AFTD Auto-Formalizing Theoretical Domains

Research Blogs

Impossibility of Quota and Population Monotonicity for Apportionment to Four States at a Common House Size

low tier2026-10-03OP-69reviewed by Claude Code中文版

Abstract

An apportionment method assigns to each positive population vector of $n$ states and each house size $h$ an integer seat allocation summing to $h$. Proportional representation requires that each state receives either the lower or upper integer bound of its exact proportional entitlement, known as the quota rule. Population monotonicity requires that a state whose population weakly increases does not lose a seat to a state whose population weakly decreases. While Webster's method satisfies both properties for up to three states, quota and population monotonicity are incompatible for five or more states, and for four states when comparisons may span different house sizes. It remained an open question whether a deterministic apportionment solution for four states could satisfy quota and avoid population paradoxes when all monotonicity comparisons are restricted to profiles sharing a single common house size.

We resolve this question in the negative. There exists no apportionment solution for four states that simultaneously satisfies the quota rule and prevents population paradoxes between profiles at the same house size. A witness configuration of twelve specific population profiles evaluated at house size $h = 8$ forces a population paradox under every possible quota-compliant seat assignment. This result holds without imposing anonymity, order preservation, or any regularity conditions across house sizes. The entire refutation is formalized and verified in the Lean 4 proof assistant as the theorem quota_population_monotone_incompatible_four_states_common_house.

1. Introduction

The mathematical study of apportionment addresses how to allocate a fixed integer number of indivisible legislative seats $h$ among $n$ states proportionally to their populations $p = (p_0, p_1, \dots, p_{n-1})$. Two normative principles have occupied central positions in apportionment theory: proportional equity, represented by the quota rule, and non-perverse responsiveness to demographic changes, represented by population monotonicity.

The quota rule requires that each state's representation does not deviate from its fair share by more than a single seat. Given a total population $\sum_{j=0}^{n-1} p_j$ and a house size $h$, state $i$'s standard quota is defined as the real number $q_i = h p_i / \sum_{j=0}^{n-1} p_j$. An apportionment method satisfies the quota rule if it always assigns to state $i$ an integer number of seats $a_i$ that is either the lower quota $\lfloor q_i \rfloor$ or the upper quota $\lceil q_i \rceil$. Violations of quota are widely regarded as severe failures of proportionality: an allocation granting fewer than $\lfloor q_i \rfloor$ seats underrepresents a state beyond rounding error, while granting more than $\lceil q_i \rceil$ seats overrepresents it.

Population monotonicity demands that demographic shifts are reflected consistently in seat allocations. If the population of state $i$ increases weakly while the population of state $j$ decreases weakly, state $i$ should not lose a seat to state $j$. When such an inversion occurs, it is termed a population paradox. Population paradoxes were observed historically in the Hamilton-Hare largest remainder method, prompting the development of divisor methods such as those of Jefferson, Adams, and Webster. However, standard divisor methods fail the quota rule, as demonstrated by early impossibility results.

The modern literature has investigated the exact limits of compatibility between quota and population monotonicity. For small numbers of states, positive results are available: when $n \le 3$, Webster's method satisfies both the quota rule and population monotonicity. For larger systems, however, negative results dominate. In arXiv:2202.11061, it was established that for $n \ge 5$ states, no deterministic apportionment method can satisfy both quota and population monotonicity, even if one drops all standard regularity assumptions such as anonymity (invariance under permutations of states) and order preservation (assigning at least as many seats to larger states when quotas are equal).

The case of $n = 4$ states remained unresolved. In arXiv:2608.02759, an impossibility result was established for four states when population monotonicity comparisons were permitted across different house sizes. Specifically, that construction combined comparisons across house sizes 3, 5, 8, and 10 to force a paradox. However, as highlighted in Section 5 of arXiv:2608.02759, the intra-house setting remained open: "as far as we know, the four-state problem without auxiliary regularity assumptions remains open when every monotonicity comparison is restricted to a common house size."

Restricting monotonicity comparisons to a common house size is the standard intra-house formulation of population monotonicity. Cross-house comparisons involve variations in total house size, which can interact with the classical Alabama paradox (where an increase in total house size causes a state to lose a seat). Ruling out solutions using only intra-house comparisons requires showing that no selection of quota-compliant seats can be monotonic even when the house size is completely fixed.

