The Making of Mind. A R Luria

Mental Development in Twins

THE IDEA of studying identical and fraternal twins in order to separate the contributions of heredity and environment to a particular human characteristic was by no means original with us. At the time we undertook this work in the early 1930s, we were familiar with the work of K. J. HoIzinger, Cyril Burt, and others who had begun to exploit the possibilities for exploring the origins of human intellectual functions that were inherent in the existence of identical and fraternal twins.

The logic of this approach is by now sufficiently familiar that a sketchy review is all that is needed to show the foundations on which we built. The simplest comparisons involve identical and fraternal twins raised at home. Here one can assume that the social environment for each member of a twin pair is more or less homogeneous, although among pairs there may be a great deal of environmental homogeneity as well, depending on the particular life circumstances of the families, such as educated versus non-educated parents, urban versus rural settings. However, the biological similarity between two twins in the same family will differ according to whether they are monozygotic (identical) or heterozygotic (fraternal). With constant environmental influences, one can assume essentially identical intellectual abilities for identical twins since both environmental and biological causes are more or less identical. For fraternal twins one can assume greater variability owing to differences in genetic makeup. More complex comparisons involving twins separated shortly after birth, which would vary the environmental as well as the biological antecedents of behavior, are also possible, but in practice such work will be complicated by the many unknown factors arising from differences of the environments in which these children are placed.

In the early 1930s an extremely propitious opportunity to investigate these issues arose as a result of research on genetics that was then going on at the Medico-Genetic Institute in Moscow. It was proposed that we set up a research program at the institute, which had outstanding boarding facilities, a fine educational program, and virtually unlimited access to twins from all over the USSR.

We undertook this work from our own theoretical perspective. The logic of varying environmental and biological factors using twins was clear enough, but we felt that previous research had been severely handicapped, not only in the limited numbers of subjects that were usually included from each of the crucial groups, but also in the weak measures of intellectual function that were used to assess environmental and biological influences. We were particularly unhappy with the use of standardized IQ tests as the indicators of intellectual development. These tests, which were developed on a purely pragmatic basis to predict school performance, seemed to us then, as now, to be a hopelessly atheoretical and opaque means of observing the structure of higher psychological functions.

Drawing on the many pilot studies of the late 1920s, we conceived a more complex set of relations between the performance of identical and fraternal twins which depended jointly on the nature of the specific task presented, the theoretical contributions of natural (biological) and environmental (cultural) factors to each kind of task performance, and the age of the child.

Natural and cultural processes not only arise from different sources but also change in different ways in the course of development. As a child grows older, natural processes change quantitatively. His muscles become larger, his brain grows more myelinated, his limbs change in size; analogously, his mental processes increase in power. But the basic principles of their action remain unchanged. Testing memory span, for example, is a means of gauging natural memory, because the child simply has to reproduce the stimuli without changing any of the information presented. Involuntary remembering also shares the property that the child does nothing special to accomplish recall; the material simply “impresses itself” on him.

Cultural processes, on the other hand, change qualitatively. To use memory as an example again, it is not just that the growing child's natural capacity to register and retrieve information increases; as a result of the increasing influence of the social environment, changes occur. In the principles whereby the information is registered and retrieved. Instead of remembering naturally by retaining impressions and involuntarily reproducing them, the child gradually learns to organize his memory and to bring it under voluntary control through use of the mental tools of his culture.

Our research assumed that genetic contributions to behavior will be more directly reflected in tasks that call for natural cognitive processes than in ones that evoke culturally mediated processes. Drawing on our ideas concerning the developmental course of natural processes, we assumed a stable relation between heredity and behavior in natural cognitive tasks as the child matures. But for cultural processes we assumed a changing relation. We reasoned that among young children, in whom cultural processes still play a subordinate role, children who are genetically similar will behave similarly because their behavior is based on natural processes. But as culturally determined forms of information processing come to be relied on more and more, the children's environment will have a greater effect on behavior than does their genotype. Thus in older children, a similar environment will lead to similiar performance in tasks that require mediated, culturally influenced modes of cognition, even if the children differ genetically.

The logic of our approach required us to arrange that children be exposed to tasks that varied in the extent to which natural and cultural psychological processes were brought into play. We also wanted to work with children whose ages ranged from the period in which natural processes are dominant (5-7 years) up to the period where cultural processes are normally dominant. (11 - 13 years). Finally, we needed a way to vary genetic factors while holding environmental factors constant. Comparing identical and fraternal twins seemed the ideal technique.

Our previous research helped us to determine the age range of the children needed. Our younger group was composed of five-to-seven-year-olds. Although we had been able to see the early stages of the emergence of culturally mediated information processing in children of this age under specially arranged conditions, we knew they were unlikely to apply cultural processes in most tasks. Our older group was made up of eleven- to thirteen-year-olds, who we knew were likely to use cultural processes if the task permitted. Within each age group we had approximately fifty pairs of twins, half identical and half fraternal.

One series of experiments consisted of three memory tasks studied extensively by Leontiev in his research on the development of memory. In the first task we presented nine geometric figures to the children and later asked them to recognize these figures from a set of thirty-four. We considered the visual recognition required by this study a good example of natural, direct remembering.

In our next task fifteen difficult-to-remember words were presented one at a time to each subject. The subject was asked to recall them all following presentation of the entire set. This task could be done either by simply remembering the words directly or by using complicated mediated processes to remember the words. For example, the subject could think of a word to help him remember each word presented or could conjure up an image to help him recall the necessary word. Because words could be recalled either directly or in a mediated way, we believed that this task allowed either natural or cultural processing. However, the difficulty of the task made it seem likely that natural processes would dominate in all but the oldest, most sophisticated children.

In the third task we asked each child to remember another set of fifteen words. As we presented each word to the child, we showed him a picture that he could use as an auxiliary sign to help him remember the word. The pictures were not connected with the words in any obvious way, so the child had to establish artificial links between them in order for them to be effective aids in helping him to remember. As in the second study, we presented the words and their associated pictures repeatedly until the child could remember all the words on the list. Then, when the child had memorized all the words in this way, we showed him the pictures one by one and asked him to recall the word that went with it.

It would have been difficult to ascertain anything like a structural change in the way that remembering was carried out if we had considered a quantitative analysis of the data alone. There were the usual age-related increases in remembering associated with each of these three tasks. But a qualitative analysis of the data revealed some important facts.

We found that the structure of visual memory for geometric figures was as elementary and natural in the older group as in the younger group. Almost none of our subjects used indirect or logical processes to any noticeable degree in memorizing the geometric figures. By contrast, a qualitative analysis of the results of the third task, in which each word presented to the children was accompanied by a picture, yielded quite different results. Most of the younger children remembered the words in the same direct way that they remembered either the visual figures in the first task or the orally presented words in the second. They were unable to use the auxiliary stimuli to build a logical connection between the picture and the words to be remembered. In many cases the memorized word was recalled in the absence of any connection between the word and the picture cue. The children were unable to give any information at all about the connection between the words recalled and the pictures being seen. When we questioned these children, they usually answered, “I simply remembered it.” I do not think that their answers reflected a lack of introspection. Their remembering simply had not been mediated by the pictures.

We also found that the picture was a good auxiliary device for remembering for some of the children, not because of any logical connection between the words and the picture, but because of visual similarity. When using the picture as a cue to help them to remember, such children did not form a logical connection but tried to see the word in the picture. For instance, one child remembered the word “sun” when presented the picture of an ax. When we asked how he remembered the word, the child pointed to a small yellow spot on the picture of the ax and said, “Look, here is the sun.” In one form or another, natural processes dominated.

When we studied the qualitative features of the recall of older children, we found the process of remembering by establishing similarities between the word and the picture changed to remembering by creating logical connections between the word and the picture. In these children we rarely found a word that had been remembered in an elementary, direct manner or by means of visual connection with the picture. They remembered the word “sun,” for example, when using the auxiliary picture ax by creating such logical connections as, “We work with an ax and the sparks glitter in the sun” or “A man worked with an ax on a sunny day.”

These observations established that our tasks were successful in evoking both cultural, mediated responding in some cases and direct, natural responding in others. These results provided the background for the next step in our analysis, to see if we could demonstrate that natural remembering is most closely related to the genetic makeup of the child, whereas the cultural form is related to his environment.

We reasoned as follows: the genetically determined, natural process should be similar in identical twins; that is, the difference in performance for such twins ought to be small. In particular, it ought to be smaller than the difference in scores for fraternal twins who share a similar environment but relatively dissimilar genetic makeup. If we call the difference in scores for a given test D, we can calculate two different D scores: D_i is the difference between scores of identical twins, and D_f is the difference between fraternal twins. Our reasoning led us to concentrate on the ratio of these differences. In particular, we expected D_f > D_i for natural processes, and D_f = D_i for cultural processes, since genetic similarity is not influential and the environments of the different kinds of twins ought to be equally susceptible to cultural influences.

The quantitative results in terms of the difference scores D_f and D_i for the natural and mediated tasks in the two age groups were as follows:

The pattern of results fitted our prior expectations rather neatly. For the task where natural, direct remembering dominated, the identical twins performed almost three times more similarly than the fraternal twins at both age levels (e.g. D_f > D_i). In the mediated task, the same pattern emerged for the younger children, although the superiority of D_f over D_i was much less; but among the older children, D_f and D_i were almost identical. This is exactly what we had predicted would happen on the assumption that for this task the older children would come to use mediated, culturally patterned forms of remembering. A report of this research was published in the American Journal Character and Personality in the late 1930s.

Psychologists for the most part have been interested in twins in order to separate environmentally influenced processes from those influenced by heredity, as we attempted to do in the study just described. But twins are of enormous importance for psychological research over and above that set of questions. Because identical twins usually share a very similar home environment, more so than that of other siblings by virtue of their identical ages, and because their physical similarities often induce adults to treat them alike, identical twins also represent an interesting opportunity to study the effects of environmental variation because psychologists can introduce such variation in a planned manner against an unusually constant background, thus making their studies maximally sensitive to detecting environmental influences.

V. N. Kolbanovsky, A. N. Mirenova, and I collaborated on a second series of studies to see if we could design educational games that would develop children's ability to engage in constructive activity. We chose to concentrate our study on constructive play because games designed to develop such activity are widely used in preschool. They usually consist of a set of blocks of various shapes which the child can use to build different kinds of structures. Most educators argue that in addition to being entertaining, constructive play helps develop a child's imagination and elementary mental processes, including the ability to discriminate shapes and to estimate visually.

We began by surveying the kind of educational material used to stimulate constructive activity in preschoolers and the way these materials were used by teachers. We distinguished two main methods of inducing children to engage in constructive activity. In the first, the teachers required the children to construct block structures using a model. On occasion these models were drawings of structures in which outlines of all the elements necessary to build the structure were shown, so the children could copy the model by putting the blocks together step by step. This kind of construction was supposed to teach the child to pursue specific goals, to concentrate on the task, and to analyze the patterns of the model and to discriminate its component parts.

However, we found that it rarely held the attention of preschoolers. The work of putting the blocks in place according to a preselected pattern was often so boring that the teacher had to coax children to complete the task.

Perhaps as a result of the shortcomings of this highly structured method, many teachers allowed preschoolers to play freely with blocks, building whatever they wanted. While this free play may be conducive to the development of children's creative irnagination, we doubted that it had any further educative effects.

In fact, our analysis suggested that both these educational strategies had shortcomings. From our point of view, real constructive activity should give the child a carefully defined goal. This goal should be either stated verbally or presented in the form of a model which the child has to copy. In carrying out this goal, the child should have to analyze the problem and find ways to solve it by selecting blocks which match the distinctive features of the particular structure and by rejecting those not suited to the task. Neither of the two forms of constructive activity we had observed met all of these conditions. A child who is given a detailed model to copy does not really need to analyze the problem. All the child has to do is select the blocks that are in the drawing or the model one by one and put them in place. There is nothing in the task presented in this way that requires reasoning. About the best we can hope from it is that it gives the child an opportunity to make elementary discriminations, which most children of this age have already practiced extensively.

Free block building gives the child an appealing goal to work toward, but it does not tell the child how to achieve it. The child has to find his own way by selecting those building blocks and procedures suited for the task. Both the task and the means used to complete it are flexible. As the child builds, the task often becomes more subtle and refined, and new details are added. Because of these features, free building activity is usually interesting to the child and holds his attention for a relatively long time.

Despite these positive features, free building activity also has a number of drawbacks from a pedagogical point of view. In formulating a construction task in free play situations, the child usually is not concerned with whether he can actually build the structure with the materials at hand. Another drawback of the free play situation is that it often drifts from true constructive activity into ad hoc creative play. By ad hoc creative play I mean that the child assigns meanings to things on the spur of the moment for the purposes of the game, like putting down a block and calling it a car. To some extent, such play can get along without any analysis of the objective properties of the materials with which the child is playing. He can put a block out and say, “Let this be a tree.” Another block may be a car, and a third, smaller block, a dog. Using the blocks, children, especially the younger preschoolers, give free rein to their creative imaginations. This kind of play is often entertaining, but it does not develop the child's observational skills or his ability to analyze the problem and determine the elements and combinations of elements that are best suited for the constructive task.

With these objections in mind, Mirenova and I developed the following constructive task. We presented the child with a model which he was asked to copy with a set of blocks. The models we presented differed from those normally used in preschools in that the contours of the individual elements needed to build the structure were hidden. This was accomplished either by giving the child a two-dimensional outline of the model he was to construct or by presenting him with a three dimensional model covered with thick white paper so that the general contours were visible but the individual components were not.

The typical model, in which each of the individual blocks in the structure are visible, can be copied by the child simply by matching, visually, the blocks before him with the individual blocks in the structure he is trying to copy. But the outline model gave the child a specific goal without giving him explicit information on how to accomplish this goal, because he could not see how the components were put together. In contrast to the normal type of constructive task, the outline model required that the child himself figure out which of many possible blocks were best suited to the task. This challenge made the work interesting and helped to hold his attention.

Our construction task also differed from free play block building by requiring the child to remain within the framework provided by the model, against which he could continually monitor his progress. Even slight discrepancies in dimensions or shapes were noticeable when the child compared the structure he had built to the model. This requirement kept the child within the framework of constructive thinking and prevented him from drifting off into ad hoc play.

We felt that play of this sort would develop complex forms of perceptual activity in children. Their direct, unanalyzed perception of the building model would not suffice; they had to organize their examination of the model to distinguish critical elements and relations. They had freedom of choice in ordering some parts of the task, but they also had to work within definite constraints.

We tested these ideas about constructive play on five pairs of identical twins attending school at the Medico-Genetic Institute. We began by conducting a number of psychological tests to determine if the twins' perceptual and cognitive development was in the normal range expected for children of this age. We also observed their visual constructive thinking. After these examinations, the two groups were given control tasks to perform. In one control task the child was asked to build from diagrams showing individual elements that made up the structure; in another control task he was asked to build from diagrams that showed only the general outline; and in the last task he was asked to play freely with the blocks. These control tasks gave us a baseline against which we could later measure the psychological changes brought about by our training program.

We gave one twin of each pair a training regimen based on the model in which all of the elements were clearly depicted. This group was designated the “build-from-elements group.” The second group of twins was given a training regimen using the outline model, in which the elements making up the model were not visible. This group was called “the build-from-model group.”

To ensure that all the children had the same level of experience with the kind of constructive activities we had designed for them, we chose children who were living at the Medico-Genetic Institute and attending the special kindergarten there. The two groups of children lived in different rooms, attended different kindergartens, and met each other only on walks or at times when play materials were not available. Each group of children was given training sessions with their respective construction tasks twice a day for two and a half months. In all, each twin completed a total of about fifty building sessions.

At the beginning, both groups of children had considerable difficulty with their constructive activity. They would often use the wrong size or the wrong shape of blocks and end up with structures that markedly diverged from the models they had been given to copy. One of the main difficulties exhibited by all the children was to choose a block arbitrarily and announce that it was going to represent a certain feature of the model without checking to see if it actually did resemble that feature. For example, if the model had a pointed roof, a child might use a pyramid or lean two elongated blocks against each other and call it a pointed roof without looking to see if his construction was like the pointed roof of the model. As a result, the children's buildings often bore little resemblance to the models they had been given to copy. Yet the children described these structures as if they were indeed replicas of the model and showed little or no awareness of the discrepancies.

After two months of training, we gave each of the groups some tests to determine whether the two training programs had differentially affected the development of their visual thinking and constructive activity. In the first test we gave all the children a model to copy whose elements were concealed as in the three-dimensional model. We found that the children who had trained using such models could build identical structures far more sucessfully than could children who had constructed their buildings from models in which all the elements were depicted. At first it seemed possible that this difference merely reflected a specific practice effect. However, when we presented both groups with new models depicting all the elements and asked the children to copy them, we found that those children who had practiced on models which gave only the general outlines were still superior.

What was it about the practice in the build-from-model group that had produced differences in the children's behavior? We tried to answer this question by analyzing the mistakes made by the children in the course of constructing the different models. We found that the children who had been trained in the build-from-model group planned their structures. Their first response to the task was to stop and analyze the overall pattern, whereas the children trained in the build-from-elements group just plunged ahead. We also found the build-from-model-group superior in the way in which they related elements of the structure to the whole, and they were more articulate in describing the way in which their structure differed from the model when they reached the stopping point. When the children in the build-from-elements group had completed their structures, they often maintained that their structures were identical to the model even when discrepancies were apparent. They seemed to be referring to the fact they had picked the correct elements and seemed not to notice that these elements did not bear the proper relation to the whole.

After completing an analysis of the children's performance on the training tasks, we designed a variety of new problems to try to explore the basis of the initial differences observed. One such task was to present the children with a model that had one or two pieces missing. The children in the build-from-model group seemed to have little difficulty picking out the proper elements and putting them in their places. But the children who had learned to rely on models that specified all the elements could not deal with the task at all.

This result led us to consider the possibility that the training program for the build-from-elements group had only exercised the children's elementary perception, so that they had difficulty when trying to apply these perceptual skills to more complex problems. When we tested the two groups of children on their ability to discriminate elementary figures, we found no difference between them. We also tested the children's ability to concentrate by asking them to study the difference between pairs of blocks in order to determine which were necessary to complete the next steps in a task. The children showed no difference in the amount of time they were able to concentrate.

From these observations we concluded that the differences in their performance on the criterion construction task were not the result of differences in elementary skills or attention span. Rather they were located in the children's ability to analyze complete models into their component elements and to relate these elements to the whole. We further tested this notion by presenting the children with complex figures and asking them to reproduce them from memory. The figures were made of irregularly shaped blocks which, when put together, formed a recognizable whole.

We found that the children who had trained in the build-from-model group were able to reproduce the general shape of the figures using the correct elements, whereas the children trained in the build-from-elements group were able to reproduce only the individual elements in the task and could not grasp the whole.

We then presented the children with what we called the honeycomb task. To do this task, the children had to recognize that the side of each diamond making up the honeycomb was also the side of the adjacent diamond. Like all five- to six-year-olds, the twins with whom we were working found this task difficult. However, the children in the constructive modeling group showed that they could generate rules for reproducing the honeycomb in their copies even though they made errors, while the children in the build-from-elements group were completely unable to handle the problem.

To determine how general was the analysis of the relationship between objects and their spatial configuration, we asked each child to imitate the movements of the experimenter who was facing him. If the experimenter raised his right hand, the child was supposed to raise his right hand, although the hand nearest the experimenter's right hand was the child's left. In other words, in order to respond correctly, the child had to transpose the movement in space. Consistent with our analysis, the children who had been trained in the build-from-model group were better able to do the spatial transpositions than were the children in the build-from-elements group.

Finally, we examined the quality of both groups of children's free constructions. We found that the children in the build-from-elements group had not progressed at all in their free constructive play. They often became distracted in the middle of building something and wandered off into fantasies which transformed the meanings of the elements and of the task itself. The children in the build-from-model group first formulated a plan for the building and than carried it out. We saw this as reflecting a general change in their ability to plan and execute a constructive task. In follow-ups on this work we found that there was a significant difference between the children in the two groups as much as six months later. This research was interesting not only for its general theoretical and educational implications but also for its demonstration that identical twins could serve as especially useful subjects in intensive work with small groups of children.

The special attraction of working with twins in the two series of studies described so far rested on the opportunities they offered for controlling environmental and genetic influences on intellectual development. A third study of ours with twins was based on the special social relation that sometimes arises between twins and its effect on their language and subsequent intellectual development.

It is not unusual for twins to spend most of their waking hours together. When they are very young, a great many of these hours are spent in semisupervised play; an adult is present to see that no harm comes to the children, but they are left to amuse themselves. In the course of their time together twins commonly develop figures of speech that are “private,” that is, words or phrases that have a special meaning to them but to no one else. In its simple form such private speech is common in any social group, but it is usually restricted in scope. Occasionally, however, special circumstances that further isolate twins can lead to a special “twin situation” in which the two children begin to rely heavily on the shared meanings of private words which have no identifiable significance to the adults around them.

In the course of our work we encountered such a pair of twins. Yura and Llosha, identical twin boys, had a complex phonetic impairment. They did not speak at all until they were two years old. At two and a half their only identifiable words were “mama” and “papa". At four they made only a few additional sounds when playing together. These sounds did not seem to be related in any systematic way to objects or actions; at least their mother could not detect any patterns. By the time the boys were five years old they could say a few conventional words when speaking to adults, but their play with each other did not involve much talk, and when they did speak, the sounds they made (aga, nu, ntsa, a, bulbul) seemed to be guided not by grammar but by actions and gestures. Some stable names were detectable: their own slightly distorted, made-up words, such as pi pi for “chicken,” and a small store of Russian words for common objects, for parts of the body, and for elementary actions. The twins could also understand common speech that referred to them. But when the speech did not refer to them directly or when it was grammatically complicated, it seemed to pass them by.

At home the twins spent most of their time playing together. They were not read to or intellectually stimulated in any special way. Despite their anomalous speech, they appeared to be intellectually normal. They were mischievous, energetic, and friendly. They dressed themselves, ate at the table, and helped with small chores.

The facilities of the Medico-Genetic Institute offered a unique opportunity to study these twins and the relation of speech to the development of cognitive activities. The children were enrolled in the institute's residential kindergarten and quickly adjusted to their new setting. In comparison with their classmates, both Yura and Llosha's constructive play was somewhat retarded. They never made buildings or complex structures. When playing with large blocks, they spent most of their time hauling them around the room. They played together most of the time, rarely playing with or speaking to other children. When they did play with others, it was usually in simple chasing games. They never played with other children in creative activities such as drawing or modeling, or in role play.

During their initial period in the kindergarten we were able to record a great deal of their speech. Although by this time, at five and a half years old, their vocabulary had grown some, their speech was still agrammatical. They were never seen initiating conversations with adults, and attempts to draw them into conversation usually failed, although they would sometimes point to an object mentioned or use some idiosyncratic word to name it.

Their speech with each other was closely bound to action and marked the emotional character of the activity. An analysis of their speech during eight play sessions showed that about 80 percent of all their talk was made up of amorphous expression sentences that were incomprehensible if you could not see what the children were doing. Even though they used many common words, the meaning of the words remained unstable and resembled the speech of normal two-and-a-half to three year olds. These characteristics of Yura and Llosha's speech contrasted sharply with that of their peers, who had progressed far in their mastery of the semantics and grammar of Russian.

On the basis of our belief that speech is the mechanism for constructing and maintaining culturally determined, mediated cognitive activities, we saw the situation-bound, undifferentiated character of the twins' speech as the primary cause of their primitive play. Furthermore we expected to observe other differences in the children's intellectual activities as well. So long as the twins' speech remained diffuse and action-bound, it could not serve to regulate their behavior effectively. Without adult word meanings they were, in a sense, shut off from the culture's tools of thought. This feature of the children's speech was apparent in the virtual absence of narrative speech and of speech that served a planning function, that is, could guide the children's actions. The most sophisticated planning or regulative speech we heard during this initial period consisted of a few short phrases, such as, “Llulla (Yura), throw here, you here,” and even these examples were connected with what they were doing at the time.

Once the children were accustomed to the kindergarten and we had made these baseline observations, we undertook an experiment to determine if we could modify the level of the children's speech in a short time and by so doing produce a change in their mental functions. We began with a ten-month program during which time the twins lived apart, a circumstance which in itself helped to break up the twin situation and promote the rapid acquisition of more adult speech and the parallel development of mental functions in both children. In addition, we gave Yura, the weaker and less developed twin, special speech training which was designed to help him discriminate and articulate sounds and to master adult speech. The lessons were as follows: first the child was encouraged to answer questions, then required to name objects, and finally to answer questions actively, repeat complicated phrases, and describe pictures. The instruction continued for three months and then, after a two-month break, resumed for another six months.

Following are two examples of such dialogues taken from the beginning and end of Yura's training program:

In the first of these examples Yura responded silently or not at all to the adult. Ten months later his speech, although somewhat distorted phonetically, had acquired the adult form semantically, grammatically, and functionally.

We tested the children at three months into the program and again at ten months to see if separating them had produced changes in their speech and in their mental functions. We also wanted to see if the special training that we had given Yura made an appreciable difference over and above the experience of being separated from his brother. At first the boys remained silent most of the time. But nonverbal communications were inadequate for communicating with the children and teachers in the kindergarten, and as the boys began to participate in the general life of the group, they began to speak in order to express their wishes, to participate in play, and to avoid being excluded. By the time we made our first systematic evaluation of the experimental program at three months, speech bound tightly to the specific situation and agrammatical speech had become less common than forms of speech that were appropriate for children of their age. Now, although tney still made errors in grammar and pronunciation, the children could produce utterances that were extended phrases, such as, “I wanted a house, didn't get it” or “Llosha's making a table.”

After three months the verbatim records of the children's speech showed that both boys had made great progress, but Yura had advanced more than Llosha. About 40 percent of their talk was devoted to planning functions, as in, “May we have the cubes?” or “I am going to get the truck.” But Llosha's speech was rarely narrative - that is, he rarely described his own actions or those of others - and more of it was situation bound and agrammatical. Speech referring to future or past events was virtually absent. This pattern suggested that the development of planning speech grows out of the activities and interactions that normally occur in kindergarten, but the development of narrative speech seems to need special training at this stage in the child's language development.

We discovered a second, more subtle difference in the twins' speech at this time. Although both twins used planning speech in about 40 percent of the phrases recorded, Yura was more likely to formulate plans that applied to objects and actions that were not in the immediate environment. After ten months we found that Llosha had increased the amount of his planning speech, but he continued to be less able to plan things that were not immediate.

This difference was also evident in the twins' narrative speech. Yura learned to use narrative speech before Llosha did, but after ten months of separation the total percentage of narrative speech phrases was greater for Llosha (28 percent) than for Yura (21 percent). However, even here Llosha demonstrated a continued dependence on the immediate context. For the most part, his narrative speech described his own actions and immediate perceptions, in contrast to Yura whose narration was usually about events that were not connected to the immediate situation. From these findings, I formulated the respective roles of normal experience and of special practice as follows. The necessity to communicate led to the development of objective speech, but special training was required for the children to produce differentiated, well-developed sentences.

Examining the development of the twins' ability to understand adult speech, we could see an analogous pattern of development. In normal discourse it was virtually impossible to tell the twins apart because the immediate circumstances reinforced the meaning of their speech. But in special interviews we found that Yura's ability to analyze complex grammatical constructions and inflections, which are so important in Russian, was more developed than his twin's, who experienced some difficulty in comprehension.

Although these changes in speech behavior were of no small significance, our basic concern was to determine how changes in speech affected the structure of the children's thought processes. In particular, we were interested in determining whether the same qualitative changes that we had seen in the pilot studies by Vygotsky's students would occur in the short time span of our experimental intervention with the twins. At the beginning of our intervention the twins engaged in primitive play in which they attached conditional or play meanings to objects. But their play never integrated such objects into a system that incorporated verbally formulated rules, such as, “You be driver, I'll be passenger,” or “Let's build a castle.” When the twins were allowed to play together after three months of separation, there were remarkable changes in the structure of their play. The children began by singling out a project and verbally formulating it. Object meanings no longer changed in the course of the game. At last the game had an agreed-upon objective toward which the children oriented as they played.

Next we observed how the children constructed objects of their own choosing out of clay. In contrast to their earlier behavior, each child announced what he intended to build as he began, and he more or less kept to his plan. The architectural accomplishments of our twins were remarkable not for their beauty but for the forethought that guided their creation. Being able to plan, or to preconceive, the building also made the children less distractable and, more actively determined to carry the activity to its conclusion.

While both twins improved, there were divergences their intellectual development that corresponded to the linguistic differences between them. Before they were separated, Llosha generally initiated their joint activity. While he continued to lead their games based on motor activity after their separation, Yura became the acknowledged leader in activities requiring verbal formulation.

Yura's advantage in such situations was clearly demonstrated in a game in which objects were given play names and the children had to make up a story using the objects. For example, we called a pencil “mama,” a vase “the tree,” and a spoon “the wolf.” We then played out a little game with “the tree,” “the wolf,” and “mama.” Yura, using the pencil, the vase, and the spoon to represent characters in the story, immediately caught on and played the game. But even with gestural cues from the adult, Llosha did not participate and refused to call the objects by their play names. He simply could not use the labeling function of words in a flexible way. The conflict between the words' meanings and the objects' conventional names prevented him from entering the game. Llosha also had difficulties when asked to classify objects or to point out the absurdities in a picture, while Yura quickly mastered these tasks. In each case Yura demonstrated that he had learned to use culturally assimilated schemes to organize his thinking, while Llosha's behavior resembled that of younger children. I could provide many other examples to substantiate these conclusions, but the generalizations can easily be verified simply by referring to the monograph that I wrote with F. Y. Yudovich entitled “Speech and the Development of Mental Processes in the Child.”

Psychology and Marxism | Luria Archive | Chapter 6