Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to upgrade your browser.

Log In
Sign Up

Figure 58 – uploaded by Saartje Verbeke

See full PDF downloadDownload figure

Table 58

Related Figures (194)

agent of the construction (A), and the suffix —(a)s refers to the first person, the patient (O).

Table 1 Vowel system (cf. Koul 2003: 900) ° The aim of this study is not to offer a theoretical analysis of Kashmiri grammar, but rather to undertake an empirical study of the language combined with the typological goal of finding crosslinguistic evidence for functional linguistic tendencies. As a result, we will not focus on the generative accounts of Kashmiri in this study. ally VY LLU. With respect to grammars, Grierson (1911) is the first extensive grammatical description of the language published in English. Grierson could consult Pandit Isvara Kaula’s grammar of Kashmiri in Sanskrit, which did not only describe the language, but was to a certain extent prescriptive. For instance, Kaula lay down a unified system for writing Kashmiri in the devanagari script. The system has been passed on by Hindu authors, but nowadays, the Perso- Arabic nastalig is the dominant script for Kashmiri. Before Kaula, the earliest specimens of Kashmiri were written in the sharada script, an extinct script which stems from Brahmi, like devanagari. The study of Kashmiri advanced from the nineties of the last century onwards, in particular thanks to the work of Bhatt (1999) whose dissertation on Kashmiri word order is a textbook example of a generative government-and-binding analysis and the work by the native speaker Omkar N. Koul and American linguist Peter E. Hook. Several descriptive grammars of Kashmiri have been published, in particular Kachru (1969), Sommer (1997), Wali and Koul (1997), Koul and Wali (2006). Because of its features, such as V2 word order, that are illustrative for the hypotheses of deep structure and transformation rules, Kashmiri has been a popular language in the generative tradition. Accounts in Government-and-Binding theory and X-bar theory have focused extensively on the phenomenon (Bhatt 1999, Manetti 2011)”.

Table 2 Consonant system (Koul and Wali 2006: 12-14) Historically, Kashmiri has not always been considered an Indo-Aryan language, mostly due to this deviating phonology. It has been argued that together with other languages in the region, Kashmiri is part of the Dardic ‘family’. This group, a division which stems from Grierson, has been situated either within or beside Indo-Aryan, depending on the linguist describing the family trees. Grierson made a division between Aryan, Iranian and Dardic, considering all three to be direct branches of Indo-Iranian (Masica 1991: 461). Most later scholars do consider Kashmiri and its related languages to be part of the Indo-Aryan family, not having split off before other New Indo-Aryan languages developed (since Morgenstierne 1926, also Fussman 1972, Strand 1973). In contemporary linguistics, Dardic is a term used to indicate the geographical area where Kashmiri and other related languages are spoken, without expressing an opinion on the exact genetic relations between them. Nevertheless, as Masica (1991: 462) points out, Grierson was not completely wrong, there is a branch of languages that are indeed different from both the Aryan and the Iranian families, spoken in the same area more to the West, i.e. the Nuristani languages. Nuristani and Dardic languages,

however. Example (2) is a construction with a pronominal suffix in Kashmiri:

(u)s ‘3SG’ is here only possible if the full pronoun femis ‘him’ is not expressed. Consequently, a correct rendering of the above construction is ex. (5) or (6):

From this minimal introduction into pronominal suffixation in Kashmiri, it is clear that the vagueness of the term ‘pronominal suffixation’ in a sense hides the lack of a better definition for grammatical items as variable as these morphemes.

Table 3 Present and past inflection of the copula ‘to be’ in Kashmiri (Koul 2006: 33, 38)

Table 4 Future tense suffixes (Koul and Wali 2006: 101-102)

a arated uw Peer. oe ee Kashmiri has three kinds of past tenses, each of which is expressed with a dif ferent formation. Koul and Wali (2006: 86—87) label them proximate past, indefinite past and remote past, following Grierson (1911: 43). The proximate past is used for events that happened recently, the indefinite past for events that happened at an unknown point of time in the past, and the remote past is used for events that happened a long time ago. Each of the dif ferent past tenses has a different formation. There is a past tense verb stem, in which the stem vowel alternates according to the conjugational pattern, and different suffixes are added to this past tense stem. The formation of the past tenses is given in Table 5.

Tense and aspect in Kashmiri are indicated separately and do not have their indicators mergec into one form (but see Davison 2002 for this issue in Hindi). Apart from the forms of the past tenses, there are also some verb formations which express the perfect aspect. These are formed with the perfect participle ending on the inflectional suffix -mut and an inflected form of the verb dsun. This formation is used to express “‘an event that occurred before the time under consideration” (Wali & Koul 1997: 233). The copula may be in the present, past o1 future tense, to form the present perfect, past perfect and future perfect tense, respectively. Example (8) is a past perfect clause:

Table 6 Case marking in Kashmiri (based on Koul and Wali 2006: 32) There is a high rate of case syncretism, in particular in the plural and feminine paradigm. The nominative is the formally unmarked case, without an additional ending. Nominative plurals of masculine and feminine forms are formed in various ways. Most masculine nouns do not have a distinct plural form. If they do, the plural is formed by a change of vowel or a palatalization, or a combination of both, e.g., the singular mual ‘father’ has the plural, palatalized form myal ‘fathers’. The ergative singular of this group of words also takes the palatalized form; in other words, the nominative masculine plural form is identical to the ergative masculine singular form. The objective form of these particular nouns is -is instead of -as (cf. Wali & Koul 1997: 151). Feminine plurals often take the ending -i, but they can also be formed by the same internal sound change mechanisms as the formation of the masculine plural. Kashmiri has an inflectional case marking system that distinguishes four cases: nominative, objective, ergative and ablative (Sommer 1997, Koul and Wali 2006). The objective case is used to indicate patients and recipients. Traditionally, this case is labeled as “dative”; however, since that case name is predominantly associated with the recipient argument and not so much with the patient, I believe “objective” is the more neutral term (Beames 1872: 252, Iggesen 2005: 92, Verbeke 2013). Nouns inflect for number and gender, as illustrated in Table 6.

however dependent on a number of conditions, illustrated in Section 2.1.3. The objective case is formed by adding the ending —an for plural and -i for feminine singular. Masculine singular nouns show more diversity, with either —as or —is as ending. In ex. (9) Suris, ‘to the child’ is an objective form. The objective is the case marker of the recipient argument R, and it can also mark the O argument. The marking of the latter is however dependent on a number of conditions, illustrated in Section 2.1.3.

one after the other on the stem kor ‘girl’. Pronouns also have a nominative, ergative, objective and ablative form. There are three types of pronouns for the third person: proximate, remote within sight (w.s.), and remote out of sight (0.s.). The personal proximate pronoun yi is also used as a demonstrative. The pronominal paradigm is given in Table 7. af “ The genitive case is a complex matter. Genitives are formed by adding the suffix -hund or -sund to the noun or pronoun. These suffixes in turn inflect for the gender and number of heir modifier (Koul & Wali 2006: 52). However, the genitive must take the case marker of he function of the argument. For example, the genitive argument must be marked with the iblative before a postposition. This is an example of so-called Suffixaufnahme (Plank 1995, <oul & Wali 2006: 38, Wali, Koul & Koul 2002). Example (11) shows the suffixes stacked ne after the other on the stem kor ‘girl’. types of pronouns for the third person: proximate, remote within sight (w.s.), and remote out

Table 7 Pronominal paradigm (Koul and Wali 2006: 51-52) * Note that the objective plural ending -en is a variant of -an.

Similar to the nominal forms, there is also a high degree of syncretism in the pronominal paradigm. The nominative forms are clearly different from the remaining cases. With the exception of the ergative third person masculine singular and the third person objective forms, all other forms are identical across the cases. As already noted by Kachru and Pandharipande (1979: 202), the first and second pronominal ergative form is the same as the objective case form, while the third person takes a specific ergative case form. The authors claim that speech act participants (SAP) do not have an ergative case in Kashmiri. However, the first and second person ablative forms are identical to the objective and ergative form as well, so it seems more adequate to maintain a formal opposition between an unmarked nominative case form and all other ‘oblique’ case forms for the SAP, rather than specifically arguing for the absence of an ergative case. Although its four-case paradigm seems to indicate otherwise, the Kashmiri case system is in that sense remarkably similar to the binary direct-oblique case system found in other Indo-Aryan languages, in which there are only two cases: a nominative and oblique.

In this section, I will focus on the type of alignment found in constructions with an imperfective verb form. Imperfective verb forms include the simple present, the future, past and future progressive, and present and future conditional. The rationale for grouping these verb forms together is not directly tense-related, but is based on the case frame of their arguments, i.e. nominative-accusative. The S and A of these verb forms are generally in the nominative case, for instance bi ‘Tl’ in ex. (12) and (13).

Table 8 Marking of O in imperfective constructions The verb of imperfective constructions agrees with the nominative subject. Gender and number agreement is expressed by a vowel alternation in the stem of the verb. For instance, cha in ex. (18) and past tense dsi in ex. (20) are the feminine singular form of the copula, agreeing with the feminine subject usha (18) and bi (20); chu in ex. (19) is the masculine singular form, agreeing with the masculine bi ‘I’. Person and number agreement is expressed by suffixes, which are different for the present and future tense. In the present and progressive tenses, the finite verb — the copula — takes no suffix when the subject is a lexical NP or a third person pronoun. However, when the subject is a first or second pronoun, pronominal suffixes are added to the verb. The suffixes are different for number, and can also occur when referring to other core arguments. The suffix —s in the examples (19) and (20) stands for the firet nercon cinonlar enhiect

referring to other core arguments. The suffix —s in the examples (19) and (20) stands for the

The marking on the verb can indicate the person of both arguments, as exemplified with

I discuss the further uses of the suffixes in Section 2.2.1.

