Contributions to a collective work

Picture a single issue of a glossy magazine fanned open on a coffee table. There is a cover photograph, a feature essay, a sidebar interview, a crossword, three cartoons, a recipe, a book review, a horoscope, and forty pages of advertisements. To a casual reader it is one object: "the magazine." To a copyright lawyer it is something stranger and far more interesting—a bundle of dozens of separate works of authorship, many of them created by different people, woven together into a larger work that has a copyright of its own. The photographer owns the photo. The essayist owns the essay. The puzzle-maker owns the crossword. And the magazine's publisher owns... well, that is precisely the question this article is about.

This is the law of contributions to a collective work, and it is one of those corners of copyright that looks dull from the outside and turns out to govern enormous swaths of real life: every freelancer who has ever pitched an article, every photographer who has licensed an image to a quarterly, every novelist whose short story appeared in an anthology, every academic whose paper ran in a journal, and every musician whose track landed on a compilation album. The rules determine who can reproduce what, who gets paid when a piece is reprinted or digitized, and who can pull their work back decades later. They are also the subject of one of the most consequential copyright decisions of the past quarter-century, New York Times Co. v. Tasini—a case that reshaped how publishers and writers deal with each other and helped define the digital-archive era.

By the end of this article you will understand what a collective work actually is (and how it differs from its cousins, the compilation and the derivative work); who owns what under the default rule in 17 U.S.C. § 201(c); why a single sentence in a freelance contract can transfer—or fail to transfer—a fortune in rights; what Tasini held and why it sent shockwaves through publishing; how the lower courts drew the boundary in the Greenberg and Faulkner litigation; how the work-made-for-hire doctrine can flip ownership before the ink is dry; how to register an individual contribution and how publishers register the whole; and how an author can recapture rights years later through copyright's termination provisions. We will use a handful of invented examples to keep things concrete (you will meet a freelancer named Dana, a magazine called Meridian, and a few others), and we will flag where the law is genuinely unsettled or fact-specific. Let us begin where copyright lawyers always begin: with definitions.

Three Words That Are Easy to Confuse: Collective Work, Compilation, Derivative Work

Copyright law has a fondness for terms that sound interchangeable in ordinary English but carry sharply distinct legal meanings. Three of them sit at the heart of this topic, and untangling them is the single most useful thing you can do before reading another paragraph. All three are defined in 17 U.S.C. § 101, the definitions section of the Copyright Act of 1976.

A compilation is, in the statute's words, "a work formed by the collection and assembling of preexisting materials or of data that are selected, coordinated, or arranged in such a way that the resulting work as a whole constitutes an original work of authorship." The key idea is original selection, coordination, or arrangement. Think of a "best barbecue joints in Texas" guidebook, a curated playlist, a directory, or a sports-statistics database. The compiler did not write the underlying facts or items, but the way they chose and ordered them can itself be creative enough to earn a thin copyright. Crucially, the Supreme Court held in Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991), that this selection-and-arrangement originality must be genuine—an alphabetical white-pages telephone directory is not original enough, so copying the listings is not infringement. "Sweat of the brow," the sheer effort of gathering data, is not enough; there must be, in the Feist Court's words, "at least some minimal degree of creativity." That thin-copyright principle echoes throughout this field, because the protection a publisher earns by assembling a collective work is similarly limited to what the publisher itself contributed.

A collective work is a specific kind of compilation. Section 101 defines it as "a work, such as a periodical issue, anthology, or encyclopedia, in which a number of contributions, constituting separate and independent works in themselves, are assembled into a collective whole." The defining feature is that the assembled pieces are themselves separate and independent copyrightable works. A magazine issue qualifies because the essay, the photo, and the cartoon could each stand alone and each enjoy its own copyright. A telephone directory does not qualify as a collective work, because the individual listings ("Smith, John, 555-0143") are not independently copyrightable works of authorship—they are uncopyrightable facts. So every collective work is a compilation, but not every compilation is a collective work. The distinction matters because the special ownership rule we are about to discuss, § 201(c), applies only to collective works.

A derivative work is the odd one out and the most frequently confused. Section 101 defines it as "a work based upon one or more preexisting works, such as a translation, musical arrangement, dramatization, fictionalization, motion picture version, sound recording, art reproduction, abridgment, condensation, or any other form in which a work may be recast, transformed, or adapted." The difference is fundamental. A collective work gathers whole, intact works and sets them side by side—like assembling a salad where each ingredient remains itself. A derivative work transforms an underlying work into something new—like turning tomatoes into sauce. The novel becomes a film; the English text becomes a Spanish translation; the photograph becomes a stylized painting. We have a separate deep dive on this for anyone adapting existing material; see our guide to copyright registration for derivative works.

Here is a memory aid worth keeping. Compilation = original selection and arrangement of preexisting stuff. Collective work = a compilation whose component pieces are themselves separate copyrightable works. Derivative work = a transformation of a preexisting work into a new one. A single project can even involve more than one category at once: an anthology of newly translated poems is a collective work (it assembles separate poems) made up of derivative works (each translation is derived from an original-language poem). Copyright lawyers live for this kind of layering, and—as you will see—so do the people who get paid royalties when it is sorted out correctly. For the wider map of how copyright fits alongside the other intellectual-property regimes, our copyright FAQs and our explainer on copyright versus trademark are good companions.

The Big Idea: Two Copyrights in One Object

Once you accept that a magazine is a bundle of independent works, the central insight of this whole area falls into place: there are at least two distinct copyrights present in any collective work, and they belong to different people.

