L.Kleine-Horst: Empiristic theory of visual gestalt perception. Hierarchy and interactions of visual functions. (ETVG), Part 10, II
a) The new facts
"Furthermore, you should acquaint yourself with a much newer development in America, England, Australia, and Japan, the investigation and theoretical interpretation of so-called 'figural after-effects', which throws new light onto human perception."
Figure 10-1. A pattern for producing figural after-effects (see text)
So I read "Dynamics in psychology" by W. Köhler (1940), and found very interesting new facts there. Köhler proposed drawing two equal black outline oblongs on a large piece of white cardboard, as shown in Fig. 10-1, focusing on a point midway between them, and covering, for example, the right oblong for three minutes with another piece of white cardboard. When this oblong, as the test pattern, is uncovered after this interval, the left oblong that had been observed for the three minutes, appears smaller, less black, and positioned farther back in space than the right one. In another experiment (Fig. 10-2), an oblong (A) is at first presented for several minutes, and after it is taken away, three small squares (B) are presented, the retinal projections of which fall onto the previous oblong projection area as is shown in Fig. 10-3.
Figure 10-2. Another pattern for producing figural after-effects (see text)
The three above mentioned changes in the appearance of the squares show a spatial gradient of the effects: they are greatest within the area of the previous infield, less in the outfield closest to the figure contour, and least farther away in the outfield. According to Köhler's theory, this is caused by electric currents originating from the contour and spreading outwards to slowly form a functional field of the figure, which is strongest in the infeld, slightly weaker in the outfield nearest the contour, and weakest farther away in the outfield. The establishment of the field during the inspection phase proceeds slowly according to Köhler, because it is conditioned by the diffusion of ions through the brain tissue, which requires more time.
b) Köhler's failure to account for most after-effects
It surprises me that even in 1963 B1 still believed that Köhler had developed a remarkable theory capable of accounting for figural after-effects, or even human perception at all. His current-theory of perception refers only to two of seven classes of phenomena regularly resulting from such experiments on figural after-effects. Köhler reported five of them, listed as follows (the sixth and seventh phenomenon is dealt with below):
1. a general after-effect, i.e. a perceptual change in a (black outline)
figure and its outfield, due to a lengthy previous stimulation of the retinal
area on which the figure is projected.
2. a spatial gradient of the after-effect starting from the figure contour, and spreading outwards (i.e. into the infield and the outfield).
Furthermore, he reported three special after-effects on a figure:
3. a decrease in the contour blackness of the figure
4. the figure's withdrawal into depth, and
5. its decrease in size.
Köhler's (1940) procedure for developing an explanatory theory was very simple. He established an hypothesis, used it to explain the phenomena that agreed with it, and disregarded any phenomena that contradicted it.
"We shall merely assume that currents which are conducted in the tissue tend to alter this medium, and thereby to impede their own further passage" (p.74).
With this assumption he was able to explain phenomena No.1 and 2 which did not contradict the long term inspection and fixation condition, and left out the three special after-effects (No. 3-5), which regularly occur under the same conditions, by claiming:
"Our primary interest is not concentrated on the after-effects as such. They are facts which probably have no great importance in ordinary vision; we do not as a rule fixate a given point for any length of time" (p.105).
Moreover, an electric field requires a great amount of time to be established, as it is formed via the diffusion of ions from cell to cell. Therefore, Köhler instructed his subjects to stare for long periods of time - 3 to 12 minutes - at the indicated fixation point. After this, five phenomena did occur. But Köhler did not report what would happen after only a short inspection period; he stressed using only an extended period of time for the first stimulation. He explained that, for experimental purposes, longer periods of time were an advantage: the longer the first stimulation, the longer the after-effects remain, and the longer the after-effects remain, the greater the number of different test-patterns that can be used after a single stimulation. Köhler also did not mention, in respect to Fig. 10-1, the length of time he used as the interval between the offset of the inspection and the onset of the introduction of the test pattern, or the onset of comparing the oblongs. He only claimed, "When after this period the screen is removed, and the second oblong becomes visible, the two figures no longer look alike."
