The Precision Ratio, Revisited: Why "Ratio" Is Half Right, and What to Call the Whole
- Does "ratio" mislead? Partly. In the Ernst & Banks (2002) model the cue weights are normalized and sum to 1, so a fixed-sum ratio is literally correct for relative weighting. But the same model has the combined precision equal to the sum of the separate precisions. Integration makes the whole more precise than any single sense, which is the opposite of zero-sum.
- Should a new umbrella term be coined? Yes. The concept needs at least five dimensions: (1) relative weighting, (2) absolute or global gain, (3) integration vs. segregation mode, (4) the balance between attending inward to the body and outward to the world, and (5) persistence over time. No existing term covers all five without colliding with an established field. "Precision landscape" and "sensory reweighting" come closest, and both are already taken.
Research document · Canopticon · received 2026-10-04 · McLuhan & media theory · condensed from chat; summary, findings and recommendation verbatim, tables shortened
The research says "ratio" is accurate but incomplete. The relative weights the brain gives its senses do sum to one, so that piece really is fixed-sum. But the total precision available, the overall gain, whether signals are fused or kept apart, and how long a setting lasts all vary on their own, and some of them are positive-sum. So "precision ratio" should be kept as the name for that one measurable piece, and a new umbrella term should name the whole configuration. My top recommendation is sensory krasis (a new coinage). The runner-up is the precision mix (also new). [The author later chose syncrasis, from the same Greek root via idiosyncrasy.]
1Key Findings
- Zero-sum dynamics are real, but they're narrow and depend on conditions. Visual dominance (the Colavita effect) is robust in adults but not in children (Hirst, Cragg & Allen 2018). Heartbeat-evoked potentials index competition with touch in somatosensory cortex but facilitate self-relevance encoding elsewhere (Loescher, Haggard & Tallon-Baudry 2025).
- Cooperation is the textbook case for sensory weighting. Reliability-weighted fusion lowers variance below the best single sense (Ernst & Banks 2002; Alais & Burr 2004).
- Global gain is a separate axis. The locus coeruleus–noradrenaline system is modeled as an adaptive gain control (Aston-Jones & Cohen 2005). Gain on all channels can rise or fall together without the ratio between them changing.
- Weightings persist. Priors shift with experience and the change carries over to other tasks (Adams, Graf & Ernst 2004).
- Name collisions are serious. "Precision landscape" is already used (Lao-Rodríguez, Cacciato-Salcedo & Malmierca 2026). "Sensory reweighting" belongs to the posture literature. "Sensory profile" is Dunn's clinical instrument. "Sensorium" belongs to McLuhan, Ong and Howes.
2Minimal specification
| # | Dimension | Fixed-sum? | Key evidence |
|---|---|---|---|
| 1 | Relative weighting (precision ratio) | Yes | Ernst & Banks 2002; Peterka 2002 |
| 2 | Absolute/global gain | No | Aston-Jones & Cohen 2005; Yu & Dayan 2005 |
| 3 | Integration mode | Categorical | Körding et al. 2007; Stein & Stanford 2008 |
| 4 | Inside–outside balance | Partly | Ren et al. 2024; Suzuki et al. 2013; Aspell et al. 2013 |
| 5 | Persistence | N/A (temporal) | Adams et al. 2004; Gori et al. 2008; Assländer & Peterka 2014 |
The most useful point for the paper: under optimal integration, the combined precision is π_combined = Σπᵢ. Adding or sharpening a channel raises total precision even as it lowers the other channels' share. So the zero-sum and positive-sum readings are both true at once, of different quantities.
3Conditions a new term should meet
- It names a gap. 2. It can be operationalized. 3. It is distinct from its neighbours. 4. It earns its keep through predictions. 5. It comes with a lightweight hierarchy: one umbrella term and one measurable subcomponent (precision ratio).
4Caveats
- Several citation details were supplied from memory and should be checked against the originals.
- The collision checks were not conclusive exact-phrase searches.
- The "volume knob" quote comes from the McLuhan Substack; cite it as that Substack.
- Effects of long-term exposure to a whole medium on sensory weighting have not been directly shown. That remains the paper's hypothesis.