THEORY-876: Phenomenal consciousness is the coherent functioning of an embodied agent's generative model of its sensorium, coherent in space, time and cause and run either coupled to sense data (perception) or decoupled from it (imagination); embodied grounding, the model's capacity to inform and be informed by action–perception cycles at behavioural timescales and the history of such coupling by which it was learned, is necessary, so that integrated information or a functional global workspace without it does not suffice

Source

Safron, Klimaj & Sheikhbahaee (2026), LIT-1345, read in full in NOTE-1057 from the publisher’s PDF: the abstract (p. 345) and the sections on coherent modelling, embodiment and machine consciousness (pp. 346–349). The earlier IWMT papers were not read.

What was actually shown

Nothing is shown in this paper; the theory is stated.

• The identity. Phenomenal consciousness is what it feels like to be “the spatiotemporally and causally coherent functioning of a probabilistic generative world model for an embodied agent” (p. 345), “entailing either perception or imagination when respectively coupled to or decoupled from sense-data” (p. 346).
• The necessity of embodiment. “Critically, IWMT requires (body-)world models to be capable of both informing and being informed by action-perception cycles at behaviorally relevant timescales” (p. 345). Against IIT, systems could have “arbitrarily large amounts of integrated information and still not be conscious without embodied grounding”; against GNWT, they “could instantiate functional global workspaces without entailing any kind of subjective experience” (p. 346).
• Historicity. A feedforward workspace “would be divorced from the environmental couplings and historicity by which adaptive predictive (world-)models are learned” (p. 349).
• What could have come out otherwise. The paper proposes that IWMT could be tested by building systems that meet it and “examining whether various conscious functions are realized” (p. 349). That tests conscious functions, not experience.

What this does not say

• It does not require current sensory input. Imagination is the model decoupled from sense data, and it counts. The requirement is a capacity and a history.
• It does not set a threshold. Embodiment comes in “sufficient degrees (and kinds)” (p. 345), and “virtual” embodiment is allowed (p. 346).
• It does not settle substrate. A simulation “would recapitulate the phenomenology of wetness within the simulated world” (p. 349), yet realisation “may practically (and perhaps also in principle) require” physical recurrence (p. 349). The paper holds both.
• It is not argued here. The paper restates and aligns; the case for the theory is in work not read.

Connections

• CLAIM-413. The claim’s occurrence half concerns current veridical grounding, which this account does not require. The two part on developmental grounding and the capacity for coupling, on which the claim says it is silent. This THEORY is the enactive objection that claim leaves open, stated as a positive account.
• QUESTION-254. Answers it in part: coherence, egocentric organization, recurrence and embodied grounding together, set against integration alone and broadcast alone. Its verdict on language models comes from the degree of embodiment, which is not specified.
• CLAIM-412. A rival identity of the same form with a different organization, an embodied world model in place of interpretive organization.
• THEORY-063, TERM-072. It allows that a minimal self-model may be required, against THEORY-063’s second clause, and makes perspective egocentric. THEORY-882 (Laukkonen et al.) shares the coherent world model and drops both embodiment and perspective.