Jump to text
On this page
Copy and source
Published source confirmed at last check

Source changed 2026-11-22 17:28:01 UTC · snapshot created 2026-11-22 17:28:59 UTC · last check 2026-11-22 18:05:10 UTC

Foundational.Reference:7.3 - Keep the strength of each ground visible

A programme can use several kinds of support together. Preserve the inference each supplies.

GroundSupported consequenceWhat remains to establish for a stronger use
A derivation from stated premisesThe consequence holds if the premises and derivation holdWhether the subject satisfies the premises; whether the required construction is available
A model with specified inputs and operationsA construction from which a conditional behavior can be derived or computedThe derivation or computation giving the particular conclusion, and the correspondence to the subject if that conclusion is used there
Output tied to an identified simulation runWhat that implementation returned for those inputs and settingsNumerical accuracy, whether other runs meet the claim, and the physical conditions needed to use the result for the subject
Evidence of a performed interventionThe action taken on the identified subject under the established conditionsWhat happened in response, including whether the response was measured or recorded
An observation or a retained indication tied to the subject and procedureA conclusion about the observed quantity within the instrument’s coverage and uncertaintyWhich further conditions, objects or explanations the observation can support
A conjecture or analogyA candidate account or connection worth developingIts precise correspondence, distinct consequences and grounds for relying on them

The hypotheses can be substantial. Constructor theory proposes physical principles about possible and impossible tasks, including repeatable performance. This suggests asking whether the means retain the ability to perform again and what makes an inverse transformation available. PHY.3 carries that useful distinction while retaining the particular physical premises of each argument. Accepting the complete proposed theory is a further physical question.

Fields, Glazebrook and Levin’s minimal physicalism proposes an information-based account extending cognition across physical scales. Its subject and assumptions matter: the proposed extension can guide questions about interaction, memory and persistence, but adopting those questions does not settle attribution of cognition or experience to every system. A programme investigating an adaptive machine can identify its observation, internal update and action without first resolving that broader attribution.

Vanchurin and colleagues’ multilevel-learning account, especially its stated principles and treatment of conflicting interactions, motivates questions about learning across different temporal scales and changes of organization. The account assumes a loss function, scale separation and further conditions on information flow and change. A proposed application must supply those counterparts or state the altered hypothesis. An engineering conflict alone does not establish an inevitable increase of complexity or an obligation to continue development indefinitely.

These are productive lines of inquiry with different premises and targets. Similar equations can suggest a correspondence, but the correspondence must explain which quantities, operations and consequences are preserved. In particular, an objective called free energy needs its own definition and units before it can enter a physical energy account. Keep a successful conditional model or engineering method usable while a wider unification remains unresolved. A failure of use then reopens the implicated implementation, correspondence or premise; it does not identify the ontology as the unique cause.