| paper | d4-conf-atal-2014-0324 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The proposed distributional maximum-support problem is well-defined and is a faithful population version of \(\delta_{\mathrm{RM}\Sigma}\); the abstention objection does not defeat that objective. But Proposition 9 proves a likelihood characterization, without an algorithmic guarantee or a computational formulation used to compute it, so the paper supplies no qualifying computational anchor.
fails bit a — no named computational result to mirror
The proposed mirror covers Proposition 9's maximum-support selection rule; it leaves the paper's impossibility, rationality-property, and other truth-tracking results aside.
The paper has no named computational-complexity result of the kind your anchor rule requests: no theorem or proposition classifies a problem as polynomial-time, NP-hard, or similar. Its propositions establish rationality properties, impossibility results, and truth-tracking characterizations. I can’t honestly present one as a complexity result.
The closest anchor for a positive case is Proposition 9, proved in this paper. It states that, under independence of agents and issues, homogeneous reliability with each agent more likely to judge an issue correctly than incorrectly, and the paper’s other stated assumptions, the judgment aggregation method \(\delta_{\mathrm{RM}\Sigma}\) gives a maximum-likelihood estimate of the true potential solution. In particular, it selects the consistent resolute judgment sets \(M=\langle\alpha_1,\ldots,\alpha_m\rangle\) maximizing \(\sum_k g_P^+(\alpha_k)\), the total support for their chosen issue judgments. This is a truth-tracking characterization, not a complexity classification.
The one mirror I would put forward is Continuous Maximum-Support Judgment Aggregation. An instance consists of a finite propositional agenda \(X\) and a list of \(\tau\) distinct, logically consistent judgment types \(t:X\to\{0,1,\star\}\), each assigned a binary rational mass \(\mu_t\ge 0\), with \(\sum_t\mu_t=1\). Here \(t(\phi)=\star\) means the type abstains on \(\phi\). The task is to return all consistent, resolute collective judgments \(\gamma:X\to\{0,1\}\) maximizing
\[ \sum_t \mu_t \sum_{\phi\in X} \mathbf{1}\!\left[t(\phi)=\gamma(\phi)\ne\star\right]. \]
The objective is the fraction-weighted total support for the selected judgments; abstentions contribute no support. The output remains a logically consistent collective judgment, while the society is represented by masses over types.
A plausible regime is a large public consultation on a standard, finite agenda, where many respondents share one of a relatively small number of recurring, coherent response patterns. There, \(N\) respondents may number in the hundreds of thousands while \(\tau\), the number of distinct patterns, is much smaller. A type is the complete judgment vector used by this aggregation problem; mass is the share of respondents with that vector. This is a credible high-multiplicity setting, though it depends on response patterns actually recurring.
The bridge to Proposition 9 is direct: for a finite profile with \(n_t\) voters of each type, set \(\mu_t=n_t/n\). The continuous objective is then exactly the finite total-support score divided by \(n\), so it has the same maximizers as \(\delta_{\mathrm{RM}\Sigma}\). Conversely, any rational mass distribution can be represented by a finite profile using a common denominator. That makes this a precise high-multiplicity version of the paper’s maximum-support choice, rather than a change to the outcome space.
I would expect the exact problem’s difficulty to come from consistency constraints on the agenda, so I would investigate hardness transfer (Class B) first. For arbitrary propositional agendas, choosing the best-supported consistent set has the shape of a weighted satisfiability optimization problem. The paper itself does not prove that this problem is NP-hard, so that classification needs a separate reduction; the continuum formulation alone does not establish it. The natural follow-up questions are its complexity in \(m,\tau,L\) and formula size, whether agenda structure yields tractable cases, and whether useful approximation or parameterized algorithms exist.
The scope is deliberately narrow: it mirrors the \(\delta_{\mathrm{RM}\Sigma}\) maximum-likelihood/max-support result, not the paper’s broader impossibility and rationality results. Its weakest point is also clear: Proposition 9 is not a computational result, and rational masses can be expanded into a finite profile. The positive case is that the paper supplies a natural optimization objective and the continuum gives its high-multiplicity input regime; it does not supply the complexity anchor your rubric ideally calls for.
The strongest negative point is that this paper offers no named computational result to continuize. Its propositions concern rationality properties, impossibility, and truth tracking. Proposition 9 is a maximum-likelihood characterization, not a complexity result or an algorithmic theorem. So the proposed problem is not a continuous mirror of a computational result in the paper.
There is also a mismatch in the stated anchor. Proposition 9’s likelihood calculation treats judgments as binary: each issue contributes either a correct judgment or an incorrect one. The proposed society includes abstentions and assigns them zero support. The paper defines \(\delta_{\mathrm{RM}\Sigma}\) for partial judgments, but Proposition 9 does not establish its likelihood claim for that model. Restricting the types to resolute judgments would fix this mismatch. But for agendas with logical dependencies, the paper itself says issue independence may fail; then the result no longer supplies the stated maximum-likelihood justification.
That does not defeat the best version of the mirror. Given a distribution \(\mu\) over complete, consistent judgment types, one can ask for a consistent resolute judgment maximizing expected support. For resolute types, this is exactly the finite maximum-support objective divided by the population size. A large consultation on a fixed agenda, where many respondents give the same judgment set, is a plausible high-multiplicity regime. The objective does not vanish as the population grows, and using whole judgment sets as types preserves the relevant information. The finite-to-continuous equivalence is the high-multiplicity bridge the programme is meant to study, not by itself a collision or a flaw.
So I can make a case against presenting Proposition 9 as a computational anchor: it is not one, and repairing the model leaves a new algorithmic question rather than a complexity result the paper already proves. I cannot honestly turn that into the stronger claim that no worthwhile continuous mirror exists. The repaired maximum-support problem survives the proposed objections; its value would need to be argued as a new computational question motivated by this paper’s aggregation rule, not as a continuization of one of its computational 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.