Modeling Lexical Decision as Ordinary Retrieval

We present an ACT-R model of signal-to-respond lexical decision. This model is based on a mental lexicon constructed from a complete list of English four-letter words with associated word frequencies. A number of (signal-to-respond) lexical decision phenomena are explained by the model, without having to rely on mechanisms specific to lexical decision. Besides explaining these phenomena, the model also illustrates the necessity of a competitive latency mechanism. According to this mechanism, the time necessary for retrieving an element from declarative memory is a function of the activation of all other elements in declarative memory. We argue that the data presented here can only be explained using a competitive latency