saying that in some parts of the nervous system, between its optical
and acoustical sectors, for instance, paths of nervous activity are
not fixed once for all in the youth of an individual. Either no paths
will be ready for conduction at first, or else from one point of the
tissue several paths will conduct processes equally well in several
directions, so that disorderly diffusion will occur. In the adult,
however, we observe a great many associations, let us say, again,
between the optical and the acoustical centers, and very seldom is
there any confusion in the play of reproductions. The thing we have
before us now is called a book, its parts pages, and so forth. It is a
serious symptom if some one does not call things by their right names.
Normally, the connection between definite visual processes and definite
acoustical and motor processes works astonishingly well. What other
explanation can be offered? Where we had at first no conducting path or
perhaps several indifferent paths, we might assume that with time one
single path has become differentiated and so much more susceptible to
excitation than all the others, that now processes must follow this one
path. This would mean that, disregarding the _genesis_ of this dominant
one-way function, the perfect order of association and reproduction is
again explained altogether in terms of the properties of topographical
conditions. Though these conditions are not supposed to exist in
the same manner in infancy, and though the changes by which they
are produced remain obscure at the present time, when once they are
established the direction and order of processes is as rigidly enforced
by them, and any influence of dynamics will be as utterly excluded, as
is the case supposedly in simple sensory conduction. As the railroad
train remains on its tracks because these determine one way of least
resistance, and as the enormous power of the engine has no influence
upon direction, so in reproduction as well as in sensory processes all
order and direction is a result of prearrangement, independently of the
actual properties and dynamical forces of the process itself.
If for the sake of order dynamics shall not take part in the
distribution of processes, and if distribution shall be merely an
effect of given topographical conditions, important consequences follow.
First of all, excluding only the somewhat obscure _genesis_ of
associations and habit, what happens in the nervous system will
depend either upon inherited machine arrangements or upon secondarily
_acquired_ arrangements. Therefore, where an actual performance is
not an instance of the learning process as such, it must be explained
either by original topographical conditions, or by past learning, i.e.,
by acquired changes in those conditions.[17] Now, this alternative is
nothing but the old dualism of nativistic and empiristic explanation.
No reader of all the famous discussions between nativists and empirists
can have a serious doubt that a nativistic explanation has always meant
the assumption of a given anatomical basis for the actual fact in
question. If such explanation did not seem to be acceptable, then only
one other possibility was left open, that of learning. These authors
never entertain the idea that some specific and orderly function
might occur without being controlled either by special arrangements
preëstablished _ad hoc_ or by arrangements acquired in learning. What
may this third alternative be? Vitalism? We shall see.
In well-established one-way streets what happens at the end of them
will mainly depend upon what has happened at the entrance. Sensory
experience will therefore consist of purely local elements of
experience, the genuine properties of which must depend upon local
stimuli exclusively. If, for the maintenance of order, the isolation
of processes in each pathway and in each final cell (of the brain)
must be absolute, then no influence of processes in other parts of the
nervous system will be able to alter sensory experience, and so it
must remain the same whatever the changes of attitude. By enumerating
the actual properties of all elements at a given time we give an
exhaustive account of the presented field. This is what has been
called the summative, or mosaic, character of sensory experience, as
it is understood from this viewpoint. Evidently, in terms of it, the
sensory field becomes “inflexible,” exactly as its physiological basis
is mainly determined by rigid topography. It also becomes “poor,”
because the variety of experiences is restricted to those indifferent
patterns of elements which we may find by varying independently the
properties of the local elements. Any organization of processes in
the field is excluded because order must be explained by functional
separation. Specific function, dynamically extended over an area of the
field, is excluded for the same reason; we have to deal with a purely
_geometrical_ pattern of _local_ processes.
How “empty” and “dead” does the organism appear in this theory!
Dynamically, it has nothing to contribute to the monotonous elementary
currents conducted compulsorily from a point of stimulation to a point
of reaction. So it becomes an indifferent stage for actors indifferent
to the stage as well as to each other. As an object of research in
dynamics it is less interesting than a molecule or a soap-bubble which
are both of them functional wholes.
If between the field of sensory processes and the effector organs
conduction is determined in the same manner as, in this theory, it is
between local stimuli and the elements of that field, a thoroughly
adequate formula for research in psychology will be: to find out what
stimuli produce what reactions in the effector organs. The statement
that stimuli, on the one hand, and reactions, on the other, are the
only points of interest in psychology, corresponds absolutely to the
picture of the organism, and especially the nervous system, as lacking
any characteristic process of its own as a whole. Since other schools
do not pretend to have new positive ideas about the functioning of
the nervous system, that seductive formula of behaviorism has found a
rather general assent as expressing the viewpoint of natural science
in psychology. Unfortunately, in its present usage, it is not only
seductive, but also ambiguous and superficial.
