The insistence on the pervasive dialectic of the relation between individual and society informs all of Simmel's sociological thought. Incorporation into the network of social relations is the inevitable fate of human life, but it is also an obstacle to self-actualization; society allows, and also impedes, the emergence of individuality and autonomy. The forms of social life impress themselves upon each individual and allow him to become specifically human. At the same time, they imprison and stultify the human personality by repressing the free play of spontaneity. Only in and through institutional forms can man attain freedom, yet his freedom is forever endangered by these very institutional forms.

To Simmel, sociation always involves harmony and conflict, attraction and repulsion, love and hatred. He saw human relations as characterized by ambivalence; precisely those who are connected in intimate relations are likely to harbor for one another not only positive but also negative sentiments. Erotic relations, for example, "strike us a woven together of love and respect, or disrespect . . . of love and an urge to dominate or the need for dependence . . . . What the observer or the participant himself thus divides into two intermingling trends may in reality be only one."

An entirely harmonious group, Simmel argued, could not exist empirically. It would not partake of any kind of life process; it would be incapable of change and development. Moreover, Simmel stressed, it is naive to view as negative those forces that result in conflict and as positive those that make for consensus. Without, for example, "safety valves" allowing participants "to blow off steam," many social relations could not endure. Sociation is always the result of both categories of interaction; both are positive ingredients, structuring all relationships and giving them enduring form.

Simmel differentiated sharply between social appearances and social realities. Although a given conflictive relationship might have been considered wholly negative by participants or by outside observers, it nevertheless showed, upon analysis, to have latent positive aspects. Only a withdrawal from a relationship could be considered wholly negative; a conflictive relationship, though possibly painful for one or more participants, ties them to the social fabric through mutual involvement even in the face of dissensus. It is essential to recognize, Simmel argued, that social conflict necessarily involves reciprocal action and therefore is based on reciprocity rather than unilateral imposition. Conflict can serve as an outlet for negative attitudes and feelings, making further relationships possible; it can also lead to a strengthening of the positions of one or more parties to the relationship, thereby increasing the individual's dignity and self-esteem. Because conflict can strengthen existing bonds or establish new ones, it can be considered a creative, rather than a destructive, force.

Simmel never dreamed of a frictionless social universe, of a society from which clashes and contentions among individuals and groups would be forever banned. For him, conflict is the very essence of social life, an ineradicable component of social living. The good society is not conflict-free; it is, on the contrary, "sewn together" by a variety of crisscrossing conflicts among its component parts. Peace and feud, conflict and order are correlative. Both the cementing and the breaking of custom constitute part of the eternal dialectic of social life. It would therefore be a mistake to distinguish a sociology of order from one of disorder, a model of harmony from one of conflict. These are not distinct realities but only differing formal aspects of one reality.

Throughout his work Simmel considered the individual's social actions not in themselves but in relation to actions of other individuals and to particular structures of processes. In his famous chapter on "Superordination and Subordination," he shows that domination does not lie in the unilateral imposition of the superordinate's will upon the subordinate but that it involves reciprocal action. What appears to be the exercise of absolute power by some and the acquiescence by others is deceptive. Power "conceals an interaction, an exchange . . . . which transforms the pure one-sidedness of superordination and subordination into a sociological form." Thus, the superordinate's action cannot be understood without reference to the subordinate, and vice versa. The action of one can only be analyzed by reference to the action of others, since the two are part of a system of interaction that constrains both. Attempts at analyzing social action without such reference would have been rejected by Simmel as examples of what he called the fallacy of separateness.

Moreover, he does not rest his case after demonstrating that, contrary to first appearance, domination is a form of interaction. He proceeds to show in considerable detail the particular ways in which various types of groups structure are associated with different forms of subordination and superordination--distinguishing, for example, between levelling and gradation. If a number of individuals are equally subject to one individual, he argued, they are themselves equal. Such levelling, or "negative democratization" to use Karl Mannheim's term, favors and is favored by despotic rulers. Despots try to level their subjects and, conversely, highly developed levelling easily leads to despotism. On the other hand, strong intermediated gradations among a ruler's subjects tend to cushion his impact and weaken his hold over them. Although intermediate powers may increase inequalities in the subject population, they shield the individual from the direct powers of the ruler. A pyramidal form of social gradation, whether it develops under the plan of the ruler or results from the usurpation of some of his power by subordinates, gives every one of its elements a position both lower and higher than the next rungs in the hierarchy. In this way, each level--except the very highest and the very lowest--is subordinate to the authorities above and, at the same time, is superordinate to the rungs beneath. Dependence on some persons is compensated by authority over others.

From Coser, 1977:183-188.