NOTE-1019: Sotomayor — Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith

Contribution

It narrows “transformative use”, the doctrine that grew from Campbell v. Acuff-Rose (1994) and Leval’s 1990 article. In this case the District Court had treated a new meaning or message, seen side by side, as enough to make the use transformative (maj. pp. 9–10), and AWF argued the same (p. 12). The Court holds that the first factor asks about the purpose and character of a specific use, that new meaning or message bears on it only as evidence of a different purpose, and that commerciality and the reasons for copying are weighed against it (maj. pp. 12–20, 28–35). It also ties the doctrine to the derivative-work right, so that “transformative” must mean more than “transformed” (p. 16). It decides only the first factor, for only the 2016 licence (pp. 2, 21).

Key insight

The same copying is judged afresh at each use. “The same copying may be fair when used for one purpose but not another.” (maj. p. 20). Warhol’s silkscreen, licensed to illustrate a magazine about Prince, does what the photograph was licensed to do. So whatever the silkscreen means, that use supersedes the photograph’s objects: “Both are portraits of Prince used in magazines to illustrate stories about Prince.” (p. 22). Meaning would matter if it changed the purpose, for example by commenting on the photograph itself, and here it does not.

Assumptions

• US statute and precedent. It construes §107(1) of the Copyright Act of 1976, with §101’s definition of derivative works and §106(2), read through Campbell, Harper & Row, Sony and Google v. Oracle (2021).
• Only factor 1 was before the Court. AWF did not challenge the Second Circuit’s holdings that the works are substantially similar and that factors 2–4 favour Goldsmith (maj. pp. 12, 38). Goldsmith abandoned every claim but the 2016 licence and similar future licensing (n. 9).
• The use is the licence. The Court takes the challenged use to be AWF’s licensing, not Warhol’s making of the work (pp. 20–21). The dissent rejects that framing (Kagan n. 7) and Gorsuch adopts it (Gorsuch p. 2).
• Facts as found below. It grants, without deciding, the District Court’s finding that Orange Prince “reasonably can be perceived” to portray Prince as iconic where the photograph is photorealistic (maj. p. 33). The expert record on meaning is taken as given by both sides.

Key results

The results are holdings and their reasons, in the order the opinions give them. “Maj.” pages are the opinion of the Court.

