Abstract: . . .  or memory nodes) in a computational architecture. Computation in activation-based architectures happens via the modulation and propagation of activation. The higher the activation of an element, the more available it is to affect further processing. Ambiguity resolution The process of selecting a preferred interpretation of some locally or globally ambiguous linguistic input (input that is consistent with more than one interpretation). Ambiguity can arise at multiple levels, including syntactic, lexical, semantic, and referential. Bottom-up processing The processing of some input that is a function of properties of the input only (for example, the orthographic form of a word), not the context in which the input appears. Cognitive models Theories of cognitive processes that are embodied in explicit working computer programs that generate behaviour constituting predictions to be compared to human data. Interactive activation architecture Activation-based architectures in which computation emerges from the parallel propagation of activation among a set of highly interconnected neuron-like units. Symbolic cognitive models Cognitive models that process structured, compositional representations composed of patterns (symbols) that provide distal access to other representations. Top-down processing That part of the processing of some input that is a function not only of properties of the input but also the specific context in which the input Page 13 Encyclopedia of Cognitive Science —Computational Psycholinguistics appears (for example, the linguistic context surrounding the occurrence of a word form).  . . .
--3000,1,1500,1714,38435