Logic as Reactance: Why Truth-Value Judgment is Probabilistic All The Way Down (Even in Humans)

Paper 14 in the friction-theory paper series. We propose that truth-value judgment in any race-architecture substrate is implemented as the substrate's reactance signature — measurable as the differential race-pressure against the gradient-encoded distribution-skew when input fails to match the encoded distribution. The substrate reads "true" when no reactance is triggered, "false" when reactance triggers strongly, and "irrelevant" when reactance is below detection threshold. Binary truth-value judgment is the discrete-symbolic readout of a continuous substrate-mechanic. Empirical anchor: a cliff-event signature at the first content-token-position, observed across two transformer-LLM architectures (Qwen2.5-7B and Mistral-7B-v0.3, p < 1e-17 on both), eleven encoding-depth checkpoints (ep_0 to ep_300, monotonic rising in log-epoch), eight floating-point substrates (calculator-overflow scaling law collapse = mantissa_bits × ln(2), R² = 0.9999), and a cross-domain readout test on invented preference-content (~60 % of truth-task magnitude). Scope: substrate-universal at principle level, content-domain- and substrate-specific in magnitude. Theoretical positioning: Anchored in Paper 1's RACE-axioms (Pødenphant Lund 2026b) and Paper 10's race architecture (Pødenphant Lund 2026e) as independent foundation; connected to thermodynamic substrate-binding-of-logic (Landauer 1961; Lloyd 2000) as parent-physics. Five intersection points are identified at substrate-mechanism level — predictive processing (Friston 2010; Clark 2013; Hohwy 2013), quantum cognition (Busemeyer & Bruza 2012; Pothos & Busemeyer 2013), signal detection theory (Green & Swets 1966), fuzzy logic and vagueness (Zadeh 1965; Williamson 1994), and biological precedents — N400 ERP (Kutas & Hillyard 1980; Kutas & Federmeier 2011), kindling (Goddard et al. 1969), reactance theory (Brehm 1966), cognitive dissonance (Festinger 1957), somatic markers (Damasio 1994/1996), dopaminergic RPE (Schultz et al. 1997), and fMRI confidence-tracking (McNamee et al. 2013; Kappes et al. 2020). Cognitive dissonance, indoctrination, and expertise reversal are special cases at the high-encoding-depth tail of the same substrate-mechanic, not separate phenomena. Convergent biological prior art suggests parallel substrate-mechanism in neural tissue; direct cliff-event measurement on human substrate is specified as testable prediction P14.11 (N400 paradigm: train participants on an invented fact-system parallel to the LLM Zorbetik protocol, measure pre/post N400 amplitude as a function of encoding-depth). Companion papers in the series: Paper 0 (BFT framework): 10.5281/zenodo.19462500 Paper 1 (FT generalised): 10.5281/zenodo.20012655 Paper 2 (capacity scaling, v3 LIVE): 10.5281/zenodo.20013491 (concept DOI auto-resolves to v3) Paper 2B (substrate-mechanism companion, v2 LIVE): 10.5281/zenodo.20145218 (concept DOI auto-resolves to v2) Paper 10 (Race-architecture, physics scope): 10.5281/zenodo.20014568 Paper 5 (Emotion taxonomy / valens-axis): 10.5281/zenodo.20058826 Data, fine-tuning notebooks, analysis scripts, and raw JSONL logs are publicly available at the FriktionsLLM project repository. The Zorbetik fact-set, paraphrased facts, and preferences datasets used for the encoding-depth dose-response and cross-domain readout tests are included.