• Facts (maj. pp. 1–12). Goldsmith photographed Prince for Newsweek in 1981. In 1984 her agency licensed one studio portrait to Vanity Fair as an “artist reference for an illustration”, for one use, for $400andasourcecredit(p.3).WarholmadetheVanityFairimageandfifteenmoreworksfromthephotograph(p.4;theAppendixcountsfourteensilkscreenprintsandtwopencildrawings).In2016Cond\acute{e}NastpaidAWF$10,000 to put Orange Prince on the cover of a tribute issue; Goldsmith got neither fee nor credit (p. 6). People paid Goldsmith $1,000 for a photograph for its own tribute issue (p. 7). On seeing the cover she said “It’s the photograph” (p. 8). The Court describes Orange Prince as one that “crops, flattens, traces, and colors the photo but otherwise does not alter it” (p. 8). The District Court found the works transformative and fair; the Second Circuit reversed on all four factors (pp. 9–12).
• What factor 1 asks (pp. 14–16). Its central question, from Campbell quoting Story in Folsom v. Marsh, is whether the use supersedes the objects of the original or “instead adds something new, with a further purpose or different character” (p. 15). It relates to the problem of substitution, “copyright’s bête noire” (p. 15). The §107 preamble (criticism, comment, news reporting, teaching, scholarship, research) gives the paradigm of a different purpose (p. 15; n. 4, a book review). Difference is “a matter of degree” (p. 15).
• Transformative must exceed derivative (p. 16). The owner has the right to derivative works, which §101 defines to include “any other form in which a work may be recast, transformed, or adapted”, and such transformations “may be substantial, like the adaptation of a book into a movie.” So “the degree of transformation required to make “transformative” use of an original must go beyond that required to qualify as a derivative.” Footnote 5 adds that transformative purpose and transformation of the work are separable in theory and overlap in practice.
• Two further points from Campbell (pp. 17–20). First, commerciality is an element of factor 1, relevant and not dispositive, weighed against the degree of difference (p. 18; n. 6 on the 1976 House Report). Second, factor 1 concerns justification: broadly, a distinct purpose serves copyright’s goal; narrowly, copying is justified when “reasonably necessary to achieve the user’s new purpose” (p. 19). Parody “needs to mimic an original to make its point”; satire “can stand on its own two feet and so requires justification for the very act of borrowing” (p. 17, quoting Campbell). The rule: “If an original work and a secondary use share the same or highly similar purposes, and the secondary use is of a commercial nature, the first factor is likely to weigh against fair use, absent some other justification for copying.” (p. 20). Footnote 8 reads Google the same way: a distinct computing environment, copying only as needed.
• The specific use (pp. 20–25). Several uses were made of the photograph; only the 2016 licence is alleged to infringe, and the Court says nothing about the creation, display or sale of the Prince Series (p. 21, nn. 9–10). Had the use been for teaching, “that clearly would affect the analysis” (n. 10). The typical use of a celebrity photograph is to accompany stories about the celebrity; there was uncontroverted evidence “that photographers generally license others to create stylized derivatives of their work” (p. 22), and Warhol himself paid for some reference photographs. “Such licenses, for photographs or derivatives of them, are how photographers like Goldsmith make a living.” (p. 22). In the tribute issue the purposes are substantially the same, and the use is commercial; “both elements point in the same direction” (p. 25).
• Purpose described specifically (n. 11). The Court does not reduce the purpose to “commercial licensing”. It compares the specific purposes, and notes that People, Rolling Stone and Time chose varied portrait photographs, black-and-white and colour, for their memorial issues, all serving “the same essential purpose”. It rejects the dissent’s “magazine editor” test: if an editor’s aesthetic preference sufficed, “then every derivative work would qualify” (n. 11).
• Factors 1 and 4 (n. 12). Factor 1 asks whether purposes are substitutable; factor 4 asks about actual or potential market substitution. They are positively associated and not identical. Under both factors “the analysis here might be different if Orange Prince appeared in an art magazine alongside an article about Warhol” (n. 12).
• Soup Cans (pp. 25–27). Google had suggested an artistic painting might fairly replicate an advertising logo to comment on consumerism. The Soup Cans’ purpose is “orthogonal to advertising soup”, and they target the logo: “It is the very nature of Campbell’s copyrighted logo—well known to the public, designed to be reproduced, and a symbol of an everyday item for mass consumption—that enables the commentary.” (p. 27). Licensing Soup Cans to a soup company as its logo would be a different case (n. 15). The difference in purpose there is “nearly absolute” (n. 16).
• New meaning or message is not the test (pp. 28–32). Campbell’s phrase “new expression, meaning, or message” paraphrased Leval, but cannot mean that any such addition favours the user, or ““transformative use” would swallow the copyright owner’s exclusive right to prepare derivative works” (p. 28). Arrangements, adaptations, sequels and spinoffs all add new expression (p. 28). A commercial remix of Prince’s Purple Rain would not be favoured just for new expression or aesthetic, and a film or musical of The Color Purple may alter a novel’s meaning and still need a licence (p. 29). In Campbell new meaning “was simply relevant to whether the new use served a purpose distinct from the original” (p. 29). Courts should not judge artistic worth (Bleistein, p. 31 and n. 19, which rejects a privilege resting on “It’s a Warhol”), and neither the user’s intent nor a court’s interpretation fixes purpose. But “the meaning of a secondary work, as reasonably can be perceived, should be considered to the extent necessary to determine whether the purpose of the use is distinct from the original” (pp. 31–32), as when it comments on, criticizes or provides otherwise unavailable information about the original.
• Applied to Orange Prince (pp. 32–35). Granting the iconic against photorealistic difference, the purpose remains to illustrate a magazine about Prince with a portrait of Prince, and the difference “is not enough for the first factor to favor AWF” (p. 33). “The application of an artist’s characteristic style to bring out a particular meaning that was available in the photograph is less likely to constitute a “further purpose” as Campbell used the term.” (p. 33). The asserted comment on the dehumanizing nature of celebrity “has no critical bearing on” the photograph, so is at “Campbell’s lowest ebb” (p. 34). AWF never claimed the works target the photograph, which was unpublished and chosen by Goldsmith’s agency, not by Warhol (n. 20). Targeting “is not always required”, but AWF must justify the taking with some reason other than “I can make it better” (n. 21, contrasting a film of Gone With the Wind with The Wind Done Gone). Copying that is merely helpful for a new meaning does not distinguish AWF from a sampler, a playwright adapting a novel or a filmmaker making a sequel (p. 35), and the dissent’s footnote 5 does not show the Court’s test favours them (n. 22).
• The reply to the dissent (pp. 36–37). Copyright is “replete with escape valves”: idea and expression, unprotected facts, originality, the standard for actionable copying, limited duration, and fair use’s other factors (p. 37). Holding: factor 1 favours Goldsmith; affirmed (p. 38).
• Gorsuch, concurring (pp. 1–6). Factor 1 looks at the challenged use, not at “the purpose the creator had in mind” or the aesthetic character of the work (pp. 1–2). It asks whether the use is different from “(and thus complements)” or the same as “(and thus substitutes for)” a copyrighted work (p. 2). Three clues: the preamble, which speaks of using a work for purposes such as criticism or comment (pp. 2–3); the derivative-work right, which would otherwise put the statute “at war with itself” (p. 3); and the fourth factor’s market question (pp. 3–4). AWF used its image “as a commercial substitute” (p. 4). The Court does not decide infringement: originality, substantial similarity, and the rule that two works drawing on a common third source generally do not infringe each other remain open (p. 5). Display “in a nonprofit museum or a for-profit book commenting on 20th-century art” might well be fair (p. 6).
• Kagan, dissenting (pp. 1–36). Warhol “reframed and reformulated—in a word, transformed” others’ images (p. 1). All the experts agreed the silkscreen differs in “composition, presentation, color palette, and media” and in meaning: Goldsmith’s subject was Prince’s “unique human identity”, Warhol’s “not the private person but the public image” (p. 9). An editor choosing between the two “would care”, and the publishers in fact wanted the Warhol (p. 10). Fair use exists because creation borrows (pp. 11–12, Story in Emerson v. Davies). Factor 1 is the one factor that asks what the copier accomplished (p. 13), and Campbell and Google both held commercial, transformative copying favoured (pp. 14–16). Licensors may refuse or overcharge, as Orbison’s successor-in-interest refused 2 Live Crew (p. 16). The majority reads “character” out of the statute (pp. 19–20, Webster’s), “transplants factor 4 into factor 1” and so double-counts the owner’s interest (pp. 21–22), and invents a targeting requirement that Campbell (n. 14 there) and Google do not support (p. 24). Under its test Shakespeare’s Romeo and Juliet (after Brooke), Lolita (after von Lichberg), Treasure Island, the blues, Stravinsky to Charlie Parker, Titian after Giorgione, Manet’s Olympia and Bacon after Velázquez would get no help from factor 1 (pp. 25–34). Footnote 5: the fourth factor, not the first, stops unlicensed film adaptations, citing Google. Footnote 8: artists make art for money, so a use-by-use rule still reduces the incentive to create.

