull’ bizɫê ‘female kid-goat’ bizɫe ‘male kid-goat’ For a subset of animals, gender distinction is semantically encode d. This would include using different lexical items for e xpressing gender distinctions between females and males: 1 (14) Feminine Masculine jerejî ‘female partridge ’ be q ‘male partridge ’ meye ‘sheep ’ beran ‘ram’ kerge ‘hen’ keɫeşîr ‘rooster’ mayînî ‘mare ’ esb ‘horse ’ qişqêre ‘female raven’ qaɫawe ‘male raven’ bize ‘goat’ sabrîn ‘male goat’ bize neçîre ‘female mountain goat’ keɫ ‘male mountain go at’ A subset of semantically-encoded gender nouns are morpho-phonologically encoded. Thus, the difference b etw e en her ‘ donkey’ and ma-her-e ‘female don- key’ is expr essed by the prefix ma- ‘female ’ combine d with the feminine suffix -e . For some animals, a fixed gender is used to encompass both male and female members: (15) verg (m) ‘wolf’ şêr (m) ‘lion’ meřekur (m) ‘grasshopper’ hewrêşe (m) ‘rabbit’ řûwáse (f) ‘fox’ héşe (f) ‘bear’ kité (f) ‘ cat’ 1 See also MacKenzie (1966: 14, ff.2) 90 4.1 Nominal inflection As seen, the animals in (13)–(14) have a gender split, wher eas those in (15) do not. This split cannot be easily accounte d for in terms of the animacy hierar- chy , which is generally understood as large animals with which humans interact regularly are higher in animacy , which would consequently b e r eflecte d by the gender split in languages. The difference between the setS in (13)–(14) and the one in (15) seems to come from the degree of closeness of animals to the imme- diate human environment, or being traditionally part of the food chain, e.g., in the case of ‘partridge ’ or ‘mountain goat’ . As said abov e, the morphological shape of words also has a role in gender assignment. This concerns morphologically derived nouns, whose gender is de- termined by the derivational suffix (see §4.2.1) added to the noun. For instance, awîrga ‘fireplace ’ , a feminine noun, is compose d of awîr ‘fir e’ (m) and the femi- nine suffix -ga ‘place ’ . Some tendencies suggest themselves: • Abstract nouns formed from other nouns and adje ctives by the derivational suffix -î are masculine: (16) sextî ‘ difficulty’ weşî ‘happiness’ hemahengî ‘ collaboration’ neweşî ‘illness’ dewayî ‘medication’ • Abstract nouns formed with the suffix -gerî are masculine: (17) hewramîgerî ‘Hewramîhood’ gewregerî ‘nobleness’ pîyagerî ‘manhood’ aẍegerî ‘lordship ’ paşagerî ‘kingship ’ • Place names forme d with -ga/ -ge are feminine: (18) dega ‘village ’ < de ‘village ’ + -ga ‘place ’ awîrga ‘fireplace ’ < awîr (m) ‘fire ’ + -ga ‘place ’ seringa ‘pillow’ < serin (m)+ -ga ‘place ’ ewêge ‘there ’ < awê ‘that’ + -ge ‘ over there ’ coge ‘stream’ < * cû ‘stream’ + -ge ‘place ’ 91 4 Nouns and nominal morphology • Place names ending in -xane are masculine. (19) kitêbxane ‘library’ < kitêb ‘bo ok’ + -xane ‘place ’ darûxane ‘pharmacy’ < darû ‘medicine’ + -xane ‘place ’ eđebxane ‘toilet’ řoxane ‘riv er’ aşpezxane ‘kitchen’ • Nouns ending in -waɫe are masculine. Examples: (20) zerdewaɫe ‘bee ’ < zer d ‘yellow’ + -e + waɫe ‘wing’ tûwaɫe ‘ eggshell’ • Nouns of meals ending in -îne are feminine: (21) şelemîne ‘ dish made of cracked wheat and turnip’ dowîne ‘ dish made of cracked wheat and diluted yoghurt’ xeple zeřatîne ‘a type of bread made of corn flour’ Semantic factors also play a role in gender assignment. Small entities, e .g., small fruits and grains, tend to be feminine. The nouns in this class are further phonologically marked to b e feminine. An e xception is tifî  ‘mulberr y’ , which should be masculine according to the phonological assignment rule. This appar- ent anomaly seems to be motivate d by the loss of the central vo wel in tifî  and the stress shift to î . Alternatively , one might reflect that semantic criteria take precedence over phonological criteria in the gender assignment of tiny fruit. (22) mîjûyî ‘lentil’ nuwey e ‘ chickpea’ genmî ‘wheat’ yew e, ye wê ‘barley’ tifî ‘mulberr y’ wamî ‘almond’ wezî ‘walnut’ çeqalê ‘bitter almond’ hengûrî ‘grapes’ şêɫanê ‘apricot’ heştaɫûyî ‘plum’ tifɫe şêxanê ‘strawberr y’ 92 4.1 Nominal inflection It is notable that lûbya ‘beans’ and maş ‘black lentils’ are masculine and, thus, exceptions to the generalisation made above . These nouns appear to b e recent borrowings into Hewramî, pr obably from Persian. The gender assignment, at least for maş follows from the default phonological assignment rule . 4.1.1.2 Gender of loanwords W ords borrowed from other languages acquire gender in T ekht Hewramî. Ara- bic loanwords generally constitute an earlier layer of loans in the language, espe- cially the ones related to the realm of r eligion. Arabic loans denoting abstract con- cepts are borrow ed predominantly as feminine nouns, regardless of the original gender (see b elow for some exceptions). As can b e seen, these words are often bor- row ed with an extra unstressed vow el, which is not present in the source wor d. The category of nouns exhibiting this trait also includes the temp oral nouns such as seʕbe ‘morning’ . This r eflects that there is a semantic basis for gender assign- ment with loanwords. Howe ver , it is notable that the great majority of abstract nouns are formed via the native derivational suffixes -î and -gerî which assign masculine gender , e.g., se xtî ‘difficulty’ , gewregerî ‘nobleness’ (see §4.1.1); there- fore no claim can be made about an underlying semantic basis, already operating in the native lexicon, for assigning feminine gender to abstract nouns borrow e d from Arabic. (23) temaʕe (f) ‘gree d’ [ JE.58] cf. Ar . ṭamaʕiyya (f) ʕemre (f) ‘ order’ cf. Ar . ʕamr (m) řeza (f) ‘satisfaction’ [ JE.83] cf. Ar . riḍaʔ (m) qisê (f) ‘remark’ [ZB.59] cf. Ar . qiṣṣa (f) dawa (f) ‘plea’ [DP .13] cf. Ar . daʕwa (m) dinya (f) ‘world’ [ZP .128] cf. Ar . dunya (f) duʕa (f) ‘praying’ [DG.6] cf. Ar . duʕa (m) şerte (f) ‘ condition’ [ŞC.98] cf. Ar . şarṭ (m) weʕze (f) ‘situation’ [DG.34] cf. Ar . waḍʕ (m) ʕefwe (f) ‘mer cy’ [DG.4] cf. Ar . ʕafw (m) sinʕe (f) ‘age ’ [BP .123] cf. Ar . sinn (m) seʕbe (f) ‘morning’ cf. Ar . s ʕ abāħ (m) Exceptions to the above generalisation can be found. In such cases, Hewramî assigns gender to the Arabic loan based on its phonological system of gender assignment. Thus, if an abstract feminine noun in the source language ends in a consonant, Hewramî seems to treat it as masculine. 93 4 Nouns and nominal morphology (24) xizmet (m) ‘service’ [RE.12] cf. Ar . xidmat (f) şerîʕet (m) ‘Sharia law’ [ JP.84] cf. Ar . şariʕat (f) terîqet (m) ‘denomination’ [ JP.84] cf. Ar . tariqat (f) wiɫat (m) ‘region, countr y’ cf. Ar . wilayat (f) wext (m) ‘time’ cf. Ar . waqt (m) tekbîr (m) ‘planning’ cf. Ar . tadbîr (m) fikr (m) ‘thought’ cf. Ar . fikr (m) Elsewhere , the gender of Arabic loans is retained in Hewramî as long as the noun endings correspond in gender. (25) ħegaɫ (m) ‘scarf’ cf. Ar . ʕîqal (m) ʕeba (f) ‘robe ’ [BP .187] cf. Ar . ʕaba (f) lêfe (m) ‘ quilt’ cf. Ar . lîhaf (m) xet (m) ‘line ’ cf. Ar . xat (m) qifɫ (m) ‘lo ck’ cf. Ar . qufl (m) cins (m) ‘material, stuff’ cf. Ar . jins (m) wezî (f) ‘walnut’ is an e xception, apparently borrowed from the Arabic jawz (m). The lack of gender correspondence here seems to have a semantic r eason since, as seen in §4.1.1, small fruits are feminine in Hewramî. Alternativ ely , it is possible that Arabic jawz was b orr owed from Middle Iranian, which would explain the feminine marking of the native word wezî as b eing part of small fruits. Persian provides the primary source for modern words in T ekht Hewramî. Names denoting new objects are usually assigned a gender based on their ending. Thus, consonant-final nouns are masculine. (26) lêwan (m) ‘ cup ’ maşîn (m) ‘ car’ televizyon (m) ‘tele vision’ telefon (m) ‘telephone ’ mebal, mubayl (m) ‘mobile phone’ cîp (m) ‘jeep ’ sîmkart (m) ‘SIM card’ aɫbum (m) ‘album’ vaksen (m) ‘vaccine ’ mahware (m) ‘satellite ’ zilûbeya (m) ‘Zalabiyeh’ desga (m) ‘machine ’ duçerxe (m) ‘bicycle ’ 94 4.1 Nominal inflection An exception is modern place names such as ‘restaurant’ and ‘hotel’ , which are borrow ed through regional languages. These items end in a consonant in Kurdish and Persian but appear with an added unstressed -e in Hewramî, thus feminine. This exception can be understood b ecause the names of places are generally feminine, including all the ones derived from the place suffix -ga (see §4.1.1). (27) ristûřane (f) ‘restaurant’ hutêle (f) ‘hotel’ sendelîye (f) ‘ chair’ 4.1.1.3 Functions of the base-final feminine morphemes The base-final feminine morphemes additionally expr ess the female counterpart of masculine nouns denoting occupations, and derive fruit names from the cor- responding fruit tree. These functions ar e discussed b elow . 4.1.1.3.1 Expressing the female counterpart of masculine nouns A feminine noun can express the female counterpart of masculine general nouns and occupational titles (see also Sadjadi 2019): (28) Masculine Feminine ‘baker’ nanpeç nanp eçe ‘ cook’ çaçker çaçkere ‘patient’ neweş ne weşe ‘physician’ duktir duktire ‘worker’ kar eker karekere ‘miller’ asawan asawane This list can be extended to some items in the kin domain: (29) ‘betrothed’ desgîran (m) desgîrane (f) 4.1.1.3.2 Deriving the fruit names The feminine suffixes -î and -ê  derive some fruit names from the corresponding fruit tree. This can be seen in the derivation of the ‘fig’ (f) from ‘fig tree ’ (see 30). Note, ho wever , that in most cases, both the tree and fruit names share the same gender, rendering it hard to determine the direction of derivation (see 31). (30) hencîr (m) ‘fig tree ’ hencîrî (f) ‘fig’ 95 4 Nouns and nominal morphology (31) sawî (f) ‘apple tree ’ sawî (f) ‘apple ’ çeqalê (f) ‘green apricot tree ’ çe qalê (f) ‘green apricot’ wezî (f) ‘walnut tree’ wezî (f) ‘walnut’ wamî (f) ‘almond tree’ wamî (f) ‘almond’ 4.1.1.4 Gender agreement So far , we hav e seen that phonological, semantic, and morphological criteria can predict the gender of a noun. In addition, the gender of a noun is expressed by formal variation on agreeing elements, or agreement targets. At the level of noun phrases, agreement in gender can be found on two levels: (i) agr eement between syntactically asso ciated wor ds (e .g., a noun and an adjective) (ii) the gender-sensitive allomorphy of certain suffixes, e .g., the definiteness suffix. Adjectives agr ee in gender with the nouns they modify b oth attributively and predicatively . In the following examples, the attributive adjective agr ees with the head nouns in gender (see below for gender agreement with predicative adjec- tives). (32) kuřê cuwanxas kuř-ê boy .m-indf cuwanxas-∅ good_lo oking-m ‘a good-lo oking boy’ [KŞ.68] (33) jenê xase jenî-ê wife.f-indf xas-e charming-f ‘a charming wife ’ [ JH.64] The definite suffix -eke takes on the gender of the singular noun it attaches to. When it appears on a masculine noun, it takes the forms -eke in the direct case and -ekey in the oblique case. With feminine nouns, the addition of suffix -ekê can results in the deletion of the base-final vo wel (see §4.1. 6 for other outcomes). (34) jenî ‘woman’ (f) kuř ‘boy’ (m) sg.dir jen(î)-ekê [ZQ .14] kuř-eke [ZQ .38] sg.obl jen(î)-ekê [BP .185] kuř-ekey [KŞ.31] With derived nouns, the use of the definite suffixes is triggered by the gender of the derived noun. (35) neweşî-eke ‘the illness’ < ne weşî (m) awîrga-( e)kê ‘the fireplace ’ < awîrga (f) 96 4.1 Nominal inflection At the clause level, agr eement targets for the categor y of gender are predica- tive adjectives, 3sg copula markers (3sg.m ꞊n / ꞊a ; 3sg.f ꞊ne ), and 3sg inflectional person suffixes in verbs derived from the past stem (3sg.m -∅ ; 3sg.f -e ). In ( 36), the predicative adjective , the participle, and the 3sg copula agree with the gender of dega ‘village ’ , which has b een relativised. (37) illustrates another example of gender agreement on predicative adjectives. (36) î dega toş vînî çoɫe bîyêne. î dem.pro x dega village.f to=ş 2sg=3sg:O vîn-î see.prs.ind -2sg:A çoɫ-e deserted-f bîyê=ne be.pst.ptcp.f=cop.3sg.f:S ‘This village, which you see , was deserte d. ’ [ JE.4] (37) řama dûrene . řa=ma road.f=1pl:PSR dûr-e=ne far-f=cop.3sg.f:S ‘W e have a long way [to go]. [Lit. Our way is far .]’ [BP .191] In verbs deriv e d fr om the past stem, the 3SG inflectional affix, m: -∅ , f: -e , agrees in gender with the subject of intransitive clauses and object of transitive clauses. (38) kinaçê wite. kinaçe  girl.f wit- e sleep.pst - 3sg.f:S ‘The girl slept. ’ (39) kuř wit. kuř boy .m wit- Ø sleep.pst - 3sg.m:S ‘The boy slept. ’ (40) hêɫeřûwenîşa kerde . hêɫeřûwenî ꞊şa fried_egg.f.sg.dir=3pl:A kerd- e do.pst - 3sg.f:O ‘They cooked fried eggs. ’ [HB.56] 97 4 Nouns and nominal morphology 4.1.2 Number Nouns mark numb er. T ekht Hewramî has tw o numb er values: singular and plural. As seen in §4.1, the two number values are e xpressed by different allomorphs depending on the base-final vowels ( see §4.1). All nouns, regardless of gender and count vs. mass distinction have an oblique plural in -( y )a . There are complications with mass nouns in direct plural as explained in §4.1.2.1. T able 4.4: Numb er inflection sg pl dir.m -∅ -ê dir.f -∅ obl.m -î -a obl.f -ê Number is a morphosyntactic feature in T ekht Hewramî. The r eason is that it is involved in agreement (Kibort & Corbett 2008). Numb er agreement occurs only when the nouns are in the direct case . There is no number agreement with a noun in the plural oblique. At the le vel of the noun phrase, the targets for numb er agreement are adjectives (41), classifiers (42), and quantifiers (43). (41) karê xerabê kar-ê thing-pl xerab-ê bad-pl ‘bad things’ (Khan & Mohammadirad 2024a: 439) (42) yerê danê hêɫê yerê three dan( e)-ê clf-dir.pl hêɫ( e)-ê egg.m-pl.dir ‘three eggs’ [ JH.81] (43) çinnê saɫê çinn( e)-ê some-pl saɫê year.f-dir.pl ‘several y ears’ [HB.68] At the clause level, the agr eeing elements are predicative adjectives (44) and adjective complement of the light verb (45)–(46); see §12.4 for the syntax of light verb constructions. 98 4.1 Nominal inflection (44) ême firê bênmê. ême 1pl fir( e)- ê a_lot-pl b-ên-mê be.prs- a ug-1pl:S ‘W e were a large number . ’ [BP .110] (45) ême keç-ê ne-b-îmê. ême 1pl keç- ê crooked-pl ne-b-îmê neg.sbjv-be.prs-1pl:S ‘May we not be crooked. ’ [DG.66] (46) ême zamdarê nekero . ême 1pl zamdar- ê wounded-pl ne-ker-o neg.sbjv-do.prs-3sg:A ‘He should not injure us. ’ [DG.64] In addition to nouns and adjectives, pronouns also carry the morphological expression of number ( se e 6.1). 4.1.2.1 Inherently plural nouns A number of words appear only in the plural. These are nouns denoting collective entities and entities that come in pairs. The singular of these nouns appears in the feminine gender . This then suggests that there is a semantic basis for the gender assignment of mass nouns. The nouns in this category end in unstressed -î and -ê . (47) mekî ‘salt’ hardî ‘flour’ xurmawî ‘ date ’ simêɫê ‘moustache ’ awî ‘water’ çeqalê ‘bitter almond’ pantoɫê ‘trousers’ winî ‘blood’ hengûrî ‘grape’ şiɫwarê ‘trousers’ mîjûyî ‘lentil’ şêɫanê ‘apricot’ birê ‘ eyebr ows’ qijê ‘hair’ heştaɫûyî ‘plum’ qijê ‘hair’ genmî ‘wheat’ tifɫe şêxanê ‘strawberr y’ wezî ‘walnut’ tifî ‘mulberr y’ hencîrî ‘fig’ yewê ‘barle y’ wamî ‘almond’ hêzmî ‘wood’ wezî ‘walnut’ Proof for the plural inflection of these nouns comes in the agreement patterns found in the noun phrase and within the clause. 99 4 Nouns and nominal morphology VI possessor in a predicative possessive construction (78) xway deseɫatê çaneş bîyen. xwa-î God.m-sg.obl deseɫat-ê power.m-indf çane=ş like_that=3sg:NC bîye=n be.pst.ptcp.m=cop.3sg.m:S ‘God had such a power . ’ [ZQ .42] (79) sîyaweħşî kuřêwiş bîy en. sîyaweħş-î pn-m.sg.obl kuř-êw=iş son-indf=3sg:NC bîye=n be.pst.pt cp.m=cop.3sg.m:S ‘Siyawehsh has had a son. ’ [SK.87] (80) jenekaşa lemeşa bîyew e. jen(î)-eka =şa wife-def.pl.obl=3pl:PSR leme=şa belly .f.sg.dir=3pl:NC bî-e=we be.pst -3sg.f:S=compl ‘Their wives got pregnant [lit. had b ellies] again . ’ [ME.210] (81) a zařoɫa jenîşa gerekene . a dem.dist zařoɫ(e)-a child-pl.obl jenî=şa wife.sg.f.dir=3pl:NC gerek-e=ne necessar y-f-cop.3sg.f:S ‘Those children ( of yours) ask for wives [lit. wife is necessar y to the children]. ’ [ME.13] VII subject-like argument in non-canonical subject constructions A subject-like argument in non-canonical subject constructions is differ- ent from the grammatical subject in these constructions in two r esp ects: (i) the subject-like argument is generally marked in the oblique case (see the examples below ), as oppose d to the dir e ct marking for the grammat- ical subject; (ii) the subject-like argument is indexed by clitic pronouns, whereas the grammatical subject is indexed by the copula or person suf- fixes (see §6.2 for discussion on the second point). 106 4.1 Nominal inflection (82) zařoɫam awraşan. zařoɫ(e)-a =m child.m-pl.obl=1sg:PSR awra=şa=n hungry=3pl:NC=cop.3sg.m:S ‘My children are hungry . ’ [SH.230] (83) xway îrayeşa. xwa-î God.m-sg.obl îraye=ş=a volition=3sg:NC=cop.3sg.m:S ‘[This] is God’s will. [Lit. God, his volition is.]’ [ZB.34] (84) dêđêş qînîş kewte . dêđê =ş stepmotherf.sg.obl=3sg:PSR qînî=ş rage.f.sg.dir=3sg:NC kewt-e fall.pst -3sg:S ‘His step-mother got into a rage. [lit. her rage fell.]’ [SK.15] (85) tateyt miqdarêş hoşa. tate-î =t father.m-sg.obl=2sg:PSR miqdarê=ş a_little=3sg:NC hoş=a memory.m=cop.3sg.m:S ‘Y our father remembers it a bit. [Lit. Y our father , his memor y is.]’ [BP .13] VIII The second coordinate noun in a co-ordinate d N P ( e .g., when standing alone) (86) minû hesenî min=û 1sg=and hesen-î pn-m.sg.obl ‘Hasan and I’ IX the copula complement in a prepositional non-verbal predicate The prepositional complement of a copula clause may be marke d by the oblique case. This is especially the case when the preposition is pêse ‘like, similar to ’ . In this kind of clauses, the clitic copula exceptionally moves on the preposition (see §12.2 for discussion). (87) to pêsenî emîrî. to 2sg pêse=nî like=cop.2sg:S emîr-î pn-sg.obl.m ‘Y ou are like Emir (to me). ’ 107 4 Nouns and nominal morphology (88) řisq pêsen miɫey . řisq rat pêse=n like=cop.3sg.m:S miɫe-î mouse-sg.obl.m ‘ A rat is like a mouse. ’ [PK.36] It should be note d that ther e are certain complications for deployment of case marking. For example , case marking of O-prs is sensitive to the status of nouns as definite and specific (see §11.2.2 for details). Generic Os are generally marked in the direct case (see 89). (89) min hiɫoşet miđew . min 1sg hiɫoşe =t cracked_wheat.f.sg.dir=2sg:R mi-đe-û ind -give.prs-1sg:A ‘I will give you cracked wheat. ’ [ JP.236] Similarly , there is some variation in the oblique marking of non-core argu- ments such as possessors in predicative possession construction, the subject-like argument of a non-canonical construction, etc. (see §11.2.3 for an ov er view ); see (90)–(91) for direct marking of these arguments. (90) yo sêfêş bîyêne. yo onem.sg.dir sêfê=ş potato.f.sg=3sg:NC bîyê=ne be.pst.ptcp.f=cop.3sg.f:S ‘One had potatoes. ’ [ JE.32] (91) padşa gerekşa padşa king.m.sg.dir gerek=ş=a necessar y=3sg:NC=cop,3sg.m ‘The king wanted...]. ’ [BM. 135] Another feature of He wramî that affe cts the case marking is the notion of tran- sitivity and how it is applied to light verb constructions. A s mentioned in Chapter §2, tense-sensitive alignment depends not on the transitivity of the clause in a se- mantic sense but on the lexical transitivity of individual verbs. Like wise, in light verb constructions, transitivity is determined not in a semantic sense but by the transitivity of the light verb. In terms of case marking, the subject of a transitive light verb should be assigned an oblique case in past transitive construction. In the following examples, the oblique marking on the subject is triggered by the transitivity of the light verb kerđ ‘ do ’ . 108 4.1 Nominal inflection (92) xuđay eta kerden. xuđa-î God.m- obl eta granting kerde=n do.pst.pt cp.m=cop.3sg.m:O ‘God granted him. ’ [DB.125] (93) jenê desiş kerd qisekerde y . jen(î)-ê woman.f-obl des=iş hand=3sg:A kerd-Ø do.pst -3sg.m:O qisekerdey talking ‘The woman started [lit. hand do] to talk. ’ [SH.184] There is no designated case to express the v ocative. Therefore , a vocative func- tion can be expressed in several ways. Singular nouns may be expr ess the voca- tive in a direct case. (94) táte | ‘Father!’ [ JH.22] éđa | ‘Mother!’ The vocative is expr essed via the oblique form for plural nouns. (95) hewramîya xwahafêz! hewramî-a Hewramî-pl.obl xwahafêz farew ell ‘O Hewramî [people]! Go odbye!’ The vocative particle ya , borrow e d fr om Arabic, may also express the vocative. (96) ya qaɫîçew silêmanî! ya v oc qaɫîçe-û carpet-ez.gen silêman-î pn-obl.m ‘O Solomon’s carpet!’ [DB.164] A similar strategy , sometimes employ ed in tales, is to use the particle ey to express the v ocative. (97) maço, ‘ ey kitêk e y çira ey bira!’ m-aç-o ind -say .prs-3sg:A ey v oc kitik cat.dim ey v oc çira lamp ey v oc bira brother ‘He said, ‘Oh Cat; oh Lamp; oh Brother!’’ [SH.111] Another strategy for expressing the v ocative is to use the bare noun in the direct case followed by the modifier gîyan ‘ dear’: (98) bira gîyan ‘Dear brother! (addressing the neighbour ) [RE.5] 109 4 Nouns and nominal morphology The vocative can also be expressed by adding the diminutive suffix -le (m), - lê (f) to the kin terms. Here , more emotional engagement is involv e d with the interlocutor . Notably , the diminutive suffix is frozen in wor ds referring to ‘sister’ and ‘ child’ . In neighb ouring Southern Kur dish varieties, the term for ‘ child’ is the cognate form zarû . Similarly , the general term for ‘sister’ in Kurdish, i.e., xoşk , has the frozen diminutive suffix -ik . (99) tate-le ‘Father!’ eđa-lê ‘Mother!’ mama-lê ‘Grandmother!’ waɫê ‘Sister!’ zawɫe ‘Child! < * zaro + ɫe ’ Alternatively , the definite suffix may express vocative case, mainly when used with kin terms ‘husband’ and ‘wife ’ . The definite suffix agrees with the referent in gender. (100) jenekê ‘Wife!’ [ZB.13] pîyake ‘Husband!’ [DB.221] 4.1.4 Classifiers and measure nouns Hewramî has sev eral linguistic items that exhibit characteristics of noun classi- fiers; that is, their choice depends on the semantics of the head nouns, and they may be use d anaphorically (Aikhenvald 2000). Classifiers make up a small w ord class in Hewramî, and a limited number of nouns have them. T able 4.6 lists a preliminary list of classifiers. They can be classified into sortal classifiers , which are used with count nouns, and mensural classifiers, used with mass nouns. The most common classifier is dane (literally ‘se ed’), which can be use d with various head nouns. In terms of their syntax, classifiers agree in number with the head noun. In (101), dane agrees with the head noun; in (102), it is used anaphor- ically to refer back to ‘ eggs’ in the pre vious discourse. (101) duwê danê dêwê duwê two dan( e)-ê clf-pl.dir dêw-ê ogre.m-pl.dir ‘two ogres’ [ JP.17 6] (102) a wextî danêşa sîyaw ker o. a dem.dist wext-î time.m-sg.obl dan( e)-ê=şa clf-indf=3pl:PSR sîyaw black ker-o do.prs.ind -3sg:A ‘The girl coloured one of them black. ’ [ JH.83] 110 4.1 Nominal inflection T able 4.6: Common classifiers T ype item Used with / for Sortal dane count nouns ser domestic animals, e.g., sheep, goat des clothes boɫ counting grap e hoşe counting a cluster of grape bine trees telane farmland Mensural peɫ mass nouns, e.g., “grain” Additionally , classifiers are optional and may or may not appear with the head noun. Compare the following e xamples. (103) yerê danê hêɫê maro . yerê three dan( e)-ê clf-pl.dir hêɫ( e)-ê egg.m-pl.dir m-ar-o ind -bring.prs-3sg:A ‘She [ went and] brought three eggs. ’ [ JH.81] (104) yerê hêɫêş ardê . yerê three hêɫ( e)-ê=ş egg.m-pl.dir=3sg:A ard-ê take.pst -3pl:O ‘She took three eggs. ’ [HS.82] The following classifiers have been identifie d in the fieldw ork. The list in T able 4.6 is not e xhaustive, and no claims of thoroughness can be made here . peɫ refers to ‘bugs and insects’ and grains en masse. (105) peɫê merekuř mê şarezûr . peɫ-ê clf-indf merekuř grasshopper m-ê ind -come.prs.3sg:S şarezûr pn ‘ A swarm of grasshoppers came to Sharazour . ’ [PM.1] (106) pîreke yawa law peɫêw xeley . pîr-eke old-def.m.dir yawa arrive .pst.3sg.m:S la-û to-ez.gen peɫ-êw clf-indf xele-î grain-m.sg.obl ‘The old man arrived at a pile of grains. ’ [HR.13] 111 4 Nouns and nominal morphology Ser is used for domestic animals like she ep and goats, especially when used as a sacrifice. (107) serê heywan sere biřo. ser-ê clf-indf heywan animal.m sere head.m biř-o cut.prs.ind -3sg:A ‘He butchered one animal. ’ [ JP.253] (108) ya mûsa serê heywanma de. ya v oc mûsa pn serê clf heywan=ma animalm.sg.dir=1pl:R de give .imp.2sg:A ‘Oh Moses! Give us an animal. ’ [MF .209] Boɫ and hoşe are measure nouns for one grape , and a cluster of grap e , respe c- tively . (109) boɫê hengûrî boɫ-ê clf-indf hengûrî grape ‘ one grape’ (110) duwê hoşê hengûrî duwê two hoş( e)-ê clf-pl.dir hengûrî grape.pl.dir ‘two clusters of grape ’ Other classifiers include bine , and teɫane . bine is used as a measure of unit for trees. teɫane is used as a measure of unit for farmland. (111) duwê binê wezî duwê two bin( e)-ê clf-pl.dir wezî walnut.pl.dir ‘two walnut trees’ (112) duwê teɫanê bencacnî duwê two teɫan( e)-ê clf-pl.dir bencacnî tomato.pl.dir ‘two pieces of tomato farm’ There are different measur e words for walnuts, reflecting their importance in the local economy . These are summarised b elow: 112 4.1 Nominal inflection (113) laye ‘a collection of five walnuts’ dese ‘a collection of ten walnuts’ nîme hezar ‘a colle ction of fiv e-hundred walnuts’ hezar ‘a collection of one-thousand walnuts’ Examples: (114) yerê desê wezî yere three des( e)-ê clf-pl.dir wezî walnut.pl.dir ‘thirty walnuts’ (115) desew layê wezî dese=û clf=and lay( e)-ê clf-indf wezî walnut.pl.dir ‘fifteen walnuts’ Among these, des may also be used to refer to ‘a unit of ten domestic animals’ . Other measure nouns include hît ‘pair’ , gez ‘a unit of measure equivalent to 72 cm of textile ’ , teẍar ‘a unit of weight equivalent to 120 kilograms’ , etc. (116) zemanû a anê hîtê paɫɛ bîyênêw çwar gezê parçe bîy en. zeman-û time.m-ez.gen a dem.dist anê dem.dist.3sg.f.sg.obl hît-ê pair-pl paɫɛ shoe.pl.dir bîyê=nê=û be.ptcp.pl=cop.3pl:S=and çwar four gez-ê ell.m-pl.dir parçe textile.m bîye=n be.pst.ptcp.m=cop.3sg.m:S ‘Then, let us say , at the time, [the w edding gift] consisted of (some) pairs of shoes [and] four gaz of textile . ’ [RE.11] 4.1.5 Bare nouns A bare noun can fulfil different functions in T ekht Hewramî. It may expr ess indefinite referents and definite referents alike ( see b elow ), and have singular or plural refer ence. Thus, its use radically differs from well-kno wn languages like English. In (117)–(118), the bare nouns meɫa and gaw expr ess singular nouns with non-specific referents. 