# Introduction a) Clitics The study of clitics has its origins in the tradition of Greek grammarians. This long tradition of the linguistic interest in clitics implies that the phenomenon has been studied from a number of philological and theoretical angles. The term is used to refer to items which have no accent and which cause a secondary accent to be laid on the last syllable of the word which they follow (Zwicky, 1994). It is generally agreed in the literature that we can characterize clitics as follows: a) Clitics have syntactically-determined distributions. They differ from affixes in that they may attach to words of many different classes, including words like prepositions or adverbs that do not usually take affixes. b) They are small lexical elements that cannot form an independent prosodic domain. They never bear stress lexically, and for reasons of prosody they always form an accentual unit with the preceding or the following word. They may phonologically attach to a host which may be a word or a phrase. c) They fit together into larger units and form "clitic clusters" with multiple clitics in different functions. d) They typically (tense/aspect/mood/polarity), or certain pragmatic functions (question, information status).code nominal features (person/number/ case), auxiliaries or verbal features. Author: University of Kufa, Iraq. e-mail: ahjsultan@gmail.com e) They may collapse distinctions made in non-clitic forms. As is well known, for instance, Italian 1st and 2nd person clitics collapse accusative/ dative/reflexive together; they are distinguished in 3rd person (but 3rd person dative collapses gender and 3rd person reflexive collapses gender and number contrasts) (Grimshaw, 1997). f) The word they attach to is termed the "clitic host". A clitic that precedes its host is proclitic while one that follows is enclitic. There is also a mesoclitic which appears between the stem of the host and other affixes. The endoclitic splits apart the root and is inserted between the two pieces. 1 It is a common point in linguistics that it is difficult to classify clitics and it is even more difficult to draw a demarcation line between clitics and affixes. According to Zwicky and Pullum (1983) and Zwicky (1985), it is a matter of debate whether such a line exists at all. Clitics themselves form a heterogeneous group. However, clitics can be classified into three different classes: Simple clitics are phonologically reduced allomorph of full lexical items, e.g. 'll in English. Special clitics are weak or atonic forms attached to hosts, not phonologically reduced and often formally different from strong forms, e.g. Italian object clitics. Bound word clitics are independent phonological words (not atonic) with no non-clitic counterparts, e.g. Serbo-Croatian li (cf Napoli, 1996). Clitics are one the most debated topics in syntax in general and they are very interesting phenomenon in Shabaki due to periphrastic nature of the lexical category in this language. The goal of this paper is threefold, namely to explore the different types, distribution and properties of clitics in Shabaki, secondly to investigate the various clitics-related phenomena, such as climbing, doubling and clustering and thirdly to test its sensitivity to Wackernagel's law. Wackernagel (1892) formulated a phonological principle specifically for Proto-Indo-European languages whereby clitics tend to occupy the second position in the clause. The syntactic distribution of clitics in this special clausal position is called Wackernagel position (Wackernagel, 1892) -second position in the clause, defined as after first word or first syntactic constituent. The element in first position, hosting the enclitic, could be a word or a phrasal constituent. Wackernagel position can apply to any kind of clitic. This paper is organized as follows. In section 1.2., we present a brief description of Shabaki. Section 2 discusses the pronominal clitics in Shabaki. Section 3 securitizes the possessive pronominal clitics, 'hin' constructions and ezafe construction. Section 4 discuses the demonstrative clitics. Section 5 deals with complementizer clitics. Expletive (existential) clitics are discussed in section 6. A detailed scrutiny of diminutive clitics and conjunctive clitics is offered in section 7 and 8 respectively. Tense, aspect and negation are discussed in section 9. Section 10 discusses some clitic-related phenomena such as clustering in 10.1. and climbing in 10.2. In section 11, finally, we reflect upon the theoretical implications of the present investigation, particularly in terms of Wackernagel's law. # b) Background to Shabaki The term Shabaki refers to both the people and the language of the Shabaki minority. It is an Indo-European language of the Indo-Iranian branch spoken mainly in the eastern region of Mosul province in Iraq. In western research, Shabaki is mostly described as a Zaza-Goran dialect of northwestern Kurdish language. The Median branch of modern Iranian northwest includes Zaza-Dimli, Gorani, Gaspian dialects, south Tati, Hawramani (or Horami) and Shabaki (cf Izady 1992). Along the same line, Shabaki people are claimed to be the shabankara (or shawankara) Kurds of Fars district in Iran. 2 The constituent order is syntactically SOV; i.e. predominantly verb-final like most of its Indo-Iranian sisters. However, it does not adhere to a strict word order and sentence components may occur in various positions in the clause, depending on focusing and topicality. Like Turkish, Japanese or Finnish, Shabaki is an agglutinating language (with some fusion) where morphemes have single semantic meanings and are simply connected linearly (yâna.gal.mân: house.s.our: our houses). If the subject is an NP, it is cross referenced on V1. If it is a pronoun, then it is expressed on V1. Subsequent verbs are usually marked with the same pronominal vowel clitics. For these and other reasons, Shabaki can be said to have a complex morphology. Another feature of this language is the presence of complex predicates, such as light verb constructions, onomatopoeic light