Electronic Journal of Sociology (2000)

ISSN: 1198 3655

Timeless Moral Imperatives in Causal Analysis of Social Functioning

Emanuel Smikun
American Social Indicators

The neofunctionalist movement in sociology adopted functionalism as a tradition rather than a method (Alexander 1985). Its general discourse and research programs are mainly concerned with rehabilitating Parsons, further elaborating and proliferating his ideas and conceptual schemes, and, more recently, with a revision and reconstruction of Parsons’ intellectual legacy (Alexander and Colomy 1990a; Colomy 1990, pp. xi-xli; Alexander 1998). Its epistemology locates social-scientific thought in many relatively autonomous dimensions - from metaphysical to empirical environments - but without a guiding method. Sociological method is replaced with “methodological and meta-methodological assumptions and techniques” referring primarily to the philosophical alternatives of idealism vs. materialism or positivism, and their possible synthesis (Alexander 1982, 1983). The slight of specifically functionalist methodology is based on a general dismissal of Merton’s contribution to it as well as on the fact that Parsons preferred “action theory” as a name for his sociology (Alexander 1985). Yet, the concepts of functions and functional equivalents are still kept alive, including manifest and latent functions, except that now their use comes with disclaimers and quotation marks (see Alexander and Colomy 1990b, pp. 331, 361n1, and passim). This creates an impression of a need to apologize as if these terms were something to be ashamed of. On the other hand, to my knowledge, the charge that neofunctionalism is “not really functional” so long as it does not address the question of functional imperatives (Turner and Maryanski 1988) has never been answered.

Following Alexander’s theoretical logic in sociology, many substantive contributors to neofunctionalist literature show high sensitivity to multidimentional epistemology, but methodological questions of functional analysis are simply ignored. In Stephen Fuchs’ contribution to sociology of science featured in the neofunctionalist showcase collection (Colomy 1990, pp. 224-242), institutionalized standards of scientific truth in the form of logic and method are replaced with local-situational criteria of academic reputation and power within elitist networks of professional knowledge producers. Method is seen as an otherwise useless embellishment invoked pro forma, for presentation purposes only. In principle, anything goes in this pedestrian vision of science so long as convenient authorities are cited, and reported results are approved by those in position of power at the moment (Fuchs 1986). One can only wonder what makes Fuchs’ essentially ethnomethodological writing functionalist other than citations from Luhmann. As for Luhmann, his treatment of functional analysis is hardly relevant here since he defines it epistemologically - as a comparative study of functional equivalents in solving problems of science in general. No sociological specification is offered, the only context being Luhmann’s general abstract ideas of systems differentiation (1995, pp. 52-58).

My task here is to return to sociological functionalism as a method, and to suggest that it has important redeeming values, that sociological structural-functional analysis can be improved and used without shame and equivocation as a solid foundation of fruitful empirical research. A reexamination of Merton’s paradigm of functional analysis provides a clue to the difference between instrumental-naturalistic and value-rational functional analysis. Within the latter mode, I highlight the importance of moral functional imperatives as responsible for all social structural change. Coupled with a distinction between action and behavior, this makes it possible to redefine structural-functional method as a macrosociological method of causal analysis of moral (dys)functioning, and to find a way to avoid teleological explanations. Systemic conceptual schemes are discussed as relevant for causal explanation of social functioning and development. A connection is established between Parsons’ unfulfilled idea of dynamic structural-functional analysis and modern statistical causal modeling. The question of causal order is explored in the context of theories of developmental-evolutionary change with a special reference to neofunctionalist theories of social differentiation.

Functional Analysis: Cutting the Umbilical Cord

Merton prefaced his paradigm of functional analysis with a detailed critique of the ways this method was practiced in the social sciences of his day, notably in the cultural anthropology of Radcliffe-Brown and Malinowski. When examining what he called the postulate of indispensability, Merton distinguished between two incompatible positions. On the one hand, there were assertions by social analysts that each of the existing social institutions and social structures in general were irreplaceable for the fulfillment of a variety of social functions, and that such functions could alternate and substitute for one another - hence the notion of functional alternatives, substitutes, and equivalents. Along with that, there were assertions or assumptions of the indispensability of certain crucial social functions - called functional prerequisites or imperatives - whereas the social structures through which they were fulfilled could vary and change ([1949]1967, pp. 86-90). Merton’s subsequent discussion shows, however, that he did not fully appreciate the implications and the power of this distinction.

Unlike cultural anthropology up to the 1950’s that studied preliterate societies having no written history and therefore no traceable social development, sociology studies developed societies with their rich historic backgrounds. No one can deny the tremendous institutional and behavioral structural changes that Western societies have undergone throughout their history, especially during the last three centuries. If sociohistoric continuity is not located in ever-changing social structures, then it must be in certain enduring moral values capable of creating and maintaining social bonds. Without assuming such values as imperatives of macrosocial functioning, it is impossible to account for the continuity of Western societies as cohesive units of human self-organization despite their sometimes radical structural changes. If we do assume the existence of timeless moral functional imperatives, then structural changes become necessary adaptations of social institutions and behavior patterns to better fulfill the moral imperatives of any normal social functioning. Macrosocial functional analysis must postulate such moral imperatives first, demonstrate their persistence across as well as within sociohistoric epochs, and only then examine various and variable institutional, behavioral, and concomitant spacial and temporal social structures through which the functional imperatives have (or have not) been fulfilled. This can be done by identifying typical phases of social functioning as well as the structural constraints that cause social dysfunctioning. With that, we must take the notion of functional equivalents, substitutes, alternatives to mean alternative structural arrangements fulfilling the same functional imperatives.1

There are fundamental reasons why the two methodological alternatives in Merton’s postulate of indispensability are not of equal merit, and why social scientists should not falter in choosing the right one in their assumptions. The notion of social functioning and dysfunctioning has referents in the concrete phenomenal reality of everyday successes and problems as reflected in hourly news broadcasts. By contrast, we can speak of social structures only by abstracting from the processes of change in which they are always embedded, whether in diachronic social development or in synchronic social functioning, as it is well demonstrated by recent sociological studies of economic institutions (Granovetter 1985, Zelizer 1985, Zukin and DiMaggio 1990, Fligstein 1996). When social scientists try to impute functions to an element of social structure, they unlawfully hypostatize this abstract analytical idea. The rationale of determining the contributions institutions or behavior patterns make to society’s survival is false. The term function may well be applicable to the natural world of living organisms and species whose survival may be problematic and a legitimate subject-matter of a detached naturalistic study. By contrast, the idea of all social institutions or behavior patterns performing some “useful functions” can only serve as an instrument of manipulative or exploitative, i.e., asocial causes. The classic paradox of crime considered “functional” is an extreme and obvious example that belies the general value-neutral logic of seeing all existing social structures as somehow useful. Following this logic, Merton ([1949]1967, pp. 126-136) found a number of “latent positive functions” in political machines. What better example of consistent findings of naturalistic functional analysis running against the moral foundations of American democracy?2

The content of moral imperatives may change historically and across cultures, yet they always remain the most essential ingredients of human society and civilization. The notion of moral imperatives may sound idealistic only because of their peculiar existence. Indeed, when we recognize and uphold collective moral ideals as values, and make them a source of inspiration, a guiding light of our lives, when we live lives meaningful in their terms, nothing special happens to us or to the ideals and values themselves. The imperatives and values continue to live within us to the extent that we appropriate them, i.e., make them our second nature prompting us to good deeds of which we may often be simply unaware. This is just normal behavior. But, as Durkheim suggested ([1902-1903]1961, pp. 58-94; [1924]1953, pp. 45, 54-55, 59, 94), when we forgo them, when our conduct is guided by egotistic self-interest or individual desires and appetites, we easily become disillusioned, frustrated, and, in extreme cases, depressed and despondent even to the point of suicide. Our failure to adhere to genuine moral ideals and values has grave consequences for them, too. Since moral values, beliefs, and sentiments of which our civilization is composed are realized through the acts of our daily behavior and through fostering them in our children, they can never be taken for granted.

