1 Why do we need yet another study on Parsons? After all, his importance is widely recognized, to such an extent that we call him the last of the “classics.” However, his work is situated within a historical context and could therefore be outdated. This reservation is unfounded, at least for the problem that interests us here. It should be remembered that for Parsons, one of the epistemological priorities in The Structure of Social Action is to demonstrate the legitimacy of abstraction in scientific research. Moreover, this interest leads Parsons to position himself firmly regarding “the extent to which analytical abstraction can be pushed” and in formulating a suitable response to this question for the social sciences.

A First Level of Abstraction: The Frame of Reference

2 In the introductory chapter of The Structure, Parsons introduces the notion of “descriptive frames of reference” and emphasizes that all sciences require them.

3 In support of this statement, he develops a vigorous and compelling argument: the simple description of phenomena for scientific purposes already implies “a selective ordering” based on explicit frames of reference. To cite Lawrence Henderson [1] (as Parsons himself does), a frame of reference is “a conceptual schema.” All empirical observation requires, refers to, and takes its meaning from such a schema. Scientific activity, for example, relies on explicit articulation and systematic application of the scheme. In some ways, this constitutes a preliminary stage, preceding the explanation of phenomena, and is indispensable insofar as it makes their description possible.

4 However, there are numerous frames of reference, with different degrees of generality and different scientific objectives. The key issue in Parsonian thought is clearly choosing the most appropriate frame of reference for social sciences. Before doing this, however, Parsons invites us to understand certain characteristics of frames of reference, without which the scholar—and particularly the social scientist—risks serious error.

5 All frames of reference are abstract schemata, but this is often forgotten or poorly understood, particularly in the field of social sciences. Parsons severely frowns on this absence of epistemological vigilance, writing of “the fallacy of misplaced correctness,” which consists in mistaking the abstract framework for an expression of concrete reality. Here, Parsons uses the words of Alfred North Whitehead, who demonstrated the existence of such a confusion regarding the spatiotemporal frame of reference underlying classical mechanics. To generalize somewhat, it could be said that frames of reference fall within the domain of “models,” to use the word in its widest sense, and that they should not be mistaken for primary principles of reality.

6 Parsons emphasizes this theme in order to set himself apart from a view of economic theory that was widespread in the 1930s. Thus, when discussing Lionel Robbins’s prominent work An Essay on the Nature and Significance of Economics, he criticizes Robbins for falling into “the fallacy of misplaced correctness,” which is all the more damaging for being committed by a sophisticated thinker: Robbins does in fact recognize the abstract nature of economic theory, but nevertheless presumes a “separate existence of economic forces” supposedly ensuring the independence of economics as a science. [2] The basic error recurs here in the assumption that economic forces are a separable subgroup of phenomena. It would be more appropriate to establish an analytical distinction between economic elements and other elements. This can be seen as the first stage in the elaboration of economic theory, in that, according to Parsons, “true scientific abstraction” implies “always making certain factors vary (experimentally or analytically) independently of others.” [3] However, it is the choice of a frame of reference that allows this essential distinction to be made. It is accompanied by a process of selecting phenomena that varies according to the dominant scientific interest, but which essentially aims to specify, by abstraction, the phenomena for explanation. Other phenomena are not considered to be relevant variables for the problems under examination, but are treated simply as data. Parsons illustrates this point in The Structure, using the example of a man who commits suicide by jumping off a bridge. The sociologist sees this as an act for which the causes and reasons must be identified. The physician sees it as the fall of the body, requiring explanation. However, the sociologist cannot completely ignore this fall, even if it is not one of the phenomena to be explained: if the jump did not lead to a fall and to a high probability of death, it would not be possible to speak of suicide (or even attempted suicide). The jump must therefore be included as data, but considered, in Parsons’s words, as “unproblematic” by the “social scientist.” [4]

7 It seems to me that these general considerations concerning the frame of reference have a twofold objective: they help to sustain an epistemological vigilance about the different variants, naive or disguised, of empiricism, and they aim to emphasize the full implications of the frame of reference—and its choice—in research. These observations will be kept in mind as this paper moves on to Parsons’s specific proposal in the field of the social sciences.

8 Parsons has great faith in the frame of reference of action and defends it resolutely throughout his book. Certainly, he indicates that it is not the only possible frame of reference and, referring to Znaniecki’s The Method of Sociology, he cites four: “social action” of course, but also “social relationships,” “social groups,” and “social personality” (which deals with the individual in society). But he immediately specifies that he gives priority to the scheme of action, with its emphasis on the relationship between means and ends. Moreover, in the final chapter of The Structure, which is devoted to epistemological questions, he asserts that the three other conceptual schema are “secondary” to that of action, while also recognizing their usefulness for the description of complex phenomena.

