The term ‘social change’ will be taken to mean a change in the institutional structure of a social system; more particularly, a transformation of the core institutional order of a society such that we can speak of a change in type of society. I do not believe that it is necessary to reach agreement on what is meant by the ‘core institutional order’ of a society or on how a typology of societies is to be differentiated before there can be meaningful discussion of how the process of change takes place. That is, unless there is some a priori commitment to a ‘dominant factor’ theory of social change; in which case the wrangle about whether change has ‘really’ taken place can be endless.
The main purpose of this chapter is to discuss some of the implications of recent criticisms of functionalism, especially those which have a bearing on how social change is internally generated in a society. The thesis is that, in concentrating their fire on a special, albeit prominent, version of functionalism (‘normative functionalism’) critics have become over involved with what may be called the problems of ‘social integration’. As a result, they have tended to ignore what is just as relevant to their central interests in conflict and social change, namely, the problem of ‘system integration’. And here the perspective of general functionalism would still seem to be the most useful instrument.
In a recent article, Kingsley Davis has proposed such a catholic definition of functionalism as to make it virtually indistinguishable from the most basic presuppositions of contemporary sociology. This is all very comforting. But if by functionalism nothing more were meant than seeing society as a system of interdependent parts, and an aversion to ‘reductionism’, then most of those who have been engaged in criticism of functionalism would be proselytized overnight. How many would accept the attendant ideas, such as that of ‘functional requisites’, is more debatable, and would probably depend on how they were interpreted. Again, exactly what elements are included as ‘parts’ of a social system, and the exact implications of the idea of ‘interdependence’ itself, are obviously areas of potential disagreement.
// 400 But, omitting these considerations, surely the ‘general’ functionalist standpoint which Davis has restated must be distinguished from its more specific and controversial form. Davis avoids mentioning precisely those characteristics which are now widely associated with, though not logically entailed by, a functionalist orientation: first, the emphatic role attributed to ‘common value elements’ in the integration of social action; and, second, the unwarranted assumption that the study of social stability must precede the analysis of social change. Both these predispositions, but especially the first, typify what we wish to speak of from now on as normative functionalism.
Before going on to examine the position to which we are led by the critics of normative functionalism, one further distinction is relevant to the subsequent argument. It is the wholly artificial one between ‘social integration’ and ‘system integration’. Whereas the problem of social integration focuses attention upon the orderly or conflictful relationships between the actors, the problem of system integration focuses on the orderly or conflictful relationships between the parts, of a social system.
It may be said at once that the connection between these two aspects of integration is neatly made by normative functionalism. The logic is simple. Since the only systematically differentiated parts of a society are its institutional patterns, the only source of social disorder arising from system disorder is that which takes the form of role conflict stemming from incompatible institutional patterns. If, however, it is held that such institutional patterns do not exhaust the general relevant ‘parts’ of a social system, then this particular articulation of system and social integration is only one way of relating the phenomena of ‘deviance’ and ‘conflict’ to the operation of the system as a functioning entity. To this point we shall return later. For the moment, what needs stressing is that the critics of normative functionalism have devoted their critique entirely to the wav in which this theory handles the problem of social integration; and particularly to the ambiguities of the concept of ‘institution’
The leading exponent of the general functionalist school, Robert K Merton, has already drawn attention to the static connection of the term // 401 institution: It is not enough’, he writes, ‘to refer to the “institutions” as though they were alt uniformly supported by all groups and strata in the society. Unless systematic consideration is given to the degree of support of particular ‘institutions” by specific groups we shall overlook the important place of power in society.’ The major criticism of normative functionalism which has frequently been made is that it treats institutions primarily as moral entities, without rigorously exploring the interplay between norms and power that is universally present in major institutional contexts. This weakness has been seized upon by such writers as Dahrendorf and Rex. Their basic theses are sufficiently similar to be treated jointly For the sake of convenience, their ideas may be called ‘conflict theory’.