Before Parsons presented AGIL as four universal functional requisites, he included in such list elements of natural nonhuman environment, human biological substratum, as well as the personality, the social, and the cultural “subsystems of action” (1951, pp.26-36; cf. Aberle et al.1950). In addition, Parsons spoke about “imperatives of compatibility” among various behavioral and institutional structural elements, such as value-orientations, occupational and kinship roles, economic and power structures (1951, p.167). It is perfectly clear that Parsons’ functional imperatives of compatibility had the same meaning as Merton’s structural constraints. This confusion of functional imperatives, or prerequisites, with the structures they operate through was multiplied in a parallel notion of “structural prerequisites” synonymous with imperatives of compatibility (Parsons 1951, pp.177-180). It is ironic that Parsons’ false rationalizations of functional imperatives came as his early insistence on ultimate moral values as central to social order - the true imperatives - began to be diluted and was eventually completely lost in these categories.

Sociology cannot accept imputation of ad hoc social functions to all given social structures as its method, nor the instrumental logic behind it. None of this means, however, that functional analysis should be abandoned. By incorporating the alternative, by recognizing the reality of moral functional imperatives as responsible for structural change we obtain a specifically social science having its own method different from the methods of the natural sciences. Instead of useful social functions, we can speak of social functioning in the sense of properly working social systems where moral imperatives materialize in appropriately changing social structures. Societies are sustained by timeless moral imperatives, not by immutable social institutions. Macrosocial functioning in this sense can and should be studied with a view of discovering social structural constrains and finding better substitutes for them.

Despite his critique of Radcliffe-Brown and Malinowski, Merton was unable to cut the umbilical cord connecting sociological functional analysis to its origins. For all his attempts to distance himself from anthropological functionalism, Merton was still in its grips. This is evident from his thesis of the neutrality of functional analysis with respect to its basic assumption of conservatism versus social change ([1949]1967, p. 92), from his adopting the naturalistic version of functionalism when speaking about it generally (pp. 100-102), and from his use of a naturalistic physiological model for sociological functional method in particular (pp. 102-103). Just like the anthropologists, Merton advised sociologists to begin with “the items to which functions are imputed,” and then proceed to the objective consequences of those items, i.e., their functions. At the same time, he listed functional imperatives - as the fifth of his ten steps - and recommended validating them in comparative cross-cultural rather than sociohistoric analysis. Merton continued to discuss functional methodology in terms of field observation, field notes, magic performances and rites like the Hopi rain dance, Chiricahua puberty ceremonials, and the Australian aborigines’ responses to obscenities - all the paraphernalia of cultural anthropology.

Equally notable is the fact that Merton included social dynamics in his paradigm almost as an afterthought, as something desirable rather than an indispensable precondition of sociological functional analysis, which it is. Nowhere do we find among Merton’s refined methodological observations a statement to the effect that social functioning and functional imperatives make sense only in the context of historic social development - and vice versa. Merton’s codification of functional analysis did little to discourage vulgar teleology in explanations of synchronic social functioning of the type, “The functions of incarceration are to deter crime and rehabilitate offenders.” Such functional explanations corrupted, in turn, explanations of origin (“The modern nuclear family evolved from the extended family to better provide for parents’ personal fulfillment rather than secure private property”). Merton himself contributed to functionalist teleology in his early theory of anomie (1938) where he used the term “cultural goals” for macrosocial needs or interests, and “legitimate or illegitimate means” of achieving them, for macrosocial practices. In “Manifest and Latent Functions,” he appears to have drawn the distinction between action and behavior when he wrote,

Social function refers to observable objective consequences, and not to subjective dispositions (aims, motives, purposes). And the failure to distinguish between the objective sociological consequences and the subjective dispositions inevitably leads to confusion of functional analysis [...] ([1949]1967, p.78. Emphasis in the original.)

But then, only twenty-four pages later, Merton listed “concepts of subjective dispositions (motives, purposes)” as the second step in his paradigm of functional analysis. On one early occasion, Parsons (1954, pp. 217-218) did concede functional teleology as positive contributions that elements of social structures make to the system as a whole. However, there was no teleological reasoning involved in the macrosocial mechanisms of dynamic equilibrium centered on normative values that he later postulated. If he appeared to be saying that social systems or their elements have goals, it was the unnecessary and avoidable microsociological language of social action that was betraying him. Indeed, it is difficult to accept the language of action applied to what is intuitively perceived and rationally understood as impersonal (macrosocial) functioning. The basic assumption of the action frame of reference is directly relevant to the vision of functional analysis, and to Merton’s central distinction between manifest and latent functions in particular.

Action and Behavior

Parsons’ greatest achievement was in introducing systemic models in sociology. Such models do not have to be necessarily functionalist. While empirically-grounded sociologies such as interpretive symbolic interactionism and ethnomethodology continue to rely heavily on dichotomous ideal types and non-model conceptual language indistinguishable from substantive sociological notions, the need for systemic models in sociology is as great today as ever. As sociology was only in its inception in Marx’s and Weber’s time, the elements of its conceptual language had innovative empirical meanings while also having general social-scientific meanings at the same time. Those meanings were not yet touched by shifts and expansions due to other social scientists’ repeated use of the same words and expressions. Both Marx and Weber made their sociological discoveries by introducing new concepts and deductively drawing implications for a broad variety of empirical fields using their encyclopedic knowledge of world history as well as world cultures. Their new seminal concepts, such as class struggle and surplus-value, or legitimate domination and monocratic bureaucracy, allowed them to reinterpret old concepts of bourgeois political economy or political, economic, legal, and religious social history. The meanings of these original concepts themselves were not problematic. They were considered self-evident, and an unambiguous, monosemantic understanding of these terms was taken for granted. The situation was quite different for Parsons.

By the time Parsons had come onto the sociological scene, the concepts of system, structure, function, rationality, action, socialization, social institution, social role, social class, social status, social control, and others had already lost their novelty. They had already been used by several authors and radiated different, at times inconsistent meanings depending on the context and the author with whom they were associated. Most of the conceptual terms Parsons used appeared to be the same as those used by his predecessors and contemporaries - Pareto, Weber, Durkheim, Radcliffe-Brown, Malinowski, Freud, Thomas, and Park. While the denotations of these concepts were identical, they had already accumulated additional significations and connotations, and in some cases their meanings had been transferred to new semantic domains. There was no choice but to engage in a clarification and ordering of the sprawling meanings of all these concepts. And that is what Parsons’ intellectual career was devoted to. Following Whitehead’s and Pareto’s lead, Parsons insisted on separating the tool of general conceptual organization - the ontological frame of reference - from particular sociological theories it was used in.

The simple fact is, however, that the institutional domains of economy, polity, fiduciary culture, and societal community cannot be interpreted in terms of social action as Parsons claimed. Even if Parsons’ personality, culture, and social systems are interpreted as the internal environment of action (Alexander 1998, pp. 211-233; cf. Lidz 1981), that still does not make them action-like.3 Not only do these categories lack any content of Weber’s idea of social action, but their systemic, functional meanings are far from being clear. In his early writings Parsons ([1935]1991, [1937]1968) argued that moral values as ultimate ends of social action provided a clue to understanding the institutions and processes of developed Western economies and to overcoming the difficulties experienced by the social sciences of his time. By the end of his intellectual career, action was not necessarily social, it was associated with functions of “all living systems.” If the problems inherent in Parsons’ project have to be reduced to a central one, it may be said that he tried to achieve something impossible. He attempted to use in macrosociology what ostensibly was a microsociological concept of voluntaristic social action. From the beginning of his career, Parsons was following Weber’s strategy of using categories of social action as a frame of reference for macrosociological concept formation, notably the concepts of institutional structural relations and social evolution. However, Weber’s concept of social action was treacherous.

All social action is characterized by a subjective planning of ends and means, and by a past-in-the-future consciousness of acts contemplated in such plans. Social behavior, on the other hand, is characterized by objectively existing results of acts already accomplished and by a future-in-the-past consciousness of such accomplishments. The unique intended sense or, simply, intention of social action can be intersubjectively understood by others, including microsocial scientists. The meaning of accomplished overt acts constituting social behavior can be interpreted and reinterpreted from various perspectives, and their likelihood can be estimated in macrosocial research from aggregated observational data as statistical probability. Weber packed all this together under the name of social action (cf. Schutz [1932]1967).

