Oh! Inverted Ratio!
On Monday, the National Music Publishers’ Association (NMPA) and the Recording Industry Association of America (RIAA) jointly filed a friend-of-the-court34Do you prefer Latin? The cool lawyers call it “amicus curiae.” Except, when pressed by a Roman centurion, they can never decline “amicus”: amicus, amici, amico, amicum, amico; amici, amicorum, amicis, amicos, amicis. And if you really need the vocative, you can say, “Et tu, amice?” brief with the Ninth Circuit, urging the court to reconsider its recent decision to send the “Stairway to Heaven” case back for a new trial. As I explained previously, the case was remanded mostly because of bad or misleading jury instructions. Though I felt bad for Led Zeppelin, I was excited that the Ninth Circuit was using the opportunity to fix its crazy test for substantial similarity.
The NMPA and RIAA didn’t read the case that way at all. Although the organizations usually advocate for stronger35And, to be honest, longer. copyright, they don’t advocate for broader copyright. They are, appropriately, worried that too much copyright can interfere with creativity, and thus, with the profitability of the industries they represent. They also recognize that music is particularly tricky in this regard. Therefore, they ask the Ninth Circuit to (1) reject the “inverse-ratio” rule entirely; (2) limit the application of the rule that copyright can protect the selection and arrangement of otherwise non-protectable elements; and (3) reinstate the trial court’s refusal to let the jury watch Jimmy Page listening to a recording the asserted song. I’m only going to address the first two points because the third point involves the intersection of an evidentiary rule with an artifact of the copyright act that preceded the current copyright act and thus doesn’t shed a lot of light on current copyright issues.
The organizations’ interest in reconsideration of these issues might seem obvious, but it’s worth tracing, if only because the music industry’s relationship with copyright law is kind of opaque. Most popular music that you listen to is actually subject to two separate copyrights, one for the songwriters and one for the recording artists36And maybe sound engineers, producers, etc.. The songwriters have copyright in the musical composition, so their rights extend to all performances and recordings of the song.37E.g., if band X and band Y both performed songwriter A’s song, both X and Y would owe A royalties. The artists’ copyright is limited to the artistry of the particular sound recording.38E.g., if band X and band Y both recorded songwriter A’s song, and bar R played X’s version publicly, then R would owe royalties to X and A, but not to Y. NMPA represents “publishers,” which is a non-intuitive way of saying the owner of the songwriter’s copyright.39For historical reasons, songwriters give away their copyright to these publishers in exchange for 50% of revenue generated by exploitation of the song, though publishers have no obligation to try to exploit a given song. This is so ingrained in the industry that even when smart songwriters want to control their own songs, they form their own publishing companies that they control and assign the copyright to those publishers, rather than just keep the copyrights themselves. Publishers aren’t songwriters, but they definitely want songwriters to keep writing songs. The RIAA represents the “music labels,” which are the owners of copyrights in the sound recordings.40Artists routinely assign their copyrights to their labels. This makes a bit more sense because labels do much more to exploit the copyrights, though they don’t do nearly as much as they used to. The record labels don’t make the music, but they want songwriters to keep writing good songs.
Don’t Blame the Inverse-Ratio Rule. It Came from a Bad Framework.
At its most general, the “inverse-ratio” rule holds that, in determining infringement, the more similarity there is between the accused work and the copyrighted work, the less access that must be proved, and vice-versa. Depending on the framework you’re using to determine infringement, the inverse-ratio rule can be justifiable or disastrous.
There are two frameworks to determine copyright infringement. I’m not even going to beat around the bush about which I prefer. There’s the Good one, and there’s the Bad one. Both frameworks try to reconcile two fundamental concept of copyright law. First, you must copy the underlying work to infringe it, no matter how similar your work is to the underlying work. In theory, you can by utter coincidence exactly reproduce someone’s work (though that’s unlikely, as explained below). Second, not all copying is illegal. You can, for example, copy facts and ideas.
