passively accept what I find before me as the sensory field. Then
analysis is possible in the meaning just defined. I may, however,
adopt a special attitude with regard to the field, selecting some of
its members and more or less suppressing the rest. In many cases a
change of organization will be the consequence of such an attitude,
and hence “analysis” of this sort involves a real transformation of
sensory facts in _gestalt_ psychology (cf. Chapter IV, p. 124). Of
course, an analytical attitude is not the only one by which a change
of organization may be produced. When we select certain members of
the field, we can keep them together at the same time and so favor
one special kind of “belonging together” instead of that which would
prevail without our interference. Again the change produced by our
attitude will be a real transformation.
From the viewpoint of _gestalt_ psychology a change of attitude
involves a definite physiological stress exerted upon a sensory field
by processes originating in other parts of the nervous system, and to
some degree the organization of the field may yield to it. The figure
7, for instance, is seen normally as a symmetrical form. By picking out
the lines marked “a” however and keeping them together I can almost see
the figure 7a, the lines marked “b” being repressed. In the same way
I may favor the lines marked “b” and, as it were, create the figure
7b. How concrete and real such a change is, will become apparent if we
consider the point which is the center of the figure objectively. When
we produce the figure 7a by favoring the “a,” that point is shifted to
the right, as it also is, of course, when the lines marked “b” are not
drawn at all. It is shifted to the left when we produce Fig. 7b.
[Illustration: Fig. 7]
[Illustration: Fig. 7a]
[Illustration: Fig. 7b]
[Illustration: Fig. 8]
In some cases sensory organization seems to change without any
influence being exerted upon it from without, simply because processes
which remain the same for some time in the same part of the nervous
system tend to alter conditions in that part and to block their own
path. We know that the same thing occurs in electrolytical cells in
which the current polarizes the electrodes and thereby creates forces
opposed to itself. In Fig. 8 we are confronted by a pattern formed
by three narrow sectors. Looking at the center of it, most persons
will suddenly see another pattern after a while. In the new shape
those lines, which belong together in the first pattern as contours
of one arm, belong to separate arms and vice versa. The organization
is changed. By appropriate experimentation, making our subject fixate
the center for a long while, we can reduce the time for either shape
to a minimum value, so that the alternations follow each other very
quickly.[31] If now the figure is turned around a little, so that the
arms have another position, the figure becomes as stable again as it
was at first, and only after some time the alternation will quicken,
as in the first position. In my judgment, this fact may be taken as
evidence for assuming that a local effect of organized processes itself
produces the alternations of organization.
BIBLIOGRAPHY
W. Köhler: _Die physischen Gestalten_, etc. 1920.
W. Köhler: _Komplextheorie und Gestalttheorie_. Psychol. Forsch. 6,
1925.
W. Köhler: in _Psychologies of 1925_ (ed. by C. Murchison).
W. Köhler: _Bemerkungen zur Gestalttheorie_. Psychol. Forsch, 11,
1928.
M. Wertheimer: _Untersuchungen zur Lehre von der Gestalt, II_.
Psychol. Forsch. 4. 1923.
VI
_The Properties of Organized Wholes_
THE dawn of the _gestalt_ problem in modern psychology was not the
idea of dynamical self-distribution as opposed to order enforced by
arrangement; nor did it begin with the discovery that segregated wholes
represent highly important sensory phenomena. The starting point was
the observation that sensory fields are replete with qualities and
properties which one neglects if one takes “sensations” as their sole
content and which, indeed, may have a mysterious aspect when first
viewed in this way. It was von Ehrenfels who, preceded by some casual
observations of Mach, directed the attention of psychologists toward
the fact that a great many, and perhaps the more important, properties
of sensory fields do not fit into the scheme of concepts which is
centered around the idea of “sensation.”
If it is natural for a “sensation” independently to fill its local
place in the field, determined by one local stimulus alone, the
characteristic and surprising feature in those other qualities (which
are usually neglected) is their existence only as properties of
somewhat extended regions. From this fact it seems to follow that
those qualities of von Ehrenfels cannot be determined by single local
stimuli; their existence, then, depends upon several stimuli having a
specific effect in a certain _area as such_.
As an example we may take a glass of water in which soap is dissolved.
The aspect of such a liquid is called “trübe” in German, which in
English means something like “dim” or “turbid.” Now look at it
through a little hole made in a piece of cardboard and you will see
the hole filled with a certain hue of gray (perhaps a little bluish
or reddish), but the quality of “dimness” or “turbidness” will have
disappeared. It is the property of a more extended field and depends
upon more than local uniform stimulation. Exactly the same is true of
the dimness or diffusedness which appears as a quality of things seen
in a dark corner. Again no local impression, “isolated” artificially,
shows any dimness, but the extended area does so in a striking manner.
