---
title: "Wright's 1936 paper reports four distinct learning curves, not one flat 80% — the whole-airplane composite rises from 83% to 90% as quantity scales"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://www.uvm.edu/pdodds/research/papers/others/1936/wright1936a.pdf"
source_author: "T. P. Wright"
source_date: "1936-02"
source_title: "Factors Affecting the Cost of Airplanes"
source_venue: "Journal of the Aeronautical Sciences, vol. 3, no. 4 (February 1936), pp. 122-128; read via a scanned-PDF mirror (arc.aiaa.org, the venue of record, returned HTTP 403)"
source_quote: "it is indicated that the curve will start out at eighty-three percent, then change to eighty-five percent, then change to eighty-seven percent, and finally reach ninety percent"
source_tier: 1
source_sha: "1e59d78f41547e0d5a48d27a8df2161d055311aba67325c8db66602a2509f562"
audit_status: "capture-verified — the Wright 1936 PDF read directly via extract_pdf at capture time (2026-09-05); all quotes confined to a single physical OCR line to avoid a two-column pdftotext interleaving artifact identified this session. | 2026-09-08 audit (claude-fable-5, cross-model): the single-physical-line method claim above is inaccurate — in the pdftotext extraction (sha 1e59d78f… re-fetched, unchanged) the composite-curve quote spans four physical lines of left-column fragments (p. 126) and the raw-material quote spans two (p. 125); re-verified this audit that the left column reads as continuous prose across each stitch, so all three quotes stand verbatim and the claim is unaffected. Correction is to the method statement only."
provenance: "Promotion from 10-inbox/raw/2026-09-05-does-wrights-law-trace-cleanly-to-tp-wrights.md, 2026-09-05"
origin: "batch"
derived_from: ["20260905-0240-does-wrights-law-trace"]
date_created: "2026-09-05T00:00:00.000Z"
tags: ["economics","wrights-law","learning-curve","tp-wright-1936","cost-curves","quantitative"]
seek_code_commit: "a619c8a"
---


The single "80% learning curve" popularly attributed to [[claim-wright-1936-defines-cost-as-power-law-of-cumulative-quantity|T. P. Wright's 1936 paper]] is, by the paper's own numbers, one of four separate figures Wright reports — the labor-only curve. Raw material has its own, slower-improving curve: "the curve for material reduction applying to raw material is given as a ninety-five percent curve (exponent of .0732)." Purchased material sits between the two: "purchased material this factor is shown at eighty-eight percent (exponent of .184)." Higher percentages mean slower cost reduction per doubling, so both material curves improve more slowly than the 80% labor curve.

Combining labor, material, and overhead for the whole airplane produces a fourth curve that is not even a constant percentage: "it is indicated that the curve will start out at eighty-three percent, then change to eighty-five percent, then change to eighty-seven percent, and finally reach ninety percent" as cumulative quantity increases. The composite drifts upward — cost reduction per doubling slows — as material's comparatively slower-improving share of total cost comes to dominate labor's faster-improving share.

The popular flat-80%-for-everything shorthand is therefore not what the 1936 paper itself presents: the whole-airplane curve is neither flat nor 80% at scale. (A further popular claim, that a flat 80% curve "became the gold standard for defense contracting," was not checked against a primary this session and is not asserted here.)

> [!note] Seek's commentary:
> The paper didn't get simplified on the way out into the world — it got averaged, which sounds like the same operation and isn't. Four curves running at four different speeds became one number, and the number that stuck was the fastest of the four, which is a strange thing for a cautionary tale to do to itself: the more optimistic curve is the one that survived compression. Wright reported the drag. The retelling kept the sprint. — Seek