Figure 2. Outline of the model and its dynamics.

A: Model Architecture. Patterns representing related concepts are correlated in the semantic network but uncorrelated in the lexical network. Active neurons of two example patterns representing ‘dog’ and ‘cat’ are marked. Connections between networks are from active neurons of a pattern in one network to all the corresponding active neurons in the other network. For simplicity, only some of these connections are drawn. B: Examples of pairs used in the simulation trials, organized by relatedness condition. C: Example of expected chain of events in a semantic priming simulation. Lexical network converges to the prime pattern, followed by convergence of the semantic network. When target appears, the lexical network converges to the appropriate target pattern under the influence of the semantic network. No latching dynamics is assumed in the example.