Reference Sheet 8: Social Choice Theory for Collective Decisions#

Target audience: Forge auditor who knows S5 modal logic, CEM, FOL, basic game theory, mechanism design (Sheet 3), and dynamical systems (Sheet 5) but needs social choice theory to understand impossibility constraints on collective decision-making in the Jubilee-System.

1. Orientation#

Mechanism design (Sheet 3) asks: “can we design a game that achieves outcome X?” Social choice theory asks a prior question: “can any aggregation procedure achieve outcome X, regardless of implementation?” The answers are often “no” — and the impossibility results are among the most robust in mathematical economics. The JUB model’s ax17 (Non-Coercive Guidance) forbids dictatorship, ax25 (Jubilee-System Recalibration) requires collective redistribution decisions, and th9 (Social Ergodicity) requires long-run fairness. Arrow’s theorem says these three goals generically conflict. Social choice theory maps the impossibility landscape — telling the forge auditor where the hard walls are, so that model development does not waste effort trying to walk through them.

2. Key Concepts#

Social welfare function (SWF). A function F that maps a profile of individual preference orderings (≻₁, …, ≻ₙ) to a single social preference ordering ≻_F. This is the “aggregation rule” — how individual preferences become collective decisions. Matheology use: If the Jubilee-System requires collective decisions about redistribution parameters (cycle length, redistribution fraction, exemptions), an SWF determines how individual preferences over these parameters aggregate into a collective choice.

Pareto criterion (unanimity). If every individual prefers x to y (x ≻ᵢ y for all i), then the social ordering must also prefer x to y (x ≻_F y). This is the weakest reasonable condition: unanimous agreement is respected. Matheology use: If all agents in the economy prefer a specific Jubilee-System configuration over the status quo, the SWF should reflect this. Pareto is uncontroversial — the binding constraints come from combining it with other conditions.

Independence of Irrelevant Alternatives (IIA). The social ranking of x vs. y depends only on individual rankings of x vs. y — not on how individuals rank other alternatives z. Adding or removing options should not change the relative ranking of existing options. Matheology use: When choosing between two Jubilee-System designs, the choice should not depend on whether a third (irrelevant) design is also on the table. IIA prevents “spoiler” effects.

Non-dictatorship. There is no individual i such that ≻_F = ≻ᵢ for all possible profiles. No single agent’s preference determines the social outcome regardless of what others prefer. Matheology use: ax17 (Non-Coercive Guidance) forbids coercion, which includes dictatorship. The non-dictatorship condition is a formal expression of ax17 applied to collective economic decisions.

Single-peaked preferences. A domain restriction where all individuals’ preferences have a single “peak” along a common ordering of alternatives. If alternatives can be arranged on a line (e.g., redistribution rate from 0% to 100%) and each agent has a most-preferred point with preferences declining on both sides, preferences are single-peaked. Matheology use: Single-peakedness is the escape hatch from Arrow’s impossibility. If agents’ preferences over Jubilee-System parameters are single-peaked (each agent has an ideal redistribution rate, with less preferred rates on either side), then majority rule produces a consistent social ordering — and Arrow’s impossibility does not apply.

Condorcet winner. An alternative x that beats every other alternative in pairwise majority voting. A Condorcet winner may not exist (the Condorcet paradox: A beats B, B beats C, C beats A in cyclic majority). Matheology use: If a Condorcet winner exists among Jubilee-System configurations, it is the natural focal point for collective choice. If no Condorcet winner exists, the collective decision is inherently unstable — different procedures select different outcomes, and cycling is possible.

Manipulability (strategic voting). A social choice function is manipulable if some agent can sometimes obtain a better outcome by misreporting their preferences. A strategy-proof function is one where honest reporting is always optimal (connecting to incentive compatibility in Sheet 3). Matheology use: If agents vote on Jubilee-System parameters, can they game the system by voting dishonestly? Gibbard-Satterthwaite says: for three or more options, either the system is dictatorial or it is manipulable. The only way out is domain restriction (single-peaked) or accepting some manipulation.

Liberalism (minimal liberty). Sen’s condition: at least two individuals are each “decisive” over at least one pair of alternatives — each person gets to determine the social ranking of at least one pair based solely on their own preference. Matheology use: In the matheology system, ax15 (Human Genuine Agency) and ax17 (Non-Coercive Guidance) imply that individuals should have some domain of genuine choice — a form of minimal liberalism. Sen’s paradox shows this conflicts with Pareto efficiency.

