The role of inhibition in a spreading-activation model of language production. I. The psycholinguistic perspective

Spreading-activation models of language production are only workable to the extent that they manage to solve the “heat death” problem, i.e., the danger that too many nodes in the network are overactive at the same time. Therefore, a delicate balance between activational and deactivational forces has to be struck. Of the three prevailing dampening mechanisms of decay, self-inhibition and other-inhibition, the latter has been selected for closer scrutiny. The key proposal of this two-part article is that activation-based models of language production cannot afford to do without an inhibitory component, in particular lateral inhibition among nodes of the same level. Psycholinguistic evidence is reviewed in an attempt to insulate inhibitory from excitatory mechanisms. Although it is difficult in normal adult language use to distinguish between the effects of excessive activation and insufficient inhibition, some patterns from language acquisition and aphasia can be shown to follow from inhibitory rather than excitatory problems, thus demonstrating the reality of inhibition. In a system of activational and deactivational forces, other-inhibition is claimed to have the excitatory mechanism of syntax as its natural opponent. It is finally argued that other-inhibition offers an explanation for some puzzling findings from the experimental literature.