Main idea

Changing discipline is not enough: you can travel through intelligence analysis, medicine and archival science while staying inside a single corpus, Western and English-speaking. The next shift is linguistic, and it is the one that pays off most.

Searches run in Japanese, Chinese, Russian and Arabic turned up what English had not given. Jirō Kawakita's KJ法 describes a bottom-up method in which field data is written on movable cards, brought together and recombined until a structure emerges. As early as 1969, Tadao Umesao describes a system of independent, recombinable cards whose resulting knowledge re-enters the system. The Hall for Workshop of Meta-Synthetic Engineering, developed around Qian Xuesen in the early 1990s, explicitly combines human experts, machine systems and knowledge systems, and distinguishes general knowledge from knowledge specific to the object under study which accumulates through its use.

These traditions were not ignored because they were obscure. They were ignored because they were not indexed in the language being searched.

Why it matters

This shifts the confidence you can place in a search: an exhaustive review in English is not an exhaustive review. It is an exhaustive review of one corpus, and the corpus has a boundary that nothing displays.

And it gives a concrete, cheap move, now within reach: re-run the same questions in three or four languages before concluding anything at all about the originality of an idea.

Nuances and limits

Access remains uneven: whatever has not been digitized, indexed or translated stays out of reach whatever the language you search in.

And reading a tradition in a language you do not practise exposes you to mistaking a surface resemblance for a real kinship — the risk rises precisely where the discovery is most seductive.

Open questions

• Which languages do you choose, and on what criterion, given that you cannot know in advance which one carries the tradition you are missing?