The conflict theorists have pointed out first that norms and power must be considered as general alternative modes of ‘institutionalizing’ social relationships. To quote Rex:
We have also to recognize that some of the ends which the actors in our system pure may be random ends from the point of view of the system or actually in conflict with it. If there is an actual conflict of ends, the behaviour of actors towards one another may not be determined by shared norms but by the success which each has in compelling the other to act in accordance with his interests. Power then becomes a crucial variable in the study of social systems.
Second, potential conflicts of interests are seen as endemic in all social systems which ‘institutionalize’ power relationships, because power (authority) over others is the most general form of ‘scarce resource’ and one that is inherent in society itself. ‘The distribution of authority in associations’, writes Dahrendorf, ‘is the ultimate “cause” of the formation // 402 of conflict groups.’ Thus, if potential conflicts of interest between those who exercise authority and those over whom authority is exercised are anormal’ feature of social organization, the deinstitutionalization of power, and the use of power to maintain institutions, are ever present possibilities. In any realistic and dynamic view of institutionalization, the role of power, both in the generation and control of conflict, is of prime concern.
At first sight, it would seem that the image of society constructed by normative functionalism has given rise to counter-arguments which bring us round full circle to the polemical starting-point of modern sociology, namely, the debate on social contract. But fortunately both normative functionalists and conflict theorists are not prepared to recognize as a real issue the Greenian dichotomy of ‘Will’ versus ‘Force’. The themes of norms-consensus-order, and power-alienation-conflict are not regarded as viable sociological alternatives.
It is, therefore, a little surprising to find that both Dahrendorf and Rex consider it necessary to develop their antitheses to normative functionalism in a systematic form. These take the shape, respectively, of a ‘coercion theory of society’ and a ‘conflict model of society’. For this strategy they give reasons which are even more surprising. The first is that they both feel their ‘models’ or ‘frames of reference’ are specially suited to certain problem areas in sociology, particularly to the study of industrial socie- // 403 ties. And, second, Dabrendort feels that the unification of the ‘integration theory’ (normative functionalism) and the ‘coercion theory’ is unlikely and probably impossible.
Neither of these reasons is very compelling. You cannot assert that society is unthinkable as either a purely moral or a purely coercive entity, and then suggest that a vocabulary built around one or the other of these unthinkable premisses is necessary because some societies are manifestly more orderly or conflictful than others. To be sure, the degree to which power enters into social relationships is a factor indispensable for the understanding of both the ‘imperfection’ of consensus and the propensity to conflict. But even in situations where power is very evident and conflict endemic, it is doubtful whether the phenomena of conflict can be adequately grasped without incorporating into conflict theory many of the concepts and propositions concerning the dynamic properties of value systems (or ideologies) which have been developed, or taken over, by normative functionalism. For, given the power structure, the nature of the value system is of signal importance for the genesis, intensity, and direction of potential conflict. Particularly crucial is the way in which it structures the levels of aspiration of different social strata. It may, of its own accord, create aspirations which generate demands for change, or add fuel to the fire of conflicting material interests. It may be sufficiently open and ambiguous to be exploited simultaneously by different conflict groups; or, contrariwise, be capable of absorbing counter-ideologies within itself. Or, sudden change in the relative material positions of different groups may result in widespread conflict as a consequence of what Durkheim calls ‘moral declassification’ It could, therefore, be argued that even the analysis of that facet of social integration to which Dahrendorf and Rex consider their theories to be especially relevant-namely, social conflict-requires nothing less than a systematic extension of their framework to take explicitly into account the variable properties of value systems that have been the focus of normative functionalism. To the extent that this is done, their conflict theory ceases to be a ‘special’ approach. That status is reserved for the unmodified version of normative funcationalism.
