L.Kleine-Horst: Empiristic theory of visual gestalt perception. Hierarchy and interactions of visual functions. (ETVG), Part 6,III
Refinement of the interpretation
1. Further agreement of the theory with fact
After having indicated the postulated actualization states of four gestalt factors that, as assumed, exist in a particular hierarchical relationship to one another, we now wish to ask to what extent our theoretical expectations correspond to fact.
According to the actual-genetic concept, the development of a percept with increasing stimulus strength (here the enlargement of the stimulus pattern), must constitute a bottom-up progression of the actualization of the hierarchically ordered gestalt factors. Since we have indicated a progression by drawing an ellipse around the symbols of those gestalt factors that exist in a progressing state of actualization, the question may be answered immediately: each stage is distinguished from the previous stage by a progression of the actualization of the gestalt factors, provided that one considers the development of the outer and the inner contours together.
Two stages depicted in Fig. 6-4 are not characterized only by progression; in these stages, a regression occurs, too. Whereas the progressions concern the inner contours, the regressions concern the outer contour. Weighting the instances of progression and regression, one sees that here, too, instances of progression predominate: in Stage 5, progression consists in the introduction of a further line and a new gestalt quality (Fl), but regression consists in only the transition from a veridical to a formative actualization of a single (outer contour) factor. The total regression in Stage 6 is stronger than in Stage 5, consisting in the cessation of the actualization of an outer contour factor. In contrast to this, however, we have a progression that equally consists in the introduction of a further line and a new gestalt quality (Fl) as well as the introduction of two new gestalt qualities for both inner contours (M,R). Instances of progression thus predominate in accordance with theoretical expectations - i.e. they do not contradict them at least.
Counting instances of progression and regression (i.e. the number of ellipses and squares, respectively), one obtains a progression to regression ratio of 24:3. One could easily accept this excellent result as a confirmation of our theoretical expectation of progressive development with increasing stimulus magnitude. Yet with an additional assumption, the three instances of regression here, too, may be interpreted as a confirmation of theoretical expectations. This shall be demonstrated in Chapter V.
It can be assertained from the actualization schema that the actual genesis of a percept is not only a purely quantitative increase in actualization. The actualization also turns out to be the actualization of particular gestalt factors, each of which introduces a particular gestalt quality into the visual percept. Thus, the sequence of pre-percepts also fulfills theoretical expectations regarding the postulated hierarchical relationships between gestalt factors. Whenever a new quality was introduced into the percept, it was always a quality that corresponds to a higher-level gestalt factor (higher-level with respect to the levels of the already actualized factors).
Correspondingly, in the single instance in which a gestalt factor became completely deactualized, the factor was the highest-level factor among those that had been actualized. Only the highest-level factors can become deactualized; it is not possible for a lower-level factor to become completely deactualized, so long as a higher-level factor that is built up upon this lower-level factor remains actualized.
If in the course of concrete interpretation one believes that it is necessary to deviate from this law of succession, it follows that a mistake has been made in interpretation. It is possible that an incorrect hierarchical position for a particular gestalt factor has been assumed. It is also possible, however, that one is merely dealing with an apparent deviation from the law of succession, especially if a more complex perception, with interlocking partial hierarchies, is involved (to give examples here would lead us here too far astray). Since the ETVG postulates strict laws, rather than rules with exceptions of every kind, it is empirically confirmable to a high degree, as is fitting for a good theory.
All those actualization states in the schema that are marked by neither an ellipse nor a square, i.e. those states that have not changed with respect to the previous stage, equally do not contradict the ETVG predictions of bottom-up actualization and top-down deactualization, respectively.
It is also not a contradiction of the theoretical requirements that the actualization of gestalt factors takes place in a particular sequence when, in the course of a concrete actual-genetic experiment, a successive stage is characterized by the new actualization of not one, but several, gestalt factors simultaneously, as occurs repeatedly in the Wohlfahrt sequence. The "fineness" of the succession of gestalt qualities is determined by the fineness with which the stimulus magnitude is increased. Wohlfahrt increased his stimuli by 25% from stage to stage. Finer increments of enlargement probably would have resulted in a more finely differentiated course of actual percept development.
Even the smallest stimulus pattern size chosen by Wohlfahrt for his experiments was very large in the case of Ot, who plainly responded very sensitively to the actual genesis; Ot did not see the pre-percepts given Stages 1 and 2 in our actualization schema, although this would have been possible with a yet smaller stimulus pattern.
