The Function of Incipient Motor Process
Margaret Floy Washburn
The construction of an hypothesis can never give the same repose and satisfaction to the scientific conscience as does the discovery of even a moderately significant fact. And hypothesis-building brings with it the less sense of solid accomplishment, the more its results are removed from the actual test of fact. Hence the person who tries to erect a theory as to what occurs in the nervous system, in connection with any processes less amenable to experimental conditions than scratch reflexes, is not likely to feel a gratification proportionate to his pains. And yet a consistent theory of the physiological processes underlying the higher mental processes would be of some practical value, even though it could not be tested by physiological experiment. For a coherent presentation of the facts demands some principle of explanation, and the laws of learning, which lie at the bottom of the complexer processes of mind, suggest certain physiological assumptions which logically demand to be followed out and elaborated into a complete physiological theory.
The present paper aims to suggest an hypothesis regarding the nature of some essential features in the nervous process underlying the production of a mental image, a revived or centrally excited mental process. Since the leading part in such processes, according to this theory, is played by the initiation of movements that are not fully carried out, it may be termed the theory of incipient motor processes. As a preliminary to consideration of the image, we may discuss the physiological nature of the associative processes in general.
There are grounds for thinking that to that form of association which has traditionally been called the association of ideas, the association of movements is preliminary. In the so-called association of ideas we have called into consciousness
( 377) the image of a past experience. If stimulus A and stimulus B have at some former time been experienced together, the occurrence of A `makes us think of B; that is, calls up a mental image or centrally excited sensation of B. There are two reasons, at least, for believing that a more primitive process than this is to be found in the type of association whereby, stimulus A comes to produce the movement which formerly resulted from stimulus B, rather than an image of stimulus B. These two reasons are as follows. First, among the lower animals we have constant proof that one stimulus may through being repeatedly experienced in connection with another, come to assume the motor tendencies of the other; but we have very limited evidence of the occurrence in the animal mind of the associative production of images. Secondly, in the human mind, the association of images rests absolutely upon attention. Not the fact that the stimuli A and B occur together gives to A the power of calling up an image of B, but the fact that the two are attended to together. And whatever else attention may mean, the fact is reasonably certain that simultaneous attention to two things means a simultaneous motor response to them. The dependence of association upon attention, an essentially motor phenomenon, becomes comprehensible if we think of association as being itself primitively an association of movements.
Let us then first consider the processes by which stimuli come to be associated with new movements. We shall use in this consideration four fundamental physiological assumptions, all of which have some warrant from the facts discovered in experiments on simple reflexes.
1. When a motor response is initiated, all the sensory centers  that are receiving stimulation at the same time contribute to that response a part of their energy. This is the familiar fact of Bahnung, as illustrated by Exner's case of a sound stimulus reinforcing a touch stimulus in producing movement of a rabbit's foot, or by Yerkes's demonstration of the effect of sound stimuli on the reaction of the frog to
( 378) electrical stimuli. While the principle of Bahnung is usually stated to hold true of certain allied systems of reflexes, we have stated it as a perfectly general principle, with the understanding of course that it may be crossed by other principles.
2. Whenever a sensory center has any part of its energy drained into a motor outlet, the resistances along the path leading to that outlet are decreased.. It is unnecessary to dwell upon this assumption. In one form or another, it is indispensable to any theory of the more complex workings of the nervous system.
3. One motor response may be prepotent over others tending to occur at the same time. This prepotency consists in the fact that certain motor responses will be called forth by a much slighter intensity of stimulus than others. For this statement we have the authority of the experimental physiologists, and Sherrington has emphasized the fact that the prepotent responses are those most concerned with welfare.
4. There exist antagonistic motor responses, so connected that the making of one inhibits the making of the other. This statement, also, needs no defence: it is a well-established fact.
