9 - Pattern Variables Revisited: A Response to Robert Dubin

I am grateful to Professor Dubin for the careful attention he has given to the somewhat neglected pattern variables and for his considerable effort in exploring their potential usefulness. His article has led to a serious reconsideration of the problems he has raised - in particular, the relation between what he refers to as Model I (the pattern variables as formulated in Toward a General Theory of Action¹) and Model II (the paradigm of four functional problems of systems of action from Working Papers,² and later publications). Dubin suggests that the usefulness of Model II is impaired by too drastic a condensation, and that it cannot be reconciled with Model I. The Editor's invitation to comment on his paper has given me the opportunity to work out an overdue clarification of the ways in which Model II builds on and goes beyond, rather than replaces, Model I.

Dubin is essentially correct in characterizing the pattern variables as a model that uses the unit act as its building block. The unit act involves the relationship of an actor to a situation composed of objects, and it is conceived as a choice (imputed by the theorist to the actor) among alternative ways of defining the situation. The unit act, however, does not occur independently but as one unit in the context of a wider system of actor-situation relationships; this system - including a plurality of acts - is referred to as an action system. The unit act is the logically minimal unit of analysis, but as such it can be conceived empirically only as a unit of an action system. Even for analysis of one discrete concrete act, an extended set of similar acts must be postulated as part of the action system - for example, those comprising a particular role. Figure 9.1 below is a paradigm for any such action system, not only the unit act.

The Frame of Reference

The pattern variables first emerged as a conceptual scheme for classifying types of roles in social systems, starting with the distinction between professional and business roles. In this sense, the concept "actor" referred to individual human beings as personalities in roles and the analysis - as Dubin puts it - "'looks' out to the social system from the vantage point of the actor." In Toward a General Theory, the scheme was substantially revised and its relevance extended from role-analysis in the social system to the analysis of all types of systems of action.

Pattern Variables Revisited  183

                Action is thus viewed as a process occurring between two structural parts of a system - actor and situation.  In carrying out analysis at any level of the total action system, the concept "actor" is extended to define not only individual personalities in roles but other types of acting units - collectivities, behavioral organisms, and cultural systems.  Since the term actor is used here to refer to any such acting unit, I attempt to avoid - except for purposes of analogy or illustration - psychological reference, for example, "motivation," attributed to actors as individuals. 

Thus "actor" can refer to a business firm in interaction with a household, or, at the cultural level, the implementation of empirical beliefs interacting with the implementation of evaluative beliefs.

                Both the pattern variables and the four system-problems are conceptual schemes, or sets of categories, for classifying the components of action.  They provide a frame of reference within which such classification can be made. The figures presented below indicate the methods, sets of rules and procedures, that state how these categories may be used analytically; they imply theorems - propositions that admit of logical, not empirical, proof - which state a set of determinate relationships among the categories and, in so doing, outline a theory of action.  The theory, then, is a set of logical relationships among categories used to classify empirical phenomena and, in empirical reference, attempts to account for whatever may be the degree of uniformity and stability of such phenomena.

                The pattern variables are a conceptual scheme for classifying the components of an action system - the actor-situation relational system which comprises a plurality of unit acts.  Each variable defines one property of a particular class of components.  In the first instance, they distinguish between two sets of components, orientations and modalities.  Orientation concerns the actor's relationship to the objects in his situation and is conceptualized by the two "attitudinal" variables of diffuseness-specificity and affectivity-neutrality.  In psychological terms, orientation refers to the actor's need for relating to the object world, to the basis of his interest in it.  For other levels of analysis, of course this psychological reference must be generalized.  Modality concerns the meaning of the object for the actor and is conceptualized by the two "object-categorization" variables of quality-performance and universalism-particularism. It refers to those aspects of the object that have meaning for the actor, given the situation.  The orientation set of pattern variables "views" the relationship of actor to situation from the side of the actor or actors; the modality set views it from the side of the situation as consisting of objects.  As Dubin suggests, the pattern variable of self-collectivity orientation does not belong at this level of analysis; it is placed in proper perspective below.

                In classifying the components of the actor's relation to a situation, the pattern variables suggest propositions about any particular action system in terms of those components and the type of act their combination defines; thus a particular role can be characterized by the properties of universalism, performance, and so on.  An action system, however, is not characterized solely by the actor's orientations and the modalities of objects significant to the actor; it is also a structured system with analytically independent aspects which the elementary pattern variable combinations by themselves do not take into account.

184  Sociological Theory

                In such a structured system both actor and object share institutionalized norms, conformity with which is a condition for stability of the system.  The relation between the actor's orientations and the modalities of objects in the  situation cannot be random.  The Working Papers established a nonrandom relationship between the two sets by matching the functionally corresponding categories on each side - universalism with specificity, particularism with diffuseness, performance with affectivity, and quality with neutrality.  This matching yielded Dubin's Model II. 

It turned out that this arrangement converges with the classification of functional problems of systems that Bales had earlier formulated.³  This convergence, the main subject of the Working Papers, opened up such a fertile range of possibilities that for several years my main attention has been given to their exploration rather than to direct concern with the scheme out of which it grew.   However, it is now clear that "Model II" is not a substitute for the earlier version, in the sense that it represents the whole scheme, but rather a formulation of one particularly crucial part of a larger scheme. The following discussion places that part in the context of the larger scheme as the formulation of "integrative standards," those aspects of the action system shared by actor and object and that make the system a stable one.

                In analyzing the components of any particular action system, one must also consider the larger system within which that action system is embedded.  The action system is related to the "external system" beyond it, which I refer to here as the environment of the system, as distinguished from the situation of the acting unit. The following analysis treats this relation of action system to environment as mediated mainly through the adaptive subsystem.  The combinations of pattern variable components in that subsystem were foreshadowed in the Working Papers by the "auxiliary" combinations of neutrality-performance, particularism-specificity, and so on.⁴  The present paper, I believe, establishes the analytical independence of these combinations from those of the integrative standards in Model II, and goes considerably beyond the Working Papers in setting forth their significance for action systems.

                Finally, the pattern variables - although they designate the properties of actor's orientations and objects' modalities in an action system - do not as such classify types of actors and objects.  Such a typology cannot be derived from any particular action system, but only from the analysis of a range of such systems. It is this typology of actors and objects with which Dubin's left- and right-hand columns in his Table 1 is concerned.  Figure 9.2 below has incorporated this important aspect of Dubin's problem.

