gibson (1963) the useful dimensions of sensitivitywexler.free.fr/library/files/gibson (1963) the...

THE USEFUL DIMENSIONS OF SENSITIVITY J

JAMES J. GIBSON

Cornell University

WHAT I am going to talk about is therelation of sensing to perceiving. Wehave all believed that we understood

the process of sensation fairly well and that onlythe process of perception gave us difficulties. ButI am going to suggest on the contrary that astraightforward theory of perception is possibleand that it is our understanding of sensationwhich is confused.

First let us make sure that there is really aproblem in how to treat sensing and perceiving.Some psychologists now maintain that there is nodifference between them in fact. The distinctionhas broken down; they say it has no validity andwe should forget it. I think that what they meanis this. An individual can make discriminationsin many ways. We can say either that he issensitive to many variables of stimulation or thathe can experience many kinds of differences be-tween things but what has importance, the argu-ment goes, are only the facts of discrimination,not whether they are called sensory or perceptual.There is something valid in this argument. Iwould call it the experimentalist's position—stickto the facts and cut the cackle! It is enoughto determine just what differences an animal, achild, or a man can respond to and what othershe cannot. This limited aim of psychology mightbe called simple psychophysics (not metric psy-chophysics) and it is good experimental science.But it provides no explanation of how the in-dividual keeps in touch with the environmentaround him. The problem of perception, then,the problem of contact with the environment, stillremains.

The variables of sensory discrimination areradically different from the variables of perceptualdiscrimination. The former are said to be dimen-sions like quality, intensity, extensity, and dura-tion, dimensions of hue, brightness, and satura-tion, of pitch, loudness, and timbre, of pressure,

1 This work was supported in part by the Office ofNaval Research under Contract 401 (14) with CornellUniversity.

warm, cold, and pain. The latter are dimensionsof the environment, the variables of events andthose of surfaces, places, objects, of other animals,and even of symbols. Perception involves mean-ing; sensation does not. To see a patch of coloris not to see an object. To see the extensityof a color is not to see the size of an object, noris seeing the form of a color the same as seeingthe shape of an object. To see a darker patchis not to see a shadow on a surface. To see themagnification of a form in the field is not to seean approaching object, and to see the expansionof the whole field is not to observe one's ownforward locomotion. To have a salty taste isnot to taste salt, and to have a certain olfactoryimpression is not to smell, say, a mint julep. Tofeel an impression on the skin is not to feel anobject, nor is having sensations of strain andpressure to feel the weight of an object. To feela local pain is not to feel the pricking of a needle.To feel warmth on one's skin, is not to feel thesun on one's skin, and to feel cold is not to feelthe coldness of the weather. To hear sound isnot the same thing as to hear an event, nor isto hear an increasing loudness to hear the ap-proach of a sounding object. Finally, let us notethat having a difference of sound sensation in thetwo ears is by no means the same as to hearthe direction of a sound. The last case is in-structive, for we do not in fact have such binauraldifferences in sensory experience but we do localizesounds.

Having sensations is not perceiving, and thisfact cannot be glossed over. Nevertheless, perceivingunquestionably depends on sensing in some mean-ing oj that term. That is, it depends on sensitivityor the use of the sense organs. To observe, onemust sense. The question I wish to raise iswhether or not it is true that to observe onemust have sensations.

I realize that any inquiry into the relation ofsensing to perceiving raises the ghosts of formi-dable men. It is disconcerting to feel that Locke,Berkeley, and Hume are looking over one's

1

AMERICAN PSYCHOLOGIST

FIG. 1. Momentary cross section of the light entering a human eye.

shoulder, or that Kant and two generations ofMills are raising their eyebrows. A perceptualtheorist can get into staggering muddles, and hedoes well to be cautious. Nevertheless, I have aset of hypotheses to propose and you may judgeit both for internal contradictions and for con-formity with the facts. My first suggestion, thegeneral thesis, is that the useful dimensions ofsensitivity are those that specify the environ-ment and the observer's relation to the environ-ment. There are other dimensions of sensitivitywhich do not specify such facts and relations, but

FIG. 2. Longitudinal section of the effective sector ofan optic array.

they are not useful in this way, being only in-cidental to the activity of perception.

A whole set of correlative hypotheses go alongwith this radical thesis. They need to be under-stood before it begins to have plausibility, and thetheory should be considered as a whole. Thefacts of sensory psychology and sense physiologyare so varied and voluminous that it is not easyto stand back and take a fresh look at theevidence. Moreover, each of us is apt to havehis own private opinion about the data of hissenses. But if you will suspend belief in thestandard doctrine of sensation and question yourfavorite introspections, I hope to convince youthat the explanation of sense perception is notas difficult and roundabout as it has always ap-peared to be.

Consider first the puzzle of perceptual con-stancy. I will not attempt to review the experi-ments measuring the tendency toward constancywhich are limited to vision and which, in 3113'case, are indecisive. Instead I will point to thegeneral evidence for an invariance of perceptionwith varying sensations. This invariance appearsnot only in vision but also in other senses, notablythose excited by mechanical energy, hearing, andtouch. The paradox of constancy—the "distalfocusing of perception" as Egon Brunswik (19S6)

USEFUL DIMENSIONS OF SENSITIVITY

put it, is more than a matter of color, size, andshape constancy; it is the heart of the problemof useful sensitivity.

