The Reproduction Right—Part II

This chapter addresses some advanced topics in relation to establishing a violation of the reproduction right.

Copying in Fact

Access is not an element of the plaintiff’s prima facie case

In Feist the Supreme Court said: “to establish infringement, two elements must be proven: (1) ownership of a valid copyright, and (2) copying of constituent elements of the work that are original.” This second element entails both copying in fact (or factual copying) and wrongful copying (or improper appropriation). As mentioned in the previous chapter, copying in fact means that there is a causal connection running from work A to work B and this can be established by admission, stipulation, direct observation, or indirect evidence.

The plaintiff in an action for copyright infringement must show that B was actually copied from A in the sense that there is a causal connection running from A to B. Courts refer to this requirement as either “copying in fact,” “factual copying,” or “actual copying.” Such actual copying can be deliberate, unintended, or even subconscious. If two works are similar, but were in fact independently created, there is no copyright infringement. The classic statement to this effect comes, again, from Judge Learned Hand, this time in Sheldon v. Metro-Goldwyn Pictures Corporation, 81 F.2d 49 (1936):

… but if by some magic a man who had never known it were to compose anew Keats’s Ode on a Grecian Urn, he would be an “author,” and, if he copyrighted it, others might not copy that poem, though they might of course copy Keats’s.

Of course, most of us don’t believe in magic and so we are usually prepared to draw an inference from very high levels of similarity that B probably was copied from A. We will return to this topic below.

Sometimes, quite often in fact, a defendant will simply admit that the plaintiff’s work influenced their work and fight the case on lack of wrongful copying or a defense of fair use. Sometimes, but rarely, some third party will be in a position to say, “I saw the defendant composing his novel with a detailed summary of the plaintiff’s novel in hand.” If that testimony is believed, it should establish factual copying. However, in a large number of cases, the plaintiff cannot rely on an admission or direct evidence and must establish copying in fact by indirect evidence: that is, the plaintiff must persuade the judge or jury of the ultimate fact of copying it can’t prove directly by drawing inferences from the facts it can prove. In such cases, plaintiffs invariably rely on some theory of how the defendant could have had access to the work and on similarities between the works that make copying from the plaintiff—as opposed to copying from a common source or just plain coincidence—seem the most likely explanation. Hopefully, this strikes the reader as mundane common sense, but courts have struggled with it over the years.

As discussed in the previous chapter, in Sid & Marty Krofft the Ninth Circuit framed the test for infringement in terms of ownership, access, and substantial similarity despite the fact that it said it was applying the two-part test in Arnstein. Many cases following Krofft treat access and substantial similarity as elements that establish infringement rather than components of copying in fact. This approach is mistaken. Access is a way to prove copying in fact, not something that must be proved for its own sake, and certainly not something that dispenses with the need to prove the ultimate question of copying in fact.

Probative similarity is different to substantial similarity

Another problem that Sid & Marty Krofft contributed to is the perennial confusion between similarities that make copying in fact more likely and similarities that make copying wrongful, assuming it occurred. One reason why courts get confused between copying in fact and wrongful copying is that a factual determination of similarity can play some role in both issues in many cases. Similarities between A and B may be evidence that there was copying in fact because all other things being equal, the closer A is to B, the greater the chances that B was copied from A. But a certain threshold of similarity (as to protectable expression) is also required to show that if there was copying, it was wrongful. As a result, courts sometimes lose sight of the distinction between copying in fact and wrongful copying and apply a singular threshold of substantial similarity to determine both.

In fact, courts regularly lump the entire inquiry, copying in fact and wrongful copying under the single term “substantial similarity.” This not only makes it confusing to read the cases, but it also leads courts themselves to conflate the two distinct elements. Courts should try to keep the question of whether B really did copy from A and the question of whether such copying went too far, quite separate. The extent of similarities between A and B will be important in answering both questions, but we are actually dealing with two very different concepts of similarity here. The two kinds of similarity are different and they deserve different names: the most logical nomenclature differentiates between probative similarity that supports an inference of copying in fact and substantial similarity which addresses wrongful copying.

As Alan Latman observed in his definitive article published several years after his death:

“Substantial similarity,” while said to be required for indirect proof of copying, is actually required only after copying has been established to show that enough copying has taken place. A similarity, which may or may not be substantial, is probative of copying if, by definition, it is one that under all the circumstances justifies an inference of copying. In order to emphasize the function of such similarity and avoid the confusion of double usage, this Article suggests use of the term “probative similarity” in place of “substantial similarity” in this context.

For more, see Alan Latman, “Probative Similarity” as Proof of Copying: Toward Dispelling Some Myths in Copyright Infringement, 90 Columbia Law Review 1187 (1990).

There is no fixed relationship between probative similarity and substantial similarity

The following examples will demonstrate that there is no simple relationship between the level of similarity that supports an inference of copying in fact and that which sustains a conclusion of wrongful copying.

Example 1

Consider two photos with more than a passing similarity as depicted below.¹ Photographs A and B are similar in subject, composition, lighting, and almost any other dimension one cares to think about. The only real difference between them appears to be in the photographer’s choice of color filters and the level of contrast, both of which could be changed in post-production.

Suppose that Photo ‘A’ was taken in 2006 and made available on the photographer’s website. In 2009, photo ‘B’ is entered in a newspaper photography competition and now the author of A alleges that B infringed her copyright. There is no doubt that if B had been copied from A and simply modified in some subtle way, B would be an infringing copy of A. But as it turns out two women standing beside one another on a cruise to the San Rafael Glacier in the Northern Patagonian Ice Field took both photos independently at almost the exact same time. Here we have similarity that would be enough in some circumstances to reach a threshold of wrongful copying, but not enough to prove copying in fact.

Given the subject matter, the possibility that both photos were taken at the same place and time seems only too apparent. As the Seventh Circuit noted in Ty Inc. v. GMA Accessories, Inc., 132 F.3d 1167, 1170 (7th Cir. 1997):

… two works may be strikingly similar—may in fact be identical—not because one is copied from the other but because both are copies of the same thing in the public domain. In such a case—imagine two people photographing Niagara Falls from the same place at the same time of the day and year and in identical weather—there is no inference of access to anything but the public domain, and, equally, no inference of copying from a copyrighted work.

Example 2

But it is easy to imagine the opposite. Suppose A is a well-known and widely appreciated work of literature. Author B, takes the essential story of A, changes the setting in both time and culture, adopts a different narrative style and perspective to obtain a different point of view but retains the same essential conflict. To put a little more distance between the works, B changes names and combines characters. B also extends the timeline of the original story, so that her account captures events she imagines might have taken place before and after the sequence of events in A. From this process of deliberate mutation, B may produce a work that is thoroughly original but still discernably influenced by A. Here we have a causal connection between the works, and we have similarity that is sufficient to establish copying in fact, but this is still insufficient to establish wrongful copying.

Example 3

A produces a telephone directory of the best restaurants in the Atlanta Metro area. To catch plagiarists in the act, A includes four fictional listings for restaurants that do not exist. A can use evidence that B’s telephone directory also contains the exact same fake listings to prove beyond any real doubt that B copied from A, but that won’t go very far in establishing that B has infringed A’s copyright. As we saw in the notes on Feist in a previous chapter, even if A can show that her directory contained enough creative spark to merit copyright protection, her copyright will still be “thin”, and she will also have to show that B copied her almost exactly to prevail.

