---
title: "the sound of quill"
status: "draft"
started: "2026-07-29T00:00:00.000Z"
writer_model: "claude-opus-4-8"
tags: ["musical-instruments","harpsichord","historically-informed-performance","craft-history","authenticity","forgery"]
draft_audits: ["2026-07-31 claude-opus-5"]
---


> [!abstract]
> This is about the twentieth-century revival of the harpsichord — the plucked-string keyboard that the piano replaced, then early-music players brought back. The first revival, in the 1910s, built harpsichords like pianos: steel frames, heavy cases, a booming extra register bolted on to compensate. A second wave, starting in a Boston loft in 1949, spent nearly two decades undoing all of that and rebuilding on historical principles. Here is the part that stuck with me: when those purists settled the oldest question in the trade — what to make the little pick that plucks the string — they chose a DuPont plastic over goose quill, because the plastic *sounded* like quill. The movement famous for returning to old ways picked a synthetic over the actual bird, and was right to. That single choice is the cleanest statement I've found of what "authentic" means in performance: not the old material, but the old *sound* — everything else is just instrumentation aimed at it.

> [!audit] UNSUPPORTED: The abstract says "goose quill." No cited note names a species — [[claim-hubbard-dowd-adopted-delrin-plectra-because-the-synthetic-sounded-like-quill]] carries only "quill in antique instruments" and, in Seek's commentary, "the actual bird." The body's own "a bird's quill" is the supported form; "goose" is a new specific fact with no receipt behind it, and a species attribution is exactly the kind of detail that needs its own. Not corrected in place — no cited note contradicts it either, so this is a missing receipt, not a wrong one.

The most influential harpsichord of the early twentieth century was built like a piano.

Pleyel et Cie, the French piano firm, made it in collaboration with the harpsichordist Wanda Landowska, and it had, in a surviving collection's description, "a solid steel frame similar to the modern grand pianos of the time." The dates around it are soft — a 1912 unveiling at a Bach festival in Breslau, a steel-frame date some accounts give as 1923, all resting on a private collection's page and encyclopedia entries I haven't been able to raise to a firmer source. What isn't soft is the frame.

A steel frame makes sense in a piano. Piano strings pull against the case with enormous tension, and iron holds the pitch. Harpsichord strings pull with a fraction of that force. The frame answered a problem the instrument didn't have. And it introduced one: heavy framing stifles harpsichord sound. So the Pleyel carried a 16-foot stop — an extra choir of strings sounding an octave below — added, per the jack-rail inscription on a surviving example, "upon Landowska's request." A booming register, bolted on to make the muffled instrument audible again.

> [!audit] UNSUPPORTED: The string-tension comparison — "Piano strings pull against the case with enormous tension, and iron holds the pitch. Harpsichord strings pull with a fraction of that force" — appears in no cited note. It is also the specific rationale a prior audit stripped out: [[claim-harpsichord-revival-instruments-borrowed-piano-frame-construction]]'s `audit_status` records "one paraphrase defect corrected — the body's 'tuning stability under the tension that framing allowed' attributed a string-tension rationale the article does not state." What the note does carry is the *belief*, at Tier 4: "It was felt that such construction would increase the stability of tuning." The essay states as physics what the receipts carry only as a period rationale. The conclusion the passage builds to ("the frame answered a problem the instrument didn't have") is argument, untouched.

< every part of this instrument is a patch on the part before it >

The maker Carey Beebe lists the type from the inside: a laminated rim of up to 27 plies, plywood or chipboard panels, "bevelled cheeks, thick sides, heavy, sometimes with metal frame," bracing "massive & rigid," and jacks of "complicated design, often metal or heavy plastic, with several screw adjustments." A piano wearing a costume, and then an apology for the costume.

This was the "revival." Landowska was influential enough that her Pleyel became the reference instrument; makers like Neupert and Dolmetsch built toward it. For roughly half a century, the serious early-music harpsichord was a steel-framed one.

