The Conjunctive Participle in Dakkhini Hindi-Urdu: Making the Best of Both Worlds

Karumuri V. Subbarao
University of Delhi
&
Harbir K. Arora
Galgotia Institute of Technology

Dakkhini Hindi-Urdu (hereafter, Dakkhini) is a transplanted variety of Hindi-Urdu (Indo-Aryan) in southern India where Telugu (Dravidian) is spoken. Due to prolonged contact with Telugu for more than five centuries, several changes took place in the syntax of Dakkhini. In this paper we discuss contact-induced syntactic changes in conjunctive participles in Dakkhini. We shall show that (i) lexical NPs are permitted to occur as the subject of a conjunctive participle with volitional and nonvolitional predicates violating the PRO Theorem and that such occurrences prove that contact overrides syntactic constraints/principles, (ii) the phenomenon of Backward Control not found in Hindi-Urdu is added to the grammar of Dakkhini, (iii) the negative morpheme occurs in the embedded adverbial clauses with no semantic import just as in Telugu and unlike in Hindi-Urdu, (iv) the conjunctive participle occurs in the predicate position of a clause, unlike in Hindi-Urdu with stative and nonstative verbs, (v) Dakkhini retains the 'simpler' Hindi-Urdu negative form and is not affected by the various 'complex' forms of Telugu in spite of prolonged contact, (vi) Dakkhini does not permit a negative to occur conjunctive participle to avoid 'syntactic homonymy', (vii) analogy plays a role in syntactic change, and (viii) Dakkhini optimally utilizes a borrowed syntactic construction from Telugu with regard to the occurrence of affirmative and negative conjunctive participles in predicate position.

Conjunctive participles are typical of Indian languages where a verbal form devoid of the phi (person, number and gender) features links the main clause and the subordinate clause (cf. Masica, 1976, Subbarao 1996). The conjunctive participial bound morpheme in Hindi-Urdu is kar/ ke which is a nonfinite form and is added to the right of the verb stem. For example, sun kar/sun ke 'having heard' or khaa kar/khaa ke 'having eaten'. One of the primary functions that the conjunctive participle performs is to denote sequential actions. Hindi-Urdu also has a perfect participle that is formed by adding - yaa ( huaa ) to the verb stem. This can occur as a prenominal modifier, as a predicate adjective and as an adverb (cf. Subbarao, 1984). Corresponding to the perfect and conjunctive participial forms of Hindi-Urdu, there is only one form in Telugu with - i (or - in when followed by a morpheme starting with - a) following the verb stem which is normally termed as conjunctive participle as well as past participle in Telugu grammars (cf. Arden, 1873, Krishnamurti & Gwynn 1985, Subbarao 2004). For example, win-i 'having heard' tin-i 'having eaten' are instances of conjunctive participles and win-in-a 'the one who heard' or 'the one that is heard' and tin-in-a 'the one who ate' or 'the one that is eaten') are instances of perfect participles in Telugu. In Dakkhini the conjunctive participle is formed by the addition of ke/ko to the verb stem. For example, sun ke/sun ko 'having heard', khaa ke/khaa ko 'having eaten'. The perfect participial form in Dakkhini is V-ye so and this too performs some of the functions that the conjunctive participle in Dakkhini performs (Arora & Subbarao 1989). Conjunctive participles in all three languages under consideration are invariant for person, number and gender.

In Hindi-Urdu and Telugu the conjunctive participle functions as time, reason and manner adverb (cf. Abbi 1984, Davison 1981, Kachru 1981, Kellogg 1938, Subbarao 1996). It also imparts the meaning of instead of when preceded by the negative and even though when followed by the inclusive particle in Hindi-Urdu and Telugu. The Dakkhini conjunctive participle too performs all the functions that the conjunctive participle of Hindi-Urdu performs. However, in course of time not only some new functions that are not found in Hindi-Urdu have been added but also a new syntactic phenomenon 'backward control' (Polinsky & Potsdam 2002, Subbarao 2004) too has been included in the grammar of Dakkhini due to convergence with Telugu. The constraints that operate on the conjunctive participial construction in Dakkhini are different from those that operate in Hindi-Urdu. We shall show that these differences can largely be attributed to Telugu influence.