With reference to Dubin's Table 1, the pattern variables themselves are discussed under what he terms the "actor's evaluation of objects."  The column headed "Modalities of Objects" is admittedly redundant, for in addition to the redundancies noted by Dubin, the terms "classificatory" and "relational" are synonymous with "universalism" and "particularism," respectively, as I acknowledged in The Social System.   In my Figure 9.2, Dubin's "motivational orientation" towards objects is covered by the pattern-maintenance or orientation subsystem; his "value-orientation" by the adaptive subsystem; and his "action-orientation" is characterized by the types of output of the system as a whole (see below).

Pattern Variables Revisited 185

186  Sociological Theory

                Thus the conceptual scheme of the four system-problems has added a set of rules and procedures - the basis of theorems - whereby the analysis of components of action in terms of pattern variables can be carried out by "looking down" on them, as Dubin has aptly put it, from the perspective of the action system.  The action system is presented in Figure 9.1 above so as to establish the analytical independence of the four subsystems: orientations (pattern-maintenance); modalities (goal-attainment); their combination characterizing the conditions of internal stability of a relational system shared by both actor and object (integration); their combination characterizing the ways in which that system is stably related to the environment (adaptation).

                Following the presentation of these four subsystems, the same information is displayed in tabular form different form the more familiar functional "layout."  This second presentation (Figure 9.2) is designed to "look down" on any particular action system from the perspective of the more inclusive system.  At this level, the analysis of types of actors and of objects can be carried out. In addition, Figure 9.2 highlights the distinction between the control of action - that is, the scale of priorities assigned to various ways of regulating action - and the implementation of action - the analytical relevance involved in the distinction between structure and process.

                This then is the main frame of reference of the paper's approach to the classification and analysis of the components of action. We now turn to the paradigm itself, which is altogether newly formulated from the point of view of the internal relations between its components, and is presented in Figure 9.1.  Its form is essentially that of Dubin's Table 4, which was derived from the Working Papers.5 "Model II" is treated in the paradigm as the integration subsystem of the general system. The pattern variable scheme as formulated in Toward A General Theory, that is, the two "attitudinal" and "object-categorization" sets, are incorporated into the "pattern-maintenance" and the "goal-attainment" subsystems, respectively. To avoid terminological confusion we follow Dubin in referring to the two sets of pattern variables as the orientation set and the modality set. The fourth block of cells, representing the adaptation subsystem, is also entirely new, and is explicated below.

                We have noted above that the primary reference of the concept "actor" is to the individual personality, but that in secondary respects, collectivities, behavioral organisms, and cultural systems may be conceived as actors.

It is important to remember that our scheme concerns the generalized components of action, so that such psychological terms as "cathexis" and "identification" and "need," as used here, stand for more generalized concepts than would be applicable to actors and objects on these other levels; their reference is not confined to the personality level.

Pattern Variables Revisited  187

The Orientation Set (Pattern-Maintenance)⁶

The orientation base of a system of action may be categorized in terms of the two pattern variables, affectivity-neutrality and specificity-diffuseness.  The relevant characteristic of the actor in defining his (or "its") orientation to an object or  category of objects may be an "interest" in the object as a source of "consummation." 

This may be defined as an interest in establishing a relation to an object, which the actor has no incentive to change. In psychological terms, this may be phrased that the actor has a "need" for such a relationship, which can be "gratified" by its establishment.  The alternative to the need for a consummatory relationship is the "need" for help toward the attainment of such a relationship to an object.  Therefore, besides the consummatory, there is an instrumental basis of orientation to the object-world.  At this point a pattern-variable "dilemma" arises because it is a fundamental assertion of our theory that consummatory and instrumental interests in objects cannot be maximized at the same time. 

The instrumental and consummatory bases are analytically independent.

                The very discrimination of different bases of orientation of actors to objects implies that actors are conceived as systems; they are never oriented to their situations simply "as a whole," but always through specific modes of organization of independent components.  From this point of view, it is always important whether the primary reference is to the relation of the acting system to its environment or to its own internal properties and equilibrium.  The situation, or object-world, is in the nature of the case organized differently from the actor as system.  Hence, in orientation directly to the situation, the specificities of differentiation among objects and their properties become salient.  On the other hand, where internal "needs" of the acting system are paramount, the salience of these specificities recedes, and   the orientation to objects becomes more diffuse.  This is the setting in which the specificity-diffuseness variable fits. 

It indicates that where the "interaction surface" between actor and situation is approached, the actor's interests in objects must be more highly specified than where internal states of the acting system itself are in the forefront.

                There is a pattern-variable dilemma here as well as in the instrumental- consummatory case.  This is to say that the imperatives of specificity and of diffuseness cannot be maximally satisfied at the same time.

                The cross-classification of these two orientational pattern-variables yields a four-fold table which is presented as the pattern-maintenance subsystem (L) of Figure 9.1.  As distinguished from the pattern variables themselves, which are rubrics of classification, this constitutes a classification of types of orientation to objects.  This distinction has not always been clear, I believe, neither in my own work nor in that of other writers.

                It will be seen that the pure type of "consummatory needs" combines affectivity and specificity of interest; it is "pure" because it can focus on the actor's relation to the specific discretely differentiated object.   But where the basis of interest is diffuse, there must be generalization to a broader category of objects, so the basis of the interest is the establishment of a relation between the acting system and a wider sector of the situational object-system.  We have called this a "need for affiliation," for example, for a relation of mutual "solidarity" between diffuse sectors of the acting system and the object-system. 

188  Sociological Theory

                On the instrumental side, it is apparent that the same order of distinction applies to specifically differentiated bases of interest in objects and diffuser bases.  Manipulation of objects in the interest of consummatory gratification or

even passive adaptation to them requires concern with the specificities of their properties.  Hence the "interest in instrumental utilization," though affectively neutral, is also specific; interest in the category is not enough.  Where, however, the problem is not utilization, but the place of the orientation in the internal structure of the acting system, this level of the specification of interest not only is unnecessary but, because of the independent variability of the object-situation, becomes positively obstructive.  Commitment to the specifics of object-situations introduces a rigidity of orientation which can be highly constrictive. Commitment can be and, functionally speaking, is better organized on a diffuser level. We therefore speak of "needs for commitment" as oriented to diffuse categories of objects and their properties rather than to specific objects and properties, and as engaging more diffuse sectors of the acting system than do "interests in instrumental utilization."

