| paper | d4-conf-ecai-2014-0077 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The paper’s named results are axiomatic impossibility and preservation claims; none gives an algorithm, complexity result, or computational formulation used to compute. A distribution over graph types gives a recognizable population analogue, but it remains an axiomatic question, so bit (a) fails.
fails bit a — no named computational result to mirror
Theorem 8, Lemma 4, Propositions 1, 2, and 7, and the applications concern axiomatic collective rationality; the paper has no computational results for the mirror to cover.
There is no named computational result in this paper that meets the required anchor test. Theorem 8, proved here, is the main result: for \(|V|\ge3\), no unanimous, grounded, IIE, non-dictatorial graph aggregator can be collectively rational for a property that is contagious, implicative, and disjunctive. Proposition 7 and Lemma 4 are also axiomatic results. None classifies a problem by computational complexity, so I can’t honestly use one as an anchor for a ChoCo computational mirror.
The strongest nearby population-continuum question would be: for \(|V|\ge3\) and such a graph property \(P\), does there exist a rule \(F:\Delta(2^{V\times V})\to2^{V\times V}\) that takes a binary-rational distribution \(\mu\) over graph types to a collective graph, decides each edge using only its population share \(p_e=\sum_{G\ni e}\mu_G\), includes edges with \(p_e=1\), excludes those with \(p_e=0\), and returns a graph in \(P\) whenever \(\mu\) is supported on \(P\)? This is plausible for a large consultation where many people submit the same graph of relations among a fixed set of issues. Theorem 8’s impossibility transfers to this setting by restricting \(F\) to empirical distributions of finite electorates. But that gives a population-continuous axiomatic analogue, not a computational result or a complexity question. Without a qualifying anchor, the paper cannot support the requested ChoCo case.
The negative case is strongest at the level of the paper’s results: none of its named results is computational. Lemma 4 and Theorem 8, the two propositions, and the applications characterize when an aggregator can preserve graph properties under axioms such as unanimity and independence. They give no algorithm, complexity classification, or optimization problem to mirror. The case for therefore raises no qualifying anchor.
Its proposed continuous version is still a coherent model. A large consultation could have a distribution \(\mu\) over graph submissions, and one could ask whether an aggregator preserves \(P\) on societies supported on \(P\). But that remains an axiomatic existence question. The finite theorem carries over by restricting the input to empirical distributions: each finite profile gives a rational \(\mu\). That makes the continuum analogy clean, but does not turn the result into a computational question.
A stronger computational proposal might ask for the minimum mass of reports to change so that a fixed aggregator returns a graph in \(P\), or for the closest \(P\)-graph to some aggregate. Those could be worthwhile new problems, but they add an intervention, cost model, or objective that the paper does not supply; none of its results anchors their complexity. So the case against is strong that this paper is outside ChoCo’s stated computational scope. It is weaker as the universal claim that no useful continuous computational project could be built around graph aggregation at all.
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.