The claim
THEORY-101 and THEORY-090 are declared rivals on what the first consciousness did. For Veit it weighs options in a common currency. For ALARM it gives an extreme signal “100% weight” through behavioural programmes whose cost avoidance “does not involve a choice”. THEORY-101’s body names the condition under which the rivalry would dissolve: “if ALARM’s “100% weight” were recast as a weight in a common currency rather than an override of it.”
On the owner’s reframing (CLAIM-440) both are potential of an ordering of the animal’s own states.
- Veit’s currency is compensatory potential. The potential ordering trades off against the others in proportion: a larger resource buys a larger cost. As an ordering of cost–resource pairs it is continuous, and has a continuous numerical representation (CLAIM-436, part 3).
- ALARM’s override is overriding potential. Past a threshold the alarm decides, and other concerns decide at most its ties. ALARM’s “no choice” is the limiting case in which the potential ordering leaves one option at the top.
Whether the override has a currency depends on its form (CLAIM-436, part 3a).
- A threshold override, alarm on or off, ranks everything above the threshold below everything under it. It has a currency, a discontinuous one: in that form it is an extreme weight, and THEORY-101’s dissolution goes through.
- A graded override, the alarm deciding at every level of its own intensity with other concerns breaking only exact ties, has no currency at all. In that form the rivalry does not dissolve.
Either way the two accounts agree that the earliest valence is potential by an ordering of the animal’s own states. They disagree on whether that ordering is continuous, and, if it is not, ALARM’s form decides whether any currency remains.
The test
THEORY-101’s promote_when names the experiment: “an animal accepting a noxious cost for a resource, in proportion to both”. Fit the animal’s choices between cost–resource pairs.
-
Veit predicts a continuous indifference curve: for every cost some resource buys it, and the price rises smoothly.
-
ALARM predicts a discontinuity at a threshold: above it no resource buys the cost. Whether resource still breaks ties among options above the threshold that differ in cost tells the threshold form from the graded one.
-
A realistic override has a third signature. If the alarm decides only past a discrimination threshold, the rule is a lexicographic semiorder (CLAIM-436, part 3b), and a chain of options each a little costlier and a little richer than the last produces cyclic choices: each step taken, the endpoints reversed. Tversky’s intransitivity experiments (LIT-1472, unread) are the design. Cycles are evidence for an override with a threshold and against a currency of either kind, which cannot cycle.
-
Fit a model that can jump. Meuleman et al. (LIT-1463, NOTE-1104) report appraisal “threshold” effects from regression splines, whose curves are continuous by construction, and note that “Sigmoidal curves can also be viewed as continuous approximations of binary threshold functions” (ms p. 10). A smooth fit cannot tell a steep currency from an override; the test must fit a model that allows a discontinuity and compare.
-
Sweep the cost both ways. Raising and then lowering the cost tells a fixed threshold from a hysteretic one, which the reframing does not place.
What the data so far show. Gibbons et al. (LIT-1466, NOTE-1114) found bumblebees’ feeding from a noxious (55 °C) feeder graded with the sucrose offered elsewhere: “Bees avoided noxious feeders when the unheated feeders contained high sucrose concentrations, but progressively increased feeding from noxious feeders when the sucrose concentration at unheated feeders decreased” (p. 1). The companion hermit-crab study (LIT-1470, from its abstract) found crabs in preferred shells evacuating “at a higher shock level than those in poor quality shells”. Both look compensatory at the costs tested. Neither reaches the question: the bee study used one cost, ran the resource levels in separate groups and fitted a smooth model that assumes gradedness, and the crab studies used two shell qualities, too few to show a ceiling. Within-subject cycles of Tversky’s kind are also hard to establish; the replication record is contested (Birnbaum and Gutierrez 2007; Regenwetter et al. 2011; not read).
Both regimes can occur in one animal, a currency below the threshold and an override above. On this reframing that is expected, one ordering whose potential changes regime past a point, unless the two are shown to be separate mechanisms, which is the first defeater.
What it does not say
- That either account is right about the origin of consciousness, or that either regime is felt. The recasting is functional.
- That the mathematical contrast settles the empirical one. A finite set of tested pairs can always be fitted by a currency (CLAIM-436, part 1), so the test needs enough pairs to tell a steep curve from a discontinuity, and a design showing that the discontinuity does not move with the size of the resource.
Note of 2026-10-29: what holds it back
It rests on CLAIM-440, the owner’s reframing of valence as ordering-potential, which the record does not hold. It grounds on THEORY-101 for Veit’s common currency and on THEORY-090 for ALARM’s override, readings still Proposed on whether each account is true, and no held reading states either author’s proposal alone.