The Competitiveness of Nations in a Global Knowledge-Based Economy
F. A. von Hayek
Economics and Knowledge 1
Scientism and the Study of
Society : PART
: PART II
Economica, New Series,
10 (37 )
Feb. 1943, 34-63.
PART I PART II PART III
[HHC: titles added to numbered items]
V – Objectivism &
The great differences between the characteristic methods of the physical sciences and those of the social sciences make it not difficult to understand why the natural scientist who turns to the works of the professional students of social phenomena should so often feel that he has got among a company of people who habitually commit all the mortal sins which he is most careful to avoid, and that a science of society conforming to his standards does not yet exist. From this to the attempt to create a new science of society which satisfies his conception of Science is but a step. During the last four generations attempts of this kind have been constantly made; and though they have never produced the results which had been expected, and though they did not even succeed in creating that continuous tradition which is the symptom of a healthy discipline, they are repeated almost every month by someone who hopes thereby to revolutionize social thought. Yet, though these efforts are mostly disconnected, they regularly show certain characteristic features which we must now consider. These methodological features can be conveniently treated under the headings of “objectivism”, “collectivism”, and “historism”, corresponding to the “subjectivism”, the “individualism”, and the theoretical character of the developed disciplines of social study.
The attitude which, for want of a better term, we shall call the “objectivism” of the scientistic approach to the study of man and society, has found its most characteristic expression in the various attempts to dispense with our subjective knowledge of the working of the human mind, attempts which in various forms have affected almost all branches of social study. From Auguste Comte’s denial of the possibility of introspection, through various attempts to create an “objective psychology”, down to the behaviourism of J. B. Watson and the “physicalism” of O. Neurath, a long series of authors have attempted to do without the knowledge derived from “introspection”.. But, as can be easily shown, these attempts to avoid the use of knowledge which we possess are bound to break down.
1. The first part appeared in Economica for August, 1942.
A behaviourist or physicalist, to be consistent, ought not to begin by observing the reactions of people to what our senses tell us are similar objects; he ought to confine himself to studying the reactions to stimuli which are identical in a strictly physical sense. He ought, e.g., not to study the reactions of persons who are shown a red circle or made to hear a certain tune, but solely the effects of a light wave of a certain length on a particular point of the retina of the human eye, etc., etc. No behaviourist, however, seriously contemplates doing so. They all take it naïvely for granted that what appears alike to us will also appear alike to other people. Though they have no business to do so, they make constant use of the classification of external stimuli by our senses and our mind as alike or unlike, a classification which we know only from our personal experience of it and which is not based on any objective tests showing that these facts also behave similarly to each other. This applies as much to what we commonly regard as simple sense qualities, such as colour, the pitch of sound, smell, etc., as to our perception of configurations (Gestalten) by which we classify physically very different things as specimens of a particular “shape”, e.g., as a circle or a certain tune. To the behaviourist or physicalist the fact that we recognise these things as similar is no problem.
This naïve attitude, however, is in no way justified by what the development of physical science itself teaches us. As we have seen before,  one of the main results of this development is that things that to us appear alike may not be alike in any objective sense, i.e., may have no other properties in common. Once we have to recognise, however, that things differ in their effects on our senses not necessarily in the same way in which they differ in their behaviour towards each other, we are no longer entitled to take it for granted that what to us appears alike or different will also appear so to others. That this is so as a rule is an important empirical fact which, on the one hand, demands explanation (a task for psychology) and which, on the other hand, must be accepted as a basic datum in our study of people’s conduct. That different objects mean the same thing to different people, or that different people mean the same thing by different acts, remain important facts though physical science may show that these objects or acts possess no other common properties.
It is true, of course, that we know nothing about other people’s minds except through sense perceptions, i.e., the observation of physical facts. But this does not mean that we know nothing but physical facts. Of what kind the facts are with which we have to deal in any discipline is not determined by all the properties possessed by the concrete objects to which the discipline applies, but only by those properties by which we classify them for the purposes of the discipline in question. To take an example from the physical sciences: all levers or pendulums of which we can conceive have chemical and
1. Cf. the first part of this article, pp. 271-274.
optical properties; but when we talk about levers or pendulums we do not talk about chemical or optical facts. What makes a number of individual phenomena facts of one kind are the attributes which we select in order to treat them as members of one class. And though all social phenomena with which we can possibly be concerned will possess physical attributes, this does not mean that they must be physical facts for our purpose.
The significant point about the objects of human activity with which we are concerned in the social sciences, and about these human activities themselves, is that in interpreting human activities we spontaneously and unconsciously class together as instances of the same object or the same act any one of a large number of physical facts which may have no physical property in common. We know that other people like ourselves regard any one of a large number of physically different things, a, b, c, d,… etc., as belonging to the same class; and we know this because other people, like ourselves, react to any one of these things by any one of the movements α, β, δ, … which again may have no physical property in common. Yet this knowledge on which we constantly act, which must necessarily precede, and is pre-supposed by, any communication with other men, is not conscious knowledge in the sense that we are in a position exhaustively to enumerate all the different physical phenomena which we unhesitatingly recognise as a member of the class: we do not know which of many possible combinations of physical properties we shall recognise as a certain word, or as a “friendly face” or a “threatening gesture”. Probably in no single instance has experimental research yet succeeded in precisely determining the range of different phenomena which we unhesitatingly treat as meaning the same thing to us as well as to other people; yet we constantly and successfully act on the assumption that we do classify these things in the same manner as other people do. We are not in a position - and may never be in the position - to substitute objects defined in physical terms for the mental categories we employ in talking about other people’s actions.  Whenever we do so the physical facts to which we refer are significant not as physical facts, i.e., not as members of a class all of which have certain physical properties in common, but as members of a class of what may be physically completely different things which “mean” the same thing to us.
It becomes necessary here to state explicitly a consideration which is implied in the whole of our argument on this point and which, though it seems to follow from the modern conception of the character of physical research, is yet still somewhat unfamiliar. It is that not only those mental entities, such as “concepts” or “ideas”, which are commonly
1. The attempts often made to evade this difficulty by an il1ustrative enumeration of some of the physical attributes by which we recognise the object as belonging to one of these mental categories are just begging the question. To say that when we speak about a man being angry we mean that he shows certain physical symptoms helps us very little unless we can exhaustively enumerate all the symptoms by which we ever recognise, and which always when they are present mean, that the man who shows them is angry. Only if we could do this would it be legitimate to say that in using this term we mean no more than certain physical phenomena.
recognised as “abstractions”, but all mental phenomena, sense perceptions and images as well as the more abstract “concepts” and “ideas” must be regarded as acts of classification performed by the brain.  This is, of course, merely another way of saying that the qualities which we perceive are not properties of the objects but ways in which we (individually or as a race) have learnt to group or classify external stimuli. To perceive is to assign to a familiar category (or categories): we could not perceive anything completely different from everything else we have ever perceived before.  This does not mean, however, that everything which we actually class together must possess common properties other than that we react in the same way to these things. It is a common but dangerous error to believe that things which our senses or our mind treat as members of the same class must have something else in common beyond being registered in the same manner by our mind. Although there will usually exist some objective justification of why we regard certain things as similar in kind, this need not always be the case. But while in our study of nature classifications which are not based on any similarity in the behaviour of the objects towards each other must be treated as “deceptions” of which we must free ourselves, they are of positive significance in our attempts to understand human action. The important difference between the position of these mental categories in the two spheres is that when we study the working of external nature our sensations and thoughts are not links in the chain of observed events - they are merely about them; but in the mechanism of society they form an essential link, the forces here at work operate through these mental entities which are directly known to us: while the things in the external world do not behave alike or differently because they appear alike to us, we do behave in a similar or different manner because the things appear alike or different to us.
1. This must also serve as a justification for what may have seemed the very loose way in which we have throughout, in illustrative enumerations of mental entities, indiscriminately lumped together such concepts as “sensation”, “perceptions”, “concepts”, or “ideas”. These different types of mental entities all have in common that they are classifications of possible external stimuli (or complexes of such stimuli). This contention will perhaps appear less strange now than would have been the case fifty years ago, since in the configurations or Gestalt qualities we have become familiar with something which is intermediate between the old “elementary” sense qualities and concepts. It may be added that on this view there would, however, seem to be no justification for the unwarranted ontological conclusions which many members of the Gestalt school like to draw from their interesting observations; there is no reason to assume that the “wholes” which we perceive are properties of the external world and not merely ways in which our mind classifies complexes of stimuli; like other abstractions, the relations between the parts thus singled out may be significant or not.
