Search

This article explores the phenomenon of pied-piping within the context of two assumptions. The first assumption is that a realistic natural language parser must assign incoming words incrementally into the syntactic representation as it is built during the parsing process. The second assumption is that this process makes use of Merge, the core recursive operation in syntax. These assumptions lead us to consider an extension of the standard theory of Merge, in which phrase structure is derived from a linear string of words in a left-to-right/top-down fashion (e.g. Phillips, 1996, Phillips, 2003). We propose a generalization of the Phillips architecture that describes and correctly parses simple and complex pied-piping constructions in three languages: English, Italian and Finnish. We then show that the proposed hypothesis derives (without any additional assumptions) several well-known constraints to pied-piping and operator movement, such as Condition on Extraction Domains (CED), the Left Branch Condition, the extended projection principle (EPP), the complex properties of picture nouns and certain differences between relative and interrogative clauses. A computational implementation and full formalization of the hypothesis together with a test corpus is used to demonstrate the viability of the approach. We conclude that the top-down approach provides a useful tool for grammatical analysis.

In this paper I argue that the syntax of Mi’gmaq (Eastern Algonquian) is underlyingly configurational despite displaying surface characteristics of non-configurationality. Three subject–object asymmetries are presented that cannot be accounted for in a principled manner unless: (1) DPs are base-generated in Argument (A) positions, and (2) the subject A-position asymmetrically c-commands the object A-position. First, I show that wh-phrases are strictly ordered in multiple wh-questions which is accounted for as a Superiority effect. Next, I present disjoint reference effects within a clause which shows that Binding Condition C effects are present. Finally, I present a restricted pattern of Long-Distance Agreement with embedded declaratives that is limited to the structurally highest embedded argument. The presence of a Restricted LDA pattern is proposed as a diagnostic for the relative structural height of DPs, which has implications for Algonquian languages with a similar pattern. I conclude that Mi’gmaq is configurational. I propose that an investigation of discourse factors is the most promising possibility to explain the non-configurational appearance of Mi’gmaq.

In this article, three types of data from syntax, semantics, and morpho-phonology are presented to argue that Ionin and Matushansky's (IM, 2006) non-constituency analysis of complex numerals may not work in Mandarin Chinese and that the traditional constituency analysis is more plausible. In order to address these data under the constituency analysis, we propose and justify a fine-grained phrase structure for complex numerals in Mandarin Chinese, which can accommodate all these data satisfactorily. We also investigate the cross-linguistic implications of our argumentation by showing that similar morpho-phonological data in many minority languages of China also support the constituency analysis for complex numerals.

This article proposes that Holmberg's Generalization, the well-known constraint on Scandinavian object shift, can and should be extended as a general constraint on scrambling past c-commanding heads, and applies to the scrambling phenomena of German, Yiddish, Japanese, Korean, and other languages. In addition to evidence from the synchronic typology of scrambling and object shift, the existence of such a constraint can be shown most conclusively by careful observation of languages undergoing phrase structure changes over time. Specifically, we use the change in the position of Tense over the history of Yiddish as an experimental domain for testing the hypothesized constraint, and show that the change in Tense restricts the scrambling options in Yiddish precisely in the predicted manner. Once we have shown how scrambling is constrained empirically, we argue that the constraint stems from a condition on the interface between the narrow syntax and LF, the “Conservation of C-Command”, which ensures that the semantic scope of heads can still be calculated after various movement operations have applied in the syntax. Finally, we suggest that the empirical and theoretical results of the study provide indirect support for an antisymmetric view of phrase structure.

With the crystallization of the “generative enterprise” half a century ago, two concepts became salient: the initial state and final states of the language faculty, respectively, UG (the genetic component) and I-languages. Since then inquiry has gained far greater scope and depth. It has also led to sharpening of fundamental principles of language. At first, descriptive adequacy appeared to require rich and complex assumptions about UG. A primary goal has always been to overcome this deficiency. Core properties of concern have included compositionality, order, projection (labeling), and displacement. Early work assigned the first three to phrase structure rules and the last to the transformational component. Simplification of computational procedures suggests that compositionality and displacement (along with the “copy theory”) fall together while order may be a reflex of sensorimotor externalization, conclusions that have far-reaching consequences. As for labeling, minimal computation restricts options to the few that have considerable empirical support.

