Why it is a question
TERM-076’s restriction, to an ordering of the system’s own states by their bearing on its maintenance, keeps any potential goal from counting as valence. But a system can rank its own states by a proxy that once tracked maintenance and has come apart from it.
- The organizational case. In Wells Fargo’s Community Bank (CASE-151), cross-sell was treated as the measure of how the bank was doing, and it won the bank’s conflicts for a decade against risk, HR and complaints. It was not the bank’s viability, which the practices it rewarded endangered. CLAIM-432 reads that case as threat without remorse; this question asks what kind of valence, if any, the metric’s potential was.
- The personal case. In addiction a reward signal that once tracked the organism’s maintenance comes to govern its choices against it. The record holds Ainslie on the breakdown of will (LIT-586), which may bear on this.
What would count as an answer
A criterion that says when an ordering over a system’s own states is “by their bearing on its maintenance” and when it is by a proxy, applicable to both cases; and a verdict on whether proxy-keyed potential is valence of a distinct kind, mis-keyed valence, or not valence at all.
What would not answer it
The observation that the proxy and maintenance diverged, which defines the case without answering it.
Note of 2026-10-23: a mechanism, and a coupling
A mechanism on the personal side. Incentive sensitization (THEORY-895, from Berridge and Robinson, LIT-1461, NOTE-1109) re-keys the strength of ‘wanting’ to cues, lastingly (“Once induced, sensitization is very long lasting, and possibly even permanent”), gated by context (“not always hyper-active, but it is hyper-reactive to drug cues and contexts”), and cut loose both from pleasure and from the person’s own goals. That is proxy-keyed potential with each of Frijda’s three marks accounted for. It also sharpens this question: the cue-keyed ordering is no more an ordering of the system’s own states than ‘liking’ is (CLAIM-433’s note of the same day).
A coupling between a firm and its users. Flayelle and colleagues’ taxonomy of design features (LIT-1465, NOTE-1096) describes a provider-side selection: “design features that increase the frequency of online behaviour, time-on-device or spending (and therefore revenue) will be maintained, whereas those that do not will be abandoned” (p. 6). Read with the record’s account, two proxy potentials are coupled: the firm’s engagement metric prevails in its negotiation, and the features it keeps install cue-keyed potential in its users. The mapping of features onto the account is the record’s, not the authors’:
- reinforcement schedules and personalized cues act on strength;
- moving goalposts act on the reference course, which “keep individuals engaged for longer than planned” (p. 8), and predict mildly negative valence during use;
- near-misses and losses disguised as wins “contribute to the false belief that one is progressing” (p. 9), a user-side mis-modelled valence, counterpart of CASE-151’s misinformed board;
- autoplay wins without any conflict being posed, which a conflict record (QUESTION-268) would miss.
This joins Doctorow’s account of platform decay (NOTE-041) to CLAIM-432’s threat without remorse: the harm to users need not be represented anywhere in the firm. The taxonomy’s evidence is mostly cross-sectional or qualitative, and its authors warn against pathologizing ordinary use, so it is a description, not grounding. A documented case of a platform whose own records show features kept for engagement despite internal signs of harm would test the coupling; candidates exist, but none is read, and their standing (leak, complaint) would need settling first.
Note of 2026-10-23: the proxy as a protective lock-in
From the comparison of the record’s will and affect lines, at the owner’s direction. The question is the will line’s distinction between stability and viability (CLAIM-079) moved into the affect line, and the will line already has a term for a regime that stays stable while viability erodes: protective lock-in (TERM-020). Incentive sensitization (THEORY-895) fits its five steps closely. A drug or a cue-keyed reward activates a pursuit regime; the regime reorganizes ordinary activity; the reorganization suppresses what would restore the rest, as ‘liking’ falls and other concerns lose their contests; cue-triggered ‘wanting’ sustains the activation; and the capacity to leave it falls. QUESTION-045 asks the same of mental disorder at the individual scale, and TERM-020’s dynamical form, a bistable activation with hysteresis, is CASE-112’s.
So a candidate answer: a proxy-keyed persistent potential is valence on the account, of the pursuit perspective, and it is a protective lock-in when its continued prevalence reduces the capacity to revise it. Metric fixation in a firm would be the organizational case. The fit is by analogy; it does not yet show that the maintaining link, and not the proxy’s content, is what keeps either in place, which is QUESTION-045’s test.