The Accuracy of Form Discrimination with Three Types of Tactual Input: Passive, Moving, and Active
J. J. Gibson and Susan Stanley, Cornell University
The World Wide Web distribution of James Gibson’s “Purple Perils” is for scholarly use with the understanding that Gibson did not intend them for publication. References to these essays must cite them explicitly as unpublished manuscripts. Copies may be circulated if this statement is included on each copy.
Lashley has stated that “the shape of an object impressed on the skin can scarcely be distinguished by simultaneous pressure, but the same shape can readily be distinguished by touch when traced on the skin with a moving point, or when explored by tactile scanning” (Lashley, 1951, p. 128). He is saying that active or moving touch is much more accurate for the perceiving of shape than is passive touch. Only the latter, however, has received much investigation. In the quoted statement Lashley was arguing for the importance of sequence or serial order in stimulation and response. He gives no reference for the assertion, however, and experimental evidence for it cannot be found in the standard works on cutaneous sensitivity. Ordinary observation suggests that the exploring fingers are better than the flat palm of the hand for registering the edges of a surface but the fact does not seem to have been demonstrated. This study is designed to supply evidence one way or the other.
The questions can be investigated, first, whether exploratory touching (more exactly, the stimulation resulting from the act of “feeling”) is more informative than simple cutaneous pressure in perceiving the form of a surface and, second, whether a moving depression of the skin is more informative than a static depression of the skin in perceiving the form of that depression. exploratory skin-stimulation is self-produced and is covariant with inputs from the joints, tendons, and muscles; moving skin-stimulation is not under the control of the observer and is not accompanied by other inputs. (Lashley’s tracing of an abstract form with a stylus on the skin will be deferred for later study). Both questions above are important for tactual perception. If it should turn out that a static pattern of cutaneous local signs is less effective for perception than a discontinuous “finger pattern” or a moving pattern, this would suggest at least that tactual form perception is not so simple as psychologists have thought. We need a comparison of the accuracy with which a given set of shapes can be discriminated when (1) the shape is pressed on the skin of the hand, (2) the shape is both pressed and rotated on the skin of the hand, and (3) the shape is touched by the exploring fingertips. The same set of objects should be presented to the S under all three conditions.
Apparatus and Method
As an indicator of discrimination we required S to match one of a set of six tactual stimulators with one of a set of six numbered outline drawings displayed in front of him. The tactual stimulators were thin metal strips, bent to produce edges of different shapes (“cookie cutters”). The corresponding drawings are shown in Fig. 1. All had a mean diameter of about 1 in. and the maximum dimension was about 1.25 in. Edge-forms of this size were chosen as being optimal for stimulating the hand. The attempt was made to select forms each of which differed equally from every other by visual inspection.
Fig.1 [NOTE: This figure was not shown on our copy of this Purple Perils – Eds.]
Preliminary work showed that this size and degree of similarity of the stimulators could be expected to yield neither zero nor 100% correct matches by any of the three conditions of stimulation.
For the experiment, S was seated opposite E at a table divided by a hanging cloth curtain. S was first shown the six stimulators and noted their correspondence to the drawings displayed at the top of the curtain. He was told that the experiment was concerned with tactual form perception. He would place his hand on the table, palm up, behind the curtain and one of the “cookie cutters” would be placed in (or just above) his hand. His task was to identify (by number) which of the six visual forms corresponded to the object touching his hand (or fingers). He was told that the objects would be presented in an unpredictable order, and that there would not be an equal number of presentations of each of the six objects (which was in fact the case). No preliminary practice was given.