Claims

id	claim	strength	support
C1	Factor 1 asks whether and how far a specific use has a purpose or character different from the original’s, weighed against commerciality	strong (as law)	holding of a 7–2 Court; Campbell at 579 and §107(1) text (maj. pp. 14–20)
C2	New expression, meaning or message is relevant only as evidence of a distinct purpose, and is not by itself enough	strong (as law)	holding (maj. pp. 12, 28–32); read from Campbell’s parody analysis
C3	Transformative use must transform more than a derivative work does, or §107 would swallow §106(2)	moderate	structural argument from §§101, 106(2) (maj. p. 16, pp. 28–29; Gorsuch p. 3); the dissent answers that factor 4 can do that work (Kagan n. 5)
C4	Copying is justified most clearly when the copied work is the target; commentary aimed elsewhere must justify the borrowing	moderate	Campbell’s parody/satire distinction (maj. pp. 17–19, 34–35, n. 21); the dissent shows Google did not rely on targeting (Kagan p. 24)
C5	Photographers’ licensing of photographs and of stylized derivatives is a market the owner holds, and AWF’s licence entered it	moderate	uncontroverted record evidence cited by the Second Circuit (maj. p. 22); the magazine covers (pp. 6–8)
C6	Bringing out with a characteristic style a meaning already available in the source is unlikely to be a further purpose	weak to moderate	asserted with the human face as example (maj. p. 33); the record’s experts, cited by the dissent, say the meaning was new (Kagan pp. 9–10)
C7 (dissent)	Factor 1 should measure the copier’s added expression and meaning, and the majority double-counts market harm	moderate	Campbell and Google text, expert record (Kagan pp. 13–22)
C8 (dissent)	Under the majority’s test much of the canon of art, literature and music would get no factor-1 credit	weak	illustrative examples not litigated, outside US copyright and its term (Kagan pp. 25–34); the majority answers that other doctrines cover them (maj. p. 37, n. 10)

