| paper | d4-journals-geb-2012-0016 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
Theorem 2 is an axiomatic characterization, and the proponent agrees it is not a computational result. Computing the characterized rule for a finite set of population types is a plausible extension, but it cannot supply the missing computational result to mirror.
fails bit a — no named computational result to mirror
The proposed computation covers the constrained outcome-egalitarian branch of Theorem 2; it leaves the resource-egalitarian branch and the axioms themselves uncomputed.
The strongest honest case is that the paper’s allocation problem has a plausible continuous-population computational version. But the paper contains no named computational result: it proves no theorem about polynomial time, hardness, or another complexity class. Theorem 2, proved in this paper, is the closest possible anchor. It says that a rule on the paper’s finite-agent economies satisfies priority, solidarity, and composition exactly when it is either resource egalitarian or constrained outcome egalitarian. That is an axiomatic characterization, not a computational result, so it is not an eligible anchor under your criteria.
A concrete mirror of the constrained outcome-egalitarian allocation would use a large population of residents grouped into \(\tau\) health-response types. Each type \(t\) has mass \(\mu_t\), with \(\sum_t\mu_t=1\), and a shared, rational piecewise-linear utility function \(u_t\) describing how health investment affects outcome, including its starting point \(u_t(0)\). The input also gives a rational per-capita budget \(w\). A solution assigns each type a per-person amount \(a_t\geq 0\), using total resources \(\sum_t\mu_t a_t=w\), and computes the constrained outcome-egalitarian allocation: find a threshold \(\lambda\) such that \(a_t=u_t^{-1}(\max\{\lambda,u_t(0)\})\) for every type. This is the paper’s outcome-equalizing rule applied to type masses; equivalently, it lexicographically maximizes the population’s outcome levels subject to the budget.
That regime is plausible when many residents share the same modeled response curve and baseline, while the number of distinct curves is modest. The type description includes everything the model uses, so grouping people with identical modeled curves is consistent with high multiplicity. The authors would likely recognize the question as computing their allocation rule in a population with repeated types. The continuous population is the change; the paper already treats the resource and outcome as real-valued.
I would expect this finite-input problem to be tractable. For the stated piecewise-linear utilities, aggregate resource demand as a function of \(\lambda\) is piecewise linear. Sorting its breakpoints and solving for the threshold gives an algorithm polynomial in the number of types, total number of utility segments, and input bit length. That is a derived expectation, not a result proved in the paper. The mirror covers the constrained outcome-egalitarian branch of Theorem 2’s allocation problem; it does not continuize the axioms themselves or claim that Theorem 2 establishes an algorithmic result.
The soft spot is substantial: this is a new computational question about a rule the paper characterizes, not a continuous mirror of a named computational result. Since the source result is axiomatic and the authors study which rules satisfy principles rather than the complexity of computing them, a referee could accept the model as a natural computational extension while still rejecting it as a continuization of the paper’s result. Further questions include how the allocation behaves under rational finite-population approximations, and how computation changes when response curves are given by a richer or implicit representation.
The decisive point is that this paper has no named computational result to mirror. Theorem 2 characterizes which allocation rules satisfy priority, solidarity, and composition; Theorem 1 characterizes the rules satisfying priority and solidarity. Neither is a result about the complexity of a computational problem. The proponent’s allocation example is therefore a computational question built from a rule in the paper, not a continuous mirror of a computational result the paper establishes.
The proposed health-response setting is nevertheless a plausible high-multiplicity model. If residents share the same fully specified response curve and baseline, representing them by type masses is faithful to the paper’s anonymity assumption. I would not reject it on the ground that agents lack multiplicity, that allocations are fractional, or that the continuous answer might be simple.
The stronger versions of the mirror still miss the programme’s target. One could put a distribution over response-curve types and ask for a continuous-population analogue of the axiomatic characterization. But that remains an axiomatic question, rather than the computational question ChoCo is organized to study. Or one could give finitely many types rational piecewise-linear curves and ask for an algorithm to compute the characterized allocation. That is a well-posed computational extension, but its object is implementing a rule; the paper’s result is about why particular rules are characterized by particular principles. Allowing arbitrary continua of response curves would also require a finite representation before the computational question had a defined input.
So the negative case rests on the scope gate, not on a fundamental defect in the proposed model. If ChoCo wants to include computationalizing rules from axiomatic papers, this is a credible candidate. Under the stated criterion—mirroring a named computational result—the paper supplies no anchor, and no choice of population scenario repairs that gap.
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.