This paper examines register-based language internal variation, focussing on subject omission in English diaries. This register-specific pattern might be seen as some kind of ‘extragrammatical’ culturally-determined stylistic convention associated with this particular register, but a survey of the relevant data shows that the omission of the subject in diary styles is subject to the core syntactic constraints that have been identified in formal syntax. Importantly, the observed restrictions on subject omission do not follow from a purely functional account according to which recoverable subjects can be omitted: while recoverability certainly plays a role, there are precise constraints on the syntactic positions in which recoverable subjects can be omitted. The empirical generalisation that emerges is that subjects can be omitted in root clauses. Moreover, apart from fronted adjuncts no other constituent can precede the non-overt subject. The generalisation applies both to English and to French. The paper develops an account for subject omission which, in addition to standard assumptions about phrase structure, makes use of (i) the Phase based theory of truncation, (ii) the hypothesis of the articulated subject field. It is shown that other instantiations of subject omission such as that found, for instance, in note style journalese or in Samuel Beckett's poem Rockaby ( Bianchi, 2007 ), are governed by the same principles, suggesting that the pattern is subject to grammatical constraints which are not exclusively tied to the specific register. That the type of subject omission identified here should be analysed in terms of core grammatical principles is confirmed by the fact that subject ellipsis in second conjuncts, a phenomenon which is independent of register variation, is subject to the same restrictions as subject omission in the diary style and can be accounted for using the hypotheses developed here. The conclusion I draw from this discussion is that the grammatical patterns displayed by what might seem a culturally determined linguistic system are fully amenable to core principles and parameters of universal grammar.

In this paper we offer a syntactic description of Spanish exceptive constructions headed by excepto, salvo or menos (‘except’). Framing our hypothesis in an adjunction analysis of coordination, we argue that these exceptive markers head a Boolean Phrase, like other coordinating conjunctions. Two types of exceptive phrases can be identified, depending on the level of the constituents conjoined. In connected exceptives two DPs are conjoined. In free exceptives two CPs are conjoined; the exceptive markers select for a full-fledged CP as complement, whose null head (C) triggers a process of ellipsis in which all the syntactic material inside TP is marked for PF-deletion, except the remnant constituent(s). Our proposal supports a structural approach to ellipsis whereby elliptical constituents are in fact fully fledged though non-pronounced syntactic structures. It also supports the hypothesis that the differences in the syntactic behaviour of coordinate sentences and subordinate adverbial clauses cannot be derived from their phrase structure geometry but are instead due to the properties of individual conjunctions.

This paper explores a set of surprising contrasts between two major classes of adjectives in Tagalog. It is concerned, in particular, with the discovery that adjectival passives in Tagalog cannot occur as the main predicate of various types of impersonal clauses—i.e., clause types in which the adjectives sole argument cannot be promoted to subject. I argue that this fact follows ultimately from the claim that adjectival passives are unaccusative. This claim is compared with the observation that, in many other languages (e.g., Hebrew), adjectival passives apparently pattern as unergative. I explore the hypothesis that differences relating to the argument structure of adjectival passives can be related to larger architectural differences among languages, relating specifically to the structure of VP.

Salishan lexical suffixes resist straightforward classification since their meaning makes them look like nouns, but their form makes them look like a grammatical or derivational category (i.e., they are bound forms). It is argued that neither meaning-based nor form-based criteria can be utilized to determine categorial identity. Instead, the distribution of lexical suffixes indicates that these forms do not belong to the class of nouns. It is argued that lexical suffixes are √roots, and as such they are devoid of any functional or categorial information. At the same time, it is also argued that this does not necessarily set lexical suffixes apart from noun incorporation, as there are at least some cases of apparent noun incorporation that are better analyzed as √root incorporation.

This paper presents a research program for normalization-driven syntax. It takes the Minimalist research agenda as a starting point (Chomsky, 1995b) and explores the question of how the CI interface determines syntactic operations. The proposal provides specific content to the notion of bare output conditions and the nature of the CI interface. It does so by drawing on the tools provided by Relational Theory, a branch of set-theoretic mathematics, and Database Theory, a branch of computer science. It is demonstrated that core components of Narrow Syntax (phrase structure, selection and AGREE) are all definable in terms of Relational Theory. Then, it is shown that the process of relation optimization, or normalization, can derive chain formation. The article concludes with two speculations on the implementation of phases within a normalization-driven grammar and the implications of such a system for the learnability of the lexicon.

Tungus proprietives, regularly derived from nouns by affixation, demonstrate a mixed behaviour. As adnominal modifiers, they have the full syntactic and morphological distribution of an adjective. Yet, the base noun retains some nominal properties: it can head its own syntactic phrase, control agreement on its modifier, trigger various anaphoric processes, pluralize, and take derivational affixes. These forms present a problem for the traditional view of syntactic categories, as they have the morphosyntactic properties of more than one part of speech. They are also important for the syntax/morphology interface, because the word structure and the phrase structure do not match: although the base noun selects modifiers as if it were an independent phrasal head, it is not an independent word in morphology. The paper shows that syntactic affixation analyses cannot adequately capture the Tungus data. Instead, I propose a constructional analysis based on the idea that relationships among classes of words are expressed in the lexicon by means of cross-cutting hierarchical types, where more specific types inherit information from more general types. The idiosyncratic behaviour of proprietives is lexically specified by the cross-classification of head values. Categorial information is inherited from adjectives and determines the external distribution of the phrase. At the same time, proprietives inherit the underlying noun's semantics (referentiality) and selectional properties. This analysis raises further questions about the nature of the traditional syntactic labels such as nouns or adjectives, namely, whether we are dealing with a new category type in each case of mixed categories.