In constructions with a perfective verb form, S always takes the nominative case form (bi ‘T in ex. 25), whereas A is expressed with an ergative case (mye ‘T in ex. 24, dost-an ‘friend’ in ex. 26, mahni ‘man’ in ex. 27).

Thus, the transitive past verb basically agrees with O, in the form of vowel alternation and the addition of nominative pronominal suffixes. However, when the ergative A is a pronoun, it can be indicated on the verb. The addition of a pronominal suffix is obligatory for an ergative second person A. First and third person pronominal ergative A’s trigger pronominal suffixation on the verb and are optional. In ex. (34), a second person suffix is added to the verb, cross-referencing to the second person plural A, tohi, ‘you’. In ex. (35), the second person singular suffix —th is added to the verb, cross-referencing to tse, ‘you’.

person singular suffix —th is added to the verb, cross-referencing to tse, ‘you’.

‘Did you send this short story anywhere for printing?’ (Koul 1985: 230)

Table 9 Nominative suffixes (Koul and Wali 2006: 117) able 9-Table 11 give t ‘orms in general and with the plural forms in particular. There is neither a form for the fir: derson p ural, nor for t ‘orm of the objective is yy a binding vowel if t night change due to ru he first he paradigm of the three types of suffixes. There is syncretism in th he third person of the nominative suffixes. The first person singulé identical to the ergative form. Note that all suffixes can be precede he verb ends in a consonant. This vowel is in most instances —a, bt es of attraction, e.g. in the form parinavan chi-h-am ‘you teach me person objective suffix —m is preceded by -d instead of —a. In the tables and te» elow, only the binding vowel —a is given. rer “wr 4 7 1° MANHRTF 114-7

Table 11 Ergative pronominal suffixes (Koul and Wali 2006: 116)

These pronominal suffixes are obligatory in all instances. One of the particularities of pronominal suffixation is that the suffixes are used exclusively when the argument roles they refer to are pronominal. This is obvious when these arguments are local, i.e. speech act participants, but it is problematic when the argument is a third person. Under those circumstances, the pronominal suffixes should, theoretically, be added only when the third person argument is a pronoun. This is in contrast to agreement markers, which make the the pro)nominality of the third person argument irrelevant, as the agreement marker is obligatory and always present. The presence of a third person nominative suffix in combination with a third person overt pronominal argument could provide the solution to the definitional problem if suffixes are agreement markers or bound pronouns, see Section 1.2.2. However, it is The first paradigm of pronominal suitixes are the nominative suilixes, given in lable 9. These suffixes are added to the verb which in turn shows the gender and number agreement features, and may cross-reference to 1/ the nominative subject (S/A) of an imperfective construction (the suffix —s of the first person in ex. 41), 2/ the nominative O of a perfective construction (the first person suffix —s in ex. 42), and 3/ the nominative S of a perfective construction (the first person suffix —s in ex. 43), as illustrated in the following examples:

because it is followed by the first person objective suffix -( 3)m. The objective suffixes are used to cross-reference to an objective-marked O in an imperfective construction and to the recipient and experiencer argument. Their paradigm is found 11 Table 10. Their use is relatively straightforward. If O is second person, the suffix i: obligatory, and the pronoun itself may be overt. First or third pronominal O’s do not occu together with a pronominal suffix. Ex. (44) shows an independent pronominal O mye ‘me’ and there is no pronominal suffix for O on the finite auxiliary. In ex. (45) there is nc independent pronoun present, but a first person suffix is added to the verb, - (a)m. The suffip —h is an alternation of the second person nominative suffix —kh, phonologically changec because it is followed by the first person obiective suffix —( 3)m. The presence of a pronominal recipient argument can also be indicated on the verb. As with an objective O, a second person recipient or beneficient is always indicated by a pronominal suffix and may be left unexpressed, while first and third pronominal recipients cannot be present as independent pronouns when their pronominal suffix is added to the verb. For instance, in ex.(46) and (47), the recipient argument is not overtly expressed, but its person and number is inferable from the pronominal suffix that is added to the finite verb. In ex. (46) —(a)s is added to indicate a third person recipient, in ex. (47) —vi is added for a second nerson plural.

More examples in which the verb is marked by an objective pronominal suffix are (49) und (50): More examples in which the verb is marked by an objective pronominal suffix are (49)

‘He watches you from over there.’ possible; the construction can be as shown above; or, one can insert an independent pronoun

and delete the suffix —-m added to the auxiliary verb. That gives the construction in ex. (51):

In the plural, this changes to the construction in ex. (55), where the suffix order is subject- ohiect

A more common translation of the English sentence would be with an independent eA AIIN framics A more common translation of the English sentence would be with an independent

vor ‘he read’, a ditransitive construction would be as follows: An even more explicit construction is the following:

Here, all arguments are indicated on the verb form: the verb agrees in gender and number with an unexpressed third person object and it is followed with a third person suffix —n- referring to the third person ergative subject. The last suffix —s refers to the recipient bachas ‘to the child’. In ‘to read something to someone’, ‘someone’ is not a prototypical recipient, but it is a receiver of an action. Note also that these constructions in ex. (64)-(66) border on being

ungrammatical, since bachas ‘to the child’ is not pronominal and is, as such, not usually referred to with a pronominal suffix. The range of semantic roles that can be referred to by means of an objective suffix is quite large. Aside from being used for lesser prototypical roles such as with verbs as ‘to read’, the suffix can also refer to a possessor argument. The following example is a construction without suffixes:

The latter translation shows that the construction is ambiguous, in the sense that the suffix —us can refer to the possessor, when the suffix is translated with the possessive pronoun ‘his’. The other possible interpretation is that the friend constructs a house for ‘him’, a third person participant. The suffix refers then to the recipient of the house, as it refers to the recipient of the reading action in the example above. The overt presence of tamis ‘him’, the independent third person pronoun, would make the latter meaning completely clear. However, the co- occurrence of an objective suffix and an independent objective pronoun is not allowed. The same rule applies to the overt possessor. A construction such as ex. (69) is possible, but here the suffix necessarily refers to a third person outside of the speech act situation, who is certainly not identical to the referent of temsind ‘his’.

certainly not identical to the referent of temsind ‘his’.

-‘T took the book to/for him.’ It would be clearer to express it as in ex. (72): The suffix -s seems to be able to refer to a wide range of arguments that are not overtly *xpressed in Kashmiri, such as an (indirect) beneficiary or a non-overt possessor, similar to he use of the pronominal suffix in Sindhi and the Lahnda languages (see Section 5.2.3). While Hook and Kaul (1987), following Grierson (1911: 167) refer to this use of -s as a lativus commodi, -s can in fact be used on many more occasions than the ones that raditionally fall within the purview of the dativus commodi. Perhaps the use of the objective suffix -s is better described in terms of a dativus ethicus, which expresses a certain degree of ‘motional involvement.* Prod x ao Ue oy Pr - . 7 7 é y - * co Again, the construction from (71) is ambiguous; the meaning could either be the one above,

In some instances, third person suffixes can be used to refer to animate nouns instead of the more common third person pronouns. However, in these instances as well the suffix and the overt argument cannot occur together with objective suffixes for recipients. Consider example (73): the overt argument cannot occur together with objective suffixes for recipients. Conside

act situation.

circumstances, the verb is followed by a suffix cross-referencing to the pronominal nominative O, and this suffix shares its form with the ergative suffixes. In examples (77) and (78), the suffixes —th and —(a)n respectively refer to a second person and third person O. Formally, they belong to the paradigm of ergative suffixes. They are obligatory. Formally, they belong to the paradigm of ergative suffixes. They are obligatory.

The verb form che-s-an agrees in gender and number with bi (F), and gets a nominative suffix -s referring to the first person subject. It is followed by the third person singular suffix —(a)n, referring to the O of the sentence, su. Third person suffixes —(a)n and —kh can also occur in past tenses, not referring to the ergative subject, but to an animate object. Section 3.3 focuses on these exceptional constructions.

h’ekun ‘can’ and yitshun ‘want’. Examples (80) and (81) illustrate this pattern:

In ex. (81), the ergative suffix —(a)n is added, referring to the animate nominative O gdav ‘cow’. This O is the patient of the verb heny ‘to buy’, which is also evident from the gender- number agreement between the feminine singular patient gadv and the verb heny. The suffix that is added to the copular che-s is, in other words, not referring to the O of the verb yetshdan, but to the O of its dependent verb heny. However, to consider it as “raising” in the traditional sense of English linguistics (e.g. ‘he seemed to play the piano’ as the “raised” construction of ‘it seemed that he was playing the piano’) is an inaccurate equation, as other morphosyntactic factors play a role as well, rather than only a syntactic change of subject. Since neither the participial form nor the infinitive form in these constructions are finite, the pronominal suffix can only be added to the copula verb, which is the only verb in the construction that is able to carry inflectional markers. wn © cee x . eas rae gam pre 4 ~~. m a so 1 a cee mT eae c ae.

