The Making of Mind. A R Luria
IN THE LATE 1940s and early 1950s we worked on the role of speech in the formation of normal behavior in young children and on the failure of speech to assume normal regulative functions in retarded children. The role of human language in the formation and regulation of human activity has fascinated me from the very beginnings of my career. It is an issue to which I have returned again and again, each time in a different form. I have described, albeit briefly, the experiments I did while still a young man in Kazan which attempted to use verbal suggestion to influence reaction time. At that period in my life I was interested in modifying states of fatigue among workers. A little later I designed studies where overt motor responses to meaningful verbal stimuli were used to investigate the dynamics of hidden psychological complexes. In the mid-1920s, when we were just beginning to do clinical work, Vygotsky and I began to explore ways in which language could be used to reorganize the mental processes of patients suffering from neurological disorders, such as Parkinson's disease, so they could compensate for some of the symptoms. In the next decade our studies focused on the development of higher mental processes in children, especially twins, and the role of language in the formation of these processes.
In the years following World War II there was a great resurgence of interest in the use of Pavlovian physiology as a means of explaining all forms of human and animal behavior. This trend influenced my work for a number of years. The main strengths of the Pavlovian school lay in its extensive use of laboratory models of behavior and the sophisticated forms of experimentation that it had evolved through the years. Although my use of laboratory models was in many ways similar to that of Pavlovian scholars, I had several reservations about Pavlovian methods as they were then applied. In particular, I thought they offered an overly mechanistic and simplified explanation of human behavior which relied too heavily on the concepts of reinforcement and conditioning, that is, on the formation of temporary connections between stimuli and responses. The most dogmatic Pavlovians applied these concepts as if the behavior of children at different ages represented only the quantitative complication of simple stimulus-response principles, whereas I and others who had worked vith Vygotsky believed that children's behavior underwent qualitative changes as they grew older.
In the early 1950s I moved my base of operations to the Institute of Defectology which Vygotsky had founded many years before. There I initiated a series of experiments in which the child's own speech was used to organize simple movements in response to arbitrary physical stimuli.
We were interested in how the verbal regulation of behavior develops. To improve our understanding of the way in which the organization of behavior in normal children changes from being natural and unmediated to being instrumental aid mediated as they grow older, we made our experiments comparative. We compared the influence of speech on the organization of behavior of normal children at different ages, as well as comparing normal behavior with that of children who were suffering from various forms of mental retardation.
When we began this work, the major Pavlovian theorist in the area, A. G. Ivanov-Smoletisky, was using a version of the combined motor method which I had employed in my early research as summarized in The Nature of Human Conflicts. Ivanov Smolensky used the technique as follows: a child would be presented a long series of trials in which he had to learn to press or to refrain from pressing a rubber bulb when one of a set of lights came on. The child, who was not told ahead of time what the task was, had to discover it through what Pavlovian theorists claimed was a kind of “verbal reinforcement” in the form of the instructions “press” or “don't press”. Ivanov-Smolensky drew a strong parallel between the words “press” and “don't press” and the presentation of food to a dog following a signal of one type or another, and he viewed the child's mastery of this problem as a form of Pavlovian conditioning.
It should come as no surprise that I rejected this interpretation of the child's behavior and was not particularly happy with the way in which the experiments were conducted. In my opinion, Pavlovian scholars neglected the fact that each stimulus given to the child, especially a stimulus of the sort “press” or “don't press”, evoked a conceptually derived generalization. Following as few as one or two trials, most human beings began to formulate a general rule of the sort, “I have to press whenever there's a red light” or “I have to refrain from pressing whenever the light is green.” If my guess about the subject's reaction to this situation was correct, then the child was not reacting to the separate verbal reinforcements in these conditioning experiments. Rather, he was trying to discover a general strategy appropriate to the particular experiment. Believing that these kinds of experiments with verbal reinforcement were misguided, I decided to make a study of the real mechanisms underlying the formation of such motor responses.
