What This Series Isn't Claiming

This series describes an architecture, not a discovery. The table below states every substantive claim made across the four pages and its actual confidence level, so a reader arriving at any single page through search can check the whole series' position in one place rather than reconstructing it from four separate pages.

Claim Status Where
Object boundaries and behavioural expectations are perceived together, not sequentially Established, direct empirical finding Page 1, Spelke (1990)
Shared traits across categories work through composition, not rigid inheritance Established in cognitive science; independently established in software engineering Page 1, Rosch (1975)
A system can model itself using the same machinery it uses for external objects Established as a working engineering result in at least one real robot Page 1, Bongard, Zykov & Lipson (2006)
Self-modelling and other-modelling may share cognitive machinery Established as a serious, evidenced hypothesis; not universally accepted as settled Page 1, Gallese & Goldman (1998)
A self-model can operate in a transparent mode (invisible to the system's own reasoning) or an opaque mode (explicit and inspectable) An architectural design choice this series proposes, built from an existing mechanism Page 1
The transparent/opaque distinction matches Metzinger's account of what makes a self-model feel like a perspective A proposed mapping between an engineering mechanism and a philosophical theory, not a proof that the mapping is correct Page 2
Global Workspace Theory can be implemented with the same shared-resource mechanism already used for inter-agent coordination A proposed implementation strategy, not a claim that this is the only or correct way to implement it Page 2
Building a system with these signatures produces consciousness Not claimed anywhere in this series Pages 1–2, explicitly
Building a system with these signatures produces a functional and reporting profile resembling a claim of experience The actual engineering target this series argues for Page 2
Whether that profile requires a further, harder explanation (Chalmers) or is all there ever was to explain (Dennett, Frankish) Genuinely contested in philosophy of mind; this series takes no position Page 2
A transparent self-model might carry real moral stakes A serious, published argument (Metzinger, 2021), not a settled conclusion Page 3
That argument's specific conclusion (a global moratorium) is correct Disputed in published literature; this series does not adjudicate it Page 3
Genuine uncertainty about moral status is itself a reason for caution A separate, independently argued position (Schwitzgebel & Garza, 2015) this series finds persuasive as a general principle, not as proof any current system meets the relevant bar Page 3
Whether the transparent mode is currently in use in any real system can be verified from outside Explicitly argued against — this series holds it cannot, by the same logic as ordinary agentic verification failures Page 3

The one distinction that governs everything above

Searle's weak/strong AI distinction, stated once on page one, applies to every row in this table without exception: everything above describes a model or analogue of a process, argued on functional and structural grounds. Nothing above claims the model is the process itself, and nothing above resolves whether that further claim is even a coherent one to make.

Where this connects