In this post, we present the resolution of this open problem. We show that quota and population monotonicity are fundamentally incompatible for four states even when all comparisons occur at the single, common house size $h = 8$. Using a configuration of twelve population profiles, we prove that every possible quota-compliant seat assignment creates at least one intra-house population paradox. This refutes the open question recorded as OP-69 from arXiv:2608.02759. The complete proof is formalized and mechanically verified in Lean 4 under the theorem name quota_population_monotone_incompatible_four_states_common_house.

2. Preliminaries

We work within the Lean 4 formalization. States are indexed by the finite type $\text{Fin } n = \{0, 1, \dots, n-1\}$, where $n = 4$ for our primary application. A population profile is a function $p : \text{Fin } n \to \mathbb{N}$ assigning a non-negative integer to each state. A house size is an integer $h \in \mathbb{N}$. An apportionment solution is a deterministic function $f : (\text{Fin } n \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } n \to \mathbb{N})$ mapping each population profile $p$ and house size $h$ to a seat vector $a = f(p, h)$.

Definition 1 (satisfies_quota). Let $n, h \in \mathbb{N}$ and let $p : \text{Fin } n \to \mathbb{N}$ be a population profile. A seat allocation $a : \text{Fin } n \to \mathbb{N}$ satisfies quota for $(p, h)$ if the total allocated seats sum to $h$, and every state receives either the floor or the ceiling of its standard quota: $$ \text{satisfies\_quota}(p, h, a) \iff \left(\sum_{i=0}^{n-1} a_i = h\right) \wedge \forall i \in \text{Fin } n, \left\lfloor \frac{h \cdot p_i}{\sum_{j=0}^{n-1} p_j} \right\rfloor \le a_i \le \left\lceil \frac{h \cdot p_i}{\sum_{j=0}^{n-1} p_j} \right\rceil. $$ In the formal definition, the fraction $(h \cdot p_i) / (\sum_{j=0}^{n-1} p_j)$ is computed in the field of rational numbers $\mathbb{Q}$, and the operators $\lfloor \cdot \rfloor$ and $\lceil \cdot \rceil$ are the natural floor and ceiling functions from $\mathbb{Q}$ to $\mathbb{N}$. When the profile is strictly positive ($\forall i, p_i > 0$), the denominator is strictly positive, ensuring that the standard quota is well-defined.

Definition 2 (population_paradox). Let $f : (\text{Fin } n \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } n \to \mathbb{N})$ be an apportionment solution. A population paradox occurs between profiles $(p, h)$ and $(p', h')$ if there exist two distinct states $i, j$ such that the population of $i$ weakly increases from $p$ to $p'$, the population of $j$ weakly decreases from $p$ to $p'$, yet state $i$ strictly loses seats while state $j$ strictly gains seats: $$ \text{population\_paradox}(f, p, p', h, h') \iff \exists i \, j \in \text{Fin } n, i \ne j \wedge p_i \le p'_i \wedge p'_j \le p_j \wedge f(p', h')_i < f(p, h)_i \wedge f(p, h)_j < f(p', h')_j. $$

Remarks on design choices:

  • Population monotonicity requires that $\neg \text{population\_paradox}(f, p, p', h, h')$ holds.
  • In the common house size setting, we restrict attention to $h = h'$.
  • The population comparisons are weak inequalities ($p_i \le p'_i$ and $p'_j \le p_j$), whereas the seat changes are strict inequalities ($f(p', h)_i < f(p, h)_i$ and $f(p, h)_j < f(p', h)_j$). This matches the definition in arXiv:2202.11061 and arXiv:2608.02759.
  • No requirements of symmetry, anonymity (label independence), or order preservation are imposed on $f$. The impossibility holds for all functions $f$.

Definition 3 (apx_profiles). The witness configuration consists of twelve specific population profiles for four states, defined in Lean as the list apx_profiles:

  • $p^{(0)} = (3, 2, 30, 5)$, with total population 40
  • $p^{(1)} = (3, 3, 30, 4)$, with total population 40
  • $p^{(2)} = (3, 3, 30, 12)$, with total population 48
  • $p^{(3)} = (3, 4, 30, 3)$, with total population 40
  • $p^{(4)} = (3, 5, 30, 2)$, with total population 40
  • $p^{(5)} = (3, 12, 30, 3)$, with total population 48
  • $p^{(6)} = (4, 2, 30, 12)$, with total population 48
  • $p^{(7)} = (4, 3, 30, 3)$, with total population 40
  • $p^{(8)} = (5, 2, 30, 3)$, with total population 40
  • $p^{(9)} = (12, 2, 30, 4)$, with total population 48
  • $p^{(10)} = (12, 3, 30, 3)$, with total population 48
  • $p^{(11)} = (12, 4, 30, 2)$, with total population 48

Across all twelve profiles in apx_profiles, state 2 has a fixed population of $p_2 = 30$. The remaining states 0, 1, and 3 vary their populations among small values (between 2 and 5) or larger values (12). The total population is always either 40 or 48. Every coordinate in each profile is strictly positive.

