Lexical contrast maintenance and the organization of sublexical contrast systems

Finished · Usage-Based Theories

Summary

Wedel argues that multiple biases exerting pressure on speech representations can result in the emergence of stable contrastive sublexical categories (i.e., phonemes). Biases that are important are (i) a bias toward sublexical category similarity toward the center of a mass of exemplar representations and (ii) a lexical contrast maintenance bias at the word level that seeks to enhance phonetic cues that aid in contrasting minimal pairs. The combined effect of both biases is the maintenance of sublexical contrasts to retain lexical contrasts (brought about by the lexical contrast maintenance bias) with the clustering of sublexical productions into stable and distinct categories, even in locations in which they do not support lexical contrast (brought about by the similarity bias but restrained from complete collapse by the lexical contrast bias). This is a good argument for how contrastive sublexical categories emerge as stable constituents.

However, one notion that is not touched on in how sublexical units form in the first place and what determines their association (or lack thereof) to a particular category. Context-free sound change, in which one sound category changes in some phonetic property in all contexts, suggests that there needs to be a more abstract level of representation. The big question is how abstract can that level be?