The Geography of Memory
Memory has geography. L. P. Hartley's famous remark ("The past is a foreign country—they do things differently there") used a geographic metaphor to describe the strangeness of past events, the discomfort and excitement we feel when we take seriously the exotic intellectual worlds that earlier human beings—in all other respects exactly like ourselves—inhabited. Hartley meant to equate defamiliarization over time, the traditional province of the historian, with defamiliarization over space, the subject of explorers and travel writers (and more recently of anthropologists). Both kinds of travel make the passenger rethink settled cultural assumptions. Yet memory is geographical in another sense. It is a kind of map on which individuals and societies locate past events relative to one another. As Maurice Halbwachs pointed out many years ago, the only area of human memory not rooted in social experience is the dream, which for that very reason lacks structure, continuity, and orderly progression.1 The dream, that is, lacks a map.
Until recently the study of historical memory has also lacked guideposts. The prevailing scheme is the monograph, with memory studied among discrete social groups, in particular locales, or as expressed in single memorial forms such as monuments or textbooks. Sometimes such social memories are contrasted, implicitly or explicitly, with a presumably more objective or accurate "history," of which memorial narratives are seen as corruptions.2 But more often each historical memory is left to stand on its own, as a "version" of the past, on which the historian does not presume to pass judgment. Having absorbed the postmodern lesson that we cannot surgically remove information from the story in which it comes embedded without embedding it in some other story, we are too often content to line the stories up next to each other, like pieces of a dream, without considering their interrelation. This approach has the considerable side benefit of simplifying research, since it is easier (and safer) to report what some group