2. In this section we shall show how conjunctive participles in Dakkhini provide evidence in support of a principle which we shall label as the Principle of Optimal Utilization of an existing construction.

2.1 The subject of the nonfinite conjunctive participle clause in Hindi-Urdu is PRO and it is in general an unCase-marked, ungoverned empty element though there is counterevidence to this (Subbarao 2003). The subjects of the matrix clause and the embedded clause must be identical and this is generally referred to as the Identical Subject Constraint.

Hindi-Urdu

Telugu too obeys the Identical Subject Constraint.

Telugu

However, in Telugu the Identical Subject Constraint does not hold when the embedded sentence denotes a nonvolitional act and the embedded subject is [-animate] as in (3a).

In (3a) and (3b) the embedded conjunctive participle denotes a nonvolitional act. However, in (3b) Karuna, the embedded subject, is [+human] where as the embedded subject waanalu 'rains' in (3a) is [-animate]. Hence, the violation of the Identical Subject Constraint is permitted in (3a).

In contrast, Hindi-Urdu does not permit any violation of the Identical Subject Constraint with non-volitional predicates.

Hindi-Urdu

When the subject of the embedded conjunctive participle clause and the matrix clause are nonidentical, Hindi-Urdu permits only an oblique infinitival complement clause in place of the conjunctive participle clause. In such sentences with non-identical human subjects, the genitive occurs with the embedded subject and such occurrence of the genitive is optional when the embedded subject is [-animate] (Subbarao 1984). In (5a) the embedded subject is baariS 'rain' and it is [-animate] and hence, in (5b) the occurrence of the genitive is optional.

In (5a) baariS 'rain' gets inherent Case from the infinitive/gerund ho-naa 'to be'. The inherent case marker that is assigned is the invariant genitive ke 'of' which is in its oblique form. Thus the genitive ke assigns structural accusative Case to the embedded subject baariS 'rain' in (5a). In (5b) there is no Case assigner to the embedded subject. Hence, it appears that the infinitive assigns structural nominative Case to the embedded subject baariS 'rain'. Another way to look at the issue of Case to the embedded subject in (5b) is to relate derivations and say that the genitive in (5a) is optionally dropped in (5b) after assigning structural Case to the DP. However, in the Government and Binding framework and in the Minimalist Program of Chomsky (1995) such a solution is not welcome as each sentence is treated as unique and it cannot be related to another sentence.

The question that needs to be answered is: How and why does Telugu permit a lexical subject in the embedded subject position of the conjunctive participial clause in (3a)? The conjunctive participial marker - i is an allomorph of the past tense marker -in in Telugu and it acquired the status of a conjunctive participial marker due to syntactic reanalysis. Since the conjunctive participial marker is originally derived from the past tense marker, it retained its [+finite] tense feature. It is the [+finite] nature of the conjunctive participle that assigns structural nominative Case to the embedded subject w aanalu 'rain'. In contrast, the conjunctive participial marker kar/ke in Hindi-Urdu is devoid of the tense feature and hence, it cannot assign Nominative Case to the embedded subject in Hindi-Urdu in (4). Hence, the ungrammaticality of (4).

Just as in Telugu, Dakkhini too permits a lexical DP as the subject of the conjunctive participial clause though the conjunctive participle is [-tensed ] as in (6).

Dakkhini

The question that needs to be addressed is: Why is sentence (6) grammatical though the conjunctive participial marker is [-tensed] and hence, [-finite] in Dakkhini just as it is Hindi-Urdu and consequently, it cannot assign structural nominative Case to its Subject? There does not seem to be any structural explanation. The only plausible nonstructural explanation we could think of is that syntactic constraints are 'overridden' in language contact situations and the recipient language does not 'hesitate' to add to its grammar a phenomenon where a lexical subject can occur as the subject of the embedded clause though it is otherwise a violation of universal principles of Case assignment/Checking (Chomsky 1995).