- The Good Framework posits that copying is a separate element from what we might call “misappropriation” (i.e., taking stuff you shouldn’t take), and similarity plays different role in each. For copying, the copyright owner must prove that you at least had access to the copyrighted work, or else how could you have copied? The copyright owner must also prove some similarity between the two works, or else why are we even in court? Courts sometimes call this similarity “probative similarity.” Once copying has been proven, then we can look at what was taken, and whether it’s too much of material that copyright law actually protects. Historically, this inquiry has been called “substantial similarity.” Also, historically, substantial similarity is
- The Bad Framework kind of mushes up the two applications of similarity, I guess because the word “similarity” is used in both instances. Under the bad framework, the copyright owner need prove only access and substantial similarity. Substantial similarity ends up doing a kind of double duty, both establishing copying and misappropriation.
- Oh, there’s an Even Worse Framework, which is used exclusively by the Ninth Circuit. It’s like the Bad Framework, but it splits substantial similarity into “extrinsic” and “intrinsic” tests, both of which must be satisfied. You can think of the extrinsic test as probative similarity, looking longingly at its partner, Access, across a chasm, always knowing it’s in the wrong place. The intrinsic test is just what everyone else in the world would call “substantial similarity.”
If you apply the inverse-ratio rule to the Bad Framework (or the Even Worse Framework), you wind up with some absurdities. If I own a copy of a movie but you do not, and we both write screenplays that bear a certain amount of similarity to the movie, I might be liable while you might not be, even though the only difference is the degree of access, not on what was taken.
Applying the inverse-ratio rule to the Good Framework is justifiable. Indeed, it’s natural. If we’re only worried about the actual fact of copying—and not with taking—then it makes sense that we require less proof of access with greater proof of probative similarity and vice-versa. If I not only owned a copy of a book but was seen reading it while I was working on my own novel, then it would be enough to prove copying if my novel included many of the same themes as the book. It’s an entirely different question, however, whether taking themes is infringement: I might have copied but what I took was OK to take. Similarly, if my novel was nearly word-for-word the same as the book (what copyright lawyers would call “striking similarity”41I.e., so many similarities that they can’t be coincidences., so it would take very little access indeed to prove copying, perhaps just proof that the book was available in a nearby town’s public library. I would probably also be infringing (I probably took a lot more than just a few themes!), though that’s still a separate question that just happens to rely on the same proof.
In my last post on this subject, I said the inverse-ratio rule was “appropriate” under the Good Framework (and “perverse” under the Bad Framework). Thinking about it, I’d say it’s more accurate to say the inverse-ratio rule is justifiable under the Good Framework. I don’t think it makes much difference whether you specifically instruct the jury about the inverse-ratio rule or not. If you don’t, the jury will probably apply something very much like it anyway, just without the sheen of mathematical precision implied by the word “ratio.” All you really need is an instruction that you must take into account both degree of access and amount of probative similarity (appropriately defined), and common sense will lead the jury to something pretty close to the inverse-ratio rule.
Throwing the Rule Out with the Framework.
The NMPA and RIAA’s joint friend-of-the-court brief does not argue for the correct application of the inverse-ratio rule. It argues for its complete abolition. That’s because they do not distinguish between the Good and Bad Frameworks. At first, you might think this isn’t surprising. After all, courts often don’t bother to keep the infringement analysis straight. That’s how we ended up with the Bad Framework in the first place.
Then again, in this context, it is surprising that the brief doesn’t even acknowledge the distinction between the Good and Bad Frameworks. Reading the “Stairway to Heaven” opinion, it’d be a little hard for copyright lawyers to miss! It’s one of the opinion’s major themes: rejecting the Bad Framework and establishing the Good Framework.