A compensation for the lack of a clear conceptual distinction between action and behavior in Weber’s work came in his theme of value-neutrality based on the separation of (supposedly objective) facts and (supposedly subjective) value-judgements ([1904]1949). In his earlier conceptual statements, Weber did occasionally differentiate between subjective and overt processes, and between actor’s subjective motivations and expectations of objective and rule-governed behavior of others ([1906]1949, [1913]1981). However, his subsequent elaborations and changes of terminology, especially in Economy and Society (Weber [1922]1978), proved more confusing than clarifying. In particular, the distinction between (objective) causal adequacy, or adequate causation, on the one hand, and (subjective) adequacy with respect to meaning, or chance causation, on the other, clashed with the meaning of the likelihood of overt action (actually, behavior) as objectively given. In his actual world-historic reconstructions, Weber used only the macro-institutional-behavioral aspect of what he called social action - as cultural, economic, and political orientations of the affectual, traditional, or rational type buttressed by an even stronger reliance on (doubly-objective) structural factors of social status. The confusion in Weber’s sociological program does not directly show in his substantive historic-typological writings. You do not notice it unless you compare the word and deed of his methodology. Indeed, it was fortunate for sociology that “Weber did not in practice fully follow his own methodological precepts” (Fulbrook 1978, p. 81; see also Udehn 1981, Turner 1983, Smikun 1990). By contrast, Parsons took Weber’s programmatic idea of social action literally, and applied it in its full scope. No wonder his project was torn between his system and action concerns since he incorporated Weber’s error.

Parsons repeatedly complained of his difficulties in translating Weber ([1922]1978, pp. 57n3, 59n12, 59n13, 59n16, 60n20, 61n31) into English. Indeed, Weber’s conceptual ambiguity shows in his translation of Weber’s key terms signifying the difference between action and behavior, notably, Sinn as “meaning”(better translation: intended sense), but also of verstehen as “interpretation” and “interpretive” (better translation: understanding as extraction of intended senses), and especially of Deutung and deuten as “interpretive understanding” (better translation: interpretation as attribution of meaning to extracted intended senses). The expression “interpretive understanding” coined by Parsons (see Weber [1922]1978, pp. 3-4, 57n3) is another contradiction in terms. What can be understood in social action exists only as an intersubjective amalgamation of the senses intended by acting subjects. The senses mentally intended by actors and tentatively-hypothetically understood by others cannot be interpreted unless and until they are objectified in an external fixity, e.g., verbalized in speech, and become social facts. By the same token, social action becomes social behavior. Similarly, the senses of goals, projects, designs, strategies, or schemes subjectively planned by complex organizations and less formal collective actors crystallize in the meanings of objectively existing impersonal social institutions. Once they are understood and objectified, the intended senses can be interpreted in various ways. Interpretation begins where hypothetical understanding ends. Interpretation is assertive articulation of new meanings different from intended senses about whose correct understanding one can never be completely sure.4

What Parsons called action frame of reference included both macro-institutional-behavioral, and micro-voluntaristic-action elements at the same time. Hence such self-contradictory terms as “orientation of action,” “teleological orientation,” “actor’s behavior,” or “purposive behavior” that mixed together attributes of subjectively intended purposeful action and descriptions of objectively observed social behavior. While normative moral values had an impersonal, macrosocial meaning in Parsons’ usage, his attempts to attach a similar meaning to ends and means were not fruitful.5 Parsons’ blending of macro and micro-levels of social analysis in an all-embracing, pansociological perspective proved problematic. This nineteenth-century idea that Parsons himself summarized in The Structure of Social Action ([1937]1968) has been rejected by the combined efforts of symbolic interactionists, phenomenological sociologists, and ethnomethodologists essentially because of the violence it does to the notion of social action.

Merton distinguished between manifest and latent social functions having in mind the difference between functions intended and recognized by the actor and those neither intended nor recognized ([1949]1967). This was the way in which Merton perceived the difference between action and behavior. However, subjective microsocial intentions cannot fulfill macrosocial functional imperatives. The road to hell is paved with the best of them, as the old proverb goes. It makes no sense to speak of “manifest functions” in macrosociological functional analysis. The idea of “latent functions,” or unintended consequences of social action, does not fare any better. Unintended consequences of action are nothing other than constituents of social behavior. In their aggregated form, they fall under the category of social structures. Thus, the expression “manifest functions” is internally self-contradictory, and “latent functions” turn out to be elements of social structures. Neither notion can be sustained.

While single acts of social behavior can be seen as unintended consequences of specific social action, functional analysis does not involve itself with this microsocial aspect of their production, but rather with their macrosocial aggregates. Aggregate data on social behavior in various institutional spheres are not consequences of anyone’s intentional action, but impersonal social structures, and for that reason, legitimate - and necessary - objects of macrosocial analysis. It was this admixture of the language of social action as well as the naturalistic methodology of attributing ad hoc functions to existing social institutions that made functional explanations circular and slipping into bad teleology. Successful sociological functional analysis will not be guaranteed by abandoning the inadmissible practice of treating social structures as enduring and as the primary objects of study. We must also exclude social action from it. This is another impediment to overcome. Being a purely macrosociological method, structural-functional analysis must leave social action to (analytically separate) microsociology, such as ethnomethodology, including all action’s aspects and extensions - ends and means, goal-attainment, intelligibility, or accountability.

Of course, both action and structures of social behavior are analytical abstractions. In its concrete manifestations, all social action is really interaction, whether as role-taking or as negotiated order, and all behavioral structures are aspects of social functioning and development. Their respective levels of social analysis - microanalysis of social action and macroanalysis of behavioral and institutional social structures - are not fully independent epistemological platforms. Studies of concrete social reality take place on intermediate levels where micro and macro perspectives interpenetrate. By analogy with the terms of lower-middle and upper-middle class, we can speak of the micro-meso level of competitive interactive reinterpretations and institutionalization, as well as of the macro-meso level of behavior in public places and negotiated order. The terms of micro, macro, micro-meso, or macro-meso signify partial, holistic, and combined partial-to-holistic (inside-out) or holistic-to-partial (outside-in) perspectives. No relationship with the size of units of analysis can be presumed for them. While a nation-state can be seen as interacting with other states on the micro and micro-meso levels, an organization can be a subject of macro or macro-meso study as an functioning social system adapting to its environments. The adoption of macro-meso and micro-meso perspectives in social theory and research is a promising way to resolve the continuing difficulties of combining system and action theories directly (cf. Joas 1996, Schwinn 1998).

Systemic Causal Modeling in Functional Analysis

Good, value-rational structural-functional analysis proceeding from moral imperatives of normal social functioning can produce retrospective explanations of origin as well as explanations of present functioning and dysfunctioning without falling into the teleological fallacy. For example, the parliament may have been an answer to the complexities of feudal land and serf tenure in late medieval Europe, of multilevel vassal obligations to the king, a king’s arbitrary rule, and overabundance of lawyers. Institutions of child care (from orphanage to adoption) may have developed as a historic answer to the problem of child neglect, urban poverty, capitalist exploitation of women’s labor, lack of family planning, and so on. Similar causal explanations can be found for the origins of new types of social behavior within institutional domains. Alternative lifestyles of the 1960s and 1970s may have developed as an answer to the problem of status inconsistency as well as differential needs and interests of the baby-boom generation. Consumer or voter expectations and preferences may have been sociohistoric answers to the problem of beliefs and opinions about deceptive advertising, overflow of cognitive psychologists in academe and their employment in business. All these are legitimate macrosocial explanations of origin where the teleological language of microsocial action, like purpose or intent, is not appropriate. These are explanations of past development from our point of view of today. They cannot be attributed to anyone’s good or ill intentions in the past.6

In addition to having their evolutionary-developmental origins, institutional and behavioral social structures exist in the present time. Explanations of one set of social phenomena with another such set in the present proved problematic in old functionalism. This is synchronic, cross-sectional functional analysis proper. An empiricist version of synchronic functional analysis has been practiced for some time in sociology but without any connection with functionalist theory and methodology. We know it as causal or structural equation modeling. Remarkably, Parsons’ idea of dynamic structural-functional analysis was not too far removed from statistical causal modeling except that he became immersed in its qualitative taxonomic-conceptual aspect, developing a system of categories for motivational functions and those of social control. Writing half a century ago, Parsons occasionally sounded very much like today’s causal model builders:

[The] essential feature of dynamic analysis in the fullest sense is the treatment of a body of interdependent phenomena simultaneously, in the mathematical sense. [...] The ideal solution is the possession of a logically complete system of dynamic generalizations which can state all the elements of reciprocal interdependence between all the variables of the system (1954, pp. 215-216, emphasis in the original).
The essential question is how far the state of theory is developed to the point of permitting deductive transitions from one aspect or state of the system to another, so that it is possible to say that if the facts in A sector are W and X, those in B sector must be Y and Z (1951, p. 21).