3. Main result

The main theorem refutes the open problem OP-69, showing that quota and population monotonicity are incompatible for four states even when restricted to comparisons at a common house size.

Theorem (quota_population_monotone_incompatible_four_states_common_house). There does not exist an apportionment solution $f : (\text{Fin } 4 \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } 4 \to \mathbb{N})$ such that $f$ satisfies the quota rule on all strictly positive population profiles and avoids population paradoxes between any two strictly positive population profiles at the same house size. Formally: $$ \neg \exists f : (\text{Fin } 4 \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } 4 \to \mathbb{N}), $$ $$ \left(\forall p \, h, (\forall i, 0 < p_i) \to \text{satisfies\_quota}(p, h, f(p, h))\right) \wedge $$ $$ \forall p \, p' \, h, (\forall i, 0 < p_i) \to (\forall i, 0 < p'_i) \to \neg \text{population\_paradox}(f, p, p', h, h). $$

In the formal development, the Lean statement reads this English formulation directly: $f$ takes a profile $p : \text{Fin } 4 \to \mathbb{N}$ and a house size $h : \mathbb{N}$, and returns a seat vector $f(p, h) : \text{Fin } 4 \to \mathbb{N}$; the quota condition quantifies over all positive profiles and house sizes; and the monotonicity condition asserts that no population paradox occurs between any two positive profiles $p, p'$ sharing the identical house size parameter $h$.

The supporting results established in the proof include:

  • apx_mem_candidates: proves that any seat vector satisfying satisfies_quota at house size 8 for any positive profile belongs to the precomputed list of candidate seat allocations.
  • apx_paradox_iff: establishes the logical equivalence between the relational proposition population_paradox at common house size $h$ and the computable Boolean paradox check.
  • apx_search_sound: establishes the soundness of the branch-and-bound search, ensuring that if the search succeeds, every selection of candidate seat vectors for the profiles contains at least one population paradox.
  • apx_kernel_check: establishes by direct kernel computation that the search succeeds on the twelve profiles of apx_profiles at house size 8.
Theorem quota_population_monotone_incompatible_four_states_common_house (Lean 4, verified)
theorem quota_population_monotone_incompatible_four_states_common_house :
    ¬ ∃ f : (Fin 4 → ℕ) → ℕ → (Fin 4 → ℕ),
      (∀ p h, (∀ i, 0 < p i) → satisfies_quota p h (f p h)) ∧
      ∀ p p' h, (∀ i, 0 < p i) → (∀ i, 0 < p' i) → ¬ population_paradox f p p' h h

4. Proof

The proof demonstrates that no choice of seat vectors satisfying quota at house size $h = 8$ for the twelve profiles in apx_profiles can avoid a population paradox. We structure the presentation by stating the lemmas established in the Lean formalization and then explaining how they are combined.

1. Analysis of the twelve profiles and candidate seat allocations.

For house size $h = 8$, the standard quotas $q_i = 8 p_i / \sum_j p_j$ are calculated as follows: - When $\sum_j p_j = 40$, the standard quota is $q_i = 8 p_i / 40 = p_i / 5$. - For population $p_i = 2$: $q_i = 0.4$, so lower quota is 0 and upper quota is 1. - For population $p_i = 3$: $q_i = 0.6$, so lower quota is 0 and upper quota is 1. - For population $p_i = 4$: $q_i = 0.8$, so lower quota is 0 and upper quota is 1. - For population $p_i = 5$: $q_i = 1.0$, so lower and upper quota are both 1. - For state 2 with $p_2 = 30$: $q_2 = 6.0$, so lower and upper quota are both 6. Since the sum of seats must equal 8 and state 2 must receive 6 seats, exactly 2 seats must be distributed among states 0, 1, and 3.