In this paper, we propose a lexicalist account of constructions which initially appear to be nominalisations of phrasal structures. We argue that lexical nominalisation of verbs preserves their form-selectional properties, creating nouns which select for dependents in exactly the way that the source verbs do. These properties hold in the ‘sub-phrasal’ syntax – essentially a domain of complex predicate formation – but do not hold in the regular phrasal syntax, where the form-selectional properties of verbs and nouns are disjoint. The overall view of grammar is one in which all morphological derivation and inflection is in the lexicon, interacting with a syntax in which there is a crucial distinction between words – items from the lexicon which do not project phrases, but create sub-phrasal constituents – and phrases, which are standardly projected from words.

This paper re-examines the doubling of pronominal clitics in Greek. It is argued that clitics are not affixes but full words which move in the syntactic component and (ultimately) target the head of T. As for their position in the phrase marker, it is claimed, following Kayne, R., 1975. French Syntax: The Transformational Cycle. MIT Press, Cambridge, MA and (Philippaki-)Warburton [J Ling 13 (1977) 259], that clitics do not head a functional (clitic) projection [in the sense of Sportiche, D., 1992/1996. Clitic Constructions. Ms, UCLA/In: Rooryck, J., Zaring, L. (Eds.), Phrase Structure and the Lexicon. Kluwer Academic, Dordrecht, pp. 213–276, for instance], but are merged in the internal argument position(s) of V. Being both X 0 and X max (Chomsky, N. 1995. Bare phrase structure. In: Webelhuth, G. (Ed.), Government and Binding Theory and the Minimalist Program. Blackwell, Oxford, pp. 383–439), clitics can undergo movement avoiding the Head Movement Constraint. Two alternative solutions of how this movement proceeds are examined and they are both shown to be consistent with our proposal. Furthermore, it is proposed that the relation between the clitic and its DP-double is that of coindexation, with the double occupying an adjunct position, either a remote one (clitic left/right dislocation), or a v P-internal one (doubling without comma intonation). In fact, we attempt to unify the Clitic Left/Right Dislocation and clitic doubling in the sense that both of them involve an argumental clitic. Finally, we show that despite the fact that Greek has clitics in argumental positions, it does not display a number of properties characteristic of polysynthetic languages.

This paper aims to illustrate how we might proceed to attain repeatability and falsifiability in generative grammar, on the basis of discussion of concrete empirical materials in Japanese, such as bound variable anaphora, quantifier scope, and reconstruction effects. After critically examining and improving on the major empirical arguments for the standard view of the phrase structure of Japanese, I propose that a certain type of bound variable anaphora is based on a formal relation, Formal Dependency (FD), between two argument positions. The structural condition on FD and its interpretive consequence proposed here not only enable us to account for the fluctuation and instability of speakers' judgments but also lead us to predict a number of hitherto unnoticed correlations with regard to the (un)availability of a bound variable construal. The paper attains a significantly higher level of repeatability than before, states explicitly how its proposal can be falsified, and stresses the critical importance of making a rigorous attempt to articulate how every theoretical concept is related to the rest of the theory as well as to the native speaker's linguistic intuitions so as to be able to make definite predictions and ensure the falsifiability of our hypotheses.

One of the major challenges of language acquisition is the fact that the auditory signal received by a child underdetermines the structural description of the utterance. This paper approaches the problem by capitalizing on the optimizing structure of Optimality Theory. The learner uses a hypothesized grammar to make a best guess at the full structural description of an observed overt form, filling in the hidden structure not apparent. The learner uses it to modify their grammar, despite the fact that the full description is based in part on the previous (most likely wrong) grammar. The claim is that the learner can go back and forth between estimating the hidden structure and estimating the grammar, eventually converging on the correct grammar. The results of some simulations are presented in support of the claim.

The recent syntactic literature contains three major proposals concerning the syntactic status of adnominal adjectives: 1. (1) that they are heads, 2. (2) that they are phrasal adjuncts, and 3. (3) that they do not form a uniform class. Given this lack of consensus, it is not surprising that the distribution of adnominal adjectives had remained an unsolved problem in the study of intrasentential codeswitching. In this paper, we propose a solution by exploiting the formal model of intrasentential codeswitching developed in Mahootian (1993), which is based on the null hypothesis that codeswitching sequences are not subject to structural constraints beyond the general principles of phrase structure that govern monolingual sequences. We first show that a blanket treatment of adnominal adjectives as heads is untenable, and then argue that the distribution of adnominal adjectives described in the codeswitching literature provides a new source of evidence in favor of the view that there are two distinct classes of adjectives. We review a number of alternative treatments proposed in the codeswitching literature and show that our analysis is preferable.