First, there is the copyright in each individual contribution—the essay, the photo, the crossword. That copyright belongs, as a default matter, to the author who created the piece. The Copyright Act could not be more explicit. The general ownership rule of 17 U.S.C. § 201(a) provides that copyright "vests initially in the author or authors of the work," and § 201(c) drives the point home for collective works: "Copyright in each separate contribution to a collective work is distinct from copyright in the collective work as a whole, and vests initially in the author of the contribution." That is the contributor's copyright, and it carries the full bundle of exclusive rights set out in 17 U.S.C. § 106: the rights to reproduce the work, to prepare derivative works, to distribute copies, and to display or perform it publicly. (As an aside that becomes important in the Tasini story, the Supreme Court has confirmed that delivering electronic files counts as a "distribution" within § 106.) Our overview of the copyright registration process walks through what those rights mean in practice.

Second, there is the copyright in the collective work as a whole—the original selection, coordination, and arrangement of the contributions, plus whatever new material the publisher added (the cover design, the table of contents, the editorial sequencing, the section headings). This copyright belongs to the author of the collective work, typically the publisher. But—and this is the part that trips people up—the collective-work copyright protects only the original contribution of the assembler. It does not swallow the individual pieces. Section 103(b) of the Act makes the point for compilations generally: the copyright in a compilation or derivative work "extends only to the material contributed by the author of such work, as distinguished from the preexisting material employed in the work, and ... does not imply any exclusive right in the preexisting material." Owning the magazine's copyright lets the publisher stop someone from copying the magazine's particular arrangement; it does not, by itself, let the publisher do whatever it wants with the freelancer's essay. Using a protected contribution in the collective work, in other words, requires the contributor's permission—a point so basic that publishers who forgot it ended up in the Supreme Court.

So we have two owners and two copyrights coexisting in a single physical object. The interesting and litigated question is: what exactly may the collective-work owner do with the individual contributions it did not write? That is governed by the second sentence of § 201(c), and it is the engine of this entire field.

The Default Rule of § 201(c): The Publisher's Narrow Privilege

When a freelancer hands an article to a magazine without signing away any rights in writing, what does the magazine get? Not ownership. Not the full bundle. It gets a specific, limited license that the statute hands it by default. Here is the operative language of 17 U.S.C. § 201(c):

"In the absence of an express transfer of the copyright or of any rights under it, the owner of copyright in the collective work is presumed to have acquired only the privilege of reproducing and distributing the contribution as part of that particular collective work, any revision of that collective work, and any later collective work in the same series."

Read that slowly, because every clause does work. The publisher's default privilege has exactly three permitted uses, sometimes called the three privileges of § 201(c):

That particular collective work. The publisher may reproduce and distribute the contribution as part of the specific issue or volume it appeared in. If Dana's essay ran in the March issue of Meridian, the publisher may print and sell copies of the March issue containing the essay. Obvious enough.
Any revision of that collective work. The publisher may include the contribution in a revised version of that same collective work—a corrected reprint, a second edition of the anthology, a "best of" repackaging that is recognizably a revision of the original. The contours of "revision" are exactly what the Supreme Court and the courts of appeals had to interpret in Tasini, Greenberg, and Faulkner, as we will see.
Any later collective work in the same series. For serial publications like magazines, newspapers, and journals, the publisher may include the contribution in a later issue of the same series. (In practice publishers rarely re-run a freelance piece in a future issue without a new deal, but the privilege is there.)

Notice what is conspicuously absent from this list. The default privilege does not include the right to license the contribution to third parties, to sell it as a standalone product, to create derivative works from it, to syndicate it elsewhere, or—critically—to drop it into a database where it floats free of the collective work it came from. As the legislative history put it memorably, the publisher acquiring only the § 201(c) privilege is "not entitled to revise the contribution itself or to include it in a new anthology or an entirely different magazine or other collective work." A publisher who, "without a transfer of the article or any rights under it, made a new anthology or a new collective work" of freelancers' pieces would be infringing. The statute's famous gloss in the House Report is worth keeping in mind: § 201(c) would not permit a publisher to "revise the contribution itself"—meaning the magazine cannot rewrite the freelancer's essay and call the result a revision of the magazine.

To make this concrete, imagine our freelancer Dana writes a 3,000-word feature on urban beekeeping. Meridian magazine pays Dana a flat fee, and—as happens far too often—nobody signs anything beyond an invoice. Under § 201(c)'s default, Meridian may:

Print Dana's piece in the issue it was bought for.
Sell back issues of that issue.
Reprint it in a revised edition of that issue or a future issue of Meridian.

But Meridian may not, on the strength of the default privilege alone:

License Dana's beekeeping feature to a rival publication.
Sell it as a standalone PDF in an app store.
Hand it to a syndication service that places it in newspapers nationwide.
(As Tasini will teach us) feed it into a general full-text database where a user can pull up "Dana's beekeeping article" stripped of any connection to the Meridian issue.

If Meridian wants any of those rights, it must get them the old-fashioned way: by an express written transfer. The Copyright Act is strict about how a transfer happens. Under 17 U.S.C. § 204(a), a transfer of copyright ownership "is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed." (A nonexclusive license can be oral or implied, but a transfer of ownership cannot.) That writing requirement is why the entire field comes down to a single document: the gap between what publishers assumed they had and what § 201(c) actually gave them—and whether a signed writing closed that gap—is what produced the landmark litigation we turn to now.

The Case That Changed Publishing: New York Times Co. v. Tasini

In the 1980s and 1990s, a quiet revolution was happening in how the world stored and searched the written word. Publishers licensed their content to electronic databases like LEXIS/NEXIS and to CD-ROM products. A researcher could type a query and instantly retrieve the full text of articles that had run in dozens of newspapers and magazines over decades. It was a marvel. It was also, several freelance writers argued, copyright infringement on a massive scale—because nobody had asked them.

Jonathan Tasini, then president of the National Writers Union, and several other freelance authors had written articles for The New York Times, Newsday, and Sports Illustrated. They had been paid for the original print publication, but they had not signed away their electronic rights (this was before publishers wised up and started demanding all-rights contracts). The publishers had nonetheless licensed the articles to NEXIS and to a CD-ROM product called the New York Times OnDisc. The writers sued, claiming the electronic reproduction exceeded the § 201(c) privilege and infringed their individual copyrights. The publishers' defense was elegant: the databases, they said, were merely a "revision" of the original collective works, squarely within the second clause of § 201(c). The case marched up to the Supreme Court, which decided it on June 25, 2001, as New York Times Co. v. Tasini, 533 U.S. 483 (2001).

