Science, in the more restricted sense in which it is normally employed in English-speaking countries, is that activity by which today we attain the great bulk of our knowledge of and control over the facts of nature. This activity, like other human activities, has developed and evolved, and by no means all the stages in its evolution have merited the title of scientific. In remote prehistoric times, our early ancestors worked by trial and error combined with simple, intuitive common sense. This pre-scientific approach, however, was combined with the non-scientific methods of control that we call magic, and with equally non-scientific rationalizations in the field of explanation.
Once agriculture had given the possibility of settled civilizations, with written record and specialized social classes, the hand-to-mouth methods of common sense could be replaced by something much more scientific. Science was born—witness the astronomy and geometry of ancient Mesopotamia and Egypt. But science in this phase was still, to our modern view, unscientific in two major respects—it was traditional and it was esoteric. Scientific knowledge was confined to a limited group among the priesthood and it was cast in a mold of tradition which rendered change and progress slow. Being associated with the priesthood, it was also intimately bound up with non-scientific practice and non-scientific interpretation—magic and theology.
The era of groping trial and error lasted from the first dawn of essentially human intelligence, as marked by true speech, to the beginnings of settled civilization—perhaps a million, perhaps half a million years. The next, or traditional-esoteric phase, lasted for thousands instead of hundreds of thousands of years. After some three or four millenniums, the Greeks suddenly burst free of the prison of secrecy and traditionalism, and proclaimed the freedom of intellectual enquiry. The “birth of science” is usually dated here. This is but a half-truth. At best it was the acquisition of freedom and self-consciousness by the scientific spirit, not the emergence of a wholly new activity to be called science. And in the second place, the type of science which it inaugurated differed radically from modern science in several ways. It was almost entirely divorced from industry and practical application; it was exceedingly speculative and did not lay the stress on experimental verification that we do; and, correlated with this, it had not invented the modern methodology of publication of data and methods as well as conclusions.
A few centuries later the combination of Greek intellect and ingenuity with the practical spirit of the Roman im-perium made Alexandrian science something much more like modern science in outlook and methods of working. But this was swallowed up in the anti-scientific Christian flood and the general collapse of Roman civilization.
During the Dark Ages of the West, the Arabs kept the scientific spirit alive, and by means of their mathematical inventions paved the way for immense improvements in the technique of scientific research.
Natural science, in its modern form, can fairly be said to date back no further than the seventeenth century. With Bacon as its St. John the Baptist, it developed its gospel and its ministry. Curiosity for its own sake, but also interest in industrial techniques and practical control; freedom of enquiry; experimental verification in place of authority; full publication and abundant discussion—with these a truly new phase was inaugurated.
Today it seems that we are again in the process of launching a new phase—one in which social as well as natural phenomena are to be made amenable to scientific understanding and rational control.
As with natural science, social science too has had its earlier stages. It too passed through the stage of trial and error, in which social organization shaped itself under the influence of unconscious adjustment together with non-rational rules of conduct and non-scientific interpretations of human destiny. It also had its traditional phases, often tightly bound up with philosophical and theological interpretative principles, as witness, for example, the climax of the Middle Ages. And it had its birth of free speculative enquiry, its parallel to the Greek phase of natural science—but two thousand years later, in the philosophers of the seventeenth and especially the eighteenth centuries.
Finally, its modern stage now dawning has had, like the modern stage of natural science, its scattered precursors, its Roger Bacons and Leonardos—and it has had its precursor in the restricted sense, its equivalent of Francis Bacon in the Renaissance. Many, I am sure, would put Herbert Spencer in this position; but I believe that the true John the Baptist of social science is Karl Marx. Herbert Spencer, for all his academic knowledge—or perhaps because of it—was more in the position of an Old Testament prophet. His work was essentially analogical. He demonstrated that social science was an inevitable development, but his notions of what form it would actually take and what methods it should employ were vague and essentially erroneous.
Marx, on the other hand, developed a system directly based on social facts and directly applicable to them. He did not just prophesy a messiah; he indicated the messiah. As natural scientists tend to undervalue Bacon because he himself did not make discoveries or work out experimental techniques, so social scientists tend to underrate Marx because his system is a dialectical one, ready-made and complete with answer to any problem, not sufficiently empirical and inductive for their scientific taste. It is doubtless true that the social scientists must go their own way to work, regardless of doctrine or theoretical system; a precursor cannot take the place of the messiah or the gospel that he indicates. But at least Marx, like Bacon, gave expression to a new outlook and a new method of attack, and helped to alter materially the intellectual climate so as to make it propitious for scientific work in his field.