Parsons did not believe an adequate mathematical apparatus existed for such research procedures, but his initial program envisaged systems of simultaneous equations and tests of (functional) significance (1954, pp. 215-218; 1951, pp. 21-22). Apparently, he did not know that Sewell Wright had proposed the idea of path analysis already in the 1920’s (see Blalock 1971, Part II). If we do make a connection between statistical techniques of causal modeling and Parsons’ unfulfilled idea of dynamic structural-functional analysis, and if we place both within the context of moral functional imperatives as accounting for synchronic as well as diachronic dynamics of ever-changing social structures, we can acquire a seamless and natural unity between general value-rational social theory and advanced quantitative empirical research, i.e., between what was once dismissed as useless grand theory and unguided abstract empiricism (Mills 1959).7

Parsons’ taxonomic technique was closely related to his understanding of system and system functioning as a holistic unity of interrelated structural elements in a dynamic process of restoring equilibrium, or self-organization. Parsons conceived of a prototypical social system as comprising four elements of which two were products of a bilateral interpenetration between the other two that were pure opposites. This is the way the AGIL categories were initially justified - by subsuming pure Gemeinschaft and pure Gesellschaft pattern-variables of orientation under Adaptation and Integration, respectively, and their mutual interpenetrations, under Goal-attainment and Latent tension-management. Parsons conceived the AGIL categories as four paradigmatic social functions, but also as phases of a dynamic procedure of social functioning. While he postulated that none of the AGIL elements could be reduced to any other as orthogonal dimensions of a four-dimensional Euclidian space (Parsons, Bales, and Shils 1953, pp. 63-109), he distinguished between their earlier and later functional phase movements, and spoke about their (recursive) sequence in the direction of motivational energy flow toward Goal-attainment. Parsons also spoke of the reverse (non-recursive) symbolic-controlling movement from Goal-attainment to Latent pattern-management, for example, in performance and learning. In the evolutionary scheme (1966, pp. 16-19), the energizing strength came from Adaptation while the controlling strength came from Latent pattern-maintenance. In mid-1950’s, Parsons (and Bales 1955) reconceptualized interpenetration as external relations among subsystems, i.e., between system and environment, notably, personality, cultural, and social. He introduced the notion of boundary zones of interchange among autonomous subsystems essentially equating the constructive principle of interpenetration with exchange.8

In using this vision of a dynamic, functioning social system to produce models in various substantive fields - from family and socialization to symbolic media of interchange - Parsons made systemic model-building a norm that sociology cannot afford to abandon. We may rightfully disagree with, and discard the specific contents of Parsons’ models along with his action-based AGIL categories themselves as a proven dead end (see Historic note below), but there is no reason why they cannot be replaced with other multilevel systemic categories of normal social functioning. Such models are systemic in the sense that the meanings of all their elements are mutually interrelated as parts of a linguistic and visual whole. Fortunately, Richard Münch (1981, 1982, 1987a, 1987b, 1990) keeps this technique alive. His transparent architecture of consistent fourfold concept formation allows him efficiently to reduce empirical complexities to limited and manageable numbers of systematically interrelated ideal types. One can argue about the specific contents of Münch’s models as well as its underlying constructive principles,9 but it is clear that anything less than similar consistent systems of categories for (institutional and behavioral) social structures and procedures of social functioning will not do. It will leave the task of explaining macrosocial functioning and dysfunctioning and discovering deep structural constraints unfulfilled. Only systemic causal models will help to rebuild the abandoned Parsonian building from its foundation up, anything less will remain cosmetic renovation.

To be sure, sociological systemic models and conceptual schemes are ideal-typical analytical tools that have no direct referents in social reality. Parsons knew this. He always warned his critics against “the fallacy of misplaced concreteness.” By the same token, conceptual schemes cannot be falsified in the sense of being empirically proven wrong. They can be challenged and replaced by another set of systemically interrelated concepts if the latter is found to have more overall interpretive power, or to be more comprehensive, or more internally consistent. When and if macrosocial research problems and hypotheses are formulated with an orientation to established conceptual schemes of social functioning and typical institutional and behavioral social structures rather than to the endless variety of raw variables produced for a myriad of diverse research projects, and when both sets of such categories are associated with the content of moral imperatives, all research results will necessarily become mutually relevant, mutually referential, and cumulative. While this may shift some of our collective effort to constructing and specifying conceptual schemes themselves, the overall savings of effort and improvement of research efficiency seem beyond doubt. In this way, macrosociology will become focused on timeless moral imperatives of social functioning as its core subject-matter rather than episodic and ephemeral fads and foibles. Parsons had similar albeit motivation and action-based considerations in mind when he wrote,

[...] completely raw empiricism is overcome by describing phenomena as parts of or processes within systemically conceived empirical systems. The set of descriptive categories employed is neither ad hoc nor sheer common sense but is a carefully and critically worked out system of concepts which are capable of application to all relevant parts or aspects of concrete system in a coherent way (1951, p.20).

Parsons apparently adopted the systemic idea of interpenetration from Weber’s categories of the relation between religious ethics and the world, and he invoked it as early as 1935 (p. 257). Münch (1981) attributes this idea to Kant, but it may go even further back, to Aristotle (Dubrovsky 1999). Indeed, Marx ([1857-1858]1973, p. 89) presented exchange and distribution as (mutually interpenetrated) middle terms of a standard syllogism where general-societal production and singular-individual consumption were pure polar terms. Mediating bilateral interpenetration of polar opposites is a constructive systemic principle of greater complexity than either Kantian polar dichotomies or Hegelian triads of (unilateral) transcendental supersession (Aufhebung), and it can be seen as incorporating them both. If Parsons’ idea of dynamic structural-functional analysis is combined with the systemic principle of internal bilateral interpenetration, it becomes fully congruent with the paradigmatic four-variable just-identified recursive structural equation model where variables x3 and x4 carry the interpenetrating effects of variables x1 and x2 (see Blalock 1971, pp. 18-32; Duncan 1975, pp. 25-50). To be sure, as causal modeling is not limited to four variables, so is the constructive principle of systemic interpenetration open to higher-order recombinations of pure dichotomies.

Causal Order and Developmental Change

Systemic causal modeling involves the problem of the causal order of analytical categories under which successive subsystems of present social structures are to be subsumed until their empirical analysis becomes meaningful. What determines cross-sectional causal order among social phenomena of interest? While the rich evolutionary-anthropological literature may be helpful in producing systemic hierarchical taxonomies of the kind Parsons proposed in the false context of action theory, what criteria do we have to determine the causal order, for instance, between family status and political socialization? We know that economy and the state evolved from family and cultural institutions (Engels [1884]1972, Levi-Strauss 1956, Poggi 1978). In the present, however, it is political, economic, and cultural factors that affect family institutions, not the other way around (Wilson 1990, Goode 1963). It appears that synchronic cross-sectional causal order is the reverse of the evolutionary-developmental order, that is, where later-developed elements are likely to lead to earlier ones (cf. Heise 1975, pp. 5, 8).

