Making Room in Copyright Law for Creativity

A Critique of the NMPA’s and RIAA’s Joint Brief in the “Stairway to Heaven Case”

As I explained last time, in their joint friend-of-the-court brief in the “Stairway to Heaven” case, the National Music Publishers Association (NMPA) and the Recording Industry Association of America (RIAA) raised two substantive points: 1) the “Inverse Ratio” rule should be abolished because it is more distracting than useful, and 2) the copyright in the arrangement of non-protectable elements should be “thin,” i.e., can only be infringed if the accused work is nearly identical with it.

Although I’m persuaded by the brief’s reasoning on the “Inverse Ratio rule,” I don’t understand why the NMPA and RIAA were so concerned about the rule. Their interest in the “Stairway to Heaven” case is to make sure copyright in musical compositions isn’t so broad that it discourages the writing of new songs. The members of both organizations depend on the production and recording of new music, after all. But the “Inverse Ratio rule” only increases the chances of infringement on the margins and doesn’t directly affect the incentive to create. Perhaps your new song is more likely to infringe, but your new song will also have broader scope.

But the other issue goes directly to the process of writing new songs because it relates to the practice of using what’s already “out there” in your own way. The NMPA and RIAA propose that, when you make a new song out of non-protectable elements, you can still have a copyright in your selection and arrangement of those elements, but that copyright is “thin.” We say copyright is “thin” because we’re only protecting the selection and arrangement of the elements, not the elements themselves. Such a copyright is naturally a lot less robust than if the elements were also included.

Normally, when we talk about making room for creativity in copyright law, we’re talking about fair use—and for good reason. Fair use is all about building on the works of others. But, as you’ll see, how we treat substantial similarity also plays a role in balancing incentivizing creativity against stifling creativity.

The Barnes Foundation: A Thought Experiment

Have you ever been to the Barnes Foundation art museum Philadelphia? Albert Barnes was genius at collecting fine art, particularly impressionist, post-impressionist and early modernist works. He also had eccentric and anti-elitist ideas about art and how it should be experienced. He didn’t believe in categorizing artwork by style or period. Instead, he mixed different styles in the same room: old masters juxtaposed with Native American crafts, American furniture and American folk decorative art (like weather vanes). These he’d arrange in a way he thought deepened the viewer’s experience with the works. The arrangements might have been eccentric, but they were neither arbitrary nor systematic¹, but reflected Barnes’ own aesthetic and beliefs.

Barnes didn’t have a copyright any for the works he purchased and displayed. But he did have a copyright in his own original way of selecting artwork and arranging them on a wall and in a room.² More examples are here and here and here. If you could get ahold of copies of the artwork, you could (probably) arrange them in your own way without infringing Barnes’ copyright. To infringe Barnes’ copyright would have to look a lot like Barnes’ selection and arrangement. Some minor changes wouldn’t prevent infringement, for example, swapping one weather vane for another, or one Navajo rug for another, or one Old Master for another of a similar size. Altering the arrangement slightly—say, moving two paintings in closer by six inches—also wouldn’t prevent infringement. But changes much more significant than that would probably be infringing.

There are a couple of other forces that “thin” out the copyright in Barnes’ otherwise highly original selection and arrangement of artwork. First, he can’t own the idea of combining, say, Old Masters with American folk ironwork. If you have some Old Masters and American folk ironwork laying about, you can arrange them on a wall almost however you like—just not in a way similar to Barnes’ way. Second, there are conventional ways to display artwork, and no one owns the copyright in those conventions. They are, to use the misleading term of art³, scènes à faires. Thus, if your arrangement was straight (i.e., not crooked or placed at a jaunty angle) and had at least one row of artwork at eye level, those elements wouldn’t contribute meaningfully to any infringement analysis.

But Barnes had an advantage that a songwriter doesn’t. At least the elements Barnes was using were themselves complex and distinctive from one another. By contrast, the songwriter’s toolbox is much more bare. A painting for Barnes would be an individual note for a songwriter.

Imagine, instead, if Barnes was limited to squares, which can only be one of twelve colors. They can be any size larger than, say, 6 inches a side, so long as their sizes are ratios of each other. Also assume that the squares had to fit into larger squares, which are all the same size. Also assume that some of the colors look really bad next to each other, but others look good. And the wall is relatively small, say ten feet by 20 feet. There are, of course, millions of combinations of colored squares that would fit on the wall. And Barnes’ arrangements would still be fairly eccentric because he really, really didn’t care what the market or the academy thought of him. So let’s remove Barnes from the equation—thank you for your work, Mr Barnes!—and replace him with someone much more conventional—in fact, someone whose livelihood depends on getting as many people to like his black-and-white arrangements as possible. And also assume making these black-and-white arrangements is common, potentially hugely profitable, and has been practiced for decades, if not longer.