The Modality Set (Goal-Attainment)

With reference to the obverse side of the action relationship, that of the modalities of objects, the modality set of pattern variables constitutes the classificatory framework  particularism and universalism, and performance and quality.  Particularism in this context means that from the point of view of the action system, the most significant aspect of an object is its relation of particularity to the actor: as compared with other objects which can "intrinsically" be classified as similar to it, the significance of this object to the actor lies in its inclusion in the same interactive system. In the contrasting case of universalistic modalities, the basis of an object's meaning lies in its universalistically defined properties, hence its inclusion in classes which transcend that particular relational system. For example, when a man falls in love, it is this particular woman with whom the love relationship exists. He may, like some other gentlemen, prefer blondes, but he is not in love with the category, but with one particular blonde. Thus the same kind of dilemma exists here as for the two pattern variables described above - it is impossible to maximize the particularistic meaning of objects and their universalistic meaning at the same time. A man sufficiently in love with blondeness as such, who therefore pursues any blonde, cannot establish a very stable love relationship with a particular woman. That there is an important "matching" between consummatory bases of interest and particularistic meaning of objects is clear; its significance is discussed below.

                A basic postulate of action theory is that the status of acting systems and those of the situational object-world in which they act are independently variable. At their "interface," then, an especially important property of objects is their probable performance in respect to the actors oriented to them.  Recall that the prototype of the actor-object relation is social interaction, in which the "object" is also in turn an actor who does something. Thus physical objects, which do not "act," are the limiting case of objects to which the term "performance" is inherently inapplicable.        

In contrast with this situation, is the meaning of objects in terms of what they "are," of their qualities defined independently of performances, which are

Pattern Variables Revisited  189

inherently relative to situations. The internal reference of the acting system matches with interest in the qualities of objects rather than their performances, since these are presumptively more independent of direct situational exigencies.

                These two classificatory rubrics - performance-quality and universalism- particularism - yield a four-fold typology of objects (or of components), seen from the perspective of their meaning to actors. This is the goal-attainment subsystem (G) in Figure 9.1. This terminology is also adopted from the prototypical case of interaction of persons. Thus an object whose primary meaning is particularistic and based on its actual and expected performances, following psychoanalytic usage, may be called an "object of cathexis." It is "looked at" in terms of its potentialities for gratifying specific consummatory needs. However, if an object is defined in universalistic terms, but at the same time as a source of performances significant to the actor, it can be said to be an "object of utility," for it is viewed with respect to its potentialities in helping to bring about consummatory states of the acting system.

                In contrast with both these types, objects may be treated as "objects of identification" if their meaning is both particularistic and refers essentially to what they "are" rather than what they "do." Here the objects' meaning to actors is not subject to the more detailed fluctuations which go with the meaning of cathexis.

                Finally, the universalistic case, the fourth type, is called an "object of generalized respect." Here the object is categorized by the actor in universalistic terms, but also with relation to its qualities. This is the type of object which in a social context Durkheim speaks of as generating attitudes of "moral authority."7

Problems of Integration and Adaptation

The argument so far may be summarized: We have outlined, in terms of the present conceptual scheme, the elementary components of action and certain aspects of their interrelations. Essentially these are the components of unit acts but do not yet comprise systems of action.

First, we have assumed that all action involves the relating of acting units to objects in their situation. This is the basis for the fundamental distinction between components belonging to the characterization of orienting actors and those belonging to the modalities of the objects to which they are oriented - that is, between the two "sets" of elementary pattern variables.  Second, we have used the elementary variables to classify types of elementary combination. The underlying assumption here is that on this level they are always analytically independent; hence the orientation set (cluster L of Figure 9.1) and the modality set (cluster G) are treated as mutually exclusive, each type being composed of components drawn only from one of the two sets.  Third, each cell within each cluster is composed of only two pattern variable designations.  Fourth, what elsewhere are defined as "pattern variable opposites" never occur in the same cell.  Subject to these rules, the classifications designated by the four cells in each cluster are logically exhaustive of the possibilities. We consider the fourth assumption to be the application of a fundamental theorem concerning the

190 Sociological Theory

conditions of the stability of orientation, namely, that neither the same orientation nor the same object can be successfully defined, in a particular context or orientation, in terms of both alternatives without discrimination, for example, universalistically and particularistically or specifically and diffusely at the same time.

                Subject to these constraints, however, we see no reason why the composition of possible types of unit acts do not exhaust the range of logically possible independent variation of the components thus formulated. But such a definition does not tell us anything about the conditions of the existence of a system of such unit acts other than that there are such limiting circumstances as physical and biological conditions of survival. In other words, this level of analysis describes a population of action-units and certain of the ways in which they are empirically ordered in relation to each other. It cannot provide an analysis of the relations of their interaction, which constitute a system subject to mechanisms of equilibration and change as a system through "feedback" processes - in one sense, the organization of the system.

                To take the step to this organizational level, it is necessary to attempt to conceptualize two basic sets of "functions" which cannot be treated either as the orientations of actors or as the meanings or modalities of the objects to which they are primarily oriented. These are, first, the modes of internal integration of the system, that is, of the interrelations of the elementary actor-object units. This means, within our frame of reference, the normative standards on the basis of which such relations can be said to be stable. Second, there are the mechanisms by which the system as a whole is adapted to the environment within which it operates. Since from the point of view of orientation this environment must consist in some sense of objects, the problem is that of conceptualizing the relation between objects internal to the system and those (albeit in some sense meaningful) external to the system.

                To repeat, those reviewed above constitute the full complement of elementary components of action systems. Therefore, in dealing with these two additional system functions or subsystem clusters, we do not propose to introduce additional elementary components, but rather to suggest new combinations of these components. On this basis the I and A clusters of cells in Figure 9.1 are constructed on the hypothesis that each cell of the two clusters should be defined by one pattern variable component drawn from each of the two elementary subsets. If this policy and the general rules formulated above are followed, the combinations represented in the two clusters will be logically exhaustive of the possibilities.