Figure 1 is a picture of a patchwork of visualsensations. Note that it is a cross section of awide-angle cone of light rays which might entera human eye, the left eye in this case. Figure 2is a longitudinal section of such a wide-range cone.It is stationary and momentary as represented inthe picture, but whenever the eye moves to a newfixation point it will take in a new cone of rays.At that station point in the room there exists acomplete optic array of available stimulation, thearray being sampled and explored by new fixations.Figure 3 shows that if the man moves, instead ofhis eye moving, the pattern of the entering arrayis transformed, that is, every patch of color in thearray changes form, and the patchwork as a wholeis altered.

All this is simply the outcome of the laws ofambient light, or what may be called opticalperspective (Gibson, 1961). The laws of pictorialperspective with which we are more familiar area special case involving the sheaf of rays at apicture plane (Gibson, 1960b). Figure 4 is anillustration of so-called linear perspective on apicture.

The sensations of the visual field shift withevery movement of the eye, and transform withevery movement of the head. But, the perceptionof the room remains constant throughout. Thereis invariance of perception with varying sensations.

There are two kinds of seeing, I argue, oneresulting in the experience of a visual field andthe other in the experience of a visual world(Gibson, 19SO). The field is bounded; theworld is unbounded. The field is unstable; the

FIG. 3. Perspective transformation of the patchwork of anoptic array due to change of viewpoint.

FIG. 4. The special case of a ray sheaf at a hypotheticalpicture plane.

world is stable. The field is composed of adjacentareas, or figures; the world is composed of surfaces,edges, and depths, or solid objects and interspaces.The field is fluid in size and shape; the world isrigid in size and shape. As pure cases, they aredistinct, although in many experimental situationsthe observer gets a compromise experience be-tween the two extremes. However, these experi-mental situations are seldom ones in which theobserver is free to explore a complete optic arraywith his eyes, and are never ones in which he isallowed to move about so as to obtain a series orfamily of perspectives.

The visual field ahead of the observer duringlocomotion expands in a sort of centrifugal flowgoverned by the laws of motion perspective. Thevisual world during locomotion is phenomenallyquite rigid. Sensation varies but perception isinvariant. To be sure, the observer sees hislocomotion. The expansion of the field aheadspecifies locomotion. This suggests a strange andradical hypothesis—that the visual sensation inthis case is a symptom of kinesthesis, havingreference to the self instead of the world, and thatit has nothing to do with the visual perception ofthe world (Gibson, 1958).

Another case is that of the perception arousedby the perspective transformation of a silhouettein an otherwise empty field of view. As an experi-ment, this does not require a panoramic motionpicture screen, and it can be carried out in alaboratory. There results a perception namedstereokinesis, or the kinetic depth effect, or simplyrigid motion in depth (Gibson & Gibson, 1957).Behind the translucent screen in such experiments,at an indefinite distance, there appears a virtualobject moving in space. The form of the silhouettechanges; the form of the phenomenal object re-mains invariant. The observer can see a changeof form if he attends to the flat screen, but what


he spontaneously reports is a rigid object. Ordi-narily the transformation is seen as motion of theobject, not as a sensation.

Another example is the familiar one that thecolor of the surfaces of the environment, includingthe white to black series, do not change as theillumination goes from brilliant to dim. Thecorresponding sensations, however, the film colorsobtained by seeing a surface through an aperture,vary widely with illumination. The perception ofwhiteness is quite a different matter from thesensation of brightness. With the available stim-ulus of a complete optic array, the ambient lightreflected from a whole layout of surfaces, one candetect the actual physical reflectance of eachsurface. The absolute luminous intensity of acolor patch determines the sensation of bright-ness, but only if it is taken in isolation.

Finally, 1 remind you of the difference be-tween the binocular sensations of objects in depthand the binocular perception of the depth of ob-jects. When one attends to his visual sensationsone can notice the doubling or diplopia of imagesin the field of view; crossed diplopia from here tothe fixated object, and uncrossed diplopia beyondthat point. This doubling changes with everychange in convergence, especially as we look toor away from what our hands are doing. Weought to see nothing as single except what lieson the momentary horopter. But of course wesee everything as single, that is, we perceive it so.There is a phenomenal unity of each objectdespite an ever-varying doubleness of its sensation.

Auditory perception, we say, is based on adifferent mode or department of sense from visualperception. But the paradox of invariant per-ception with varying sensations holds nevertheless.Consider those very peculiar and special sounds,the phonemes of speech. They are acousticallyanalysable, it is true, in terms of intensity, fre-quency, and the frequency spectrum, but theirdistinctive nature consists of higher-order variableswhich are now beginning to be specified. Phonemesare the same at quite different levels of pitch andloudness, and hence are phenomenally constantfor the voices of men, or women, or children.Speech cannot only be voiced; it can also bemurmured, shouted, whispered, or sung. It canbe emitted in falsetto, or even by a sort of whis-tling, without completely destroying the distinctivefeatures which define the phonemic units of speech.

They are invariant with changes of auditorysensation.