There are still lots of ways the squares can be arranged, but 1) not nearly as many as before, and 2) they’re going to start clustering in patterns. There will be instances of innovation, but mostly the square arrangers will be working within industry conventions. When livelihoods are involved, people will play it safe.

Now, let me ask you:

1. How many squares will it take to make an “element” of the arrangement? We’ve agreed that a single square is too small. Ten squares? Could it be as small as three, if they’re unusual? Does it depend on more than just the number of squares?
2. If any of those elements are conventional, how many must be taken before it’s too much? Does it matter if the elements are themselves in a pattern, and that pattern was or wasn’t also taken?

The NMPA and RIAA answer question (1) as “about seven.” They answer question (2) as “Lots, and the pattern must be taken as well.” In other words, if a musical element isn’t convention—a scène à faire, in the parlance—then it must be at least seven notes long before it can be considered by a jury as something others shouldn’t be allowed to take. But if the elements are conventional, then nearly the entire song must be taken to be considered infringement.

Right Passengers, Wrong Vehicle

You can see why the NMPA and RIAA are concerned. Songwriters, perhaps more than other creators (with the possible exception of architects) must draw from a rich tradition that tells them and their audience what is pleasing and what isn’t, at least, if they want to make a living. If songwriters start to worry that drawing from this tradition might be infringing, they won’t write as many, or as good, songs. And those with the resources to fight a copyright lawsuit will have an edge over those who don’t, reducing the diversity (and probably thus the quality) of songs.

You can also see why the NMPA and RIAA don’t really see the benefit of broader protections for musical compositions. They’re concerned about the identical copying of songs. What they want more than anything are more effective ways of stopping copying. They have little interest in the marginal stuff. If Sam Smith lifts a melody from Tom Petty and Jeff Lynne, Messrs. Petty and Lynne—or, more precisely, their publishers—can take care of themselves.

I might be sympathetic with the NMPA’s and RIAA’s argument here, but this case isn’t the right vehicle for it. (Alas, what case would be?) Although they argue that protection for the selection and arrangement of non-protectable elements should be “thin,” that’s not what they ask the Ninth Circuit to do in their brief. Instead, they ask the Ninth Circuit to reverse itself⁴ and re-instate the trial court’s jury instructions. If they wanted to implement their “thin” argument, they’d actually need to ask the Ninth Circuit to uphold the reversal but change the instructions back to the trial court.

Allow me to explain. As you know, the Ninth Circuit reversed the trial court in part because the trial court failed to instruct the jury that copyright protects the selection and arrangement of non-protectable elements, then compounded the error by striking out an important instruction, found in the pattern instructions, that copyright will protect works even if they aren’t “new or novel.” As I explained, the trial court was flat-out wrong to do this. It’s almost inexplicable, except that copyright law is hard. Further, the NMPA and RIAA know the trial court was wrong. Remember, they agree that there is copyright in the selection and arrangement of non-protectable elements, just that the copyright should be “thin.” They’d have to agree that it was wrong to leave out the instructions.⁵

Anyway, when a case is reversed because of bad jury instructions, it gets “remanded” (i.e., sent back to the trial court) for a new trial, this time with better jury instructions. Of course, the point of the remand is to avoid repeating mistakes, so the appellate court must issue directives to the trial court. Usually, the directives take the form of not being “inconsistent with this opinion. In this case, however, the Ninth Circuit was considerably more specific: ”[T]he district court erred both in the formulation of the originality jury instructions and in withholding a selection and arrangement instruction.”

Instead of asking the Ninth Circuit to reverse its own reversal—i.e., affirm the trial court—the NMPA and the RIAA should have asked the Ninth Circuit to modify its directive to the trial court on remand. It should have asked the Ninth Circuit to direct the trial court to include an instruction on the selection and arrangement, but to also include an instruction that infringement of the selection and arrangement of non-protectable elements requires near identical copying.

