THE COST IS NOT THE SOURCE: Virtue Signalling, Costly Signals and the Better Forgery of a Virtue
Abstract The recent philosophical debate on virtue signalling has two sides. Tosi and Warmke treat moral display aimed at status as a vice of public discourse. Levy answers that virtue signalling is virtuous: signals that are costly or hard to fake are reliable, and they supply higher-order evidence about moral questions. This paper accepts Levy’s framework and contests its central inference, that paying a cost makes a moral signal credible. Signalling theory itself does not support the inference: what keeps signals honest is cost that differs between those who have the quality and those who do not, not cost as such. For moral acts whose cost is the same for everyone who performs them (a donation, a boycott, a public stand), the paper proves a Reward Ceiling. A costly moral act separates the committed from the uncommitted only while the reward the audience pays for it stays below its cost. For any log-concave distribution of costs, the probability that a signaller is committed falls steadily as that reward rises, while the act’s form and cost stay unchanged. Audiences that learn can correct this; scores, ranking algorithms and reward models, which pay by rule, cannot. Joined to the programme’s Competence Theorem, the result yields a second claim. Where audiences read cost as evidence and relax their checking as cost rises, the most expensive forged signals are the most damaging. The paper replaces both cost (Levy) and motive (Tosi and Warmke) with a structural criterion of source: a moral display belongs to the commitment it signals to the degree that it persists when the audience’s reward is removed, and again when it is reversed. Two predictions are registered with their falsifiers. The economics of image motivation (Bénabou and Tirole) already contains the dilution effect; the paper concedes this and claims only its application to the moral-philosophical debate, the competence extension, and the perturbation criterion. The consequence matters beyond the seminar room. Societies that score conduct, and AI systems trained on approval, raise the reward for moral display as a matter of design. By the Reward Ceiling they make that display less informative the more they reward it. Keywords: virtue signalling; moral grandstanding; costly signalling; honest signals; image motivation; higher-order evidence; forgery; Competence Theorem; sourcehood; social credit; AI alignment Highlights · Levy’s defence of virtue signalling rests on the claim that cost makes a signal credible; signalling theory says honesty comes from cost that differs between types, not from cost as such. · Reward Ceiling: a moral act of equal cost to all separates the committed from the uncommitted only while the audience’s reward stays below the cost; above it, everyone signals and the signal says nothing. · For any log-concave cost distribution, the probability that a signaller is committed falls steadily as the reward rises, while the act’s form and cost stay unchanged. · With the Competence Theorem: where audiences trust cost and check less as cost rises, the most expensive forged signals do the most damage. · Source is judged neither by cost nor by motive but by persistence when the audience’s reward is removed and when it is reversed. Relevance The question whether moral display can be trusted is no longer only a question about manners on social media. A social credit system pays citizens, in access and standing, for recorded displays of good conduct. Platforms pay in visibility, and employers in reputation. AI systems trained on human approval are paid, in the gradient, for the display of care. In each case the reward for the form of a virtue is set by design, and it is often set high. Levy’s argument implies that this is harmless as long as the displays remain costly. This paper shows why it is not. Once the reward for a display exceeds its cost, the display certifies nothing about the person, or the system, behind it. The better the display is performed and the more it costs, the more damage it does when the commitment is absent. The protection of freedom and of the virtues therefore depends on keeping open places where good acts are not paid for: the conditions under which an act can show that it is one’s own.
Authors
- José Caetano de Mattos (ORCID: https://orcid.org/0009-0002-1567-6567)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.5281/zenodo.24176934
- Primary Topic
- Philosophical Ethics and Theory
- Type
- preprint