In the brief, the NMPA and RIAA argue (correctly) that three Circuits have rejected the inverse-ratio rule. What they don’t say is that the three Circuits rejected the rule in three very different ways:
- Second Circuit: In Arc Music v. Lee, which was decided way back in 1961, rejected what plaintiff described as “the Inverse Ratio Rule.” “We fear that counsel with that semantic proclivity natural to our profession have allowed themselves to be seduced by a superficially attractive apophthegm42I had to look this one up, too. I just means a maxim or aphorism. which upon examination confuses more than it clarifies.” What a sentence! What a burn! Importantly, the court noted that an inverse-ratio rule is unnecessary: “And it is not an unnatural step in inference of fact for ease of access to suggest a deduction of copying when similarity is found.” In the end, the court upheld this jury instruction: “To find for the plaintiff you must find that by a fair preponderance of the evidence that the similarities between the compositions and the access inferred are sufficient to reasonably base a finding of copying.” Key here is that the court was using the Good Framework.
- Seventh Circuit: In Peters v. West, curmudgeonly Judge Easterbrook makes a point of deciding not to understand the difference between the Good and Bad Frameworks. He surveys the law, sees that there are two different types of similarity, plus something called “misappropriation,” threw up his hands, and without knowing it, settled on the Bad Framework.43Judge Easterbrook is often linked to his longtime Seventh Circuit colleague Judge Posner. Both joined the appellate bench at around the same time, and both came out of the “Chicago School” of thought that emphasized an economic approach to jurisprudence. But whereas Judge Posner was fascinated by copyright law, Judge Easterbrook seems bored and annoyed by it. To his credit, however, he did not compound his error by introducing the inverse-ratio rule. As we’ve discussed, the inverse-ratio rule makes the Bad Framework even worse.
- Eleventh Circuit: In Beal v. Paramount Pictures, the court rejected the inverse-ratio rule without a lot of discussion. This was just as well, because the Eleventh Circuit used the Bad Framework.
Of these three rejections, only the Second Circuit’s is really worth paying attention to, because only the Second Circuit was applying the Good Framework. And what it said is: you don’t need the inverse-ratio rule. You just need the jury to be aware that they need to weigh both access and probative similarity in determining copying, and leave it to them to weigh the evidence as they see best. The inverse-ratio rule is at best harmless and at worst confusing. It might confuse a jury into thinking there need be no evidence of probative similarity if there is overwhelming evidence of access, which would be absurd. The opposite—so much similarity that access may be assumed, often known as “striking similarity”—is often said to be acceptable.
I think the balance between access and similarity is more symmetrical than the law says, by the way. Yes, striking similarity assumes access because you can’t imagine the existence of the similarities without one work copying the other. But actually, you’re just demanding such a low degree of access that just existing in the culture will be enough. But imagine if the defendant was hermetically sealed and cut off from the world during the production of the allegedly infringing work? Or if the plaintiff’s work was just a handwritten manuscript locked in the bottom drawer of a desk somewhere?
But, in the End, it’s Not Really Worth Saving.
I doubt the NMPA and RIAA are really ignorant or obtuse about the Good and Bad Frameworks. I’ll bet they made the rational decision that the dangers posed by the misapplication of the inverse-ratio rule are outweighed by any benefit of the correct application of the rule. And, let’s be honest, it’s exhausting trying to explain the difference between the Good and Bad Frameworks. Better to just be rid of the inverse-ratio rule in all instances. And, thinking about it, I don’t think they’re wrong!
All the same, I hope the Ninth Circuit ignores them. The “Stairway to Heaven” decision does important work in replacing the Bad Framework with the Good Framework, in one of the two most important circuits.44The other most important circuit, the Second, already uses the Good Framework, but only about half the time. It’s so easy to slip into the Bad Framework. Besides, songwriters aren’t thinking about anything like the inverse-ratio rule when they complain about the chilling effect of the “Blurred Lines” and “Stairway to Heaven” cases. They’re talking about the right to use non-protectable elements in their own compositions, and that is much more like the other substantive legal issue the NMPA and RIAA raise in their joint brief. I’ll examine that issue in a later post.
Thanks for reading!