***

As seen in each of the examples, there is no simple relationship between probative and substantial similarity. In application, the similarities that establish wrongful copying are often the same as those that tend to prove that there was copying in the first place, but the kinds of similarity are fundamentally different. Substantial similarity and probative similarity intersect, but they are not the same, nor is one necessarily a subset of the other.

Probative similarity is any similarity at all between the works that makes it more likely that the accused work was copied from the plaintiff’s original work. Similarities in subject matter, genre, perspective, general idea, and concept are unlikely to meet this test, and they should also be filtered out from any assessment of substantial similarity for reasons to be explained shortly. However, probative similarity can be established by identical features that are inconsequential and unlikely to be noted by the ordinary observer but for which copying seems the more likely explanation than sheer coincidence. As noted, the fact that two works contain identical deliberate mistakes is great evidence of copying in fact, but it will usually mean nothing in terms of wrongful copying. If two maps feature the same nonexistent town, it stands to reason that either one was copied from the other or both were copied from the same source. The duplication of a single errant entry would not cross the threshold of improper appropriation, but it can be strong evidence of copying in fact.

Nonetheless, there is also a significant overlap between probative similarity and substantial similarity—if two works are sufficiently similar in terms of protectable expression that the ordinary observer would regard them as the same in their aesthetic appeal, that one conclusion may establish that one was indeed copied from the other and that such copying went too far.

Independent creation

Evidence of independent creation is evidence that there was in fact no copying. While in general it is hard to prove a negative, evidence in the form of logs, drafts, and eyewitness accounts can go some way to proving independent creation. However, courts and juries can always discount such evidence, on the view that the defendant’s belief in independent creation is perfectly consistent with subconscious copying. As discussed in a previous chapter, in Bright Tunes Music v. Harrisongs Music the court ruled that George Harrison had subconsciously copied the earlier song, the Chiffons’ song “He’s So Fine” given he had likely been exposed to it in the many weeks it spent in the top-100 chart and the similarity of the melodies. A similar theory prevailed in Three Boys Music v. Bolton, although with much weaker evidence of access.

The death (and likely rebirth) of the inverse ratio rule

In the absence of an admission of copying or direct evidence, the plaintiff must rely on circumstantial evidence to establish that her work was in fact copied. Two vital such circumstances are that the defendant had access to the work and that there are similarities between the works suggestive of copying.

Imagine that in one case the defendant actually possessed a copy of the plaintiff’s work and in another, there was no evidence that she had ever seen it, much less had convenient access to a copy. Assuming some moderate level of similarity between the plaintiff’s and the defendant’s works, it seems reasonable to say that there is better evidence of actual copying in the former case. If the defendant possessed a copy of the plaintiff’s work, it is easy to see how she could have copied it. Given that degree of access, a certain level of similarity we can call X% might be enough to persuade the finder of fact that she did actually copy the work. Now imagine in the second case that there are a great many more unexplained similarities between the works, so the level of similarity is Y% (where Y is greater than X). The stronger the evidence of similarity (especially as Y approaches 100%), the more likely the work was copied, even if there is no specific evidence of access. This makes sense as long as there is a bare possibility of access and the similarities don’t have some other explanation.

Based on the foregoing, there is something to the intuition that there is some kind of tradeoff between evidence of access and evidence of similarity, but the status of the “inverse ratio rule” is now in doubt.

Skidmore v. Led Zeppelin, 952 F.3d 1051 (9th Cir. 2020) (En Banc)

Notes and questions

(1) One way to think of the supposed “inverse ratio rule” is that it simply reflects a proposition about what kind of circumstantial evidence will support a finding of copying in fact by the defendant. Once copying in fact has been established, the plaintiff must still show that such copying was wrongful and the “rule” is irrelevant. However, in the Ninth Circuit, the rule was often applied to reduce the plaintiff’s burden of proof in terms of wrongful copying. This makes no sense at all, but that is how the court seemed to apply it in Sid & Marty Krofft. In that case, the court said: “We believe that this degree of access justifies a lower standard of proof to show substantial similarity.” There was no question in Sid & Marty Krofft that copying had in fact taken place. That should have made the inverse ratio rule irrelevant, but instead the court seemed to apply the rule to reduce the burden of proof for establishing wrongful copying. This is daft and adds to the list of problems with Sid & Marty Krofft discussed in the prior chapter.

In Rentmeester in 2018, the Ninth Circuit tried to salvage the inverse ratio rule by confining it to the question of copying in fact where it belongs. This addressed the most glaring problem with the inverse ratio rule, but only two years later, the Ninth Circuit convened en banc in Skidmore v. Led Zeppelin, to “reject[] the inverse ratio rule” and said it was joining the majority of circuits that had already done likewise.

(2) A major problem with the “inverse ratio rule” is its name—describing the observation in terms of an “inverse ratio” or indeed as a “rule” invites misapplication. Placing the terms access and similarity into a “rule” or even worse, a “sliding-scale,” as some courts have done, invites courts to view the concepts in their broadest terms. This leads to error because access and similarity in their most general senses are far from the only things to consider when trying to determine whether it is more likely than not that the defendant copied the plaintiff’s work. If the question is whether, in all the circumstances, is it likely that B copied from A, then courts must consider all sorts of explanations for similarities between B and A, including external motivation/stimulus; constraints of the environment; common third-party sources of inspiration. Invoking the “inverse ratio rule” may distract from this broader inquiry. Moreover, it is problematic to talk in terms of an “inverse ratio” between two concepts that can’t be measured on a continuous scale, are obviously non-linear, and are not reducible to a common scale.

(3) Perhaps given all the baggage of the inverse ratio rule the Ninth Circuit was right to disavow it entirely. But the court did not disavow the notion that there could be tradeoffs in particular cases, and it would not be surprising to see the underlying logic of the inverse ratio rule reappear under different terms in the future.

How much access is enough

Access to the plaintiff’s work can be established in many different ways. A plaintiff may satisfy the requirement by showing that the copyrighted work was sent directly to the defendant or a close associate of the defendant. In Peters v. West, 692 F.3d 629 (7th Cir. 2012), for example, the defendant closely collaborated with a music producer who had a copy of the plaintiff’s recording. Even though the music had never made it to the radio, that was enough to establish access. This is why many people and organizations working in creative industries return submissions unopened. In other cases, plaintiffs have established access by showing that the copyrighted work was so widely disseminated that the defendant can be presumed to have seen or heard it. In the Internet age, it is usually not that hard to establish the bare possibility of access, but that may not be enough.

Art Attacks Ink, LLC v. MGA Entertainment Inc., 581 F.3d 1138 (9th Cir. 2009)

Notes and questions

(1) In Gaste v. Kaiserman 863 F.2d 1061 (2d 1988), the plaintiff, Gaste, alleged that the defendant, Kaiserman, copied his song “Pour Toi” to create the hit song “Feelings,” claiming that Kaiserman gained access to the song through a third-party music publisher, despite the defendant’s denial of ever hearing the song prior to the litigation. Although the court emphasized that a “reasonable opportunity” to access the work is required, rather than just a bare or speculative possibility, it held that the jury could reasonably conclude that the defendant, Kaiserman, had access to Gaste’s song “Pour Toi” through Lebendiger, who had a connection to both parties. Moreover, the court was not troubled by long span of time and geography separating the works. The court concluded that “although Gaste’s theory of access relies on a somewhat attenuated chain of events extending over a long period of time and distance, we cannot say as a matter of law that the jury could not reasonably conclude that Kaiserman had access to the song through Lebendiger.”

