physiologically the inner dynamics of optical processes, we cannot
do more at present than draw conclusions from the properties of the
visual field in comparison with given retinal constellations. In the
course of this attempt we find that “form,” wherever it exists, is a
_supra_local property of that part of the field; so the property of
the underlying process must be a supralocal phase of it. We find also
that “form” is a property exclusively of segregated wholes; so, with
or after the organization of these wholes, that special supralocal
property of physiological process must develop which is the basis of
“real form” in direct experience. As it is an attribute of detached
wholes, this phase of the process exists only in definite areas, which
are determined by certain relations of stimulation. These relations
must be decisive _dynamically_ for “form,” as they are for “wholes”
which have “form” as their most important visual property. Evidently,
one cannot understand “form” from the viewpoint of mosaic theory;
dynamical self-distribution with its functional coherence and structure
of process is the only principle known at present, which may give us an
explanation of “form” in the future.[38]
I shall not try to demonstrate with the same care the concrete reality
of definite forms in _time_. What has been said would have to be
repeated more or less in the case of melodies, of rhythms, of seen
movements and so forth. The “form” of a musical _motif_ begins at a
definite point and ends at another; then another _motif_ may follow.
But there is no real form extended from the second tone of the first
figure to the third of the second figure; and between the two figures
there is what is called a “dead” interval (corresponding, as “empty”
time, to the mere extension or ground outside a visual form). Again
when in the dark room a bright spot in motion describes the path of
Fig. 6, we see certain definite forms of movement, 1 → 2, 3 → 4, 5
→ 6, for example. We do not see other forms, such as, for instance,
a form corresponding to a fraction of 1 → 2, the horizontal, and a
fraction of II, taken together. Real form coincides with the wholes and
sub-wholes segregated in this case as in that of melodies and rhythms.
In passing we may remark that the behavior of others is continually
seen in _forms_ of events which correspond to the natural grouping of
what we see or hear of them.
As form is a property of segregated wholes, what has been said against
the explanation “by meaning,” or other indirect explanations, in the
case of segregated wholes, will be more or less to the point with
respect to their forms, though, probably, for most readers this is yet
a rather critical issue. Therefore, the problem shall be treated once
more in the following paragraphs.
[Illustration: Fig. 11]
1. What is the effect of our past experience of certain definite
forms upon visual experience in subsequent life? Drawings like Fig.
11 and Fig. 12 contain a great many geometrical lines the outline
of which, when given alone, would make us see other forms than those
which we see naïvely. Thus, in both of them the outline of Fig. 13
is present geometrically. If now we have a large number of drawings,
seen naïvely in a certain way, and besides them certain other figures
geometrically contained in the first, will training or repeated
experience with regard to the latter change the way we see the first,
so that the learned forms become “real” when the larger drawings are
shown? Gottschaldt has recently made such experiments.[39] Since past
experience is supposed to have its effects upon form independently
of our knowledge about the presence of corresponding outlines, i.e.,
automatically, the subjects were not told to analyze or to look for
the “learned” forms. The larger drawings were given them simply for
description. Under these circumstances, with three previous exposures
of the smaller forms, the training had no effect upon subsequent
perception of form in more than 90 per cent. of the cases. When the
number of previous exposures of the smaller forms was increased to 520
with new subjects, the result was still the same, the drawings being
seen unchanged by the training in 95 per cent. of the trials. Not even
the few cases, which showed a positive result, can be explained by
the mere fact of previous training, because all those subjects, who
occasionally saw the “learned” forms in the larger drawings, had some
suspicion about the aim of the experiment and asked the experimenter
whether they should look for the “learned” forms in the larger
drawings. Though they were not instructed to do so, of course they
went into the test with an attitude of definite expectation. So the
few positive results do not prove an automatic after-effect of past
experience upon forms seen subsequently, since, apart from training as
such, there was this other factor of expectation.
[Illustration: Fig. 12]
[Illustration: Fig. 13]
Gottschaldt’s larger drawings were “difficult,” i.e., of a highly
stable organization. In some of them, in spite of strong efforts, I
cannot really _see_ the smaller form, even when I know about its
geometrical presence and position. However, if any one criticizes this
“difficulty,” by his own argument he recognizes the reality of stable
visual form as something stronger in such cases than a very large
amount of previous experience. Furthermore, the larger drawing (Fig.