The question immediately poses itself as to why the emergence of social science into large-scale and efficient operation has been so long delayed. The triumphs of natural science both in discovering radically new knowledge and in applying it practically to satisfy human needs have been so spectacular and so fruitful that it would seem natural and obvious to extend the same methods to the field of social phenomena.
The answer is a very simple one: the methods are not the same. The scientific spirit remains unaltered whether it is contemplating a nebula or a baby, a field of wheat or a trade union. But the methodology of social science is inevitably different from that of natural science. It is different and must be different for one basic reason—the investigator is inside instead of outside his material. Man cannot investigate man by the same methods he uses to investigate external nature. He can use the methods of natural science to investigate certain aspects of man—the structure and working of his body, for instance, or the mode of his heredity; but that is because these are shared with other organisms and because they are partial aspects which can be readily externalized. But when he starts investigating human motive, his own motives are involved; when he studies human society, he is himself part of a social structure.
What consequences does this basic difference imply? In the first place, man must be his own guinea pig. But this is impossible in the strict sense. He is unable to make fully controlled experiments. Even if an absolute despot or a fully totalitarian state were to subject a group of people to rigorous experimentation—by depriving them of alcohol, for instance, or by adopting a new form of education—the results would have only a limited application. The smallness of the group, the compulsion involved, the inevitable limitations on the contacts and full social activity of the group: these factors would make it impossible to apply the results directly to an entire normal society, however regimented; and the difficulties are of course infinitely greater in any free society.
Also, large-scale social phenomena are unique not only in detail but in essence as well. Geological phenomena are unique, but they do not show the same quality of progressive change: they are in most features repetitive. Thus even the method of comparative study is only partially applicable, while that of repeatable and controlled experimentation is not applicable at all to major social phenomena.
A second, more technical difficulty is in a sense a consequence of the first. Causation in social science is never simple and single as in physics or biology, but always multiple and complex. It is of course true that one-to-one causation is an artificial, affair, only to be unearthed by isolating phenomena from their total background. None the less, this method is the most powerful weapon in the armory of natural science: it disentangles the chaotic field of influence and reduces it to a series of single causes, each-of which can then be given due weight when the isolates are put back into their natural interrelatedness, or when they are deliberately combined (as in modern electrical science and its applications) into new complexes unknown in nature. This method of analysis is impossible in social science. Multiple causation here is irreducible.
The problem is a two-fold one. In the first place, the hu- j man mind is always looking for single causes for phenomena, I The very idea of multiple causation is not only difficult, but definitely antipathetic. And secondly, even when the social scientist has overcome this resistance, extreme practical difflculties remain. Somehow he must disentangle the single causes from the multiple field of which they form an inseparable part. And for this a new technique is necessary.
Next, and in many ways more important than the first two together, comes the question of bias. Under this head I include anything appertaining to the investigator’ which may deflect his scientific judgment. It is the equivalent of experimental and observational error in natural science. In natural science, there are statistical methods for discounting both sampling error and personal error; the limits of accurate measurement are determined for different types of instrument; the procedure of controlled experimentation has been reduced to a fine art; chemical reagents and biological material are tested for their purity and uniformity. The procedure of thus discounting errors in natural science has proved difficult enough. But to discover how to discount bias in social science is proving very much harder. The social scientist absorbs certain ideas from the culture in which he has grown up. If he sets himself to investigate that culture, he must discount his class bias, his social bias, his nationalist bias. If he is investigating an alien culture, he must use a great effort of imagination to overcome the very fact of its alien quality, which may appear illogical or inefficient to him, while the familiar methods of his own culture will tend to be accepted as more natural and feasible than they really are.
Then there is the inherent genetic bias imposed by his own temperament. For certain purposes, investigators in social science are their own instruments to a much greater extent than they are in natural science—and the instruments differ in their very construction.