Concepts

• Fair use. An affirmative defence to infringement, codified in §107 with four factors weighed together; the user bears the burden (maj. n. 21).
• Purpose and character of the use. As the majority reads it, the reasons for and nature of a specific use, including whether it is commercial; “character” includes commercial or nonprofit character (n. 14). As the dissent reads it, a thing’s “main or essential nature”, which for art includes its artistry (Kagan pp. 13, 20).
• Transformative. A use with a further purpose or different character, by degree (maj. p. 16). Distinct from being transformed, which any derivative is.
• Derivative work. A work based on one or more preexisting works, including “any other form in which a work may be recast, transformed, or adapted” (§101, quoted in part at maj. p. 16); preparing one is the owner’s exclusive right (§106(2)). The opinion’s examples are book to film, musical arrangements, stage adaptations, sequels and spinoffs (p. 28). The part of §101 not quoted lists translation first.
• Supersede the objects / supplant. Story’s phrase for a use that serves the original’s ends and so substitutes for it (p. 15).
• Justification. Broad: a distinct purpose serves copyright’s goal. Narrow: copying is reasonably necessary to the new purpose (p. 19).
• Targeting. The copied work is the object of the comment, as in parody and the Soup Cans, rather than its subject or society (pp. 17, 27, 34).
• Parody and satire. Parody targets an author or work; satire ridicules society and does not necessarily target a work (p. 17).

Connections

It applies and narrows Campbell v. Acuff-Rose (1994), and reads Google v. Oracle (2021) as consistent with it (maj. n. 8, n. 18); the dissent reads both the other way. It draws on Leval’s Toward a Fair Use Standard (1990), from which “transformative” comes, and on Leval’s own later opinion in Authors Guild v. Google (2d Cir. 2015). It affirms the Second Circuit (11 F.4th 26), whose “celebrity-plagiarist privilege” phrase it adopts (n. 19). The dissent’s tradition of borrowing (Story, Twain, Lethem’s The Ecstasy of Influence, Eliot in The Sacred Wood) is the one Lessig’s Remix (LIT-1298) argues from, though neither opinion cites Lessig.

Bearing on the record

• Lessig: remix’s meaning comes from the reference (NOTE-1023). Lessig argues that remix needs the original because its meaning lies in the reference, which only the original carries (Remix, p. 74). The opinion keeps a version of that and confines it. The law credits the need for the original only when the original is what the new work is about: parody “needs to mimic an original to make its point”, and the Soup Cans work because of what the logo is (maj. pp. 17, 27). Where the reference serves comment on something else, the new work is satire, and the borrowing must be justified another way (pp. 34–35). So Lessig’s argument transfers to law as a justification only for remix that targets its source. His own examples split. Söderberg’s Bush/Blair video, which Lessig reads as political comment (Remix, pp. 73–74), uses a love song to mock two politicians; on the majority’s terms that is satire as to the song. The opinion’s list of insufficient reasons names “a musician who finds it helpful to sample another artist’s song” (p. 35), and its Purple Rain remix example (p. 29) makes the same point. Three limits on the transfer: the case is a professional, commercial licence, not amateur remix, which Lessig wanted taken out of copyright altogether (free use, not fair use) and which the Court never considers; the Court analyses each use, which is close to Lessig’s fourth reform (regulate uses, not copies) though not derived from it; and the Court’s finding that the owner holds the market for stylized derivatives (p. 22) is the point at which Lessig’s unshown claim that remix does not compete with its sources (NOTE-1023, C3) meets a legal answer that assumes it can.
• CLAIM-002 (identity across retellings: artifact, transmission, function). The law runs CLAIM-002’s three relations separately, and that separation is the useful datum. Identity of the artifact is the infringement question, conceded here (substantial similarity, maj. p. 12; “It’s the photograph”, p. 8). Continuity of transmission is copying in fact, and Gorsuch notes that two works drawing from a common third source generally do not infringe each other (Gorsuch p. 5). Similarity of function is what factor 1 judges. But the opinion’s “purpose” is the purpose of a use in a market (illustrating a magazine story about Prince), not the communicative function of the work. The dissent’s “meaning or message” is the closer analogue of CLAIM-002’s function, and the majority rules it out as the test. So the case does not support or test CLAIM-002’s thesis that descent is not similarity. It shows that one institution judging reworkings needs all three relations and keeps them apart.
• CLAIM-034 (fidelity one goal among adaptation, critique and parody). The opinion grades exactly these goals, by the source’s role. Parody and critique of the source are favoured because the source is their object (pp. 17–19, n. 7). Adaptation, arrangement, sequel and, by §101, translation are the owner’s derivative right (pp. 16, 28). Satire, which uses the source as material for comment elsewhere, must justify the taking (p. 34). This gives CLAIM-034 an external criterion it lacks: whether the source is the target or the material of the rendering. It does not bear on CLAIM-034’s evaluative claim. The law asks who may make a rendering, not how to judge one, and it scores high similarity of purpose against the user, the reverse of fidelity as a goal. The majority’s sentence that a style bringing out “a particular meaning that was available in the photograph” adds no further purpose (p. 33), set against the new meaning the record’s experts found, as the dissent cites them (Kagan pp. 9–10), is a dispute over whether a rendering’s meaning lies within the source’s possibilities. That is the question CLAIM-034’s one-way simulation poses, but the Court decides it by assertion and does not measure it.
• CLAIM-210 (sharing a kind against being a rendering). The opinion has the same structure. Factor 1’s comparison of purposes is a kind-sharing test, symmetric in itself: the People, Rolling Stone and Time covers share the photograph’s purpose too (maj. n. 11), and that wrongs no one because none copied it. Purpose-sharing matters only given a copying, the directed relation, and Gorsuch’s common-source rule (Gorsuch p. 5) states the other half. This is an independent instance of the distinction, from law, and supports CLAIM-210’s claim that the two relations are different. It does not touch its claim about fidelity, which the law does not measure.
• QUESTION-002 (what survives translation, retelling, rendering). It does not answer it. It is a documented case of an institution deciding, on an expert record, what of a source survives in a reworking: the composition and pose (fig. 6; “crops, flattens, traces, and colors”, p. 8), against a changed meaning that the majority grants for argument and the record’s experts, cited by the dissent, affirm. The two sides measure different things: the majority the purpose of the use, the dissent the aesthetics and meaning of the work. That is a clean example of QUESTION-002’s demand that an answer say what is held fixed. It also illustrates CLAIM-220’s task-relativity, with the use standing in for the task: the same rendering is judged differently in a tribute magazine, an art magazine (n. 12) or a museum (Gorsuch p. 6). That is an analogy only; the law’s “use” is a market deployment, not a population of receivers recovering distinctions.
• Recognizability as a target. The Lessig reading asked whether recognizability of the source artifact could be a preservation target. The opinion splits recognizability in two. Recognizability of the source is what justifies parody’s borrowing (it must “conjure up” the original, p. 19). Recognizability of the renderer, a work “immediately recognizable as a ‘Warhol’”, was rejected as a ground (pp. 10–11, n. 19). The record’s question should keep the two apart.
• No bearing on CASE-100: the X-Men film took the line, not the fans’ redub, so no copying use of the redub is in question. No bearing on CLAIM-332, QUESTION-119 or QUESTION-023. No instruction for machine-learning practice; not held by the anthology.

