Call Now To Get Started (615) 734-1188 [email protected]

Songwriters: The Ninth Circuit Might Actually Have Heard You

When the Blurred Lines verdict was reached—and again when it was upheld, twice—there were many stories about how the decision cast a pall on songwriting. What if I, a songwriter, accidentally wrote the something kind of sort of the same as a song from the past, even if I never heard the song (that I can recall)? Can I be liable for copyright infringement by accident? If I don’t know what a “constellation” of elements is, how can I avoid it?

The early answer was: Well, yes, there is such a thing as “unconscious” infringement. Copyright law doesn’t have—and never has had—any kind of “intent” requirement. But don’t worry because copyright only protects against copying, and if your work isn’t that similar and no one can show you had any real access to the underlying work, you’re safe. Blurred Lines was a perfect storm of access because one of the songwriters admitted to wanting to replicate the underlying song. You won’t do that.

Then came—no, not the Stairway to Heaven case—but the Dark Horse case. That case was the nightmare scenario. Minimal proof of access (maybe heard it on the radio?), with only minor similarities between the songs, yet the jury somehow found infringement. The judge, thank goodness, threw out the jury’s verdict, though on different (and not entirely reassuring) grounds. But it’s cold comfort when even Katy Perry must go all the way to a jury verdict and beyond. Few songwriters have those kinds of resources.

This stock photograph is entitled “Songwriter Tools of the Trade.” The whiskey is for understanding copyright law. ID 126277677 © Ttempleman |

Innovation at The Cheesecake Factory

What struck songwriters the most was how little overlap it took to constitute infringement. Though many of the articles referenced the Sam Smith/Tom Petty dust-up, at least that was a recognizable tune. Blurred Lines was about a “constellation” of high-level musical elements. Stairway to Heaven was a few measures, but the overlap was minimal. Dark Horse was about even less. The smaller the sample size, the easier to accidentally “infringe” on a previous work.

So: what is the least number of notes that two songs can have in common before we say one song might infringe the other? Eight?

The Stairway to Heaven decision partially addressed this problem. As I explained last time, it’s no longer enough for two songs (or any works) to just have a set of random elements in common. Instead, the elements being asserted by the copyright holder must be aesthetically related through the concept of “selection and arrangement,” and that same “selection and arrangement” will have to be present in the accused work. Thus, it won’t be enough for the two songs to have, say, five otherwise non-protectable musical elements in common. The two songs will have to have selected and arranged those elements to achieve a similar aesthetic effect.

But there’s another problem. In popular music, the songwriters said, you just can’t get that much separation from other songs. Indeed, one might (somewhat cynically or snobbishly) point out that popular music is defined by its familiarity, not its innovation. Or, perhaps more kindly, the creativity lies in coming up with something that’s both fresh and familiar. It’s like making 100 different desserts out of the same five or six common ingredients—for The Cheesecake Factory. They won’t all be winners, and many will be difficult to distinguish from each other, but every once in a while, some slight change to one ingredient in relation to another will produce something special (but not too special).

One challenge is that any rule we come up with must apply to all forms of creative endeavor. This is because copyright law doesn’t generally distinguish between different types of creativity. And they are diverse! Music all by itself is very diverse. Copyright law has to cover not just popular music (whatever that is) and jazz, but also a Balinese gamelan compositions, Chinese operas and any other type of music made by human beings anywhere. But copyright law must also cover paintings, television shows, novels, product guides, computer programs, sculpture, industrial design, architecture, photography—all of which have their own (often massive) diversity.

Minting Thin Copyright

So our answer can’t be: “eight notes.” Or even “eight musical elements.” We need a rule that will scale with any work, in any genre, in any “type.” And, thanks to the some of the amici, the Ninth Circuit in the Stairway to Heaven case gave us one, I think (page 48, footnote 13):

But to be clear, we do not recognize a separate, heightened standard for proving actionable copying. The standard is always substantial similarity. Of course the degree of overlap in original expression that is required for the similarity to be substantial is determined by the range of possible protectable expression. More similarities are required to infringe if the range of protectable expression is narrow, because the similarities between the two works are likely to cover public domain or otherwise unprotectable elements. Thus, for works where there is a narrow range of available creative choices, the defendant’s work would necessarily have to be “virtually identical” to the plaintiff’s work in order to be substantially similar. We have at times described this result as the work having a “thin” copyright. (“A factual compilation receives only limited copyright protection.”). A selection and arrangement copyright is not always thin.

(I’ve removed all internal citations.)