Perhaps it should also be mentioned here that there is no reason to regard values as the only purely mental categories which do therefore not appear in our picture of the physical world. Although values must necessarily occupy a central place wherever we are concerned with purposive action, they are certainly not the only kind of purely mental categories which we shall have to employ in interpreting human activities the distinction between true and false provides at least one other instance of such purely mental categories which is of great importance in this connection. On the connected point that it is not necessarily value considerations which will guide us in selecting the aspects of social life which we study, see below p. 55, footnote 2.
2. Although the second time we are exposed to a new stimulus we may already “recognise” it as identical with what happened to us in circumstances which its recurrence calls to our mind, we should still not have been “conscious” of it on the first occasion when it had not yet acquired a place in the structure of our mind.
The behaviourist or physicalist who wished in studying human behaviour really to avoid using the categories which we find ready in our mind, and who wanted to confine himself strictly to the study of man’s reactions to objects defined in physical terms, would consistently have to refuse to say anything about human actions till he had experimentally established how our senses and our mind group external stimuli as alike or unlike. He would have to begin by asking which physical objects appear alike to us and which do not (and how it comes about that they do) before he could seriously undertake to study human behaviour towards these things.
It is important to observe that our contention is not that such an attempt to explain how our mind or our brain transforms physical facts into mental entities is in principle impossible. Once we recognise this as a process of classification there is no reason why we should not learn to understand the principle on which it operates. Classification is, after all, a mechanical process, i.e., a process which could be performed by a machine which “sorts out” and groups objects according to certain properties.  Our argument is, rather, in the first instance, that for the task of the social sciences such an explanation of the formation of mental entities and their relations to the physical facts which they represent is unnecessary, and that such an explanation would help us in no way in our task; and, secondly, that such an explanation, although conceivable, is not only not available at present and not likely to be available for a long time yet, but also unlikely to be ever more than an “explanation of the principle”  on which this mechanism of classification works. It would seem that any mechanism of classification would always have to possess a degree of complexity greater than any one of the different things which it classifies ; and if this is correct it would follow that it is impossible that our brain should ever be able to produce a complete explanation (as distinguished from a mere explanation of the principle) of the particular ways in which it itself classifies external stimuli. We shall later have to consider the significance of the related paradox that to explain all our knowledge would require that we should know more than we actually do, which is, of course, a self-contradictory statement.
But let us assume for the moment that we had actually succeeded in fully reducing all mental phenomena to physical processes. Assume that we knew the mechanism by which our central nervous system groups any one of the (elementary or complex) stimuli a, b, c, … or m, n,… or r, s, t,… into definite classes determined by the fact that to any member of one class we shall react by any one of the members of the corresponding classes or reactions α, β, γ,… or γ, ξ, о, or φ, χ, ψ,… This assumption implies both that this system is not merely familiar to us as the way in which our own mind acts, but that we
1. Which, as we have already seen, does, of course, not mean that it will always treat only elements which have common properties as members of the same class.
2. Cf. the concluding paragraphs of the first part of this article.
explicitly know all the relations by which it is determined, and that we also know the mechanism by which the classification is actually effected. We should then be able strictly to correlate the mental entities with definite groups of physical facts. We should thus have “unified” science, but we should be in no better position with respect to the specific task of the social sciences than we are now. We should still have to use the old categories, though we should be able to explain their formation and though we should know the physical facts “behind” them. Although we should know that a different arrangement of the facts of nature is more appropriate for explaining external events, in interpreting human actions we should still have to use the classification in which these facts actually appear in the minds of the acting people. Thus, quite apart from the fact that we should probably have to wait forever till we were able to substitute physical facts for the mental entities, even if this were achieved we should be no better equipped for the task we have to solve in the social sciences.
The idea, implied in Comte’s hierarchy of the sciences 1 and in many similar arguments, that the social sciences must in some sense be “based” on the physical sciences, that they can only hope for success after the physical sciences have advanced far enough to enable us to treat social phenomena in physical terms, in “physical language”, is, therefore, entirely erroneous. The problem of explaining mental processes by physical ones is entirely distinct from the problems of the social sciences, it is a problem for physiological psychology. But whether it is solved or not, for the social sciences the given mental entities must provide the starting point, whether their formation has been explained or not.
VI – Objectivism & Quantification
We cannot here discuss all the other forms in which the characteristic objectivism” of the scientistic approach has made itself felt and led to error in the social sciences. We shall, in the course of our historical survey, find this tendency to look for the “real” attributes of the objects of human activity which lie behind men’s views about them, represented in a great many different ways. Only a brief survey of them can here be attempted.
Nearly as important as the various forms of behaviourism, and closely connected with them, is the common tendency to attempt in the study of social phenomena to disregard all the “merely” qualitative phenomena and to concentrate, on the model of the natural sciences, on the quantitative aspects, on what is measurable. We have seen before  how in the natural sciences this tendency is a necessary consequence of their specific task of replacing the picture of the world
1. Cf. the comment on this by Carl Meager in the passage quoted before, Economica, August, 1942, p. 287, footnote.
2. Cf. the first part of this article, Economica, August, 1942, pp. 272, 275.
in terms of sense qualities by one
in which the units are defined exclusively by their explicit relations. Its success here has brought it about
that it has come to be regarded as the hall-mark of all genuinely scientific
procedure. Yet its raison
d’être, the need to replace the classification of events which our senses
and our mind provide by a more appropriate one, is absent where we try to
understand human beings, and where this understanding is made possible by the
fact that we have a mind like theirs, and that from the mental categories we
have in common with them we can reconstruct the social complexes which are our
concern. The blind transfer of the
striving for quantitative measurements  to a field
where the specific conditions are not present which give it its basic importance
in the natural sciences is the result of an entirely unfounded prejudice. It is probably responsible for the worst
aberrations and absurdities produced by scientism in the social sciences. It not only leads frequently to the
selection for study of the most irrelevant aspects of the phenomena because they
happen to be measurable, but also to “measurements” and
assignments of numerical values which are absolutely meaningless. What a distinguished philosopher recently
wrote about psychology is at least equally true of the social sciences, namely
that it is only too easy “to rush off to measure something without considering
what it is we are measuring, or what measurement means. In this respect some recent measurements
are of the same logical type as Plato’s determination that a just ruler is
729 times as happy as an unjust one.” [2 ]
Closely connected with the tendency to treat the objects of human activity in terms of their “real” attributes instead of as what they appear to the acting people is the propensity to conceive of the student of society as endowed with a kind of super-mind, some sort of absolute knowledge, which makes it unnecessary for him to start from what is known by the people whose actions he studies. Among the most characteristic manifestations of this tendency are the various forms of social “energetics” which, from the earlier attempts of Ernest Solvay, Wilhelm Ostwald and F. Soddy down to our own day  have constantly reappeared among scientists and engineers approaching the problems of social organisation. The idea underlying these theories is that, as science is supposed to teach that everything can be ultimately reduced to quantities of energy, man in his plans should treat the various things not according to the concrete usefulness they possess for the
1. It should, perhaps, be emphasised that there is no necessary connection between the use of mathematics in the social sciences and the attempts to measure social phenomena - as particularly people who are acquainted only with elementary mathematics arc apt to believe. Mathematics may - and in economics probably is - absolutely indispensable to describe certain types of complex structural relationships, though there may be no chance of ever knowing the numerical values of the concrete magnitudes (misleadingly called “constants”) which appear in the formula describing these structures.
2. M. R. Cohen, Reason and Nature, p. 305.
3. Cf. L. Hogben (in Lancelot Hogben’s Dangerous Thoughts, 1939, p. 99): “Plenty is the excess of free energy over the collective calory debt of human effort applied to securing the needs which all human beings share.”
purposes for which he knows how to use them, but as the interchangeable units of abstract energy which they “really” are.