9 This enables an understanding of why Parsons gives so much importance to characterizing the relevant “basic unit” or “unit act” in the frame of reference of action. The unit act requires an “actor,” is oriented toward an “end,” and takes place in a “situation” composed of two distinct elements: those over which the actor has no control (the “conditions” of the action) and those over which the actor exerts control (the “means”). Parsons’s originality resides in the addition of an extra characteristic to these traditional elements. He calls this the “normative orientation” of the action and states that it influences the “choice of means” (that is, the choice of certain means over others) “in so far as the situation allows alternatives.” Responses to this analysis of the unit act, and its ramifications, vary: Jeffrey Alexander overvalues it, seeing it as the basis for a “multidimensional approach to action,” whereas Anthony Giddens undervalues it in an attempt at controversial differentiation. In any case, it has the merit of describing the essential categories of the frame of reference of action and sketching out their relationships, although it does this with some ambiguity: as demonstrated by later developments, the “normative orientation” cannot be reduced to the choice of certain means, but is also expressed in the nature of certain ends, pursued and desired in their own right. [5] In The Structure, Parsons does not yet offer the enlightening distinction between values (as criteria for the desirable) and norms (as rules of action). This reservation aside, Parsons’s treatment of categories lays the groundwork for a second use of the frame of reference, no longer only descriptive, but also analytical. The first scenario involves concrete description of the ends, means, and conditions. In the second scenario, these categories, along with the normative orientation, become elements of the action. In this sense, the frame of reference of action constitutes the departure point required to build a large-scale theoretical construction, such as that conceived by Parsons.

10 Generally, the priority that Parsons accords to action appears to be justified. It should be noted, as the argument developed in The Structure demonstrates, that this priority does not exclude recourse to the notion of the system: at a certain level of complexity, the phenomena confronting the sociologist must be analyzed in terms of systems of action. Throughout his career, Parsons was to maintain this position, meaning that the extent of the “transition” from The Structure of Social Action to The Social System should be qualified: it would be excessive to say that there is a radical opposition between the action perspective and the system perspective, at least in Parsons’s work. It should also be added that since sociology takes an interest in “systems of social action,” the “relational schema” can be considered entirely suitable, without contradicting the priority status of the schema of action. In this case, we are not faced with the inherent problem found in any paradigm of “utilitarian” inspiration (in the wider sense of the term, where utilitarianism sees actors as juxtaposed, making the passage from action to interaction problematic).

11 It thus seems that Parsons’s application of the frame of reference of action is neither narrow nor dogmatic. It should nevertheless be remembered that as a result of its generality, this framework is compatible with diverse variants.

A Second Level of Abstraction: The Formulation of Concepts

12 Parsons’s thought on this matter is twofold. On the one hand, it relies on the concept of the Weberian ideal type, recognizing its importance. On the other hand, it is guided by a desire to clarify and improve this concept. Parsons can only be favorable to the general orientation of a theory presenting the ideal type as a “construction of elements abstracted from the concrete, and put together to form a unified conceptual pattern.” [6] However, reading the Wissenschaftslehre in the light of the works of Alexander von Schelting, [7] he (like von Schelting) immediately finds himself uneasy with the way in which the ideal type groups together two heterogeneous conceptual categories: “individualizing” concepts and “generalizing” concepts. Given the epistemological “line” that he takes throughout The Structure, his attention is focused on “generalizing” ideal types.

13 Similarly, Parsons is in agreement with Weber on one key point: the necessity of recourse to general concepts within the framework of the social sciences, particularly those that, like sociology, attempt to identify “regularities.” However, he believes that Max Weber remains ambiguous regarding the degree of abstraction achieved in and by the “generalizing” ideal type. The concept in question can, in fact, be interpreted in two very different ways. Parsons specifies these in a text published before The Structure, which is devoted to a critical analysis of the work of Alexander von Schelting. [8] In the first interpretation, the generalizing ideal type “designates a fictive entity, ‘objectively possible’ [in a hypothetical sense], a ‘unit’ or a ‘part’ of a historical individual, [which] is logically analogous to ‘the machine without friction’ or to the ‘ideal gas” in physics.” However, in the second interpretation, “the element [in question]”—it should be noted that the reference to the ideal type has disappeared—“constitutes a ‘general property’ of historical individuals such as ‘economic rationality,’ analogous to ‘mass’ or ‘velocity’ in mechanics” (Camic [ed.], 1991, 127–8). Parsons emphasizes that both interpretations concern “abstract and general concepts,” but that it is “vital” to distinguish between them. This distinction is implied (but never made explicit) in Weber’s theoretical work. [9]

