L.Kleine-Horst: "Empiristic theory of visual gestalt perception. Hierarchy and interactions of visual functions" (ETVG). Köln
2001 (Preface and Introduction)
Some similarities and differences between Berlin gestalt psychology
and the ETVG
The ETVG is a "new" gestalt psychology of visual perception. It is most easily understood when compared with the "old" Berlin gestalt psychology. Common to both is, first of all, the inductively obtained belief that every visual experience is not a mosaic of "sensations", as the early structuralists in their "constancy hypothesis" believed (see Part 10) due to the mosaic of the stimulated photoreceptors. Instead, experience is holistic, which means that "the whole is different from the sum of its parts". Second of all, the assumption that a number of "factors" are responsible for this difference; and third of all, the importance of describing visual phenomena before trying to explain them. There are further points of partial agreement between the two theories, but also a number of fundamental differences, as shown in the following. The following quotations stem from Koffka (1935); references to Parts of the ETVG are designated with "Part 0" to "Part 10". In Part 10, some differences between the Berlin gestalt theory and the ETVG are presented in greater detail, encouraged by criticisms that the ETVG has received.
According to the Gestalt Theory, a visual percept is conditioned by two kinds of forces, one of them being the "external forces" or "outer conditions", terms that are synonymous with proximal sensory stimulus energy, the "forces issuing from the retina" (p.139). Koffka distinguishes between these "nerve energies which are...liberated by stimulation" and such liberated "by intra-organic processes" (p.100), i.e. the "internal forces" or "inner conditions" "which will tend to impress on this [stimulus] distribution the simplest possible shape" (p.138). If the external forces would operate alone, then "our perception were nothing but a geometrical projection of the retinal stimulus pattern" (p.139). When the external forces are weak, such as in the cases of "(1) short time of exposure, (2) low intensity, (3) small size, (4) after-images" (p.141), then the internal forces are strong enough to produce dislocations, and "simple, well-balanced figures are perceived when irregular figures are actually exposed" (p.141). Thus, a figure can be "strongly deformed compared with the stimulus pattern, being simpler, more symmetrical, with rounded instead of pointed corners, gaps closed, and even lines which were demanded by the general shape but absent in the stimulus filled in" (p.143).
These deformations caused by internal forces follow a principle that is extremely important to Berlin gestalt psychology, the "Law of Prägnanz". It is "a general, though admittedly somewhat vague, principle", and means: "psychological organization will always be as 'good' as the prevailing conditions allow. In this definition the term 'good" is undefined. It embraces such properties as regularity, symmetry, simplicity and others" (p.110). Koffka reports a number of other "forces", "factors", or "laws": shape-giving forces (p.132), the law of unification and segregation and the forces of cohesion (p.135), the factor of closure and the factor of good shape (p.151), the law of good continuation (p.153), the factor of proximity (p.164), and others. Following is a brief outline of the differences between the Berlin school and the ETVG:
1. Koffka (p.682) quotes Köhler (1929, p.192) when he states that the word gestalt "has the meaning of a concrete individual and characteristic entity, existing as something detached and having a shape or form as one of its attributes." This would define that for which we have chosen the term "formed figure", i.e. part of the visual field, bound off from the rest by means of a contour, and having a certain shape or form. But Koffka broadens the meaning by including the "nature of organization, as it was expressed in the law of prägnanz" (p.682) in its implication. "Gestalt" means "good gestalt" as well, according to Koffka. Thus in psychology there are two gestalt-concepts, that cannot clearly be differentiated: gestalt as a "figure" with any shape, and gestalt as a "prägnant" figure, i.e. a figure with such shape "properties as regularity, symmetry, simplicity, and others"; in short, a "good gestalt". (One can sometimes find a third meaning of "gestalt" in early gestalt psychology: sensory stimulus.)
