pith. machine review for the scientific record. sign in

arxiv: 2604.13539 · v1 · submitted 2026-04-15 · 📊 stat.AP

Recognition: unknown

Relative plausibility versus probabilism: A level-of-analysis error in juridical proof

Stanley E. Lazic

Authors on Pith no claims yet

Pith reviewed 2026-05-10 12:40 UTC · model grok-4.3

classification 📊 stat.AP
keywords relative plausibility theoryjuridical proofprobabilistic reasoningMarr levels of analysislegal evidenceposterior oddsexplanatory comparison
0
0 comments X

The pith

The debate between relative plausibility theory and probabilistic accounts of legal proof rests on a level-of-analysis error, making the two approaches compatible.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that framing relative plausibility theory against probabilistic models as rivals in juridical proof confuses two different kinds of description. Relative plausibility theory states the computational task: compare competing explanations in light of the evidence to determine whether a standard of proof has been satisfied. Probabilistic approaches instead supply the algorithmic details of how those comparisons can be represented and carried out numerically. When plausibility judgments meet basic coherence requirements, the two line up because relative plausibility matches posterior odds. Recognizing the distinction removes the appearance of conflict and shows how explanatory comparison and probability can each contribute to legal reasoning without one displacing the other.

Core claim

RPT provides a computational-level description of juridical proof, characterizing the task of comparing explanations in light of the evidence and assessing whether a standard of proof has been met. Probabilistic approaches supply algorithmic-level accounts that specify how such comparative assessments can be represented and computed. When plausibility judgments satisfy minimal coherence conditions, relative plausibility corresponds to posterior odds.

What carries the argument

Marr's distinction between computational and algorithmic levels of analysis, used to separate the task description of comparing explanations from the procedures for computing their relative strength.

If this is right

  • Relative plausibility theory and probabilistic accounts become complementary rather than opposed.
  • Legal reasoning can integrate explanatory comparison at the task level with probabilistic computation at the method level.
  • Relative plausibility judgments track posterior odds when minimal coherence conditions hold.
  • Longstanding disputes about the proper role of probability in law are resolved by clarifying the level at which each account operates.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Legal education could separate instruction on proof standards from training on evidence-weighting procedures.
  • Empirical tests of juror decisions could check whether their plausibility orderings match Bayesian updating on the same facts.
  • The same level distinction might clarify analogous debates in scientific hypothesis evaluation.

Load-bearing premise

Marr's levels of analysis can be applied directly to the debate over juridical proof without distortion.

What would settle it

A case in which coherent plausibility judgments and probabilistic calculations yield systematically different verdicts on identical evidence would falsify the claim that the approaches operate at separate levels and align under coherence.

read the original abstract

Debates about juridical proof are often framed as a conflict between probabilistic approaches and relative plausibility theory (RPT). This paper argues that this opposition rests on a level-of-analysis error. Drawing on Marr's distinction between levels of analysis, we show that RPT and probabilistic approaches operate at different conceptual levels and are therefore compatible rather than competing theories. RPT provides a computational-level description of juridical proof, characterizing the task of comparing explanations in light of the evidence and assessing whether a standard of proof has been met. Probabilistic approaches supply algorithmic-level accounts that specify how such comparative assessments can be represented and computed. When plausibility judgments satisfy minimal coherence conditions, relative plausibility corresponds to posterior odds. Recognizing this distinction clarifies longstanding disputes and highlights the complementary roles of explanation and probability in legal reasoning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims that debates framing relative plausibility theory (RPT) as opposed to probabilistic accounts of juridical proof rest on a level-of-analysis error. Invoking Marr's distinction between computational and algorithmic levels, it positions RPT as supplying a computational-level description of the task (comparing explanations in light of evidence and assessing whether a standard of proof is met), while probabilistic approaches provide algorithmic-level specifications of how such comparisons can be represented and computed. The paper asserts that, when plausibility judgments satisfy minimal coherence conditions, relative plausibility corresponds to posterior odds, rendering the approaches compatible rather than competing.

