≡ Menu

Copyrighting all the melodies to avoid accidental infringement | Damien Riehl

Various people have pointed me to this video since it came out two years ago.

Here’s the description:

In the litany of copyright infringement lawsuits, technology lawyer and musician Damien Riehl demonstrates that music is merely math, and has a finite number of possible melodies. If you’ve ever thought a song you like sounded similar to another, the culprit may not be an unethical forger, but rather the limited mathematical musical equations that our favorite artists have to work with. Current copyright law is at risk of severely limiting future music creation and future human creativity. This talk suggests a new way to handle these legal cases. Damien Riehl is a technology lawyer with a B.S. in music. After beginning to code in 1985, and for the web in 1995, he has worked for the chief judges of state and federal courts; litigated for a decade; taught law-school copyright classes; and led teams in software development, digital forensics, proactive cybersecurity, reactive cybersecurity incidents, and world-scale investigations. Damien’s combined experience in the law, technology, and music has inspired his most recent project—copyrighting billions of unique melodies.

See also Riehl’s Tweet about this, and Alexis C. Madrigal, “The Hard Drive With 68 Billion Melodies,” The Atlantic (Feb. 26, 2020). Here’s the beginning of the Atlantic piece:

In an era when millions of songwriters upload music to the internet—and just about any song can be plucked from obscurity by TikTok teens—it seems inevitable that the same melodies end up in different songs. There have been a number of high-profile music copyright-infringement cases, including a multimillion-dollar decision against Katy Perry for her song “Dark Horse.” A jury found that she’d infringed upon the copyright of Flame, a Christian rapper who’d posted a song with the same melody to YouTube, even though Perry insisted that she’d never heard of the song or the rapper. For some musiciansmusicologists, and lawyers, the verdict felt scary; after all, large numbers of songs now live on SoundCloud and YouTube. It became thinkable to ask: Could the world run out of original melodies?

Damien Riehl and Noah Rubin were two of those worried musicians. Riehl is a lawyer who has worked on copyright. Rubin is a coder. They were hanging out after a long day at work when a “a lark, a thought experiment” occurred to Riehl: Maybe they could exhaust all possible melodies—and in so doing, protect musicians from being sued for copying songs they don’t remember hearing.

Soon, they had a hard drive filled with almost 69 billion melodies. …

Now Riehl and Rubin want to release the fruits of that brute-forcing into the public domain. They figure that in a future suit where a musician is hit with copyright infringement, she could point back to the melody on that hard drive as her uncopyrighted inspiration. Their point, ultimately, is that melodies could be seen as math, which is to say facts, and facts cannot be copyrighted.

Yes, sigh, copyright law is problematic, but this video doesn’t really do a good job of showing that it is or really proving anything. Here are some comments, drawn from a Facebook post and some private email discussion.

This story and the coverage is very confused. First: the problem with copyright can’t be solved with a trick or gimmick. Second, Riehl and his partner, Rubin (R&R), don’t seem to oppose copyright on any coherent or principled ground. So what exactly are they trying to prove by this stunt?

Consider: if their ploy works, it shows how stupid copyright is. But if such a workaround is possible, everyone will do it. And then the IP interests and the state would find a way to block it. The problem with copyright is not that we haven’t come up with enough gimmicks to evade it yet. The problem is the system itself.

The problem is it would not protect anyone. Suppose Katy Perry comes up with a new song, Dark Skank. Years later she sues Britney Spears, claiming her song Shiny Pole has a similar melody. Britney’s defense? Well the melody that Shiny Pole and Dark Skank share in common was already on R&R’s thumb drive, melody #78,328,482,221. So what? Under copyright, if Perry came up with Dark Skank independently she has a copyright in it. If she can show Spears copied that tune then that’s copyright infringement. I suppose R&R could claim that they have a copyright also on the melody that their computer generated (or maybe there is no copyright in it; who knows?),1 and grant Spears a license. But maybe they wouldn’t. Or maybe they would charge her a lot for it. Who knows? Did they open-source the melodies their computer generated?

