 The cohort model was shaped to account for psychological data on the perception of language. It assumes that there is a separate active processing element associated with each item in the lexicon. As soon as there is a match between the input and the perceptual specifications of these elements, syntactic and semantic specifications associated with a respective lexical item are brought into operation. The following animation shows the gradual shrinking of the cohort of potential candidates with increasing input information. In an extended version of the cohort model, several cohorts that share the main perceptual specifications of the access unit are activated simultaneously. This extended version accounts for our ability to identify even incomplete input.