Significance. If the result holds, the re-framing offers a constructive way to reconcile explanatory and probabilistic perspectives in legal reasoning, potentially resolving apparent conflicts by clarifying their distinct roles. This could advance interdisciplinary work at the intersection of law, epistemology, and decision theory. The argument is purely conceptual with no empirical tests, formal derivations, or parameter fittings supplied, so its significance rests on the soundness of applying Marr's framework without distortion to normative legal standards.

major comments (2)
  1. [Abstract and correspondence section] Abstract and the section developing the correspondence claim: The statement that 'when plausibility judgments satisfy minimal coherence conditions, relative plausibility corresponds to posterior odds' is asserted without supplying the derivation, mapping, or proof that would show how legal plausibility judgments reduce to or align with posterior odds under those conditions. This mapping is load-bearing for the compatibility thesis.
  2. [Marr's levels application] Section applying Marr's levels to juridical proof: The separation treats normative criteria (such as the definition of 'beyond reasonable doubt') as external constraints on an information-processing task whose inputs and outputs can be specified independently. If those criteria are constitutive of what counts as a valid comparison of explanations, the computational/algorithmic distinction collapses and the claimed compatibility does not follow. The manuscript provides no argument addressing this risk.
minor comments (1)
  1. [Introduction] The introduction could include more explicit citations to foundational works in relative plausibility theory (e.g., the specific papers by Allen or others) to better delineate the precise disputes being reframed.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for these constructive comments, which highlight areas where the manuscript's central claims require greater explicit support. We respond to each major comment below and note the revisions we will undertake.

read point-by-point responses

  1. Referee: [Abstract and correspondence section] Abstract and the section developing the correspondence claim: The statement that 'when plausibility judgments satisfy minimal coherence conditions, relative plausibility corresponds to posterior odds' is asserted without supplying the derivation, mapping, or proof that would show how legal plausibility judgments reduce to or align with posterior odds under those conditions. This mapping is load-bearing for the compatibility thesis.

    Authors: We agree that the correspondence claim is central to the compatibility thesis and that the manuscript currently states the result without an explicit derivation or mapping. In the revised version we will add a short formal subsection (or appendix) that supplies the mapping: under the stated minimal coherence conditions (transitivity of the plausibility ordering together with respect for evidential entailment), the relative-plausibility ratio between two explanations is ordinally equivalent to their posterior-odds ratio. The argument relies on the standard representation theorem that any coherent qualitative probability ordering can be extended to a quantitative probability measure, thereby aligning the two frameworks at the level of the comparison itself. revision: yes

  2. Referee: [Marr's levels application] Section applying Marr's levels to juridical proof: The separation treats normative criteria (such as the definition of 'beyond reasonable doubt') as external constraints on an information-processing task whose inputs and outputs can be specified independently. If those criteria are constitutive of what counts as a valid comparison of explanations, the computational/algorithmic distinction collapses and the claimed compatibility does not follow. The manuscript provides no argument addressing this risk.

    Authors: This objection correctly identifies a potential vulnerability in applying Marr's framework to a normative domain. We maintain that the computational level can incorporate the normative standard as part of the task's output specification (the task is to decide whether the evidence renders one explanation sufficiently more plausible than its rivals to meet the proof standard), while the algorithmic level concerns only the representational and computational procedures used to reach that output. The norms therefore define the goal of the computation without dictating its internal form. We will revise the relevant section to include an explicit paragraph addressing this distinction and the risk of collapse, clarifying that constitutive norms shape what counts as a successful output but leave open multiple algorithmic realizations. revision: partial

Circularity Check

0 steps flagged

No circularity; applies external Marr framework to reframe debate without internal reduction

full rationale

The paper invokes Marr's levels of analysis as an external conceptual tool to classify RPT as a computational-level description (task of comparing explanations and assessing standards of proof) and probabilistic accounts as algorithmic-level (how to represent and compute comparisons). The stated correspondence—that relative plausibility corresponds to posterior odds when plausibility judgments satisfy minimal coherence conditions—is presented conditionally without any derivation, equations, or parameter fitting internal to the paper that would make the result equivalent to its inputs by construction. No self-citations are load-bearing, no ansatzes are smuggled, and no known results are renamed as novel derivations. The argument is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The claim rests on the applicability of Marr's levels to legal proof and on the existence of minimal coherence conditions that link plausibility to posterior odds; no free parameters or new entities are introduced.