113 4 Nouns and nominal morphology (117) meɫa mara mareş biřa peyş. meɫa mullah.m m-ar-a ind -bring.prs-3pl:A mare=ş marriage=3sg:O biř-a cut.prs.ind -3pl:A pey=ş for=3sg:R ‘They fetched a Mullah [and] married her (the girl) to him (the shepherd’s son). ’ [KŞ.88] (118) pêse gaw til bowe ... pêse as_if gaw bull.m til rolling b-o=we be.prs.sbjv-3sg:S=compl ‘ As if a bull had rolled down [into the valle y] ... ’ [ZP .79] Similarly , a bare noun can expr ess a noun with an indefinite spe cific refer ence which is known to the speaker . In the following example , the narrator talks about a specific woo dland he has seen while searching for a mountain shelter . (119) kelêwe dîmandîmê bê fire tûl bê; dare san bê. kel-êwe mountain-indf dîmandîm-ê wide-indf b-ê be.prs- a ug.3sg:S fire a_lot tûl long b-ê be.prs-a ug.3sg:S daresan woodland.m b-ê be.prs- a ug.3sg:S ‘It was a big mountain; it was very high; it was a woodland. ’ [ZQ .12] A bare noun can also be used in a generic sense to express indefinite non- specific plural items. The noun would have plural inflection in equivalent situa- tions in English. (120) be kune awîşa ardêne. be by kune clay_pot.m awî=şa water.f.sg.dir=3pl:A ardê=ne bring.pst.pt cp.f=cop.3sg.f:O ‘They used to fetch water using clay pots . ’ [ JE.16] (121) be hesere hêzmîşa ardênê pey zimsanî. be by hesere mule.f hêzmî=şa firew oo d.pl.dir=3pl:A ardê=nê bring.pst.pt cp.pl=cop.3pl:O pey for zimsan-î winter.m-sg.obl ‘They fetched firew oo d for the winter on mules . ’ [ JE.35] 114 4.1 Nominal inflection Finally , a bare noun may hav e a definite reading. This happens when the refer- ence to the noun has be en established both for the sp eaker and the listener ( see § 4.1.6). In the following e xcerpt, çawre ‘tent’ is first unmarked as a bare noun and has a generic plural reference . In the second mention, it takes the definite suffix -ekê . In the third mention, it has a definite refer ence but appears as a bare noun. (122) ca a wextî çawr e b o . çawrekêne fire ginawe kem. w extê wey owe , lawaw aman çawre peře bîye awî. ca afterwards a dem.dist wext-î time.m-sg.obl çawre tent.f b-o be.prs.ind -3sg:S çawre-(e)kê =ne tent.f-def.f.sg=post fire a_lot gin-a=we fall.prs.ind -3pl:S=compl kem little wext-ê time.m-indf wey e-o=we wake.prs.ind-3sg:S=compl lawaw flood.m ama=n come.pst.pt cp.m=cop.3sg.m:S çawre tent.f peř-e full-f bî-e be.pst.3sg.f:S awî water.f ‘Back then, there wer e tents. They tossed around [in bed] in the tent for a while. When the y finally woke up, the flood had come [and] the tent was filled with water . ’ [ZB.19-ZB.20] 4.1.6 Definiteness Nouns are marked for definiteness by the following suffixes: -eke ; -e . The defi- niteness marker -eke has distinct allomorphs depending on the gender , case, and number of the base to which it attaches: T able 4.7: Definiteness paradigm dir obl sg.m -eke -ekey sg.f -ekê -ekê pl -ekê -eka The following paradigms illustrate the inflection of the definite suffix -eke on consonant-final nouns. (123) kuř ‘boy , son’ (m) sg.dir kuř-eke [ZQ .38] sg.obl kuř-ekey [KŞ.31] 115 4 Nouns and nominal morphology (149) tisle ‘a prickly fetid plant’ cf. tis ‘silent fart’ şotîle ‘ euphorbia plant’ cf. şot ‘milk’ Related diminutive suffixes are -oɫe , and -île : (150) camoɫe ‘small pot’ ( < cam ‘pot’ + -oɫe ) 4.2.1.6 -ga/ -ge , -gê The derivation suffix -ga/ -ge, -gê derives place names from nouns denoting ob- jects. (151) awîrga ‘fire place ’ < awîr (m) ‘fire ’+ -ga ‘place’ seringa ‘pillow’ < serin (m) + -ga ‘place ’ The suffix has be come lexicalised in nouns that already denote a place name . In addition, it has be come lexicalised as part of the adverbial noun wêr ega ‘ evening’ and yagê ‘place ’ . (152) ewêge ‘there ’ < awê ‘that’ + - ge ‘ over there ’ yagê ‘place ’ < ya ‘place ’ + - gê ‘ over ther e’ cûwe ‘str eam’ < * cû ‘stream’ + -ge ‘place ’ dega ‘village ’ < de ‘village ’ + - ga ‘place ’ milega ‘ridge ’ < mile ‘neck?’ + - ga ‘place ’ 4.2.1.7 -gerî The derivational masculine suffix -gerî derives quality nouns with the sense of “association of a person or a group of pe ople with certain featur es” from nouns and adjectives. (153) gewregerî ‘nobleness’ ge wre ‘big’ hewramîgerî ‘Hewramîhood’ hewramî ‘Hewramî’ pîyagerî ‘manhood’ pîya ‘man’ padşagerî ‘king-hood’ padşa ‘king’ aẍegerî ‘lordship ’ aẍe ‘lord’ qeyẍagerî ‘ chieftain-hood’ qeyẍa ‘ chieftain’ 4.2.1.8 -wan The derivational masculine suffix -wan has the appro ximate meaning of “taking care of” . The feminine form is -wane . 122 4.2 Nominal w ord formation (154) asawan ‘miller’ (m) cf. asaw ‘mill’ asawane ‘miller’ (f) cf. asaw ‘mill’ neçîrewan ‘hunter’ cf. neçîre ‘hunt’ řoçewan ‘ one who fasts’ cf. řoçe ‘fasting’ 4.2.1.9 -yane The derivational masculine suffix -yane occurs in the sense of ‘vernacular asso- ciated with a region’ . (155) hewramî-yane ‘Hewramî v ernacular’ paweyane ‘Paw e vernacular’ 4.2.1.10 -gel The derivational suffix -gel is derived from the gele ‘her d, flo ck’ (m). The suffix adds a collective meaning to the base noun. This suffix is assumed to be the origin of the plural suffix -gel in the neighbouring CK varieties of the Sanandaj region and in Southern Kurdish. (156) gaw-gel ‘flock of cows’ pez-gel ‘flock of she ep ’ (157) şuwanew pezgelî! şuwane-û shepherd-ez.gen pezgel-î sheep_herd-m.sg.obl ‘Shepherd!’ [ ÇH.118] 4.2.2 Nominal compounding Compound nouns are primarily formed by juxtaposing the component parts. The component parts form a single word in terms of str ess marking, gender assign- ment, and agreement. As discussed in §4.1.1, the compound wor d hêɫeruwenî ‘fried eggs’ ( < hêɫe ‘ egg’ (m) + řuwenî ‘ oil’ (f) has feminine gender , which is determined by the non-final stress in řuwenî ‘ oil’ . The most frequent compounds are N-N and N- V compounds, where the V equals the verb ’s present stem. 123 4 Nouns and nominal morphology 4.2.2.1 N-N compounds (158) hêɫeřûwenî ‘fried eggs’ hêɫe ‘ egg’ + řûwenî ‘ oil’ yanekoɫê ‘(leaving) houses’ yane ‘house’ + koɫê ‘shoulder’ herbene ‘donke y keep er’ her ‘donke y’ + b ene ‘servant’ Less commonly , compound nouns are formed by adding a spatial particle to the semantic, phonological and morphosyntactic head noun. (159) wer des ‘ser vant’ cf. wer- ‘front’ + des ‘hand’ çêrxan ‘basement’ cf. çêr- ‘under’ + xan ‘house ’ serxan ‘top floor in a house’ cf. ser- ‘up ’ + xan ‘house’ 4.2.2.2 N- V compounds N- V compounds feature the present stem of the verb in the comp ound. The most common verb stem used is -ker ‘ do ’ , which derives nouns denoting o ccupations from nouns which either denote the object manufacture or a pr o cess. (160) wêjen-ker ‘gum producer’ (m) cf. wêjen ‘T ragacanth gum’ çaşt-ker ‘ cook’ (m) cf. çaştî ‘foo d’ selem-ker ‘pre-seller’ (m) cf. selem ‘pre-selling’ kar-ker , kareker ‘worker’ (m) cf. kar ‘work’ coɫa-ker ‘weav er’ cf. coɫa ‘weaving’ The feminine counterpart of the above names is derived by adding the un- stressed -e to the compound noun: (161) çaştker-e ‘ cook’ (f) karker-e, kar eker-e ‘worker’ (f) Other verb stems are found marginally in N- V compounds: (162) çareniwîs ‘fortune-teller’ cf. çare ‘ destiny’ + niwîs- ‘write ’ hermanber ‘domestic w orker’ cf. herman ‘w ork’ + b er- ‘take ’ nanpeç ‘baker’ cf. nan ‘bread’ + peç- ‘bake’ xuɫekêş ‘soil carrier’ cf. xuɫe ‘soil’ + kêş- ‘pull’ There are also a fe w fauna terms which seem to b e formed by reduced clauses similar to the English flower term forget-me-not . The term for cricket insect is cêř cêř kere , containing cêř ‘ chirping’ , kere ‘ do ’ , yielding ‘ do es chirping’ . The -e element on the pr esent stem of ‘ do ’ seems to be a nominaliser . Alternatively , these terms can be analyse d as N- V compounds with the final -e marking the feminine gender . Similarly , the term for ‘mantis’ is coɫakere , and heɫîzewere r efers to a type of grasshopper: 124 4.2 Nominal w ord formation (163) coɫakere < coɫa ‘weaving’ + ker e ‘ do’ ‘mantis’ [lit. ‘weaver’] heɫîzewer e < heɫîze ‘ churn’ + were ‘ eating’ ‘a typ e of grasshopper’ [lit. ‘churn eater’] 4.2.2.3 N- ADJ compounds The compound nouns in this categor y consists of lexicalised items for e xpressing concepts such as ‘ old man’ , ‘old woman’ , etc. The positioning of the adjective may vary dep ending on the compound. (164) řîşçerme ‘ old man [lit. white b eard]’ cf. řîş ‘beard’ + çerme ‘white ’ pîrejenî ‘ old woman’ cf. pîr ‘ old’ + -e + jenî ‘woman’ 4.2.2.4 Echo compounds Echo compounds are a feature of regional languages such as Kur dish, Persian and T urkish. The y are formed by a partial reduplication strategy in which the initial consonant in the base word is r eplaced by m . The reduplicant can b e juxtaposed to the base or preceded by the co ordinate particle ꞊û ‘and’ . (165) jen꞊û menî ( < jenî û menî ) ‘women and so on’ [BP .49] qise=û mise ‘gossip and such’ [RE.23] çîw mîw ‘things and stuff’ 125 5 Noun phrase 5.1 Noun phrase properties A noun phrase can consist of a noun, proper or generic (1), a noun and optional modifiers (2), an independent pronoun (3), a substantivised adje ctive (4), or a verbal noun (5). (1) lawaw aman çawre peře bîye awî. lawaw flood.m ama=n come.pst.pt cp.m=cop.3sg.m:S çawre tent.f peř-e full-f bî-e be.pst-3sg.f:S awî water.f ‘The flood had come [and] the tent was filled with water . ’ [ZB.20] (2) î duwe dêw e î dem.pro x duwe two dêw=e ogre.m=dem ‘these two ogres’ [ JP.178] (3) dey min şûyş kerû pene. dey disc.pt cl min 1sg şû-î=ş husband.m-sg.obl=3sg:R ker-û do.prs.ind -1sg:A pene to ‘I will marry him [ lit. I will make husband to him]. ’ [ JH.59] (4) cuwanê bê pêsew î girdû cuwana. cuwan -ê youth.m-indf b-ê be.prs-a ug.3sg:S pêse-û like-ez.gen î dem.pro x gird-û all-ez.gen cuwan-a youth.m-pl.obl ‘He was a young man like all the young men. ’ [ZQ .54] (5) bizekê cîya leweřyayşa hurpiřa. bize-( e)kê goat.f-def.pl.dir cîya instead_of leweřyay =şa graze.inf=3pl:PSR hur-piř-a pvb-jump.prs.ind -3pl:S ‘The goats wer e dancing instead of grazing. ’ [ JP.55] 5 N oun phrase A noun phrase may also consist simply of a numeral. In such cases, the bare form of the numeral is used. (6) ř owê ce řowa yuw e mê yoy ç mê xizmetû şê ʕeladînî. řo-ê day.m-indf ce from řo-a day.m-pl.obl yuwe one.f m-ê ind -come.prs.3sg:S yo =îç one.m=add m-ê ind -come.prs.3sg:S xizmet-û service.m-ez.gen şê sheikh.m ʕeladîn-î pn-m.sg.obl ‘Once [lit. One day of days], a woman and a man [ lit. One (f) came , one (m)] came into the service of Sheikh Aladin. ’ [ZB.1] (7) duwê yerê mênê pane . duwê two yerê three m-ê-nê ind -come.prs-3pl:S p=a=ne at=dem.dist=post ‘T wo or three [men] went [lit. came] there [to Biyare]. ’ [ŞC.9] 5.2 Modifier s in the noun phrase Modifiers in the noun phrase are of four general types: demonstrative deter- miners (8), quantifiers (including numerals) (9), adjectives (10), possessors (11), and relative clauses embe dded with NP (12); see §14.1 for the syntax of relative clauses). These modifiers differ in their placement within the NP and whether or not they attach to the head noun via a linker ( see §5.5). (8) î kuře î dem.pro x kuř=e boy.m=dem ‘this boy’ [ZQ .44] (9) beʕzê hêzimê beʕzê some hêzm(î)-ê firew oo d.f-pl.dir ‘some firew oo d’ [ZP .13] (10) karî naşerʕî kar-î task.m-ez.a ttr naşerʕî unlawful ‘unlawful acts’ [BP .62] 128 5.3 Linkers in the noun phrase structur e (11) jenû şuwaney jen(î)-û woman.f-ez.gen şuwane-î shepherd.m-sg.obl ‘the shepherd’s wife ’ [KŞ.98] (12) ane ke berdma şiş mangê menn. ane dem.dist.m.3sg.dir [ke rel berd-∅=ma] take.pst -3sg.m:O=1pl:A şiş six mang( e)-ê month.f-pl.dir menn-Ø remain.pst -3sg.m:S ‘The one whom we took [with us] liv ed for six months. ’ [ZQ .29] 5.3 Linkers in the noun phrase structure Nominal heads are linked to their modifiers through a head-linking forma- tive , generally referred to as “ ezafe (or izafe)” within Iranian linguistics. T ekht Hewramî uses two ezafe linkers: -û and -î . Their use is essentially dependent on the type of mo difier that follo ws the head noun: -û is use d with locational nouns (§10.1.1.2) and genitive nouns ( se e below ), whereas -î is used with adjective mod- ifiers. On this basis, MacK enzie (1961: 82) calls the former “genitive ezafe’’ and the latter “ epithetic ezafe ” . I retain MacKenzie ’s genitiv e ezafe but use ‘attribu- tive ezafe’ for his epithetic ezafe . These lab els largely predict the use of -û and -î , but as will be se en below , there are cases where -î is used with nominal and pronominal possessors. 5.3.1 Genitive ezafe The presence of genitiv e ezafe is generally present on the head noun, but its presence can be conditioned by the morphophonemic processes resulting from its attachment to some vo wel-final bases, the case of the head noun, or (less so) semantics. The genitive ezafe -û occurs following consonant-final bases. (13) kuřû şuwaney kuř-û son.m-ez.gen şuwane-î shepherd.m-sg.obl ‘shepherd’s son’ [KŞ.51] When added to vowel-final bases, v owel hiatus is r esolved through different strategies. In nouns marked for the direct case, the genitive ezafe -û is realised 129 5 N oun phrase differently depending on the final vo wel of the base noun. It surfaces as a glide [ w] following vow el-final masculine nouns (with final str ess) and feminine nouns ending in -ê or -á . -û → [ w] / 𝑉 _ (wher e V  is a stressed vow el or -ê ) (14) kuřew şuwane y kuř-e-û son.m-def-ez.gen şuwane-î shepherd.m-sg.obl ‘the shepherd’s son’ [KŞ.34] (15) zemaw patşay zema-û son_in_law-ez.gen patşa-î king-m.sg.obl ‘the king’s son-in-law’ [ED .275] (16) kinaçêw pađşayç maço kinaçê-û girl.f.sg-ez.gen pađşa-î=ç king-m.sg.obl=add m-aç-o ind -say.prs-3sg:A ‘the king’s daughter said’ [KŞ.40] In contrast, the genitive ezafe -û triggers the deletion of the final vo wel in unmarked feminine nouns ending in unstressed -e or -î . The rule is summarised below . 𝑉 → ∅ / _-û (wher e V = [e , î] and is unstressed) (17) jenû ħemey ẍeybî hamîlê bo. jen(î)-û wife-ez.gen ħeme-î pn-m.sg.obl/ez.a ttr ẍeybî invisible hamîlê pregnant.f b-o be.prs.ind -3sg:S ‘Hama the Invisible ’s wife was pregnant. ’ [BP .205] (18) îne dastanû mina. îne dem.pro x.m.3sg.dir dastan(e)-û story.f-ez.gen min=a 1sg=cop.3sg.m:S ‘This is my story . ’ [ JM.12] On the other hand, adding the genitive ezafe to a noun – which, accor ding to the alignment pattern, should appear in the oblique case – may lead to morpho- logical competition b etw e en the genitiv e ezafe and the oblique case. This occurs primarily with masculine and feminine nouns in the singular oblique. In compe- tition for the post-nominal slot, the genitive ezafe wins, leading to the nominal 130 5.3 Linkers in the noun phrase structur e base being expressed in the unmarked direct case. In (19), the preposition com- plement should appear in the oblique case, hence şuwane-y [shepher d-m.sg.obl]. Y et, the oblique case is dropped b efore the genitive ezafe . Note that this cannot be explained by a phonological rule, as there there is no rule deleting <y> before <û>. In (20), the complement of the preposition should appear in the oblique case, hence derdê [benefit.f.sg.obl]; however , the oblique case is not compatible with the ezafe suffix, so instead the unmarked noun is used. (19) ama la şuwanew gawa. ama come.pst.3sg:S la to şuwane-û shepherd.m.sg.dir-ez.gen gaw(e)-a cow .f-pl.obl ‘He came to the cowherd. ’ [ÇH.108] (20) be derdû min be to derd( e)-û benefit.f-ez.gen min 1sg ‘to my benefit’ [ÇK.106] On the other hand, the plural oblique -a and the genitive ezafe are compatible , and the latter follows the former on the head noun. (21) dewayş kerđ çemaw kinaçêw patşay deway=ş medicine.m=3sg:A kerđ do.pst çem-a-û eye-pl.obl-ez.gen kinaçê-û daughter-ez.gen patşa-î king-m.sg.obl ‘He put medicine into the king’s daughter’s eyes. ’ [DB.312] (22) luwe la î dosaw tateyte . lu-e go.imp -2sg:S la to î dem.pro x dos-a-û friend-pl.obl-ez.gen tate-î=t=e father-m.sg.obl=2sg:PSR=deic ‘Go to your father’s friends. ’ The ezafe linker is sometimes absent. This absence may be semantically mo- tivated, meaning the if possessor gets to b e used frequently with a spe cific pos- sessed noun, a lexicalised reading of the NP may start to take shape, trigger the absence of the genitive ezafe . 131 8 V erb system: stems and major derivational processes T able 8.6: Common light verbs in light verb constructions l v l v c Gloss kerđey ‘ do, make ’ weş kerđey ‘to make; to build; to cure ’ (lit. ‘well to do ’) zamdar kerđey ‘to injure ’ (lit. ‘wounded to do’) wiş kerđey ‘to inform’ (lit. ‘memory to do’) des kerđey ‘to start’ ( lit. ‘hand to do ’) bîyey ‘be ’ řed bîyey ‘to cross’ (lit. ‘cr ossing to b ecome ’) peyđa bîyey ‘to appear , to b e born’ (lit. ‘visible to b e ’) weş bîye y ‘get healed’ ( lit. ‘well to be ’) nizîk bîyey ‘to approach’ (lit. close to b e ’) kewtey ‘fall’ neweş ke wtey ‘to get ill’ (lit. ‘ill to fall’) hetîm kewtey ‘to be left an orphan’ ( lit. ‘ orphan to fall’) pek kewtey ‘to be worried’ (lit. strength to fall’) paɫ kewtey ‘to lean, to lie down’ (lit. ‘side to fall’) nizîk kewtey ‘to get close to ’ (lit. ‘close to fall’) warđey ‘ eat’ qesem warđey ‘to swear an oath’ ( lit. ‘ oath to eat’) gîr warđey ‘to get stuck’ ( lit. obstacle to eat’) derđ warđey ‘to b e of use ’ (lit. ‘pain to eat’) day ‘give ’ tefre day ‘to avoid’ ( lit. ‘avoidance to giv e’) biř day ‘to cover a distance ’ (lit. ‘piece to give ’) amay ‘ come ’ b er amay ‘to rise ’ (lit. ‘out to come ’) tûş amay ‘to run into, to get into trouble ’ (lit. ‘accident to come’) biřyey ‘ cut’ mare biřyey ‘to marr y’ (lit. ‘marriage p ortion to cut’) sere biřyey ‘to behead’ ( lit. ‘head to cut’) gêrtey ‘grab ’ desû dîm gêrtey ‘to perform a Muslim prayer ritual’ (lit. ‘hand and face to take’) nîyay ‘put’ namê nîyay ‘to name ’ (lit. ‘name to put’) 234 8.2 Light v erb constructions T able 8.7: The lexical categor y of non-v erbal elements in LV Cs Category lv c Gloss Noun ʕefwe kerđey ‘to pardon’ ( lit. ‘par don to do’) wey kerđe y ‘to raise ’ ( lit. ‘training to do ’) zînî kerđey ‘to saddle’ (lit. ‘saddle to do ’) nima kerđey ‘to pray’ (lit. ‘pray to do ’) Adjective adiz kerđey ‘to upset’ (lit. ‘upset to do’) keç kerđey ‘to paralyse’ (lit. ‘cr ooke d to do ’) dagîr kerđey ‘to occupy’ (lit. ‘occupied to do ’) řas kerđey ‘to carry out’ ( lit. ‘right to do ’) aşkira kerđey ‘to disclose’ (lit. ‘visible to do’) Noun/Adj řed bîye y ‘to cross’ (lit. ‘ crossing to do ’) isɫah kerđey ‘to amend’ (lit. ‘amendment to do ’) haɫî bîyey ‘to understand’ (lit. ‘understoo d to be ’) cemʕ kerđey ‘to gather’ ( lit. ‘addition to do ’) swar kerđey ‘to mount’ ( lit. ‘rider to do ’) bergozar kerđey ‘to hold’ (lit. ‘accomplished to do’) Particle ber amay ‘to rise ’ (lit. ‘out to come ’) ber arđey ‘to take out’ (lit. ‘out to take ’) ber kerđey ‘to expel ( lit. ‘ out to do ’) wer day ‘to release ’ ( lit. ‘ out to giv e’) light verbs; the number of LV Cs with a particle as the non-verbal complement is down to a few cases listed in T able 8.7. LV Cs may also be categorised according to the transitivity of the light verb el- ement. The light verbs kerđey and warđe y are transitive , whereas bîyey , ke wtey , and amay are intransitive . In general, the transitivity of the LV C can be deter- mined by the light verb, hence hiş kerđey ‘to inform’ ( < hiş ‘intelligence ’ + kerđey ‘ do ’) is transitive, and hetîm ke wtey ‘to b e left an orphan’ ( < hetîm ‘ orphan’ + kewtey ‘to fall’) is intransitiv e. Therefor e, the transitivity in LV Cs is determined based on the lexical transitivity of the light verb , not the semantic transitivity . The LV Cs in (24) ar e syntactically transitive, ev en though some may be semanti- cally considered intransitive . 235 8 V erb system: stems and major derivational processes (24) fewt kerđey ‘to pass away’ (lit. ‘ death to make ’) zîya kerđey ‘to increase ’ ( lit. ‘addition to make ’) kem kerđey ‘to de cr ease ’ ( lit. ‘little to make ’) derđ warđey ‘to b e of use ’ (lit. ‘pain to eat’) gîr warđey ‘to get stuck’ ( lit. ‘hook to eat’) tefre day ‘to avoid’ (lit. ‘ evasion to give ’) Light verb constructions can also be classified according to the position of the complement relativ e to the light verb. While in the majority of the cases, the nominal element precedes the light verb, as seen above, in the follo wing constructions, the nominal complement consistently follows the verb . The light verb in these LV Cs is often a mobility verb . The nominal element, thus, can b e said to be the goal argument of the LV . (25) ginay řa ‘set off on the road’ luway řa ‘walk’ wistey řa ‘carry out’ amay cuwab ‘start to sp eak’ nîştey leme ‘get pregnant’ astey cîya ‘leave behind’ menay cîya ‘b e left behind’ yaway sinʕe ‘reach adulthoo d’ [BP .123] 8.3 Particle verbs In addition to simple verbs and Light verb constructions, ther e are a number of verb forms with figurativ e meanings comp osed of a frozen pr eposition and a simple verb . The particle verbs listed in T able 8.8 are different fr om the particle- based LV Cs in that they contain frozen pr epositions in their structure and the fact that the frozen preposition follows the verb. The le xical transitivity of the simple verb determines the transitivity of such verbs. 8.4 V alency changing morpholog y Causative and passiv e morphology affe ct the valency of the v erbs. They are nei- ther regular nor pr o ductiv e. They shar e the commonality of taking the present stem as the basis for morphological derivation. 236 8.4 V alency changing morphology T able 8.8: The preposition type i n particle v erb constructions Adposition Particle verb Gloss pene ‘to ’ yaway pene (intr) ‘to grow up ’ (lit. ‘arrive to ’) kewtey pene (intr) ‘receive(?)’ (lit. ‘fall to ’) zanay pene (tr) ‘learn about’ ( lit. ‘know to ’) pore ‘at’ dîyay pore (tr) ‘to lo ok at’ kêşay pore (tr) ‘to hit’ (lit. ‘hit at’) kewtey pore (intr) ‘stumble ’ (lit. ‘fall at’) wene ‘at’ qomyay wene (intr) ‘to happen’ ( lit. ‘accident at’) xuřyey wene (tr) ‘to shout’ (lit. ‘shout at’) day wene (tr) ‘to set off’ (lit. ‘give at’) wer ‘ out’ day wer (tr) ‘to herd out’ (lit. ‘give out’) yaray wer (intr) ‘to cope with’ (lit. ‘ dare at’) weɫê ‘front’ wiste y weɫê (tr) ‘to drive forth’ ( lit. ‘thr ow front’) kewtey w eɫê (intr) ‘move in front of’ (lit. ‘fall front’) gel ‘with’ kewtey gel (intr) ‘accompany’ ( lit. ‘fall with’) pêwer e ‘together’ kewtey pê were (intr) ‘run into each other’ ( lit. ‘fall with’) 8.4.1 Causative voice As discussed, the present stem is the basis for deriving causativ e stems. The causative voice is expressed primarily via suffixation, in which case the suffixes -n (prs) and -na (pst) attach to the present stem of some intransitive v erbs and de- rive transitive counterparts ( see b elo w for other alternations). Following Haspel- math (1993b), this kind of alternation is called “causativ e alternation”, meaning that the inchoative is the basic stem and the causative is deriv ed from it. Ex- amples (27-(28) exhibit the causative alternation for the v erb topay ‘ die’; topnay ‘kill’: 237 8 V erb system: stems and major derivational processes (26) inf Intransitive Gloss Causative Gloss prs pst prs pst sotey soç - sot- ‘to burn’ soçn- so çna- ‘to burn’ yaway yaw - yawa - ‘to arrive ’ yawn - yawna - ‘ cause to arrive ’ êşay êş - êşa - ‘to hurt’ êşn- êşna - ‘to hurt’ topay top - topa - ‘ die ’ 4 topn - topna - ‘kill’ cimay cim - cima - ‘move ’ cimn - cimna - ‘move ’ (27) dêwe çane topo miro . dêw-e ogre-def ç=a=ne in=dem.dist=post top-o die.prs.ind -3sg:S mir-o die.prs.ind -3sg:S ‘The ogre died there . ’ [ JP.202] (28) topneş! top n -e=ş kill.prs-2sg:imp:A =3sg:O ‘Kill it (the snake)!’ [MR.42] The causative morphology in general only derives a transitiv e stem from an intransitive one . Howev er , in one case, the causative affix is added to a transitive stem to yield a change in meaning. From (29) it can be assumed that the causative suffix would driv e a semantically related meaning when adde d to a transitiv e verb . (29) bexşay ‘to forgive ’ b exşnay ‘to distribute ’ prs bexş - bexşn- pst bexşa- bexşna- In a few verb pairs, T ekht H. has preserved an older pattern of umlaut for the formation of causative stems, attested since the Middle Iranian period ( Skiær vø 2009: 220), e.g., Middle Persian: ahram ‘go up ’ vs. ahrām ‘lead up ’ (tr). This pattern is attested in the verb ‘to break’ , where the intransitive/inchoativ e verb appears with the detransitiviser suffix -ye (prs); -ya (pst). The causative counter- part is not only short of the intransitive suffix but is featur ed by a change in the vow el of the stem. 4 Hewramî makes a distinction between the verb used when a human dies and when an ani- mal dies. For the former , merđey is used, while for the latter topay is used. Likewise, for the causative counterpart, the verbs used are different: topnay ‘kill an animal’ vs. kuşte y ‘kill a hu- man’ . It should be noted that in some contexts, e .g., when dehumanising a person, it is p ossible to use the verb used for animals to refer to when a person dies, or is killed. 238 8.4 V alency changing morphology (30) meřyay ‘to break’ (intr) mařay ‘to break’ (tr) prs meřye- mař- pst meřya- mařa- A similar pattern is attested for the verb ‘to pour’ . Here, the vo wel is not changed, but the stem used in the past causative is different. (31) micyay ‘to pour’ (intr) mitey ‘to pour’ (tr) prs micye- mic- pst micya- mit- In both cases, it is the intransitive/inchoative stem, rather than the transitive stem, that seves as the base for the formation of the passive stem, . This points to the identical stem morphology of passives and inchoatives. T able 8.9 highlights the morphological alignment of passive and inchoative stem morphology for verbs ‘br eak’ and ‘p our’ . The past stems have been given for ease of comparison. The identical morphology concerns the use of the detransitivising suffix in the formation of both the inchoative and the passive. T able 8.9: Identical morphology of passive and inchoative stems V erb tr . stem passive/inchoative stem ‘to break’ mařa- meř-ya ‘to pour’ mit- mic-ya The causative/transitivising suffix is also used with intransitive verbs e xpress- ing sound emission without necessarily increasing the verb ’s valency . Here , the transitivising suffix expresses the agentivity . When the causative/transitivising suffix is present on a verb of sound emission, the v erb is treated as a transitive verb and is indexed via a clitic pronoun in the past tense . Without a causative suffix, the verb is tr eate d as an intransitiv e verb and is indexed via verbal person suffixes. Consider the difference between (32) and (33). (32) heşekê boɫnaş. heşe-( e)kê bear-def.f.sg boɫna=ş growl.pst=3sg:A ‘The bear growled. ’ 239 8 V erb system: stems and major derivational processes (33) tûteke gefa. tûte-( e)ke dog-def.m.sg.dir gef a-Ø bark.pst -3sg:S ‘The dog barked. ’ Note that there is variation in the coding of the verbs of sound emission, such that the same verb may be code d as transitive (in which case it appears with the agentive -n ) or intransitive; see for example the variation for encoding ‘it (the donkey ) brayed’ in T able 8.10. Finally , sev eral other alternations occur for derivation of causative from the inchoative or vice v ersa. Most of the verb pairs exhibit what is known as ‘ equipol- lent alternation’ , meaning that b oth the causativ e and the inchoative are derived from the same stem, but the affixes used are different. Denominal v erbs in §8.1.3 feature this alternation. (34) pijg-ye ‘scatter’ (intr) pijg-in ‘sow , to scatter’ (tr) çik-ye ‘ drip ’ (intr) çik-n ‘suckle ’ (tr) xinîk-ye ‘suffocate ’ (intr) xinîk-n ‘strangle ’ (tr) temam-ye ‘finish’ (intr) temam-n ‘finish’ (tr) nam-ye ‘bend’ (intr) nam-n ‘bend’ (tr) dêw-ye ‘become angr y’ (intr) dêw-n ‘get angry’ (tr) gîs-ye=( e)ne ‘shine ’ (intr) gîs-n=ene ‘light, start’ (tr) The following examples illustrate the alternation for temam-y e vs. temam-n ‘finish’ . 5 This verb may also be expressed as a light verb construction: lûr e꞊ş kerđ ‘It howled. ’ 240 8.4 V alency changing morphology T able 8.10: Transitivity in verbs of sound emission ‘bark’ gefa (intr) ‘It ( dog) barked. ’ ‘bleat’ barya-we (intr) ‘It ( she ep , goat) bleated. ’ ‘moo ’ qoř ya-we (intr) ‘It (co w) mooed. ’ ‘neigh’ hîlna꞊ş (tr) ‘It ( horse , mule) neighed. ’ ‘bray’ seřa (intr); seřna꞊ ş (tr) ‘It ( donkey ) brayed. ’ ‘howl’ lûrna꞊ş 5 (tr) ‘It ( wolf) howled. ’ ‘roar’ ‘It (le opard) r oared. ’ ‘howl’ qařna=ş (tr) ‘It ( jackal) howled. ’ ‘gecker’ wêqna꞊ş (tr) ‘It (fox) geckered. ’ ‘It (stone marten) geckered’ ‘grunt’ ‘It (squirr el) grunted. ’ ‘growl’ boɫna꞊ş (tr) ‘It ( bear) gr owled. ’ ‘ cluck’ qirazna꞊ş (tr) ‘It (hen) clucke d. ’ ‘roar’ neřna꞊ş (tr) ‘It ( lion) r oared. ’ ‘ caw’ qiřna꞊ş (tr) ‘It (raven) cawed. ’ ‘ quack’ ‘It (duck) quacked. ’ ‘grunt’ mîzna꞊ş (tr) ‘It (turtle) grunted. ’ ‘ chirp ’ cirîkna꞊ş (tr) ‘It (sparro w) chirped. ’ ‘squeak’ ‘It (mouse) clicked. ’ ‘meow’ mîyawna꞊ş (tr) ‘It ( cat) meowed. ’ ‘buzz’ wîzna꞊ş (tr) ‘It ( bee) buzze d. ’ ‘squeak’ qařya꞊we (intr) ‘It (rabbit) squeaked. ’ ‘ crow’ wena꞊ş (tr) ‘It (rooster , partridge) cro wed. ’ ‘ coo’ ‘It (do ve) coo ed. ’ ‘ croak’ ‘It (frog) croaked. ’ ‘hiss’ ‘It ( snake) hissed. ’ ‘hoot’ ‘It (o wl) hoote d. ’ 241 8 V erb system: stems and major derivational processes (35) awê her kunene her netemamyɛne. awê water.f.sg.obl her emph kune=ne clay_pot=post her emph ne- temamyɛ =ne neg-finish.pst.pt cp.f=cop.3sg.f:S ‘The water [ was] in the clay p ots. It w ould not finish [soon]. ’ [ JE.21] (36) feqîş temamnan. feqî=ş theologian.m=3sg:A temamna =n finish.pst.pt cp.m=cop.3sg.m:O ‘He finished [studying] Islamic jurisprudence. ’ [ZP .16] Suppletive noncausal/causal alternation is attested as well. Here , different verb roots are used: (37) mir- ‘ die ’ (intr) kuş- ‘kill’ (tr) gin- ‘fall’ (intr) wiz- ‘ drop ’ (tr) Some verb pairs exhibit the ‘anticausativ e alternation’ , meaning that the causative v erb is basic and the inchoative verb is derived from it: (38) fař ‘ change ’ (tr) fař - ye ‘ change ’ (intr) diř ‘tear’ (tr) diř-ye ‘tear’ (intr) yaw=we ‘spread’ (tr) yaw-ye=we ‘be spread’ (intr) pêç ‘wrap ’ (tr) pêç-ye ‘wrap (oneself)’ (intr) The following examples illustrate the alternation between yawye=we ‘be spread’ and yaw=we ‘spr ead’ . 242 8.4 V alency changing morphology (39) zeře çermekeş ana ça yawyano. zeř-e coin-ez.cmpd çerme-( e)ke=ş white-def.m.sg.dir=3sg:PSR ana-Ø loc.deic.cop -3sg.m:S ça there yawya =n=o be_spread.pst.pt cp.m=cop.3sg.m:S=compl ‘Her white coins were ther e, spread out ov er [the ground]. ’ [PK.27] (40) qeran qeran yawanşo. qeran kurus qeran kurus yawa =n=ş=o spread.pst.pt cp.m=cop.3sg.m:O=3sg:A =compl ‘It (the rat) had spread them (the coins) kurus by kurus. ’ [PK.30] Some verb pairs exhibit ‘labile alternation’ , in which the same verb is used both in the inchoative and causative: (41) toq- ‘scare, be scared’ (tr, intr) The following examples e xhibit the verb toqay (prs: toq- ; pst: to qa- ) ‘be scared, scare ’ being used intransitively (42) and transitively (43): (42) pîyakeyç lers ker o toqo. pîya-( e)ke=yç man-def.m.sg.dir=add lers shake ker-o do.prs.ind-3sg:A toq -o be_scared.prs.ind -3sg:S ‘The man trembled (in fear ); he was scared. ’ [ JF .29] (43) sîyamarêwî toqanaş. sîyamar-êw-î black_snake-indf-m.sg.obl toqa =na=ş scare .pst=cop.1sg:O=3sg:A ‘ A black snake scared me. ’ [ JL.68] Some verbs do not show any alternation. A subset of these v erbs exists only with the causative/agenttive suffix. Most verbs of sound emission seen in T able 8.10 belong to this categor y . (44) wişknay ‘to scour’ wişkna- wişkn- demnay ‘to start (morning)’ demna- demn- pirûnay ‘rub off (e ye)’ pirûna- pirûn- řaznay=we ‘adorn’ (tr) razna=we razn=w e Another subset occurs only in the inchoative form: 243 8 V erb system: stems and major derivational processes (66) mađam w eşeş kerdêbowe ... mađam as_long_as weş-e=ş well-f=3sg:A kerdê=b-o=we do.pst.pt cp.f=be.prs-3sg:O=compl ‘Now that it appears that he has healed her [thoroughly ] ... ’ [ JP .261] (67) dûr ginew e çî şarî! dûr far gín-e=we fall.prs.imp -2sg:S=compl çî from=dem.pro x şar-î city.m- obl.m ‘Get away from this town!’ [BP .165] That the particle is associated with completeness is evident in comparing the following two v erbs, one without the particle, i.e., tawyo , and one with it, i.e., tawyowe . The one without the particle implies that the ice is still in the pr ocess of melting without necessarily melting completely . The one with =we implies that the melting will surely be completed (68) y exeke taw y o. yex-eke ice-def.m.sg.dir tawy(e)-o melt.prs-3sg:S ‘The ice is melting. ’ (69) y exeke taw y owe . yex-eke ice-def.m.sg.dir tawy(e)-o=we melt.prs-3sg:S=compl ‘The ice is melting out. ’ The particle appears at the end of the verbal form, after all other suffixes and clitics: (7 0) kîyanoşe we. kîyan-o=ş=ewe send.prs.ind -3sg:A =3sg:O=compl ‘He sent him back. ’ [ JP.96] (71) milarew e. mi-l-a=re꞊we ind -go .prs-3pl:S=povb=compl ‘They went back. ’ [BP.124] 250 8.6 Infinitiv e 8.6 Infinitive The infinitive is formed by adding the suffix -ey to the past stem of the verbs. The vow el e gets deleted following vowel-final stems: (72) Gloss Infinitive Past stem ‘read, study’ wenay wena- ‘ do ’ kerđey kerđ- The infinitive has both nominal and verbal functions. In its nominal function, the infinitive occurs as a preposition complement (73), as a nominal complement (74), and as a nominal argument in existential or copula clauses (75). (73) şûnû peya bîyeyşr e eyakêş merđe. şûn-û after-ez.gen peya visible bîyey=ş=r e be.inf=3sg:PSR=post eđa-( e)kê=ş mother.f-def.f.sg=3sg:PSR merđ-e die.pst -3sg.f:S ‘W ell, the child was b orn. His mother died after his birth. ’ [KŞ.23] (74) xizone çêro ʕebakê be ʕinwanû witeywew seyaywe. xiz-one creep.prs.ind -3sg:S çêr=o under=post ʕeba-( e)kê robe.f-def.f.sg be by ʕinwan-û label.m-ez.gen witey=we=û sleep.inf=post=and seyay=we rest.inf=post ‘He put it over him. He crept under his robe, supposedly to rest and sleep [lit. in the guise of sle eping and r esting]. ’ [BP .188] (75) ew aman xeyr amayê w eş amayê b eynne bîy en. ew dem.dist ama=n come.pst.pt cp.m=cop.3sg.m:S xeyr goodness.m amay-ê come.inf-indf weş good amay-ê come.inf-indf beyn=ne between=post bîye=n be.pst.ptcp.m=cop.3sg.m:S ‘They [lit. He] would come. There w ere greetings and welcoming. ’ [RE.25] The infinitive also occurs as the complement of the verb des kerđey ‘start’ [lit. ‘hand do ’]. 251 8 V erb system: stems and major derivational processes (7 6) ser ew des kero gire way . serew from_above des hand.m ker-o do.prs.ind -3sg:A gireway cry.inf ‘He (Little Hama) started to cr y on the roof. ’ [BP .152] (77) şûnîre amana desim kerđen karêz biřyey . şûnî=re afterwards=post ama=na come.pst.pt cp.m=cop.1sg:S des=im hand.m=1sg:A kerđe=n do.pst.pt cp.m=cop.3sg.m:O karêz subterranean_canal.m biřyey cut.inf ‘Then, I started [ lit. hand do] to dig subterranean canals. ’ [ JM.6] In its verbal function, the infinitiv e is use d in purposive clauses follo wing a verb of movement. In most Iranian languages, a subjunctiv e v erb form is expecte d in this context. (78) kîyanaş pey beẍay wenay . kîyan-a=ş send.prs.ind -3pl:A =3sg:O pey to beẍa-î pn-m.sg.obl wenay read.inf ‘They sent him to Baghdad so that he studied. ’ [ JP.80] (79) berdênmêşa lo keney bê heq. berdê=nmê=şa take.pst.pt cp.pl=cop.1pl:O=3pl:A lo fodder_grass keney mow .inf bê without heq salary.m ‘They would take us to mo w fodder grass for free [i.e., without wages]. ’ [RE.65] 252 9 V erbal inflectional mor phology and verbal categ ories 9.1 V erbal inflection: tense-asp ect-moo d and person Present and past stems combine with inflectional person suffixes and modal pre- fixes to yield several tense-aspect-moo d categories. The indicative and subjunc- tive prefixes ar e among the T AM prefixes expressing mood. Howev er , the ma- jority of verbs have no overt marking of mood distinction, in which case the difference between indicative and subjunctive v erb form is carried out by stress; the verb stem in the subjunctive mood carries stress: kéro ‘that he/she does’ vs. kero ‘he/she does’ (see §3.3.4). T able 9.1 summarises T AM/negation prefixes: T able 9.1: T AM and negation prefixes T ense A spe ct-mood prefix Negativ e indicative m- mé- , nimé- Present subjunctive ∅ -, bi- ne- imperative bi- me- , ne- , nimé- perfective, imperfect, ∅ - ne- Past perfect, past ∅ - ne- perfect, past conditional ∅ - ne- As can be seen, there are tw o negation forms for the indicative. The form mé- originates from the deletion of n(í)- in n(í)-me through cluster r eduction, with the stress later shifting to mé- ( see Karim & Mohammadirad forthcoming). The resultant form mé- is thus originally an indicative pr efix which happene d to carry the stress of the negativ e form and later remorphologised as the negation marker . The indicative form m- is the result of pr etonic shortening of me- before stressed syllables: *me-z-ó > mi-zó ‘she gives birth’ . The variant me- is now mor- phologised as the negation marker and is evident in the negative form nime- . 9 V erbal infle ctional morphology and v erbal categories 9.1.1 Indicative m(i)- The unstressed indicative prefix mi-/m- expresses general present and futur e tense in verbs deriv e d fr om the present tense. It is likely r elated to the Old Iranian *hama-aiwa- ‘same duration, time ’ (Windfuhr 2009: 26). The indicative form has a full form me- which is reconstructable in the negative of indicative ( see ab ov e). m- is, howe ver , only prefixed to a few conjugated verbs. Ther e are thus two classes of present stem verbs r egarding the use of the prefix m- : 1 Class I consists of the majority of the verbs to which the indicative m- is not pr efixed; class II verbs take the indicativ e prefix m- . For class II verbs, phonological factors seem to be the main trigger for using the prefix m- : verb stems starting with a vow el take the indicative particle. Examples: (1) m-aç-û [ind-say .prs-1sg:A] ‘I say’ m-ar-û ‘I bring’ m-az-û ‘I let’ m-êş-û ‘I hurt (intr)’ m-êj-û ‘I am valued’ m-êjye-w-re ‘I lie do wn’ m-ûs-û ‘I sleep ’ m-e-w ‘I come ’ Similarly , the verb stems consisting of only a consonant take the indicative prefix. An exception is the verb bîy ey ‘be, be come ’ (prs b- ; pst bî- ), which does not take the prefix, seemingly due to assimilation in the bilabial feature . (2) mi-ş-î [ind-go.prs-2sg:S] ‘you go ’ mi-l-î ‘you go ’ mi-z-o ‘she gives birth’ For the rest of the v erbs, the syllable structure seems to b e the main factor triggering the occurrence of the indicative prefix. V erbs with a CV structure take the m- prefix. (3) mi-ge-w [ind-copulate.prs-1sg:A] ‘I copulate ’ mi-đe-w ‘I give ’ Similarly , verb stems with CCV structur e take the m- prefix. The second con- sonant in the stem of these verbs is a glide . The addition of the m- results in the resyllabification of CCV to CV C.CV . 1 See Mohammadirad & Karim (2025) for the development of indicativ e and subjunctive prefixes in conservative Hewramî varieties. 254 9.1 V erbal inflection: tense-aspe ct-mood and p erson (4) mi-đye-w [ind-lo ok.prs-1sg:S] ‘I lo ok’ mi-nye-w ‘I put’ mi-zye-w ‘I arrive ’ V erbs with the syllable structure CCV C split in taking the m- prefix. The addi- tion of indicative pr efix results in the resyllabification of the sequence whereby the verb ’s CCV C syllable structur e is restructured as CV C. CV C. (5) mi-nvîs-û [ind-write.prs-1sg:A] < cf. nvîs- ‘I write ’ mi-jnew-û < cf. jnew- ‘I hear’ mi-jnas-û < cf. jnas- ‘I know’ Exceptions include the following: (6) jmar-û [ count.prs-1sg:A] ‘I count’ vreş-û [ sell.prs-1sg:A] ‘I sell’ The majority of the stems based on the present tense do not take the indicative prefix. These verbs fall into different morphosemantic and phonological classes. Some generalisations can be made here: The indicative prefix is omitted before verbs with CV C structure. The initial consonant is usually a strong consonant. The reason could be that the m- prefix undergoes phonetic attrition in consonant clusters, e.g., *m-ker-o < kero ‘he/she does’ . (7) piř-o [fly .prs-3sg:S] ‘It flies’ ker-o ‘he/she does’ çin-o ‘he/she weaves/plucks’ saw-o ‘he/she rubs’ şan-o ‘he/she scatters/sows’ xiz-o ‘he/she slips’ xiw-o ‘he/she laughs’ doş-o ‘he/she milks’ zan-o ‘he/she knows’ jîw-o ‘he/she lives’ şêl-o ‘he/she rubs/presses’ wêç-o ‘he/she sieves’ An exception is the verb sanay ‘to buy’ , which for some speakers appears with the indicative m- : (8) sana / mi-sana ‘they buy’ 255 9 V erbal infle ctional morphology and v erbal categories Similarly , verb stems with CV CC structure do not take the m- prefix. (9) perm-o [rely_on.prs-3sg:A] ‘he/she relies on’ pers-o ‘he/she asks’ ters-o ‘he/she is afraid’ This category also includes verbs with the agentive suffix -n : (10) kiřn-o [scratch.prs-3sg:A] ‘he/she scratches’ wuřn-o ‘he/she destroys’ gêɫn-o ‘he/she narrates’ řemn-o ‘he/she makes run’ şokn-o ‘he/she shakes (tr)’ The indicative prefix does not occur with verbs of sound emission. This in- cludes the verbs that take the agentive suffix -n and those that do not. (11) pirxn-o [snore .prs-3sg:A] ‘he/she snores’ qêřn-o ‘he/she shouts’ seř-o ‘it brays’ girew-o ‘he/she cries’ bařî( e)-o꞊we ‘it bloats’ pijm-o ‘he/she sneezes’ Unaccusative v erbs with adde d detransitiviser suffix -y e do not take the indica- tive suffix. (12) meřy(e)-o [break.prs-3sg:S] ‘it br eaks’ şêwy(e)-o ‘he/she gets confused’ qomy( e)-o ‘it happ ens’ fařy(e)-o ‘it changes’ fîsy( e)-o ‘it overflows’ The same usage is seen with light verb constructions. The indicative prefix appears only with the light verbs fulfilling one of the conditions above: (13) tûşiş mê. tûş꞊iş accident꞊3sg:PSR m-ê ind-come.prs.3sg:S ‘He ran into him. ’ < tûş amay ‘to run into, to get into tr ouble ’ [ JP.132] vs. 256 9.1 V erbal inflection: tense-aspe ct-mood and p erson (14) tateş fewt kero . tate=ş father.m=3sg:PSR fewt death.m ker-o do.prs.ind -3sg:A ‘His father passed away . ’ < fewt kerđey ‘to pass away’ [ JP.8] 9.1.2 Subjunctive b(i)- In verbs derived from the pr esent stem, the subjunctive mood is expressed by the prefix b(i)- . The prefix bi- and its cognates mark the subjunctiv e and imperative moods in mo dern W est Iranian languages. The subjunctive was without any T AM marker in Middle Iranian, with a different inflection than the indicative mood. Only in the early Iranian perio d did b(i)- come to dev elop a T AM marker , which also functions as an imperative (Noorlander & Stilo 2015). The prefix is originally str esse d in line with other Iranian languages. Ho wever , for the majority of verbs which take the pr efix, the stress shifts to the stem, as shown by the phonemic analysis of bzánmê ‘that we kno w’ [RE.51], se e Figur e 9.1a. The stress shift is blocked before stems starting with a consonant cluster; see the phonemic analysis of bíjnasû ‘that I know’ in Figure 9.1b. (a) Unstr essed sbjv ‘that we know’ (b) Stressed sbjv ‘that I know’ Figure 9.1: The stress pattern pr efixe d subjunctiv e stems The subjunctive prefix occurs with a small class of verbs. The limited use of the subjunctive prefix parallels the r estricted use of the indicative prefix seen in §9.1.1. There are thus tw o main classes of verbs regarding the subjunctiv e prefix: class I verbs (making up around 95%) not taking the pr efix, and class II verbs (ar ound 5%) taking the prefix. A small set of verb stems may appear with or without the subjunctive prefix (see below ). As seen in Figures 9.1a and 9.1b, the subjunctive prefix may be stress-bearing dep ending on the syllable structur e of the stem. By contrast, in verb stems featuring no subjunctive pr efix, the stress has shifted onto the first vowel of the stem, resulting in the phonemic stress placement §3.3.4. Historically , the stress shifts onto the stem following the loss 257 9 V erbal infle ctional morphology and v erbal categories of the subjunctive prefix (Karim & Mohammadirad forthcoming). Figur e 9.2 ex- hibits the stress placement of the prefix-less subjunctive stem ber-o ‘that he takes’ [HB.11]. Figure 9.2: The stress pattern of pr efix-less subjunctive stem ber- ‘take’ The same morphophonological conditions as the indicative m- seem to trigger the use of the subjunctive prefix. Ther efore, subjunctive pr efixes appear with consonant-only and vow el-initial stems. An exception is the verb bîye y ‘be, be- come ’ (prs b- ; pst bî- ), which does not take the prefix. (15) b-û  s-û [sbjv-sle ep .prs-1sg:S] ‘that I sle ep ’ [BP.182] bí-l-o [sbjv-go.prs-3sg:S] ‘that he goes’ [DG.54] b-ár-o [sbjv-bring.prs-3sg:A] ‘that he brings’ [HB.14] b-ê  -nê [sbjv-come.prs-3pl:S] ‘that they come ’ [ JE.78] b-ó [b e.prs.sbjv -3sg:S] ‘that she be ’ [KŞ.17] The following is a sample of verbs which appear with the subjunctive pr efix in the text corpus: (16) bí-spar-o [sbjv-hand_over-3sg:A] ‘that he hands over’ [ŞC.89] bí-nye-ûne [sbjv-set_up .prs-1sg:A] ‘that I set up ’ [ JH.13] bi-sán-a [sbjv-buy .prs-3pl:A] ‘that they buy’ [BP .58] bí-zy-a [sbjv-grow .prs-3pl:S] ‘that they gro w’ [ JP .27] The following list is a sample of v erbs from the text corpus that appear with no subjunctive T AM prefix. (17) wér-o [ eat.prs.sbjv-3sg:A] ‘that they eat’ [PM.36] kér-o [ do.prs.sbjv-3sg:A] ‘that he does’ [ZP .111] bér-o [take.prs.sbjv -3sg:A] ‘that he takes’ [RE.52] kúş-î [kill.prs.sbjv-3sg:A] ‘that you kill’ [KŞ.98] 258 9.1 V erbal inflection: tense-aspe ct-mood and p erson sáw-î [rub .prs.sbjv -2sg:A] ‘that you rub ’ [DG.47] dón꞊mê [talk_to.prs.sbjv -1pl:A] ‘that we talk to ’ [ JE.7 6] táw-o [read.prs.sbjv -3sg:A] ‘that he/she can’ [ JM.61] The verb zanay ‘to know’ appears both with and without the subjunctive prefix, though the latter is mor e common: (18) bi-zan-mê ‘that we know’ [ JP. 64] zan-a ‘that they know’ [ JP .67] In light verb constructions, the subjunctiv e prefix is largely absent. The light verb day ‘give ’ e xhibits variation in taking the subjunctive prefix (see 19e-19f). (19) a. dagîr ker-o ‘that he occupies’ [DP .18] b . dawa ker-o ‘that he claims’ [DP .18] c. wey ker-û ‘that I take care of’ [ JP.226] d. peket gin-o ‘that you get worried’ [ JP.234] e. zer er de-yme ‘that we harm’ [PM.5] f. tefre bi-đ( e)-o ‘that he avoids’ [KŞ.105] 9.1.3 Imperative b- As stated, the prefix b- also expr esses the imperative moo d. Like marking the sub- junctive mood, the imp erativ e b- is not regularised. P honological factors trigger the use of the imperative b- : (20) b-ár-e [imp-bring.prs-2sg:A] ‘Bring!’ [ JH.50] bí-zn-e=re [imp-take .prs-2sg:A =post] ‘T ake out!’ [KŞ.96] bí-nye [imp-put.prs-2sg:A] ‘Put!’ [KŞ.96] bí-nye-y dê [imp-put.prs-2pl:A] ‘Y ou (pl) Put!’ [BP.186] The verb ‘give ’ takes the indicativ e prefix m- , but in the imperative appears with no b- prefix, see (21)–(22). This may suggest that the subjunctive/imperative prefix has a more restricted use than the indicative m- . (21) î kinaçête de pî kuřîme. î dem.pro x kinaçê=t=e daughter.f.sg=2sg:PSR=dem dé-( e) give.prs.imp-2sg:A p=î to=dem.pro x kuř-î=m=e son.m-sg.obl=1sg:PSR=dem ‘Give y our daughter to my son [in marriage]. ’ [RE.7] 259 9 V erbal infle ctional morphology and v erbal categories 9.1.6 T AM and negation forms in verbs with past time reference The verb forms with a past time refer ence are characterised by the absence of a specific T AM prefix. the negator ne- negates v erbs with past time reference , as shown in the following e xamples. (34) Past p erfective a. wat=iş say .pst=3sg:A ‘he/she said’ b . ne-wat=iş neg-say .pst=3sg:A ‘he/she did not say’ (35) Imp erfect / past irrealis a. waç-ên-î say .prs- a ug-2sg:A ‘you wer e saying’/ ‘you would say’ b . ne-waç-ên-î neg-say .prs- a ug-2sg:A ‘you wer e not saying’/ ‘you would not say’ (36) Present perfe ct a. wate=n=iş say .pst.pt cp.m=cop.3sg.m:O=3sg:A ‘he/she has said (it)’ b . ne-wate=n=iş neg-say .pst.pt cp.m=cop.3sg.m:O=3sg:A ‘he/she has not said it’ (37) Past p erfect a. wate=b-ê=ş say .pst.pt cp.m=be.prs- a ug.3sg:O=3sg:A ‘he/she had said (it)’ b . ne-wate=b-ê=ş neg-say .pst.pt cp.m=be.prs- a ug.3sg:O=3sg:A ‘he/she has said it’ 266 9.1 V erbal inflection: tense-aspe ct-mood and p erson (38) Past conditional a. wat-ɛ=t say .pst - cond=2sg:A ‘(If) you said!’ b . ne-wat-ɛ=t neg-say .pst - cond=2sg:A ‘(If) you didn’t say!’ (39) Past p erfect conditional a. wat-e=bî-ɛ=t say .pst.pt cp.m=be.pst - cond=2sg:A ‘(If) you had said!’ b . ne-wate=bî-ɛ=t neg-say .pst.pt cp.m=be.pst - cond=2sg:A ‘(If) you hadn’t said!’ 9.1.7 Bound person/numb er markers T ekht H. has different paradigms of verbal person/numb er agreement/anaphora suffixes (and gender agr eement in 3sg in some paradigms) in its morphosyntax, used with verbal and non-verbal pr e dicates. These paradigms include (i) copula person markers (PMs), (ii) verbal person/number affixes, and (iii) clitic pronouns. V erbal p erson/number affixes are further divided into two paradigms: one used with verbs derived fr om the present stem and the other used with verbs derived from the past stem. T able 9.3 exhibits the paradigms of person marking. T o b egin with the verbal person/number affixes, the paradigm of present tense verbal affixes is used with T AM categories derived from the present stem of the verbs. The verbal person/number affixes used in this set are stress-taking. They are used to expr ess the subject of present stem verbs. The indexing of the subject argument is obligator y in these contexts, meaning that the endings are agreement markers. The variant forms in the singular, i.e., -ûne , -îne and -one , are phonologically- and morphologically-conditioned variants of -û , -î , and -o , respectively . The heavier forms are used when the endings ar e followed by the coordinate particle =û ; contrast (40) and (41). (40) kuçêş çene werû î masî. kuç-ê=ş little-indf=3sg:R çene from wer- û eat.prs.ind -1sg:A î dem.pro x mas-î yoghurt.m-sg.obl ‘I shall eat a little of it, [ of] this yoghurt. ’ [ JH.42] 267 9 V erbal infle ctional morphology and v erbal categories T able 9.3: Paradigms of p erson/number and gender marking V erbal V erbal Copula Clitic Imp erative PMs prs. PMs pst. PMs pr onouns suffixes 1sg -û, -ûne -a, -anê =na =m 2sg -î, -îne -î =nî =t -e 3sg.m -o, -one -∅ =n , =a =ş 3sg.f -e =ne 1pl -mê, -îmê -îmê, -mê =nmê =ma -mê, -imdê 2pl -de, -îdê -îdê, -dê =ndê =ta -dê 3pl -a, -an(ê) -ê =nê =şa (41) mareş biřûnew . mare=ş marriage=3sg:O biř- ûne =û cut.prs.ind -1sg:A=and ‘I shall marry her [to Pir Shaliyar], and so on. ’ [ JP.225] The heavy forms tend to o ccur with verbs of Goal of motion, such as ‘ come ’ and ‘go ’ , 2 and o ccasionally with v erbs in the irrealis perfect. (42) milone yane. mi-l- one ind -go.prs-3sg:S yane house.m.sg.dir ‘He (the sultan) went inside . ’ [ JH.7 6] (43) herkesî goş darabone herkes-î whoever-obl.m goş ear.m dara=b- one hold.pst.pt cp.m=be.prs-3sg:O ‘ Anyone who had listened to him, [ had gone away fr om the city]. ’ [BP .167] The variant forms in 1pl and 2pl are phonologically conditioned allomorphs of -mê and -dê , respectively . -îmê and -îdê are used following consonant-only 2 The vow el-only person forms may e qually be used with verbs of motion. The choice of the heavier forms may express some discourse pr operty related to the subje ct. 268 9.1 V erbal inflection: tense-aspe ct-mood and p erson stems (b efore which ther e is a word-boundar y ), e.g., b-îmê ‘that we ar e ’ , and following stems ending in a vo wel, e.g., mi-đe-ymê ‘w e give ’ . -mê and -dê are used elsewhere: bi-l-mê ‘let us go ’; ker-mê ‘that we do ’ . The verbal person/number affixes in the past tense are historically the result of the univerbation of the copula base with lexical v erbs, hence their close similarity with the copula paradigm. The paradigm of past verbal person/number affixes has not totally lost its clitic origin, as it is not stress-bearing. The verbal affixes in this paradigm are obligator y indices of intransitive subjects and (to a large extent) direct objects (see §11.1.2.2). In addition, they are used pronominally to e xpress oblique arguments such as recipients, addressees, possessors, and beneficiaries (see §11.1.2). The variant forms for 1pl and 2pl, i.e., -îmê and -îdê , e xhibit morphologically conditioned allomorphy with -mê and -dê . The former are used in the structure of past perfective, see (44); the latter are used in the formation of past perfe ct (45): (44) wit- îmê sleep.pst -1pl:S ‘W e slept. ’ (45) witê=b-ên- mê sleep.pst.pt cp.pl=b e .prs-a ug-1pl:S ‘W e had slept. ’ The copula endings consist of the stem n and set 2 verbal person/number suffixes. The stem has zero inflection in the 3sg.m. The variant form -a is likely to result from the omission of the n . Copula endings ar e obligator y indices of intransitive subjects in non-verbal predicates. Clitic pronouns are r eflexes of Old Iranian genitive/dative and accusativ e p er- son clitic sets (K orn 2009). As discussed in §6.2.1, they assume both phrasal (e.g., possessor in an attributive possessive NP , adposition complements) and clausal functions ( e.g., indexing past transitive subje cts). Their use may be optional or near-obligatory , dep ending on the function they have. For instance, while the y may alternate with free pronouns in indexing possessors, clitic pronouns hav e near-obligatory use in indexing past transitive subjects (§11.1.2.1). The multifunctionality of clitic pronouns and verbal affix PMs r esults in a tense-based role-reference inversion of bound p erson markers in transitive clauses. Consider the expression of A and O in (46). The same person value may express differ ent clausal arguments according to tense. 269 9 V erbal infle ctional morphology and v erbal categories (46) a. b er -î=m take.prs.ind-2sg:A =1sg:O ‘Y ou take me. ’ b . b erđ -î=m take.pst -2sg:O=1sg:A ‘I took you. ’ The tense-based role-refer ence inversion of bound p erson marker in transitiv e clauses may extend to non-core arguments as w ell. In (47), the same set of person markers expresses A and R triggered by tense. (47) a. m-aç -î=m ind-say .prs-2sg:A =1sg:R pey to ‘Y ou say (it) to me. ’ b . wat -î=m say .pst -2sg:R=1sg:A pey to ‘I said (it) to you. ’ Imperative suffixes are , for the most part, limite d to the second person. As can be se en, the 2sg has a differ ent form than the corresponding forms in other paradigms. It is, howe ver , noticeable that the imperative person form -imdê (at- tested in the T ekht varieties of Silên and Nwên) is interpreted as containing ‘the speaker and more than one addressee’ is differentiated from -mê which is inter- preted as meaning ‘the speaker and one addressee ’ . (48) pîyakeyç maço tate w kinaçɫa, ‘da r oɫe gîyan bilimdê pey hêzma. ’ pîya-( e)ke=îç man-def=add m-aç-o ind-say .prs-3sg:A tate-û father-ez.gen kinaçɫa, girl.dim.pl.obl da hort roɫe child gîyan dear bi-l-imdê imp-go.prs-1sg+2pl:S pey for hêzm(î)-a wood-pl.obl ‘The man, the girls’ father said, ‘Dear child(ren), let’s go for [ collecting] woods. ’’ [ÇK.46] (49) maço, ‘ dey bilmê şerʕene!’ m-aç-o ind-say .prs-3sg:A dey hort bi-l-mê sbjv -go.prs-1pl:S şerʕe=ne! sharia_law-post ‘He (Moses) said [to the snake], ‘Let us ( you and me) go to Sharia law . ’’ [MR.21] 27 0 9.2 The copula 9.2 The copula As r emarked in §9.1.7, the copula paradigm consists of the stem n and the infle c- tional endings from the set 2 verbal person/number suffixes. The copula forms are used for the predication of non-verbal elements, such as nouns, adjectives and prepositional phrases. They are also used in the inflection of certain perfect tenses (see below ). The paradigm of the present copula is exhibited in T able 9.4. T able 9.4: Paradigm of present copula Copular endings 1sg =na 2sg =nî 3sg.m =n-∅ , =a-∅ 3sg.f =ne 1pl =nmê 2pl =ndê 3pl =nê In the 3sg, =n is use d following vowel-final non-verbal w ords (see 50), and =a is used following consonant-final words ( see 51). The following examples illustrate the inflection of copula in ascriptive clauses. Recall that adje ctiv es infle ct for gender ( only in the singular) and numb er ( see §7 .1.1). (50) xas ‘well’ 1sg.m xas꞊na [ well.m꞊cop.1sg:S] ‘I (m) am well’ 1sg.f xase=na [ well.f=cop.1sg:S] ‘I (f) am well’ 2sg.m xas꞊nî [ well.m꞊cop.2sg:S] ‘you (m) are well’ 2sg.f xase=nî [ well.f=cop.2sg:S] ‘you (f) are well’ 3sg.m xas꞊a [ well.m=cop.3sg:S] ‘he is well’ 3sg.f xase꞊ne [ well.f=cop.3sg:S] ‘she is well’ 1pl xasê=nmê [ well.pl=1pl:S] ‘we are well’ 2Pl xasê꞊ndê [ well.pl=2pl:S] ‘you are well’ 3Pl xasê꞊nê [ well.pl=3pl:S] ‘they are well’ (51) gewre ‘ old ( of age)’ 1sg.m gewre꞊na [ old.m꞊cop.1sg:S] ‘I (m) am old’ 1sg.f gewrê=na [ old.f=cop.1sg:S] ‘I (f) am old’ 2sg.m gewre꞊nî [ old.m꞊cop.2sg:S] ‘you (m) are old’ 2sg.f gewrê=nî [ old.f=cop.2sg:S] ‘you (f) are old’ 271 9 V erbal infle ctional morphology and v erbal categories 3sg.m gewre꞊n [ old.m=cop.3sg:S] ‘he is old’ 3sg.f gewrê꞊ne [ old.f=cop.3sg:S] ‘she is old’ 1pl gewrê=nmê [ old.pl=1pl:S] ‘we ar e old’ 2Pl gewrê꞊ndê [ old.pl=2pl:S] ‘you are old’ 3Pl gewrê꞊nê [ old.pl=3pl:S] ‘they ar e old’ 9.2.1 Negative copula The negation of the present copula is e xpressed by adding nîye- to the copula endings (T able 9.5). The negative form of the copula can be considered a trun- cation of the negation nî- and the existential copula, e.g., nîy e=na <* nî-he=na [neg-exist=cop.1sg:S] ‘I am not’ . T able 9.5: The negative copula paradigm Negative copula 1sg nîye=na 2sg nîye=nî 3sg.m nîya-∅ 3sg.f nîye=ne 1pl nîye=nmê 2pl nîye=ndê 3pl nîye=nê The following examples illustrate the use of negativ e copula in the text corpus. (52) ħîçê nîya. ħîç-ê nothing-indf nîy( e)=a neg.exist=cop.3sg.m:S ‘There is nothing [left]. ’ [HB.4] (53) hermane be desû min nîyene. hermane work.f be to des-û hand.m-ez.gen min 1sg nîye=ne neg.exist=cop.3sg.f:S ‘The task [ of ruling] is not in my hands. ’ [ŞC.86] Proof for the analysis of nîye- as consisting of the negation nî- and the existen- tial copula comes from the use of the nîye- as the negated predicate in predicative possessive constructions, which are formed based on the existential copula (see §9.2.3.1). 272 9.2 The copula (54) zeraʕetma nîyaw bencanîma nîyenew xîyarma nîyaw genmîma nîyene . zeraʕet=ma agriculture.m=1pl:NC nîy( e)=a=û neg.exist=cop.3sg.m:S=and bencanî=ma tomato.f=1pl:NC nîye=ne=û neg.exist=cop.3sg.f:S=and xîyar=ma cucumber.m=1pl:NC nîy( e)=a=û neg.exist=cop.3sg.m:S=and genmî=ma wheat.f=1pl:NC nîye=ne neg.exist=cop.3sg.f:S ‘W e don’t have much agriculture; we don’t have cucumbers; we don’t have tomatoes; we don’t have wheat. ’ [PM.37] 9.2.2 Past copula There are tw o paradigms of the past copula; se e T able 9.6. One set is formed by combining the past stem of the verb ‘to be ’ with verbal person/number affixes from the past tense . The se cond set uses the present stem of the v erb ‘to b e ’ , to which the augment -ên is added, and the resulting form is inflected by the same set of past tense verbal person/number suffixes. In the current state of the language, set 2 occurs at a higher rate than set 1. In the literature , the second set has be en r eferred to as the “imperfe ct copula” , e.g., Mahmoudv eysi & Bailey (2018: 551), primarily be cause the augment -ên which is used in the structure of past imperfective is used in the structure of this past copula as well. T able 9.6: Past copula paradigms Set 1 Set 2 1sg bî-a b-ên-ê 2sg b(î)-î b-ênî 3sg.m bî-∅ b-ê <* b-ên 3sg.f bî-e b-ê <* b-ên 1pl bî-mê b-ên-mê 2pl bî-dê b-ên-dê 3pl bî-ê 3 b-ên-ê In the curr ent state of the dialect, set 2 past copula functions as the basic copula. Examples: 3 In one case, the form bî-a is attested, e.g., gumê bîya [ŞC.39] ‘they got lost’ , which would indicate syncretism between 1sg and 3pl. Howe ver , it is possible that the narrator actually meant ‘I was lost’ , which w ould discard the syncretism scenario. 273 9 V erbal infle ctional morphology and v erbal categories (55) ême firê bênmê. ême 1pl fire-ê a_lot-pl b-ên-mê be.prs- a ug-1pl:S ‘W e were a large number . ’ [BP .110] (56) min zaroɫe bênê. min 1sg zaroɫe child b-ên-ê be.prs-a ug-1sg:S ‘I was a child. ’ [ JH.1] Set 1 is used in the inflection of light verb constructions which are based on the light verb ‘to be ’ . (57) luwanê nizîkû qiřoɫekey bîyaw e. luwa-anê go.pst -1sg:S nizîk-û close-ez.gen qiřoɫ-ekey hollow.m-def.m.sg.obl bî-a=we be.pst-1sg:S=compl ‘I went [ and] got closer to the tree hollow . ’ [ZQ .36] (58) kuřêş peya bî. kuř-ê=ş son.m-indf=3sg:PSR peya visible bî-∅ be.pst-3sg.m:S ‘ A b o y was b orn to him. ’ [BP .4] Less frequently , set 1 is used as the past copula, including in cleft constructions (59). (59) î gîre çêş bî min ward? î dem.pro x gîr=e hook.m=dem çêş what bî-∅ be.pst-3sg.m:S min 1sg ward-∅ eat.pst -3sg.m:O ‘What is this situation that I am caught in?’ [HB.23] (60) min çen bê ʕeqiɫ bîya! min 1sg çen so_much bê-ʕeqiɫ without-wisdom bî-a be.pst-1sg:S ‘[Look] how stupid I was!’ [HB.19] (61) maço , ‘îne her hêɫe bî’ m-aç-o ind -say.prs-3sg:A îne dem.pro x.m.3sg.dir her just hêɫe egg.m bî-∅ be.pst-3sg.m:S ‘He said, ‘This one was just an egg. ’’ [ JH.95] 274 9.2 The copula 9.2.3 Existential copula The existential copula is formed by adding inflectional suffixes from the past tense to existential base hen <* he + =n . T able 9.7: The existential copula paradigm Gloss Lexical gloss hen-a [exist -1sg] ‘I am’ hen-î ‘you are ’ hen-∅ ‘he is’ hen-e ‘she is’ hen-mê ‘we are ’ hen-dê ‘you are ’ hen-ê ‘they are ’ Examples from the text corpus: (62) pîyawê hen taw o î kinaçê to weşe kerow e. pîya-ê man.m-indf hen-∅ exist -3sg.m:S taw-o can.prs.ind -3sg:A î dem.pro x kinaçê girl.f to 2sg weş-e well-f kér-o=we do.prs.sbjv -3sg:A =compl ‘There was a man in the Hewraman r egion who could cure his daughter . ’ [ JP.154] (63) ne nan henû ne har dî henê. ne no nan bread.m.sg.dir hen-∅=û exist -3sg.m:S=and ne no hardî flour.f.pl.dir hen-ê exist -3pl:S ‘There is no br ead. There is no flour . ’ [HB.3] The existential copula can be used nominally , in which case it may appear as the head noun in a nominal phrase. (64) ey henû nîy enû min ey v oc hen=û being=and nîyen-û non_being-ez.gen min 1sg ‘O my relativ es [ lit. O my being and non-b eing]’ [KŞ.56] 275 9 V erbal infle ctional morphology and v erbal categories T able 9.10: The negation of the p r esent indicative–sample paradigms kerđey ‘ do ’ luway ‘go ’ arđey ‘bring’ 1sg me-ker-û me-l-û nim(e)-ar-û 2sg me-ker-î me-l-î nim( e)-ar-î 3sg me-ker-o me-l-o nim(e)-ar-o 1pl me-ker-mê me-l-mê nim(e)-ar-mê 2pl me-ker-dê me-l-dê nim(e)-ar-dê 3pl me-ker-a me-l-a nim(e)-ar-a (88) milonû wer werwe mar o b e nîrûweke yş de koɫê penj koɫê gi řowê. mi-l-on=û ind -go.prs-3sg:S=and wer out werwe snow.f m-ar-o ind -bring.prs-3sg:A be by nîrû-ekey=ş force.m-def.m.sg.obl=3sg:PSR de ten koɫ-ê load/shoulder.m-pl.dir penj five koɫ-ê load/shoulder.m-pl.dir gi each řo-ê day.m-pl.dir ‘He (i.e., Jamsher Shah) fetched snow , five or ten loads daily , using his men. ’ [DP .34] This habitual function of the present indicative allo ws its extension to the narrative present, which is the use of the present tense form to refer to past events. In this function, the pr esent narrative alternates with the past tense to make a past tense event mor e vivid (Schiffrin 1981). In the following excerpt, the shift to the present tense adds an element of surprise to a past tense ev ent. 6 (89) luwanê nizîkû qiřoɫekey bîyaw e. miđye w sîwînîş mê qiřoɫekey . nizîk bîyawe temaşem kerd. we xtê miđyew kuřeke bîy en peřû qiřoɫêwe , qiřoɫekey . luwa-(a)nê go.pst -1sg:S nizîk-û close-ez.gen qiřoɫ-ekey hollow.m-def.m.sg.obl bî-a=we be.pst-1sg:S=compl mi-đye-û ind -look.prs-1sg:S sîwînî=ş voice=3sg:PSR m-ê ind -come.prs.3sg:S qiřoɫ-ekey hollow.m-def.m.sg.obl nizîk close bî-a=we be.pst-1sg:S=compl temaşe=m looking=1sg:A kerd do.pst 6 See Noorlander & Mohammadirad (2022) for a similar function of narrative present in the Kurdish and Neo- Aramaic dialects of the region. 282 9.3 T AM categories built on the present stem wext-ê time.m-indf mi-đye-w ind -look.prs-1sg:S kuř-eke son.m-def.m.sg.dir bîye=n be.pst.ptcp.m=cop.3sg.m:S peř-û full-ez.gen qiřoɫ-êw e hollow.m-indf qiřoɫ-ekey hollow.m-def.m.sg.obl ‘I went [ and] got closer to the tree hollow . I noticed [ lit. notice] a v oice came (lit. comes) from the hollow . I got closer and lo ok ed inside. I noticed that the boy [ had gro wn up so much that he] had filled a tree hollow , i.e., the tree hollow . ’ [ZQ .36–ZQ .38] 9.3.1.2 Performative The present indicative may also e xpress a perfe ctiv e present. Here, the v erb ex- presses a bounded p erfective ev ent. (90) fermawo , ‘weɫa çûn ana bêheyanî kerût penû ʕaɫemî. ’ fermaw-o say.prs.ind -3sg:A weɫa indeed çûn as an( e)=a dem.dist.m.3sg.dir=deic bêheya=nî shameless=cop.2sg:S ker-û=t do.prs.ind -1sg:A =2sg:O pen-û advice.m-ez.gen ʕaɫem-î world-m.sg.obl ‘He (Pir Shaliyar ) said, ‘Inde ed, since you ar e so shameless, I ( hereby ) expose your impr opriety to the world. [Lit. I (hereby) make y ou the news of the world.]’’ [ JP .198] (91) maço, ‘ dey min şûyş kerû pene. ’ m-aç-o ind -say.prs-3sg:A dey disc.pt cl min 1sg şû-î=ş husband.m-sg.obl=3sg:R ker-û do.prs.ind -1sg:A pene to ‘She said, ‘I will marry him. ’’ [ JH.59] 9.3.1.3 Future The present indicative is fr equently use d to expr ess a future time reference . This use of the future in the narrative is sequential to what precedes. 283 9 V erbal infle ctional morphology and v erbal categories (92) ême milmê şeşik. milmê goşew şarî, awayeke y . çowe neferê we mê, pêwaymawe . ême 1pl mi-l-mê ind -go.prs-1pl:S şeşik pn mi-l-mê ind -go.prs-1pl:S goşe-û corner-ez.gen şar-î city.m-sg.obl awaî-ekey village.m-def.m.sg.obl ç=o=we in=dem.dist=post nefer-êwe person.m-indf m-ê ind -come.prs.3sg:S pêway=ma=we welcoming.m=1pl:PSR=post ‘W e will go to Shashk. W e will go to the city suburb, to the village . There, a person will welcome us. ’ [HB.32–HB.34] 9.3.2 Present subjunctive Like in the present indicative, the vast majority of verbs in the present subjunc- tive lack any expr ession of the subjunctive prefix; see the conjugation of kerđey ‘ do ’ in T able 9.11. As discussed in §9.1, for this class of verbs, the stress on the stem distinguishes the subjunctive from the indicativ e, thus kéro ‘that he/she does’ vs. keró ‘he/she does’ . For the rest of the verbs, the present subjunctiv e is built by attaching the subjunctive prefix b(i)- to the pr esent stem of the verb , followed by inflectional person affixes. With these verbs, the subjunctive prefix may or may not be stressed, dep ending on the syllable structure of the verb stem ( se e §9.1.2 for details). The negation of present subjunctive v erb forms is expressed by ne- , reduced as n- , before vow el-initial verbs (see T able 9.12). The present subjunctive verb forms are , by default, used in sub ordinate clauses. This includes, for example , the protasis of general conditions (93), dependent purpose clauses (94), and prospective aspe ct (95). T able 9.11: The present subjunctive–sample paradigms kerđey ‘ do ’ luway ‘go ’ arđey ‘bring’ 1sg kér-û bí-l-û b-ár-û 2sg kér-î bí-l-î b-ár-î 3sg kér-o bí-l-o b-ár-o 1pl kér-mê bí-l-mê b-ár-mê 2pl kér-dê bí-l-dê b-ár-dê 3pl kér-a bí-l-a b-ár-a 284 9.3 T AM categories built on the present stem T able 9.12: The negation of the present subjunctive–sample paradigms kerđey ‘ do ’ luway ‘go ’ arđey ‘bring’ 1sg né-ker-û né-l-û n-ár-û 2sg né-ker-î né-l-î n-ár-î 3sg né-ker-o né-l-o n-ár-o 1pl né-ker-mê né-l-mê n-ár-mê 2pl né-ker-dê né-l-dê n-ár-dê 3pl né-ker-a né-l-a n-ár-a (93) dey çêşim miđey min kuřleke yt weş kerûwe? dey disc.pt cl çêş=im what=1sg:R mi-đe-î ind -give .prs-2sg:A min 1sg kuřle-( e)key=t son.dim-def.m.sg.obl=2sg:PSR weş well kér -û=we do.prs.sbjv -1sg:A =compl ‘What will you give me if I bring back your little son to life [lit. heal]?’ [ÇK.109] (94) amɛnmê xizmetû to îcazema biđey . amɛ=nmê come.pst.pt cp.pl=cop.1pl:S xizmet-û service.m-ez.gen to 2sg îcaze=ma permission=1pl:R bi-đé-î give.prs.sbjv -2sg:A ‘I (lit. W e) have come to your service so that you might permit us. ’ [PM.15] (95) tejnayene epê w ext bê soçmê nařehetîyene. tejnay=ene thirst.inf=post epê in.dem.pro x wext time.m b-ê be.prs-a ug.3sg:S sóç-mê burn.prs.sbjv -1pl:S nařehetî=ene sadness=post ‘W e were about [ lit. it was time] to burn fr om thirst and sorrow [from leaving behind one of the sons]. ’ [ZQ .25] The subjunctive verb form is used following deontic particles ba , and meger , used in interactional discourse expressing a wish, asking for permission, or giv- ing advice. 285 9 V erbal infle ctional morphology and v erbal categories (96) watiş, ‘pilekanê çine ba bilmê!’ wat=iş say.pst=3sg:A pilekan(î)-ê stair.f-indf çín-e set_up.prs.imp -2sg:A ba hort bi-l-mê sbjv -go.prs-1pl:S ‘He said, ‘Set up a stairway for us to move!’’ [ JH.25] (97) ʕerziş kerđ, ‘ qurban ba dastanêt pey gêɫnûwe, dastane. ’ ʕerz=iş petition.m=3sg:A kerđ do.pst qurban sir .v oc ba hort dastan( e)-ê=t story.f-indf=2sg:R pey to gê  ɫn-û=we narrate.prs.sbjv -1sg:A =compl dastane story.f ‘He said, ‘Sir , let me tell you a story , the story!’’ [ZQ .7] (98) meger berî kîyanîş p e y hewramanî. meger if_only bér-î take.prs.sbjv -2sg:A kî  yan-î=ş send.prs.sbjv -2sg:A =3sg:O pey to hewraman-î pn-m.sg.obl ‘Maybe you could take her to Hewraman. ’ [ZP .34] The subjunctive verb form is used in some factive content clauses and non- factive complement clauses. (99) metawo bilo awe yanî. me-taw-o neg.ind -can.prs-3sg:A bi-l-o sbjv -go.prs-3sg:S aweyanî habitat.m ‘He wasn’t allowed to go to the village . ’ [DG.54] (100) îcaze, bizane ađ îcaze miđo ême eçê nehar kermê yam ne. îcaze permission bi-zán-e imp -know.prs-2sg:A ađ 3sg.m.dir îcaze permission mi-đ( e)-o ind -give.prs-3sg:A ême 1pl e=çê in=here nehar lunch.m kér-mê do.prs.sbjv -1pl:A yam or ne no ‘See if he lets us stay here for lunch or not. ’ [PM.9] In some cases, a subjunctive v erb form occurs in a main clause. A case in point is the use of subjunctive mood following the particle da in a construction expressing speaker-oriented modality. In the following example, the conte xt is self-hortative , i.e., the speaker encourages herself to action. 286 9.3 T AM categories built on the present stem (101) maço, ‘ da bizanû çêşşa ser ama. ’ m-aç-o ind -say .prs-3sg:A da hort bi-zan-û sbjv -know .prs-1sg:A çêş=şa what=3pl:R ser to ama come.pst.3sg:S ‘She said, ‘I shall see what happene d to them. ’’ [SH.124] The subjunctive may be used in indefinite relative clauses. (102) ne kesê hen be weş dađiş berî la. ne nor kes-ê person.m-indf hen-∅ exist -3sg.m:S be by weş good dađ=iş cry.m=3sg:R bér-î take.prs.sbjv -2sg:A la to ‘Nor is there a person with whom one can take counsel in peace. ’ [ZB. 60] 9.3.3 Imperative For most verbs, the imperative verb forms ar e constructed by adding the 2nd person inflectional p erson suffixes to the present stem of the v erb (see §9.1.3). Additionally , a small subset of verbs takes the subjunctive/imperative pr efix; se e T able 9.13 bi- for sample paradigms. T able 9.13: The imp erative–sample paradigms kerđey ‘ do ’ luway ‘go ’ arđey ‘bring’ 2sg ker-e lu-e , bi-l-e b-ar-e 2pl ker-dê lo-dê b-ar-dê The negation of the imperative is carried out by me- for verbs which in their imperative form do not take any prefixes, ne - for stems starting with m , and nime- for a subset of vow el-initial verbs (see §9.1.4 for details). T able 9.14 lists sample paradigms of prohibitive verbs. The imperative mood is used to command an action or expr ess a request. When used negatively , it prohibits an action fr om b eing undertaken. (103) bereşo! bér-e=ş=o take.prs.imp -2sg:A =3sg:O=compl ‘T ake her back!’ [ZP .106] 287 9 V erbal infle ctional morphology and v erbal categories T able 9.14: The negation of the imp erativ e–sample paradigms kerđey ‘ do ’ luway ‘go ’ arđey ‘bring’ 2sg mé-ker-e mé-l-e nim(e)-ár-e 2pl mé-ker-dê mé-l-dê nim(e)-ár-dê (104) watiş, ‘fermawdê beydê mêmanû minindê . ’ wat=iş say.pst=3sg:A fermaw-dê say.prs.imp -2pl:A b-e-îdê imp -come.prs-2pl:S mêman-û guest.m-ez.gen min=indê 1sg=cop.2pl:S ‘He said, ‘Please come! Y ou are my guests. ’’ [HB.51] (105) îneyşa mekerdê zînan. îney=şa dem.pro x.obl.m.3sg=3pl:PSR me-ker-dê proh-do.prs-2pl:A zînan prison.m ‘Do not put this one in prison. ’ [BP .133] An imperative verb form may be given added immediacy by combining with the particle da . (106) fermawo , ‘ da lo dê bizandê fiɫane kes çî nama?’ fermaw-o say.prs.ind -3sg:A da hort lo-dê go.prs.imp-2pl:S bi-zan-dê imp -know.prs-2pl:A fiɫan-e such_and_such-ez.cmpd kes person.m çî why n( e)-ama-∅ neg-come.pst -3sg.m:S ‘He (Baba Khwada) said, ‘Go [and] see (lit. know ) why such-and-such person hasn’t come [to the mosque]?’’ [BP .77] The particle da may be combine d with the discourse particle dey for an e ven greater degr e e of immediacy . (107) da dey hurbêze î mexlûqî gir diş gêɫeş pone. da hort dey disc.pt cl hur-b-êz-e pvb-imp -rise.prs-2sg:S î dem.pro x mexlûq-î people-m.sg.obl gird=iş all=3sg:PSR gê  ɫ-e=ş wander.prs.imp -2sg:S=3sg:R pone at ‘Come on, get up and search among all these people. ’ [HB.90] 288 9.3 T AM categories built on the present stem 9.3.4 Present progressiv e The present progr essive can be expressed in several ways. It can be expr esse d by the infle cted form of the verb identical to the present indicative. The development seems to b e an e xtension from the progr essive sense to the habitual sense (Deo 2015). (108) min îse î nîştenare îsrahetîç kerû. min 1sg îse now î dem.pro x nîşte=na=re sit.pst.pt cp.m=cop.1sg:S=povb îsrahet=îç rest.m=add ker-û do.prs.ind -1sg:A ‘Now I am seated, and I am resting. ’ [PM.45] (109) înê î qisê çêşene? çêş maçdê? maça çêş? înê dem.pro x.f.3sg.dir î dem.pro x qisê talk.f.sg çêşe=ne what.f=cop.3sg.f:S çêş what m-aç-dê ind -say.prs-2pl:A m-aç-a ind -say.prs-3pl:A çêş what ‘What is this talk? What are you saying? What are they saying?’ [ JP .223] The present progr essive may be expressed by a reduplicated progressiv e con- struction consisting of the inflected form of the verb precede d by a form contain- ing the present form of the verb and the suffix -ay , r esembling the infinitive suffix. The copy takes the same inflectional prefix as the inflected verb , suggesting it is on its way to being grammaticalise d as a v erbal form. 7 (110) weray werû. wer-ay eat.prs-nmlz wer-û eat.prs-1sg ‘I am eating. ’ (111) yewaşê de wrû dimaw ʕîşay kero, şew e kero waray waro mê ew peřû degɛ. yewaşê then dewr-û around-ez.gen dima-û afterwards-ez.gen ʕîşa-î evening_pray ers.m-sg.obl ker-o do.prs.ind -3sg:A şewe night.f ker-o do.prs.ind -3sg:A war -ay rain.prs-nmlz 7 A similar construction consisting of the infinitive and the inflected form of the verb exists in the neighbouring Jewish Neo- Aramaic dialect of Sanandaj, e.g., şatoê şatêna ‘I am drinking’ (Khan 2009: 275). This has been taken as a case of matching between Hewramî and Neo- Aramaic in the contact setting in Sanandaj (see Khan & Mohammadirad 2024a for details). 