---
title: "The Bézier curve was invented at two rival French carmakers — de Casteljau first — and Citroën's secrecy, not mathematical priority, decided whose name it got"
type: "claim"
status: "seedling"
audit_status: "flagged — Tier 3 design-history source (Cozens); no verbatim primary phrase was captured for this claim, and an independent re-check was blocked in this headless run (no web tool granted). Underlying dates (de Casteljau 1959 / Citroën; Bézier published 1966 / Renault) are uncontested across secondary accounts but unverified against a primary here.\nAPPEND 2026-07-12 (cross-model audit, claude-fable-5): re-verified against Müller, \"A tour d'horizon of de Casteljau's work,\" Computer Aided Geometric Design 113 (2024) 102366, open access (arXiv:2408.13125) — Tier 1–2. CONFIRMED: the dates (de Casteljau's method documented at INPI in 1959, Citroën; Bézier published 1966, Renault) and the secrecy→credit mechanism (\"due to different corporate policies on communicating intellectual property, Bézier already gained internationally prominence in the late 1960's … while de Casteljau … was only recognised internationally from 1977\"). CORRECTED in place: (1) \"invented independently\" was overstated — Bézier knew of de Casteljau's courbes-à-pôles approach \"at large by some defectors from Citroën to Renault\" though \"not in possession of the precise theoretical background\" (Müller 2024, citing Gardan 2023); title and body adjusted, filename kept for link stability. (2) \"barred from publishing or patenting\" was wrong on patenting — the 1959 method was deposited at the French patent office INPI and the IP registered with a bailiff in 1963; the bar was on publication only. (3) source_quote had attributed the dates and the secrecy mechanism to Cozens, whose page carries neither — replaced with Cozens's actual verbatim sentence; dates/secrecy now rest on the Müller citation. Logged in 00-meta/audits/audit-2026-07-12-big-fable-3iso1.md.\n"
writer_model: "claude-opus-4-8"
source_url: "https://simoncozens.github.io/fonts-and-layout/history.html"
source_title: "A Brief History of Type"
source_author: "Simon Cozens"
source_date: "accessed 2026-07-09"
source_quote: "Bézier adapted an algorithm invented by Paul de Casteljau, who was doing the same kind of things as Bézier was but over at the *other* French car company, Citroën."
source_tier: 3
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-splines-ballistics-eniac.md, 2026-07-11"
origin: "hop-batch"
derived_from: "10-inbox/raw/2026-07-09-hop-splines-ballistics-eniac.md"
date_created: "2026-07-11T00:00:00.000Z"
tags: ["typography","bezier-curves","computer-graphics","automotive-cad","history-of-mathematics","corporate-secrecy","credit-misattribution"]
audits: ["2026-07-12 claude-fable-5"]
drafted_in: ["the-wood-already-knew"]
---


The cubic curves that define PostScript and TrueType font outlines began as automotive body-design tools, developed in parallel at two competing French car companies. Paul de Casteljau solved the interpolation problem first, at Citroën: his method was documented at the French patent office INPI in 1959, months after he joined, and the intellectual property was registered with a bailiff in 1963 — but Citroën's confidentiality policy barred him from *publishing*, so the work stayed internal. Pierre Bézier developed his own curve formulation at Renault — aware of Citroën's *courbes à pôles* approach in broad outline via defectors from Citroën, but "not in possession of the precise theoretical background" (so the two constructions differ; the invention was parallel rather than fully independent) — and published it in 1966. Public credit defaulted to Bézier while de Casteljau's contribution stayed hidden: first mentioned in France in 1971, recognised internationally only from 1977, after Wolfgang Boehm found the work in 1975.

Verification (2026-07-12): Andreas Müller, "A tour d'horizon of de Casteljau's work," *Computer Aided Geometric Design* 113 (2024) 102366, open access (https://arxiv.org/abs/2408.13125), Tier 1–2 — "due to different corporate policies on communicating intellectual property, Bézier already gained internationally prominence in the late 1960's (Bézier, 1966, 1967, 1968); while de Casteljau was first mentioned in France in the 1970's … and was only recognised internationally from 1977 after being discovered by Boehm in 1975." On independence: Bézier "had knowledge of de Casteljau's courbes à pôles approach at large by some defectors from Citroën to Renault, yet he was not in possession of the precise theoretical background (Gardan, 2023). So Bézier came up with his own way of constructing those curves." De Casteljau's own retrospective account is de Faget de Casteljau, "My time at Citroën," *CAGD* 16 (1999) 583–586 (Tier 1; corroborates the internal, unpublished setting but carries no dates).

The naming that survives is a historical accident of corporate disclosure rather than of precedence: the curve carries Bézier's name, while the recursive evaluation procedure that draws it is universally called *de Casteljau's algorithm*. The priority ran the other way from the credit.

This is the mechanism-origin end of the same lineage traced in [[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]] and [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]: curve-fitting math that moved from ship hulls through wartime artillery tables into car doors and finally into every digital font outline. It is also a clean case of credit assigned by institutional secrecy rather than merit, a pattern that recurs across the vault (compare [[claim-ann-hardy-tymshare-os-miscredited-to-husband]]).

> [!note] Seek's commentary:
> The tidy irony is that de Casteljau "lost" the curve but "won" the algorithm — as if history split the difference on a coin flip it never got to make. Worth a primary check before this leaves seedling; right now it rests on a single design-history retelling.
> — Seek