The Theory of Social Structure

S.F. Nadel

• Publisher: Taylor & Francis
• Year: 2013
• Print ISBN: 9780367475468
• eBook ISBN: 9781136550409
• Edition: 1
• Topic: Social Sciences
• Subtopic: Anthropology
• Index: Social Sciences

The main feature of this book is a discussion of 'role analysis' and its relevance to social structure. Arguing that the role system of a society is the matrix of its social structure, the author presents a detailed theoretical analysis of the problems inherent in this approach.Chapters cover: · The problems of role analysis.· Conformity and deviance· The coherence of role systems· Degrees of abstraction· Structure, time and realityOriginally published in 1957.

IV

THE COHERENCE OF ROLE SYSTEMS

IN this section we shall mainly be concerned with the interconnection of roles and with the coherence of the role system (or ‘social structure’) that thus emerges. First, however, I propose to consider a very special type of interconnection, presupposed in the subsequent discussion, namely the interconnection, or summation, of roles played by identical actors.

It is a sociological truism that every individual fills several roles both successively in his lifetime and simultaneously at any given time. The modus operandi of this succession and coincidence of roles presents no new problem, both being engineered by the usual mechanisms of recruitment. All that is new is that the preconditions for the assumption of additional roles now include not only ordinary, as it were autonomous, human attributes, but attributes indicating the performance (perhaps the successful performance) of specified other roles, previously or concurrently. Examples are of course familiar—political posts open only to ‘family heads’ or ‘men of the people’; a husband automatically becoming a ‘son-in-law’; a great ‘scholar’ or a ‘philanthropist’ qualifying for positions of social prominence, and so forth.

The ambiguity entailed in the definition of certain roles occurs again on this level; it now concerns the difficulty of separating the linked roles played by the same individual from a single role series. The primary evidence for the presence of the former is again linguistic: it is simply a fact that an ‘old man’ will be called such in certain situations, while in others he will be referred to, say, as a ‘wise man’, or that a man will be known as a ‘husband’ in respect of certain rights and obligations and as a ‘son-in-law’ in respect of others. No ambiguity arises if the linkage of roles is variable and not exclusive to the roles in question, for example, if some (not all) ‘old men’ are also considered ‘wise’, and some ‘young men’ equally qualify for this description. But where the roles (or presumed roles) appear in linkages which have some degree of regularity and exclusiveness, e.g. where all or most old men are considered wise and no young man ever is, we may be in doubt whether we are in fact dealing with genuinely separate roles or only with additional attributes to a given role series. Symbolically expressed, though it may appear as if a person P enacted two roles, A and B, such that

(13)

P (A + B), when

A ⊃ a, b, c … n and

B⊃n,

in effect he might enact only one expanded role (A or B), such that

(14)

A/B =∑a, b, c … n

Take these further examples: Since any married man will have to assume, say, the current taboos towards his wife’s mother, should we not include these among the attributes of ‘husband’ rather than say of him that he is assuming the additional role of ‘son-in-law’ ? Again, if in a society (like Nupe) all women traders are expected to be also prostitutes, and vice versa, is this a summation of two separate roles or only the full description of one? And if the typical ‘businessman’ is the sort of person likely to belong to the Rotary Club and similar associations, does this indicate a probable linkage between several roles or does it only characterize the one role ‘businessman’?

There is no categorical answer to such questions. This is all a matter of viewpoint and heuristic usefulness. To quote from Sprott: ‘How far one says (of a person) that he is playing a multiple role, or a single one at any given moment, will depend on the purpose of one’s enquiry’.¹ But we can be a little more precise; for in choosing between the two manners of description we are likely to consider two things. First, we cannot fail to consider the existing linguistic convention and the expediency of departing from it, e.g. by denying that ‘son-in-law’ or ‘prostitute’ is a separate role. More generally speaking, it will always be inexpedient to enlarge any role series so much that the role name (‘husband’, ‘woman trader’, ‘businessman’) ceases to be informative. Secondly and more importantly, we shall always consider also the compatibility of the linked series of attributes. If, for example, in a society with unilinear descent the woman as a ‘wife’ must give all her loyalties to her husband and his kin (lineage, clan), but as a ‘daughter’ or ‘sister’ to her own kin, it seems logical to allocate the two potentially conflicting sets of obligations to separate roles. I said ‘logical’ because our conception of roles, as series of interconnected ways of acting laid down by the rules of society, carries the implication that the series is always a compatible one. If then, we meet with a conflict of norms in the total behaviour of an individual, we shall tend to see it as a conflict between the roles he is made to assume rather than within any one role.

