Categorical processing of visual stimuli in relation to geometrical, graphemic, or lexical context.

Categorical perception was first demonstrated in studies of speech sounds (Liberman, Harris, Hoffman, & Griffith, 1957). The present work employed visual stimuli to explore categorical responding in relation to the context in which the stimuli were embedded. The target stimulus was a vertical line whose length was varied from 20 to 31 min (approximately) in steps of 1.2 min. Experiment 1 examined the effect of a geometrical context on the subjects' ability to discriminate between pairs of lines. The context improved performance, but produced no evidence of categorical responding. In Experiment 2 a graphemic context depressed performance, but failed to show clear evidence of categorization. By contrast, strong evidence of categorical responding was obtained in Experiment 3, in which the graphemes used in Experiment 2 were embedded in meaningful words. From this pattern of results it is argued that categorical responding is reflective not of relatively peripheral perceptual activity, but of higher-order decision processes.