Talcott Parsons (1937)

The Structure of Social Action

Source: Structure of Social Action, publ. McGraw Hill, 1937. Introduction reproduced here.

INTRODUCTORY

THE PROBLEM

"Who now reads Spencer? It is difficult for us to realise how great a stir he made in the world.... He was the intimate confidant of a strange and rather unsatisfactory God, whom he called the principle of Evolution. His God has betrayed him. We have evolved beyond Spencer." Professor Brinton's verdict may be paraphrased as that of the coroner, "Dead by suicide or at the hands of person or persons unknown." We must agree with the verdict. Spencer is dead. But who killed him and how? This is the problem.

Of course there may well be particular reasons why Spencer rather than others is dead, as there were also particular reasons why he rather than others made such a stir. With these this study is not concerned. But in the " crime," the solution of which is here sought, much more than the reputation of, or interest in, a single writer has been done to death. Spencer was, in the general outline of his views, a typical representative of the later stages of development of a system of thought about man and society which has played a very great part in the intellectual history of the English-speaking peoples, the positivistic-utilitarian tradition. What has happened to it? Why has it died?

The thesis of this study will be that it is the victim of the vengeance of the jealous god, Evolution, in this case the evolution of scientific theory. In the present chapter it is not proposed to present an account either of what has evolved or of what it has evolved into; all that will come later. It is necessary to preface this with a tentative statement of the problem, and an outline of some general considerations relevant to the way the present task is to be undertaken, and how the present study should be judged.

Spencer's god was Evolution, sometimes also called Progress. Spencer was one of the most vociferous in his devotions to this god, but by no means alone among the faithful. With many other social thinkers he believed that man stood near the culminating point of a long linear process extending back unbroken, without essential changes of direction, to the dawn of primitive man. Spencer, moreover, believed that this culminating point was being approached in the industrial society of modern Western Europe. He and those who thought like him were confident that evolution would carry this process on almost indefinitely in the same direction cumulatively.

A good many students have lately become dubious of these propositions. Is it not possible that the future holds in store something other than "bigger and better" industrialism? The conception that, instead of this, contemporary society is at or near a turning point is very prominent in the views of a school of social scientists who, though they are still comparatively few, are getting more and more of a hearing.

Spencer was an extreme individualist. But his extremism was only the exaggeration of a deep-rooted belief that, stated roughly, at least in the prominent economic phase of social life, we have been blest with an automatic, self-regulating mechanism which operated so that the pursuit by each individual of his own self-interest and private ends would result in the greatest possible satisfaction of the wants of all. All that was necessary was to remove obstacles to the operation of this mechanism, the success of which rested on no conditions other than those included in the conception of rational pursuit of self-interest. This doctrine, too, has been subjected to increasingly severe criticism from many quarters, by no means all relevant to the purposes of this study. But another article of faith about the workings of the social world has been breaking down.

Finally, Spencer believed that religion arose from the prescientific conceptions of men about the empirical facts of their own nature and their environment. It was, in fact, the product of ignorance and error. Religious ideas would, with the progress of knowledge, be replaced by science. This was only a phase of a much wider deification of science. Indeed the interest of the Spencerian type of social scientist in religion has thus been virtually confined to primitive man - the question was, how has science developed out of primitive religion? In this field, too, there is increasing scepticism of the Spencerian view. It has been possible above to cite views on only a few questions. It is, however, enough to indicate that a basic revolution in empirical interpretations of some of the most important social problems has been going on. Linear evolutionism has been slipping and cyclical theories have been appearing on the horizon. Various kinds of individualism have been under increasingly heavy fire. In their place have been appearing socialistic, collectivistic organic theories of all sorts. The role of reason and the status of scientific knowledge as an element of action have been attacked again and again. We have been overwhelmed by a flood of anti-intellectualistic theories of human nature and behaviour, again of many different varieties. A revolution of such magnitude in the prevailing empirical interpretations of human society is hardly to be found occurring within the short space of a generation, unless one goes back to about the sixteenth century. What is to account for it?

It is, of course, very probable that this change is in considerable part simply an ideological reflection of certain basic social changes. This thesis would raise a problem, the answer to which would be difficult to find in terms of Spencerian thought. But to deal adequately with this problem would far transcend the limits of this study.

