Contenu du sommaire : Réanalyse et changement linguistique

Two points of view on language change Nowadays. Numbers of authors approach language change in a deterministic point of view through grammaticalization scales leading from “lexical” to “grammatical”, from “less grammatical” to “more grammatical”, from “semantically full” to “semantically empty”. This unidirectional view of change finds an early criticism in F. de Saussure's thoughts, for whom grammatical change is unpredictable, non-teleological, due to accidental (re) analyses of the language material that occur hic et nunc in the mind of language users, under the exclusive pressure of contemporary forms. In the light of these two antagonist points of view, we examine a few cases of syntactic change and eventually raise the question where the supposed “tendencies” should be situated as regards language evolution.

Syntactic Reanalysis and the Formalist-Functionalist Conflict in Linguistics. Two positions with respect to reanalysis are widespread in the literature. The first is a central feature of the predominant principles-and-parameters approach to formal syntax. In this view, the sudden resetting of a parameter leads to a cascade of structural reanalyses. The second has been developed primarily by functional linguists. In this way of looking at things, diachronic change is “fuzzy”, in the sense that categories and syntactic constructions undergo gradual reanalysis. In this article, I argue that neither position is correct. Reanalysis is in fact discrete and sudden. However, one must abandon the idea that reanalysis can be captured in terms of parameter-resetting.

Reanalysis and linguistic change. The purpose of this contribution is to examine the contact point which it is possible to determine between reanalysis and diachronic study of language. We begin with the main characteristics of reanalysis, the question of the nature of the process and the problem of the role that plays analogy. We observe then how reanalysis is submitted to the main tendencies of the evolution: change from synthetic system to analytical system, hierarchical organization of syntagms and specialization of categories. We finally envisage the nature and the role of ambiguity, the phase of actualization, which entitle to consider that the reanalysis is entirely completed.

Reanalysis and analogical attractions: the example of French infinitive constructions. In the transition from the “old” word order Je le dois faire into the “modern” one Je dois le faire, we may have seen the sign of a reorganisation of infinitives constructions: first considered as verbal compound lexia [auxiliary+infinitive] (= C1 construction), they would supposedly then have been reanalyzed as verb phrases [governing verb+infinitive argument clause] (= C2 construction). A careful examination of the data shows that in fact, each of the two constructions exerted an analogical influence over the other, giving rise to a hybrid variant of its rival. If C2 and its hybrid result have imposed over time, it is for optimality reasons. The principal diachronic change that resulted from it is the birth in French of an original predicate auxiliaries' category. The reconversion of verbal auxiliaries into predicate auxiliaries heads in some respects towards a degrammaticalization.

Reanalysis of the et (and) conjunction. This study focuses on two classes of constructions articulated by et (and), one in Old French, the other in contemporary French: (i) constructions “in combination” in Old French as (1) Quant vit abatu son escu, / Et lors n'i a plus atendu (Sone de Nansai, quoted by Antoine 1958); (ii) constructions with a subject inversion, and articulated by et (and) as (2) La tension monte-t-elle à une table, et sa voix de velours suffit à l'apaiser (printed news). In both cases, the reanalysis of the structure causes a change in the value of et (and). On the basis of these two micro-phenomena, I will make several observations on the theoretical understanding of the linguistic change.

Reflections on the origin of insubordination. The case of three hypothetical insubordinates in French The paper examines the status and the origin of French hypothetical subordinates that are not governed by an explicit matrix clause (e.g., Oh si j'étais riche!). In the literature, such constructions are generally considered as instances of “insubordination” (Evans 2007), i.e. as subordinates freed from their matrix clause through a process of grammaticalization. The article suggest a description of the three hypothetical insubordinates of French, which confirms some degree of autonomy acquired by [si p] from a semantico-pragmatic, prosodic and textual point of view. Nevertheless, I show that these insubordinates do not exhibit clear signs of grammaticalization and that an explanation in terms of reanalysis only is better able to account for the linguistic change.

Graphical reanalyses: ‘spontaneous writing' in text messages and the status of pronominal clitics in contemporary French. A manual analysis of the first 400 French text messages of the Swiss SMS corpus according to their graphical realization of clitic subjects and object clitics (agglutination / complete ‘fusion' with the finite verb, like in chu à la bourre, je suis en train de travailler, or complete drop, like in suis désolée pour ta sale nuit) leads to the following results: Clitic subjects are often heavily modified at the graphical level (agglutination, fusion), but also regularly dropped, when e.g. topical, i.e. for discourse-pragmatic reasons. That is a contradictory result, as the first supports, whereas the second contradicts their status as reanalyzed agreement affixes/markers. This holds also for object clitics: they are only rarely modified graphically (= no affixes?), but almost never dropped (= agreement markers?).