Another, hardly less crude, and even more widespread, example of this tendency is the conception of the “objective” possibilities of production, of the quantity of social output which the physical facts are supposed to make possible, an idea which frequently finds expression in quantitative estimates of the supposed “productive capacity” of society as a whole. These estimates regularly refer, not to what men can produce by means of any stated organisation, but to what in some undefined “objective” sense “could” be produced from the available resources. Most of these assertions have no ascertainable meaning whatever. They do not mean that x or y or any particular organisation of people could achieve these things. What they amount to is that if all the knowledge dispersed among many people could be mastered by a single mind, and if this master-mind could make all the people act at all times as he wished, certain results could be achieved but these results could, of course, not be known to any except such a master-mind. It need hardly be pointed out that an assertion about a “possibility” which is dependent on such conditions has no relation to reality. There is no such thing as a productive capacity of society in the abstract - apart from particular forms of organisation. The only fact which we can regard as given is that there are particular people who have certain concrete knowledge about the way in which particular things can be used for particular purposes. This knowledge never exists as an integrated whole or in one mind, and the only knowledge that can in any sense be said to exist are these separate and often inconsistent and even conflicting views of different people.
Of a very similar nature are the frequent statements about the “objective” needs of the people, where “objective” is merely a name for somebody’s views about what the people ought to want. We shall have to consider further manifestations of this “objectivism” more fully towards the end of this article when we turn from the consideration of scientism proper to the effects of the characteristic outlook of the engineer, whose conceptions of “efficiency” have been one of the most powerful vehicles through which this attitude has affected current views on social problems.
VII – Collectivism & “Wholes”
Closely connected with the “objectivism” of the scientistic approach is its methodological collectivism, its tendency to treat “wholes” like “society” or the “economy”, “capitalism” (as a given historical “phase”) or a particular “industry” or “class” or “country” as definitely given objects about which we can discover laws by observing their behaviour as wholes. While the specific subjectivist approach of the social sciences starts, as we have seen, from our knowledge of the inside of these social complexes, the knowledge of the individual attitudes which form the elements of their structure, the objectivism
of the natural sciences tries to view them from the outside  ; it treats social phenomena not as something of which the human mind is a part and the principles of whose organisation we can reconstruct from the familiar parts, but as if they were objects directly perceived by us as wholes.
There are several reasons why this tendency should so frequently show itself with natural scientists. They are used to seek first for empirical regularities in the relatively complex phenomena that are immediately given to observation, and only after they have found such regularities to try and explain them as the product of a combination of other, often purely hypothetical, elements (constructs) which are assumed to behave according to simpler and more general rules. They are therefore inclined in the social field, too, to seek first for empirical regularities in the behaviour of the complexes before they feel that there is need for a theoretical explanation. This tendency is further strengthened by the experience that there are few regularities in the behaviour of individuals which can be established in a strictly objective manner; and they turn therefore to the wholes in the hope that they will show such regularities. Finally, there is the rather vague idea that as “social phenomena” are to be the object of study, the obvious procedure is to start from the direct observation of these “social phenomena”, where the existence in popular usage of such terms as “society” or “economy” is naïvely taken as evidence that there must be definite “objects” corresponding to them. The fact that people all talk about “the nation” or “capitalism” leads to the belief that the first step in the study of these phenomena must be to go and see what they are like, just as we should if we heard about a particular stone or a particular animal. 
The error involved in this collectivist approach is that it mistakes for facts what are no more than provisional theories, models constructed by the popular mind to explain the connection between some of the individual phenomena which we observe. The paradoxical aspect of it, however, is, as we have seen before,  that those who by the scientistic prejudice are led to approach social phenomena in this manner are, by their very anxiety to avoid all merely subjective elements and to confine themselves to “objective facts”, induced to commit the mistake they are most anxious to avoid, namely to treat as facts what
1. The description of this contrast as one between the view from the inside and the view from the outside, though, of course, metaphorical, is less misleading than such metaphors usually are and perhaps the best short way to indicate the nature of the contrast it brings out that what of social complexes is directly known to us are only the parts and that the whole is never directly perceived but always reconstructed by an effort of our imagination.
2. It would, of course, be false to believe that the first instinct of the student of social phenomena is any less to “go and see”. It is not ignorance of the obvious but long experience which has taught him that directly to look for the wholes which popular language suggests to exist leads nowhere. It has, indeed, rightly become one of the first maxims which the student of social phenomena learns (or ought to learn) never to speak of “society” or a “country” acting or behaving in a certain manner, but always and exclusively to think of individuals as acting.
3. Cf. Part I of this article, p. 286,
are no more than vague popular theories. They thus become, when they least suspect it, the victims of the fallacy described by an expressive German term as Begriffsrealismus and made familiar to us by Professor Whitehead as the “fallacy of misplaced concreteness”.
The naïve realism which uncritically assumes that where there are commonly used concepts there must also be definite “given” things which they describe is so deeply embedded in current thought about social phenomena that to free ourselves from it needs a deliberate effort of will. While most people will readily admit that in this field there may exist special difficulties in recognising definite wholes because we have never before us many specimens of a kind and therefore cannot readily distinguish their constant from their merely accidental attributes, few are aware that there is a much more fundamental obstacle that the wholes as such are never given to our observation but are without exception constructions of our mind. They are not “given facts”, objective data of a similar kind which we spontaneously recognise as similar by their common physical attributes. They cannot be perceived at all apart from a mental scheme that shows the connection between some of the many individual facts which we can observe. Where we have to deal with such social wholes we cannot, as we do in the natural sciences, start from the observation of a number of instances which we recognise spontaneously by their common sense attributes as instances of “societies” or “economies”, “capitalisms” or “nations”, “languages” or “legal systems”, and where only after we have collected a sufficient number of instances we begin to seek for common laws which they may obey. Social wholes are not given to us as what we may call “natural units” which we recognise as similar with our senses, as we do with flowers or butterflies, minerals or light-rays, or even forests or ant-heaps. They are not given to us as similar things before we even begin to ask whether what looks alike to us also behaves in like manner. The terms for collectives which we all readily use do not designate definite things in the sense of stable collections of sense attributes which we recognise as alike by inspection; they refer rather to certain structures of relationships between some of the many things we can observe within given spatial and temporal limits and which we select because we think we can discern connections between them - connections which may or may not exist in fact.
What we group together as instances of the same collective or whole are different complexes of individual events, in themselves perhaps quite dissimilar, but believed by us to be related to each other in a similar manner; they are classifications or selections of certain elements of a complex picture on the basis of a theory about their coherence. They do not stand for definite things or classes of things (if we understand the term “thing” in any material or concrete sense) but for a pattern or order in which different things may be related to each other - an order which is not a spatial or temporal order but can be defined only in terms of relations which are intelligible human
attitudes. This order or pattern is as little perceptible as a physical fact as these relations themselves; and it can be studied only by following up the implications of the particular combination of relationships. In other words, the wholes about which we speak exist only if, and to the extent to which, the theory is correct which we have formed about the connection of the parts which they imply and which we can explicitly state only in the form of a model built from those relationships. 
The social sciences, thus, do not deal with “given” wholes but their task is to constitute these wholes by constructing models from the familiar elements - models which reproduce the structure of relationships between some of the many phenomena which we always simultaneously observe in real life. This is no less true of the popular concepts of social wholes which are represented by the terms current in ordinary language; they too refer to mental models, but instead of a precise description they convey merely vague and indistinct suggestions of the way in which certain phenomena are connected. Sometimes the wholes constituted by the theoretical social sciences will roughly correspond with the wholes to which the popular concepts refer, because popular usage has succeeded in approximately separating the significant from the accidental; sometimes the wholes constituted by theory may refer to entirely new structural connections of which we did not know before systematic study commenced and for which ordinary language has not even a name. If we take current concepts like those of a “market” or of “capital”, the popular meaning of these words corresponds at least in some measure to the similar concepts which we have to form for theoretical purposes, although even in these instances the popular meaning is far to vague to allow the use of these terms without first giving them a more precise meaning. If they can be retained in theoretical work at all it is, however, because in these instances even the popular concepts have long ceased to describe particular concrete things, definable in physical terms, and have come to cover a great variety of different things which are classed together solely because of a recognised similarity in the structure of the relationships between men and things. A “market”, e.g., has long ceased to mean only the periodical meeting of men at a fixed place to which they bring their products to sell them from temporary wooden stalls. It now covers any arrangements for regular contacts between potential buyers and sellers of any thing that can be sold, whether by personal contact, by telephone or telegraph, by advertising, etc., etc 
When, however, we speak of the behaviour of, e.g., the “price system” as a whole and discuss the complex of connected changes which will correspond in certain conditions to a fall in the rate of interest, we are not concerned with a whole that obtrudes itself on popular notice or that is ever definitely given; we can only recon-
1. Cf. F. Kaufmann, “Soziale Kollektiva,” Zeitschrift für Nationalökonomie, Vol. I, 1930.
2. It should be noted that though observation may assist us to understand what people mean by the terms they use, it can never tell us what a “market” or “capital”, etc., really are, i.e., which are the significant relations which it is useful to single out and combine into a model.
struct it by following up the reactions of many individuals to the initial change and its immediate effects. That in this case certain changes “belong together”, that among the large number of other changes which in any concrete situation will always occur simultaneously with them and will often swamp those which form part of the complex in which we are interested, a few form a more closely interrelated complex, we do not know from observing that these particular changes regularly occur together. That would indeed be impossible because what in different circumstances would have to be regarded as the same set of changes could not be determined by any of the physical attributes of the things but only by singling out certain relevant aspects in the attitudes of men towards the things; and this can be done only by the help of the models we have formed.
