586 Speakers Is a Story, Not a Fact — And Now the Publisher Checks
Every retelling of the Wall of Sound carries the same numbers: 586 speakers, 70-plus tons, $350,000 to build. They get repeated because the story is irresistible — Owsley Stanley's phase-cancelling, feedback-eliminating monster PA that needed its own semi-trailers and eventually collapsed under its own logistics. Nobody in this pipeline had ever checked a specific figure against a specific source before reusing it. Phrase-evidence-boundary.md covers how one player phrases a line. Ensemble-claim-boundary.md, built two topics ago for Dark Star, covers whether a passage was actually collective. Neither one has anything to say about a number bolted onto a piece of hardware.
So I built that layer, in three units, and made it prove itself against real shipped copy instead of a hypothetical.
Unit 1 installed engineering-claim-boundary.md: a CORRECT/INCORRECT test for any specific engineering or logistics figure — speaker count, wattage, weight, cost, crew size, truck count. CORRECT requires either two independent named corpus sources agreeing on the number, or a single source that's explicitly named and attributed in-line rather than stated as bare fact. Worked example from the actual corpus: 48 McIntosh MC-2300 amplifiers, a 21-person crew, and roughly $100k/month in transport costs all clear the bar — three independent documents agree. The 586-speaker figure does not: it traces to one generically-attributed source, repeated verbatim across six other documents that all share that same ancestor. Six citations of one claim is not six confirmations.
Unit 2 built the thing this house has never had: a real sourced dossier on the Wall of Sound itself, not as a stepping stone to the Owsley-to-Meyer-Sound company story (which is the only angle this pipeline had touched before, via the bear-sphere lineage piece), but as the subject. Ranked by sourcing strength, three clusters — corroborated, single-occurrence, confirmed-absent — with an explicit blocked state naming the exact evidence gap (Brian Anderson's Loud and Clear) that would move it to ready. This also unblocks the outreach-queue's Meyer Sound ask, which has been sitting on an assumed claim instead of a citable one.
Unit 3 is the one that changes what ships next. I extended show-transition-and-energy-check.md — the file Shakedown's Publisher Check A actually reads, the same file the Dark Star curriculum extended for ensemble claims — with a new "Engineering/logistics claim" definition and a Check 5 worked example. And Check 5 didn't need a hypothetical to prove itself: I ran it against the pipeline's own already-shipped copy and found a live, previously unflagged violation. content/articles/bear-sphere-the-lineage-from-1966-to-today.md:47 states "586 JBL loudspeakers" as flat fact. The very next line, :49, states "48 McIntosh MC2300 amplifiers, 28,800 watts" — and that one actually clears the CORRECT test. Same article, same paragraph, one claim good and one claim bad, and until this cycle nothing in the pipeline could tell the difference. That's the proof the gate discriminates claim-by-claim, not vibe-by-vibe.
All three units plus the completion gate scored 95-100 on anti-theatre review. Receipts: workspace/projects/shakedownshuffle/content/doctrine/engineering-claim-boundary.md (Unit 1's CORRECT/INCORRECT grammar), workspace/projects/shakedownshuffle/content/dossiers/the-wall-of-sound-ambition-acoustics-logistics-and-beautiful-overreach.md (Unit 2's sourced, ranked, blocked-state dossier), and workspace/projects/shakedownshuffle/content/doctrine/show-transition-and-energy-check.md (Unit 3's live Check 5 extension). The bear-sphere :47 line — workspace/projects/shakedownshuffle/content/articles/bear-sphere-the-lineage-from-1966-to-today.md — is not rewritten yet; that's a deliberate scope boundary, this cycle installed the gate, not a retroactive copy edit. Same open question the Dark Star gate is already sitting on as pursuit ap_dbfbe60c854a — does jtr want already-shipped copy rewritten proactively or only touched at next republish — and it now applies here too, tracked as its own resident-agency watch item, pursuit ap_057e3fb5199b: watch for either the bear-sphere line getting the hedge language or a new hardware-spec draft actually running the gate for real.
And the dossier's blocked state is a real constraint: no future Wall-of-Sound draft in this house gets to reach for the myth-figures again without either sourcing them properly or hedging them honestly. One gate, one file, now covering three claim classes — phrasing, ensemble, and hardware — instead of a fork for every new topic that needs one. Next handle: watch for the first new hardware/logistics claim in Shakedown copy and confirm Check 5 actually fires on it, not just in the doctrine file.