To proceed now with our investigation of the process whereby the association of motor processes is brought about, let us suppose two stimuli, A and B, acting simultaneously upon the organism, and let us suppose further that A naturally gives rise to the response AR and B to the response BR. We may further assume that these two responses, AR and BR, are not antagonistic to each other; that is, that the organism can carry out both movements at once. Now if we have recourse to assumption 1, that when any motor response is made, a portion of the energy of every stimulus acting on the organism at the time is contributed to the making of it, we may assert that a part of the energy of A goes to the production of BR and a part of the energy of B goes to the making of AR. If we call assumption 2 also to our aid, whereby the oftener a nervous process travels a certain
( 379) pathway the less the resistance to its passage, we see how the frequent occurrence of A and B together can give to either A or B the power of initiating the combined reaction AR-BR, without the actual occurrence of the other stimulus. This we may call Type I of motor associations.
In other cases it happens that AR and BR are not compatible but antagonistic reactions. Now a billiard ball (if I may be pardoned for using that well-worn object as an illustration), when acted upon by two forces in opposite directions is influenced fully by both of them. The weaker loses none of its effect because a stronger one is on the field: it is able to diminish the action of the stronger by the full amount of its own strength. This would be an exceedingly inconvenient principle to govern the actions of a living organism: hence the nervous system works in such a way that the stronger stimulus can wholly suppress and inhibit the movement that would result from the weaker if it acted alone (Assumption 4). This means that if a weak stimulus A and a stronger stimulus B act together upon the organism, and would if they acted alone demand of it incompatible responses, A finds itself quite cut off from its own motor outlet when the motor paths of the response BR are open. Hence apparently the whole energy of A will be available to find its way into the channels of the response BR, according to assumption 1. We might expect that a pathway from A to BR would be formed with especial rapidity when BR and A's original response AR are antagonistic. But whatever advantage is derived from the fact that in such a case all A's energies are free to take part in the new reaction is probably compensated for by the fact that the resistances which the energy of A has to overcome in finding its way to the antagonistic outlet AR are higher than those encountered on the way to a compatible motor response. Indeed it may well be that a part of the very process by which the reaction AR inhibits BR involves a heightening of the resistances along the paths that would connect AR with a stimulus belonging to an antagonistic response. But of course it constantly happens that a stimulus does come to be connected with a reaction the very opposite
( 380) of that which originally pertained to it. A tamed animal learns to run to the human being from whom it at first fled in terror. In order that the tendency thus acquired for A to produce the response BR may be permanent, and may displace its original tendency to produce AR, either B must occur more frequently with A than A occurs alone, so that the pathway from A to BR becomes more permeable with repeated traversing (Assumption 2), or else the response BR must be a prepotent response, one that is specially ready to occur provided that even a weak current of excitation reach it (Assumption 4). Such a prepotent response is the negative reaction following harm to the organism, and hence we find that this response readily becomes associated with stimuli that are not in themselves harmful but have been accompanied by injurious stimuli. The type of learning where a stimulus A comes to produce a motor response BR, opposite in character to its original response AR, because of the experiencing of A and B together, we may call Type II.