14 As we know, Parsons considers it appropriate (at least when dealing with the theoretical social sciences) to extend the process of abstraction to include the second type of concepts, “elements.” As is made clear in the cited text, these elements correspond to general properties, or even combinations of properties, and not to constituent (or rather, hypothetically constituent) parts of a social phenomenon. The recommended process is thus to deconstruct concrete and singular wholes into analytical elements, which could be qualified as fundamental. In Parsonian theory, these elements are seen as being ideally universal, or at least universalizable. This is the basis of what Parsons himself calls his “analytical realism.” The emphasis on analytical elements as privileged concepts of sociological theory thus leads him to depart radically from the Weberian position and to refuse to subscribe to the theory of the fictional nature of concepts.

15 The distinction between the two types of concepts is accompanied, in the above passage, by a further differentiation concerning laws. If the conceptualization arises from the logic of the ideal type, “the law [is thus] a generalization about the behavior of this hypothetical identity under certain conditions.” If, however, the conceptual development brings to light analytical elements, “the ‘law’ represents a uniform method of relation between the specific ‘values’ of two or more such elements” (Camic [ed.], 1991, 128). In the first scenario, the law sets out a borderline case, which may be more or less approximate. In the second scenario, its ambition and demands reflect an essentially Newtonian model of science. However, it is justifiable today to see such a model as dated: it seems less and less well adapted to the specific requirements of the social sciences. Parsons himself, who admits “having played with the possible applications and the relevance, for our purposes, of the Newtonian model for a theoretical system,” gradually moved away from it, coming to consider that “the theoretical logic of theory in social sciences should be closer to the Mendelian model than to the Newtonian model.” [10] However, reservations about Parsons’s long-held notion of laws and, more widely, about the definition of priorities for the theoretical social sciences should not mask the interest of his thinking for the construction of concepts. The “program” that aims to overcome the inherent limits of the ideal type, in order to bring out the analytical elements, is of interest, and Parsons’s implementation of it merits examination.

16 In this respect, The Structure cannot help but leave the reader dissatisfied: although it clearly outlines the program, it somewhat lacks analytical concepts. However, Parsons does not forget the principles laid out. He adheres to them fully, by deconstructing the classical notions of “community” and “society” to formulate pattern-variables (these, along with the four functions, are his best-known concepts).

17 It seems justifiable to see “community” and “society” as ideal types, even if certain precautionary remarks are necessary. Firstly, the conceptual status that Tönnies attributed to them changed in the course of his long career. Secondly, Weber himself, in his 1920 “systematic,” reformulates the two concepts as “communitarization” (Vergemeinschaftung) and “societization” (Vergesellschaftung). Finally, it is important to note Weber’s ambivalence toward Tönnies, combining respect and reluctance. Therefore, community and society are not, stricto sensu, Weberian ideal types. Nonetheless, from an epistemological point of view, they still belong to the general category of ideal types.

18 Parsons’s work of analytical decomposition allows him to demonstrate that community and society are composite forms, relying on a combination of analytically independent variables. Thus, the distinction between professional roles and parental roles, which reformulates and transposes the well-known dichotomy in more modern sociological language, is based on the term-for-term opposition of two groups of variables. The universalism, achievement (performance), affective neutrality, and specificity that are characteristic of professional roles are opposed, to particularism, ascription (quality), affectivity and diffuseness in parental roles. It is hard to deny that the use of “pattern variables” has clarifying value, and that it has helped to enrich sociological analysis. This is demonstrated by Parsons’s work (in his “intermediary” period) and by the multiple applications of the “variables,” even if they are no longer in fashion today.

19 In recognizing this contribution, it should be remembered that, as for any conceptual tool, there are limits to the relevance of variables. The different groupings of variables (or rather, of their alternative poles) do not all have the same empirical scope: in the two examples given above, it is manifest, but in other cases it can be very limited. However, Parsons sometimes systematically “rotates” his concepts, as if the possible combinations were all of the same importance. He particularly does this when attempting to establish the correspondence between different “sets” of variables and the associated functional problems linked (according to Robert Bales) to the activity of small groups. [11] In doing so, he does not avoid the kind of conceptual scholasticism for which he is often criticized, but which should not be allowed to become a caricature of his work as a whole.