OK, so some context and background. Led Zeppelin and (especially) some of the amici1I.e., friends of the court. were arguing that (1) selection and arrangement is entitled to only “thin” copyright, and (2) “thin” copyright requires a heightened showing of similarity. The Ninth Circuit seems to reject that, but really, it accepts it—it just positions that argument in a broader rule related to “range of expression.”

The classic instance of “thin” copyright is something like a phone book, but arranged in an order only slightly more creative than alphabetical. Or consider the hypothetical I gave last time where I select some but not all 14th Century English literature based on what I think is best or what I think is more important for students to know. That’s a protectable selection (even without any arrangement!). But we say the protection is “thin” because the protection applies only to the arrangement, not to the “meat” of work you produced.

Let’s say the anthology has 14 works of 14th Century English literature. If you publish an anthology of the same 14 works, you’ll probably infringe my copyright. But how many works must overlap before you’re no longer infringing? 12? Ten? Eight? What else would you like to know before answering that question?

Home on the Range of Expression

What we’d would like to know is: what is the range of expression? If there were only 14 works to choose from, then obviously I don’t even have a copyright because there would be no selection! What if there were, say, 20? What if ten of them were generally regarded as the most important and are nearly always anthologized? If you also chose those ten, then your range of expression is severely restrained: just four out of ten works. I think mathematically, you have 5040 different combinations (10 x 9 x 8 x 7), which sounds like a lot, but consider if we could choose 10 works out of 100: 100 x 99 x 98 x 97 x 96 x 95 x 94 x 93 x 92 x 91 is bigger than my calculator can do, but it’s roughly 1020 combinations. And that’s just for an anthology of public-domain works! When it comes to “creative choices” in the arts, they run into the millions, billions, unfathomably large numbers.

So, returning to the 10 + 4 anthology, I’d say you can avoid infringing my copyright in the anthology by swapping out just one of the four “in play” works. I.e., you would only infringe my copyright by using the exact same 14 works. The copyright would be that thin.

Now let’s consider if there were more works for me to choose from. Or if the consensus about the ten “most important” works were much weaker. My range of expands, and my copyright gets “thicker.”

You’ll notice that there were a number of factors that might impinge on your freedom of expression (so to speak). First was just the sheer number of works to choose from—a kind of maximum possibility. Second, though, was social: what the market expected from your anthology. Ten of the works I “chose” were dictated to me by the market (or social/professional expectation), so I didn’t really “choose” them at all.

The same with music. As songwriters kept pointing out, there are only so many notes, and popular music can only have so many measures. And not all note combinations sound pleasant. And not all pleasant combinations are acceptable to a wide, popular—or, at least, to the gatekeepers. The same with rhythm. And what you end up with has to be sing-able, probably on a first or second try.2Do you think Katie Perry is going to have any patience with your dotted-triplet-grace-note-octave-jumping stuff?

The Ninth Circuit rejected a rule that selection-and-arrangement always means thin copyright. Indeed, by my examples above, you can see that “thinness” and “thickness” are gradients, not a category. And you can certainly imagine very thick copyrights in selection and arrangement.

Ditch the Metaphor

The rule seems to be: the wider the range of expression, the “thicker” the copyright, and the “thicker” the copyright, the broader the scope of protection. You could almost say: the wider the range of expression, the wider the scope of protection. You can leave out the “thick/thin” metaphor if you want.

And this makes a certain amount of sense. If there’s a wide range of expression, then similarities are less coincidental; there’s more “space” for legal work; there is perhaps more hunger for variety.

This rule is scalable and flexible. It works in popular music, as much as it works in industrial design (where similar restrictions are present), as much as it works in representational art, abstract art, novels, “genre fiction,” dance, architecture (ditto), maybe even software (ditto, but in spades).

Juries will intuitively understand how most artists work, so there won’t be much need for special proof about range of expression. The parties themselves can supply much of what the jury needs to know. Experts can be brought in to tell juries about less familiar types of works. And experts will sometimes still be needed to explain how what a range of expression is narrower than it seems because of commonalities and external pressures (e.g., popular music, certain types of dance, software).

Big note of caution: The bit I quoted above is in a footnote and is explicitly dicta (which is to say, it’s not binding authority). I do think it’s going to be highly persuasive authority, though it needs some fleshing out.

Thanks for reading!

Rick Sanders

Rick is currently General Counsel for Software Freedom Conservancy. Previously, he has been practicing law as an intellectual-property litigator since 2000.


    1 I.e., friends of the court.
    2 Do you think Katie Perry is going to have any patience with your dotted-triplet-grace-note-octave-jumping stuff?