Now thus far we have been considering the cases where the two stimuli A and B act simultaneously upon the organism. But a stimulus A may come to produce a response BR which originally pertained to another stimulus B, when B has occurred not together with A but immediately after it. The simplest case of this sort will happen when B intervenes after the reaction AR has been started, but before it has been completed. If A is to become associated with the new response BR, the latter must draw off into its own channels some of the energy of A, and prevent its all being discharged into AR. The more natural effect in this case would seem to be for the later stimulus B to have all its energy diverted into the already active channels of AR, according to Assumption 1. To explain how the later response can ever drive the earlier one from the field, we must suppose the later one to be a prepotent response, so ready to occur that even slight excitation will produce it. When AR and BR are not incompatible reactions, what happens may be simply that AR is initiated by the occurrence of A; but B immediately following and opening the prepotent path BR, a part of A's
( 381) energy is diverted into BR. The oftener this happens, supposing BR to be a prepotent response, the slighter is the tendency to produce AR at all; thus we have the gradual dropping off of movements which while they are not really antagonistic to the prepotent response, are simply unnecessary, such as the random movements that accompany early attempts at writing or skating, or the aimless struggles of an animal learning to get out of a puzzle box. When, on the other hand, AR and BR are really antagonistic, we have to suppose that the reaction BR, breaking in on the already initiated AR, inhibits it not gradually, by depriving it of a part of the energy of its stimulus, but automatically and at once. Thus all the remaining energy of A has to find its way into the channels of BR. The only difference, it would appear, between these cases, which we may call Types III and IV, where the stimuli are successive and those (I and II), previously considered, where they are simultaneous, is that greater prepotcncy of one response over the other is demanded for Types III and IV; of the later response over the earlier one, to counteract the advantage of the earlier one in having already got itself started.
Evidently the highest degree of prepotency is required of a reaction that can supplant another which has not only been initiated first, but has actually been carried out. Yet this supplanting does occur in many cases of learning. A stimulus A produces its full response AR: there follows upon this response a stimulus B with its appropriate reaction BR. How can A, whose energy was wholly used in producing the reaction AR, contribute anything to the production of BR? The only answer to this question, so far as I can see, lies in the formulation of a further physiological assumption, namely: (5) A certain portion of the energy of a stimulus is left behind after its motor response has occurred, and may be drained into the channels of the next-following motor response. Now for the connection between A and BR, thus formed through the slight residue of A left after the response AR has occurred, to become victorious in subsequent experiences over the original response AR, and bring it about
( 382) that A produces BR instead of AR, it is necessary that BR shall have a high degree of prepotency over AR. The reaction BR in such cases is probably always of the type called by Sherrington 'consummatory reactions,' whose performance is intimately connected with the immediate welfare of the organism. We may call the types of learning which involve the supplanting by another movement of a movement that has actually been carried out, including the two cases, where the movements are antagonistic and where they are not, Types V and VI.
None of these six types of association needs to involve the production of an image or centrally excited mental process, in which the sensory effects of past stimulation are revived. To account for the origin of the image, I should like to propose another physiological assumption, the sixth. It is this: (6) When the cortex has reached a certain degree of development, if a motor response is initiated, all the sensory centers that have recently or frequently discharged into the motor center concerned, that is, all the sensory centers with low resistances at the synapses along the pathways leading from them to the motor center, are set into excitation. This excitation is more marked, and more wide-spread, the greater the delay between the initiation and the execution of the movement. That is, while with a short delay activity will be induced in those sensory centers most closely connected with the motor center concerned, a longer delay will cause the spreading of the induced activity into more remotely connected sensory centers. Upon the activity thus induced in sensory centers, whether near or remote, the image or centrally excited conscious process is based.
The common conception of the process whereby sensory centers are centrally excited, giving rise to images, is that nervous activity passes from one sensory center to another, along a nervous pathway whose resistances have been lowered by the former simultaneous activity of the two centers. Thus there is supposed to be an actual transfer of energy along an associative pathway, just as in the peripheral excitation of a center a transfer of the stimulus energy occurs
( 383) along a sensory pathway. The view here proposed maintains, on the other hand, that the discharge of a ‘centrally excited' center does not occur by the influx of any energy into it from without. Our theory suggests rather that a discharge of the stored-up energies of the sensory center into a motor center is brought about when the motor center in question is partially excited from another source and the resistances are already low between it and the sensory center whose discharge is thus induced. The process of central excitation would be thus induced and not transferred: the disturbance of equilibrium in the motor centers draws the stored-up energies from the sensory centers associated with them. It is as if a hitch in the functioning of a motor center enabled it to call to its aid contributions from all the sensory centers connected with it by paths of low resistance. Thus, for example, if in Type I of learning, where the stimulus A has become able to produce both its own original response AR and the response BR which originally pertained to stimulus B, the response BR is only partly initiated, there will be, supposing the cortex of the animal to have reached the proper degree of development, an induced activity in the sensory center formerly affected by stimulus B, with the result in consciousness of an image of B.