2.2. In this section we shall show that Dakkhini has added a new phenomenon of Backward Control to its grammar and the addition of this involves not only having new syntactic structures but also a violation of the rules of the source language Hindi-Urdu.

Let us first look at the phenomenon of control in Telugu. Telugu like other South Asian languages permits nominative as well as nonnominative subjects as controllers of PRO. The nonnominative subjects include dative, locative and instrumental subjects (Subbarao & Bhaskararao 2004). When a t ime expression occurs as the predicate of a sentence in Telugu, the time expression requires a dative Case-marked subject.

Telugu

However in control structures involving the dative Case-marked DP as the matrix subject, the time expression cannot occur in the matrix subject position and it is only the subject of the embedded sentence that can occur overtly. Such cases are instances of Backward Control in contrast to the more widely spread phenomenon of Forward Control in almost all the languages of the world. In such cases we shall show that there are no corresponding Forward Control structures and there exist only Backward Control structures in Telugu. Hence, there is a 'syntactic pattern gap' in Telugu (Subbarao 2004).

When a time expression occurs in the matrix VP in Telugu, it requires a dative Case-marked subject as (8) below from Telugu shows. An important feature of this construction is that the DP denoting the time expression does not trigger agreement though it is in plural and is the grammatical subject of the dative subject construction. Hence, in (8) the verb ayyindi / paTTindi 'took' is in third person nonmasculine singular which is default agreement. The verb neither agrees with the dative subject nor with the grammatical subject padi roojulu 'ten days' (Subbarao & Bhaskararao in press for details).

Telugu

Telugu exhibits both Forward Control and Backward Control (Subbarao 2004 for details). However, when a conjunctive participle occurs in the embedded clause and a time expression in the matrix clause as in (9) and (10), only Backward Control is permitted and not Forward Control. " is used to denote the empty position of the matrix subject.

Telugu

Sentences corresponding to (9) and (10) with Forward Control in Telugu are ungrammatical as a dative Case-marked DP cannot occur in the matrix subject position.

In contrast, Hindi-Urdu does not permit Backward Control and hence, does not permit a nominative or dative Case-marked DP as the subject of the embedded clause in sentences (13) and (14).

Hindi-Urdu

Hindi-Urdu permits only a perfect participle occurring in such cases which are instances of Forward Control as in as in (15) and (16), the only permissible option in Hindi-Urdu. In fact, Dravidian speakers learning Hindi often commit an error by using the conjunctive participle in place of the perfect participle. We shall not go into pinpointing the reasons here.

Hindi-Urdu

In Dakkhini too when a time expression occurs as the predicate of the matrix sentence, a conjunctive participle occurs when the subjects of the embedded clause and matrix clause are differently Case-marked. That is, Dakkhini too permits only Backward Control which is an unHindi-like pattern in such constructions (Subbarao 2005 for further details). The following examples are illustrative.

Dakkhini

Thus, we observe that Dakkhini has incorporated a new phenomenon of Backward Control that involves not only having new syntactic structures but it also violates the rules of the source language, Hindi-Urdu.

Dakkhini however did not adopt the Telugu pattern in toto. Recall that in Telugu the matrix verb does not exhibit agreement with the time expression ((9) and (10)) and it manifests default agreement. Dakkhini retains the agreement pattern of Hindi-Urdu and hence, the matrix verb agrees with the time expression. Hence, Dakkhini goes along with Hindi-Urdu in terms of agreement with the non-nominative subject.

2.3. Dakkhini permits unidentical subjects when a time adverbial occurs in the matrix clause (19) and neither the donor language Telugu (20) nor the source language Hindi-Urdu (21) permits unidentical subjects in such cases.

Dakkhini

A sentence corresponding to (19) of Dakkhini is ungrammatical in Telugu (20) though there is a nonvolotional predicate in the conjunctive participial clause. As mentioned earlier, Telugu requires the embedded subject to be [-animate] and hence, (20) is ungrammatical.

