L.Kleine-Horst: Empiristic theory of visual gestalt perception. Hierarchy and interactions of visual functions. (ETVG), Part 6, II
Actual-genetic pre-percepts of a figure
1. Wohlfahrt's discovery
With his dissertation of 1925, Erich Wohlfahrt (1932), a student of Friedrich Sander, opened a new field of research. He presented his subjects with bright line figures against a dark background. At first, these figures were so small that the subjects were often able to perceive nothing more than a bright patch with a fixed, and often circular border. It was only after the pattern was enlarged somewhat that the subjects were able to make further distinctions: here, the originally unitary figure contour becomes divided into several segments and then inner contours emerged. Finally, after the stimulus had reached a certain size, the perception was completely differentiated, and the figure was perceived as it would be under "normal" perceptual conditions. This development of the perceptual configuration, from a diffuse-holistic configuration to the final percept, the "end-gestalt", through several intermediate percepts, the "pre-gestalts", was termed "Aktualgenese" by Sander (1928), "actual genesis" by Kragh (1955), and "microgenesis" by Werner (1956). Sander was familiar also with the counterprocess of "Aktualgenese", which he termed "Morpholyse". I prefer to designate this process a "counterterm": "actual lysis". Furthermore, I prefer to avoid the term "gestalt" as a single concept, and use it only in word combinations because, first, the early gestaltists used the term with several meanings: sometimes it was used to designate a "figure", sometimes a "good gestalt", and sometimes the "sensory stimulus". Second, actual genesis and actual lysis do not only occur in the percepts from the fifth level (Fl) onwords, but also effect percepts for which the Fl-factor has not yet been actualized. Thus I prefer to speak of "(actual-genetic) pre-percepts" instead of "pre-gestalts", and of "end-percepts" instead of "end-gestalts". By the way, Metzger (1966) critizised, quite correctly, the use of the prefix "pre", as this is valid only for the actual genesis and not for the counter course process, the actual lysis.
Figure 6-4. Actualization schema of the gestalt factors
involved in one of Wohlfahrt's actual-genetic series (A),
for outer contour (B), and inner contours (C), Stages 1 to 9, of the target
R=rectangularity, M=measurement equality, Q=quantity, Fl=figuredness, f=formative actualization; v and v v=informative (more veridical) actualization, ellipse=progressive stage, square=regressive stage, compared to the previous.
In Fig. 6-4A, the reader will find such a sequence of pre-percepts. The stimulus, depicted above, and as percept A9, was presented not as a black line figure against a white background as shown here, but as a bright line figure against a dark background. The subjects, however, then proceeded to draw what they had perceived as a black figure on white paper. In his work, Wohlfahrt numbered the line ends of the stimulus so that he could indicate which line of the pre-percept led to a given line in the end-percept.
The actual-genetic sequence of percepts of Fig. 6-4A is based on sketches made by a single subject, to describe the sequence of pre-percepts resulting from the stimulus (Stage 9). I, however, have expanded this series. Wohlfahrt's subject, "Ot", began by sketching the figure shown here at Stage 3. As the stimulus was optically enlarged, Ot then sketched Stages 4 to 8. At Stage 8, the subject had the impression that the figure was perceived as it "is". Such a percept was termed an "end-gestalt" by Sander and his students. In fact, the subject's perception of the presentation was not yet as completely differentiated as it would have become after even further enlargement, namely as we perceive it in 6-4A at Stage 9. In order to include this last possible percept in my interpretation, I added the stimulus itself as "Stage 9" to the sequence. (This last-possible percept shall be referred to here as the "end-percept").
Wohlfahrt's subject Ot sketched a comparatively differentiated figure, a regular pentagon, as the first configuration perceived. The author reported, however, that in one third of the cases (the perception of all subjects with all stimuli) a "completely undifferen- tiated bright patch" was seen at first, which "was usually sketched as a small filled-in circle" (p. 364). In order to be able to make use of this common pre-percept, it has been incorporated in Fig. 6-4A as Stage 2 in the form of the outline of a circle; it has not been filled in because we shall, at first, examine the actual genesis only of contours, not of fields.
Butzmann (1940), another Sander student, identified a different stage in his experiments, experiments that were carried out using a different actual-genetic procedure. This stage preceded even the "circlelike" stage. He described the percept appearing in the first phase as follows:
"There remains more or less an impression of brightness, regarding which one can neither say where it is bounded nor whether it is bounded at all. It is something that possesses an extention. To some degree, it even merges with the ground so that a clear ground-figure differentiation is impossible" (p.148)
This contourless configuration has been introduced here as Stage 1.