When I shortened the inspection time, I obtained both the same phenomena as reported by Köhler, and also such that Köhler did not report. I did not follow the rules that Köhler felt were necessary to prove his theory valid: I need only several seconds to establish the after-effects, for example, in the experiment of Fig. 10-1, when I stared at the fixation point very rigidly, and then suddenly uncovered the right oblong. In this case, I even experienced a new phenomenon, the suddenness of the occurrence of the three figure-weak phenomena: a sudden constriction of the left oblong, its sudden withdrawal toward the background, and a sudden graying of its black contour. Thus, two further after-effects that could not be accounted for by Köhler's theory, and that were therefore withheld by the author, are:
6. the occurrence of the after-effects after short-term inspection as well,
7. the suddenness of the after-effects with the sudden introduction of the test figure.
Apart from its failure to explain the three figure-weak phenomena, with only these two facts Köhler's current-theory breaks down because
1. the phenomena reported by Köhler were observed under temporal conditions not necessary for establishing them, they are only necessary for Köhler's explanation: a long-term inspection period, long enough to allow for the sufficient diffusion of ions from the contour into its adjacent fields.
2. Köhler used the second pattern exclusively as a "test-pattern" to demonstrate that the after-effect is due to the long-term inspection of the first pattern. However, the second pattern not only exhibits a long-term influence on itself (as Köhler mentioned as well), but also a short-term influence on the first pattern, which is seen when the second pattern is presented quickly.
I did not trace the further development of figural after-effect research, so I do not know whether any of the investigators in America, England, Australia, and Japan did not also find these short-term phenomena. I only read somewhere that the predicted current has proved to be non-existent, and I wondered, based on the past where a remarkable number of facts that could not be explained by or even contradicted the current-theory were simply swept under the carpet, why somebody bothered to search for a current at all.
Another question: What told Köhler about the influence of attention on figural after-effects? Nothing at all. In his experiments, he neutralized the spatial gradient effect of voluntary attention by instructing the subjects to focus upon a point midway between the figures to be compared. In Fig.10-2 there are three "second figures". In respect to the previously presented figure/outfield pattern, they can be distinguished as (one) "infield square" and (two) "outfield squares". As the subjects focus on a point midway between the outfield squares (which is approximately also midway between the infield square and the farthest outfield square), it proves that voluntary attention is not responsible for the differences in the strength of the three figure-weak phenomena.
Köhler, however, did not consider any involuntary attention (and its spatial gradient effects). He did not even mention either involuntary or voluntary attention, both of which possess a spatial gradient. If he had done so, he would have had to describe the "attentional field" and its spatial gradient effects (as is roughly drawn in Fig.10-3). Furthermore, he would have needed to explain why it was necessary, for the purpose of experimental exactness, to neutralize the spatial gradient effects of at least voluntary attention. He just silently neutralized it. As the Berlin gestaltists, and other scientists, had and have great difficulties with "attention" (see Part 6), Köhler simply neutralized the spatial gradient of voluntary attention (instead of investigating its influence on the relevant phenomena), neglected the spatial gradient of involuntary attention, and therefore had to search for another entity that possessed a spatial gradient. And so he hypothesized an "electric field".
c) Accounting for the figural after-effects using the ETVG
In 1963, when B1 suggested that I deal with figural after-effects and their theoretical interpretation, I was not able to account for these facts using the ETVG, as I had only developed my 1961 version of the theory, which was not elaborate enough for achieving such an aim. This book, however, provides the principles and laws that can account for the phenomena listed above, except No. 5, which can be explained only by an additional hypothesis. One of the effects is the "general after-effect", i.e. the after-effect of stimulation itself. The ETVG, in its present state, does not explicitly provide an hypothesis on why the effect of sensory stimulation does not cease when the stimulation itself has ceased. However, the antagonism between the positive and negative functions of the gestalt factors as well as the bottom-up and top-down influences between the factors at different levels leads to the assumption of circular processes that trigger each other so that the functions are capable of still being in operation when the light stimuli have already ceased to impinge upon the receptors. This process is responsible for both the "positive afterimages" and its counter- process, the recreation-process, which allows the "negative afterimages" to occur.