While using a general methodological framework that was compatible with Pavlovian techniques, we developed an experimental method which, in my opinion, was more appropriate to an understanding of the psychological system with which we were concerned. We began each experimental session by giving the subject a set of verbal instructions that were meant to evoke a simple motor reaction. Then the limits within which the child was able to follow these instructions were studied, and the tasks were modified in various ways so that we could learn how children of different ages and different neurodynamic characteristics came to master or failed to master problems of this kind.
We found that normal children of two to two-and-a-half years of age are unable to follow even the simplest direct verbal instructions if the instructions are given prior to the task itself. When we instructed children of this age, “When the red light appears, squeeze,” they reacted to the verbal instruction itself and began to press the bulb at once instead of waiting for the red light to appear. The first part of the verbal instruction - “When the red light appears” - evoked what we, using Pavlovian terminology, called an “orienting reflex.” That is, the child began to search for the light. The second part of the verbal instruction “squeeze” evoked an immediate motor reaction, and the child pressed the bulb. The intended stimulus, the red light, actually became a distracting factor, and children who had already begun to press the bulb at the mention of the word “squeeze” often stopped responding altogether when the light came on. In addition, each verbal command “squeeze” evoked, not a unitary bulb-squeeze, but a whole series of involuntary motor reactions which only gradually exhausted themselves. Even the direct negative instruction “stop,” frequently led to excitation and to stronger, less controlled motor responses.
This picture began to change when we observed normal children between the ages of three and four. They could follow the instruction “squeeze” with only a few, if any, extraneous responses. In the course of the most simple experiments they learned to listen to the instructions and to wait until the appropriate stimuli appeared. We called this ability to withhold a response and to organize it in terms of preliminary verbal instructions a “functional barrier.” We believed that the children were verbally formulating a general rule for themselves which served as a barrier against the tendency to respond directly to the verbal instruction.
Although there is a clear improvement at this age, the verbal regulation of the three to four-year-old's motor responses are still easily disorganized. In order to produce such disorganization, we needed to change the conditions of the experiment only slightly. Instead of asking the child to respond to a single stimulus - to squeeze or not squeeze when a red light came on - we asked him to make a choice: “When you see a green light, do nothing. When you see a red light, squeeze.”
We found two kinds of disorganization in the performance of children between the ages of three and three-and-a-half when they were asked to make such a choice. One group of children continued to respond when a negative stimulus, the green light, followed a positive stimulus, the red light. That is, the negative stimulus evoked an impulsive motor reaction which in Pavlovian terms could be explained as “irradiation of excitation.” This is another way of saying that the initial verbal instruction was no longer controlling the children's actions because the first part of the response, pressing a button when a red light came on, spilled over into the response to the green light.
Another group of children failed to respond when a red light, the positive stimulus, followed a green light, the negative stimulus. Again, using Pavlovian terminology, we spoke of this as “irradiation of inhibition,” evoked by the preceding negative stimulus. An analogous phenomenon appeared when we asked children to react with their right hand to a red light and their left hand to a green light. After pressing the button with their right hand, the children continued to use their right hand to respond to any stimulus even when the left hand was specified by the preliminary instructions.
None of these mistakes occurred because the children forgot the instructions. Following each experimental session, we asked the children to repeat the verbal instructions. All were able to do so adequately, even though they were still unable to do what they said they were supposed to do. The behavior of these normal children began to be brought under verbal control only after the age of four. By the age of six they had no difficulty at all with these kinds of tasks. Then they made mistakes only when we asked them to react as quickly as possible, or when fatigue set in. We summarized this pattern of age-related changes in response to verbal instruction in Pavlovian terms by speaking of the gradual development of selectivity and plasticity in the nervous processes. Within the theory developed by Vygotsky we spoke of the gradual formation of complex models of verbally controlled programs of motor behavior.
When we began to conduct similar experiments with mentally retarded children, we found that even at the age of seven, when Russian children normally start school, severely feeble-minded children could not follow the simplest direct instructions. If we presented our single stimulus experiment in which children had to press or refrain from pressing when a single light came on, we found that saying the words “When the red light appears” prompted them to try and find the light, while the order “squeeze” evoked an immediate motor response and they squeezed the bulb. Each link of these instructions evoked a direct orienting or motor response. In fact, the children's uncontrolled motor impulses evoked by the word “squeeze” often did not stop after the order “stop” was given. In some cases the second command increased the excitation in the child's motor system, and he responded more intensely.