12) cannot be said to have its stable organization in consequence
of much experience in previous life outside the laboratory. What we
naïvely see there is not better known from previous experience than the
small figure 13. After these results, whoever defends the automatic
influence of past experience upon our seeing definite forms, will
have incumbent upon him the task of supporting his theory by other
experiments. If such an influence exists, it must be restricted to
rather special cases.
2. Where “form” _exists_ originally, it acquires a meaning very
easily. But here a whole with its form is given first and then the
meaning “creeps into it.” That meaning automatically produces a form
where beforehand there is none, has not been shown experimentally in
a single case, as far as I know. It may be that in a very _unstable_
constellation, in which a certain form can be seen or organized, past
experience of such a form will tend to produce it really, whereas
without that previous experience, this would not happen. Even in this
case, however, we should still have to explain what factors produced
that form in previous life. It is only assumed, then, that conditions
were more favorable, and the question remains open whether they were
not favorable for its originating directly. In any case, even granting
that previous experience of a certain form favors its appearance in the
future, we ought to realize that such an occurrence will be limited to
definite cases, namely those in which the actual constellation does
not tend decidedly toward other more stable wholes and forms. The
number 4, for instance, is certainly a well-known form; but when Fig.
10 is shown to hundreds of people without any special instruction or
warning, only a few of them will see the 4 and mention it in their
description of the figure. If any one protests that we have never seen
the 4 in such an environment before, he misses the point, because
no explanation in terms of previous experience is needed when 4 is
given upon a homogeneous ground, as it usually is. If past experience
has an influence, it is exactly in such a case as that of Fig. 10,
in which its influence ought to be effective. And after all, it is
not the “unusual” environment as such which prevents our seeing the
4 immediately. In the following figure (14) the 4 _is_ seen at once,
though in my judgment the environment is not less unusual than in the
first case. Why, then, is it seen now? Because the relation between the
added lines and the geometrical parts of the 4 is not such that these
parts are absorbed in the formation of other wholes. But this is a
principle of original organization. So we may say that any influence
of past experience upon the organization of forms in the future depends
upon the degree and the stability of sensory organization in the actual
case. I shall give a few more examples which show the absorption or
destruction of a well-known form by actual organization into other less
known forms.
[Illustration: Fig. 14]
[Illustration: Fig. 15]
Look at Fig. 15 before you read the next lines. No one would use the
name of letter E in describing it as long as he is a naïve observer,
though the well-known E is present geometrically, and the form really
seen is less known. Fig. 16 may be beheld as an ornament for months
before one first finds two H’s in it. Similarly the letter K is
hidden in Fig. 17. If the reader is already a little sophisticated, he
would be wise to show these figures to some of his more naïve friends,
for five seconds perhaps, asking them only to give a spontaneous
description of them. I do not think the enormous experience we all have
with respect to these letters will produce any considerable effect upon
the result, as long as the subjects do not begin to look for hidden
forms. If they do, of course, the result will be produced by actual
attitude apart from previous experience. When we give our subjects a
hint about the camouflaged form, we create in them what the physicist
might call a “_special vector_,” favoring that concrete form. But when
our subject, adopting an indifferent attitude, looks unaided at some of
the unknown figures, this vector does not exist, despite a great deal
of “past experience.” Without it, the hidden forms do not appear.
[Illustration: Fig. 16]
[Illustration: Fig. 17]
3. Again, some psychologists will go so far as to derive all form in
the visual field from tactual or motor experience. Our answer in this
case is once more simple enough: Real form is a concrete property of
certain things. If its history is to be traced to other than visual
experiences, we must have concrete real form in those other regions
in which it has its origin. Nothing can be imported from other fields
into vision without having originally occurred in these other fields.
So the whole problem would only be shifted from one sensory field to
another, and the explanation would have to assume original forms in the
latter. This reasoning applies also to our sensations of eye-movements.
If visual form shall be derived from them, real form must be a property
of some of these kinesthetic experiences. Therefore, such an hypothesis
does not seem to be a scientific gain, and we may as well acknowledge
the problem of visual form _as such_.
4. I find a decisive argument against explanations of form in terms
of past experience, in the following consideration. No doubt, parts
of the visual field can reproduce past experiences associated with
them in previous life. But when we ask what factors are the main
reproductive properties in these cases, we see that in 99 out of 100
cases _reproduction depends upon the form of a segregated whole in the
field_. Take the concrete real form away, so that reproduction can
be aroused by only the color or the place or, perhaps, the size of an
area, and you will realize that for the most part, without form, visual
realities are not specific enough to reproduce definite experiences.