Finally, both normative functionalism and conflict theory quite // 404 obviously utilize many sociological concepts which are the property of neither the one perspective nor the other for the solution of their respective problems. Witness only Dahrendorf’s extensive use of the concept of ‘multiple group relationships’ to account for the variability of class conflict in a way that is not at all dissimilar from the way it is used, for example, by Williams. Surely it is in the active use of precisely such common concepts and propositions, rather than in procuring an agreed definition of ‘institution’ or ‘society’, that the desired unification of which Dahrendorf is so sceptical is constantly being achieved. In actual fact, the divergence between what he calls ‘integration theory’ and ‘coercion theory’ is much more evident in defining problems than in solving them.
Why1 then, the concentration on the development of alternative conceptual schemes in which the ideas of power and conflict play a central role? Partly because the recognition given by normative functionalism to the arguments put forward along these lines has so far amounted to nothing more than lip-service. More fundamentally, perhaps, it is because, in seeing equilibrium analysis combined in normative functionalism with a focus on shared value elements, Dahrendorf and Rex, with their manifest interest in social change, have as a consequence sought the key to this problem in the area of power and conflict. If this is so, how far do the conflict theorists take us in the analysis of social change?
Dahrendoff and Rex assert that social change is a result of the shifting balance of power between conflict groups.’ Now, while social change is very frequently associated with conflict, the reverse does not necessarily hold. Conflict maybe both endemic and intense in a social system without causing any basic structural change. Why does some conflict result in change while other conflict does not? Conflict theory would have to answer that this is decided by the variable factors affecting the power balance between groups. Here we reach the analytical limits of conflict theory. As a reaction to normative functionalism it is entirely confined to the problem of social integration what is missing is the system-integration focus of general functionalism, which, by contrast with normative functionalism, involves no prior commitment to the study of system stability.
// 405 This is exceedingly interesting, because both Dahrendorf and Rex arrive at their respective positions through a generalization of Marx. Yet it is precisely Marx who clearly differentiates social. and. system integration. The propensity to class antagonism (social integration aspect) is generally a function of the character of production relationships (e.g. possibilities of intra-class identification and communication) But the dynamics of class antagonisms are clearly related to the progressively growing ‘contradictions’ of the economic system. One might almost say that the conflict’ which in Mancian theory is decisive for change is not the power conflict arising from the relationships in the productive system, but the system conflict arising from ‘contradictions’ between ‘property institutions’ and the ‘forces of production’. Though definitely linked, these two aspects of integration are not only analytically separable, but also, because of the time element involved, factually distinguishable. Thus it is perfectly possible, according to this theory, to say that at any particular point of time a society has a high degree of social integration (e.g. relative absence of class conflict) and yet has a low degree of system integration (mounting excess productive capacity).
Further interest attaches to the fact that the idea of structural contradictions is central to the general functionalist view of change:
The key concept bridging the gap between statics and dynamics functional theory is that of strain, tension, contradiction, or discrepancy between the component elements of social and cultural structure. Such strains may be dysfunctional for the social system in its then existing form; they may also be instrumental in leading to changes in that system When social mechanisms for controlling them are operating effectively, these strains are kept within such bounds as to limit change of social structure.
The vital question is, of course: what are the ‘component elements’ of social systems which give rise to strain, tension, or contradiction? General functionalism, as I understand it, does not attempt to formulate an answer to this question.’
It is, by contrast, in normative functionalism that institutional patterns emerge as the only generally identified and systematically differentiated // 406 components of a social system between which there can be conflict and resultant strain. Since social systems are differentiated only along the institutional axis, there can be no place for the kind of contradictions which Marx envisaged, contradictions which are obviously relevant to the problem focus of conflict theory. We may ask, therefore, does the Marxian view contain the elements of a more general sociological formulation?