Limitations

• Narrow by design. One factor, one use; the making, display and sale of the Prince Series, the other factors, and infringement itself are not decided (maj. p. 21, n. 10; Gorsuch pp. 5–6). Its reach to other uses is left to the lower courts, as the dissent notes (Kagan p. 22).
• “Purpose” is chosen at a level the Court picks. The Court describes the purpose as illustrating a magazine about Prince with a portrait of Prince, and concedes a more specific description is possible (p. 33). Why that level, and not “commercial licensing” or a Warhol for a tribute cover, is argued by example (n. 11), not by a rule. The dissent’s charge that the test reduces to commercial substitution (Kagan pp. 18–22) turns on this.
• Meaning is both counted and not. Meaning “as reasonably can be perceived” is relevant to purpose (p. 32), yet the commentary on celebrity is set aside “even if” perceptible (p. 34). The opinion does not say what meaning would have changed the purpose short of targeting, beyond its examples (art magazine, museum, teaching).
• The targeting rule is softened in a footnote. “targeting is not always required” (n. 21), but no other sufficient justification is described except Google’s technical necessity (n. 8).
• Factor 1 and factor 4. The relation is called a positive association that is not absolute (n. 12). The dissent’s double-counting objection is answered by assertion.
• The dissent’s canon is outside the case. Its historical examples were made without US copyright or under expired terms; the majority says so (n. 10), and says other doctrines would cover most of them (p. 37). Neither side shows how they would come out.
• The images were not seen in this reading, so nothing here judges the visual differences the opinions argue over (fig. 6; Kagan n. 3).

Open questions

• What, short of targeting the source, makes a new meaning a new purpose under factor 1? The lower courts’ answers since 2023 would settle it; this reading did not look at them.
• Does the derivative-work floor (p. 16) leave any room for a commercial translation or adaptation to be favoured under factor 1? The majority’s examples suggest not; the dissent says factor 4 should carry that load (Kagan n. 5).
• For the record: can the source’s role, target or material, be made a coded variable of renderings next to the goals CLAIM-034 lists, and does it predict how readers judge a rendering’s fidelity or license?
• For the record: is “purpose of the use” a usable stand-in for the task in CLAIM-220, or does a market deployment differ from a receiver task in a way that blocks the analogy?