| paper | d4-conf-ijcai-2018-0307 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
The paper’s named results are representation or impossibility claims about axioms, not algorithms, complexity bounds, or computational hardness. The proposed weighted population question changes the identity-sensitive dictatorship claim and remains axiomatic, so no anchor passes the computational gate.
fails bit a — no named computational result to mirror
No computational mirror survives for Theorem 3 or the paper’s representation and limitation results; the proposed weighted axiomatic feasibility question is a new problem.
The strongest honest positive case is an adjacent population version of the paper’s aggregation question, but it does not meet ChoCo’s computational-anchor requirement.
The closest named anchor is Theorem 3, stated as the main impossibility theorem in this extended abstract; its proof is deferred to Mata Díaz and Pino Pérez (2017). It says that any ES basic complete merging operator satisfying Standard Domain \((\mathrm{ESF\mbox{-}SD})\), Pareto \((\mathrm{ESF\mbox{-}P})\), and Independence \((\mathrm{ESF\mbox{-}I})\) is dictatorial, meaning it satisfies \((\mathrm{ESF\mbox{-}D})\). That is a substantive impossibility result, but it is axiomatic: it does not establish a problem’s complexity or give an algorithm. The paper has no named computational-complexity result to serve as a qualifying anchor.
A plausible population setting would be a large pool of medical or policy information sources. A type would be a complete epistemic state, including every feature the merging rule uses; \(\mu_t\) would be the fraction of sources of type \(t\), with many sources but relatively few types. For a finite type set \(T\), a candidate problem is: given the propositional language, epistemic space, admissible types \(T\), and integrity constraints, does there exist a rule
\[
F:\Delta_{\mathbb Q}(T)\times\mathcal K\to\mathcal E
\]
that maps each rational mass distribution and constraint to a merged epistemic state, and satisfies population versions of Standard Domain, Pareto, and Independence? A solution would be such a rule; a negative answer would be an impossibility proof. This is a coherent axiomatic-feasibility question, but specifying the population versions of those postulates is new work, and the problem is not a computational mirror of Theorem 3. The theorem’s “dictator” is a named individual in a profile \(N\to\mathcal E\); a distribution erases individual identity. Replacing that person with a dictatorial type or class changes the claim. Keeping identity as part of each type, meanwhile, can make the number of types grow with the population.
So I would not classify this candidate as tractable, hardness-transferring, or continuum-specifically hard. If the axioms were adapted successfully, an Arrow-style impossibility is plausible, but that would be a new axiomatic result, not a computational consequence of this paper. The weakest point—and the reason this cannot carry a ChoCo case—is that the population model may be sensible for merging many sources, while the named result’s core is an identity-sensitive axiomatic impossibility rather than a computational problem.
Theorem 3 is the only plausible anchor, and it fails ChoCo’s computational gate: it is an axiomatic implication from Standard Domain, Pareto, and Independence to dictatorship. Theorem 1’s representation result and Theorem 2’s limitation are axiomatic or structural too. None states a computational problem, complexity bound, or algorithm whose complexity landscape a continuous mirror could chart.
The strongest population version would take rational masses over complete epistemic-state types and ask whether a merge rule satisfies weighted versions of the paper’s social postulates. That could be a coherent research question, but it changes the theorem’s central object. In Theorem 3, a dictator is a named agent \(d_N\). A distribution over types forgets which individual supplied each state. Keeping identities in the types makes the support grow with the population; replacing the individual dictator with a dictatorial type or cohort changes the property being studied. That is not a better encoding of Theorem 3. It is a new axiomatic problem, and it still supplies no computational question unless one adds a separate task such as computing an outcome for a specified operator.
A large pool of repeated information sources may be a sensible high-multiplicity setting. That makes a weighted Arrow-style study plausible as adjacent work, but it does not turn this paper’s impossibility theorem into a computational result. I would reject this as a worthwhile *ChoCo mirror*. The universal claim that no useful continuous study could arise from the paper is less secure: a new weighted-axiom project might be worthwhile on its own. Its weakness as a negative case is precisely that it rules out a mirror of the paper’s results, not every future problem inspired by them.
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.