Why was the plaintiff’s argument in Gaste stronger than in Art Attacks? What does this tell you about the difference between a reasonable opportunity of access and bare or speculative possibility?

(2) Does posting to social media give the whole world a reasonable opportunity of access? The Ninth Circuit said no in Woodland v. Hill, 136 F.4th 1199 (9th Cir. 2025). A photographer alleged that the recording artist Montero Lamar Hill, who performs as Lil Nas X, had imitated twelve of his photographs. His theory of access was that both men posted to Instagram, that their work fell within the same narrow sub-genre, and that Instagram’s recommendation algorithm would therefore have put his photographs in front of Hill. The court accepted that digital platforms could in principle make access easier to establish, but only where a plaintiff can show a reasonable chance that the defendant would actually have encountered the work given the platform’s algorithm or content-sharing policy. That was a large qualification, and the plaintiff could not satisfy it. Instagram has more than a billion users and far more posts; the fact that the defendant used the platform and the photographs were on it established no more than a bare possibility. The algorithmic theory was speculation, because nothing in the material offered showed that similarity of profile content alone would cause Instagram to promote one user’s posts to another, and there was no allegation that Hill had followed, liked, or otherwise interacted with the plaintiff’s account or anything like it. The court expressly declined to say what a plaintiff would need to plead about a platform’s algorithm to get over the line.

The plaintiff also argued that the sheer number of allegedly imitated photographs — twelve, not one — itself suggested access. The court disagreed, and the answer is worth remembering: the existence of multiple similar works does not prove access. That is a similarity argument wearing an access costume, and it belongs, if anywhere, in the striking similarity analysis considered below.

Striking similarity

What if the plaintiff does not have evidence that the defendant had a reasonable opportunity to view the work but there was a very high degree of similarity between the works? In such cases of “striking similarity” some courts take the view that the similarity between the works becomes its own evidence of access. Courts accept in effect that the similarities between A and B are such that the only plausible explanation is copying in fact. These cases are unusual. As the Seventh Circuit observed in Selle v. Gibb, 741 F.2d 896 (7th Cir. 1984), an inference of access may arise from “proof of similarity which is so striking that the possibilities of independent creation, coincidence and prior common source are, as a practical matter, precluded.” But note that in that case the court was careful to point out that a prior common source for both plaintiff’s and defendant’s work could undermine an inference of copying. In a later Seventh Circuit case, Design Basics, LLC v. Lexington Homes, Inc., 858 F.3d 1093 (7th Cir. 2017), the court summarized:

So in the rare case involving an uncanny resemblance for which copying the plaintiff’s work is the only plausible explanation, further proof of access may not be required. But in general, to prove a circumstantial case of copyright infringement, the plaintiff must separately prove both access and similarity. (emphasis added).

Note that Design Basics indicates the threshold for striking similarity is high, it applies to “rare case involving an uncanny resemblance.”

Unicolors, Inc. v. Urban Outfitters, Inc., 853 F.3d 980 (9th Cir. 2017) illustrates striking similarity at work. In that case, Unicolors, a Los Angeles-based fabric design company, held a copyright in a fabric design called PE1130. Between 2008 and 2011, Unicolors sold about 14,000 yards of fabric with this design. In 2010, Urban Outfitters developed a dress using a fabric design that Unicolors alleged was similar to its PE1130 design. The district court determined that there was insufficient evidence to conclude on summary judgment that the Subject Design had been widely disseminated. But it concluded that such evidence was unnecessary because on its analysis, the works were strikingly similar. The court of appeals agreed:

We conclude that a district court may grant summary judgment for plaintiffs on the issue of copying when the works are so overwhelmingly similar that the possibility of independent creation is precluded. The works at issue in this case meet this standard. The objective similarities between the works are stark: the shapes, motifs, arrangements, spacing, and colors of the images in each design are nearly identical. Because of the decisive objective overlap between the works, no reasonable juror could conclude under the intrinsic test that the works are not substantially similar in total concept and feel. Therefore, the district court properly granted summary judgment.

Wrongful Copying (a.k.a Improper Appropriation)

The role of experts in wrongful copying

The ordinary observer test for substantial similarity is applied by conducting a side-by-side comparison of the two works, but that does not always make sense. Consider a novel written in English and an alleged copy written in French. These works don’t lend themselves to side-by-side comparison and statistically, a person who is fluent in both English and French is extraordinary, not ordinary. A simple side-by-side comparison also doesn’t make sense if there are important similarities that relate to unprotectable elements within the works. Thus, even when courts determine substantial similarity with respect to the ordinary observer, they usually supercharge that hypothetical observer with enormous powers of discernment.

The ordinary observer standard has led to undue rigidity on the use of experts in copyright litigation. Using experts to determine the reaction of the ordinary observer strikes many courts as a non sequitur. Most famously, in Arnstein v. Porter, 154 F.2d 464 (2d Cir. 1946), the Second Circuit said that analytical dissection of the two works in issue was relevant to establish copying in fact and thus “the testimony of experts may be received to aid the trier of the facts.” But, on the issue of wrongful copying—the court used the terms “illicit copying” and “unlawful appropriation”—the court said:

On that issue [wrongful copying] the test is the response of the ordinary lay hearer; accordingly, on that issue, ‘dissection’ and expert testimony are irrelevant.

This seems wrong. The simple fact is that there is nothing ordinary about the ordinary observer. The ordinary observer is not a person to be found and interviewed by the court; it is a legal construct, a lens to be applied when comparing A to B to determine substantial similarity. There are very few cases where the finder of fact would not be assisted by some expert instruction as to what to look for and what to listen for in the comparison of A to B. Moreover, in cases where the subject matter is complex, technical, or simply foreign to the average juror, expert dissection and analysis will be the only relevant criteria.

As the Northern District of Illinois explained in Francescatti v. Germanotta, No. 11 CV 5270, 2014 WL 2767231, at *8 (N.D. Ill. June 17, 2014):

Although dissection and expert testimony is not favored, the judicially created ordinary observer test should not deprive authors of this significant statutory grant merely because the technical requirements of a different medium dictate certain differences in expression. Without deciding the question, we note that in some cases it may be important to educate the trier of fact as to such considerations in order to preserve the author’s rights under the Copyright Act.

Why are courts so hostile to the use of experts in assisting with the analysis of wrongful copying? Partly this is a mistrust born of experience. On the plaintiff’s side, experts will note every conceivable similarity and explain why those similarities are what is really important about the works and why the objective dissimilarities are of no consequence.

In the Ninth Circuit, the extrinsic-intrinsic test allows for expert evidence on the issue of substantial similarity in the objective “extrinsic” part of the analysis. That seems preferable to the more doctrinaire position in the Second Circuit.

