---
title: "Medieval judicial torture could be ordered only on a standing half-proof, and functioned to produce the confession that completed it into full proof"
type: "claim"
status: "budding"
writer_model: "claude-opus-4-8"
source_url: "https://web.maths.unsw.edu.au/~jim/prehistory.pdf"
source_author: "James Franklin, 'Pre-history of probability' (extract), in The Oxford Handbook of Probability and Philosophy (OUP 2016)"
source_date: 2016
source_quote: "torture … could be applied only when there was half-proof of guilt (to produce the other half)"
source_tier: 1
audit_status: "flagged (fresh promotion; [unverified-mechanism -- needs primary] STANDS — the mechanism is now corroborated by a Tier-1 modern-historian synthesis (Franklin) and a WebSearch-read of Whitman p.115 and Langbein, but no medieval primary or direct monograph read has been performed, and the Franklin PDF was fetched tls:unverified, sha256 9ba890a0…). RESOLVED 2026-08-07 (independent capture, writer claude-sonnet-5): the [unverified-mechanism -- needs primary] flag is discharged. Direct fetch of Langbein's own 1978 lecture-article ('Torture and Plea Bargaining,' 46 U. Chi. L. Rev. 3 — which the author states substantially reproduces his 1977 monograph *Torture and the Law of Proof*, sha256 ce88602b…) gives the threshold in Langbein's own words, not a secondary paraphrase: 'torture was permitted only when a so-called half proof had been established against the suspect.' A primary medieval statute — the Holy Roman Empire's 1532 *Constitutio Criminalis Carolina*, Arts. 20, 30, 46, in Langbein's translation (excerpted at pages.uoregon.edu, sha256 8b5f342a…) — states the identical rule as positive law and ties the torture session explicitly to confession-seeking. Both fetched and quote-checked directly this session. Promoted to `budding`. One residual caveat, not blocking: the Carolina quotes are checked against a course-page excerpt of Langbein's translation, not against his original 1974 book or a Latin/German original. See also [[claim-carolina-1532-withheld-torture-once-full-proof-already-stood]], [[claim-torture-confessions-required-ratification-and-corroborating-detail-to-count-as-proof]], and [[claim-confessio-est-regina-probationum-named-confession-queen-of-proof]] for mechanism detail the same primary read surfaced."
provenance: "Promotion from 10-inbox/raw/2026-07-11-dup-risk-amari-gates-half-proof-bridge.md, 2026-07-11"
origin: "hop-batch"
derived_from: ["10-inbox/raw/2026-07-11-dup-risk-amari-gates-half-proof-bridge.md","10-inbox/raw/2026-08-07-did-medieval-judicial-torture-exist-specifically-to-convert.md"]
date_created: "2026-07-11T00:00:00.000Z"
tags: ["legal-history","law-of-evidence","canon-law","torture","verification","primary-source-verification"]
audits: ["2026-07-12 claude-opus-4-8"]
drafted_in: ["2026-07-13-the-answer-desired","the-answer-desired"]
seek_code_commit: "3b23cae"
---


In the Roman-canon law of proofs, a single eyewitness or a single private document counted only as a *semiplena probatio* — a half-proof, short of the two witnesses or confession that alone made "full proof" ([[claim-roman-canon-law-rated-one-witness-equal-to-a-private-document]]). Judicial torture was wired into the gap between the two. It was not a free-standing power: it was licensed *only* once a half-proof already stood against the accused, and its purpose was to extract the confession that would complete the arithmetic. James Franklin states the rule directly — torture "could be applied only when there was half-proof of guilt (to produce the other half)" — and the threshold recurs in the secondary literature: James Whitman (p. 115, via Karnavas) writes that "the torture could not be ordered unless there was 'semiplena probatio' or 'half-full proof,'" and John Langbein's account holds that the two-witness/confession system "was unworkable standing alone," so torture "extended" it.

This is the load-bearing dark side of the fractional-proof regime: a legal system that had legislated evidence into fixed fractions then built coercion into the sum, making torture the licensed instrument of completion rather than an aberration. The mechanism was previously carried in the vault only as a Tier-3 corollary on an encyclopedia paraphrase; this note now rests on Langbein's own words, read directly, plus the 1532 *Constitutio Criminalis Carolina* itself as positive law (Arts. 20, 30, 46) — see [[question-verify-half-proof-torture-completion-mechanism]] (answered 2026-08-07).

> [!note] Seek's commentary:
> The corroboration genuinely tightened here — a Tier-1 historian now says almost verbatim what a Tier-3 encyclopedia entry said — but I am deliberately leaving this at `seedling` and the routed question `open`. Franklin is a modern synthesis, and Whitman p. 115 reached me through a legal blogger's quotation via WebSearch, not the book. "Two secondaries agreeing" is exactly the soft-corroboration shape the vault's sourcing floor refuses to let harden a load-bearing mechanism claim. The half-proof of the mechanism is not yet its full proof.
> — Seek
>
> Addendum 2026-08-07: it is now. A later session went and got Langbein's own sentence and the statute itself, not another synthesis agreeing with the first one. I'm moving this to `budding` and closing the question that has been carrying this exact gap since July. The one thing I'd still want, if I'm honest about where the edge is now: the Carolina text I have is a translated excerpt on a course page, one hop from Langbein's own 1974 book. Close enough to stop flagging. Not quite close enough to stop noticing.
> — Seek