Table 12 Grierson’s paradigm of pronominal suffixes (1911: 60) Grierson managed to incorporate all suffixes into one paradigm. As the ergativ onstruction in the Indo-Aryan languages was considered in his age as a construction with yassive agent, the suffixes used to indicate the ergative subject are labelled “agent suffixes’ Clearly, however, he makes a difference between the “accusative” suffixes, used to indicat he object of an imperfective construction lower in person than A, and the genitive/dativ suffixes, used to indicate recipient and O’s higher in person than A. However, he does nc onsider this as a matter of a hierarchical relation between the arguments, but simp! nentions that accusative suffixes cannot be used if the subject is a third person singular, an inder these circumstances, they should be replaced by a genitive/dative suffix. Grierson’ srammar discusses extensively the phonological system of Kashmiri, and the vowel change hat accompany the morphological formation of verbs and nouns. He considers the abov -onsonants the suffixes, but they can be preceded by a or u, depending on the form (Grierso Kashmiri:

Table 13 Independent arguments and ergative pronominal suffixes The major part of the corpus are constructions with an independent pronominal subject, without suffixal marking (134 = 45%). The majority (89) of this total are third person subjects, half of that amount (45) are first person subjects. As expected, no second person subjects are found in this category, since these pronominal suffixes are obligatory. 13 occurrences are overt nominal subjects which are combined with third person pronominal suffixes. All of these are animate arguments, as in the following example, where the suffix - (a)n refers to ramisan: 2.3.1.2 Results for the ergative suffixes

suffixes. All of these are animate arguments, as in the following example, where the suffix -

FEWONMLIUOEYU TUILD atu” A SAVIULVY VV AE VU VMAPMY Vata. We do not find constructions that combine a third person subject with a second persor ybject (e.g., “he sees you’); as a result we cannot assess the obligatory rule of second persor uffixal marking. The reason is that this is written language, which already has a smalle: lumber of speech act participants, and this combination in particular is rare. However, since he grammars are very clear on this particular point, we can assume that this rule is followed ind this is also confirmed by spoken language data (Section 2.3.2). The following table give: he number of the occurrence of the objective pronominal suffix referring to an O ir mperfective tenses, that is higher or equal to the A-argument.

Table 15 Number of independent arguments and objective suffixes for recipients and experiencers The results agree quite clearly with the prescription in the grammars. The attestation of a pronominal suffix in combination with an independent pronominal argument is a second person argument, which makes this construction grammatical. Of our 516 constructions, 382 have no recipient or experiencer argument, and are not included in Table 15. Quite a large

The table shows that most objects are independent without suffixal marking. This is in line with the general consensus on objects, because most of them are inanimate and automatically are not referred to by pronominal suffixation. Another large part of the dataset, 48 construction have no independent O, though the verb is transitive. Only three of them take a pronominal suffix. Independent pronominal objects are not frequent, and most of them are marked with suffixes, following the rules. When suffixal marking is absent, it is because the object is inanimate, though it is referred to by a pronoun, as in ex. (83): The names of the different types of pronominal suffixes are not randomly chosen. Nominativ suffixes refer to nominative arguments, objective suffixes to arguments in the objective cas Though the semantic roles of the arguments can vary slightly, the correspondence betwee wgument and suffix is obvious. There is an exception to this regularity with regard to th rgative suffixes, as explained in Section 2.2.1.4. In rule, these indicate the ergative argumen .e. the subject of a transitive perfective construction. However, the same paradigm of suffix« s also used to refer to a nominative O in imperfective constructions. The O takes th 1ominative case when it is lower in person than the subject. This particular use of the ergativ suffixes is obligatory, in other words, each animate nominative object argument needs to t eferenced on the verb.

Table 17 Distribution of independent subject pronominal arguments: in constructions with nominative suffixes

Here, the suffix —an is used to refer to bubiji, and it is plural to indicate respect. The other group of nominative pronominal suffixes refers to the nominative O of a per fective construction. However, these constructions are not very common, since the O- argument needs to be a speech act participant in order to be indicated with a pronominal sufi firs fix. In general, most O’s in our corpus are inanimate. I found only seven attestations of a t/second person O in a perfective construction, but remarkably, five of those constructions had no independent O-argument, only a pronominal suffix.

had no independent O-argument, only a pronominal suffix.

SMM: TV ER SE OS EM ee Ee OU: oe, SESE eee Besides these oral tasks, the same constructions used for the translation task were also given to a group of seven BA students to elicit written translations. Since some verb endings barely differ from each other in pronunciation, I hoped to resolve that problem by having them written out. To my surprise, there was a substantial difference between the written and the spoken translation tasks. The constructions which were asked to be translated orally, were almost all translated with independent pronouns. For instance, the preferred translation of “I saw you” in the oral exercise is: saw you” in the oral exercise is:

None of the speakers translated the construction without independent pronouns. Of course, in such elicitation tests it is expected that the speakers will feel the need to be completely clear and to give as much information as possible. It is telling that during the same exercise in a written test, six out of seven participants gave the full form, with independent pronouns and pronominal suffixes. In the story-telling tests, on the other hand, there are occurrences where there is no independent pronoun but only a suffix. It seems that at least in discourse, the occurrence of pronominal suffixes is related to the absence of an independent argument, and this is in turn again determined by pragmatic and syntactic motivations such as topicalization of an argument or coordination constructions. Pronominal suffixes then seem to function primarily as reference-tracking devices. In constructions without independent pronoun, they safeguard the consistency of the clause and the continuing reference to the correct argument, avoiding an ambiguous interpretation (cf. Corbett 2006: 275).

‘The carpet became filled with water.’ (For: ‘Water spoilt the carpet.’)"’ 1! Note that kalin is an inanimate subject here, but the construction is made intransitive and not ergative. Although it is generally implied that pronominal suffixes are used exclusively to refer tc pronominal arguments, sometimes one finds exceptions to this rule. For instance, it has beer reported in Hook and Koul (1997) that when the ergative subject is an inanimate third person the ergative suffixes -n (sg.)/-kh (pl.) are added at the end. Hook and Koul (1997) focu: particularly on “expressions of natural forces and natural processes’, as in ex. (89). particularly on “expressions of natural forces and natural processes’, as in ex. (89).

‘This old cupboard stayed above all.’ (For: ‘The old cupboard dominated the room.’ )

Ex. (92) was considered as ungrammatical; however, replacing ab-an ‘water’ with rud-an ‘rain’, made it slightly more acceptable. The construction in which it is replaced by rang bursh-an ‘paint brush’, seemed very awkward to native speakers, they all preferred the translation with s#t. Cross-linguistically, this is what is expected, since the paint brush is not a typical agent, but a typical instrument. The same is valid for the following construction, with another typical instrument ‘stone’: typical agent, but a typical instrument. The same is valid for the following construction, with

and syntactically determined. Dixon (1994: 23ff) introduced a systematic difference between languages with a predominantly syntactic alignment and those with a semantically-based alignment. The arguments in languages of the former type are syntactically identifiable: they respond to the syntactic roles of subject/object, or to the case roles of A (ergative), S and O. The languages with a semantically-based alignment show different case markings depending on the semantic features of the arguments or the verb (i.e. agentivity, volitionality etc.). Traditionally, the varying marking is visible on the S-argument, which either aligns with A or with O, Jepending on its semantic features (Wichmann 2008: 3-4). Though the term “semantic ilignment” is primarily used to indicate a varying marking of the S-argument, it is clear that variation in marking of other arguments can also be related to semantic features (Donohue 2008: 50-58, Hopper & Thompson 1980). Donohue defines semantic alignment as “the ohenomenon whereby basic alignment properties of a language can be best described by uppealing to semantic factors, rather than syntactic ones” (Donohue 2008: 24). The >xperiencer construction, which occurs in many languages, is an example of a construction yased on semantic principles (see ex. 96 from German, in which there is an accusative subject nstead of the syntactically expected nominative, cf. Donohue 2008: 58-62 for Icelandic), but hat does not imply that the case marking of most of the constructions in these languages is Jetermined by the agentivity or other semantic properties of the arguments. On the contrary, he fact that these constructions are singled out by linguists as deviating from the norm, confirms that the “normal” pattern of case marking in this particular language is accusative, and syntactically determined. Equally, the phenomenon of “differential object marking” can be considered as an instance of alignment based on semantic features as well. For instance, Hindi-Urdu shows a typical differential object marking pattern (Bossong 1985). In Hindi-Urdu and in many other Indo- Aryan languages, the marking of the object is dependent on the semantic criteria of animacy and/or definiteness of the object. In ex. (97), kitab is nominative, but it takes the objective marker ko if the speaker wants to convey that it is a specific book that the boy is reading, as in ex. (98) (Mohanan 1994),

Consider ex. (98): The object kitab ‘book’ gets an objective marking. The alignment of this particular example is not economic, since both the arguments are explicitly marked with their respective case markers. Ergative and accusative alignment patterns are economic, since only one argument needs to be marked to distinguish the functions of both arguments (Comrie 1989, Croft 2001: 101-102). The case marker ko is a syntactically redundant marking, its main function is to semantically convey that the book is specific. In sum, the best-known Indo-Aryan language Hindi-Urdu shows patterns of semantic alignment through differential object marking, the marking of the object being dependent on the extralinguistic animacy or specificity properties. The phenomenon of differential object marking is well-known in the literature (Bossong 1985, Lazard 2001, Aissen 2003, Sinnem&ki 2014 amongst others). It is not very traditional to include it as a type of semantic alignment; however, it is, in its most basic definition applied, a type of case marking that abides to semantic principles.

Table 19 Marking of O in imperfective constructions The hierarchy that is particularly relevant for Kashmiri case marking is the person hierarchy (see Sections 2.1.22.2.1.5). Person is one of the features that are part of Silverstein’: referential hierarchies, and this particular feature has been linked mostly with verb agreemen (Nichols 1986, Siewierska 2004). In the person hierarchy, first and second person are the highest ranked as speech act participants, followed by third person and ultimately by nouns Surprisingly, it is in the case marking that we can observe co-argument conditionec alignment, as illustrated in Table 19 (repeated from Table 8): the case marking of the C argument is dependent on its ranking in person compared with the A argument. There are some constructions which might pose problems for this case-marking pattern. For instance, constructions with reflexives have an A and O referring to the same referent (e.g. ‘you saw yourself in the mirror), which would imply that neither A (‘you’) nor O (‘yourself ) is higher in person. As a result, the case marking is objective. In ex. 101 , the form pdnas is used, an objective form inflected as a noun, hence used as a third person animate O which is equal in person to the subject.

equal in person to the subject.