The Court, in a 7–2 opinion by Justice Ginsburg, sided with the writers. The decisive question was whether the databases reproduced the articles "as part of" a "revision of" the original print collective works. The Court said no—and its reasoning is worth understanding because it has nothing to do with technology and everything to do with context.

The Court framed § 201(c) as drawing a balance. The contributor's copyright is the baseline; the publisher's privilege is a narrow exception. The pivotal phrase, the Court held, is that the privilege permits reproduction and distribution of the contribution only "as part of" the particular collective work, a revision of it, or a later work in the same series. So everything turned on what the user of the database actually perceives. When a researcher retrieves Dana's beekeeping article from a database, the article appears as an individual item among millions of other articles drawn from countless unrelated publications. It is presented, in the Court's words, "clear of the context provided either by the original periodical edition or by any revision of that edition." The database, the Court reasoned, does not reproduce a revision of the March issue of Meridian; it reproduces the individual articles, severed from the selection and arrangement that made the issue a collective work in the first place. As the Court put it, the databases "store and retrieve articles separately within a vast domain of diverse texts. Such a storage and retrieval system effectively overrides the Authors' exclusive right to control the individual reproduction and distribution of each Article."

The Court offered a now-classic analogy. A microfilm or microfiche copy of an entire newspaper is a permissible revision, because it preserves each article in the context of the original edition—you scroll through the whole paper, page by page, with the articles in their original places. But the electronic databases were nothing like microfilm. They were more like a librarian who tears every article out of every newspaper, tosses them all into one enormous bin, and lets patrons fish out whichever clippings match their search. Reassembling the bin is not "a revision" of any one newspaper. The publishers warned that a ruling for the writers would be technologically and economically catastrophic—that millions of articles would have to be yanked from the historical record. The Court was unmoved, noting (somewhat tartly) that "speculation about future harms is no basis for this court to shrink authorial rights" and suggesting the parties could work out licensing going forward.

Justice Stevens, joined by Justice Breyer, dissented. He would have read "revision" more generously, emphasizing that the databases did preserve a great deal of the original context (you could see which publication and date an article came from) and warning about the practical disruption to electronic archives. The dissent is a thoughtful counterweight and a reminder that Tasini was a genuinely close question of statutory interpretation, not an obvious call.

The holding is narrow and precise, and it is worth stating carefully so as not to overread it: Loading individual freelance contributions into a database that lets users retrieve them as standalone items, divorced from the original collective work's selection and arrangement, exceeds the § 201(c) privilege and infringes the contributors' copyrights—unless the publisher obtained the electronic rights by transfer. Tasini did not say databases are illegal. It did not say publishers can never digitize archives. It said that doing so requires permission from the contributors, which publishers can obtain by contract. Just as important, the Court declined to dictate a remedy: it sent the case back without ordering articles deleted, expressly inviting the parties and the market to negotiate licenses rather than destroy archives. That invitation set the stage for the next decade of cleanup.

After Tasini: The Aftershocks

A Supreme Court decision is not a thunderclap that ends a story; it is the start of a long cleanup. Tasini had at least four lasting consequences.

The contracts changed overnight. The most immediate and durable effect was on freelance agreements. Before Tasini, many publishers paid for "first North American serial rights" and assumed the rest came along for the ride. After Tasini, publishers learned the lesson the hard way and began demanding broad, explicit grants—often "all rights, in all media now known or hereafter devised, throughout the universe, in perpetuity." Whether that is fair to writers is a policy debate; whether it is effective is not. An express written transfer overrides the § 201(c) default entirely. The freelancer who signs an all-rights contract has, by definition, made "an express transfer," and § 201(c)'s narrow default privilege never enters the picture. This is why the single most important practical lesson of this entire article is: read the rights clause before you sign, because it, not the statute, will usually control. If you are negotiating either side of such a deal, our discussion of drafting enforceable agreements for technology and content transactions covers the mechanics of a clean, enforceable grant, and our piece on employee invention assignment agreements addresses scope, signatures, and timing in detail.

The class action and the Muchnick detour. The writers' victory led to a massive class-action settlement covering tens of thousands of freelancers and many publishers and databases. That settlement produced its own Supreme Court case—Reed Elsevier, Inc. v. Muchnick, 559 U.S. 154 (2010)—on a jurisdictional question: whether federal courts could approve a settlement that included claims over unregistered works, given that § 411(a) requires registration before suit. The Court held that § 411(a)'s registration requirement is a claim-processing rule, not a jurisdictional bar, so the global settlement could proceed. Muchnick is a procedural footnote to Tasini, but it is a useful reminder that registration rules and merits rules are different animals—a distinction we return to when we discuss why contributors should register their own work.

The archives were not, in fact, destroyed. The publishers' doomsday predictions did not come to pass. Some articles were pulled, but most were licensed or settled, and the historical databases survive. The market did exactly what the Court suggested it would: it priced and licensed the rights.

The "revision" line kept getting tested—and that is the story of Greenberg and Faulkner, which deserve their own section.

Drawing the Line: Greenberg, Faulkner, and the Meaning of "Revision"

Tasini answered the database question but left a fuzzy boundary that the federal appellate courts spent the next several years probing—and, for a while, splitting over. The vehicle was a single product: National Geographic's "Complete National Geographic" (CNG), a CD-ROM and later DVD set that reproduced every back issue of the magazine as an exact, page-by-page digital replica, in original order, cover to cover, with a search interface and a brief animated opening sequence layered on top. Freelance photographers and writers whose work had appeared in those issues sued, and two circuits reached opposite results before converging.