Criticism of the Marxian interpretation of society and social change has focused on the meaning and importance attributed to the ‘material mode of production’. Sometimes, this has been simply and erroneously interpreted as technology. Yet it is quite obvious that in the Marxian schema technological change is not regarded as the prime mover, but as a force which operates interdependently with the productive relations of the society, that is, the prevailing organization of property and labour. The inclusion of productive relationships in the concept ‘mode of production’ lays the theory open to the criticism that the degree of differentiation and independence of such relationships from other social structures in the same society varies very considerably; and that, in particular, the saliency of the economic system under capitalism is not at all characteristic of most historical societies, in which the mode of political organization heavily conditioned.. the structure and potential change of productive relationships. Marxian theory has not, for fairly obvious reasons, been overmuch concerned to rebut such criticism of its basic sociological assumptions. Given its premisses about the general long-run decisiveness of the economic order for social change, it has quite logically confined its discussion of system integration to the internal dynamics of the mode of production itself-to the economic theory of the contradiction between ‘forces of production’ (technological potential) and the ‘relations of production’ (property institutions).
While this narrowing down of the problem of system integration is highly questionable, the idea of a contradiction between the material conditions of production and the productive institutions of the economic system has a more general relevance that should not be ignored.
First, contradiction implies that the material means of production (e.g. industrial technology) favour a set of potential social relationships (social- //407 ist ownership) which constitutes a threat to the existing social relation-ships institutionalized in the property system (private ownership). Now, whatever reservations one may have about the specific linkage of industrial production with socialist property relationships, there is nothing metaphysical about the general notion of social relationships being somehow implicit in a given set of material conditions. Material conditions most obviously include the technological means of control over the physical and social environment and the skills associated with these means. They include not only the material means of production, but also what Weber frequently refers to as the material means of organization and violence. Such material conditions must surely be included as a variable in any calculus of system integration, since it is clear that they may facilitate the development of ‘deviant’ social relationships which run counter to the dominant institutional patterns of the system. Michels’s study of oligarchical tendencies is only the classic example.
Second, according to Marx, the actualization of these potential counter-relationships is determined by the success with which those with vested interests in the existing order are able to resolve the functional incompatibility between the material means of production and the property framework. In the capitalist case, this incompatibility arises from the
inability of private property institutions to accommodate the productive capacity of the industrial system. The focal point of strain is ‘overproduction’. The argument, of course, goes further than this. The theory of the ‘crisis mechanism’ not only postulates dysfunctionality but attempts
to demonstrate how the internal contradictions of the mode of production are endogenously intensified to the point of system breakdown by the inherent development of productive forces. This mechanism, most fully elaborated in the case of capitalist societies, is the conveyor belt which moves a society from one stage of its historical evolution to the next. But, in order to use the idea of a functional incompatibility between the dominant institutional order of a social system and its material base, it is not necessary to assume that the system must inevitably break down or that it must inevitably be succeeded by another system of a given type.
We now have a view of system integration, particularly relevant to conflict theory, which may be summed up as follows:
One generally conceivable source of tension and possible change in a social system is that which arises from a ‘lack of fit’ between its core institutional order and its material substructure.
The material substructure in such a case facilitates the development // 408 of social relationships which, if actualized, would directly threaten the existing institutional order.
The system will be characterized by a typical form of ‘strain’ arising from the functional incompatibility between its institutional order and material base.
The actualization of the latent social relationships of the system will depend on the success with which groups having vested interests in the maintenance of the institutional order are able to cope with the dysfunctional tendency of the system in the face of particular exigencies.
If these exigencies lead to an intensification of the functional incompatibility of the system, and if compensating measures by vested interest groups lead (unintentionally) to a further actualization of the potential social relationships of the system, a vicious circle of social disintegration and change of the institutional order is under way. If, on the other hand, compensating measures are effective, the institutional order will remain intact, but the focal point of strain will continue to be evident so tong as the functional incompatibility of the system persists.