axioms (1)
  • domain assumption Marr's distinction between computational, algorithmic, and implementational levels of analysis applies without distortion to the domain of juridical proof.
    Invoked to separate RPT from probabilistic approaches.

pith-pipeline@v0.9.0 · 5427 in / 1259 out tokens · 55706 ms · 2026-05-10T12:40:48.724128+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages

  1. [1]

    doi: 10.48550/arXiv.2008. 05912. C. Aitken, F. Taroni, and S. Bozza. Evidence, probability and relative plausibility.The International Journal of Evidence & Proof, 26(4):309–324, July

    work page doi:10.48550/arxiv.2008 2008

  2. [2]

    doi: 10.1177/13657127221114508

    ISSN 1740-5572. doi: 10.1177/13657127221114508. 8 Relative plausibility versus probabilism C. Aitken, F. Taroni, and S. Bozza. The role of the Bayes factor in the evaluation of evidence.Annual Review of Statistics and Its Application, 11(1):203–226, Apr

    work page doi:10.1177/13657127221114508

  3. [3]

    doi: 10.1146/annurev-statistics-040522-101020

    ISSN 2326-831X. doi: 10.1146/annurev-statistics-040522-101020. R. J. Allen and M. S. Pardo. Relative plausibility and its critics.The International Journal of Evidence & Proof, 23 (1–2):5–59, Jan. 2019a. ISSN 1740-5572. doi: 10.1177/1365712718813781. R. J. Allen and M. S. Pardo. Clarifying relative plausibility: A rejoinder.The International Journal of Ev...

    work page doi:10.1146/annurev-statistics-040522-101020

  4. [4]

    doi: 10.1177/13657127231155798

    ISSN 1740-5572. doi: 10.1177/13657127231155798. P. G. Bissiri, C. C. Holmes, and S. G. Walker. A general framework for updating belief distributions.Journal of the Royal Statistical Society Series B: Statistical Methodology, 78(5):1103–1130, Feb

    work page doi:10.1177/13657127231155798

  5. [5]

    and Holmes, C

    ISSN 1467-9868. doi: 10.1111/rssb.12158. L. Bovens and S. Hartmann.Bayesian Epistemology. Oxford University Press, Oxford, UK,

    work page doi:10.1111/rssb.12158

  6. [6]

    URLhttps://doi.org/10.1214/ss/1009212673

    ISSN 0883-4237. doi: 10.1214/ss/1009212673. K. M. Clermont. Common sense on standards of proof.SSRN Electronic Journal,

    work page doi:10.1214/ss/1009212673

  7. [7]

    doi: 10.2139/ssrn.3007129

    ISSN 1556-5068. doi: 10.2139/ssrn.3007129. L. J. Cohen.The Probable and the Provable. Oxford University Press, Oxford, UK,

    work page doi:10.2139/ssrn.3007129

  8. [8]

    doi: 10.1119/1.1990764

    ISSN 1943-2909. doi: 10.1119/1.1990764. R. T. Cox.The Algebra of Probable Inference. Johns Hopkins Press, Baltimore,

    work page doi:10.1119/1.1990764 1943

  9. [9]

    doi: 10.1177/13657127231209019

    ISSN 1740-5572. doi: 10.1177/13657127231209019. D. Deutsch.The Beginning of Infinity. Penguin Books, London, UK,

    work page doi:10.1177/13657127231209019

  10. [10]

    doi: 10.2307/1339656

    ISSN 0017-811X. doi: 10.2307/1339656. J. B. Gelbach. It’s all relative: Explanationism and probabilistic evidence theory.The International Journal of Evidence & Proof, 23(1–2):168–175, Jan

    work page doi:10.2307/1339656

  11. [11]

    doi: 10.1177/1365712718816747

    ISSN 1740-5572. doi: 10.1177/1365712718816747. A. Ghosh and A. Basu. Robust Bayes estimation using the density power divergence.Annals of the Institute of Statistical Mathematics, 68(2):413–437, Jan

    work page doi:10.1177/1365712718816747

  12. [12]

    doi: 10.1007/s10463-014-0499-0

    ISSN 1572-9052. doi: 10.1007/s10463-014-0499-0. I. J. Good.Probability and the Weighing of Evidence. Charles Griffin & Company, London,

    work page doi:10.1007/s10463-014-0499-0

  13. [13]