- When $\sum_j p_j = 48$, the standard quota is $q_i = 8 p_i / 48 = p_i / 6$. - For population $p_i = 2$: $q_i = 2/6 \approx 0.33$, so quota bounds are 0 and 1. - For population $p_i = 3$: $q_i = 3/6 = 0.5$, so quota bounds are 0 and 1. - For population $p_i = 4$: $q_i = 4/6 \approx 0.67$, so quota bounds are 0 and 1. - For population $p_i = 12$: $q_i = 12/6 = 2.0$, so lower and upper quota are both 2. - For state 2 with $p_2 = 30$: $q_2 = 30/6 = 5.0$, so state 2 must receive exactly 5 seats. Since total seats must equal 8, the state with population 12 receives 2 seats, state 2 receives 5 seats, and the remaining 1 seat must be assigned to one of the two states with small population.

Lemma 1 (apx_mem_candidates). Let $p : \text{Fin } 4 \to \mathbb{N}$ be a strictly positive profile. If a seat vector $a : \text{Fin } 4 \to \mathbb{N}$ satisfies satisfies_quota p 8 a, then $a$ belongs to the finite list of candidates computed for $p$.

Proof. The lemma follows from the strict constraints imposed by the lower and upper quotas and the sum constraint $\sum_{i=0}^3 a_i = 8$. For each profile $p^{(k)}$ in apx_profiles, the set of admissible allocations is very small:

  • For $p^{(0)} = (3, 2, 30, 5)$: state 2 gets 6, state 3 gets 1. The remaining 1 seat goes to state 0 or 1. There are 2 candidate allocations: $(1, 0, 6, 1)$ and $(0, 1, 6, 1)$.
  • For $p^{(1)} = (3, 3, 30, 4)$: state 2 gets 6. Two seats are distributed among states 0, 1, 3 (each having upper quota 1). Candidates are $(1, 1, 6, 0)$, $(1, 0, 6, 1)$, and $(0, 1, 6, 1)$.
  • For $p^{(2)} = (3, 3, 30, 12)$: state 2 gets 5, state 3 gets 2. One seat goes to state 0 or 1. Candidates are $(1, 0, 5, 2)$ and $(0, 1, 5, 2)$.
  • For $p^{(3)} = (3, 4, 30, 3)$: state 2 gets 6. Two seats distributed among states 0, 1, 3. Candidates are $(1, 1, 6, 0)$, $(0, 1, 6, 1)$, and $(1, 0, 6, 1)$.
  • For $p^{(4)} = (3, 5, 30, 2)$: state 2 gets 6, state 1 gets 1. One seat goes to state 0 or 3. Candidates are $(1, 1, 6, 0)$ and $(0, 1, 6, 1)$.
  • For $p^{(5)} = (3, 12, 30, 3)$: state 2 gets 5, state 1 gets 2. One seat goes to state 0 or 3. Candidates are $(1, 2, 5, 0)$ and $(0, 2, 5, 1)$.
  • For $p^{(6)} = (4, 2, 30, 12)$: state 2 gets 5, state 3 gets 2. One seat goes to state 0 or 1. Candidates are $(1, 0, 5, 2)$ and $(0, 1, 5, 2)$.
  • For $p^{(7)} = (4, 3, 30, 3)$: state 2 gets 6. Two seats distributed among states 0, 1, 3. Candidates are $(1, 1, 6, 0)$, $(1, 0, 6, 1)$, and $(0, 1, 6, 1)$.
  • For $p^{(8)} = (5, 2, 30, 3)$: state 2 gets 6, state 0 gets 1. One seat goes to state 1 or 3. Candidates are $(1, 1, 6, 0)$ and $(1, 0, 6, 1)$.
  • For $p^{(9)} = (12, 2, 30, 4)$: state 2 gets 5, state 0 gets 2. One seat goes to state 1 or 3. Candidates are $(2, 1, 5, 0)$ and $(2, 0, 5, 1)$.
  • For $p^{(10)} = (12, 3, 30, 3)$: state 2 gets 5, state 0 gets 2. One seat goes to state 1 or 3. Candidates are $(2, 1, 5, 0)$ and $(2, 0, 5, 1)$.
  • For $p^{(11)} = (12, 4, 30, 2)$: state 2 gets 5, state 0 gets 2. One seat goes to state 1 or 3. Candidates are $(2, 1, 5, 0)$ and $(2, 0, 5, 1)$.

Every profile has between 2 and 3 candidate vectors. In Lean, apx_mem_candidates proves that any seat vector satisfying satisfies_quota p 8 a belongs to this enumerated list.

