Theories disagree on the proposed mechanism underlying activation of semantically similar words during word recognition. Attractor-based views uniquely predict that the effect of semantically similar words will depend on the topology of semantic space. Consistent with attractor-based models and challenging traditional accounts, words with many distant neighbors (moderately similar meanings) were processed faster, but words with many near neighbors (highly similar meanings) were processed slower. Eye-tracking data suggest that near neighbors show a transient peak in activation, but distant neighbors show relatively constant activation. These patterns are consistent with attractor dynamics and challenge traditional views of semantic structure and processing.

Contribution to SYMPOSIUM Meaning in Mind (Pexman)