There were as many avenues of social development as social structures. Thus there were innumerable relatively independent possible paths of social development - and there are as many possible reconstructions of them today. The three significant theories of Western social development - due to new cultural attitudes as religious world-outlooks (Vico, Comte, Weber), due to class struggle bringing new production relations (Marx and Engels), and due to increasing complexity of institutional social structures (Spencer, Tönnies, Durkheim) - do not have to be taken as negating each other. Speaking of the economic line of Western social development, Marx’s succession of socioeconomic formations can be corrected based on the historic development of factors of production rather than on forms of class struggle. Social classes are historic occupational groups that can no longer be treated as general sociological or transhistoric categories without an inadmissible modernization of past history. As for factors of production, they developed from land to labor to capital to information. Based on that, we can speak of a historic line of socioeconomic development from ancient colonialism to slavery to capitalism to professionalism.

One can argue that wealth was created mainly by foreign conquest and colonization in Antiquity rather than by exploitation of slave labor. In class terms, this would mean Athenian or Roman conquerors vs. barbarians rather than patricians vs. plebeians (see Pseudo-Xenophon in Toynbee 1953). Substituting ancient slavery with colonialism has the additional advantage of covering Marx’s residual category of the Asian mode of production. Slavery, or serfdom, was much more typical of medieval Western Europe. It was primarily the slave labor of serfs that sustained the Carolingian and other West-European kingdoms economically whereas feudalism was an administrative-political system of regulating vassal obligations (see Bloch 1975, Bush 1996, Drescher and Engerman 1998). Medieval slavery did not disappear with the advent of industrial capitalism in Europe, but only changed into alienated hired labor as classically described by Marx ([1844]1975). Maritime colonialism, too, was revived by major European powers of the past three centuries. Classic laissez-faire capitalism ended with the Great Depression. It has been replaced by professionalism as the new socioeconomic system.

Wealth is now increasingly created by the expert knowledge of managers, attorneys, accountants, physicians, architects, engineers, as well as entertainers, sportsmen, and other professionals with lengthy specialized training and ethical standards enforced by autonomous self-monitoring associations. This phenomenon is often expressed naturalistically, as social change due to changes in technology. In sociological terms, technological progress means shifts in social structures in favor of groups with higher cultural status due to better education and training, and the effects of these shifts on other institutional domains including the economic one. To be sure, our space exploration is a new form of colonialism, but we also use our long arm here on Earth to guarantee overseas sources of raw materials such as crude oil or international waterways. Alienated labor, too, continues to be widely used in mass production and secondary labor market jobs, especially those done by immigrant or foreign workers. And somewhat regulated capitalism is alive and well in many manufacturing, construction, and mining industries. However, investment capital is fragmented by comparison with the concentration of industry and blind without professional financial management. Professionalism is the leading factor of today’s developed economies.

Social change has always been driven by social movements spurred by unfulfilled moral imperatives and ideals such as social justice. European workers’ class struggles of the nineteenth and early twentieth centuries were specific instances of such movements. But all wars and other large-scale social upheavals that punctuate human history can also be seen as outcomes of morally driven movements, as delayed violent reforms of dysfunctional social structures or failed social-structural substitutes. Changes in structures of social behavior also took place at all times in the form of generational transition and transregional diffusion, or by intensification of social space and time as in urbanization-suburbanization and industrialization-professionalization. In such incremental behavioral change, some social strata become relatively smaller and some others, relatively larger. These cumulative quantitative changes make certain strata too small and too insignificant to be counted separately, and certain others, too large and internally too heterogeneous to be counted as one. Structural differentiation and amalgamation mark only quantitative social change, but it cannot account for qualitative changes in world history. Institutions may differentiate internally, but what is of more interest is why they replace other institutions as functional substitutes, and why they are abolished when they get in the way of moral exigencies of their time.

When Durkheim deplored the consequences of organic solidarity in anomie and egotistic behavior outside occupational groups, he only scratched the surface of bitter moral socioeconomic wars highlighted by Marx. Embracing Durkheim’s line of thinking about division of labor and attempting to develop it further by adding a few empiricist modifications in response to critiques of Parsons - such as contingency on concrete social groups and broader array of possible outcomes (see Alexander and Colomy 1990b) - still means staying within the same old Spencerian-Darwinian, basically biological (and politically safe) functionalist imagery of social change. It is meaningless to speak of social structural differentiation and functional specialization in the context of historic social change because this implies the naturalistic paradigm of enduring ever-useful institutions and their changeable instrumental functions, e.g., “pariah groups that fill crucial social roles” (Alexander and Colomy, 1990b, pp. 268-269). Too much attention to empirical complexity may simply indicate a lack of parsimonious conceptual schemes consisting of limited numbers of essential types and subtypes. For all his pan-systemic approach, even Luhmann (1990) is only concerned with removing the obscurity of social complexity. Genuine, qualitative social-structural change comes only with a change in the operation of moral functional imperatives. Sociology must be primarily concerned with this kind of change.

Past sociohistoric evolution is a sequence of transitions of one significant social structure to another in which all major elements of present-day institutional and behavioral forms can be traced back to the earliest stage through an uninterrupted process of transformations or substitutions. As such, evolutionary social development can be described as proceeding or moving structure. By contrast, social functioning at different stages of sociohistoric development is a structured process, or a procedure, consisting of distinct recurring phases. Both functioning and development are sustained by moral functional imperatives operating successfully and prevailing over dysfunctional trends, thus creating relatively stable periods, e.g., early periods of reconstruction and consolidation after qualitatively new stages of social development begin. From this perspective, every new social development is a consequence of moral imperatives unfulfilled during previous sociohistoric epochs. It consists in structural compensations for lack of adequate social-structural forms or of their proper maintenance. Critical social situations occurred when proven structural substitutes of the past failed, thus marking the conclusion of one developmental stage and the beginning of the next one.

Recurring phases of synchronic social functioning and dysfunctioning also form irreversible successions within every aspect of social behavior. For that reason, every functional phase can be seen as a potential (synchronic, yet dynamic) cause of all subsequent ones. Generally speaking, crime and other social problems are manifestations or symptoms of macrosocial dysfunctioning. The task is to identify specific structural constraints for moral imperatives of normal social functioning. In going beyond the symptomatic phenomena of criminal justice to deep structural causes of crime, we will need to specify causal relationships between likely (dysfunctional) elements of social structures, on the one hand, and the incidence or rates of crime and victimization, on the other. To do that, we will need a theory of causal chains among procedural phases of such social dysfunctioning. To determine their order in causal explanations of criminal behavior, let us say by high levels of unfair social advantages and disadvantages, we will need to formulate a theory of origin for all typical phases of social functioning, and then take the reverse of that order. It can be proposed, for example, that if lifestyles have their societal-historic origins in social status, attitudes, and socialization (as they do in individual social development), in synchronic analysis socialization patterns are affected by lifestyles as well as by social status and attitudes, not the other way around.


Functionalism is more than a valuable classic tradition in social theory worthy of being continued. The adoption of neofunctionalism rather than neo-Parsonianism as a name for a movement in sociology imposes certain obligations with respect to the methodology of structural-functional analysis. Centered on moral imperatives, structural-functional analysis becomes causal analysis of moral (dys)functioning. This is a powerful method that no macrosociologist can afford to ignore. Equipped with this method, social science can reconstruct specific structural causes of present social functioning and dysfunctioning in quasi-experimental cross-sectional social research. Significant causal factors of social dysfunctioning can be identified as characteristic behavioral structural constraints for the operation of moral imperatives in particular institutional domains and in their particular (extensive and intensive) spatial and temporal segments. Systemic causal models of synchronic functioning and dysfunctioning social structures can be built and quantitatively estimated taking the reverse of evolutionary-developmental order as a guide. Alternative theories of macrosocial functioning and dysfunctioning can then be tested in field experiments with new or improved social-structural arrangements.

