Foucault begins by outlining recent trends in two branches of historical method. Firstly, historians have come to address the 'great, silent, motionless bases' that lie beneath the political successions, wars, and famines with which traditional historical practice has been concerned. Examples include studies like 'the history of sea-routes, the history of corn or of gold-mining,' which seek to address the deeper, underlying processes of history. This trend has led to a shift in the theory of history as well, with older questions about the unifying causal connections between events giving way to questions about the isolation of certain 'strata' of history and about the possible 'systems of relations' in which those strata may be understood.
Secondly, there has been a shift in disciplines that address the history of ideas (of science, of philosophy, of literature, etc). The shift here has been from a focus on 'continuities of thought' toward a focus on 'disruptions,' moments of transformation or threshold when ways of thinking have undergone large-scale changes. This view of history is most interested in discontinuities, with historical relations taking the form of passing or contingent affinities or 'compatibilities' (Foucault calls these 'architectonic unities'). The historical problem in these fields, then, 'is no longer one of tradition, of tracing a line, but one of division, of limits; it is no longer one of lasting foundations, but one of transformations that serve as new foundations, the rebuilding of foundations' (any idea of ultimate origins, then, becomes irrelevant). Historical causality is also problematized, as these new methods uncover what Foucault calls 'recurrent distributions,' the multiplicity of frameworks that must be applied to any one area of history: in the case of the history of science, for example, there are always 'several pasts, several forms of connexion, several hierarchies of importance, several networks of determination, several teleologies, for one and the same science.'
In short, then, history proper seems to be seeking deep, hidden, stable structures, while the history of thought seems to be discovering ever more discontinuities and ruptures. But Foucault says that this apparent contrast is a false one: both kinds of historical practice pose 'the same problems,' and they have only 'provoked opposite effects on the surface.' In fact, all of the new problems that Foucault has just outlined stem from one process: 'the questioning of the document.' Instead of its traditional role as a mere vehicle for history as a kind of memory, the document is now becoming important in and of itself. This change stands to redefine the entirety of the historical practice: 'history is one way in which a society recognizes and develops a mass of documentation with which it is inextricably linked.' This new view of history, in which documents become artifacts or 'monuments,' means that history now aspires to be a kind of 'archeology.'
This change has four major consequences. First, there is an intensive questioning of received ideas about the various kinds of series that constitute history; rather than taking for granted certain kinds of progressive series (primarily the assumption of a 'continuous chronology of reason invariably traced back to some inaccessible origin') and then fitting events into that series, historians are questioning the series themselves. This process has resulted in the 'surface effects' detailed above in history and in the history of ideas. Second, the notion of discontinuity assumes a major and pervasive role in historical practice across the board. Discontinuity precedes the work of the historian, as he or she tries to select between discontinuous levels of analysis and types of periodization in which to address their (documentary) material; and, paradoxically, it also results from their description, because they are showing historical limits and moments of breakdown.
Third, there ceases to be any possibility of a 'total history,' a history that depends on a united frame for all history or on the essential spirit or 'face' of a given period. Totalizing history is replaced by 'general history,' in which no continuities are presumed in the open field of documentary evidence. We cannot even posit the traditional 'parallel histories' of law, economics, the arts, etc.; we must rather accept a much more heterogeneous 'form[s] of relation.' Finally, this 'questioning of the document' raises a host of new methodological problems for the historian: how should one construct and delimit bodies ('corpora') of documents? What levels of analysis and what 'principles of choice' inform such constructions? What kinds of limits should be drawn to define groups, regions, or periods? These problems existed before in the field of the philosophy of history, but they now characterize the methodological field of history itself.
Foucault asks why this massive and pervasive change has not been noted before. His answer is a largely psychoanalytic one: the idea of an ordered, teleological, and continuous history serves to make 'human consciousness the original subject of all historical development and all action.' Essentially, we have insisted on a whole, centered notion of the human subject, and therefore on the continuous history that goes hand in hand with such a subject. Marx (by founding a purely relational analysis), Nietzsche (by replacing original rational foundations with a moral genealogy), and Freud (by showing that we are not transparent to ourselves) all challenged this tradition of keeping history in a 'tranquilized sleep' by introducing a radical discontinuity to history and its human subject.