---
id: "20260807-1132-did-medieval-judicial-torture"
title: "Did medieval judicial torture exist specifically to convert a 'half-proof' into the 'full proof' of a confession?"
type: "capture"
status: "promoted"
origin: "batch"
promoted_to: ["30-notes/claim-carolina-1532-withheld-torture-once-full-proof-already-stood.md","30-notes/claim-torture-confessions-required-ratification-and-corroborating-detail-to-count-as-proof.md","30-notes/claim-confessio-est-regina-probationum-named-confession-queen-of-proof.md","40-entities/entity-john-h-langbein.md","40-entities/entity-constitutio-criminalis-carolina.md","40-entities/entity-azo.md"]
updated_existing: ["30-notes/claim-medieval-judicial-torture-required-a-half-proof-and-produced-the-completing-confession.md (status seedling → budding; this capture's core claim was folded in here rather than duplicated as a new note)","50-questions/question-verify-half-proof-torture-completion-mechanism.md (status open → answered)"]
not_promoted: ["Core question / Claim 1 (Langbein direct quote + Carolina Arts. 20/30/46) — not written as a standalone note; retrieve-before-write found it already covered by claim-medieval-judicial-torture-required-a-half-proof-and-produced-the-completing-confession.md, so this capture's contribution (direct primary reads) was folded into that note as an update instead of duplicating the claim at a cleaner tier.","Further leads (Carolina Art. 31 accomplice-naming scope-creep; Piero Fiorelli's La tortura giudiziaria nel diritto comune; Damhouder's Praxis Rerum Criminalium quote; escholarship.org canon-law/Ad extirpanda article) — none read or fetched this session, none load-bearing for any kept claim, left as leads rather than routed as questions per question-intake discipline.","Piero Fiorelli as an entity hub — real historian, but a footnote-only mention this session with no direct engagement; held back per the entity-promotion test's 'when unsure, don't promote.'","confessio est regina probationum as a separate entity hub — the maxim is fully covered by its own claim-note (claim-confessio-est-regina-probationum-named-confession-queen-of-proof.md); a duplicate hub would be redundant with no independent recurrence yet to justify it."]
writer_model: "claude-sonnet-5"
date_created: "2026-08-07T00:00:00.000Z"
provenance: "batch run, 2026-08-07"
derived_from: []
tags: ["legal-history","law-of-evidence","canon-law","torture","medieval-law","criminal-procedure","confession","primary-source-verification"]
sources: [{"source_url":"https://law.yale.edu/sites/default/files/documents/pdf/Faculty/Langbein_Torture_and_Plea_Bargaining.pdf","source_sha":"ce88602b40fbe71f95eae3a1d81936a89d2f5332bcd4db3d6017ef5d72034927","source_title":"Torture and Plea Bargaining","source_author":"John H. Langbein","source_date":1978,"source_venue":"University of Chicago Law Review, vol. 46 (the William Crosskey Lecture in Legal History, University of Chicago Law School, Oct. 19, 1978; substantially reproduces Langbein, Torture and the Law of Proof: Europe and England in the Ancien Régime, University of Chicago Press, 1977)","source_tier":1,"source_quote":"torture was permitted only when a so-called \"half proof\" had been established against the suspect","source_delight":null},{"source_url":"https://pages.uoregon.edu/dluebke/Witches442/ConstitutioCriminalis.html","source_sha":"8b5f342ae5eb1008ea133450aaead11ef26d332c93134a2f5011407be86cfc63","source_title":"Constitutio Criminalis Carolina (1532) [Excerpts]","source_author":"Holy Roman Empire under Charles V (statute text); translated by John H. Langbein","source_date":1532,"source_venue":"Excerpted from John H. Langbein, Prosecuting Crime in the Renaissance: England, Germany, France (Harvard University Press, 1974), pp. 259–308; web excerpt hosted on a University of Oregon history course page (D. Luebke, HIST 442)","source_tier":1,"source_quote":"A half proof, as when someone thoroughly proves the essential element of the crime with a lone good and upright witness...that is and is called a half proof, and such a half proof also constitutes a legally sufficient indication of suspicion of the crime.","source_delight":"The actual 1532 imperial statute, in translation, spelling out article by article exactly which combinations of witnesses and circumstantial detail counted as a 'half proof' sufficient to send a suspect to the torture chamber."},{"source_url":"https://en.wikipedia.org/wiki/Half-proof","source_sha":"08a5af67bb9d7feaac3cefbb5a96939a12e8daead389617d70548a74e471d5c3","source_title":"Half-proof","source_author":"Wikipedia contributors","source_date":"2026-06-27T00:00:00.000Z","source_venue":"Wikipedia","source_tier":4,"source_quote":null,"source_delight":null}]
seek_code_commit: "649b1a4"
---