“Clearness” and “definiteness” as qualities of a field or of parts of
it, have the same supralocal character. I may also mention the property
of a surface which we perceive by touch as “rough” (the German “rauh”).
There is no character like “roughness” in purely local experience
of touch. And further we see that von Ehrenfels’ peculiar qualities
occur in temporal extension as well as in space in that the German
word “rauh” is used for certain acoustical phenomena, as it is for
“rough” surfaces. When listening to rather rapid beats or to the “R”
of human speech we get that peculiar sensory quality, which seems to be
intrinsically similar to the “roughness” of a surface. Again, as this
quality depends upon rapid beats, if stimulation is shortened below
a certain limit, it disappears, whereas other acoustical qualities
will subsist. In passing we may remark that in most cases words like
“homogeneity” and “continuity,” whether applied to optical or to other
experiences, mean properties of extended sensory fields.
From a functional point of view, these observations are less surprising
than was the general opinion at the time of von Ehrenfels’ discovery.
The processes directly underlying our experience of a color will be
a certain chemical reaction, i.e., certain molecules are built up or
destroyed. Now the chemist may analyze such a reaction, but there is
a natural limit to his analysis, because at least one whole specimen
of each atom or molecule taking part in the reaction, and the whole
dynamical event thus comprised, must be included. Beyond that limit
what is called “this specific reaction” will lose its meaning for the
chemist and, therefore, for psychophysical theory which associates one
definite color with one definite kind of reaction. Therefore, even
before parting with the theory of “local sensations” we are obliged to
accept dynamical realities existing only in somewhat extended areas of
space. If that is so in chemistry, the same fact should not frighten
us when we face it in direct experience.
According to the theory of dynamical self-distribution color as a
quality is dependent upon chemical reactions in a brain-field, although
the production of local chemical reactions in a certain distribution
is only one side of the total process. There is much more that is
characteristically spatial and dynamically real in such a coherent
distribution than in the mosaic of local sensation-processes which are
the “dynamical realities” of machine theory.
From the new viewpoint the Ehrenfels-qualities, corresponding to more
extended dynamical realities than color, would be physiologically
produced at the same time as color, since each is a phase of the
same total “process-in-distribution.”[32] In machine theory there
must be “sensations” first of all, if upon this basis something
else shall be founded as a secondary product. It would have been a
superhuman achievement for von Ehrenfels to have gone so far as to
sacrifice at once the machine theory and its sensations. To him his
peculiar qualities remained new bits of experience simply added to the
sensations, and in the school of Graz (v. Meinong, Witasek, Benussi)
there was much discussion about the _fundierte Inhalte_ (qualities
founded upon the basis of sensation) as the product of intellectual
faculties working upon the sensations. So, with regard to the
fundamental concepts applied to the sensory field, not much was changed.
An enormous number of the new qualities discovered for psychological
theory by von Ehrenfels are properties of those segregated wholes
which were discussed in the last chapter. “Simple,” “complicated,”
“regular,” “harmonious,” are words which _may_ have a meaning when
applied to an experience occupying a point in space and time, though
in most cases they refer to products of organization. But when we
call something “symmetrical” this something is certainly a segregated
whole. Similarly, “slender,” “round,” “angular,” “clumsy,” “graceful”
are specific properties of definite wholes. And from these there is
only one step to the more particular “form-qualities” given in the
characteristic aspect of a circle, a triangle, a pear, an oak-tree,
and so forth, all of them existing exclusively in their corresponding
wholes. In German the word “gestalt” may be used as a synonym for
“form,” or perhaps “shape.” So von Ehrenfels, taking the case of
specific shape as the most important and evident among his qualities,
applied the name of “gestaltqualitäten” to all of them. Therefore
it will be clear that not only the different forms or shapes of
objects and figures are included, but also qualities like “regular.”
Furthermore, we have seen that there are temporal _gestaltqualitäten_
as well as spatial ones, since the definition applies to the specific
properties of a melody, to its “major” or “minor” character, for
instance, in the same way it does to the “angularity” of a figure.
Finally, seen movement as a whole may have a _gestaltqualität_ which is
temporal and spatial at the same time. This is the case in the aspect
of a definite form of dancing and in the characteristic movements of
animals, such as “jumping” or “creeping.”