Only those figures represented in Stages 3 to 8 of Fig.6-4A are thus from Ot. These figures were published by Sander (1928, p.58) in his well known conference report on Gestalt psychology, which is where I found them. What I did not know at that time was, however, that there were several intermediate stages, in the case of Ot, which Sander simply did not mention at all. In this way, he obtained a more salient gestalt sequence. I shall not deal with these intermediate stages, at present, although they will be taken up and interpreted separately in Chapter V.
2. The actualization schema
Reading Sander's report in the late 1950's, I immediately "saw" the course of development of the percept sequence, and proceeded to label the gestalt qualities that were successively introduced, with a couple of letters. I immediately saw that several "Prägnanz tenden- cies", which I was familiar with from other literature, tendencies such as closedness, perpendicularity, regularity, and so forth, were involved. Why these "gestalt factors", as I called them, expressed themselves successively in the percept, with increasing stimulus magnitude, and not simultaneously, was a question I soon asked. I answered it by assuming that the gestalt factors must be hierarchically ordered, and that the higher a gestalt factor is located in this hierarchy, the higher its actualization threshold. Accordingly, increasing the magnitude of the sensory stimulus would cause the gestalt factors to be succesively actualized in their hierarchical order, "from the bottom up".
It is primarily for the four factors Fl, Q, M, and R that the state of actualization is changed during the actual genesis of the target. The actualization schema given in Fig.6-4B and C demonstrates these changes. To the left of the schema, the reader will find the symbols of the gestalt factors involved in the development of form perception, including the uppermost factor of figure perception, Fl. The upper actualization schema shown here concerns the outer contour, the lower actualization schema the inner contours.
A few points regarding the general phenomenal effect of the gestalt functions: the reader is already familiar with the factor Fl, the uppermost factor of the figure perceptual system. By means of this factor, one can detect the closedness (Fl+) of a figure's outer contour (see the example of "the moon in the sky" in Parts 1+2). Moreover, one can perceive also the non-closedness, the "openness", of a contour with this factor. A 3/4-circle does not actualize the factor "closedness" to the same degree as a fully closed stimulus does. There are thus two possible outcomes of actual genesis: either I see the 3/4-circle as a 3/4-circle, in which case I perceive it "veridically", as one says; or I perceive it as a closed circle, as a full circle. The gestalt factor "Fl" is involved in both cases, but in the latter case, it produces a "distorting", i.e. an "illusory", effect, as it causes the stimulus to appear more closed than it actually is. An "intermediate solution", too, is possible here; this results when the factor "closedness" causes the 3/4-circle that constitutes the stimulus to appear "more closed" than it "objectively" is, but not completely closed; an example of a percept thus produced is a 7/8-circle. In a case, in which a figure appears more closed than corresponds to the stimulus relationships, we speak of a "formative" effect of the gestalt factor "closedness".
Assuming that the sensory energy is large enough to actualize the gestalt factor "Fl" in the first place, relatively small energies will actualize the factor only "formatively", if at all, whereas large sensory energies will reproduce the closedness/openness relationships more "veridically", due to the quantitative-informative phenomenal effect of the factor Fl. What is true for the factor "Fl" is true also for every other gestalt factor.
I have termed the factor M the "factor of measurement equality" because its formative effect consists in equalizing the lengths of, and the location differences between, lines and figures, making them more "equal" than they actually "are". (The factor M could partially correspond to the "tendency toward regularity" of the early Gestaltists.) In this way, one can experience a rectangle as a square if the factor M supplies an experience of isometric sides. If the stimulus energy is then increased (by increased size or increased brightness of the stimulus, for example, or if one looks at the stimulus more carefully, i.e. one directs more attention to the stimulus), then one perceives the "objectively" unequal lengths of the sides, by virtue of the same factor M.
The factor "R" is termed the "rectangularity/parallelism" factor because one of its formative effect consists in causing all angles, between converging or intersecting lines, to appear more orthogonal, more perpendicular than corresponds to the actual angle measurements in the stimulus pattern. An angle of 70° may thus appear to measure 80° or 86°, indeed, perhaps even 90°, in which case it appears to be a genuine right angle. An obtuse angle is likewise "distorted" so that it bears more resemblance to a right angle and is not so obtuse, indeed, perhaps even so that it appears to be a right angle.
One can thus speak of both the "(non-veridical) formative effect" and of the "(veridical) informative effect" of each and every gestalt factor. One should, however, not attach too much importance to such designations. There is no such thing as an "absolutely veridical" perception: every (phenomenal) perception is a gestalt perception (Part 1), and thus by definition non-veridical. Nevertheless, we shall retain the terms "veridical" and "non-veridical", as the preceeding discussion has defined them with sufficient precision, even if they are theoretically questionable.