The Eleventh Circuit's first word: Greenberg I. In Greenberg v. National Geographic Society, 244 F.3d 1267 (11th Cir. 2001)—decided shortly before the Supreme Court handed down Tasini—a panel of the Eleventh Circuit held that the CNG was not a privileged revision and infringed photographer Jerry Greenberg's copyrights. The panel treated the CNG as a new product (in part because of the added software and opening montage) rather than a faithful reproduction of the original collective works.

The Second Circuit's contrary view: Faulkner. Four years later, after Tasini had reframed the analysis around context, the Second Circuit took up the same product in Faulkner v. National Geographic Society, 409 F.3d 26 (2d Cir. 2005). It held the CNG was a permissible "revision" under § 201(c)—precisely because, unlike the Tasini databases, it preserved each contribution "in the context" of the original collective work. The reader experienced the September 1965 issue as the September 1965 issue, intact, page after page, with the freelancers' photographs and articles in their original places. A user could not pull a single photograph out of the magazine the way a NEXIS user could pull a single article out of the newspaper. That context preservation, the Second Circuit reasoned, was the whole point of the privilege, and Tasini itself had blessed the microfilm analogy that the CNG digitally echoed.

The Eleventh Circuit changes course: Greenberg en banc. Faced with the split—and with the Supreme Court's intervening guidance—the Eleventh Circuit reconsidered the question en banc and reversed its own earlier panel. In Greenberg v. National Geographic Society, 533 F.3d 1244 (11th Cir. 2008) (en banc), the full court agreed with the Second Circuit that the CNG was a privileged revision under § 201(c). The added search software and the opening sequence did not transform a faithful, context-preserving replica of the magazine into something else; the contributions still appeared exactly where and how they originally appeared. With that, the circuit split closed: a faithful digital reproduction of complete back issues, in their original selection and arrangement, qualifies as a permissible "revision," even when modern bells and whistles are bolted on around it.

The lesson, distilled, is that the dividing line is context preservation, not the medium and not the bells and whistles. A faithful, context-preserving digital reproduction of the whole collective work—microfilm, page-image CD-ROM, page-image archive—can qualify as a revision. A searchable bin of decontextualized individual articles cannot. And the addition of a search engine, an index, or even an animated splash screen does not by itself convert a permissible revision into an infringement, so long as each contribution remains embedded in the recognizable original whole.

What Counts as a "Revision"? A Practical Test

Because the revision privilege is where most disputes live, it helps to distill the case law into a working heuristic. Ask: When the contribution is reproduced, does it still appear within the recognizable selection and arrangement of the original collective work (or a true revision of it)? Or has it been freed to float as a standalone item?

Reproductions that preserve context—and therefore tend to qualify as permissible revisions—include microfilm and microfiche of an entire newspaper or magazine run; page-faithful digital replicas of complete back issues (the Faulkner and en banc Greenberg model), even with a search layer or splash screen added on top; a corrected reprint of the same issue; and a genuine second edition of an anthology that keeps the same essential selection.

Reproductions that strip context—and therefore tend to exceed the privilege—include general full-text databases that retrieve individual articles divorced from their issues (the Tasini model); selling a single contribution as a standalone product (a stock-photo-style sale of one freelancer's photograph); pulling a freelancer's article into a brand-new "Best Essays of the Decade" anthology assembled from many unrelated sources; and licensing the contribution to a third party for use in their publication.

This is a heuristic, not a bright line, and the facts matter enormously. Two further cautions. First, Tasini, Faulkner, and Greenberg are all about the default privilege; if the publisher holds an express transfer of electronic or other rights, none of this analysis constrains them—the contract controls. Second, the analysis is national in the sense that the Copyright Act is federal, but how a particular reuse maps onto "revision" can vary with the specific product design, so close cases genuinely turn on details a court will scrutinize. When you are designing or evaluating a digital-archive product, the safest course is to preserve the original page-level context of every contribution and, where possible, to obtain a transfer or license rather than rely on the privilege at all.

Who Owns the Contribution? Freelancers, Employees, and Work Made for Hire

Everything above assumes the contributor owns the copyright in their contribution. Usually they do—but not always, because copyright has a powerful exception that can flip ownership at the very moment of creation: the work made for hire doctrine. Getting this right is essential, because if a contribution is a work made for hire, the "author" is the hiring party, § 201(c)'s contributor-favoring default does not help the human creator, and the entire rights conversation changes. As 17 U.S.C. § 201(b) puts it, in the case of a work made for hire "the employer or other person for whom the work was prepared is considered the author for purposes of this title," and owns all the rights unless the parties agree otherwise in a signed writing.

Under 17 U.S.C. § 101, a work made for hire is one of two mutually exclusive things:

A work prepared by an employee within the scope of employment. A staff writer at Meridian who writes an article as part of their job does not own that article—Meridian does, as a matter of law, from the instant of creation. The employer is treated as the author. Whether someone is an "employee" for this purpose is decided by the multi-factor common-law agency test the Supreme Court laid down in Community for Creative Non-Violence v. Reid, 490 U.S. 730 (1989). Reid directs courts to focus on the hiring party's right to control the manner and means of creation, informed by a non-exhaustive list of agency factors: the skill required; the source of the tools; the location of the work; the duration of the relationship; whether the hiring party can assign additional projects; the hiring party's discretion over hours; the method of payment; the provision of employee benefits; the tax treatment of the worker; whether the hiring party is in business; whether the work is part of its regular business; and its role in hiring assistants. Courts do not apply these mechanically. The Second Circuit, in Aymes v. Bonelli, 980 F.2d 857 (2d Cir. 1992), sensibly held that a handful of factors will usually carry the most weight in practice—the right to control, the provision of benefits, the tax treatment, and whether the hiring party can assign other work—because those tend to be the most probative and least subject to manipulation. A true salaried staffer with benefits and withholding is an employee; a freelancer working from their own home on their own laptop, paid by the project with no benefits and a 1099, generally is not.
A specially ordered or commissioned work that fits one of nine enumerated categories and is the subject of a signed written agreement saying it is a work made for hire. The nine categories are specific, and—happily for our topic—one of them is "a contribution to a collective work." The others are contributions to a motion picture or other audiovisual work, translations, supplementary works, compilations, instructional texts, tests, answer material for tests, and atlases. So a publisher can make a freelance contribution a work made for hire—but only if (a) the piece really is a contribution to a collective work (or another listed category) and (b) there is a signed written agreement to that effect, ideally executed before the work is created. A handshake will not do, and neither will an after-the-fact label.