                Within these rules the problem is that of the basis of allocation of the components as between the two clusters, and within each as between the cells. The governing principles for treating this problem are more fully elucidated below, following a review of the allocations themselves and some problems of the system as a whole. Here, suffice it to say, first, that internal integration is dependent on the matching of the function of the object for the "needs" of the orienting actor with the functional meaning with which the object is categorized. Thus in some sense the gratification of consummatory needs is dependent on the possibility of categorizing appropriate objects as objects of cathexis, and so

Pattern Variables Revisited  191

on.  Why only two of the four components which might define this matching are involved, and which two, are also explained below.

                 Secondly, the significance of objects external to the system is not their actual meaning in the system, but rather their potential meaning for the system - the ways in which taking cognizance of this meaning or failing to do so may affect the functioning of the system. With these preliminaries, we may now review schematically the actual content suggested for the cells.

The Integrative Subset

How are the formal characteristics of the I and A cells in Figure 9.1 to be interpreted? The integrative subset states the primary conditions of internal stability or order in an action system. These conditions may be formulated as follows:

(1) In so far as the primary functional problem of the system, conceived either in terms of structural differentiation or temporal phases, is adaptive, stability is dependent on the universalistic categorization of the relevant objects, regardless of whether or not they have certain particularistic meanings, and on sufficient specificity in the basis of interest in these objects to exclude more diffuse considerations of orientation.

(2) In so far as the primary functional problem is the attainment of a goal for the system, stability is dependent on attention to the potentialities of performance of the object in its relation to the actor, and on affective engagement of the actor in the establishment of the optimal (consummatory) relation to the object - hence the lifting  of "inhibitions" on such engagement. 

(3) In so far as the primary functional problem is integration of the system, stability is dependent on particularistic categorization of the relevant objects (that is, to the extent that they are also actors, their inclusion in the system), and the maintenance of a diffuse basis of interest in these objects (that is, one which is not contingent on fluctuations in their specific performances or properties).

(4) In so far as the primary functional problem for the system is the maintenance of the pattern of its units, stability is dependent on maintaining a categorization of the objects in terms of their qualities independently of their specific performances, and an affectively neutral orientation, one that is not alterable as a function of specific situational rewards.

                In terms of the regulation of action, these combinations of pattern variable components define categories of norms governing the interaction of units in the system. Norms themselves must be differentiated. It is in the nature of an action system to be subject to a plurality of functional exigencies; no single undifferentiated normative pattern or "value" permits stability over the range of these different exigencies. Hence norms constitute a differentiated and structured subsystem of the larger system. They constitute the structural aspect of the relational nexus between actors and objects in their situations.

                Precisely because the above propositions state conditions of stable equilibrium involving the relations between a plurality of elementary components, I believe that they go beyond description to state, implicitly at least, certain theorems about the consequences of variations in these relations. These theorems are considered following the discussion of the system itself.

192  Sociological Theory

The Adaptive Subset

 In the adaptive subset, the formal bases of selection of the component combinations, as we have noted, are antithetical to those used in the integrative subset. This is to say that they combine both external and internal references, and both instrumental and consummatory references.

                We have termed these combinations as defining "mechanisms" for ordering the adaptive relations of a system of action to the environment in which it functions. To clarify this problem an important distinction must be made. When we referred above to the orientation of actors to objects and the related modalities or meanings of objects, we were indicating components internal to a system of action. Objects that are constituents of the system must, however, be distinguished from objects that are part of the environment of the system. The boundary concept which defines this distinction is "particularism;" an object categorized particularistically is defined as belonging to the system. Adaptation concerns the relations of the whole system to objects which, as such, are not included in it.

                Adaptive mechanisms, then, must be conceived as ways of categorizing the meanings of objects universalistically, that is, independently of their actual or potential inclusion in a given system. These mechanisms are "symbolic" media, including language as the prototype, but also empirical knowledge, money, and so on. Use of the media for referring to objects and categories of objects does not ipso facto commit the actor to any particular relation of inclusion or exclusion relative to the objects concerned. By use of the media, however, meanings may be treated as internal to the system, whereas the objects themselves may or may not remain external. This is the basic difference from modalities, which are meanings wherein the objects themselves are defined as internal.

                In this context, the pattern variable combinations of the adaptive subset may be explicated as follows: 

(1) In order to symbolize the adaptive significance of objects in the environment of an action system (for example, to "understand" them cognitively), it is necessary to categorize them in terms of what actually or potentially they "do" (performance), and to orient to them with affective neutrality, that is, independently of their potentialities for gratifying the actor. This "pattern" is defined as a condition for stability of an orientation to the external environment which can maximize  objective" understanding of the objects comprising it; adopting a term from personality analysis we may term the pattern empirical "cognitive symbolization."

(2) In order to symbolize and categorize objects that are external to the system according to their significance for goal-attainment, it is necessary to focus their possible meaning on specific bases of interest or "motivation" (specificity), and on their potential "belongingness" in a system of meanings which also defines the system of action (particularism). This we call "expressive symbolization," the generalization of particularistic meanings to a universalistic level of significance.

(3 In order to symbolize and categorize the significance of norms that are external to the system, it is necessary to treat them as aspects of an objectively "given" state of affairs or "order" (quality), and to treat them with affectivity - that is, the

Pattern Variables Revisited 193

actor cannot be emotionally indifferent to whether or not he feels committed to the norms in question. This we name "moral-evaluative categorization." 

(4) In order to symbolize and categorize the significance of "sources of normative authority," it is necessary to combine a universalistic definition of the object, as having properties not dependent on its inclusion in the system, with a diffuse basis of interest, so that the meaning in question cannot be treated as contingent on the fluctuating relations between the orienting actor and the environment. This we call "existential interpretation."

                Here another version of the external-internal distinction is important. For the first two of these  the adaptive and goal-attainment categories - refer to objects considered as such, irrespective of whether or not they are included with the acting system within a more comprehensive system. In the latter two cases, however, this question of common membership in a more comprehensive system is central. A norm is binding on a unit only in so far as the unit shares common membership with other units similarly bound. An object is a source of normative authority only so far as its authority extends to other units, defined universalistically as similarly subject to that authority. It is on these grounds that we emphasize "symbolization" in the first two cases and "categorization" in the second two.

                Note that the differentiation of symbolic media according to functional significance parallels the differentiation of integrative standards. They too are results of a process of differentiating the components involved in the elementary pattern-variable sets and of integrating the selected components across the orientation-modality line. As distinguished from the internal integration of the system, the adaptive subset refers to the system's integration with its environment as part of a more comprehensive system of action.

