There, in anonymous fashion, someone who signs “An engineer” gives a brief history of a case when “institutional archaeology” was needed because of the loss of “institutional memory” in a large corporation that required “reverse smuggling” of forgotten secrets by a person who knew the secrets but could not have the secrets told him. Specifically, the corporation and its engineers had become ignorant of how one of their factories worked—even though it did work, and actually worked so well that they wanted to expand it! But expansion required a knowledge that had been lost through several generations of reorganizations, partial digitization, retirements, etc. So the corporation brought out of retirement one of the engineers who had designed parts of the machines. The author of this brief case study has a good sense of all the ironies involved as the business, legal, and engineering parts of the corporations worked together, though in ways they could not mutually acknowledge, to keep the whole working (which it never ceased doing). As Sacks would say “everyone has to lie” (1975).
This case is another nice demonstration that:
1) “things have agency” (Latour 2005);
2) writing instruction manuals is impossible because, as Garfinkel pointed out one cannot imagine all the settings within which the manual will be read (2002). One has to add a temporal dimension to this: one cannot imagine which is the part of the manual that will become inaccessible as the manual, as object, disintegrates;
3) Terminator-like, the machines keep working, industry hums along, the humans scratch their head (that this is mostly the case is what Durkheim tried to capture that Latour and other critics cannot quite explain unless they join him in accepting that complex systems have a life of their own, that is they have “agency as things” and there must be mechanisms through which the machines “tell” or “instruct” the humans about what the machines need in order to keep functioning);
4) as the live humans scratch their heads to develop a meta-knowledge about the machines (above the everyday knowledge they may gain by interacting with the machine), they find ways (à la Rancière) to develop a new kind of situated knowledge (since the new version of how the plant currently works is not the same as the earlier version of how the plant would work). For example, they hire retired engineers as consultants who themselves figure out where they put in their garage partial blue-prints they were not supposed to have taken out of the factory, and how to smuggle them back into the factory …);
5) those who read Lévi-Strauss (1966  ) must be told again that the distinction bricolage/engineering is a critique of rationalism: real engineers are always bricoleurs.
Latour, Bruno Reassembling the social: An introduction to actor-network theory. New York: Oxford University Press. 2005
Sacks, Harvey “Everyone has to lie.” in Sociocultural dimensions of language use. Edited by M. Sanches and B. Blount, 57-79 . New York: Academic Press 1975