has some incredible writing on this:

One of the most remarkable things about such moments is how they can seem to burst out of nowhere — and then, often, dissolve away as quickly. How is it that the same “public” that two months before say, the Paris Commune, or Spanish Civil War, had voted in a fairly moderate social democratic regime will suddenly find itself willing to risk their lives for the same ultra-radicals who received a fraction of the actual vote? Or, to return to May ‘68, how is it that the same public that seemed to support or at least feel strongly sympathetic toward the student/worker uprising could almost immediately afterwards return to the polls and elect a right-wing government? The most common historical explanations — that the revolutionaries didn’t really represent the public or its interests, but that elements of the public perhaps became caught up in some sort of irrational effervescence — seem obviously inadequate. First of all, they assume that ‘the public’ is an entity with opinions, interests, and allegiances that can be treated as relatively consistent over time. In fact what we call “the public” is created, produced, through specific institutions that allow specific forms of action — taking polls, watching television, voting, signing petitions or writing letters to elected officials or attending public hearings — and not others. These frames of action imply certain ways of talking, thinking, arguing, deliberating. The same “public” that may widely indulge in the use of recreational chemicals may also consistently vote to make such indulgences illegal; the same collection of citizens are likely to come to completely different decisions on questions affecting their communities if organized into a parliamentary system, a system of computerized plebiscites, or a nested series of public assemblies. In fact the entire anarchist project of reinventing direct democracy is premised on assuming this is the case.

To illustrate what I mean, consider that in America, the same collection of people referred to in one context as “the public” can in another be referred to as “the workforce.” They become a “workforce”, of course, when they are engaged in different sorts of activity. The “public” does not work — at least, a sentence like “most of the American public works in the service industry” would never appear in a magazine or paper — if a journalist were to attempt to write such a sentence, their editor would certainly change it. It is especially odd since the public does apparently have to go to work: this is why, as leftist critics often complain, the media will always talk about how, say, a transport strike is likely to inconvenience the public, in their capacity of commuters, but it will never occur to them that those striking are themselves part of the public, or that whether if they succeed in raising wage levels this will be a public benefit. And certainly the “public” does not go out into the streets. Its role is as audience to public spectacles, and consumers of public services. When buying or using goods and services privately supplied, the same collection of individuals become something else (“consumers”), just as in other contexts of action they are relabeled a “nation”, “electorate”, or “population”.

All these entities are the product of institutions and institutional practices that, in turn, define certain horizons of possibility. Hence when voting in parliamentary elections one might feel obliged to make a “realistic” choice; in an insurrectionary situation, on the other hand, suddenly anything seems possible.