Next we have the bias introduced by the peculiar psychological development of human beings. They can resolve their inevitable conflicts during childhood and adolescence only by relegating a great deal to their unconscious, whether by the psychological mechanism of suppression or by that of repression. Roughly speaking, the former introduces bias by leaving gaps in a man’s knowledge and outlook, whereas in the latter the gaps are accompanied by strong emotional distortions and resistances. The scientific study of sex, for instance, has been much retarded by repressional bias—witness the reception originally given to Havelock Ellis’s great work and the extraordinary resistance still offered to Freud’s ideas.
Bias of this type has the additional danger that those who make an effort to discount it may readily swing into overcompensation—a bias of opposite sign. The investigator whose youth was tormented by intolerant religion is apt to discount the social importance of religion far too much; the convert to Freudian methods is liable, in discounting his own early sexual repressions, to underestimate the social value of repression in general.
Bias has been encountered in natural science also, but only when its findings come up against emotionally-held convictions—only, that is, when it has had social entanglements. We may cite the prohibition of anatomical dissection, the proscription of Galileo’s findings, the hostility to the Darwinian theory, the Nazi distortion of racial anthropology, the Soviet attack on modern genetics. The present course of general anti-scientific feeling, so noticeable during the past decade, has been due in part to an emotional reaction against the social misuse of science, as in the employment of poison gas and war technique in general; partly to a general feeling that scientific findings, by sapping the traditional view of man’s place in the universe and in society, are undermining the basis of ordered society—again a resistance to change in the emotionally-charged foundations of our own existence. i
Finally there comes the most fundamental difference of all, a difference which is the parent of the difficulty caused by bias. Values are deliberately excluded from the purview of natural science: values—and all that they connote of motive, emotion, qualitative hierarchy, and the rest—constitute some of the most important data with which the social scientist must deal. But how can science deal with them? Science must aim at quantitative treatment: how can it deal with the irreducible absolutes of quality? Science must be morally neutral and dispassionate: how can the social scientists handle the ethical bases of morality, the motives of passion? Bias arises from his being called upon to do so; but the primary difficulty is a much more fundamental one.
Let us be frank with ourselves. There is a sense in which, because of this qualitative difference between its data and those of natural science, social science can never become fully and rigorously scientific. To understand and describe a system involving values is impossible without some judgment of values, and still more impossible without such value-judgment is the other scientific function, that of control.
However, this is not quite so serious as it at first sight appears. Even in natural science, regarded as pure knowledge, one value-judgment is implicit—belief in the value of truth. And where natural science passes into control, a whole scale of values is involved. The application of natural science is guided by considerations of utility. Utility is a value-judgment; but there are different kinds of utility— utility for profit, for war, for food production, for health, for amusement, for education. The application of science through the instrument of laissez faire economic systems has brought us to a position at which we are being forcibly reminded that these different utilities may conflict, that we are faced, in other words, with a difficult problem of value-judgment.
Put in another way, this is because natural science, by the fact of being applied, becomes a social problem and so a subject for social science. To set up a new value-system is in certain ways analagous to advancing a new hypothesis in natural science, and to demonstrate that such a new system is desirable or necessary is to discover and formulate some of the “laws of nature” for the coming phase of social evolution.
Thus, rather crudely, we may say that in respect of the problem of values, social science in its aspect of knowledge is faced by the same difficulties as is natural science in its aspect of control. The value problem begins earlier for social science, in the pure instead of the applied stage.
The difficulty is thus in a sense an artificial one. Its consideration has reminded us that natural science is not such a pure disembodied activity as is often assumed. Language is in part responsible for the assumption. There is no such thing as natural science per se. The phrase is a shorthand description of those activities of human beings which are concerned with understanding and controlling their natural environment. And, just as simple one-to-one causation is a fiction, only approximated in artificially isolated systems, so the emancipation of natural science from considerations of value is a fiction, approximated by the possibility of temporarily and artificially isolating scientific activity from other human activities. Nevertheless, both of these are useful fictions, dangerous only if we erect them into realities over against the true realities of the universal interdependence of phenomena and the unitary nature of human activity—the one a field of material interaction, the other a field of interrelated values.
The essential differences between natural and social science thus boil down to this—that the phenomena with which the latter deals are less readily isolated, and that as an activity it is more closely entangled with human values. These differences, however, even if only qualitative, are very real, and it remains true that social science must develop its own methodology if it is to become an efficient instrument.
