The paper discusses some possible interpretations of the notion ‘argument sharing’ in the analysis of Oriya serial verb constructions (SVCs). The paper in particular addresses the notion of ‘token-sharing’. We argue that the concept of token-sharing, although, due to its rigour, theoretically the most interesting construct, cannot be maintained for object sharing in Oriya VP-serialization. We therewith aim at exposing salient properties both of the construction type and the notion itself.
This paper develops a syntactic account of pseudoclefts that is consonant with current attempts to develop a strictly derivational account of syntactic relations. In what follows, I show that recent proposals in the realm of the higher functional field, remnant movement, and binding conspire to yield a simple view on (specificational) pseudoclefts.
This paper argues that in cleft constructions in Malayalam, the clefted constituent is (overtly) moved to the cleft focus, and that the ‘gap’ in the cleft clause is not created by movement of an empty operator as was proposed in Madhavan (1987) and some later studies. The cleft clause is shown to be non-finite; it has an Aspect Phrase dominating the vP, and no Tense node. This move solves the problem of moving a constituent out of the cleft clause without violating any constraint on movement. Non-copular clauses in general are argued to be in the domain of a Mood Phrase wherein resides finiteness. The paper also shows that the T node of the matrix clause of a cleft construction hosts a [+FOCUS] feature whereby it is able to attract an XP from within the clause. The idea in Jayaseelan (2000) that there is a Focus Phrase immediately above vP is integrated into the proposed analysis in such a way that the obligatory clefting of embedded interrogatives receives a principled explanation.
When question-words in Malayalam are subjected to long distance dependencies, direct extractions across finite clausal complements are restricted by complement/ non-complement asymmetry due to subjacency. However, clausal piedpiping rescues such long distance question-word dependencies. Long distance dependencies of question-words across infinitive complements, on the other hand, are possible adopting both strategies and are not affected by the complement/non-complement asymmetry. This state of affairs obtains largely because finite clausal complements are held on the left periphery of the matrix clause, which is an adjunct position, while infinitival clausal complements are held in situ. Therefore, direct extraction from the former is sensitive to subjacency, while from the latter, it is not. The present article looks at this phenomenon from the minimalist point of view in terms of the Minimal Link Condition (MLC).
Many and few are adjectival modifiers of an unpronounced noun NUMBER and the same holds for much and little as modifiers of AMOUNT. This reflects a broader property of UG, namely that UG excludes the possibility that a single adjective could simultaneously express what is expressed by large/small and what is expressed by number. This property in turn can be seen as a special case of UG imposing a maximum of one interpretable syntactic feature per lexical item (Principle of Decompositionality). The analysis is extended to a few andquite a few, to numerous and to questions of scope and polarity.
Jayaseelan (1992) aims to account for the fact that ‘inverse copula sentences’ such as My best friend is John may exclude coreference between a Name and a pronoun in the absence of the c-command configuration which triggers the application of Principle B of the Binding Theory. For example, the inverse copula sentence His teacher is John’s brother disallows coreference between his andJohn, while the corresponding predicative sentence, in which the structural relation between pronoun and name is identical, allows it: His brother is John’s teacher. Jayaseelan proposes an LF movement which creates a configuration subject to Principle B and which applies in a natural way to the inverse sentence but not to the corresponding canonical predicative sentence. I adopt Jay’s idea that LF movement provides the key to an account of coreference violations ‘beyond Principle B’ in the inverse copula sentence. I propose that the Focus constituent of the inverse sentence is preposed and subsequently 'deconstructed' into a Focus of Focus (F/F) and a presupposition of Focus (p/F). The rules which determine the interpretation of NPs at the syntaxsemantics interface then suffice to account for coreference violations both in the inverse copula sentence and in other sentence types.
In this article I address the question as to why natural language has special principles governing the binding of pronouns and anaphors and the licensing of reflexive predicates. That is, why do we have principles such as conditions A, and B of the binding theory? Based on facts from Malayalam discussed by Jayaseelan (1997), I show that the role of licensing reflexivity and the obligation to be (locally) bound should be fundamentally distinguished. A crucial property of complex anaphors is that they allow the language system to express reflexivity, and yet prevent an arity violation from arising.