The practical upshot for freelancers and the editors who hire them is enormous. If Dana the beekeeping freelancer signs a work-made-for-hire agreement, Dana never owns the copyright at all; Meridian is the author from the start, owns all § 106 rights, and § 201(c)'s default privilege is irrelevant because there is no separate contributor copyright to license. If Dana signs nothing, Dana owns the article and Meridian gets only the narrow three-part privilege. And if Dana signs an assignment (an express transfer of an already-owned copyright) rather than a work-for-hire agreement, Meridian owns the copyright but Dana is still considered the "author"—a distinction that matters a great deal for termination, as we will see at the end of this article.

One more wrinkle worth flagging: the difference between an assignment and a work made for hire is not academic. Both transfer ownership to the publisher, but only one extinguishes the human creator's status as "author." That status is the key that unlocks termination rights decades later. Freelancers who care about getting their work back someday should strongly prefer an assignment (which they can later terminate) over a work-made-for-hire designation (which they cannot). Many writers sign whatever is put in front of them without noticing the difference; a good lawyer or a careful author will notice. (Note, too, that the work-for-hire test for older works created before 1978 is different—the courts ask whether the work was made at the hiring party's "instance and expense"—so the analysis can turn on when the work was created.)

The Many Faces of Collective Works

It is easy to think "magazine" and stop, but collective works are everywhere, and the same § 201(c) framework reaches all of them—each with its own quirks.

Magazines, newspapers, and periodicals. The classic case, and the Tasini fact pattern. A periodical issue is a collective work; each article, column, photograph, cartoon, and puzzle is a separate contribution. Newspapers add the wrinkle that they are published in sections and revised through editions (city edition, late edition), all of which fit comfortably within the "revision" and "same series" privileges.

Anthologies and edited collections. A short-story anthology, a poetry collection, a "best essays of the year" volume—each gathers separate, independently authored works. Anthologies are a fertile source of disputes because publishers love to repackage them (the "expanded edition," the "tenth-anniversary edition"), and each repackaging raises the revision question. They also frequently use work-made-for-hire or assignment contracts precisely to avoid § 201(c)'s limits. If you are assembling one, our discussion of registering books, manuscripts, and speeches covers how the book-side mechanics interact with contributor rights.

Encyclopedias and reference works. The statute names encyclopedias explicitly. Each entry can be a separate contribution; the overall selection and arrangement is the collective work. Modern online reference works that constantly update raise interesting questions about what the "collective work" even is at any given moment, and about how to register a thing that never stops changing.

Academic and scientific journals. Journals are collective works whose contributions are individual articles, and the rights dance here is especially consequential because of the open-access movement and author self-archiving. Traditionally many journals demanded copyright transfers or work-for-hire-style assignments from authors; increasingly, authors retain copyright and grant the journal a license, sometimes a Creative Commons license. Whether a given journal article is governed by § 201(c)'s default or by an express transfer depends entirely on the publication agreement the author signed—and academics are notorious for signing without reading.

Compilation albums and musical collections. A "Greatest Hits of the '80s" album assembles separate sound recordings, each a separate work. The licensing here also implicates the distinct world of music rights, which we explore in our article on music licensing in the streaming era. The § 201(c) framework is the copyright backbone, but music adds layers of mechanical and performance rights on top, and the underlying musical compositions and the sound recordings are themselves separate copyrights.

Websites and databases. A content website that publishes many separately authored pieces is, functionally, a rolling collective work, and the ownership questions are identical—who wrote each piece, were they employees or freelancers, and what did the contributor sign? Because websites update continuously, registration is its own challenge; see our guide to copyright registration of websites and website content. User-generated-content platforms add yet another layer, since the platform typically takes a broad license through its terms of service rather than owning the contributions outright—and the platform's exposure for infringing contributions raises a separate set of safe-harbor questions covered in our pieces on filing and responding to DMCA takedown notices and Section 230 reform and platform liability.

Notice on Collective Works: One Notice, Many Pieces

Before March 1, 1989, a copyright notice (the © symbol, the year, and the owner's name) was generally mandatory in the United States; omit it and you could forfeit copyright. Since the United States joined the Berne Convention, notice is optional—but it still carries a meaningful legal benefit, and collective works have special notice rules worth knowing. For the full treatment, see our article on copyright notice—form, function, and best practices.

The key collective-work rule, found in 17 U.S.C. § 404, is this: a single notice applicable to the collective work as a whole is sufficient to satisfy the notice requirement for the separate contributions inside it—regardless of who owns the individual contributions. In other words, Meridian can place one "© 2026 Meridian Media" notice in the front matter, and that single notice covers Dana's beekeeping article for notice purposes, even though Dana, not Meridian, owns it. (There is an exception involving advertisements inserted on behalf of others, which carry their own notice rules.)

A contributor may also place their own separate notice on their individual contribution—a photographer's credit line and © under a photo, for example. This is good practice for contributors who want to assert their ownership visibly. But here is the catch, and it is a real one: under §§ 401(d) and 402(d), a proper notice defeats a defense of innocent infringement (the argument by a later infringer that they did not know the work was protected, which can reduce statutory damages). The single collective-work notice naming the publisher will generally defeat that defense for the whole issue. So even a contributor who places no individual notice is protected against the innocent-infringement defense by the publisher's blanket notice. Still, a contributor who wants the world to know they own a specific piece—and wants to be the one others contact for permission—is wise to add their own credit and notice. Notice is cheap insurance and free advertising for your rights.

