Does spreading activation summate?

SummaryNetwork models of semantic memory assume implicitly or explicitly that the degree of activation of a node is a monotonic function of the total amount of excitation reaching that node from all sources. For example, the activation level of the node representing “apple” should be greater if it is receiving excitation due to the activation of the nodes for “fruit” and “pear” than if only one of these neighboring nodes is activated. This notion was tested by presenting semantic primes 80 ms or 320 ms before a letter string (e.g., apple or ipple) requiring a lexical decision. The prime stimuli consisted of a pair of simultaneous items that were identical in the single prime condition (e.g., fruit/fruit or pear/pear) or different in the double prime condition (e.g., fruit/pear or pear/fruit), and were either related (target = apple) or unrelated (target = copper) to the target, or neutral. As predicted by the summation of semantic activation assumption of network models there was a larger priming effect (in the 320 ms SOA condition) following the simultaneous activation of two related nodes than following the activation of only one node.