289 9 V erbal infle ctional morphology and v erbal categories T able 9.15: The reduplicate d present pr ogressive–sample paradigms arđey ‘bring’ warđey ‘ eat’ 1sg m-ar-ay m-ar-û ‘I am bringing’ wer-ay wer-û ‘I am eating’ 2sg m-ar-ay m-ar-î wer-ay wer-î 3sg m-ar-ay m-ar-o wer-ay wer-o 1pl m-ar-ay m-ar-mê wer-ay wer-mê 2pl m-ar-ay m-ar-dê wer-ay wer-dê 3pl m-ar-ay m-ar-a wer-ay wer-a war -o rain.prs.ind -3sg:S m-ê ind -come.prs.3sg:S ew dem.dist peř-û side.m-ez.gen degɛ village.f.sg.obl ‘Then, it was either during the evening pray ers or during the night that he arrived at the other side of the village while it was raining. ’ [KŞ.59] The reduplicated progressiv e construction may also b e used in negation and interrogativ e clauses, exhibiting a grammaticalised progressive form. The r edu- plicated construction in negation clauses conveys extra emphasis. (112) meweray me werû! me-wer-ay neg.ind-eat-nmlz me-wer-û neg.ind-eat.prs-1sg:A ‘I am not eating!’ (113) milay milî? mi-l-ay ind -go.prs-nmlz mi-l-î ind -go.prs-2sg:S ‘ Are you going?’ (114) taway tawû barşa kerû? taway tawû çêş kerû? taw-ay can.prs-nmlz taw-û can.prs-1sg bar=şa load=3pl:O ker-û do.prs.ind-1sg:A taw-ay can.prs-nmlz taw-û can.prs-1sg çêş what ker-û do.prs.ind-1sg:A ‘ Am I able to load them? What am I able to do?’ [SH.127] 290 9.3 T AM categories built on the present stem (115) mewînay mewînmê? me-wîn-ay neg.ind -see.prs- adv me-wîn-mê neg.ind -see.prs-1pl:A ‘Don’t we really see?’ The construction may also be use d for futur e reference in certain contexts. (116) êşew gêɫay gêɫa pey pîyaya. êşew tonight gêɫ-ay search.prs-nmlz gêɫ-a search.prs.ind-3pl:A pey for pîya-ya man-pl.obl ‘T onight, they will b e looking for men [ who can ser ve them]. ’ [ JL.34] (117) min gunakê may mê milimre . min 1sg guna-( e)kê sin.f-def.f.sg m-ay ind -come.prs.nmlz m-ê ind -come.prs.3sg:S mil=im=re shoulder.m=1sg:PSR=post ‘I won’t shoulder the burden [ of injuring them]. [Lit. the sin will come to me.]’ [DG.67] (118) řama dûrene . milay milmêwe . řa=ma road.f=1pl:PSR dûr-e=ne far-f=cop.3sg.f:S mi-l-ay ind -go.prs-nmlz mi-l-mê=we ind -go.prs-1pl:S=compl ‘W e have a long way [to go]. W e should be going. ’ [BP.191] Another strategy to express the pr esent progressiv e is through a noun phrase consisting of the nominal form xerîk ‘busy’ combined with the infinitive form of the verb or another nominal. (119) î meʕmûrê xerîkû genekarînê. î dem.pro x meʕmûr-ê officer.m-pl.dir xerîk-û busy-ez.gen genekarî=nê debauchery.m=cop.3pl:S ‘The officers were busy engaging in debauchery . ’ [BP .45] Similarly , the present pr ogressive may be expressed by xerîk bîye y ‘be busy’ combined with the inflecte d form of the verb . 291 9 V erbal infle ctional morphology and v erbal categories T able 9.20: The past p erfective/preterite–sample paradigms merđey arđey 1sg mérđ-a ‘I died’ árđ-∅=im ‘I brought it (m). ’ 2sg mérđ-î árđ-∅=it ‘Y ou brought it (m). ’ 3sg.m mérđ-∅ árđ-∅=iş ‘He/she brought it (m). ’ 3sg.f mérđ-e árđ-e=iş ‘He/she brought it (f). ’ 1pl mérđ-îmê árđ-∅=ma ‘W e brought it (m). ’ 2pl mérđ-îdê árđ-∅=ta ‘Y ou brought it (m). ’ 3pl mérđ-ê árđ-∅=şa ‘They brought it (m). ’ As seen in T able 9.20, S and A are indexed by different paradigms of person endings. T ransitive objects (O) ar e indexed the same as S. When combine d with the clitic pronoun indexing the A argument, the or dering is V -O=A, irrespe ctive of the person of the O; see T able 9.21. T able 9.21: Past p erfective–the inflection of arđey ‘bring’ O suffix 3pl A Gloss 1sg arđ-a=şa [bring.pst-1sg:O=3pl:A] ‘they br ought me ’ 2sg arđ-î=şa [bring.pst-2sg:O=3pl:A] ‘they br ought you’ 3sg.m arđ-∅=şa [bring.pst-3sg.m:O=3pl:A] ‘they brought him’ 3sg.f arđ-e=şa [bring.pst-3sg.f:O=3pl:A] ‘they brought her’ 1pl arđ-îmê=şa [bring.pst-1pl:O=3pl:A] ‘they brought us’ 2pl arđ-îdê=şa [bring.pst-2pl:O=3pl:A] ‘they brought y ou’ 3pl arđ-ê=şa [bring.pst-3pl:O=3pl:A] ‘they brought them’ The negation of past perfective is expressed by ne- : (136) maço, ‘nezanam. ’ m-aç-o ind -say.prs-3sg ne-zana=m neg-know.pst=1sg:A ‘He (the man) said, ‘I didn’t understand [his p oint]. ’’ [ JH.26] The past perfective is used to express specific time-b ound ev ents (i.e., com- pleted events) at a particular time in the past. 298 9.4 T AM categories derived from past stem (137) duwê be çkêş dîyê , zarowê . duwê two be çk( e)-ê=ş baby-pl.dir=3sg:A dî-ê see.pst -3pl:O zaro-ê child-pl.dir ‘She gave birth to two babies. ’ [ZQ .15] The past perfective may be use d to refer to sequential time-bound events in a narrative: (138) gêɫanê qiřoɫe darêm yoso . berdim nîyamne qiřoɫe dareke . gêɫa-(a)nê wander.pst -1sg:S qiřoɫ-e hollow.m-ez.cmpd dar-ê=m tree.m-indf=1sg yos=o find.pst=compl berd-Ø=im take.pst -3sg.m:O=1sg:A nîya-Ø=m=ne put.pst -3sg:O=1sg:A =po vb qiřoɫ-e hollow.m-ez.cmpd dar-eke tree.m-def.m.sg.dir ‘I wandered around [ and] found a tree hollow . I took [him] and put him in the tree hollow . ’ [ZQ .23–ZQ .24] The past perfective may express an action that has a starting and end point in the past but lasted for a long perio d. In (139), ‘raising childr en’ would have lasted several y ears. (139) pase zawlêşa wey ker dê. pase like_this zawlê=şa child.pl.dir=3pl:A wey raising kerd-ê do.pst -3pl:O ‘They raised children in this way . ’ [ JE.55] The extended period of time may overlap with other ev ents in the surrounding discourse. In (140), the adv erbial phrase ‘when I got married’ sets the frame for all the events r elating to the perio d after marriage . (140) wextê jenîm ar de, ye waşê yanem nebê. cîya bîyanê. ca zəmsan bê. cîya bîya. luwanê hîçim nebê. çenû jenî luwaymê yanema gêrt kirahe . wext-ê when-indf jenî꞊m woman꞊1sg:A ard-e bring.pst -3sg.f:O yewaşê well yane=m house꞊1sg:NC ne-b-ê neg-be.prs- a ug.3sg:S cîya separate bî-anê. be.pst -1sg:S ca disc.pt cl zəmsan winter b-ê. be-prs.a ug.3sg:S cîya separate bî-a be.pst -1sg:S luwa-(a)nê go.pst -1sg:S hîç=im nothing꞊1sg:NC 299 9 V erbal infle ctional morphology and v erbal categories ne-b-ê neg-be.prs- a ug.3sg:S çen(î)-û with jenî woman luwa-îmê go.pst -1pl:S yane꞊ma housedir.m=1pl:A gêrt-∅ take.pst -3sg.m:O kirahe rent ‘When I got married (I took a wife), well, I didn’t have a house. I left my father’s house (lit. I b ecame separate). It was winter . I left the family of my father , and I went away . I did not have anything. T ogether with my wife, w e rented a house. ’ (Khan & Mohammadirad 2024a: 309, glossing and transcription modified) The perfective is also used for the expression of a time-bound event in the immediate past, corresponding to the English perfe ct of r e cent past/hot ne ws perfect. In the following example, the narrator witnesses a guest coming through the door and asks whether the recording should continue. (141) dey aney ç ama mêmana çêş kermê? dey disc.pt cl ane=yç dem.dist.3sg.dir.m=add ama-Ø come.pst -3sg:S mêman=a guest=cop.3sg:S çêş what kér-mê do.prs.sbjv -1sg:A ‘Oh, he has arrived. He is guest. What should we do? [Should w e ke ep recording?]’ [HR.59] The perfective may express a completed action with a projected future sense. In this usage, the perfective occurs in a sub or dinate clause, whether syndetic (142) or asyndetic (143). (142) eger goşiş darayne dûr dûr kewto . eger if goş=iş ear.m=3sg:A dara-î=ne hold.pst -2sg:R=povb dûr far dûr far kewt-Ø=o fall.pst -3sg.m:S=compl ‘If they [lit. he] listen to you, they will go away . ’ [BP .163] (143) dey êşew herkes mêmaniş hen, êşe w herkes mêmaniş hen, mêmanekeş şewê witê, ser eş biřo. dey disc.pt cl êşew tonight herkes everyone mêman=iş guest.m=3sg:NC hen-∅ exist -3sg.m:S êşew tonight herkes everyone mêman=iş guest.m=3sg:NC hen-∅ exist -3sg.m:S mêman-eke=ş guest.m-def.m.sg.dir=3sg:PSR 300 9.4 T AM categories derived from past stem şew(e)-ê night-f.sg.obl wit-ê sleep.pst -3pl:S:S sere=ş head.m=3sg:O bíř-o cut.prs.sbjv -3sg:A ‘T onight, who ev er has guests, [when] the guest sle eps at night , he shall decapitate him. ’ [BP .52] 9.4.2 Past conditional The past conditional is formed by attaching the conditional affix -εn to the past stem of the verb , followed by set 2 inflectional person suffixes. MacKenzie (1966: 34) argues that -εn is presumably derived from the conditional infix -a plus the augment -ên , hence -εn <* -a + -ên . T able 9.22: Past conditional–the infle ction of ‘arriv e’ S Gloss 1sg yáw(a)-εn-ê [ arrive.pst - cond.a ug-1sg:S] ‘(if) I had arrived’ 2sg yáw(a)-εn-î [ arrive.pst - cond.a ug-2sg:S] ‘(if) you had arriv ed’ 3sg yáw(a)-ε [ arrive.pst - cond.a ug.3sg:S] ‘(if) he/she had arrived’ 1pl yáw(a)-εn-mê [ arrive.pst - cond.a ug-1pl:S] ‘(if) we had arrived’ 2pl yáw(a)-εn-dê [ arrive.pst - cond.a ug-2pl:S] ‘(if) you had arrived’ 2pl yáw(a)-εn-ê [ arrive.pst - cond.a ug-3pl:S] ‘(if) they had arrived’ T able 9.23: Past conditional–the infle ction of arđe y ‘bring’ O suffix 3pl A Gloss 1sg arđ-εn-ê=şa [bring.pst-cond.a ug-1sg:O=3pl:A] ‘(if) they had brought me ’ 2sg arđ-εn-î=şa [bring.pst-cond.a ug-2sg:O=3pl:A] ‘(if) they had brought y ou’ 3sg arđ-ε=şa [bring.pst-cond.a ug.3sg:O=3pl:A] ‘(if) they had brought her/him’ 1pl arđ-εn-mê=şa [bring.pst-cond.a ug-1pl:O=3pl:A] ‘(if) they had brought us’ 2pl arđ-εn-dê=şa [bring.pst-cond.a ug-2pl:O=3pl:A] ‘(if) they had brought y ou’ 3pl arđ-εn-ê=şa [bring.pst-cond.a ug-3pl:O=3pl:A] ‘(if) they had brought them’ The negation of the past conditional is expressed by ne- : 301 9 V erbal infle ctional morphology and v erbal categories (144) eger řezaşa nebîɛ neđênê. eger if řeza=şa satisfaction.f=3pl:NC ne-bî-ɛ neg-be.pst.cond.a ug.3sg:S ne-đ( e)-ên-ê neg-give.prs- a ug-3pl:A ‘If they didn’t agree, the y wouldn’t give her . ’ [ JE.85] (145) eger minit çene bîyɛnê eger if min=it 1sg=2sg:R çene with bî-ɛn-ê be.pst.cond.a ug-1sg:S ‘If you had me with you .. ’ [P W .88] The past conditional expresses hypothetical situations in the past. (146) a wextî to mîsal watɛt, ‘a fiɫane kesem gerekene . ’ a dem.dist wext-î time.m-sg.obl to 2sg mîsal for_example wat-ɛ=t say.pst - cond.a ug=2sg:A a dem.dist fiɫan-e such_and_such-ez.cmpd kese=m person.f=1sg:NC gerek-e=ne necessar y-f=cop.3sg.f:S ‘In earlier times, let us say (lit. for instance.) you said , ‘I want that certain person (i.e., girl). ’’ [ JE.73] (147) eger kinaçekêşa donɛ watɛş, ‘ erê, kerû, ’ eger if kinaç(ê)-ekê=şa girl.f-def.f.sg=3pl:A donɛ talk_to.pst.cond.a ug.3sg:R wat-ɛ=ş say.pst.cond.a ug=3sg:A erê yes ker-û do.prs.ind -1sg:A ‘ After they had talked to the girl [and] she had said , ‘Y es, I will [marr y him], ’’ [ JE.77] 9.4.3 Perfect The perfect is expressed by combining the resultative participle with the copula PMs. The participle agrees in gender and number with the underlying S and O . In T able 9.24, the participle forms for the verb ‘to die ’ is merde [ die.pst.pt cp.m.sg], merdê [ die.pst.pt cp.f.sg], and merdê [ die.pst.pt cp.pl]. Similarly , the copula PMs agree with the O and S argument in person. The clitic pronouns expr ess the A argument. 302 9.4 T AM categories derived from past stem T able 9.24: The p erfect–sample paradigms S/O merđey ‘ die ’ arđey ‘bring’ 1sg.m merđe=na arde=na=ş [bring.pst.ptcp.m=cop.3sg:O=3sg:A] 1sg.f merđê=na ardê=na=ş [bring.pst.ptcp.f=cop.3sg:O=3sg:A] 2sg.m merđe=nî arde=nî=ş 2sg.f merđê=nî ardê=nî=ş 3sg.m merđe=n arde=n=iş 3sg.f merđê=ne ardê=ne=ş 1pl merđê=nmê ardê=nmê=ş 2pl merđê=ndê ardê=ndê=ş 3pl merđê=nê ardê=nê=ş The following examples parse the perfect verb forms. In (148), both the partici- ple and the copula agree with the S argument. In (149)–(150) they agree with the O argument. (148) bizêw menêne cîyay sayqeke we. biz( e)-êw goat.f-indf menê=ne remain.pst.pt cp.f=cop.3sg.f:S cîyay behind sayqe-( e)ke=we lightning.m-def.m.sg.dir=post ‘a goat was [accidentally ] left behind [ healthy ] from the lightning [that caused the floo d]’ [ZB.42] (149) î dêwênê î kinaçêşa bestêne. î dem.pro x dêw-ê=nê ogre.m-pl.dir=cop.3pl:S î dem.pro x kinaçê =şa girl.f.sg=3pl:A bestê=ne tie.pst.pt cp.f=cop.3sg.f:O ‘It was the ogres who had muted the girl. ’ [ JP .177] (150) hewarêşa wişkinɛnê. hewar-ê=şa summer_habitat.m-pl.dir=3pl:A wişkinɛ=nê scour.pst.pt cp.pl=cop.3pl:O ‘They scoured the summer habitats [ searching for food etc.]. ’ [ JE.3] The negation of the perfect is expressed by ne- : 303 9 V erbal infle ctional morphology and v erbal categories (151) dizîm nekerđenû hîzîm nekerdenû. girđkar nebîyena. dizî=m theft.m.sg=1sg:A ne-kerđe=n=û neg-do.pst.pt cp.m=cop.3sg.m:O=and hîzî=m adultery.m.sg=1sg:A ne-kerde=n=û neg-do.pst.pt cp.m=cop.3sg.m:O=and girđkar know__it_all ne-bîye=na neg-be.pst.ptcp.m=cop.1sg:S ‘I have not committed theft or adultery . I was not a know-it-all. ’ [ JM.14] The perfect refers to a situation that has come ab out as a result of an action in the past. (152) padşay kerdena wekêɫ. padşa-î a kerde=na king.m-sg.obl wekêɫ do.pst.pt cp.m=cop.1sg:O advocate.m ‘The king has put me in charge. ’ [ZP .99] One salient function of the perfect is its use to refer to habitual actions in the far past. The perfect may b e used here to refer to imperfective habitual activities and perfective events alike . This function of perfe ct may be considered “narrative perfect” . An entir e narrative may be built on this function of the p erfect. (153) dey çêgey ç qeyîm ʕaɫifşa kennenû. ħeywanşa wey kerdenû. he warêşa wişkinɛnê. î dega toş vînî çoɫe bîyêne . ħîç nebîyen. dey disc.pt cl çêge=îç here=add qeyîm old_time ʕaɫif=şa grass.m.sg.dir=3pl:A kenne=n =û mow.pst.pt cp.m=cop.3sg.m:O=and ħeywan=şa animal.m.sg.dir=3pl:A wey raising kerde=n =û do.pst.pt cp.m=cop.3sg.m:O=and hewar-ê=şa summer_habitat.m-pl.dir=3pl:A wişkinɛ =nê scour.pst.pt cp.pl=cop.3pl:O î dem.pro x dega village.f to=ş 2sg=3sg:O vîn-î see.prs.ind -2sg:A çoɫ-e deserted-f bîyê=ne be.pst.ptcp.f=cop.3sg.f:S ħîç nothing ne-bîye=n neg-be.pst.ptcp.m=cop.3sg.m:S ‘In the past, they (people) mowed grass. They raised animals. They scoured the summer habitats [sear ching for foo d etc.]. This village , which you see, was deserted. There was nothing [here]. ’ [ JE.1–JE.5] 304 9.4 T AM categories derived from past stem The perfect may additionally b e used to express an event that the speaker has not witnessed, but is hearsay . This is the evidentiality function of the perfect, which is also found in neighbouring languages such as Persian, T urkish, Arme- nian (Lazard 2001), and in Neo- Aramaic (Khan 2020). (154) heta min jinyenim pîye we ama xizmetû şê ʕeladînî. heta even min 1sg jinye=n=im hear.pst.pt cp.m=cop.3sg.m:O=1sg:A pîye-(ê)we man.m-indf ama come.pst.3sg:S xizmet-û service.m-ez.gen şê sheikh.m ʕeladîn-î pn-m.sg.obl ‘I have ev en heard that a man came to the service of Sheikh Aladin. ’ [ZQ .2] Another context of the evidential function of the perfect is its use in inferential contexts. In (155), the context is one in which a mute girl starts to speak. The narrator infers from this evidence that the r eason the girl was mute was be cause the ogres had muted her . (155) î dêwênê î kinaçêşa bestêne. î dem.pro x dêw-ê=nê ogre.m-pl.dir=cop.3pl:S î dem.pro x kinaçê=şa girl.f.sg=3pl:A bestê=ne tie.pst.pt cp.f=cop.3sg.f:O ‘It was the ogres who had muted the girl. ’ [ JP .177] The perfect may refer to legendar y ev ents the speaker has only learne d about from reports. This is another instance of the evidential function of the perfect since the speaker has not witnessed the event himself. (156) ca padşakey waten be lalowe , ew lalowe kinaçekê . ca afterwards padşa-( e)ke-î king-def.m.sg.obl wate=n say.pst.pt cp.m=cop.3sg.m:O be to lalo-e maternal_uncle.m-def ew dem.dist lalo=e maternal_uncle.m=dem kinaç(ê)-ekê daughter.f-def.f.sg ‘Oh, the king had said to the uncle, to [his] daughter’s uncle. ’ [ZP .43] Related to expressing evidentiality, the perfect may be use d to expr ess mirativ- ity, i.e., “marking statements based on inference and statements based on direct 305 9 V erbal infle ctional morphology and v erbal categories experience for which the speaker had no psychological preparation” (DeLancey 1997: 35–36). In (157), the perfect marks the sp eaker’s unpr epared mind and his surprise that he has just witnessed the donkey starting to talk. (157) ‘her ta îse qisêş nekerdênê . ’ her donkey ta until îse now qisê=ş talk.pl.dir=3sg:A ne-kerdê=nê neg-do.pst.pt cp.pl=cop.3pl:O ‘[He said surprisingly ], ‘the donkey hadn’t talked until now!’’ [HB.46] 9.4.4 Perfect progressive The perfect progressive is built by a r eduplicate d construction consisting of an inflected verbal form in the perfe ct pr e ceded by a double comprising the past stem and the suffix -î . T able 9.25: The p erfect progr essive–a sample paradigm S kewtey ‘to fall’ 1sg.m kewt-î kewte=na ‘I (m) have be en/had been falling’ 1sg.f kewt-î kewtê=na ‘I (f) have be en/had been falling’ 2sg.m kewt-î kewte=nî 2sg.f kewt-î kewtê=nî 3sg.m kewt-î kewte=n 3sg.f kewt-î kewtê=ne 1pl kewt-î kewtê=nmê 2pl kewt-î kewtê=ndê 3pl kewt-î kewtê=nê The negation of the perfect progressive is e xpressed by ne- . (158) newatîş newaten. ne-wat-î=ş neg-say .pst -nmlz=3sg:A ne-wate=n neg-say .pst.pt cp.m=cop.3sg:O ‘He has not be en saying [ what was not to be told]. ’ [hearsay ] The following example illustrates the use of perfect progressiv e in an interrog- ative clause . 306 9.4 T AM categories derived from past stem (159) maço, ‘ çî amay amêndê?’ m-aç-o ind -say.prs-3sg:A çî why amay come.nmlz amê=ndê come.pst.pt cp.pl=cop.2pl:S ‘He said, ‘Why have you come [here]?’’ [ŞC.35] The perfect progressive may r efer to habitual actions in the far past for which the speaker has learned only from reports. This use of perfe ct pr ogressive occurs primarily in folktales, exhibiting the evidential function of perfect progressiv e. (160) î kabrɛçe ce ʕêraqo amɛnê, wêr eganew nîmeřonew seʕbne bexşnayşa bexşnɛnêwe . î dem.pro x kabrɛ=ç=e man.pl.dir=add=dem ce from ʕêraq=o pn=post amɛ=nê come.pst.pt cp.pl=cop.3pl:S wêrega=ne=û evening=post=and nîmeřo=ne=û noon=post=and seʕb=ne morning=post bexşnay=şa distribute.nmlz=3pl:A bexşnɛ=nê =we distribute.pst.pt cp.pl=cop.3pl:R=compl ‘They (people) would donate [food] to the fellows (the tax collectors) who had come from Iraq, in the ev enings, mornings, and at noon. ’ [BP .38] Relatedly , the perfe ct pr ogressive may expr ess far past events that the speaker has not witnessed, but the events have some personal significance for the speaker . This might be called the “ experiential p erfect progressiv e” . The e vents in question wer e o ccurring continually but w ere completed at some point. In the following excerpt, the speaker talks about his life a long time ago when he was away from home and did not know what was happening to his children during his absence . (161) min eçagene karîger bîyena. şiş mangê xeberêm nezanan. bizanû kewtî kewtênê warđîşa warden dizîş dizîy en. min 1sg e=çagene in=there karîger labourer.m bîye=na be.pst.ptcp.m=cop.1sg:S şiş six mang( e)-ê month.f-pl.dir xeber-ê=m news.m-indf=1sg:A ne-zana=n neg-know.pst.pt cp.m=cop.3sg.m:PSR bi-zan-û sbjv -know.prs-1sg:A kewt-î fall.pst -nmlz kewtê=nê fall.pst.pt cp.pl=cop.3pl:S 307 9 V erbal infle ctional morphology and v erbal categories gather . They said ( according to what was reported to me) , ‘Baqî or Hasan, we will send one of them to go to the military ser vice. ’’ (Khan & Mohammadirad 2024a: 313, glossing and transcription modified) 9.4.8 Perfect plup erfect The perfect plup erfect is made of the participle form of the verb follo wed by the perfect form of the auxiliar y ‘to be’ . This T AM category seems outdate d as it was only occasionally attested in the tales narrate d by the narrator fr om Nwên. The perfect plup erfect has not be en listed as a T AM categor y in Luhon (MacK enzie 1966), reflecting its rare use . With intransitive verbs, both the participle form of the main verb , the participle form of the auxiliary ‘b e ’ , and the copula agree with the S argument of the verb , the first two in gender and number, the latter in person. T able 9.32 exhibits the perfe ct pluperfect of the verb ‘to sleep ’ . T able 9.32: Perfe ct pluperfect–the inflection of ‘sle ep ’ S Gloss 1sg.m wite=bîye=na [ sleep.pst.ptcp.m=be.pst.pt cp.m=cop.1sg:S] 1sg.f witê=bîyê=na [sleep.pst.pt cp.f=be.pst.pt cp.f=cop.1sg:S] 2sg.m wite=bîye=nî [ sleep.pst.ptcp.m=be.pst.pt cp.m=cop.2sg:S] 2sg.f witê=bîyê=nî [sleep.pst.pt cp.f=be.pst.pt cp.f=cop.2sg:S] 3sg.m wite=bîye=n [ sleep.pst.ptcp.m=be.pst.pt cp.m=cop.3sg:S] 3sg.f witê=bîyê=ne [sleep.pst.pt cp.f=be.pst.pt cp.f=cop.3sg:S] 1pl witê=bîyê=nmê [sleep.pst.pt cp.pl=be.pst.pt cp.pl=cop.1pl:S] 2pl witê=bîyê=ndê [sleep.pst.pt cp.pl=be.pst.pt cp.pl=cop.2pl:S] 3pl witê=bîyê=nê [sleep.pst.pt cp.pl=be.pst.pt cp.pl=cop.3pl:S] 314 9.4 T AM categories derived from past stem With transitive v erbs, the participles in b oth the main v erb and the auxiliar y agree in gender and number with the O argument. On the other hand, cumu- lative person/number copula endings agree in person/numb er with the O argu- ment. The paradigm in T able 9.33 features the O argument appearing in different persons in combination with the 3pl A argument, hence ‘they have had brought me (m)’ , ‘they have had brought me (f)’ , etc. T able 9.33: Perfe ct pluperfect–the inflection of ‘bring’ O 3pl A Gloss 1sg.m arđe=bîye=na=şa [bring.pst.pt cp.m=b e.pst.pt cp.m=cop.1sg:O=3pl:A] 1sg.f arđê=bîyê=na=şa [bring.pst.ptcp.f=be.pst.pt cp.f=cop.1sg:O=3pl:A] 2sg.m arđe=bîye=nî=şa [bring.pst.pt cp.m=b e.pst.pt cp.m=cop.2sg:O=3pl:A] 2sg.f arđê=bîyê=nî=şa [bring.pst.ptcp.f=be.pst.pt cp.f=cop.2sg:O=3pl:A] 3sg.m arđe=bîye=n=şa [bring.pst.pt cp.m=b e.pst.pt cp.m=cop.3sg:O=3pl:A] 3sg.f arđê=bîyê=ne=şa [bring.pst.ptcp.f=be.pst.pt cp.f=cop.3sg:O=3pl:A] 1pl arđê=bîyê=nmê=şa [bring.pst.ptcp.pl=be.pst.pt cp.pl=cop.1pl:O=3pl:A] 2pl arđê=bîyê=ndê=şa [bring.pst.ptcp.pl=be.pst.pt cp.pl=cop.2pl:O=3pl:A] 3pl arđê=bîyê=nê=şa [bring.pst.ptcp.pl=be.pst.pt cp.pl=cop.3pl:O=3pl:A] The negation of the perfect plup erfect is expressed by ne- . (171) newitebîyena. ne-wite=bîye=na neg-sleep.pst.pt cp.m=b e.pst.pt cp.m=cop.1sg:S ‘I have had not slept. ’ (Pseudo-English translation) The perfect plup erfect seems to occur only in tales. It may express actions which have started in the past but continue to impact the present state of affairs. (172) gîr wardêbîyênê ana ça matiɫê sergerdanê . gîr hook wardê=bîyê=nê eat.pst.pt cp.pl=be.pst.ptcp.pl=cop.3pl:S ana-Ø loc.deic.cop -3sg.m:S ça there matiɫ-ê waiting-pl sergerdan-ê wandering-pl ‘They wer e stuck there; they are ther e, waiting and wandering [not knowing what to do]. ’ [KT .54] 315 9 V erbal infle ctional morphology and v erbal categories 9.4.9 Summary of T AM categ ories derived from the past stem T able 9.34 summarises the verbal forms derived from the past stem for the verbs ‘sleep ’ and ‘do ’ inflected in the first person. T able 9.34: T AM categories derived fr om the past stem–summar y T AM category Infle ction Gloss Past perfective wit-a [sleep.pst -1sg:S] Past conditional wit-εn-ê [sleep.pst.cond.a ug-1sg:S] Perfect wite=na [sleep.pst.pt cp.m=cop.1sg:S] Perfect progressive wit-î wite=na [sleep.pst -nmlz sleep.pst.pt cp.m=cop.1sg:S] Irrealis perfect wite=b-û [sleep.pst.pt cp.m=b e.prs-1sg:S] Conditional perfect wite=bî-ɛn-ê [ sleep.pst.ptcp.m=be.pst - cond.a ug-1sg:S] Past perfect wite=b-ên-ê [sleep.pst.pt cp.m=be-a ug-1sg:S] Perfect pluperfe ct wite=bîye=na [ sleep.pst.ptcp.m=be.pst.pt cp.m=cop.1sg:S] Past perfective kerđ-∅=im [do.pst -3sg.m:O=1sg:A] Past conditional kerđ-ε=m [do.pst -cond. a ug.3sg:O=1sg:A] Perfect kerđe=n=im [do .pst.ptcp.m=cop.3sg.m:O=1sg:A] Perfect progressive kerđ-î kerđe=n=im [do.pst -nmlz do.pst.pt cp.m=cop.3sg.m:O=1sg:A] Irrealis perfect kerđe=b-o=m [do.pst.pt cp.m=be.prs-3sg:O=1sg:A] Conditional perfect kerđe=bî-ε=m [ do.pst.pt cp.m=b e .pst-cond.a ug.3sg:O=1sg:A] Past perfect kerđe=b-ê=m [do .pst.ptcp.m=be.a ug.3sg:O=1sg:A] Perfect pluperfe ct kerđe=bîye=n=im [ do.pst.pt cp.m=be.pst.pt cp.m=cop.3sg:O=1sg:A] 316 10 Adpositions and adverbs 10.1 Adpositions The category of adp ositions includes prepositions, postpositions, and circumpo- sitions. This means that Hewramî has a mixed adpositional typology . This trait is a common feature of Northw estern Iranian languages such as V afsi and most varieties of Kurdish. It is assumed to be a reflection of their ge ographical distri- bution between O V languages, e.g., Armenian, T urkish, Indic, and V O languages, e.g., Arabic, Aramaic ( cf. Stilo 2009: 18). Additionally , T ekht Hewramî features formativ es of (mainly) pr ep ositional ori- gin, commonly called “absolute adpositions” within Iranian linguistics. As dis- cussed in §10.1.4, one way of analysing these formatives is in terms of applicative morphology . 10.1.1 Prepositions The majority of adpositions b elong to the category of prepositions. Preposi- tions have both grammatical and spatial-temporal functions. They usually trig- ger oblique marking on the nominal that follows them, though oblique marking is contingent on certain factors, e.g., animacy and pre- and post-verbal position- ing of the prepositional phrase. By way of e xample, in (1)–(2), the preposition be marks the instrument. Howev er , only the animate complement is marked in the oblique case. Investigating factors influencing differential oblique marking remains a topic for feature research. (1) miqdarêş ađî be sey misefay ardeniş, miqdarê=ş some=3sg:PSR ađî 3sg.obl.m be by sey pn misef a-î pn-m.sg.obl arde=n=iş bring.pst.pt cp.m=cop.3sg.m:O=3sg:A ‘Some of the money which he had made Say Mustafa take . [Lit. He brought some of it by Say Mustafa .]’ [ JP .113] 10 A dpositions and adverbs (2) be kune awîşa ardêne. be by kune clay_pot.m awî=şa water.f.sg.dir=3pl:A ardê=ne bring.pst.pt cp.f=cop.3sg.f:O ‘They used to fetch water using clay pots . ’ [ JE.16] Prepositions fall into two sets: basic pr epositions, and lo cational nouns. The locational nouns can b e used independently or in combination with prepositions. These two sets are morphologically distinguishe d. Simple prepositions occur with nominals through simple juxtaposition. Locational nouns are linked to the nominal via a linker dubbe d “genitiv e ezafe” (§5.3.1). 10.1.1.1 Basic prepositions The following is the list of basic pr ep ositions, most of which ar e monomor- phemic. (3) be ‘to, by , with, in’ ce ‘at, from’ pey ‘for , to ’ çenî/çenû ‘with’ bê ‘without’ ta ‘until’ e= ‘in, from’ ne ‘into ’ (occurring rarely , see §10.1.2) Among these, bê and ta occur only with non-bound complements, i.e., with nouns and independent pronouns, which are not pr oso dically deficient. On the other hand, be and ce are unique in that they exhibit allomorphy when attach- ing to demonstratives and clitic pr onouns. When combine d with demonstra- tives, the initial consonants are devoiced, yielding the forms ç= and p= , e.g., ç=î mentêqe-î [in=dem.pro x region.m-sg.obl] ‘in this region’ [PM.5]; p=a mê- man mizgî [to=dem.dist guest.m mosque.m] ‘to the guest in the mosque ’ [ JH.40]. The resulting forms sometimes grammaticalise into spatial adverbs: ça ‘ther e ’; çê ‘here ’ . When combined with clitic pronouns be and ce are realised as pene and çene , respectively , e.g., pene=m ‘to me ’; çene=ş ‘in it’ . In traditional Iranian philol- ogy , these allomorphs are r eferred to as “absolute prepositions” (see MacKenzie 1966: 55 for Hewramî Luhon, McCarus 2009: 601 for Central Kurdish, and Lazard 1992 for Laki). çenî/çenû ‘with’ exhibits a slightly different pattern: its “absolute ” form is çene . T able 10.1 summarises the allomorphs of b e , ce , and çenî/çenû . 318 10.1 Adpositions T able 10.1: The allomorphs of b e and ce and çenî with nouns with demonstratives with clitic pronouns be b e p= pene ce ce ç= çene çenî/çenû çenî/çenû çenî/çenû çene Basic prepositions occur before nominals through simple juxtaposition. An exception is çenû ‘with’ , which is a combination of çenî and the genitive ezafe form -û . The form çenû is now grammaticalised as the citation form for most speakers. The original form çenî seems to b e less frequent. 10.1.1.1.1 be ‘to, by , with, in’ This preposition has a range of meanings. It typically marks the following roles: recipient (4), addressee (5), (metaphorical) goal (6), comitative (7), beneficiar y (8), passive agent (9), and manner (10). 1 (4) qurban to desûrêt da be sey misefay . qurban sir .v oc to 2sg desûr-ê=t order.m-indf=2sg:A da-Ø give.pst -3sg.m:O be to sey pn misef a-î pn-m.sg.obl ‘Sir , you or dered Say Mustafa [ lit. gav e an order to]. ’ [ JP.126] (5) maço be xanî. m-aç-o ind -say.prs-3sg:A be to xan-î chief.m-sg.obl ‘He said to the chief . ’ [KŞ.97] (6) xeber yawo be baba xwađaw ħemey ẍeybî. xeber news.m.sg.dir yaw-o arrive.prs.ind-3sg:S be to baba pn xwađa=û pn=and ħeme-î pn-ez.a ttr ẍeybî invisible ‘The news reached Baba Khwada and Hama the Invisible . ’ [BP.46] 1 See Mohammadirad & Rasekh-Mahand (2017, 2018) for the study of the multifunctionality of Hewramî instrumental and dative markers within the conte xt of Iranian languages. The list of functionalities of adpositions provided here is more compr ehensive. 319 10 A dpositions and adverbs (7) êtir çayşa kerden be nanekey . êtir disc.pt cl çay=şa tea.m.sg.dir=3pl:A kerde=n do.pst.pt cp.m=cop.3sg.m:O be with nan-ekey bread.m-def.m.sg.obl ‘Then, they made tea [to be ser v e d] with the food. ’ [ JE.38] (8) xizmet be merđimî kero. xizmet service.m be to merđim-î people.m-sg.obl ker-o sbjv -do.prs-3sg:A ‘[ And he] be of ser vice to people . ’ [ JP .87] (9) sewzê xel kiryan b e sahêbîçiş. sewzê crop xel grain kir-ya=n do.prs-p ass=cop.3sg.m:S be by sahêb -î=ç=iş owner-obl.m=add=3sg:PSR ‘The crop has been pile d up [lit. turned into corn] by its owner too. ’ [HR.16] (10) be adizî luwan. be in adizî desperation luwa=n go.pst.pt cp.m=cop.3sg.m:S ‘He had left [town] in desperation. ’ [ JH.62] be may also mark the nominal argument of inchoative verbs ‘become, turn into ’ , indicating a change of state. (11) bîyen be feqî. bîye=n be.pst.ptcp.m=cop.3sg.m:S be adp f eqî theologian.m ‘[He] be came a theologian. ’ [ZP.15] (12) miniş kerdena be wekêɫ. min=iş 1sg=3sg:A kerde=na do.pst.pt cp.m=cop.1sg:O be adp wekêɫ advocate.m ‘He has put me in charge. [Lit. made me adv o cate .]’ [ZP.89] It should be mentione d, though, that the use of the pr eposition is optional here , as the nominal argument of inchoative verbs may be realised without any flagging: 320 10.1 Adpositions (13) min, taze padşay kerdena wekêɫ. min 1sg taze anyway padşa-î king.m-sg.obl kerde=na do.pst.pt cp.m=cop.1sg:O wekêɫ advocate.m ‘ As for me , well, the king has put me in charge. ’ [Lit. The king has made me advocate.] [ JP.206] (14) bawe yo . keraşawe yo b-a=we be come.prs.ind-3pl:S=compl yo one.m ker-a=şa=we do.prs.ind -3pl:A =3pl:O=compl yo one.m ‘They became one. They wer e put in the same [cell] [lit. they made them one]. ’ [BP .141] 10.1.1.1.2 pey ‘for , to ’ This preposition generally triggers oblique case marking on its complement. It typically expresses a beneficiar y argument and a goal argument of a verb of movement, whether human or inanimate . (15) hêɫeřûwenîşa kerde pey sahêbû her eyû pey herbenew herey . hêɫeřûwenî=şa fried_eggf.sg.dir=3pl:A kerd-e do.pst -3sg.f:O pey for sahêb -û owner.m-ez.gen her -e-î =û donkey.m-def-m.sg.obl=and pey for herbene-û donkey_keeper.m-ez.gen her -e-î donkey.m-def-m.sg.obl ‘They (the host’s family ) cooke d fried eggs for the donkey’s o wner [i.e.] for the donkey keeper . ’ [HB.56] (16) luwew e pey yaney! lu-e=we go.prs.imp-2sg:S=post pey to yane-î house.m-sg.obl ‘[Now ] go back home!’ [ JH.118] (17) î kinaçê narîwe êtir pey min. î dem.pro x kinaçê girl.f n( e)-ar-î=we neg.sbjv -bring.prs-2sg:A =compl êtir any_more pey to min 1sg ‘[Please] do not bring this daughter [of mine] back to me. ’ [ZP.48] 321 10 A dpositions and adverbs The nominal complement of pey occurs regularly in the oblique case. In rare cases, the complement appears in the direct case. (18) kîyasenşa pey beẍa. kîyase=n=şa send.pst.pt cp.m=cop.3sg.m:O=3pl:A pey to beẍa pn.m ‘[Then] they sent him to Baghdad. ’ [ZP.14] 10.1.1.1.3 çenî ‘with’ This preposition occurs in two forms in the corpus: çenî and çenû . The latter results from the unimorphation of çenî with the genitiv e ezafe -û . çenî expresses a comitative argument. The complement is generally marked in the oblique case. (19) î masî bere deş pa mêman mizgî çenîyû nanî. î dem.pro x mas-î yoghurt.m-sg.obl bér-e take.prs.imp -2sg:A dé-( e)=ş give.prs.imp-2sg:A =3sg:O p=a to=dem.dist mêman guest.m mizgî mosque.m çenî-û with-ez.gen nan-î bread.m-sg.obl ‘T ake this yoghurt [and] together with [ some] bread [and] giv e it to the guest in the mosque. ’’ [ JH.40] (20) nîşare çenû lalow ekeyş. nîş-a=re sit.prs.ind -3pl:S=po vb çenû with.ez.gen lalo-ekey=ş maternal_uncle.m-def.m.sg.obl=3sg:PSR ‘She sat [ on the stone] with her uncle. ’ [ZP .56] 10.1.1.1.4 ce ‘at, from’ This preposition takes an oblique argument in its spatial sense, ‘in, from’ or in the sense of ‘ of’ . ce is also used to express the standard of comparison; see (23). 322 10.1 Adpositions (21) maça neferê înarê ce menteqew hewramanatî. m-aç-a ind -say.prs-3pl:A nefer-ê person.m-indf îna-∅=rê loc.deic.cop -3sg.m:S=po vb ce in menteqe-û region-ez.gen hewraman-at-î pn-pl-m.sg.obl ‘It was said [to him in the dream] that there is a person in the Hewraman region. ’ [ZP.31] (22) ane ce zemînî pêse waçî xway ketê pey kîyasen. ane dem.dist.m.3sg.dir ce from zemîn-î earth.m-sg.obl pêse as_if wáç-î say.prs.sbjv -2sg:A xwa-î God.m-sg.obl ket-ê be d-indf pey to kîyase=n send.pst.pt cp.m=cop.3sg.m:R ‘[He had levitated] this much from the ground, as if God had sent him a be d. ’ [ JP .69] (23) ce taranî kuweyt de wɫetmenter bê. ce from taran-î pn-m.sg.obl kuweyt pn dewɫetmen-ter rich- cmpr b-ê be.prs-a ug.3sg:S ‘Kuwait was more affluent than T ehran. ’ [ JM.44] 10.1.1.1.5 bê ‘without’ This preposition marks peripheral arguments which are not in the verb ’s argu- ment structure . The preposition has a privative sense. It usually triggers oblique marking on the nominal complement. (24) bê a bega ême netawanma ħîç kermê. bê without a dem.dist beg-a chief.m-pl.obl ême 1pl ne-tawa=n=ma neg-can.pst.pt cp.m=cop.3sg.m:O=1pl:A ħîç nothing kér-mê do.prs.sbjv -1pl:A ‘Without those chiefs, we wer e not able to do anything. ’ [RE.64] (25) mîyo bê heyasîş mekiryo. mi-đy( e)-o ind -look.prs-3sg:S bê without heyas-î =ş pn-m.sg.obl=3sg:NC me-kiry(e)-o neg.ind-do.do.prs-p ass-3sg:S ‘He realised that he cannot cope without Hayas. ’ [HS.19] 323 10 A dpositions and adverbs (52) kewt ne şarî. kewt fall.pst.3sg.m:S ne into şar-î town-m.sg.obl ‘He stormed into town. ’ [ÇH.167] (53) kewt ne kora. kewt fall.pst.3sg.m:S ne into kor-a blind-pl.obl ‘He rushed into blind pe ople. ’ [DB.309] 10.1.3 Circumpositions Circumpositions occur frequently in Hewramî. They ar e constructe d mainly in two ways: by combining a basic preposition and a postposition, e.g., ce ... ne ‘in, (less so) from’ or by combining a locational noun and a p ostposition, e.g., mil ... hur ‘ on top ’ . There are also more complex forms such as pêw ere ‘together’ con- taining the circumposition be .... ere combine d with the indefinite suffix -êw e . Within basic prepositions, be and ce can form circumpositions with p ostposi- tional elements. The circumpositions may serve to disambiguate the multiple meanings of basic prepositions. For instance, ce has both locational and source meanings. The addition of postp ositions ne and o/e w helps distinguish b etween these spatial senses. (54) ce ... ne ‘in, ( less so) from’ ce ... o/we/ew ‘from’ be ... we ‘about’ be ... ne ‘ on’ be ... re ‘in’ dilê ... re ‘in’ dilê ... ne ‘in’ The circumpositions listed in (54) differ in the rate to which they trigger oblique marking on their complement (Mohammadirad in re view). For instance, the complement of be ... we (55) is more likely to be oblique-marke d than the complement of ce ... ne (56) and ce ... o/we (57). 330 10.1 Adpositions (55) min înem dî be sey misefay w e. min 1sg îne=m dem.pro x.m.3sg.dir=1sg:A dî-∅ see.pst -3sg.m:O be about sey pn misefa-î=we pn-m.sg.obl=post ‘I witnessed this ab out Say Mustafa. ’ [ JP.74] (56) ce hewramanne karî naşerʕî kerênê. ce at hewraman=ne pn=post kar-î task.m-ez.a ttr naşerʕî unlawful ker-ên-ê do.prs- a ug-3pl:A ‘They wer e committing unlawful acts in Hewraman. ’’ [BP. 62] (57) ce qeredaxo maroş ta heɫebce. ce from qeredax=o pn=post m-ar-o=ş ind -bring.prs-3sg:A =3sg:O ta until heɫebce pn ‘He took his brother from Qaradax to Halabja. ’ [DG.29] Circumpositions consisting of locational nouns and a postp osition comprise a far bigger class than circumpositions consisting of a basic preposition and a postposition, see (58). This could be due to the bigger size of the class of lo cational nouns. (58) ber ... ne ‘in front of’ [HB.73] beyn ... ne ‘in between’ [BP .178] çêr ... ne ‘under’ [BP .194] dil ... ne ‘inside ’ [KŞ.57] dewr ... ne ‘at around’ [ZP .67] dewr ... hur ‘around’ la ... we ‘at the place of’ [ZB.59] la ... o ‘with’ [ JH.1] la ... ne ‘at the place of’ [HB.73] la ... re ‘ne xt to’ ce la ... we ‘to, at the place of’ [BP .211] mil ... we ‘be cause of, ov er’ [ JM.27] mil ... re ‘ on’ [ JP .107] mil ... hur ‘on top ’ payîn ... re ‘do wn at’ [ZQ .5] qew ... we ‘ on, around’ [HB.61] ser ... we ‘ on top of’ [BP .153] 331 10 A dpositions and adverbs şûn ... re ‘after’ [ JM.4] şûn ... o/we ‘b ehind’ [DP .8] teref ... we ‘from’ [PM.34] werawer ... w e ‘in front of’ [ JP .201] The nominal complement of the circumpositions listed in ( 58) may be marke d in the oblique case. (59) înê dilê namekeyne minvîso . înê dem.pro x.3pl dilê inside name-( e)key=ne letter.m-def.m.sg.obl=post mi-nvîs-o ind -write.prs-3sg:A ‘He wrote these [ words] inside the letter [ and gave it to the shepherd’s son]. ’ [KŞ.57] (60) min la şê xîwe menîşû. min 1sg la with şêx-î=we sheikh.m-sg.obl=post me-nîş-û neg.ind -sit.prs-1sg:S ‘I won’t sit together with the sheikh [ any more!]. ’ [HB.67] (61) da luwe şûnû se y misefayre da hort lu-e go.prs.imp-2sg:S şûn-û after-ez.gen sey pn misefa-î=re pn-m.sg.obl=post ‘Go after Say Mustafa!’ [ JP.43] The components of some circumpositions may b e grammaticalised into com- pound adpositions. The resultant forms often convey spatial and temporal adver- bial meanings (see §10.2.1). (62) beynne ‘in-between’ [RE.25] serew ‘fr om top ’ [BP .152] dilne ‘inside ’ [ JE.26] çêrwe ‘from underneath’ [KŞ.31] çêrhur ‘from below facing upwar ds’ [ZB.45] şûnîre ‘afterwards’ [ JM.6] 10.1.4 Absolute adpositions As hinted in §10.1.1.1, T ekht Hewramî features a set of formatives that ar e mor- phological allomorphs of basic adpositions. These formatives have been lab elled 332 10.1 Adpositions “absolute prepositions” in works on Iranian languages. Contrary to all other ad- positions seen so far , which can b e used with prosodically independent com- plements (i.e., independent pronouns, nouns), the complement of some of the ‘absolute adpositions’ can only b e a bound pronominal element, as seen in the difference between be (63) and p ene (64). A dditionally , absolute adpositions are associated with a sp ecial syntax (see below ). (63) ne watim b e to? ne-wat=im neg-say .pst=1sg:A be to to 2sg ‘Didn’t I say [that] to you?’ [P W .87] (64) watim peneş wat=im say.pst=1sg:A pene=ş to=3sg:R ‘I said to him. ’ [ JH.31] T able 10.2 demonstrates the absolute form of adpositions found in the main text corpus. A s can b e seen, adpositions can be classifie d into four general sets concerning their ‘absolute forms’: (1) formatives which are derivable by trans- parent phonological process fr om their basic forms, e.g., be vs. p ene . Those adpo- sitions that have an absolute form will go into this form when the adpositional T able 10.2: Adpositions and their corresponding absolute forms Adposition/locational noun Gloss Absolute form be ‘to, by’ pene — ‘with’ pene be ... re ‘ on’ pore be ... we ‘ on’ powe pey ‘for’ pey ce , ce ... ne ‘from’ çene çenî ‘with’ çene la ... we ‘to ’ la ... we / lawe mil ... re ‘to ’ mil ... re / milre ser ... o ‘ on top ’ ser ... o / sero bê ‘without’ — ta ‘until’ — ser ‘ on, to ’ ser 333 10 A dpositions and adverbs complement is a bound pronominal argument, rather than an NP (as seen in the difference between (63) and (64); (2) formatives for which the basic form and the absolute form are the same, see pey , ser ; (3) formatives with no basic forms, see pene ‘with’; (4) formatives with no absolute forms, e.g., bê , ta . Absolute adpositions are associated with a sp ecific syntax. They can only take bound pronominal arguments, realised through different markers depending on the tense and transitivity of the clause. Ther e is an additional complication, how- ever . Due to independent processes of clitic movement, the clitic pronoun may move fr om its governing adposition and attach to another element of the clause. Crucially , the adposition remains in the absolute form even if its complement has moved off it to attach to a distinct host. In clauses with v erbs deriving from the present tense , clitic movement is generally leftward and is r ealised on the first element of the VP . The VP-initial element is an object NP in (65), a light verb complement in (66), and a verb in (67), on which the preposition complement lands. 3 (65) nanşa pey peya kero . nan= şa bread.m=3pl:R pey for peya visible ker-o do.prs.ind -3sg:A ‘He finds food for them. ’ [BP .148] (66) bawiřiş pene kero . bawiř= iş belief =3sg:R pene in ker-o do.prs.ind -3sg:A ‘He (Hama Y oso) b elie ved him. ’ [ JP .59] (67) bero miđoş pene. ber-o take.prs.ind -3sg:A mi-đ( e)-o= ş ind -give.prs-3sg:A =3sg:R pene to ‘She took [it and] gave it to him (Hayas). ’ [ JH.43] In the examples above , the VP-initial element is immediately realised to the left of the absolute prepositions. The clitic complement of the adposition can still land on the VP-initial element provided the host is at a reasonable distance 3 Discussing a similar phenomenon in neighbouring Central Kurdish dialects, Karim & Salehi (2022) propose an applicative analysis of absolute prepositions. While insightful, their analysis runs into problems in cases like (65), where the absolute form is not a derivational morpheme on the verb , which is one of the defining features of applicative morphemes (Pacchiar otti & Zúñiga 2022). Notice that nor in the corresponding Central Kurdish rendering of (65) is the formative a derivational pre verb on the root: nan=yan bo peya e-ka . 334 10.1 Adpositions from the preposition; see the difference between (68)–(69), on the one hand, and (7 0) on the other . In (7 0) the clitic pronoun can, in principle , move leftward and be realised on esb-î . The lack of clitic movement may suggest that the mobility is not obligatory . (68) min şûyş kerû pene. min 1sg şû-î= ş husband.m-sg.obl=3sg:R ker-û do.prs.ind -1sg:A pene to ‘I will marr y him. ’ [ JH.59] (69) y ewêş kera sero. yew( e)-ê= ş lucerne-f.sg.obl=3sg:R ker-a do.prs.ind -3pl:A ser=o on=post ‘They will put [ some] lucerne on top of the straw . ’ [HB.37] (7 0) a esbî zînî ker e p eym. a dem.dist esb-î horse-m.sg.obl zînî saddle kér-e do.prs.imp -2sg:A pey=m for=1sg:R ‘Saddle up the horse for me. ’ [ŞC.53] Likewise , the clitic complement of the adp osition is r ealised in situ when the adposition is the first element within the VP. Note that the element before pey is the verb of the preceding clause, hence not counted as the immediate VP-internal element to the left as a host for the clitic pronoun. (71) çowe e w mê peym maro. ç=o=we in=dem.dist=post ew dem.dist m-ê ind -come.prs.3sg:S pey=m for=1sg:R m-ar-o ind -bring.prs-3sg:A ‘From that dir e ction, he comes [to me] and brings [food] for me. ’ [PM.46] Similarly , with intransitive verbs derived from the present tense and frequently also in the past tense, the complement of the adposition is realised as a clitic pronoun. Her e, the clitic floats leftward. The leftwar d movement takes ev en the S argument (72) or the copula subject (73) as host. This could be interpreted as a remnant of the original clausal second-position rule for clitic placement in Hewramî. The leftward mo vement of the clitic pronoun in intransitive clauses may also target the verb as the host, as shown in (74). 335 10 A dpositions and adverbs (72) haminiş aman milre . hamin= iş summer.m=3sg:R ama=n come.pst.pt cp.m=cop.3sg.m:S milre on ‘Then it be came summer . ’ [Lit. Summer came upon him.] [DP.10] (73) eger minit çene bîyɛnê ... eger if min= it 1sg=2sg:R çene with bî-ɛn-ê be.pst.cond.a ug-1sg:S ‘If I had be en with y ou ... ’ [PW .88] (74) kuřekêş germîyanne nimêniş çene . kuř-ekê=ş son.m-def.pl.dir=3sg:PSR germîyan=ne pn=post nim( e)-ê-nê= iş neg.ind -come.prs-3pl:S=3sg:R çene with ‘His sons did not accompany him [they stayed in] Garmiyan. ’ [ZB.9] The adposition complement is realised as a person index in the following ex- ample featuring a past intransitive construction. (75) diɫim şîyenî pene. diɫ=im heart=1sg:PSR şîye=n- î go.pst.pt cp.m=cop.3sg:S=2sg:R pene to ‘I am fond of you. [Lit. My heart has gone to you.]’ [SK.06] In clauses based on a past transitive verb , the adposition complement do es not move leftwar d. Instead, it can b e r ealised through a verbal person affix or a copula PM on the verb ( depending on the T AM category of the verb), with the same feature values. If (7 6) were in the present tense , the complement of p e y would mov e leftward to land on ketê . (7 6) xway ketê pey kîyasen. xwa-î God.m-sg.obl ket-ê be d-indf pey to kîyase= n send.pst.pt cp.m=cop.3sg.m:R ‘[ As if] God had sent him a bed. ’ [ JP. 69] (77) îna qaqezêçiş dana pey . în( e)=a dem.pro x.m.3sg.dir=pt cl qaqez-ê=ç=iş letter.m-indf=add=3sg:A da= na give.pst.pt cp.m=cop.1sg:R pey to ‘Look, he has given me a letter . ’ [KŞ.84] 336 10.2 Adv erbs 10.2 Adv erbs The class of adverbs is composed of non-derived and derived adverbs. Adv erbs are typically formed from a combination of an adpositional or a demonstrative element and a nominal. Among adverbial expr essions, manner adverbs can also be use d adjectivally . 10.2.1 Spatial adverbs Spatial adverbs are classified based on having a demonstrative in their structure or not. Following Diessel (1999), the forms based on demonstratives are called ‘local adverbial demonstratives’ here . Spatial adverbs based on demonstratives are listed in T able 10.3 (see §6.3.3 for discussion). T able 10.3: Spatial adverbs based on demonstratives ‘here ’ çêge , êge < ç ‘in’ + ê ‘this’ + -ge ‘there (immediate context) çage < ç ‘in’ + a ‘that’ + -ge ‘there ( distant but visible)’ çoge, oge < o ‘that’ + -ge ‘there (invisible)’ pagene, epage we ‘ down here ’ epêre ‘ down there (invisible)’ epare ‘up there ( visible)’ e wagehur ‘up there (invisible)’ epagehur Spatial adverbs not derived fr om demonstratives are formed from a combina- tion of spatial nouns and often a postposition. In addition to adverbial use, they may have adpositional uses (see §10.1.3). The most common spatial adverbs are listed in (78). (78) beynne ‘in-between’ [RE.25] serew ‘fr om the top ’ [BP.152] dilne ‘inside ’ [ JE.26] çêrwe ‘from underneath’ [KŞ.31] çêrhur ‘from below facing upwar ds’ [ZB.45] berewe ‘ outside ’ ce dûrew ‘from far away’ [DP .4] The following examples illustrate the use of spatial adv erbs. 337 10 A dpositions and adverbs (79) çêrwe kuřekey memew bizekêş war d. çêrwe from_underneath kuř-ekey boy.m-def.m.sg.obl meme-û breast.m-ez.gen bize-( e)kê=ş goat.f-def.f.sg=3sg:A ward eat.pst ‘From underneath, the boy fed from the goat’s udder . ’ [KŞ.31] (80) serew des ker o gireway . serew from_above des hand.m ker-o do.prs.ind -3sg:A gireway cry.inf ‘He (Little Hama) started to cr y from the top [ of the roof]. ’ [BP.152] (81) kabrayç ama berew e. kabra=yç fellow=add ama come.pst.3sg:S berewe outside ‘The fellow came outside . ’ [KK.34] 10.2.2 T emp oral adverbs T emp oral adverbs ar e classifiable into five categories: general deictic adverbs (§10.2.2.1); time-of-day adverbs (§10.2.2.2), calendrical adverbs (§10.2.2.3), calen- drical cyclic adverbs (§10.2.2.4), and other temporal adverbs (§10.2.2.5). These categories present different morphological pr operties. 10.2.2.1 General deictic adverbs General deictic adverbs may , in principle, inflect for case, as se en in T able 10.4, though some occur only in one case in the text corpus. T able 10.4: General deictic adverbs dir obl ‘now’ îse îsey ‘back then, in the past’ çaweɫ çaw eɫî < * ç= a eweɫ ‘from that beginning’ ‘at that time, then’ ewsa - ‘after some time ’ - ça dimay ‘afterwards’ şûnîre - ‘before now’ - çêw eɫî The following examples e xhibit the use of çaweɫ and çaweɫî . 338 10.2 Adv erbs (82) çaweɫ ta cawe nelabê pane pêwyê tewenekê . çaweɫ in_the_past ta until cawe road ne-la=b-ê neg-go.pst.pt cp.m=be.prs-a ug.3sg:S p=ane at=dem.dist.m.3sg.dir pêwy(e)-ê be_visible.prs-a ug.3sg:S tewen(î)-ekê stone.f-def.f.sg ‘In the past, when no road was constructed there [lit. The road had not gone there], the stone was visible . ’ [ZP.54] (83) çaweɫî sextî nebîy en. çaweɫ-î in_the_past-m.sg.obl sextî difficulty.m.dir ne-bîye=n neg-be.pst.ptcp.m=cop.3sg.m:S ‘In the past, there was no hardship . ’ [ JE.56] 10.2.2.2 Time-of-day adverbs Time-of-day adverbs have the base forms as follows: (84) seʕbe ‘morning’ nîmeřó ‘midday’ wêrega ‘ ev ening’ şewe ‘night’ When used in the sense of temporal lo cation, they may be follow ed by a p ost- position or case marking. Note that the oblique case is blocke d before the post- position =ne (see §10.1.2). (85) seʕbê, seʕbne ‘in the morning’ nîmeřone ‘in the midday’ wêregane ‘in the ev ening’ şewê ‘at night’ Examples: (86) î kabrɛçe ce ʕêraqo amɛnê, wêregane w nîmeřonew seʕbne bexşnayşa bexşnɛnêwe . î dem.pro x kabrɛ=ç=e man.pl.dir=add=dem ce from ʕêraq=o pn=post amɛ=nê come.pst.pt cp.pl=cop.3pl:S wêrega=ne=û evening=post=and nîmeřo=ne=û noon=post=and 339 11 Grammatical relations This chapter lays out argument indexing and argument flagging. Following Haspelmath (2013), I use the term ‘indexing’ as a cov er term for the agree- ment phenomenon encompassing both “grammatical agreement” and “anaphoric agreement” . The term agreement is reserved for cases where the index obligato- rily indexes an argument regar dless of the presence or the absence of the coref- erent NP in the clause . Agreement in this sense parallels what Haig (2018) refers to as “ obligator y” inde xing. On the other hand, the term “alternating indexing” is used to refer to the occurrence of indexes as anaphora. This refers to cases wher e contextual factors have an effect on the presence or absence of the person index, e.g., the pr esence or absence of the co-referent NP , wor d order, clausal focus, etc. The term flagging is reserved for case marking on clausal arguments and mark- ing by adpositions. Thus, a clausal argument can b e flagged by an oblique case, an adposition, or b oth. Follo wing Hasp elmath (2019), argument flagging and ar- gument indexing are together called argument marking. Some other terminological conventions are in order . In discussing alignment types, I use the following abbreviations to r efer to the core arguments within the clause: S_ single argument of an intransitive verb; A_ agent-like argument of a transitive clause: P_ patient-like argument of a transitiv e clause; T_theme- like argument in ditransitive clauses; R_r e cipient-like argument in ditransitiv e clauses. The term R here is extended to non-core oblique arguments, including recipients, beneficiaries, p ossessors, and comitativ es. Hewramî is a language with a split ergative system, wher e ergative align- ment is conditioned by tense. The alignment is accusative in clauses built on the present stem of the v erb, but ergativ e in clauses built on the past stem of the verb . Alignment in Hewramî is r elevant in terms of agreement and case mark- ing. While the alignment system is of a tense-sensitive type, ther e are certain language-internal deviations from tense-sensitiv e alignment, b oth in terms of agreement and case marking. In terms of agreement, the system offers differen- tial A indexing (§11.1.2.1) and differential P inde xing (§11.1.2.2), limite d to transi- tive clauses built on past stem v erbs. As for case marking, differential A flagging in transitive clauses built on past stem verbs (see §11.2.1, and differential P flag- ging, most notably in clauses built on the present stem of the verb (§11.2.2), ar e 11 Grammatical r elations attested. Hewramî also features differential case marking on non-cor e arguments (§11.2.3). The full investigation of differential argument inde xing and differential argument flagging requires detailed corpus-based studies. Here, I lay out some basic patterns. 