From the fact that in this theory dynamics is excluded from the
determination of order and distribution, one more consequence follows
immediately. Everywhere in nature dynamical events depend upon the
properties of those processes and materials which exert influences
upon one another. In a solution, containing Na_{2}SO_{4} and BaCl_{2},
BaSO_{4} will be precipitated because of certain properties of Ba,
SO_{4} and H_{2}O which, in their mutual relations, determine the
dynamics of the process. Two electric currents will produce mutual
attraction of their conductors if both have the same direction;
repulsion, on the other hand, if the direction of one current is
opposite to that of the other. The rule is general, that “relative
properties” as exemplified in these cases are decisive for dynamical
interaction. A theory, therefore, which excludes dynamics from the
determination of distribution, allowing it the production only of
elementary nervous current, will have to draw the inference that the
properties of local processes have no influence whatever upon the total
distribution in the field. A given local process will be altogether
indifferent to its neighbors, running its course uninfluenced by their
existence. All possible patterns may be produced by appropriate sets of
peripheral stimuli; no mutual forces are admitted which, in the whole
field, would produce certain definite distributions rather than others.
A similar consideration in the case of association and reproduction
will occupy us later on.
When confronted with such an unretouched picture of current assumptions
about physiological functions, most psychologists will protest. They
will declare that one should not take too literally what has been
used rather as an analogy in a preliminary endeavor to develop ideas
about the processes of the nervous system. Every one admits, moreover,
that there are cases of “irradiation” in some parts of the tissue! To
this criticism I must answer that if _all_ analogies chosen, in the
first tentative picturing of orderly nervous function, are of the same
type, using topographical arrangements as the basis of order, this
may be taken as evidence that other analogies do not occur to these
authors. Preliminary though it may be, it remains a machine-picture,
and no other has been developed, which is different in principle. As to
irradiation of current, this concept as such does not mean more than a
lack of definiteness and accuracy in the machine; it presupposes order
enforced by strict isolation as the normal case, a slight deviation
from which is the whole content of the idea. I admit that it makes
our ideas about nervous functioning a little more nebulous than they
should be according to the extreme viewpoint of machine theory; but I
deny that in this manner _order_ of distribution can be or has been
explained by any one. Granted that all conductors may “leak” a little
at certain points and that, therefore, local processes may become
interfused mutually to a certain degree, I am unable to deduce orderly
distribution or organization from such an indifferent “spreading” or
interpenetration of essentially isolated units.
Let us compare the consequences of the theory with observation.
Some relevant data have already been mentioned. Other and much more
important data will occupy us in the next chapters.
That constancy of brightness and of size cannot be explained by the
assumption of one-way conduction determining local sensory experience
in terms of local stimulation does not need further discussion, since,
because of their incompatibility with that assumption, those facts
are commonly believed to exemplify the influence of meaning. But now
that experimentation has shown that the meaning theory does not seem
to explain them either, neither the empiristic nor the nativistic
assumptions help us in these cases. So we must try to conceive a third
kind of nervous function other than the types in which processes are
strictly directed either by inherited or by acquired arrangements.
If there is a third reasonable assumption, it will be necessary to
apply it also in those other cases, such as constancy of form, speed,
localization, and so forth, which are so similar to constancy of
brightness and size, that the same explanation should obviously be
given in all these instances. Quite generally then the alternative
between empiristic and nativistic hypotheses must be misleading.
The introspective theorem that changes of attitude cannot influence
“true” sensory experience does not agree better with the facts.
It seems rather to be an arbitrary definition of “true” sensory
experience. In observation, at least, if I transform the white in
the shadow and the black in full light into two similar grays, _by
“introspection_,” there can be no more radical influence of attitude
upon sensory experience than this transformation which occurs under
constant conditions of stimulation. And the same holds wherever
introspection, destroying natural experience by an artificial
technique, finds its “true” sensations instead. Fortunately in one
part of our science at least, this seems to be the prevailing opinion.