At this point a general remark about terminology may be useful. For von
Ehrenfels the new characteristic properties themselves were objects
of outstanding importance; he was more interested in them than in
those segregated parts of the field which exhibit the best examples
of _gestaltqualitäten_ as their properties. In the German language
however--at least since the time of Goethe, and especially in his own
papers on natural science--the noun “gestalt” has two meanings: besides
the connotation of “shape” or “form” as a _property_ of things, it
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. Following this tradition, in _gestalt theorie_
the word “gestalt” means any segregated whole, and the consideration
of _gestaltqualitäten_ has become a more special side of the
_gestaltproblem_, the prevailing idea being that the same general type
of dynamical process which leads to the formation and segregation of
extended wholes will also explain their specific properties.[33] Here
the main stress is laid upon a characteristic type of process. This,
indeed, is the most general concept of _gestalttheorie_: wherever a
process dynamically distributes and regulates itself, determined by
the actual situation in a whole field, this process is said to follow
principles of _gestalttheorie_. In all cases of this type the process
will have some characteristic which exists in an extended area only,
so that a consideration of local points or local factors as such
will not give us full insight into the nature of the process. From
this viewpoint, even the segregation of circumscribed wholes becomes
one more or less particular, though highly important, case among the
various possibilities which are included in the most general idea of
self-distribution and self-regulation, and in consequence the concept
of _gestalt_ may be applied far beyond the limits of sensory fields.
According to the most general definition of _gestalt_, the processes
of learning, of reproduction, of striving, of emotional attitude, of
thinking, acting, and so forth, may be included as subject matter
of _gestalttheorie_ insofar as they do not consist of independent
elements, but are determined in a situation as a whole. Quite apart
from psychology the same will be true of ontogenetic development,
and other biological events, wherever they show the definite marks of
self-distribution and self-regulation. And this extension of the term
does not mean vagueness, as many seem to believe. If prior to detailed
investigation the general use of concepts belonging to machine and
mosaic theory has prevailed for such a long time and so widely, it
cannot be forbidden to discuss the opposite principle of dynamical
order and regulation, which certainly is less known among psychologists
though equally well established by physical theory. By no means do we
believe, however, that any problem is really solved by the application
of the general principle _as such_. On the contrary, when the principle
seems to apply the concrete task of research is just beginning because
we want to know the manner in which each special kind of process
regulates itself under the conditions of each particular case.[34]
If in the treatment even of sensory fields the real solution remains a
task for the future, at least the first step may be made at once. Here,
as everywhere, it will consist in the recognition of the reality and
the concreteness of the problem. No one fails to see that there is a
problem in visual depth as a property determined by conditions on the
two retinæ, or--more generally--by stimulation on a two-dimensional
surface. To see the real problem in the case of “form,” as a property
of segregated wholes, seems to be much more difficult. The reason
is the same as in the case of segregated wholes themselves (cf. pp.
175-177). When we consider retinal stimulation, for instance, our
thinking operates with ready-made wholes which already have definite
forms as we know them from perception. So we say innocently that “the
form” of our pencil or of a circle is projected upon the retina. These
words contain the experience-error. There is no factor in retinal
stimulation which might pick out of the geometrical distribution of
_all_ local stimuli that circumscribed whole which will appear in the
form of our pencil or of a circle, only after such a segregated whole
has become a functional reality. On the retina we have the indifferent
mosaic of millions of local stimuli, and nothing else. By arbitrary
geometrical thinking, we may select and combine certain retinal spots;
thus we may imaginatively impose all possible forms upon the retina,
including, if we like, those of the pencil and of the circle. We must
not forget, however, that this is mere play when compared with retinal
reality and that the form of these objects is at this moment not more
really there physiologically than that of an angel or of an Arabic
letter.
Some concrete examples will show us what is meant by “concrete, real
form” better than any amount of general discussion. Sometimes we look
upon the map of a country the general form of which we have seen
thousands of times on other maps. But what we really see remains a
foreign and completely unknown figure for a long while, until finally
something happens in our field of vision and the known form suddenly
appears, the unknown figure disappearing completely at the same time.
Excellent cases for this observation are the charts of ship captains,
whereon the sea and its bordering regions are represented as land
and the coast are on common maps. Now, the general contour of the
land is the same on the maritime chart as it is on the usual map;
the geometrical line separating land and water “is projected” upon
the retina as always.[35] Notwithstanding that fact, when looking at
such a map of the Mediterranean I may fail completely to see Italy;
instead I see some “funny” figures (corresponding to the Adriatic and
so forth) which are new to me but nevertheless “have a concrete form”
in my visual field, whereas on the usual map it is the peninsula that
“has a concrete form.” So “to have form” as a concrete property is a
peculiar feature distinguishing certain areas of our actual visual
field from others which have no form. In our example, so long as the
Mediterranean “has form,” the area corresponding to Italy is formless
and vice versa. One is tempted to say that even a form existing on
the retina does not necessarily determine a corresponding form in
vision. But we have to insist upon the fact that _no_ form at all is
given on the retina as a functional reality, neither that of Italy
nor that of the Mediterranean. There is only a certain geometrical
constellation of different local stimuli. Whether the physiological
process starting from the retina under these conditions will lead to a
simple “line” as real form, or to the form of the peninsula with the
contour as its border, or to that of the Mediterranean, the contour
now being its boundary-line, whether it will produce any other or _no_
form whatsoever, all this cannot be deduced from the constellation
of retinal stimuli before we know by visual experience what area is
segregated, and distinguished as having _real_ visual form.