    doi: 10.1093/biomet/66.2.393

    ISSN 1464-3510. doi: 10.1093/biomet/66.2.393. C. C. Holmes and S. G. Walker. Assigning a value to a power likelihood in a general Bayesian model.Biometrika, 104 (2):497–503,

    work page doi:10.1093/biomet/66.2.393

  14. [14]

    doi: 10.1093/biomet/asx010

    ISSN 1464-3510. doi: 10.1093/biomet/asx010. P. Horwich.Probability and Evidence. Cambridge University Press, Cambridge, UK,

    work page doi:10.1093/biomet/asx010

  15. [15]

    doi: 10.1198/016214503388619229

    ISSN 1537-274X. doi: 10.1198/016214503388619229. E. T. Jaynes.Probability Theory: The Logic of Science. Cambridge University Press, Cambridge, UK,

    work page doi:10.1198/016214503388619229

  16. [16]

    doi: 10.2307/1227491

    ISSN 0038-9765. doi: 10.2307/1227491. D. Kaye. The laws of probability and the law of the land.The University of Chicago Law Review, 47(1):34,

    work page doi:10.2307/1227491

  17. [17]

    doi: 10.2307/1599414

    ISSN 0041-9494. doi: 10.2307/1599414. S. P. Lyddon, C. C. Holmes, and S. G. Walker. General Bayesian updating and the loss-likelihood bootstrap.Biometrika, 106(2):465–478, Mar

    work page doi:10.2307/1599414

  18. [18]

    doi: 10.1093/biomet/asz006

    ISSN 1464-3510. doi: 10.1093/biomet/asz006. 9 Relative plausibility versus probabilism D. J. C. MacKay. Bayesian interpolation.Neural Computation, 4(3):415–447,

    work page doi:10.1093/biomet/asz006

  19. [19]

    Neural Computation4(3), 415–447 (1992) https://doi.org/10.1162/neco.1992.4.3.415

    ISSN 1530-888X. doi: 10.1162/neco.1992.4.3.415. D. J. C. MacKay.Information Theory, Inference, and Learning Algorithms. Cambridge University Press, Cambridge, UK, 22nd printing edition,

    work page doi:10.1162/neco.1992.4.3.415 1992

  20. [20]

    doi: 10.3389/fnins.2018.00734

    ISSN 1662-453X. doi: 10.3389/fnins.2018.00734. D. McManaman.Introduction to Plausible Reasoning. Lightning Source Inc, Nov

    work page doi:10.3389/fnins.2018.00734 2018

  21. [21]

    doi: 10.1177/1365712718816256

    ISSN 1740-5572. doi: 10.1177/1365712718816256. M. S. Pardo and R. J. Allen. Juridical proof and the best explanation.Law and Philosophy, 27(3):223–268, Sept

    work page doi:10.1177/1365712718816256

  22. [22]

    doi: 10.1007/s10982-007-9016-4

    ISSN 1573-0522. doi: 10.1007/s10982-007-9016-4. C. S. Peirce.Illustrations of the Logic of Science. Open Court, New York,

    work page doi:10.1007/s10982-007-9016-4

  23. [23]

    doi: 10.1177/1365712718819509

    ISSN 1740-5572. doi: 10.1177/1365712718819509. D. S. Schwartz and E. Sober. What is relative plausibility?The International Journal of Evidence & Proof, 23(1–2): 198–204, Jan

    work page doi:10.1177/1365712718819509

  24. [24]

    doi: 10.1177/1365712718816259

    ISSN 1740-5572. doi: 10.1177/1365712718816259. S. P. Sullivan. A likelihood story: The theory of legal fact-finding.University of Colorado Law Review, 90(1),

    work page doi:10.1177/1365712718816259

  25. [25]

    doi: 10.1111/j.1539-6924.2005.00699.x

    ISSN 1539-6924. doi: 10.1111/j.1539-6924.2005.00699.x. J. R. Welch. Rebooting the new evidence scholarship.The International Journal of Evidence & Proof, 24(4):351–373, Aug

    work page doi:10.1111/j.1539-6924.2005.00699.x 2005

  26. [26]

    doi: 10.1177/1365712720943329

    ISSN 1740-5572. doi: 10.1177/1365712720943329. 10

    work page doi:10.1177/1365712720943329