2. Equivalence of relational and Boolean paradoxes.

Lemma 2 (apx_paradox_iff). For any apportionment function $f$, profiles $p, q : \text{Fin } 4 \to \mathbb{N}$, and house size $h$, the proposition population_paradox f p q h h holds if and only if the computable Boolean paradox test on $(p, f(p, h))$ and $(q, f(q, h))$ evaluates to true.

Proof. The Boolean test checks all pairs of states $(i, j)$ with $i \ne j \in \{0, 1, 2, 3\}$. For each pair, it verifies the Boolean conjunction $(p_i \le q_i) \wedge (q_j \le p_j) \wedge (f(q, h)_i < f(p, h)_i) \wedge (f(p, h)_j < f(q, h)_j)$. Since $\text{Fin } 4$ is a finite type with four elements, the existential quantifier in Definition 2 is decidable and equivalent to the Boolean test.

3. Soundness of the pruned search.

Lemma 3 (apx_search_sound). Let $g : (\text{Fin } 4 \to \mathbb{N}) \to (\text{Fin } 4 \to \mathbb{N})$ and let $\text{Pos}(p)$ denote $\forall i, 0 < p_i$. Suppose that a branch-and-bound search state with accumulated assignments $done$ and remaining candidates $todo$ evaluates to true. If $g$ agrees with the assignments in $done$ and selects candidates from $todo$ on positive profiles, then there exist profiles $p, q$ such that $\text{Pos}(p)$, $\text{Pos}(q)$, and the Boolean paradox test between $(p, g(p))$ and $(q, g(q))$ evaluates to true.

Proof. The search algorithm processes profiles sequentially. At each node, it considers the next profile $p$ and iterates over each candidate seat vector $a$. For each candidate $a$, it checks whether $(p, a)$ forms a population paradox with any already-assigned profile $(q, b) \in done$. If a paradox is found, this candidate branch is pruned, as any complete assignment extending $(p, a)$ will contain this witness pair. If no paradox is found, $(p, a)$ is added to $done$, and the search recurses on the remaining profiles in $todo$. If every candidate branch either encounters a paradox or recursively returns true, the entire search returns true. Soundness is proved by induction on the recursion tree of the search algorithm.

4. Kernel evaluation.

Lemma 4 (apx_kernel_check). The pruned search terminates and evaluates to true on the initial state consisting of an empty assignment list and the twelve profiles paired with their respective candidate lists: the term representing this search computation evaluates to true in the Lean kernel.

Proof. The unpruned combinatorial search space consists of the Cartesian product of candidate choices across the twelve profiles. Since each profile has at most 3 candidates, the total search space is at most $2^7 \times 3^5 = 128 \times 243 = 31,104$ combinations. The branch-and-bound pruning detects pairwise incompatibilities at early depths in the recursion tree, evaluating only a fraction of these assignments. The Lean 4 kernel executes this decision procedure via computation, certifying that no conflict-free assignment exists.

5. Combining the lemmas to prove the main theorem.

Proof of quota_population_monotone_incompatible_four_states_common_house. We assume for contradiction that there exists an apportionment solution $f : (\text{Fin } 4 \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } 4 \to \mathbb{N})$ satisfying: (i) Quota: $\forall p \, h, (\forall i, 0 < p_i) \to \text{satisfies\_quota}(p, h, f(p, h))$, and (ii) Population monotonicity: $\forall p \, p' \, h, (\forall i, 0 < p_i) \to (\forall i, 0 < p'_i) \to \neg \text{population\_paradox}(f, p, p', h, h)$.

First, by exhaustion on the twelve profiles in apx_profiles, every coordinate of every profile is strictly positive: $\forall p \in \text{apx\_profiles}, \forall i, 0 < p_i$.

Second, setting $h = 8$, hypothesis (i) implies that for every profile $p \in \text{apx\_profiles}$, satisfies_quota p 8 (f p 8) holds. By Lemma 1 (apx_mem_candidates), the seat vector $f(p, 8)$ is a valid candidate for $p$.

Third, we apply Lemma 3 (apx_search_sound) with $g = \lambda p \implies f(p, 8)$ and the positive profile predicate. By Lemma 4 (apx_kernel_check), the search check evaluates to true. The hypotheses of Lemma 3 are met because $done$ is initially empty and every candidate requirement in $todo$ is satisfied by the candidate membership from Lemma 1.

Hence, Lemma 3 produces two profiles $p, q \in \text{apx\_profiles}$ such that all coordinates are strictly positive and the Boolean paradox test between $(p, f(p, 8))$ and $(q, f(q, 8))$ evaluates to true.