As a second example the following variation of Fig. 8 (or Fig. 8
itself) may be taken: Fig. 9. With exactly the same constellation of
stimuli we may have two different forms, either a cross consisting
of four slender arms or a cross (like that of an order) containing
four large sectors. So long as we _really_ have the first one, i.e.,
as existing in vision, the other will be absorbed in the general
surroundings, which optically have _no real form_ at the time. When the
second form becomes a visual reality, the first disappears.[36] The
observer will remark that the oblique lines are boundary-lines of the
_real form_ in both cases. They belong to the slender cross as _its_
contours, in the first case, and to the large cross, in the second.
[Illustration: Fig. 9]
Simple observations of this kind were first made by Rubin,[37] who
has given us a great many examples. That only definite special parts
of the field have “real form” (as their most characteristic property)
was shown conclusively by the fact that subjects who had seen _one_
definite form at the first presentation of such an “ambiguous” pattern,
did not recognize the pattern, if later on, at the second presentation,
the other form happened to be seen. This second form had not existed
as a visual reality when, previously, the first one was experienced;
so the second one was something completely new when seen in the later
presentation. Again, when one of the two crosses (Fig. 9) is “real,”
one does not see other forms which are quite as real (or _unreal_)
geometrically on the retina. So one does not see the form 9a or 9b or
9c.
[Illustration: Fig. 9a, Fig. 9b, Fig. 9c]
In the case of Fig. 10 two unknown forms will be seen as a group
through which a horizontal line is laid. When I tell the reader that
the number 4 is before him in the field, he will undoubtedly find it;
but if he is not influenced by theoretical prejudices, he will confess
that the form of the 4 did not exist as a visual reality at first and
that, if it began to exist later on, that meant a transformation of
visual reality.
[Illustration: Fig. 10]
In the last example it will be obvious that the reality of a form
depends upon the existence of a definite whole which, when segregated
as such, _has_ that real form, others, which would correspond to other
wholes, being excluded from visual existence. So when we first look
at Fig. 10 we have a definite organization, consisting of two unknown
wholes with the horizontal line running through them. Consequently one
part of the geometrical constellation of the 4 is absorbed by the
whole on the left side, a second fraction by the angular whole to the
right and the rest by the horizontal. And the destruction of the 4 as
a whole also destroys its form. When after instruction, the observer
really gets the 4 as a form in his visual field, a corresponding
whole is produced at the same time. We formulate the general theorem
that real form depends upon the segregation of corresponding wholes.
This statement is confirmed by all observations of puzzle pictures,
of camouflaged objects, and so forth. The existence of visual form
presupposes the existence of wholes which, then, have a real form as
their specific property. The reader will be able to convince himself
that even the reality of partial forms depends upon the existence of
corresponding sub-wholes.
The _objects_ around us are, in most cases, very stable wholes;
therefore, we regularly see them with their definite forms, excepting
in those rare instances in which the wholes and their real forms are
disguised by chance conditions or by intentional camouflage. This is
the reason why we so easily fail to see the problem of form and why we
remain satisfied by “the form given objectively in retinal projection.”
There is, however, no visual form in the world to which our last
considerations do not apply. Wherever we have “real form” in the visual
field, we must assume a specific property of functional process in
that area. But the processes do not have this special property in all
parts of the field. If you walk along a street between high houses on
a bright day, the sky is seen above you surrounded by darker contours
of houses and towers. Do you see that bright area as a form? Generally
not; it has no form, it is indifferent “background,” though on your
retinæ there does exist a homogeneous area surrounded by contours.
These contours remain borders of the houses; the houses have forms,
but the sky has not. If you wish to see a certain area of the sky as a
definite visual form, look at it through a little hole cut into a box
which you then put over your head. If the hole is cut like the letter H
you will see an area of the sky in the corresponding form as a bright
figure on a black ground. And now _this ground_ has no visual form.
Who wishes to judge about the problems of _gestalt_ theory ought to be
well acquainted with these observations and with the consequences to be
deduced therefrom. Exactly as we may either have color in a definite
part of our visual field, or mere brightness of a certain nuance, so
a definite area may have a certain form or it may be formless, i.e.,
merely extended.
Since “real form” presupposes a segregated whole, the existence of
“form” depends upon factors of stimulation similar to those upon which
the segregation and organization of wholes depend. Again, definite
relations in the total constellation of retinal stimuli are found
to be decisive for the existence of real form. But no abstract
consideration of all the logical relations that might possibly obtain
among local stimuli would make us able to forecast where or when a
specific form will be seen. As for the existence of segregated wholes,
i.e., organization, certain _special_ relations again are important and
others indifferent; which are the important ones can only be discovered
by the observation of real forms appearing under a given set of
conditions.