Macrosociology can be advanced if such analysis excludes social action from its frame of reference, whether voluntaristic, strategic, communicative, or collective, and if it is confined to the macrosociology of institutional and behavioral social structures. The notion of teleological functional explanation is a naturalistic misconception. Instead, systemic causal explanations of macrosocial functioning and dysfunctioning are this method’s central concern. While microsocial action always appeals to prevailing social norms, and indeed revolves around them, such norms are discovered only in the realm of macrosocial (dys)functioning and underlying social structures with their moral standards. The idea of functional interdependence must be relegated to the subsidiary task of constructing systemic conceptual schemes serving the needs of causal modeling of macrosocial functioning and dysfunctioning. To conceive macrosocial functioning and dysfunctioning in relational rather than in substantial terms (Emirbayer 1997), postulated elements of conceptual schemes must be relevant to timeless moral imperatives, notably, social justice. At issue in structural-functional analysis are procedures of macrosocial functioning and development as fulfilling or not fulfilling moral imperatives that are normative - through the mediation of ideology, labeling, and institutional behavior - for all social actors. Such analysis presupposes also a continuous emergence of social norms - through creative reinterpretation of meanings, institutionalization, and negotiated order in new social structures of family, cultural, economic, political, regional, generational, socioecological, occupational, and other types of social behavior overcoming inherited structural constraints.

The idea of opposite causal order in the analysis of social functioning and evolutionary development also applies to visions of social systems in general - as ontological assumptions involved in the instrumental-naturalistic and value-rational modes of structural-functional analysis. Instrumental-naturalistic functionalism originated in biological visions of the world where the functioning of living systems is fully determined by their past phylogenetic evolution. For that reason, living systems can be subjected to value-neutral conceptualization and instrumental manipulation. This relationship is reversed in human social relations. The fabric of our social world is not a pregiven fact. It consists of norms and values contested and shaped by numerous social groups representing powerful social movements. Our moral imperatives help us make sense of the past and see inherited social structures as falling short of the requirements of the present. In contrast to naturally evolved living systems, human social development is an artifact determined by the ways in which systems of social relations function and dysfunction. This is why both synchronic and diachronic sociological explanations always contain a value element. This is also why modernization theories of social change have been discarded as invalid. As opposed to the contemplative-positivist norm of value-neutrality, good social science guided by structural-functional method is the best means for testing possible forms of social development (cf. Marx [1845]1975). A demonstrable method is one way for value-rational social scientists to uphold high standards of objectivity and truth, and avoid being ideological servants rationalizing someone’s partisan utopias, plans, or schemes whether well-intentioned or otherwise.

Historic Note: The Erosion of Meaning in Agil Categories

In the concluding chapter of The Structure of Social Action, Parsons became alerted to the fact that macrosocial reality could not be conceptualized as unit acts combined into more complex structures by analogy with atoms and molecules in physics. In view of this obstacle to his planned sociology of action, Parsons set aside the microscopic-voluntaristic aspects of social action - ends and means - and attacked the problem from its macroscopic behavioral side, turning to the study of its typical normative orientations of behavior. Following this road during the 1940’s, Parsons (1954) specified Weber’s general category of value-rational orientations of action - of social behavior, really - with four dichotomous pairs of what he came to call pattern-variables: diffuse vs. specific, instrumental vs. expressive, particularistic vs. universalistic, and ascriptive vs. achievement. Although he cautioned in The Structure of Social Action against an analysis of social relations in terms of any dichotomous categories, and specifically, in terms of Gemeinschaft and Gesellschaft, Parsons later recognized that they indeed lay behind his pattern-variables (1954, p. 360; Parsons and Shils 1951, p. 49; Parsons and Smelser 1956, pp. 33-34). Universalistic, functionally specific, affectively neutral, and achievement roles and orientations all belonged to the artificial-cultural-cold-Gesellschaft type, while their opposites - particularistic, functionally diffuse, affective, and ascriptive orientations were all of the natural-motivational-warm-Gemeinschaft type.

By 1951, Parsons had devised a method of structural-functional analysis where value-orientations and motivations were treated as interrelated social functions determining the relevance and significance of structural elements of what he continued to call systems of action. When he proceeded in The Social System (1951) to explore the mutual relationships between the two sets of pattern-variables, Parsons found that he did not have special succinct categories to name their combinations, and that created a problem. It boggled Parsons’ endeavor to develop a system of general sociological categories in an increasingly verbose and tortuous analytical discourse of never diminishing abstractness. Bales’ categories of small-group interaction analysis offered a solution to this problem. In Chapter 3 of The Working Papers (1953), Bales’ twelve categories of social-emotional problems arising in task-oriented small groups were reduced to six symmetrical positive and negative pairs and then to four. Parsons associated each of the four interaction problems - Adaptive, Instrumental, Expressive, and Integrative - with a pair of pattern-variables as “different ways of conceptualizing the same thing.” In Chapter 5, the names of the two middle categories, Instrumental and Expressive, were changed to Goal-gratifying and Latent, respectively. Then Bales’ categories were interpreted dynamically, as phases in the change of system state, i.e., its functioning, and a necessary change was made in the associations of Parsons’ pattern-variables with Bales’ newly-renamed categories.

In this way, pure Gemeinschaft and Gesellschaft orientations were subsumed under Adaptation and Integration, and their interpenetrations, under Goal-gratification and Latent tension-management as paradigmatic social functions, essentially, functional imperatives. However, while helping to make his conceptualizations less verbose and more readable, this reorganization of Parsons’ analytical apparatus did not come without a price. What became known as the AGIL four-function paradigm added new complications to Parsons’ thought. Bales’ categories commanded their own meanings that implied certain usages and precluded certain others. This raised a problem of justifying the identification of combinations of pattern-variables with particular AGIL categories. Parsons also had to demonstrate that these combinations were consistent with Bales’ understanding of process as a linear succession of phases each related only to its preceding and following ones. Parsons began speaking about process as intermittent, but also as circular and spiral

Initially, Parson was flexible and shrewd in making a disclaimer that the AGIL sequence of the four functions was provisional only, that it had been obtained as a “certain generalization” from Bales’s empirical research in small-group dynamics, and that besides a simple reversal, “other phase [sequences are] possible.” Moreover, the names of the four functional phases were also all incidental, literally appropriate only to face-to-face interaction in small groups (Parsons, Bales, and Shils 1953, pp. 72, 172, 179, 188, 193). But in the avalanche of new conceptual discoveries triggered by the AGIL formula, all these caveats were soon forgotten. In particular, why did Parsons place the Integration phase that accommodated all pure Gemeinschaft orientations in the middle position rather than in the end - or the beginning - of the process where it properly belonged? In the excitement of all the changes of focus and terminology, the sequential order of the third and fourth phases in Bales’ original functional problems, detailed on pages 64, 74, and 79 (Chapter 3), became reversed on page 180. It was apparently an oversight as there is no reason to believe that it happened by design. In any event, the erroneous AGIL order of functional phases was accepted “as a handy idealized model,” and in this form it entered into all Parsons’ conceptualizations and theories that followed - rather than the more correct one - AGLI. On second thought, the choice of AGIL over AGLI may well have been deliberate, based solely on the euphonic qualities and connotations of the two acronyms. This hardly noticeable error compounded even more the inconsistencies of Parsons’ subsequent work. The enormous remedial effort Parsons spent during the 1960’s to reconcile those inconsistencies can in part be traced to this small error.

The difference between the polar Gemeinschaft and Gesellschaft-type orientations in all the four pattern-variables became quickly obscured by the new crosscutting distinction between motivational, actor-related pattern-variables, on the one hand, and situational, object-related ones, on the other. This loss of connection with pattern-variables and, therefore, with (macrosocial) action content in AGIL already happened in Economy and Society - notwithstanding the “Technical Note” to the contrary (Parsons and Smelser 1956, pp. 33-38), and except for one revisit ([1960]1967), it soon became all but history. In his subsequent increasing use of the AGIL paradigm, Parsons abandoned even the limited content of macrosocial action - read: behavior - that had been carried over by the original meanings of the pattern-variables. As a lifesaver, he arbitrarily relapsed into the voluntaristic-microsocial language of actors, ends, and means without as much as one sentence of justification. Five years after he introduced the macrosociological functionalist AGIL formula, Parsons ([1958]1960, pp. 165-168) simply identified Goal-attainment with goals, Adaptation with means, Integration with norms, and Latent pattern-maintenance with values. While he was apparently unconcerned with the difficulty of conceiving his subsystems, especially the cultural one, and the later added “behavioral organism” in such action terms, Parsons’ sociology as a general theory of action remained a promise not kept.

