The claim
Let M(t) = f(m(t)) for a macro variable M and micro variables m, as Seth’s centre of mass is a function of the boids’ positions at the same time (NOTE-935, §4). Eichler and Didelez define an effect of M only for an intervention σ_M(t) independent of the other components at t, m(t) among them (Assumption 2.3(i), THEORY-704). Under supervenience, any change to M(t) is a change to m(t). So σ_M(t) cannot satisfy (i), and “the effect of M on m” in their sense is undefined, not zero.
What Granger causation from M to mᵢ does show is predictive, as TERM-058 defines it: M’s past improves the prediction of mᵢ beyond the past of the other variables in the regression. In a linear regression this can happen when f is nonlinear and the regression cannot form f(m(t − 1)) from the m’s pasts. Seth’s preprocessing makes f nonlinear (NOTE-935, Limitations). That is a fact about the predictor, which is what Seth says G-emergence is relative to (§5.2).
The step from “every micro variable is measured” to “the Granger result is interventional” is exactly what CLAIM-333 denies in general. Here it fails for a different reason than the sampling step: the intervention is ruled out by the definition of M, not by the data.
Why it is filed
Seth claims no more than prediction. The disclaimer is for readers of the record. NOTE-726 and NOTE-791 compare Seth’s measure with Hoel et al.’s and Barnett and Seth’s, and NOTE-010 and THEORY-497 discuss downward causation in ΦID. “Downward causation” in Seth’s sense should not be read as the interventional kind when the accounts are compared.
What it does not say
- That macro variables cannot have interventional effects in any sense. An intervention defined on the micro variables, with M as a summary of its effects, is not ruled out. Nor is an account of causation that does not require holding the micro base fixed.
- That Seth’s measure is wrong. It is a measure of prediction, as stated.
- Anything about ΦID’s downward-causation atom D⁽ᵏ⁾, whose interventional reading has its own conditions (NOTE-010, C4).
Note of 2026-10-27: the edge says what the claim uses
The status review of this date (ADR-057, ADR-058) replaced grounds: THEORY-703 with Seth’s paper itself, LIT-1170 (read in full in NOTE-935). The claim quotes Seth’s definitions and label. THEORY-703’s contested part, that G-emergence tracks weak emergence, is not used.
Note of 2026-11-02: contextual emergence denies the premise
In the exchange “Assess SDT Coverage” (2026-11-02), A5 proposed testing Jennings’s attention as free will by perturbing slow oscillations “while holding relevant local inputs … constant”. Under supervenience that is an intervention on the micro base, which is this claim’s point. Jennings’s abstract distinguishes contextual emergence (Bishop and Atmanspacher, LIT-1987) from supervenience. So if her paper succeeds, it falls outside this claim’s scope rather than refuting it. That is the precise point of contact, and A5 missed it. Bishop and Atmanspacher’s contextual emergence claims stability conditions for a well-defined higher-level description, not new causal powers. Whether Jennings departs from that is a question for a reading of the paper.
The claim is unchanged.
Read on 2026-11-03 (LIT-1963, NOTE-1508). The paper denies supervenience and says it needs neither downward nor circular causation (fn. 14). Its reason that the boundary resists redescription is that the boundary’s properties “are macro-level ones” (p. 20), and it offers no intervention on the macro level that holds the base fixed. So the note stands: the premise is denied, not refuted (THEORY-1314).