In Maithili there exists a rule which is analogous to the Subject-to-Subject Raising rule in English. This rule is interesting because it raises the subject of a tensed embedded clause, whereas the Subject-to-Subject Raising rule raises to the matrix subject position the subject of a tenseless embedded clause. The DPtrace in the former case apparently violates one of the principles of Universal Grammar, namely, the binding principle for anaphors, proposed in Chomsky (1981,1995), since it is not bound in its governing category. We demonstrate, however, that the DP-trace does obey the binding principle. What is required here is to investigate this phenomenon from the perspective of the parametric variation that distinguishes Maithili-type languages from English-type languages. This article consists of four main Sections. In Section 1, we analyze the raising construction in Maithili. We also look into the rule of i-insertion, which optionally applies to fill the empty subject position of the matrix IP, in case Subject-to-Subject Raising does not apply. In Section 2, we present evidence in support of the rule of Subject-to-Subject Raising in Maithili. Section 3 shows how the stipulation of this rule complies with the theta-criterion, proposed in Chomsky (1981). Section 4 is addressed to the detailed analysis of the binding relation that holds between a DP-trace and its antecedent. In this part, we argue that INFL in Maithili, unlike in English-type languages, is generated inside V and does not govern the subject DP. The subject DP is rather governed and Case-marked by the preceding COMP, though it does not happen in raising constructions, since we assume that a raising verb may optionally trigger CPdeletion. We also show that in Maithili-type languages the notion of ‘governing category’ must dispense with “accessible SUBJECT”, since AGR in these languages cannot be treated as accessible SUBJECT. Finally, we show how binding principle (A) holds for the antecedent/DP-trace relation in Maithili.
Complement and adjunct clauses in Bangla often exhibit internal complementizers. Some of these, the present paper suggests, are particles base-generated in situ that covertly move their features to C at LF. This suggestion forms the core of a reasonably complete account of the placement of various clause particles relative to finite verbs in clause structure.
The structure of finite CP shows some unexpected syntactic variation in the marking of finite subordinate clauses in the Indic languages, which otherwise are strongly head-final. Languages with relative pronouns also have initial complementizers and conjunctions. Languages with final yes/no question markers allow final complementizers, either demonstratives or quotative participles. These properties define three classes, one with only final CP heads (Sinhala), one with only initial CP heads (Hindi, Punjabi, Kashmiri) and others with both possibilities. The lexical differences of final vs initial CP heads argue for expanding the CP projection into a number of specialized projections, whose heads are all final (Sinhala), all initial, or mixed. These projections explain the systematic variation in finite CPs in the Indic languages, with the exception of some additional restrictions and anomalies in the Eastern group.
This paper examines the particle ne in direct yes-no questions in Assamese. The particle ne occurs in the clause final position. When overt it is optionally followed by a negative element nai ‘is not’ or the k-word ki ‘what’. Constituent wh-questions have an abstract Q morpheme in the C position. Our study shows that the yes-no questions too have an abstract Q morpheme in the C position. The particle ne is not a question particle but a [+wh] disjunctive marker. The negative nai and the k-wordki are not constituents of the Disjunctive Phrase. They occur outside the matrix clause and are constituents of the second clause, which is conjoined to the matrix clause by the disjunctive particle ne.
The sequence of a nasal-plus-voiceless-obstruent (plosive) is a less favoured cluster cross-linguistically. Languages adopt various strategies to avoid such a cluster. These include nasal substitution, nasal deletion, post-nasal voicing, denasalization etc. The Dravidian language Malayalam exhibits post-nasal voicing to avoid NC̥. Interestingly, it also allows nasal gemination for the same purpose; and both the processes share identical domains: morpheme internal as well as across morphemes. Presumably this is a pointer towards some subtler facts at work inviting serious critical engagement. In course of the present study, it has been found that: (a) nasal gemination is a marked strategy to avoid NC̥ and is restricted to a limited set of morphemes; and (b) post-nasal voicing is a more productive and hence unmarked strategy in this language. A principled account, in terms of constraint interaction, of these discrete processes and their underlying functional unity emerges if the theory of underspecification is roped in while positing the input form. Additionally, some predictions follow automatically about the direction of language change.