It is no less probable that a considerable part has been played by an "immanent" development within the body of social theory and knowledge of empirical fact itself. This is the working hypothesis on which the present study has been made. The attempt will be made to trace and evaluate the significance of one particular phase of this process of development which can be discerned and analysed in detail in the work of a limited group of writers in the social field, mostly known as sociologists. But before entering upon this enterprise it is necessary to make a few preliminary methodological remarks about the nature of a ''body of social theory and knowledge of empirical fact." What are the main relations of the principal elements in it to each other, and in what sense and by what kind of process may such a "body" be thought to be undergoing a process of development? Only then can it be stated explicitly what kind of study is here proposed and what order of results may reasonably be expected from it.

THEORY AND EMPIRICAL FACT

In the following discussion some fundamental methodological propositions will be laid down without any attempt to give them a critical foundation. It will, however, turn out that the question of the status of these views will form one main element of the subject matter of the whole study. Their soundness is to be judged not in terms of the arguments brought forward in their defence in the present introductory discussion but in terms of the way they fit into the structure of the study as a whole and its outcome.

There is, more often implicit than explicit, a deep-rooted view that the progress of scientific knowledge consists essentially in the cumulative piling up of "discoveries" of "fact." Knowledge is held to be an entirely quantitative affair. The one important thing is to have observed what had not been observed before. Theory, according to this view, would consist only in generalisation from known facts, in she sense of what general statements the known body of fact would justify. Development of theory would consist entirely in the process of modification of these general statements to take account of new discoveries of fact. Above all, the process of discovery of fact is held to be essentially independent of the existing body of "theory," to be the result of some such impulse as "idle curiosity."

It is evident that such terms as "fact" are much in need of definition This will come later. At the present juncture against the view just roughly sketched may be set another, namely, that scientific " theory " - most generally defined as a body of logically interrelated "general concepts" of empirical reference - is not only a dependent but an independent variable in the development of science. It goes without saying that a theory to be sound must fit the facts but it does not follow that the facts alone, discovered independently of theory, determine what the theory is to be, nor that theory is not a factor in determining what facts will be discovered, what is to be the direction of interest of scientific investigation.

Not only is theory an independent variable in the development of science, but the body of theory in a given field at a given time constitutes to a greater or less degree an integrated "system." ' That is, the general propositions (which may be, as will be seen later, of different kinds) which constitute a body of theory have mutual logical relations to each other. Not, of course, that all the rest are deducible from any one - that would confine theory to the one proposition - but in the sense that any substantive change in the statement of one important proposition of the system has logical consequences for the statement of the others. Another way ff of putting this is to say that any system of theory has a determinate logical structure.

Now obviously the propositions of the system have reference to matters of empirical fact; if they did not, they could have no claim to be called scientific. Indeed, if the term fact is properly interpreted it may be said that a theoretical proposition, if it has a place in science at all, is either itself a statement of fact or a statement of a mode of relations between facts. It follows that any important change in our knowledge of fact in the field in question must of itself change the statement of at least one of the propositions of the theoretical system and, through the logical consequences of this change, that of other propositions to a greater or lesser degree. This is to say, the structure of the theoretical system is changed. All this seems to be in accord with the empiricist methodology sketched above.

But, in the first place, it will be noted that the word "important " used above was italicised. What does an important change in our knowledge of fact mean in this context? Not that the new facts are vaguely "interesting," that they satisfy "idle curiosity, " or that they demonstrate the goodness of God. But the scientific importance of a change in knowledge of fact consists precisely in j its having consequences for a system of theory. A scientifically unimportant discovery is one which, however true and however interesting for other reasons, has no consequences for a system of theory with which scientists in that field are concerned. Conversely, even the most trivial observation from any other point of new - a very small deviation of the observed from the calculated position of a star, for instance - may be not only important but of revolutionary importance, if its logical consequences for the structure of theory are far-reaching. It is probably safe to say that all the changes of factual knowledge which have led to the relativity theory, resulting in a very great theoretical development, are completely trivial from any point of view except their relevance to the structure of a theoretical system. They have not, for instance, affected in any way the practice of engineering or navigation.