Consider also the hypothetical sensations that ahearer would get during auditory localization—the different sense impressions or sense data fromthe two ears. The main stimulus differences areones of intensity and time of onset. As we knowfrom the experiments of Wallach (1940) andothers, the hearer turns and tilts his head fromside to side, as if exploring, when he hears anunseen event. For a repeating sound, this meansthat the relative loudnesses and onsets of sensa-tion are continually changing during the headturning. But the perception is that of a fixedor constant direction of the sound in space. Asa matter of fact there is no evidence to showthat any man or animal ever heard the changesof binaural sensations when turning his head.There is no awareness of such a flux. Binauraldisparity never becomes conscious as binoculardisparity can (Rosenzweig, 1961). I prefer tobelieve that the binaural mechanism is an activesystem which responds to disparity and tends toreact by nullifying it, that is, by pointing thehead toward the source of sound. The systemresponds to the sound field in the air, and weare misled when we consider only the wave trainentering each ear separately.

So much for hearing. It is the sense of touch,so-called, that provides the clearest examples ofthe invariance of perception with varying sensa-tions. In the last two or three years I have beenrunning a series of experiments to test the limitsof what an observer can do by touching or feelingwithout vision, that is, to discover what he candetect or discriminate about surfaces, edges, inter-spaces, objects, and motions in the neighborhood ofhis body. In these experiments we can comparethe classical results obtained with passive punctatestimulation of the skin (intensity, locus, duality,and motion of a cutaneous impression or a pat-tern of such impressions) and the results ob-tained with the self-produced stimulation oftouching. In general, an observer can perceivethe properties of an object by active touch withquite surprising success. So also, of course, cana blind person. The following results come froma long series of observations (Gibson, 1962).

Rigidity. For example, when pressing on arigid object with a finger, or squeezing it withthe hand, there is an increase of sensation and

USKFUL DIMENSIONS OF SENSITIVITY

then a decrease, or usually a flow of changingintensities. The perception, however, is of a con-stant rigidity of the surface. One simply feelsthe object. The impression on the skin as suchis hard to detect. When one is touched by thesame object instead of touching it, however, thevariation of intensity is easy to detect. An ob-server can distinguish correctly between two pro-tuberant surfaces, one rigid and the other yielding,when he presses them, but not when they arepressed on his passive skin.

Unity. When feeling one object between twofingers, only one object is felt, although twoseparated cutaneous sensations occur. This is asurprising fact when you consider it. The differ-ent local signs of these impressions have seeminglydropped out of the experience. The result is thesame whether the object is held with two, three.four, or five fingers; the multiplicity of impressionson the skin has no effect on the perception ofspatial unity of the object. It can be held bytwo hands and still be one object. It can befelt bv manv combinations of all 10 fingers, in

rapidly changing combinations, and the percep-tion of the object is all the better for it.

Stability. Active touch is exploratory and theobserver tends to slide his finger over a corneror protuberance of a hidden object. The im-pression is then displaced over the skin and afeeling of tactile motion would be expected tooccur. But the object is perceived to be stationaryin space, and the tactile motion is not noticed.The perception is stable although the sensation ismoving.

Weight. When one holds or lifts an object, thejudgment of its weight is easier than when it isallowed simply to press downward against the skinof the supported resting hand. In active lifting,a whole set of additional inputs is involved. Be-sides the end organs of the skin and the deepertissue, the receptors of the finger joints, wristjoints, and arm joints are excited, and the wholeneuromuscular feedback system of the arm isactivated. The flux and array of pressure sensa-tions and articular sensations from a dozen or sojoints ought to be of bewildering complexity. It

FIG. S. The visible object and the tangible object.


server looking at one object and feeling another.

probably would be if introspection could detectall that goes on in hefting a weight. But whatthe observer perceives is the mass of the object,unchanging despite the changing sensations. Aweight comes to be as well or better perceived, infact, when the object is shifted back and forthfrom one hand to the other. Something invariantemerges from this seeming mishmash of excita-tion. The perception is equivalent to that whichaccompanies the controlled and isolated sensoryimpressions of the standardized weight-liftingexperiment.

Shape. A method of investigating the percep-tion of unfamiliar shape by active touch is illus-trated. Figure 5 shows an object behind a curtainwith another identical (or different) object visibleon a turntable. Figure 6 shows an observerfeeling the object with both hands and judgingwhether the visible one is the same or different.Alternatively he might be required to match itwith one of 10 visible objects, as shown in Figure 7.The degree to which these sculptured free formsdiffer among themselves is illustrated in Figure 8,where two are identical and the third different, and

also in Figure 9, a view from the side. All theseobjects have six protuberances in front, and arounded back. The ordinary observer, after verylittle practice, can distinguish among the tangibleobjects and match them to their visible replicaswith little error (Caviness & Gibson, 1962).

The haptic system of the exploring hand is sensi-tive to the variables of solid geometry, not thoseof plane geometry. It gets nothing of a flat pic-ture, but it gets a great deal of the shape of asolid object. The hand can detect all of thefollowing properties: the slant of a surface, theconvexity or concavity of a surface, the edge orcorner at the junction of two or more surfaces,and the separation of two edges, as our experi-ments demonstrate. Now it has always been as-sumed that the skin must be analogous to theretina—that it is a sensory mosaic which registersthe form or pattern of the receptors excited. Theskin and the retina can, in fact, do so when theyare passively stimulated, and this has been takento be their basic or sensory function.