A form of grammar called generalized generalized phrase structure grammar or G 2 PSG is presented. G 2 PSG is a generalization of generalized phrase structure grammar (GPSG) in which the head feature convention, the foot feature principle, the control agreement principle and the system of feature specification defaults are subsumed into a single mechanism which is an extension of the feature co-occurrence restriction mechanism of standard GPSG. A parsing algorithm for G 2 PSGs is described.

Linguistic creativity is a traditionally-recongnized trait of natural language. Certain proposals in transformational linguistics seek to explain linguistic creativity with the help of the metamathematical notion of recursive rules. It is argued in this paper that a close investigation of this aspect of transformational theory reveals little empirical or conceptual support for inducting these rules from metamathematics into linguistic theory. Several major assumptions of transformational theory interact significantly with the axiomatic character of its recursive rules; some of these assumptions have been examined in this paper. In this connection, a brief comparison between transformational grammar and Pānini's linguistic theory — in the context of a prefactory remark on Pānini's grammar in Chomsky's Aspects — also appears to throw valuable light on the possible form the principle of linguistic creativity may take in the grammar of a natural language.

The notion of valence and the relation of dependency connected with it were introduced into the theory of grammar by Tesniere. Later, Hays, Gaifman, and Baumgartner showed that dependency grammar and constituent grammar are not only complementary but (at least weakly) equivalent. Robinson worked out a model of a generative grammar with a deep structure built on dependency relations rather than on phrase structure relations. Robinson argues that the concept of head cannot be formalized within the framework of a phrase-structure categorial component, but that it can be formally specified for each phrase, if dependency rules generate the structural strings of categories, thus supplying additional information needed for some of the transformations. In this paper, 1 an attempt has been made to overcome the shortcomings in Robinson's model by modifying her dependency rules and adding semantic specifications to the dependents of V, taking into account some of the considerations that led Fillmore to make up his ‘cases’.

This paper compares the predictions made by two competing general models of phrase structure for cross-language word order patterns. The first general model is categorial grammar, in which constituents contract ‘function-argument’ relations. The second is exemplified by theories such as the X-bar theory in which constituents are either heads of phrase or (various types of) modifiers of these, and hence in which the ‘modifier-head’ relation holds. It is the general differences between these two types of relations which will concern us, for they result in partially different structural groupings being assigned to the same surface constituents, and in very different node labellings. In particular ‘they embody different claims about the kinds of syntagmatic relations that hold between constituents. Functionargument relations are primarily semantically motivated, and the syntactic rules of phrasal construction are assumed to mirror these isomorphically. Modifier-head relations are typically syntactically motivated, within the context of autonomous theories of syntax. The set of word order data to be used is taken from Hawkins (1983). The question we shall address is: how well does one or the other theory succeed in describing and explaining these universal word order data? There are, of course, numerous exemplars of each of these models. Compare, for example, the categorial grammars of Ajdukiewicz (1935), Montague (1974), Vennemann (1976, 198 l), Vennemann and Harlow (1977), Keenan (1979), Keenan and Faltz (1978), Flynn (1982). Modifier-head

Conditions on rules affecting the positioning of Spanish clitic pronouns within a transformational framework are discussed. It is argued that the Specified Subject Condition and the Tensed Sentence Condition are irrelevant to the analysis. The primary conditions which must be satisfied for a clitic pronoun to move from one verb to another are that the verbs must be adjacent to each other and that the target verb must be ‘superior’ to the source verb. It is proposed that the work done by transformations can be greatly reduced if clitic pronouns are generated adjacent to their verbs by the phrase structure rules. This phrase structure analysis allows the straight-forward explanation of some data which are troublesome for the full transformational analysis. The Clause Reduction analysis of clitic positioning is discussed and rejected. Finally, some still unexplained facts are presented.

This study presents a generative phonological grammar of a child's productive competence in terms of adult model and child replica. The inventory of monoremes produced by the child at 0 ; 11 and the mother's models for them are described in terms of syllables, a set of distinctive features, and the process of reduplication. The grammar proper consist of a set of rules characterizing both the set of possible words which the child at that stage was able to use as models and the set of possible monoremes which were ‘pronounceable’ by the child at that stage. The possible models and replicas generated include those actually attested as well as additional items assumed to be ‘accidentally’ absent from the data. Observations are made on the kind of evidence which would confirm or disconfirm the grammar, and further observations are made on the validity of this kind of grammar and the nature of its rules.