The dative arguments are cross-linguistically drawn on either the syntactic notion of ‘indirect object’ or the semantic notion of ‘recipient’ (Lambert, 2010). The case prototypically marks a recipient phrase in GIVE construction (Newman, 1996) and encoding indirect object (Blake, 2001) appears in the dative case. The indirect object appearing in dative case are expressed by clitics in many languages. The same syntactic and semantic notions are found in Pashto for dative arguments. Pashto has nominative, ergative, accusative and dative cases based on their function and thematic roles. The indirect object and the subject with the thematic role of the experiencer appear in the dative case (Rahman & Bukhari, 2014). The present study is concerned with the description of pragmatic propositions for dative argument

realizations in Pashto. The discourse-pragmatics component is the realization of information structure having a role in linking the semantic and syntactic representations of the clause. However, the role of the information structure varies across languages (Van Valin, 2005). These responsible conditions are mostly clause-external, not clause-internal in nature. The referents identified by dative arguments can be cognitively active, cognitively less accessible or they can be inactive. The dative case in Pashto is the oblique case preceded by the preposition da and pa or followed by the postposition 'ta’ (Rahman & Bukhari, 2014).

da 'of (POSS)'

pa 'by means of, with; at (time expressions)'

ta 'to (DAT)'

The postposition as a whole is considered a lexical category and the preposition is a functional one in order to get a unified treatment of the directionality of heads in the language (Roberts, 2000, p. 62; Rahman & Bukhari, 2014). The following examples show the dative case in Pashto.

1) a) za hagha ta khat likam

1.SG.NOM 3.SG.OBL DAT letter.ABS write.PRES.IMP.NOM

I am writing him a letter.

b) Ahmad ma ta kanzal wukral

Ahmad.ERG 1.SG.OBL DAT abuse.ABS do.PST.PF.ABS

Ahmad abused me.

The dative argument above realizes the indirect object. It may also realize the subject of the thematic role is experiencer, for example. The dative arguments (realizing subject) taking the preposition (or type of preposition) and postposition are influenced by the volition on the part of the subject as well. If there is no volition on the part of the subject, postposition is used (2a). If the volition increases a bit, preposition is used (2b), and if the volition is less than the volition in the (2b) example, then ambiposition is used (Rahman & Bukhari, 2014).

2) a) Maa ta chaa kar pa hawala karay

1.SG DAT someone work PREP assign do.PST.PF

I have been assigned work by someone.

b) pa Ahmad da kar boj day

PREP Ahamd PREP workload be.PRES

Ahmad is very busy.

c) pa Ahmad bande khatal ghwari

PREP Ahamd POST.over climb want.PRES

Ahamd should be beaten.

The above examples show that postposition is used when there is no volition on the part of the subject. In the first example (2a), the subject has been assigned a work without his volition. The postposition ‘ta’ is used after the first NP realizing the subject in the clause. In the second example (2b), the subject (Ahmad) has an increased volition and the preposition ‘pa’ is used before the subject. The third example has no volition at all on the part of the subject (Ahmad cannot be ready to be beaten) and so, the ambipositon is used in such constructions.

There are two types of VCs (the morphological realization is Pashto). One of these functions as an agreement marker, while the other can replace full dative arguments or other arguments in the clause. The following examples (3a,b) show these types of VCs. In the first example (3a), the VC (agreement markers) occurs with the corresponding dative argument realized by VC; in the second example (3b), the VC realizes the full NP in a postpositional phrase. If both these types of VCs occur in a single clause, it causes clitic doubling in the language.

3) a) Za DAR ta paise DAR legam

1.SG.NOM 2.VC.ERG DAT money 2.VC send.PST.PF

I am sending you the money.

b) Za WAR ta waim

1.SG.NOM 3.VC DAT tell.PRES.IMP

I am telling him/her/them.