The index "f" shall be used here to indicate a predominantly formative effect of a gestalt factor. This formative effect decreases to the same degree to which the stimulus magnitude (here the size of the stimulus) increases: as a consequence thereof, the perception becomes more "veridical". When the stimulus is seen as in Stage 9, one can call it a completely veridical percept. A perceptual stage with more veridicality than that before shall be indicated by the index "v".
The actual-genetic development of the percept toward greater veridicality, and with a final result of completely differentiated perception, is manifest functionologically as a progression of actualization, as follows:
1. as the "bottom up" actualization of the hierarchically ordered
2. as the transition from a predominantly formative to a predominantly veridical effect of one or more gestalt factors
Conversely, a percept decomposition, that can ultimately result in the disappearance of the percept, is manifest functionologically as a regression of actualization, as follows:
1. as the transition from a predominantly veridical to a predominantly formative
effect of one or more gestalt factors,
2. as the "top down" deactualization of the hierarchically ordered gestalt factors.
In the actualization schema of Fig.6-4, those actualization states that represent an actualization progression, in respect to the preceding stage, have been enclosed in an ellipse. Those actualization states that represent an actualization regression, have been set in a square. In this way, one can recognize the path of progression or regression at a glance.
When I describe individual stages in the following, I shall, on one hand, describe the experience itself, and on the other hand, the functional conditions of the experience that themselves lie beyond the experience; the latter shall be represented as the actualization states of the four particular gestalt factors Fl, Q, M, and R. We shall thus engage in both descriptive phenomenology and explanatory functionology, and thus find ourselves on ground circumscribed by the three-sphere concept. We shall first investigate the actual genesis of the outer contour, then the actual genesis of the inner contours, and finally, the actual genesis of the inner fields, always applying the expanded version of Wohlfahrt's sequence of pre-percepts.
3. Actual genesis of the outer contour
In Stage 1, all that is perceived is "something bright" at a certain location (Pml), a spot of "bright fog". The border between this fog and the dark outfield is still quite diffuse; the brightness gradually merges into the darkness. Under no circumstance is there a definite solid contour between bright and dark. We may thus assume that the gestalt factor Fl has not yet been actualized; indeed, the functions Gml and Ll that cause one to experience a sharp (Gml+) border (Ll+) between two homogeneous (Gml-) sections of the visual field (Ll-), have not yet been actualized either, at least not veridically.
With an increase in the size of the stimulus, what previously was merely "something bright", obtains a definite contour and becomes a part of the entire perceptual field that is bordered off against the remaining perceptual field, namely a "figure". Consequently, all gestalt factors, up to and including the factor "closedness" (Fl), are actualized in Stage 2, for "closedness of the contour" or "enclosing contour" is one gestalt quality that is produced by this gestalt factor. Since gestalt quality "Fl" has only just appeared in this stage, its symbol has been enclosed in an ellipse.
The stimulus in this actualization schema exhibits "openings"; it is not completely closed. Nevertheless, a closed figure is perceived here at Stage 2. We must interpret this perceptual quality as a formative actualization of the gestalt function Fl+; Fl+ "closes" the figure, makes the figure appear more closed than it "veridically" is. (This formative effect of Fl remains present through Stage 7.)
When we want to describe the development of the quantity factor`s actualization, we have to be aware that the "end-percept" is a second-order figure, consisting of seven first-order "line figures", five of which constitute the outer contour of the second-oder figure. Because we assumed that the factor Fl has not been actualized in Stage 1, we begin to count the number of the line figures with Stage 2. Here there is only one (Q1) outer contour, in Stages 3 to 9, there are five (Q5). In both Stage 2 and 3 the percept is more differentiated, in respect to the figure number, compared to the respective previous stage. Thus the Q-symbols are in ellipses; no ellipses are drawn for the following stages, where the actualization state of R remains the same.
In respect to this this example (Fig.6-4A), it is not possible to give plausible particulars about the state of actualization of the three orientation factors, V, H, and T. Factor E ("elongatedness") detects the presented stimulus pattern, as a whole, veridically right from the start, as a (practically) "not elongated" (second-order) figure. From Stage 3 onwards, it is perceivable that the pattern consists of "elongated" (E) figures, namely "line figures". Furthermore, these lines are, without exception, objectively "straight", so that with respect to the state of actualization of both the gestalt factors E and S, no changes occur in the process of the actual genesis. Thus these factors were not considered in the actualization schema.