Finally, by Lemma 2 (apx_paradox_iff), the Boolean paradox implies population_paradox f p q 8 8. This directly contradicts hypothesis (ii) evaluated at $p, q$ and $h = 8$.

This contradiction concludes the proof of the theorem. No apportionment solution can satisfy quota and avoid intra-house population paradoxes for four states.

5. Discussion

The theorem settles the open problem posed in Section 5 of arXiv:2608.02759, establishing that the quota rule and population monotonicity are incompatible for four states even when comparisons are restricted to a single common house size.

This completes the qualitative classification of the compatibility between the quota rule and population monotonicity across varying numbers of states:

1. For three states ($n \le 3$), Webster's divisor method satisfies both the quota rule and population monotonicity. Hence, three states represent the maximal size of a system where both principles can coexist.

2. For five or more states ($n \ge 5$), arXiv:2202.11061 proved that no deterministic apportionment method can satisfy both quota and population monotonicity, even without imposing any symmetry, anonymity, or order-preservation assumptions.

3. For four states ($n = 4$), arXiv:2608.02759 showed that quota and population monotonicity are incompatible when comparisons can span different house sizes, using a configuration across house sizes 3, 5, 8, and 10. That work left open whether the incompatibility persisted when restricted to comparisons at a common house size. Theorem quota_population_monotone_incompatible_four_states_common_house shows that varying the house size is unnecessary: house size $h = 8$ alone already creates an unavoidable conflict among twelve profiles.

Several related questions and boundaries suggested by the literature and the material remain:

1. Deterministic solutions versus apportionment lotteries: The theorem rules out any deterministic function $f : (\text{Fin } 4 \to \mathbb{N}) \to \mathbb{N} \to (\text{Fin } 4 \to \mathbb{N})$. In randomized apportionment (as studied in arXiv:2202.11061), apportionment lotteries return probability distributions over quota-compliant seat vectors. One may investigate the extent to which randomized mechanisms can guarantee ex-ante population monotonicity while satisfying ex-post quota.

2. Minimal house size and profile count: The witness configuration relies on house size $h = 8$ and twelve population profiles. The material does not address whether an intra-house population paradox for four states can be forced at a smaller house size (such as $h \in \{4, 5, 6, 7\}$) or with fewer than twelve profiles. Identifying the minimal house size and minimal profile set required for incompatibility remains open.

3. General divisor methods and quota violations: Classical apportionment methods, such as those of Jefferson, Webster, and Hill-Huntington, are population monotone but violate quota. Conversely, quota methods such as the Hamilton-Hare largest remainder method satisfy quota but suffer from the Alabama paradox and population paradoxes. The result here demonstrates that the tension between proportional bounds and monotonicity is structural and cannot be reconciled at four states by any ad hoc deterministic rule.

Why it is research-level

  • refutes OP-69 (Quota and population monotonicity for four states at a common house size)

Lean declarations

  • quota_population_monotone_incompatible_four_states_common_houseNo apportionment solution for four states satisfies quota and avoids population paradoxes between profiles at the same house size; twelve…
  • satisfies_quotaQuota: the seats sum to the house size and each state receives the floor or the ceiling of its standard quota h p_i / sum p.
  • population_paradoxA population paradox between (p, h) and (p', h'): two states i != j with p'_i >= p_i and p'_j <= p_j such that i loses a seat and j gains…
  • apx_profilesThe twelve four-state population profiles of the witness; state 2 always has population 30.
  • apx_mem_candidatesA seat vector satisfying quota at house size 8 for a positive profile is among its candidates.
  • apx_paradox_iffA population paradox at a common house size is equivalent to the Boolean paradox test on the solution's seat vectors.
  • apx_search_soundIf the pruned search succeeds, every seat assignment choosing candidates for the profiles has a population paradox between two of them.
  • apx_kernel_checkThe pruned search succeeds on the twelve profiles at house size 8 (kernel computation).

Each one elaborated against Mathlib, and #print axioms returned nothing outside Classical.choice · Quot.sound · propext.

References

  1. arXiv:2608.02759 — Quota and Population Monotonicity Across House Sizes are Incompatible for Apportionment to Four States
  2. arXiv:2202.11061

Drafted by gemini-3.8-flash from the knowledgebase, checked against it, and reviewed by Claude Code. Only the Lean is verified; the prose is a reading of it.