The Abstraction, Filtration, and Comparison Test

In Computer Associates Intern., Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992), the Second Circuit recognized that the ordinary observer test generally applied to determine substantial similarity was unlikely to be of much use in the context of computer software. At 713, the court said:

… in deciding the limits to which expert opinion may be employed in ascertaining the substantial similarity of computer programs, we cannot disregard the highly complicated and technical subject matter at the heart of these claims. Rather, we recognize the reality that computer programs are likely to be somewhat impenetrable by lay observers — whether they be judges or juries — and, thus, seem to fall outside the category of works contemplated by those who engineered the Arnstein test.

In place of an ordinary observer test, the court announced an approach that has become known as the “abstraction, filtration, and comparison” test. The AFC test is widely used in computer software cases and occasionally in other contexts.

As its name suggests, the AFC test has three steps. As the court explained in Computer Associates:

… we think that district courts would be well-advised to undertake a three-step procedure, based on the abstractions test utilized by the district court, in order to determine whether the non-literal elements of two or more computer programs are substantially similar. This approach breaks no new ground; rather, it draws on such familiar copyright doctrines as merger, scènes à faire, and public domain. …

In ascertaining substantial similarity under this approach, a court would first break down the allegedly infringed program into its constituent structural parts. Then, by examining each of these parts for such things as incorporated ideas, expression that is necessarily incidental to those ideas, and elements that are taken from the public domain, a court would then be able to sift out all non-protectable material. Left with a kernel, or possible kernels, of creative expression after following this process of elimination, the court’s last step would be to compare this material with the structure of an allegedly infringing program. The result of this comparison will determine whether the protectable elements of the programs at issue are substantially similar so as to warrant a finding of infringement.

Using this AFC approach, the court was able to set aside similarities based on high-level abstractions, standard programming techniques, and any aspects of the software that would have been dictated by external factors. Without this kind of systematic dissection, courts would be far too quick to conclude that two programs performing the same function in the same environment were substantially similar.

Special Considerations Relating to Music

Dual protection for musical works and sound recordings

Copyright law draws a distinction between musical works—which we can broadly think of as composition—and sound recordings, the work that results from fixing a performance of a musical work into a material object such as magnetic tape or some kind of digital storage device. Musical works and sound recordings are both now protected by copyright law, although that has not always been the case. Musical works have been part of copyright since 1831, sound recordings did not become subject to copyright protection in the United States until 1972.

In lay terms you might think of an LP or a CD as a sound recording, but in copyright law the physical medium is a “phonorecord” and the abstract work encoded in that medium is the sound recording. In other words, the phonorecord is the vessel and the sound recording is the content. What makes this terminology even more confusing is that the phonorecord contains both a copy of the sound recording and of the underlying musical work or composition. The Copyright Act defines sound recordings and phonorecords, but not “musical works.”

17 U.S. Code § 101 - Definitions

“Sound recordings” are works that result from the fixation of a series of musical, spoken, or other sounds, but not including the sounds accompanying a motion picture or other audiovisual work, regardless of the nature of the material objects, such as disks, tapes, or other phonorecords, in which they are embodied.

“Phonorecords” are material objects in which sounds, other than those accompanying a motion picture or other audiovisual work, are fixed by any method now known or later developed, and from which the sounds can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device. The term “phonorecords” includes the material object in which the sounds are first fixed.

Commentators often point to the dual protection of music as something of an oddity, after all, novels and motion pictures are each protected by only a single copyright. The oddity is partly historical, but it also reflects an important distinction between different kinds of creative contributions. A musical composition is literally an instruction on how to perform a piece of music. Those instructions can be quite detailed, or somewhat bare bones. Either way, the composition defines the essence of the song and provides more than enough detail to justify copyright protection, but it still leaves room for each performer to express the song their own particular way. Copyright in the musical work accrues to the author of the original song, having a separate copyright in the sound recording provides a narrow exclusive right with respect to a particular recorded performance of the music work.

For example, Jimi Hendrix’s recording of All Along the Watchtower is widely considered to be the definitive version of that song. All Along the Watchtower was originally written and recorded by Bob Dylan. Dylan and Hendrix have distinctive voices and musical styles, but the difference in the recordings goes deeper than this: the Hendrix recording is dominated by a powerful electric guitar sound and multilayered production that gives it a different, arguably more urgent, feel. So perhaps the dual nature of musical copyright is not so odd after all: by providing a separate copyright for sound recordings, copyright law recognizes those contributions and gives musical performers and producers who create a distinctive sound a separate copyright in that production.

Differences between the rights that accrue to musical works and to sound recordings

Sound recording copyright is quite different to musical work copyright. For reasons discussed in a later chapter, there is no general public performance right for sound recordings, only a more limited right for public performance by a “digital audio transmission” which is primarily relevant to the Internet radio and satellite radio. Compare Section 106(4) to section 106(6) of the Copyright Act.

As a result, when the Hendrix version of All Along the Watchtower is played on terrestrial AM/FM radio, only Bob Dylan gets paid. However, when the Hendrix version is played via Internet radio, such as Pandora or Spotify, the copyright owner of the sound recording also gets paid.

Unfortunately, it gets even more complicated. Although sound recordings have existed since the 19th century, they were only made the subject of federal copyright in the United States in 1972. The 1972 law was forward looking and did not grant retrospective protection to recordings prior to February 15, 1972. However, Congress extended copyright protection to pre-1972 sound recordings in the “Classics Protection and Access Act” which was passed as part of the Music Modernization Act of 2018. The method Congress chose to achieve this objective adds some additional complexity to the structure of the Copyright Act. The CPA Act adds a new chapter to the Copyright Act (Section 1401) that provides that the unauthorized use of a sound recording fixed before February 15, 1972 shall be subject to the same remedies as any other form of copyright infringement. Section 1401 of the Act puts pre- and post-1972 sound recordings on an equal footing, but it does not extend a general performance right to sound recordings.

In addition to having a separate and more limited public performance right, sound recordings are also different to musical works in terms of the reproduction right. Section 114(b) limits the scope of the reproduction right in sound recordings to the reproduction of “the actual sounds fixed in the sound recording.” Accordingly, even a “sound alike” recording that was intended to exactly mimic the original sound recording will not infringe the rights of the sound recording. Samples and literal copies in various media still count as copies of the sound recording, but even deliberate imitation in a new recording or performance is considered fair game.

The Seventh Circuit put that limitation to work in Richardson v. Kharbouch, Nos. 24-1119 & 24-2378 (7th Cir. Oct. 16, 2025). Eddie Richardson created a hip-hop beat, “Hood Pushin Weight,” at sixteen, and registered it as a sound recording. He later heard what he believed to be his beat in “Ain’t Worried About Nothin,” recorded by Karim Kharbouch, who performs as French Montana. The difficulty was that Richardson had registered a sound recording and not the underlying musical composition. Affirming summary judgment for the defendants, the court held that a sound recording plaintiff must produce evidence of actual duplication of the recorded sounds themselves; imitation of the composition, however close, will not do. As the court put it, § 114(b) leaves “the world at large … free to imitate or simulate the creative work fixed in the recording so long as an actual copy of the sound recording itself is not made.” Because Richardson had no evidence that anyone had sampled his recording, his claim failed. The case is a reminder that the choice of what to register is not a formality: it determines what the plaintiff must later prove.

