ABSTRACT
The evolution of the author thinking about morphological processing that has occurred over the last 20 years or so. The original framework in which the author thought about the issue was that of a prelexical morphological decomposition procedure taking place prior to a search for the stem morpheme within the lexicon. The author acknowledges the fact that the materials used in the Taft and Forster (1975) experiment were very poorly controlled. He repeated the experiment with an improvement in the construction of the materials. In more recent times, however, he have favoured a multilevel interactive-activation framework, where one of the levels contains morphemic units. The interactive-activation account is more flexible than the prefix-stripping/search framework, in that the equivalent of the prefix-stripping procedure is an integral part of the access process itself rather than a discrete stage of processing that takes place prior to access.