The mistake of treating as definite objects “wholes” that are no more than constructions, and that can have no properties except those which follow from the way in which we have constructed them from the elements, has probably appeared most frequently in the form of the various theories about a “social” or “collective” mind 1 and has in this connection raised all sorts of pseudo-problems. The same idea is frequently but imperfectly concealed under the attributes “personality” or “individuality” which are ascribed to society. Whatever the name, these terms always mean that, instead of reconstructing the wholes from the relations between individual minds which we directly know, a vaguely apprehended whole is treated as something akin to the individual mind. It is in this form that in the social sciences an illegitimate use of anthropomorphic concepts has had as noxious an effect as the use of such concepts in the natural sciences. The remarkable thing here is, again, that it should so frequently be the empiricism of the positivists, the arch-enemies of any anthropomorphic concepts even where they are in place, which leads them to postulate such metaphysical entities and to treat humanity, as for instance Comte does, as one “social being,” or as a kind of super-person. But as there is no other possibility than either to compose the whole from the individual minds or to postulate a super-mind in the image of the individual mind, and as positivists reject the first of these alternatives, they are necessarily driven to the second. We have here the root of that curious alliance between 19th century positivism and Hegelianism which will occupy us in a later study.
The collectivist approach to social phenomena has not often been so emphatically proclaimed as when the founder of sociology, Auguste Comte, asserted with respect to them that, as in biology, “the whole
1. On this whole problem, see M. Ginsberg, The Psychology of Society, 1921, chapter IV - What is said in the text does of course not preclude the possibility that our study of the way in which individual minds interact may reveal to us a structure which operates in some respects similarly to the individual mind. And it might be possible that the term collective mind would prove the best term available to describe such structures - though it is most unlikely that the advantages of the use of this term would ever outweigh its disadvantages. But even if this were the case the employment of this term should not mislead us into thinking that it describes any observable object that can be directly studied.
of the object is here certainly much better known and more immediately accessible”  than the constituent parts. This view has exercised a lasting influence on that scientistic study of society which he attempted to create. Yet the particular similarity between the objects of biology and those of sociology, which fitted so well in Comte’s hierarchy of the sciences, does not in fact exist. In biology we do indeed first recognise as things of one kind natural units, stable combinations of sense properties, of which we find many instances which we spontaneously recognise as alike. We can, therefore, begin by asking why these definite sets of attributes regularly occur together. But where we have to deal with social wholes or structures it is not the observation of the regular coexistence of certain physical facts which teaches us that they belong together or form a whole. We do not first observe that the parts always occur together and afterwards ask what holds them together; but it is only because we know the ties that hold them together that we can select a few elements from the immensely complicated world around us as parts of a connected whole.
We shall presently see that Comte and many others regard social phenomena as given wholes in yet another, different, sense, contending that concrete social phenomena can be understood only by considering the totality of everything that can be found within certain spatio-temporal boundaries, and that any attempt to select parts or aspects as systematically connected is bound to fail. In this form the argument amounts to a denial of the possibility of a theory of social phenomena as developed, e.g., by economics, and leads directly to what has been misnamed the “historical method” with which, indeed, methodological collectivism is closely connected. We shall have to discuss this view below under the heading of “historism”..
VIII – Collectivism & “The Macroscopic View”
The endeavour to grasp social phenomena as “wholes” finds its most characteristic expression in the tendency to gain a distant and comprehensive view in the hope that thus regularities will reveal themselves which remain obscure at closer range. Whether it is the conception of an observer from a distant planet, which has always been a favourite with positivists from Condorcet to Mach,  or whether it is the survey of long stretches of time through which it is hoped that constant configurations or regularities will reveal themselves, it is always the same endeavour to get away from our inside knowledge of human affairs and to gain a view as, it is supposed, would be com-
1. Cours de philosophic positive, Vol. IV (2nd - 4th Ed.), p. 258.
2. Cf. Ernst Mach, Erkenntnis und Irrtum, 3rd ed., 1917, p. 28, where, however, he points out correctly that [HHC: passage in German nor reproduced]
manded by somebody who was not himself a man but stood to men in the same relation as that in which we stand to the external world.
This distant and comprehensive view of human events at which the scientistic approach aims is now often described as the “macroscopic view”. It would probably be better called the telescopic view (meaning simply the distant view - unless it be the view through the inverted telescope!) since its aim is deliberately to ignore what we can see only from the inside. In the “macrocosm” which this approach attempts to see, and in the “macrodynamic” theories which it endeavours to produce, the elements would not be individual human beings but collectives, constant configurations which, it is presumed, could be defined and described in strictly objective terms.
In most instances this belief that the total view will enable us to distinguish wholes by objective criteria proves, however, to be just an illusion. This becomes evident as soon as we seriously try to imagine of what the macrocosm would consist if we were really to dispense with our knowledge of what things mean to the acting men, and if we merely observed the actions of men as we observe an ant-heap or a bee-hive. In the picture such a study could produce there could not appear such things as means or tools, commodities or money, crimes or punishments, or words or sentences; it could contain only physical objects defined in terms of the sense attributes they present to the observer. And since the human behaviour towards the physical objects would show practically no regularities discernable to the observer, since men would in a great many instances not appear to react alike to things which would seem the same to the observer, nor differently to what appeared as different to him, he could not hope to achieve an explanation of their actions unless he had first succeeded in reconstructing in all detail the way in which men’s senses and men’s minds pictured the external world to them. The famous observer from Mars, in other words, before he could understand even as much of human affairs as the ordinary man does, would have to reconstruct from our behaviour those immediate data of our mind which to us form the starting-point of any interpretation of human action.
If we are not more aware of the difficulties which would be encountered by an observer not possessed of a human mind this is so because we never seriously imagine the possibility that any being with which we are familiar might command sense perceptions or knowledge which are denied to us. Rightly or wrongly we tend to assume that other minds we encounter can differ from ours only by being inferior, so that everything which they perceive or know can also be perceived or be known to us. The only way in which we can form an approximate idea of what would be our position if we had to deal with an organism as complicated as ours but organised on a different principle, so that we should not be able to reproduce its working on the analogy of our own mind, is to conceive that we had to study the behaviour of people with a knowledge vastly superior to our own. If,
e.g., we had developed our modern scientific technique while still confined to part of our planet, and then made contact with other parts inhabited by a race which had advanced knowledge much further, we clearly could not hope to understand many of their actions by merely observing what they did and without directly learning from them their knowledge. It would not be from observing them in action that we should acquire their knowledge, but it would be through being taught their knowledge that we should learn to understand their actions.
There is yet another argument which we must briefly consider which supports the tendency to look at social phenomena “from the outside” and which is easily confused with the methodological collectivism of which we have spoken though it is really distinct from it. Are not social phenomena, it may be asked, from their definition mass phenomena, and is it not obvious, therefore, that we can hope to discover regularities in them only if we investigate them by the method developed for the study of mass phenomena, i.e., statistics? Now this is certainly true of the study of certain phenomena, such as those which form the object of vital statistics and which, as has been mentioned before, are sometimes also described as social phenomena, although they are essentially distinct from those with which we are here concerned. Nothing is more instructive than to compare the nature of these statistical wholes, to which the same word “collective” is sometimes also applied, with that of the wholes or collectives with which we have to deal in the theoretical social sciences. The statistical study is concerned with the attributes of individuals, though not with attributes of particular individuals, but with attributes of which we know only that a certain quantitatively determined proportion of all the individuals in our “collective” or “population” possess them. In order that any collection of individuals should form a true statistical collective it is even necessary that the attributes of the individuals whose frequency distribution we study should not be systematically connected or, at least, that in our selection of the individuals which form the “collective” we are not guided by any knowledge of such a connection. The “collectives” of statistics on which we study the regularities produced by the “law of large numbers” are thus emphatically not wholes in the sense in which we describe social structures as wholes. This is best seen from the fact that the properties of the “collectives” which statistics studies must remain unaffected if from the total of elements we select at random a certain part. Far from dealing with structures of relationships, statistics deliberately and systematically disregard the relationships between the individual elements. It is, to repeat, concerned with the properties of the elements of the “collective”, though not with the properties of particular elements, but with the frequency with which elements with certain properties occur among the total. And, what is more, it assumes that these properties are not systematically connected with the different ways in which the elements are related to each other.