In regard to multiple causation, we may look forward to the extended use of the techniques of mathematical correlation. These have already been developed to a high pitch for dealing with problems of multiple causation in physical science (for instance, the relation of crop yield to the numerous environmental factors concerned) and special methods have been worked out by Spearman and his school for dealing with psychological questions.
The use of probability methods is also indicated. Here again, these have been developed to a high pitch for use in natural science, notably in physics and biology, and it is only a question of adapting them to the problems of social science. Mathematical methods also enter into another technique which is now being rapidly developed in social science, that of the questionnaire, and especially the set of questions asked by a trained interviewer. The questionnaire method is widely used, but the reluctance or inability of large sections of the public to fill in elaborate forms restricts its sphere and impairs its sampling accuracy. The method of set questions in the hands of interviewers is being employed with great success, notably in the United States, to gauge public opinion, both with regard to commercial products and to questions of public interest. The success of the method in this form depends chiefly on two things—the proper framing of the questions and the selection of a truly representative sample of the population to answer them. It is in this latter regard that the mathematical theory of sampling errors plays an important role.
The framing of the questions to be asked is an important matter of skill and common sense rather than science. Some questions do not admit of a significant answer, or any answer at all; others will defeat their own ends by influencing the form of the answer. The investigators who interview the questionnees must also be properly trained and a constant check must be kept on their work, if the results are to be accurate. In any case, the method of questioning a representative sample of a large population can be applied only to a restricted set of problems, though within limitations it may become extremely efficient. In one field, that of the straw ballot, it is developing such uncanny accuracy that it is impinging upon practical politics. Some people are asking whether a properly conducted straw ballot could not be profitably substituted for the trouble and expense of a full election; while others feel that the announcement of a straw vote may itself influence the course of the subsequent election.
Psychologists are busy devising modifications of the questionnaire method so as to build up objective rating scales (objective, that is, for the population of which the questionnees are a representative sample) for various value-judgments. In addition, they are essaying to assess the distribution among the population of various human qualities. Intelligence-testing has long been practiced, and is now approaching full scientific validity. Attempts are also being made to assess temperament and even more elusive qualities. The method of mass observation constitutes an attempt to attain objective information on various aspects of public opinion and behavior which elude the method of yes-and-no questioning. Enquiries may concern the reaction of the public to a particular place, like the zoo or the National Gallery; to a particular event, like the coronation; to a particular activity, such as smoking or the time of rising; or to a general situation, like that of war. Numerous investigators take part, and note down either the remarks they overhear, or the answers to questions in discussion made by their acquaintance or by casual encounters.
In some cases, composite pictures which could have been obtained in no other way have resulted from the use of this method. But in general its technique, both as regards sampling and questioning, will have to be refined a good deal before it can claim to be scientifically dependable.
Another set of methods which are being developed to cope with the complexity of social problems are those of anonymous group working, repeated drafting, and circulation of I the preliminary draft results for comment and criticism. A combination of all three seems to yield the best results when tackling large and many-sided problems, such as the structure of a national agency like a health service or a big industry like steel or agriculture;, the organization of leisure, international adjustments, et cetera.
Joint work is on the increase in natural science, but here largely because of the quantitative burden of routine procedures in subjects like biochemistry or genetics. We may distinguish such work as team work. The term “group” is here used in the sense of a body of people pooling their different knowledges and skills to cope with qualitatively differentiated problems. Group work in this sense is also to be found where geneticists, ecologists, and statisticians make a united attack on some problems of microevolution, or an em-bryologist and a biochemist combine to study development. But it is far more necessary in social science. Anonymity is often desirable in group work to enable the participation of public servants or well known men whose opinions might be distorted or discounted in advance. It may also be desirable for an essentially opposite reason, to give the weight of a recognized study organization to the work of young and unknown men whose findings would otherwise tend to be disregarded. In both these ways anonymous group working, in addition to securing greater efficiency, helps to discount bias of one sort or another.
Provided that a good drafter is available, together with a chairman and a small core of members who will give regular attendance, group membership can be fluid, and specialists invited for one or a few meetings as required.
Repeated drafting is a substitute for experimentation in problems where the experimental attack is ruled out. As soon as a preliminary survey has been made of the problem in its entirety, a draft is circulated for discussion at the next meeting. The gaps and errors thus brought to light form the subject of the next period of work, when the process is repeated. Three, four, or even more complete drafts may be required before publishable conclusions are reached, just as new sets of experiments must be planned and executed to deal with tentative conclusions and new facts arising in a piece of research in natural science, before it can be written up.