These feeble-minded children were completely unable to create the more complex programs of behavior required by the second task, in which the child had to choose between responding and not responding. They could not shift between the required negative and positive motor responses and continued to react to the second stimulus as they had to the first. Thus, if a positive red light came on following a negative green light, they continued not to respond. In the situation where they were told to change hands when the light changed, they continued to use whatever hand they had begun with. Mildly retarded children were able to respond correctly in the simplest version of this kind of problem. Their symptoms were sometimes less clearly expressed and appeared only in the more complex versions of the task.
Explaining these phenomena within a Pavlovian framework was not at all a simple matter. At first glance there seemed to be two possibilities. Perhaps the children's difficulties were caused by an imbalance between excitation and inhibition, or perhaps they were caused by insufficient plasticity of the nervous processes. Although these explanations arose in connection with Pavlovian physiological concepts, the terminology should not obscure the main ideas. By speaking of a balance between excitatory and inhibitory processes, psychophysiologists were referring to the possibility that the nervous system might have a general bias so that either excitation or inhibition would tend to be pervasive. If inhibition were dominant, the child would fatigue rapidly and would be unable to respond; if excitation prevailed, the child would over-respond to the verbal instructions.
Although in our opinion the notion that mental retardation is caused by an insufficient balance between excitatory and inhibitory neural processes does not explain the phenomena we observed, we could not rule it out as a factor. A lack of balance between excitation and inhibition was long ago described by Pavlov as the basic symptom of neurosis and was studied in a number of experiments by B. M. Teplov and V. D. Nebylitsen, and its applicability to phenomena of immature children's behavior and mental retardation was a distinct possibility when this work began.
Of the two explanations, we much preferred the one that focused on the lack of plasticity and the inertia in the neural processes of feeble-minded children. As experienced teachers of the retarded know, changing from one lesson to another is not easy for feeble-minded pupils. After an hour of spelling, retarded children often continue to spell even when the lesson changes to arithmetic. We thought that the same kind of explanation might apply in our experiments. However, unlike several of the more dogmatic Pavlovian physiologists, who thought that a combination of a lack of plasticity and an imbalance between inhibitory and excitatory processes explained mental retardation, we thought that such an explanation was insufficient and that a more sophisticated approach to the problem was necessary.
As might be anticipated, we used the line of reasoning advanced by Vygotsky to discriminate between different forms of behavioral retardation and to provide a firmer basis for experirnents on the neurodynamic basis of retardation. The primary distinction was between behavior organized on the basis of higher, mediated processes and behavior based on natural processes. We acknowledged that while it was possible for the neurodynamic mechanisms posited by Pavlovian theory, such as the interplay between excitation and inhibition, to operate equally at both levels, it was also possible for a pathological condition to be present predominantly at either the higher or lower level. We hypothesized that in cases where the lower level has suffered, it is possible to compensate for the difficulties by changes in the organization of the activity using the higher, better preserved levels. In other cases we supposed that the psychophysiological situation is reversed. The higher level of organization of behavior is itself disturbed and hence cannot be used to compensate for behavioral defects. In such cases only compensation that uses lower functions would have a chance of working. To confirm this hypothesis, we needed techniques that would allow us to carry out studies of the neurodynamic features of children's behavior at both the lower and higher levels.
Although our approach was theoretically sound, to prove the correctness of our general formulation would be no easy matter. The unity of lower and higher levels of behavior in human beings does not allow a full separation of the two levels. Rather, the best we could hope for would be to construct experimental situations that would allow us to vary the relative importance of the two levels with respect to any particular activity.
