Lingvisticæ Investigationes - Volume 17, Issue 2, 1993

In this article we have stressed the treatment of the genitive case for a syntactic classification of sentences containing frozen complements: the genitive presents a problem to the extent that several syntactic functions can be assigned to it. Thus, on the one hand we examine sentences whose complement in the genitive is frozen and, on the other hand, we examine frozen sentences whose genitive complement is free.In the first case, we use three tests to determine the syntactic status of the genitive in question:(i) the alternation of the genitive complement with a prepositional phrase;(ii) a comparison with free sentences having an equivalent structure;(iii) the paraphrase of the genitive complement by an adverbial, which is frequently prepositional.In the second case, when the base form is N0 V C (accus) N (gén), we have made use of such properties as the following:(i) the pronominalisation in the form of a Ppv,(ii) the pronominalisation in the form of a Poss,(iii) the alternation of the genitive with a prepositional phrase (à N (accus)), which brings out distinct structures quite clearly.In this way, taking syntactic criteria into account for the analysis of (free or frozen) genitive forms allows us to set up classes that are more homogeneous from the point of view of their syntax. We have also been able to observe that cases, as morphological markers, play no essential role in the criteria that constitute the basis of our classification.

We present a detailed account of the shapes of prepositional noun phrases which complement verbs of movement. These complements are interpreted as destinations of the movement. We use the example of names of countries and islands to show the complexity of the problem. Attempts to predict the choice of a preposition according to the morphology of names is only partially correct. In fact, we show that in order to describe exactly the use of prepositions, we have to display all the quadruples {Verb, Preposition, Article, Noun}. Subregularities can then be accounted for, by means of a representation by finite automata. This representation has the advantage to provide directly a recognition algorithm which has been applied to a corpus. We present a sample of the results thus obtained.

Rigorous definitions are proposed for three important syntactic-morphological concepts: agreement, government and congruence. They are defined as particular cases of morphological dependency between wordforms of an utterance (distinguished from semantic and syntactic dependencies between wordforms). Definitions are based on the intermediate concepts of agreement class and related inflectional categories, as well as on the concepts of inflectional category, grammeme, syntactics feature and syntactics feature value. AGREEMENT is defined (roughly speaking) as a morphological dependency where a grammeme of the target, which is not a substitute pronoun, is selected depending either 1) upon a grammeme of a related category of the controller, or 2) upon its agreement class, pronominal person or pronominal number (syntactics features), or else 3) upon some of its semantic properties. GOVERNMENT is defined as a morphological dependency where a grammeme of the target is selected depending either 1) upon a grammeme of an unrelated category of the controller or 2) upon one of its syntactics features, which is not agreement class, pronominal person or pronominal number. CONGRUENCE is defined as a morphological dependency where a grammeme of the target, which is a substitute pronoun replacing an occurrence of the controller, is selected depending upon any property of the controller. Numerous examples of agreement, government and congruence are cited and analyzed, a comparison of agreement and government is presented, and relationships between these concepts and other types of dependencies are examined.

This work deals with the morphosyntactical relations between a noun complement and an adjective in Dét V-n (de N1 + Adj) noun phrases in which V-n is a noun derived from a transitive verb. These noun phrases are similar to transitive verb structures in which the noun N1 is the direct object of the verb. We show that the operations that relate the two structures N0 V N1 and Dét V-n (de N1 + Adj) consist in zeroing the N0 argument through a passive or a middle transformation.

Our project is to describe French, using a finite vocabulary and a finite set of grammar rules. As far as noun phrases are concerned, we have to decide which noun phrases are to be treated by grammar rules, and which ones should be put in a lexicon. We present four criteria that have been used to build the electronic dictionary for compound nouns, (DELAC), which contains over 100,000 entries. The first two criteria (not compositional meaning, and institutionalized terms) yield purely lexicalized entries; the last two criteria (distributional restrictions and exceptional transformational analysis) concern certain noun phrases which could probably be treated by grammar rules; we prefer to put them in a lexicon for methodological reasons.