This deprived Parsons’ symbolic media of interchange (other than money) of their foundation. In addition, the media were conflated with exchanged values every time Parsons spoke about the media as circulating. Parsons was forced to give up the two-way passage route of systemic phase movement for a linear hierarchy of conditions and controls. A third modification of the AGIL scheme came in “Some Problems of General Theory in Sociology” ([1970]1977) where Parsons subsumed the AGIL categories under two crosscutting dimensions, or axes - action space and time, both understood naturalistically, as those of “any living system.” Adaptation was now characterized as external-instrumental, Goal-attainment as external-consummatory, Integration as internal-consummatory, and Latency as internal-instrumental. In “Pattern-Variables Revisited” ([1960]1967), this idea was still obscured by the complicating cross-classification of the four system phases themselves in a four-by-four fashion. Now only two of the four “general systemic” functions clearly turned out to be internal, i.e., directly relevant to any system on the same level of analysis, and only two of the four had any temporal order to them.

This final rationalization of the AGIL scheme came the closest to the way Parsons had actually used it since Economy and Society where he treated the three non-economic subsystems as economy’s external environment. The same one-internal-to-three-external scheme was played out on all sublevels of analysis. The AGIL categories were no longer universal functional requisites of social systems, but rather categories of space and time! With this, Parsons recognized, in fact, that the four AGIL functions had lost all meaning of fundamental categories of social action. In their encounters with substantive fields like economics, the abstract programmatic meanings of Parsons’ major concepts did not acquire new sociological content. On the contrary, those encounters resulted in the destruction of Parsons’ abstract categories themselves. In the end, what was left of the four-function paradigm was the thin biological and general-systems idea of organism-and-environment, the equally naturalistic notion of homogeneous linear time in the context of unit act, and the evolutionary conception of structural differentiation, functional specialization, and integration. The AGIL scheme had lost all sociological meaning envisioned in The Structure of Social Action.


Aberle, D.F., A.K. Cohen, A.K. Davis, M.J. Levy, Jr., F.X. Sutton. 1950. “The Functional Prerequisites of a Society.” Ethics 60 (Jan):100-111.

Alexander, Jeffrey C. 1982. Theoretical Logic in Sociology. Volume One. Positivism, Presuppositions, and Current Controversies. Berkeley: University of California Press.

Alexander, Jeffrey C. 1983. The Modern Reconstruction of Classical Thought: Talcott Parsons. Berkeley: University of California Press.

Alexander, Jeffrey C. (ed.) 1985. Neofunctionalism. Beverly Hills: Sage.

Alexander, Jeffrey C. 1998. Neofunctionalism and After. London: Blackwell.

Alexander, Jeffrey C., and Paul Colomy. 1990a. “Neofunctionalism Today: Reconstructing a Theoretical Tradition.” Pp. 33-67 in Frontiers of Social Theory, edited by George Ritzer. New York: Columbia University Press.

Alexander, Jeffrey C., and Paul Colomy (eds.) 1990b. Differentiation Theory and Social Change. Comparative and Historical Perspectives. New York: Columbia University Press.

Blalock, Hubert M., Jr. (ed.) 1971. Causal Models in the Social Sciences. Chicago: Aldine.

Bloch, Marc Léopold Benjamin. 1975. Slavery and Serfdom in the Middle Ages: Selected Essays. Berkeley: University of California Press, 1975.

Bush, Michael L. 1996. Serfdom and slavery: studies in legal bondage. New York: Longman.

Colomy, Paul (ed.) Neofunctionalist Sociology. London: Elgar.

Coser, Lewis. 1956. The Functions of Social Conflict. New York: Free Press.

Drescher, Seymour, and Stanley L. Engerman (eds.) 1998. A historical guide to world slavery. New York: Oxford University Press.

Dubrovsky, Vitaly. 1999. “Beyond Duality: From Opposition to Constructive Attribution.” Proceedings of the Forty-Second Meeting of the International Society for Systems Sciences. Louiseville, Kentucky.

Duncan, Otis Dudley. 1975. Introduction to Structural Equation Models. New York: Academic Press.

Durkheim, Emile. [1902-1903]1961. Moral Education. New York: Free Press.

Durkheim, Emile. [1924]1953. Sociology and Philosophy. Glencoe: Free Press.

Emirbayer, Mustafa. 1997. “Manifesto for a Relational Sociology.” American Journal of Sociology 103(2):281-317.

Engels, Frederick. [1884]1972. The Origin of the Family, Private Property, and the State. New York: International Publishers.

Fligstein, Neil. 1996. “Markets as Politics.” American Sociological Review 61:656-673).

Freedman, David A. 1991. “Statistical Models and Shoe Leather.” Sociological Methodology 21, pp. 291-313.

Fuchs, Stephen. 1986. “The Social Organization of Scientific Knowledge.” Sociological Theory 4(2):126-142.

Fulbrook, Mary. 1978. “Max Weber’s ’Interpretive Sociology’: a comparison of conception and practice.” British Journal of Sociology 29(1):71-82.

Gans, Herbert. 1973. More Equality. New York: Vintage.

Granovetter, Mark S. 1985. “Economic action and social structure: the problem of embeddedness.” American Journal of Sociology 91:481-510.

Goode, William J. 1963. World Revolution and Family Patterns. New York: Free Press.

Heidegger, Martin. 1962. Being and Time. New York: Harper & Row.

Heise, David 1975. Causal Analysis. New York: Wiley.

Hempel, Carl G. 1959. “The Logic of Functional Analysis.” Pp. 271-307 in Symposium on Sociological Theory, edited by Llewellyn Gross. Harper & Row.

Joas, Hans. 1996. The Creativity of Action. University of Chicago Press.

Levi-Strauss, Claude. 1956. “Family.” Pp. 142-170 in Harry L. Shapiro (ed.) Man, Culture, and Society. New York: Oxford University Press.

Lidz, Victor. 1981. “Transformational Theory and the Internal Environment of Action Systems.” Pp.205-233 in Advances in Social Theory and Methodology. Toward an Integration of Micro- and Macro-Sociologies, edited by K.Knorr-Cetina and A.V.Cicourel. Boston: Routlage and Kegan Paul.

Liska, A.E. and B.D. Warner. 1991. “Functions of Crime: a Paradoxical Process.” American Journal of Sociology 96:1441-1463.

Luhmann, Niklas. 1995. Social Systems. Stanford University Press.

Luhmann, Niklas. 1990. “The Paradox of System Differentiation and the Evolution of Society.” Pp. 409-440 in Differentiation Theory and Social Change. Comparative and Historical Perspectives, edited by Jeffrey C. Alexander and Paul Colomy. New York: Columbia University Press.

Marx, Karl. [1845]1975. “Concerning Feuerbach.” Pp. 421-423 in Early Writings. New York: Vintage.

McKim, Vaughn R., and Stephen P. Turner (eds.) 1997. Causality in Crisis? Statistical Methods and the Search for Causal Knowledge in the Social Sciences. University of Notre Dame Press.

Merton, Robert K. 1938. “Social Structure and Anomie.” American Sociological Review 3.

Merton, Robert K. [1949]1967. “Manifest and Latent Functions.” Pp. 73-138 in On Theoretical Sociology: Five Essays, Old and New. New York: Free Press.

Mills, C. Wright. 1959. The Sociological Imagination. New York: Oxford University Press.

Münch, Richard. 1981. “Talcott Parsons and the Theory of Action. I. The Structure of the Kantian Core.” American Journal of Sociology 86:709-739.

Münch, Richard. 1982. “Talcott Parsons and the Theory of Action. II. The Continuity of Development.” American Journal of Sociology 87:771-826.

