Ingest: Foundation and History of the PDSA Cycle

Type: kb/sources/types/ingest-report.md

Source: foundation-and-history-of-the-pdsa-cycle.md Captured: 2026-07-22 From: https://deming.org/wp-content/uploads/2020/06/PDSA_History_Ron_Moen.pdf

Classification

Genre: scientific-paper -- an ANQ conference proceedings paper with a full 18-item citation apparatus (books, journal articles, archival transcripts, personal letters). It sits at the documentary-history end of the genre rather than the empirical end: no methodology or data, and its evidential weight comes from participant testimony and primary-source citation, not experiment. The limitations lens below is read accordingly. Domains: pdsa, scientific-method, discovery-lifecycle, quality-improvement Author: Ronald Moen (Associates in Process Improvement). High access, high interest: a Deming "helper" at ~70 four-day seminars (1982-1993), a direct correspondent holding personal letters from Deming, and co-author of the Moen/Nolan/Provost line the paper cites approvingly. Authoritative on the lineage and an advocate for it.

Summary

The paper traces the ancestry of W. Edwards Deming's Plan-Do-Study-Act (PDSA) cycle from Galileo and Bacon through American pragmatism (Peirce, James, Dewey, C.I. Lewis) to Shewhart's 1939 specification-production-inspection circle, Deming's 1950 "Deming Wheel," the Japanese recasting into Plan-Do-Check-Act (PDCA) around 1951, and Deming's own PDSA formulations of 1986 and 1993. Its through-line argument is that PDSA formalizes the scientific method as an iterative learn-and-improve loop, and that PDCA and PDSA are related only through that shared method — Deming himself, in cited letters and transcripts, rejected PDCA as a "corruption" and disowned its authorship. The load-bearing addition for this KB is the 1991 Moen/Nolan/Provost move: the Plan step must state a prediction and its associated theory, and the Study step compares observed data against that prediction, supplying "the deductive/inductive interplay necessary for learning."

Connections Found

Casebook notes do exist for this material, and the source's settled role is corroborating evidence, not new theory. It is an independent, applied instance of the same conjecture-consequence-test learning core that discovery-lifecycle adapts from Peirce's abduction/deduction/induction — that note currently cites only the Peirce SEP entry as support for its core, so this snapshot is a second, non-philosophical tradition (Deming-lineage quality improvement) landing on the same loop. It also corroborates first-principles-reasoning-selects-for-explanatory-reach-over: the source's "one improving test is not enough; build theory that predicts improvement under future conditions" argument mirrors that note's reach-over-adaptive-fit polarity. Secondarily, it is one of a three-snapshot Moen/Norman cluster (with evolution-of-the-pdca-cycle and circling-back-clearing-up-myths-about-the-deming-cycle) that covers the same PDCA→PDSA lineage at different depths and is best cross-linked once all three are analysed.

Extractable Value

  1. Independent-tradition corroboration of the discovery-lifecycle core. A Deming-lineage improvement cycle deliberately builds prediction+theory into Plan and observation-versus-prediction into Study — the same abduction/deduction/induction interplay the lifecycle note adapts from Peirce. This is the highest-reach item: convergence of two unrelated traditions is evidence the learning core is real and general, not an artifact of one philosophy. [experiment]
  2. Reach-over-adaptive-fit, restated from improvement science. "It is not enough to determine that a change resulted in improvement during a particular test... you will need to be able to predict whether a change will result in improvement under the different conditions you will face in the future" is near-verbatim support for the first-principles note's explanatory-reach claim. [quick-win]
  3. The Check→Study shift as a gate-vs-critique datapoint. Deming rejected "check" because it "means to hold back" and renamed the step Study — compare observed data to prediction as a basis for learning. This distinguishes a pass/fail compliance check (PDCA, "implementation and compliance") from a learning step (PDSA, "testing and learning"), which maps onto the review system's verdict/gate versus report/critique distinction and is reusable framing for why a learning step is not a pass/fail gate. [just-a-reference]
  4. A convergence synthesis worth a note. Peircean philosophy of science and Deming-lineage quality improvement independently arriving at predict-then-test-against-prediction is a higher-order claim not yet captured; it would argue the lifecycle's reach from independent convergence rather than a single source. Flagged, not written. [deep-dive]

Limitations (our opinion)

This is editorial opinion. Read as participant-historian advocacy, not neutral history. The author is a Deming insider with a stake in the thesis: the paper is structured to establish PDSA as the correct lineage and PDCA as a "corruption," and much of its strongest evidence is Deming's own interested testimony (personal letters to the author, a GAO transcript, archival remarks). The central claim that PDSA "formalizes the scientific method" and supplies "the deductive/inductive interplay" is asserted by mapping cycle steps onto abduction/deduction/induction — an analogy, never tested, and vulnerable to the "naming something is not explaining it" failure. The account of the Japanese PDCA's origins is secondhand and admittedly unresolved ("no one person claims authorship... I found no evidence to dispute Imai's translation"). None of this undercuts the corroboration value above — an independent tradition converging on the same loop is useful precisely because the author was not reading our KB's Peircean framing — but the paper should not be cited as an empirical test of the learning core, only as documentary evidence that the core recurs.

Update discovery-lifecycle to add an evidenced-by link pointing at this snapshot, corroborating the conjecture-consequence-test core from the Deming-lineage tradition — the note's Relevant Notes currently rest that core on the Peirce SEP entry alone, and this is its first independent, applied support.