If the cutaneous form sense is the basis for thefeeling of objective shape, however, an impossible


FIG. 7. The 10 sculptured objects used.

FIG. 8. Close-up of two identical and one different object.


FIG. 9. Side view of three objects.

paradox arises. The series of cutaneous pressurepatterns with a pair of exploring hands is some-thing like that of a kaleidoscope; it seemingly hasno rationale, and no single pattern is ever likethe shape of the object. Nevertheless, from theinputs of the skin and the joints together, fromthe sensory system if not from the sensations, aremarkably clear perception of shape arises. Thephenomenal shape of the object is invariant al-though the phenomenal patterns of sense datafluctuate and vary from moment to moment.

Conclusion. From all these facts of vision,hearing, and touch we ought to conclude thatsensations are not the cause of perceptions. Thisis a strange statement. But I am willing to drawthis conclusion. Conscious sensory impressionsand sense data in general are incidental to percep-tion, not essential to it. They are occasionallysymptomatic of perception. But they are not evennecessary symptoms inasmuch as perception maybe "sensationless" (as for example in auditorylocalization). Having a perception does not entailthe having of sensations.

The difficulty in accepting this conclusion ishow to explain sense perception unless by way ofsensations. But there is a way out of this dif-ficulty, and that is to distinguish two meaningsof the word "sense." Sensitivity is one thing,sensation is quite another.

The first meaning refers to the effects of stimu-lation in general. The second refers to consciousimpressions induced by certain selected variablesof stimulation. We can now assert that in thefirst meaning sensory inputs are prerequisite toperception, but that in the second meaning sensoryimpressions are not prerequisite to perception. Inother words the senses are necessary for percep-tion but sensations are not. In order to avoidconfusion it might be better to call the sensesby a new term such as esthesic system. We canthen distinguish between sensory perception andsensory experience, between perception as a resultof stimulation and sensation as a result of stimu-lation. The variables of stimulation that causethe first must be different from those that causethe second. Likewise the dimensions of sensitivity


to informative stimuli must be different fromthose to uninformative stimuli.

How is the invariance of perception with vary-ing sensations to be explained? By higher-ordervariables of stimulation which are themselvesinvariant, and by the sensitivity of esthesic sys-tems to such invariant information. This kind ofsensitivity is useful to animals. It may be innate,or acquired, or a little of both—that is a questionfor experiment. We can study it directly. We donot have to solve the puzzle of how there can beinvariance of perception despite varying sensa-tions. We do not have to inquire how sensationsmight be converted into perceptions, or corrected,or compensated for, or how one set of sensationsmight reciprocally interact with another set. Ifthe sensations are disposed of, the paradox ofperceptual constancy evaporates. Clearly thehypothesis of stimulus invariants is crucial forthis explanation, and I will have to return to itlater. Note that with this approach, a seeminglyuseful tool of experimenters, the index of con-stancy, loses its meaning. It ceases to be a meas-ure of perceptual achievement. The supposedbaseline of this ratio, the "retinal" size, shape, orbrightness, cannot be used in a computation ofthe achievement if it is not the basis of the percep-tion. It falls in a different realm of discourse, andit simply is not commensurable with perceptualsize, shape, or brightness.

What are sensations? We might well pauseat this stage to consider what is being discarded.Just what are these experiences that the perceptualtheorist should no longer, appeal to? I suggestedat the beginning that our understanding of sensa-tions has always been obscure. The reason forthis, I think, is that the term sensation has beenapplied to quite different things. Let us examinethe various meanings of the word to be found inphilosophy, psychology, and physiology.

1. The theoretical concept. Theories of per-ception, as already noted, have always assumedthat sensations were the necessary occasions ofperception; that they were entailed in perceiving.This is precisely the assuption that is being chal-lenged by my distinction between sensation andsensitivity. Its plausibility comes only from theevidence that stimuli are the necessary occasionsof perception. I shall argue that none of the

kinds of experience which have been called sensoryrequires this theoretical assumption.

2. The experimentalist's concept. In psycho-physical experiments the variables of sensationhave been taken to be correlates of the variablesof physical energy which the experimenter couldapply to his observer. In the past, the latterhave tended to be those which were fundamentalfor physics proper, and which were controllableby borrowing the instruments of optics, acoustics,and mechanics. The favorite physical variableswere intensity and frequency for wave energy,along with simple location or extension, and timeor duration. But these dimensions of stimula-tion have little to do with the environment. Theyare fundamental for physics but not necessarilyso for sense organs. The dimensions of availablestimulation in a natural physical environment areof higher order than these, being variables ofpattern and change. We are beginning to be ableto control these natural stimulus variables. Notealso that the stimuli of classical psychophysicsare applied to a passive observer by an experi-menter whereas the stimuli in perceptual psycho-physics are obtained by an active observer (al-though the opportunities for obtainable stimulationare provided by the experimenter). The experi-ences resulting from these two situations are apt tobe different, as the experiments on active touchdemonstrate.