A corresponding sentence in Hindi-Urdu (21) below too is ungrammatical due to two reasons:

(i) the matrix and embedded subjects are different, and

(ii) a lexical subject occurs as the subject of the conjunctive participial clause.

Hindi-Urdu

In a sentence corresponding to (19), Hindi-Urdu (22) employs a sentential infinitival clause and the clauses do not require the Subject Identity Constraint and hence, the matrix and embedded subjects need not necessarily be coreferential.

The infinitival clause in Hindi-Urdu modifies the noun phrase aaTh din 'eight days' and the entire noun phrase is followed by the adverbial marker ke baad 'after' in (22).

Hindi-Urdu

Telugu too does not permit a conjunctive participle in such cases and in place of a conjunctive participle a relative participle is used which is formed by adding -a, the adjectivalizer, to the conjunctive participle puTT - in 'having been born' in (23).

The conjunctive participial marker /- in / in (23) is an allomorph of the conjunctive participial marker /- i/. It occurs when followed by a vowel across morpheme boundary. Examples such as (19) from Dakkhini show how the recipient language exploits the new rule that is added to the grammar to the maximum possible extent by permitting different subjects in the matrix and embedded clauses going even a step ahead of the language from which it borrowed the rule. However, Telugu does have a construction of the type (24) given below in which a time expression can occur as the subject of the embedded clause and the matrix subject is unidentical with the embedded subject. The embedded sentence is in italics.

Telugu

Note that the conjunctive participial clause in Telugu in (24) involves a nonvolitional predicate and the subject is [-animate] and hence a lexical subject is permitted in the embedded clause. In Dakkhini, such sentences are ungrammatical.

Dakkhini

The ungrammaticality of (25) indicates that Dakkhini does not go all the way in following Telugu in the violation of the Subject Identity Constraint and there are a few exceptions as in sentence (25) above. Hindi-Urdu however does not permit unidentical subjects at all as (26) and (27) illustrate.

Hindi-Urdu

Such time adverbial constructions are expressed by an oblique infinitival complement construction in Hindi-Urdu as in (22) above. A grammatical sentence corresponding to (26) is expressed in Hinidi-Urdu by an oblique infinitival complement as in (22) above and the one corresponding to (27) is expressed by two independent clauses as in (28).

Hindi-Urdu

To summarize, Backward Control is a new phenomenon that is added to the grammar of Dakkhini. Dakkhini permits the occurrence of a conjunctive participle involving Backward Control when a time expression occurs as the predicate of the matrix clause. By permitting Backward Control Dakkhini violates the 'rules' of the source language Hindi-Urdu.

Dakkhini however does not permit a time expression to occur as the subject of the embedded clause with a conjunctive participle and the matrix and embedded subjects are different (25) though Telugu permits such sentences (24). However, the ungrammaticality of sentences such as (25) indicates that there are instances where Dakkhini does not follow the rules of Telugu in toto and such cases show how Dakkhini retains its unique and distinct identity. Thus Dakkhini retains the uniqueness of its own syntax and demonstrates that it is not a replica or a carbon copy of the donor language Telugu with heavily borrowed syntax. In the following section we shall discuss the cases where Dakkhini adopts Telugu morphology in the formation of concessive clauses, participles in predicate position, presumptive clauses, counterfactuals and conjunctive participle s of compound verbs.

3.1 Dakkhini uses a conjunctive participle followed by an inclusive particle bii 'also' in the formation of concessive clauses (29) and in such cases Telugu too has an identical construction (30). In contrast, Hindi-Urdu does not permit a conjunctive participle followed by an inclusive particle (31) and instead, it uses an imperfect participle (32). The following examples from Dakkhini, Telugu and Hindi-Urdu are illustrative. The concessive clause is in italics.

