From Syllables to Stress: A Cognitively Plausible Model

摘要

The underlying goal in our work is to develop a model of phonology whichincorporates many of the insights of current phonological theories while at the same time being faithful to known constraints on processing in the human brain. If we can't possibly be going through long derivations when we are speaking, then what exactly are we modelling with our phonological analyses. One can draw a distinction between competence and performance here, but that is just begging the question. If we want to have a cognitively plausible model, we minimally need to constrain, if not eliminate, sequential and iterative application of rules, both of which results in long derivations with numerous intermediate stages. Quite simply, the problem is that there isn't time for the brain to perform long phonological derivations under normal circumstances. In this paper we will address a range of issues pertaining to syllabification and stress. Our overall model is implemented in a connectionist framework, and this has imposed significant constraints on the nature of our phonological representations and rules. (KR)