An immediate problem is determining a core meaning that is consistent with usage. Unfortunately, the literature shows little agreement on what constitutes class.¹ Classes have been defined by property ownership (Marx), position or role (Mosca, 1939; Pareto, 1963), status rank (Warner, 1960; Lenski, 1966), prestige (Barber, 1957); or by intermarriage (Schumpeter, 1951). In spite of this denotative variance, however, underlying the discussion and use of the concept are common assumptions which are also shared by my definition.² First, class membership is not biologically determined, but is a form of social stratification based on laws, esteem, wealth, or power. Estates or castes may comprise types of class stratification, or classes may be defined by different status-ranks. In all cases, however, class is a vertical division into superior-inferior or, in my terms, superordinate-subordinate. Second, in general a difference in class is a difference in wealth, power, or esteem. All definitions do not assume an identity between class and rank, but all assume a correlation. Class is a division in privileges. Thus, in my view superordinates benefit from the status quo; they are well situated in the system of rewards and privileges. Third, class is a dividing line between different kinds of behavior. What these behaviors are depends on the actual definition, but nonetheless all definitions assume that different classes can be separated empirically according to different patterns of behavior, such as intermarriages, social mixing, organizational memberships, travel, etiquette, and mannerisms.

Fourth, all definitions assume that class membership has continuity. There is disagreement as to whether membership may be frozen, as for castes, but in no case are classes considered ad hoc or momentary aggregations, like a crowd. There may be a circulation of members between classes, but it is not so rapid as to homogenize their distinctive behaviors nor to change in a generation more than a small proportion of class membership.

Fifth, all definitions assume that interests are associated with class membership. Interests may be directed toward status mobility or maintenance, overcoming or protecting the privileges of the upper class, or changing or defending the status quo (as in my definition). In any case, class divisions form latent or manifest interest groups. A class is considered potentially or actually a group which shares similar interests vis-à-vis some other class.

Sixth, a class is always defined within a system of relations constituting an opposing class or classes, analogous to hot being defined relative to cold, up relative to down, virtue relative to vice, and so on. Class as rank, esteem, privilege, or position, requires an opposition to complete the definition. This is because all definitions assume a superior-inferior stratification, and an inferior or superior cannot be defined without the other.

Finally, all assume that class constitutes a basic component of the social order (as I have made it a common component of social space). To understand the major social divisions in society, patterns of interaction, and conflict is to understand class.

Such are the common assumptions in the literature. These notwithstanding, since definitions of class do vary, some markedly different assumptions must exist.

One difference is whether classes can be latent groups or must be manifest in behavior or consciousness. I treat class as manifest within an organization's empirical division of power, but common class interests and consciousness may be latent. Authoritative roles structurally divide people into classes which constitute different life situations. Out of these different structural relationships and resulting experiences develop similar attitudes (latent interests), which are the seat of common class interests.

A second difference has to do with whether one can indeed define "class." Because of the confused state of the literature, some simply pose a definition "for research purposes," or treat class as a construct, a stepping-stone concept towards empirical analysis. A confused literature, however, does not mean a confused reality. Clearly, there is some important aspect of society that all definitions have captured. I believe my definition reflects this common reality in a way consistent with our scientific and normative interests, while focusing specifically on the power-conflict-process complex of specific concern to me.

Another difference in assumption concerns the significance of class. Some argue that class is a functional division of society based on talent or aptitude. The upper class, being those who contribute scarce resources required for society to function, are thus rewarded more than the lower class. Others argue as I do that classes are divisions resulting from historical conflicts of power. The privileges of an upper class result from their dominance, usually through monopolization of force.

A final difference concerns whether class is a term applicable only to capitalist societies, and not to traditional and communist systems. Clearly, my definition applies to all societies at all times. All societies are held together by law-norms, an aspect of which is to define authoritative roles that divide people everywhere into those who can command and those who obey. This is true no less for communist societies, than for the so-called capitalist. In any society this division determines who gets what.

In sum, class as I use the term shares those assumptions commonly underlying usage. Moreover, I also assume that the concept captures a meaningful aspect of society, and that the idea of a latent class does this. Finally, I clearly see classes as engaged in a struggle and the superordinate class as having won its privileges through power.

First, I must clarify my definition of class at this level. A Great Class is a quasi or partially organized class traversing society. Class interests are active; members have some consciousness of a class identity; and leaders use class slogans and interests in a struggle for power at the state level, thus forming the conflict front dividing field and antifield.

Such Great Classes have in one society or another comprised:³

• the peasant or farmer class

• large landowner class

• proletariat (industrial labor) class

• bourgeoisie class

• managerial class

• managed class

The managerial class manage and control organizations and the state. In socialist countries, this class is unified, since those who manage the state also manage all the other major organizations. Clearly today in the worldwide movement towards socialism, the bourgeoisie are being transformed into a managerial class.⁴ By this I do not mean a division of ownership and control of industries through the growth of common stock corporations. I mean that common ownership is assumed by the state through nationalization, or the management of privately owned companies is completely regulated by the state (as in Hitler's Germany). Thus, he who manages the state controls all the major organizations. The division is now between the party (the elite) and the ruled, between managers and managed.

How do we get to such Great Classes from my definition? Great Classes are large clusters of individuals at different ends of the class component of social space. That is, when society is divided so that the class distances between individuals are either large or small (some will always be superordinate, most will always be subordinate), then the class component will have two clusters of individuals at opposite ends with few in between. A great Class represents such a distinctive cluster.

If there are many classes in a society, individuals will be spread more or less evenly along the class component in their distances. Then, a bipolar distribution of individuals on the class component is identical with a division of society into two Great Classes.

A final note on class and the elite. My notion of class generalizes the conception and analysis of the elite made by Mosca (1939) and Pareto (1963), who restricted their definition to those who formally control the government.⁵ Thus, at this level my superordinate-subordinate distinction based on authoritative roles is consistent with their view. I abstract the concept, however, to every organization, and then define the generalized distance between individuals taking all organizations into account. This is because I am interested in the struggle for power and the process of conflict at all levels, while Mosca is concerned wholly with state power.

In terms of the Great Classes, however, who are the elite? At one time the large landowners (aristocracy, the feudal barons), at another the bourgeoisie who controlled the state machinery, and now increasingly, the managers (bureaucrats, party members, planners). When society becomes divided into two classes, the theoretical distinction between class and elite empirically disappears.