These propositions do not limit the analysis of system integration to the productive system of a society. Nor do they imply a differentiation of types of societies primarily in terms of their modes of production. Such problems cannot be settled a priori. Consequently, the ‘dominant’ or ‘core’ institutional orders may vary from one type of society to another; and the identification of such institutional orders would seem to be first and foremost a way of defining what is meant by saying that a society has changed. There are, however, certain problems which arise when the concepts of ‘dominant’ institutional order and material base are applied to social systems. It may make sense to apply such a distinction to some particular subsystem of a society or to some particular type of corporate group; is it equally relevant, in the case of a society, to regard, for example, the productive system as a ‘material base’ from the point of view of the ‘dominant’ political system, even though the productive system manifestly includes institutional elements? In so far as the predominant concern is with the way in which the material preconditions of a certain type of political action are, or are not, to be found in a given economic order, there would appear to be good reason for answering this question in the affirmative. Such an answer would, of course, in no way // 409 prejudice the further explanation of how such a given economic order came about; the problem of the ‘causes of the type of system instability under consideration is, anyway, a quite separate issue. It should also be noted that the degree of institutional differentiation of economic and political structures varies very considerably. In cases where the relations of production and the relations of political power are not institutionally very distinct, and especially where the relations of production are institutionalized to a considerable extent around political goals, it would seem reasonable to regard the economic order much more directly as a ‘material base’ of the ‘dominant’ political institutions. A brief reference to Weber’s discussion of patrimonialism may serve to illustrate these points as well as the propositions previously advanced.
Although Weber S concept of patrimonialism, and especially that of patrimonial bureaucracy, refers. primarily to a type of political structure, it is clear from his remarks that this structure might well be regarded as the ‘core’ institutional order of the society and as a major point of reference for societal change. Moreover, Weber’s analysis of the material preconditions of bureaucratization clearly indicates the nature of the functional problems facing societies of the patrimonial bureaucratic type. These centre on the relationship between the institution of bureaucracy and the material substructure of a subsistence economy. After setting out the general rule that ‘A certain measure of a developed money economy is the normal precondition for the unchanged and continued existence, if not for the establishment, of pure bureaucratic administration,’ Weber goes on to note that historical cases of ‘distinctly developed and quantitatively large bureaucracies’ may be found which ‘to a very great extent, partly even predominantly, have rested upon compensation of the official in kind’. This he explains by arguing that, ‘even though the full development of a money economy is not an indispensable precondition for bureaucratization, bureaucracy as a permanent structure is knit to the one presupposition of a constant income for maintaining it’, and that ‘a stable system of taxation is the precondition for the permanent existence of bureaucratic administration’. But again: ‘For well-known and general reasons, only a fully developed money economy offers a secure basis for such a taxation system.’
The strategic functional problem, then, is one of maintaining a taxation system that can effectively meet the material needs of a bureaucracy in // 410 the context of a subsistence, or near-subsistence, economy. The centralizing goal of bureaucratic institutions is constantly liable to sabotage by the potential social relationship structure of the subsistence economy which favours the decentralization and ‘feudalization’ of power relationships. As Weber himself says: ‘According to all historical experience, without a money economy the bureaucratic structure can hardly avoid undergoing substantial internal changes, or, indeed, turning into another type of structure.’ The relationship between bureaucracy and taxation is a highly interdependent one. The efficiency of the bureaucracy depends upon the effectiveness of its taxation system; and the effectiveness of the taxation system depends on the efficiency of the bureaucratic apparatus. Thus, for whatever reason, any increase in the bureaucratic load or decrease in taxation capacity may generate a vicious circle of decentralization of power. Indeed, it might be argued that the ‘taxation’ crisis of patrimonial bureaucracy is essentially analogous to the ‘production’ crisis of capitalism. At any rate, the focal point of strain in this type of society is taxation capacity relative to bureaucratic needs.