This matter of the importance of facts is, however, only one part of the picture. A theoretical system does not merely state facts which have been observed and that logically deducible relations to other facts which have also been observed. In so far as such a theory is empirically correct it will also tell us what empirical facts it should be possible to observe in a given set of circumstances. It is the most elementary rule of scientific integrity that the formulator of a theoretical proposition must take into account all the relevant known facts accessible to him. This process of verification, fundamental to science, does not consist merely in reconsideration of this applicability to known facts by others than the original formulator of the theory, and then simply waiting for new facts to turn up. It consists in deliberately investigating phenomena with the expectations derived from the theory in mind and seeing whether or not the facts actually found agree with these expectations.

This investigation is one of situations which have been studied either never at all before or not with these particular theoretical problems in mind. Where possible the situations to be investigated are experimentally produced and controlled. But this is a matter of practical technique, not of logic.

In so far as the expectations from the theory agree with the facts found, making allowance for "errors of observation," etc., the theory is "verified." But the significance of the process of verification is by no means confined to this. If this does not happen, as is often so, either the facts may be found to disagree with the theoretical expectations, or other facts may be found which have no place in the theoretical system. Either result necessitates critical reconsideration of the system itself. There is, then, a reciprocal process: direction, by the expectations derived from a system of theory, toward fields of factual investigation, then reaction of the results of this investigation on the theory.

Finally, verification in this sense is not the only important relation of a theoretical system to the direction of empirical investigation. Not only are specific theoretical propositions which have been directly formulated with definite matters of fact in view subject to verification. But further, a theoretical system built up upon observations of fact will be found, as its implications are progressively worked out, to have logical consequences for fields of fact with which its original formulators were not directly concerned. If certain things in one field are true, then other things in another, related field must also be true. These implications also are subject to verification, which in this case takes the form of finding out what are the facts in this field. The results of this investigation may have the same kind of reaction on the theoretical system itself.

Thus, in general, in the first instance, the direction of interest in empirical fact will be canalised by the logical structure of the theoretical system. The importance of certain problems concerning the facts will be inherent in the structure of the system. Empirical interest will be in the facts so far as they are relevant to the solution of these problems. Theory not only formulates what we know but also tells us what we want to know, that is, the questions to which an answer is needed. Moreover, the structure of a theoretical system tells us what alternatives are open in the possible answers to a given question. If observed facts of undoubted accuracy will not fit any of the alternatives it leaves open, the system itself is in need of reconstruction.

A further point is of importance in the present connection. Not only do theoretical propositions stand in logical interrelations to each other so that they may be said to constitute "systems" but it is in the nature of the case that theoretical systems should attempt to become "logically closed." That is, a system starts with a group of interrelated propositions which involve reference to empirical observations within the logical framework of the propositions in question. Each of these propositions has logical implications. The system becomes logically closed when each of the logical implications which can be derived from any one proposition within the system finds its statement in another proposition in the same system. It may be repeated that this does not mean that all the other propositions must be logically derivable from any one - on the contrary, if this were true scientific theory would be sheer tautology.

The simplest way to see the meaning of the concept of a closed , system in this sense is to consider the example of a system of simultaneous equations. Such a system is determinate, i.e., closed, when there are as many independent equations as there are independent variables. If there are four equations and only three variables, and no one of the equations is derivable from the others by algebraic manipulation then there is another variable missing. Put in general logical terms: the propositions stated in the four equations logically involve an assumption which is not stated in the definitions of the three variables.

The importance of this is clear. If the explicit propositions of a system do not constitute a logically closed system in this sense it may be inferred that the arguments invoked rest for their logical cogency on one or more unstated assumptions. It is one of the prime functions of logical criticism of theoretical systems to apply` this criterion and, if gaps are found, to uncover the implicit assumptions. But though all theory tends to develop logically closed systems in this sense it is dangerous to confuse this with the " empirical " closure of a system. To this issue, that of "empiricism," it will be necessary often to return.