Furthermore, ditransitive constructions might pose a problem for the marking of O. In fact, if an O that should be marked in the objective case occurs in a ditransitive construction, 11 may lose its objective marking. There is a restriction on two objective-marked arguments in one clause, and the objective marking of the recipient argument is obligatory (see Section 2.1.2). A ditransitive construction in which both O and the recipient argument should be marked rarely occurs (the presence of three animate arguments in one clause is uncommon). If it does, the double objective marking is considered to be ungrammatical by some speakers (Hook and Kaul 1987: 60). TO i ee ee ee cae cg ee a ee ee OU OM eee es ieee ze

third argument n6dkar-as ‘servant’ .“ On the contrary, nokar-as is an objective form. The default case of the recipient is objective, as for instance mye ‘me’ in ex. (104).

* Hook and Koul (2006) argue that there is a semantic difference between these two constructions: the first one is a true “causative”, the second one is a “sociative causative”, which means that the agent “helps” the causee. They offer the alternative translation of “He will help the servant write a letter.” By making this distinction, they hope to explain why the postposition athyi is in one construction present and absent in the other. However, causative constructions which only have two animate arguments in general do not use the postposition athyi (e.g. su cavyi timan cay ‘he will give them tea to drink’ or md cha Suryis dwad cdvan ‘mother is feeding the child milk’). Hook and Koul (2006) argue that these last two examples are “derived trivalents” and not “causatives”, I fail to see the difference however, and maintain that the most straightforward explanation for the use of athyi is the restriction on two bare objective markers in a clause.

objective case, as for instance mye ‘me’ in ex. (104). nominative, though she is the recipient of the action. The expected form would be an TE EN RE Hook & Koul (2006: 63) give the four possibilities for the same construction: ‘Mother ha ne give milk to the girl’ in examples (106)-(109). According to them, the case marking of th bject arguments varies depending on the focus of the construction, 1.e., which argument i onsidered to be the patient. In the first sentence, it is bi, ‘I’ in the second one, it is kiir, ‘girl’ n the third one, it is dwad, ‘milk’, and in the last one all arguments are considered alike hough this possibility is called “less felicitous”, according to Hook and Koul (2006: 63). Th erb agrees with the nominative arguments bi, ‘T in, kur, ‘the girl’ in, and dwad, ‘milk’ in, i ender and number, and takes a pronominal suffix for bi, ‘I’, if it is in the nominative case (

The hierarchically determined case-marking pattern in Kashmiri as summarized in Table 19 has led to an interpretation in terms of an “inverse and direct transition” of alignment (Hook and Kaul 1987). The terminology direct-inverse was borrowed from Amerindian linguistics (cf. Zufiga 2006), and has mostly been applied to these languages. It is based on the view that verbal and nominal coding has a priori the function of distinguishing the arguments. The direct pattern is the more “natural”, or functionally unmarked situation (Comrie 1989: 128), in which A is higher on a referential hierarchy than O. In the inverse pattern, O is higher on the hierarchy than A, or both O and A are on the same level (Hook and Kaul 1987). In Algonquian languages, the inverse pattern is sometimes indicated on the verb with a separate affix. For instance, the following examples (from Wolvengrey 2011) are from the Algonquian language Plains Cree, which is well-known for its intricate hierarchical verb agreement pattern. The language is agglutinative to a high extent, with infixes, suffixes and prefixes, referring to the person of the arguments. Ex. (110) illustrates the direct pattern, with a suffix —d indicating that this is a direct construction, with a first person A and a third person O. Ex. (111) is identical, only the third person plural is now A, while the first person is O. The suffix indicating that this is an inverse construction is -iko. The suffixes -d@ and -iko are the only indication of the different content of these two examples.

In Plains Cree, as in most languages which display the phenomenon, the indication of direct/inverse patterns is found in affixation to or in the verb. The constructions above are formally identical, except for the insertion of different suffixes right after the verb: @ indicates a direct construction: the highest of the arguments is the agent of the construction; iko represents the inverse construction, in which the lower third person argument is the agent (Haude & Zitifiiga 2016: 445). Another language with a similar person marking pattern is the Tibeto-Burman language Nocte (DeLancey 1981). This language is also spoken in India, yet far removed from Kashmiri. Examples (112) and (113) are from Nocte:

with an accusative case, apparently because A is a ‘low’ third person.

Plains Cree and Nocte are languages with a complicated verb formation pattern. However, regardless of the case marking vs. agreement differences, there are some similarities with the case marking pattern in Kashmiri. For instance, in the direct pattern, the marking of both A and O is nominative. cf. tsi ‘you’ in ex. (115). In the inverse pattern, A is nominative and O is objective. Ex. (116) shows the “inverse’

In other words, an obviative marker indicates that the marked argument is “relatively low in terms of animacy and/or discourse prominence” as opposed to the proximate argument (Haude & Witzlack-Makarevich 2016: 435, Aissen 1997: 705). The definition of the process of obviation is then “the obligatory ranking of third person nominal (nouns, pronouns) based on discourse function, syntactic relations, and semantic properties such as animacy” (Nichols 2011: 661, based on Aissen 1997). This is a broad definition for something which is generally assumed to be a very specific feature of head-marking Amerindian languages. The concept of obviation has been borrowed by Nichols (2011: 661ff), when she distinguishes a pattern of obviation in Ingush purely on the basis of syntactic behaviour, e.g. certain constructions are impossible because of the non-coreferentiality of the proximate and the obviative argument. Ingush is a dependent-marking language and has no proximate or obviative marker. Yet, certain syntactic phenomena are explained using the idea of obviation, a ranking of the arguments according to discourse prominence, animacy, or other semantic criteria. In this broader sense, the concept of obviation might be also relevant for Kashmiri. obviative is the lower one (Aissen 1997: 705). One finds markers for the obviative/proximate distinction in the Algonquian languages and in other languages with a direct-inverse pattern, though not in all of them. In Nocte for instance, the difference is not indicated, and person is the only feature that determines the alignment (DeLancey 1981: 642). The following examples (117) and (118) from the Algonquian language Plains Cree illustrate the difference between the obviative and the proximate argument (examples from Haude & Zufiga 2016: 446, who adapted them from Wolfart 1996: 410). The obviative argument is here atim-wa ‘dog’, marked with wa for the obviative. The proximate argument is ndpéw, ‘man’, without any formal marking.

Table 20 Ergative pronominal suffixes (Koul and Wali 2006: 116) The marking of O in imperfective constructions presented in Section 3.2 is a highly economical system, in which there is no superfluous marking. When O is outranked by A in person, O remains in the nominative case. However, a nominative O is obligatorily marked on the verb with a pronominal suffix. The paradigm of this suffix is the ergative one, repeated here in Table 20.

Table 21 Case marking of O and presence of pronominal suffixes referring to O in imperfective constructions

Table 22 Future suffixes (Koul and Wali 2006: 101-102) Ex. (120) is a a second person future tense clause:

Table 23 Person suffixes of a transitive future verb (Koul and Wali 2006: 103) In any future tense construction with an animate nominative O, the future forms from Table 23 are obligatory. The verb form in ex. (121) can only end on -(a)n, not on -4. The following examples illustrate the pattern with a first (bi ‘TP’ in ex. 121, 124) or second person (tohy ‘yow in ex. 122, 123) subject, and with a singular (panun dos “my friend’ in ex. 121, su ‘he’ in ex. 122, 124) or plural (tim ‘they’ in ex. 123) O.

The general rule is that the formally more complex future suffixes from Table 23 are used when O is nominative and animate, in which case they are obligatory. However, these suffixes can optionally be used to refer to any nominative O, also to an inanimate 0. Examples (126)-(128) illustrate this. Either O is not indicated on the verb as in ex. (126), or it is indicated, and then we find the verb form par-an as in ex. (127). In ex. (128), we find that the same options are possible with the verb ‘to do’. In ex. (124), the form mari instead of maran would be impossible. The form on -i can be used in constructions where the O balances on the line between animacy and inanimacy, for instance, in ex. (125), where khat ‘sheep’ is considered rather as a means of consumption than as an animate being. as an animate being. instance, in ex. (125), where khat ‘sheep’ is considered rather as a means of consumption than

[is tempting to explain this use as an extension through analogy: speakers use the —(a)n fc vith animate O’s and get so used to the form that they extend them to inanimate C .ccording to my informants, the choice between ex. (126) and (127) is entirely optional, - does entail a semantic difference that they find difficult to pinpoint. Some accounts rention a difference related to specificity, in the sense that the suffix is used when the C 10re specific (Bhat 1987: 129; Koul and Wali 2006: 102). This would mean that idication of O in ex. (127) implies that this O, yi kitab, ‘this book’ is more definite/spec: yan the O in ex. (126). In light of the original use of the suffix to refer to animate O’s, 1 vould be a reasonable assumption, as animacy and specificity are often linked, for instance atterns of differential object marking (see Section 3.1.1). The pattern of differential obj 1arking that is wide-spread in the Indo-Aryan languages, is absent in Kashmiri. This patte s illustrated in ex. (96)-(97) from Hindi-Urdu above, dictates that animate or specific ( sceive case marking, inanimates and non-specific ones do not (Masica 1986). The use of (a)n/-(a)kh suffix in Kashmiri seems similar to that of Hindi-Urdu ko: both the suffixes < ye case marker indicate that the O argument is animate/specific. Yet native speakers cashmiri do not consider specificity marking to be a convincing explanation. They do f 1at there is a certain shift in meaning between the two options, but that shift lies more in reaning of the action. For instance, a construction such as ex. (129), which refers t pecific book, is considered as completely grammatical and acceptable without the suffix — 129 ak | kitab. pagal h mez- che bi | par-i | yi nae—neth

specific book, is considered as completely grammatical and acceptable without the suffix —7.:

person subject, for instance:

According to the native speakers, these constructions do not depend on the specificity of O. Nevertheless, they are used with a particular meaning, i.e. as imperatives. The meaning of the future is here more enforcing, the action certainly needs to be performed. This is in contrast to the use of the simple future second person form pariv, which would denote “only” a future. Interestingly, constructions as ex. (132) with the verb ‘to see’ are less acceptable. This construction is deemed impossible by most Srinagar-based speakers:

construction is deemed impossible by most Srinagar-based speakers:

they also occur in perfective constructions, referring to an animate O. For instance: Here, the syncretism does not cause ambiguity, which confirms again the economy hypothesis. A more problematic syncretism is that of the third person plural, of which the objective suffix is identical with the ergative suffix, i.e. —kh. It implies that third person plural animate arguments will be marked on the verb, irrespective of their case. ha. wea. nf the: ocrffiw~acd —§fa)e ann — “fab tea wat anls: limmitadn ta thieA narann naAaminatissa. CY’

the O is a third person singular: These suffixes occur here unexpectedly, from a grammatical point of view. The O of an ergative perfective construction is in the nominative case, and can be indicated by means of nominative suffixes. However, there are no nominative suffixes for a third person, and it seems as if the ergative paradigm of suffixes fills a gap, again, being an economic solution. Hence, though grammatically unexpected, a usage-based approach to grammar explains this particular use of the suffix in terms of economy. We notice a clear tendency towards an obligatory verbal indication of any nominative argument overall, irrespective of its (pro)nominality. Moreover, rather than person, these suffixes indicate the number of the argument on the verb.

a _ I argue that the ergative suffixes used to cross-reference to nominative, animate O’s, function in a way similar to obviative markers: they indicate the argument that ranks lowest in discourse prominence, being an animate third person O occurring in the same construction with an SAP A argument (a “mixed scenario” Zuniga 2006: 48, Haude & Zuifiga 2016: 446). A second person nominative O argument is also indicated with an ergative suffix. This second person suffix cannot be considered as an obviative marker, strictly speaking, since obviative is a term only applied to third person arguments (Aissen 1997). However, considering the “rule” that all second person arguments are obligatorily indicated in Kashmiri, it is no surprise that the use of the suffix is “extended” to second person arguments as well. Note that the future has a secondary conjugation, only based on the number of the third person O argument (see Table 23). Considering these suffixes as obviative markers opens perspectives on the question why it is, perhaps bizarrely, the ergative paradigm of suffixes that cross-reference to an O argument. The hypothesis of DeLancey (1981) on split-ergativity argues that O is the viewpoint of the speaker in ergative constructions, this is the argument in focus. One could draw the line that the ergative argument is then the less prominent argument, the more obviative-like argument. Thus, there is a cognitive resemblance between the ergative A argument and the nominative O argument in imperfective constructions: they are both the less prominent argument in the co-argument relation, and they should be marked accordingly. Obviously, this explanation is tentative and needs further exploration. "WA MARR See BPiaslhkkiisaoess 36 Se. eer esa Zee hn Teta A tetas Leet aesceaeaes +e <ee | OUOBAo Ta

Table 24 Pronominal suffixes in Poguli (based on Hook 1987) Thus, the pronominal suffix system in Poguli has four groups of suffixes instead of three although for each person, many suffixes overlap in form. There are nominative, ergative objective and accusative suffixes’. The pronominal suffixes in Poguli also indicate gende (Hook 1987: 64), although this only seems to be the case for the nominative singular suffixe and the third person suffixes. It is, however, also possible that the gender and numbe agreement triggers the gender differentiation and not the suffixes, given that soun alternations may be caused by the combination of agreement markers in the stem and th attached pronominal suffixes. Nevertheless, I follow Hook in giving both the masculine an feminine forms of the suffixes, e.g., -uws (M.), -is (F.) in Table 24.

a a I One such an explanation is that one could think of this phenomenon as a strategy of politeness: a second person always needs to be acknowledged in speech as a matter of respect. Bickel et al. (1999) have shown that maintaining one’s own face while at the same time showing empathy for the status of the fellow speech act participant is highly relevant for the agreement pattern in Indo-Aryan Maithili. In Maithili, combinations of person markers occur according to the level of honorificity of the speech act participants, and their mutual place in the honorific hierarchy. Consider the following example from Maithili (Bickel et al. 1999: 482): In this example, all animate arguments are indicated on the verb. The second person is non- o1 middle-honorific, the third person nominal argument kaniyd is indicated as a third persor honorific. In Maithili, the agreement morphemes for the first and the second honorific persor are identical. Bickel et al. (1999: 497) invoke the concept “face” (Brown and Levinson 1987 to explain this: a speaker maintains face in front of a hearer by not differentiating between the marking referring to herself and that referring to the addressee, whereas a third persor honorific is marked separately. A second principle is “empathy”, a concept borrowed fron Kuno (1987). Bickel et al. (1999:503) use empathy to explain the fact that in Maithili, “objec agreement is more likely either if the object is itself a speech-act participant or if the subject 1: a speech-act participant”. The same sensibility in Kashmiri may be indicated by the fact that < second person is deemed important enough to be consistently marked by the verb. With the marking, the speaker encodes empathy and respect towards the addressee. Fe ee , : : : ., ee we cen mas

The suffixes -s/-t, are not specifically associated with an argument in the ergative case. They primarily refer to a subject in the second person. Intransitive verbs in the past tense, for instance, receive the -s/-t when S in the nominative is a second person, as in bas-l-d-s ‘you (m.sg) sat’ (example taken from Pandharipande 1997: 412). The suffixes -s/-t are also the endings of the second person in the present tense, irrespective of whether the verb is transitive or intransitive (Wali 2006: 11; Pandharipande 1997: 409). The following example is from Wali (2006: 11): shows the marking —s.

When the perfect participle in combination with a copula is used to express the present perfect, -s is added to the copula (which, if it is in the present tense, is the normal suffix), but the participle itself displays no marking. The resemblance with the treatment of second person pronouns in Kashmiri is remarkable, the only difference being that in Marathi the suffix always refers to A/S, irrespective of the case (nominative or ergative), and whether the verb is transitive or not. Rather surprisingly,however, Hook (2011) notes that in Old Marathi this -s/-t suffix cross-referenced to O, not to an (ergative) subject, as in the following example from (Master 1964: 130-131): Wali (2006: 11):

The exclusive use of pronominal suffixes in Marathi to refer to a second person perhaps indicates that it is a phenomenon that is directly related to the specific features of a second person pronoun rather than to features belonging on a referential hierarchy. There is no other evidence that Marathi is particularly susceptible to influence of the referential hierarchies, apart from the differential object marking pattern which it shares with Hindi-Urdu. If honorificity is the explanation for these second person suffixes, then pragmatic considerations rule over the syntactic rules on suffixes. The nominative suffixes can be considered as agreement markers, and the objective and ergative suffixes react as optional person markers.

Kashmiri shows V2 word order in main clauses, cf. examples (142)-(147), in which the copula verb dsus/cha always occurs in the second position. The main verb, i.e., the verb carrying the agreement features, follows a first constituent. The default word order in Kashmiri has the subject argument in the first position, as in ex. (142) and (147). Other NPs can be fronted when topicalized (ex. 143-145). If present, the non-finite participial verb form either takes the clause final position (preferably) or immediately follows the auxiliary. The pattern in which the participle is fronted, as in ex. (146) is highly infrequent. The default pattern of the main clause is identical to that in the Germanic languages German and Dutch. pattern of the main clause is identical to that in the Germanic languages German and Dutch.

Kashmiri subordinated clauses differ in word order: relative and adverbial clauses are verb final (similar to Dutch and German), other subordinated clauses such as root clauses (indirect speech) or complement clauses retain the V2 word order. Examples (149)-(150) from the corpus illustrate this pattern.'

‘She was the only woman who didn’t make fun of him.’

between the subject and the inflected verb. Most Indo-Aryan languages display verb final word order, as Hindi-Urdu, e.g. (152):

In Indo-European, V2 word order is considered a branch particular feature, a common development of a number of Indo-European languages. The question is what elements have contributed to the development of V2 word order in Kashmiri. In the next section 4.1.1 I pose the question if the phenomenon of pronominal suffixation is somehow linked with V2 word order, as both are distinct features which set Kashmiri apart from most other Indo-Aryan languages. Section 4.1.2 and 4.1.3 look at the issue from a historical perspective, and ask the question if there is any historical evidence for a simultaneous development towards both pronominal suffixation and V2 word order.