The Perspective of the System as a Whole

So far we have considered the elementary components which make up a system of action and two main ways in which they are related across the orientation modality line. These components and relations, however, constitute a system which in turn functions in relation to what we call an "environment." We now consider a few aspects of the properties of this system in its environmental context. The main reference point for this analysis is a rearrangement or transformation of the items of Figure 9.1, as presented in Figure 9.2.

                The components in Figure 9.2 are the same sixteen pattern variable combinations represented in Figure 9.1. However, there are two new features of the arrangement: First, each of the four major blocks of cells of Figure 9.1 is set forth as a column of Figure 9.2. Within each column the cells in turn are arranged from top to bottom in the order L-I-G-A. This constitutes a cybernetic hierarchy of control,⁸ that is, each cell categorizes the necessary but not sufficient conditions for operation of the cell next above it in the column, and in the opposite direction, the categories of each cell control the processes categorized in the one below it. For instance, definition of an end or goal controls the selection of means for its attainment.

194  Sociological Theory

STRUCTURAL CATEGORIES                                      CATEGORIES OF PROCESS

Units of Orientation                Integrative                            Symbolic Representation                Internal Meanings                Outputs

to Objects (L)                            Standards (I)                       of External                 of Objects (G)                to

(Properties of Actors)                                                   Objects (A)                                (Inputs-Outputs)                Environ

L              Neut                                       Qual                                       Dift                                          Univ

                Diff                                          Neut                                       Univ                                        Qual

Ground-of-meaning

Ancnorage

                NORMATIVE                                                                        EXISTENTIAL

                COMMITMENTS                   PATTERN-                            INTERPRETATION                “RESPECT"

MAINTENANCE

I               Aft                                            Part                                         Aft                                            Part                Responsib          Dift                                          Manifold of  Dift                     Qual                                       Qual                Action

evaluative

                                                                selections

                AFFILIATIONS                       INTEGRATION                    MORAL-                 IDENTIFICATION

                                                                                                                EVALUATION

                                                                Allocative selection

G             Aft                                            Pert                                         Spec                                       Pert                         Expressive

                Spec                                       Rangeof                Aft                            Part                                         Part                         Action

                                                                action-choice

                CONSUMMATORY                  GOAL (attainment)                EXPRESSIVE                       CATHEXIS

                NEEDS                SELECTION                SYMBOLIZATION

A             Neut                                       Empirical-Univ                   Neut                                       Pert                         Instrument

                Spec                                       cognitive                  Spec                     Pert                                         Univ                        Action

                                                                field

                INSTRUMENTAL                                ADAPTATION                       COGNITIVE

                CAPACITIES                        Means-Selection                              SYMBOLIZATION                UTILITY

^ ---------- Direction of Limiting Conditions

ç===== Direction of Implementation vis-a-vis Environment

                Direction of Environmental “Stimulation”  =====è

Figure 9.2 The action system in relation to its environment

                The second difference from Figure 9.1 is the arrangement of the columns from left to right in a serial order which, stated in functional terms, is L-I-A-G.

The two left-hand columns designate the structural components of the system. 

The L column formulates the properties of units conceived as actors;

the I column formulates the structural aspect of the relational nexus between units, that is, the norms which function as integrative standards.

The two right-hand columns categorize the elements of process by which the system operates.

The G column shows the modalities of objects from the point of view of change of meaning as a process of relating inputs and outputs; it brings into the system meaning-categorizations generated by the system.

The A column formulates the components involved in the symbolic mechanisms mediating the adaptive aspect of process.

Whereas the hierarchy of control places the A subset at the bottom of each column, as a column itself it is placed "inside" the system because it consists of a set of symbolized meanings (or "representations") of the environmental object-world outside the system, or the categorization of objects independently of their inclusion in or exclusion from the system. It therefore constitutes the internal environment of the system, the environment to which units must adapt in their relations to each other, but the actual objects symbolized constitute the external environment to which the system as a whole must adapt.

                We have suggested that the outputs of action systems consist in changes in the meanings of objects. It follows that the inputs also consist in meanings of objects. What the process of action accomplishes, then, is change in these meanings. We assume of course that new objects and categories of objects are created in the process; these presumably are themselves action systems and their "cultural" precipitates. The distinction between changing the meaning of an old object and creating a new object thus appears to depend on the point of observation.

                The modalities of objects in the G column of Figure 9.2 therefore may be treated as a classification of the outputs of internal action process, in a sense similar to the usage in economics of "value-added."9 Thus action process, so far as it is effectively adaptive internally, may be said to add utility to objects  for example, utility in the economist's sense, the relevant category for social sys- tems, also is a category of meaning in the present context. Action which is successfully oriented internally to goal-attainment leads to the enhanced cathec- tic value of objects in the system. Action which is successfully integrative leads to increased ~~identification-meaning" - in social systems, to solidarity with and among objects. Finally, processes of "pattern-maintenance" maintain or restore the "respect" in which the relevant system itself is held as an object in the social system; here is Durkheim's "the integrity of moral authority."

                The designations to the right of the G column in Figure 9.2 are the "action- orientations" in the Orientation column of Dubin's Table 1. We suggest that these can be treated as categories of output to its environment of the system as a whole (as distinguished from the outputs of internal process). Thus instrumental action by a system may be treated as resulting in increase in the instrumental values to it of objects within its environment or more inclusive system. Similarly, expressive action produces enhanced cathectic meaning of objects in the envir- onment; and responsible action increases the integrative identification category of meaning (for example, in the social system, "moral" value). In accord with principles we have used consistently,'0 we suggest that there is no category of output for the L subsystem except in cases of change in the structure of the system.

The Classification of Objects

One further set of categories which play a part in Dubin's Table 1 needs to be accounted for - the classification of types of object as physical, social, and

196  Sociological Theory

cultural. This problem can most conveniently be treated at the environmental level. If a given system is conceived as an actor or an action system, then a system with which it interacts is a social object. We have explained why this category should be differentiated into at least two subcategories: the system organized about the single human individual, namely, personality; and the social system constituted by the interaction of a plurality of individuals. A physical object, then, is one with which the system does not in this sense interact, and which, standing below the action system in the hierarchy of control, is condi- tional to it; a cultural object is also one with which it does not interact, but which stands above it in the hierarchy of control, and therefore is a focus of its own control system.

