From Historical Amnesias: An Interview with Paul Connerton
We need to distinguish cultural memory from historical reconstruction. Knowledge of all human activities in the past is possible only through knowledge of their traces. It might be the bones buried in Roman fortifications, or a pile of stones that is all that remains of a Norman tower, or a word in a Greek inscription whose use reveals a custom: in all these cases what the historian deals with are traces, that is to say, the marks which some phenomenon has left behind. Simply to apprehend these marks as traces of something is to have gone beyond the stage of making statements about the marks themselves; to consider something as evidence is to make a statement about something else, that is to say about that for which it is taken as evidence.
Historians, in other words, investigate evidence in much the same way as lawyers cross-question witnesses in a court; they extract from that evidence information which it does not explicitly contain or even information which was contrary to the overt assertions contained in it. Historians are able to reject something explicitly told to them in their evidence and to substitute their own interpretation of events in its place. And even if they do accept what a previous statement tells them, they do this not because that statement exists but because that statement is judged to satisfy the historian’s criteria of historical truth. Far from relying on authorities other than themselves, historians are their own authority; their thought is autonomous vis-à-vis their evidence, in the sense that they possess criteria by reference to which that evidence is criticized. Historical reconstruction is therefore not dependent on social memory. It is autonomous with regard to social memory.
HT: MEH