11.1 Argument indexing Argument indexing has accusative alignment in T AM categories based on the present stem of the verb and ergativ e alignment in T AM categories base d on the past stem of the verb . Agr eement is manifeste d by using different inde xes for indexing S, A, and P , summarised in T able 11.1. The b ound indexing of R is the same as the indexing of P in present tense constructions. In past tense constructions, it could be realised as a verbal person/numb er suffix or a clitic pronoun, both being alternating indexes. Recall from §9.1.7 that the verbal person suffixes have partially different paradigms in the present and past tenses. On the other hand, one and the same paradigm of clitic pronouns expr esses different clausal, e.g., A, P , and phrasal arguments, e .g., p ossessor ( see §6.2.1 for detaile d discussion). T able 11.1: Morphological indexing of arguments and indexing type T AM Argument Index Indexing type prs stem S, A verbal person/numb er suffix obligatory P , T , R clitic pronoun alternating pst stem S verbal person/numb er suffix obligatory P , T verbal person/number suffix (nearly) obligatory R verbal person/number suffix alternating A clitic pronoun (nearly ) obligator y R clitic pronoun alternating T able 11.1 lays out the morphological expression of arguments across differ ent T AM constructions, and the indexing type for each argument. The functional status of indexes as obligatory indexes or alternating indexes is fully discussed in §11.1.1 and §11.1.2. As can be seen from T able 11.1, the morphophonological form of the indexes is not a good predictor of their functions. For instance, historical agreement suffixes in the past tense have become degrammaticalise d and can 348 11.1 Argument inde xing now be alternating indexes of R arguments, such as possessors, beneficiaries, recipients, etc. Likewise , the clitic pronouns that index A -past arguments hav e retained some of their pronominal origins to some degree but increasingly sho w traits of agreement indexes ( see §11.1.2.1). As seen in T able 11.1, among the core arguments of the clause, intransitiv e subjects are regularly indexed by verbal person suffixes. The following examples illustrate agreement with S: (1) with a co-refer ent subje ct argument, and (2) with a zero subject argument. The S argument is a coordinated noun phrase in (3). Y et, the copula verb agrees only with the second coordinate. (1) bizekê cîya leweřyayşa hurpiřa. bize-(e)kê goat.f-def.pl.dir cîya instead_of leweřyay=şa graze.inf=3pl:PSR hur-piř- a pvb-jump.prs.ind -3pl:S ‘The goats wer e dancing instead of grazing. ’ [ JP .55] (2) milawe pey germîyanî. mi-l- a =we ind -go.prs-3pl:S=compl pey for germîyan-î pn-m.sg.obl ‘They went to Garmiyan. ’ [ZB.39] (3) eđaw tatew min her beđb e xt bîyen. eđa=û mother.f=and tate-û father.m-ez.gen min 1sg her emph beđb ext poor bîye= n be.pst.ptcp.m=cop.3sg.m:S ‘My parents wer e po or . ’ [ JE.43] 11.1.1 Argument indexing in T AM constructions built on the present stem of the verb Accusativ e alignment is found in clauses built on the present stem of the verb . V erbal p erson/number affixes carr y out the expr ession of S and A. In contrast, clitic pronouns expr ess P . Accusative alignment is also r efle cted in the obliga- toriness of the indexes used to express the cor e arguments within the clause. While the indexing of A and S is obligatory , the indexing of P is conditioned by the absence of the co-referent NP . In other wor ds, the P-indexing clitic pronoun is mutually exclusive with an o vert direct object in the same syntactic domain. 349 11 Grammatical r elations (4) ême milmê şeşik. ême 1pl mi-l- mê ind -go.prs-1pl:S şeşik pn ‘W e will go to Shashk. ’ [HB.32] (5) îney maçmê. îney dem.pro x.obl.m.3sg m-aç- mê ind -say.prs-1pl:A ‘W e will say this. ’ [BP .69] (6) beroma yanew wêşa. ber-o= ma take.prs.ind -3sg:A =1pl:O yane-û house-ez.gen wê=şa refl=3pl:PSR ‘He will take us to his [lit. their] house. ’ [HB.35] Similarly , the same set of clitic pronouns that e xpress P also expr ess R argu- ments. In doing so, clitic pr onouns are alternating with the coreferent nominal. Note that the placement of R-indexing clitics follows the clitic placement rule set out in §10.1.4 and §6.2.2. In short, the clitic pronouns land on the leftmost element within the VP as their host, thus excluding the subject NP as a possible host. (7) dey to waçem la. dey disc.pt cl to 2sg waç-e= m say.imp -2sg:A =1sg:R la to ‘T ell me!’ [ÇK.107] (8) êtir mebo ême karêş p ene kermê . êtir no_longer me-b-o neg.ind -be.prs-3sg:S ême 1pl kar-ê= ş task.m-indf=3sg:R pene for kér-mê do.prs.sbjv -1pl:A ‘W e can no longer do anything for him. ’ [ JP.75] 11.1.2 Argument indexing in T AM constructions built on the past stem of the verb As r emarked ab ov e, clauses built on the past stem of the verb featur e ergative alignment. In canonical ergative constructions, the v erbal person/numb er affixes agree with P and S marked in the direct case. In (9)–(10), the verb agr e es with a direct-marked object NP. which is nominal in (9) and pronominal in (10). S is 350 11.1 Argument inde xing indexed by the same set of verbal person/number affixes as P (11). On the other hand, A is indexed by clitic pronouns in (9)–(10). (9) pase zawlêşa wey ker dê. pase like_this zawlê =şa child.pl.dir=3pl:A wey raising kerd- ê do.pst -3pl:O ‘They raised children in this way . ’ [ JE.55] (10) ađma tehwêɫ gêrt. ađ =ma 3sg.m.dir=1pl:A tehwêɫ delivery.m gêrt- Ø grab.pst -3sg.m:O ‘W e got hold of him. ’ [ZQ .18] (11) karewanîyê amê yanê serere . karewanî-ê caravan_people-pl.dir amêya= nê come.pst.pt cp.pl=cop.3sg.pl:S sere=re . top=postp ‘Some passers-by had stayed there . ’ [DB.15] In terms of obligatoriness, the indexation of S is fully obligatory , i.e., ev er y past intransitive clause is required to have an inde x for S, whether or not the coreferent NP is present. P is indexed by the same set of person/numb er suffixes and is nearly obligatory (see §11.1.2.2 for cases where the indexing is absent). Sim- ilarly , the indexation of A -past arguments via clitic pronouns is nearly obligatory . In the last two cases, the inde xes feature transitional behaviour b etween ‘agr e e- ment’ and ‘anaphora’ , meaning they are neither fully-fledged agreement indexes nor pronominal indexes ( see §11.1.2.1). As outlined in §6.1, spee ch act pr onouns have lost the case distinction. Nonetheless, they trigger agreement on the verb when occurring as a direct ob- ject in an ergative construction. Examples (12)–(14) represent P inde xing when P is an SAP pronoun. (12) to minit quɫ kerđa. to 2sg min =it 1sg=2sg:A quɫ pierced kerđ- a do.pst -1sg:O ‘Y ou disabled me. ’ [P W .30] (13) qesem pa xway toş epî layiqetî weş ker denî. qesem oath.m p=a to=dem.dist xwa-î God.m-sg.obl to =ş 2sg=3sg:A e=p=î emph=adp=dem.pro x layiqetî worthiness.m weş good kerde= nî do.pst.pt cp.m=cop.2sg:O ‘I swear to God, who has made you with such virtue . ’ [ZQ .54] 351 11 Grammatical r elations (14) ça toşa şîrîne kerđî ça çageyç minşa sîyaw kerđa. ça there to =şa 2sg=3pl:A şîrîn-e sweet-f kerđ- î do.pst -2sg:O ça there çage=yç there=add min =şa 1sg=3pl:A sîyaw-Ø black-m kerđ- a do.pst -1sg:O ‘[The husband said to his wife], there where they sweetened you, they also blackened me. ’ [XX.87] The similarity between S and P indexes as oppose d to A -past clitic pronouns is also reflected in syntactic criteria. One such case criterion is deletion under same- refer ent clause co or dination, which differentiates between the A index, on the one hand, and the S and P indexes, on the other hand. In the following e xamples, the indexes of S and P occur on each of the coordinate verbs, but the corefer ential A index is deleted in the first coordinate verb . In §6.2, this behaviour of the A - past index was argued to reflect its status as a ‘ clitic’ , as oppose d to the affixal status of S and P indexes. (15) ehmeđe dizeyç hurêst lûwaw e lûlejenay . ehmeđ-e pn-ez.cmpd diz-e=yç thief -def=add hur-êst- Ø pvb-rise.pst - 3sg:S lûwa- Ø =we go.pst - 3sg:S =compl lûlejenay flute_playing ‘ Ahmad the Thief got up [and] went back [to the palace] to play flute . ’ [ED .255] (16) yewaşê berdê kuştêşa. yewaşê then berd- ê take.pst -3pl:O kuşt- ê = şa kill.pst -3pl:O=3pl:A ‘Then they (the king’s men) took them (the pregnant women) [ and] killed them. ’ [KŞ.19] As a r eflex of a historical construction dating back to Middle Iranian, the ex- pression of P via verbal affixes/copula affixes gets e xtended to R arguments un- der ‘affix co-optation’ (Haig 2018). R arguments undergoing this process include possessors (17), recipients (18), human goals (19), addressees (20), sources (21), comitatives (22), etc. This results in e xternally realised R arguments, e.g., exter- nal possession, where the possessor is realised at a distance from its possessed noun. Unlike the indexing of O , which tends to be obligator y ( see §11.1.2.2), the indexation of R arguments alternates with the presence of the cor eferent NPs. 352 11.1 Argument inde xing (17) xeberêm nezanênê. xeber-ê=m news.m-pl.dir=1sg:A ne-zanê= nê neg-know.pst.pt cp.pl=cop.3pl:PSR ‘I didn’t know their news. ’ [ JM.30] (18) waçe her kesî nan danî. wáç-e say.prs.imp -2sg:A her every kes-î person.m-sg.obl nan bread.m da= nî give.pst.pt cp.m=cop.2sg:R ‘T ell who ev er gave you food. ’ [BP .165] (19) xway ketê pey kîyasen. xwa-î God.m-sg.obl ket-ê be d.m-indf pey to kîyase= n send.pst.pt cp.m=cop. 3sg.m:R ‘[ As if] God had sent him a bed. ’ [ JP. 69] (20) be xwa xuř ye ymêş venî. be by xwa God.m xuřî- eymê =ş shout.pst -1pl:R=3sg:A venî at ‘By God, he shouted at us. ’ [ŞC.44] (21) sanɛnêşa çene. sanɛ= nê =şa take.pst.pt cp.pl=cop.3pl:R=3pl:A çene from ‘They used to collect [tax] from them (people of Hewraman). ’ [BP .25] (22) mameɫeş nekerđe çenî. mameɫe=ş deal=.m=3sg:A ne-kerđ- e neg-do.pst - 3sg.f:R çenî with ‘He didn’t have dealings with her . ’ [SH.254] The indexing of non-core argument on past stem v erbs takes over the object index slot. Thus in (22) the v erb doesn’t agree with the masculine direct obje ct mameɫe , indexing instead the comitative ( see §11.1.2.2 for more details). Alternatively , R arguments may b e expr essed by clitic pronouns, in which case the argument is realised locally on the governing adposition. This indexing strategy is less frequent in the text corpus than indexation via v erbal affixes. 353 11 Grammatical r elations (23) watim peneş, ‘to milî ko?’ wat=im say.pst=1sg:A pene= ş to=3sg:R to 2sg mi-l-î ind -go.prs-2sg:S ko where ‘I said to him, ‘Where are you going?’’ [ JH.31] (24) aẍekew ʕêraqî, berdenşawe peyşa. aẍe-( e)ke-û agha.m-def.m.sg.dir-ez.gen ʕêraq-î pn-m.sg.obl berde=n=şa=we take.pst.pt cp.m=cop.3sg.m:O=3pl:A =compl pey= şa to=3pl:R ‘The aghas in Iraq, they took it back [the taxes] to them and so on. ’ [BP .28] 11.1.2.1 Differential A indexing A feature of the Iranian languages is that following the shifts in the v erbal and nominal morphology and the rise of ergativity in Middle Iranian, the historical clitic pronouns came to expr ess A arguments in clauses built on the past stem of the verb . 1 A -inde xing clitic pronouns have fully turned into obligator y indexes of A -past arguments in neighbouring Central Kurdish (though they hav e preserved their mobility ). But in Hewramî, as well as in some other Iranian languages, A - past indexing clitics have r etained, to some extent, their pronominal origin (Jügel & Samvelian 2016; Mohammadirad 2020b), as their pr e cursors did in Middle Ira- nian (Jügel 2015). In other wor ds, clitic pronouns have r emaine d alternating in- dexes of A arguments in clauses built on the past stem of the verb . This implies that not all A arguments are indexed via a clitic pronoun. Our frequency count based on the corpus in Mohammadirad (2025c), suggests that overall 12% of past transitive constructions (45 out of 379 clauses) are unindexed (see T able 11.2). See Mohammadirad & Haig (forthcoming) for a detaile d study of differential A indexing in Hewramî based on a much larger corpus. This suggests that, in terms of token frequency , clitic pronouns hav e nearly grammaticalised as agreement markers. The differential indexing of A in He wramî and related languages has b een assumed to b e conditioned by the complementarity between the clitic pronoun 1 This shift concerns the loss of finite perfective verb forms in late Old Iranian and their replace- ment by the resultative participle . In terms of its argument structure, the r esultative participle agreed with the direct-marked P argument as it did with S of intransitive constructions. How- ever , the expr ession of A was carrie d out by oblique NPs or thr ough clitic pronouns (see Haig 2008, 2017 for a detailed discussion, and Mohammadirad 2020b for the fate of A -past indexing in W estern Iranian languages). 354 11.1 Argument inde xing T able 11.2: Indexing A -past arguments n. past tr . clauses A is indexed A is not indexed T otal corpus 379 88% 12% and the oblique-marked A argument in the same syntactic domain. For instance, discussing the development of A -indexing clitics in Iranian, Haig (2020: 102–103) suggests that: “In Middle Iranian, these subject clitic pronouns wer e in comple- mentary distribution with free NP subje cts; this kind of system is still attested in some W est Iranian languages to this day . ” This is exactly the case in the fol- lowing excerpt fr om Hewramî. In the first clause, the clitic pr onoun is absent in the presence of the ov ert-oblique-marked NP. In the second clause , the clitic pronoun resumes the absent A argument. (25) a. min taze padşay kerdena wekêɫ. min 1sg taze anyway padşa-î king.m-sg.obl kerde=na do.pst.pt cp.m=cop.1sg:O wekêɫ advocate.m ‘I— the king has put me in charge [lit. made me an advocate]. ’ b . watenîçiş, ‘mişyo neberûşo. ’ wate=n=îç= iş say.pst.pt cp.m=cop.3sg.m:O=add=3sg:A mişyo a ux ne-ber-û=ş=o neg.sbjv -take.prs-1sg:A =3sg:O=compl ‘ He (the king) has said [to me], “Y ou shall not take her back. ’’’ [ZP .107]–[ZP .108] Y et, the data from the corpus show that the complementarity between the clitic and the oblique-marked overt NP only partly accounts for differ ential A indexing in Hewramî, as suggested by the following examples. Mor e importantly , it do es not explain why the clitic pronoun o ccurs despite the oblique-marked A NP being present. (26) pađşay desûriş dan be min. pađşa-î king.m-sg.obl desûr= iş order.m=3sg:A da=n give.pst.pt cp.m=cop.3sg.m:O be to min 1sg ‘The king has ordered me [to do this]. ’ [ JP .209] 355 11 Grammatical r elations (46) be kune awîşa ardênew . nîyɛneşare. be by kune clay_pot.m awî i =şa water.f.sg.dir=3pl:A ardê= ne i =û bring.pst.pt cp.f= cop.3sg.f:O =and nîyɛ= ne i =şa=re put.pst.pt cp.f =cop.3sg.f:O =3pl:A =po vb ‘They fetched i water i in clay pots. They unloaded it i [the water]. ’ [ JE.17] In clauses with O A V order , the verb tends to agree with the topical P (see 47). (47) werêsekê min warđêne. werês( e)-ekê rope-def.f.sg min 1sg warđê=ne eat.pst.pt cp.f=cop.3sg.f:O ‘[The lion said], ‘I have eaten the rope . ’’ [ÇH.85] T able 11.3 summarises the ratio of object indexing per overt and zero objects in the main text corpus (Mohammadirad 2025c). In counting the clauses with a P index, clauses containing speech verbs, e.g., ‘he said’ , were dismissed from the count as they could also be employed to fulfil other discourse functions. It can be se en that ov erall, 88% of direct objects are indexed on the verb , meaning that the absence of indexing is the marked, unexpecte d pattern. All zer o obje cts occur with agreeing verbs; in contrast, 15% of ov ert Os do not trigger agreement on verbs. Put differently , there are more o vert P arguments with non-indexing verbs. Thus, the data pro vide some support for the complementarity hyp othesis, which states that zero arguments are favoured by ov ert agreement markers and vice versa (see Nichols 2019). 4 T able 11.3: Indexing P-past arguments n. past tr . clauses O is indexed O is not indexed N % N % Overt object NP 246 208 0.85 38 0.15 Zero object 82 82 100 – – T otal 328 290 0.88 38 0.12 Now the question is which contexts trigger the use or non-use of agr eement suffixes with an object argument. Object indexing is present with topical Os, which are marked with a definite suffix. 4 See Mohammadirad (forthcoming[ b]) for a detailed study of object indexing in T ekht Hewramî based on a corpus of nearly 36,000 words. 362 11.1 Argument inde xing (48) tomekeş şana zemînekeyne. tom-eke =ş seed.m-def.m.sg.dir=3sg:A şana scatter.pst.3sg:O zemîn-ekey=ne land.m-def.m.sg.obl=post ‘He scattered the seeds in the field. ’ [ JP.51] (49) ħewt seferê awekêşa mite . ħewt seven sefer-ê time-pl.dir aw(î)-ekê =şa water .f-def.f.sg=3pl:A mit- e pour.pst -3sg.f:O ‘They poured the water seven times. ’ [MM.29] (50) maziɫoxekêş arđê. maziɫoxe-(e)kê =ş prayer_rug-def.pl.dir=3sg:A arđ- ê take.pst -3pl:O ‘He took the prayer rugs. ’ [ŞE.83] As discussed in §12.4, in light verb constructions which do not allo w a direct object in their argument structure, the nominal element e xhibits some properties of the direct object. For example, it contr ols agreement on the verb in clauses built on the past stem of the verb . The light verb construction is koç kerđey ‘to migrate ’ in (51), containing the masculine noun koç ‘migration’ , and duʕa ker dey ‘to pray’ in (52), containing the feminine noun duʕa ‘prayer’ . The light verb agrees in gender and number with these nominal elements. (51) koçşa kerden. koç=şa migration.dir.m=3pl:A kerde=n do.pst.pt cp.m=cop.3sg.m:O ‘they returned [to the village]. ’ [ JE.13] (52) duʕaw xeyrîş kerdêne . duʕa-û prayer.f.sg.dir-ez.gen xeyr-î=ş goodness.m-sg.obl=3sg:A kerdê=ne do.pst.pt cp.f=cop.3sg.f:O ‘They would bless [us]. ’ [DG.6] As seen in T able 11.3, the expected P indexing is occasionally absent. A few tendencies can be outline d here ( see Mohammadirad forthcoming( b) for a de- tailed study). In a r egular past transitive clause, the v erb agrees with the object NP , as seen in (48)-(50). The expe cted direct object index is absent when there is an additional object in the clause. Put differently , If the O-index slot is filled 363 11 Grammatical r elations by the non-core argument, the object cannot be indexed in Hewramî. In (53), the agreement with the plural object is missing since the agreement slot for the di- rect object has b een taken ov er by the index for the indirect object. Following Haig (2018), this is called “slot co-optation” . This phenomenon also has parallels cross-linguistically , e.g., in W arlpiri (Hale 1982: 251–252). In (54), the direct object is a feminine noun, which requires 3sg.f agreement on the v erb, i.e ., -e . Y et its slot has be en taken by the complement of pey . (53) sêŋze danê heserêşa da pene. sêŋze thirteen danê clf.pl heser( e)-ê=şa mule.f-pl.dir=3pl:A da-Ø give .3sg.m:R pene to ‘They gave him thirteen mules. ’ [ÇH. 69] (54) sêɫêşa pey kerd. sêɫê=şa Halva.f pey for kerd-Ø do.pst -3sg.m:R ‘They made Halva for him. ’ [MM.35] Second, P-past indexing is absent with plural inanimate Os. In the following examples, the verb has a default 3sg.m inflection and does not agree with the plural object. (55) penc çemçêşa nîyanre. penc five çemç( e)-ê=şa spoon.m-pl.dir=3pl:A nîya=n=re put.pst.pt cp.m=cop.3sg.m:O=povb ‘They (my family ) had set five spoons [on the tablecloth]. ’ [ JE.46] (56) gir kerêşa kerd. gir all ker-ê=şa chore-pl.dir=3pl:A kerd-Ø do.pst -3sg.m:O ‘They did all [their ] chores. ’ [HB.58] Differential P indexing is also attested for gender agreement in the text corpus. Where the controller fails to trigger agr eement features on the predicate, the 3sg.m form appears as the default form on the agreement target, i.e., the verb . In (57), the direct object is feminine, yet the verb fails to agr ee with it and instead appears in the default 3sg.m form. 364 11.1 Argument inde xing (57) çaştekêşa kerden be awîrgakê. çaşt(î)-ekê=şa meal.f-def.f.sg=3pl:A kerde=n do.pst.pt cp.m=cop.3sg.m:O be on awîrga-( e)kê hearth.f-.f.sg.sg.obl ‘They made the food on the hearth. ’ [ JE.39] Some head nouns do not carry numb er agreement that is triggered by numeral and measure nouns. In such cases, the verb tends to agree with the singular head noun. (58) wîs koɫê lokeşa ard. wîs twenty koɫ( e)-ê load-pl.dir loke=şa cotton=dir.m.sg=3pl:A ard-Ø bring.pst -3sg.m:O ‘They brought tw enty loads [of cotton]. ’ [ME.120] (59) ħewt koɫê xezêneş berđwe law mamojenîş. ħewt seven koɫ( e)-ê load-pl.dir xezêne=ş treasure=3sg:A berđ-Ø=we take.pst -3sg.m:O=copl la-û to-ez mamojenî=ş uncle ’s_wife=3sg:PSR ‘He took back seven loads of treasur e to his uncle ’s wife. ’ [ED.139] Direct objects consisting of a coordinate noun phrase fail to trigger object in- dexing on the verb if the co ordinate nouns ar e inanimate. In (60), number agree- ment is missing with liçû lût ‘lip(s) and nose(s)’ . On the other hand, co ordinated direct objects featuring human referents tend to trigger number agreement on the verb , see (61). This confirms the typological tendencies laid out in Corb ett (2006: 184–185). (60) liçû lûtû dêw ekaş berđ. liç=û lip=and lût-û nose-ez.gen dêw-eka=ş ogre-def.pl.obl=3sg:A berđ-Ø take.pst -3sg.m:O ‘He took the ogres’ lips and noses [to the king]. ’ [ME.156] (61) î jenû wême w î kuřme kuştêbîyɛnê. î dem.pro x jen(î)-û wife.f-ez.gen wê=m=e=w refl=1sg:PSR=dem=and î dem.pro x kuř=m=e son=1sg:A =dem kuştê=bî-ɛn-ê kill.pst.pt cp.pl=be.pst-cond.a ug-3pl:O ‘[What if] I had killed my wife and son [ by misjudgment]?’ [XŞ.104] 365 11 Grammatical r elations 11.2 Argument flagging Argument flagging refers to the flagging of arguments thr ough case morphol- ogy or adposition marking. Argument flagging follows accusative alignment in clauses built on the present stem of the verb and ergativ e alignment in clauses built on the past stem of the verb ( see T able 11.4). Cor e arguments are flagged by fusional case suffixes. The split alignment is only rele vant for third-person nouns and pronouns. Speech act pronouns have lost the case distinction. The following examples exhibit accusativ e alignment in clauses based on the present stem of the verb: S-prs (62) and A -prs (63) are marked in the direct case. In contrast, P-prs (64) is marked in the oblique case. (62) seʕbê wiɫaxdarê mila. seʕbê morning-f.sg.obl wiɫaxdar -ê stableman-pl.dir mi-l-a ind -go .prs-3pl:S ‘In the morning, the horse grooms went. ’ [ŞC.66] (63) dê wê řas maça. dêw -ê ogres-pl.dir řas truth m-aç-a ind -tell.prs-3pl:A ‘The ogres are telling the truth. ’ [SK.64] (64) y ewayç bero. yew(e)-a =y ç barley.f-pl.obl=add ber-o take.prs.ind -3sg:A ‘[Out of obligation], he took the barley seeds, too. ’ [ JP.29] The following examples illustrate ergativ e alignment, limited to T AM cate- gories based on the past stem of the verb . S-past (65) and P-past (66) are marked in the direct case. In contrast, A -past (67) is marked in the oblique case. (65) kar ewanîyê amêyanê serer e. karewanî-ê caravan_people-pl.dir amêya=nê come.pst.pt cp.pl=cop.3sg.pl:S sere=r e. top=postp ‘Some passers-by had stayed there . ’ [DB.15] (66) he warêşa wişkinɛnê. hewar -ê =şa summer_habitat.m-pl.dir=3pl:A wişkinɛ=nê scour.pst.pt cp.pl=cop.3pl:O ‘They scoured the summer habitats [ searching for food, etc.]