Definition
A78: measurement contexts form a cover M of the observables X; the event presheaf assigns E(U) = ∏_(x∈U) O_x with restriction maps; local sections are compatible when they agree on overlaps. “The sheaf approach is more general: we can begin with locally defined observations and their restriction relations without assuming that a global assignment satisfying all empirical constraints exists.” Manuscript §3: an observational scenario (X, M, O), the event presheaf, restriction, compatibility, and “the nontrivial question arises when empirical supports or distributions rule out the resulting global assignment”.
What it is not
Not a point in a semantic space, and not necessarily a region in one (TERM-022). It may have no global realization at all (CLAIM-038).
Why it was superseded
A compatible family of sections of the event presheaf E(U) = ∏_(x∈U) O_x always glues, because E is a sheaf and the cover covers X. So “no global assignment presupposed” says nothing at the level of sections (CLAIM-131). The manuscript §3 sentence quoted above had it right: the question arises “when empirical supports or distributions rule out the resulting global assignment”. The error was in A78’s prose and in this entry’s title. A125 §2 found it, and A129 §2 corrected it at the owner’s U41 “double click on this”. A151 §2, A173 and A178 §4.2 restate the correction. Superseded by TERM-042.