In this article we examine the requirement that the minimal word be a disyllabic trochee being satisfied partially or fully in different ways in three unrelated languages spoken in India, namely, Punjabi, Bangla and Tamil. In Bangla, which does not have a phonemic vowel length distinction and where closed syllables count as bimoraic and heavy attracting prominence, monosyllables whether closed or open invariably have a long vowel. We argue that the monosyllabic lengthening is due to a catalectic syllable making these minimal words ‘virtual’ disyllables. Punjabi has a three-way distinction in syllable weight, a monomoraic light syllable, a bimoraic heavy syllable and a trimoraic superheavy syllable. Following the assumption in the literature, we assume that trimoraic syllables are virtual disyllables. We find that, by and large, monosyllables are trimoraic with a geminate final consonant augmenting the monosyllable to a superheavy status satisfying the disyllabic requirement. Finally, we examine an intriguing pattern of optional, final epenthesis in monosyllables in Tamil, once again arguing for a disyllabic word minimum. Between the three languages, we observe that the disyllabic minimality requirement is met in different ways, exhausting the three logical possibilities of stem augmentation, namely, vowel lengthening (in Bangla), consonant gemination (in Punjabi) and both consonant gemination and vowel epenthesis (in Tamil).
This article deals with phonological awareness which is explicit or conscious knowledge of the phonological structure that speakers of a language exhibit. It seeks to establish whether biliterate speakers of Telugu and English who are exposed to an alphabetic writing system and a semi-syllabic writing system are able to manipulate words in terms of syllables and phonemes. This experimental study involves 30 native speakers of Telugu who also know English. The results show that the two languages in question are not treated alike by the speakers, although there is some evidence of transference from one language to another. It is apparent in many cases that rather than the sounds, the item that is processed mentally is a visual representation of the word. Moreover, each language seems to be processed in its own script although there is also evidence to show that the native language is more dominant and the second language is processed in the script of the native language.
The paper discusses some possible interpretations of the notion ‘argument sharing’ in the analysis of Oriya serial verb constructions (SVCs). The paper in particular addresses the notion of ‘token-sharing’. We argue that the concept of token-sharing, although, due to its rigour, theoretically the most interesting construct, cannot be maintained for object sharing in Oriya VP-serialization. We therewith aim at exposing salient properties both of the construction type and the notion itself.
This paper develops a syntactic account of pseudoclefts that is consonant with current attempts to develop a strictly derivational account of syntactic relations. In what follows, I show that recent proposals in the realm of the higher functional field, remnant movement, and binding conspire to yield a simple view on (specificational) pseudoclefts.
This paper argues that in cleft constructions in Malayalam, the clefted constituent is (overtly) moved to the cleft focus, and that the ‘gap’ in the cleft clause is not created by movement of an empty operator as was proposed in Madhavan (1987) and some later studies. The cleft clause is shown to be non-finite; it has an Aspect Phrase dominating the vP, and no Tense node. This move solves the problem of moving a constituent out of the cleft clause without violating any constraint on movement. Non-copular clauses in general are argued to be in the domain of a Mood Phrase wherein resides finiteness. The paper also shows that the T node of the matrix clause of a cleft construction hosts a [+FOCUS] feature whereby it is able to attract an XP from within the clause. The idea in Jayaseelan (2000) that there is a Focus Phrase immediately above vP is integrated into the proposed analysis in such a way that the obligatory clefting of embedded interrogatives receives a principled explanation.
When question-words in Malayalam are subjected to long distance dependencies, direct extractions across finite clausal complements are restricted by complement/ non-complement asymmetry due to subjacency. However, clausal piedpiping rescues such long distance question-word dependencies. Long distance dependencies of question-words across infinitive complements, on the other hand, are possible adopting both strategies and are not affected by the complement/non-complement asymmetry. This state of affairs obtains largely because finite clausal complements are held on the left periphery of the matrix clause, which is an adjunct position, while infinitival clausal complements are held in situ. Therefore, direct extraction from the former is sensitive to subjacency, while from the latter, it is not. The present article looks at this phenomenon from the minimalist point of view in terms of the Minimal Link Condition (MLC).