                However, a further principle is involved, not developed here, of inteipenetra- tion ~f5y5~~~5.ii The crucial case of physical systems with which the personality interpenetrates is the behavioral organism, the physical system which consti- tutes the fundamental facility-base for the operation of the personality system. At the other extreme are "acting" cultural systems, implemented through social and personal actions, which constitute the operating normative control systems of social systems. At each "end" of the control series, then, is a set of limiting conceptions of nonaction "reality." At the lower end is "purely physical" reality with which the action system does not interpenetrate, but which is only conditional to it. At the upper end is "nonempirical," perhaps "cosmic," reality with which, similarly, there is no significant interpenetration, and which is thus conceived only as an "existential ground" of operative cultural systems.

                A similar classification can be worked out for the alternative case where the system in question is conceived as acting, and not as an object. Here it seems that the parallel to a cultural object is the conception of the "subject" as "knowing, feeling, and willing." At the social level, this is our concept of "actor" in the sense of participation in interaction. At the interpenetrating subsocial level, it is the concept of organism, as "functioning" in relation to an environment. Perhaps at a still lower level should be placed the "hereditary constitution" of a species (as distinguished from the particular organism in phylogenetic, not onto genetic terms).

Combinations of the Components

We now return to the question of the bases of combination and allocation of the pattern variable components. A maximum number of types could be generated of course by treating the potential combinations as all those randomly possible. This procedure, however, would mean the sacrifice of connections referred to above as the organization of systems of action and the determinate theoretical generalizations associated with them.

                We have restricted random combinations, first, by composing two cell clusters (L and G) exclusively from one or the other of the elementary sets; second, by never placing both members of a "dilemma" pair in the same cell; third, by placing only one component from each elementary set in each cell of the I and A

Pattern Variables Revisited  197

clusters; and, finally, by drawing these from "functionally cognate" cells of the elementary combination paradigms. (See Figures 9.1 and 9.2.) Within these rules of organization we have followed a further policy of selection in the allocations to the I and A clusters. In terms of the "geometry" of Figure 9.1, this policy involves two procedures: (1) for the I cluster, the distribution of the modality components is derived by keeping the "functionally cognate" refer- ence constant and then rotating clockwise the modality axes one quarter turn, and the distribution of the orientation components is similarly decided by rotating the orientation axes in the counterclockwise direction; (2) for the A cluster, the direction of rotation is the reverse in each case. Thus, in the G cluster the distinction between universalism and particularism defines the hor- zzontal axis of the paradigm, in the I cluster it assumes the diagonal. Put other- wise: of the two occurrences of each component in the G table only one of each is included in the I table, and these are placed in a diagonal position. The effect of this is to "shift" the relevant category from one to the other of the two positions in which it could be placed in the elementary set. The procedure never leads to "crossing over" into a "forbidden" cell; for example, universalism and particu- larism never "change places."

                What is the meaning of these patternings? It is inherent in the organization of Figure 9.2 that integrative functions stand higher in the order of control than either goal-attainment or adaptive functions, which follow in that order. On grounds that cannot be fully explained here, I suggest that the horizontal and vertical axes of the paradigm state the location of the processes, conceived as inter-unit interchanges, which, respectively, have primarily internal adaptive significance in providing facilities to the units in question, and internal goal- attainment significance in providing rewards. Thus, the "rotation" brings about an involvement of the pattern variable components in integrative interchanges along the axes of Durkheim's "mechanical" (L-G) and "organic" solidarity (A4).12

                The suggestion, then, is that, relative to the elementary clusters, both I and A clusters have integrative significance. The I set states internal integrative stand- ards, departure from which is associated with those realistic internal con- sequences known in interaction theory as "negative sanctions." The A set states standards of meanings of external objects ("cultural standards"), departure from which is associated with cultural selectivity and distortion, although not with immediately felt "sanctions."

                What of the obverse "directions" of rotation? There is a double incidence of these directionalities. Within the clusters the rotations of the axes of the orienta- tions and of the modalities are in opposite directions. The modalities of objects, from the point of view of a system of action, constitute ways of relating not only the acting unit but the system to the environment external to it. Hence it is an imperative of integration that, from the modality side, priority should be enjoyed by the category of meaning of the object (internally, as defining the actor-object relation) which is of primary functional significance for the system in the relevant context. From the orientation side, the imperative is that priority goes to the mode of orientation of primary significance to the actor in terms of its "needs." Thus, if the system function in question is adaptive, universalistic

198  Sociological Theory

meanings take precedence over particularistic. For the actor, then, the primacy of specificity may be regarded as protecting his interest in other contexts of meaning of the same and other objects by limiting his commitments to the more immediately important ones.

                These two designations are "functionally cognate" in that they share the characteristics of external orientation and instrumental siguificance. Here the rotation means that on the A-I axis of the integrative cluster (not of the system as a whole) the modality component in the adaptive cell is related to what in the G cluster is its consummatory "partner," whereas the orientation component is related to its internal partner. This is simply another way of stating the obverse directions of rotation. Put in general functional terms: the obverse relationship protects the system by giving primacy to instrumental over consummatory considerations in the adaptive context, while it protects the actor by giving primacy to external over internal considerations.

                Another example from the adaptive cluster pairs the integrative cell with affectivity. From the viewpoint of the system, the significance of the object as "internalized" or institutionalized must clearly take precedence over its varying performances as oriented to the external situation. For it to serve as a standard of moral-evaluative categorization, however, there must also be affective in- volvement. The rotation in this case means that categorization in terms of quality is specifically distinguished from the performance component in its application to cognitive symbolization, whereas affectivity is contrasted (and thus integrated) with neutrality in the cognitive context. The formula for evaluative categorization on the modality side therefore designates internal significance, on the orientation side, consummatory significance.

                The "diagonal" relations of the pattern variable pairs in the I and A clusters thus formulate the relations of combined discrimination and balance between the modality components and the orientational components. In each case the balance "protects" the categorization from confusion with its pattern variable opposite.

                The same essential principles hold when the functioning of the system as a whole is considered. Here rotation in the clockwise direction designates what psychologists often call "performance" process, that is, change in the relations of the system to its environment on the assumption that its internal structure remains unchanged. The primary focus of change in this case lies in the adaptive subsystem. The counterclockwise direction of process designates "learning" processes. Here the primary focus of change centers in the internal structure of the system, in the first instance in the integrative system producing a change in its standards.