The first VC in (3a) above functions as a proper argument, while the following one identifies this argument in terms of personal features. The dative clitic can occur alone by identifying empty arguments (Rahman, Anees & Khan, 2020). This study explains the different encoding possibilities of the dative argument which are discourse-pragmatic in nature. But in some languages like Spanish and Pashto, the argument position for dative and accusative are doubled by clitics. The clitics even if considered as agreement markers like the dative clitics are different from agreement markers. For example, the agreement markers are suffixes in Pashto but the dative clitics are proclitics as they precede the verb and attach (or incorporate) to the verb if not intervened by perfective aspect marker or by particle showing negation, both of which are considered clitics by Roberts (2000). According to this assumption, the dative clitics are considered inflections and the accusative clitics as determiners (Bleam, 1999). The indirect object doubling is more frequent than the direct object doubling in all languages if they have clitic doubling. Jaeggli (1986) analyzes the asymmetry of direct and indirect object doubling in terms of the different nature of clitics. He assumes that the accusative clitics are ‘obligatory case absorbers’ and dative clitics are ‘optional case absorbers’ since the accusative clitics absorb the case features assigned by the verb, a coreferential accusative phrase was only possible when case features were assigned by some other means, specifically the preposition ‘a’. The special clitic ‘a' is not a preposition in French but it is considered a case marker, not assigning but realizing a dative case. The dative doubling is more frequent than the accusative doubling because it is not constrained by the referential features of the lexical phrase co-referring with it. The dative clitic is possible with ±animate and ±specific (Suner, 1988). The dative clitic is banned nowhere and it is obligatory in many structures, for example, if the dative phrase is realized by a strong pronoun, if the coreferential phrase occurs preverbally, with experiencer thematic role, if the dative is realized by a proper name, or with any of cocinar-type verbs which may incorporate a dative argument (Company, 2006, p. 49; Mayer, 2003) or if there is an inalienable possession display (Soriano, 1993). The Spanish datives have been described in terms of their semantic content and morphosyntactic properties. In terms of the predicate structure, some datives are associated with verbs of transfer like dar ‘give’ and some with verbs of action like cocinar ‘cook’. In some of the verbs of the dar ‘give’ type, the nominal exclusively marks the dative argument. The following example shows the exclusive nominal.

4) Carolina Ø dio mucha importancia a la pelicul Carolina gave much importance Dat the movie

Carolina gave a lot of importance to the movie.

The clitic is the head of a dative clitic phrase (DCIP) in a VP shell-type structure and the associated non-clitic full NP occupies the specifier position of this dative clitic phrase (Demonte, 1995). The verbal clitics have dative features to them (Roberts, 2000, p. 64). In many languages, the predicates with two complements show alternation in their argument structure. In English and other Germanic languages, verbs can project into [V NP1 P+NP2] or into [V NP2 NP1].

5) a) John gave a book to her.

NP1 (ACC) NP2 English

b) John gave her a book.

NP2 (ACC) NP1

c) John aghe la kitab warkro

John 3.SG DAT book.ACC give.PST.PF

John gave her a book. Pashto

d) Jan gaf haar hef boek. Dutch

e) Hans gab hhr das Buch. German

But in some languages like Dutch, German and Pashto, the NP2 takes the dative case, unlike the accusative in English. This complement alternation in some languages appears in their morphology, for example, in Chichewa, the [V NP P+NP] structure has another form which is expressed by a complex verb with an applicative suffix (Demonte, 1995). So, the language that assigns structural case by the preposition ‘to’ as in English unlike the oblique case assigned by the preposition a ‘to’ in these languages (Romance languages) behaves as a root and not affix- like preposition (as in English) and they assign oblique case (Baker, 1988). We find only [V NP P+NP] argument structure in Romance languages. In clitic doubling construction in these languages, the dative clitic co-occurs with the full lexical a + NP constituent complement. The dative alternation in Spanish is assumed to have similar syntactic and semantic properties like dative alternation in English. The goal and benefactive arguments in clitic doubling constructions have similar semantic and syntactic properties to double object structure in English, and constructions without clitic doubling are more like prepositional ditransitive. The indirect object anaphor is possible in sentences without the dative clitic in Spanish. The presence of the affected clitic blocks the rising of the internal argument (Demonte, 1995).