Likewise, an enlargement beyond the "end-percept" (Stage 8) would have allowed the subject Ot to perceive the stimulus as we perceive it here in Stage 9: with a small opening at corner 1. Part of Wohlfahrt's method of investigation consists in breaking off the experiment as soon as the subject had the impression that the perception was veridical, regardless of whether this was objectively true or not. This was the stage (here: Stage 8) he termed "endgestalt".
2. Actual genesis of the infield
Above, the factor "Fl" was treated as "closedness of the contour"; the phenomena resulting from this factor were interpreted accordingly. But as the reader is aware, "contour closedness" is not the only gestalt quality of "Fl" (Part 2); this factor possesses also infield and outfield properties. We shall neglect the outfield here, but a few points regarding the infield need to be added. Both qualities, "infield" and "contour", belong together and form the unity "figure". Let us then complete both the functionology and especially the phenomenology of the actual genesis of the "figure" experience, by describing the actual genesis of the infield.
The drawings of Stages 3 to 8 in the actualization schema stem from the subject Ot. I have added drawings 1 and 2, on the basis of general observations of actual-genetic experiments. Wohlfahrt (1932, p. 364) himself reported a stage of "completely undifferentiated bright spots" which he often obtained as a first stage, although it did not appear with "Ot" in the case of this particular stimulus.
The brightness fills the entire interior of the perceived figure, although only a bright (luminous) line drawing was presented. It is thus not Stage 2 of Fig. 6-4 that is seen, but a configuration such as that depicted in Stage 2 of Fig.6-5. This configuration, which was actually perceived in actual-genetic development, corresponds exactly to that configuration which we termed "figure" in our virtual actual genesis, and is thus a percept of the type "moon in the sky". The emergence of a filled-in figure actual-genetically earlier than that of a line-figure is in agreement with the statements in Chapter I, as a filled-in figure is a first-order figure, and a line figure is a second- order figure. In Part 8, an additional cause is described: the thresholds within a figure, i.e. in the infield, are higher than those in the outfield. Thus, the same factors that produce both the outer contours and the inner contours are, when increasingly activated, able to overwhelm the outer contour threshold more easily, i.e. earlier, than that of the inner contours. The process of filling-in itself is explained in Part 3.
Figure 6-5. Sequence of pre-percepts in the actual genesis of the inner contours and fields
From the following reports given by the subjects, the reader will see that the phenomenal state of the infield being "filled-in", and thus the non-veridical/formative effect of the gestalt factor Fl (with the actualization of Ll- and Gml- for the infield), remains up to and including Stage 4 of Fig.6-5, after an increase in stimulus magnitude and before more veridical brightness relationships begin to appear, in the form of the emergence of inner contours.
The existence of the predicted perceptual stage, in which only the factors below Fl are actualized, can be demonstrated also from actual-genetic experiments, specifically the experiments of Butzmann (1940). Butzmann, too, worked with luminous line drawings, but used a different stimulus amplification procedure. He first presented the figures far out in the periphery of the perceptual field, and then moved them toward the focus. At first, diffuse-holistic percepts appeared, which became differentiated, step by step, as the stimulus projections approached the fovea. His subjects saw an undifferen- tiated configuration with a definite contour (Fig.6-5, Stage 2) prior to all differentiations of the outer contour. Butzman's subjects, however, saw also an earlier "first stage" which is shown here as Stage 1 in Figures 6-4A and 6-5. This first phase was described above. Butzmann's subjects described what they perceived as "a diffuse impression of brightness", "approximately round fog", "circularly blurred without definite borders", and "a dispersed gleam of light" (p.248). In the actual-genetic Stage 1, we thus find a certain confirmation of the assumption that in the perceptual actual genesis of a complex stimulus source, a phase must occur in which one can see brightness (Pml) and brightness difference (Dm), but not (yet) clearly distinguishable inhomogeneous (Gml+) and homogeneous (Gm-) areas, nor lines (Ll+), which are clearly to be distinguished from fields (Ll-).
