TL;DR
This paper tackles the fundamental question of how a group of agents can establish a self-governing community by agreeing on a decision-making rule (a constitution) and a mechanism to amend it. They propose a set of intuitive axioms that logically derive a simple majority rule as the initial constitution and a consistent amendment process, offering a formal framework for bootstrapping democracy.
What Problem Does This Paper Address?
Imagine a group of individuals wanting to form a community where they collectively make decisions. The immediate challenge is agreeing on how to make those decisions – what voting rule or decision-making process should they adopt? This leads to an infinite regress problem: to decide on the decision rule, they'd need another decision rule, and so on. This paper addresses this by proposing a way to derive an initial "constitution" (a decision rule) from a set of fundamental, agreed-upon principles (axioms). Crucially, this constitution must also specify how it can be changed or amended, ensuring the community can adapt over time. The authors aim to provide a formal, logical foundation for this process, moving beyond simply assuming a pre-existing rule.
Methods
The paper employs an axiomatic approach. They start with a set of "intuitive and self-evident axioms" that describe desirable properties of a decision-making process. These axioms are then used to logically infer the form of the decision rule.
Key axioms for the initial decision rule include:
- Decisiveness: Every proposal must be either accepted or rejected.
- Monotonicity: If a proposal is accepted, it remains accepted if more agents approve it.
- Anonymity: The decision doesn't depend on who approves, only how many.
- Concordance: If two proposals are incompatible, they cannot both be accepted.
- Minimality: Among rules satisfying the above, choose the one requiring the fewest approvals.
These axioms, particularly Minimality, lead to the conclusion that the decision rule must be a "d-supermajority approval rule" (accept if at least a fraction 'd' of agents approve). The initial rule, derived from these axioms, turns out to be simple majority (d=1/2).
For amending the constitution, additional axioms are introduced, focusing on:
- Posterior Consistency: An amendment is accepted only if it's also accepted under the new rule. This acts as an "anti-hypocrisy" measure.
- Condorcet Consistency / Conservatism: These axioms guide the choice among permissible amendments, either by selecting the "most-preferred" rule (Condorcet) or the "closest" rule to the current one (Conservatism).
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Bootstrapping Democracy from First Principles: The core idea is that a functional decision-making system (a constitution) can be logically derived from a minimal set of universally agreeable principles, rather than being imposed or chosen arbitrarily.
- The Power of Minimality: The "Minimality" axiom is crucial. It acts as a tie-breaker, forcing the selection of the simplest rule that satisfies the other criteria, which in this case leads to simple majority.
- Self-Updating Constitutions: The paper doesn't just define an initial rule but also a mechanism for changing that rule, creating a dynamic, self-updating constitution. This is a significant step beyond static constitutional design.
- Formalizing "Fairness" and "Adaptability": The axioms for amendment (Posterior Consistency, Condorcet Consistency/Conservatism) provide a formal way to capture notions of fairness and adaptability in constitutional change.
- Two Paths to Amendment: The paper presents two distinct but related amendment mechanisms (Condorcet and Conservative), showing that while they differ in their immediate application, they can lead to similar outcomes, especially when starting from simple majority.
Key Results
- Axioms 1-5 logically imply that the decision rule must be a d-supermajority approval rule, with the initial rule being simple majority (d=1/2).
- Axioms 1-7 (or 1-7a for conservatism) combined with assumptions on agent preferences (single-peaked preferences over decision rules) lead to a unique amendment process.
- The Condorcet Amendment rule dictates increasing 'd' if a higher supermajority prefers it and it's maximal, decreasing 'd' if a lower supermajority prefers it and it's minimal, and retaining the current rule otherwise.
- The Conservative Amendment rule prioritizes closeness to the current rule, with similar conditions for increasing or decreasing 'd'.
- Iterative application of the Conservative Amendment rule, starting from simple majority, converges to the same outcome as the Condorcet Amendment rule.
Limitations and Future Work
- Fixed Agent Set: A key assumption is that the set of agents is fixed. The authors acknowledge that a real self-governed community must also decide how to determine membership and manage additions/removals.
- Agent Preferences: The paper assumes "single-peaked" preferences over decision rules. Exploring other preference structures (e.g., metric preferences) is suggested as future work.
- Alternative Axioms: The authors mention that other sets of axioms could lead to different decision rules and amendment processes, inviting further exploration.
- Practical Implementation: While the framework is formal, its direct application to real-world political systems would require careful consideration of how to elicit preferences and implement the derived rules.
Summary updated