verbs and serial verb constructions which are very productive (Sultan, 2010(Sultan, & 2011)). # II. Pronominal Clitics in Shabaki: Their Morpho-Syntactic Properties Let us first present the clitic paradigm in Shabaki in table (1) below. The table above shows that Shabaki pronominal clitics appear in 36 member paradigm. It is important to note that all forms of personal pronouns are subject to cliticization. They are only derived for person and number. Dative strong pronouns differ from accusative strong pronouns only by the addition of the preposition 'ba' (to). When clitics are used, the preposition becomes 'bana'. Clitics seem to retain their case marking and dative and accusative clitics have different weak forms. They invariably appear in strong forms post-verbally in mono-transitive canonical sentences. The choice of full versus clitic pronoun is presumably driven by discourse considerations. The weak nominative, dative and genitive clitics seem to be in line with Wackernagel law as it will be seen later. (1) a. kitâb-am dâ ba Ali. Book.ACC-1SG.POSS give.PST to Ali.DAT 4 I gave Ali the book. (1) b. Kitâb-am dâ ba-na-?. Book.ACC-1SG.POSS give.PST to-towards-3 SG.DAT I gave him the book. It is interesting to note that the mesoclitic 'na' appears between the preposition 'ba' and the nominal indirect object clitic '-?'. No reason except the phonological one is behind its occurrence in this position because it does not have any specific meaning. In both (1a) and 1(b), the dative object appears in a prepositional phrase. In table (1) above the strong first, second and third person singular, the accusative, dative and genitive clitics in sentences containing heavy verbs are syncretic because they all have the same form (-i). The weak nominative and genitive clitics in light verb constructions are also syncretic (-m), (-mân), (-at), (-tân), (-a?) and (?an) in first, second and third person singular and plural respectively. The strong nominative and genitive clitics in heavy verb sentences are only syncretic in third person singular. In second and third person plural in heavy verb sentences, the weak dative and genitive are syncretic. Pronominal clitics are placed after the first constituent in a clause or in a Wackernagel position. So far, it has been clear that subject clitics agree in number with the subject of the clause except for the first and second person singular forms. In nontransitive canonical sentences (SV), clitics occur in second positions and are attached to the leftmost element of a constituent phrase (verb phrase) whether the strong pronoun is present (2a-7a) or absent (2b-7b). The subjects in (2a-7a) are pronouns and the subjects are also expressed clitically on verbs. Subjects in (2b-7b) are only clitically expressed on the verbs. Therefore, it seems safe and sound to claim that these clitics obey the Wackernagel's law. See table 2. In Shabaki the OSV word order is considered canonical in monotransitive sentences as in (8.a & b). It is conspecuous that the subject pronominal clitic cliticizes post-nominally to the object (whether the object be a noun or a pronoun). However, it does not lose its status as a clitic when the object is pronominalized, and never becomes a free-standing pronoun. Enclitics are obligatory whenever the verb is in first position in Shabaki as shown in (9a & b). Shabaki clitics never appear in initial position in such clauses. For example, the typical ordering of the clitic and verb (as in 9a & b) systematically switches to prevent the clitic from appearing clause-initially. There is of course an important interaction between enclisis and the Wackernagel law. As shown in (9) above, Shabaki clitics always appear to the right of the verb when the verb is clause initial, and can generally be claimed that Shabaki shows enclisis when the verb is clause-initial. The traditional explanation for this is that enclitic elements by definition require a host to the left, so naturally when they would otherwise appear clause initially, they must switch positions so that a left host is available. However, given that Shabaki clitics have been widely assumed to be both proclitic and enclitic and therefore dependent on both the verb and object for positioning, explaining the unexpected switch to enclisis for verb-initial contexts and the shift to encliticize to accusative object pro(noun) for object-initial contexts. I schematize the occurrence of clitic pronouns in Shabaki as follows: (10) a. V [+finite] + Enclitic Object b. Object + Enclitic V [finite] Clitic placement in Shabaki as shown in the structures (10 a & b) above depends on the constituent immediately preceding the finite verb: X in (b). X is an object in the accusative case. The clitic can occur in enclisis to the verb in (10 a) when the object follows the verb. In (10 b) when X is sentence-initial, the clitic leaves the verb and attaches to the right of the host, i.e. it remains an enclisis. In Shabaki, clitics can not be placed in sentence-initial position as it is assumed above. When there is no element preceding the verb in sentence-initial position, the clitic has to occur post-verbally. In Shabaki, the verb agrees with the subject or the object in number and person in all tenses and aspects. In (11a), for example, the verb 'kat' (fell) takes the morphological default form. In (11b) the post-verbal pronominal clitic form agrees with the nominative subject in number and person. This is true with light verbs (as in 12a & b) where the agreement in person and number is attached per-verbally to the preverbal element, be a noun or adjective. Ali and Ahmad talked. Clitic doubling is a process whereby a clitic and a non-clitic referring to the same argument are allowed to co-occur in one clause (Halpern 1998: 107-8). Pronominal clitics cliticized to verbs, for instance, may co-occur in clauses with nominal phrases co-referential with them. In Shabaki, doubling involves exclusively subjects and there are no doubling 'object' clitics. Strong subject pronouns and their coreferential clitics co-occur within the same clauses. Interestingly, the glide 'y' with either '-i-' or '-e-' are inserted between the verb and the pronominal subject clitic. Consider the following examples in (13a &b): (13) a. Am amâ-yi-m. 1SG come.PST.1SG. I came. (13) b. Atu amâ-ye-t. 2SG come.PST.2SG You came. Meso-clitics can also be detected in Shabaki as in (14a) below. The first subject clitic '-m-' is preceded by simple past clitic '-a-' and followed by the third person object clitic. This is a treble clitic sequence. In present progressive, accusative clitics are prepronominal clitics first position for topicality. The clitic and the non-clitic referring to the same argument cooccur in one clause as (14). ( 14) a. Am (an) pek-â-m-a?. (Aman = I am who, it is me who ...) (-m-is a meso-clitic) 1SG beat.PST.1SG.3SG. I beat him. (14) b. Am(an) ma-peki-?. 