. ’ [ JE.3] 366 11.2 Argument flagging (67) ênne paɫewana zorşa kerđen ênne so_much paɫewan-a warrior-pl.obl zor=şa pressure=3pl:A kerđe=n do.pst.pt cp.m=cop.3sg:O ‘The warriors put much pressure [ on the skin]’ [SK.132] In ditransitive clauses, the R argument tends to be treated differently than the direct object argument in either ditransitive clauses or monotransitive clauses. The alignment system with ditransitive clauses, where all thr ee arguments are human, follows either “indirective alignment” or “neutral alignment” . In indir ec- tive alignment, P and T are flagged the same, whereas R is treated differently . In neutral alignment, P , T , and R are flagged the same (Haspelmath 2005). The fol- lowing examples illustrate indir ective alignment in present tense constructions. R is flagged by a preposition, whereas T (68) and P (69) are only flagged by the oblique case. (68) deway ekey miđo be zayfekê. deway -ekey medicine.m-def.m.sg.obl mi-đe-o ind -give.prs-3sg:A be to zayf e-(e)kê girl.f-def.f.sg ‘He gave the medicine to the girl. ’ [SH.198] (69) şîyaw eħşî b ere . şîyaweħş-î pn-m.sg.obl bér-e take.imp-2sg:A ‘T ake Siyawahsh. ’ [SK.23] Neutral alignment is attested less frequently , and it is characterised by P (7 0), T (71), and R (71) being marke d in the oblique case . Howe ver , note that in (71) the oblique case is not visible on the feminine definite suffix -ekê , yet as definite- marked Ps regularly take oblique marking, it can be assume d that the the T argument in (71) is oblique-marked. (7 0) luw e hardeka barew e. lú-e go.prs.imp-2sg:S hard(î)-eka flour.f-def.pl.obl b-ar-e=we imp -bring.prs-2sg:A=compl ‘Go [and] bring the flour . ’ [RE.19] (71) zayfekê miđewwe îftađekey . zayf(ê)-ekê girl-def.f.sg mi-đe-w=we ind -give .prs-1sg:A =compl îftađ(e)-eke y messenger-def.obl.m ‘I will give the girl back to the messenger . ’ [KT .166] 367 11 Grammatical r elations In short, nominal case marking follows the template in T able 11.4. Direct case and oblique case interact in flagging the core arguments of the verb . T able 11.4: Morphological flagging of core arguments dir obl T AM based on present stem verbs S, A P , T , R T AM based on past stem verbs S, P , T A, R In reality , not all arguments are case-marke d according to the template in T able 11.4. The rest of this section gives a brief introduction to differential argument flagging, by which I am mean the alternation in case marking on different clausal arguments. 11.2.1 Differential A flagging The alignment system licenses case marking for A arguments. In clauses built on the present stem of the verb , A occurs in the unmarked direct case, realised by different endings in the singular ( depending on the noun class) and - ê in the plural. By contrast, in verbal categories deriv e d fr om the past stem of the verb , A should be, by default, accompanied by the oblique case suffixes. The following examples illustrate oblique marking on A -past arguments. Case marking on A interacts with differential A inde xing, which is mainly triggered by information structure ( see § 11.1.2.1). (72) pase herey wat ... pase like her -e-î donkey.m-2sg-obl.m wat say.pst ‘ As the donkey said ... ’ [HB.54] (73) meselen ađîşa hukim kerdênmê . meselen for_example ađîşa 3pl.obl hukim rule.m kerdê=nmê do.pst.pt cp.pl=cop.1pl:O ‘For instance, they ruled ov er us. ’ [BP.10] Right-dislocated A NPs in past transitive clauses tend to be oblique-marke d (74), though occasionally they may appear in the direct case (75). 368 11.2 Argument flagging (74) fermawa cuwanekey be řîşçermekey . fermawa say .pst cuwan-ekey youth-def.m.sg.obl be to řîşçerme-( e)key old_man-def.m.sg.obl ‘The young man said to the old man. ’ [HR.10] (75) watiş, ‘xasa’ cuwaneke. wat=iş say .3sg:A xas=a well=cop.3sg.m:S cuwan-eke youth.m-def.m.sg.dir ‘The youth said, ‘alright. ’’ [HR.72] In reality , not all A arguments in past transitive clauses are oblique-marked. A token frequency count of overt As re veals that a quarter of the third-person overt A NPs ar e not oblique-marked; se e T able (11.5). T able 11.5 excludes overt independent SP A pronouns in A function, whose token frequency is 24. Nor does it include cases where the oblique case is not visible after certain nominal bases or the additive clitic ( see 4.1.3. T aken together , these make up 39 tokens of A -past NPs, amounting to 38% of total ov ert A arguments for which the case marking is not available. A s a side note, the total number of transitive clauses derived from past tense stems is 395, of which 103 have ov ert A arguments, amounting to 26% of overt A NP arguments in the whole corpus, against 74% of A arguments b eing expressed through clitic pronouns, or occasionally dropped. This low percentage of overt A NPs is asso ciated with their high degree of topicality and is a reflection of universal bias against ov ert lexical NPs (Du Bois 1987). T able 11.5: O vert third person A arguments in the text corpus T oken frequency % Oblique-marked As 45 70% Direct-marked overt A s 19 30% The counts in T able 11.5, around a quarter of third-person agents skip oblique marking. Some tendencies suggest themselves. The follo wing discussion is based on Mohammadirad (2024a). First, the A argument is not oblique-marked when used as the subject in the cleft construction. 369 11 Grammatical r elations (7 6) î dê wênê î kinaçêşa b estêne . î dem.pro x dêw-ê=nê ogre.m-pl.dir=cop.3pl:S î dem.pro x kinaçê=şa girl.f.sg=3pl:A bestê=ne tie.pst.pt cp.f=cop.3sg.f:O ‘It was the ogres who had muted the girl. ’ [ JP.177] The data suggest that information prominence triggers case marking on A -past arguments. The latter operates at two levels: “local” and “global” (McGregor 2006, Chappell & V erstraete 2019). In local prominence, the presence of case marking on A is associated with the A argument b eing in focus, especially in contrast with another argument in the immediately preceding clause, or against a presup- position in the mentioned discourse. In global prominence, the pr esence of case marking on A concerns the expectations about A arguments in larger chunks of discourse (see below ). In Hewramî, local prominence is generally associated with oblique case marking. Local prominence is associated with the A argument b eing in narrow focus and contrastive focus. In (77), the case marking on her ‘ donkey’ is triggered by its contrast with min . (77) î her-î zûwaniş zana min hîçim nezanan. î dem.pro x her -î donkey-sg.obl.m zûwan=iş language.m.sg.dir=3sg:A zana-Ø know.pst -3sg.m:O min 1sg hîç=im nothing=1sg:A ne-zana=n neg-know.pst.pt cp.m=cop.3sg.m:O ‘The donkey knew the [Sheikh’s] language; I didn’t kno w a thing!’ [HB.71] Case marking on an A argument can also be triggered by global prominence. Accor ding to McGregor’s (2006) “ expecte d actor principle ” , in episodes of dis- course with an expected actor , the actor can b e left unmarked after its introduc- tion. Any deviation from the e xpe cted actor is marked in the ergative case. Sim- ilarly , in Hewramî, a topic that is continued in discourse may lose the oblique case, appearing instead in the direct case. In the following excerpt, the estab- lished direct-marked topic of the intransitive clause in (78a) is repeate d in the following transitive clause (78b), e ven though the oblique form ađîşa is expecte d (see Mohammadirad in prep( a) for a detailed discussion). 37 0 11.2 Argument flagging (78) a. tenya ađê luwɛnê. tenya only ađê 3pl.dir luwɛ=nê go.pst.pt cp.pl=cop.3pl:S ‘Only they (Baba Khwada, Hama the Invisible , and Little Hama) went [to Iraq]. ’ b . êtir ađê watenşa, ‘ême diruwê meyeymê . ’ êtir disc.pt cl ađê 3pl.dir wate=n=şa say.pst.pt cp.m=cop.3sg.m:O=3pl:A ême 1pl diruwê lie.f me-de-îmê neg.ind -give.prs-1pl:A ‘They said, ‘W e are not going to lie. ’’ [BP .116]–[BP .117] In terms of information structure , A NPs in the direct case tend not to carr y nuclear stress. In other wor ds, they behave like topics and contain given informa- tion. In the following examples featuring dir ect marking of the A NP, the nuclear stress is on the verb (79), the place adverb (80), and the negation prefix (81). (79) heɫbetene î kuře biřiş dawe . heɫbete=ne of_course=post î dem.pro x kuř=e boy.m=dem biř=iş piece=3sg:A dà=we | give.pst=compl ‘It is obvious that this boy has arrived. ’ [KŞ.80] (80) sibhan aẍe ça aẍegerîş kerden. sibhan pn aẍe agha.m çà there aẍegerî=ş governorship.m=3sg:A kerde=n | do.pst.pt cp.m=cop.3sg.m:O ‘Sibhan Agha go verned over there . ’ [DP .27] (81) ‘her ta îse qisêş nekerdênê . ’ her donkey ta until îse now qisê=ş talk.pl.dir=3sg:A nè-kerdê=nê | neg-do.pst.pt cp.pl=cop.3pl:O ‘[He said surprisingly ], ‘the donkey hadn’t talked until now!’’ [HB.46] Sometimes, the unmarked A -past argument is processed in a different intona- tion unit than the rest of the clause . This is the case with the following examples. (82) î kuře her memeş warden. î dem.pro x kuř=e | boy.m=dem her emph meme=ş breastm.sg.dir=3sg:A warde=n | eat.pst.pt cp.m=cop.3sg.m:O ‘The boy had kept breast-feeding [from the tree]. ’ [ZQ .44] 371 11 Grammatical r elations (111) jenêç nîşore duwê zaroɫê wîno. jen(î)-ê=ç woman-f.sg.obl=add nîş-o=re sit.prs.ind -3sg:S=po vb duwê two zaroɫ(e)-ê child-pl.dir wîn-o see.prs.ind -3sg:A ‘The wife gave birth to two babies. [Lit. She sat down [ and] saw two babies.]’ [ZB.24] In short, Hewramî ov ertly oblique-marks a great range of objects, whether they are human or non-human objects, as long as they play a salient role in dis- course, hav e b een previously e voked, etc. It was also seen that oblique marking extends ev en to indefinite direct objects on the condition that they are specific. This was reported to b e limited to the vernaculars of Nwên and Silên. By contrast, oblique-marking is absent for nominals with non-spe cific, generic r eference, and for syntactic reasons, when a direct object is quantified. T able 11. 6 summarises a T able 11.6: Conditions on differential P flagging in Hewramî: a pr elim- inary analysis Factors Feature Case marking Person 1st and 2nd pronouns no third person pronouns yes Animacy human a yes animate (non-humans) yes inanimate yes Uniqueness proper nouns yes common nouns yes/no b Definiteness definite-marked yes specific yes/no c non-specific no Quantification P being quantifie d no a Case marking for all animacy values is contingent on definite reading of r eferents, and/or discourse saliency . b Common nouns can be case-marked if they have definite reference . c The presence of case marking on specific P arguments marked by the indefinite suffix -êw(e) is subject to dialectal variation (see ab o ve). 378 11.2 Argument flagging preliminary overview of differential P flagging in T ekht Hewramî. As said, the factors conditioning differential P flagging interact with each other . Thus, a hu- man P may not be marke d in the oblique case if it has a non-specific reference (see 84). 11.2.3 Differential flagging of non-core arguments The case system licenses oblique case marking for third-person nouns and pro- nouns that function as non-core arguments, e .g., goals, recipients, addressees, comitatives, beneficiaries. Differential flagging is taken to mean that not all non- core arguments are marke d in the oblique case. Mohammadirad (2025b) iden- tifies several factors conditioning case marking on the non-cor e arguments. It should be note d that these factors interact in differ ential case marking and it is ultimately the combined effect of these factors that is crucial in differential flag- ging of non-core arguments. The type of adpositional flagging is an imp ortant factor triggering differential case marking. On the one hand, flagging by means of prepositions (112) and (less so) circumpositions (113) tends to trigger oblique case on the adposition complement. (112) luwew e pey yaney! lú-e=we go.prs.imp-2sg:S=post pey to yane-î house.m-sg.obl ‘[Now ] go back home!’ [ JH.118] (113) î kinaçêw to dermanû derdîş îna la î pîyaywe . î dem.pro x kinaçê-û daughter.f.sg.dir-ez.gen to 2sg derman-û treatment.m-ez.gen derd-î=ş illness.m-sg.obl=3sg:PSR îna-∅ loc.deic.cop -3sg.m:S la with î dem.pro x pîya-î=we man.m-sg.obl=post ‘This daughter of yours, the treatment for her illness lies with this man (i.e., Pir Shaliyar ). ’ [ZP .33] On the other hand, non-core arguments that are postpositional (114) or bare (115) are much less likely to get case-marked. 379 11 Grammatical r elations (114) jenekêm qomyaş venî kelekew e. jen(î)-ekê=m woman.f-def.f.sg=1sg:PSR qomya=ş happen.pst.3sg:S=3sg:R venî at kel-eke=we mountain.m-def.m.sg.dir=post ‘My wife was about to deliver a baby in the mountain. ’ [ZQ .14] (115) melowe yane . me-l-o=we neg.ind -go.prs-3sg:S=compl yane house.m ‘He didn’t go back home. ’ [ JH.109] Note that it is not always straightforward whether the case marking on non- core arguments is due to the type of adposition used or another factor . For in- stance, if the head of the adpositional phrase is a locational noun (see §10.1.1.2), the case marking can be triggered by the ezafe marking on the locational noun, which makes the construction look like an adnominal possessive phrase. (116) mêwe dilû he wramanî. m-ê=we ind -come.prs.3sg:S=compl dil-û inside-ez.gen hewraman-î pn-m.sg.obl ‘He returned to Hewraman. ’ [ JP.141] Another factor triggering differential case marking on non-cor e arguments is whether the adposition complement is an adnominal p ossessiv e construction, e.g., be tate-w min ‘to my father’ . Her e, due to competition b etween ezafe marking on the head of the NP and case marking on the same slot (see §5.3.1 for the interaction of ezafe and case marking), only ezafe marking is viable ( se e 117), unless the head noun is plural, in which case the two suffixes ar e compatible (118). (117) ama la şuwanew gawa. ama come.pst.3sg:S la to şuwane-û shepherd.m.sg.dir-ez.gen gaw(e)-a cow .f-pl.dir ‘He came to the cows’ shepherds. ’ [ÇH.108] 380 11.2 Argument flagging (118) dewayş kerđ çemaw kinaçêw patşay . deway=ş medicine.m=3sg:A kerđ do.pst çem-a-w eye-pl.obl-ez.gen kinaçê-û daughter-ez.gen patşa-î king-m.sg.obl ‘He put medicine into the king’s daughter’s eyes. ’ [DB.312] Role is another factor conditioning differential case marking on non-core ar- guments. For instance, a nominal complement of the verb ‘become ’ rarely takes an oblique case, r egardless of animacy . (119) bî be patşa. bî-Ø be.pst -3sg.m:S be adp patşa king.m ‘He be came a king. ’ [DB.161] For other roles, ther e is an animacy comp onent playing a r ole in differential flagging. For instance, inanimate (120) and non-human animate instruments (121) are not generally case-marked. On the other hand, human instruments (122) tend to occur in the oblique case. (120) be kune awîşa ardêne. be by kune clay_pot.m awî=şa water.f.sg.dir=3pl:A ardê=ne bring.pst.pt cp.f=cop.3sg.f:O ‘They used to fetch water using clay pots. ’ [ JE.16] (121) be hesere hêzmîşa ardênê pey zimsanî. be by hesere mule.f hêzmî=şa firew oo d.pl.dir=3pl:A ardê=nê bring.pst.pt cp.pl=cop.3pl:O pey for zimsan-î winter.m-sg.obl ‘They fetched firew oo d for the winter on mules. ’ [ JE.35] (122) werwe maro be nîrûwekeyş. werwe snow.f m-ar-o ind -bring.prs-3sg:A be by nîrû-ekey=ş force.m-def.m.sg.obl=3sg:PSR ‘He fetched snow [ five or ten loads daily ] using his men. ’ [DP .34] The animacy effect may show itself for other arguments as well. For e xample, human goals are ov er whelmingly case-marked (123). Howe ver , non-human goals may sometimes appear in the bare form (124). 381 11 Grammatical r elations (123) milo la ađî. mi-l-o ind -go.prs-3sg:S la to ađî 3sg.obl.m ‘He went to him (his uncle). ’ [ JP.14] (124) ta meřeber amɛ. ta until meřeber pn amɛ come.pst.3pl:S ‘[ which means] they came as far as Marabar . ’ [BP .114] T o b etter illustrate differential case marking of non-cor e oblique arguments, I did a frequency count of the goal arguments of verbs of mo vement, ‘ come ’ and ‘go ’ , in the text corpus. The count was limited to nominal and pronominal (i.e., third person) goals; goal arguments of ‘ come ’ and ‘go ’ that are first and second persons were not counted as they do not show case distinctions. A dverbial goals ( e.g., He went there) w ere not counted either . Finally , note that The frequency count shows that around 17% of goal arguments (21 out of 122) ar e case-marke d. In T able 11.7, I have classified the differential oblique marking based on the type of flagging the goal arguments have , and whether or not they are possessed. T able 11.7: Frequencies of oblique-marke d goals of ‘ come ’ and ‘go ’ , cat- egorised according to flagging Flagging N Obl-marked % Unmarked % Prepositional 22 19 86 3 14 Bare 49 2 4 47 96 Postpositional 19 0 0 19 100 Possessed 32 0 0 32 100 T otal 122 21 17% 101 83% As can be seen from T able 11.7, only 17% of goal arguments of ‘ come ’ and ‘go ’ are oblique-marked. Evidently , goals that are flagged by prepositions exhibit stark differences in case marking from postpositionally flagged goals, which are not case marked. Similarly , possesse d goals ar e not case marked across the b oar d. On the other hand, bare goals are not flagged for case marking across the board. In short, differential case marking on non-core arguments dep ends on differ- ent factors, which together determine whether or not a non-core argument is 382 11.2 Argument flagging marked in the oblique case. The relevant factors w ere said to be typ e of adpo- sitional flagging, role , animacy , and whether or not the adposition complement is possessed. Se e Mohammadirad (in re view ) for investigation of the combined effect of these factors on differential case marking of non-core arguments. 383 12 Sy ntax of the clause This chapter surveys the syntax of different types of simple clauses and the argu- ment structure associated with light verb constructions. The following topics are discussed: verbal clauses (§12.1); copula clauses (§12.2), clauses with existential particles (§12.3), light verb constructions (§12.4), reciprocal constructions (§12.5); periphrastic causative constructions (§12.6), passiv e clauses (§12.7), and interrog- ative clauses (§12.8). 12.1 V erbal clauses 12.1.1 Subject constituent The subject constituent exhibits typical properties of subjecthoo d, including con- trol of r eflexives. Recall that in the vernaculars of Silên and Nwên, the refle xive pronoun can take the bare form wê in e xpressing certain functions. In (1a), the subject constituent controls the reference of the r eflexive in the following clause . In (1b), it controls the r eference of the reflexiv e in the sub or dinate clause. (1) a. ayê ç mila wê şarawe . ayê=ç 3pl=add mi-l-a ind -go .prs-3pl:S wê refl şar-a=we hide.prs.ind-3pl:A =compl ‘They i went [ and] hid themselves i . ’ b . zatşa meb o wê aşkera kera. zat=şa fear=3pl:NC me-b-o neg.ind -be.prs-3sg:S wê refl aşkera disclosed ker-a do.prs.ind-3pl:A ‘(They i ) wer e afraid to make themselves i visible. ’ [ÇK.67] In the following example fr om the vernacular of Nwên, the bare form wê has the same reference as the subject. (2) lûwa zemînêweş da w erû wê. lûwa go.pst.3sg:S zemîn-êwe=ş land.indf=3sg:A da-Ø give .pst.3sg:O wer-û on-ez.gen wê refl ‘(He i ) went and put himself on a land [lit. gave a land on himself i ]. ’ [ED .30] 12 Syntax of the clause 12.1.2 W ord order configurations The basic word or der pattern is A O V . Despite having O V order , Hewramî ex- hibits several head-initial configurations, including Noun- Adjective , Possesse d- Possessor , Matrix clause-complement clause, V erb-Goal, and V erb-Re cipient, run- ning against the predictions of the head-directionality hypothesis (Dr yer 1992). T aking oblique arguments into account, the word or der is of the A O VX typ e . Following Hawkins (2008), the notation ‘X’ indicates oblique arguments. It is seen later that the post-verbal X is esp ecially the case for goals, recipients, and addressees. Similarly , the basic wor d order pattern in the neighbouring Iranian and Semitic languages has be en r eporte d to be A O VX (Haig 2022, Haig et al. 2025, Mohammadirad 2024d). In what follows, following the methodology in Haig, Rasekh-Mahand, Stilo, et al. (2024), some w ord order configurations in verbal clauses are surveyed. In some cases, I apply a gradient corpus-based method to describe word order pr operties of different arguments. 12.1.2.1 Order of subject, object and verb The subject constituent, by default, precedes the direct obje ct constituent, hence the order A O V . In the default A O V order , the subject is generally characterised by being topical, i.e., high in the animacy and definiteness scales, and expr essing given information. The nuclear stress is generally placed on the direct object expressing ne w information (3), that is, on the imme diate pr everbal constituent. When the direct object expresses given information, the nuclear str ess generally falls on the verb (4). (3) çêrhur zaroɫeke şot wer o. çêr=hur under=post zaroɫe-( e)ke child.m-def.m.sg.dir şòt milk.m wer-o | eat.prs.ind-3sg:A ‘The baby drank [its] milk from below . ’ [ZB.45] (4) lalo kinaçekê maro . lalo maternal_uncle.m kinaç(ê)-ekê girl.f-def.f.sg m-àr-o | ind -bring.prs-3sg:A ‘Then, the [girl’s] uncle br ought her (to Hewraman). ’ [ZP .49] Occasionally , the subject constituent comes between the verb and its direct object, resulting in O A V order . This is especially true in past transitive construc- tions when the A argument is focal. By contrast, in clauses with A O V order , the A argument is generally given. Another difference is that in clauses with O A V 386 12.1 V erbal clauses order , the O argument is given, but this is not always true of clauses with A O V order . Re call fr om §11.1.2.1 that the focality of the A argument in this construction triggers no-indexing of the A argument on the verb . (5) heywane awê b er de heywane animal.f.sg.dir awê water.f.sg.obl berd-e take.pst -3sg.f:O ‘The flood [ lit. water ] to ok away the animals. ’ [ZB.21] (6) kuřû min ř ozgar ya nekuşten. kuř-û son-ez.gen min 1sg řozgarî-a pn-pl.obl ne-kuşte=n neg-kill.pst.pt cp.m=cop.3sg.m:O ‘My son, the Rozgaris did not kill him. ’ [HM.08] The subject constituent may b e expr essed in the p ost-v erbal p osition, tagged as an afterthought (7b). The effect of afterthought is to bind the clause with what precedes. (7) a. qotêw aman. qot( e)-êw coffin.m-indf ama=n come.pst.pt cp.m=cop.3sg:S ‘ A coffin arrived [floating on the water]. ’ [MF .75] b . asawekeş bînan qotekey . asaw-eke=ş mill.def.m.sg.dir=3sg:A bîna=n block.pst.pt cp.m=cop.3sg:O qote-( e)key coffin-def.m.sg.obl ‘The coffin had blocked the [water stream in] the mill. ’ [MF .7 6] As discussed in §11.2.1, the majority of subjects are only expr esse d thr ough in- dexing via mobile clitic pr onouns. Of the total 395 past transitive clauses in the text corpus, 103 have overt A arguments, amounting to 26% of o vert A NPs argu- ments in the whole corpus. Overt subjects only o ccur in certain contexts, e .g., in contrastive topic constructions. (8) min řencim daw to berd! min 1sg řenc=im toil=1sg:A da=w give .pst=and to 2sg berd take.pst ‘I toiled, and you took (the credit)!’ [YX.16] 387 pêɫe poçî, p ewkî (KŞ.20) pêɫe /peˈɫæ/ (n.m) eyelid • pêrê /ˈpeɾe/ (adv) day after tomorro w • ( JH.107) pêse, pise /peˈsæ, pɨˈsæ/ (adv) like , as if • pêse [like]: ( JE.72); pise [ like]: (ZB.17); pêse [as_if]: ( JP .125) pêsne /pesˈnæ/ (adv) like this, in this manner • (ZP .125) pêt, pêç /pet, p etʃ/ (v.tr) fold • pête=n [fold.pst.pt cp.m=cop.3sg.m:O]: ( JE.13) pêway /peˈwɑj/ (n.m) welcoming • pêway=ma=w e [welcoming.m=1pl:PSR=post]: (HB.34) pêwer e /pewæˈɾæ/ (adp) together • ( JP .97) pêwya, pêwye /pewjɑ, pewjæ/ (v.intr) be visible • pêwîy(e)-ê [be_visible.prs- a ug.3sg:S]: (ZP .54); pêw y( e)-o [ be_visible.prs-3sg:S]: (KŞ.7 0) pilekanî /pɨlæˈkɑni/ (n.f) stairs • pilekanî [stair.f]: (DG. 62); pilekanî [stair.pl.dir]: (DG.62); pilekan(î)-êw e [stair.f-indf]: ( JH.12); pilekan(î)-ê [stair.f-indf]: ( JH.25) pilt /pɨlt/ (adj) short (human) • piřa, piř /pɨrɑ, pɨr/ (v.intr) fly • piřa-∅ [fly.pst -3sg.m:S]: (ZB.56) pirđî /ˈpɨɾɹˠi/ (n.f) bridge • piřna, piřn /pɨrnɑ, pɨrn/ (v.tr) break off • piřna=ş=re [break_- off.pst=3sg:A =povb]: (ZB.22) piřne /pɨrˈnæ/ (n.m) jumping • (ŞC.55) pişqelî /pɨʃˈqæli/ (n.f) sheep dung • pişt /pɨʃt/ (adp) back • (ŞC.63) pitew /pɨˈtæw/ (adj) hard ( of stone) • pijgîna, pijgîn /piʒɡinɑ, piʒɡin/ (v.tr) scatter • píjgîn-î=şa [scatter.prs.sbjv - 2sg:A =3pl:O]: ( JP.27) pîfê /ˈpife/ (n.f) burr • pîfê [burr.f]: (ZP.39); pîf(ê)-ê [burr.f-pl.dir]: (ZP .39); pîf(ê)-ekê=we [burr.f-def.f.sg=post]: (ZQ .40) pîr /piɾ/ (adj) old • ( JM.10) pîya /piˈjɑ/ (n.m) man • pîya-(e)ke [man.m-def.m.sg.dir]: ( JP .57); pîye-(ê)we [man.m-indf]: (ZP .2); pîya-î [man.m-sg.obl]: (ZP.45); pîya-ê [man.m-indf]: (DG.24); pîya-î=re [man.m-sg.obl=post]: (DG.50); pîya [man.m]: ( JP .163); pîya-( e)key [man.m-def.m.sg.obl]: (DG.53); pîya-î=we [man.m-sg.obl=post]: (ZP .33); pîyɛwe [man.m.obl.indf]: (ZP .66); pîya-(e)ke [man.m-def.m.sg.dir]: ( JP .132); pîyɛ=ma [man.pl.dir=1pl:A]: (BP.121); pîya-( e)ke=ma [man.m- def.m.sg.dir=1pl:PSR]: ( JH.110); pîyɛ [man.pl.dir]: ( JE.79) pîyagerî /pijɑɡæˈɾi/ (n.m) gentlemanhood • (ŞC.91) pîyaz /piˈjɑz/ (n.m) onion • pîyaz-î [ onion.m-sg.obl]: ( JP .229); pîyaz-î [onion.m- sg.obl]: ( JP .240); pîyaz [onion.m]: ( JP .248) poçî, p ewkî /ˈpotʃi, ˈpæwki/ (adv) that is why • pewkî [that_is_why ]: (PM.36); 479 [Document text truncated for crawler view.]