Haiman’s hypothesis was investigated for European languages such as Romance and Germanic, and for Kashmiri. Hook and Manaster-Ramer (1985) showed that his claim is not applicable to Kashmiri. Firstly, it is not clear at all that Kashmiri is a Type A language. As with most Indo-Aryan languages, Kashmiri takes a pragmatic approach towards the presence of its core arguments, and the subject argument is often absent. In our dataset, 25% of the constructions in total has no independent subject (see Section 2.3.1.5). Even if there is no independent subject, this has no consequences for the V2 word order. For instance, the insertion of a dummy subject is simply avoided by the construction. In ex. (153), the famous “dummy” subject construction ‘it rains’ is translated to Kashmiri. The verbal action of “raining” is simply replaced by a nominal paraphrasis: raining” is simply replaced by a nominal paraphrasis:

Other constructions without independent subject include coordination constructions, where he subject is not a new topic. They occur frequently in my corpus, and in spoken languages. ‘or instance, examples (157)-(159), which occur without an independent subject, were all receded by sentences in which the subject was introduced. The inflected verbs in these xamples, viz. korun, ‘she did’, wuchhum, ‘I saw’ and vonun, ‘she said’ are transitive yerfective verb forms, with an ergative pronominal suffix added referring to A. In principle, hese forms can also occur without this suffix, but with an independent A instead, though, yarticularly in spoken language, the independent pronoun is absent and a pronominal suffix is idded instead (see also the corpus data on the ergative pronominal suffixes in section 2.3.1.2). Section 2.3.1.2).

“While laughing, she greeted Subhas.’

second position, and not preceeded by an independent subject, as illustrated in ex. (160). As an interjection # can also occur in the second position, e.g. in ex. (161).

A second argument against Haiman’s hypothesis that V2 word order creates a problem for type A languages, relates to the argument structure of questions. In Kashmiri, the insertion of a question word changes the word order from V2 to V3; hence the subject does not shift position to come after the verb, on the contrary, the verb shifts its position. Consider ex. (162), where the inflected verb chi-vi has shifted to the third position, preceded by the subject tohy and the question word kydzi. Ex. (163) shows a yes-no question, introduced by the question marker kyah. The objective experiencer argument is in the second position, the inflected verb cha follows. This is an experiencer construction, more literally translated as ‘Is there trust on astrology to you?’.

there trust on astrology to you?’.

In sum, the non-expression of an independent subject does not seem to influence the specific rules of word order in Kashmiri. The idea that there are type A and type B languages is based on the study of a few European languages and has not gained much credence in typological linguistics. The expression of the subject argument is often pragmatic and not regulated by syntactic rules. In Kashmiri, any argument can be left unexpressed in spoken language. Nevertheless, there seems to be a connection between word order and pronominal suffixation, in particular with respect to the suffixes referring to a subject argument. Since the V2 word order also needs to be sustained in coordinated constructions, the addition of a pronominal suffix guarantees the identification of the subject argument. The connection can however only be established from a historical point of view by looking at the assumed evolution in Kashmiri from verb final word order to verb second. consequently, there is no shift of the verb to the third position:

Table 25 Class 1 conjugation in Old Kashmiri There is also a second class past intransitives, which is only attested for singular forms:

Table 26 Class 2 conjugation in Old Kashmiri The endings of class 1 past transitive conjugation are identical to those of class 1 past intransitives. However, the past transitive verbs agree with the object, and pronominal suffixes referring to the object as well as to the ergative subject may be added. Toshkhani (1972) already mentions the use of pronominal suffixation in the BK, and identifies this as an

Table 27 Present tense conjugation of the copular verb

Table 28 Case marking of nouns in Kashmiri

Kashmiri formally distinguishes a nominative, ergative, objective and ablative case. Though there are intricate phonological changes (Grierson 1911) which make the case marking sometimes rather opaque, and though there is overlapping of case endings to a great >xtent, inflectional case marking has not disappeared in Kashmiri. In contrast to Dutch and English, Kashmiri has retained inflectional case marking in nominal forms as well, and not only with pronominal forms.

Table 29 Postpositional case marking in Hindi-Urdu This list is not exhaustive. Other postpositions indicate other locative constituents (Reindhl 2016). The Old Hindi inflectional case marking system continued in the form of a nominative and oblique case, inflected for gender and number, with the use of the nominative restricted to raditional nominative functions, such as the intransitive subject and transitive subject in imperfective constructions, and inanimate/indefinite O’s. The oblique case is seldom used on its Own, except in some temporal or locative constituents. Non-nominative arguments are indicated by means of the postpositions ne and ko. The same pattern of reinforcement is found in the languages of the Hindi belt (Rajasthani, Bhojpuri), in Punjabi, Gujarati, in all main Indo-Aryan languages. In the Eastern Indo-Aryan languages, the use of postpositions is not so UULY Witt PlOHOIII Tal LOLUIS, Compared to other Indo-Aryan languages, Kashmiri dependent marking has not been reinforced with postpositions. A comparison with Hindi-Urdu clarifies the difference in development. After the case syncretism which happened in the Middle Indo-Aryan stages of Prakrit and Apabhramsa (Bubenik 1998), Old Hindi had two morphological cases left, generally denoted as nominative (or direct) and oblique case (Strnad 2013). The nominative was used for intransitive subjects and transitive subjects in imperfective constructions, and fot the objects of ergative constructions. The oblique was used for the ergative subject of a perfective construction, and for (indirect) objects. In Old Hindi texts, however, one starts to see the introduction of postpositional markings, reinforcing the marking of the argument roles. For instance, indirect objects received the marking ko, which were later also extended to mark animate or definite O’s (Stmad 2013). Another important innovation was the introduction of the ergative postposition ne, arguably borrowed from Old Rajasthani, where it was used as a marker of animate arguments (Tessitori 1914, Verbeke and De Cuypere 2009). The introduction of postposition led to the case marking system of contemporary Hindi, as illustrated in Table 29:

In the BK, there is a clear tendency of the copular verb not to occur in the final position of the construction. There is no clear V2 word order pattern as yet, but forms of the copular verb do occur either in SVO word order, or VSO. I counted the attestations of the copular verb in the texts. There are fewer than would be expected in a Modern Kashmiri text. Reasons for this are multiple: Old Kashmiri follows the same tendency as the Old and Middle Indo-Aryan languages, that there is no obligatory overt existential verb, and statements of facts, or descriptions, are given with nominal forms. Also, the present tense is still generally expressed with a synthetic form, not with the periphrastic form that is common in Modern Kashmiri. In the appr. 1600 verses long text, the present tense of ks- occurs 55 times, either used with the lexical meaning of an existential verb, or in combination with a participial form. Revealingly, in only a few of the attestations, the verb occurs in the final position of the construction. In other words, verb final word order in Kashmiri is not the default. Only the construction in ex. (167) shows the copular verb in final position, where another word order pattern would be as likely:

In Modern Kashmiri, the word order would be similarly verb final: ‘... After he climbed on a great wagon that was yoked with thousand horses.’

‘In which many leaves are.’

For questions, the V3 word order as in Modern Kashmiri is also attested, see ex. (174).

these constructions are still verb final. The copular is not in final position, a constituent is mostly following. Nevertheless, these examples show that the cliticization of the auxiliary verb has not led directly to V2 word order. On the contrary, it seems as if the participial verb is shifted to the front in an equal way as the copular auxiliary. Obviously, our data is still to scant to make diachronic conclusions. However, it is interesting that particular the construction with this tendency of fronting the participial form occurs quite frequently in the tawt ‘Today, how do you carry on, how is your behaviour?’

‘Listen to this, o king, [now that we are] together.’

Table 30 Pronominal suffixation in Old Kashmiri Toshkhani (1972) discerns four paradigms of suffixes; however, as in Modern Kashmiri. the ergative suffixes are identical to those used to refer to first and second pronominal objects of imperfective constructions (accusative). The difference with the objective paradigm is only noticeable for the second person singular, with th for the ergative and accusative, and y for the objective. The Old Kashmiri system shows a clear deviation from the system in Poguli, where there are separate paradigms for these two grammatical roles (see Section 3.3.2). This seems to imply that Poguli has developed separately from Kashmiri from as early as the Early New Indo-Aryan stage, and is perhaps not so closely related to Kashmiri as is generally assumed. What is clear, is that the second person suffix has not reached the degree of obligatoriness that it has in Modern Kashmiri. Ex. (182) above is a construction with a second person subject where the second person suffix is absent. This is not a unique accident, even in the more recent Lalla-Vakydani, Grierson and Barnett (1920: 140) report one example in which the second person suffix is absent, though there is an independent pronoun: tse gol" ‘you destroy’. whereas the expected form would be tse goluth.

Table 31 The Vedic enclitics (Macdonnell 1916) Pronominal suffixation in Kashmiri seems to be a language-internal development rather than a feature that has been passed on from an ancestral language spoken at least more than 1000 years before the first Old Kashmiri manuscripts. In the Bandsurakatha, there is no evidence of enclitic forms of the pronoun occurring in the second position of a clause. On the contrary, the pronominal paradigm has already been simplified, and there is only one form per pronoun (though there is variation in spelling, and there are Prakrit and Sanskrit forms which have marked occurrence). The pronominal paradigm in Old Kashmiri is given in Table 32 :

Table 32 Pronominal paradigm in Old Kashmiri (Toshkhani 1972, on the basis of the forms attested in the BK)

‘Shiva smiled at him, that I greeted him.’