There are languages where the verb agrees with the object, the object here is marked for some of its features. The Turkish objects, for example, are marked for specificity (Enc, 1991), the Uralic languages for definiteness (Comrie, 1989) and Persian for identifiability (Roberts, Roberts2005). However, these agreement markers cannot replace the arguments not carrying the pronominal specification of the object arguments (Belloro, 2007). The dative arguments doubling is considered obligatory in Albanian which occurs with wh-words, indefinites and definites (Belloro, 2007) and it is obligatory for both IOs and DOs involving definite NP in Macedonian. The clitic doubling in Slovenian, though mostly occurs with the first and second singular person, yet it is not limited in terms of number, person or case; and all pronouns can be doubled. But mostly the pronouns in genitive, accusative and dative case have corresponding clitic forms and so can be doubled. Though clitic doubling here shows the DP to be the topic, it is not the defining characteristic of clitic doubling because the DP has always a discourse background and the participants are present in the context. Since only the pronouns are doubled here, so, the specificity condition for doubling cannot be tested in Slovenian. The DP in Slovenian can be doubled but the wh-word can be doubled only in the dative, not in the accusative or genitive. If the accusative (direct object) doubling occurs at all, it is more restricted than the dative (indirect object clitic) doubling. The nature of these restrictions may be syntactic, semantic or pragmatic. In Tamazght (Ouali, 2006), the IO-doubling is optional and the dative clitic may occur with specific animate or inanimate referents. Doubling construction is more common with IOs than with DOs. It is more common with animate than inanimate, with specific than with non-specific and with definite than with indefinite (Belloro, 2007).

Bleam (1999) considers dative clitics as agreement markers (inflexions), heading the doubled object construction containing the indirect object and accusative as determiners, having the semantic and morphology of determiners. In Spanish, the clitic is obligatory in doubled object constructions but it is optional in Greek doubled object constructions and so, has different syntax in both languages unlike that proposed by Demonte (1995) and Bleam (1999).

The data in these studies suggest that dative clitic doubling is greater than accusative clitic doubling. The dative argument appears as a personal pronoun, while the direct object as NPs. Direct object doubling happens with definite, while indirect object doubling with indefinite objects (Barrenechea, 1987). We have direct and oblique forms to represent ergative, nominative (or absolutive), accusative and dative cases in Pashto. The direct cause is the unmarked nominative case and the oblique case may be accusative (in the present) and ergative (in past), or it may be the dative followed by a case assigning postposition. According to Roberts (2000), the ergative NP has the agent role in Pashto as the experiencer role, for example, is commonly expressed by the dative.

6) a) Maa la yakh shaway wo

1.SG.OBL DAT cold.M.ABS do.PST.PF.M.ABS be. PST

I was feeling cold.

d) Maa la yakh kege

1.SG.OBL DAT cold.M do.PRES

I am feeling cold

e) Da Kiran na zamaa pars wrak sho

POSS kiran DAT 1.POSS purse.NOM lose.NOM do.PST.PF

Kiran lost my purse.

f) Da hagha kiran khwakha da

POSS 3.SG.M kiran.F.NOM like.PRES.F.NOM be.PRES

He likes Kiran.

The above examples show that the experiencer role is expressed by dative case and so the verb in turn agrees with the absolutive in the past; while in present, the verb shows default agreement as the subject here appears as dative and verb does not agree with object in accusative from. The verbal clitics have dative features to them, they correspond to NPs that are marked dative or other sort of indirect arguments (Tegey, 1977; Tegey & Robson, 1996; Babrakzai, 1999; Roberts, 2000). They can double overt NP, which has dative features. The following examples taken from Roberts (2000) show it clearly.

7) a) ahmad (taa ta) yaw kitaab DAR kaw-i

Ahmad PN.2.SG DAT one book 2SG(DAT) do.3SG

Ahmad is giving you a book.

b) meene (mug ta) mewa RAA wa legel-a

Meena PN.1.PL DAT fruit.F.M 1.PL(DAT) PERF sentFEM3SG

Meena sent us fruit.

(Babrakzai, 1999, p. 82)

The Spanish datives have been described in terms of their semantic content and morphosyntactic properties. In terms of the predicate structure, some datives are associated with verbs of transfer like dar ‘give’ and some with verbs of action like cocinar ‘cook’. In some of the verbs of the dar ‘give’ type, the nominal exclusively marks the dative argument. The following example shows the exclusive nominal.

8) Carolina Ø dio mucha importancia a la pelicula Carolina gave much importance Dat the movie

Carolina gave a lot of importance to the movie.

The third participant with the cocinar ‘cook’ may be a dative argument where the co-occurrence of the clitic is obligatory and the coreferential NP is optional; or a prepositionally marked adjunct where dative clitic is ungrammatical. The following examples show them dative argument and prepositionally marked adjunct respectively.

9) a) Carolina les dio a los chicos un libro de cuentos. Carolina DAT3PL gave.1s DAT the children a book of short stories Carolina gave the kids a book of short stories.

b) Carolina se lo dio a los chicos

Carolina DAT3ACC 3MS gave. 3s DAT the children Carolina gave it to the kids.

c) Carolina le dio a Esteban un libro de cuentos Carolina DAT3S gave. 3s DAT Esteban a book of short stories Carolina gave Esteban a book of short stories.

The verb like dar, unlike the concinar, allows clitic alone, NP alone and doubling but coninar does not allow NP alone. The proposition does not assign a case and when there is no NP, it is realized by a pro, which shows that dative doubling is always obligatory (Belloro, 2007). There is commonly no agreement about this sharp division into two types by all authors and some authors include some verbs in one group but others include them in the other group. For example, the verb hacer 'make' is included in the transfer group and in the creation type by others. The verb poner 'put', though a three-place verb and considered a type of transfer verb, is included in the creation type of verb (Demonte, 1994).

The occurrence of clitic alone is more than the doubling which in turn is more than the NP alone. The dative case role is associated with topical referents having a high degree of discourse saliency and individuation (Greenberg, 1974; Givon, 1984, 2001) and if there is a maximal deviation for the dative participant, then the NP is used exclusively. The datives typically refer to the singular definite referents and not to those with generic references and the indefinite in Spanish (Company, 2006, p. 503). But clitic exclusive and clitic doubling can be analyzed with the same verb classes where the clitic exclusive is more frequent than the clitic doubling showing the correlation between datives and active referents involved in the topical chains. Though the clitics may be doubled by the datives, yet clitic exclusive is the unmarked manifestation of the dative arguments in the history of Spanish (Company, 2006, p. 504).

The clitic occurs if it refers to the referent already referred to in the previous clause; but if the antecedent is not present or is farther away in the context of the discourse, then the clitic doubling construction occurs. Particularly, the doubling occurs when the pronominal fails to identify the intended referent by the hearer; if there are more participants, or a new participant is introduced in the context of the discourse. The doubling occurs when the referent is less accessible than for the normal expectation for dative arguments. The accessibility of the referents is encoded differently by the dative arguments (the dative argument can be focal NP or a VC itself) in Pashto showing a greater account of the activation level of referents in the language. If the datives are topical, active, and definite, then doubling may have a marked meaning. If the referents are neither accessible nor identifiable, then the NP-exclusive occurs (Belloro, 2007, p. 145).

The nominative (direct) case can occur with the dative in past, the verb agrees with the nominative in past (10a). But if the subject is ergative, for example, the verb shows default agreement (10b). Both the above dative arguments can be realized by VCs (10a, b).

10) a) Za sari ta

qaar shum

1.SG.NOM man.OBL DAT anger do.PST.PF.NOM

I rebuked the man.

b) Maa sari ta owe

1.SG.ERG man.OBL DAT tell.PST

I told the man.

11) a) Za WAR ta qaar shum

1.SG.NOM 3.VC DAT anger do.PST.PF

I rebuke the man.

b) Maa WAR ta owe

1.SG.ERG 3.VC DAT tell

I told him.

In the (11a) example, the verbal clitic is not because of the presence of a compound verb. In the second example (11b), we have verbal clitic with a simple verb.

The primary aim of the present study was to present a discourse-pragmatic account of the dative argument in Pashto. There are different encoding possibilities for dative argument. All these encoding possibilities of dative argument and the corresponding verbal clitic are discourse-pragmatic in nature. The current study is concerned with these discourse-pragmatic conditions responsible for different encoding possibilities of dative argument. The study aimed to investigate the different encoding possibilities of the dative argument and investigate the discourse-pragmatic conditions responsible for different encoding possibilities of the dative argument. The study also tried to find out the theoretical justification of different realizations of the dative argument.