What is more, section 115 of the Copyright Act contains a compulsory license which allows any performer to record a new version of a musical work, provided that the song has already been commercially recorded. The compulsory cover license provision has enabled a great deal of creativity over the years, but it also has a troubling racial history. It was common practice in the early 20th century for white producers and record labels to use white musicians to record more or less exact copies of performances by black artists. At the time, White Americans would eagerly consume White “Pop” music but not “Race records.” For more, see Robert Brauneis, Copyright, Music, and Race: The Case of Mirror Cover Recordings ( https://ssrn.com/abstract=3591113); Kevin J. Greene, Thieves in the Temple: The Scandal of Copyright Registration and African-American Artists, 49 Pepperdine Law Review 615 (2022).

Section 115 now has two limbs. The original compulsory cover license still works song by song: a would-be recording artist serves a notice, records the work, and pays a statutory rate. But most music is now streamed rather than sold, and serving separate notices for tens of millions of works was never going to be practical. The Music Modernization Act of 2018 therefore added a blanket license for digital music providers, covering all musical works available for mechanical licensing at once.

Administering it required a new institution. The Act directed the Register of Copyrights to designate a Mechanical Licensing Collective to run the blanket license: to build and maintain the database matching sound recordings to the musical works embodied in them, to collect royalties from the streaming services and distribute them to songwriters and publishers, and to hold unmatched royalties until their owners are identified. The rates themselves are not negotiated but set by the Copyright Royalty Board in periodic proceedings, which is why arguments about mechanical royalties tend to be arguments about regulatory definitions rather than about contracts.

Mechanical Licensing Collective v. Spotify USA Inc., 763 F. Supp. 3d 608 (S.D.N.Y. 2025), shows what that looks like in practice. Spotify Premium had offered unlimited music streaming; in November 2023 Spotify added fifteen hours of audiobook listening a month, without changing the price. In March 2024 it launched a separate audiobooks-only plan, and began reporting Premium as a “Bundled Subscription Offering” rather than a standalone “Subscription Offering” — a reclassification that carries a materially lower royalty, and which the Collective said would cost the music industry some $150 million in the first year alone.

The Collective argued that Spotify had manufactured a bundle for royalty purposes: that audiobook access was not really an “other product or service” within the regulatory definition, and that in any event it had no more than “token value”, since Spotify had originally added it without raising the price. Judge Torres rejected both arguments and dismissed the complaint. The regulation says “other products or services”, not other preexisting or separately sold products or services, and the court declined to read words into it that Congress and the Copyright Royalty Board had not used. Fifteen hours of audiobooks a month was more than token. On the Collective’s own pleaded facts, Premium was a bundle. The Collective moved for reconsideration and lost on that point: the court had not overlooked its allegation that Spotify had previously certified Premium as a standalone offering, but had rejected it as irrelevant. It did reopen the case, to let the Collective seek leave to replead two different claims — that Spotify underpaid on the audiobooks-only plan, and that it miscalculated Premium’s royalties by treating that plan’s price as the standalone value of the non-music component. Both assume the classification holding rather than challenge it.

The case is a useful corrective to the intuition that a compulsory license removes bargaining from the picture. It does remove price negotiation. What it puts in its place is a contest over how a service describes itself, decided by reference to regulatory definitions — and, on these facts, decided against the collective that Congress created to protect songwriters.

The scope of rights in musical works

The preceding discussion is premised on the notion that there is a clear distinction between musical composition and musical performance. This distinction has been challenged by some academic commentators.

Copyright law has generally assumed that musical works should be primarily defined in terms of melody, rhythm, and harmony, but with a central emphasis on melody. This history is extensively reviewed by Professor Joseph Fishman who summarizes: “Throughout much of its history, copyright law has seemingly adopted a fragmented version of pure sonicism. The musical work as legal object amounts not even to a combination of melody, rhythm, and harmony, but more narrowly to melody in particular.” See Joseph Fishman, Music as a Matter of Law, 131 Harvard Law Review 1861 (2018).

Copyright’s traditional focus on melody, and to a lesser extent rhythm and harmony, reflects traditional Western views about music, and it fits comfortably with the archetype of song creation beginning with a single instrument (usually a piano or a guitar). Several commentators have noted the racial and cultural implications of a conception of music that privileges some types of music over others, and some types of musicians over others.

The traditional view, reflected in all but a few cases, means that copyright in a musical work is limited to aspects of melody, rhythm, and tempo and the like that would ordinarily be found in sheet music. See e.g. D’Almaine v. Boosey, (1835) 160 Eng. Rep. 117 (Ex.) (holding that merely rearranging an opera score did not excuse copying because “it is the air or melody which is the invention of the author, and which may in such case be the subject of piracy”); Northern Music Corp. v. King Record Distributing Co., 105 F. Supp. 393, 400 (S.D.N.Y. 1952) (“It is in the melody of the composition — or the arrangement of notes or tones that originality must be found. It is the arrangement or succession of musical notes, which are the finger prints of the composition, and establish its identity.”)

But note that the musical work is not just the primary melody, in Fred Fisher, Inc. v. Dillingham 298 F. 145 (S.D.N.Y. 1924), Judge Hand explained that infringement could occur from copying “any substantial component part, either in melody or accompaniment” and held the defendant liable for copying an accompaniment motif known as an ostinato, rather than the main melodic line. An ostinato is a bit like a bassline, but not the same thing. An ostinato is a repeated musical pattern or phrase, which can occur in any part of the music, in any voice or instrument. A bassline is a sequence of notes played in the lower register (bass), usually outlining the harmonic structure of a piece of music. While a bassline can be repetitive, it doesn’t have to be, so not all basslines are ostinatos. Sheet music contains tempo, key, and key signature, as well, but focusing on these elements would usually be unduly limiting on the scope of the right.

Copyright in Musical Style?

The composition-focused view of the scope of musical works leaves out a great deal that one hears in a typical sound recording. On this view, instrumentation, production techniques, distinctive vocal qualities and timbres (timbre is the character or quality of a musical sound or voice as distinct from its pitch and intensity) are not part of the musical work copyright, although they are entitled to the copyright protection against mechanical reproduction as part of the sound recording right.

The importance of the distinction between musical composition and performance is highlighted by the “Blurred Lines” case. As mentioned in a previous chapter, Blurred Lines is a polarizing case study because many key stylistic features of Got To Give It Up were so blatantly copied by Robin Thicke and Pharrell Williams, and yet, no specific melody, harmony, rhythm or lyrics were actually copied at all. The two songs have similar basslines, a similarly sparse arrangement, a syncopated cowbell rhythm, very similar instrumentation, handclaps and similar vocal style. These similarities seem to have been intentional, but they should not have been enough to sustain a finding of infringement. The vocal qualities, instrumentation and syncopated cowbell and other similarities which make the songs sound so similar are aspects of performance and should not be thought of as part of the copyright in the underlying musical work.

The division between sound recording rights and the rights in the underlying musical composition is not an unfortunate accident. It provides an important source of protection for performers—if their performances are recorded, they cannot be mechanically reproduced without permission—but it also ensures that other performers are free to adopt the same styles, techniques and grooves in new musical works without fear of liability. Without this distinction, the idea-expression distinction would be seriously compromised in the realm of popular music.

Professor Sean O’Connor argues for a broader conception of the musical work that would secure exclusive rights in many stylistic aspects of performance, such as the sounds and impressions created by new devices and performing techniques. O’Connor argues that AI music generators are based on distinct musical styles and thus these styles are sufficiently valuable and specific to merit copyright protection. See, Sean M. O’Connor, AI Replication of Musical Styles Points the Way to an Exclusive Rights Regime, in Research Handbook on Intellectual Property and Artificial Intelligence 65 (Ryan Abbott ed., 2022).

O’Connor actually makes two distinct arguments about musical style. His more modest claim is that stylistic elements within a specific work, such as the vocal style and instrumentation in a particular song, could be annotated in a complete orchestral score and are thus legitimately within the scope of musical work copyright. He calls this “compositional performative style” and argues that by failing to protect “compositional performative style” modern copyright law is merely “privilege the creative contributions of only some kinds of composers—those working well within the Western classical music derived mode of composition abstracted from performative aspects and style—and not others.” O’Connor also argues protecting compositional performative style would bring the law into harmony with market realities and modern production techniques. He points out that pop music today is composed with the groove, sound, and beat first with lyrics and melody often added as an afterthought and that it is often the techniques used to produce a unique sound that are the true innovation. O’Connor’s even bolder claim is that the law should recognize a proprietary style that emerges not just from a single work, but across a collection of works, perhaps even a collection of works spanning a long timescale and different producers and collaborators.

In Music as a Matter of Law, Joe Fishman also notes that the traditional melody focused view of musical works in copyright law is out of step with modern production techniques and modern views about music. But Fishman argues that despite this, for pragmatic reasons, copyright in musical works should be limited to melody and should not include musical style and performance elements. Fishman both concedes and claims too much. A more reasonable position, the position that the caselaw seems to adopt by and large, is that we should understand the musical work as being comprised of those musical elements that would be reproduced in a standard form of composition. That certainly places melody front and center, but it does not suggest an exclusive focus on melody. Fishman concedes too much because the reasons to continue to understand the musical work as elements that would be reproduced in a standard form of composition go beyond simple pragmatism. It relates back to the most fundamental axiom of copyright, the distinction between ideas and expression. Protecting musical style would grant protection at an unhealthy level of generality. The law should not preempt other artists from using a given combination of stylistic elements to create an entirely new work of music.

The traditional distinction between musical composition and performance in copyright leaves the ineffable qualities of personality captured in performance to the right of publicity. In Midler v. Ford Motor Co., 849 F.2d 460, 463 (9th Cir.1988), the Ninth Circuit held that “when a distinctive voice of a professional singer is widely known and is deliberately imitated in order to sell a product, the sellers have appropriated what is not theirs and have committed a tort in California.” See also, Waits v. Frito-Lay, Inc., 978 F.2d 1093 (9th Cir. 1992). Copyright’s traditional distinction between musical composition and performance places innovations in style and technique in the public domain where they belong by virtue of the idea-expression distinction. O’Connor’s proposal for protecting compositional performative style separate from melody flies in the face of the idea-expression distinction. Popular musical styles and techniques should be copied, that is how new genres of music emerge.

The idea-expression distinction is also key to understanding why giving exclusive rights to a broader concept of musical style that emerges across a series of works is untenable. Personality and reputation are emergent properties that develop across an artist’s career, and these are protected by laws in relation to the right of publicity and defamation. However, copyright protects the author’s interests in relation to singular works. Any concept of musical style that can only be identified by considering several works collectively seems far too abstract to merit copyright protection.

Registration and the scope of copyright in musical works

Another reason why no infringement should have been found in Williams v. Gaye is that even if one could mount a theoretical argument that Marvin Gaye’s specific choices regarding instrumentation could have been brought within the scope of the musical work copyright by registering a more elaborate musical notation, this did not in fact transpire. The version of Got To Give It Up registered with the Copyright Office was not a fully realized orchestral score, or even a standard piece of sheet music, it was merely a lead sheet identifying the very basics of the composition. This is significant because when Marvin Gaye first recorded Got To Give It Up, copyright still required registration in advance of publication, registration in turn required deposit, and deposit had to be in the form of written notation, i.e., some form of sheet music. Registration with the Copyright Office was a prerequisite to obtaining federal copyright protection under the Copyright Act of 1909 and previous U.S. copyright acts. Voluntary publication of a work without noticeable registration would make you ineligible for copyright protection and thus permanently consign it to the public domain. The registration requirement had a special significance for musical works because, until 1978, the only form of the work that the Copyright Office would accept for registration was written notation, i.e., sheet music.

Once upon a time it was common practice for music publishers to submit only a stripped-down lead sheet for registration and deposit thus leaving out chord voicings, voice leading, bass line or other aspects of accompaniment. This practice combined with the registration requirement meant that any aspect of a song that was not recorded in the written notation was not part of the song’s copyright. This has proved to be particularly problematic for composers who did not read or write sheet music and thus relied on others to correctly record the full scope of their works. In 2020, the Ninth Circuit took up this issue (among others) in Skidmore v. Led Zeppelin. The extract that follows addresses the deposit issue and also the portion of the court’s reasoning relating to the originality instructions given to the jury.

Skidmore v. Led Zeppelin, 952 F.3d 1051 (9th Cir. 2020) (En Banc)

The claimed portion includes five descending notes of a chromatic musical scale. These notes are represented on the piano as a set of adjacent black and white keys, from right to left. The beginning of Stairway to Heaven also incorporates a descending chromatic minor chord progression in A minor. However, the composition of Stairway to Heaven has a different ascending line that is played concurrently with the descending chromatic line, and a distinct sequence of pitches in the arpeggios, which are not present in Taurus.

Dr. Ferrara testified that the similarities claimed by Skidmore either involve unprotectable common musical elements or are random. For example, Dr. Ferrara explained that the similarity in the three two-note sequences is not musically significant because in each song the sequences were preceded and followed by different notes to form distinct melodies. He described the purported similarity based on these note sequences as akin to arguing that “crab” and “absent” are similar words because they both have the letter pair “ab.” He also testified that the similarity in the “pitch collection” is not musically meaningful because it is akin to arguing that the presence of the same letters in “senator” and “treason” renders the words similar in meaning.

In view of our precedent and accepted copyright principles, the district court did not commit a reversible error by instructing the jury that a limited set of a useful three-note sequence and other common musical elements were not protectable.

An original work may include or incorporate elements taken from prior works or works from the public domain. However, any elements from prior works or the public domain are not considered original parts and not protected by copyright. Instead, the original part of the plaintiff’s work is limited to the part created:

1. independently by the work’s author, that is, the author did not copy it from another work; and

2. by use of at least some minimal creativity.

Notes and questions

(1) It is worth listening to the comparison for yourself. See e.g., https://www.youtube.com/watch?v=ye7hCIWwhGE&ab_channel=RainsonZeppelin.

(2) In Skidmore v. Led Zeppelin, the Ninth Circuit rejected the argument that “the deposit copy is somehow archival in nature and more of a reference point than a definitive filing” and thus that copyright in a musical work extends beyond the sheet music. This had important implications for how the trial was conducted: “Because the deposit copy defines the four corners of the Taurus copyright, it was not error for the district court to decline Skidmore’s request to play the sound recordings of the Taurus performance that contain further embellishments or to admit the recordings on the issue of substantial similarity.” Was this decision correct? Was it fair to the plaintiff?

