This Ph. D. thesis addresses the issue of accessing the content of digital texts, in so as it is linked to the expression of temporal location. The aim of this work is twofold: it consists in conciliating a linguistic approach and an applied approach, to participate in the development of new tools for text mining, information retrieval and knowledge management - new tools that would able to take advantage of the semantics of temporal information expressed in texts. It is thus both about implementing interaction systems and modeling the semantics of one the most salient textual units that contributes to anchor in time the situations described in texts: temporal locating adverbials. The linguistic analysis aims to show that language engineering can benefit from considering in a new way the phenomena of temporal location, by uncovering the different operations at work in temporal locating adverbials. Pinpointing the semantic values of these operations, we show that it becomes possible to develop new Information Retrieval systems, able of processing queries involving both a calendar criterion and a set of keywords, such as "universities in the early twelfth century", for instance. We show that it also becomes possible to interact with structured data describing temporal information, both for search process and for data semi-automatic enrichment.