Some or all of the successive drafts may also be circulated to a comparatively large number of outside experts for written criticism. The collation of such comments often brings to light new details and unexpected points of view which the group, in its preoccupation with its own trend of thought, has overlooked. It affords a method of enlarging the group without the time-consuming business of large-scale discussion.
In other cases, the actual investigator may be a single man, while the group element is provided by interviews and by circulation of drafts. This method is best adapted to problems which are of large geographical scale and local diversity, though it may also be used for those which are qualitatively diversified in themselves. It was adopted, for instance, in Lord Hailey’s African Research Survey, and in the supplementary volumes to it on research and capital investment in Africa. Where the subject matter is highly technical, as in the case of African research, repeated circulation of drafts for criticism by a large number of experts is imperative.
It may be expected that the working out of various techniques, made necessary by the nature of the data of social science, will have fruitful repercussions in certain fields of natural science, such as evolution and comparative biological study in general, where the present bias in favor of experimental work and specific results is leaving vast bodies of published data awaiting the synthetic treatment which only organized group attack can provide.
I have already mentioned certain substitutes for the controlled experimentation of the natural sciences. But experimentation as a method is not ruled out in social science, though it must take different forms. Regional or group experimentation is the most obvious method. Two regions or groups are chosen which are as similar as possible, and certain measures are introduced in the one, while the other serves as control. The Carlisle experiment on liquor control in Britain was an early essay in this method, but unfortunately it has been allowed to drag on without any serious attempts to draw theoretical conclusions or to frame practical policies on the basis of its operation. The T.V.A. in America is perhaps the largest social experiment ever undertaken in a non-totalitarian country; the area involved, however, is so large that strict controls are difficult to find. Education is a field in which regional experimentation could be very fruitful. We hear a great deal about educational experiments; but for the most part these are not experiments in the scientific sense, merely new ventures. They lack the crucial feature of scientific experimentation—blank controls.
As the spirit of scientific planning extends into government, we may expect to see regional experiments tried out in many fields. Medical and health services would afford another excellent field for this; new techniques, like that of Stapledon for converting rough grazings and moorland into pasture, should be tested on a regional scale for their social effects as well as their technical efficiency. The social results of cheap electric power could be made the subject of local experiments much more rigorous than that of the T.V.A. Different methods of developing backward tropical territories—by international or national chartered companies, by public works schemes under the local administration, by the establishment of cooperatives—could and should be made the subject of carefully planned regional experiments.
The fact that in social science man is his own guinea pig has a number of methodological consequences both for social science research and for its practical applications. In a sense, the scientist must always secure the cooperation of his material, however inert. The chemist must discover the conditions in which his pure substances will oblige him by reacting; the geneticist’s results will be doomed if his cages or culture bottles interfere with the fertility of his mice or flies; the student of growth will be misled if he cannot coax his animals to maximum performance. But the social scientist often requires true cooperation from his material in the sense of an understanding of the reason for his work and voluntary participation in its course. Education as a social experiment can never succeed without properly equipped teachers, specially trained in pedagogy. The interview method will give entirely misleading results without interviewers skilled in the technique of their job.
In the field of application, propaganda and public relations may be of prime importance. A good example is the cancer campaign recently instituted in the United States. Cancer has been presented to the public in such a way as to create a real interest in it as a social problem, and the public is collaborating in the attack upon it. The vast problem of malnutrition will never be solved unless the public is made to take a similar interest in it. The British Medical Association has made a beginning in this field with its milk campaign ; but it is a beginning only.
In general the whole technique of propaganda, persuasion, and public relations needs the most intensive study before the findings of science can be socially applied. When does propaganda defeat its own ends by setting up counter-resistance? What are the relative values of reiteration and of variety of appeal? Of the printed word, the poster, the cinema, or the radio? Of rational persuasion as against mere suggestibility? Of intellectual comprehension as against a sense of active participation? We simply do not know, and until we know, our progress towards efficient social structure and a fuller life will be fitful and slow. In many ways, the enlistment of public cooperation is to social science what the enlistment of capital investment is to natural science: it provides motive power for application.