The implications of these considerations justify the statement that all empirically verifiable knowledge even the commonsense knowledge of everyday life - involves implicitly, if not explicitly, systematic theory in this sense. The importance of this statement lies in the fact that certain persons who write on social subjects vehemently deny it. They say they state merely facts and let them "speak for themselves." But the fact a person denies that he is theorising is no reason for taking him at his word and failing to investigate what implicit theory is involved in his statements. This is important since " empiricism " in this sense has been a very common methodological position in the social sciences.

From all this it follows what the general character of the problem of the development of a body of scientific knowledge is, in so far as it depends on elements internal to science itself. It is that of increasing knowledge of empirical fact, intimately combined with changing interpretations of this body of fact - hence changing general statements about it - and, not least, a changing a structure of the theoretical system. Special emphasis should be laid on this intimate interrelation of general statements about empirical fact with the logical elements and structure of theoretical systems.

In one of its main aspects the present study may be regarded as an attempt to verify empirically this view of the nature of science and its development in the social field. It takes the form of the thesis that intimately associated with the revolution in empirical interpretations of society sketched above there has in fact occurred an equally radical change in the structure of theoretical systems. The hypothesis may be put forward, to be tested by the s subsequent investigation, that this development has been in large part a matter of the reciprocal interaction of new factual insights and knowledge on the one hand with changes in the theoretical system on the other. Neither is the "cause" of the other. Both are in a state of close mutual interdependence.

This verification is here attempted in monographic form. The central focus of attention is in the process of development of one coherent theoretical system, that to be denoted as the voluntaristic theory of action, and the definition of the general concepts of which this theory is composed. In the historical aspect the primary interest is in the process of transition from one phase of its development to another, distinctly different, one. Of the first phase Spencer may be regarded as a late, and in some points extreme, but nevertheless a typical representative. For convenience of reference and for no other purpose this has been designated as the "positivistic" system of the theory of action, and its variant, - which is most important to the present study, the "utilitarian." Both these terms are used in technical senses in this work and they will be defined in the next chapter, where the main logical structure of the positivistic system is outlined.

It is, however, a striking fact that what is in all essential respects the same system may be found emerging by a similar process of transition from the background of a radically different theoretical tradition which may be designated as the "idealistic." One dominant case of this latter transition, the work of Max Weber, will be dealt with at length. It goes without saying that this convergence, if it can be demonstrated, is a very strong argument for the view that correct observation and interpretation of the facts constitute at least one major element in the explanation of why this particular theoretical system has developed at all.

As has been said, interest will be focused in the process of emergence of a particular theoretical system, that of the "voluntaristic theory of action".' But the above considerations indicate the great importance of dealing with this in the closest connection with the empirical aspects of the work of the men whose theories are to be treated. So for each major thinker at least a fair sample of the major empirical views he held will be presented, and the attempt made to show in detail the relations of these to the theoretical system in question. In each case the thesis will be maintained that an adequate understanding of how these empirical views were arrived at is impossible without reference to the logical structure and relations of the theoretical concepts employed by the writer in question. And in every case except that of Marshall the attempt will be made to demonstrate that the conspicuous change in his empirical views from those current in the tradition with which the writer in question was most closely associated cannot be understood without reference to the corresponding change in the structure of his theoretical system from that dominant in the tradition in question. If this can be demonstrated it will have important general implications. It will be strong evidence that he who would arrive at important empirical conclusions transcending common sense cannot afford to neglect considerations of systematic theory.

The choice of writers to be treated here has been dictated by a variety of considerations. The central interest of the study is in the development of a particular coherent theoretical system, as an example of the general process of "immanent" development of science itself. This process has been defined as a matter of the logical exigencies of theoretical systems in close mutual interrelation with observations of empirical fact and general statements embodying these facts. Hence a choice of authors is indicated which will serve to isolate these elements as far as possible from Others, such as influence of the general "climate of opinion," irrelevant to the purposes of this study.

The first criterion is actual concern with the theory of action Among those who are satisfactory in this respect it is desirable to have represented as great a diversity of intellectual tradition, social milieu and personal character as possible. The inclusion of Marshall is justified by the fact that economic theory and the question of its status involve a crucial set of problems in relation to the theory of action in general and to the positivistic system, especially its utilitarian variant.