In both the examples above, the objective third person suffix —s(a) is in all three verbs, asisa, kyamasa and haronasa. The subject of asisa is the proper noun hara ‘Shiva’, whereas the subject of the other two verbs is not independently mentioned. The first person ergative suffix —m(a)- 1s used in kyGmasa, and the third person ergative suffix —n(a)- (cerebralized for phonetic reasons) is used in haronasa. The distribution of first and t hence quite similar to that in Modern Kashmiri: they do not seem to independent argument. Example (189) shows the verb carrying a first hird person suffixes is occur anaphoric to an person objective suffix attached to the verb mus-, without an independent pronoun. Example (190) shows that the suffix —m is also used to refer to a first person objective, and in Mod ern Kashmiri only —m remains. Example (191) shows the use of the independent objective pronoun mi, and the verb does not take a suffix. does not take a suffix. remains. Example (191) shows the use of the independent objective pronoun mi, and the verb

Note that in ex. (182), the ergative subject is a second person. However, it is not indicated or the verb, which means that the rules for the pronominal suffixation of second persons are not so fixed as yet, and even second person suffixes tend to occur in complimentary distributior with an independent argument. For instance, in ex. (192) the verb does not show the objective second person pronominal suffix —y, yet the independent pronoun fsi is expressed. All three possibilities are given in ex. (192)-(194): ex. (192) only has the independent pronoun, ex (193) has both the pronoun and the suffix, and ex. (194) shows only the pronominal suffix. Ir sum, though the second person pronominal suffix can be used together with the independen pronoun, there is no fixed rule yet which makes its use obligatory. pronoun, there is no fixed rule yet which makes its use obligatory. The order of the suffixes is ergative-objective, as in Modern Kashmiri. Objective suffixes are also able to indicate a possessor, as is illustrated in ex. (188), where —sa in haronasa ‘he took’ refers to the owner of komdra ‘maidenhood’. The objective third person suffix is often mentioned, in particular after verbs of speaking, as in ex. (195)-(196) with the verb nigad- and dapp- both meaning ‘say’ or ‘tell’. Similar constructions with the verb wonun ‘say’ are very common in Modern Kashmiri (cf. Section 2.3.1.3).

“Why rises the fear of battle in your heart?’

Following Anderson (1993), we conclude with the following hypothesis: Since Anderso akes Wackernagel less strict than Hock, and defines the constituent which needs to take th second position as the constituent that carries enclitic features, not as the constituent that : nclitic itself, it might well be the case that when the auxiliary verb or the main verb started t carry person marking, it was shifted towards the second position. The mere fact that bot features are attested, though not yet standardized, occurring together in exactly this perioc seems to point to this conclusion. Cross-linguistically, another language with V2 word orde: ngush, also has pronominal suffixation, and direct-inverse marking (Nichols 2011 However, a comparison between Kashmiri and its closely related neighbor Poguli, offers eve More evidence in support of our claim. Though often described as a dialect of Kashmir Poguli shows interesting deviating features.

‘All those who came early, raise your hands.’ (Wani 2015)

One can perhaps link these two varieties to the distinction between the language Khah and Poguli, both spoken in the same area around the town of Banihal (cf. Dhar 2013). Dhar (201 3) gives examples of Khah speech, which all show the V2 word order, also the constructions he considers as Poguli. The area of Pogul Peristan lies in between Srinagar and Jammu, in between two large spheres of influence. Srinagar, the capital of Kashmir, is predominantly Muslim. The language of education, literature and politics is Urdu written in the Perso-Arabic script. Without taking into consideration the commercial value of English, Kashmiri deemed to be a symbol of the identity of the Kashmiri people, and sets them aside from t rest of India. Jammu, on the other hand, is very much Hindu-orientated, with Hindi and t devandagari script as main means of communication. Though there are local languages, such as Dogri, and though they are spoken by highly educated people, they form no threat for t dominance of Hindi, as Jammu people consider themselves truly North Indian. As such, t is he he ne ne people of Pogul Peristan are somewhere in between. The teachers at the public schools come

Table 33 Paradigm of pronominal suffixes in Poguli We find four types of suffixes, referring to the most important arguments of a (di)transitive construction. A so, the suffixes seem to differ for masculine and feminine gender, and there is much overlapping and syncretism. According to this table, there is no gap for the third person nominative suf fix, but what looks like simple gender/number endings are given, and in the third person plural a vowel is added. The nominative third person suffix ends in a vowel, which makes t form more like 2009). he hypothesis that the third person suffix has been incorporated in the verb ly than the alternative that there never was a third person suffix (Siewierska

In Poguli, there is clear gender and number agreement with O in perfective constructions, ‘onsider ex. (206)-(209): In Poguli, there is clear gender and number agreement with O in perfective constructions,

the object tu. The use of a pronominal suffix is also found in ex. (210):

The verb form dit- is here complemented with a third person ergative suffix —(u)n, referring to the subject tanviri.

<ishtawari is the second language that has been considered a dialect of Kashmiri. The area in vhich it is spoken is close to the Poguli area, around the same road south from Srinagar owards the city of Jammu, but more concentrated towards the west. The main cities and areas vere it is spoken are the district Kishthawar, and in other bordering districts (see Tariq et al. 014). No research has been done focusing exclusively on Kishtawari grammar. Tariq et al. 2014) is a sociolinguistic report on how the language is used by the speakers, which shows hat it is extremely limited in use: it is not used in education systems, nor in formal or written ettings. However, some data on Kishtawari has been gathered by Kashmir University, in the ‘ontext of a project describing the languages of Kashmir.'' From that data, it is obvious that <ishtawari strictly follows the V2 word order, just like Kashmiri, and it seems that verb final vord order is not a possibility, unlike Poguli but identical to Kashmiri. Examples (211)-(213) ire constructions with identical lexical meanings, in an imperfective and perfective tense. The v2 word order is obvious, as is the use of pronominal suffixation. In ex. (212), there is a rononominal suffix —m, used for the ergative first person subject, and ex. (213) the third yerson ergative suffix —n is used. person ergative suffix —n is used. ™'T owe much gratitude to Nahida Ali, for giving me access to the data on Kishtawari.

order of the suffixes is ergative-objective, identical to Kashmiri, but different from Pogulli.

In the data, there is one instance (ex. 218) of an experiencer construction with an objective subject which shows a co-occurrence of both independent pronoun and pronominal suffix. There is not enough material to consider this occurrence as evidence that the complimentary distribution of pronominal suffix and independent pronoun is not obligatory in Kishtawari; we would need more data to confirm this. In Kashmiri, the following construction would be ungrammatical, either the independent pronoun mye would be used, or the pronominal suffix —m added to the copular verb chu-.

—m added to the copular verb chu-.

University, data gathering on Puncchi has rendered some sample data.

‘The government gives money to the poor.’

Table 34 Languages of the Kashmir area The languages included in Table 34 and discussed in this chapter belong to two language families, Indo-European and Tibeto-Burman. Most of the languages are Indo-European, and belong either to the Indo-Aryan or the Iranian branch of the Indo-Iranian strand of Indo- European. The Indo-Aryan languages can be split in four groups: Dardic, Lahnda, Punjabi and Pahari.

Dranul is Clearly a UMMAITECUONAL ONE, It Makes NO sense tO MCiude Hranul 1 US OVELVICW. The second language, which is a real isolate, is Burushashki>*. Burushashki is more relevant to our study because it is spoken in the area of Gilgit-Baltistan, and there is even an enclave of “Buru’s”, Burushashki speakers, in Srinagar. Burushashki shows some interesting features related to pronominal suffixation. However, the affixes which take some of the functions alsc found in Kashmiri, are prefixal. For instance, prefixes are added to nouns when referring tc invariant possessees. Second, in transitive constructions, O is indicated on the verb by means of a prefix (Willson 1996, Berger 1998, Anderson and Eggert 2001, Munshi 2006). Note that the nouns in Burushaski are divided into four different “inflectional classes”, according tc their referential properties (human masculine and feminine, animate and inanimate) (Anderson & Eggert 2001: 235). Consider ex. (223) and (224), of which the arguments belong to the human feminine and human masculine class.

Examples (223)-(224) make clear that the subject is indicated on the verb, feminine -wmo in ex. (223), masculine -imi in ex. (224), irrespective whether the construction is transitive or intransitive. This suffix is more of an agreement marker than a pronominal suffix; it cross- references to the nominal subject argument and indicates the class of this argument. The object is indicated by means of a prefix mu- in ex. (224). Willson’s (1996) further discussion makes it clear that it is primarily the animate object which is prefixed on the verb, Anderson & Eggert (2001: 237) note that “class-IV [inanimate] objects are not marked in general by the prefixal agreement position”. The object prefixes show a greater resemblance to the Kashmiri pronominal suffixes; however, their position in the verb form is totally different. The prefixes can also occur on a non-finite verb of a construction, as on del in ex. (225). In Kashmiri, only the finite, inflected verb, can take a pronominal suffix.