Münch, Richard. 1987a. “Parsonian Theory Today: In Search of a New Synthesis.” Pp. 116-155 in Social Theory Today, edited by A.Giddens and J.Turner. Stanford U Press.

Münch, Richard. 1987b. “The Interpenetration of Microinteraction and Macrostructures in a Complex and Contingent Institutional Order.” Pp. 319-336 in The Micro-Macro Link, edited by J.Alexander, B.Giesen, R.Münch, and N.Smelser. Berkeley: University of California Press.

Münch, Richard. 1990. “Differentiation, Rationalization, Interpenetration: The Emergence of Modern Society.” Pp. 441-464 in Differentiation Theory and Social Change. Comparative and Historical Perspectives, edited by Jeffrey C. Alexander and Paul Colomy. New York: Columbia University Press.

Parsons, Talcott. [1935]1991. “The Place of Ultimate Values in Sociological Theory.” Pp. 231-257 in Talcott Parsons: The Early Essays, edited by C. Camic. Chicago: University of Chicago Press.

Parsons, Talcott. [1937]1968. The Structure of Social Action. New York: Free Press.

Parsons, Talcott. 1951. The Social System. New York: Free Press.

Parsons, Talcott. 1954. Essays in Sociological Theory. Revised Edition. New York: Free Press.

Parsons, Talcott. [1955]1960. “Social Strains in America.” Pp. 226-247 in Structure and Process in Modern Society. Glencoe: Free Press.

Parsons, Talcott. [1958]1960. “Some Reflections on the Institutional Framework of Economic Development.” Pp. 98-131 in Structure and Process in Modern Society. Glencoe: Free Press.

Parsons, Talcott. [1960]1967. “Pattern Variables Revisited: A Response to Robert Dubin.” Pp. 192-219 in Sociological Theory and Modern Society. New York: Free Press.

Parsons, Talcott. 1966. Societies: Evolutionary and Comparative Perspectives. Englewood Cliffs: Prentice-Hall.

Parsons, Talcott. [1970]1977. “Some Problems of General Theory in Sociology.” Pp. 229-269 in Social Systems and the Evolution of Action Theory. New York: Free Press.

Parsons, Talcott, and Edward A. Shils. 1951. “Values, Motives, and Systems of Action.” Pp. 47-275 in Toward a General Theory of Action, edited by T. Parsons and E. Shils. Cambridge: Harvard University Press.

Parsons, Talcott, Robert F. Bales, and Edward A. Shils. 1953. Working Papers in the Theory of Action. Glencoe: Free Press.

Parsons, Talcott, and Robert F. Bales. 1955. Family, Socialization and Interaction Process. Glencoe: Free Press.

Parsons, Talcott, and Neil J. Smelser. 1956. Economy and Society. A Study in the Integration of Economic and Social Theory. New York: Free Press.

Radcliffe-Brown, A. R. 1952. Structure and Function in Primitive Society. New York: Free Press.

Schutz, Alfred. [1932]1967. The Phenomenology of the Social World. Northwestern University Press.

Schwinn, Thomas. 1998. “False Connections: Systems and Action Theories in Neofunctionalism and Jurgen Habermas.” Sociological Theory 16(1):75-95.

Smikun, Emanuel. 1990. “Word and Deed in Weber’s Methodology.” Michigan Sociological Review No. 4.

Sobel, Michael. 1995. “Causal Inference in the Social and Behavioral Sciences.” Pp. 1-39 in: Handbook of Statistical Modeling for the Social and Behavioral Sciences, edited by G. Armiger et al. New York: Plenum Press.

Toynbee, Arnold. 1953. Greek Civilization and Character: The Self-Revelation of Ancient Greek Society. New York: Mentor.

Turner, Jonathan H., and Alexandra Maryanski. 1979. Functionalism. Menlo Park, CA: Benjamin/Cummings.

Turner, Jonathan H., and Alexandra Maryanski. 1988. “Is ’Neofunctionalism’ Really Functional?” Sociological Theory 6:110-121.

Turner, Stephen P. 1983. “Weber on Action.” American Sociological Review 48:506-519.

Weber, Max. [1904]1949. “’Objectivity’ in Social Science and Social Policy.” Pp. 49-112 in The Methodology of the Social Sciences, translated and edited by E. Shils and H. Finch. New York: Free Press.

Weber, Max. [1906]1949. “Critical Studies in the Logic of the Cultural Sciences.” Pp. 113-187 in The Methodology of the Social Sciences, translated and edited by E. Shils and H. Finch. New York: Free Press

Weber, Max. [1906]1978. “The Logic of Historical Explanation.” Pp. 111-131 in Max Weber: Selections in Translation, edited by W. Runciman. Cambridge: Cambridge University Press.

Weber, Max. [1922]1978. Economy and Society. An Outline of Interpretive Sociology. Berkeley: University of California Press.

Weber, Max. [1913]1981. “Some Categories of Interpretive Sociology.” Sociological Quarterly 22(2):151-180.

Wilson, William Julius. 1990. The Truly Disadvantaged: The Inner City, the Underclass and Public Policy. University of Chicago Press.

Zelizer, Vivian A. 1985. Pricing the Priceless Child: The Changing Social Value of Children. New York: Basic.

Zukin, Sharon, and Paul DiMaggio (eds.) 1990. Structures of capital: The social organization of the economy. Cambridge University Press.


1. Merton ([1949]1967, p. 106) himself occasionally used the phrase “functional imperatives” in this sense where, in his words, one deals with “the range of possible variation in the [structural] items which can, in the case under examination, subserve a functional requirement.”

2. All exercises attempting to attribute useful functions to social problems, such as conflict (Coser 1956) or poverty (Gans 1973, pp. 102-126), are products of the same perverted instrumental logic of naturalistic functionalism. Empirical tests of the usefulness-of-crime hypothesis (e.g., Liska and Warner 1991) arrive at similar conclusions.

3. According to Turner and Maryanski (1979, pp. 71-72), Parsons’ subsystems were elaborations of the structural elements of his 1937 concept of social action. “Ideas have become a cultural system; actors’ goal-seeking and calculating activities are conceptualized as a ’personality system’; and the missing element in the analysis of ’unit acts,’ interaction, is now conceptualized as a ’social system.’“ This reconstruction of Parsons’ intellectual history is not supported by evidence. Parsons introduced culture, personality, and the social system in separate works spanning several years - in 1942, 1945, and 1950 (see Parsons 1954). Nor can the subsystems-from-action hypothesis account for Parsons’ preoccupation with pattern-variables during the same period and for their central role in the development of the AGIL paradigm.

4. As Heidegger (1962, pp. 55-58, 182-203) noted, interpretation and understanding are never the same.

5. In his early work, Parsons ([1937]1968, pp. 85-86) distinguished between voluntaristic action and behavior, but he talked about behavior in a psychological rather than sociological sense. From the 1950’s on, he blended the distinctive features of the two and used the term social action for both.

6. Hempel (1959) whose thinking was deeply involved with natural sciences recognized the legitimacy of naturalistic functional explanations. Still, he did so reluctantly and with reservations. He saw them as weak and relativistic with respect to specific conditions for survival or normal social functioning.

7. The fact that statistical techniques of causal modeling are not without their pitfalls and problems (see Freedman 1991and discussion; Sobel 1995; McKim and Turner 1997) should not deter us from this promising strategy.

8. This often produced place holders, a formalistic interpretation of a set of abstractions in terms of themselves. Thus, in Economy and Society (Parsons and Smelser 1956, pp. 51-66), investment and political encouragement of economic activity were presented as boundary zones of interchange between economy and polity by referring to the same abstract meanings of Adaptation and Goal-attainment in terms of which the pure opposites - economy and polity - had been interpreted in the first place. In principle, there is no reason why a recombination of a pure opposite with a product of its interpenetration should not be possible. However, for this procedure to make sense, the second- and higher-order polar opposites must be explicitly redesignated as such, and the change of their meanings specially noted.

9. Münch does not differentiate between subsystem exchange, two-by-two cross-classifications, and interpenetration proper. Nor does Luhmann.