The ETVG explanation of the No. 3 and 4 phenomena that resulted from the experiment, is quite simple. First, assume that the left oblong in Fig.10-1 is the first, and the right follows as second. When the left oblong is presented alone for a certain time, its black contour must become gray, according to the gestalt law Dtic1 dmio1, if the contour itself is considered to be "contour-figure" (ic1), so that the white fields separated by this figure constitute its outfield (o1). It does not matter if these fields form the infield and outfield of the "oblong-figure" (ic2). Second, the longer a figure is presented, the more it must "retire" back in space, due to the gestalt law Dtic Fdo. (As defined in Part 7, the distance of a figure from the eye is one of the two "depth outfields", Fdo.) Thus, two special figural after-effects (3 and 4) agree with the ETVG gestalt laws.
The third special after-effect (5), however, does not seem to agree with any gestalt law. But let us try, in spite of this, to explain the observed change in size of the left oblong using the ETVG. We now deal with the left oblong as a single figure. According to the ETVG gestalt law Dtic2 Dlic2, the oblong must phenomenally become larger with increasing time and not smaller, as is true in the figural after- effect experiments. Since the ETVG gestalt laws are all expected to be true, we must find another influence that overlaps the gestalt law Dtic Dlic. This influence must follow a law with the form "Dtic dlic". However, according to the principles for forming gestalt laws with the known gestalt factors, such a gestalt law is "impossible". For such a situation I propose returning to the natural situation where an infant implicitly learns both the relationships among the objects of its environment and the relationship between itself and the objects. Here we find the solution for our problem. Infants implicitly also learn the following "size/distance relationship": the closer in depth an object is, the larger it "is" (i.e. appears), as well as: the farther away in depth an object is, the smaller it "is" (i.e. appears). They furthermore learn: the larger an object "is", the nearer it "is". And they learn: the nearer an object is, the more interesting, or frightening, (i.e. in any case, "attractive") it is. This means: great attentional attractiveness, perceived nearness, and perceived largeness of an object (a figure, or a part of it) belong together; these "properties" are associatively closely connected with one another according to the fundamental Empiristic Association Hypothesis. In accordance with the everyday experience that every object loses its attraction, the longer it is presented (Dtic aic), we can conclude, that the longer the left oblong is presented, the less attractive, and consequently the farther away, the smaller, and less black/white contrastive it "is". Ebenso gilt: aic Dlic, i.e an increase in size when attention decreases.
If we assume that the laws that lead to a decrease in size are more effective than the laws that lead to an increase in size, and both groups of laws superimpose their effects, the resulting net effect would be a decrease in size.
If under certain experimental conditions, in which a figure loses in figure properties and takes on outfield properties, a certain group of phenomena occur, the same phenomena must occur when any other condition is established that decreases the figure properties of this figure. Such a condition is given for the left oblong in Fig. 10-1 when the right oblong is presented, particularly when it is presented suddenly. This suddenly occurring "new" figure is a very strong figure which establishes a respective strong outfield, and therefore decreases the figure property of the "old" figure, the left oblong that lies in the new figure's outfield, to a great extent.