1SG PROG-beat.PRES-3SG. I am beating him. Serial verb construction (SVC) is a sequence of two or more verbs which act as a single but complex predicate. A sequence of verbs qualifies as an SVC if there is no marker of syntactic dependency between the components. In Shabaki, clitics tend to attach to both components of the serial verb constructions as in (15). ( 15) Çam-at okar bitit-a. # Eye-GEN-2SG open-IMP see-IMP-2SG. Open your eyes and understand. As for transitive verbs, the preverbal subject clitic agrees with the subject in SOV word order (as in 16a & b). Here, the verb reverts to the default unmarked form and carries agreement with nominative and not with accusative noun. In SVO, the verb agrees both in person and number with the subject (as in 17a & b). ( 16) a. Ali dâr-a? bar-i. (SOV) Ali-NOM-3SG tree-3SG-NOM cut-PST Ali cut the tree. Numerous studies have recognized that cliticization of pronouns has both syntactic and phonological dimensions. The general rule for cliticization in Shabaki is that clitics attach to the end of the first word of the clause or phrase which they are syntactically constituents of, regardless of the syntactic category of that word whether a noun or a verb. Therefore, they seem to obey the Wackernagel law. In the examples above, the third person singular and plural clitics are attached to a noun (as in 16a) and to a verb (as in 17b). In examples (8 & 9) above and (16 and 17) below, Shabaki language seems to have retained Wackernagel-type object clitics which Iranian had originally Middle Persian/Parthian (cf. Boyce 1964;Brunner 1977). Shabaki does not lose them, but has them affixed on the verb in different word orders. The word order in ( 17) is possible because the verb carries the accusative marker '-i' of the direct object. In the shabaki di-transitive sentences in (1 repeated below in 18), the direct object must appear pre-verbal while the indirect object occur post-verbal whether they consist of full NPs or clitics. When the subjects of these sentences are full pronouns or full NPs, they must appear clause-initial and the object appears second-position and the co-referential subject clitics always appear at the end of object, be it a full NP or a clitic. In ditransitive sentences, the word order is (OSV-PP). The dative object is always a prepositional object. (18) a. kitâb-am dâ ba Ali. Book.ACC-1SG.POSS give.PST to Ali.DAT I gave Ali the book. (18) b. Kitâb-am dâ ba-na-?. Book.ACC-1SG.POSS give.PST to-?-3SG.DAT I gave him the book. It is important at this junture to mention that present tense nominal sentences obey the canonical word order Complement VS. The verb To BE in Shabaki seems to have been morphologically shortened and therefore it appears as 'a' at the end of the sentence. In Shabaki, copular verbs are particles always required after the predicates that have first, second and third person subjects. They are fully grammaticalized for tense, negation and person. Examples (19)(20)(21)(22) in the table above contain present and future copular clitics which always appear clause-final when the subject is null. The copular clitic is context-sensitive. It is '-na' when it is added to a vowelfinal complement. It is only '-a' when added to a consonant-final complement. When the vowel is short front '-i', the copular clitic becomes '-yâ' with the glide 'y-.' When the subject is animate, the subject pronominal clitic appear after the copular clitic. In the negative form, the negative clitic 'na' is inserted between the complement and the copular clitic which is 'â' with glide 'w' separating them. The affirmative and negative copular verb clitic in simple past is the inchoative 'bi'. The pronominal subject clitic is cross-referential on copular inchoative 'bi'. The negative clitic is inserted between the complement and the inchoative 'bi'. Consider the following examples: (23) a. Çarma-bi-m. White-1SG-BE.PST I was white. (23) b. Çarma-na-bi-m. White-1SG-NEG-BE.PST I was not white. To express the progressive aspect, the vowel 'o-' is added to the end of the complement and then the copular clitic is added. The negative clitic is inserted between the vowel '-o-' and the copular clitic. See the example in (24). Heart-wide-NEG-BE.PRES.PROG He was not getting generous. (26) a. Hâli-yo-ma-bo. Empty-BE.PRES.PROG He was getting empty. # III. # Possessive Clitics The notion of possession in Shabaki can be expressed in three ways: the possessive clitical pronoun, hin construction and ezafe construction. The following subsections will discuss in much detail these three ways. # a) Possessive pronominal clitics The use of a clitic pronoun to express the possessor is shown in table (4) below. The genetive clitics are stated in table (1). The possessive pronouns in shabaki are always enclitics. They are attached to the end of the word. If there are adjectives or plural markers, the possessive clitic is always rightmost item. # Table 4 : Paradigm Possessive Pronouns in Shabaki Singular possessive pronouns are preceded by '-a-' when they are attached to consonant-final nouns. When they are attached to vowel-final nouns, '-a-' does not appear. It is worth to note that the plural morpheme precedes the possessive pronominal clitics. It seems clear that in Shabaki the main difference between