One might be tempted, however, to believe that the "roundness" or "curvedness" of the (circular) contour has something to do with the factor "straightness". If, however, one contemplates such a relationship, one will quickly reach the conclusion that it does not work, not unless we abandon those principles already established: it is in fact straightness that constitutes the formative effect of the factor S. One must thus expect that, due to the small stimulus strength present in Stage 2, curved stimulus lines will appear straight, instead of the reverse (that straight lines will appear curved). On the other hand, also a veridical effect of the gestalt factor "straightness" is not present here, as the stimulus (No.9) is not curved, instead, it consists of only straight lines. An informative effect of the gestalt factor S would necessarily appear as straightness of the contour. We can only resolve this dilemma by postulating the non-actualization of the factor S. That the Fl-factor, under these conditions, necessarily leads to a kind of "circle"-experience shall be demonstrated later in the section entitled "The formless disk". After the brief discussion of the unproblematic contribution of the factors Q, E, and S, the state of actualization of the factors M and R, and also Fl, will now be dealt with in greater detail.
Once the stimulus reaches a certain size, a regular pentagon is experienced. The contour has become more complex, as we no longer have a single unified contour, but a five-part contour. We thus must assume that, for the outer contour, the quantity factor is from now on veridically actualized at value 5 (indicated by Qv).
On the other hand, the gestalt factor "measurement equality", too, has been actualized, here predominantly formatively, as the (elongated and straight) contour segments are of almost identical length, in contrast to the properties of the stimulus. This is indicated by "Mf".
At this point, one might contemplate assuming the actualization of the gestalt factor "rectangularity" R (more exactly: the function R+), because the contour segments converge at a certain angle. Here, however, difficulties similar to those encountered in the case of the "curved" circle arise. A veridical actualization of the R-factor is not present here, for the angles do not actually correspond to those of the stimulus. A formative effect of the factor R is certainly also not present, as there are hardly "more right angles" in Stage 3 than in the stimulus; in fact, they do not even possess the partial perpendicularity present in the stimulus. How is this dilemma to be resolved? In the same way as the dilemma encountered in Stage 2 was resolved: we must postulate the non-actualization of the "tiresome" factor, in this case the R-factor. The resulting interpretation of Stage 3 as an "angle-less pentagon" shall be clarified later, too.
If one now compares the length of the contours in Stage 4 with those of the stimulus, one will discover that the formative isometry of the sides, which we had found to exist in Stage 3, has vanished; in this stage, the length ratios of the contours have become much more like those of the stimulus. A differentiation of the percept toward greater veridicality, with respect to the gestalt factor M, has thus taken place; this is indicated by writing "Mv".
In this stage, we must assume that the gestalt factor "R" is actualized: the angles here between those lines that had already been perceived in Stage 3, i.e. the angles between the lines, represented by the point sequence 1-2-3-6-5-4-1, largely correspond to the angles in the target. We thus write "Rv".
Both expressions (Mv and Rv) are written within an ellipse, which represents a progression in the actualization of both factors, with respect to the previous stage. In the case of R, the progression consists in that the gestalt factor has been actualized for the first time; in the case of M, the progression consists in that a transition from a predominantly formative effect to a predominantly veridical one has taken place.
Here, the most striking phenomenal change in the sequence of pre-percepts, with respect to the previous stage, is the appearance of an inner contour due to the actualization of factor Q at value 6. The actualization state of the inner contours, however, shall be dealt with later. At present, we wish to concern ourselves with the change of only the outer contour. Phenomenally, this change consists in a reduction in complexity of the pre-percepts, with respect to Stage 4; functionally, it consists in a regression in the actualization of a gestalt factor: as compared to those of Stage 4, the angles between the outer contour lines in Stage 5 are experienced as "more perpendicular". This is indicated by writing "Rf". To indicate this regression in actualization, this expression is written within a square. This regression seems to concern only the factor "R"; in contrast, properties of length may be considered to correspond to those of Stage 4, indicated by Mv without square. Not here, but later, we shall account for the regression in the actualization of R, despite the size of the stimulus being increased.
By the same line of reasoning that we used in Stage 5, we must assume here that the factor R is not (i.e. is no longer) actualized. Accordingly, we are dealing with an outer contour (which is all that concerns us at the moment) that exists in an "angleless" state, as in Stage 3. The symbol "R" is thus, at this stage, dropped from the actualization schema, written as "--" within a square.