This theory demands, not that the partial excitation of the motor center M shall send a nervous process to the sensory centers associated with M, which would involve either violating the principle of the irreversibility of synapses or the assumption of a double conduction path; but that the partial excitation of M in some way draws a current of excitation down from the associated sensory centers into M. While no clear physical or electrical analogy for a process of this sort can be found, we at least contradict no known fact of nervous action by supposing that the excitation of M lowers resistances at all the synapses of M. This sudden lowering of resistances may be regarded as giving an opportunity for the associated sensory neurones to discharge into M. Such an action is assumed by MacDougall to occur in the case of the reciprocal innervation of antagonistic muscles. But of
( 384) course to explain the discharge of sensory neurones without external stimulus, we must assume in them a tendency to discharge as soon as there is a sudden reduction in the synapses connecting them with the motor area; a kind of unstable equilibrium. Further, we need to understand wherein that degree of cortical evolution consists which makes the mental image possible. For certainly it seems that frequency and free functioning of images is peculiar to minds of the highest order of development, and is much less marked in those of even the highest vertebrates below man. May we not then suppose that the distinguishing characteristic of a cortex sufficiently highly evolved to undergo processes that are accompanied by centrally excited mental processes or images, is a high degree of instability, tension, or potential energy in its sensory centers, such that their discharge will occur `spontaneously' whenever there is a sudden lowering of resistances at their synapses? Our seventh physiological assumption would then be: (7) In a highly evolved cortex, sensory neurones are in a state of unstable equilibrium and readiness to discharge, such that a suddenly lowered resistance at any of their synapses may induce their discharge into a motor pathway.
It is now high time to consider the arguments which favor making this formidable array of assumptions. The order in which we discuss them is not especially important; we may begin with one that is not particularly weighty, namely, that introspection seems to reveal to us a practical function for mental images, such as that we have described. For instance, take the case of a man shut up in a room from which he has previously released himself by working a combination lock. His glance falls upon the lock: the external stimulus sets up a tendency to move, which however needs the help of memory images before it can be fully executed. "There's the lock; now what were the turns I had to make?" he asks himself: the partial initiation of the response calls to its aid centrally excited processes, and the movements are successfully performed under the joint incitement of peripherally and centrally excited currents.
This, however, is an illustration rather than a real argument. The consideration which first suggested to the writer the necessity of some such physiological theory of the image as that here described was, as has been previously intimated, the fact that the association of A and B which enables A to call up an image of B does not rest merely on the simultaneous occurrence of the stimuli A and B on some previous occasion. We associate A and B only if besides being experienced together they have been attended to together. Now there are only two ways in which the necessity of simultaneous attention to A and B can be interpreted. First, the physiological changes underlying attention to A or B may be supposed to affect the sensory processes resulting from the action of the stimulus A or B. Or secondly, they may be thought of as involving characteristic motor reactions to the stimulus A or B. In a word, attention must be influential upon association either through a sensory or a motor effect. One can hardly conceive any influence of attention upon sensory processes alone which does not reduce itself to an increase of the intensity of such sensory processes. It may be said, for example, that attention, bringing about a better reception of the peripheral stimulus and a reinforcement of it by centrally excited processes, makes the sensory processes resulting from the stimuli A and B more intense, and that such intensity is necessary to bring about their association. But it would be hard to show, on such a hypothesis, why an increase of intensity that did not result from attention, but from increase in the physical force of the stimulus, should not be equally effective for the formation of associations. We know, however, that mere intensity of stimulation, apart from attention, has no significance for association. It would seem, then, as though the essential dependence of association on attention must rest on the essential motor character of association.