This question is as will be seen, the most important single link between utilitarian positivism and the later phase of the theory of action. Pareto also was deeply concerned with the same set of problems, but in relation to distinctly different aspects of the positivistic tradition, and in the midst of a strikingly different climate of opinion. The comparison of the two is most instructive.

Durkheim's starting point was also positivistic, indeed by far the most explicitly so of the three. But it was the variant of the positivistic system most radically foreign to that of utilitarian individualism in which Marshall was primarily immersed, and Pareto also, though to a less extent. In personal character and background more violent contrasts are scarcely imaginable than between Marshall, the strongly moralistic middle-class Englishman; Durkheim, the Alsatian Jewish, radical, anticlerical, French professor; Pareto, the aloof, sophisticated Italian nobleman; and, finally, Weber, a son of the most highly cultured German upper middle class, who grew up on the background of German idealism and was trained in the historical schools of jurisprudence and economics. These intellectual influences were of no real importance in the formation of the thought of any of the other three. Moreover, Weber's personal character was radically different from any of the other three.

Another point strongly in favour of this choice is that although all four of these men were approximately contemporary, there is with one exception not a trace of direct influence of any one on any other. Pareto was certainly influenced by Marshall in the formulation of his technical economic theory, but with equal certainty not in any respect relevant to this discussion. And this is the only possibility of any direct mutual influence. In fact, within the broad cultural unit, Western and Central Europe at the end of the nineteenth and beginning of the twentieth century, it would scarcely be possible to choose four men who had important ideas in common who were less likely to have been influenced in developing this common body of ideas by factors other than the immanent development of the logic of theoretical systems in relation to empirical fact.

Certain other considerations are relevant. The main concern of the study is with the outline of a theoretical system. Its minor variations from writer to writer are not a matter of concern to this analysis. It is, however, necessary to work out its logical structure and ramifications in the clearest form attainable. Hence the choice has been made of intensive analysis from the relevant point of view of the work of a small number of the most eminent men. Marshall was, by many in his field, thought to be the most eminent economist of his generation. But the interest of the present study in him is more limited than in the others. The other three are all generally known as sociologists. There can be little question of their eminence in their generation in their field. A list of the first six sociologists of the last generation which failed to include all three names could hardly be taken seriously. This is not to say they are the only equally eminent ones, but for the purposes of this study they are distinctly the most suitable.

In order to avoid all possibility of misunderstanding, it should be reiterated: This study is meant to be a monographic study of one particular problem in the history of recent social thought, that of the emergence of the theoretical system which has been called the "voluntaristic theory of action." It follows that there are a number of related things which this study is not and is not meant to be. In the first place it is not a history of sociological theory in Europe in, roughly, the last generation. It deliberately avoids the inclusiveness with regard both to problems and to men which such a task would require. If there is anything at all in its results, it follows that the process under investigation is one element of the history of European sociological theory in that period. Then this study will constitute a monographic contribution to this history, but that is all.

In the second place, it is not a general secondary interpretation of the work of any or all of the men dealt with. Its aim is neither secondary exposition as such nor critical evaluation of them. With respect to each of the theorists the aspects which this study rests are of great, sometimes of central, importance to their work " a whole. But in the treatment of none will the attempt be made to evaluate this importance relatively to that of other aspects. That must be left to other studies. Finally, in harmony with all this, there has been no attempt to discuss all aspects of the work of these men or all the secondary literature about them. Practically all the existing secondary literature about them has been read, but has been cited only where it seemed particularly relevant to the immediate context. Failure to cite is not to be interpreted as implied criticism, only lack of important bearing. Also, with the texts themselves, encyclopedic completeness has not been aspired to. Nor has every passage that could be construed as relevant to the purpose in hand been cited but only enough, taken in terms of the structure of the writers' theories as a whole, to establish the points at issue.

Perhaps one more word with reference to interpretation may be permitted. This study is conceived to be an organic whole, concerned with ideas which are logically interrelated and permeate the whole study. The reader should keep this in mind in weighing whatever critical remarks he may be inclined to make. Particularly in a study of this character, it is legitimate to ask that a fact cited or a statement made be taken not only in its immediate intrinsic character and meaning but also in relation to the total structure of which it forms a part.

