| paper | d4-conf-ijcai-2009-0187 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
Theorem 4.4 is a constructive anchor, but its object is a choice rule on tournaments whose vertices are distinct alternatives. A distribution over voter rankings merely induces the tournament, leaving the theorem unchanged and the voter masses unused. Giving mass shifts a strategic role would change the actors and manipulation studied, so no anchor yields a continuous-population computational mirror.
fails bit b-none — no population with multiplicity to relax
No result is mirrored as a population-computation problem; Theorem 4.4 constructs a rule on tournaments, and Theorems 4.2 and 4.3 give impossibility results about rule properties.
The best positive lead is a majority election generated by a continuous voter population. Fix a finite set \(A\) of alternatives; let each voter type be a complete ranking of \(A\), and let \(\mu_t\) be the fraction of society with type \(t\). The society induces a tournament \(T_\mu\), with \(x\) dominating \(y\) when more than half the mass ranks \(x\) above \(y\). This is plausible in a large electorate with many voters but relatively few distinct rankings.
A nearby question would ask whether there is a rule \(r\) on the tournaments induced by such societies that is monotonic, pairwise nonmanipulable, and onto. The solution would be an explicit rule. But this does not make a good mirror of the paper’s problem: once \(T_\mu\) is formed, the voter masses disappear. The paper’s manipulators are alternatives that reverse match outcomes, not voters who change rankings or transfer mass. So the continuous society is just a way of generating the paper’s input; it does not turn the paper’s question into a population-computation problem.
The paper’s named results are Theorem 4.2, proved here: for \(|A|\ge 3\), every PNM choice rule fails Condorcet-consistency; Theorem 4.3, proved here: with \(|A|=3\), every PNM choice rule fails non-imposition; and Theorem 4.4, proved here: for every \(|A|\ne 3\), there exists a monotonic, PNM, non-imposing choice rule. These are combinatorial possibility and impossibility results. None asserts a computational problem is in P, hard, or belongs to another complexity class, and the paper gives no named result analyzing the runtime of computing the rule. Its enumeration for four alternatives, which finds 146 qualifying rules, does not supply such a result either.
So I cannot give a qualifying continuous problem for an anchor: this paper contains no named computational result of the kind the programme asks us to mirror. The weak positive intuition is the continuous-electorate wrapper above; its weakness is decisive for this assignment, because it puts mass in the input while leaving both the paper’s strategic actors and its results unchanged.
The strongest objection is that none of the three named theorems is a computational result. Theorems 4.2 and 4.3 are impossibility statements about axioms; Theorem 4.4 is an existence construction. They concern functions from the finite set of tournaments on \(A\) to \(A\), not algorithms acting on a society. The enumeration for four alternatives—reducing the search to 2,048 rules and finding 146—shows a finite computation, but the paper gives no complexity result for a computational problem.
A continuous electorate is an understandable way to *generate* a tournament: put mass on rankings and orient each edge by the majority. But the paper’s rules and manipulation tests see only the resulting edge directions. The alternatives are the strategic actors, and a manipulation flips an edge; voter mass neither defines the manipulation nor enters the axioms. Asking whether those axioms hold on tournaments generated by a distribution therefore adds a society to the input without making the society the object of the problem. Theorems 4.2 and 4.3 still have no population-computation question to mirror, and Theorem 4.4’s construction still concerns choosing a rule on tournaments.
The stronger candidate would give mass a role by asking how much voter mass must shift to reverse an edge or change the selected alternative. That is a plausible continuous computational problem, but it changes the strategic actors and the manipulation being studied: it asks about voters or campaigns changing a majority relation, not alternatives reversing their match. It could be worthwhile adjacent work, but it is not a mirror of these theorems. Likewise, treating alternatives themselves as a large population would make mass obscure the named alternatives and pairwise relations the paper’s axioms require.
So the negative case is strong for this paper: its results offer no computational anchor for the programme, and making the electorate continuous leaves the relevant object outside the population. The universal claim is harder to establish absolutely—one can always formulate adjacent questions—but those questions would begin a different computational study rather than continuize this paper’s results.
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.