Taking into account the theoretical assumptions that praise the description of the use of the system and not only of the system in abstract, we will defend the need for a grammar of uses for European Portuguese. Such a description implies dealing with the complexity, the gradualness of linguistic phenomena, the compulsory contextualization of discourses, both in discursive genres and in their enunciative circumstances. This description testifies that regularities also exist in use and there should not be an excessive cut between describing the language and describing the use of language. Unlike what happens in Brazilian Portuguese, in European Portuguese there are few researches that focus on these uses, and even less researches that deal with oral productions, namely interactional and informal, based on corpora. It is necessary to pay more attention to the uses, above all to the description of the functioning of the language as discourse updated in certain discursive genres, as the informal conversation. These studies are crucial both for the teaching of Portuguese as a foreign language and as a mother tongue, since describing only the system, the standard variety, generally written, is scientifically reductive and ineffective from the point of view of teaching. But descriptions of uses are also essential for translation studies or contrastive analyzes of both different languages and different varieties of Portuguese. Starting from small research experiences that take place with university students, we propose a research route with corpora of familiar conversations, which try to account for the unguarded oral uses of the language, in an interaction context. Finally, two examples of this working plan will be advanced: that of the presentational marker "é assim" and that of the vague nominal quantifiers, with a function of attenuation and diminution of the speaker's enunciative responsibility, "Um bocado = a bit" and "Um bocadinho = a little." We will conclude by showing how the corpora analysis of informal conversations even allows us to better understand some characteristics of literary texts.
Exploring content selection strategies for Multilingual Multi-Document Summarization based on the Universal Network Language (UNL)