singular and plural possessive pronouns is that 'ân' is a clitic that can be seen as a pluralizing marker of pronouns. A simple comparison between singular and plural possessive pronouns reveals that the pluralizing item '-ân' is simply added to the singular forms to constitute the plural form. # b) Hin construction The second way to convey possession in Shabaki is accomplished via using 'hin' construction. The origins of 'hin' in Shabaki can be traced back to the Old Persian relative/demonstrative hya/tya. Darmesteter (1883) and Meillet (1931) point out that hya (tya-) is not a simple linker, but that it further has a demonstrative value. Hya (tya-) becomes -i/-y in Middle Iranian (Partian) and progressively loses its demonstrative value to end up as a simple linker and became specialized as a device for nominal attribution. In Shabaki, the grammaticalization of this linker has not proceeded in the same way and at the same rate of other Iranian languages. This resulted in a complex picture for the different uses of the Ezafe. It is hard to pinpoint when exactly this development process occurred. The most appropriate scenario for the grammaticalization of this linker in Shabaki is the following: (27) hya/ tya- hin-i hin -i By looking at examples in (28) in comparison to those in section 3.3., we can deduce that 'hin' is optional and it is undergoing the process of grammaticalization. In Shabaki, the use of 'hin' at present is considered more formal and emphatic than iezafe. More emphatic still is when 'hin' and '-I' of ezafe are linked. 'Yâna' in (28a) is definite even though it does not receive any defining particle. In indefinite form, it becomes 'yân-e'. The enclitic attaching to 'trombel' in (28b) is the definite article and the indefinite counterpart is 'trombel-e'. The second '-I' attached to 'Jassim' in ( The possessive pronouns can occur in hinconstruction which is comparable to of-genitive in English as in (30). In Shabaki, when the possessive pronoun is added to the nominal source, the result is a reflexive pronoun as in (38). This construction is a type of ezafe in Iranian languages. Interestingly, a possessive pronominal clitic may stick to 'hin' to express objective case. More interestingly, ezafe -i is added to the composite construction. # c) Ezafe construction In general, there three types of ezafe in Shabaki. Ezafe often behaves strangely. The possessive pronoun is '-i' which can be a trait of Old Iranian. Consider the examples in (45). 5 (32) a. Trombel-i Jassim-i. Car-DEF Jassim-EZF Jassim's car (32) b. Trombel-aka Jassim-i. (-aka denotes diminutive sense) Car-DIM Jassim-EZF Jassim's (little) car Interestingly, -i also functions as a definite article and '-e' functions as indefinite article in Shabaki. Shabaki neither distinguishes between masculine and feminine nor between animate and inanimate. Nouns after numbers are not pluralized. Therefore, ezafe construction functions comparatively simpler. There are two basic patterns: N-ezafe Adj and N-ezafe N. In case of the N-ezafe N construction, the relationship is usually one of Possessee and Possessor or kinship. Definiteness is marked by the vowel -i vowel. Interestingly, the definite marker and the Ezafe are incompatible in Shabaki, i.e. you cannot have both at the same time. The element '-i' can be considered a genitive marker. It can also be added to 'hin' as in 'hini'. The solid evidence we have for the claim that '-i' can function as ezafe in Shabaki is that it is never used in isolation. Ezafe is a linking clitic phonologically realized as /i/ after a consonant and /yi/ after a vowel. It appears after noun phrases. Phonologically it forms a unit with the preceding element. In possessive constructions, ezafe is affixed to the possessor. Interestingly, ezafe clitic is added to diminutive noun. It is always the last item in the noun phrase. Also, complex ezafe construction is possible. It may link three nouns together as in (34). (33) a. Qaç-i kursi-la-i. Leg-DEF chair-DIM-EXF The leg of the chair. (33) b. Tok-i sây-aka-I. Skin-DEF apple-DIM-EXF The skin of the apple. (34) Qâp-i trombel-i Jassim-i. Door-DEF car-DEF Jassim-EXF The door of Jassim's car. # IV. # Demonstrative Pronoun Clitics In this section, we provide a brief description of the properties of Shabaki demonstrative pronoun clitics that will be relevant to the discussion in this paper. Overall, the Shabaki demonstrative pronoun clitics show very unfamiliar syntactic properties. The three properties of the Shabaki demonstrative clitic paradigm that are unique, and relevant for our purposes here, are sketched below. The first notable property concerns the proximity and remoteness of the object. In most cases, there is a four-degree gradation among proximity and interesting property of the Shabaki demonstrative pronoun clitics, be they singular or plural, is that they have diminutive forms. The third, and most importantly, notable property of the Shabaki demonstrative clitics is the fact that they have a slight complex constructions such as doubling, trebling or even severalfold. Shabaki demonstrative clitics are doubled by comparative clitics, trebled by superlative clitics, and severalfolded by ezafe clitics or diminutive clitics as shown in the table (5) below. # Table 5 : Demonstrative pronominal paradigm in Shabaki Table 6 : Strong and weak demonstrative clitics in In addition to the noun and adjective demonstrative pronouns, there is a series of adverbial demonstratives meaning 'this place (i-ndâ = here)', 'that place (u-ndâ = there)', 'this time (i-stâ = now), 'that time (u-stâ = then)', this way (i-tar = in this way), 'that way (utar = in that way)', (ç-i-do), (ç-u-ndo), (ç-i-ndâ), (ç-undâ), etc. in (35) in contradistinction to the specifier '-i' is used to convey the meaning of that specific place. (35) a. Arn-a gul-i sari u-n-dâ-i. Put-2SG rose-DEF on that-place-DEF Put the rose on there. (35) b. Sari u-n-dâ-i arn-a gul-i. On that-place-DEF put-2SG rose-DEF Put the rose on there. (36) a. Arn-a gul-i u-n-dâ. Put-2SG rose-DEF that-place Put the rose there. (36) b. U-n-dâ arn-a gul-i. That-place put-2SG rose-DEF Put the rose there. Shabaki demonstrates some unique constructions such as combining full demonstrative pronouns with clitic personal pronouns when the former is the direct object in accusative case and the latter is the subject of the sentence in the nominative case. Nonetheless, the clitic form of the demonstratives alone renders the construction ungrammatical. Likewise, these pronouns with the same forms can be used with plural nominals as (39) because they are not sensitive to the pluralization of the nouns they define. (39) a. I-trombel-gal These-cars These cars (39) b. U-trombel-gal Those-cars Those cars V. # Complementizer Clitics Complementizers are known to have evolved from other grammatical words such as demonstrative pronouns, conjunctions, adpositions or case-markers, or from lexical words such as nouns and verbs. Complementizers in Shabaki are introduced with the clitic '-ki' (which or that or who). In the list in (40), the complementizing clitic '-ki' follows nouns, full noun phrases, quantifiers and adverbs. '-Ki' gains its specific meaning from its co-text and, therefore, it may indicate a variety of meaning such as near and far in time and place, this or that person, thing, quality and manner, and few, little, a lot and much in number and quantity. A glance over the list of complementizers in (40) reveals that they have emerged through grammaticalization. (40) a. Time Istâ-ki: this time when (used in present and future time sentences) Ustâ-ki: that time when (used in past time sentences) Asâ-ki: during, while (refers to the synchrony of two actions) Hatâ-ki: until the time when (refers to the end-point of an action) Çani-ki: at the moment when (refers to the commencement-point of an action) Waxti-ki: when, during (denotes all above complementizers) (40) b. Place Indâ-ki: this place where Undâ-ki: that place where (40) c. Person, thing Harke-ki (anybody who) Ina-ki: this which, this who Una-ki: that which, that who (40) d. Manner and quality I-tar-ki: this manner that, this way that U-tar-ki: that manner that, that way that (40) e. Quantity Êqa-ki: to the degree that, to the number that, to the amount that VI. # Expletive Clitics (Existential Construction) The expletive subjects in shabaki existential sentences do not have the same lexical form across tenses. They also vary with the polarity of the sentences. This complicated state can be ascribed to the fact that the expletive subject clitic is homophonous with the possessive verb 'hand' (to have) on the one hand and homophonous with the verb 'be' (to be) on the other hand. In past tense, the possessive clitics, expletive clitics and copular clitics lump together to formulate the entire expletive clitical subject. What is salient about the future is that the inchoative construction is used to express the existential sentences. The pivot in existential sentences in shabaki may be inverted to sentence-initial position. The negative clitic 'ni-' or 'na-' procliticize the expletive clitics. The tenor of examples the table (7) below is that expletive subjects are in fact a cluster of two or more clitics. Inchoative verbs describe a change of state (the process of beginning or becoming). The striking difference among the items used to express expletion can not be a matter of mere chance. The present form stands at stark contrast to past and future. This issue leads us straight into the question why does this variety occur? Most strange, the affix '-o' is a productive inchoative clitic usually added to nouns and adjectives or sometimes adverbs. Ma-b-o consists of three elements: future-denoting clitic 'ma-', copular '-b-' and the inchoative clitic '-o'. In (41b), the glide'-y-' is inserted to ease pronunciation. In (42b), the combination of negation and the verb 'to be' is rather conspicuous without the interference of any glide. # VII. # Diminutivization The commonest word-formative device among nouns is the creation of a diminutive (a form which indicates smallness) from a base noun. The striking feature of this morphological phenomenon in Shabaki is that it never occurs without either the definite (-a) or indefinite article (-e). Included in the table (8) are diminutive clitics. # Table 8 : Diminutive clitics in Shabaki The referents of demonstrative pronouns can be diminutivized 'ina-la' (this little) and 'una-la' (that little). 'al' is encliticized to singular possessive pronouns referring both to items near and far. There are two wordlevel diminutive clitics in Shabaki: (-a) k and -l. when -k attaches to consonant-final words the intervening vowel -a is inserted between the word and the clitic, then the definite or indefinite clitics are added. It is important to note that -a is a specific article defines the noun it attaches to. The (in) definite clitics are always the outermost ones. Related to diminutive are degradation clitics in Shabaki. It is usually common in Shabaki to add the enclitic '-o' (that) to any proper noun to degrade his or her person. If the noun ends in a vowel, the enclitic replaces the final vowel and forms a prosodic word. If the noun ends in any consonant, it encliticizes to that consonant and forms a prosodic word. (44a) conveys a polite or neutral meaning whereas (44b) carries a degrading tone or belittling of the person concerned. Surprisingly, the diminutive clitic '-la' can be added to this degradation clitic '-o' as in (44c). The first striking property of the conjunctive clitic 'u' (and) in Shabaki is its rather unusual distribution. It always appears after the first word in the sequence of conjoined items, i.e. in second position within a phrase. Note that its distribution does not rely on the word class of the preceding word, to which it is attached: in (45a) it attaches to a noun, in (45b) to an adjective or adverb, and in (45c) it attaches to a verb. The evidence for the bound status of this element comes from prosodic phonology. Here, the host word 'kura' and the particle -u form a unit with respect to stress assignment. In Shabaki, stress is placed on the penultimate syllable of a word if it is heavy; otherwise it is placed on the antepenultimate syllable. Therefore, 'kura' would be stressed on the antepenultimate 'kura'. If this word happens to be followed by the unstressed particle -u, however, stress is assigned to the syllable immediately preceding -u yielding kura-w. Note that this rule of stress assignment does not make reference to the weight of the syllable preceding -u. Even if this syllable happens to be light, it will receive stress in this case. This behavior in stress assignment justifies the postulation of a domain in prosodic phonology which encompasses both the host word and the clitic particle in a clitic group (cf. Nespor & Vogel 1986, Jacobs 1997, Anderson 1992: 202). This constellation of properties can best be captured by saying that elements like the Shabaki -u "are words from the syntactic point of view but are phonologically dependent *?