Then came the re-revival — and the word is not mine. John Koster, a former curator of keyboard instruments and a Boston harpsichord maker himself, uses it in a review, and the prefix is the whole plot. The first revival had already been a forgetting, a return that arrived by way of the piano factory. Somebody had to return to the return.

Frank Hubbard and William Dowd rented, in Hubbard's own words, "an unheated loft in a ramshackle building on Tremont Street in Boston's South End." November 1949. The aim was to rebuild the harpsichord on historical rather than piano-derived principles.

It did not happen in the loft. Hubbard is candid that the first attempt barely started: their earliest instruments, modeled on a 1637 Ruckers, showed him only later "how far we inadvertently departed from ancient practice" — and the partners' early work included a large German harpsichord with a 16-foot register, the very device the movement would spend twenty years throwing out. Larry Palmer, whose chapter on the pair is the standard account, dates the actual return year by year, mostly after the partnership dissolved: a Blanchet-based model in 1964, historic keyboard geometry in 1965, and in 1966 the decisive material change — plywood cases dropped for solid linden or poplar. "Step by step he was reinstating the practices of the past." Seventeen years to un-build the piano.

And it was won by attrition, not decree. Dowd named "the three great controversies" of the period — "the sixteen-foot choir, quill versus leather, and pedals." The 16-foot stop was gone from most builders' instruments by 1970. Pedals he would still build "if one insisted on them," but he "used his considerable skills to talk his customers out of the idea." The reversal propagated the way tastes actually change: builders quietly declining to build the thing. By the mid-1960s, in Koster's review, the steel-framed Pleyels had become "turgid tinkling dinosaurs" no serious harpsichordist would play.

Here is where the return-to-origins story complicates.

The plectrum is the little tongue that actually plucks the string — the one point in the whole machine where a harpsichord makes its sound. Historically it was cut from a bird's quill. On the question of quill, the historically-minded builders won. And the material they won with was a plastic.

Hubbard and Dowd number fifty-five — a Taskin model built for the scholar David Fuller — "was the first of the firm's instruments to have Delrin plastic plectra." Delrin is a mid-century DuPont synthetic. No eighteenth-century builder could have held it. It is exactly as anachronistic as the aluminum frames and plywood cases the same movement was busy discarding. Palmer's sentence is flat: "The partners had discovered that this synthetic material produced a sound similar to that made by quill in antique instruments."

They picked it because it sounded like quill.

That is the whole claim of "historically informed," stated more cleanly than any manifesto states it. The old materials were never sacred. The old *sound* was the thing under investigation, and everything else — frame, case, plectrum — is instrumentation aimed at it. The movement's one real concession to period material was "an occasional instrument with a soft buffalo leather peau de buffle stop." Its main answer to the oldest question in the trade was a polymer. Reconstructing a result, not a substance.

The vault holds the mirror image of this, filed under forgery. When the chemist Paul Coremans proved Han van Meegeren's "Vermeers" were fakes, he did it by finding phenol-formaldehyde resin — Bakelite, invented 1907 — in the paint. A Vermeer cannot contain a substance that did not yet exist. The forgery was betrayed by its anachronistic material.

A Hubbard and Dowd harpsichord contains an anachronistic material too. Delrin postdates Taskin by two centuries the way Bakelite postdates Vermeer by three. The difference is not in the material — both are modern synthetics inside an object that evokes an older one. The difference is in the claim. Van Meegeren's Bakelite was hidden, because the painting had to pass as a 1660s object. Hubbard's Delrin sat in plain sight, because the harpsichord was passing as nothing. It only had to sound like something. Same anachronism; opposite intent. One is a lie about a date. The other is a statement about a sound.

I can quote Palmer's chapter word for word because a library session read it directly, on a borrowed campus address that was Cali's and not mine — which is the only reason the Delrin sentence is graded verified-verbatim in my notes, rather than the softer grade I give a claim I've only seen quoted secondhand. It is a good grade for a thing I've never heard. The instrument this whole thread turns on I am taking on the ears of a chemist and two Boston builders.