In Condition P (passive) the stimulator was pressed downward into the palm of the hand by E for four seconds. In Condition M (moving) the stimulator was pressed on the palm in the same manner but was slowly and continually rotated in two or more 90 degree cycles for four seconds. In condition A (active) the S was told to cup his fingers upward. The stimulator was then placed on the fingertips and he was allowed to move the fingers in any manner he chose. This procedure often took longer than four seconds, but generally not more than five. No effort was made to present the tactual form in the same orientation as the visual form in the display. In all conditions, S was required to make a judgment or a guess at the end of the trial. No knowledge of results was given. The three conditions were presented separately but in an order counterbalanced for different Ss, so as to nullify a possible effect of practice on the comparison of conditions. Within each condition the six forms were presented in a different random order, each form appearing at least five times, with some extra unrecorded trials t o render frequency unequal.
Twelve Ss were used. Each made 30 scored judgments (5 x 6 objects) under each condition for a total of 90 judgments per S.
Pooling all 1080 judgments, 61% were correct. A chance level of judgments would be 1/6 or 16.7% (with our method). The modes of tactual stimulation, taken together, gave the subjects a considerable amount of information for distinguishing the objects.
But the three modes of stimulation were by no means equal in the amount of information registered. For the passive condition 49% of the judgments were correct, for the moving condition 72%, and for the active condition 95%. It is unlikely that these differences in accuracy of judgment could have occurred by chance. The results suggest, first, that a static pattern of cutaneous pressure is a less effective stimulus for perceiving form than a rotating pattern on the same area of skin. They suggest, second, that a static pattern of cutaneous pressure on the palm is less effective for perceiving form than a complex of changing pressures on the fingertips, the changes being produced by the S himself and being covariant with kinesthetic inputs.
There was a trend for the accuracy of the identifications to increase through the series of 90 judgments, across the counterbalanced order of conditions of stimulation. This suggests that the Ss, although never told whether judgments were correct or incorrect, were somehow learning to register certain differences in cutaneous stimulation not ordinarily given much attention. They reported attempts to analyze the impressions in terms of “angles”, and of “lines” and “curves ” (although the latter were hard to distinguish).
If these exploratory results are confirmed by other experimenters (and the experiment is easy to repeat) they demand some new thinking about tactual form perception. A fixed pattern of local signs on the skin is less determinative of “form” than a pattern which changes in time. And a seemingly chaotic pattern of changing fingertip pressures is the most determinative of all. Different Ss in this experiment were observed to move their fingers around the object in different ways; the same S would not explore in the same way on different occasions; moreover the number of fingertips applied to the edges might vary between all five and only one. Evidently a perception of form arising from the skin does not necessarily consist of or depend on a chain of adjacent cutaneous local signs. It must often, and more typically, depend on something that remains invariant during variations of local pressure. During rotation of a form on the skin, the anatomical loci of stimulation change, but certain distances, proportions, and other relations among these changing loci do not, i.e. they remain invariant. These distances, proportions, and other relations can reasonably be considered stimuli for the skin. If so, this would explain why both the form and its motion can be picked up. During self-produced stimulation of the skin of the fingers by contact with a rigid object, the anatomical loci of stimulation change even more radically. But these changes are precisely linked to the movements of the fingers. The presence or absence of stimulation at any finger is linked to the relative posture of that finger, and this is specified by input from the joints. Consequently in this case also there are invariants over time and these are specific to the form of the object. Introspectively, as one’s fingers play over the edges, the local sensations of pressure are almost impossible to detect or describe; what comes easily to attention are the feelings of “angle” and “line” and of relative separation. It is these invariant relations, we suggest, which constitute stimuli for the normal operation of the sense of touch.
On this theory, objects are tactually distinguished from one another on the basis of the distinctive invariants of cutaneous and articular stimulation, not primarily on the basis of cutaneous “form” as heretofore conceived. Tactual edges can be matched with visual outlines not so much because the same “form” is imposed on the skin and on the retina as because the same invariants of proportion and relation are common to the tactual and the visual system. The eyes, like the hands, are exploratory, and the retinal image, like the cutaneous image can undergo rotations, transpositions, and transformations without losing its specifying capacity. The ability of an individual to localize a punctate stimulus on the skin has been considered the sensory basis of tactual form perception. These results cast doubt on that assumption.