This strategic functional problem sets the stage for the characteristic conflicts of interest that arise between the bureaucratic centre, the officialdom, landed magnates, and peasantry. The points of tension are those which represent an actualization of the potential for ‘feudalization’:
the tendency of officials to appropriate the economic and political resources of the office; the struggle of large landowners to gain immunity from taxation and/or usurp fiscal and political functions; and the local relationships of economic and political dependency into which the peasantry are forced in seeking protection against the tax burden of the bureaucratic centre. These centrifugal tendencies may be seen as both a cause and a consequence of the possible failure of mechanisms for maintaining effective taxation capacity and central control. The outcome of such struggles, and the success with which the functional problem is solved by the bureaucratic centre, is, of course, decided in each historical case by the particular circumstances facing the patrimonial bureaucracy. These may vary very considerably; but whether they make for stability or breakdown or bureaucratic institutions, all societies of this type may be studied from the point of view of their common contradiction.
Another example of a not too dissimilar kind is that of the functional tensions arising from the relationship between the totalitarian political system and the industrial economy of the Soviet Union. It is noteworthy in this connection that many who would deny the relevance of the idea of ‘internal contradictions’ to capitalist societies have only too readily exaggerated the incompatibility of industrialism and the institutions of a one-party state. Be this as it may, it would seem that the type of contradiction envisaged here is one which those having an interest in the dominant political institution have thus far successfully controlled, but which nevertheless is likely to remain as a focal point of strain and potential change. It arises from the tendency of an industrial mode of production to create latent interest groups of a class character. This tendency must be ‘dysfunctional’ for a totalitarian political system, one precondition of which is a ‘classless’ society, i.e. an absence of bases of potential social organization outside the party bureaucracy
Such a contradiction could manifest itself either by such latent interest groups striving for an autonomous corporate existence (which seems unlikely given the nature of party control) or by their subversion of the party organization from within. Of such groups, associated with industrialization, the least potentially threatening is that of worker opposition. Using Weber’s typology of class formation, worker protest hardly advanced beyond the stage of ‘mass reactions of a class character’ (labour turnover and so on) in the early phase of Soviet industrialization; and, while disruptive to the economy, it was not allowed to develop into a more politically dangerous ‘societal’ action. More of a threat from this point of view, however-and this is the element of truth in Burnham’s otherwise extravagant thesis of a ‘managerial revolution’-is the so-called ‘Soviet bourgeoisie’: the functionally important quasi-group of predominantly industrial bureaucrats which has emerged as a result of rapid industrialization.
The focal point of strain for the totalitarian political system is not simply that this latent class tends to develop vested interests in its position and privileges, but that it has an organizational capacity and cohesiveness that could form the basis of a political opposition. And, given the nature // 412 of the political system, such an interest group would be most likely to take the form initially of cliques within the party bureaucracy. Therefore, the strategic functional problem of the dominant institutional order, from this point of view, is that of maintaining the control of the party bureaucracy over the industrial bureaucracy, and more especially of securing the party against infiltration by vested interest groups of the managerial elite (which includes insulating the latter from any wider support in the society). Most fundamentally the party must develop means by which it can systematically ‘dc-classify’ the lines of stratification and interest-group formation that have their basis in the industrial substructure. At the same time, however (and here arises the point of system tension), such declassification must not undermine the conditions of industrial efficiency.
The foregoing examples have been all too sketchy, but perhaps they may serve the purpose of illustrating the viewpoint advanced in the main body of the chapter. It has not been the intention to claim that this perspective is the only possible way to approach the problem of social change, still less to imply that there is anything other than a polemical advantage to be gained by focusing on system integration as opposed to social integration. What has been suggested, however, may be summed up as follows:
The propensity to social change arising from the functional incompatibility between an institutional order and its material base has been ignored by normative functionalists because of their concentration on the moral aspects of social integration.
It has been equally ignored by conflict theorists, who, in concentrating on the weakness of the normative functionalist approach to social integration, have failed to relate their interest in social change to the problem of system integration.
 K. Davis, ‘The Myth of Functional Analysis as a Special Method in Sociology and Anthropology’, American Sociological Review, 24(1959).
