Given its logic of accumulation, the archive would appear to be a location where knowledge of the past is safely stored. The very existence of a state archive, such as the colonial one, gives the historian the impression that continuity of government goes hand in hand with continuity of memory. Yet the memory of states is often extremely short: what is a source for a historian is a piece of intelligence for a state that has been produced for a certain function. Once this has been accomplished, the file can be closed and safely forgotten, with the certainty that it can be retrieved any time. The purpose for which a certain item of information has been collected often goes unrecorded, however. It may or may not be retrieved at a later date, but even if it is, the reason for that recovery usually has very little to do with the context in which it was first recorded. Most often it provides answers to problems that have little to do with the original issues or the circumstances under which it was created.

A process of selection, omission, and reinterpretation is applied to the available information for it to be of use where it is relocated. This process implies assigning a new meaning to a past event. This, for instance, is what Shahid Amin demonstrated in his study of the Chauri-Chaura uprising in 1922: it became a ‘metaphor’ for the national history of postcolonial India. But the meaning of an event can be altered much earlier than that: a single year may suffice. As early as 1925, John Murray Ewart picked out certain pieces of information from the previous year's intelligence records to bring some order to the chaos of 1924, and thus the meaning of the revolution began to change. Ewart not only affected the subsequent course of events by providing powerful justifications for the hardening of Indirect Rule with the clear objective of containing nationalism, but he also wrote perhaps the most influential version of the 1924 Revolution.