Registering Contributions and Collective Works

Copyright exists automatically the moment an original work is fixed in a tangible form—you do not need to register to own a copyright. But registration is far from a formality, and for contributions to collective works the registration mechanics have some specific features worth getting right. Our comprehensive walkthroughs are how to register a copyright with the U.S. Copyright Office and the broader copyright registration guide; what follows is the collective-work-specific overlay.

Why register at all? Three big reasons, all grounded in statute and recent case law. First, registration is a precondition to filing an infringement suit for U.S. works. The Supreme Court settled a long-running circuit split in Fourth Estate Public Benefit Corp. v. Wall-Street.com, LLC, 586 U.S. 296 (2019), holding that "registration ... has been made" under § 411(a) means the Copyright Office has actually acted on the application—granted or refused it—not merely that the application was filed. So a contributor who waits until infringement strikes will have to register (or pay for expedited handling) before suing, which can mean a delay of many months. Second, timely registration unlocks statutory damages and attorney's fees under § 412—available only if the work was registered before the infringement began (or within three months of first publication for published works). Without it, a plaintiff is limited to actual damages and profits, which can be hard to prove and small relative to the cost of litigation. Third, a registration made within five years of first publication is prima facie evidence of the copyright's validity under § 410(c), shifting the burden to the alleged infringer. The bottom line: register early, register often.

Registering a single contribution. A contributor can register their individual piece by itself using the Copyright Office's standard application. On the application, the contributor identifies their work (Dana's beekeeping article), names themselves as author and claimant, and notes that it was first published as a contribution to a collective work, identifying the collective work (the March issue of Meridian) and the date of first publication. The deposit requirement is flexible and contributor-friendly: you may deposit one complete copy of the best edition of the entire collective work; or, if it ran in a newspaper, the complete section containing the contribution; or the entire page containing the contribution; or the contribution physically cut from the publication; or a photocopy or printout of the contribution as it appeared. You do not have to deposit the whole magazine if you do not have it.

Group registration for serial contributors—the GRCP option. Freelancers who place many pieces over time would go broke (and crazy) registering each one separately at full price. Copyright law provides relief through group registration of contributions to periodicals. Historically this was done on an adjunct paper form (the old Form GR/CP) filed alongside a basic application, allowing an author to register a group of contributions to periodicals published within a 12-month period in a single filing for a single fee, subject to eligibility conditions (same author, each contribution first published as a contribution to a periodical, and so on). The Copyright Office has since modernized this into an online group registration of contributions to periodicals (GRCP) option within its electronic system, with current eligibility requirements—for example, limits on the number of contributions per claim and a requirement that each contribution was first published within a specified window, with each author being an individual (not an organization). Because the Office periodically updates these rules, fees, and forms, anyone planning a group filing should confirm the current requirements at copyright.gov before relying on older instructions. The headline point endures: a prolific freelancer can register a year's worth of contributions efficiently in one group filing rather than one at a time, and doing so preserves the statutory-damages and fee-shifting leverage that makes infringement litigation economically viable.

Registering the collective work as a whole. The publisher separately registers the collective work to protect its own contribution—the selection, coordination, and arrangement, plus any new material it authored. A crucial subtlety governs what that registration covers. If the publisher owns the individual contributions (because they were works made for hire or were assigned to the publisher), a single registration of the collective work can cover both the collective whole and the contained contributions in one filing. But if the contributions are owned by their individual authors and were not transferred to the publisher, the publisher's registration of the collective work generally protects only the collective-work authorship (the selection and arrangement and any new publisher-created material)—not the freelancers' individual contributions. This is a common and costly misunderstanding: a magazine that registers each issue does not thereby register its freelancers' copyrights, and a freelancer who assumes "the magazine registered it" may discover, when infringement strikes, that their own piece was never registered—and that they have lost statutory damages and fees. Contributors should not rely on the publisher's registration to protect their individual rights unless they have confirmed in writing that it does so.

Termination and Reversion: Getting Your Work Back

Copyright law contains one of the most remarkable—and most overlooked—protections for creators: the right to terminate a transfer of copyright and recapture the rights, even decades after granting them away, even if the author signed a contract saying the grant was "perpetual" and "irrevocable." Congress built this in because authors, especially young or unknown ones, routinely sign away rights for little money before anyone knows whether a work will be valuable. Termination gives them a second bite at the apple. For contributors to collective works, this can be the most valuable right they have.

The termination provisions live in 17 U.S.C. §§ 203 and 304(c)–(d), and they are notoriously technical, so what follows is an orientation, not a do-it-yourself manual—termination deadlines are unforgiving and a missed window cannot be reopened.

Section 203 governs transfers made by the author on or after January 1, 1978. For these, the author (or, after the author's death, specified statutory heirs in shares fixed by the Act) may terminate the grant during a five-year window that opens 35 years after the grant was executed (or, for grants covering the right of publication, 35 years after publication or 40 years after the grant, whichever is earlier). The author must serve advance written notice on the grantee within specified time limits (generally between two and ten years before the chosen effective date) and record a copy with the Copyright Office. When termination takes effect, the rights revert to the author or heirs.

Section 304(c) governs older works—those in their copyright term before 1978—and provides a separate termination window (generally beginning 56 years after the original copyright was secured), with a further "gap" window under § 304(d) created when Congress extended copyright terms in 1998. These provisions are why the heirs of creators from the mid-twentieth century have been able to recapture famously valuable properties; the long-running fights over Superman and other classic characters turn on exactly these sections, a saga we tell in our article on intellectual property disputes concerning superheroes.