20 We are thus confronted with a common risk, inherent in any process of abstraction: in losing its heuristic value, it can lead to “emptiness.” If, however, as seen above, such a risk is unavoidable, the flexibility offered by the multiple combinations of variables provides a further advantage. It allows unexpected pairings, in the form of counterintuitive and possibly enlightening combinations. Thus, the combination of universalism and quality, which assumes an ideal model to be pursued or defended, gives Parsons a departure point to identify an inconsistency in the German value system, at once the source of tensions and anomie (normlessness), making it particularly vulnerable to the rise of Nazism. Here, the use of variables allows for the development of an original, if debatable, interpretation. [12]

21 The conceptual part of the program outlined in The Structure thus culminates with the development of the pattern variables. Their diverse applications and the various elucidations that they offer give solid justification to Parsons’s reasoning.

A Third Level of Abstraction: The Search for Mechanisms

22 It seems that undogmatic readers will readily accept the usefulness of Parsonian reasoning, with its dual focus on the frame of reference and on “analytical elements.” They will also appreciate the significance of the results obtained. However, such readers might be tempted to add that attempts to refine theory are now moving in other directions, particularly that of the search for mechanisms. Undoubtedly, scientistic enthusiasm would be inappropriate here, as the formal theories remain incomplete. Moreover, it is not possible to envisage going (as Sørensen advocates) [13] beyond the “language of variables” (in a Lazarsfeldian sense), unless quantitative data are available to the researcher. The path is thus difficult, and even narrow, especially on the level of general theory. Nevertheless, it is worth exploring.

23 In his way, Parsons did explore it, contrary to the controversial image representing him as trapped in the construction of taxonomies. His contribution was not negligible, if only because of his work on mechanisms of social control.

24 At the beginning of The Social System, Parsons proposes a general and somewhat complex characterization of mechanism, which highlights three dimensions: a dynamic aspect, which can be analyzed in terms of processes; an individual component, in terms of motivation; and a reference to the consequences for the system of action in question—in this case, the social system. [14] In its generality, this formulation presents a major disadvantage: it does not make clear the essential character of the mechanisms best demonstrated by Parsons. In fact, their potential effectiveness depends on their relational dimension. At the very heart of the mechanism lies an expected result of the action (or the attitude) of alter on ego in the context of the system of interaction, which tends to favor its maintenance (or rather its continuity).

25 In this perspective, the question of social regulation takes a central position. This is why, in chapter 7 of The Social System, Parsons examines its various forms—from informal social regulation to institutionalized methods of regulation—as well as the appropriate mechanisms used. [15]

26 This gives him a different perspective on the relationship between doctor and patient—one of his preferred empirical subjects—and makes him add to the now-established analyses in chapter 10, which particularly define the respective roles of physician and patient using “pattern variables.” The originality of this approach arises mainly from the fact that Parsons is able to fully consider the asymmetry of the relation between physician and patient;the question now is to assess whether and how it is possible to compensate for this asymmetry.

27 This is where the mechanism comes into play, taking on this compensatory function. It has four components, which can also be understood as complementary mechanisms. Firstly, the physician (understood in a broad sense and including psychotherapists as well as ordinary practitioners) gives support to the patient, in order to help him toward recovery. Such support undeniably has a moral dimension. The physician also shows permissiveness regarding the emotions sometimes manifested by the patient, or even his lack of control. Nevertheless, the physician exhibits a refusal to reciprocate in the “game” of seduction or manipulation that the patient may play: he “blocks” any attempt at diversion or escape from the shared task. Finally, he uses the social “rewards” available to him in a manner adapted to the patient’s behavior (manipulation of rewards): he gives his approval or his disapproval regarding the patient’s behavior and attitudes, according to the patient’s significant “efforts” and to his ability to remain on the path to recovery. The mechanism (or the mechanisms) thus aim(s) to ensure the patient’s progressive “social reintegration” and in the specific role of the physician, becomes institutionalized, thus highlighting each of the components. However, it should be noted that the same mechanism (or at least the same type of mechanism) is found in other areas of social life, with varying degrees of institutionalization.

