Evaluating integrated NLP in foreign language learning: technology meets pedagogy

In this paper I would like to present the pedagogical perspective upon a 3-year research project with direct implications for the field of Foreign Language Teaching (FLT), and English for Professional Purposes (ESP) in particular. As a project in Computer Science, the basic aim of LARFLAST 1 was to investigate the integration of several innovative software components into a harmonized environment for learning English financial terminology. In order to ensure the pedagogical relevance of the work, user involvement and endorsement of the research prototype were sought at every stage of the project. The working bench consisted of the following resources: 1) a training corpus of texts for selfstudy 2) a collection of additional relevant texts, gathered by an intelligent Web-agent 3) a terminology testing section of exercises in the popular CALL formats, including a NLP component 4) online financial dictionaries 5) GenWeb, generating dynamic web pages, intended as immersive contexts for the learners, making use of the capabilities provided by web browsers. LARFLAST studied and integrated language technologies like semantic analysis, information retrieval and elements of natural language generation. The paper discusses the difficult negotiation process between user requirements, pedagogical rationale and the specificity of technological problems that participants sought to solve and subsequent work, trying to cope with some of the problems.