Third International Conference on Spoken Language Processing (ICSLP 94)

Yokohama, Japan
September 18-22, 1994

Frames, a Unified Model for the Representation of Reference and Space in a Man-Machine Dialogue

Daniel Schang, Laurent Romary

CRIN-CNRS-INRIA Lorraine, Vandoeuvre-les-Nancy, France

This paper aims at instanciating the notion of "frames" (from now on "reference context"1) which seems to emerge from recent work on space modeling but has until now been essentially seen as a too general concept to be usable within an understanding system. This notion, which may be subsumed by Langacker's "cognitive domains", still places itself at a more general level than what he introduces as "search domains". We will show that a reference context is essential for any general view of spatial reasoning, especially when considering the classical distinction between landmark and trajector which really takes its meaning within a surrounding "portion" of space such as the one we will present here.

Full Paper

Bibliographic reference.  Schang, Daniel / Romary, Laurent (1994): "Frames, a unified model for the representation of reference and space in a man-machine dialogue", In ICSLP-1994, 951-954.