Wertheimer: The Perception of Movement and the “Natural” Organization

The work of Wertheimer (1912a and 1923) is correctly considered to lay bare the experimental foundations of Gestalt psychology. However, it is interesting to note the phenomenological tenets of his research that are to some extent independent of the Berlin School of Gestalt theory.

It is well known that the first article deals with the perception of movement. Wertheimer says that a motion can be described as follows: an object X at the time t1 is found in the position L1 or in the place P1; from the time tn onward X is found at the position Ln or at the place Pn, whence in the interval t1-tn X had to occupy successively the positions L1-Ln continuously in both space and time. Yet this description admits physical motions that are not really perceived. For example, the clock pointers are simply seen located at two distinct points before and after so that motion is actually inferred. Since the object of study is perceived movement, this description needs to be restricted. Otherwise physical motion would provide the model to study the perception of movement. However, it might imply assumptions that are inconsistent with the observation of how and when a movement really appears. Instead the questions of research have to be derived from clear-cut and conspicuous appearances of movement if one aims at discovering their constituents and structure (1912a: 166).

According to this phenomenological tenet, the research uses stroboscopic movement. Wertheimer (1912a: 167-168) defends the choice of this experimen­tal paradigm by arguing that it enables him to reject preconceived interpreta­tions of movement perception. If the clock pointer description is accepted as a model, for instance, one can believe that perceivers have to integrate the inter­mediate positions through which a body passes when it physically moves from the starting point to the end. Then if the physical motion is the model of the research, the stroboscopic movement is an illusion due to the misapplication of the inference that is alleged to integrate the appearances of motion in ordi­nary experience. However, if this subjective inferential integration is assumed, the conclusion ensues that the stroboscopic, but also every movement, is not perceived at all. Even in ordinary life, subjects could unawares integrate the phases of external motion on the basis of past experience or implicit knowl­edge. On the contrary, Wertheimer argues that stroboscopic movement makes it possible to address the description and the explanation of what is “psycho­logically” given because it allows for separating the phenomenal from the physical determinants of movements. Through the variation of experimental conditions, a series of clear-cut appearances of movement can be obtained so that the study is conducted without prejudice and on the grounds of the sole phenomenal evidence that as such may be extended to ordinary experience.

Wertheimer employs the tachistoscope and a sliding frames device. The experimental design consists in presenting two stimulations at rest that are parallel or oblique bars and stripes located in different positions at distances of 1, 3, 5 cm, which are successively exposed to subjects with a varying inter­val t. The experimental questions are: is there an optimal time interval for a clear-cut appearance of movement to occur? Are there qualitatively distinct appearances for t values differing from the optimal interval? Does something appear for a smooth change of these latter t values? What are the constituents of the appearance of movement? Is the field, that is, the space between the first and the last position involved in the seen movement, perceived by sub­jects? Furthermore: how does the movement perception arise? Does it develop from the appearances occurring at different t values? Does it break up in them, and how? (1912a: 166).

Wertheimer discovered three t values for the interval between the exposi­tion of the stimuli A and B at rest, which set the boundaries of three primary stages (Stadien) of qualitatively distinct phenomena. The 30 msec. ca. interval marks the boundary of the stage where the simultaneous rest appears. Two stripes A and B are seen at once at rest in their respective places. The interval of 60 msec. ca. marks the boundary of the optimal stage where the movement ap­pears. The clear perception arises of one and the same object that moves from the place of A to the place of B, while the in-between space appears as a contin­uous ground across which the movement takes place. With the tachistoscope, a stripe is seen to move back and forth, if A and B are parallel, or to rotate if A and B are the vertical and horizontal of a right angle. If a repeated exposition of the two stimuli at rest is realized for the same optimal t, then one and the same object appears to move back and forth several times or to rotate towards either side of A and B (1912a: 186, 213). Finally, the interval of 200 msec. ca. marks the boundary of the stage where the succession appears. Two stripes A and B are successively seen located at rest in distinct places. The observation of the phe­nomenon at the optimal stage allows drawing important conclusions (1912a: 186). Firstly, it shows that movement is a perceptual datum that is given be­tween the positions of A and B, while it does not need stimulation in the space between them. Secondly, the appearance of movement does not require an object that occupies successively distinct places (1912a: 212-213). Wertheimer reports protocols in which subjects claim to see an object that is already in mo­tion or is beginning to move as early as its first appearance in the initial posi­tion of A. They don’t see an object first in this position, then the same object in motion and finally at rest in the position of B. He argues that this evidence has real theoretical value, because what is perceived is already caught in motion; hence the movement is not a phenomenal property that is added to an object or to A and B; rather it is a feature that groups the visual elements or empha­sizes the way the object appears in starting or end positions. By shortening the exposition time of the stimulation, the subjects can see the object beginning to move from the start, still in motion at the other end, or continuing uninter­ruptedly to move from A to B. If A and B are successively presented with fixed t, the subjects no longer report seeing a back and forth movement or a repeated rotation in which they see the initial and final positions of A and B. Instead, they report seeing a moving object that repeatedly crosses a spatial area from one end to the other without being able to discern the initial and final places in the appearance of this continuous movement. In other words, the object is seen as something purely in motion for the whole field.

Wertheimer defends the generalization of the findings on the phenomenal features of movement. The experimental conditions are modified so that a physical and a stroboscopic movement are concurrently presented one below the other, through the tachistoscope and the sliding frames device, while the subjects do not actually know which one is the physical motion (1912a: 168, 173-174). In most cases, the subjects cannot distinguish the two, even if they have become skilled in accurate observation by a month’s training with the ta- chistoscope. Sometimes, after a repeated exposition of the sliding frames and with a longer observation, the subjects succeed in identifying correctly the two movements that are yet distinguished by a qualitative difference. Moreover, Wertheimer recorded several protocols in which the stroboscopic movement is the best appearance of movement, that is to say the case in which the char­acter of being in motion is much more conspicuous than in the case of the physical one.

Wertheimer reports another phenomenon to support the view that strobo­scopic movement provides the evidence to study the perception of movement, because the appearances are clear-cut, conspicuous and can be observed without the assumptions deriving from knowledge of the physical motion. If t is a little shorter than the optimal stage, the subjects report seeing either a motion across the visual field or a rotation, which are clearly seen as unitary movements although they cannot be attributed to an object because no stripe or bar appears (1912a: 222, 226-277). If A and B are the vertical and horizon­tal lines meeting at an angle, a rotation around the vertex with a downward direction toward the horizontal line is seen, although both A and B appear to have remained still. The subjects report seeing a motion across the field, rather than the motion of A to B or vice versa. They cannot tell anything about whether the object has moved or rotated, but they report that a passage across the field or a rotation has appeared. To emphasize this feature of being object­less, Wertheimer calls this appearance “pure movement” or phi phenomenon. He remarks that if subjects are asked to pay attention to the field area wherein the movement is going to occur, the phi phenomenon becomes more and more conspicuous.

Thus Wertheimer claims the theoretical value of his experimental design. Phenomena arising at well-determined t values are observable data with specific features. Appearances at each stage are as objective phenomena as col­ors or forms, in the sense that they too refer to or are about something (1912a: 227). Among them the phi phenomenon is equivalent to the pure appearance of motion. This research paradigm provides both a theoretical means to test various hypotheses on movement perception and an experimental proof of the features of the objective appearance of movement. The first conceptual result is that physical movement is not an adequate model because there is no need for stimulation at each intermediate position. The phi phenomenon shows that there is no need even for the perception of an object as the bearer of mo­tion against the naive interpretation of movement. Moreover it is experimen­tally proved that movement perception does not depend on non-perceptual psychological functions, whether association or judgement. There is no room for a subjective integration of data in the intermediate positions between A and B. In the phi phenomenon, neither the properties of an object nor the object itself appears either in the starting and final positions or in the in-be­tween places, yet still the movement appears. Besides, if a third object is placed somewhere between A and B, it is not integrated in the movement appearance (1912a: 201-202, 223-224). Wertheimer presents a smaller stripe C with alter­natively the same or different form and color either after exposing A and B or through successive expositions A-B, A-C. In the optimal stage, if subjects are asked to concentrate their attention on C, it appears to be at rest while another object keeps uninterruptedly passing from A to B. Sometimes subjects report seeing C in motion, yet C either moves in a separate field area or snaps in a direction which is neither A-C nor A-B. Then the apparent movement of C is never phenomenally connected with the A-B movement. In the stage of the phi phenomenon with A and B meeting at a right angle and C located at 45° near the vertex, the unitary objectless movement still appears to span the 90° field region, while C neither appears to be as it were lengthened by taking part in a motion that spans a more extended region, nor does it present a shimmer that flits from it to the vertex. The field between C and the vertex belongs still to the black ground at rest.

The hypothesis of the subjective integration can also be formulated thus: subjects now see A here and then B there at a distinct distant position; hence since A and B appear as the same, subjects believe with certainty that they are an identical object and judge that one and the same thing must have moved from the position of A to that of B (1912a: 187k). This hypothesis rests on the assumption that identity is a necessary feature of the perception of movement and that the belief and the judgement are the non-perceptual functions that supplement the sensory content of what occurs in-between the stimulations of A and B. As regards the role of identity, Wertheimer concedes that actually in the optimal stage one and the same object is seen moving from A to B, but he argues that the feature of identity is phenomenally separable from the appear­ance of movement. Indeed, the identity is perceived only for a limited range of t. Beyond this range, nothing enables the subjects to presume or to suppose that something identical is somehow present, while the motion appearance remains compelling. At neighboring t values, if t is shortened a little, the move­ment is seen, but A and B appears as two objects. If t is smoothly varied start­ing from the optimal value, the identity is lost, although the movement is still perceived. If t is varied in the reverse direction starting from the simultaneity stage, the movement occurs even if the identity is not yet perceived. On the other hand, within the same range of t the perception of movement is pre­served even if A and B are clearly different in forms or colors. If A is a longer stripe than B, one and the same stripe is seen to move along a curved trajectory from A to B. If A is red and B blue, the subjects see either a red stripe in motion that is blue at the arrival or, with successive expositions, a stripe that moves back and forth while changing color (1912a: 189-190). Wertheimer reports that through this variation of t, even for non-optimal values, it is never the case that the identity is perceived while the movement appearance is uncertain or dubi­ous, rather the opposite is often true. In conclusion, there is robust evidence that the appearance of identity is detachable from the movement appearance. Therefore belief in the identity of the moving object is not required. As regards the role of judgement, Wertheimer claims that the unprejudiced observation of data proves that the movement is actually seen, namely it cannot be de­scribed as “something alike occurs here and there, then it must be one and the same object that has moved” (1912a: 187). There are cases in which appearances do not allow for the judgement of equality A = B; hence, its postulation has no evident grounds. Even if the subjects are told that two different stimuli A and B are employed and they know that the experimental paradigm consists of a varying succession of A and B at rest, they cannot but see the movement of an object.

However, another hypothesis can be made that accounts for the perception of movement as a judgement. The subjects only believe they see the move­ment, while in reality they judge erroneously that the sensory material of A and B is in motion because they are deceived by the values of t that would be the same if an existent object was physically in motion (1912a: 240-241). Optimal intervals bring about an illusion of movement and induce the sub­jects to misapply a judgement that could be correct in similar circumstances. Wertheimer proves this hypothesis to be false. In general, if the stroboscopic movement were the outcome of an error of judgement, the illusion would be weakened and vanish in conditions of repeated and accurate observations in which the subjects are asked to concentrate their attention, even for a pro­longed time, or have been previously trained to the experimental conditions. As the experimental findings show, however, this is not the case. In particular, if the attention is focused on the field region where the illusion is presumed to occur, the phenomenal evidence of movement is not weakened, rather it is increased and strengthened. Moreover, since the phi phenomenon does not involve an object, it should show a weakening of the alleged illusion of movement. Yet a variant of the phi phenomenon disproves this claim. If A is an inclined stripe touching the horizontal stripe B, the subjects see a double pure rotation of 90° to the right and the left at the same time. They do not see two separate stripes A and B rotating in opposite directions, but rather A and B that remain still and are distinguished from two pure rotations whose last phase involves the horizontal stripe. Wertheimer calls this phenomenon “phenomenon 2-phi’ (1912a: 205). How could an illusion arise of two opposite movements even if the stripe A does not split into two or cannot seem to move in two opposite directions at the same time?

Finally, the hypothesis cannot explain the whole series of appearances. The illusion yielded by an error of judgement may account for a single appearance. Yet it cannot account for each and every appearance that arises gradually and systematically according to the smooth variation of t through all the stages of the simultaneity, the optimal motion, the pure motion, the succession.

