Abstract
Modern studies of perceptual grouping or coherence are described and their theoretical implications discussed. the concept of “Prägnantz” or “good figure” and its modern derivatives are shown to be inadequate in accounting for grouping data. It is shown that grouping is not an all‐or‐none phenomenon in that contour grouping may occur with respect to some perceptual responses while independence is found for others. In general the notion that grouping arises from “emergent features” is not considered useful in accounting for the data in simple two‐line figures and an alternative view, based on cooperative neural connections, is proposed. Global emergent features are, however, shown to be a factor in grouping within some complex configurations. It is suggested that lateral relationships among contours may function independently of terminator relationships in influencing perceptual coherence. However, subjective contours form more readily across the collinear terminators of poorly grouped lines than across the same terminators of coherent lines. This “paradox” makes ecological sense. Finally, converging operations among different methods of measuring perceptual coherence are illustrated.

This publication has 18 references indexed in Scilit: