The Making of Mind. A R Luria
BY THE END of the war we had obtained a far better understanding of the “morphological organization” of higher psychological processes. Our progress had occurred along two fronts. First, we knew a great deal more about the structure of those inner activities, hidden from view, that we term “psychological processes.” Second, we had greatly increased our understanding of the role played by separate areas of the brain in fulfilling these activities. The time was long past to consider psychological processes as the result of either strictly localized brain activities or the “mass action” of the brain in which all of its parts were equipotential.” It was time for us to begin the next step in our work: to explain the neurophysiological or, to use the Russian phrase, “neurodynamic” mechanisms underlying the activity of brain loci implicated in specific syndromes. If this step could not be taken, we would remain on the level of pure description. Such descriptions, though valuable, are not the endpoint of scientific enterprise.
Progress depended upon advances in both of the areas that had concerned me all of my life. On the one hand, I had to move from brain structures to a deeper understanding of the neurophysiological mechanisms that were operating in these structures. On the other hand, our psychological analysis of higher cortical functions was by no means complete, and we needed improved psychological analyses as well. To signify the combination of these two enterprises, the “neurological” and the “psychological,” the term neuropsychology was coined. Developing this field of science has taken a long time and the help of many people. I have been most fortunate to receive strong support from a group of young co-workers (former students from the Psychology Faculty at Moscow University), friends, and trusted colleagues. The work that has occupied me in recent years has benefited immeasurably from the assistance of E. D. Homskaya, with whom I carried out many investigations and who has become an independent and gifted scholar, rich in psychological experience and precise in experimental technique; N. A. Filippycheva, who contributed her outstanding skills as a clinical neuropsychologist and neurophysiologist; A. I. Meshcheryakov, who helped me to begin this line of research before moving on to his classical investigations of the deaf-blind child; L. S. Tsvetkova, who participated closely in work on the recovery and rehabilitation of functions; N. A. Bernshtein, who before his death in the early 1960s provided invaluable theoretical knowledge about dynamic systems: E. N. Sokolov, one of the most experienced and gifted psychophysiologists in the world, who became closely involved in our studies of memory and orienting activity; and of course Olga Vinogradova, a brilliant experimenter and highly gifted thinker, who played a key role in a good deal of my work.
Among the younger generation E. G. Simernitskaya contributed her knowledge of interhemispheric functions, N. K. Kiyashchenko provided important insights into memory defects, and many other people provided indispensable assistance at every step of the way. To all of these people and many more whom I cannot name for lack of space I owe the deepest debt of gratitude. They made possible what was beyond the grasp of one man.
During recent years we have focused on two main problems, each of which illustrates a different path toward specification of the mechanisms underlying complex psychological functions. In dealing with these two lines of research, I shall emphasize their underlying logic. The first area of research was the neurophysiology of the frontal lobes.
During the 1940s and 1950s important progress was made in the study of brain organization. This research provided a basis for investigations into the neurophysiology of brain processes. which we had heretofore studied only at a more global level. H. Magoun. G. Morruzzi, N. Jasper, Donald Lindsley, and Wilder Penfield added enormously to our understanding by their studies of the brain stem, particularly the role of the ascending and descending reticular formation.
As Magoun and Morruzi showed in 1949, the reticular formation is a formation in the brain stem that is specially adapted, both in its structure and in its functional properties, to play a role in the regulation of the state of the cerebral cortex. Unlike the cortex, this formation does not consist of isolated neurons, capable of sending single impulses along their axons. Instead it is constructed like a nerve net with the bodies of nerve cells that are connected with each other by short axons distributed throughout it. Excitation spreads over the net of this nervous structure gradually, not in an all-or-nothing fashion. Consequently, the level of excitation in the system as a whole can change gradually, thus modulating the whole state of the nervous system.
Whereas formerly we could characterize approaches to the brain as “horizontal,” that is, as concerned with processes organized at a given level of organization, this new wave of research turned our attention to “vertical” relations between the structures deep within the brain and nearer the surface. This reorientation clarified the ways in which the brain both generates and controls the level of its own activation.
Initially it was shown that the reticular formation is crucial to activating the brain. This activation was at first thought to be quite nonspecific. That is, any stimulation, such as a loud sound, the smell of sausage, or the sight of a butterfly, was assumed to have a generalized effect on the reticular formation, which in turn activitated other parts of the brain stem without regard to specific characteristics of the initiating stimuli. Another way of putting matters is that the reticular formation was assumed to control the quantitative level of activation in the brain, but was not responsible for the qualitative differences in activitation, which depended on characteristics of the specific stimuli. But specific characteristics of the stimuli to which the person, or animal, was exposed were soon demonstrated to exert an effect on the kinds of activitation that resulted; sausages and butterflies activated other parts of the brain in different ways. It became necessary to consider both a nonspecific and a specific activating function of the reticular formation.
A second important distinction that had to be made concerning brain activation was the direction of the excitatory influence. In the earliest work, investigators were impressed by the way that excitation proceeded from lower to higher brain formations, corresponding to the way in which the environment impinges on the organism. But it soon became apparent that it was also necessary to recognize that there were fibers which reversed the direction of neural activity. Upper levels of the brain's organization activated (or modulated the activity of) lower or more peripheral levels.
In due time neuroanatomists discovered that the structure of the reticular formation included both ascending and descending fibers, some of which were discharged only by specific forms of stimulation, while others were activated in a nondifferentiated way that seemed to affect the brain as a whole.
Because there are many different indices of neural activity, we experimented a good deal to determine which indices were best suited to our needs. Like many other investigators of intellectual functions and the brain, we make extensive use of the electroencephalogram, a device that records electrical activity of the brain from electrodes placed on the surface of the skull. The electrical activity or “brain waves” of the normal, awake adult who is not being exposed to any special stimulation, such as an adult sitting in a comfortable chair in a dark room, is dominated by small, rapid brain waves known as alpha waves. When a stimulus is introduced and the person is altered, alpha activity is suppressed, giving us a key indicator of activation.
When these electrophysiological techniques began to make their appearance in the 1950s, we decided that special attention needed to be paid to the role of the frontal lobes in the organization and support of adequate levels of brain activation. We hypothesized that frontal lobe symptoms that we had previously described might be linked to disturbances in the brain's ability to coordinate levels of activation among its subsystems. It was to this problem that Homskaya turned her attention in the mid-1950s, initiating a line of research that has continued for more than twenty years. We established a special neurophysiological section in our laboratory at the Burdenko Institute of Neurosurgery, where our neuropsychological and neurophysiological research could be carried out with the close coordination that they demanded.
A cornerstone of our work was Sokolov's research on the orienting reflex, or as Pavlov.had called it many years earlier, the “what is it” reflex. The orienting reflex was particularly useful because it manifested both the specific and nonspecific characteristics that we knew to be central to the brain's activation mechanisms. The basic experimental model for our work came from a technique I had developed with Vinogradova in the early 1950s. Although in that case we had used galvanic skin response and peripheral blood flow as indicators of specific and nonspecific activation, the logic remained the same when desynchronization of alpha waves was used, as countless subsequent experiments have shown.
Subjects were first adapted to sitting quietly in a chair in a bare room where nothing special was happening. Then they were read a list of common words over a speaker system from a control room. The words were presented at intervals of a minute or so with just enough variability in time between words so that there was no point in the subject trying to anticipate when the next word would come.
The initial response to the first word was a clearly marked nonspecific activation, which manifested itself as an increase in skin resistance to electricity, a decrease in peripheral blood supply, and desynchronization of the alpha rhythm. As new words were presented, the magnitude of the activation, or orienting reflex, decreased. When the orienting reflex had almost disappeared, after presentation of ten to fifteen words, we followed presentation of one word, say “house,” by a mild electric shock to the person's hand. Needless to say, this shock produced a new orienting reflex and a high level of activation. More important, it produced the higher activation not only in response to the shock but in response to later words in the sequence as well.
When we completed the experiment without further shocks, we discovered that the gradual presentation of spoken words would again fail to evoke any appreciable level of activation. But by presenting shocks selectively to certain words, we imbued these words with special significance, which then enabled us to trace the selective activation controlled by word meaning. For example, if we included in our series of common words the word “home” to go along with the word “house,” we discovered that practiced subjects demonstrated no special activation to any of the words in the sequence except those two; both words evoked high levels of activation. We could demonstrate that word meaning was the basis of selection by including words that were like “house” in other ways. For example, we could include “louse” to see if acoustic similarity controlled activation. In normal adults it did not.
With this experimental model we could carry out further neurophysiological investigations of the way in which the frontal lobes affect activation of the brain as a whole. We could trace the nonspecific effect of all kinds of stimuli, and we could imbue certain stimuli with special significance in order to distinguish between specific and nonspecific activation. It was not always necessary to use conditioned reflex techniques, as we had in the 1950s; a great variety of methods, including simple instructions, were often just as useful.
The results of the work carried out by Homskaya and her colleagues were of the greatest significance for our understanding of frontal lobe functions. Working with either normal subjects or patients who had suffered damage to posterior portions of the brain, such as patients with a parietal lobe lesion, they reliably found the patterns of specific and nonspecific activation. Working with a technique based on preliminary instructions, such as “Listen for the word house” or “Listen to the words as I say them,” Homskaya found that the instructions which required selective activation produced both a higher level of activation overall and appropriate changes when the target stimuli occurred. With instructions of the second type that required no special selective attention, activation was less durable as well as nonselective.
A completely different picture emerged in research on frontal lobe patients. These people as a rule exhibit diminished active behavior, and lack of spontaneity is a basic symptom of their pathology. In this new research their response to stimulation as measured by activation in the EEG pattern was significantly opposite to that obtained in normal subjects or patients with posterior lesions. In frontal patients, stimuli which had no special significance evoked marked EEG changes which were more or less like those observed when normal subjects were presented such stimuli. But no change occurred in the EEG pattern of frontal lobe patients when stimuli were imbued with special significance by means of verbal instruction. If we dealt with patients who had suffered really massive bilateral lesions, we even saw inhibition of the cortical processes in response to significant words, reflected as a slowing of EEG activity, where we might have expected it to increase. These results indicated that the frontal lobes are responsible for modulating the tonus, or, the level of neural activity, of the cerebral cortex. In normal individuals, the stabilizing effect of the frontal cortex in accordance with verbal instructions is reflected directly in the EEG patterns. Once we had developed the basic technique for tracing the level of activation using EEG indicators of physiological processes, we were able to repeat and extend a number of our basic psychological observations, tracing their physiological bases. The subsequent research, reported in monographs by Homskaya and in a number of my own publications, demonstrated that in the normal adult the frontal lobes exert control over behavior in part as a result of their control over the level of activation aroused by different kinds of verbal stimuli. For example, we repeated many of the basic experiments modeled on the combined motor method. Sometimes we asked the patient to “press the key when the red light comes on.” At other times we introduced elementary choices into the task: “When you see the red light, press with your right hand; when you see the green light, press with your left hand.” In such cases the frontal lobe patient might be able to respond correctly for a few trials, but cortical tonus quickly broke down, and as it did so, errors began to occur. Sometimes the patient began to perseverate, continuing to make his responses even when the light was no longer on. Sometimes in the choice experiments he began to respond only with one hand no matter what the stimulus, which indicated that selectivity was lost.
The ability of the patient to retain verbal instructions was not lost in these cases. He could repeat the instructions. But the instructions had lost their controlling function. Purely imitative verbal responses were also maintained. Such patients could say “red” whenever a red light appeared, but they could not control their motor responding in accordance with their verbal behavior.
We think it of real significance that these results mirror the results obtained in our earlier research with children. In the case of three to three-and-a-half year old children we were dealing with young people whose brains were still developing. It is at this period that mylenization of the neurons of the frontal lobes begins to reach completion; and it is at this age that young children begin to control their behavior in accord with verbal instructions. In both cases, there is evidence that the complex organization of human conscious action depends critically on the operation of neurophysiological mechanisms in the frontal lobes. In the case of children, the brain is developing at the same time that the child is acquiring higher forms of behavior, social in origin and verbally mediated in structure. In the adult who suffers a lesion of the frontal lobes, these higher forms become inaccessible as the result of insult to the basic neural structures of the brain.
Progress toward the explanation of higher psychological functions required progress along two fronts. The work extending “downward” into neurophysiology pointed in one direction. But we must also achieve a more detailed understanding of the psychological processes that are organized as a part of the interaction between the brain and man's social environment.
Illustrative of the effort required to explore this second dimension of neuropsychology are the forty years of research on the psychology and brain organization of language. So complex is this enterprise that I have coined a special phrase, “neurolinguistics,” to indicate the problems it poses for scientific analysis, problems that are critically bound up in our understanding of human language. Although linguistic phenomena have played a role in much of my research, I have given little of the background which led me to use particular linguistic devices in my diagnostic work. Nor have I considered the psychology of, and brain organization of, language per se. However, the problem of language and the brain has been of great concern to me in recent years, and it illustrates the general approach to psychological analysis that is essential to neuropsychology.
My interest in linguistic phenomena grew naturally out of my early research using the combined motor method and Vygotsky's theory, which emphasized language as a key tool, unique to human beings, for mediating their interactions with the world. But a serious study of language as a highly organized system of human behavior began in earnest only after I had begun work on the problem of the neuropsychology of sensory and semantic aphasia.
Sensory aphasia is a condition in which patients can speak but are unable to understand spoken language. I knew that their defect could not be attributed to a loss of hearing or any general decrement in intelligence. I also found that patients suffering from semantic aphasia could understand isolated words but were at a loss when presented with relational terms, such as “brother's father” or “the circle under the triangle.”
While these observations were interesting and of potential diagnostic value, I could not be sure of what the diagnoses meant. As a clinician, I was busy studying O. Potzl's Die Apbasielebre vom Standpunkt der Kliniscben Psycbiatrie and Head's Aphasia and Kindred Disorders of Speech, along with the work of Gelb, Goldstein, and others. The work of these neurologists suggested that I had to understand the way information about spatial or quasi-spatial information is stored in the brain and how people come to construct a synopsis of a scene with many details. To learn more about these phenomena, I knew that I also had to master what the linguistics of that period, the late 1930s, could teach me.
One of the first important influences in my early linguistically oriented work was N. S. Trubetskoy's “Grunzuge der Phonologie,” which was published in Prague in 1939. It was immediately recognized as a revolutionary account of the mechanisms of the acoustic organization of language. Trubetskoy argued that language processing depends critically not only on the physical features of sound, such as pitch, but also on the value of sounds as a means of distinguishing word meaning. He emphasized this phonemic aspect of speech instead of its phonological aspect because the organization of sounds into different phonemes is what distinguishes different languages, not the physical or phonological aspects of sound alone. If I was to understand aphasia, I knew that I would have to analyze the breakdown of the system of phonemes, not just the physical deterlorization of acoustic analysis. When I first began this line of work, the lesson I learned from Trubetskoy was not at all obvious; after all, I was observing patients who failed to distinguish between b and p, or d and t, who suffered what was then called “alienation of word meaning” (For example, a patient might repeat quizzically: “Posture ... what does it mean? ... bosture ... bolstul ... ?”).
I knew that patients who had suffered temporal-parieto-occipital lesions suffered severe comprehension difficulties, but I did not know what it was about certain grammatical functions that implicated these areas of the brain. Here I was helped by V. V. Vinogradov, who drew my attention to the publication of a young Swedish linguist, C. Svedellus. In his analyse de language, published in Uppsala in 1897, Svedelius divided all forms of verbal communication into two basic classes: communication of events, the contents of which can be expressed in images, and communication of relations, such as “Socrates is a man” or “Kathy is prettier than Mary,” in which special linguistic devices such as prepositions or variations in word order are necessary to express the ideas being communicated. I needed to undertake a special study of these constructions, an enterprise to which I devoted two years.
I began by studying model constructions of the type “brother's father” or “father's brother,” constructions that included the attributive form of the genitive case. My patients could not understand these constructions, which in Russian take the form of otets brata and brat otsa respectively, but they could understand other forms of the genitive case, such as the genitive of parts, as in a “piece of bread,” kusok kbleba.
I began to understand that in the attributive genitive constructions there was a conflict between the two words that required the person to make a mental transformation in order to overcome the conflict and to understand the phrase. One had to abstract the immediate, concrete meaning of the word “brother's” and convert the semantic content of a noun into the semantic content of an adjective; mentally, the sequence of words had to be reversed. This was true because in Russian adjectives precede nouns, as in vkusnyi kbleb, “good bread.” But in a construction like brat otsa, “father's brother,” the genitive form of the noun “father” performs the function of an adjective yet follows the noun it modifies. This kind of transformation can be made only if the relational meaning of the whole expression is grasped. But that transformation in the service of relational meaning was exactly what the patients in question found difficult to make.
When I looked into the area of historical linguistics, I found that relational constructions of this kind appeared late in the development of the Russian language. They did not appear at all in old Slavic chronicles. Instead, one finds simple appositions: not dett boyar, “children of the Boyars,” but the simpler expression boyare deti, “the boyar children.” The same lack of attributive genitive constructions appeared to be true of old German and old English texts as well. In German, for example, instead of mit Leidschaft der Liebe, there appeared the apposition mit Leidschaft und Liebe. This evidence suggested that the attributive genitive, which was especially disturbed in cases of semantic aphasia, was a construction of relatively recent historical origin and required special mental work. The attributive genitive is necessary for the communication of relations of a special kind that are not involved in either the genitive of parts or the communication of events.
The two years I spent in the study of linguistics early in my career stood me in good stead when I began to work seriously on the problem of semantic aphasia because I could understand more fully the different mental requirements that seemingly similar linguistic acts placed on people. I was thus in a beter position to carry out differential diagnosis of pathological symptoms which previously had been lumped together in the neurological literature. As my work continued to involve me in attempts to understand the brain basis of language-related behavior, I found it necessary to continue to study the psychology of language at the same time that I searched for its neurological bases. And just as advances in neurology and neurophysiology were instrumental to our study of brain mechanisms, advances in the study of linguistics were crucial to advancing our understanding of those phenomena of speech which brain pathology was interrupting; the two enterprises are inextribably bound together. Time and again I found myself returning to old data, armed with new insights from advances in linguistics.
One of the distinctions which began to appear in linguistics with the work of de Saussure in the 1920s and which I came to use heavily in the 1940s owing to the work of Roman jakobson was the difference between the “paradigmatic” aspects of language, which refers to the placing of words and the things they denote into categories, enabling people to use language to make comparisons and generalizations, and the “syntagmatic” aspect of language, which enables people to join words together into coherent expressions. The paradigmatic function of speech allows the codes of language to be used to separate out significant cues in the environment, and it also makes it possible to take whole categories of cues into consideration at a single moment, which is what we do when we use categories. The paradigmatic function of speech is intimately related to the basic motives that direct activity. It links our intentions to our thoughts.
In trying to understand the cortical organization of language, we must recognize the existence of both the categorizing and the intention-fulfilling functions which intermingle in every utterance. Moreover, recognizing that they carry out different, If related, functions, we can expect their cortical localization to be different. Looking back into the history of neurology, we can see that as early as 1913 Pick was pointing to the syntagmatic function when he sought to determine how condensed thought patterns can be expanded into smooth, sequentially organized statements, and Jackson's reference to the “propositionizing” aspect of speech showed that he too recognized the importance of this function.
Working with this distinction, which was anticipated and used by Vygotsky in Thought and Language and in his preliminary articles on brain localization, we found the anticipated difference in the brain localization of syntagmatic and paradigmatic language functions. Lesions in the forward parts of the left hemisphere, which are known to be closely related to motor functions, selectively impair fluent, syntagmatically organized speech, but complex verbal codes based on paradigmatic organization remain more or less intact. Patients with such lesions easily name single objects, but their speech takes on the classic “telegraphic” style that many early investigators noted, owing to a breakdown of the predicative function which is basic to fluent speech.
From the point of view of this linguistic distinction, just the opposite pattern of disturbance occurs in patients with lesions in the rear of the head. These people can speak fluently, but the relations between individual words break down. This is the linguistic basis for observations in which grammatical relations such as “father's brother” are destroyed by lesions in the parietooccipital area.
I could extend these examples, showing how a combination of linguistic, psychological, and neuropsychological analysis is required in order to understand the mechanisms underlying each particular form of speech pathology. It suffices, however, to point out that exactly the same principles for understanding individual words and simple phrases apply as well to longer, more complex phrases and to the comprehension of entire paragraphs and texts or narratives. A full discussion of these issues may be found in my Basic Problems of Neurolinguistics. Instead of multiplying examples, I shall give a single example which makes clear the ways in which the various disciplines that contribute to the understanding of an activity as complex as language must be combined to understand its brain organization. The example relates to the phenomenon known as “elicited imitation” in the literature on children's language, but in aphasiology it is referred to as “conduction aphasia.”
In 1875 Wernicke described a special form of aphasia in which the patient retained his full understanding of speech addressed to him and to some extent could produce coherent speech spontaneously, but he was unable to repeat sounds, words, or sentences. This phenomenon was considered paradoxical because the patient was heard to make very complex remarks at the same time that he failed to repeat back the simplest phrases provided by the examiner. Wernicke hypothesized that this disturbance was caused by a break in the direct connections between the sensory and motor “speech centers,” although each of these centers remained intact and each retained connections to the hypothetical “higher centers.” In the years to follow, several additional cases of this kind were reported, and the concept of a special kind of conduction aphasia was passed down through generations of textbooks.
As is often the case, in the face of such a simple schema, contradictory data were overlooked. In their encounters with so-called conduction aphasia, workers noted that in some patients the difficulty in repeating words seemed to arise from a difficulty in naming objects; in other cases, individual objects could be named but complex material could not be repeated. Difficulties in narrative speech, which hypothetically ought to be under the control of only the higher centers, were also observed.
In my opinion the beginning of an understanding of this phenomenon came from neither a strictly linguistic nor a strictly neurological source, but rather from a psychological analysis of the activity demanded of someone repeating what someone else says. As early as the 1870s Jackson suggested that naming objects and repeating individual words are not the most elementary or natural forms of speech activity. In a series of investigations, Goldstein directed special attention to the importance of making a psychological analysis of speech repetition. His argument is that neither the naming of individual objects nor the repetition of words is the basis of most natural speech activity. Instead, the basic form of speech communication is the formulation of ideas as whole propositions which are intimately bound up with the motives and conditions of the activity in which the individual is engaged. When the neurologist begins to ask the patient to repeat phrases that are arbitrary and totally unrelated to anything he is supposed to do, the patient is really being asked to engage in an abstraction of speech from action at the same time that he is being asked to speak.
The major finding of Goldstein's analysis received support from quite a different source. Long ago Piaget, as well as Vygotsky and his students discovered that after young children had learned to speak, they still experienced difficulty in carrying out presumably simple imitation tasks in which all they were required to do was to repeat some action or phrase after an adult. Quite recently, Daniel Slobin and his colleagues in America made a study of a child's spontaneous utterances around the home. From time to time they would ask the child to repeat something that he had said just a few minutes earlier. In addition to confirming observations made half a century earlier but appearing now in the context of the burgeoning field of developmental psycholinguistics, Slobin pointed out that the child's spontaneous speech was organized by the motives which guided his activity as a whole; deprived of the organizing motive, the child's speech lost its guiding principle. The child's failure represents the phenomenon that would have been called conduction aphasia if Slobin had been working with an adult who had a brain injury instead of a healthy two-and-a-half year old.
Because the general approach initiated by Vygotsky formed the basis for our research, we adopted the fundamental proposition that a change in the goal of a task inevitably leads to a significant change in the structure of the psychological processes which carry it out. A change in the structure of activity, in other words, implies a change in the brain organization of activity. Therefore the transition from spontaneous to elicited speech, whether in dialogue or in monologue, not only changes the task and the structure of the speech process but also changes the functional systems of the brain that support these activities. To believe that conduction aphasia constitutes a loss of abstract attitude – the erroneous direction toward which Goldstien's psychological theorizing carried him – or that it reflects a mere severing of connections between two brain centers is to misunderstand the structure of the task, the nature of the activity, and the significance of the brain injury.
Once the true complexity of language phenomena are understood, there is no further need to ignore the seemingly anomalous cases of conduction aphasia, since they are no more than indications that conduction aphasia is not a single syndrome but a set of disturbances which express themselves differently according to the demands placed on the patient and the particular areas of the brain that are affected. just as we found that sensory, motor, and semantic aphasias were terms covering a multiplicity of related phenomena, so our later analysis has shown that several subcategories of disturbance are covered by the loose term “conduction aphasia.”
These examples illustrate a process which has no endpoint. If one wants to understand the brain foundations for psychological activity, one must be prepared to study both the brain and the system of activity in as great detail as contemporary science allows. In many cases important clues can be gotten from specialists in related fields. This was true in our studies of neurolinguistics. It was true as well in our studies of the disturbance of memory and problem solving. But in each of these cases, we found that we must use the work of specialists as a starting point, modifiying tasks and theories as we want, because the conditions of clinical work do not permit the well-controlled application of many experimental methods. And in dealing with patients, we must never forget that an individual human life is at stake, not a statistical abstraction which, on the average, supports a theory.