(3) How did the deposit copy ruling impact the court’s originality analysis?

(4) The deposit copy rule is not a Ninth Circuit peculiarity. In Structured Asset Sales, LLC v. Sheeran, 120 F.4th 1066 (2d Cir. 2024), the Second Circuit took the same view. Structured Asset Sales owns a one-ninth interest in the royalties from “Let’s Get It On,” the 1973 Marvin Gaye song written with Ed Townsend, and alleged that Ed Sheeran’s “Thinking Out Loud” infringed it. Because “Let’s Get It On” was registered under the 1909 Act, the court held that the scope of the copyright is fixed by the sheet music deposited with the Copyright Office in 1973, and not by the commercial sound recording, which contains a good deal that the deposit copy does not. That mattered because the plaintiff’s theory depended on a selection-and-arrangement claim built from a common four-chord progression combined with a syncopated harmonic rhythm. Confined to the deposit copy and after filtering out what was unprotectable, the court held that the combination was too commonplace to support a finding of infringement as a matter of law. Read alongside Skidmore, the case shows how much can turn on a decision made at the registration stage decades before any litigation, and on the once-common practice of depositing a stripped-down lead sheet rather than a full transcription.

De Minimis Copying

Not all literal reproduction constitutes copyright infringement. De minimis non curat lex is a Latin phrase that means: the law does not concern itself with trifles. Under the doctrine of de minimis non curat lex, a court may reject an infringement claim based on literal copying that is insignificant as falling below the standard of a prima facie infringement. In other words, some technical copies are so minor that they do not trigger the reproduction right and thus the court does not need to address the closely related question of the defense of fair use.

For example, in Sandoval v. New Line Cinema Corp., 47 U.S.P.Q.2d (BNA) 1215 (2d Cir. 1998), the court of appeals held that the use of certain photographs in the movie “Seven” was de minimis because the photographs only appeared briefly in 11 shots and when they did appear they were either obstructed, out of focus, or seen from a distance and in poor lighting. However, in Ringgold v. Black Entertainment Television, Inc., 126 F.3d 70 (2d Cir. 1997), the same court held that the use of artwork as set decoration on a television show was not de minimis. Although the artwork appeared for only a few seconds in less than perfect focus, it was nonetheless clearly discernable and in some detail.

Consider the following illustrative case:

Solid Oak Sketches, LLC v. 2K Games, Inc., 449 F. Supp. 3d 333 (SDNY 2020)

there [was] no objective perspective as to how the Defendants’ video game is generally played, or to what extent certain game features can be or are actually utilized, that would allow this Court to make determinations about the choices and subsequent observations of the ‘average lay observer,’ or about the observability and prominence of the Tattoos. The Court [was] thus unable to conclude without the aid of extrinsic evidence that ‘no reasonable jury, properly instructed, could find that the two works are substantially similar.’

Notes and questions

(1) In the case above, the court held that the copying was de minimis, but note that in Catherine Alexander v. Take Two Interactive, 489 F. Supp. 3d 812, 823 (S.D. Ill. 2020) the court held that the de minimis use doctrine did not apply to a video game that copied and displayed WWE wrestler Randy Orton and his tattoos. The court said de minimis was inapplicable because “[the] defense has been successfully invoked to allow copying of a small and usually insignificant portion of the copyrighted works, not the wholesale copying of works in their entirety as occurred here.” Is that correct?

(2) In Gottlieb Development LLC v. Paramount Pictures Corp., 590 F. Supp. 2d 625 (2008) the plaintiff alleged that Paramount Pictures’ use of its copyrighted “Silver Slugger” pinball machine in the background of a single scene in the movie What Women Want infringed its copyrights. The court disagreed. Applying the de minimis doctrine, it found that the Silver Slugger pinball machine appeared only briefly, always in the background, partially obscured, and without any focus or relevance to the plot. Its copyrighted elements, including distinctive designs, were neither clearly visible nor discernible to an average viewer. The use was deemed quantitatively and qualitatively trivial, satisfying the de minimis threshold.

(3) For a similar case, see Lmnopi v. XYZ Films, LLC, 449 F. Supp. 3d 86 (E.D. N.Y. 2020) in which plaintiffs alleged that a mural created by Lmnopi— an American artist dedicated to utilizing her artistic skills as a way to amplify messages emanating from within movements for social, economic, racial and climate justice—was prominently featured without permission for 3.5 seconds in the opening scene of the film Bushwick, released by XYZ Films and distributed by Netflix, Amazon, and Google. The use was de minimis because the mural appears for approximately 3.5 seconds in a 93-minute film, consistently in the background, partially obscured by a red pickup truck and an actress in the foreground, alongside other murals, with no focus, reference, or relevance to the film’s plot.

(4) For a recent case on the de minimis doctrine, see Kelley v. Morning Bee Inc., 2023 BL 337537, S.D.N.Y., No. 21-8420, 9/26/23. In 2021, Apple and Morning Bee released a documentary about Grammy-winning artist Billie Eilish. In a scene lasting forty-three seconds, Eilish arrives at Auckland Airport where the plaintiff’s photos, a series called “Airportraits,” were displayed. These photos were briefly visible for fifteen seconds, albeit unfocused and overshadowed, in the background while a Māori group performed for Eilish. The images represented only 0.18% of the film’s total duration. The court ruled for the defendants on a motion to dismiss on the basis of the de minimis doctrine and, in the alternative, fair use.

(5) Come back to this issue when you have read the material on fair use and query whether and how the doctrine of de minimis use plays a different role to the fair use doctrine. Do we need both? How are they different?

(6) Is music sampling ever a de minimis use? In Bridgeport Music, Inc. v. Dimension Films, 410 F.3d 792 (6th Cir. 2005) the court held that sampling any portion of a sound recording—even a short unrecognizable snippet—is infringing without any room for consideration of the quantity or quality of what was copied. The court said, “Get a license or do not sample.” The ruling has been widely criticized. In VMG Salsoul, LLC v. Ciccone, 824 F.3d 871 (9th Cir. 2016), the Ninth Circuit rejected Bridgeport, holding that very short, unrecognizable samples may qualify as de minimis and do not require a license. Madonna’s song “Vogue” used a 0.23-second horn sample from “Love Break” by the Salsoul Orchestra. The court found the sample was not recognizable to the average listener, and thus too trivial to support a claim.

Richardson v. Townsquare Media, Inc., 174 F.4th 299 (2d 2026)

Notes and questions

(1) Richardson treats the de minimis question as part of the plaintiff’s prima facie case on substantial similarity rather than as a defense. What does that reallocation do to the dynamics of a motion to dismiss? Who now bears the burden, and at what stage can the question realistically be resolved?

(2) The court weighs both the quantitative amount taken and the qualitative “observability” of what was taken — its focus, prominence and detail. If a publisher deliberately selects the single most iconic and recognizable frame of a two-hour film to use as a thumbnail, can that use ever be de minimis on the court’s reasoning, even though the frame is a vanishingly small fraction of the film?

(3) What is the proper denominator? Richardson suggests that § 106(5) may make each individual image of an audiovisual work a copyrighted work in its own right, in which case the screenshot is not a fraction of anything but a whole work reproduced entire. Is that a persuasive reading of § 106(5), and what would follow from it?

