Studies of visual synthesis: integration of fragments into forms.

On each of a series of trials, subjects indicated whether two sequential visual forms were the same or different. The first form was presented on some trials as a whole within a square frame and on others in fragments--with its perimeter distributed over two or three frames. The subject was instructed to visualize the first form as a whole in any case, and study time (ST) for the form was recorded. The same/different reaction time (RT) for the second, test, form (always presented in one frame) was also recorded. Experiment 1 revealed that variables affecting ST had little effect on RT and indicated that subjects can process fragments of forms in parallel. Experiment 2 demonstrated that the representation encoded from fragments differed from one generated from long-term memory. Whereas the first experiments used wholes as test stimuli, in Experiments 3 and 4 on some trials, fragments were used. In Experiment 3, RT was facilitated when a test fragment coincided with an intact fragment of the first stimulus. This indicated that first stimuli were not encoded as complete wholes. In Experiment 4, irregular forms were used as stimuli, and the RT data departed from predictions of the parallel model. Taken as a whole, the results place constraints on the codes produced by constructive processes acting to synthesize fragments into wholes.