Wertheimer takes the three primary stages as “regions” of possible appearances, which are exemplified by a distinguished phenomenon representative of succession, simultaneity and movement. Accordingly, along with characteristic appearances for each region there are more or less deviant appearances that may arise as the optimal t for each region shifts and approaches the t values that set the boundaries of other regions. In addition to the deviant appearances that are obtained within or at the boundaries of a region, Wertheimer observes that a smooth change of the presentation conditions may bring about the abrupt occurrence of appearances that radically depart from those of the primary stages. If the t values are smoothly manipulated in order to allow a transition among the three stages from the optimal stage of motion to the simultaneity or the succession and in the reverse direction, the subjects initially see a bad appearance of motion that is difficult to describe, then a motion that is no longer unitary but rather seems to be a jolt or split in an up and down motion, and finally a well-defined appearance that Wertheimer calls “partial-motion” (1912a: 192). If A and B meet at a right angle and t is between the values of the optimal and the simultaneity stages, two stripes appear to move separately spanning a 30° field region. With t values neighboring the simultaneity stage, there occur appearances that do not involve the field, but rather qualitative changes of the moving object. Wertheimer (1912a: 196, 234) calls them “internal movements.” The subjects report seeing changes of brightness or a flickering in the two stripes that are no longer at rest, because their upper end appears first followed by the remainder, while the brightness change spreads downward over it. Sometimes the middle part of the stripe appears before its upper and bottom ends. Finally, with t values between the optimal movement and the succession, the phenomenon of “singular movement” occurs, in which A or B remains still while the other moves (1912a: 199f.). If A and B meet at a right angle and the vertical stripe corresponding to A is completely at rest, the horizontal stripe moves independently with a 45° rotation. Wertheimer remarks that it is possible to obtain independent rotations of varying amplitude, from 30°, 15° to a minimal movement of B, but B never reaches the region of A nor does it have an effect on it. In the minimal motion, B appears to begin to move or to snap in its own place. If the exposition of the stimuli is reversed, the stripe and its direction of motion are inverted accordingly. Wertheimer emphasizes that were the alternative explanations of movement perception true, only a weakening or worsening of the same appearances of movement should take place for t values that vary smoothly, instead of abrupt appearances that are characterized by distinguishable qualitative features.

Wertheimer (1923) develops the line of research that embeds the phenome­nology in the experimental design and the interpretation of data. At the begin­ning of his article, Wertheimer lays bare his phenomenological method (1923: 301). If one stares outside through a window, he might describe what he sees as, for instance, a house, the sky, some trees. It is true that the properties of wavelengths might figure in the description, whose values can be found in the visual scene according to the physical knowledge that is applied to the study of vision. Yet the physical units into which the visual scene is decomposable can­not replace the descriptive units corresponding to the ordinary visual things. Wertheimer claims that the description of what appears avoids the arbitrary partition of the visual scene. The subject at the window might believe that or­dinary visual things are in reality physical properties and try to compute their number in the scene, concluding that there are 327 values of some physical properties X or Y. However, no one is able to observe this number of values as such or to specify which values of X or Y correspond to the distinct things in the scene. Even if one succeeded in classifying the 327 values into distinct groups consisting of 120, 90, 117 values and assigned them to the house, the trees, the sky, what would be the reliable criterion to prove this assignment correct? Since the assignment does not derive from an observable rule of correspondence, it is arbitrary because it does not provide any conclusive reason to decide between this distribution and (127, 100, 100) or (150, 117). In fact, the validity of the account in terms of physical properties depends on the description of vi­sual things. The distribution of the physical units into which the visual scene is decomposed needs to match the grouping of visual things and qualities; hence, it depends on the description of how appearances are unified and segregated with respect to one another. The description of the visual scene decomposes it into independent or dependent, unified or separated, connected or detached units. This decomposition represents the phenomenal articulation of the scene assuming that subjects cannot perceive arbitrarily different unifications and connections of parts in it and at the same time perceive the same mean­ing. For example, a prolonged observation could induce someone to see a piece of the dark window frame along with a tree branch as the character “N.” Yet the unification of these two elements of the visual scene is actually a discovery that is made through an attitude of inspection of the scene that is very far from the reality of what is actually given, so that the surprise for what else could have been perceived is not incidental. The same argument holds for auditory perception. Suppose one listens to a melody that is composed of 17 tones and accompanied by 32 tones. The melody and the accompaniment are the real perceptual objects, which are not correctly described either as 49 tones or as an arbitrary partition, for instance two groups of 20 and 29 tones.