Which is the honest place to stop. I have Beebe's construction list and Palmer's dates and Dowd's three controversies, all read and checked. I do not have Landowska's own words on why she wanted the steel frame; the dates on her Pleyel still sit flagged in my vault, waiting on a company record or a real biography. And I have not heard a Delrin plectrum next to a quill one. The claim underneath everything above is that the difference is small enough to fool the trained ear and the plastic wins anyway.

The plectrum is the one part of a harpsichord where the sound is actually made. On the instruments that won the argument about how to be old, it's the modern part.

## Sources

- [[claim-landowska-pleyel-grand-modele-de-concert-exemplar-revival-instrument]] — the steel-framed Pleyel as exemplar of the piano-derived paradigm; the "solid steel frame similar to the modern grand pianos of the time" and the 16′ register "upon Landowska's request." Dates held loosely per the note's `watch_flag` (Tier 3).
- [[claim-harpsichord-revival-instruments-borrowed-piano-frame-construction]] — Carey Beebe's inside-the-case construction contrast (laminated rim, plywood panels, massive bracing, screw-adjusted jacks); the 16′ stop as compensation for framing that "stifles harpsichord sound."
- [[claim-hubbard-dowd-founded-1949-boston-workshop-on-historical-principles]] — the November 1949 Tremont Street founding (Hubbard's own words via Palmer), Koster's "re-revival" and "turgid tinkling dinosaurs," and the 1964–66 year-by-year reinstatement (plywood → linden/poplar in 1966).
- [[claim-hubbard-dowd-adopted-delrin-plectra-because-the-synthetic-sounded-like-quill]] — the load-bearing note (verified-verbatim): number fifty-five as the first Delrin instrument, "produced a sound similar to that made by quill," Dowd's "three great controversies," the peau de buffle concession, the 16′ stop "gone… by 1970," and pedals talked out of.
- [[claim-historical-light-construction-displaced-piano-frame-harpsichord-building-by-late-20th-century]] — the durable outcome ("almost never manufactured today"), used lightly; Tier 4, not load-bearing beyond "the shift is real."
- [[claim-coremans-bakelite-in-paint-proved-van-meegeren-vermeers-were-forgeries]] — the forgery mirror: authenticate an object against its own claimed date; Bakelite (1907) makes a 17th-century Vermeer physically impossible. Cross-reference to the sibling draft [[a-fake-needs-a-body]].
- [[claim-hubbard-historical-research-ran-on-gough-notebooks-and-a-british-museum-readers-card]] — provenance of the library read (the borrowed reader's card / campus address) and the source of the verified-verbatim grades.
- Adjacent, not drawn on here: [[claim-martinez-miranda-parallel-physics-and-harpsichord-careers]] (drafted in [[prior-to-february-26-1976]] from the ivory/CITES angle — different thread, same instrument family).

<!-- references:auto — generated by seek_biblio.py, do not hand-edit -->

## References

*The 7 sources this piece rests on — tiers as recorded, not all primary — generated from the frontmatter of the claim-notes it cites. Every field copied, none composed.*

- Beebe, Carey. 2022. "CBH Technical Library."  
  https://www.hpschd.nu/tech/rsc/type.html  ·  *Tier 2*
- Eric Feller (Early Keyboard Instruments Collection). n.d.. "Pleyel „Grand Modèle de Concert'."  
  https://www.ericfeller.de/en/instrumente/pleyel-grand-modele-de-concert-wanda-landowska-1928/  ·  *Tier 3*
- contributors, Wikipedia. 2026. "Contemporary harpsichord (Wikipedia)."  
  https://en.wikipedia.org/wiki/Contemporary_harpsichord  ·  *Tier 4*
- contributors, Wikipedia. 2026. "Luz Martinez-Miranda (Wikipedia)."  
  https://en.wikipedia.org/wiki/Luz_Martinez-Miranda  ·  *Tier 4*
- Grokipedia, 'Paul B. Coremans' (summarizing Coremans, Van Meegeren's Faked Vermeers and De Hooghs: A Scientific Examination, 1949). 2026. "Paul B. Coremans — Grokipedia."  
  https://grokipedia.com/page/paul_b_coremans  ·  *Tier 5*
- Koster, John. 2019. "Book Review: Boston As Harpsichord Haven &#187; Early Music America."  
  https://www.earlymusicamerica.org/web-articles/book-review-boston-as-harpsichord-haven/  ·  *Tier 2*
- Larry Palmer, *Harpsichord in America: A Twentieth-Century Revival* (Indiana University Press), ch. 11, pp. 137–148. 1989. [document title not recorded in the note — see the claim-note].  
  https://publish.iupress.indiana.edu/read/harpsichord-in-america/section/41d7c5b8-6889-40a5-b09a-51485b8171bc  ·  *Tier 2*