My co-worker E. D. Homskaya and I used the combined motor method in a series of three experiments to study the controlling functions of speech on motor and verbal processes. We hoped that by using speech responses in one case and motor responses in the other, we would achieve the desired differentiation between levels of behavioral organization. In one situation the children were told to respond to a verbal instruction with a simple motor response, squeezing a recording device, as in the previous studies. In the second situation they had to respond by saying the word “yes” to the red stimulus and “no” to the green one. By comparing the children's responses in these two modalities, we could see if there were differences in the plasticity of the nervous system at the higher, verbal level of behavior and at the lower, motor level. In a third experimental situation the verbal and motor responses were combined: the children had to say “yes” and press a recording device at the same time or say “no” and refrain from pressing the device.
The exact psychological consequences of each kind of task had to be analyzed carefully. In the first experimental condition, where only motor reactions were required, the child needed both a well-balanced relation between excitatory and inhibitory processes and a high degree of plasticity in the motor system to complete the task successfully. When only a verbal reaction was required, excitatory-inhibitory balance and plasticity in the motor system were irrelevant. The required reaction would be hindered only if the higher level at which verbal behavior is organized were disturbed. The last experimental arrangement was of course more complex. In order to cope adequately with this problem, it was necessary for the child to establish a functional system coordinating both motor and verbal components. If such a functional system were not formed, the verbal and motor components would represent no more than parallel actions, and it was entirely possible that they would interfere with each other.
We found that normal children from two to two-and-a-half years of age were unable to respond appropriately in any of these experimental situations. Excitation of their motor impulses was so generalized that they were unable to carry out the required program of motor activity. Their verbal reactions were also disorganized. They inertly repeated “yes, yes” or “no, no,” depending on the initial stimulus. And it was totally impossible for them to combine motor and verbal responses. As a rule these two actions inhibited each other.
I hasten to caution that in natural situations when a two to two-and-a-half year old's actions are meaningful to him and are bolstered by relevant previous experience, he will not behave inertly. A two-year-old child who stretches out his hand for candy will not continue to stretch out his hand once the candy is in it. But under artificial laboratory conditions, when pressing a button or saying “yes” is not accompanied by immediate rewards and occurs in response to an arbitrary verbal instruction, there is a certain inertia of both motor and verbal systems.
A different picture emerges when three to three-and-a-half year olds engage in these experiments. At this age the child's motor system in the artificial laboratory situation becomes more plastic and loses its earlier inertia. The verbal system also begins to become more flexible. The child who responds by saying “yes” or “no” to the conditioning stimuli no longer keeps on repeating these words. Combining the verbal and motor responses usually only slightly improves the child's performance on the motor component of the task. In a few cases we observed a definite improvement in the child's performance when these two response modes were combined. Saying “yes” or commanding himself “go,” the child's motor responses begin to be better organized and controlled, and he overcomes the inertia that has been typical of his motor responses at an earlier age.
We noted a remarkable phenomenon in children between three-and-a-half and four years of age. Although saying “yes” or “no” and pressing a button is a double response, it shares a single positive direction. Both motor and verbal response systems are excited. However, when a child has to say “no” or “stop” and simultaneously must block a motor reaction, the excitation of the verbal system is positive while the meaning is negative. Since every vocal response tends to produce a motor response, even though the meaning of the response “no” is negative, a psychophysiological conflict is evoked when the negative is used. We noted a certain degree of disassociation between verbal and motor reactions during this three-and-a-half to four year old transitional stage. While saying “no,” the child often presses the button and does not inhibit his movements. When the child is a few months older, or if we institute special training in which the meaning of the verbal response is emphasized by explicit rewards, a new functional system is formed which begins to regulate the child's motor reactions. His motor activity comes under the control of word meaning rather than the primitive response to the mere sound produced by speech. I viewed this result as an indication that we had created a model of how the child's language comes to control his behavior under special laboratory conditions.
