Technical texts present interesting and so far poorly researched linguistic characteristics. In this article, a research project is described, carried out by a multidisciplinary group of linguists and computer scientists, which aims at devising and realising prototypes of computer programmes for extracting information from technical texts. This research, as is illustrated by concrete examples, has led to computer programmes that have the form either of networks between concepts or of phrases taken from the analysed texts, and that are, if necessary, accompanied by automatically assigned semantic information.

Listening and reading comprehension of several syntactic structures was analysed with second to fifth grade readers. Three different structures were compared in both reading and listening: simple sentences, simple sentences with morphological markers for number and gender, and complex relative sentences. The difference between reading and listening was analysed systematically. Data showed that this difference changes with school level and with the type of syntactic structure and their specific constraints. The constraints that differentiate most between reading and listening were shown to be the identification of morphological markers necessary for identifying words (for simple sentences with markers) or the handling of working memory resources (for complex sentences).

In this paper, the author describes some of the linguistic aspects of bilingual speakers' development and emphasizes particularly the cognitive data. In the first section, a number of neurophysiological mechanisms underlying language acquisition are sketched out. In the second section, some implications for the neurophysiology of bilingualism are examined. Lastly, the main trends of acquisition and processing of bilingual mental lexicon in children are put forward.