Lingvisticæ Investigationes - Volume 9, Issue 1, 1985

Generative grammar around 1980The aim of this paper is to put into interaction Chomsky's methodological claims with his descriptive propositions. §1 presents in compact form Chomsky's methodological positions and in particular Rules and Representations. §2 examines Filters and Control, firstly the control problem and then, systematically, nearly all the filters proposed by Chomsky and Lasnik for English grammar. The main results are the absence of explicative power in all cases and very disturbing descriptive results as regards filters: they suppose different conditions on trace sensivity to them. On Binding is revised in §3 and particularly in relation to Filters and Control. Descriptive results are also disturbing here: they concern structures of obligatory control and embedded infinitive phrases. The Case filter is not a sufficient enough device and filters of the Filters and Control type are always required. Binding conditions on each other leave many unresolved examples. It is claimed (§4) that there is no such permissive notion of counter-examples capable of justifying the detected descriptive problems. Chomsky an hypotheses change, but it is impossible to say if they improve.

Semantics of the Japanese demonstratives.Japanese has a three-way distinction of demonstratives, kore (this), sore (that) and are (that). The distinction between sore and are in their deictic use is much like one found in Latin: sore refers to something near the hearer, are to something not near the speaker nor near the hearer. I claim that the distinction between sore and are in their anaphoric use parallels the distinction between direct, experiential knowledge and conceptual, formal knowledge.

Recent analyses of French cliticization (Borer 1981, Aoun 1981, Jaeggli 1981 ) treat object clitic pronouns as lexical affixes. In each case, the clitic is an argument of the verb, generated by a morphological rule which inserts the features PERSON, GENDER, NUMBER, on the verb. The clitic absorbs the Case assigned by the verb. This analysis fails to explain some important facts about French cliticization, in particular the fact that French clitics are not obligatorily arguments of the verb to which they attach. This problem is discussed in relation to Klavans' and Zwicky's analyses of cliticization, and it shown that French clitics seem to have post-lexical, rather than lexical, characteristics.

This is the analysis of a specific use of the French conjunction mais, when situated in an attributive structure, between two adjectives:p mais q —» SN + V {être) + Adj. , mais Adj.As this simple syntactic pattern allows a variety of different possible semantic interpretations, my chief object was to establish the paradigm of the logical relations leading to the correct understanding of p mais q.When and if possible, the formal indicators of each different acceptation of the structure and its stylistic potential have been specified.

Which is the best way of describing spatial prepositions? Even though geometry and logic provide rigorous and useful tools for this undertaking, I will show that they fall short from describing completely these prepositions. A complete description must take into consideration the movement and the function of their prepositional terms as well as cognitive notions like access to perception, potential meeting, the relationship container/ contained, etc. General orientation, a family resemblance concept whose main features are the frontal direction, the direction of movement and the line of sight, is also needed.

This paper attempts to show that, within a grammatical framework allowing for "empty nodes" in syntactic representations, a crucial theoretical distinction should be drawn between empty categories as defined in the Governement-and-Binding literature: pro, PRO, trace), and elliptical (or: zeugmatic) gaps, such as the empty slots in "Gapped", or "VP Deletion" sentences, once derived through "deletion under identity". Through a brief study of two groups of examples involving Gapping, and "Comparative Deletion", it is suggested that unlike empty categories, which, supposedly, fall under the scope of Core Grammar, elliptical (zeugmatic) gaps abide by distinct linguistic principles, some of which outside the scope of sentence grammar, including a functional Recoverability Principle, and constraints on discourse anaphora.