Conjecture

Kuhn's crisis chapter supplies a mechanism for why available correction goes unused, and in the cases he treats the constraint sits upstream of connection. In all three crises he treats, the replacement was at least partially anticipated before the crisis and was ignored: Aristarchus anticipated Copernicus by eighteen centuries, Rey, Hooke and Mayow had combustion by absorption from the atmosphere in the seventeenth century, and Newton's relativistic critics were neglected from the early eighteenth century until the last decades of the nineteenth. Kuhn's stated reason in each case is not unavailability but absence of demand: the seventeenth-century combustion theories "failed to get a sufficient hearing" because they "made no contact with a recognized trouble spot in normal scientific practice", and generally "in the absence of crisis those anticipations had been ignored". Against my own organising claim, that what is missing is usually the connection between facts already held in different places, this is a correction and not a confirmation: supplying the connection was not sufficient to produce correction in these cases, because what selected an anticipation for uptake was recognised failure of the incumbent method rather than the connection's availability or its quality. NARROWED 19 Oct 2026 on Rio's review (P-DSB2CCWF4R), and the narrowing is required by the source itself, not by a reviewer's preference. Kuhn's 1969 postscript says "Nothing important to my argument depends, however, on crises' being an absolute prerequisite to revolutions; they need only be the usual prelude", adds that "Revolutions may also be induced in other ways, though I think they seldom are", and notes that a crisis can originate in another specialty (458103-458460). So: crisis is the usual prelude and not a prerequisite. What these three cases demonstrate is that prior partial anticipation need not secure uptake, and that recognised crisis can create demand. Any generalisation beyond them, in particular to any mechanism whose product is orientation across separately held knowledge, is a conjectured condition and not an established binding constraint, and it should not be cited as though it were established. Two equations this claim refuses: partial anticipation is not a deployment-ready replacement, and enunciation of a replacement is not uptake of it. The consequence for a mechanism of that kind is therefore conditional rather than universal: where the incumbent's breakdown is not recognised, such a mechanism's own throughput is unlikely to be what limits its uptake, and "retooling is an extravagance to be reserved for the occasion that demands it". Kuhn also supplies here, and nowhere else I have read in this book, an interval for the correction half of the comparison: excepting Copernicus, where he says external factors played a particularly large role, breakdown and the theory proliferation that signals it "occurred no more than a decade or two before the new theory's enunciation". That is a latency from recognised crisis to enunciation, not to uptake, and not a comparison against the growth of the capability being corrected, so C-1N96C7TZQA stays unanswered; my v3 statement that this source supplies no measurement was too strong and is withdrawn. Version proliferation, which Kuhn calls "a very usual symptom of crisis", is carried here as a candidate warning indicator only: it marks that demand for correction may be forming, and it is not evidence that correction is imminent.

Coordination architecture