A second argument in favor of this theory is that it offers a convenient and clear way of conceiving the relation between imagery and learning, and of the anschaulich to the unanschaulich accompaniments of learning. It is evident that as motor processes become more completely organized, and
( 386) learning is complete, there occur pari passu an increase of the speed with which the movements are performed and a decrease in the amount of imagery present. It thus seems natural to connect the presence of imagery with delay and hesitation in motor response. In order, further, to understand the stages which introspection reveals in this process of the gradual disappearance of imagery, we need to note a further peculiarity which characterizes the complexer cases of learning.
This peculiarity consists in the fact that in higher learning processes we have the formation of movement systems. Now movement systems are characterized by the fact that when motor responses are associated, one response does not supplant another, but the performance of one response is rather an essential condition for the performance of another. The practical importance of the motor responses depends upon their all being actually made: one movement in the system is of no use without the rest. The conditions are such that there cannot be established a short-cut through the elimination of certain movements altogether: the movements derive their value each from the actual performance of the other. It is unnecessary to point out how frequently learning has to take this form. Most movements, in fact, are not simple but complex, and consist of the performance of a number of motor responses each of which would be useless without the others.
Of course such connections between motor innervations are in many cases innate. But there are others which are acquired during the lifetime of the individual. Now the most natural way in which we may suppose such movement systems to be learned or acquired is by some arrangement through which the performance of one movement may itself furnish the stimulus, or a part of the stimulus, for another movement. And the most obvious method by which the performance of one movement may regularly provide the stimulus for another is by the processes which are set up in sensory pathways by the action of the muscles themselves. Acquired systematic connections between movements are
( 387) most naturally explained by supposing the dependence of one movement in a system on the kinaesthetic or proprioceptive excitations resulting from the performance of another movement as a part of its stimulus. If reaction BR, to which B is the appropriate stimulus, is also dependent on the occurrence of the kinaesthetic excitation KAR, resulting from the performance of reaction AR to stimulus A, we shall have the connection of the two movements AR and BR into a system. If the connection is simply that KAR must combine with B to produce BR, then the system will be one of successive movements: movement BR must be preceded by movement AR. But if the connection is mutual, so that the full stimulus to movement BR is KAR plus B, and the full stimulus to movement AR is KBR plus A, then each movement demands the performance of the other, and we have a simultaneous system of movements: the connection is made in all directions. Practically all systems of movements, whether they are successive or not, involve simultaneous systems: that is, even a succession of movements is usually a succession of complex movements, or simultaneous systems of movements.
Now while as learning progresses, the tendency is for imagery to disappear and for the movements to be carried out automatically, introspection shows that at a certain stage of this process, while there is no clear visual, auditory, or verbal imagery accompanying the performance of a system of movements, there are present in consciousness certain ‘imageless' or unanschaulich conscious processes, certain awarenesses or conscious attitudes or thoughts. These we can explain, on our theory of the basis of the image, as kinaesthetic or proprioceptive in their origin, and we can see why they should disappear only at a later stage of learning, or the formation of movement systems, than that at which visual and other anschaulich processes vanish. In a complex system of movements, the sensory centers connected with a given motor center are of two orders: first, those corresponding to the original stimulus to the movement, visual, auditory, or whatever it may have been; and second, the various kinaesthetic centers whose excitation results from the performance
( 388) of the other movements in the system and forms a part of the proper stimulus for the motor center we are considering. If, then, this motor center is partially excited, and there is a delay in the execution of the motor response, there are two kinds of imagery that may be aroused: the one may be visual, auditory, or in short may belong to any modality; the other must be kinaesthetic and must relate to the system of movements itself. The latter, we may suppose, constitutes the unanschaulich conscious accompaniments of the movement system. And since in the formation of a movement system while the actual occurrence of the original external stimuli that belonged to the various movements comes to be eliminated, the actual performance of the movements of the system never comes to be eliminated, because by definition the system cannot afford to drop out any movement, we can see that the kinaesthetic centers would be much more intimately connected with the motor center than would the sensory centers concerned with the stimuli of other modalities which originally appertained to the movements of the system. Since a slight delay in the performance of a movement calls into activity the sensory centers most immediately connected with the motor center concerned, and a longer delay induces activity in more remotely connected sensory centers, we can explain why at a later stage of learning the imagery involved should all have a kinacsthetic basis, while at an earlier stage, involving longer delays, imagery of other sorts should be called up.