Dakkhini

Telugu

Hindi-Urdu

The ungrammaticality of (31) in Hindi-Urdu is due the fact that Hindi-Urdu does not permit a lexical DP to occur as the subject of the conjunctive participle as it cannot be nominative Case-marked whereas Telugu and Dakkhini do permit lexical subjects to occur in the embedded subject position with such predicates due to the phenomenon of Backward Control. Unlike in Telugu, the subject of the concessive clause in (29) sab kuch 'every thing' cannot get nominative Case marking as the conjunctive participle in Dakkhini is [-finite]. This is another instance which shows tha language contact overrides syntactic constraints. Thus, we observe that in the formation of concessive clauses Dakkhini adopts the morphology as well as the syntactic pattern of Telugu.

The conjunctive participle of a stative as well as non-stative verb can occur in the predicate position in Telugu (sentences (33) & (34)) whereas in Hindi-Urdu, a conjunctive participle is not permitted in predicate position ((sentence (35) ) and only a perfect participle is permitted to occur ((sentence (36)).

Telugu

Hindi-Urdu

Dakkhini too, just as in Telugu, permits both stative and nonstative verbs in predicate position.

Dakkhini

Telugu permits a stative verb in the negative conjunctive participle in predicate position ((39)), but not a nonstative verb ((40)).

Telugu (Telangana dialect)

Dakkhini on the other hand, has no such restriction. Both stative and non-stative verbs can occur as conjunctive participles in predicate position with the negative having the predicate in its scope.

Dakkhini

Sentences (41) and (42) illustrate how a particular construction of the donor language is optimally utilized by the recipient language by generalizing the borrowed construction to both stative a nonstative verbs in negative conjunctive participles.

3.3 We shall now discuss the formation of presumptive clauses in Telugu and Dakkhini. They both exhibit a one to one correspondence in morphological structure. The verb in the presumptive in Telugu is formed by the conjunctive participial form of the verb followed by verb be in future tense.

Telugu

As sentence (44) below illustrates, in Dakkhini too the presumptive is formed just as in Telugu. The conjunctive participial form in the presumptive is nonfinite.

Dakkhini

44.

In contrast, in Hindi-Urdu the presumptive cannot be formed with a conjunctive participle in the predicate position as (45) illustrates.

Hindi-Urdu

Instead, the presumptive in Hindi-Urdu consists of the verb in perfect aspect and verb ho 'be' in its future tense form.

The use of a conjunctive participle in the formation of the presumptive in Dakkhini on the same pattern of Telugu deserves a comment. It is generally agreed that morphology is 'conservative' and is neither borrowed nor is that vulnerable to change. Thus, Dakkhini has adopted the morphology of Telugu

(i) in the formation of the presumptive and,

(ii) with regard to the occurrence of negative and affirmative participles.

This shows that morphology is vulnerable to change in contact situations. We shall show below that the morphology of counterfactuals in Dakkhini too is identical to that of Telugu and is different from Hindi-Urdu, the source language.

3.4 In the formation of counterfactuals in Dakkhini and Telugu the verb stem is followed by a conjunctive participial marker. The counter-factual is underscored.

Verb stem + cpm + the verb be in conditional form

Dakkhini

In Telugu, the conjunctive participial form of the verb occurs in the embedded clause as well as in the matrix clause in counterfactuals.

Telugu

In contrast, in Hindi-Urdu, the main verb in perfect aspect and verb be in imperfect aspect constitute the counterfactual.

Hindi-Urdu

The use of the conjunctive participial form of the verb in counterfactuals in Hindi-Urdu results in ungrammaticality.

3.5 South Asian languages permit compound verbs where the main verb (V 1 ) is followed by another verb (V 2 ) which imparts aspectual meaning. The second verb V 2 loses its original meaning due to the process of grammaticalization and V 2 is called the vector verb. A conjunctive participle in Telugu permits a vector verb to occur with it while in Hindi-Urdu the conjunctive participle does not. Dakkhini follows the Telugu pattern. The verb vector wees 'drop' in Telugu has the meaning of 'drop' when it occurs as a main verb.