RESIDUAL CATEGORIES

Two or three further preliminary questions should be taken up so as not to leave the reader in doubt on some matters that are bound to arise in his mind. In the first place, one further conclusion about the character of scientific development follows from the position already taken. It is possible to have scattered and unintegrated bits of knowledge`, and to assent to the "truth" of further scattered bits as they are called to one's attention. This type of knowledge does not, however, constitute "science " in the sense in which this study is interested in it.

The latter is present only in so far as these bits of knowledge have become integrated with reference to fairly clear-cut theoretical systems. In so far as this has happened, two things can be said. It is at least unlikely that such a system should play an important part in canalising the thought of a considerable number of highly intelligent men over a period of time, if it were not that the propositions of the system involved empirical references to phenomena which were real and, within the framework of the conceptual scheme, on the whole correctly observed.

At the same time the structure of the conceptual scheme itself inevitably focuses interest on a limited range of such empirical facts. These may be thought of as a " spot" in the vast encircling darkness, brightly illuminated as by a searchlight. The point is, what lies outside the spot is not really "seen" until the searchlight moves, and then only what lies within the area into which its beam is newly cast. Even though any number of facts may be "known" outside this center, they are not scientifically important until they can be brought into relation with a theoretical system.

This fact is of the greatest importance as a canon of interpretation. In studying a man's empirical work the questions asked will not merely be, what opinions did he hold about certain concrete phenomena. nor even. what has he in general contributed to our knowledge of these phenomena? The primary questions will, rather, be, what theoretical reasons did he have for being interested in these particular problems rather than others, and what did the results of his investigation contribute to the solution of his theoretical problems? Then, in turn, what did the insights gained from these investigations contribute to the restatement of his theoretical problems and through this to the revision of his theoretical system? Thus, in connection with Durkheim the real point of interest is not in his having established the fact that the suicide rate in the French army was, during a certain period, considerably higher than in the civil population. Those interested in this fact for its own sake can consult his study. The present interest is, rather, why did Durkheim study suicide anyway, and what is the significance for his general theory of this and the other facts he established in the course of his investigation of it?

Something should also be said about the general character of the process by which this awakening of new scientific interest in fields of fact proceeds, and theoretical problems shift. Every system, including both its theoretical propositions and its main relevant empirical insights, may be visualised as an illuminated spot enveloped by darkness. The logical name for the darkness is, in general, "residual categories." Their role may be deduced from the inherent necessity of a system to become logically closed. On whatever level it operates, a theoretical system must involve the positive definition of certain empirically identifiable variables or other general categories. The very fact that they are defined at all implies that they are distinguished from others and that the facts which constitute their empirical reference are thereby, in certain aspects at least, specifically differentiated from others.

If, as is almost always the case, not all the actually observable facts of the field, or those which have been observed, fit into the sharply, positively defined categories, they tend to be given one or more blanket names which refer to categories negatively defined, that is, of facts known to exist, which are even more or less adequately described, but are defined theoretically by their failure to fit into the positively defined categories of the system. The only theoretically significant statements that can be made about these facts are negative statements - they are not so and so. But it is not to be inferred that because these statements are negative they are therefore unimportant.

It is true that in the work of the mediocre proponents of a theoretical system the qualifications of their empirical deductions from theory which are necessitated by the existence of these residual categories are often ignored, or so vaguely stated as to be virtually meaningless. In the case of the dogmatists of the system their existence, or at least their importance for the system, may even be vehemently denied. Both procedures are vastly encouraged by an empiricist methodology. But in the work of the ablest and most clear-headed proponents of a system these residual categories will often be not merely implicit but explicit, and will be quite clearly stated. In this sense, the best place to go to find the starting points of the breakdown of a system is to the work of the ablest proponents of the system itself. This more than any other reason is the explanation of why the work of so many of the greatest scientific theorists is "difficult." Only the lesser lights can bring themselves to dogmatise about the exclusive importance and adequacy of their own positively defined categories.