Why was the size/distance law not integrated into the basic row of interaction factors? Because it is not purely visual, as the ETVG gestalt factors are. The ETVG does not deal with any memory contents in which other senses are also involved. It is, however, not possible for an infant to distinguish only visually whether an object is closer or farther away (they may be able to purely functionally, but not phenomenally, distinguish it). It also cannot only visually distinguish whether an object really becomes larger or if it only appears larger as it comes nearer, just as it cannot only visually distinguish between left and right, above and below. The association between largeness and nearness involves other senses than only the visual sense; the sense of touch, for example, when the milk bottle touches the lips of the baby and is thus extremely near after it has become extremely large; and the sense of smell, when the baby smells the milk more, when the bottle becomes larger, and when it is extremely large the smell is extremely strong, at which point the bottle even comes into contact with its mouth. If one wants to understand the prevalence mentioned above, one has to be aware of the infant's vital interest in the milk bottle, and in its smell and nearness: the most attractive objects "are" large, and not small. This is, indeed, much more important for the infant than the comemorized relationship between the small object and its large environment, measured in the frontal plane, which leads to the Xdl-laws.
In the experimental design depicted in Fig. 10-2 other conditions are given. Here, the stimulus that is presented first, the oblong, hits a "virginal" retina, so to speak, but the set that is presented second does not, because, to a certain degree, the visual system (with all its visual functions) is still activated by the first stimulus. Thus the oblong establishes stimulation conditions for the presentation of the squares different from those it had found itself.
As will be done in the following, when interpreting the effects of stimulation, not only the stimulation of the retinal areas that mediate figure experience, but also that of the retinal areas that mediate outfield experience must be taken into consideration.
When a figure is projected onto the retina, it attracts a certain amount of attention toward itself, according to the gestalt law Flic Aic. When a figure is projected onto a retinal area that is functionally "empty", i.e. at that moment activated neither by a present nor by a prior stimulation whose effect has not worn off due to its "general after-effect", its attentional attraction is very large. Hence, the figure shows, for example, very black contours, it is in the foreground, and it is almost of "veridical" size. In the case of two figures being consecutively projected onto the same retinal area, the attraction of all or parts of the second figure is smaller the more the visual system is still acativated by the first figure's stimulation, i.e. the second figure loses in figure-properties such as blackness, nearness, and (in the case of these after-effects) largeness, and tends to take on more outfield-properties.
As the infield of a figure is more attractive than its outfield, a figure projected onto a former infield is not as attractive as a figure projected onto the former outfield. This is because a figure that occurs on a location where another figure has just been before is not as interesting and attractive as a figure that occurs where no other figure has previously been. Which explains why in Fig. 10-2 the square that falls onto the area of the first figure's infield shows the three figure-weak phenomena more strongly than the squares falling onto the first figure's outfield.
The attentional relationships in the case of Fig. 10-2 are shown in Fig.10-3. When the gaze is focused on the fixation point and the visual field is homogeneously illuminated, voluntary attention might be spread across the field with a spatial gradient as shown by the broken curve in Fig.10-3. The inspection figure, once it is projected, is a second source of attention, in this case involuntary attention. Here, the degree of attention continuously decreases from the figure outwards, i.e. into its outfield, according to its own spatial gradient. Both attentional spreads superimpose: that of voluntary ("subject- based") attention starting from the fixation point, and that of involuntary ("object-based") attention starting from the figure. The solid curve in Fig.10-3 shows both the "estimated net-effect" of the two kinds of attention and the different values of attention paid to the locations where the three test figures are projected after the inspection figure has been removed. One finds a certain amount of attentional effect on the inspection figure's projection location, a lesser degree in the immediate vicinity of the initial figure, and even less farther away from the initial figure. The right side of the (broken) curve could move somewhat to the left, as soon as the oblong appears on the left of the previously homogeneous field. Certainly, all this is "theory" and must be tested by experiments.
Figure 10-3. Estimated spatial distribution of attention, depending on
the position of both the fixation point (x)
and the inspection pattern using the example of Fig.10-2
Voluntary attention is directed toward an object by a "conscious voluntary act", and is identical to the "A" in the AX-laws. (With "X", any gestalt factor at Levels 1 to 5 plus attention itself is meant.) Involuntary attention directed toward an object is that which is "unconsciously attracted" by the object, according to the gestalt law Flic Aic, and is identical to the "A" in the XA-laws. Attention of both kinds possesses a spatial gradient.
back to contents