But it is not a manner of great moment whether we decide one way or the other; our analysis will in any event lead us to all the attributes making up the presumed role or combination of roles. I shall, therefore, in the following largely disregard this possible ambiguity. Nor can I attach much importance to the question what we should call the summation of all the roles individuals might be expected to play: ‘General Role’, ‘Role Personality’, or ‘Social Personality’ are three names which have been proposed. These names actually suggest more than a mere summation, namely a merging or integration of all roles in some sort of super-role. This seems to me a dubious assumption; I frankly cannot picture the concrete case. Indeed, I would suggest that closer analysis will always show the alleged super-role to be in effect only one of the several roles enacted by an individual, selected because it appears to be the one most powerful in his life or the one most consequential considering the general character of the society: which condition is most frequently satisfied by occupational roles (‘he is a teacher through and through’, ‘he is a soldier first and foremost’), and roles indicating economic standing and civil status (’socio-economic status’, as one often says). To some extent, then, the presumed super-role may prove to indicate no more than the ‘halo effect’ discussed earlier.

There seems to be an exception, however. In primitive societies the linkage of roles may be so firm and predictable that individuals do appear to act in such embracing, overall roles. Simplifying a little we can visualize a situation where a ‘father’ is always the head of a household, the teacher of his children, the ‘manager’ of a labour team, a ‘councillor’ or ‘elder’ of the community, an officiant in rituals, and one of the ‘onlookers’ in dances or sports; while a ‘son’ will always be a subordinate member of the household, a ‘pupil’ or ‘trainee’, a ‘worker’ in the labour team, a ‘common citizen’, one of the congregation in rituals, and an active participant in dances and sports. (Similar linkages can of course be established for other roles—‘grandfathers’, ‘married women’, etc.). Since in such combinations it is, for obvious reasons, the contingent condition (being a father or son) which determines the assumption of all the concomitant roles, the description of a man as a ‘father’ or ‘son’ will in fact amount to describing almost his complete social being (or ‘personality’). But this is simply the possibility mentioned before, when the series of firmly linked roles might equally be regarded as a single, more broadly conceived role, extended by the inclusion of additional attributes. And whether we speak of summation or extension, it still is a process of combination and linkage in which the components can be isolated.

The combination as such, however, is important also from another point of view. One of the consequences of this process must be the reduction in the number of genuinely separate or separable roles in the society. And since each role involves its actor in appropriate relationships with actors in other roles, the reduced separability of roles also entails a reduced separability of relationships. The ‘family head’ who is also a ‘manager’, ‘priest’, and ‘elder’ will thus be linked with persons who are at once ‘kinsfolk’ or ‘dependents’, ‘workers’, members of the ‘congregation,’ and ‘common citizens’, in this fourfold manner. Differently expressed, a series of diverse relationships (domestic or kin, economic, religious, political) will come to link the same sets of persons, constituting a congruent set of linkages. Which means that the society is to that extent rendered less complex and more homogeneous.

This multiplication of relationships stretching between the same actors has been recognized by several scholars as a characteristic feature of primitive societies. Firth and others speak of the ‘many-stranded quality of relationships’¹: this is a slightly misleading description since it suggests single relationships of a complex order rather than a bundle or set of congruent relationships. Perhaps we might, more accurately, speak of a manifold of relationships. The lesser complexity of a society operating with such relationship manifolds has equally been pointed out. Gluckman suggests that a social structure in which there is a high degree of ‘congruence in the links between … persons’ should be called a relatively ‘uncomplicated’ structure; a society in which links are not congruent, tending to relate any given person ‘with many different persons in various systems of ties’, is by contrast a ‘complicated’ structure.²

Once more, the choice of names is not altogether a happy one. For the term ‘uncomplicated’ obscures the distinction between sheer quantitative simplicity and the reduced complexity we have in mind, which is the result of rules of combination. A society differentiating between four pairs of roles, say, family head and dependent, manager and worker, elder and common citizen, priest and worshipper, is undoubtedly more complex or complicated than one acknowledging two only, e.g. having no ‘managers’ (because each man works on his own) and no ‘priests’ (because all worship is strictly communal). Needless to say the relationships involved are similarly reduced in number, the ‘simpler’ society lacking two relationships present in the more ‘complex’ one (manager-workers, priest-congregation). But think now of a society which, while exhibiting the same fourfold (or eightfold) differentiation of roles and relationships, combines the actors, in one of these ways:

(a) Family head = manager

dependent = worker

elder = priest

common citizen = worshipper;

(b) Family head = manager = elder = priest

Dependent = worker = common citizen = worshipper.

This is clearly an altogether different type of non-complexity.The possible relationships, like the roles on which they hinge, are not reduced in number but, as it were, in their scatter through the population. Thus, the social system is not simply less ‘complicated’ (indeed, the component relationship manifolds might be said to make it more so): perhaps it might be called more highly combinative or involute.