Types of Action and the Organization of Components

Another theoretical issue requires brief comment. This concerns the fact that the present analysis is mainly an analytical classification of components of any system of action, including the "unit act" as the most elementary building block of action systems.13 Dubin, however, speaks of types of act. From the present

Pattern Variables Revisited  199

 point of view types must be constructed of varying combinations of compon- ents. In addition to composition - in terms of the presence or absence of com- ponents, or different "weights" assigned to them - there is organization of these components. We interpret the restrictions on random combination, and the clustering of pattern variable combinations in the four functional sets, to be statements of organization. The state of a system is never, in our opinion, adequately described by its "composition" - that is, by what components are present in what quantities; the patterns of their relationships are equally essen- tial. These considerations should be taken into account in attempts to develop a typology of acts from a classification of components in the act.

                Another relevant point concerns the status of the pattern variable, self versus collectivity orientation. My present view is that this was an unduly restricted formulation of an element in the organization of action components at the level next above that designated by the primary pattern variables. In fact, Figure 9.1, I believe, documents four levels of organization. The first of these is represented by the L and G cells, characterized by pairs of elementary pattern-variable components - resulting in orientations and modalities, respectively. The second level is represented by the cross-combinations of elements from each pattern variable set, as shown in the I and A cells; as noted above, these are necessitated by the exigencies of differentiation and integration of the elementary combina- tions. The third level is the combination in turn of all of these elements into the four subsystems which have functional significance for the system as a whole, while the fourth is the organization of the system as a whole in relation to its environment.

                The problem of the self-collectivity variable arises at the point where the I and A cells are organized into their respective subsystems. Subunits are organized into higher order "collective" units, the prototype being the organization of "members" into social collectivities. This organization takes place along the axis which distinguishes the "external" and "internal" foci in these cells. The inference is that there is another concept-pair which formulates the other axis of differentiation. In the I and A cells this is termed the "instrumental- consummatory" axis, which should be placed on the same analytical level of generality as the former pattern variable.

                The difference, I believe, between the two primary pattern variable sets and this other "secondary" set - internal-external and instrumental-consummatory - is one of level of organization. The secondary set formulates the bases of relationship across the two primary sets, as distinguished from relations within each.

Some Theoretical Propositions

These restrictions on combinatorial randomness logically imply certain general propositions about the modes of inter-connecting the components of a system of action. As distinguished from the exposition of a frame of reference, these are theoretical propositions or theorems. We are not sure that all propositions which can be derived from the logical structure of the system have been exhaustively

200  Sociological Theory

worked out, even at this very high level of generality. But the following propositions seem to be the most significant:

                1. The nature of the hierarchy of control, running from the cultural reference at the top of Figure 9.2 to the physical at the bottom, indicates that the structure of systems of action is conceived as consisting in patterns of normative culture. The ways in which types of action system are differentiated, then, means that these patterns may be conceived as internalized in personalities and behavioral organisms, and as institutionalized in social and cultural systems.

                2.                It follows from this first proposition, plus the exposure of any system of action to plural functional exigencies, that the normative culture which constitutes its structure must be differentiated relative to these functional exigencies. These differentiated parts must then be integrated according to the four standards formulated in the I cells of Figure 9.1, and action oriented to the four different standards must be appropriately balanced, if the system is to remain stable. This is to say that process in the system, if it is to be compatible with the conditions of stability, must conform in some degree with the rules of a normative order, which is itself both differentiated and integrated.

                3.                For this "compliance" with the requirements of normative order to take place, the "distance" must not be too great between the structure of the acting unit and the normative requirements of its action necessitated by the functional exigencies of the system. It follows that the structure of acting units (which are objects to each other), as well as of norms, must incorporate appropriate elements of the system of normative culture - involving the internalization of "social object systems" in personalities, and the institutionalization of culturally normative systems in social systems.

                4.                Coordinate with the importance of order as formulated in the hierarchy of control and the place of normative culture in action Systems, is the pattern of temporal order imposed by the functional exigencies of systems. Coordinate with the normative priority of ends is the temporal priority of means; only when the prerequisities of a consummatory goal-state have been established in the proper temporal order can the goal-state be realistically achieved. In both Figures 9.1 and 9.2 process is thus conceived in temporal terms as moving from left to right, the direction of "implementation."

5.                   A "law of inertia" may be stated: Change in the rate or direction of process is a consequence of disturbance in the relations between an actor or acting system and its situation, or the meanings of objects. If this relational system is completely stable, in this sense there is no process which is problematical for the theory of action. Whatever its source, such disturbance will always "show up" in the form of "strain" or difficulty in the attainment of valued goal-states. From this point of reference may be distinguished two fundamental types of process:

                (a)                "Performance" processes: These are processes by which the disturbance is eliminated or adequately reduced through adaptive mechanisms, leaving the integrative standards - the most directly vulnerable aspect of the structure of the system - unchanged. The process may be adaptive in either the passive or the active sense, that is, through "adjusting to" changes in environmental exigencies or achieving "mastery" over them. The basic paradigm of this type

Pattern Variables Revisited 201

of process is the means-end schema. In Figure 9.1 the directionality of such process is clockwise relative to the goal-focus, from A to G.

                (b)                'Learning" processes or processes of structural change in the system: here, whatever its source, the disturbance is propagated to the integrative standards themselves and involves shifts in their symbolization and categorization and in their relative priorities. Whereas in performance processes goals are given, in learning processes they must be redefined. Relative to the goal-focus, then, the directionality of such process is counterclockwise, from I to G in Figure 9.1.

6.                   To be stable in the long run, a system of action must establish a generalized adaptive relation to its environment which is relatively emancipated from the particularities of specific goal-states. To preserve its own normative control in the face of environmental variability, it must be related selectively to the environment.

There are two primary aspects of this adaptive relationship:

(a) the level of generality of symbolic or "linguistic" organization of the orientation to environmental object-systems (the higher the level of generality the more adequate the adaptation); and

(b) the ways in which the boundary of the system is drawn in terms of inclusion-exclusion of objects according to their meanings. The latter is synonymous with the conception of "control" in relevant respects.