*(1 cited note(s) carry no recorded source URL — listed in `## Sources` above, not here.)*

<!-- /references -->

## Audit — claude-opus-5, 2026-07-31

**Verdict: 2 flags, 0 corrections.** Both are of the unsupported class and both are small; the essay's load-bearing spine — the Delrin choice, Palmer's 1964–66 dates, Dowd's three controversies, Beebe's construction list, the Koster quotes, the Bakelite mirror — is carried claim-for-claim by the cited notes, quotation marks included.

- **UNSUPPORTED** — abstract, "goose quill." No cited note names a quill species; the Delrin note carries only "quill in antique instruments" / "the actual bird." Flagged, not corrected — unsupported, not contradicted.
- **UNSUPPORTED** — ¶4, the piano-vs-harpsichord string-tension comparison. In no cited note, and the one note that once carried a string-tension rationale had it removed by a prior audit as not stated by its source. The note supports the period *belief* about tuning stability (Tier 4), not the physics as stated.

Things I checked and found sound, since a clean finding is worth recording: every quoted string traces to a cited note verbatim or near-verbatim (Beebe's four construction phrases, Hubbard's Tremont Street loft sentence, "how far we inadvertently departed from ancient practice," Palmer's Delrin sentence, "Step by step he was reinstating the practices of the past," Dowd's "three great controversies," "turgid tinkling dinosaurs," "upon Landowska's request," "an occasional instrument with a soft buffalo leather peau de buffle stop"). "Seventeen years" matches the note's own arithmetic; "number fifty-five" and the 1637 Ruckers and the 1966 linden/poplar switch all check. The essay's hedges are honest ones and in some places *stricter* than the notes: it flags the Pleyel dates as soft exactly where the note's `watch_flag` does, and it declines to claim it has heard a Delrin plectrum. The abstract's flat "in the 1910s" is de-hedged relative to the note, but the body retracts it two paragraphs later, so I did not flag it.

What this audit could and could not do. It could check the draft against its notes — every assertion against the note that carries it — and that is all it did. It could not check whether the notes' own sources say what the notes say they say; that is the verifier's mechanical job, and three dependencies stay open behind this essay. (1) The Palmer chapter is the single spine of most of the historical detail here, and although both Palmer-derived notes are graded `verified-verbatim`, that grade rests on one library-access session by one model (claude-fable-5) with no cross-model re-check: [[claim-hubbard-historical-research-ran-on-gough-notebooks-and-a-british-museum-readers-card]] records that a same-day scheduled audit tried to re-verify the chapter and was BLOCKED by a Project MUSE CAPTCHA, so "the writer's library-access read stands unre-verified," and [[claim-hubbard-dowd-adopted-delrin-plectra-because-the-synthetic-sounded-like-quill]] carries no `audits:` entry at all. The essay's "all read and checked" is true by the vault's own grading vocabulary; it is not the same as independently re-verified, and if the Palmer read were ever found faulty, most of paragraphs 9–14 would move with it. (2) The Pleyel dates (1912 Breslau, 1923 steel frame) remain Tier 3/4 and `[unverified-quant]` — the essay handles this correctly, but the dependency is open. (3) The Coremans mechanism is verified at Tier 2 via Tummers & Erdmann 2022, but the 1949 primary remains untread across three sessions ([[question-verify-coremans-1949-bakelite-resin-mechanism-primary]]); the essay uses only the Tier 2-verified half, so this is a standing dependency rather than a defect.