Wertheimer’s argument, then, is that the phenomenological description is the starting point of research because it provides a sort of standard to gauge the scale and the units for the study of vision. The descriptions capture the fact that the outside world does not appear as either an undistinguishable sensory flow or a collection of bundles of qualities that are tied together by chance. Instead, the things and qualities of the world are grouped in the visual scene and the most “natural” grouping among those that are possible is forced upon subjects. Indeed, the same grouping is preserved when the stimulation changes, for example if a continuous stimulation is substituted with a discrete one. The tones of a melody and its accompaniment are unified in two distinct auditory objects if they are played by a musical instrument or reproduced by the loud and short single tones of a musical box. Likewise the figures made up of dots are visual unities that still stand out against the homogeneous ground just as figures made up of continuous lines do. Wertheimer holds that in every natural or artificial condition, like those designed for experimental research, many unifications and segregations of units may arise. Among them there are groupings that are comparatively easier to see because they appear as “natu­ral” forms of unification and segregation. Other groupings are more difficult to observe, though they still arise or are artificially produced but only sometimes under determinate circumstances.

Wertheimer aims at discovering the factors underlying the natural-like grouping of units in any visual scene and their possible combinations. In ac­cord with the phenomenological tenets of his research into movement per­ception, he remarks that these factors must be extracted by the observation of the appearances rather than inferred from the physical properties of the stimulation. Wertheimer emphasizes that there is no lawful principle from which to derive a rule that for a given number of stimulations that together are effective for perception, the same number of corresponding phenomenal units occur (1923: 302). Since perceptual things are dependent on the relations of unification and segregation underlying grouping, these relations may have a greater magnitude than the number of single stimuli. If the stimuli A, B, C, D, E are given and act together, there have to be some principles that parse the corresponding perception by grouping the appearances into (a – b – c)/(d – e) rather than (a – b)/(c – d – e). This means that self-ruling laws govern the ap­pearances as observable rules of connection of the visual objects and events. It is well known that Wertheimer discovers that these principle are the grouping factors of proximity, similarity (Gleichheit), common fate, good continuation, closure and the past experience that is yet acknowledged to play a limited role contingent upon well-determinate perceptual conditions in opposition to a Hume-like theory of perception (1923: 33of.). Wertheimer capitalizes on previ­ous researches (cf. Schumann, 1900; Muller, 1904), but he gives the factors a phenomenological treatment, trying to specify their cooperation and competi­tion. A meaningful implication is that the same factors account for the struc­ture of perception and the spontaneous, natural-like and ordered form of the outside world.

Wertheimer shows that these grouping factors underlie the inherent com­position of visual objects, namely the relations between their parts, in connec­tion to which the relations among objects as a whole can be perceived. Let’s consider the proximity factor. However simple a factor it may be, it provides the foundation of the unity and inner structure of visual objects, as the follow­ing example shows.

Starting from row 7(a), let’s add a pair of equidistant dots, one above and one below, for each dot starting from the second on the left, in such a way that the dot pair is not collinear with the dots in the row as in figure 7(b), in which one is likely to see three distinct parallel rows. Now let’s repeat this operation twice starting respectively from the third and the fourth dot in the row, un­til we obtain the figure 7(c), in which the rows are no longer visible. Instead one is likely to see a whole consisting of the leftmost dot and four subgroups, like those marked by the lines in figure 7(d), which are nonetheless connected in a unitary visual object. One might also have seen a different visual object obtained through other subgroups connected as in figure 7(e). Yet the prox­imity factor makes, ceteris paribus, the visual organization shown in 7(d) the most “natural” to be seen with respect to the whole object and its visible parts corresponding to its inner structure. According to Wertheimer, on the grounds of the experimental research into the form of combinations of the grouping factors, the psychological science of perception can account for the structure of the phenomenal world.

Source: Calì Carmelo (2017), Phenomenology of Perception: Theories and Experimental Evidence, Brill.

Leave a Reply

Your email address will not be published. Required fields are marked *