 A W. Gouldner, Reciprocity and Autonomy in Functional Theory’, in L. Gross (ed~), Symposium on Sociological Theory (New York, ‘959).
 Gouldner quite properly points out that this tendency has amounted to what is in fact ‘implicit factor-theorizing’: ‘Although the methodological position of the earlier functionalists commonly affirmed an amorphous interdependence of parts within a social system, it does not follow that the specific empirical analysis in which they engaged actually utilized this principle. In particular, the classic contributions, from Comte to Parsons, have gone out Of their way to stress the significance of ‘shared. value elements’ in maintaining the equilibrium of social systems.’ (‘Reciprocity and Autonomy in Punctional Theory’, 265.)
 Merton, Social Theory and Social Structure, p. [22.
 Dahrendorf, C/ass and Class Conflict in Industrial Society, and Rex, Key Problems of Sociological Theory
 Ibid. 112
 Briefly, to define authority as institutionalized power is to beg exactly the question that Merton raises, if the line between authority and power is drawn in terms of the presence or absence of a claim to legitimacy, not in terms of the sentiments of those (principally) over whom authority is exercised. Perhaps the most general consideration which makes the deinstitutionalization’ of authority an ever-present possibility is the fact that, whereas the legitimacy of authority tends to take the form of general principles, acts of authority are always specific; and they are always more specific than derived rules of authority, no matter how well developed the latter Thus, the ‘exploitable’ ambiguity surrounding the derivation and interpretation of the legitimacy of specific acts means that authority is never given, but is always contingent upon its exercise. It is precisely with such conflicts arising within the interstices of institutionalized power that ‘conflict theory’ is concerned; and not simply with the more unusual approximations to ‘unstructured’ power conflicts.
 Dahrendorf, Class and Class Conflict, p.172.
 T. H. Green, Principles of Political Obligation (London, 1906).
 At any rate, in formal terms. For instance, Parsons: ‘I do not think it is useful to postulate a deep dichotomy between theories which give importance to beliefs and values on the one hand, to allegedly “realistic” interests, e.g., economic, on the other. Beliefs and values are actualized, partially and imperfectly, in realistic situations of social interaction and the outcomes are always codetermined by the values and realistic exigencies; conversely what on concrete levels are called “interests” are by no means independent of the values which have been institutionalized in the relevant groups.’ (Parsons, Structure and Process in Modern Society, p. i73~) See also Dahrendorf, Clavs and Class Conflict, pp.159, 163, and Rex, Key Problems of Sociological Theory, p.112 But, while there is formal agreement on this point, both the normative functionalists and the conflict theorists fail to explore in any rigorous way the interrelationships of ‘normative’ and ‘realistic’ elements of social systems.
 Both authors state their propositions in summary form (Dahrendorf, Class and Class Conflict, pp. 236-40; Rex, Key Problems in Sociological Theory, pp.129-31,195). Their premises are very similar: ‘Every society displays at every point dissensus and conflict; social conflict is ubiquitous’ (Dahrendorf, p. 162); ‘Instead of being organized around a consensus of values, social systems may be thought of as involving conflict situations at central points’ (Rex, p. 129). The major disagreement between the two would seem to be how far, in fact, lines of social conflict overlap (see Rex, pp. 117-18).
 Dahrendorf, Class and Class Conflict, pp. 161-4; Rex, Key Problems in
sociological Theory, pp.112, 114.
 Dahrendorf, Class and Class Conflict, p. 164.
 To take an actual example, compare the explicit use of the idea of the ‘exploitability’ of the common value system by Parsons (The Social System, pp. 293, 355) in accounting for the intensification of ‘deviance’ with the implicit reference to such an idea by Rex (Key Problems in Sociological Theory, p.125) in discussing class conflict.
 Dahrendorf, Class and Class Conflict, pp 213-IS, and Williams, American Society, pp. 560-I.
