Contenu du sommaire : Les auxiliaires : délimitation, grammaticalisation et analyse, sous la direction de Hava Bat-Zeev Shyldkrot

The common notion of verbal auxiliary is presented and various situations which lead to generalize the notion are discussed. It is shown that aspectual auxiliaries should have the same status as tense auxiliaries. It is also shown that numerous nominal constructions are better analyzed when their main verb is considered as an auxiliary (e.g. Bob stopped selling this books = Bob (stopped + interrupted) the sale of his books).

A historical perspective is presented by Jean-Claude Chevalier who studies the origin and the development of the concept « auxiliary ». The interpretation of verbal morphology in French, within a Latin frame, encourages the first French grammarians to try and understand the reasons for replacing Latin forms such as amavi or amor by French periphrases. The first step would be to list all the forms called « notae » . The second step would be to restrain them to forms of être and avoir named « auxiliaries ». A third step discovers in the use of the two main auxiliaries a very productive range of significations (Ramus). Henceforth, the category « auxiliary » tends to be incorporated in the verbal analysis (Maupas, Oudin, Vaugelas). Nevertheless, it is only with the Grammaire Générale de Port Royal that a complete analysis is suggested.

The aim of the paper is to compare the class of auxiliary verbs in three Romance languages, French, Spanish and Italian, in the framework of grammaticalization theory. Grammaticalization is defined as the process by which lexical items are gradually converted into grammatical markers, i.e. a dynamic operation by which elements can move over time from one linguistic component to another (from the lexicon to the grammar and eventually, to the morphological component). Auxiliaries are defined both on a syntactic and a semantic basis : formally, they are verbs followed by a non finite verb form (only infinitives are considered here) and semantically, they express one of three notional concepts : time, aspect or modality. The major claim of the paper is that French is the most grammaticalized of the three languages, and Spanish the least. The empirical evidence that supports this hypothesis relates to three kinds of data : the number of verbs that can be analyzed as members of the category AUX, the selectional properties concerning the subject NP (± human NP) and the possible alternation of the infinitival complement with NP complements.

Within a Generative framework, specifically the Minimalist Program (Chomsky 1995) combined with what is known as Distributed Morphology theory, Picabia presents a description of verbal inflections thereby offering a syntactic analysis of auxiliary verbs in French. She discusses the status of the auxiliary, which is largely studied by French grammarians. Picabia raises the question whether auxiliary verbs form a lexical or a functional category. In order to demonstrate her point of view, she develops an analysis of avoir et aller, two verbs that can be considered close to functional category, and whose auxiliarity is undoubted.

Bat-Zeev Shyldkrot describes and analyses a series of alternative passive forms, which express a meaning that is not "an active transitive one". She particularly examines structures with se voir/ se faire/ se laisser followed by an infinitive, where the conjugated verb is considered to be an auxiliary. Bat-Zeev Shyldkrot expounds the two parallel methods that are used in order to define passive forms. One method compares other structures to an unmarked passive structure with être. A second method considers the meaning of the sentence as a distinctive feature, and looks for other structures that seem to convey the same meaning. She tries to define the difference in meaning between the various non-canonical passives and the canonical ones .

Danielle Leeman evokes the resemblance existing between the preposition's fonction with respect to the noun on the one hand, and some tenses with respect to the verb, on the other hand. She rejects, like many others, the partition between lexical and grammatical items and offers to elaborate a hypothesis regarding the meaning of these forms. She presents different categories of nouns and examines their compatibility with the preposition dans. Leeman further demonstrates that dans is susceptible to forming both temporal and causal complements. She concludes that whatever the noun category introduced by dans is, the preposition presents the event as a whole. Leeman supports the hypothesis that the preposition establishes a certain aspectual feature of the noun, in the same way as the auxiliary establishes an aspectual interpretation of the verb.

In various languages, such as French, which do not have nominal assertions, a specific auxiliation has taken place for nouns, prepositions, adverbs and adjectives. This auxiliation has « grammaticalized » verbs, nouns and prepositions with aspectual, causative and « modal » values which are complementary to those of the verbal auxiliation. The grammaticalized elements involved in this auxiliation do not have a compositional lexical meaning, but a regular grammatical one, as in the verbal auxiliation. Current linguistic formalisms are unable to represent the distribution properties of the elements entering this rich auxiliation system. The analysis of its genesis in Indo-European languages makes it possible to understand its syntactic and semantic role as well as its striking similarities in Romance and Germanic languages.

In a Lexique-Grammaire framework, Annie Meunier analyzes the sequence Paul a eu l'honnêteté d'intervenir, which conveys a modality parallel to the one expressed by par N (par honnêteté). The first sequence presents certain properties. It carries tense inflections before an infinitival predicate. It also carries a negation mark extended over the entire sentence. The insertion of this modality within the predicate evoked the term auxiliary. Meunier enumerates the properties of the two modalities and the constraints imposed on their usage. It is suggested that the two modalities can be considered modal auxiliaries of a « psychological » nature in the sense of Gross 1975 (table 1). She further demonstrates that despite the differences, the two modalities display semantic and distributional similarities.