The present study has focused on VCs in the Pashto language. The dialect under investigation was the Yusafzai dialect. The previous works in the language include Penzl (1955), Tegey (1975, 1977), Hallberg (1992), Muhammad (1993), Tegey and Robson (1996), Babrakzai (1999), Roberts (2000) and Rahman (2009). This study was different from all the previous studies in the sense that it did not rely on set examples in the language only. The current study was concerned with discourse-pragmatic conditions responsible for dative argument realization in Pashto. The data was collected from different varieties of sources. The data collected included both written and spoken forms. A recent novel in the language in the same dialect under investigation was selected to represent the written form of the language. The spoken form of the language was represented by taking data from the native speakers of the same dialect. The first type of data for the present study was taken from a Pashto novel Da Dashte Luth Musafar ‘The Traveller of the Desert of Luth’ written by Muhammad (1996). This novel was selected because it was a recent novel in the language under investigation. The novelist is a native speaker of the Pashto language (Yusafzai dialect). The novel was written in the same Yusafzai dialect of the Pashto language under investigation in the current study. The data was obtained from five different groups of Pashto speakers. Each group discussion of half an hour was recorded. The participants in every group discussed different topics of their interest. The total number of participants was twenty-two in number. Four out of the five groups contained five participants while one of the groups contained just two members. The recorded data was transcribed and calculated for dative argument occurrence. This analysis helped the researchers to know that the dative argument realized by clitic-alone was greater in natural discourse than in the text. The clitic doubling, on the other hand, had more or less the same frequency of occurrence in both text and naturally occurring discourse. The examples of clitic doubling construction for dative arguments and their possible variants were compared with theoretical examples proposed in previous studies. The comparison pointed out some discourse-pragmatic conditions responsible for the phenomenon. In addition, the study is a great help in understanding the language in the light of a structural-functional framework like RRG. The study is highly significant in the sense of showing the communicative perspectives in the light of the model used here. The study is highly significant in the sense of showing the communicative perspectives in light of the model used here.

The dative argument can be full lexical NP, pronominal or realized as VC. All these possibilities have pragmatic propositions. Full lexical NPs are used when the referents are newly introduced in the discourse (Rahman, Din & Iqbal, 2017). When there are more possible referents in the context of the discourse, the addressee's state of knowledge may not be clear about any one of them but may assume the reference to the last-mentioned referent. Speakers in such conditions use full lexical NP to avoid misinterpretation of the intended referent. This is quite frequent in texts than in natural discourse because the possible referents in natural discourse are always less than the possible referents in the texts. The following examples from the Pashto text show the full lexical NP appearing as a dative argument.

12) a) Mazigar khapalo astanano ta dars warkawi

Evening own assistant DAT lesson give

He teaches his assistants in the evening.

b) Maskhutan berta naroghantoon ta ze aw tar nime

Night back hospital DAT go and til mid

Shpe da naarogho muayena kawi

Night PREP patients diagnose do.PRS.IMP

At night he visits back the hospital and visits the patients till midnight. (Tegey & Robson, 1993)

The pronominals identifying dative arguments are used for those previously introduced in the context of the discourse. The pronominal is more frequent in natural discourse than in texts, where the referents are introduced before. The following examples from the Pashto text show the pronominal identifying dative.

13) a) Dwe ta daa pukhtana aw deeni fariza khkari

3.PL DAT this of pathans and religious duty seem

They consider it a cultural and religious duty.

b) Taa ta DAR wa rasedo?

2.SG DAT 2.VC PERF reach.PST.PF

Did he reach there?

(Tegey & Robson, 1993)

The VCs identifying dative arguments are used for those which are introduced before and understood in the context of the discourse. The referents here are so understood in the context of the discourse that the number features are not realized by VCs but only the person features are identified to know about the intended referents in discourse. The following examples show the VCs identifying dative arguments.