There remains the question of bias. Here there is no ready method to hand. It took generations for natural science to work the technique of discounting experimental and observational error; it will take generations for social science to work out that of discounting the errors due to bias. The first step is obviously to make the world aware of the existence of bias and the need for its discounting. Where human affairs are still handled in a pre-scientiflc spirit, bias is apt to play a very large practical role, especially the bias in favor of one’s own group, whether class, religion, or race. Such bias produces powerful rationalizations, which are then used to justify policies of the merest self-interest. The enslavement of Negroes was justified on the basis of the Scriptural authority for the menial destiny of the sons of Ham; the brutalities of the Spanish conquest in the New World, on the basis of the unique superiority of Christianity; the Nazi persecution of Jews, on the racial superiority of “Aryans.” The group bias of the prosperous classes in early nineteenth-century England appeared in astonishing assertions about the inherent inferiority of “the poor”; the same bias is evident in certain aspects of the eugenics movement today. The general assumption that your own group standards and methods are right and sensible while other beliefs and forms of behavior are wrong, silly, or at best inferior has determined much of the relations between the great imperialist powers and the backward peoples of the earth.
Another widespread and disastrous form of bias arises from psychological conflict and tension. Censoriousness in respect of moral taboos, the desire to see the infliction of vindictive punishment, the unconscious reluctance of many parents to see the harsh school discipline under which they suffered replaced by more humane methods, the emotional basis of militarism—all these and many other undesirable determiners of human conduct are the result of bias arising from repression or emotional conflict and the inflicting of lasting distortion on the psyche.
In these fields, bias is thus an urgent subject for investigation by social science and the application here will lie in making its findings universally known and accepted by the public in general and by administrators in particular.
But even in scientific circles bias may play a surprisingly large part. A good example was the resistance of the great majority of medical men during the early part of the last war to admitting any cause for breakdown among soldiers save physical shell-shock and malingering. And the uncritical assumption, made even by scrupulously careful persons, that differences in intelligence between social classes were genetic and not due to nutrition or other social factors, is another. Again, we have the thesis of anthropologists, like Levy Bruhl, that savage mentality is in some way qualitatively different from and inferior to our own, whereas it is in fact essentially similar, but operating under different material and social conditions.
No golden rule can be laid down for the avoidance of such pitfalls, apart from the obvious step of realizing that they exist. Beyond that, special methods must be worked out in each field.
Voices are still raised proclaiming that social science is a contradiction in terms, that human affairs are not intrinsically amenable to the scientific method. They are, I believe, wrong. They are confusing the methods of natural science with scientific method in general. Social science differs inevitably from natural science in many important respects, notably in its lesser capacity for isolating problems, and more generally in its lesser degree of isolation from other aspects of human activity and its consequent greater entanglement with problems of value. It must therefore work out its own methodology, just as natural science had to do after Bacon and the eager amateurs of the seventeenth century had glimpsed natural science as a new form of human activity.
Let us not forget that the working out of this technique and this methodology by natural science took a great deal of time and is indeed still progressing. During the growth of modern science, the amateur has been largely replaced by the professional; university laboratories have been supplemerited by governmental and industrial institutions; whole-time research has become a new profession; the team has in many types of work replaced the individual; cooperative group work is beginning; and the large-scale planning of research is in the offing. The criteria of accuracy have meanwhile steadily improved, and mathematics has put new tools and indeed new outlooks at the disposal of the natural scientist.
Finally, the enormous growth of applied science has had effects of the utmost importance on pure research. It has done so partly by providing new instruments which would otherwise have been unavailable; one need only instance the gifts of the wireless industry not only to pure physics but also to such unexpected branches of science as nervous physiology. And partly by suggesting new lines of research, the needs of wireless again have revealed new facts concerning the upper atmosphere, while the study of plant pests and human diseases has brought to light new modes of evolution.
We need have no fear for the future of social science. It too will pass through similar phases from its present infancy. By the time that the profession of social science, pure and applied, includes as many men and women as are now engaged in natural science, it will have solved its major problems of new methods, and the results it has achieved will have altered the whole intellectual climate. As the barber-surgeon of the Middle Ages has given place to the medical man of today, with his elaborate scientific training, so the essentially amateur politician and administrator of today will have been replaced by a new type of professional man, with specialized scientific training. Life will go on against a background of social science. Society will have begun to develop a brain.