When we analyze a clang we hear several notes appear successively
in the mass which previously we heard as a _unity_. If in this case
we agree that under constant conditions of stimulation our attitude
transforms one sensory reality into others, and that the clang heard as
one is not less real than the partial tones apparent during analysis,
we have no right to contradict similar observations in other cases.[18]
As to the statement that sensory experience is a purely local affair,
each point of a sensory field depending upon “its” local stimulus
exclusively, we must reiterate that no grounds have ever been given
for such a radical assumption. Rather it seems to be an _a priori_
belief about what _ought_ to be the nature of things, experience to the
contrary notwithstanding. As far as observation goes the properties
of local retinal stimulation do not simply determine the size, the
form, the localization and the brightness of local experience; neither
does retinal speed determine seen speed, as it probably ought to
according to the thesis that the geometry of retinal facts determines
spatial experiences. As a matter of observation, all the well-known
“illusions” may be cited as evidence of the fact that local processes
depend upon _sets_ of stimuli. To some degree this controversy will
be settled by pragmatic principles: the decision will go to that side
whose principles prove most fruitful in the further development of
psychology. In the meantime, however, we may carry on our analysis of
the polemical situation at present.
Almost all psychologists agree, in the case of one experience, upon the
fact that local sensory experience is determined by more than merely
local stimulation. This case is that of color-contrast, which most
psychologists suppose to be an effect of interaction in the nervous
system. If here the point-to-point correlation between retinal stimuli
and sensory experience is surrendered, because the determination of
local experience by conditions in a larger area is too evident, how
can we proceed hereafter as if this discovery and concession had not
been made? It took science some time before it would accept simple
observation even in this case. Helmholtz refused to do so and, of
course, he applied the meaning theory in order to save its fundamental
belief, i.e., the point-to-point determination of local sensory fact by
local stimulation. But after the first step has been made, we should
realize not only that one theory of contrast has taken the place of
another, but also that instead of applying persistently one general
principle to all sensory experience, we have begun to accept a new
one. In the future, wherever experience does not correspond to local
stimulation, we should, at least, consider the possibility that such
an experience may depend upon the total set of conditions in a larger
field, exactly as contrast does. _If_ it does, this may help us to
understand why changes of attitude affect sensory experience in some
cases. Sensory experience, which depends upon the constellation of
stimuli in a larger area, and therefore does not correspond to merely
local, inflexible units of process, may be influenced by “processes
of attitude” as well. The most important point, however, seems to be
that if we find a type of process depending upon _sets_ of stimuli
instead of single ones, this process may be that “third possibility”
we are seeking. Certainly, it would neither be like the elementary
processes of nativistic machine theory, i.e., absolutely determined by
_inherited_ arrangements, nor like the processes of empiristic theory,
upon which a similar compulsion is exerted by _acquired_ arrangements.
In the next chapters I shall try to show that other facts, much more
important than those we have considered hitherto, point exactly in the
same direction. For the sake of order in nervous function the machine
theory excludes _organization of process_ in the field. But we shall
see that organization may be regarded as a typical fact in sensory
experience. Again the theory excludes the assumption of any specific
process dynamically extended over an area of the field. But it will
not be difficult to show that there are experiences in great number
the specific properties of which belong to _extended wholes_ and do
not exist in local isolation. This means also that the real variety
of processes occurring in the sensory field is simply enormous when
compared with those essentially indifferent patterns of local elements
which may occur according to the machine theory.
If dynamics are excluded from the determination of distribution, local
process will correspond everywhere with its stimulus; the actual
properties of stimulation in their mutual relationships will play
no rôle in the whole affair, as they would undoubtedly in the case
of dynamical interaction. Reviewing our observations, however, we
find that everywhere the aspect of sensory experience depends upon
the properties of stimuli in their mutual interrelationship. This is
well known in the case of contrast and tonal fusion, but we may cite
as well all those observations which were discussed at length in the
last chapter. Constancy of brightness, for instance, depends upon the
relation of the illumination and brightness of the neighborhood to the
brightness of the field in question. That _organization_ in the field
depends upon the definite “relative properties” of local conditions,
will be shown in the next chapter.