This transition from impulsive responses to responses controlled by the meaning of an utterance occurred in children somewhere between the age of three-and-a-half and four, although the exact timing differs greatly, depending on subtle features of the experiment and the particular children involved. It is difficult to specify exactly when and under what conditions the newly organized functional system can be observed, because it is initially quite fragile. Numerous studies conducted in various countries have both supported our findings and come up with conflicting results. The discrepancies can be explained only by careful analysis. In this kind of experiment, using children between the ages of three-and-a-half and four, even the slightest differences in the morphology of the verbal responses, such as between the instruction “go” and “no” or “press” and “don't press,” can be important. But the essential point is that each set of deficiencies can be observed at a particular period in the child's development and will disappear during subsequent periods. As I see it, the sequence of changes is the important factor and not the precise age at which the new functional system appears. It is also important to emphasize again that these experiments are no more than experimental, laboratory models of the development of control in children's behavior.
When we were carrying out this work, we knew that specialists distinguished two basic forms of mental retardation in children, excluding retardation caused by local brain lesions and the still unclear group usually labeled “minimal brain disorders.” We called one type “general asthenia.” This type of retardation is usually caused by undernourishment and some somatic discases. The other form is true feeble-mindedness or retardation, and it is caused by intrauterine brain intoxication, congenital traumas, or in some cases genotypic factors. Because the two forms of retardation often share similar symptoms, it is not an easy task to distinguish between them. We attempted to use our general characterization of the development of verbal self-regulation as a means of providing a differential diagnosis of these two forms. We hypothesized that in retardation of the sort assoclated with general asthenia, the symptoms of the disorder were the result of a dysfunction in lower, somatic processes. If this were the case, the neurodynamic features associated with excitation and inhibition in the motor system would be more deficient than the neurodynamic features of the higher, verbal system. It followed that the better preserved verbal system could be used to help overcome the neurodynamic insufficiencies of the motor system.
The situation would be different for children suffering from essential feeble-mindedness. We hypothesized that their higher functions, including their verbal system, would suffer more than their lower functions. Therefore the children's speech would be of little help in reorganizing their behavior or in compensating for their defects.
Our work on this kind of differential diagnosis began in the early 1950s and took many years to complete. It is summarized in a two-volume monograph, Problems in the Higher Nervous Activity of Normal and Abnormal Children, published in Russian in 1956 and 1958. [English summaries of this work appear in A. R. Luria, The Role of Speech in the Regulation of Normal and Abnormal Behavior (Pergamon Press, 1960).] In this work E. D. Homskaya observed that although children with an asthenic syndrome showed marked disturbances when required to make motor responses to verbal instructions, they did not show similar difficulties when asked to make only verbal responses. They could respond “yes” and “no” in an appropriate way, but they over-reacted when required to make a movement in response to a light. They failed to respond if a positive stimulus followed a negative one. They also showed inertness in the face of an initial positive stimulus, continuing to respond even when given the negative stimulus. Since their defects seemed to be localized in their motor system, we hypothesized that it would be possible to use the verbal system to help bring the motor system under control.
We were correct. We found that when we combined the verbal and motor responses, children who suffered from general excitation began to respond more regularly and more appropriately to the instructions. They stopped responding impulsively to negative stimuli altogether. And children whose behavior was inert began, with the help of their own verbal responses, to make stable motor responses to positive stimuli.
Our study of truly retarded children yielded totally different results. Members of our research group, including Dr. A. I. Meshcheryakov, Dr. V. I. Lubovsky, and Dr. E. N. Martsinovskaya, showed that the neurodynamic disturbances in the verbal processes of these children were far more pronounced than their motor processes. The difficulties that we had associated with an imbalance between excitatory and inhibitory processes or with the problem of pathological inertia were evident in both the motor and the verbal systems of these children. Since the same difficulties appeared in the same form in both the verbal and motor systems, we could confidently predict that combining verbal and motor responses would not improve the performance of the truly retarded children.
All these observations were made under artificial laboratory conditions. Although they are useful both as experimental models of the development of verbally controlled behavior and as diagnostic aids, they should not be overgeneralized. To establish the generality of the laws, we knew that we had to investigate children's natural behavior in order to understand the distortions that the laboratory conditions might have introduced. In the late 1950s we conducted a series of special observations in nurseries.
