| paper | d4-journals-jet-2012-0077 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
Propositions 1–4 establish impossibilities about which axioms a representable social welfare function can satisfy; they do not assert that a computational object can be found or that computing one is hard. Treating generations as cohorts does not add a computational task without introducing a new feasibility or optimization problem, so no anchor meets the computational-result requirement.
fails bit a — no named computational result to mirror
The proposed mirror considered Propositions 1–4, all axiomatic nonexistence results; Lemma 1 and the introductory comparisons are also axiomatic or background material, leaving no computational result to mirror.
I can make a modest cohort interpretation, but I can’t make a qualifying case for a ChoCo mirror from this paper. In a large intergenerational economy, each \(x_i\) could be read as the average utility of a large cohort. That makes the paper’s fairness questions relevant to populations. But the paper studies whether a social welfare function over utility streams can satisfy certain axioms; it does not define a mass-transfer or other computational problem over voter types, nor analyze its complexity. Treating generations as cohorts does not supply that missing problem.
The closest named results are Propositions 1–4, all proved in this paper. Proposition 1 says there is no SWF \(W:[0,1]^{\mathbb N}\to\mathbb R\) satisfying Pigou–Dalton transfers and Weak Pareto. Proposition 2 says there is no such \(W\) satisfying Hammond Equity and Weak Pareto. Proposition 3 says there is no such \(W\) satisfying Very Weak Inequality Aversion and Weak Pareto. Proposition 4 says there is no such \(W\) satisfying Independent Future, Hammond Equity for the Future, and Weak Pareto. These are axiomatic nonexistence results, not complexity results about a computational problem.
So I have no honest anchor, and therefore no continuous problem per anchor to hand to a prover. Turning the stream into a continuum of generations would continuize time, while turning each generation into a continuum of people would leave the paper’s central question—representability and compatibility of axioms—intact. Either may motivate related work, but neither gives this paper a computational population mirror under the stated scope.
The proponent names Propositions 1–4, but none is a computational result. Propositions 1–3 rule out a representable social welfare function satisfying particular pairs of axioms; Proposition 4 rules one out for Independent Future, Hammond Equity for the Future, and Weak Pareto. These are the paper’s nearest anchors, but they specify no finite input, feasible action, or algorithmic output whose complexity could be mirrored. Their proofs instead show that an entire class of functions cannot exist, using order and representability arguments over \([0,1]^{\mathbb N}\). There is no algorithmic problem here waiting for a continuous-population version.
The strongest population interpretation would treat each generation as a large cohort and \(x_i\) as that cohort’s average utility. That is a plausible high-multiplicity setting, but it does not supply the missing computational task. The propositions ask how a scalar evaluator ranks whole utility streams. They do not ask how to choose a feasible stream, transfer utility at some cost, or implement a welfare improvement. Adding those ingredients could yield a new optimization problem, but the paper gives no basis for choosing them, and an evaluator satisfying the combined axioms is precisely what the propositions rule out.
That cohort reading also does not make the population itself continuous in the relevant sense. The axioms compare particular generations: HE and the transfer principles concern two coordinates, while HEF and IF distinguish the present from the future. Representing people only by a distribution of utility types would erase those dates and distinctions. Encoding date in each type preserves them, but leaves the model as a time-indexed stream with infinitely many distinct positions; it does not turn the paper’s question into a finite-support population problem. Replacing dates with a continuum would continuize the horizon, not the population, and would still pose an axiomatic existence question.
So the case against is strong for this paper’s actual results, though not a proof that no related computational project could ever be worthwhile. A new project on allocating resources across large cohorts might be valuable. It would need its own feasibility model and computational objective, however; it would not be a continuous mirror of any of Propositions 1–4.
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.