The copy that was there

Nothing about the music separates them. Yet they are not the same historical thing. One was present at the release. The other belongs to the album's afterlife. The question is what, if anything, the object can show of that difference.

Five things in one case

It helps to separate what a sealed CD actually contains.

There is the work: the recorded music. It never changes. A copy made in 1994 and one made in 2014 carry the same performance.

There is the edition: a specification. Catalog number, barcode, artwork, track list, packaging. An edition can stay in production for decades without changing any of these.

There is the copy: this disc and this case, made at some plant on some date.

There is the retail event: this copy arriving in a particular shop, priced, shelved, perhaps sold.

And there is the moment: the weeks around the release, the radio, the reviews, the store displays, what it meant to buy this album then.

A claim about one of these is not a claim about the others. "Release-window copy" sounds like a claim about the copy. It is really a claim about the retail event, and through it, the moment.

A stack of clocks

Each layer keeps its own time.

The work keeps none. The edition's clock runs for as long as the label keeps printing it, so a barcode dates a specification, not an object.

The copy's clock is the most precise, and it is on the disc. Since 1994, discs made under the industry's Source Identification system carry codes in the inner ring identifying the mastering facility and the plant that moulded the disc.¹ A disc with those codes was made in 1994 or later. But a sealed copy hides its disc. The seal that preserves the copy covers the best evidence of when it was made.

So what can be read from outside is mostly the retail layer, written by shops: price stickers, store names, sometimes a date.

The newspaper test

The obvious comparison is a newspaper printed on the day of an event against a later reprint with the same text. It is worth checking whether that comparison holds.

The Library of Congress gets asked so often whether old newspapers are originals that it keeps a guide to a handful of famous issues. One is the New York Herald of April 15, 1865, carried the news of Lincoln's assassination and has been reprinted many times.² The originals came in at least four editions that day. The library tells them apart by the latest news bulletin each one contains: a 2 a.m. edition, a 3 a.m. edition, a 10 a.m. edition with the first mourning rules, a 2 p.m. edition reporting Andrew Johnson's inauguration.² Most copies found today are reprints, usually on dark wood-pulp paper rather than the originals' rag paper.²

Two things in that guide matter here.

First, a newspaper dates itself by what it knows. Its content is of the moment. A CD's music is not: it knows nothing about the week it was sold. So the analogy fails for the work. A later CD of an unchanged edition can be physically the same as an early one, where a Herald reprint betrays itself in its paper.

But the analogy partly holds for the packaging. Hype stickers often name events: a hit single, an award. A sticker that names an event cannot have been printed before it. A hype sticker is the closest thing a CD has to a latest bulletin. It gives a "not before" date, though never a "not after." A sticker on top of the shrink can also have been added later still, so its bound is the date of the event it names, not of the copy.

Second, the most famous edition of that Herald may not exist. Most reprints claim to be an 8:10 a.m. "Mourning Edition," and the library notes that the existence of an original of that edition "has never been proved."² Repetition of a category produced the category's main example. That is a useful warning for any trade where "first pressing" and "original" are worth more than their alternatives.

What the catalog leaves out

Collectors have a vast shared reference for telling releases apart, and its rules show where the retail layer sits.

Discogs, the crowd-built database most collectors use, does not allow cut-outs or "items that have otherwise been marked or altered after manufacture" to count as unique releases.³ Nor can a reissue that cannot be told apart from the original issue be entered separately: "unique releases must have a means of identifying them as such!"³

For a catalog of editions those are sensible rules. But they mean the retail layer, the store stickers and marks that record a copy's life after the factory, is outside the catalog by design. The main system for describing records describes everything except the part of a copy that was present at a moment. Nobody hid it. The category simply does not reach that far.

What a sticker can say

That leaves the shop's own marks.

A retail sticker can carry history even without a date. A 2020 post on the blog of UCLA's Strachwitz Frontera Collection, an online archive made from the Arhoolie Foundation's recordings, quotes the engineer who digitized them, Juan Antonio Cuellar, calling old store stickers "a mental map of where Mexican culture was bought, sold, and treasured by many immigrants."⁴ A sticker records a stop on the record's road from manufacturer to market, as the post puts it: a place, a kind of store, a community.⁴

When a sticker does carry a date, it dates the sticker. The alludio.music standards, written for the author's own archive of sealed CDs, put it exactly: a retailer sticker date "bounds when the sticker was applied, not when the disc was manufactured."⁵ One of that archive's copies is a sealed Mothership, the Led Zeppelin compilation from 2007, with a sticker dated 2015.⁶ The sticker is real evidence. It is evidence about a shop in 2015.

A dated sticker may not even date a sale. It records when a shop priced the copy, which might be the week it arrived or long before anyone bought it.

The strongest objection

The strongest objection is that "release-window copy" is mostly a story told about objects that cannot carry it. Sealed, the disc is unreadable. The sticker dates a sticker. The edition was made for years. What remains is a feeling about proximity, and feelings about proximity are what sellers turn into prices.

There is a deeper version too. A release is not a natural event like an assassination. The release date is set by a label, and the moment around it is partly built by marketing. To treasure a copy from that week is to treasure a moment the industry designed.

Both points are fair. Neither erases the thing. A designed event is still an event, and people lived it: they queued, they heard the single on the radio, they bought the album or didn't. And the objection is not that copies from the moment do not exist. It is that most cannot be shown to be one. That is a claim about evidence, not about history.

What survives

What survives is a way of speaking about one copy. Not "release-day CD" but a set of bounds, each tied to the mark that supplies it: no earlier than this, because of that sticker; priced by this shop, at this price, in a format from this period; disc unseen. The alludio.music standard requires a sticker with a full printed date within six months of release, a first-run edition and three kinds of photograph before it will call a copy "Release Window."⁵ As of its last build, its standards page reports that 0 of 102 copies carry any classification.⁵ A release day, in almost every case, is not among the bounds. Almost nothing on a sealed object records the day it was sold. Only paper about the object could: a receipt, a documented first owner.

An open edge

One odd consequence follows from how copies survive.

A CD bought in its release week was, very probably, opened and played. That is what it was for. A copy that survives sealed with a shop's sticker from that week was either bought and kept closed, or never bought at all. The industry had a whole system for the second kind: unsold copies were returned, marked with a notch or a punched barcode, and resold cheaply as cut-outs.⁷ How many of today's sealed old copies were once unsold stock, nobody has counted.

If many were, then the cleanest surviving witnesses of a release are the copies its audience passed over. They were there for the moment, in the shop, under the lights. They just went home with no one. That does not make them worse evidence. It makes them a different kind: not a record of what people bought, but of what was offered to them.