Empirical Constraints on Computational Theories of Metaphor : Comments on lndurkhya RICHARD

The goal of developing a computational theory of metaphor is an admirable one (Indurkhya, 1986, 1987). Such a theory, however, is adequate only to the extent that it can be embedded into a more comprehensive account of language processing. Empirical investigations have provided some important constraints on how the comprehension and evaluation of metaphorical language may fit into this grander scheme. My goal for these brief comments is to describe three constraints and illustrate their consequences both concretely for Indurkhya’s theory and more abstractly for other possible computational models. The three sections of this commentary discuss issues of metaphor processing that are highlighted by Indurkhya’s work.