Gordon H. Bower, working at Stanford University for over fifty years, was one of the most versatile and productive psychologists of the twentieth century. His early career focused on mathematical models of learning, collaborating with Estes and others on stimulus sampling theory and Markov models of paired-associate learning. His later work encompassed organizational factors in memory, narrative comprehension, and the pioneering study of emotion-memory interactions.
Mathematical Learning Theory
P(transition to learned state) = c per trial
P(correct | learned state) = 1
This Markov model predicts all-or-none learning:
errors before last error are Bernoulli with P = 1 - g
Bower's one-element model of paired-associate learning demonstrated that, for simple associations, learning occurs in an all-or-none fashion rather than gradually. Before the critical trial, performance is at chance; after it, performance is perfect. This model, derived from Estes' stimulus sampling theory with a single stimulus element, generated precise quantitative predictions about error distributions, trial-of-last-error distributions, and backward learning curves that were confirmed in extensive experimental tests.
Bower demonstrated that imposing hierarchical organization on to-be-learned material dramatically improves recall. His experiments on conceptual hierarchies showed that organized presentation could increase recall by factors of two to three over random presentation, providing key evidence that memory encoding involves active structuring of information rather than passive recording.
Mood-Congruent Memory
Bower's associative network theory of mood and memory proposed that emotions function as nodes in a semantic network. When an emotional state is activated, it spreads activation to associated concepts and memories, facilitating the encoding and retrieval of mood-congruent material. This framework generated the mood-congruent memory effect (better memory for material matching one's current mood) and mood-dependent memory (better retrieval when encoding and retrieval moods match).
Legacy and Impact
Bower trained dozens of students who became leaders in cognitive psychology, mathematical psychology, and cognitive science. His commitment to rigorous experimental methods combined with formal mathematical modeling exemplified the best traditions of scientific psychology. His work demonstrated that mathematical models of learning could make precise, testable predictions and that memory is an active, constructive process shaped by organization, meaning, and emotional state.