Up to and including the clear "figure" stage (Fig.6-5, Stage 2), an actualization of the first five levels of the gestalt perception system has occured. With yet further increase in the stimulus strength (regardless of how this increase is effected), other higher-level, and thus higher-threshold, gestalt factors are actualized step by step; as we have already seen in the case of the outer contour, these are form factors. First, the form factors for the perception of the outer contour are actualized. At some point in the development, the same form factors begin to allow the inner contours to appear as well. Let us consider what Wohlfahrt (1932) said regarding the stage of the definitely-contoured disk. This is not a description of a particular "Ot"-stage but a summary of frequently occurring observations by all subjects, and with all stimuli (with "gestalt" being identical to our formed or non-formed "figure" in the widest sense of the word):
"As the next step, after the direction in which the gestalt chiefly extends can be stated, the gestalt appears with a very simple, rounded outline that is closed off against its surroundings." (p.365)
We can assume that the figures up to and including Stage 4 are still solid figures and not line figures, although the subjects usually drew the solid figures perceived as line figures. The transition from a figure that is perceived as solid, to a figure that is perceived as linelike (and thus perceived more veridically), becomes clear in Wohlfahrt's general description :
"What subsequently brings a first ordering into the 'surging undividedness' of the gestalt interior are the inner fields of the gestalt, led by the largest of these. The form, the location, and the grouping of these completely govern the gestalt interior in this stage of analysis. They are not holes...they are not a 'transparent background' but 'dark nuclei', first results of an incipient crystallisation; qualitatively incomparable with the subground from which the gestalt as a whole has just closed itself off." -
"The 'bright bars' ("Stege") between the nuclei often harden as quick as lightning into lines, mostly into straight lines... The dark, extended (fieldlike) nuclei lose their importance as structuring agents ("Gestaltmomente") and cede this role to the lines." (p. 367/8)
From this description by Wohlfahrt we can subdivide Ot-Stage 5 into the Phases 5a and 5b and Stage 6 into the Phases 6a and 6b (brightness values must always be "rethought": what is white here was dark in the experiment, what is black here was bright in the experiment.) That the darkening (in the drawing the brightening) originates in the middle of the field and proceeds toward the contour and not the reverse, is based on a law as well, but shall not be dealt with here.
3. The "formless disk"
I would like to return to Stages 2 and 3 in Fig. 6-4 for which it was necessary to postulate the non-actualization of higher-level gestalt factors. This was necessary because an assumption of actualization - regardless of whether "f" or "v" - would have led to a theoretically "impossible" interpretation.
When I began to interpret Wohlfahrt's sequence of pre-percepts in the late fifties, I thought that a circle is a curved, or round, line that encloses a field, and that a circle is symmetric, more so than an ellipse; whereas the latter is bilaterally symmetric along two axes, a circle is radially symmetric. In order to be able to perceive a symmetric figure such as the circle, I thought it necessary that the factors "straightness", "measurement equality", and "rectangularity" be actualized, as a circle is "round" (S), and has equal (M) diameters that enclose all possible angles (R).
As I then attempted to enter Stage 2, interpreted in this way, into the actualization schema according to this assumption, I encountered difficulties: it was not possible without contradicting theoretical statements, which had already proved themselves to be fruitful. I thus began to search for arguments, with which the steps from Stage 1 to Stage 4 would constitute a progression. I found the following interpretation: it is true, it is possible to interpret the configuration drawn in Stage 2 as a geometric circle that can be perceived with the actualization of M and R. But this interpretation is not conclusive. If I see something, something bright, that I perceive only as being bordered off from its surroundings by a contour, and nothing more, so that I cannot perceive any peculiarities of form, no dents, no bulges, and thus no part of the entire contour that looks any different from any other part, then this experience can be assumed to be conditioned by the actualization of only the partial hierarchy up to and including the Fl-level. The circlelike "form" of this percept is not produced by any form factors, but by the factor Ll. If, as in this case, there are no other factors to dictate another route, only the factor Ll, with its law of "minimizing the sum of Dls" (Part 2), determines the route a line takes in order to border off an (in)field against its enclosing (out)field.
4. The "angleless pentagon"
We encounter a similar situation in Stage 3 of Figure 6-4: certainly, it is possible to perceive a regular pentagon under the actualization of the gestalt factor rectangularity/parallelism (which "measures" also angle size). But it is not necessary that R be actualized in order to experience "regular pentagon"; it is sufficient to actualize the factor hierarchy up to and including straightness (S) and measurement equality (M). In this case, the factors S and M would differentiate the percept of Stage 2 by simply dividing the one (Q1) contour into five (Q5) parts, and show these veridically straight and formatively of equal length. One can hardly describe or draw such lines other than a regular pentagon. Such a pentagon is "angleless", so to speak.
Here, too, it seems that percepts that are phenomenally the same are functionally different. Nevertheless, in account of the functional- phenomenal correlations, one must expect that the "angleless" pentagon looks "somehow" different from the configuration with five "genuine" angles, perceived under the additional actualization of the gestalt factor R.
The subjects did in fact make statements that point to the existence of such a difference. Almost every pre-percept, i.e. a subjective configuration that is not yet "veridical", is perceived with a shade of the "temporariness", with a feeling of "non-finality". In contrast, the "end-gestalt" is felt to be "final", to be "the completion of the assigned task" (Wohlfahrt, 1932, p 369), even when the exact "veridical" perception had not been achieved.
back to contents