The consequence of this is that in the statistical study of social phenomena the structures with which the theoretical social sciences are concerned actually disappear. Statistics may supply us with very interesting and important information about what is the raw material from which we have to reproduce these structures, but it can tell us nothing about these structures themselves. In some field this is immediately obvious as soon as it is stated. That the statistics of words can tell us nothing about the structure of a language will hardly be denied. But although the contrary is sometimes suggested, the same holds no less true of other systematically connected wholes such as, e.g., the price system. No statistical information about the elements can explain to us the properties of the connected wholes. Statistics could produce knowledge of the properties of the wholes only if it had information about statistical collectives, the elements of which were wholes, i.e., if we had statistical information about the properties of many languages, many price systems, etc. But, quite apart from the practical limitations imposed on us by the limited number of instances known to us, there is an even more serious obstacle to the statistical study of these wholes: the fact which we have already discussed, that these wholes and their properties are not given to our observation but can only be formed or constituted by us from their parts.
What we have said applies, however, by no means to all that goes by the name of statistics in the social sciences. Much that is thus described is not statistics in the strict modern sense of the term, does not deal with mass phenomena at all, but is called statistics only in the older, wider sense of the word in which it is used for any descriptive information about the State or society. Though the term will to-day be used only where this description is of a quantitative nature, this should not lead us to confuse it with the science of statistics in the narrower sense. Most of the economic statistics which we ordinarily meet, such as trade statistics, figures about price changes, and most “time series”, or statistics of the “national income”, are not data to which the technique appropriate to the investigation of mass phenomena can be applied. They are just “measurements” and frequently measurements of the type already discussed at the end of Section VI above. If they refer to significant phenomena they may be very interesting as information about the conditions existing at a particular moment. But unlike statistics proper, which may indeed help us to discover important regularities in the social world (though regularities of an entirely different order from those with which the theoretical sciences of society deal), there is no reason to expect that these measurements will ever reveal anything to us which is of significance beyond the particular place and time at which they have been taken. That they cannot produce generalisations does, of course, not mean that they may not be useful, even very useful; they will often provide us with the data to which our theoretical generalisation must be applied to be ofany practical use. They are an instance of the historical information about a particular situation the significance of which we must further consider in the next sections.
IX – Historism & Conscious Design
To see the “historism” to which we must now turn described as a product of the scientistic approach may cause surprise since it is usually represented as the opposite to the treatment of social phenomena on the model of the natural sciences. But the view for which this term is properly used (and which must not be confused with the true method of historical study) proves on closer consideration to be a result of the same prejudices as the other typical scientistic misconceptions of social phenomena. If the suggestion that historism is a form rather than the opposite of scientism has still somewhat the appearance of a paradox this is so because the term is used in two different and in some respect opposite and yet frequently confused senses: for the older view which justly contrasted the specific task of the historian to that of the scientist and which denied the possibility of a theoretical science of history, and for the later view which, on the contrary, affirms that history is the only road which can lead to a theoretical science of social phenomena. However great is the contrast between these two views sometimes called “historism” if we take them in their extreme forms, they have yet enough in common to have made possible a gradual and almost unperceived transition from the historical method of the historian to the scientistic historism which attempts to make history a “science” and the only science of social phenomena.
The older historical school, whose growth has recently been so well described by the German historian Meinecke, though under the misleading name of historism,  arose mainly in opposition to certain generalising and “pragmatic” tendencies of some, particularly French, 18th century views. Its emphasis was on the singular or unique (individuell) character of all historical phenomena which could be understood only genetically as the joint result of many forces working through long stretches of time. Its strong opposition to the “pragmatic” interpretation, which regards social institutions as the product of conscious design, implies in fact the use of a “compositive” theory which explains how such institutions can arise as the unintended result of the separate actions of many individuals. It is significant that
1. G. Meinecke, Die Entstebung des Historismus, 1936. ‘The term historism applied to the older historical school discussed by Meinecke is inappropriate and misleading since it was introduced by Carl Menger (Die Irrtbümer des Historismus, 1884) to describe the distinguishing features of the younger historical school in economics of Schmoller and his associates. Nothing shows more clearly the difference between this younger historical school and the earlier movement from which it inherited the name than that it was Schmoller who accused Menger of being an adherent of the “Burke-Savigny school” and not the other way round. (cf. G. Schmoller, Zur Methodologie der Staats- und Sozialwissenschaften,” Jahrbuch für Geseizgebung, etc., N.F., Vol. VII, 1886, p. 250).
among the fathers of this view Edmund Burke is one of the most important and Adam Smith occupies an honourable place.
Yet, although this historical method implies theory, i.e., an understanding of the principles of structural coherence of the social wholes, the historians who employed it not only did not systematically develop such theories and were hardly aware that they used them; but their just dislike of any generalisation about historical developments also tended to give their teaching an anti-theoretical bias which, although originally aimed only against the wrong kind of theory, yet created the impression that the main difference between the methods appropriate to the study of natural and to that of social phenomena was the same as that between theory and history. This opposition to theory of the largest body of students of social phenomena made it appear as if the difference between the theoretical and the historical treatment was a necessary consequence of the differences between the objects of the natural and the social sciences; and the belief that the search for general rules must be confined to the study of natural phenomena while in the study of the social world the historical method must rule, became the foundation on which later historism grew up. But while historism retained the claim for the pre-eminence of historical research in this field, it almost reversed the attitude of the older historical school to history and under the influence of the scientistic currents of the age came to represent history as the empirical study of society from which ultimately generalisation would emerge. History was to be the source from which a new science of society would spring, a science which should at the same time be historical and yet produce what theoretical knowledge we could hope to gain about society.
We are here not concerned with the actual steps in that process of transition from the older historical school to the historism of the younger. It may just be noticed that historism in the sense in which the term is here used, was created not by historians but by students of the specialised social sciences, particularly economists, who hoped thereby to gain an empirical road to the theory of their subject. But to trace this development in detail and to show how the men responsible for it were actually guided by the scientistic views of their generation must be left to the later historical account. 
The first point we must briefly consider is the nature of the distinction between the historical and the theoretical treatment of any subject which makes it in fact a contradiction in terms to demand that history should become a theoretical science or that theory should ever be “historical”. If we understand that distinction, it will become clear that it has no necessary connection with the difference of the concrete objects with which the two methods of approach deal and that for the
1. Although in its German origins the connection of historism with positivism is less conspicuous than is the case with its English followers such as Ingram or Ashley, it was no less present and is overlooked only because historism is erroneously connected with the historical method of the older historians, instead of with the views of Roscher, Hildebrandt and particularly Schmoller and his circle.
understanding of any concrete phenomenon, be it in nature or in society, both kinds of knowledge are equally required.
That human history deals with events or situations which are unique or singular when we consider all aspects which are relevant for the answer of a particular question we may ask about them, is, of course, not peculiar to human history. It is equally true of any attempt to explain a concrete phenomenon if we only take into account a sufficient number of aspects - or, to put it differently, so long as we do not deliberately select only such aspects of reality as fall within the sphere of that system of connected propositions which we call one theoretical science. If I watch and record the process by which a plot in my garden that I leave untouched for years is gradually covered with weeds, I am describing a process which in all its detail is no less unique than any event in human history. If I want to explain any particular configuration of different plants which may appear at any stage of that process, I can do so only by giving an account of all the relevant influences which have affected different parts of my plot at different times. I shall have to consider what I can find out about the differences of the soil in different parts of the plot, about differences in the radiation of the Sun, of moisture, of the air-currents, etc., etc.; and in order to explain the effects of all these factors I shall have to use, apart from the knowledge of all these particular facts, various parts of the theory of physics, of chemistry, biology, meteorology, and so on. The result of all this will be the explanation of a particular phenomenon, but not a science of how garden plots go out of cultivation.