+ upon a neighboring word" (Zwicky 1994: 572). The phonological bonding of this clitic is a purely phonological matter and relies in this case on processes of syllabification. This positioning affects the stress pattern of the first conjoined item. The enclitic 'u' provides an example for a rule which is dependent on speech rate. Here, this element assimilates to the beginning of the following item in fast speech (cf Kaisse 1985: 25ff, Borg & Azzopardi-Alexander 1997: 137). Given the fact that Wackernagel's law forbids clitics to appear in initial position due to their phonological status, the conjunctive marker '-u' (and) always encliticizes to the end of the first word and not at the beginning of the second word. # IX. # Tense, Aspect and Negation Let us now deal with the clitic forms and distributions of tense, aspect and negation one by one, and discuss their specific properties and interactions. We will start with tense. Tense in Shabaki is sometimes expressed by certain proclitics which occur before the verb. In intensive simple present tense and simple future sentences, the copula verbs have 'â' and 'na' proclitics for affirmative and 'wâ' preceded by the negation marker 'na' (not) for negative. Shabaki signals negative polarity inflectionally among other strategies. The glide 'w' is inserted to ease the transmission from 'a' of 'na' into tense marker 'a'. If the complement ends in a consonant sound, the present tense clitic appears as 'â' but appears as '-na' if it end in vowel. These two clitics are in complementary distribution. Past tense is expressed by the proclitics '-bi'. Speakers of Shabaki usually have difficulties assigning a precise meaning to -o in this position other than indicating tense. However, it attaches to a phonological host, forming a stress unit with it. Past progressive is expressed by a cluster of proclitics: '-o-' expressing change-of-state. The negation marker 'na-' precedes the progressive marker 'ma-'. Progressive past is achieved via 'a-' Negation clitic in Shabaki has the same status as the imperfect clitic (ma-) discussed above, since it is positioned before the main verb and cannot occur in isolation. It is stressless and the vowel is omitted before verbs which begin with a vowel. It is sometimes embedded into clitic clusters when the verb is preceded by imperfective aspect proclitic. Negation clitic never serves as a host for other clitics. In (46a) the copula verb is itself a clitic whereas in (46b) the copular verb is a host and the negative marker procliticizes to it. (46) a. Qarmaz-â. Red-BE.PRES.3SG It is red. (46) b. Qarmaz na-wâ. Red NEG-BE.PRES.3SG It is not red. To sum up, the negation proclitic in Shabaki is 'na'. It always precedes the verb and can occur in clusters with other clitics, such as tense and aspect clitics. It seems that 'na' in Shabaki is a clitic both phonologically and grammatically. It is unstressed. It does not occur in isolation and is always attached to the verbal host. It is inserted between tense and aspect clitics. The negative marker 'na' loses its clitical status and becomes a full lexical word when it is used to answer to a yes-no question. -w-is an epenthetic glide between the negative enclitic and the copula verb -a'. A glottal stop is attached to the end of the word in this case. Shabaki has a two-class pronominal system (strong -clitic pronouns) that is attested with morphologically distinct forms and has a different distribution. The form of the (in) definite article in Shabaki is weak and needs a host word and forms one prosodic constituent. Clitic doubling, clustering and climbing may be considered defining features of Shabaki. Clitic doubling demonstrated the slightest degree of grammaticalization and the highest degree of pragmatic significance. All subject pronouns need to be doubled by coreferential clitics usually attached postverbally. This doubling is obligatory. Contrary to full pronouns, clitics are syntactically deficient in that they cannot be modified, co-ordinated or contrastively stressed. Furthermore, they can occur neither in isolation nor in the same positions as full pronouns or NPs. With respect to their referential properties, they differ from full pronouns insofar as they can refer to non-human entities, while full pronouns cannot. In Shabaki the subject and object of a sentence occur in pre-verbal position (Shabaki is a OSV language), but they may attach themselves as clitic pronouns to the end of the verb and form a one-word sentence. Like some European languages, clitics in Shabaki never occur in a sentence initial position even when word order changes from SOV (as in 53a) to VSO (as in 53b), to SVO (as in 53c) or to VSO concantative sentence where the subject and object are encliticised to the verb (as in 53d). In Shabaki nominals, it is noteworthy to observe that clitics sometimes allow to densely pack several morphosyntactic details into just a small amount of text. Shabaki object pronoun clitics are enclitic to finite verbs and remains enclitic to non-finite forms. Shabaki shows relics of ezafe clitics which are recorded from the middle Persian. These forms were gradually used less and less, and