Usages du gérondif et du participe présent en français parlé et écrit : étude comparée basée sur corpus

The aim of the research is to better understand the use of the French gérondif (i.e. en parlant) and present participle (i.e. parlant) in spoken and written contemporary French. This is the third part of an earlier data-driven quantitative and qualitative analysis. It is based on a written corpus of about 6,800,000 words and an oral corpus of approximately 3,200,000 words. Written French refers to the language of contemporary French newspapers (i.e. national daily quality press), whereas spoken French is to be understood as spontaneous oral productions - free and guided interviews, family conversations, telephone conversations, various daily life verbal interactions – between speakers whose mother tongue (or one of their native languages) is French. Our first objective is to verify the reproducibility of the results obtained in our previous empirical studies (Floquet et al. 2012, Escoubas-Benveniste et al. 2012) by further reducing the potential bias that might have been introduced by the difference in the size of the oral and the written corpora analyzed so far. Our second goal is to enrich the description of linguistic facts by introducing a diatopic dimension of two varieties of French into the spoken data. Consequently the previous sample of spoken data was enlarged by adding one more corpus of the French spoken in France and two corpora of, respectively, Switzerland and Quebec spoken French language. We applied the previously developed quantitative analysis (Escoubas-Benveniste et al. 2012), which combines computer assisted TaLTaC2 (Bolasco, 2010) and manual exploration performed on a written newspaper corpus (L’Humanité, 2010) and on two spoken corpora – corpus Phonologie du français contemporain (PFC); Corpus du français parlé parisien (CFPP) –, to the three additional speech corpora of spontaneous oral data – Enquête socio-linguistique à Orléans (ESLO), Corpus oral de français de Suisse romande (OFROM); Corpus du français parlé au Québec (CFPQ). A qualitative analysis followed, whose purpose was to describe lexical, morphological and syntactical aspects of both non-finite verbal uses from a differential point of view (written vs spoken corpora; gérondif vs present participle). In order to examine the reproducibility of our earlier results on the status of the two verbal forms in -ant, we evaluate (a) the rarity of free (as opposed to lexicalized) verbal forms in -ant in the spontaneous spoken vs newspaper written French corpora; (b) the differential distribution of gérondifs and present participles in both types of data; (c) the presence of both verbal modes in spontaneous spoken French, taking into account the phenomena of grammaticalization. The quantitative and qualitative analysis performed on the much larger sample of spoken French confirms our previous results as for (a) the relative rarity of gérondifs in spoken data as compared to journalistic written data and the tendency shown by these forms to be used as “constructions absolues” (Floquet et al. 2012) (i.e. often initial clause syntactically unbound constructions); (b) the rarity of “free” uses (i.e. non grammaticalized uses) of both non finite verbal forms in -ant in the oral as compared to the written data and also the greater frequency of the gérondif in the spoken corpora, whereas the present participle is much more frequent than the gérondif in the written one (Escoubas-Benveniste et al. 2012). The lexical and morpho-syntactical analysis of the free non finite verbal oral occurrences shows, apart from some canonical uses also present in the written corpus, other interesting cases indicating that the distinction between the uses of these non finite forms might be diminishing in the spoken language. In quantitative terms and with respect to the comparative distribution of both verbal non-finite forms in the written and oral corpora, the data show that the gérondif plays a pre-eminent role in relation to the participle in spontaneous speech: when it is the result of a free construction, the gérondif is 3 times more common than the participle in oral data and this distribution is reversed in written French. In spontaneous speech, as illustrated by the oral corpora, the gérondif is also more often the result of a free construction (91% of occurrences) than the present participle (60% of occurrences). Written data, by contrast, show that present participles are more common than gerunds and correspond to a free form in 99% of cases. From a qualitative standpoint, the greater frequency of the gérondif in oral data could well be correlated with the lexical nature of the verbs. In fact, the data show two frequent types of spontaneous oral specific constructions that seem strongly associated with specific discursive functions: 1) en disant that introduces reported speech/thought, marking therefore changes in enunciative levels; it tends to override the participle as a noun modifier; (2) the gérondifs of moving verbs that occur in two types of constructions: moving manner constrained constructions (i.e. sortir en courant vs run out) and spatial location constructions (i.e. X est là en entrant; il y a Y en descendant). These types of gérondifs are closely related to descriptive and narrative discourses that are generally well represented in spontaneous speech. Syntactically, the spoken uses of the gérondifs are free of any constraint on their syntactic empty subject, whose referent (the agent of the process) may be either explicit or implicit (and expressed in some way outside the utterance frame) or even not expressed but pragmatically identifiable. As for the other predicate arguments of the verb, they are rarely expressed within the gérondif structure: indeed, gérondifs with more than one cliticised argument are extremely rare. The spontaneous speech data show uses of simple and compound gérondifs, that standard rules of written language would normally reserve to the present participle, such as second level causal predicates. One explanation for these uses might be found in a position conflict deriving from the canonical syntactic order for these constructions (in the preverbal zone) and the position they effectively receive during the real-time elaboration of spontaneous speech. Therefore a systematic analysis of the various structural levels is needed in order to better describe these phenomena. Finally, in the varieties of French that are spoken in Quebec and in Switzerland, the present participle tends to incorporates the uses of the gérondif and shows particular cases in which it is not easy to discriminate its exact nature. The reasons for this reverse trend remain to be elucidated. In order to account for such an opposition between the uses of gérondif and present participles in spoken and written samples of French, we will evoke the idea that each particular form of expression - journalistic formal written register as opposed to spoken informal register of private/family verbal interactions– is ruled by a distinct grammar. According to this view, the use of gérondifs and present participles would be rare in the grammar of spoken French and limited to a reduced number of lexical verb types and syntactic-semantic patterns that might undergo a form of lexicalization