Control can thus be seen to be the active aspect of the concept of adaptation. The generalization here is that only controllable elements can be included in a system. The criterion for inclusion within an organized action system state is the action theory version of the famous "principle of natural selection." This is a fundamental generalization about all living systems, and particularly important for action systems because they constitute a higher order of such systems.¹⁴

Concluding Remarks

The whole of the preceding exposition sets out a conceptual scheme, as frame of reference and as theory. 

It in no way purports to be an empirical contribution.  Dubin, however, speaks of the importance of empirical verification of these concepts, and of their promise in this respect.  There is no feature of his discussion with which I more fully agree; but the reader should not be misled to suppose that this presentation contributes to that goal.  Certainly a good deal has been accomplished in this direction at various levels in my own work and in that of my collaborators as well as of many others, above all through codification with various bodies of empirical material and the conceptual schemes in terms of which they are analyzed.¹⁵

                It should be kept in mind that the six propositions stated above are couched at a very high level of generality, deliberately designed to cover all classes of action system.  Therefore it is unlikely that these propositions as such can be empirically verified at the usual operational levels.  Such verification would require specification to lower levels, for example, the conditions of small experimental groups as a subtype of social system.  Only in so far as codification reveals uniformities in the cognate features of many different types of operationally studied system do the more general theorems have a prospect of approaching rigorous empirical verification.

202 Sociological Theory                  

This specification should not be assumed to be capable of being carried out by simple "common sense;" it requires careful technical analysis through a series of concatenated steps. I believe, however, that the theory of action in its present state provides methods for successfully carrying Out this specification, and conversely, generalization as well from lower-level uniformities to higher levels. Perhaps the most important key to this possibility is the conception of all systems of action as systematically articulated with others along system- subsystem lines. The basic system types designated here as organisms, personalities, social systems, and cultural systems must be regarded as subsystems of the general category of action system. Each of these in turn is differentiated into further subsystems at different levels of elaboration. Any subsystem is articulated with other subsystems by definable categories of input-output interchange, the processes, in sufficiently highly differentiated subsystems, being mediated by symbolic-type mechanisms such as those discussed above.

                In many respects, this possibility of dealing with multiple system references and of keeping straight the distinctions and articulations between them, has turned out to be the greatest enrichment of theoretical analysis developed from Dubin's "Model II." A "flat" conception of a single system reference which must be accepted or rejected on an all-or-none basis for the analysis of complex empirical problems, cannot possibly do justice to the formidable difficulties in the study of human action.

Notes

1              Talcott Parsons and Edward A. Shils, eds, Toward a General Theory of Acuon (Cambridge: Harvard University Press, 1951).

2              Talcott Parsons, Robert F. Bales, and Edward A. Shils, Working Papers in the heory of Action (Glencoe, IL: Free Press, 1953).

3              Robert F. Bales, Interaction Process Analysis (Cambridge: Addison-Wesley, 1950),

ch. 2.

4              Cf. Parsons, Bales, and Shils, op. cit., ch. 5, figure 2, p.182.

5              Ibid., p.182.

6              There is a pattem-maintenance subsystem below the adaptive subsystem in the hierarchy of control of any system of action and another above the integrative subsystem in the senes. In figure 9.1 we define L as the lower-level case, on the basis parallel to the usage employed in relating the household to the firm in Talcott Parsons and Neil I. Smelser, Economy and Society (Glencoe, IL: Free Press, 1956), ch. 2.

7                Particularly in L'Education Morale (Paris: Alcan, 1925). Cf. Parsons, The structure of Social Action ~ew York: McGraw-Hill, 1937), ch. 10. This classification of mean- ings of objects has been more fully set forth in Talcott Parsons, Edward A. Shils, Kaspar D. Naegele, and Jesse R. Pitts, eds, Theories of Society (Glencoe, IL: Free Press, 1961), Introduction to Part IV.

8              Cf. Parsons et al., eds, Theories of Society, op. cit., General Introduction, Part II.

9              See Parsons and Smelser, op. cit., ch. 4, for a discussion of this concept; it is further developed by Smelser in Social Change in the Industrial Revolution (Chicago: Uni- versity of Chicago Press, 1959).

Pattern Variables Revisited 203

10           Cf. Parsons and Smelser, op. cit.

11           Cf. Talcott Parsons, ~ Approach to Psychological Theory in Terms of the Theory of Action," in Sigmund Koch, ed., Psychology: A Study of A Science (New York: McGraw-Hill, 1959), vol.3.

12           On the general problem of interchanges and their paradigmatic location, see Parsons and Smelser, op~ cit. On the relation of the integrative interchanges to Durkheim's two types of solidarity, see Talcort Parsons, ~~Durkheim's Contribution to the Theory of Integration of Social Systems," in Kurt H. Wolff, ed., Emile Durkheim 1858-1917 (Columbus: Ohio State University Press, 1959).

13           The most important attempt to use essentially this conceptual scheme at the level,

as I see it, of the ~~unit act" of the behavioral organism is James Olds' interpretation of the S-R-S sequence which has figured so prominently in behavior psychology, in action theory terms; see Olds, The Growth and Structure of Motives (Glencoe, IL: Free Press, 1956), ch. 4. Another paradigm which seems to be more generalized, but even more precisely corresponding in logical structure with the unit act, is the TOTE unit presented by George A. Miller, Eugene Galanter, and Karl H. Pribram in Plans and rhe Structure of Behavior (New York: Holt, 1960).

14           These propositions represent a farther development of the set of ~~laws" of action systems tentatively stated by Parsons, Bales, and Shils, op. cit., ch. 3.

15           For example: Bales' work on small groups; the work on family structure and socialization,. including codification with psychoanalytic theory presented in Parsons, Robert F. Bales et al., Family, Socialization and Interaction Process (Glencoe, IL: Free Press, 1955); codification with economic theory in Parsons and Smelser, op. cit.; and with certain problems of economic development in Smelser, op. cit.; codification with learning theory in Olds, op. cit.; the analysis of voting behavior in Parsons,  ~Voting' and the Equilibrium of the American Political System," in Eugene Burdick and Arthur Brodbeck, eds, American Voting Behavior (Glencoe, IL: Free Press, 1958), pp. 80-120; the relation to various aspects of psychological theory in Koch, op. cit.; and the recent ssays published in Parsons, Structure and Process in Modern Societies (Glencoe, IL.: Free Press, 1960), the bibliography of which contains farther references.