The following is an attempt at constructing a theoretical model of the power structure of socialist society as distinct from the stratification model. The essential differences between these two ways of accounting for the phenomenon of social inequality can be summarized as follows:

the power-structure model aims at depicting the differentiation of opportunities of action, of influencing events, and of access to appropriate resources – as seen, in principle, by an outside observer; the stratification model attempts to systematize and make coherent the disparate assessments of these qualities, possessions, and accomplishments of members of the society, which are selected as important and significant by the members themselves. All in all, while the power-structure model starts from situations as its raw data, the stratification model is focused on evaluative patterns;

while the stratification model compares the values assumed by selected variables in differentiating groups within the population (without pre-empting the question of relations between the groups) -the power-structure model attempts to represent the web of dependencies, which both unites and opposes groups within a society. ‘Dependency’, which is the meaning ascribed by Marx to the notion of ‘social relation’, stands here for any constraints that one group exercises on the other group’s claims to either con-summatory or instrumental resources;

the intention behind the power-structure model is to reveal the conditions which are likely to generate important disparities in behaviour given a certain conjunction of circumstances. The stratification model, on the other hand, is set on describing existing variations in behaviour; the question whether observed variations have a systemic, or a contingent (even if monotonous) nature may be asked in this connexion, but it does not belong to the logic of the stratification model;

130the power-structure model employs the logic of typology: it normally sets apart distinct types of systematically determined situations, which are defined in opposition to each other and which tend to be mutually exclusive. The stratification model, by contrast, uses the logic of a continuum defined by a specific relation between any pair of points selected at random (e.g., ‘more and less prestige’) -rather than by its sharply determined poles. By detecting discontinuities in what is essentially an uninterrupted line, the stratification model arrives eventually at the logic of classification:

last, but not least, the stratification model is a linear one; the power-structure model is not. A single set of values may, of course, be applied in retrospect with the effect of arranging the units of the power structure in a linear order; but this operation would not belong to the logic of the model. The same operation is, however, the principle underlying the construction of the stratification model, and thereby constitutes its inherent element. The linearity (portrayed ordinarily as verticality) of the stratification model is therefore determined by its own logic and as such is necessarily given. An important consequence of this distinction is that the power-structure model is not distinguished from the stratification model merely by the selection of power as its central variable; indeed, one may easily construct a stratification model of power.