Minsky and Papert 1969 proved single-layer limits but only conjectured — as an "intuitive judgment" — that the multilayer extension was sterile

Perceptrons (Minsky & Papert, MIT Press, 1969) contains rigorous, uncontested proofs that single-layer perceptrons cannot compute linearly nonseparable functions (XOR, large-input parity). What it says about multilayer networks is categorically different in epistemic status. At p. 232 (per the Tier 3 carrier; primary not yet read) the authors wrote:

"There is no reason to suppose that any of these virtues carry over to the many-layered version. Nevertheless, we consider it to be an important research problem to elucidate (or reject) our intuitive judgment that the extension is sterile."

The carrier text continues: "Perhaps some powerful convergence theorem will be discovered…" — the passage is a research call, not a closure. The operative phrase is "intuitive judgment": the authors themselves marked the multilayer claim as conjecture, and explicitly invited its elucidation or rejection. In Simon's own vocabulary this is a satisficing research judgment — expert intuition (claim-simon-satisficing-and-heuristic-search-same-move) exercised in a low-validity environment where it is least reliable — argued in full at observation-simon-intuition-framework-diagnoses-perceptron-sterility-conjecture.

Corroboration at the same tier: Wikipedia's Perceptron article states the belief that they proved (or even formally conjectured) a multilayer impossibility result "is not true"; Minsky & Papert later claimed perceptron research waned "not because of their book, but because of inherent problems"; critic Jordan Pollack's middle reading is that a "small proof concerning a global issue" was "interpreted by the community as a rather successful attempt to bury the whole idea."

The conjecture stood unrebutted for ~15 years not because it was strong but because no trained multilayer counterexample existed — claim-rhw-1986-demonstration-not-invention supplied that counterexample in 1986, using an algorithm that already existed (claim-linnainmaa-priority-not-paternity, backpropagation-gap). The circulating stronger version of this story is tracked at myth-perceptrons-book-killed-connectionism. Verification of the p. 232 quote against the primary is routed at question-verify-perceptrons-p232-sterile-quote.