3. Critical Theorems#

Arrow’s Impossibility Theorem (1951). For three or more alternatives, no SWF simultaneously satisfies: (1) unrestricted domain (works for all possible preference profiles), (2) Pareto criterion, (3) IIA, and (4) non-dictatorship. At least one condition must be violated. Why it matters: If the Jubilee-System requires collective choices among three or more configurations, Arrow says: accept a dictator (violating ax17), restrict the preference domain (e.g., impose single-peaked preferences), accept IIA violations (outcomes depend on “irrelevant” alternatives), or accept that some unanimous preferences are overridden. These are the only options. The JUB model must specify which constraint it relaxes.

Gibbard-Satterthwaite Theorem (1973/1975). For three or more outcomes, any social choice function that is surjective (every outcome is possible) and strategy-proof (honest reporting is optimal) is dictatorial. Equivalently: any non-dictatorial surjective social choice function is manipulable. Why it matters: This is the mechanism design companion to Arrow. Where Arrow constrains aggregation rules, Gibbard-Satterthwaite constrains choice procedures. For the Jubilee-System: if the redistribution parameters are chosen by a non-dictatorial procedure, some agents can sometimes benefit by strategic misreporting. The system must either accept this or restrict the domain.

Sen’s Liberal Paradox (1970). No social decision function can simultaneously satisfy: (1) unrestricted domain, (2) Pareto criterion, and (3) minimal liberalism (at least two individuals are each decisive over at least one pair). Informally: individual rights and collective efficiency conflict. Why it matters: The JUB model asserts both individual agency (ax15) and collective optimization (life-trifecta, ax24). Sen says these generically conflict. Resolution strategies include: rights as constraints rather than preferences (Nozick), rights as meta-preferences (Sen’s later work), or sequential application (rights first, then Pareto within the rights-constrained domain). The JUB model should specify which resolution it adopts.

Black’s Median Voter Theorem (1948). With single-peaked preferences and an odd number of voters, majority rule selects the median voter’s peak as the Condorcet winner. This is the positive result that rescues Arrow in restricted domains. Why it matters: If Jubilee-System parameter preferences are single-peaked (each agent has an ideal redistribution rate), then the median voter’s ideal rate is the stable collective choice. This provides a concrete prediction: the Jubilee-System’s redistribution rate will converge to the median preference. Whether this median is “correct” (in the life-trifecta sense) is a separate question — the median voter theorem tells you what the collective will choose, not what is optimal.

May’s Theorem (1952). For exactly two alternatives, majority rule is the unique social choice function satisfying anonymity (all voters are treated equally), neutrality (both alternatives are treated equally), and positive responsiveness (if the outcome was a tie and one voter changes to support A, then A wins). Why it matters: For binary choices (e.g., “implement ax25 or not”), majority rule is not just one option — it is the only option that treats people and alternatives symmetrically. This provides a strong justification for democratic decision-making on binary Jubilee-System questions.

4. Common Pitfalls#

Treating Arrow’s theorem as a defect of democracy. Arrow’s theorem applies to all aggregation procedures, not just voting. Markets, committees, algorithms, and AI systems that aggregate preferences all face the same impossibility. The theorem is about the structure of aggregation, not the failure of any specific institution.

Assuming impossibility means impasse. Arrow says no perfect SWF exists. Many good SWFs exist. Majority rule with single-peaked preferences, approval voting, ranked-choice voting, and score voting all have desirable properties — just not all desirable properties simultaneously. The productive question is: which properties does the Jubilee-System prioritize?

Ignoring domain restrictions. Most impossibility results assume unrestricted domain — any logically possible preference profile can occur. In practice, economic preferences are structured: agents prefer more wealth to less, stability to chaos, and functioning institutions to dysfunctional ones. These restrictions often rescue positive results (Black’s theorem). The JUB model should identify and justify its domain restrictions.

Conflating social welfare functions with social choice functions. An SWF produces a full ranking of alternatives. A social choice function (SCF) selects a single winner. Arrow applies to SWFs; Gibbard-Satterthwaite applies to SCFs. The two theorems are related but not equivalent. Know which one applies to your context.

