| paper | d4-journals-jair-2023-0097 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
Algorithm 1 and the running-time analysis in Section 8.1 support a plausible continuous aggregation problem, but they are unnumbered. The numbered Proposition A.3 establishes social-choice properties, not a computational result, so the proposed mirror has no eligible named result to anchor it.
fails bit a — no named computational result to mirror
The proposed mirror covers recursive opinion aggregation as described in Algorithm 1 and Section 8.1; it leaves the other aggregation functions and the social-choice property results outside its scope.
The strongest substantive positive candidate is a high multiplicity version of the paper’s recursive opinion aggregation. But it does not meet your anchor rule: the paper has no numbered Theorem, Lemma, Corollary, or Proposition classifying a computational problem as polynomial time, hard, FPT, or similar. Its numbered propositions concern aggregation properties. For example, Proposition A.3, proved in the appendix, says recursive aggregation is collectively coherent, has exhaustive and coherent domain, and is anonymous and non-dictatorial; it also says it fails sided and weak unanimity, endorsed unanimity, and familiar monotonicity. That is a result about the function’s properties, not its complexity.
The nearest computational result is Algorithm 1 and the surrounding, unnumbered analysis in Section 8.1. The algorithm computes recursive aggregation bottom-up over the debate structure, averaging relationship acceptances and then weighting each statement’s descendants by those averages. The paper analyzes its running time and reports that, in its online-debate setting, computation is \(O(|R|\cdot |P|)\), where \(P\) is the opinion profile. That is a substantial computational claim, but it is not one of the specified numbered anchors.
If that algorithmic section were allowed as a lead, I would ask the following. High-Multiplicity Recursive Opinion Aggregation: Given a fixed acyclic debate structure \(\langle S,R\rangle\), a target statement \(u\), and a finitely supported distribution \(\mu\) over complete opinion types \(t=(v_t,w_t)\), where \(v_t:S\to[-1,1]\), \(w_t:R\to[0,1]\), and all values and masses are binary rationals, compute the exact collective recursive valuation of \(u\). Each type’s mass is its fraction of the population. Require every relationship used in the recursion to have positive average acceptance, \(\bar w_\mu(r)=\sum_t\mu_t w_t(r)>0\). For a statement with no descendants, use \(V_\mu(s)=\sum_t\mu_t v_t(s)\); otherwise use the acceptance-weighted average of its descendants, with weights \(\bar w_\mu(r)\). The solution is the exact value \(V_\mu(u)\); there is no optimization objective.
A plausible regime is a large participatory consultation with a short, fixed debate structure and coarse ratings. If participants rate eight statements as positive, neutral, or negative, there are at most \(3^8=6{,}561\) valuation types, even for a population of millions; if the debate structure is shared, its relationship acceptances can be fixed across types. The type is the participant’s complete opinion on the structure, and mass is the share with that opinion. When \(\mu_t=n_t/N\), the result exactly matches the paper’s finite-profile recursive aggregation. The natural expectation is tractable: type-level averaging followed by the paper’s DAG computation should run in time polynomial in \(|S|,|R|\), the support size, and the input bit length. This covers recursive aggregation and its coherence guarantee, not the paper’s full axiomatic analysis.
The weakness is that exact complete opinion vectors over a large debate may be nearly unique, especially under the paper’s real-valued opinion model. Then the support size can approach the number of participants, and the high-multiplicity formulation gives little compression. Even in the coarse, short-debate scenario, this is an easy aggregation task rather than a case where discrete hardness dissolves. Questions about threshold decisions or robustness to changes in type masses could follow, but they extend beyond the paper’s computation.
So the positive mirror is plausible, but there are zero compliant computational anchors, and hence no anchor-based problem or complexity classification I can honestly present under the stated requirement.
The negative case wins the anchor requirement: the paper has no numbered theorem, lemma, corollary, or proposition classifying the complexity of a computational problem. Its numbered propositions concern social choice properties. Algorithm 1 and the running-time discussion in Section 8.1 are the nearest computational result, but they are unnumbered. The proposed continuous aggregation problem is therefore a salvage from that section, not a compliant anchor under the stated rule.
That does not defeat the salvage as a mirror. A large participatory consultation with coarse ratings can have many participants sharing the same complete opinion type. Giving each type its population mass yields a well-posed continuous input; weighted averaging followed by the recursive computation gives the exact collective valuation. Rational masses recover finite profiles by choosing a common denominator. The setting need not track identities, and the result need not degenerate. Nor can I dismiss it because the computation may be easy: the instructions explicitly rule that out.
So the honest negative case is narrow: there is no eligible computational anchor in the paper. I cannot support the broader claim that no worthwhile continuous mirror exists in any scenario. If Section 8.1’s unnumbered algorithm is allowed as an anchor, the recursive-aggregation mirror survives.
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.