Ronald A. Finke

Creative Imagery : Discoveries and inventions in Visualization

Contents: Introduction. Visual Discoveries in Imagery. Creative Mental Synthesis. Creative Inventions in Imagery. Restricted Imagery Inventions. Preinventive Object Forms. Personal Inventions. Creative Concepts. Creative Implications.

A velocity effect for representational momentum

Observers were shown a rectangle at three orientations along a possible path of rotation. They were instructed to remember the third orientation in the sequence and then were presented with a rectangle at a fourth orientation that was either the same as, or slightly different from, the third orientation. Each observer tested was more likely to accept as “same” those distractors that were rotated slightly past the third orientation than those test items presented in the physically same position, and the degree of memory shift increased with increasing rate of presentation of the inducing displays.

Imagery, Creativity, and Emergent Structure

Recent advances in the field of creative cognition have helped to reveal the cognitive structures and processes that are involved in creative thinking and imagination. This article begins by reviewing recent studies of creative imagery that have explored the emergent properties of mental images. The geneplore model of creative cognition, which describes how preinventive structures such as creative mental images are generated and interpreted, is then discussed. In discussing this model and its implications, a distinction is made between aspects of creative imagery that reflect conscious, deliberate control and those that reflect the absence of such control, as illustrated particularly by the emergence of unanticipated structures in imagined forms. The intentional, structured qualities of creative thinking are then contrasted with its spontaneous, unstructured qualities. The article concludes by discussing the recent topics of chaotic cognition and creative realism and how they bear on the general issue of balancing structured and unstructured processes in creative endeavors.

Explorations of creative visual synthesis in mental imagery.

The mental synthesis of visual patterns has been previously studied by instructing subjects to imagine assembling the component parts in specific ways. We report two experiments that show that subjects can often discover recognizable patterns in imagery when the component parts are randomly chosen and are provided without instructions for how they might be assembled. On each trial, the subjects were given a set of three parts, consisting of geometric forms, lines, or alphanumeric characters, and were instructed to close their eyes and imagine combining the parts to make some type of recognizable pattern. They were successful in doing so on about 40 % of the trials. Many of these mentally synthesized patterns were strikingly creative, and few of them could be predicted, either by the experimenter or by the subjects themselves, simply by knowing what the parts were. On the contrary, most of the subjects reported that they had performed the task by imagining various combinations of the parts until a recognizable pattern “emerged.” These findings show that visual discoveries in imagery can be reliably induced under appropriate laboratory conditions.

Memory confusions for real and imagined completions of symmetrical visual patterns

In Experiment 1, subject ratings indicated that it was easier to imaginally complete half of a symmetrical form when the form was symmetrical about the vertical axis than when the form was rotated so that it was symmetrical about the horizontal axis. In Experiment 2, new subjects saw these same forms, some presented as wholes and some presented as halves. Compared with controls, subjects who were instructed to imagine the vertically symmetrical half forms as complete had more difficulty remembering which forms had been presented as wholes and which had been presented as halves. Imagery and control groups did not differ in ability to discriminate half from whole horizontally symmetrical forms. The results are consistent with predictions from Johnson and Raye’s (1981) reality monitoring model that discriminating memories of imagined and perceived events should be a function of the amount of cognitive operations included in the memory.

Mental imagery acuity in the peripheral visual field

Subjects made judgements of resolution on two small dots that they either imagined or acutally observed at horizontal and vertical positions away from the point of eye fixation. As the distance between these two dots increased, the size of fields of resolution in imagery increased, in proportion to increases in the size of fields of resolution in perception. For vivid imagers, fields of resolution in imagery were the same size as those in perception, whereas for nonvivid imagers, fields of resolution in imagery were smaller than those in perception. In addition, fields of resolution in imagery and perception were virtually identical in shape, exhibiting similar horizontal eccentricity and vertical asymmetry. Fields within which attention can be distributed in imagery were also measured by having subjects make judgements of resolution on pairs of dot patterns imagined simultaneously on opposite sides of the point of eye fixation. These fields were smaller than fields of resolution for images of single dot patterns and were circular, as opposed to elliptical. These results suggest that peripheral acuity in visual imagery is limited by the same types of neural constraints that limit peripheral acuity in visual perception.