We found generally that one to one-and-a-half year olds cannot follow verbal instructions if they are not linked to feeding. And even in the feeding situation the intonational components of the stimulus and their place in the situation as a whole, rather than word meaning, are clearly decisive in regulating the children's behavior. The behavior of a child crawling on the floor is almost wholly determined by his orienting reactions to the physical characteristic of commands, no matter what the verbal instructions. This same phenomenon can be seen when a child is next to a table on which there are several toys. The verbal instruction, “Please give me the fish,” may start a reaction, but it does not yet program the child's behavior. The child's gaze may turn to the fish, and his hand may start to move toward it, but both his eyes and his hand easily wander off the instructed course. The child, when he orients, is attracted by nearer, brighter, newer, or more interesting objects. Only by singling out the object named and moving it can one make it more attractive to the child. In short, the child is under the control of the physical features of the situation. What we wanted to discover were ways in which we could describe how verbal instruction comes into conflict with the influence of the immediate physical environment in very young children and finally comes to dominate it.
Our experiments were really quite simple. Two objects, a wooden cup and a small wooden wine glass, were placed on a table. A coin was put in the cup as the child watched. Then, asking the child, “Where's the coin?” we instructed him to find it. The youngest children. of one to one-and-a-half years old, reached for the cup, but their orienting response was still so strong that they usually grasped both the cup and the wine glass at the same time. Only somewhat later did children grasp the cup and find the coin. Verbal instructions became decisive only with children who were about two years old.
Next we wanted to know whether those children who correctly did as they were told would continue to remember the command after a short delay and whether it would still control their behavior. We introduced a delay of about ten to fifteen seconds between the time in which we put the coin in the cup and asked the children to find it and the time that they were free to begin their search. We found that the youngest children could not do the task at all. Although they could follow the verbal instruction immediately, the instruction lost its controlling function when there was a delay, and they resorted to grasping both objects at once. Children who were slightly older were able to retain control of their behavior despite the delay.
In both these cases our observations dealt with a combination of verbal commands and immediate stimulation: the child observed where the coin was placed and heard the instructions. What would happen if we isolated these factors from each other? Could the child follow the verbal command if it was not supplemented by the immediate visual stimulus? A screen was placed between the child and the containers on the table so that the child could not see where the coin was placed. Then the command was given: “The coin is in the cup. Please find the coin. ”
Children one-and-a-half to three years of age who easily performed the task in the previous series were unable to follow the “pure” verbal command unaided by the visual stimulus. They became confused and frequently picked up both the cup and the wine glass. Around the age of three and a half the pure verbal command began to assume a controlling role.
After we had gotten a glimpse of the early stages of the way in which verbal commands assume a controlling function, we began to investigate the stability of that function. We did so in two ways. First we created a situation in which the verbal command was in conflict with the child's prior experience. Second, we created a situation in which the verbal command conflicted with the immediate information in the visual environment. To create the first kind of conflict, the coin was placed in the cup three to four times to establish a set of expectations in the child. Then the pattern was broken and the coin was placed in the wine glass. In both the visually aided and the screened series, this complication made children who had previously been able to follow the verbal commands unable to do so. In this situation the children continued to grasp the cup. Only after a few months were they able to overcome this complication and complete the task.
The second kind of conflict that we introduced turned out to be even more complicated. We instructed children two-and-a half to three years of age, “If I lift my fist, you lift your finger” or “If I lift my finger, you lift your fist.” The youngest children had some difficulty repeating these verbal instructions and sometimes simplified them. The three to three-and-a-half year olds had no difficulty with this part of the task. But when the children tried to follow their own commands, they had a hard time. Observing the experimenter's fist, the younger children imitated him rather than following the verbal instruction. They showed no signs of being aware that there was a discrepancy between what they were doing and the verbal instructions. The older children, however, displayed some signs of conflict. In response to the experimenter's fist, they correctly raised their fingers, expressed doubts, and then replaced their fingers with their fists. Only after some months, when the children were nearly four years old and older, were they able to follow verbal instructions that conflicted with what they saw.
I am aware that this work constituted no more than a series of pilot studies and that additional extensive observations must still be made. Nonetheless, a strategy that combines artificial laboratory models with more natural kinds of observations and with quasi experiments is extremely fruitful.