It follows from this that the surest symptom of impending change in a theoretical system is increasingly general interest in such residual categories. Indeed, one kind of progress of theoretical work consists precisely in the carving out from residual categories of definite positively defined concepts and their verification in empirical investigation. The obviously unattainable, but asymptotically approached goal of the development of scientific theory is, then, the elimination of all residual categories from science in favour of positively defined, empirically verifiable concepts- For any one system there will, to be sure, always be residual categories of fact, but they will be translatable into positive categories of one or more other systems. For the empirical application of any one system these residual elements will be found to be involved in the necessary data.

The process of the carving out of positive concepts from residual categories is also a process by which the reconstruction of theoretical systems is accomplished as a result of which they may eventually be altered beyond all recognition. But this should be said: The original empirical insights associated with the positive categories of the original system will be restated in different form, but unless they entirely fail to stand up to the combined criticism of theory and renewed empirical verification, they will not be eliminated. Indeed, as has been noted above, this is unlikely to happen. This fact is the essential basis for the justification of talk of the "progress" of science. Theoretical systems change. There is not merely a quantitative accumulation of "knowledge of fact" but a qualitative change in the structure of theoretical systems. But in so far as verification has been valid and sound, this change leaves behind it a permanent precipitate of valid empirical knowledge. The form of statement may well change, but the substance will remain. The older statement will generally take the form of a "special case" of the new. The utilitarian branch of positivistic thought has, by virtue of the structure of its theoretical system, been focused upon a given range of definite empirical insights and related theoretical problems. The central fact. - a fact beyond all question. - is that in certain aspects and to certain degrees, under certain conditions, human action is rational. That is, men adapt themselves to the conditions which they are placed and adapt means to their ends in such a way as to approach the most efficient manner of achieving these ends. And the relations of these means and conditions to the achievement of their ends are "known" to be intrinsically verifiable by the methods of empirical science.

Of course this statement contains a considerable number of terms which have been, and still are, ambiguous in general usage. Their definition is one of the prime tasks of the study as a whole. This range of factual insight and the theoretical problems involved in it, and this alone, is the theme of the first analysis. The task of the first two parts of the study is to trace its development from one well-defined theoretical system to another. The process has been essentially that just sketched, a process of focusing attention on, and carving positive theoretical concepts out of, the residual categories to be found in the various versions of the initial system.

Perhaps it is permissible to state here, or to repeat in a somewhat different form, a vital canon of interpretation for a study of this kind. It is in the nature of the enterprise that many facts and theoretical considerations that are important from any one of a large number of different possible points of view will have been neglected. A specific criterion has just been laid down of what scientific "importance" is considered to mean, and the remarks just made serve further to elucidate the meaning of this criterion. If a critic is to charge neglect of the importance of such things, he should be able to show either (a) that the neglected consideration bears specifically on the limited range of theoretical problems to which this study has been deliberately limited and that its correct consideration would significantly alter the conclusions about them or (b) that the whole conception of the nature of science and its development here advanced is so fundamentally wrong that these criteria of importance are inapplicable.

[In general, pains have been taken to state legitimate lines of criticism as explicitly as possible because it is my experience, particularly in dealing with the secondary literature on these writers, that it is extraordinarily difficult for an idea or ideas which do not fit the requirements of the prevailing "system" or systems to be understood at all even by very intelligent people. These writers are persistently criticised in terms utterly inapplicable to them. The fates both of Durkheim's proposition "Society is a reality sui generis, which is still predominantly held to be merely an unusable "metphysical postulate" (it started precisely as a residual category), and of Weber's theory of the relations of Protestantism and capitalism are conspicuous examples].

THEORY, METHODOLOGY AND PHILOSOPHY

Out of these considerations grows directly another range of problems which must be commented upon briefly. It will be asked whether a study of this character will find it possible to confine itself to "science" or will not find it necessary to embark upon the perilous waters of philosophy. Such a venture will, indeed, prove necessary at certain points and it is hence advisable to make a general statement of the relevant relations of these two kinds of discipline to each other and to the kind of study here attempted. Like the other statements in this chapter it will be brief and without critical foundation.