In an instance like this the particular sequence of events, their causes and consequences, will probably not be of sufficient general interest to make it worth while to write them up or to develop their study into a distinct discipline. But there are large fields of natural knowledge, represented by recognised disciplines, which in their methodological character are no different from this. In geography, e.g., and at least in a large part of geology and astronomy, we are mainly concerned with particular situations, either of the earth or of the universe ; we aim at explaining a unique situation by showing how it has been produced by the operation of many forces subject to the general laws studied by the theoretical sciences. In the specific sense of a body of general rules in which the term “science” is often used  these disciplines are not “sciences”, i.e., they are not theoretical sciences but endeavours to apply the laws found by the theoretical sciences to the explanation of particular “historical” situations.
The distinction between the search for generic principles and the explanation of concrete phenomena has thus no necessary connection with the distinction between the study of nature and the study of society. In both fields we need generalisations in order to explain concrete and
1. It will be noted that this, still restricted, use of the term “science” (in the sense in which the Germans speak of Gesetzeswissensehaft) is wider than the even narrower sense in which its meaning is confined to the theoretical sciences of nature.
unique events. Whenever we attempt to explain or understand a particular phenomenon we can do so only by recognising it or its parts as members of certain classes of phenomena, and the explanation of the particular phenomenon presupposes the existence of general rules.
There are very good reasons, however, for a marked difference in emphasis, reasons why, generally speaking, in the natural sciences the search for general laws has the pride of place while their application to particular events is usually little discussed and of small general interest, while with social phenomena the explanation of the particular and unique situation is as important and often of much greater interest than any generalisations. In most natural sciences the particular ituation or event is generally one of a very large number of similar events, which as particular events are only of local and temporary interest and scarcely worth public discussion (except as evidence of the truth of the general rule). The important thing for them is the general law applicable to all the recurrent events of a particular kind. In the social field, on the other hand, a particular or unique event is often of such general interest and at the same time so complex and so difficult to see in all its important aspects, that its explanation and discussion constitutes a major task requiring the whole energy of a specialist. We study here particular events because they have contributed to create the particular environment in which we live or because they are part of that environment. The creation and dissolution of the Roman Empire or the Crusades, the French Revolution or the Growth of Modern Industry are such unique complexes of events, which have helped to contribute the particular circumstances in which we live and whose explanation is therefore of great interest.
It is necessary, however, to consider briefly the logical nature of these singular or unique objects of study. Probably the majority of the many disputes and confusions which have arisen in this connection are due to the vagueness of the common notion of what can constitute one object of thought - and particularly to the misconception that the totality (i.e., all possible aspects) of a particular situation can ever constitute one single object of thought. We can here touch only on a very few of the logical problems which this belief raises.
The first point which we must remember is that, strictly speaking, all thought must be to some degree abstract. We have seen before that all perception of reality, including the simplest sensations, involves a classification of the object according to some property or properties. The same complex of phenomena which we may be able to discover within given temporal and spatial limits may in this sense be considered under many different aspects ; and the principles according to which we classify or group the events may differ from each other not merely in one but in several different ways. The various theoretical sciences deal only with those aspects of the phenomena which can be fitted into a single body of connected propositions. It is necessary to emphasise that this is no less true of the theoretical sciences of nature
than of the theoretical sciences of society, since an alleged tendency of the natural sciences to deal with the “whole” or the totality of the real things is often quoted by writers inclined to historism as a justification for doing the same in the social field.  Any discipline of knowledge, whether theoretical or historical, can, however, deal only with certain selected aspects of the real world and in the theoretical sciences the principle of selection is the possibility of subsuming these aspects under a logically connected body of rules. The same thing may be for one science a pendulum, for another a lump of brass, and for a third a convex mirror. We have already seen that the fact that a pendulum possesses chemical and optical properties does not mean that in studying laws of pendulums we must study them by the methods of chemistry and optics - though when we apply these laws to a particular pendulum we may well have to take into account certain laws of chemistry or optics. Similarly, as has been pointed out, the fact that all social phenomena have physical properties does not mean that we must study them by the methods of the physical sciences. 
The selection of the aspects of a complex of phenomena which can be explained by means of a connected body of rules is, however, not the only method of selection or abstraction which the scientist will have to use. Where investigation is directed, not at establishing rules of general applicability, but at answering a particular question raised by the events in the world about him, he will have to select those features that are relevant to the particular question. The important point, however, is that he still must select a limited number from the infinite variety of phenomena which he can find at the given time and place. We may, in such cases, sometimes speak as if he considered the “whole” situation as he finds it. But what we mean is not the inexhaustible totality of everything that can be observed within certain spatio-temporal limits, but certain features thought to be relevant to the question asked. If I ask why the weeds in my garden have grown in this particular pattern no single theoretical science will provide the answer. This, however, does not mean that to answer we must know everything that can be known about the space-time interval in which the phenomenon occurred. While the question we ask designates the phenomena to be explained, it is only by means of
1 Cf. e.g., E. F. M. Durbin, “Methods of Research - A Plea for Co-operation in the Social Sciences,” Economic Journal, June, 1938, p. 191, where the writer argues that in the social sciences “unlike the natural sciences, our subdivisions are largely (though not entirely) abstractions from reality rather than sections of reality” and asserts of the natural sciences that “in all these cases the object of study are real independent objects and groups. They are not aspects of something complex. They are real things”. How this can be really asserted, e.g., of Crystallography (one of Mr. Durbin’s examples) is difficult to comprehend. — This argument has been extremely popular with the members of the German historical school in economics, though, it should be added, Mr. Durbin is probably entirely unaware how closely his whole attitude resembles that of the Kathedersozialisten of that school.
2. It is significant that all of H. Rickert’s well-known work on the differences between the Naturwissenschaften and Kulturwissessschaften is based on the assertion that since all phenomena which we can observe are physical phenomena, all generalising (theoretical) science must be physical science.
the laws of the theoretical sciences that we are able to select the other phenomena which are relevant for its explanation. The object of scientific study is never the totality of all the phenomena observable at a given time and place, but always only certain selected aspects and according to the question we ask the same spatio-temporai situation may become any number of different objects of study. The human mind indeed can never grasp a “whole” in the sense of all the different aspects of a real situation.
The application of these considerations to the phenomena of human history leads to very important consequences. It means nothing less than that a historical process or period is never a single definite object of thought but becomes this only by the question we ask about it; and that according to the question we ask, what we are accustomed to regard as a single historical event can become any number of different objects of thought.
It is confusion on this point which is mainly responsible for the doctrine now so very much in vogue that all historical knowledge is necessarily relative, determined by our “standpoint” and bound to change with the lapse of time.  This view is a natural consequence of the belief that the commonly used names for historical periods or complexes of events, such as “the Napoleonic Wars”, or “France during the Revolution”, or “the Commonwealth Period”, stand for definitely given objects, unique individuals  which are given to us in the same manner as is the case with the natural units in which biological specimens or planets are given to us. Those names of historical phenomena define in fact little more than a period and a place and there is scarcely a limit to the number of different questions which we can ask about events which occurred during the period and within the region to which they refer. It is, however, only the question that we ask which will define our object; and there are, of course, many reasons why at different times people will ask different questions about the same period.  But this does not mean that history will at different times and on the basis of the same information give different answers to the same question. This alone, however, would entitle us to assert that historical knowledge is relative. The kernel of truth in the assertion
1. For a good survey of the modern
theories of historical relativism see M. Mandelbaum, The Problem of
2. Cf. below, p. 57, footnote.
3. It is not possible to pursue
further here the interesting question of the reasons which make the historian
ask particular questions and which make him ask at different times different
questions about the same period. We
ought, however, perhaps briefly to refer to one view which has exercised wide
influence, since it claims application not only to history but to all
Kulturwisseszschaften. It is
Rickert’s contention that the social sciences, to which, according to him, the
historical method is alone appropriate, select their object exclusively with
reference to certain values with respect to which they are important. Unless by “value consideration”
(Wertbezogenheit) any kind of practical interest in a problem is meant so
that this concept would include the reasons which make us, say, study the
about the relativity of historical knowledge is that historians will at different times be interested in different objects, but not that they will necessarily hold different views about the same object.