3. The physiological concept. The earlyphysiologists discovered the receptor elements ofthe sense organs and assumed that these cells(rods, cones, hair cells, etc.) were the units of areceptor mosaic. Hence a sensation was taken tobe a correlate of a single receptor, that is, theend organ of a nerve fiber. But we are nowfairly sure, after recording from single fibers withmicroelectrodes, that the functional units of asense organ are not the anatomical cells, butgroupings of cells. It was also assumed, afterJohannes Miiller, that a specific mode or qualityof sensation corresponded to any given nerve orfiber. But this generalization too, can no longerbe supported since, for one thing, the same fibercan participate in different groupings and havethereby different receptive functions. WhenMiiller insisted that the mind had no directcontact with the environment but only with the"qualities of the sensory nerves," he was con-

10 AMERICAN PSYCHOLOGIST

fusing sensitivity with sensation. He assumedthat the function of the senses was to providesensations. He was right, surely, to maintainthat perception depends on stimulation but wrongto maintain that it depends on the consciousqualities of sense. A sense organ has to bedefined as a hierarchy of functional groupings ofcells, and they are not always adjacent ana-tomically.

4. The analytic concept. The attempt to reduceconsciousness to its lowest terms by introspectionculminated in Titchener. A sensation was takento be an irreducible experience not analysableinto components—a simple datum. It is fair tosay that the attempt failed. Sensations as com-bining elements are no longer advocated, althoughthe elegance and force of the structuralist programwas such that traces of it are still influential inpsychology. Conscious perceptions cannot al-ways be reduced to conscious sensations, as theGestalt theorists have shown. It is clear thatsensation in this meaning of the term is notprerequisite to perception.

5. The empiricist's concept. According toLocke and all the thinkers influenced by him,sense impressions are the original beginnings ofperceptual experience prior to learning. They areinnate, and pure sensations are had only by thenew-born infant. They are without meaning andprobably without reference to external objects.They are data for thought (or inference, or inter-pretation, or association, or other kinds of learning,either automatic or rational). What they arelike has been the subject of endless inquiry, andthis explains our strong curiosity about the firstvisual experiences of the congenitally blind afterthe operation for cataract. The theory thatoriginal experience was composed of sensationshas always appealed to psychologists because theavailable alternatives, nativism and rationalism,implied either a faculty of perception or a facultyof reason. But we can reject sensation as theoriginal beginning of perception and accept usefulsensitivity as something present from the start oflife without being driven into the arms of facultypsychology. We can also avoid the naggingdifficulty that infants and young animals (and thecataract patients, in my opinion) do not, on theevidence, seem to have the bare and meaninglesssensations that classical empiricism says theyshould have.

6. The concept of an experience with subjectivereference. There is still another possible meaningof the term sensation. It is the meaning used insaying that a stomach-ache is sensory rather thanperceptual. The same could be said of an after-image as compared with an object, for it seems torefer more to the observer than it does to theouter world. In cases of passive tactual experience,the observer can feel either the impression on theskin as such or the object as such, depending onhow he directs his attention. It is as if thephenomenal experience had both a subjective poleand an objective pole. Pain is ordinarily subjec-tive (although there may be some objective refer-ence, e.g., a pin) and vision is ordinarily objec-tive (although there can be a subjective aspect,e.g., dazzle), but all senses, in this view of thematter, carry both subjective and objective in-formation. The observer's body, as well as hisenvironment, can always be noted, together withthe relation between them. The body and theworld are different sources of stimulation; there ispropriosensitivity as well as exterosensitivity.Sherrington was wrong only in supposing thatthere are separate proprioceptors and exteroceptors.All organs of sensitivity, I suggest, have this dualfunction.

Note that sensation considered as the subjec-tive pole of experience is quite different from theother meanings of sensation. This is not theprovider of data for perception or of messagesor elements, nor is it the innate beginning ofperception. This is a legitimate and useful mean-ing, but not the classical one—the basis of theexperience of the external world.

Conclusion. Having examined the various kindsof experience that have been called sensory, Iconclude that no one of them is required as thenecessary occasion of perception. Several of themdo undoubtedly occur in a man who introspects,or who serves as subject in an experiment, butthe explanation of perception can dispense withall of them.

RECONSTRUCTION OF A THEORY OF PERCEPTION

If sensitivity is distinguished from sensation,and if perception depends on the former but notthe latter, we will have to make a fresh start onthe explanation of perception. We will have todiscard many cherished doctrines and formulas(like separate and distinct modalities of sense),

USEFUL DIMENSIONS OF SENSITIVITY 11

to clarify and find words for new things (likestimulus patterns and transformations), and todevise new experimental methods (such as how tocontrol stimulus information instead of traditionalstimuli). What are the requirements of a theoryof perception not mediated by sense data?

