Long-term semantic priming: a computational account and empirical evidence

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Psycholinguistics

Background:

• Semantic priming is traditionally understood as a short-lived phenomenon.
• Attractor neural network models offer a new perspective on long-term word priming.
• Previous research has not demonstrated semantic priming over extended lags.

Purpose of the Study:

• To propose and test a model of long-term semantic priming in attractor neural networks.
• To investigate conditions under which semantic priming effects can persist over long lags.
• To challenge the conventional view of rapid decay in semantic priming.

Main Methods:

• Computer simulations of attractor neural networks.
• Three experiments involving semantic-decision and lexical-decision tasks.
• Varying numbers of primes per target and list lengths (lags of 0, 4, and 8 items).

Main Results:

• Long-term semantic priming was observed in semantic-decision tasks, but not lexical-decision tasks, when multiple related primes were used.
• The long-term semantic priming effect was replicated for semantic decisions with single primes.
• Semantic priming was demonstrated across extensive word lists and lags, significantly longer than previously reported.

Conclusions:

• Semantic priming can persist over much longer intervals than traditionally assumed.
• Sufficient engagement of semantic-level processing is crucial for observing long-term priming effects.
• This study provides the first empirical evidence for semantic priming spanning numerous intervening items.