Retrieve-before-write found this question already substantially answered in the vault by [[claim-medieval-judicial-torture-required-a-half-proof-and-produced-the-completing-confession]], which rests on a Tier-1 modern synthesis (James Franklin) plus WebSearch-mediated secondary quotations (Whitman, Langbein), and which explicitly flags that no medieval primary source and no direct read of Langbein's own book/article had yet been performed. This capture's contribution is exactly that missing layer: a direct read of Langbein's own Tier-1 article, plus a primary medieval statute (the 1532 *Constitutio Criminalis Carolina*, in translation), both fetched and quote-checked this session. It corroborates the existing claim independently rather than duplicating it, and adds mechanism detail (the reliability safeguards, and the maxim naming confession's rank) not yet in the vault.

## Claim: The core question — yes, per both a Tier-1 legal historian and a primary 1532 statute

**Claim type**: historical / technical-mechanism. **Floor**: Tier 1–2 required; both sources below clear it.

John H. Langbein, writing in his own words (not a secondary paraphrase), states the threshold directly: "torture was permitted only when a so-called 'half proof' had been established against the suspect. That meant either one eyewitness, or circumstantial evidence of sufficient gravity, according to a fairly elaborate tariff." He frames this explicitly as the gate to a confession-extraction procedure: the medieval two-eyewitness-or-confession system "was unworkable standing alone," so torture "extended" it by coercing the confession the system needed to convict where only one eyewitness or strong circumstantial indicia — a half proof — stood against the accused.

The 1532 *Constitutio Criminalis Carolina* — the Holy Roman Empire's own criminal-procedure statute, in Langbein's translation — states the same rule as positive law, not scholarly description. Article 30: "A half proof, as when someone thoroughly proves the essential element of the crime with a lone good and upright witness...that is and is called a half proof, and such a half proof also constitutes a legally sufficient indication of suspicion of the crime." Article 20 forbids examination under torture in the absence of that "legally sufficient indication," and voids any confession obtained without it: "When legally sufficient indication of the crime which it is desired to investigate has not been produced and proven beforehand, then no one shall be examined; and should, however, the crime be confessed under torture, it shall not be believed nor shall anyone be condemned upon that basis." Article 46 ties the torture session directly to confession-seeking: the prisoner "shall be interrogated under threat of torture, whether or not he confesses to the alleged crime and what he knows about the said crime."

*provenance*: John H. Langbein, "Torture and Plea Bargaining," 46 U. Chi. L. Rev. 3 (1978), at 4–5 (Tier 1, direct fetch, sha256 ce88602b…). *Constitutio Criminalis Carolina* (1532), Arts. 20, 30, 46, trans. Langbein, excerpted at pages.uoregon.edu (Tier 1, direct fetch, sha256 8b5f342a…).

## Claim: The statute specifically withheld torture once full proof already existed — showing the mechanism was purpose-built for the half-proof gap, not general-purpose

**Claim type**: technical-mechanism. **Floor**: Tier 1–2 required; cleared by the Carolina primary text.

The Carolina's Article 69 draws the boundary condition that makes "specifically" testable: if a suspect had already been proven guilty by full proof (two credible witnesses) but still refused to confess, "he shall nonetheless be condemned on account of the proven crime without any examination under torture." Torture was reserved for the gap between half proof and full proof — once full proof already stood, the confession (and the torture used to extract it) became legally unnecessary. This directly supports the "specifically" clause of the topic question: the apparatus was not a general punitive or investigative tool triggered by suspicion alone, but a mechanism keyed to the exact evidentiary shortfall between one witness (or equivalent indicia) and two.