then substituted for by other simplified forms. Although Shabaki is SOV, some verbs may appear in sentence-initial positions. In these cases, proclitics are disallowed as in (53a). It is observed that Shabaki tends to have its focused components in clause-initial position. (54) a. Am-na? tit. 1SG-3SG see.PST He saw me. Finally, we can conclude that clitics in Shabaki have various distributions. They can attach to nominal, verbs, prepositions, pronouns and other clitics. This may be the rationale behind the numberless word order possibilities in Shabaki which arise out of the complex, dynamic interaction of clitics and basic clause elements. This state conveys an impression of vividness and some complication to learners of this language. This paper demonstrated that Wackernagel's law is highly operative in Shabaki. The Shabaki clitics are principally phrasal in character and obey the Wackernagel law. In Shabaki, there is a change in clitic placement that is neither connected to the clitics' phrasal character nor to their obedience to the Wackernagel's law, but that it is the parameter determining the relative order between verb and clitic that has changed its value in the development from middle Persian to present day Shabaki. Like Slavic languages, Shabaki seems to exhibit a few cases a tension between two tendencies. On the one hand, it tends to obey Wackernagel's Law and on the other hand it tends to follow the word to which they most closely relate. Personal pronominal clitics, if they are not in second position, they are almost always after the verb (Gribble 1988, 194-195). # Notes 1. Endoclitics defy the Lexical Integrity Hypothesis but evidence from the Udi, Pashto and Degema languages suggest that they do exist. 2. For more about the origin of Shabak in western research, the reader is referred to Vinogradov (1974), Leezenberg (1997), Mackenzie (1961Mackenzie ( , 1966)), Izady (1992), and Bates & Rassam (2000). 3. Â â as in apple; A a as in about; Çç as in church; ? ? as in shoe; ? ? as in vision; X x as in Loch (in Scottish). The voiceless uvular fricative in English, ? ?, corresponds a voiced uvular fricative in Shabaki. The voiced and the voiceless pharyngeal fricatives replace a andh in some Shabaki words respectively. The last two sounds are not part of Shabaki phonology. 2![Shabaki strong vs. clitic pronoun forms in subject positions The following two examples illustrate that Shabaki allows alternation between two word orders (OSV & VSO). 8b and 9b are generarted by 1a and 2a by deleting the NP. The pronoun replaces the noun in these examples. (8) a. Kitâb-am tit. (OSV) Book.ACC-1SG.NOM see.PST I saw the book. (8) b. Aw-am tit. (OSV) 3SG.ACC-1SG.NOM see.PST I saw him. (9) a. Tit-am kitâb-i. (VSO) See.PST-1SG.NOM book-DEF I saw the book.](image-2.png "Table 2 :") 9![b. Tit-m-a?. (VSO) See.PST-1SG.NOM-3SG.ACC I saw it.](image-3.png "( 9 )") 11![a. Âwel-ka kat. Boy-DEF fall-PST. The boy fell. (11) b. Âwerda katâ-ye. Boys-DEF-NOM fell-PST-3PL. The boys fell. (12) a. Ali qasa-? kard. Ali-NOM-3SG talk-3SG Do-PST-3SG. Ali talked. (12) b. Ali u Ahmad qasa-?ân kard. Ali-NOM and Ahmad-NOM talk-3PL do-PST-3PL.](image-4.png "( 11 )") 16![b. Ali u Ahmad dâr-?ân bar-i. (SOV) Ali-NOM and Ahmad-NOM tree-3PL-NOM cut-PST Ali and Ahmad cut the tree. (17) a. Ali bar-i-? dâr-i. (SVO) Ali-NOM-3SG cut-PST-3SG-NOM tree-DEF. Ali cut the tree. (17) b. Ali u Ahmad bar-i-?ân dâr-i. (SVO) Ali-NOM and Ahmad cut-PST-PERF-3PL-NOM tree-DEF. Ali and Ahmad cut the tree.](image-5.png "( 16 )") 24![a. Çarm-o-bi. White-3SG-BE.PST.PROG It was getting white. (24) b. Çarm-o-na-bi. White-3SG-NEG-BE.PST.PROG It was not getting white. Present and future progressive aspects in intensive sentences are expressed by using imperfective proclitic 'ma-' before the copular clitic 'bo'. Due to the concanative nature of Shabaki, 'mabo' is linked to the complement after adding the vowel '-o' to the end of the complement. When '-o' is added to a word ending in a vowel, it is often dropped out. It the vowel is '-I', '-yo' is added instead. The suffix '-o' indicates a change of state and it usually encliticizes to the adjectival element in the light verb constructions. (25) a. Dal-haraw-o-ma-bo. Heart-wide-BE.PRES.PROG He was getting generous. (25) b. Dal-haraw-o-na-ma-bo.](image-6.png "( 24 )") 30![N + hin + N (as in 28 a & b) Nominative source + possessive pronoun (as in 31) Nominative source + reflexive pronoun (as in 28 c)](image-7.png "( 30 )") ![31) a. Ali tit-a? hin-am. Ali.NOM see.PST.3SG of me Ali saw mine. (41) b. Ali tit-a? hin amni. Ali.NOM see.PST.3SG of mine Ali saw of mine.](image-8.png "(") ![37) a. Ina-m tit. This-1SG see.PST I saw this. (37) b. Ina-gal-m tit. This-PL-1SG see.PST I saw these.'I' and 'u' followed by singular nominals seem to escape Wackernagel restrictions for unknown reasons as in (38 a & b). One possible reason could be the fact that clitics of the same form but with different functions create ambiguity if they simultaneously procliticize to the same noun. Therefore, the ezafe clitic '-i' will win the Wackernagel position while the definite and demonstrative clitics will have the license to escape Wackernagel restrictions.(38) a. I-trombel This-car This car An Account of Clitics in Shabaki with Reference to Wackernagel's Law Volume XIV Issue I Version I](image-9.png "(") 1 3 (28) a. Yâna hin Jassim-i.House belongs to Jassim.Jassim's house(28) b. Trombel-i hin Jassim-i.Car-EZF belongs to Jassim-EzafeJassim's car(28) c. Trombel-i hin he-?.Car-EZF belongs to him-selfHis own car16(28) d. Trombel-i hin he-?-â.Car-EZF belongs to him-self-BE-PRSTIt is his (own) car.(28) e. Trombel hin-I he-a?-â.Car belonging-Ezafe REF-3SG-PRESIt is his (own) car.It is worth remarking, however, that examples in(29) sound somewhat weird and 'hin' is an integral partof the sentence and it is obligatory.(29) a. Hin-a?Belongs to hisHis (own car)(29) b. Hin-a?