The main outline of a view of the general character of empirical science has already been presented. The distinction of science from all the philosophical disciplines is vital. It will turn out to be so at every stage of the ensuing study. But this is not to be taken to mean that the two kinds of discipline are without significant mutual interrelations and that each can afford to ignore the other. For the purposes of this study. - not necessarily for others. - it is legitimate to define philosophy as a residual category. It is the attempt to achieve a rational cognitive understanding of human experience by methods other than those of empirical science.

That there are important mutual relations of philosophy and science, once the distinction between them is established, is a simple deduction from the most general nature of reason itself. The tendency of theoretical systems in science to become logically closed is a special case. The general principle is that it is in the nature of reason to strive for a rationally consistent account of all experience which comes within its range at all. In so far as both philosophical and scientific propositions are brought to the attention of the same mind, there is in the nature of the case, a tendency to bring them into relations of logical consistency with one another. It likewise follows that there are no logically watertight compartments in human experience. Rational knowledge is a single organic whole.

The methodological principles already laid down yield a canon for use in this context as well as others. Since the present concern is with the character and development of certain specific theoretical systems in science, and the interest in these systems is scientific, philosophical questions will be treated only when they become important to these systems in the sense strictly defined. Discussion will be deliberately limited to important philosophical questions in this specific sense. But equally there will be no attempt to avoid them on the plea that they are philosophical or "metaphysical" and hence have no place in a scientific study. This is often a facile way of evading the clear decision of vital but embarrassing issues.

It is important briefly to indicate a few of the main ways in which philosophical questions will be found to impinge upon the problems of this study. In the first place, while scientific knowledge is not the only significant cognitive relation of man to his experience, it is a genuine and valid one. This means that the two sets of disciplines stand in a relation of mutually corrective criticism. In particular, the evidence gained from scientific sources, observation of fact and the theoretical consequences of these facts constitutes, in so far as it is sound, valid ground for criticism of philosophical views.

If, then, scientific evidence which there is reason to believe is correct and has a bearing on important problems, is in conflict with philosophical views explicitly or implicitly involved in the works studied, this will be taken as an indication of the necessity to inquire into the basis of these views on a philosophical level. The object will be to discover whether the philosophical grounds for them are so cogent as to leave no alternative but to revise the earlier impression of the validity of what purported to be scientific evidence. A number of instances of such conflicts will be encountered where philosophical ideas do conflict with crucially important and relevant empirical evidence. However, in none of these has it been possible to discover sufficiently cogent philosophical grounds for discarding this evidence.

But this necessity of criticising philosophical positions from a scientific point of view is not the only important relation of the two sets of disciplines. Every system of scientific theory involves by implication philosophical consequences, both positive and negative. This is nothing more than a corollary of the rational unity of cognitive experience. Then it is also true that every system of scientific theory involves philosophical assumptions.

These may lie in a number of different directions. But the ones to which special attention should be called now are the "methodological." That is, the questions of the grounds of empirical validity of scientific propositions, the kinds of procedures which may on general grounds be expected to yield valid knowledge, etc., impinge directly on the philosophical fields of logic and epistemology.

Indeed it is scarcely too much to say that the main preoccupation of modern epistemology from, approximately, Locke on has been with precisely this question of the philosophical grounds for the validity of the propositions of empirical science. Since all through the study questions of validity will be of pressing importance, discussions of their philosophical aspects cannot safely be neglected. This is important especially in one context. A group of methodological views will be encountered which, again for convenience of reference and that purpose alone, have been brought together under the term "empiricism." The common characteristic of them is the identification of the meanings of the concrete specific propositions of a given science, theoretical or empirical, with the scientifically knowable totality of the external reality to which they refer. They maintain, that is, that there is an immediate correspondence between concrete experienceable reality and scientific propositions, and only in so far as this exists can there be valid knowledge. In other words, they deny the legitimacy of theoretical abstraction. It should already be evident that any such view is fundamentally incompatible with the view of the nature and status of theoretical systems which is a main foundation of this whole study. Hence discussion of the philosophical grounds advanced to support it cannot be avoided.

It is in this sense of the borderline field between science on the one hand, logic and epistemology on the other, that the term "methodology" as used in this work should be understood Its reference is thus not primarily to "methods" of empirical research such as statistics, case study, interview and the like. These latter it is preferable to call research techniques.