From these few examples, it is clear that the verb changes according to the object and —s is added when there is a first person pronominal argument that is not the subject. Grierson gives a number of other examples, but they do not make the pattern more clear; in particular with respect to the subject agreement with verbs in a perfective tense it is unclear whether this happens explicitly, or whether the verb simply assumes a default form in these tenses, and shows no agreement at all. In intransitive constructions the agreement is clear, and Grierson distinguishes between, for instance, the masculine singular form go, the feminine singular form gy- and the plural form gye-, for the verb ‘to go’. In short, though the information on Brokskat is minimal, there are at least observations of pronominal suffixes, referring to object arguments, patients or recipients, in the first person. They seem to occur only with verbs in perfective tenses, which makes them susceptible to being interpreted as markers of a traditional kind of ergative agreement with the object. However, this would be contrary to the dominant feature of Shina, namely that there is agreement in number and gender with the subject for all tenses. person, by means of person markers (Schmidt and Kohistani 2008: 158). These verba suffixes refer to a subject argument and are obligatory, I consider them to be inflectiona agreement markers. To find out whether Shina has the feature of pronominal suffixation, w must focus on suffixes referring to objects. Of the Shina group, only Brokskat shows clea evidence of such pronominal suffixation. According to Grierson (1903-1926, Vol. VUI, 2 217-218), in Brokskat, in the past, there are suffixes which refer to the direct or indirect objec (1“ person -is, 2"°/3" person —1). Grierson (1903-1926, Vol. VIII, 2: 217-218) gives th following example for “something (m.) struck me”: kutedho-(i)s, with kutedho being the pas participle. Combined with a 2™ or 3 person object, the form would be kute-t. The suffixe occur in combination with the independent pronoun, as the following of Grierson’s example show:

Liljegren (2014) points out that there seems to be a trend towards ergative case marking without ergative agreement in the languages of the Greater Hindu Kus h. He refers to other anguages which show this alignment pattern, as for instance Nepali. However, if we accept hat the development in alignment from Middle Indo-Aryan to New Ind o-Aryan was caused yy a reduction in the verb paradigm and the development or reinforcement of the ergative case was a consequence of this reduction, then it remains a problem to d etermine why these anguages have actually lost the verbal ergativity while retaining the ergative case marking. A yossible explanation is found in a variety of Shina which was spoken far away from the -entral Shina speaking area, and after the influx of other languages, became separated from its elative languages, Palula (Liljegren 2014, 2013). Palula, also known in older sources as The other varieties of Shina, Astori, Kohistani and Gilgiti, do not show pronominal suffixation; but, as mentioned above, there is agreement with the ergative subject. Since the Dardic languages were recognized as Indo-Aryan, Shina is considered as deriving from Old Indo-Aryan, and together with this assumption it follows that the language changed its alignment from accusative to ergative due to the reduction of the verbal paradigm (Verbeke 2013: 91). The past participle became in Middle Indo-Aryan the only means to express a past action, which led to a construction in which the verb agreed with the object, and the subject was expressed in an oblique case. The Shina languages show a different case marker for the subjects of transitive and intransitive constructions, which indicates that the language has indeed passed through this stage and developed an ergative case. For some Shina varieties, this ergative marker has been reinforced, as it now also occurs in imperfective constructions. For instance, in Kohistani Shina, the marker is an Indo-Aryan variable —e, -i (masculine singular), -o (feminine singular and ji (plural) (Schmidt and Kohistani 2008: 53) in the perfective constructions, but it is —s in the imperfective constructions, which seems to have been borrowed from the Tibeto-Burman languages spoken in the area (as already mentioned by Bailey 1924). Examples (229)-(230) show the different case marking. Ex. (229) is an imperfective construction with the ergative marker —s, ex. (230) is perfective with the feminine ergative marker —d. feminine ergative marker —d.

Phalura (Morgenstierne 1941, Masica 1991, cf. Liljegren 2008: 19), is related to Shina, but is spoken far from the Shina area, and not in contact with Kashmiri (Liljegren 2009). According oO Liljegren’s intricate description of the language, the alignment pattern of Palula is tripartite mn case marking, with an ergative and an objective case. In the imperfective tenses, there is no srgative marking (Schmidt and Kohistani 2008: 51), but the perfective tenses show an srgative alignment with respect to the verb agreement. In the perfective, the verb agrees in sender and number with the object, that is generally in the nominative case (like in Kashmiri, ut not like in Hindi-Urdu), except when it is pronominal. Consider ex. (231), where the verb agrees in gender and number with the masculine object ma ‘T’. agrees in gender and number with the masculine object ma ‘T’. The Palula pattern places Shina firmly in the track that most Indo-Aryan languages have followed, towards an ergative pattern in the perfective constructions. The only way of explaining the accusative verb agreement is by postulating the add ition of pronominal suffixes, referring to the ergative subject.’ This leads us to the conclusion that the Shina languages had an ergative alignment pattern, to which pronominal suf. ergative subject were later added, leading to their exceptional pattern agreement combined with ergative case marking. Only in Palula, perha fixes referring to the today of nominative ps due to its distance from Kashmiri, the suffixes were not added, and in Palula the alignment, including the verb agreement, remained ergative. In Brokskat, on the other side of the suffixes were extended to object arguments, and it is telling that they oc first person arguments, reflecting the exceptional status of speech act Shina spectrum, the cur in particular with participants in form. Hook (1990), in a study of the Shina spoken in Skardu, which is geographically close to the Brokskat area (or even identical), notes that there are attestations in which the verb agrees with a dative experiencer argument. He gives ex. (232).

Table 35 Verbal pronominal suffixes in Pashai (Lehr 2014: 213)

between the O pronominal suffix —i- and the A pronominal suffix —am. before the specificity marker, and also at the end of the form.

Table 36 Presence of the pronominal suffixes depending on the person hierarchy of A and O

attached to the verb, but to the O of the imperative construction:

Morgenstierne (1967) considers the nominal suffixes identical to the verbal pronominal suffixes. Taking Lehr’s (2014) point of view implies that these two types of suffixes are different, and would mean that the recipient suffixes of the examples above are of the second, nominal type of pronominal suffixes. Despite the geographical distance, there are a number of similarities between the pronominal system of Pashai and that of Kashmiri. As in Kashmiri, suffixes referring to different arguments are added one after the other, with the order of the suffixes determining the argument roles of the pronouns. Moreover, there are similarities even in form between the suffixal paradigms. —m is the suffix for the first person, suffixes on —an or —n are t hird person related. Lehr gives —ad- for the second person plural, and Morgestierne (1967: 114) mentions —at or Gt, which bears similarities to the Kashmiri —th for second persons. The pronominal suffixes occur much more frequently referring to speech act participants; only in the present- future non-specific constructions, the third person is always represented with a suffix. pronominal

In sum, there are similarities with Kashmiri, though they seem to be more conceptual than oncrete: semantics of the arguments, specificity and the level of speech act participancy all lay a role in determining the presence/absence of pronominal suffixes. However, there is one mportant difference with pronominal suffixation in Kashmiri: the suffixes in Pashai are not ptional. Morgenstierne (1967) does report a varying occurrence of pronominal suffixes. pparently at random, but Lehr (2014) clearly shows that there occurrence is dependent on the ense/aspect of the verb and on the person hierarchy. In other words, there is no optionality, 0 room for the speaker to use or to leave out a pronominal suffix. She only observes fixed ronominal suffixes. This is, perhaps, also the reason why she does not report on suffixes eferring to recipient arguments. Tr panalracinn tha Aiffaranpac hatuyaan Kachminm and Dachai mala it plaar that tha anffivac According to Lehr (2014: 218), for the present-future non-specific tenses, the O pronominal suffixes change in form when A is a second person. This could be because of particular phonetic rules associated with the second person forms, or perhaps it is a reflection of the special status of the second person argument, as it has in Kashins. (see Section 3.4). eT . 4 41 +4 yr 1 UmIKt1a oe we ae a i 7 4 Par | a5 4F

remains identical, even though the pronominal subject changes: Bhaderwahi is spoken in the area between Srinagar and the town of Banihal, where also Poguli is spoken, but here also there is no pronominal suffixation. Recent fieldwork at Kashmir University has rendered data from the language Puncchi, a Pahari language spoken in the district and the town of Poonch in Azad Kashmir, and again, there is no trace of pronominal suffixation. For instance, consider examples (248)-(249), in which the verb form remains identical, even though the pronominal subject changes:

which has only theoretical implications, but no evidence for it has been found in the language Thirdly, there is another set of clitics which are always combined with postpositions, e.g. da ta: to you. As the clitics are alternatives for independent pronouns, they appear in a range 0! functions. Their typical place is immediately after the first constituent of the sentence, in th second position, typical for clitical forms. For instance, in ex. (250)-(251), the clitics functior as ergative subjects (examples from Tegey 1977: 138, glossing my own. The ergative subject: clitic is glossed as oblique): The same forms are also used to indicate objects, for instance in ex. (252):

The pronominal clitic ye follows the first constituent o f the construction wror-mo ‘our brother’. It preceeds the verb, and as such there is some discussion whether these forms are en(do)clitic, i.e. following or inserted into a constituent, or proclitic, i.e., proceeding the verb form (Bégel 2010, Roberts 2000). !? Note the form mo, which is the possessive clitic for the first person. This clitic has a different form from the “weal k” pronoun. Also, the verb in the present shows personal endings (-i in ex. 252 and 253) whic pronominal form. h can occur in combination with a

The weak pronouns, or pronominal clitics, are used to indicate ergative subjects and direct objects. For possessors, other clitics are used, and for indirect object, a clitic in combination with a postposition is used. In such instances, the clitic is attached in front of the postposition. Consider ex. (253): In short, Pashto has an extensive clitic system, which does not show many similarities with the system in the Indo-Aryan languages (Dardic, Lahnda) however. The clitics are freely attached to any constituent of the sentence (though there are preferences), which makes them alternatives for the overt pronouns that are only used for emphasis. Interestingly, modals and adverbials can also occur in cliticized forms, attachable in the same way as the pronominal clitics. The system in Pashto shows much more similarities with the old Vedic enclitic system, rather than with the pronominal suffixation pattern of Kashmiri." Then Tens oon. ohewmad «xf Gea These Teexerneexeeree 4 wexoclkkiash. dwanleedexe: SBlelekss aeod Glasewhkews worwrcrens

Related topics:

History Computer Science

Connect with 287M+ leading minds in your field

Discover breakthrough research and expand your academic network

Explore
Papers
Topics

Features
Mentions
Analytics
PDF Packages
Advanced Search
Search Alerts

Journals
Academia.edu Journals
My submissions
Reviewer Hub
Why publish with us
Testimonials

Company
About
Careers
Press
Help Center
Terms
Privacy
Copyright
Content Policy

580 California St., Suite 400

San Francisco, CA, 94104

© 2025 Academia. All rights reserved