RECIPROCITY — why it works, how to run it, what it produces
Reciprocity = the test of symmetry in costs, benefits, and risks across parties, in relation to their demonstrated interests, with compensation/warranty where symmetry cannot be achieved.
Put simply: “Do you impose on others what you would not accept yourself, without compensation?”
A claim passes reciprocity when:
-
Parties and their demonstrated interests are enumerated.
-
Transfers of benefits/costs/risks are mapped (who gains, who pays, who is exposed).
-
Symmetry tests are run (would each accept the same treatment under reversal of roles?).
-
Externalities are exposed and compensated (insurance, restitution, bonding).
-
Information asymmetries are disclosed or warranted (no hidden rent-seeking).
If these conditions hold, cooperation is mutually admissible.
-
All cooperation is exchange under uncertainty.
-
Predation and parasitism arise when one party externalizes costs, conceals risks, or exploits asymmetry.
-
By forcing symmetry disclosure and compensation, reciprocity collapses the space of irreciprocal strategies, leaving only cooperative equilibria (or boycott if compensation is refused).
-
This converts “ought” into a computable test: if symmetry cannot be established, the claim/action is inadmissible.
-
Represent parties and interests as nodes in a graph.
-
Represent transfers as directed edges with annotations (benefit, cost, risk).
-
Run symmetry checks: if we invert the graph (swap roles), do transfers remain acceptable?
-
Detect externalities (unlabeled costs landing on commons) and propose compensation terms.
-
Flag informational asymmetries (one side holds hidden knowledge).
This is graph-constraint checking + counterfactual swapping — something language models can execute symbolically, with structured prompting.
-
Hidden externalities (future harms, commons degradation) → require prospective disclosure (“list foreseeable externalities”), then bind with warranties/insurance.
-
Moral hazard (actor insulated from risk) → require bonding/escrow.
-
Asymmetric information (seller knows quality, buyer doesn’t) → require disclosure or guarantee.
Decision rule:
-
If symmetry fails and no compensation is possible → Inadmissible: Irreciprocal.
-
If symmetry holds or is cured by compensation → Admissible (proceed to Decidability).
-
If parties/interests are incomplete → Undecidable: Missing Mapping.
Claim: “Impose congestion pricing on downtown drivers.”
-
Parties: City, Drivers, Residents, Businesses.
-
Demonstrated interests:
City: reduced traffic, cleaner air.
Drivers: time savings, mobility.
Residents: health, quiet.
Businesses: customer access. -
Transfers:
Cost: fee from Drivers → City.
Benefit: reduced traffic → Residents & Businesses.
Risk: economic displacement → Businesses. -
Symmetry test: If Residents had to pay drivers for clean air instead of the reverse, would that be acceptable? Yes, in principle.
-
Externalities: Risk of small business harm; addressed by fee exemptions or subsidies.
-
Compensation plan: Revenue earmarked to improve public transit (compensation to drivers) and support affected businesses.
-
Verdict: Admissible with compensation. Without compensation, irreciprocal (drivers subsidize residents unfairly).
-
Truth made the claim testifiable (what congestion pricing is, what it entails).
-
Reciprocity maps interests and audits symmetry.
-
Once irreciprocity is exposed and cured, we now have a feasible set of cooperative actions.
-
That feasible set is the input to Decidability: we can resolve the case without discretion, because the asymmetries have been normalized.
RECIPROCITY_CERT
– Parties: …
– Interests: …
– Transfers: table
– Symmetry audit: pass/fail, externalities, info asymmetries
– Compensation plan: list remedies
– Verdict: Admissible / Inadmissible / Undecidable
– Parties: …
– Interests: …
– Transfers: table
– Symmetry audit: pass/fail, externalities, info asymmetries
– Compensation plan: list remedies
– Verdict: Admissible / Inadmissible / Undecidable
Source date (UTC): 2025-08-24 03:21:33 UTC
Original post: https://x.com/i/articles/1959456016028033290
Leave a Reply