Problem of Proof

The NMPA and RIAA’s brief does nothing—and can do nothing—about the biggest problem about music and copyright exposed by the “Blurred Lines” and “Stairway to Heaven” cases: proof. You’ll notice that, in the example of the Barnes Foundation, we assumed (correctly) the elements themselves—the paintings, the ironwork, the furniture—were not protectable. That’s obvious in our example: everything is so old that it’s all in the public domain. But it’s not so obvious in the context of popular music.

When we discuss what is and isn’t protectable, and how “thick” or “thin” copyright protection should be, what we’re talking about, legally, is originality (and, to a lesser extent, expressiveness). A work is not protectable by copyright unless it is “original.” Originality just means that the work was a product of the author’s mind. It’s an easy standard to meet, but it excludes things that the author didn’t create herself, such as facts. More germane to our discussion here, it also excludes stock tropes, figures, characters, situations, subroutines, techniques, and so forth. These are known by the misleading term scènes à faire, but you can think of them as a heritage shared by members of a class of creators. Nobody can own a boastful soldier, which has been a stock character since the days of Plautus.

In the normal case, it’s actually up to the defendant to prove lack of originality.⁶ Music, unfortunately, is chocked full of its own varieties of scènes à faire, but the public—i.e., the jury—is by and large ignorant of them. We consume lots of popular music, but we don’t really study it. We really don’t see it the way songwriters see it.

Proving that a song lacks originality—or, more precisely, the bits you are accused of infringing lack originality—almost certainly means expert testimony. Someone experienced with the songwriting business, or, better yet, intimately familiar with the history of music, needs to explain to the jury that such-and-such element is so common (as of the date the underlying song was written) that all songwriters have equal access to it. This is (1) expensive, so not all defendants can afford it; and (2) harder than it looks. In the “Blurred Lines” case, the Thicke/Williams parties presented just this sort of evidence⁷, and yet, the jury seemed not to believe it. Or perhaps they just got mixed up because their jury instructions, while accurate, were terrible.

So, it’d be great if copyright protection for selection and arrangement of non-protectable elements is “thin,” but that’d be no good unless you can prove those elements are protectable. In other fields, our experience is enough to at least strongly hint at what is a scène à faire, but music seems to be an exception.

The Hobby Horse Isn’t Going to Stay in the Stables

Do you know what might really help? Separating copying from misappropriation. Exactly what the Ninth Circuit is trying to do in sending the “Stairway to Heaven” case back for another trial. Crazy, huh?

Here’s why. When we want to know whether something’s been copied, the existence of non-protectable elements in both works is actually relevant. Recall my example (from a previous post) of a witness viewing me writing on a pad of paper while reading a copyrighted book, and it turned out what I copied was exactly like a detailed table of factual findings in the book. That is powerful evidence of copying! I clearly had access to the book, and I can’t explain away the detailed table. But I didn’t infringe! Because facts aren’t protectable! That demonstrates the crucial difference between copying and misappropriation (i.e., taking what isn’t mine to take).

Most courts mix copying and misappropriation up. They think all you need to prove is copying, and in the absence of direct evidence⁸, you can prove copying by proving access and “substantial similarity.” But “substantial similarity” does double duty here: it’s made to prove not only copying, which can include non-protectable elements, but also misappropriation, which must filter out non-protectable elements.

Consider. Let’s say I go to the Barnes Foundation and make note of what Mr. Barnes had placed on one wall of the gallery. I then obtained (legally) exact reproductions of everything on that wall. I then placed the reproductions on a different wall in a totally different way than Barnes did. You can say I “copied” because I had access, and it’s impossible to explain the coincidence that I used only works found on that one wall. But I didn’t infringe because, while I took his selection, I didn’t take his arrangement: the creative way he arranged the works.⁹ I think we’d all agree that the selection of perhaps a dozen or so works isn’t protectable.

Under the majority framework for infringement, however, my selection and my arrangement of the works won’t be separated for the jury. They both go into the same bucket of “substantial similarity.” Having found “copying,” there’s nothing to stop the jury from finding infringement. Under the correct framework, however, the jury would be instructed to start over when considering evidence of misappropriation. Although you might rely on the same evidence in considering copying and misappropriation, you have to look at the evidence differently for each.

Let’s go back to “Blurred Lines.” Do you think it was possible the jury got a little confused about what Thicke and Williams supposedly did wrong? So much of that evidence went to proving copying, but I’ll bet it also went to proving misappropriation. And, since they were just following Ninth Circuit law (at the time), they weren’t necessarily wrong, either.

That’s enough for now. Thanks for reading!