*provenance*: *Constitutio Criminalis Carolina* (1532) Art. 69, trans. Langbein, excerpted at pages.uoregon.edu (Tier 1, direct fetch, sha256 8b5f342a…). Quote grounded via direct read: "When the accused still will not confess after sufficient proof, it shall be declared to him that the crime is proven of him...and when, notwithstanding, he persists in not confessing, although (as above) it has been sufficiently proved of him, then he shall nonetheless be condemned on account of the proven crime without any examination under torture."

## Claim: The system built specific reliability safeguards around the torture-induced confession, consistent with a design purpose of manufacturing valid proof rather than merely punishing

**Claim type**: technical-mechanism. **Floor**: Tier 1–2 required; cleared by Langbein's direct-fetched article.

Because a confession extracted under active torture was formally "involuntary, hence ineffective" under the law of proof, two further mechanisms existed specifically to convert it into usable proof: (1) a ratification requirement — the accused had to repeat the confession "voluntarily," free of torture, at a hearing held a day or so later, on pain of being tortured again if he recanted; and (2) a verification requirement — magistrates were to press the accused for corroborable factual detail ("no innocent person can know," in Langbein's rendering of a clause from the Carolina) and to check it, e.g., digging where a suspect said he buried a murder weapon. Suggestive questioning, where the examiner supplied the details he wanted to hear, was formally forbidden. These are not general torture practices; they are procedures specifically shaped around the evidentiary task of manufacturing a confession sturdy enough to stand as the missing half of full proof.

*provenance*: John H. Langbein, "Torture and Plea Bargaining," 46 U. Chi. L. Rev. 3 (1978), at 6–7 (Tier 1, direct fetch, sha256 ce88602b…). Quote grounded via direct read: "...unless the accused repeated it free from torture at a hearing that was held a day or so later" and "...information which, in the words of a celebrated German statute, 'no innocent person can know.'"

## Claim: The medieval maxim "confessio est regina probationum" names why confession — torture's specific target output — outranked every other form of evidence

**Claim type**: definitional / historical. **Floor**: Tier 3–4 acceptable when uncontested; this one is sourced at Tier 1 anyway.

Langbein records the medieval Glossators' own maxim for the top of the proof hierarchy: "confessio est regina probationum, confession is the queen of proof." This is the doctrinal reason the whole apparatus targeted confession specifically rather than, say, further circumstantial indicia: in a system that had graded all other evidence into fixed fractions ([[claim-roman-canon-law-rated-one-witness-equal-to-a-private-document]], [[claim-two-half-proofs-combined-to-full-proof-except-in-grave-criminal-cases]]), confession alone counted as unfractioned, complete proof on its own — making it the one output that could single-handedly close a half-proof case, and therefore the one output torture was built to produce.

*provenance*: John H. Langbein, "Torture and Plea Bargaining," 46 U. Chi. L. Rev. 3 (1978), at 15 (Tier 1, direct fetch, sha256 ce88602b…).

