{"path":"research/what-a-rule-rests-on/cases-and-principles-sources-as-read-2026-10-08.md","content":"# Agreement on cases and on principles: sources as a helper read them (2026-10-08, evening)\n\nA read-only helper (Sonnet; 14 pages opened, about a dozen more blocked) was asked to test one\nsentence from `docs/convergence.md` against published work, and to look as hard for evidence\nagainst it as for it. Its report is below, unchanged. Afterwards Claude opened one of its sources\ndirectly (Awad et al. 2020, the abstract, from Europe PMC), and it reads as reported. What Claude\nmade of the report: [../what-a-rule-rests-on.md](../what-a-rule-rests-on.md) § 13.\n\n**The sentence tested:** \"People who disagree often agree on what to do in a particular case before\nthey can agree on any principle. So agreement on cases comes first, and principles are drawn from it\nafterwards.\"\n\n**The leads the helper was given** were Claude's, from memory and marked unverified: Sunstein 1995;\nJonsen and Toulmin 1988; moral dumbfounding; Awad 2020 and Bago 2022; Prothro and Grigg 1960 and\nMcClosky 1964; the gap between principle and implementation in racial attitudes; Beauchamp against\nToulmin; Arras 1991; Nichols and Knobe 2007; Dryzek and Niemeyer 2006.\n\n---\n\nShort answer: as worded, the claim is not supported. The sources support a narrower claim. Small bodies that must decide together and share a tradition can often agree on an outcome and on low-level reasons without agreeing on high-level theory. In mass publics, the better-documented pattern runs the other way: agreement on abstractions, disagreement on application. I found no study that measures case-level and principle-level agreement in the same deliberators and also establishes which comes first.\n\nReach: I read 14 distinct pages. About a dozen further attempts were blocked (403 or login redirect), including Sunstein's article on Chicago Unbound and SSRN, PNAS, Nature, Wiley, Semantic Scholar and Springer. I did not bypass any of them. Several full texts came from third-party-hosted PDF copies.\n\n| Source | Year | What it found | Direction | Read | URL |\n|---|---|---|---|---|---|\n| Sunstein, Incompletely Theorized Agreements (108 Harv. L. Rev. 1733) | 1995 | Names three kinds: (a) agreement on a general principle with disagreement on cases; (b) a mid-level principle with disagreement on both theory and cases; (c) his focus, agreement on outcomes plus low-level principles. \"a comfortable and even emphatic agreement on a general principle, accompanied by sharp disagreement about particular cases\" | Both | Secondary: Solum's page block-quotes pp. 1739-40. The Harvard Law Review text was not reached. | legaltheorylexicon.com/2004/05/legal_theory_le_1-8/ |\n| Jonsen, \"The Birth of the Belmont Report\" (talk to NBAC) | 1998 | Congress's mandate was to identify principles. The commission read Jonas, Ramsey and five commissioned ethicists, and respect for persons emerged as dominant from that literature. \"The Commissioners believed, as principalists, they worked as casuists.\" | Cuts against (qualifies Jonsen and Toulmin) | Full transcript, read in raw text | sitearchives.georgetown.edu/kie-nbac/transcripts/jul98/belmont.html |\n| SEP, \"Theory and Bioethics\" (reporting Toulmin 1981 and Jonsen and Toulmin 1988) | n/a | Toulmin reports agreement on cases despite theoretical disagreement. The entry lists a criticism: casuistry may need \"pre-established agreement on fundamental values\". \"Agreement amongst members of the commission was relatively easy to reach even on difficult cases\" | Supports as testimony, not measurement | Secondary summary, via fetch tool. Toulmin and the book itself were not opened. | plato.stanford.edu/entries/theory-bioethics/ |\n| Beauchamp, \"Methods and principles in biomedical ethics\" (J Med Ethics) | 2003 | Accepts paradigm cases as authorities, but says analogy needs a connecting norm, so principles are not optional. \"casuistry is not a rival methodological model to the use of principles and specification\" | Cuts against | Full text of the method section | rlillo.educsalud.cl/Capac_Etica_BecadosFOREAPS/09___Beauchamp_principles__method.pdf |\n| Arras, \"Getting Down to Cases\" | 1991 | Casuistry meets conflicting \"deep\" interpretations. It is therefore \"unlikely sources of increased social consensus on controversial bioethical questions\". | Cuts against | Abstract only (PubMed, PMID 2010719) | pubmed.ncbi.nlm.nih.gov/2010719/ |\n| Prothro and Grigg, Journal of Politics 22 | 1960 | Voters in Ann Arbor and Tallahassee agreed 94.7-98.0% on abstract democratic principles. On the ten applied propositions, none reached 90%. \"consensus breaks down completely\" when principles become specific propositions. | Cuts against; same respondents at both levels | Full text | fbaum.unc.edu/teaching/POLI195_Fall09/Prothro_Grigg_JOP_1960.pdf |\n| Racial principle-policy gap paper, GSS white respondents (cites Schuman, Steeh and Bobo) | c. 2024 | Defines the gap as support for a principle without support