14) a) Charcharak WAR ta owayal che tol me

Cricket 3.VC DAT tell.PST.PF that whole 1.CL

sandare wayale songs sing.PST.IMP

The cricket told him that he was singing song the whole year.

b) megi WAR ta owayal che tol kal

ant 3.VC DAT tell.PST.PF that whole year

de sandare owyale was WARta atan kawa

2.CL songs sing.PST.IMP now 3.VC DAT dance. do

The ant replied him that the whole year you sang songs, now dance to it.

(Tegey & Robson, 1993).

VCs can occur along dative arguments i.e., the dative argument can appear in different forms following pragmatic conditions for their occurrence in different forms. The VC can occur alone or occurs along the dative argument causing dative doubling. These possibilities are pragmatically driven. Pashto is a pro-drop language and the overt subject or object may be realized by the agreement suffixes identifying the empty arguments. Like agreement suffixes, the VCs may license empty dative arguments in the clause. The cases where the overt NPs do not occur are the cases where the argument is realized by the phonological silent pro. If it is pro, the dative doubling is obligatory with such verbs either morphosyntactically realized or not. The following examples show the VCs identifying the empty dative arguments.

15) a) Maa khat WAR kro

1.SG.ERG letter.ABS 3.VC give.PST.PF.ABS

I gave the letter to him/her/them

b) Za khat DAR kom

1.SG.NOM letter.ACC 2.VC give.PRES.IMP.NOM

I am giving you the letter.

The VCs in the above example identify the empty dative arguments. Verbal clitics may also co-occur with overt dative arguments like agreement suffixes. The following examples show the co-occurrence of VCs with dative arguments. The dative argument here can be a full lexical NP either in the form of a proper name or pronoun as in the following examples.

16) a) Maa Ahmad la khat WAR kro

1.SG.ERG Ahmad DAT letter.ABS 3.VC give.PST.PF.ABS

I gave Ahmad the letter.

b) Za taa la khat DAR kom

1.SG.NOM 2.SG DAT letter.ACC 2.VC give.NOM

I am giving you the letter.

The verbal clitic identifying dative arguments may co-occur with another VC in the same clause. One of the VCs here functions as an agreement marker, while the second VC replacing the full lexical NP identifies the dative argument. The following examples show the co-occurrence of VCs with dative arguments in the form of VCs.

17) a) Maa WAR la khat WAR kro

1.SG.ERG 3.VC DAT letter.ABS 3.VC give.PST.PF.ABS

I gave him/her/them the letter.

b) Za DAR la khat DAR kom

I.SG.NOM 2.VC DAT letter.ACC 2.VC give.PRES.NOM

I am giving you the letter.

The VCs used instead of the full lexical NPs have an identical form to the VCs used adjacent to the verb showing agreement markers with dative arguments. However, these two forms differ with respect to their attachment to the host words. The VCs showing agreement markers attach to the verb host, while the VCs used instead of the proper arguments can attach to other lexical categories like postposition and adjectives etc.

In this study, it was shown the difference

between the so-called pronominal second position clitics (enclitics) and the VCs (proclitics). The pronominal second position clitics are assumed to be the second argument (Tegey, 1977) or the agreement markers of the same argument agreeing with the pro argument in the clause (Roberts, 2000). The VCs, on the other hand, are pronominal features agreeing with dative arguments or occur alone to realize the missing dative argument. The VC co-occurs with a dative argument causing clitic doubling in the language. The data from Pashto text and natural discourse showed that there are three encoding possibilities for a dative argument. The dative argument occurs alone as NP-exclusive without VC. The dative argument may also be realized by the VC as VC-only without an overt dative argument in the clause. The dative argument may occur with the VC causing clitic doubling in the language. The corpora of natural discourse and texts showed that the clitic alone was the unmarked way of encoding dative argument and clitic doubling appeared to be the marked realization when the referents deviate from the expected accessibility level. The co-occurrence of VC and proper dative argument was an instance of clitic doubling not of dislocation. Both the VC and proper dative argument occur in the same prosodic domain; there is no pause after the proper dative argument as we find in case of dislocation. Further, The VC that co-occurs with the proper dative argument agrees with this argument in terms of person features both of which are in the same prosodic domain.