Learn to Describe Objects the way ‘Ordinary’ People Do by Using Language Technology

Abstract

Speaking consists to a large extend in making statements about objects: ‘”The baby is asleep”, “The game is over”, etc. Yet objects can be described in many ways: it, the library, the first building to the right, etc. The way how an object is described or referred to depends to a large extent on the context, i.e. the set of elements (physically or mentally present in the listeners mind) with which the target can be confused (think of different kind of round objects you can purchase in a sporting goods retail store). Hence, the question: how does a speakers decide on the information to convey (content) and on the linguistic resource to use (expression)? This implies among other things pragmatic knowledge, which is often learned on the basis of correlations. People realize that changes of the situation often reflect in language: different inputs (ideas, objects of a scene) yielding different outputs, i.e. linguistic forms. We present here a setting that allows for this kind of learning. It is a web-based application that generates a scene and various descriptions of its components. Users can change the scene and watch how these choices affect (or not) the linguistic form. The descriptions are produced in English and French, and they are rated in terms of communicative adequacy. This should allow students not only to learn to produce correct sentences, but also help them to realize which one of them is, communicatively speaking, the most adequate form.