-â.Belonging-POSS.3SG-PRES.1SGIt is his. © 2014 Global Journals Inc. (US) In extensive sentences, the negative proclitic 'na' also precedes the main verb. When the sentence is VS, the proclitic is verb-initial and breaches the Wackernagel law. (47) a. Am-â. Come-PST-3SG He came. (47) b. Na-m-â. NEG-come-PST-3SG He did not come. The perfective enclitic '-na' ,which seems an independent item from tense, aspect or epistemic modality, takes on an evidential meaning insofar as it expresses indirect knowledge about an event: speaker has not witnessed the event but has acquired information about it from indirect sources by inference, hearsay, etc. It marks speaker's evaluation of the information. The verb ending with '-na' denotes an uncertain or probable action or event. (47) c. Am-â-na. Come-3SG-PERF. He has come. In (47c) above, 'am-â-na' is a present perfect sentence in Shabaki. The perfective clitic '-na' expresses an indirect evidentiality interpretation which allows both a report and an inference reading. ## X. ## Clitics-Related Phenomena a) Clitic clustering In Shabaki, a sequence of a proclitic and an enclitic forms a phonological word. Various types of clitic-clitic combinations occur in Shabaki. In (48), it seems impossible for accusative third person object and third person dative object to cluster together. (48) Dâ-m-a? ba-na-? Give.PST-1SG-3SG.ACC to-2SG.DAT I gave it to him. The reciprocal pronoun 'ef-tar' (each other) is a phonological word which consists of only proclitic and enclitic without a host. Neither constituent can stand on its own. This clitical sequence can be preceded by a set of prepositions which are fused into the sequence. 'A' in the prepositions 'ba' (with), 'ça' (from) and 'na' (towards) is dropped out when pro-cliticized to the reciprocal pronoun 'eftar'. Consider the list in (49). The most interesting point about reciprocal pronouns in Shabaki is that they can be followed by personal pronouns as in (49). In (52a), the subject clitics ? -i' and ?-ne' anaphorically refer to pronominal clitic subject ?-am'. In (52b), the clitic ?-a?' catephorically refers to the full noun ?Ali'. (52) a. Am alâ-t amâ-ne kâ ?â-t kar-i. 1SG to.2SG come.PST.1SG see.2SG do.INF.1SG (LV) I come to see you. ## Conclusions Clitics are challenging for many theories of grammar because they straddle morphology and syntax. In most theories, cliticization is considered a phrasal phenomenon: clitics are affix-like expressions that attach to whole phrases. Constituency-based grammars in particular struggle with the exact constituent structure of such expressions. Clitics are grammatical words, but lack features of a phonological word. This paper has shown that Shabaki has an intricate system of clitics which includes subject pronominal clitics, object pronominal clitics (dative and accusative), possessive pronominal clitics, demonstrative clitics, ezafe clitics, diminutive clitics, aspectual clitics, tense clitics, connective clitics, copular clitics and negative clitics. In general, Shabaki has both simple clitics and special clitics. Shabaki pronominal clitics are special clitics. Shabaki has proclitics, enclitics and mesoclitics. All clitics cliticize obligatorily. They never bear stress. * A-Morphous Morphology SRAnderson 1992 Cambridge University Press Cambridge * Peoples and Cultures of the Middle East, Pearson education DBates ARassam 2000 New Jersey * Albert & MarieBorg Azzopardi-Alexander 1997 Routledge London/New York Maltese. (Descriptive Grammars * Some Middle Persion and Parthian constructions with governed pronouns MaryBoyce Dr. J. M. Unvala memorial volume Bombay Linden House 1964 * A Syntax of western Middle Iranian ChristopherJBrunner Persian Studies Series 1977 * The status of case MiriamButt TracyHolloway King Clause Structure in South Asian Languages VDayal AMahajan Dordrecht Kluwer Academic Publishers 2004 * Etudes iraniennes, I. Paris: F. Vieweg JamesDarmesteter 1883 Caravan Books Delmar * On Clitics in Old Bulgarian and Old Russian CGribble American Contributions to the 10th International Congress of Slavists ASchenker Columbus Slavica 1988 * The best clitic: constraint conflict in morphosyntax JaneGrimshaw Elements of grammar, Liliane Haegemann Dordrecht Kluwer 1997 * Clitics, The handbook of morphology AHalpern A. Spencer and A. M. Zwicky 1998 Blackwell Oxford * Evolving Tobler-Mussafia effects in the placement of French clitics PHirschbühler MLabelle New Approaches to Old Problems: Issues in Romance Historical Linguistics, Amsterdam/Philadelphia, John Benjamins SNDworkin &DWanner 2000 * The Kurds: A Concise Handbook MehrdadRIzady 1992 Washington, Crane Russak * Latin Enclitic Stress Revisited HaikeJacobs Linguistic Inquiry 28 1997 * Connected Speech EMKaisse 1985 Academic Press New York * Between Assimilation and Deportation: History of the Shabak and the Kakais in Northern Iraq MLeezenberg Syncretistic Religious Communities in the Near East BKellner-Heinkele& KKehl-Bodrogi Leiden Brill 1997 * Kurdish dialect studies DNMackenzie 1961 Oxford University Press London * The dialect of Awroman (Hawra?m?an-i?Luho?n). Grammatical sketch, texts, and vocabulary DNMackenzie 1966 EjnarMunksgaard Copenhagen * Grammaire du vieux-perse AntoineMeillet 1931 Honoré Champion Paris * Prosodic Phonology MNespor IVogel 1986 Foris Dordrecht * DonnaNapoli Jo Linguistics Oxford Oxford University Press 1996 * A (phrasal) affix analysis of the Persian Ezafe PolletSamvelian Journal of Linguistics 43 2007 * Causatives in Shabaki ASultan Acta Linguistica Journal 4 2 2010 * An account of light verb constructions in Shabaki ASultan Acta Linguistica Journal 5 2 2011 * Serial verb constructions in Shabaki ASultan Acta Linguistica Journal 6 1 2012 * Evidentials in Shabaki ASultan 2013 Unpublished paper * Ethnicity, Cultural Discontinuity and Power Brokers in Northern Iraq: The Case of the Shabak AVinogradov American Ethnologist 1 1974 * Wackernagel Über ein Gesetz der indogermanischen Wortstellung. Indogermanische Forschungen 1 Jakob 1892/1977 * ArnoldMZwicky Clitics and Particles. Language 61 1985 * Clitics ArnoldMZwicky Oxford The Encyclopedia of Language and Linguistics RE J M YAsher Simpson Pergamon Press 1994 2 * Cliticization vs. inflection: English n't ArnoldMZwicky GeoffreyKPullum Language 59 3 1983