 Dahrendorf, Class and Class Conflict, pp 231-6; Rex, Key Problems of Sociological Theory, p. 196.
 1 may refer here once more to the excellent essay by Gouldner (‘Reciprocity and Autonomy in Functional Theory’), and especially to his idea of the ‘functional autonomy’ of parts This concept provides an obvious link between social and system integration He explicitly points out that ‘the concept of the differential functional autonomy of parts directs attention to the need to distinguish between parts having a greater or lesser vested interest in system maintenance’, and that ‘not only efforts to change the system, but also those directed at maintaining it are likely to entail conflict and resistance as a result of differential functional autonomy. What I find a little ambiguous, however, is his use of the term ‘parts’ of a system: at one stage they seem to mean structural aspects (e.g. ecological conditions); at another, actual groups (the French bourgeoisie) The ‘parts’ which may become functionally autonomous are surely groups; the ‘parts’ whose interplay conditions their functional autonomy are the structural elements of the system. I hope this will become clear in the subsequent argument.
 Merton, Social Theory and Social Structure, p.122.
 Gouldner, Reciprocity and Antonomy in Functional Theory’, pp.244-8
 especially, M. Weber, Economy and Society, 1044-51.
 See, e.g., P. A. Baran, The Political Economy of Growth (New York, 1957) and Sweezy, The Theory of Capitalist Development. For the difficulty of locating the ‘crisis mechanism’ of feudalism, see M. H. Dobb (ed.), The Transinon from Feudalism to Capitalism: A Symposium (Patna, India, ‘957).
 See the instructive remarks of R. Coulborn, Fcudalism in History (Princeton, NJ, ‘956), 254~69
 Thus differences of opinion about the endurance of Western feudal society depend very largely on whether the military, the political, or the economic aspect of this institutional complex is singled out as the ‘core’ order; see 0. Hintze, ‘Wesen und Verbreitung des Feudalismus5, Salat und Verfassung (Leipzig, 1941).
 What else does Weber imply when he writes: ‘It is clear, therefore, that the disintegration at the Roman Empire was the inevitable political consequence of a basic economic development: the gradual disappearance of commerce and the expansion of a barter economy. Essentially this disintegration simply meant that the monetarized administrative system and political superstructure of the Empire disappeared, for they were no longer adapted to the infrastructure of a natural economy.’ The Agrarian Sociology of Ancient Civilizations, trans. R. I. Frank (London, 1976), 408.
 From Max Weber, pp. 205-9
 The logic of this is succinctly argued by Rioch, Feudal Society, p.68, and L. M. Hanman, The Early Medicval State (London, 1960)19.
 From Max Wcber, p.205.
 On the particular conditions favouring the stability of patrimonial bureaucracy in Egypt and China, see Weber, Economy and Society, ii, ch. 12. The most famous instance of breakdown, that of the later Roman Empire, is a case where the ‘defence mechanisms’ introduced by the bureaucracy (aptly described by Lot as the ‘regime of castes’) intensified the trend towards subsistence economy and actualized the potential for ‘feudal’ relationships. See Weber, The Agrarian Sociology ofAncient Civilizations, Part JV; F. Lot, La Fin du Monde Antique et le debut da Moycn Age (Paris, i~si); M. Bloch, ‘The Rise of Dependent Cultivation and Seignorial Jnstitutions’, and 0. Ostrogorsky, ‘Agrarian Conditions in the Byzantine Empirc in the Middle Ages’ (both in S. H Clapham and E. Power (eds.), The Cambridge Economic History, i (Cambridge, 1942)). The general problem of ‘feudalizing’ tendencies in patrimonial bureaucratic societies is discussed in R. Coulborn, Feudalism in History (Princeton, NJ, 1956), and S, N. Eisenstadt, ‘Political Struggle in Bureaucratic Societies’, World Politics, 9 (1956).
 Feldmesser, ‘Equality and Inequality under Khrushchev’.