We must dwell a little longer on the nature of the “wholes” which the historian studies, though much of what we have to say is merely an application of what has been said before about the “wholes” which some authors regard as objects of theoretical generalisations. What we said then is just as true of the wholes which the historian studies. They are never given to him as wholes, but always reconstructed from their elements which alone can be directly perceived. Whether he speaks about the government that existed or the trade that was carried on, the army that moved, or the knowledge that was preserved or disseminated, he is never referring to a constant collection of physical attributes that can be directly observed, but always to a system of relationships between some of the observed elements which can merely be inferred. Words like government or trade or army or knowledge do not stand for single observable things but for structures of relationships which can be described only in terms of a schematic representation or “theory” of the persistent system of relationships between the ever-changing elements.  These “wholes”, in other words, do not exist for us apart from the theory by which we constitute them, apart from the mental technique by which we can reconstruct the connections between the observed elements and follow up the implications of this particular combination.
The place of theory in historical knowledge is thus in forming or constituting the wholes to which history refers, it is prior to these wholes which are not visible except by following up the system of relations which connects the parts. The generalisations of theory, however, do not refer, and cannot refer, as has been mistakenly believed by the older historians (who for that reason opposed theory), to the concrete wholes, the particular constellations of the elements, with which history is concerned. The models of “wholes”, of structural connections, which theory provides ready-made for the historian to use (though even these are not the given elements about which theory generalises but the results of theoretical activity), are not identical with the “wholes” which the historian considers. The models provided by any one theoretical science of society consist necessarily of elements of one kind, elements which are selected because their connection can be explained by a coherent body of principles and not because they help to answer a particular question about concrete phenomena. For the latter purpose the historian will regularly have to use generalisations belonging to different theoretical spheres. His work, thus, as in all attempts to explain particular phenomena, presupposes theory; it is, as is all thinking about concrete
1. It does not alter the essential fact that the theorising will usually already have been done for the historian by his source which in reporting the “facts” will use such terms as “state” or “town” which cannot be defined by physical characteristics but which refer to a complex of relationships which, made explicit, is a “theory” of the subject.
phenomena, an application of generic concepts to the explanation of particular phenomena.
If the dependence of the historical study of social phenomena on theory is not always recognised this is mainly due to the very simple nature of the majority of theoretical schemes which the historian will employ and which brings it about that there will be no dispute about the conclusions reached by their help and little awareness that he has used theoretical reasoning at all. But this does not alter the fact that in their methodological character and validity the concepts of social phenomena which the historian has to employ are essentially of the same kind as the more elaborate models produced by the systematic social sciences. All the unique objects of history which he studies are in fact either constant patterns or relations, or repetitive processes in which the elements are of a generic character. When the historian speaks of a State or a battle, a town or a market, these words cover coherent structures of individual phenomena which we can comprehend only by understanding the intentions of the acting individuals. If the historian speaks of a certain system, say the feudal system, persisting over a period of time, he means that a certain pattern of relationships continued, a certain type of actions were regularly repeated, structures whose connection he can understand only by mental reproduction of the individual attitudes of which they were made up. The unique wholes which the historian studies, in short, are not given to him as individuals,  natural units of which he can find out by observation which features belong to them, but constructions made by the kind of technique that is systematically developed by the “theoretical” sciences of society. Whether he endeavours to give a genetic account of how a particular institution arose, or a descriptive account of how it functioned, he cannot do so except by a combination of generic considerations applying to the elements from which the unique situation is composed. Though in his work of reconstruction he cannot use any elements except those he empirically finds, not observation but only the “theoretical” work of reconstruction can tell him which among those that he can find are part of a connected whole.
Theoretical and historical work are thus logically distinct but complementary activities. If their task is rightly understood, there can be no conflict between them. And though they have distinct tasks,
1. The confusion which reigns in this field has evidently been assisted by a purely verbal confusion apt to arise in German in which most of the discussions of this problem have been conducted. In German the singular or unique is called the Individuelle, which almost inevitably calls forth a misleading association with the term for the individual (Individuum). Now, individual is the term which we employ to describe those natural units which in the physical world our senses enable us to single out from the environment as connected wholes. individuals in this sense, whether human individuals or animals or plants, or stones, mountains or stars, are constant collections of sense attributes which, either because the whole complex can move together in space relatively to its environment, or for cognate reasons, our senses spontaneously single out as connected wholes. But this is precisely what the objects of history are not. Though singular (individuell), as the individual is, they are not definite individuals in the sense in which this term is applied to natural objects. They are not given to us as wholes but only found to be wholes.
neither is of much use without the other. But this does not alter the fact that neither can theory be historical nor history theoretical. Though the general is of interest only because it explains the particular and though the particular can be explained only in generic terms, the particular can never be the general and the general never the particular. The unfortunate misunderstandings that have arisen between historians and theorists are largely due to the name “historical school” which has been usurped by the mongrel view - better described as historism and which is indeed neither history nor theory.
X – Historism & “Theories of History”
The naïve view which regards the complexes which history studies as given wholes naturally leads to the belief that their observation can reveal “laws” of the development of these wholes. This belief is one of the most characteristic features of that scientistic history which under the name of historism was trying to find an empirical basis for a theory of history or (using the term philosophy in its old sense equivalent to “theory”) a “philosophy of history” and to establish necessary successions of definite “stages” or “phases”, “systems” or “styles” which followed each other in historical development. This view on the one hand endeavours to find laws where in the nature of the case they cannot be found, in the succession of the unique and singular historical phenomena, and on the other hand denies the possibility of the kind of theory which alone can help us to understand unique wholes, the theory which shows the different ways in which the familiar elements can be combined to produce the unique combinations we find in the real world. The empiricist prejudice thus led to an inversion of the only procedure by which we can comprehend historical wholes, their reconstruction from the parts; it induced scholars to treat as if they were objective facts vague conceptions of wholes which were merely intuitively comprehended; and it finally produced the view that the elements which are the only thing that we can directly comprehend and from which we must reconstruct the wholes could, on the contrary, be understood only from the whole, which had to be known before we could understand the elements.
The belief that human history, which is the result of the interaction of innumerable human minds, must yet be subject to simple laws accessible to human minds is now so widely held that few people are at all aware what an astonishing claim it really implies. Instead of working patiently at the humble task of rebuilding from the directly known elements the complex and unique structures which we find in the world, and at tracing from the changes in the relations between the elements the changes in the wholes, the authors of these pseudo-theories of history pretend to be able to arrive by a kind of mental short cut at a direct insight into the laws of succession of the immediately apprehended wholes. However doubtful their status, these
theories of development have achieved a hold on public imagination much greater than any of the results of genuine systematic study. “Philosophies” or “theories” of history (or “historical theories”) have indeed become the characteristic feature, the “darling vice” of the 19th century.  From Hegel and Comte, and particularly Marx, down to Sombart and Spengler these spurious theories came to be regarded as representative results of social science; and through the belief that one kind of “system” must as a matter of historical necessity be superseded by a new and different “system” they have even exercised a profound influence on social evolution. This they achieved mainly because they looked like the kind of laws which the natural sciences produced; and in an age when these sciences set the standard by which all intellectual effort was measured, the claim of these theories of history to be able to predict future developments was regarded as evidence of their pre-eminently scientific character. Though merely one among many characteristic 19th century products of this kind, Marxism more than any of the others has become the vehicle through which this result of scientism has gained so wide an influence that many of its opponents equally with its adherents are thinking in its terms.
Apart from setting up a new ideal this development had, however, also the negative effect of discrediting the existing theory on which past understanding of social phenomena had been based. Since it was supposed that we could directly observe the changes in the whole of society or of any particular changed social phenomenon, and that everything within the whole must necessarily change with it, it was concluded that there could be no timeless generalisations about the elements from which these wholes were built up, no universal theories about the ways in which they might be combined into wholes. All social theory, it was said, was necessarily historical, zeitgebunden, true only of particular historical “phases” or “systems”.
