Abstract
<title>Abstract</title> <p>Archaeologists routinely infer function from morphology, yet a stage-linked pattern can arise from task-specific engineering, inherited production tradition, population differentiation, selective survival, or the record-formation processes that decide what is observable—and no general protocol distinguishes admissible evidence from evidence the record cannot decide. We introduce a component-diagnostic framework for when morphology can support the causal inference of task-specific engineering. It decomposes an engineering claim into causally distinct components and screens each, before any confirmatory test, through a six-stage admissibility gate (construct validity, scale validity, observability, identifiability under missingness, precision, applicability); admissible components then pass to a diagnostic layer for functional specificity. A failing component is reported with its reason, not mistaken for a disconfirmed one. A distinctive feature is prospective: where a component is non-identified, the same bounds compute the minimum observation that would change its state, turning non-identification into a research design. We demonstrate it on a claim about monument transport on Rapa Nui (the walking-moai hypothesis). A claimed wide transport-stage base yields a real, size-robust contrast (Welch t = 3.18, d = 0.55) that is carried by the shoulder, not the base, and shows no functional specificity; the mechanically relevant D-shaped outline is non-identified, its bounds overlapping so completely that even their ordering is undetermined, its rescue frontier requiring near-complete observation of road-stage bases; and basal observability rises selectively against matched non-basal controls, making missingness itself an operational signature. Both impasses are observational equivalence, arising once in measurement and once in the record.</p>