Obviously it will have to show that sensitivity,with or without accompanying sensations, isadequate for all the manifold properties of per-ception (Gibson, 1959). It will have to showthat the afferent inputs to the nervous system ofa child or a man are rich enough to explain thedegree to which he is aware of the world (butthe inputs are taken to be those of active systems,not passive receptors or even sense organs). Itwill have to show that there is information inavailable stimulation (but the potential stimuliare taken to be limitless in variables of higherorder). It will have to show that there are con-stants in the flow of available stimulation inorder to explain constancy. It will have to showthat these invariants in the ambient light, sound,and mechanical contact, do in fact specify theobjects which are their sources—that somethingin the proximal stimulus is specific to the distalstimulus (Gibson, 1960a). It will have to sug-gest how these invariants can be discovered bythe activity of selective attention (but there arehints of such a mechanism in what we alreadyknow about sense-organ adjustments, so-called,and about the selective filtering of higher nervecenters). It will have to explain propriosensitivity(self-perception) along with exterosensitivity (ob-ject perception), but without appealing to theoversimplified doctrine of a special sense ofkinesthesis.

Moreover, the theory will have to explain all theobservations and experiments of past generationswhich seem to make it perfectly evident that theobserver contributes meaning to his experience,that he supplements the data, and that significanceaccrues to sensation. I have assumed limitlessinformation in available stimulation from thenatural environment. Therefore, the explanationmust be that the experimenter has limited theavailable information in all such experiments, orelse that, in a natural situation, the availablestimulus information is impoverished, as by dark-ness or a disadvantageous point of view. Psy-chologists are accustomed to use stimulus situationswith impoverished, ambiguous, or conflicting in-

formation. These have been devised in the hopeof revealing the constructive process taken tocharacterize all perception. In these special situa-tions there must indeed occur a special process.It could appropriately be called guessing. But Iwould distinguish perceiving from guessing, andsuggest that we investigate the first and try tounderstand the second by means of corollariesabout deficient information.

The theory will have to provide an explanationof illusions, not only the optical ones but thoseof all the other channels of sensitivity. Thepostulates of stimulus information and stimulusecology, however, suggest ways in which thevarious illusions can be, for the first time, classi-fied into types and subtypes of misperception,with the reasons therefore. A proper descriptionof the information in an optic array will necessarilyinclude a description of the information in a picture,and the ambiguous, conflicting, equivocal, or mis-leading information that can be incorporated ina picture. Note that illusions will be treated asspecial cases of perception, not as phenomenawhich might reveal the laws of the subjectiveprocess of perception.

Finally, the theory will have to be consistentwith the known facts about social perception andall the information that has accumulated aboutthe perception and learning of symbols and words.Here, you may think, a sensitivity theory of per-ception must surely fail. Even allowing thatphysiognomic and expressive character may havesome basis in complex stimulation, words canhave no meaning except that supplied by theperceiver. But this objection, cogent as it maysound, entirely misses the point. Once it is grantedthat stimuli may carry information, or havemeaning, the whole theory of meaning is revolu-tionized, and we have to make a new start on it.Once it is granted that a child or a man candevelop sensitivity to the invariants of the ecolog-ical stimulus environment it is no great step toadmit that he can also learn to respond to theinvariants of the social and the symbolic environ-ments. The laws by which stimuli specify eventsand objects are not, of course, the rules or con-ventions by which chosen events or objects standfor others, but both are lawful. If animals andchildren can register perceptual meanings it is notsurprising that children and adults can go on toregister verbal meanings. However, just as the


the modification of stimulation by reactionsof the exteroceptive sense organs

the modification of reactions by stimulationof the proprioceptive system

exploration with fingers-t

^savoring and sniffingwith mouth and nose

cocking head orpricking up ears

focusing, fixating,converging andpursuing with eyes

tendon and jointstimulation

inner ear stimulation -

visual stimulation.

FIG. 10. The feedback loops for exploring stimulation and those for controlling behavior. (The angular lines repre-sent physical action; the curved lines represent neural action.)

child does not first have a repertory of sensationsand then attach meanings, so also he does notfirst hear a vocabulary of words and then attachmeanings.

Role of Attention in Perception

An entirely different picture of the senses hasemerged. For this to happen, we had to supposethat their sole function was not to yield sensa-tions. Instead of mere receptors, that is receiversand transducers of energy, they appear to besystems for exploring, searching, and selectingambient energy. The sense organs are all capableof motor adjustment. Figure 10 is a diagramwhich supplements and alters the usual stimulus-response diagram. It shows on the left the modifi-cation of stimulation by reactions of the exterocep-tive system, and on the right the modification ofreactions by stimulation of the proprioceptivesystem. The latter is familiar nowadays under thename of feedback, that is, the neural loopsessential for the control of behavior. But theloops on the left are just as essential as those onthe right. The organism has two kinds of feed-back, not one. There are two kinds of action,in fact, one being exploratory action and theother perjormatory action. Muscles can enhanceperception as well as do work and some, like theeye muscles, have this function exclusively. Thehands, mouth and nose, ears, and eyes are all intheir own way active systems, as the body is. The

primary reaction to pressure on the skin is explora-tion with the fingers. Chemicals at the nose andmouth first elicit sniffing and savoring. Soundat the ears causes head turning. Light at theeyes brings about focusing, fixating, converging,and exploring of the light. Note that the outcomeof all these adjustments is to obtain stimulationor, rather, to obtain the maximum informationfrom the available stimulation.

This new picture of the senses includes at-tention as part of sensitivity, not as an act ofthe mind upon the deliverances of the senses.Every csthesic system is an attentional system.Attention is not an intervening process, therefore,but one that starts at the periphery. It alsocontinues to select and filter the already selectedinputs at nerve centers, as we know both fromintrospection and from the evidence obtained bymicroelectrode recording.

Pattern and Change oj Stimulation

Consider the sense organs in the old way, eachas a population of receptive units. We havethought of the retina, the skin, the tongue, andperhaps the olfactory epithelium as examples ofa sensory surface, a mosaic. Even the Organ ofCorti and the lining of the statocysts may beconceived in this way. But note, parenthetically,that the flat surface analogy does not hold atall for the articular sense, that is, the set ofreceptors for all the joints of the skeleton. The


point is that any population of receptive unitsis capable of delivering a simultaneous array ofneural inputs (although it is gratuitous or false tocall this a two-dimensional pattern or picture, aswe do for the retina and are tempted to do forthe skin). Apart from this muddle, every sense,then, is a pattern sense. Equally, they are allcapable of delivering a sequence or stream ofneural inputs or changes in the simultaneouspattern. Every sense is therefore a transforma-tion sense as well as a pattern sense.

Consider next the stimulation for these senses,the proximal stimulation. In every case it alsois a simultaneous array and a successive flux.There are two kinds of order in stimulation, asI once put it, adjacent order and sequential order(Gibson, 19SO). Pattern and change are char-acteristic of stimulation in general, unless it hasbeen sterilized by an experimenter, and here iswhere the information lies. For example, patternand change occur at the retina and the skin—evenmore at the dual retina and the two-handed skin,as the experiments reported have shown. Theyoccur at the basilar membrane of the cochlea as,respectively, the momentary sound spectrum andthe transients of sound; moreover the binauraldisparity patterns change with head movement.The simultaneous pattern of input from all jointsof the skeleton taken together is a highly intricateand interlocked configuration, yet its slightesttransformation seems to be registered when theindividual moves. Pattern and change occur atthe gustatory and olfactory surfaces, and evenfor the statocysts and the semicircular canals.Pattern and change are universal.

Now, sensory physiologists have always recog-nized the importance of patterns of stimulation

and tried to relate them to the sensory projectionareas of the brain. What they have not understoodis transformations of pattern. They have triedto imagine a cortical correlate of form, which isdifficult enough (as witness Hebb's recent attemptto explain visual form perception, 1949), butnot the changes of form which I have described.A tabulation may help to clarify the problem(Table 1).

The motionless frozen observer with his eyesfixed on a motionless frozen world gets a patternof stimulation from each of his senses (Type Istimulation) but the situation is hardly typical.The array at the eyes is comparable to a pano-ramic still picture. If he moves, or if somethingmoves, the arrays change (at the eyes, the skin,and the joints, for instance) in specific ways ordimensions (Type II stimulation). I have workedout the dimensions of transformation for theeyes, and it ought to be possible to do this forthe other systems. Subjective movement andobjective motion (A and B) normally yield dif-ferent stimuli even at the eyes. The observercan see himself moving, as one does in automobiledriving, and even see his own eye movements, asin observing the shifting of an afterimage, butthese are perceptions with subjective reference.They are "proprioceptive." We might say thatthe stimuli are propriospecific, since they carryinformation about the self.

The third type of stimulus variable is crucialsince it is taken to explain the invariance ofobject perception. Change of an array usuallyinvolves nonchange. Some order is preserved inevery transformation. Neither at the eye nor theskin nor at any other organ does the energyscintillate, as it were, like the random flashing of

TABLE I

A CLASSIFICATION OF STIMULUS VARIABLES FOR PERCEPTION

I. The unchanging stimulus array. Unvarying variablesDimensions of pattern, form, and structure as such

II. The changing stimulus array. Varying variables

A. Self-produced transformation—specifies motion of self1. With sense-organ exploration—control of attention (e.g., eye movement)2. With gross motor reactions—control of performance (e.g., locomotion)

B. Other-produced transformation—specifies motion of object

III. The invariants in a changing stimulus array. Invariant variablesUnchanging dimensions under transformation—specify rigid surfaces and objects


the fireflies in a field. There are always invariantvariables alongside the varying variables. Theyare specific to (but not copies of) the permanentproperties of external things. It is not a paradoxthat perception should correspond to the distalobject, although it depends on the proximalstimulus, if the object is in fact specified in thestimulus. The Ames demonstrations purporting toshow that optical stimulation can never specifyobjects depend on a fro/en array from which theinvariants cannot emerge.

Consider the difference between unvarying andinvariant variables of stimulation (Type 1 andType III). In the former case the stimuli thatwould be invariant do not get separated off fromthose that would vary if the array underwent trans-formation. The fro/en array, the case of continuousnontransformation, carries less information. Thecase is one that never occurs in l i fe . A prolongedfreezing of the pattern of stimulation on the retinaor the skin, in fact, yields an input which soonfades away to nothing.

The normal world is sufficiently full of motionsand events to make a stream of stimulation. Buteven without external motions a flow is produced.The normal activity of perception is to explorethe world. We thus alter its perspectives, ifevents do not alter them for us. What explora-tion does is to isolate the invariants. The sensorysystem can separate the permanence from thechange only if there is change.

In vision, we strive to get new perspectives onan object in order to perceive it properly. 1believe that something analogous to this is whathappens in the active exploratory touching of anobject. The momentary visual perspectives, ofcourse, are pictures or forms in the geometricalsense of that term whereas the momentary tactualperspectives are not. Nevertheless they are similarsince, for an object of a given solid shape, anychange of cutaneous pattern like any change ofretinal pattern is reversible. The impression ofthe object on the skin, like its impression on theretina, can recur by a reversal of the act thattransformed it. The successive patterns thus fallinto a family of patterns which is specific to theobject. I submit to Hebb the suggestion that thefirst problem in perceptual physiology is not howthe brain responds to form as such, unvaryingform, but instead how it responds to the invariantvariables of changing form. T think we should

attempt a direct physiological theory of objectperception without waiting for a successful theoryof picture perception.

Invariant Properties oj a Changing Stimulus Array

The crux of the theory of stimulation here pro-posed is the existence of certain types of perma-nence underlying change. These invariants arenot, I think, produced by the acquiring of invariantresponses to varying stimuli—they are in thestimuli at least potentially. They are facts ofstimulus ecology, independent of the observer al-though dependent upon his exploratory isolationof them. This kind of order in stimulation is notcreated by the observer, either out of his pastexperience or by innate preknowledge. Just asthe invariant properties of the physical world ofobjects are not constructed by the perceiver, sothe invariant properties in available stimulationare not constructed by him. They are discover-able b3^ the attentive adjustments of his senseorgans and by the education of his attention.

Some of these stimulus invariants are extremelysubtle. The ultimate subtleties of the informationin stimulation may well be unlimited. But otherinvariants are quite simple and easily detected. Theoptical texture that specifies a physical surface (incontrast with the textureless patch that specifiesan empty space) is invariant with illuminationand under all transformations of perspective. In-trospect ively we say that one yields a surfacecolor and the other a film color, but the spatialmeaning is what counts, not the introspection.The textures of earth, air, and water are different,and the differences are constant. So are the dif-ferences that specify to the young of any speciesthe fur, feathers, or face of the mother. Theintensity and wave length of the light are irrele-vant. The infant seems to be sensitive fromthe beginning, more or less, to such externalstimuli as these. The tablet of his consciousnessmay be nearly blank at birth, as Locke believed,but the impressions that do appear are vagueperceptions, not bare sensations. The earliestdimensions of sensitivity are useful ones.

Classical sense impressions, I think, are some-thing of which only a human adult is aware.They tend to arise when he introspects, or whenhe tries to describe the content of experience, orthe punctate momentary elements of perception,


or when simple variables of physical energy areexperimentally isolated for him by a psychologist,or when stimuli are applied to his receptors insteadof his being allowed to obtain them for himself.Far from being original experiences, they aresophisticated ones; they depend on having had agreat deal of past experience.

This is not to deny that perception alters withlearning or depends upon learning. Instead itpoints to a different kind of learning from thatwe have previously conceived. Unquestionably theinfant has to learn to perceive. That is why heexplores with eyes, hands, mouth, and all of hisorgans, extending and refining his dimensions ofsensitivity. He has to separate what comes fromthe world and what comes from himself. Buthe does not, I think, have to learn to convert sensa-tions into perceptions.

REFERENCES

BRUNSWIK, E. Perception and the representative designof experiments. Berkeley: Univcr. California Press,1956.

CAVINESS, J. A., & GIBSON, J. J. The equivalence ofvisual and tactual stimulation for the perception of

solid forms. Paper read at Eastern Psychological As-sociation, Atlantic City, April 1962.

GIBSON, J. J. The perception of the visual world.Boston: Houghton Mifflin, 1950.

GIBSON, J. J. Visually controlled locomotion and visualorientation in animals. Brit. J. Psychol., 1958, 49,182-194.

GIBSON, J. J. Perception as a function of stimulation.In S. Koch (Ed.), Psychology: A study of a science.New York: McGraw-Hill, 1959. Pp. 457-501.

GIBSON, J. J. The concept of the stimulus in psychology.Amer. Psychologist, 1960, 16, 694-703. (a)

GIBSON, J. J. Pictures, perspective, and perception.Daedalus, 1960, 89, 216-227. (b)

GIBSON, J. J. Ecological optics. Vision Res., 1961, 1,253-262.

GIBSON, J. J. Observations on active touch. Psychol.Rev., 1962, 69, 477-491.

GIBSON, J. J., & GIBSON, E. J. Continuous perspectivetransformations and the perception of rigid motion./. exp. Psychol., 1957, 54, 129-138.

HEBB, D. O. The organization of behavior. NewYork: Wiley, 1949.

ROSENZWEIG, M. R. Development of research on thephysiological mechanisms of auditory localization.Psychol. Bull., 1961, 58, 376-389.

WALLACII, H. The role of head movements and vestibularand visual cues in sound localization. J. exp. Psychol.,1940, 27, 339-368.

gibson (1963) the useful dimensions of sensitivitywexler.free.fr/library/files/gibson (1963) the...

Documents