I have grown increasingly allergic to arguments about analogies and origins in archaeology. Analogies are simply devices to stimulate the invention of ideas. The source of intuition is fairly irrelevant for the power of the ideas thus generated. The history of science is full of exciting hypotheses that had amusing sources, often far removed from the actual contexts to which the idea ultimately applied.

Archaeologists have tended to generate analogies in a kind of a typo-numerical exercise: you start with the elements/traits that you are archaeologically faced with and have questions about. You then scan the present, the ‘ethnographic present’, or other pasts, for contexts that you assume to be more complete than your starting point. That context becomes your analogy generator that offers the most trait/element congruences with your archaeological case. The more congruences, and the closer in time and space that case is to your archaeological context, the more easily the hypotheses that it generates will convince you. This method of hypothesis generation is exactly that—a method of hypothesis generation—and the hypotheses thus generated are not inherently superior to those generated any other way, with some rather amusing examples from the history of science. They now need to be as vigorously evaluated as any other idea. One should harbour the suspicion that hypotheses generated by analogy in the context of foragers might be considerably more pedestrian than those inspired by other contexts because of the many problems inherent in the ‘ethnographic present’ of foragers (after all, as with the archaeological record, it is also one where pre-contact behaviours usually were not observable any more).

Similarly, the search for origins is simply a typological dead-end. The descriptive elements of one's present (often, the ethnographic present or a select set of its material elements) are chased into the past until all formal coincidence ultimately disappears (when a ‘point of origin’ has been located). Or a complex of descriptive elements (e.g. ‘some archaeologically preservable material elements of recent Kalahari populations’) is matched with precedent co-occurrences of the same elements until the coincidences eventually fizzle out (with ‘the culture of origin’ just overlying that point). Either procedure is processually vacuous. The resultant data structure (I have called it ‘origination cone’ in a previous publication (Wobst Reference Wobst1989)) is simply the result of the mindless sorting procedure. Its shape is pre-given, and its ultimate spatial and temporal referents are the result of the number of attributes one starts with, rather than any anthropologically interesting relationships or processes. Yet forever after, these points of origin will motivate our explanatory efforts and the largesse of archaeological supporters.

There are no easy processual shortcuts that relate Pleistocene and Holocene populations to each other. Searches for analogies and for origins merely extract cultural elements from the contexts of change and variation in which they were embedded, in that way detracting from those contexts, and deflecting from the real problems of the present and its pasts.