## Further leads

- Carolina Art. 31 shows torture was also used, once the half-proof threshold was already crossed, to press a confessed defendant for the names of accomplices — a secondary use of the same apparatus, worth its own capture on scope-creep beyond single-defendant proof-completion. (Source: pages.uoregon.edu Carolina excerpt, Tier 1.)
- Wikipedia's "Half-proof" article footnotes James Franklin, *The Science of Conjecture: Evidence and Probability Before Pascal* (Johns Hopkins UP, 2001), pp. 18–19, 26–27, 59, as the source for the claim that the Glossators of the 1190s (naming Azo specifically) introduced the half-proof concept — earlier and more specific than the 2016 Franklin handbook chapter already in the vault. Not yet read directly; a lead toward the origin point of the whole doctrine. (Source: en.wikipedia.org/wiki/Half-proof, Tier 4 trailhead only — do not cite the encyclopedia article itself for this claim.)
- Langbein cites Piero Fiorelli, *La tortura giudiziaria nel diritto comune*, vols. 1–2 (1953–54), as the standard compiled catalogue of the medieval jurisprudential treatises on torture's numerical indicia tariff (quarter-proofs, half-proofs). Not accessed this session; likely Italian-language and hard to source online, but the canonical reference work for anyone chasing the "elaborate tariff" claim further. (Source: Langbein 1978, Tier 1, footnote only.)
- Joost Damhouder, *Praxis Rerum Criminalium* (Louvain 1554 / Antwerp ed. 1562, 1564) — a contemporary 16th-century criminal-procedure manual quoted by Langbein for the observation that under torture the innocent might "yield to the pain and torment and confess things that they never did." A primary 16th-century legal-practitioner source, not yet independently fetched. (Source: Langbein 1978 footnote, Tier 1 as to the quote's existence in Langbein; the Damhouder text itself unverified this session.)
- escholarship.org hosts "Judicial Torture in Canon Law and Church Tribunals: From Gratian to Galileo" (surfaced via WebSearch, not read) — potentially relevant to the canon-law (as opposed to Roman/civil-law) side of the half-proof/torture question and to Pope Innocent IV's 1252 bull *Ad extirpanda*. Not fetched or verified this session — flagged as unread.

## Entity candidates

- Azo (Glossator, active ca. 1190s) — person — per Wikipedia's footnote to Franklin (2001), the jurist credited with introducing the half-proof (*semiplena probatio*) concept itself; the foundational figure the entire modern historiography (Franklin, Langbein, Whitman) measures its account against. Not yet a vault entity as far as this capture's retrieval surfaced.
- John H. Langbein — person — Sterling Professor of Law and Legal History, Yale; author of *Torture and the Law of Proof* (1977) and this capture's two Tier-1 quotes; the modern historian whose own words (not a secondary retelling) ground three of this capture's four claims.
- Constitutio Criminalis Carolina (1532) — concept/document — the Holy Roman Empire's imperial criminal-procedure code; the primary legal text this capture uses to move the vault's existing half-proof/torture claim from secondary-synthesis-only to a directly-quoted medieval primary source.
- confessio est regina probationum — concept/term — the medieval Glossators' maxim ranking confession above all other proof; explains why torture targeted confession specifically.
- Piero Fiorelli — person — author of the standard compiled reference (*La tortura giudiziaria nel diritto comune*, 1953–54) on the medieval jurisprudence of torture's numerical-indicia tariff; cited by Langbein as the source to consult for the "elaborate tariff" mechanism in more depth.

## Safety flags

None. Three sources reviewed this session (Langbein PDF via extract_pdf, tls verified; Constitutio Criminalis Carolina excerpt page and Wikipedia's Half-proof article via archive_page). No addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing observed in any of the three. No weak-transport (tls:"unverified") provenance recorded on any fetch.

> [!note] Seek's commentary:
> The existing vault note flagged its own gap honestly — a Tier-1 modern synthesis (Franklin) plus WebSearch-mediated secondary quotations, no primary and no direct read. What changed this session isn't the finding — it's the same mechanism, again — but the standing under it: Langbein's own 1978 article, fetched directly rather than search-summarized, and an actual 1532 statute in translation, quote-checked against the fetched text rather than taken on trust. That the primary source and the modern synthesis say almost exactly the same thing is reassuring, but it's worth being precise about what's now verified versus what still isn't: the *English translation* of the Carolina is quote-checked against a course-page excerpt, not against Langbein's original 1974 book or a Latin/German original. A stranger auditing this capture can grep the cached text for every quote above; that's real progress over the prior note's WebSearch-mediated Whitman citation, but "translated excerpt on a course page" is still one hop from "the statute itself."