Computational aspects of concept invention

Abstract

In this chapter, we present a computational framework that models concept invention. The framework is based on and extends conceptual blending. Apart from the blending mechanism modeling the creation of new concepts, the framework considers two extra dimensions, namely, origin and destination. For the former, we describe how a Rich Background supports the discovery of input concepts to be blended. For the latter, we show how arguments, promoting or demoting the values of an audience, to which the invention is headed, can be used to evaluate the candidate blends created.We also address the problem of how newly invented concepts are evaluated with respect to a Rich Background so as to decide which of them are to be accepted into a system of familiar concepts, and how this, in turn, may affect the previously accepted conceptualisation. As technique to tackle this problem we explore the applicability of Paul Thagard’s computational theory of coherence, in particular, his notion of conceptual coherence. The process model is exemplified using two structured representation languages, namely order-sorted feature terms and description logic.