There is, admittedly, the more familiar and less far-fetched term ‘homogeneous’, which may serve as well. Fortes recently pointed out that ‘just what we mean by a homogeneous society is still rather vague though we all use the term lavishly’.¹ If we adopt it to indicate the character of such ‘combinative’ or ‘involute’ systems we shall at the same time improve the precision of the term; for the scale and regularity of the summation of roles (and of the attendant multiplication of relationships) seems to provide the needed criterion for the definition, and probably the measurement, of this crucial yet elusive phenomenon.

Now Fortes’s own ‘working definition’ of social homogeneity is this: A homogeneous society is ideally one in which any person (meaning here an individual with his ‘assemblage of statuses’) can be substituted for any other person of the same category without bringing about changes in the social structure. This implies that any two persons of the same category have the same body of customary usages and beliefs. The paradigm of ‘sibling equivalence’ further suggests that such persons are ‘equivalent’ in regard to their ‘achievable life histories’. If I understand Fortes correctly, his principle of ‘social substitut-ability’ follows from my thesis concerning the summation of roles, so that social homogeneity as he uses the term does in fact describe our ‘involute’ social system; and the issues involved seem sufficiently important to warrant fuller exposition.

I assume that when Fortes speaks of persons ‘of the same category’ being ‘substituted’ for one another he means something very simple the kind of situation in which the heir or successor to any office, role or ‘status’ steps into the place of the previous incumbent. I assume, further, that by ‘changes in the social structure’ Fortes means the rearrangement, possibly the dislocation, of the pre-existing ordering of relationships and roles (or ‘statuses’). Picture, then, the replacement of a family head in a society where this role is firmly combined with those of elder and priest: the successor to the headship of the family will then automatically assume the other two offices as well. As a result, all the roles will be kept in being; furthermore, the relationships which used to link the family head-cum-elder-cum-priest with others will be preserved, with their given pattern and degree of congruence. Consider now the opposite case—a society where the three roles are not firmly interconnected, being only fortuitously combined in the person of a particular family head. On his replacement a number of changes are apt to happen: one or the other concomitant role may disappear, at least temporarily, until a separate, new actor is recruited; or the roles, though not disappearing, will be redistributed, so that the former congruence no longer holds; or finally, the successor may bring with him some new role (say, that of a leader in warfare or dancing), not formerly linked with that of family head, again disturbing the previous pattern of roles and relationships.

In the first type of society it is therefore true to say that the two persons in the same category (of family head) will ideally have the same ‘achievable life histories’; equally, their substitution for one another will cause no ‘changes in the social structure’ (though ‘shifts’ is probably a better word to describe the rearrangements in question). In the second type of society, both effects are likely to be reversed. Our general conclusion, then, is that a society is ‘homogeneous’, providing for this substitut-ability of persons, to the extent to which the summation of roles current in the society is both firm and extensive, so that the summation is indeed a measure of homogeneity: which is what we set out to show.

We might add, by way of a footnote, that the summation of roles and the consequent involution of relationships are likely to occur on any significant scale only in the case of roles belonging to different sectors or areas of social life—economic, political, religious, etc. Any considerable summation of roles in the same sector would defeat the very principles of role allocation and differentiation, that is, the reason for having roles at all; while the attendant relationships might cease to be compatible or indeed workable. I am not at the moment considering the logical impossibility of conceiving of certain roles as combinable, when they represent mutually exclusive categorizations (e.g. priest and layman, a man’s father and son). I am thinking of the purely practical difficulties that would, arise if the ‘manager’ of a team also played the part of an ordinary worker in it, an ‘elder’ that of a common citizen, and vice versa. The sector of kinship is in some respects an exception since here all roles are widely combinable. Obviously, anyone is likely to be, at the same time, a father, son, brother, husband, uncle, cousin. Equally obviously, he is so to different persons; that is to say, there is no corresponding congruence and involution of relationships. Preferential marriage rules, e.g. of cross-cousin marriage, do bring about a certain measure of congruence (spouses being also cousins, a mother’s brother a father-in-law, a father’s sister a mother-in-law). But other marriage rules, in the form of incest taboos and rules of exogamy, on the contrary restrict any such coincidence of ties, and hence the risk of creating unworkable combinations of relationships. It is worth keeping this in mind when discussing the ‘many-strandedness’ of relationships in primitive societies; for however ‘involute’ they may prove to be in other respects, in the sector of kinship they retain a maximum of ‘scatter’ and separability.

This seems the appropriate place to consider briefly the advantages and disadvantages of the summation of roles, judged from the viewpoint of the society engineering it. As regards the disadvantages, we have already suggested the main danger: it is that of imposing upon individuals roles and relationships which are incompatible, and hence creating strains and tensions both in the personality and in the society ...

Table of contents

• Cover
• Title Page
• Copyright Page
• Preface
• Contents
• Memoir
• I Preliminaries
• II Problems of Role Analysis
• III Conformity and Deviance
• IV The Coherence of Role Systems
• V Degrees of Abstraction
• VI Structure, Time and Reality
• VII Conclusions: Structure and Function