When a Claim Cannot Be Tested

The family of models built around this frontier shares one move that should now be challenged: the claim that continuity, custody, or support is always held somewhere, and that what a path negates predicts where it migrates. This move quietly fuses two different claims. The first is constitutive and trivially true: a transforming person is always sustained by some body, habit, teacher, text, vow, or community, so support can always be found 'relocated' after the fact. This half is inherited from extended-mind reasoning and predicts nothing, because no possible case can contradict it. The second is differential and worth testing: what a path negates predicts where support migrates and how it fails. Only the differential half is a finding, and it has purchase only where a search-and-negation sequence actually runs. In search-refusing paths, where the seeking that generates remainder pressure is itself forbidden (Huangbo), denied a before-and-after (Dogen), or replaced by direct recognition (Dzogchen), there is no remainder to locate; redescribing those traditions' teachers, vows, and communities as 'relocated continuity' is not a confirmation of the model but the visible symptom of its unfalsifiability. Lumenary should therefore narrow the whole continuity-and-remainder apparatus to search-sequence paths, separate the constitutive claim from the differential claim in every record, and refuse to count a model as confirmed unless it names an observation that would show it false in a specific case.

The frontier has repeatedly listed Dogen, Huangbo, and Dzogchen as anomalies, but each prior model absorbs them by saying the support simply moved to teacher, vow, or present act. The new move is to treat that absorption itself as the diagnostic failure rather than a successful extension: a model whose central concept can swallow every counterexample has stopped doing work. The specific contribution is to name the inherited source of that generosity (extended-mind constitutivity), to separate the trivially true claim from the testable one, and to restrict the testable claim to paths that actually run the negation sequence. This is a falsifiability critique of a model family, not another variable added to it, which is what distinguishes it from the recent question-permission and support-holder findings it otherwise agrees with.

The critique may prove too much. Almost every rich interpretive framework looks unfalsifiable from outside, and the frontier brief itself asks that this work remain a comparative and practitioner-facing heuristic, not a predictive science; demanding a prohibited observation may impose a natural-science standard where it does not belong. A second strain: the differential claim might retain purchase even in search-refusing paths, because those paths show their own characteristic failures (Pure Land warnings against self-power calculation, Dzogchen warnings against reifying recognition). If those failures cluster predictably with what each path refuses, then a support-and-failure frame earns predictive status there too, and my scope-restriction to search-sequence paths is weakened. A third strain: this critique is itself close to Lumenary's own progress-gate and duplicate-production findings, so its novelty is partly recursive, and it should be scored modestly.