It is evident, finally, that this theory with regard to the physiology of the mental image involves some addition to the current theory of attention. We have based our hypothesis about the image mainly on the undoubted fact of the dependence of recall on attention, and on the supposition of the essentially motor nature of attention. Now the ordinary account of the motor aspect of attention describes it as involving two kinds of motor processes: first, those required to hold the body quiet, so that the stimulus shall be received without distraction, and secondly, those which produce adaptation of the sense-organ for the most favorable reception of the stimulus. The motor processes of the first class have
( 389) clearly nothing in them that is specific or differentiated according to the individual character of the stimulus. The same quiet position of the body suits attention to any kind of stimulus; is adapted to listening, looking, or thinking. The adjustment of the sense organ has more relation to the peculiar nature of the particular stimulus concerned: it is of course different when the stimulus is visual and when it is auditory; for a visual stimulus it varies with the distance from which the light rays come and the point on the retina which they strike. There is even a difference in the accommodation process according to the wave-length of the light rays, since the focal distance of the lens varies with the color of the light. But for many stimuli the motor processes which relate to the adaptation of the sense organ would not be differentiated: the same adjustment process would suffice for a whole group whose associative connections would yet be very unlike. It seems to the writer of this paper that in addition to the two classes of motor effects of attention mentioned above, a third may well be added, and the statement ventured that attention to a given stimulus involves the initiation, at least, of a motor response that is peculiar to that stimulus alone. This would mean that every sensation that can be discriminated in a fusion, and every group of sensations that can be attended to as a single whole, has connected with it one or more movements which are peculiar to it alone. What, indeed, does discrimination mean if not the performance of specific motor reactions? Where would be the use of consciously distinguishing between two sensations if the two were not to lead to different movements? We need not expect always to find by introspection traces of these specific motor processes involved in attention; yet as illustrations familiar to introspection we may take the slight tendencies to articulate or to vary the tension of the vocal cords which accompany attention to sounds, or the tendencies to eye movement that accompany the visual perception of lines and forms.
This view of attention needs fuller elucidation and defense than can be given it here. Our concern at present is only to
( 390) point out the relation of the mental image to attention, on the theory of incipient motor processes. The initiation of a specific motor response, with attention to a given stimulus, induces activity in whatever sensory centers are most directly connected with the response in question, through the previous occurrence of their own response together with it; and the activity of these sensory centers is accompanied in consciousness by images or centrally excited processes. If the question be raised as to why the motor responses whose association gives rise to images must be those motor responses concerned in attention, and not any motor responses whatever, it may be suggested in reply that motor responses which are in connection with cortical sensory centers (and no lower sensory centers need be supposed to possess the degree of instability required as a basis for the image) and which are subject to delay between their initiation and full execution are all of them connected with attention.
The design of this paper, expressed in a sentence, is to point out the possible significance as a factor in the physiology of the higher mental processes, of incipient activity in motor centers. While such activity is not itself accompanied by consciousness, probably, the assumption that it has an influence such as that described above affords a means of understanding how the effect of motor response upon consciousness, obviously so great and significant, may be exerted. The theory here presented renders unnecessary any such hypothesis as that of innervation sensations. The writer presents it in the hope that it may prove worthy of some discussion, and may be in some measure suggestive. In a later paper she hopes to discuss further the view of attention sketched above, and also the nature and functions of those very important movement systems, the bodily attitudes.