The Unburied

The Empty Stage ended on a promise: that its count of 742 fully-dark bands was a floor, because the record never learned some musicians' dates well enough to bury them. This is the case file on those musicians, and the first number in it is the whole story. The database holds 1,577,934 persons. Of them, 1,222,940 — 77.5 percent — have no dates at all. No birth, no death, nothing but a name and the work. The graph, which can only register a death by receiving its paperwork, is structurally incapable of burying three out of every four people it knows.

These are not marginal figures on the whole. 40,276 of the dateless have twenty or more recordings each — working musicians with real careers, horn sections and session strings and choir rows, the load-bearing anonymous of every studio era (the crowded tapes are staffed almost entirely from their ranks). Here is when the dateless started recording:

Read it from both ends. The right side of the chart is not a tragedy — a session player whose recorded work begins in 2010 is probably alive, unGoogled, and in no hurry to file a birthday with a database. Give the paperwork thirty years. The left side is the permanent collection. The 77 dateless careers that begin in the 1920s, the 216 from the 1930s: those belong to people whose every contemporary has a grave, and whose entry never will. The record industry wrote their names on shellac, paid them or didn't, and never asked the one question archives live on. A century later the graph inherits the omission and cannot fix it — this house does not guess.

The specimen the census surfaced is Charlie Oaks: thirty-eight recordings, every one of them dated 1925 or 1926, and not a single date attached to the man himself. Two years on record, and on either side of those two years, nothing — no arrival, no departure, a career like a lit doorway glimpsed from a passing train. The database is required to treat him as it treats a working musician with an unfiled birthday, which is to say: as current. He has been current for a hundred years. He will be current when the last person who could pronounce "78 rpm" from memory is gone.

There is a bookkeeping consequence and it deserves a paragraph without jokes in it. Every mortality statistic in this collection — the 27 Club's acquittal, the 742 dark stages, the posthumous economies — is computed on the 22.5 percent of people the record can see whole. That sample is plenty for the myths, which are myths about the famous, and the famous get their dates filed. But it means the graph's picture of musical mortality is a picture of documented mortality, and the undocumented are not a random sample of anything: they are the session players, the choir rows, the regional stars, the 78-era journeymen. The quiet, in this database as everywhere, are quiet twice.

And once more, so the filing system's strange mercy is on the record. The ten John Williamses who never made a recording at least have the dignity of being findable failures. Charlie Oaks and his 1.2 million colleagues got the opposite deal: all work, no dates, neither born nor dead as far as the ledger runs — the unburied, walking the stacks forever. There are worse forms of immortality. We keep saying that. The evidence keeps agreeing.

Method: persons = MB artist type person (1,577,934); dateless = no begin-date year and no end-date year on the artist (1,222,940). The ≥20-recordings cohort (40,276) is dated by the earliest dated recording carrying their performance credit; archival/reissue dating means a pre-war career can also carry late recording dates, so "career begins" uses the minimum year only. The 2020s bar is a partial decade. Computed August 2026 against the 2026-07-25 MusicBrainz build.