Chapter 8 (pt.1)
A MULTIVARIATE MODEL OF SOCIAL ORDER
The formalization of theoretical and even empirical statements in sociology is not always as considerable a gain as some adamant proponents of the procedure suggest. This applies particularly to the use of symbolic logic where ordinary language (used carefully) would suffice. In the building of complex theoretical arguments, however, it is often of use to attempt to state simply, directly, and all together, the central propositions involved. The gain is primarily one of visibility. It brings the propositions to the forefather than leaving them hidden in their disparate original contexts. It states them simply, and hopefully fairly discretely, making clear at least what the propositions to be empirically used or tested are. And, especially in the charting of paths of influence it makes clear what effect the presence or absence of each proposition has on the model as a whole. These are the benefits, at any rate, which I hope this chapter (essentially a more formal restatement of earlier propositions) will produce.
Unfortunately the data available on the Tallensi do not allow for extensive quantification of empirical material on any of our variables, and hence the model is only potentially mathematical. The model is built on eleven variables, and involves forty-nine separate propositions about their interrelations. Each of the variables is conceived of as continuous and is presented with reference to both of its directional tendencies. I shall briefly discuss each variable, although no quantitative empirical measures are proposed. All variables are seen as referring to a population otherwise defined, as aggregate measures. Next, the propositions will be listed, and last their interrelations will be represented diagrammatically. This will include both an overall path diagram of the model, and paths of influence involving only limited subsections of the model. It should be noted that paths imply only the existence of some influence, not either a given quantity or constant measure of influence. Thus each arrow denotes a way in which a variable may be influenced, not an extent of influence.
1. Homogeneity/Heterogeneity This variable measures simply the extent of social differentiation among the members of a given population. At the very least all socially organized populations seem to be characterized by differentiation on the parameters of age, sex, and certain skills, though these may be worked out in any number of ways. The extent of variance in wealth, the extent of division of labour, the extent of linguistic uniformity -- all these are potential factors in determining homogeneity/heterogeneity. It should be clear, for example, that although the Tallensi hardly constitute a polar homogenous population, they are far less heterogenous than the populations of neighbouring African kingdoms or certainly of modern nations. Tallensi are (traditionally) differentiated primarily on the bases of age, sex, descent and kinship. Together with the somewhat less central parameter of locality, these four factors could be seen as underlying most variation of wealth, for example. It is probable that differentiation of custom and linguistic usage was continuous from community to community, although these do not seem to have influenced much other than local identity itself (cf. Goody's comments on a neighbouring people, 1962, esp. pp. 4-5). By social differentiation I refer to distinctions which directly influence extent or manner of sociation. Blau, for example, has based an entire conceptual framework of social structure on in-group/out-group interaction frequencies (1974). His approach is inadequate because of its overemphasis on this variable and further, its neglect of the matter of types of interaction/sociation.
2. Multiplexity/monoplexity of social bonds This is a measure of the extent to which a link between any two (or more) social individuals is confined to one area or category of relation or is operative in many. As a variable it has the basic problem that it is unclear whether the analyst or the actor defines the multiplicity of categories. I think this will remain largely a moot point, since any resolution depends on some variety of inter-subjectivity which will always be open to contest. One may assert, for example, that in a highly multiplex social organization actors will not identify as many different varieties of relationship, so that a low actor-statement of multiplexity would be, paradoxically, indicative of the inverse. Be this as it may, the basic content of the variable should be clear enough. It refers to the overlap which obtains between the different role-sets implied by an individual's various statuses, and/or invoked by his various intentions (thus it is that Gluckman sometimes refers to 'multiplex' and sometimes to 'multipurposive' relations, see Chapter 5). One of the central importances of multiplexity is the constraints it puts on options in the planning of social action. This it does in two ways: first, by increasing the investment of the actor in the particular relationships which are heavily multiplex, and second, (parallel to the first) by reducing the number of other directions in which he can move. The first is the effect of the particular multiplexity itself, the second of the prevalence of multiplexity in a population.
3. Solidarity/weakness of social groups This refers to groups as constituted, that is self-identified and organized units. It has two basic aspects. First, is the density of the network of relationships among members. Second is the strength of the collectivity in relation to outside units. The former aspect involves measures of, among other things, the extent to which the actions of any single member affect or imply the actions of others. Thus a group is more solidary where there is a close relation among the actions of its members. This relation, however, is seen as one of cause and effect, however much it must be measured merely in implication (cf. Boudon, 1974). In other words, the common determination of a populations' individual actions by an outside cause does not make of that population a solidary group. It is the mutual and/or reciprocal determination of each others activities by fellow members which does so. As to collective 'strength', I have in mind the extent to which groups are collectively liable, or otherwise collectively implicated by the actions of single members (more accurately, partial populations), and the extent to which groups engage in collective action. This latter may be more or less concerted without changing the extent of solidarity. As to the former, what matters is not the formal definitions of range of liability, but, more generally, the representative function of partial populations. This, of course, is equally in evidence whether the group rallies to the defense of a member or whether outsides choose to extend the group identification regardless of the members own opinion. In most practical cases, to be sure, the two will influence each other.
