LEARNING AND MORPHOLOGICAL CHANGE

An account is offered to change over time in English verb morphology, based on a connectionist approach to how morphological knowledge is ac quired and used. A technique is first described that was developed for modeling historical change in connectionist networks, and that techni que is applied to model English verb inflection as it developed from t he highly complex past tense system of Old English towards that of the modern language, with one predominant ''regular'' inflection and a sm all number of irregular forms. The model relies on the fact that certa in input-output mappings are easier than others to learn in a connecti onist network. Highly frequent patterns, or those that share phonologi cal regularities with a number of others, are learned more quickly and with lower error than low-frequency, highly irregular patterns. A net work is taught a data set representative of the verb classes of Old En glish, but learning is stopped before reaching asymptote, and the outp ut of this network is used as the teacher of a new net. As a result, t he errors in the first network were passed on to become part of the da ta set of the second. Those patterns that are hardest to learn led to the most errors, and over time are ''regularized'' to fit a more domin ant pattern. The results of the networks simulations were highly consi stent with the major historical developments. These results are predic ted from well-understood aspects of network dynamics, which therefore provide a rationale for the shape of the attested changes.