Many and few are adjectival modifiers of an unpronounced noun NUMBER and the same holds for much and little as modifiers of AMOUNT. This reflects a broader property of UG, namely that UG excludes the possibility that a single adjective could simultaneously express what is expressed by large/small and what is expressed by number. This property in turn can be seen as a special case of UG imposing a maximum of one interpretable syntactic feature per lexical item (Principle of Decompositionality). The analysis is extended to a few andquite a few, to numerous and to questions of scope and polarity.
Jayaseelan (1992) aims to account for the fact that ‘inverse copula sentences’ such as My best friend is John may exclude coreference between a Name and a pronoun in the absence of the c-command configuration which triggers the application of Principle B of the Binding Theory. For example, the inverse copula sentence His teacher is John’s brother disallows coreference between his andJohn, while the corresponding predicative sentence, in which the structural relation between pronoun and name is identical, allows it: His brother is John’s teacher. Jayaseelan proposes an LF movement which creates a configuration subject to Principle B and which applies in a natural way to the inverse sentence but not to the corresponding canonical predicative sentence. I adopt Jay’s idea that LF movement provides the key to an account of coreference violations ‘beyond Principle B’ in the inverse copula sentence. I propose that the Focus constituent of the inverse sentence is preposed and subsequently 'deconstructed' into a Focus of Focus (F/F) and a presupposition of Focus (p/F). The rules which determine the interpretation of NPs at the syntaxsemantics interface then suffice to account for coreference violations both in the inverse copula sentence and in other sentence types.
In this article I address the question as to why natural language has special principles governing the binding of pronouns and anaphors and the licensing of reflexive predicates. That is, why do we have principles such as conditions A, and B of the binding theory? Based on facts from Malayalam discussed by Jayaseelan (1997), I show that the role of licensing reflexivity and the obligation to be (locally) bound should be fundamentally distinguished. A crucial property of complex anaphors is that they allow the language system to express reflexivity, and yet prevent an arity violation from arising.
In Maithili there exists a rule which is analogous to the Subject-to-Subject Raising rule in English. This rule is interesting because it raises the subject of a tensed embedded clause, whereas the Subject-to-Subject Raising rule raises to the matrix subject position the subject of a tenseless embedded clause. The DPtrace in the former case apparently violates one of the principles of Universal Grammar, namely, the binding principle for anaphors, proposed in Chomsky (1981,1995), since it is not bound in its governing category. We demonstrate, however, that the DP-trace does obey the binding principle. What is required here is to investigate this phenomenon from the perspective of the parametric variation that distinguishes Maithili-type languages from English-type languages. This article consists of four main Sections. In Section 1, we analyze the raising construction in Maithili. We also look into the rule of i-insertion, which optionally applies to fill the empty subject position of the matrix IP, in case Subject-to-Subject Raising does not apply. In Section 2, we present evidence in support of the rule of Subject-to-Subject Raising in Maithili. Section 3 shows how the stipulation of this rule complies with the theta-criterion, proposed in Chomsky (1981). Section 4 is addressed to the detailed analysis of the binding relation that holds between a DP-trace and its antecedent. In this part, we argue that INFL in Maithili, unlike in English-type languages, is generated inside V and does not govern the subject DP. The subject DP is rather governed and Case-marked by the preceding COMP, though it does not happen in raising constructions, since we assume that a raising verb may optionally trigger CPdeletion. We also show that in Maithili-type languages the notion of ‘governing category’ must dispense with “accessible SUBJECT”, since AGR in these languages cannot be treated as accessible SUBJECT. Finally, we show how binding principle (A) holds for the antecedent/DP-trace relation in Maithili.
Complement and adjunct clauses in Bangla often exhibit internal complementizers. Some of these, the present paper suggests, are particles base-generated in situ that covertly move their features to C at LF. This suggestion forms the core of a reasonably complete account of the placement of various clause particles relative to finite verbs in clause structure.