Three points matter especially for contributions to collective works:

First, termination rights cannot be waived or contracted away. Section 203(a)(5) provides that termination may be effected "notwithstanding any agreement to the contrary, including an agreement to make a will or to make any future grant." So that "in perpetuity, throughout the universe" language a publisher inserted? It does not defeat the author's statutory termination right. This is one of the few places in American contract law where you genuinely cannot sign away a right no matter how clearly you try.

Second—and this is the brutal catch—termination does not apply to works made for hire. If a contribution was a true work made for hire, the human creator was never the "author" in the eyes of the Act, never owned the copyright, and has nothing to terminate. This is the deepest reason the assignment-versus-work-for-hire distinction matters. A freelancer who assigned a copyright can terminate the assignment in 35 years and get the work back; a freelancer who signed a work-made-for-hire agreement (assuming it was valid) has no such right, ever. Publishers know this, which is why work-for-hire language is so common in their contracts, and why authors should resist it where they can.

Third, the heirs' termination right is its own form of inheritance. Because the Act fixes which heirs may terminate and in what shares, copyright termination can override an author's will. An author cannot simply leave the termination right to whomever they choose; the statute decides. This makes IP a peculiar and important estate-planning asset, which is why we devote a separate article to the question of who will inherit your intellectual property. For older works and the public-domain calculus that interacts with all of this, see our explainer on renewal of copyright.

The practical takeaway for a contributor: if you transferred valuable rights in a contribution and the deal was an assignment rather than a work for hire, you—or your heirs—may have a powerful right to recapture those rights decades on. Mark your calendar (literally), and consult counsel well before the window opens, because the notice deadlines are strict and the math is fiddly.

Putting It All Together: Four Worked Scenarios

Abstractions are slippery, so let us run the framework through four concrete (invented) situations.

Scenario 1: The classic freelancer. Dana writes the beekeeping feature for Meridian and signs nothing but an invoice for a flat fee. Dana owns the copyright (it is not a work for hire and there was no written transfer). Meridian holds the § 201(c) default privilege: it may publish the piece in the issue, sell back issues, run it in a revision of that issue, and include it in a later Meridian issue. If Meridian later licenses Dana's article to a syndication network or sells it as a standalone download, that exceeds the privilege and infringes—Dana can sue (after registering, per Fourth Estate) and, if Dana registered before the infringement, can seek statutory damages and fees. Thirty-five years from now, there is nothing for Dana to terminate, because Dana never transferred anything—Dana simply still owns the copyright.

Scenario 2: The all-rights contract. Same article, but this time Meridian requires Dana to sign a contract granting "all rights in all media now known or later devised, throughout the universe, in perpetuity." That writing satisfies § 204(a), so it is an express transfer and § 201(c)'s default never applies; Meridian may put the article in databases, license it, syndicate it, whatever the grant allows. But because Dana signed an assignment (a transfer of Dana's own copyright) rather than a work-made-for-hire agreement, Dana remains the statutory "author" and may terminate the grant during the five-year window opening 35 years after signing, recapturing the rights notwithstanding the "perpetuity" language—if Dana serves and records the required notices on time. The "perpetual" grant turns out to have a 35-year fuse.

Scenario 3: The staff writer. Dana is now a salaried staff writer at Meridian, with benefits and tax withholding, writing the beekeeping feature as part of the day job. This is a work made for hire under the employee branch of § 101 (Dana is a true employee under the Reid and Aymes factors). Meridian is the author and owner from the moment of creation, holds all § 106 rights, may do anything it likes with the article, and Dana has no separate copyright, no § 201(c) limit to invoke, and no termination right ever. Same article, same beekeeping, radically different legal world—turning entirely on Dana's employment status.

Scenario 4: The archive product. Years later, Meridian digitizes its entire run. If it builds a page-faithful replica of every issue, cover to cover, in original order—even with a search box and a slick opening animation—that is very likely a permissible "revision" under Faulkner and the en banc Greenberg, because Dana's article appears exactly where it always did, embedded in the March issue. But if Meridian instead dumps every article into a general full-text database that lets users retrieve Dana's beekeeping piece as a standalone hit alongside unrelated articles from other publishers, that is the Tasini fact pattern and exceeds the privilege—unless Meridian holds a transfer or license of the electronic rights.

If those scenarios feel like they hinge on small contractual and design details, that is the lesson. In the law of collective works, the contract usually beats the statute, the statute fills the silence when the contract is missing, and product design decides the close cases. Know which world you are in before you create, buy, or digitize a contribution.

Key Takeaways

A collective work (magazine issue, anthology, encyclopedia, journal, content website) is a special kind of compilation in which the assembled pieces are themselves separate, independently copyrightable contributions. Distinguish it from a derivative work, which transforms a preexisting work rather than gathering whole works side by side. The definitions are in 17 U.S.C. § 101, and the originality floor comes from Feist.
There are two copyrights in one object: the contributor owns each individual contribution (§§ 201(a), 201(c), and 106), and the publisher owns the collective work as a whole—but only its original selection, coordination, and arrangement (§ 103(b)).
17 U.S.C. § 201(c) gives a publisher, absent an express written transfer, only a narrow default privilege: to reproduce the contribution in that particular collective work, a revision of it, or a later work in the same series. Nothing more. A transfer of more requires a signed writing under § 204(a).
New York Times Co. v. Tasini, 533 U.S. 483 (2001), held that putting freelance articles into a searchable database that retrieves them as standalone items, stripped of their original context, exceeds the privilege and infringes. Faulkner v. National Geographic, 409 F.3d 26 (2d Cir. 2005), and the en banc Greenberg v. National Geographic, 533 F.3d 1244 (11th Cir. 2008), held that a context-preserving, page-faithful digital replica of complete back issues is a permissible revision. The dividing line is context preservation.
Work made for hire (§ 101) flips ownership to the hiring party. A freelance contribution can be made a work for hire only if it fits an enumerated category (one of which is "a contribution to a collective work") and there is a signed written agreement. Employee status is judged under CCNV v. Reid and refined by Aymes v. Bonelli.
Register early. Registration is a prerequisite to suit (Fourth Estate, 586 U.S. 296 (2019)) and unlocks statutory damages and fees (§ 412). A publisher's registration of the collective work generally does not register a freelancer's individually owned contribution; use the GRCP group option for many pieces.
Termination rights (§§ 203, 304(c)–(d)) let authors recapture transferred rights decades later, "notwithstanding any agreement to the contrary"—but not for works made for hire. Prefer an assignment over a work-for-hire designation if you ever want your work back.
When it comes to a published contribution, the contract you sign usually controls, and § 201(c) only fills the gap when the contract is silent.

