The inference
From. CLAIM-046: “pragmatic identity and fidelity should be defined on observable responses before any equivalence of internal states is assumed”. CLAIM-112 says what kind of claim that was: “a choice of what to measure, not a theory of what meaning is”. CLAIM-038 (A78), the operational thesis it complements: “We can begin with local observational data and investigate whether a global object exists.” The record files that object in TERM-014, “an equivalence class of situations relative to a chosen family of measurements”, and TERM-042, “a context-indexed system of empirically admissible local assignments or distributions”.
To. CLAIM-139: “A communicative kind’s identity is grounded in its relational organization rather than in any expression or encoding”. Its unit, TERM-040, is “Not an observational class (TERM-014)”, and “may be real while remaining indistinguishable under an insufficient family of empirical probes” (A218 Ch7).
The step. CLAIM-142: “It does not withdraw CLAIM-046’s methodological rule. Evidence still comes from observable responses. What changes is what that evidence is of.” The kind is offered as what best explains the observations.
The two stances disagree on a case. Two realizations that agree on every probe in the family are one object on TERM-014 and may be two kinds on TERM-040.
Critical questions
- Which rivals were considered? One, by the exchange itself, and only to concede it: A214 §13, quoted in CLAIM-139, “A traditional ontology of intrinsically individuated objects plus relational properties might reproduce the same statistical predictions.” The operational stance, which posits no kind at all, is the other, and it stands in the record too: CLAIM-046 and CLAIM-112 are Active, and TERM-014 and TERM-042 are Active.
- Does a rival explain as well? Yes, on the exchange’s own account. This is the question that fails. CLAIM-247: CLAIM-142’s sentence “keeps CLAIM-046 as a rule of method and says the evidence is now of a kind. It does not argue for the kind.” TERM-014’s “Where it went”: “No turn addresses this entry.” The owner’s turns over the change asked for the argument to be laid out (U56), tied together (U59) and turned into a proposal (U60).
- Would more evidence discriminate them? Not the probes in hand. By TERM-040’s own definition a kind can outrun the probe family, so the data that would choose between the kind and the class are the data the operational account says are all there is.
Proposed. The record holds both stances, and no claim says which bearer the line now holds. CLAIM-115, the manuscript’s thesis, uses no TERM for it. The step happened without an argument, as CLAIM-140 says of the receiver. CLAIM-142’s reading, method kept and object changed, is one coherent reconciliation, and nothing here says the realist reading is wrong. It goes Active when a turn argues the kind over the class; it is Rejected if the record holds that the class suffices.