Forgetting cardinal utility. Arrow’s theorem assumes ordinal preferences (rankings without intensity). If cardinal utility is available (how much an agent prefers x to y), then utilitarianism (sum of utilities) satisfies Pareto, IIA, and non-dictatorship — but requires interpersonal utility comparisons, which are philosophically contentious. The JUB model’s ax22 (Divine Preference for Genuine Love) may provide a framework for cardinal valuation that escapes the ordinal trap.

5. Bridge to Matheology#

Arrow’s theorem meets ax17 + ax25 + th9. The triad:

  • ax17 (Non-Coercive Guidance) → non-dictatorship

  • ax25 (Jubilee-System Recalibration) → collective choice among redistribution configurations (three or more alternatives)

  • th9 (Social Ergodicity) → the chosen configuration must achieve ergodic dynamics

Arrow says: you cannot always select the ergodic configuration by a non-dictatorial aggregation of unrestricted preferences. The JUB model must resolve this by:

  1. Domain restriction: Argue that preferences over Jubilee-System parameters are naturally single-peaked (Black’s theorem applies, median voter selects).

  2. Divine guidance as information, not coercion: ax17 allows guidance. If divine guidance informs preferences (shifting peaks toward the ergodic configuration) without overriding them, then the median voter’s peak may coincide with the ergodic optimum — not by dictatorship but by informed consensus.

  3. Reduced alternative set: If the choice is binary (“implement Jubilee-System or not”), May’s theorem applies and majority rule suffices without impossibility.

Sen’s paradox meets ax15 + ax24. ax15 (Human Genuine Agency) implies minimal liberalism — each agent has a domain of genuine choice. ax24 (Life-Trifecta) implies collective optimality — the outcome must be Stable ∧ Extensible ∧ LifeFriendly. Sen says these conflict. The JUB model’s resolution likely involves the three-domain partition (Dₓ, D_free, D_inno from the JUB axioms):

  • Dₓ (forced domain): no liberal rights (physics, not choice)

  • D_free: liberal rights apply — agents are decisive over personal choices

  • D_inno ⊆ D_free: innovation domain where Pareto optimization operates within the rights-constrained feasible set

This sequential application (rights first in D_free, then Pareto in D_inno) is a known resolution strategy for Sen’s paradox.

Mechanism design (Sheet 3) + social choice as complementary lenses. Sheet 3 asks: “given that agents are strategic, can we design a game with good equilibria?” This sheet asks: “given that we must aggregate preferences, do good aggregation rules exist?” The intersection:

  • Gibbard-Satterthwaite (this sheet) says non-dictatorial choice functions are manipulable.

  • The revelation principle (Sheet 3) says manipulable mechanisms can be converted to incentive-compatible direct mechanisms.

  • Together: there exists an IC mechanism (Sheet 3) but it requires accepting either dictatorship or restricted domain (this sheet). The IC mechanism’s existence does not contradict Arrow — it operates within Arrow’s constraints.

HELL findings and impossibility identification. Several HELL con-findings may be sharpened to impossibility-theorem form:

  • “ax25 cannot simultaneously preserve individual freedom and achieve collective optimality” → Sen’s paradox applied to the Jubilee-System context.

  • “No collective procedure can select the ergodic redistribution rate without strategic manipulation” → Gibbard-Satterthwaite applied.

Pro-findings should respond by identifying the domain restriction, information structure, or mechanism design that circumvents the specific impossibility. A generic “the system works” response is insufficient against a specific impossibility theorem.

New questions social choice theory enables:

  • Are preferences over Jubilee-System parameters naturally single-peaked? If so, Black’s theorem rescues collective choice from Arrow’s impossibility.

  • Does ax17’s “guidance” constitute information provision (shifting preferences), preference restriction (reducing the domain), or something else? Each has different social-choice-theoretic consequences.

  • Can the Jubilee-System’s redistribution rule be decomposed into binary choices (May’s theorem applies) rather than multi-alternative choices (Arrow’s theorem applies)?

  • Is the JUB model’s resolution of Sen’s paradox via the three-domain partition (Dₓ / D_free / D_inno) formally valid — does sequential application of liberal rights then Pareto optimization produce a consistent social choice function?