Lingvisticæ Investigationes - Volume 20, Issue 1, 1996

The present work is based on the methodological setting of the LADL studies.In order to establish the body subject to analysis, we selected only the fixed phrases N0 ser/estar Prep X. In this way, we have collected 1138 expressions involving estar Prep X, and 534 with ser Prep X. In order to classify these expressions, we have grouped them according to the number and form of their complements.As to the part pertaining to semantics, the alphabetical indices have allowed us to establish that, concerning ser and estar, all the formal schemes exists for both Spanish and french, at least under the form of fixed expressions.As to the part concerning translation, we have attempted to mesure the isomorphisms of the structures in the two languages; this has enabled us to verify the existence of a near-total parallelism between the structures of the phrases in ser and estar Prep X, and in the structures in être Prép X.

The Lexicon-Grammar of frozen adverbs of comparison has been established for French (M. Gross 1984b), Portuguese (Marques Ranchhod 1990, 1991) and Italian (De Gioia 1994a) in a monolingual perspective. The systematic character of such descriptions allowed us to make an explicit comparison between Italian and Portuguese.The observations we made outlined, on the one hand, that the syntactic properties of these forms are analogous to those of free comparatives, and, on the other hand, that they are largely similar among the Romance languages involved.

The "associative anaphoras" (e.g. He enters the village and goes to see the mayor), contrary to the "true anaphoras" (e.g. A policeman rings at the door; and soon the policeman leaves), are neither co-references, nor lexically identical to the antecedent. Our analysis of the associative anaphora tends to turn it into a particular case of the true anaphora. We can distinguish two main categories of associative anaphoras: those which result from an ellipsis of true anaphora, and those which result from an antecedent ellipsis of true anaphora. Here is an example of the first category: I leave a restaurant; the cook (of that restaurant, 0) is remarkable. Here is an example of the second category: He takes his temperature (with a thermometre, 0); the thermometre shows 39° C.

In this paper, we investigate some syntactic and semantic properties of a subclass of consecution discourse markers, or "connectives", in French. Consecution connectives express causal, implicative, or deductive relations between propositional entities. Typical instances of the class in English are therefore, so, then. We study the specific properties of donc (resembling therefore) in contrast with de ce fait, du coup, and alors (which can be rendered by so or then in many cases). We first describe the surface position constraints for these connectives, and relate the observations to the general problem of adverb position. Next, we appeal to a basic distinction between illocutionary force, propositional attitude, and propositional content to explain some observed semantic scope differences among the four items. Focussing on donc, we turn to the problem of the (in)compatibility of these connectives with if-sentences, illustrated by the following contrast.(A) S'il fait beau, alors j'irai me promenerIf the weather is fine, (then) I'll have a walk(B ) ?? S'il fait beau, donc j'irai me promenerIf the weather is fine, (therefore) I'll have a walkWe connect this difference with a variation in connexion "strength", a notion we substantiate in the final section by resorting to a version of generalized quantification. We also consider the influence of surfacc position on acceptability for examples of type (B), and we propose that this should be related to a difference in syntactic scope, which shows in explicative c 'est que structures as well.

This article presents the first results of a research project to develop an automated system for identifying thesaural items in electronic text (the ACRONYM Project). The main objective is the identification and use of nyms (words or groups of words occurring in similar collocational environments and, as such, deemed to share meaning reference).The thesaurus, readily updatable, is a representation of textual reality by virtue of being automatically extracted from the text itself.

In all languages, the ambiguities of simple forms pose a difficult problem in automatic text analysis. The data collected in the present work give a broad coverage of lexical ambiguities in French. They are obtained from the electronic dictionnary DELAF, which contains more than 650,000 forms.For the purpose of analysis, ambiguities are distributed into four large groups: ambiguities between two verbal forms, ambiguities between a verbal form and a non-verbal form, ambiguities between a noun and a non-verbal form, remaining ambiguities. Moreover, each group is divided into homogeneous subsets: for instance, one subset includes all forms which are ambiguous between a verb in the indicative present and a feminin noun. This classification into groups and subsets is presented with many illustrative examples.The inventory of all ambiguous French forms is an essential tool for the writing of algorithms needed for resolving ambiguities in text analysis.

In this work, we study the transformational relations between adjectival and verbal constructions in a syntactic approach. This study is composed of three parts. First, we describe the characteristics of this type of adjectives and we deal with different problems such as the collection of these adjectives, the description of their syntactic properties, etc. The second part includes the chapters 3 and 4, in which we describe the similarities and the differences between the two constructions. Syntactic and transformational regularities appear but they are not systematic, thus the construction of the lexicon-grammar of these adjectives is necessary. In the third part, which contains chapters 5 and 6, we describe adjectival constructions that have an appropriate noun in subject position. First, we classify these nouns according to semantic criteria, then we investigate the properties of the constructions obtained with different transformations, i.e., a nominal sentence containing a support verb, a sentence in which the subject is restructurated and a metonymical sentence.