28 Parsons thus expressly refers to the behaviors of a mourning person’s immediate or more distant relations. The attitudes of support and tolerance are, in fact, a shared trait of mourning rituals (not in a narrow ethnological sense, but covering a large range of different behaviors according to the place occupied by institutional formalism). The manifestations of the refusal of reciprocity undoubtedly depend on the closeness to the person in mourning. However, this refusal is generally present, even if it is latent, implicit, or subtle. Finally, the members of the meaningful social circle often make measured use of social “rewards,” with the variation being more in the extent of their consciousness in doing so. In this regard, Parsons’s logic invites the hypothesis that the conscious use of “rewards” tends to prevail when the role is strongly institutionalized.

29 This second example is in itself worthy of attention, [16] but a general conclusion applies: there are multiple fields of application for the mechanism(s) of social regulation highlighted by Parsons.

30 The two prominent cases described above are enough to give an idea of the heuristic fertility of the analyses that Parsons specifically devotes to social regulation. Nevertheless, it is useful to join Parsons in reexamining their epistemological status. Recalling the main stages of his reasoning at the beginning of chapter 11, Parsons emphasizes that in chapter 6 on socialization and in chapter 7 on deviance and social regulation, he offers not a theory in the full sense of the term (that is, “a system of laws”), but “a paradigm” and thus has to argue in terms of “mechanisms.” This recognition of what he sees as a limitation is nevertheless accompanied by a firm insistence on the theoretical “systematization” that made “recourse to the concept of the mechanism” possible (Parsons, 1951, 485).

31 It is worth pointing out that here, Parsons again displays nomothetic ambition, which with today’s emphasis on the “local” character of laws is at risk of appearing excessive. However, this is perhaps not what is most important: despite admitting that, in the current state of knowledge, an entirely general theory is impossible, Parsons strongly suggests that the path of generalization is open in the direction of systematization, if “a paradigm”—understood as “a set of canons for the statement of problems, in such terms as to ensure that the answers to the questions asked will prove to be of generalized significance” (Parsons, 1951, 485)—is available. This reflection allows a different conclusion from that reached by Parsons. Where Parsons’s ultimate objective continued to be the exposition of laws, it could justifiably be argued that the search for laws and “general theory” are not intrinsically linked. Renouncing the search for laws, as many now have (at least for universal laws), does not necessarily imply renouncing “general theory.” It is thus necessary to redefine the requirements of this theory, as well as the degree of generality sought.

32 Clearly, it is not possible to attempt such a redefinition within this article. However, it is fair to say that the search for and the elucidation of “social mechanisms” could be one of its priorities. It is therefore worth emphasizing that the “mechanisms of social regulation” identified by Parsons are fully compatible with Mario Bunge’s definition of a social mechanism as “a process involving at least two agents engaged in forming, maintaining, transforming, or dismantling a social system” (Bunge, 1997, 447). One might even add that Parsons’s analysis introduces a more subtle note: regulation helps to maintain the system, by ensuring the reintegration of people in emotionally difficult situations. Moreover, each of the submechanisms introduced by Parsons is understandable in the Weberian sense of the word: there is no “black box.” It seems that Parsons, although on different bases and by different routes, conforms to the fundamental principle proposed by Raymond Boudon. [17]

33 Overall, Parsons offers a triple lesson. Firstly, he provides a strong reminder that any description, and a fortiori any explanation, of social phenomena depends on the prior choice of a frame of reference. This frame of reference defines the perspective on reality, governing and organizing the way it is perceived. It seems to me that Parsons’s treatment of this first, indispensable transition to abstraction in The Structure of Social Action remains exemplary. Secondly, one of his original features is his invitation to elaborate analytical concepts by deconstructing ideal types and taking the process of abstraction beyond their inherent “utopian” rationalization. The multiple uses of pattern variables are enough to demonstrate the scope of the method, even if the combinations of variables are not all of the same interest for sociological analysis: some are anchored in sociohistorical reality, while others remain virtual. Undoubtedly, the Parsonian project retains some traces of the illusion that one can transcend history, despite the legitimacy of its primary aim, the elaboration of concepts and, in the long term, of a potentially generally applicable theory. Thirdly, Parsons examines a family of social mechanisms: the mechanisms of social regulation. Contrary to portrayals of him suggesting otherwise, he also shows an ability to elucidate the complex processes of readjustment (by actors) and of rebalancing (of collectivities and social systems). Admittedly, this clarification is incomplete, and the elucidation of the appropriate mechanisms is not, for Parsons, the central criterion in providing a true explanation leading to a pertinent theory.

34 Parsons does not and cannot offer a ready-made answer to the theoretical questions confronting us today. Nevertheless, he provides paths that are still worth exploring and taking seriously, and therefore worth debating. This is the type of contribution, in the end, that allows us to recognize a “classic.”