In the ETVG the noun "gestalt" is not a technical term, and therefore it also lacks "good gestalt", and has no explanatory concept of "prägnanz". "Prägnanz" is only a descriptive term, sometimes used when a percept shows those qualities described above, for example, the "closure" of a line or field, the "continuation" of dots or lines or the "straightness" of a line. In the ETVG, phenomenal aspects like "regularity", "symmetry", and "simplicity" are considered to be too complex or too vague to be suitable for the formation of a theory of visual perception. If one tries to find the true meaning of such phenomenal impressions, "symmetry", for example, could be defined as equal perpendicular location differences between two lines to a (for the most part phenomenally non-existent) third line ("symmetric axis"). Thus, it is not "symmetry" that is a technical term in the ETVG, but "location", "location difference", "line", "perpendicularity (rectan- gularity)", "measurement equality", "two" and "three". Furthermore, in this theory a fundamental distinction is made between phenomenal entities ("gestalt qualities", as just mentioned) and non-phenomenal entities ("gestalt factors" or "gestalt functions") that are the conditions of the phenomenal entities.
2. The Gestalt Theory does not fully list the internal factors that cocondition a visual percept, but instead vaguely finish references to them with "and others". In contrast, the ETVG includes 25 such factors and considers them to be the complete set.
3. The Gestalt Theory seldom includes any description of theoretically relevant interrelationships between the "forces". The ETVG assigns each of the 25 gestalt factors to a certain level of the ten-level functional hierarchy, and thus can specify which lower-level factors play a role in this creation of a particular factor (see Fig. 7-1). The five-level hierarchy containing the six factors that condition the static, two-dimensional figure/outfield perception is described in Parts 1 to 4, the five-level hierarchy containing the quantity, orientation, and form factors in Parts 5 and 6, and the eleven additional factors for depth, time, and motion perception in Part 7.
4. The Berlin school does not discuss in detail how external forces "constrain" the internal forces to impress the simplest possible shape on the stimulus distribution (p.138). The ETVG, however, defines (as precisely as possible) the stimulus conditions needed to constrain a particular gestalt factor to "its" particular "formative effect" on the visual percept.
5. The greatest deficit of Berlin (and Leipzig) gestalt psychology is the inability to explain why a gestalt perception occurs in the first place after objects have been optically projected upon the retina, i.e. why a particular phenomenal configuration takes place due exclusively to the non-configurated mosaic of retinal stimulations, independent of any secondary "dislocation" of parts from this configuration through "internal forces". This question especially arises in the case of "prägnant", i.e. "best" shapes, when a stimulus is such that no "internal forces" are able to "dislocate" any part of it to obtain a percept with a yet "better" shape.
The early gestaltists did (and still do) not provide any answer to this question. As an example: When a square is projected upon the retina, a "square" is seen. Why? Why do we experience a certain number of stimuli within the retinal mosaic as a "borderline", or even as a "straight" borderline? Why do we see "four" such borderlines, and determine them to be of "equal" or "unequal" length, and, even more fundamental, why can we perceive such a thing as "length", why the "sharpness" of a borderline, the "homogeneity" of a "field", and its "enclosedness"? These questions are not answered by the gestaltists, the only answers given are to questions such as why the edges of a slightly trapezoid "square" appear of equal length under weak stimulus conditions, and why its angles appear to be at 90° when they are really 87° and 93°, or why a nearly straight line is seen as straight, a nearly closed line or field as fully closed, a somewhat blurred borderline as sharper, and so on. But no gestaltist, or any other scientist, even tries to explain the actual "primordial" gestalt perception, i.e. the large number of deviations of the perceptual structure from the mosaic structure of stimulations. They do not explain why we can perceive an "angle" at all, or why qualities such as "straight" and "curved", "inhomogeneity", "line", "horizontal" and so forth. And this although they know that the single receptors do not have the ability to produce such qualities at all.
According to the ETVG, it is the gestalt factors that "create", or "produce", the "gestalt qualities" that "compose" a "veridical" percept, and it is these same gestalt factors that are also responsible for the ("illusory") deviations from the "veridical" experience. The primary effect of a gestalt factor is the "qualitative/quantitative-informative" effect, secondary is its "formative" effect which in turn leads to the "normative" effect, i.e. the "Prägnanz" experience, when the "demands" of the formative effects are "fulfilled". This means that the sensory stimuli actually contribute nothing to the "gestalt" perception, they only "actualize" the hierarchy of 25 specific "gestalt factors", whereby the 25 specific "gestalt qualities", or dimensions of gestalt qualities, form, level by level, a visual percept.