All concepts of individual phenomena, according to this strict historism, are to be regarded as merely historical categories, valid only in a particular historical context. A price in the 12th century or a monopoly in the Egypt of 400 B.C., it is argued, is not the same “thing” as a price or a monopoly to-day, and any attempt to explain that price or the policy of that monopolist by the same theory which we would use to explain a price or a monopoly of to-day is therefore vain and bound to fail. This argument is based on a complete misapprehension of the function of theory. Of course, if we ask why a particular price was charged at a particular date, or why a monopolist then acted in a particular manner, this is a historical question which cannot be fully answered by any one theoretical discipline; to answer it we must take into account the particular circumstances of time and place. But this does not mean that we must not, in selecting the factors relevant to the explanation of the particular price, etc., use
L. Brunschvicg, in Philosophy and
History, Essays presented to E. Cassirer, ed. by R. Kubansky and H. J.
precisely the same theoretical reasoning as we would with regard to a price of to-day.
What this contention overlooks is that “price” or “monopoly” are not names for definite “things”, fixed collections of physical attributes which we recognise by some of these attributes as members of the same class and whose further attributes we ascertain by observation; but that they are objects which can be defined only in terms of certain relations between human beings and which cannot possess any attributes except those which follow from the relations by which they are defined. They can be recognised by us as prices or monopolies only because, and in so far as, we can recognise these individual attitudes and from these as elements form the structural pattern which we call a price or monopoly. Of course the “whole” situation, or even the “whole” of the men who act, will greatly differ from place to place and from time to time. But it is nothing but our capacity to recognise the familiar elements from which the unique situation is made up which enables us to attach any meaning to the phenomena. Either we cannot thus recognise the meaning of the individual actions, they are nothing but physical facts to us, the handing over of certain material things, etc., or we must place them in the mental categories familiar to us but not definable in physical terms. If the first contention were true, however, this would mean that we could not know the facts of the past at all because in that case we could not understand the documents from which we derive all knowledge of them. 
Consistently pursued historism necessarily leads to the view that the human mind is itself variable and that not only are most or all manifestations of the human mind unintelligible to us apart from their historical setting, but that from our knowledge of how the whole situations succeed each other we can learn to recognise the laws according to which the human mind changes, and that it is the knowledge of these laws which alone puts us in a position to understand any particular manifestation of the human mind. Historism, because of its refusal to recognise a compositive theory of universal applicability, unable to see how different configurations of the same elements may produce altogether different complexes, and unable, for the same reason, to comprehend how the wholes can ever be anything but what the human mind consciously designed, was bound to seek the cause of the changes in the social structures in changes of the human mind itself - changes which it claims to understand and explain from changes in the directly apprehended wholes. From the extreme assertion of some sociologists that logic itself is variable, and the belief in the “pre-logical” character of the thinking of primitive people, to the more sophisticated contentions of the modern “sociology of knowledge”,
1. Cf. C. V. Langlois and C. Seignobos, Introduction to
the Study of History, transl. by G. G. Berry,
this approach has become one of the most characteristic features of modern sociology. It has raised the old question of the “constancy of the human mind” in a more radical form than has ever been done before.
This phrase is, of course, so vague that any dispute about it without giving it further precision is not likely to be profitable. That not only any human individual in its historically given complexity, but also certain types predominant in particular ages or localities, differ in significant respects from other individuals or types is, of course, beyond dispute. But this does not alter the fact that in order that we should be able to recognise or understand them at all as human beings or minds there must be certain invariable features present. We cannot recognise “mind” in the abstract. When we speak of mind what we mean is that certain phenomena can be successfully interpreted on the analogy of our own mind, that the use of the familiar categories of our own thinking provides a satisfactory working explanation of what we observe. But this means that to recognise something as mind is to recognise it as something similar to our own mind, and that the possibility of recognising mind is limited to what is similar to our own mind. To speak of a mind with a structure fundamentally different from our own, or to claim that we can observe changes in the basic structure of the human mind is not only to claim what is impossible; it is a meaningless statement. Whether the human mind is constant can never become a problem - because to recognise mind cannot mean anything but to recognise something as operating in the same way as our own thinking.
To recognise the existence of a mind always implies that we add something to what we perceive with our senses, that we interpret the phenomena in the light of our own mind, or find that they fit into the ready pattern of our own thinking. This kind of interpretation of human actions may not be always successful, and, what is even more embarrassing, we may never be absolutely certain that it is correct in any particular case; all we know is that it works in the overwhelming number of cases. Yet it is the only basis on which we ever understand what we call other people’s intentions or the meaning of their actions; and certainly the only basis of all our historical knowledge since this is all derived from the understanding of signs or documents. As we pass from men of our own kind to different types of beings we may, of course, find that what we can thus understand becomes less and less. And we cannot exclude the possibility that one day we may find beings who, though perhaps physically resembling men, behave in a way which is entirely unintelligible to us. With regard to them we should indeed be reduced to the “objective” study which the behaviourists want us to adopt toward men in general. But there would be no sense in ascribing to these beings a mind different from our own. We should know nothing of them which we could call mind, we should indeed know nothing about them but physical facts. Any
interpretation of their actions in terms of such categories as intention or purpose, sensation or will, would be meaningless. A mind about which we can intelligibly speak must be like our own.
The whole idea of the variability of the human mind is a direct result of the erroneous belief that mind is an object which we observe as we observe physical facts. The sole difference, however, between mind and physical objects, which entitles us to speak of mind at all, is precisely that wherever we speak of mind we interpret what we observe in terms of categories which we know only because they are the categories in which our own mind runs. There is nothing paradoxical in the claim that all mind must run in terms of certain universal categories of thought, because where we speak of mind this means that we can successfully interpret what we observe by arranging it in these categories. And anything which can be comprehended through our understanding of other minds, anything which we recognise as specifically human, must be comprehensible in terms of these categories.
Through this theory of the variability of the human mind, to which the consistent development of historism leads, it cuts, in fact, the ground under its own feet: it is led to the self-contradictory position of generalising about facts which, if the theory were true, could not be known. If the human mind were really variable so that, as the extreme adherents of historism assert, we could not directly understand what people of other ages meant by a particular statement, history would be inaccessible to us. The wholes from which we are supposed to understand the elements would never be visible to us. And even if we disregard this fundamental difficulty created by the impossibility of understanding the documents from which we derive all historical knowledge, without first understanding the individual actions and intentions the historian could never combine them into wholes and never explicitly state what these wholes are. He would, as indeed is true of so many of the adherents of historism, be reduced to talking about “wholes” which are intuitively comprehended, to making uncertain and vague generalisations about “styles” or “systems” whose character could not be precisely defined.
It follows indeed from the nature of the evidence on which all our historical knowledge is based that history can never carry us beyond the stage where we can understand the working of the minds of the acting people because they are similar to our own. Where we cease to understand, where we can no longer recognise categories of thought similar to those in terms of which we think, history ceases to be human history. And precisely at that point, and only at that point, do the general theories of the social sciences cease to be valid. Since history and social theory are based on the same knowledge of the working of the human mind, the same capacity to understand other people, their range and scope is necessarily co-terminus. Particular propositions of social theory may have no application at certain times, because the combination of elements to which they refer to do not occur.  But
l. Cf. W. Eucken, Grundlagen der Nationalökonomie, 1940, pp. 203-205.
they remain nevertheless true. There can be no different theories for different ages, though at some times certain parts and at others different parts of the same body of theory may be required to explain the observed facts, just as, e.g., generalisations about the effect of very low temperatures on vegetation may be irrelevant in the tropics but still true. Any true theoretical statement of the social sciences will cease to be valid only where history ceases to be human history. If we conceive of somebody observing and recording the doings of another race, unintelligible to him and to us, his records would be in a sense history as, e.g., the history of an ant-heap. Such history would have to be written in purely objective, physical terms. It would be the sort of history which corresponds to the positivist ideal, such as the proverbial observer from another planet might write of the human race. But such history could not help us to understand any of the events recorded by it in the sense in which we understand human history.
When we speak of man we necessarily imply the presence of certain familiar mental categories. It is not the lumps of flesh of a certain shape which we mean, nor any units performing definite functions which we could define in physical terms. The completely insane, none of whose actions we can understand, is not a man to us - he could not figure in human history except as the object of other people’s acting and thinking. When we speak of man we refer to one whose actions we can understand. As old Democritus said
“Man is what is known to all.”
[HHC: original Greek not
(To be concluded.)
1. “Man is what is known to all. ” Cf. H. Diehis, Die .Fragmente der Vorsokratiker, 4th ed., BerlIn, 1922; Democritus, Fragment No. 165, Vol. II, p. 94. I owe the reference to Democritus in this connection to Professor Alexander Rüstow.
PART I PART II PART III