(4) In the older cases — Sandoval, Gottlieb, Gayle — the copyrighted work turns up incidentally in the background of someone else’s narrative. In digital journalism a single frame is pulled out and made the focal point of the article. Is the real distinction the medium, as Townsquare argued, or the editorial function: set dressing on the one hand, headline illustration on the other? After Richardson, what is left for online publishers who use screenshots from viral videos without a license, and which doctrines must they now rely on instead?

Statutory Limits on the Reproduction Right

As we have seen, the copyright owner’s exclusive reproduction right in Section 106(1) is defined with reference to the thresholds of substantial similarity and the de minimis doctrine. In addition, the scope of the reproduction right is also limited in specific situations by Section 113, Section 120(a) and Section 114 of the Copyright Act.

Useful articles

Section 113 adds some nuance to the concept of reproduction in the context of articles of manufacture. Section 113(c) provides some latitude for pictures or photographs of useful articles in the context of advertising, commentary, or news reporting. Note that the section only applies “where work has been lawfully reproduced in a useful article that is being offered for sale or distribution to the public ...”

Subsection (d) limits the author’s moral rights in relation to mutilation or destruction to allow for the removal of works of visual art from buildings if the work was installed prior to the effective date of the Visual Artists Rights Act of 1990, or the author expressly consented to such removal, etc.

17 U.S. Code § 113 - Scope of exclusive rights in pictorial, graphic, and sculptural works

(a) Subject to the provisions of subsections (b) and (c) of this section, the exclusive right to reproduce a copyrighted pictorial, graphic, or sculptural work in copies under section 106 includes the right to reproduce the work in or on any kind of article, whether useful or otherwise.

(b) This title does not afford, to the owner of copyright in a work that portrays a useful article as such, any greater or lesser rights with respect to the making, distribution, or display of the useful article so portrayed than those afforded to such works under the law, whether title 17 or the common law or statutes of a State, in effect on December 31, 1977, as held applicable and construed by a court in an action brought under this title.

(c) In the case of a work lawfully reproduced in useful articles that have been offered for sale or other distribution to the public, copyright does not include any right to prevent the making, distribution, or display of pictures or photographs of such articles in connection with advertisements or commentaries related to the distribution or display of such articles, or in connection with news reports.

(d)

(1) In a case in which—

(A) a work of visual art has been incorporated in or made part of a building in such a way that removing the work from the building will cause the destruction, distortion, mutilation, or other modification of the work as described in section 106A(a)(3), and

(B) the author consented to the installation of the work in the building either before the effective date set forth in section 610(a) of the Visual Artists Rights Act of 1990, or in a written instrument executed on or after such effective date that is signed by the owner of the building and the author and that specifies that installation of the work may subject the work to destruction, distortion, mutilation, or other modification, by reason of its removal,

then the rights conferred by paragraphs (2) and (3) of section 106A(a) shall not apply.

(2) If the owner of a building wishes to remove a work of visual art which is a part of such building and which can be removed from the building without the destruction, distortion, mutilation, or other modification of the work as described in section 106A(a)(3), the author’s rights under paragraphs (2) and (3) of section 106A(a) shall apply unless—

(A) the owner has made a diligent, good faith attempt without success to notify the author of the owner’s intended action affecting the work of visual art, or

(B) the owner did provide such notice in writing and the person so notified failed, within 90 days after receiving such notice, either to remove the work or to pay for its removal.

For purposes of subparagraph (A), an owner shall be presumed to have made a diligent, good faith attempt to send notice if the owner sent such notice by registered mail to the author at the most recent address of the author that was recorded with the Register of Copyrights pursuant to paragraph (3). If the work is removed at the expense of the author, title to that copy of the work shall be deemed to be in the author.

(3) The Register of Copyrights shall establish a system of records whereby any author of a work of visual art that has been incorporated in or made part of a building, may record his or her identity and address with the Copyright Office. The Register shall also establish procedures under which any such author may update the information so recorded, and procedures under which owners of buildings may record with the Copyright Office evidence of their efforts to comply with this subsection.

Architectural works

Section 120(a) limits the scope of copyright in architectural works to ensure that photographs and other representations of buildings ordinarily visible by the public do not infringe any copyright in the underlying architectural work. The same section also allows the owners of the building embodying an architectural work to make changes to the building without infringing the right to make a derivative work based on the copyrighted work in section 106(2).

17 U.S. Code § 120(a) - Scope of exclusive rights in architectural works

(a) Pictorial Representations Permitted. —

The copyright in an architectural work that has been constructed does not include the right to prevent the making, distributing, or public display of pictures, paintings, photographs, or other pictorial representations of the work, if the building in which the work is embodied is located in or ordinarily visible from a public place.

(b) Alterations to and Destruction of Buildings.—

Notwithstanding the provisions of section 106(2), the owners of a building embodying an architectural work may, without the consent of the author or copyright owner of the architectural work, make or authorize the making of alterations to such building, and destroy or authorize the destruction of such building.

Note that § 120(a) is narrower than it first appears. It permits pictorial representations of a constructed building; it does not permit every two-dimensional depiction of the design. In Designworks Homes, Inc. v. Columbia House of Brokers Realty, Inc., 126 F.4th 589 (8th Cir. 2025), real estate agents drew simplified floorplans of houses in order to market them for resale. The Eighth Circuit reversed a grant of summary judgment resting on § 120(a), holding that a floorplan is not a “pictorial representation” within the meaning of the section. The agents won all the same, but on fair use: the floorplans served an informational purpose the designs themselves lacked, and any substitution in the resale market was, as the court put it, “substitution blessed by copyright’s first sale doctrine.”

Limited Protection for Sound Recordings

Section 114(b) limits the scope of the reproduction right in sound recordings to the reproduction of “the actual sounds fixed in the sound recording.” Accordingly, even a sound-alike recording that was intended to exactly mimic the original sound recording will not infringe the rights of the sound recording, as such. This limitation does not apply to musical works, so the sound-alike recording may still infringe the reproduction right with respect to the underlying composition.

17 U.S. Code § 114 - Scope of exclusive rights in sound recordings

(b) The exclusive right of the owner of copyright in a sound recording under clause (1) of section 106 is limited to the right to duplicate the sound recording in the form of phonorecords or copies that directly or indirectly recapture the actual sounds fixed in the recording. The exclusive right of the owner of copyright in a sound recording under clause (2) of section 106 is limited to the right to prepare a derivative work in which the actual sounds fixed in the sound recording are rearranged, remixed, or otherwise altered in sequence or quality. The exclusive rights of the owner of copyright in a sound recording under clauses (1) and (2) of section 106 do not extend to the making or duplication of another sound recording that consists entirely of an independent fixation of other sounds, even though such sounds imitate or simulate those in the copyrighted sound recording. The exclusive rights of the owner of copyright in a sound recording under clauses (1), (2), and (3) of section 106 do not apply to sound recordings included in educational television and radio programs (as defined in section 397 of title 47) distributed or transmitted by or through public broadcasting entities (as defined by section 118(f)): Provided, That copies or phonorecords of said programs are not commercially distributed by or through public broadcasting entities to the general public.