When we talk about the automatic influence of past experience upon
actual vision, we are tempted to extend this idea to the creation of
definite “form” by reproduction. But then we forget that there is not
much reproduction without form itself as a reproductive factor, and
that, from the viewpoint of such an hypothesis, form _as such_ must
be strictly excluded from the sensory field and, therefore, from the
reproductive features of the situation, supposedly being present in the
field _only by way of_ the reproduction of past experience. The curious
result is that we presuppose “form,” erroneously and unawares, in order
to get form reproduced from past experience! A naïve observer does not
see the 4 in our example, although it is such a well-known figure. Why?
Past experience cannot work if there is no specific factor present
which might reproduce that definite part of past experience under the
circumstances of the case. In our example the ever so well-known 4 will
not be reproduced, as, in other examples, countless other well-known
forms are not reproduced either, there being nothing in the actual
optical processes which might effect that definite reproduction. I
suspect that a great many explanations in terms of meaning and past
experience fail completely to realize this side of the question.
It is easy to say: This, of course, is so because in earlier life
certain things have happened. But the time has come for more concrete
thinking in this field. Reproduction does play a highly important rôle
in mental life and in behavior. It can play it only if the sensory
world is sufficiently endowed with concrete properties in those areas
which are said to reproduce our previous experience. If it is not,
because we regard it as a mosaic of sensations, by what factors can a
definite reproduction be determined? “Transposing” shows that a form
will remain the same independently of the color, the place and the size
of its area. If, now, a definite form is said to appear in the field
by reproduction, I do not see what characteristic of the mosaic can
possibly determine the right reproduction, since we cannot rely upon
color, place and size for it.
In order to show that “form” can never be explained by the existence
of “sensations,” von Ehrenfels laid great stress upon this very
fact--mentioned previously by Mach--that a form can be “transposed,”
i.e., will remain the same as a visual property, after its brightness,
color, size and place are changed. Of course, when shifted too far
toward the periphery of the field, a form will change its character
more or less, but, excepting this case, the possibilities of
transposing a form are very numerous. Again in this respect temporal
form behaves like spatial form, since a melody may be rendered in
different keys and still remain the same, so far as its musical form
is concerned. Indeed, nothing can show better than these facts that
the specific supralocal organization of the processes underlying form
as a sensory experience must be a physiological reality, certainly as
much as are the processes underlying chroma as a quality of sensory
experience. Since, in the case of vision, we can change not only
chroma, but brightness, localization and size of a form, without
disturbing the form itself, the usual categories of quality, intensity,
position, extension do not help us when real form is to be explained.
There is only one thing we cannot change too much without influencing
form itself, viz., that special set of relations of stimulation which
seem to be decisive for the segregation of a definite whole and for
the specific dynamical structure underlying form.[40] We do not
normally _experience_ these relations in the sensory field. But in the
determination of self-distribution of process, in the segregation and
“formation” of wholes, they must become almost all-important somewhere
between the retinæ and the psychophysical field.
Up to the present time there has been a tendency to regard the
remarkable properties of wholes, especially the possibility of
transposing their translocal properties, as the achievement of
“higher” processes. From the viewpoint of _gestalt_ theory sensory
organization is as natural and primitive a fact as any other side
of sensory dynamics. If, for von Ehrenfels and the _Grazer Schule_,
_Gestaltqualitäten_ and form are treated as a superstructure built upon
the foundation of sensations, this assumption would fail to account
for the evident rôle of _gestalt_ in such biological processes as
ontogenesis. In psychology, too, the influence of _gestalt_ has been
demonstrated by Hertz (cf. Chapter V) in very primitive behavior,
in which no one would assume the operation of “higher” processes.
Lashley seems to have been the first to find that an animal, trained
to react positively to one of two “stimuli,” will shift his reaction
spontaneously when he finds himself before two other stimuli of the
same class, the new reaction being directed to that stimulus which
plays the same rôle in the new pair as a whole, which another specific
stimulus played in the pair used during the learning process. I wonder
why these experiments failed to exert their due share of influence
upon current theoretical formulations. Without knowing Lashley’s work
I have repeated the same experiments on apes and chicks, with special
care to exclude all possibilities of indirect explanation. There is no
doubt, now, that a chick, trained with two grays, I and II (II being
darker than I), always to choose II, will, after a while, when II and
the new (darker) gray III are given, in the majority of the trials
not choose II but the unknown nuance III. The same experiments were
performed on apes with size, and also with different hues of color.