The outer contour lines have also become somewhat more isometric. This constitutes a regression in the actualization of "M". Whereas the regression in the actualization of factor "R" manifests itself in the non-actualization of this factor, in the case of the regression of "M" we are dealing with a transition from a predominantly veridical state to a predominantly formative one. We thus write "Mf", and within a square.
Once again, a progression in actualization can be determined, as the angles between the outer contour lines largely correspond to those of the stimulus. We thus write "Rv", within an ellipse. The lengths of the outer contour lines, too, lose their unveridical equality, changing toward greater veridicality. This is indicated by writing "Mv", within an ellipse.
Whereas several lines in Stage 7 were approximately the same length, here in Stage 8 they become veridically unequal lengths. Let us indicate this further progression by writing "Mvv". According to our notation rules, every expression in the actualization schema always refers to the present actualization state, as compared to the previous actualization state. The double "v" here indicates that the effect of the gestalt factor "M" in this stage represents a further advance toward higher veridicality. The angles, too, become more veridical in this stage (Rvv). In addition, the outer contour is no longer closed at 5 and at 6-3 points, where the contour was previously formatively closed; we thus write "Flv".
Had the actualization experiment not been broken off, with Stage 8 as the "end-percept", Ot. would have certainly perceived the stimulus just as we perceive it in this book. The phenomenal change here is the opening of the outer contour at 1, which we indicate by writing "Flvv".
4. Actual genesis of the inner contours
The actualization schema for the inner contours is shown in Fig. 6-4C. The first (Q1) inner contour appears in Stage 5. It is certainly true that Fl is actualized here. One can interpret this inner contour as a line-figure, as we interpreted the outer contour lines First, we are dealing with formative closedness (Flf) here, as the inner fields of the stimulus are indeed open, and not closed. Second, this single line actualizes the quantity factor formatively at value 1 (Q1).
I am unable to state unequivocally whether "M" is actualized here or not. Although the inner contour (which corresponds to the line 3-1 in the stimulus) is just as long as the line 5-4, this isometry is not necessarily due to an effect of the factor "measurement equality", which influence the length of the inner contour. It would suffice for the already assumed formative actualization of the factor R, actualized for the outer contour, with which the parallelization of the edges 5-6 and 4-1 was reached. Then, in the closed figure, the lines between 6(=3) and 1 would "automatically" reach the same length as the line 5-4 had for these edges. Thus, the assumption that M must be actualized for the inner contours, is not accurate.
If, however, R is actualized for this first inner contour, then so is M. It is not possible to actualize a factor without simultaneously actualizing all the lower-level factors, on which it is based. But even if we were to assume that "M" is actualized for the inner contour, we still would not know whether or not also "R" is actualized, i.e. whether or not nearly perpendicular and parallel line relationships between the inner contour and the outer contours result. These R-relation- ships would automatically result from the arrangement of outer contour lines, if a line were to appear between 6 and 1 via 3, as would the actualization of measurement equality. As a precaution, I shall indicate neither M nor R in the actualization schema for the inner contour. (Nevertheless, indicating "Mv", or "Mf", and "Rf" could easily be accomodated by our theoretical expectations, and thus would not produce theoretically "impossible" actualization sequences.)
A second inner contour appears (Q2); it corresponds to line 6-7 in the stimulus. One end is joined to the first inner contour, the other end is joined to an outer contour line. This second inner contour divides the lower field into two equal parts since it is joined to the middle of both the first inner contour and the outer contour line. We can assume that this is a formative effect of the measurement equality factor, indicated by writing "Mf". Perpendicularity, too, is entirely formative here. Especially pronounced is the non-veridical and thus double formative perpendicularity between the second inner contour and the contour lines, which it adjoins to; we thus write "Rf".
A further progression occurs in Stage 7. Those factors that are already actualized produce a pre-percept that allows a more exact and more differentiated interpretation of the stimulus. This is indicated with "Mv, Rv", as both the line lengths and the size of angles, with which the lines converge, have changed toward greater veridicality.
In this stage, the development of the percept is almost complete. Whereas the progression toward veridicality of lengths and angles particularly affected the second inner contour in Stage 7, in this stage both the length of the first inner contour and the angle with which the first inner contour is joined to the outer contour, become more veridical. Here we must write "Mvv" and "Rvv". In this stage, the openings at the outer-contour corners 5 and 6-3 occur; we thus write "Flv".
The additional opening at corner 1 can be best symbolized with Flvv.
Note: Wohlfahrt (1932) recognized two directions in actual genesis, which can best be seen when the inner contours are introduced into the percept.
"Refinement of the actualgenetic interpretation"
back to contents