Foucault gives us an indication of what he means by virtue in the introduction to The Use of Pleasure: The History of Sexuality, Volume Two.  At this juncture he makes clear that he seeks to move beyond a notion of ethical philosophy that issues a set of prescriptions. Just as critique intersects with philosophy without quite coinciding with it, so Foucault in that introduction seeks to make of his own thought an example of a non-prescriptive form of moral inquiry. In the same way, he will later ask about forms of moral experience that are not rigidly defined by a juridical law, a rule or command to which the self is said mechanically or uniformly to submit. The essay that he writes, he tells us, is itself the example of such a practice, “to explore what might be changed, in its own thought, through the practice of a knowledge that is foreign to it.” (9) Moral experience has to do with a self-transformation prompted by a form of knowledge that is foreign to one’s own. And this form of moral experience will be different from the submission to a command. Indeed, to the extent that Foucault interrogates moral experience here or elsewhere, he understands himself to be making an inquiry into moral experiences that are not primarily or fundamentally structured by prohibition or interdiction.
Whether there are genuine law-like causal regularities that govern social phenomena is not at all clear. In any event, no laws governing the social world have been discovered that meet the demanding criteria of natural science. To be sure, social scientists have identified many social regularities, some of which they have even dubbed social laws. Examples from the discipline of economics would include the laws of supply and demand. From political science we find Roberto Michels’ iron law of oligarchy, which holds that popular movements, regardless of how democratically inclined, over time will become hierarchical in structure. Another proposed law of politics is Duverger’s Law, which posits that two-party systems will emerge in political systems that feature simple-majority, single-ballot electoral systems. But upon closer inspection, these laws fail to meet the criteria for genuine law-like regularities. Sometimes, particularly in economics (which boasts more purported laws than the other social sciences), the laws merely describe logical relationships between concepts. These laws may be true by definition, but because they do not describe the empirical world, they are not scientific laws. On the other hand, social laws that claim to describe empirical regularities invariably turn out to be imprecise, exception ridden and time-bound or place-bound rather than precise and universal. Consider the law of demand from economics, which holds that consumer demand for a good will decrease if prices go up and increase if prices go down. Though this pattern typically occurs, it is not without exception. Sometimes increasing the price of a good also increases demand for it. This may happen when consumers interpret a higher price as signaling higher quality or because purchasing an expensive good provides an opportunity for conspicuous consumption – wasteful expenditure as a display of status. Moreover, the law of demand is a weak law; it merely specifies an inverse relationship between price and demand. Unlike the more precise laws of natural science, it does not specify the magnitude of the expected change.
In short, it seems that from the empirical observability for us of an ensemble to its historical acceptability, to the very period of time in which it is actually observable, the analysis goes by way of the knowledge-power nexus, supporting it, recouping it at the point where it is accepted, moving toward what makes it acceptable, of course, not in general, but only where it is accepted. This is what can be characterized as recouping it in its positivity. Here, then, is a type of procedure, which, unconcerned with legitimizing and consequently excluding the fundamental point of view of the law, runs through the cycle of positivity by proceeding from the fact of acceptance to the system of acceptability analyzed through the knowledge-power interplay. Let us say that this is, approximately, the archaeological level [of analysis].