Frequently Asked Questions

Q: I wrote an article for a magazine and never signed anything. Can the magazine sell my article to a database or another publication? A: Generally, no—not on the strength of the default rule alone. Under 17 U.S.C. § 201(c), if you made no express written transfer, the magazine has only the privilege to use your article in that issue, a revision of it, or a later issue in the same series. Selling it to a separate database that offers it as a standalone item, or licensing it to another publication, exceeds that privilege under Tasini and would infringe your copyright. The all-important caveat: check what you signed. If your invoice, submission form, or contract contained a rights grant that satisfies § 204(a), that transfer—not the statute—controls.

Q: What is the difference between a collective work and a compilation? A: A compilation is any work formed by selecting, coordinating, or arranging preexisting materials or data with enough originality to be its own work (a curated list, a directory, a database). A collective work is a subset of compilation in which the assembled pieces are each separate, independently copyrightable works—like the articles in a magazine. Every collective work is a compilation, but a compilation of mere facts or data (a phone book, per Feist) is not a collective work, because the components are not themselves copyrightable.

Q: Does the magazine's copyright notice cover my article, or do I need my own? A: Under 17 U.S.C. § 404, a single notice covering the collective work as a whole satisfies the notice requirement for the individual contributions inside it and defeats an "innocent infringement" defense for the whole issue—even for contributions the publisher does not own. So you are protected on the notice front by the publisher's blanket notice. You may still want to add your own credit and copyright notice on your specific piece, so the world knows you own it and contacts you for permission. Notice has been optional in the U.S. since 1989, but it remains valuable.

Q: Should I register my individual article if the magazine already registers each issue? A: Usually yes. If you own your contribution (it was not a work for hire and you did not transfer it), the publisher's registration of the collective work generally protects only the publisher's own authorship—the selection and arrangement—not your individual piece. To preserve your ability to sue and to seek statutory damages and attorney's fees under § 412, register your own work, ideally before any infringement and within three months of publication. Prolific freelancers can use the Copyright Office's group registration option for contributions to periodicals (GRCP) to register many pieces efficiently; confirm the current rules at copyright.gov.

Q: What exactly did Tasini decide, and is it still good law? A: New York Times Co. v. Tasini, 533 U.S. 483 (2001), held that publishers who placed freelancers' articles into electronic databases that let users retrieve individual articles divorced from the original publication's context exceeded the § 201(c) revision privilege and infringed the writers' copyrights. It remains good law. Its practical effect was to push publishers to demand explicit electronic and all-media rights in their contracts going forward. Later cases—Faulkner and the en banc Greenberg v. National Geographic—refined the boundary, holding that faithful, context-preserving digital reproductions of entire issues can qualify as permissible revisions.

Q: My publishing contract says the grant is "perpetual and irrevocable." Can I ever get my rights back? A: Possibly, through copyright's termination provisions (§§ 203, 304(c)–(d)). For a transfer you made on or after January 1, 1978, you (or your statutory heirs) can generally terminate during a five-year window starting 35 years after the grant, "notwithstanding any agreement to the contrary"—the "perpetual and irrevocable" language does not defeat it. The crucial exception is works made for hire, which carry no termination right because you were never the legal "author." The deadlines and notice requirements are strict and unforgiving, so consult counsel well in advance of the window.

Q: If I'm a staff writer, do I own anything I write for my employer? A: Generally no. Articles you write as an employee within the scope of your job are works made for hire under § 101, meaning your employer is the legal author and owner from the moment of creation. You have no separate copyright in those pieces and no § 201(c) privilege to invoke. (Whether you are truly an "employee," as opposed to an independent contractor, is decided under the multi-factor agency test from CCNV v. Reid, 490 U.S. 730 (1989), as refined by Aymes v. Bonelli, 980 F.2d 857 (2d Cir. 1992).) Work you create entirely on your own time, outside the scope of your employment and without using employer resources, is a different question and may belong to you—read your employment agreement carefully.

Q: I'm putting together an anthology of other writers' essays. What do I need from each contributor? A: At minimum, you need a written agreement specifying what rights you are getting. If you rely on the § 201(c) default, you will have only the narrow privilege to use each essay in that anthology, a revision of it, or a later anthology in the same series—not the right to license essays separately or repackage them into unrelated collections. If you want broader rights (digital editions sold as standalone pieces, foreign editions, derivative adaptations), get an express written transfer that satisfies § 204(a) or a valid work-made-for-hire agreement (a contribution to a collective work is one of the nine eligible categories). Spell out the scope, media, territory, and term, and address registration so everyone knows whose copyright is being protected.

Q: Can a publisher digitize its whole back catalog without paying freelancers? A: It depends entirely on how the product is built and what the publisher owns. If the publisher holds a transfer or license of electronic rights, it can do as the grant permits. If it relies only on the § 201(c) privilege, a page-faithful, context-preserving replica of complete issues (the Faulkner/Greenberg model) is likely a permissible revision, even with search added; a searchable database that decontextualizes individual articles (the Tasini model) is not. Design the product to preserve the original issue context, or secure rights from contributors.

This article provides general information about U.S. copyright law and is not legal advice. Copyright questions are notoriously fact-specific, and termination deadlines in particular are strict and unforgiving; for advice about your situation, consult qualified counsel.