for the program that implements it. Reports it persists in the General Social Survey, 1994-2018. | Cuts against | Abstract and introduction. Schuman's book and Dixon 2017 were not opened. | nsp.lse.ac.uk/articles/137/files/668df8e4b4a37.pdf |\n| Dryzek and Niemeyer, AJPS 50 | 2006 | Value pluralism can coexist with meta-consensus. They cite MacIntyre that ethicists who disagree on principles often agree in applied ethics, but note a shared liberal tradition. Citizens' jury: after deliberation 7 of 12 favoured closing the track; at the start no option was first choice of more than four. \"at the abstract level there is a high degree of agreement on the legitimacy of basic values\" | Both | Full text, selected passages. Rokeach and MacIntyre are cited, not checked. | ask-force.org/web/Discourse/Dryzek-Reconciling-Pluralism-Consensus-2006.pdf |\n| Niemeyer, Veri, Dryzek and Bächtiger, APSR 118 | 2024 | The Deliberative Reason Index measures, for pairs of deliberators, whether agreement on considerations tracks agreement on preferences. It rose in most of 19 forums, and group-building mattered most. \"We reiterate that DRI is not a measure of consensus.\" | Neutral: reasons-to-outcome alignment, not order | Open-access full text, key passages | cambridge.org/core/services/aop-cambridge-core/content/view/6558F69855ADA8B15BF2EC2E5D403E71/S0003055423000023a.pdf/how_deliberation_happens_enabling_deliberative_reason.pdf |\n| Awad et al., PNAS (42 countries, 70,000 people) | 2020 | The three sacrificial dilemmas had the same ordering of acceptability in every country, but the levels varied substantially by country. \"In every country, the three dilemmas displayed the same qualitative ordering of sacrifice acceptability\" | Both | Abstract only (Europe PMC) | europepmc.org/article/PMC/PMC7007553 |\n| Bago et al., Nature Human Behaviour (45 countries) | 2022 | The personal-force effect appeared in all cultural clusters. The intention-by-force interaction was inconclusive in the Eastern and Southern clusters. It measured case features only, not principles. | Weakly supports; mostly neutral | Abstract only (Europe PMC) | doi.org/10.1038/s41562-022-01319-5 |\n| Nichols and Knobe, Noûs 41 | 2007 | When asked abstractly, 86% said responsibility is impossible in a deterministic universe. When given a concrete wrongdoing, 72% said the agent was fully responsible. | Cuts against: levels can diverge | Secondary: a critique page reports the figures; the Wiley abstract returned 403. Between- or within-subject design not verified. | cah.ucf.edu/fpr/article/explaining-compatibilist-intuitions-about-moral-responsibility-a-critique-of-nichols-and-knobes-performance-error-model/ |\n| Royzman et al. 2015 (JDM); McHugh et al. 2018 preprint and 2023 | 2015-23 | Royzman: dumbfounding estimates approach zero once beliefs about harm are accounted for. McHugh: it persists under stricter criteria and appears in Chinese, Indian and MENA samples. All of this concerns single individuals, not agreement between disagreers. | Both, weakly relevant | Abstracts only (RePEc; Europe PMC). The McHugh \"Reasons or rationalisations\" abstract is the 2018 preprint. | ideas.repec.org/a/cup/judgdm/v10y2015i4p296-313_2.html |\n\n**Verdict.** Agreement comes more easily on cases or outcomes when a small group must reach a decision, shares a tradition or institution, and can settle for low-level, theory-neutral reasons. That is Sunstein's third kind, the Belmont story, and Dryzek and Niemeyer's MacIntyre caveat. Arras and Beauchamp add that this depends on shared background norms. The reverse is documented in mass publics. Prothro and Grigg and the principle-policy literature show near-unanimity on abstractions and splits on application, and Sunstein himself describes that pattern first.\n\nThe wording gets four things wrong:\n- It states the pattern as general.\n- It asserts an order (\"comes first\", \"afterwards\"). No source measured temporal order. Sunstein claims a level of generality, not a sequence.\n- Jonsen's 1998 account has principles in the mandate and in the pre-read literature.\n- It treats \"case\" and \"principle\" agreement as independent. Nichols and Knobe show judgments shifting with abstraction level.\n\n**Leads that did not check out, or could not be reached**\n- Sunstein's primary text: not reached (403). The three-kinds passage comes from Solum's quotation.\n- Beauchamp's own dispute with Toulmin over the commission: not verified. SEP reports no such rebuttal. A search-tool snippet suggesting commissioners disagreed about deduction was never confirmed on a page.\n- Not opened: Jonsen and Toulmin's book, Toulmin 1981, McClosky 1964, Schuman's book, Dixon 2017, Sinnott-Armstrong, Haidt's original dumbfounding paper, and the full texts of Royzman and McHugh.\n- Levi and Quine, the original two other sources, were not checked. They were not in your leads.\n- Lead 10 is only partly met. Prothro and Grigg, the GSS gap and Niemeyer measure both levels in the same people, but none tests which comes first.\n"}