Source
Shanahan (2024), “Simulacra as Conscious Exotica”, Inquiry 69(6), 2982–3010, LIT-465, read in full in NOTE-382 from arXiv:2402.12422 v2; the version of record was not reached, so sections and quotations are the preprint’s.
What was actually shown
The method (§§4–5). Following the private-language remarks, consciousness words “have meaning only insofar as they relate to what is public”: bodies, brains and behaviour. For an exotic entity, the question becomes whether we can “engineer an encounter” with it in a world we share, and how a “society-wide conversation” would then come to speak of it. There is “no more to the truth of the matter” than the consensus reached, and where there is none, “no residual philosophical mystery”.
Candidature (§§5.2–5.4). The ability to engineer an encounter, “even if only in principle, establishes an exotic entity’s candidature”. The octopus and an alien cube containing a simulated world pass. If an entity’s behaviour is wholly unintelligible, the language of consciousness does not apply: the “void of inscrutability” is empty.
Chat agents (§§6.1–6.2). “It is not possible to engineer an encounter with a simple conversational agent, even in principle”, since “we cannot be with them in a shared world”. Tools, multimodality, memory and planning legitimise talk of belief and intention, but “we still cannot engineer an encounter”, and “nothing has changed” for consciousness.
Embodied agents (§§6.2–7.2). Robots and virtually embodied agents can be encountered and so are candidates. In an extended scenario, a community might come to treat embodied simulacra as conscious, retract on learning the mechanism, find the gap closed, or invent a “consciousness adjacent” vocabulary. The paper describes these language games without taking a side.
What could have come out otherwise. Given the method, little: the verdict on chat agents follows from the candidature test and the fact that they share no world with us. What could fail is the method, and the paper does not test it.
What this does not say
- That chat agents lack experience as a matter of fact. On the method, there is no hidden fact for them to lack. The verdict is about whether the words apply, not about an inner state.
- That the method is established. It is summarised from Wittgenstein, with “no attempt here to offer a definitive interpretation” (fn 8). Chalmers’s pure-thinker view is named and set aside (fn 18), and readers who do not share the method get no argument against it (NOTE-382, Limitations).
- That embodied simulacra are conscious. They are candidates, and the outcomes of the scenario are listed, not weighed.
- That a minority view can be right. If there is no truth beyond convergence, holding a view the community rejects is “hard to reconcile” with the account, as the paper concedes (§5.5).
- Anything settled about the self. The two self passages, that a chat agent has “nothing that would qualify as a self worth preserving”, lacking body, history and autobiographical memory (§3), and that a role player has no “person behind the mask” (§7.3), are asserted. The second takes a stable human core self for granted, which the author’s 2025 paper gives up (THEORY-490). Neither is part of this account.
Connections
- THEORY-023. This account does not take a side in that dispute; it denies its presupposition. If the method holds, the deadlock over AI consciousness is not epistemic: there is no fact beyond a consensus, and for chat agents the grammar of the word answers now. Its promotion condition is untouched, since the paper offers no evidence of the kind it asks for.
- The evidential approach (THEORY-460, from Chalmers; LIT-056) is named as the rival method, and Chalmers and Nagel (THEORY-694) as statements of the dualist intuition the paper sets out to dissolve.
- Embodiment. Seth (LIT-135) is quoted on consciousness and living systems. THEORY-090 and THEORY-093 tie consciousness to the body on other grounds, physiological rather than grammatical.
- The chat agent’s self. Shanahan’s 2025 account of what a chat agent’s “I” refers to is THEORY-490.