probe-agency-knowledge-experiment

{
 "question_id": "q:brain-if-consciousness-fundamental",
 "source_role": "protocol-agency-knowledge-boundary",
 "tradition": "cross_tradition",
 "central_reading": "The filter account (Trika kañcuka ontology) treats agency-limitation (kalā-kañcuka) and knowledge-limitation (vidyā-kañcuka) as categorically distinct metaphysical dimensions — kalā contracts universal agency into limited efficacy, vidyā contracts universal knowledge into limited cognition. Predictive Processing treats both agency (precision of action-outcome predictions) and knowledge (precision of perceptual priors) as variations in a single graded precision-weighting parameter. This difference is testable: if kalā and vidyā are categorically distinct, independent factorial manipulation of agency and knowledge should produce dissociable behavioural signatures (agency selective effects on intentional binding, knowledge selective effects on metacognitive efficiency) that cannot be reduced to a single latent precision dimension. The proposed Agency-Knowledge Dissociation Task (AKDT) operationalises this as a 2×2 within-subjects factorial design with agency (75% vs 25% action-outcome contingency) and knowledge (80% vs 55% cue validity) as factors, using meta-d' as the shared behavioural readout, with Bayesian model comparison (single-parameter vs two-parameter models) as the decisive test.",
 "claims": [
  {
   "id": "cl:kalavidya-categorical-vs-graded-distinction",
   "text": "The Trika kañcuka ontology specifies kalā (agency-limitation) and vidyā (knowledge-limitation) as categorically distinct orthogonal limitation dimensions, contra the Predictive Processing framework which models both agency and knowledge effects as graded variations in precision-weighting of prediction error. This generates testably different experimental predictions at the agency/knowledge boundary.",
   "source_id": "text:tantraloka-kancuka-panchaka",
   "source_span": "Tantrāloka, Kāñcuka-pañcaka section: kalā limits universal agency (kartṛtva) to limited efficacy; vidyā limits universal knowledge (jñānatva) to limited cognition — described as distinct restriction axes, not as variations of a single parameter",
   "evidence": "textual",
   "pramana": "anumana",
   "hetu": "The six kañcukas (kalā, vidyā, rāga, niyati, kāla, māyā) are described in Tantrāloka as orthogonal limitation dimensions each constraining a distinct aspect of universal consciousness. If they were merely parametric variations of a single restriction mechanism, their distinctions would dissolve under factorial experimental manipulation that independently modulates agency and knowledge.",
   "sadhya": "Independent factorial manipulation of agency and knowledge will produce dissociable behavioural signatures — agency effects on intentional binding and sense-of-causality without corresponding shifts in metacognitive efficiency, knowledge effects on metacognitive efficiency without corresponding sense-of-causality shifts — that cannot be captured by a single latent precision parameter within standard PP models.",
   "vyapti": "If kalā and vidyā are categorically distinct limitation dimensions (orthogonal restriction axes in kañcuka ontology), then factorial manipulation of agency and knowledge should produce qualitatively distinct patterns of behavioural dissociation (main effects on different readouts, significant interaction term), whereas PP predicts that both manipulations modulate a unified precision-weighting parameter producing additive effects on whichever readout most sensitively indexes precision, with no residual variance requiring qualitatively distinct dimensions.",
   "vyapti_confidence": 0.55,
   "falsifier": "All phenomenological and behavioural variance from a 2×2 agency × knowledge factorial design is adequately captured by a single latent dimension corresponding to precision-weighted prediction error, with no evidence of categorical boundaries between agency-dominant and knowledge-dominant conditions as det