4. Inclusivity/discreteness of group structure This variable measures the extent to which groups are organized in hierarchies of ascending scale throughout a given population. In other words, it refers to the extent to which smaller groups are parts of larger groups. It is, however, multi-dimensional. The basic factor is the extent of organizational monism, that is, the extent to which a single set of organizing principles and rules of transformation order groups at all levels of size and in all areas of operation. We have discussed (see Chapter 3) the way in which the principle of agnatic descent so orders Tale corporations. Any scaling of this variable would also have to take into account another factor, however. This can be termed the relative salience of groups at different levels of organization. In other words, one must consider the inclusivity of a set of groups to be less if the higher order groups are weak or atrophied with respect to the lower order or vice-versa. This in fact applies over the entire range of groups, so that a set is more inclusive the more equally salient are all of its groups. 'Salient' may be understood to mean here that the groups are effective in concrete social action. Lastly, groups may well include each other and yet maintain a high level of discreteness where groups of one level are organized on different principles than groups of another. Thus although families are included in polities under Anglo-American law and social organization, they are organized and treated differently. This is a further refraction of the monism/pluralism distinction. This is a distinction which has been ignored by some writers who regard all dispersion of power through a group structure as pluralist, without attention to the extent of differentiation of organizing principles (see for example, Nisbet, 1973).
5. Long-term/short-term planning This, quite simply, is an aggregate variable of the extent to which the members of a given population plan on the basis of temporally near vs. distant ends. It measures solely the planning activity, not, for example, the length of time it takes actors to realize their plans. It should be clear that the time span of planning may be a period of several generations in a lineage-based society, so that the fact that single psychological individuals make plans does not mean that psychological individuals are necessarily the units on behalf of whom the plans are made. It should also be remembered, as per discussion in the last chapter, that we are concerned with the conscious plans of the actors we have defined as variably sociated. We do not, in Nadel's words "make society and culture think for the people" (1953, p. 276). The relationship between cause and intention is a complex one (cf. Köbben, 1970); in our treatment we are concerned with some aspects of intention as cause.
6. Assumes adequacy/inadequacy of information This variable is the aggregated extent to which actors search for additional information with which to make decisions. Phrased conversely, it refers to the range at which actors assume they have information adequate to their needs in a choice situation. The unit of its measurement would presumably be some sort of 'distance per decision'. That is, the decision, not the actor, matters first and foremost as the unit of aggregation. The positive end of the variable for continuity and/or order is assumed adequacy, at least in the short run and considered in isolation. One query which would have to be answered before any actual measurement of the variable could begin is to what extent it is the immediate act of searching for additional information which is of primary importance, and to what extent it is the psychical effects of the assumption which are at issue. In other words, how much does it matter that people may act on information they assume to be inadequate. Of course, a correlative question is in what situations and how much does this occur. The variable is further complicated by the problem of measuring extent of assumed inadequacy, and/or expectation for varying extent. Thus, we say that the Tallensi have a fairly high rate of assumed adequacy of information, but we are aware that they have culturally patterned statements of chance, luck. fate-- in short, of the unpredictability of life. These are phrased in terms of the free range of action of ancestors (cf. Fortes, 1959).
7. Similarity/dissimilarity of situation definitions. The concern here is with the extent to which actors share similar definitions of the social situations in which they participate. The variable is again complex, or at least its measurement must be, and must involve a large range of the arbitrary. It cannot be simply done in terms of situations, for the essence of the variable is that these have no definite boundaries, but are actor-constructs. It must be a measurement of congruence among definitions made by each actor. The catch is that it would be very difficult to assess anything but a fraction of the decisions actors make of situations, especially since many such definitions are never made explicit, even when they are in operation. A key element of the variance, however is to be the number of definitions an actor makes, and, as this implies, the duration and range of each of these (aggregated). This problem is primarily one of empirical methodology and interpretation, however. The conceptual-theoretical meaning of the variable should be clear (see also discussion in chapter 5). Actors tend either more or less to see situations as involving the same set of fellow actors under the same conditions with the same durations and the same implications.
