| paper | d4-journals-geb-2022-0075 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
The named results are structural characterizations, uniqueness, and counting, with no computational result under the stated standard. The proposed recognition problem is precise enough to describe, but it is a new task over an externally supplied partition, not a computational result of this paper. The referendum has a plausible high-multiplicity population, but that does not supply the missing computational anchor.
fails bit a — no named computational result to mirror
The proposed threshold representation concerns Theorems 2.4 and 2.7; it does not provide a computational analogue of Corollary 2.9 or of the paper's other results.
The strongest population regime is a large referendum on a binary policy. Voters have exactly three relevant types: prefer \(a\), prefer \(b\), or are indifferent. A citywide vote may have millions of residents but only these three types, so its society is naturally described by masses \(\mu=(\alpha,\beta,\iota)\), with \(\alpha+\beta+\iota=1\). This is a convincing high-multiplicity setting: the proportions are the meaningful quantities, and the paper’s thresholds already describe how those proportions affect the decision.
The paper’s best structural anchor is Theorem 2.4, proved here: every extended quota majority method is anonymous and non-manipulable, and every anonymous, non-manipulable binary social choice function on the paper’s weak-order domain is an extended quota majority method. The paper’s Theorem 2.7, also proved here, adds that every onto function in this class has exactly one proper extended quota representation. Corollary 2.9 counts the functions: there are \(2^{n+1}\) for a society of \(n\) voters. These results make the quota structure a strong basis for a continuous analogue.
However, none is a qualifying computational anchor under the requested standard. They characterize rules, establish uniqueness, and count the family; none classifies a computational problem as polynomial-time solvable, hard, fixed-parameter tractable, or similar. The exact count is not a complexity result about an input problem. So I cannot honestly claim that this paper contains a named computational result to mirror.
If a structural anchor is allowed, the concrete candidate is Limit Extended-Quota Recognition, an asymptotic version of Theorem 2.4. An instance is a finite rational polyhedral partition of the mass simplex, with each open cell labelled \(a\) or \(b\). Ask whether there is a finite quota sequence \(q_0,\ldots,q_r\in\mathbb{Q}\cap[0,1]\), with \(q_r\in\{0,1\}\), that generates those labels away from its threshold lines. At each earlier quota \(q_j\), the rule checks whether \(\alpha>q_j\) or \(\beta>1-q_j\); it selects the corresponding alternative at the first threshold reached. The terminal quota gives the constant fallback: \(q_r=0\) selects \(a\), and \(q_r=1\) selects \(b\). A solution is such a quota sequence; a negative answer requires a certificate that no sequence generates the labelled cells. I would expect this recognition problem to be tractable, since its rules are built from ordered threshold cuts in a two-dimensional simplex. The paper itself establishes no such complexity result.
This candidate follows the paper’s quota representation closely, and the authors could plausibly recognize it as a high-multiplicity limit question. It covers the representation result, with properness and uniqueness as natural follow-up questions. Its weak point is central: for a nonatomic population, an individual’s report changes zero mass, making individual strategy-proofness vacuous. Limit Extended-Quota Recognition therefore studies limits of the paper’s finite strategy-proof rules; it does not preserve their incentive condition as a substantive continuum-agent property. That makes it a plausible extension of their mathematics, but a doubtful mirror of a computational result—and the absence of a qualifying anchor remains decisive under this task’s rubric.
The strongest negative case is that this paper has no computational result to mirror. Theorem 2.4 characterizes anonymous, non-manipulable rules; Theorem 2.7 gives a unique representation for the onto rules; Corollary 2.9 counts them. None classifies an algorithmic problem by its complexity. The count \(2^{n+1}\) is a result about the size of a family, not the running time of a task. That leaves no paper-level computational anchor for the continuization programme.
The referendum scenario is still a convincing high-multiplicity regime: a large population can naturally be summarized by masses \((\alpha,\beta,\iota)\) for the three preference types. The problem is what happens to the central incentive condition. In an atomless population, one person’s report changes zero mass, so individual strategy-proofness places no meaningful constraint on the rule. Requiring incentives to hold against a positive-mass group would avoid that collapse, but it changes the property from individual non-manipulability to a form of group incentive compatibility. Retaining a voter’s ability to change the aggregate instead keeps an atom of size \(1/n\), returning to the finite model whose quotas are being rescaled.
That also weakens Theorem 2.7 as a mirror. One can normalize the finite quotas as \(k_i/n\) and ask for a canonical threshold representation on the mass simplex. But this remains a structural classification, and a limiting sequence may need new conventions about its length and boundary behavior. Asking for the shortest quota sequence from an explicitly supplied partition would be a possible algorithmic problem; it is not a computational result in this paper, and its difficulty would depend on how the rule is encoded.
Corollary 2.9 fares no better as an anchor. With real-valued thresholds, a continuous version might have uncountably many rules, making cardinality uninformative; with a rational grid, the question becomes one about the chosen discretization. Asymptotic counts or description lengths could be studied, but they would be new questions rather than mirrors of a computational result here.
The proposed Limit Extended-Quota Recognition problem is the best attempt to supply one. It can be formalized as a decision problem, but the paper gives no reason to treat a labelled polyhedral partition as a natural input or to care whether it has a quota representation. If rules are supplied by quota sequences, recognition is largely inspection; if supplied by arbitrary partitions, the problem’s encoding and motivation come from the new formulation. Either way, it does not rescue one of the paper’s anchors.
So the negative case is strongest under the programme’s stated scope: this paper offers no worthwhile *mirror of its results* because its results are axiomatic, not computational. It is weaker as a universal claim that no useful computational question could ever grow out of the paper. A high-multiplicity referendum makes the population model sensible, and a new task about learning, recognizing, or designing threshold rules might be worthwhile. The paper itself, though, supplies neither that task nor a computational result whose continuous version would answer it.
The adversarial triple: the proponent anchors on up to three named results; the opponent sees that case and must defeat every anchor; the judge decides which case convinced it. These are the pipeline’s own outputs, generated by tools/triple_run.py — no human edited them. The paper’s own text is not reproduced here beyond the quoted statement above.