| Popis: |
The positioning of the attributive adjective is a subject that has produced a substantial literature in French linguistics. In most articles and books, it is considered that the adjective may as well be before or after its head noun. However, previous studies have focused almost exclusively on written French and / or invented examples. There are few studies on spoken French and no broad description. We therefore choose to launch a systematic description of the position of adjectives in spoken French (mostly unplanned), starting by limiting to adjectives appearing on both right and left of the noun to which they relate. These adjectives are described as alternating. This is the case in examples like: une sublime jeune femme / une jeune femme sublime (‘a gorgeous young woman’). We focus on adjectives that can be pre-nominal or post-nominal in similar or identical contexts, without apparent distinctive setting (as in the example above). To carry out this study, we adopted a corpus-driven approach (Tonigni-Bonelli, 2001), restricted exclusively to occurrences met in our data. This paper presents a first exploratory study in which we mention only the first results. To highlight the influence on the results of the nature of the data taken into account, a comparison will be conducted with data of the same size from newspaper. It will be focusing specifically on the cases for which the trends are not congruent between oral and newspaper. So, we show that the detection of regularities can be strongly affected by the data taken into account. Our study is based on an oral resource of approximately 2.3 million words coming from a combination of various corpora. File formats and transcription conventions were standardized. The entire resource was automatically tagged into parts of speech and lemmatized with a tool specially trained to work on spoken French. Regarding the written data for comparisons, we also combined various newspaper corpora. The size of the newspaper part is comparable to the size of the oral part. Queries were launched on the tagged corpus to observe the main trends and to extract the list of alternating adjectives. Then we read concordances to identify and correct detection errors. However by this method, we realized that we missed a small part of relevant examples. In a second step, to have the guarantee to work on the entirety of relevant data and to control a maximum of parameters, we queried directly the raw transcriptions to lead detailed analysis. On the tagged spoken corpus, we identified a total of 33,397 occurrences being divided as follows: 16,037 pre-nominal adjectives (48%) and 17,360 post-nominal adjectives (52%). Alternating adjectives are not frequent in terms of different forms (only 162 of 2,146), but they are the majority in terms of occurrences (19,130). This phenomenon is related to the presence, in this category, of the most common adjectives like grand (‘large’), petit (‘little’), autre (‘other’), etc. Furthermore, we saw that the pre-nominal position, among alternating adjectives, are large majority and that majority is mostly related to the 10 most frequent adjectives in this position. Regarding the various adjective usages, we highlighted four cases. First, we observed non spontaneous alternation. They are adjectives that alternate in writing but very rarely in the oral data that we viewed. The only examples are encountered in recordings of planned spoken French: Frodo jouit pleinement de son actuelle suprématie en l’absence de Freud [c-oral-rom] (‘Frodo fully enjoys his current supremacy in the absence of Freud’ [tv report on chimpanzees]) or in situations where there is an effect of parallelism: c’est un long travail et un dur travail [CRFP] (‘it is a long work and a hard work’). The second case includes adjectives which the position is in complementary distributions. In this category, we met only discriminating contexts. This is the case of the adjective prochain (‘next’ in English). All contexts in which prochain is in pre-nominal position are different from those where it is in post-nominal position. The third case is similar to the previous but it contains some rare examples that are not in complementary distribution. The adjective propre (‘own’ in English) belongs to this category. We have found two post-nominal examples that are identical to the pre-nominal context (presence of a possessive determiner): le patois était le euh leur ben leur langage propre [Corpaix] (‘the dialect was their own language’). Finally, there are free alternate adjectives, that is to say adjectives can be either pre-nominal or post-nominal. This last category seems relatively marginal compared to others. But only a thorough work on the 162 alternate adjectives will demonstrate this hypothesis. We draw the tentative conclusion that the position of adjectives in unplanned spoken French is mostly intended. This point highlights the existence of a spontaneous grammar, hard to see in written data especially for stylistic reasons. |