The structure of finite CP shows some unexpected syntactic variation in the marking of finite subordinate clauses in the Indic languages, which otherwise are strongly head-final. Languages with relative pronouns also have initial complementizers and conjunctions. Languages with final yes/no question markers allow final complementizers, either demonstratives or quotative participles. These properties define three classes, one with only final CP heads (Sinhala), one with only initial CP heads (Hindi, Punjabi, Kashmiri) and others with both possibilities. The lexical differences of final vs initial CP heads argue for expanding the CP projection into a number of specialized projections, whose heads are all final (Sinhala), all initial, or mixed. These projections explain the systematic variation in finite CPs in the Indic languages, with the exception of some additional restrictions and anomalies in the Eastern group.
This paper examines the particle ne in direct yes-no questions in Assamese. The particle ne occurs in the clause final position. When overt it is optionally followed by a negative element nai ‘is not’ or the k-word ki ‘what’. Constituent wh-questions have an abstract Q morpheme in the C position. Our study shows that the yes-no questions too have an abstract Q morpheme in the C position. The particle ne is not a question particle but a [+wh] disjunctive marker. The negative nai and the k-wordki are not constituents of the Disjunctive Phrase. They occur outside the matrix clause and are constituents of the second clause, which is conjoined to the matrix clause by the disjunctive particle ne.
The sequence of a nasal-plus-voiceless-obstruent (plosive) is a less favoured cluster cross-linguistically. Languages adopt various strategies to avoid such a cluster. These include nasal substitution, nasal deletion, post-nasal voicing, denasalization etc. The Dravidian language Malayalam exhibits post-nasal voicing to avoid NC̥. Interestingly, it also allows nasal gemination for the same purpose; and both the processes share identical domains: morpheme internal as well as across morphemes. Presumably this is a pointer towards some subtler facts at work inviting serious critical engagement. In course of the present study, it has been found that: (a) nasal gemination is a marked strategy to avoid NC̥ and is restricted to a limited set of morphemes; and (b) post-nasal voicing is a more productive and hence unmarked strategy in this language. A principled account, in terms of constraint interaction, of these discrete processes and their underlying functional unity emerges if the theory of underspecification is roped in while positing the input form. Additionally, some predictions follow automatically about the direction of language change.
In this article we examine the requirement that the minimal word be a disyllabic trochee being satisfied partially or fully in different ways in three unrelated languages spoken in India, namely, Punjabi, Bangla and Tamil. In Bangla, which does not have a phonemic vowel length distinction and where closed syllables count as bimoraic and heavy attracting prominence, monosyllables whether closed or open invariably have a long vowel. We argue that the monosyllabic lengthening is due to a catalectic syllable making these minimal words ‘virtual’ disyllables. Punjabi has a three-way distinction in syllable weight, a monomoraic light syllable, a bimoraic heavy syllable and a trimoraic superheavy syllable. Following the assumption in the literature, we assume that trimoraic syllables are virtual disyllables. We find that, by and large, monosyllables are trimoraic with a geminate final consonant augmenting the monosyllable to a superheavy status satisfying the disyllabic requirement. Finally, we examine an intriguing pattern of optional, final epenthesis in monosyllables in Tamil, once again arguing for a disyllabic word minimum. Between the three languages, we observe that the disyllabic minimality requirement is met in different ways, exhausting the three logical possibilities of stem augmentation, namely, vowel lengthening (in Bangla), consonant gemination (in Punjabi) and both consonant gemination and vowel epenthesis (in Tamil).
This article deals with phonological awareness which is explicit or conscious knowledge of the phonological structure that speakers of a language exhibit. It seeks to establish whether biliterate speakers of Telugu and English who are exposed to an alphabetic writing system and a semi-syllabic writing system are able to manipulate words in terms of syllables and phonemes. This experimental study involves 30 native speakers of Telugu who also know English. The results show that the two languages in question are not treated alike by the speakers, although there is some evidence of transference from one language to another. It is apparent in many cases that rather than the sounds, the item that is processed mentally is a visual representation of the word. Moreover, each language seems to be processed in its own script although there is also evidence to show that the native language is more dominant and the second language is processed in the script of the native language.