6. Wohlfahrt (1932) and other Leipzig gestaltists did empirical "actual-genetic" investigations, as described above. The Berlin gestaltists could not integrate these new facts into their theory, and thus Koffka ignored them as far as was possible, and reported only marginal results from Wohlfahrt: "he stresses the phenomenal instability which appeared as a direct observable property of the figures" (p.143). Wohlfahrt's most important results, however, were the realization that two processes are involved in actual genesis: the "development to object adequacy" and the "tendency toward good gestalt". This experimental result coincides well with the ETVG claim that both the step by step emergence and the intensity of visual qualities increase with the increasing magnitude of the stimulus that triggers these qualities via the actualization of the respective gestalt factors "from the bottom up".
7. Based upon empirical work, Koffka (1935, p.155-159) and the other Berlin gestaltists rejected the empiristic hypothesis which states that practice in seeing a certain figure should facilitate a search task, in which it is required that this figure be detected in a complex configuration. There has been no significant difference found between a group that had practice and a second that did not. The results are indeed conclusive; however, despite all, they do not justify the rejection of memory as a crucial condition for forming percepts. These previous experiments dealt with "explicit memory", the memory bank of experiences, i.e. phenomenally given facts. But there is another kind, "implicit memory", which was unknown at the time, and still is not accepted by all of today's scientists. Implicit memory stores non-experienced, only functionally existent facts, for example, relationships between functional states that condition experiences. The ETVG is able to derive all 25 gestalt factors from one well-known and old-fashioned memory law, by applying it to certain starting conditions for an implicit learning process; the gestalt factors are thus memory contents that are implicitly acquired in early infancy. The gestalt factors for forming the figure/outfield percept with and without depth and time aspects are derived from the relationships and relation-relationships between retinal stimulations (Parts 4 and 7), whereas the derivation of the orientation and form factors is based upon the relationships and relation-relationships between eye muscle innervations (Part 5).
8. The Berlin gestaltist Wertheimer (1923) proposed a small number of factors that are involved in the forming of a "figure", or a "group" (which can be considered a second-order figure). These "factors" became "laws" for Koffka: the laws of proximity, of good continuation, of closure, of good shape, of segregation, of unification, among others. But the number of factors or laws has hardly increased since Wertheimer, i.e. for more than 75 years. All these "laws" were not clearly formulated, there does not seem to have even been any attempt to do so; all have remained vague.
As the 17 ETVG gestalt factors at the first five hierarchy levels are memory contents derived from the retinal stimulus, they are closely interconnected through "associations". Thus the ETVG, too, proposes "gestalt laws"; they represent the 18x18=324 influence relationships between the 17 gestalt factors (plus the 18th factor "attention"), and are formulated as being the relationship between a defined stimulus change and the ensuing phenomenal change (Parts 8, 9). These gestalt laws include the Berlin gestalt laws.
9. When searching for the cause of visual organization, Koffka finds that "...physical organization, spontaneous distributions of process show general characteristics of the kind we are looking for". The soap bubble, for example, "solves a maximum-minimum problem ", as a sphere's "surface is smallest for a given volume" (p.107). According to this, Metzger (1953, p.465) refuses to take into consideration "that the brain consists in a tangle of fibres and cells....more basic is the fact that it is - as the entire living body - a system of finely distributed liquids...." like the "air bubbles on the coffee cup and the blob of fat on the soup...". Certainly, the great experimental neurobiological results of Hubel and Wiesel (1962, 1965, 1968) were unknown to the gestalt psychologists at that time.
This research was also unknown to the author, when he started the ETVG in 1961. But whatever was already known of the brain's anatomical and functional structure, it was purposely not taken into consideration while developing the ETVG. The question asked did not concern the material causes of phenomena, but their possible "functional" conditions, of whatever kind these functions might be. This purely theoretical approach to asking questions was extremely successful, as it led to a system of functions that afterwards proved to be neuronal functions of which a part has in the meantime been empirically found by neurobiologists (see Part 3).
This is an absolutely new theory. Its most striking novelty is based on the theoretical evidence that an actual sensory stimulus does not immediately lead to visual perception, but at first reviews past experiences in order to allow us to not only experience the pure actual modal, spacial and temporal stimulus relationships between brightnesses and colors, but also relate them to the average of similar relationships that have most frequently occurred in the past. Furthermore, this theory is founded on a new model of reality. Whereas the Berlin gestalt theory wanted to overcome dualism by restricting it to monism, the ETVG wants to outgrow dualism by enlarging it into trialism (Part 10).
Gestalt psychology is dead - long live gestalt psychology!
back to contents