8. Centralization/decentralization of power. This variable refers very straightforwardly to the extent to which power is vested in the hand of few or many actors. It is, however, complicated by the fact that its linearity is somewhat misleading when viewed over time. The most diffuse power situation is the most likely to lead to centralization of power. Any application must therefore make a decision with regard to the temporality of the variable (which of course is true of all variables, only more obviously important here). Any actual measure will also involve a postulation with regard to the meaning of 'extent of power' at different levels in the system. Thus power at high levels can be used for purposes which power vested in the intermediate or still more diffuse loci cannot. In brief, the definition of power which is used will influence the definition of centralization which is most meaningful. The most important aspect of the variable here is a continuum from individual through intermediate associations to central rulers (see chapter 6 for discussion).
9. Accessibility/inaccessibility of ends. This deals with the extent to which members of a population are able to achieve (some part of) the ends they set of themselves. Of necessity it also includes the extent to which they are able to see their ends as in the process of realization when they are very long term plans. This is very close to Merton's chief casual factor in his discussions of anomie (1957, pp. 131-60, 161-94): the breakdown of the effective relation between goals and means (institutionalized norms) in a society. Anomie, for Merton, results to the extent to which culturally conditioned ends cannot be met by socially accepted means. In this variable, although we assume that most ends and patterns of ends are socioculturally conditioned, we also consider those which are not. Indeed the extent to which there are goals which are not is an important factor of the variable (if these are then less accessible). We are here concerned directly with the extent to which a society allows its members to achieve their ends, not only with the influence of 'norms governing conduct'. Thus the inability of a migrant farm labourer to achieve financial security and residential stability for his family is not simply a matter of 'norms governing conduct'. 'Deviant behaviour', whether or not it occurs, is unlikely to provide an adequate personal (or social) solution. To be sure, this variable applies to all ends, including those which might be part of a revolutionary movement. The inaccessibility of ends is disaffecting always, but probably more so in proportion to the extent to which there are other means in competition with any culturally (or counter-culturally) sanctioned ones. Thus, to Merton's five 'adaptations' (conformity, innovation, ritualism, retreatism, rebellion) must be added the active refusal to adapt, the attempt to radically change society.
10. Stability/instability of social relationships. This variable refers to the extent to which interpersonal relationships between concrete social actors remain stable over time. Although the one obviously implies the other, it does not refer directly to the extent of stability in the overall pattern of relations. Its assumed operation is egocentric. That is, we are concerned with its effect on the planning and actions of social individuals. It is, therefore, stability within the frame of reference of given individuals which is at issue, although one could conceivably decide to substitute an external assessment of stability as the empirical measure. This would not be without it conceptual consequences, however. One would then be measuring the impact of 'actual' stability on individual action. There would be danger then of a failure to explicitly include the necessary mechanical element in the relation -- the deciding individual's own perceptions. Perceived stability would present empirical difficulties of its own, and presumably artifacts of the measure. One would need to ask to what an individual compares his own situation in assessing relative stability, or whether, perhaps, he compares each of his relationships to his others in assessing the stability of each. Presumably 'actual' and 'perceived' stability measurements would, if handled well, show a high correlation with each other.
11. Authority/anomie. This variable refers to the extent to which the alternatives for action open to individuals are socially limited by the vesture of prior right to decision in another source. It is explicitly a matter of 'right', in Fortes' terms, of 'morality'. 'Anomie' in this usage clearly derives from (though it is not identical with) the Durkheimian meaning of 'normlessness' and from Merton's consideration of the effectiveness of institutional norms and their 'integration with cultural goals'. It is distinct in its conceptualization as a function of decisions. Non-anomie (authority) is seen as necessarily external (supra-individual). Authority thus involves not simply norms or values, but mechanisms for the making of decisions or the abrogation of the need for new decisions. If authority is then defined as a sociated mechanism for superseding individuals in right to decision, it has two central necessary conditions: consensus and continuity. Those subject to authority (and this includes those agreeing not to transgress the bounds of authority over another, as for example one city will not claim taxes from the occupants of another) must agree that it is the appropriate means for determining right, and they must continue to so agree. It is consensus as to mechanism, not as to any particular result which is the subject of the variable. It is thus also important that mechanisms 'work', that is, that they are successful in securing decisions, for authority involves precisely a limitation on the number of situations requiring new decisions. This is why continuity in support of the mechanism is important. Without such continuity there must be continual new decisions as to what mechanism (i.e., what 'authority') to support.
In so far as one of the most general functions of social structure is to provide a basis for predictability and regularity of social behaviour, it becomes increasingly limited in effectiveness as these elements of the social structure become dissociated. At the extreme, predictability is minimized and what may be properly called anomie or cultural chaos supervenes. (Merton, 197, pp. 159-160)
The Variables - The Propositions