Mental extrapolation and representational momentum for complex implied motions

A sequence of static displays implying consistent motions of a pattern induces distortions in memory for the last-observed appearance of the pattern. These memory distortions suggest that there is an internal analogue to physical momentum called representational momentum. Two experiments are reported comparing performance on tasks in which observers must remember the final display with those in which they are instructed to extrapolate the implied motions out to the next step in the sequence. In each experiment, the memory shifts were highly correlated with the actual rates at which the implied motions were extrapolated. In addition, the memory shifts were larger when the mental extrapolations occurred along the same directions of implied motion each time. These findings suggest that the memory shifts depend both on the rate at which the mental extrapolations are performed and on how consistently they are performed.

Directional scanning of remembered visual patterns

We report a set of experiments that helps to define the conditions under which mental image scanning may be used spontaneously for specific, practical purposes. Subjects were shown a dot pattern, followed by an arrow, and their task was to say whether the arrow was pointing at any of the previously seen dots. When no advance information was provided about the arrows location, reaction time for correctly verifying that an arrow was pointing at a dot increased linearly with increasing arrow-dot distance, and the subjects almost always reported scanning a mental image in order to make their judgments. However, when a cue for the arrows location was presented 2 sec beforehand, reaction time was uncorrelated with distance, and most of the subjects reported using an alternative strategy based on the determination in advance of correct directions from that location to the dots. When given only 1 sec of advance information about arrow location, most subjects reported using a combined image-scanning and advanced determination strategy, resulting in a reaction time function that increased only for the farthest distances. Because in each of these experiments instructions to form or to scan mental images were never given, these findings address the most common criticisms of the image-scanning paradigm.

Mapping the visual field in mental imagery.

We describe a simple psychophysical technique for measuring the size and shape of visual fields in mental imagery, and use this technique to compare fields in imagery and perception within which bar gratings of various spatial frequencies can be resolved. The first experiment demonstrates that the size of fields of resolution obtained when bar gratings are imagined decreases with increasing spatial frequency of the gratings, in the same way that the size of fields obtained when the gratings are actually observed decreases. This experiment also shows that imagery and perceptual fields are very similar in shape, both being elongated horizontally and extending farther below the point to which ones gaze is directed than above. The second experiment shows that experimental subjects are not able to anticipate the more subtle characteristics of these fields, such as the precise rate at which field size decreases with increasing spatial frequency--characteristics that most clearly reveal the very close correspondence between the imagery and perceptual fields. The third experiment shows that the size of the imagery fields can serve to discriminate between vivid and nonvivid imagers when the task of imagining the gratings is made sufficiently difficult. We interpret these findings as evidence that visual imagery involves the activation of mechanisms in the visual system that are specifically designed to process information about spatial frequency. In particular, we claim that constraints that these mechanisms impose on the resolution of high spatial frequencies in perception are also imposed in mental imagery, restricting how well fine details of an object can be imagined. We then discuss implications of these findings for theories about spatial-frequency analysis in vision and for the general issue of the functional value of mental imagery.

Mental extrapolation and cognitive penetrability: reply to Ranney and proposals for evaluative criteria

We propose three criteria for establishing that mentally extrapolated motions are impenetrable with respect to ones knowledge, beliefs, or expectations about the nature of corresponding physical motions, and we review recent findings on mental extrapolation and representational momentum that appear to meet these criteria. We also respond to some arguments recently proposed by Ranney (1989) and Hubbard and Bharucha (1988) that representational momentum is cognitively penetrable. We conclude that mental extrapolations are governed to at least some extent by the inherent properties of the underlying internal mechanisms.