What I would like to get free from, is the idea that political powerin all its forms and at whatever level we grasp ithas to be analyzed within the Hegelian horizon of a sort of beautiful totality that through an effect of power is misrecognized or broken up by abstraction or division. It seems to me that it is both a methodological and a historical error to consider power as an essentially negative mechanism of repression whose principal function is to protect, preserve, or reproduce the relations of production. It also seems to me wrong to consider power as something situated in a superstructural level relative to the play of forces. Finally, it is wrong to consider power as essentially linked to the effects of ignorance. It seems to me that this kind of traditional and "omni-circulating" conception of power, found in historical writing and in present day political and polemical texts, is actually constructed on the basis of a number of outdated historical models. It is a composite notion that is inadequate for the real world in which we have been living for a considerable length of time, that is, since at least the end of the eighteenth century.
From where is this conception of power borrowed that sees power impinging massively from the outside, as it were, with a continuous violence that some (always the same) exercise over others (who are also always the same)? It comes from the model of, or if you like, from the historical reality of, slave society. The idea that power has the essential function of prohibiting, preventing, and isolating, rather than allowing the circulation, change, and multiple combination of elements, seems to me a conception of power that also refers to an outdated historical model, in this case the model of caste society. By making power a mechanism whose function is not to produce but to deduct, to impose obligatory transfers of wealth and, consequently, to deprive some of the fruit of their work; in short, the idea that the essential function of power is to seal off the process of production and to make a certain social class profit from it, in an absolutely identical renewal of the relations of power, does not seem to me to refer at all to the real functioning of power at the present time, but to how we may suppose or reconstruct it as functioning in feudal society. Finally, in referring to a power that, with its administrative machinery of control, is superimposed on forms, forces, and relations of production established at the level of an already given economy, by describing power in this way, it seems to me that we are still using an outdated historical model that in this case is the model of the administrative monarchy.
In other words, it seems to me that by making the major characteristics we attribute to political power into an instance of repression, a superstructural level, and an instance whose essential function is to reproduce and preserve the relations of production, we do no more than constitute, on the basis of historically outdated and different models, a sort of daguerreotype of power that is really based on what we think we can see in power in a slave society, a caste society, a feudal society, and in a society like the administrative monarchy. It hardly matters whether this is a failure to recognize the reality of these societies; it is in any case a failure to grasp what is specific and new in what took place during the eighteenth century and the Classical Age, that is to say, the installation of a power that, with regard to productive forces, relations of production, and the preexisting social systems, does not play a role of control and reproduction but rather a really positive role. What the eighteenth century established through the "discipline of normalization," or the system of "discipline-normalization," seems to me to be a power that is not in fact repressive but productive, repression figuring only as a lateral or secondary effect with regard to its central, creative, and productive mechanisms.
Michel Foucault, Abnormal: Lectures at the College de France 1974-1975 (2003: 50-52)