All these facts make it probable that something is wrong with the
machine theory; they look as though they might be better understood by
dynamical theory. Sometimes the observer finds dynamical events quite
obviously occurring directly within the field. This is particularly
the case, when sudden stimulation or change of stimulation is followed
by a _development_ of process. If we cause a bright figure suddenly
to appear in the dark, this figure will have at once neither its full
size nor its “right” place. Instead it will appear with an energetic
movement of extension as well as of approach. At the moment of abrupt
disappearance, it will appear to have a movement of contraction and
recession. Obviously, such observations would be exceedingly strange if
considered in terms of machine theory. Or again, take the surprising
fact that in touch, as well as in vision and hearing, the distance
of objects and events may vary considerably if the stimuli are given
in a certain manner. In some beautiful experiments of von Frey two
distant points of the arm touched at the same time will appear to
be one-half the distance apart which they appear to be when touched
in slow succession. Scholz and Kester have both measured the mutual
“attraction” which lights as well as sounds show when presented under
adequate conditions. If the machine theory did not have its inordinate
historical prestige, no one would hesitate to take these observations
as evidence of dynamical interaction within the field. The stroboscopic
movement, which belongs to the same class of observations, has acquired
a unique importance by the fact that experimentation in this field led
Wertheimer explicitly to discard the purely summative theory of sensory
experience, not only for the problem in question but generally.[19] If
at some distance from each other two stimuli are successively projected
upon the retina of a subject, a movement will be seen, starting from
the locus of the first and ending in the region of the second. Under
favorable conditions there will not even be two “impressions.” One
“thing” will move from one place to the other--certainly a case of
dynamics which it would be difficult to understand by the theory of
isolated local processes. After the more important properties of
stroboscopic movement were investigated by Wertheimer, Koffka and
others, there was much discussion, mostly about minor points, which
almost tended to obscure the main facts. As was to be expected also,
meaning was rather liberally offered as an explanation. Nevertheless,
if conditions and the attitude of the observer are not too inadequate,
we undoubtedly _have_ movement in the visual field. Those who do not
believe direct experience, where it contradicts the supposed properties
of “true” sensation, may deduce the “reality” of that movement from
the fact that, when repeated, it produces an after-image of movement
in the opposite direction, exactly as “real” movement does. Though,
historically, Wertheimer’s investigation was the beginning of
_gestalt_ theory, in the proper meaning of that phrase, the following
considerations of sensory process will take another slant, because to
me another way seems more advisable as an introduction.[20]
There are several well-known arguments which favor the machine
theory. I have sometimes heard that it gives us a picture of
nervous function which, clear and simple in itself, is the more
easy to understand since in practical life we enforce order
everywhere in the same manner, i.e., by arrangements _ad hoc_. I
must confess that such a policy of the least scientific effort
seems to me unacceptable. The comfort and the habits of the
scientist do not count where he must deal with the properties of
his subject-matter. Furthermore, only the psychologist, neurologist
and physiologist will save time and effort by an assumption
which explains order by arrangement. They simply hand their
problem down to somebody else; wherever, in theory, a problem of
function is reduced to one of special arrangements, the science
of “morphogenesis” in its ontogenetic and phylogenetic branches
is amicably asked to solve it, i.e., to explain the origin of the
arrangements. So the avoidance of difficulties in our science
means that others have proportionally more difficulties. And,
by the way, at some point functional problems must be treated
as truly functional. If it is barely possible to understand the
ontogenesis of anatomical structure by the working of “special
arrangements” in the egg and germ, it would be ridiculous to
explain _phylogenesis_ by arrangements which bring it about.
Yet it will be said that we do have special arrangements
guaranteeing definite function in a great many organs of the body.
Nobody can deny it, and I shall admit at once that the existence
of the optic nerves as a conducting system between the eyes and
the brain may be regarded as an example. Still, in our body, there
is another conducting system in which a great many substances are
transported along with the _blood_. And here, though the conduits
are a general arrangement “for transportation,” we do _not_ have
special arrangements for carrying each part of the fluid to its
right place. What a definite part of the tissue pours into the
blood or takes away from it at a given time is not determined
by isolating conductors; everywhere it depends upon the actual
relationship between the state of the tissue in question and the
chemical properties of the blood, and yet normally we have order in
the whole affair. The example shows that the existence of “organs”
does not allow us to draw any inference about elementary processes
being kept in order entirely by machine-arrangements.
However, I may be forgetting that in physiology and pathology we
have much evidence for assuming a “projection” of retinal points
upon definite points of the _area striata_ in the brain! I am
not certain whether the facts in question prove that between
the retina and the _area striata_ conduction is completely a
point-to-point affair. But that does not matter, since at the
present time neurologists no longer believe that the _area striata_
is the terminus of optic processes and that visual experience is
concomitant with processes in _this_ area. If up to this region
conduction should be a matter of strictly isolated paths, the
functional problem we are occupied with would have to be solved in
parts beyond the _area striata_.
I shall be told that the nervous system consists of cells the
fibers of which _are_ isolated from each other. I answer that in
its gray ganglionic and nuclear fields conditions are different.
Here mutual influence is not only possible; it is necessary.