Several investigators have been able to confirm these experiments. We
may conclude that animals react to such pairs as to wholes, either side
of which has a definite character depending upon its “position” in the
whole. So II is “the dark side” of the first pair; then, in the new
pair, III assumes this rôle, and since the animal has learned to choose
the dark side of the pair rather than a more or less definite gray, it
tends to avoid the specific gray of the learning period and to choose
the new gray after the transposition. Again it does not matter at all
in such a case, whether or not we assume that the chick has direct
experience. The difference between choice depending upon a more or less
definite intensity of light and reaction depending upon one definite
side of a structure, is the same for either assumption. Once more the
formula of “stimulus and response” becomes altogether misleading, since
it neglects the fact that between the stimuli and reaction there is the
organization of definite wholes with specific properties, upon which
the reaction so obviously depends.
In order to prove that dynamical self-distribution will explain
“transposition,” we have to show that “transposition” is possible in
self-balanced physical systems. Nothing can be easier! First of all,
if under certain conditions all forces of a dynamical distribution
are in equilibrium with regard to each other, it is evident that the
equilibrium is not disturbed if the intensity of all those forces
decreases or increases in the same proportion. _All_ forces being
reduced to one-half of their first value or increased to three times
that amount, they will balance each other as before. The supralocal
dynamical structure is largely independent of absolute intensities.
Now let the dynamical structure be the self-distribution of current
in an electrolyte, depending upon the total form of the electrolytic
conductor. Again the intensity of current has no influence upon its
distribution. Also, if instead of ions like Na and Cl, K and Br or
any others carry the electric charges, the distribution of current
will not be changed. Or, take as an example the electromotive
phenomena developed when two solutions (I and II) of different ionic
concentration are in contact. Those phenomena, existing only in the
pair of solutions as a whole, depend upon the _relation_ of the ionic
concentrations, the absolute concentrations being indifferent. If, in
a definite case, the solution II with a concentration of 1/20 n is the
electropositive side of the pair, as against I with a concentration,
say, of 1/4 n, in a new pair with the concentrations 1/20 n (II) and
1/100 n (III) the new solution III becomes the electropositive side.
To be “the electropositive side” of such a physical system is no less
a _gestalt property_ in a definite electrochemical whole than to be
“the dark side” is a _gestalt property_ in a sensory pair.
In some respects form is the most important property which a whole
may have; the introduction of it as a specific reality will help us
to understand certain of the properties which occur only in wholes.
In Figs. 8 and 9 we observed a change of form, now one cross or star
being present as a form, and then the other. But something else
changes at the same time. When we see the slender cross, the area of
this cross has a character of solidity and coherence; we may even
say that this slender cross has the substantiality of a “thing,”
whereas the environment appears as comparatively “empty” and “loose.”
Exactly the contrary is true when the other cross dominates the
field; the large cross becomes “solid” and “substantial,” whereas the
narrow angles have become “loose” and “empty,” like the environment.
Because an area becomes “solid” when appearing as “form” or “figure,”
Rubin, who was the first to describe the difference, has called the
“solid” quality the “figure” character, whereas to the “looseness”
of the environment he gave the name of “ground” character. This term
“ground,” or “background,” is the more appropriate since we find the
“figure” protruding a little, as if in three-dimensional space, and the
environment localized behind it, so that the solid “figure” appears
as something apart, behind which the “ground” seems to extend without
interruption like a simple homogeneous plane. The sky above the houses
(cf. p. 202) has the character of ground.
I do not think we ever find that character of “solidity” where we
do not have a segregated whole. That character, then, is a typical
property of segregated wholes and belongs to von Ehrenfels’ qualities.
Some psychologists will try to derive this character from touch or
other previous experiences which we have had with things, in the
practical meaning of the word. We find that all visual things, which
as such are clearly perceived in the field of vision, have the same
character of “solidity.” But as it appears as a visible quality, it
may be such a quality _originally_; it may even belong, independently
of other experiences, to the primary constituents of the meaning of
the term “thing” in common life. Whether the empiristic explanation is
justified or whether this quality is founded directly upon the specific
dynamics of seen form, will be decided by further research. As yet
it has been shown that, functionally, “figure” and “ground” behave
differently. Constancy of color has been shown to be more striking for
“figure” than for “ground” (Rubin); the limen for a patch of color
projected upon the “figure” was found to be higher than upon “ground”
of the same objective intensity (Gelb and Granit); after-images are
more vivid when observed upon a “figure” than they are upon “ground”
(Frank).
* * * * *