And of course it won’t prove that there is no copyright at all on the melody since it’s “just a sequence of numbers,” i.e. “just a fact.” Virtually every artwork can be digitized and thus represented by a string of binary digits. So what? We already knew this. That has never killed copyright law. Yes, if you really understand how copyright works, and that all copyrightable works are digitizable, this means that copyright law covers numbers; yes, this is absurd and ridiculous; but copyright law always was and is now inherently absurd and ridiculous. We already know this. This stunt proves nothing.

In other words, their argument that “MELODY = MATH = FACTS = UNCOPYRIGHTABLE” is based on a false assumption that the very concept of intellectual property is not fundamentally an arbitrary, self-contradictory state-created scheme designed to enrich a privileged few at the expense of others. But it is. We need to oppose copyright law on principle.

Some interchanges:

Ash Navabi: “By putting all their melodies online, which must include many duplicate melodies of existing copyrighted works, are they infringing on many existing copyrights despite the disclaimer on their website that “All the Music LLC does not claim copyright in, nor does it claim Creative Commons Zero or public-domain designation of, any copyrighted work to which it does not own the copyright or any work that is uncopyrightable”?”

My response: This is a bit confused. I think that if they publish a random work that happens to coincide with a previous work, then either that publication is infringing, or it’s not. I don’t think you are saved from it being infringement, by the fact that you do not “claim copyright.” That is totally irrelevant and based on a misunderstanding of how copyright works. Copyright infringement does not arise from “claims,” it arises from actions—namely, copying (or performing or making derivative works of) a copyright-protected work. It has nothing to do with any “claims” or even “disclaimers.” That said, this experiment might be able to use a fair use defense…. but who knows. The fair use idea is vague and non-objective.

Gerrit Johan Hagelslag: “Is this any different from anyone publishing a dictionary and then thinking that Stephen King will sue them because they made a collection of all the words that he used?”

Yes, it’s different. Words and titles and phrases are too short to get copyright protection. The copyright is in longer arrangements that have some originality. The other problem is that unlike patents, you can in theory have independent copyrights as long as the similar works are both independently authored—it’s just very unlikely someone would independently write a novel that is identical to a previous one. But in that case, both authors have their own copyright. In patent law, publicizing an invention does prevent others from patenting it later, since it’s no longer an invention—even if they independently recreate it. But it doesn’t work this way in copyright. So I’m not clear on the purpose of this stunt.

Suppose R&R have a computer randomly generate, and record, a trillion melodies. So to me, it’s not even clear that there is a copyright in these trillion melodies, since arguably the computer generated it, not a human author. (See the monkey copyright dispute noted above, and recent AI-copyright cases.)

And even if there is copyright, I suppose R&R have copyrights in all trillion melodies (presuming each one is “original”). So what? Let’s suppose in 2023 some musician M writes a song S that is very similar to melody number 7billion in the database (let’s call it S’). So what? M now has a copyright in song S. He can still run around suing people for it. So what that R&R have a copyright in a similar melody S’. So what? They can’t sue M, and M can’t sue them. So what. I suppose R&R could grant free licenses to his work S’ and therefore give a decent defense to everyone M might want to sue for copying S. Hard to say. But most people probably think R&R getting a copyright somehow “blocks” M from getting one. It doesn’t. That’s not how copyright works.

In fact, I don’t care if there are a quadrillion melodies on R&R’s hard drive. It’s impossible that S’ is identical to S. They might have the same melody structure but again, this doesn’t stop M from getting copyright in song S, and there will no doubt be differences between song S and earlier melody S’. So suppose song S becomes a hit. People will be copying and pirating that song—S. M could still sue them for copyright infringement, and even if they have a license from R&R for the melody S’, that won’t protect them from liability. This entire project it seems to me is misconceived.

As I tweeted to Riehl:

Someone making a song later that happens to be similar to one of your melodies doesn’t stop them from getting a copyright, as long as they are the author and it’s original, so not sure the point of this or who it helps. The problem is copyright law, which should be abolished.

In response, he pointed me to a FAQ. But as I noted here and here:

Not a bad FAQ but ultimately you are just pointing to the absurdity of the whole notion of IP law, and copyright in particular. You say “Potential solutions”: the only one is #3: melodies are uncopyrightable. But no, there is option 4, which you fail to consider: ABOLISH ALL COPYRIGHT (AND IP) LAW. I.e., NOTHING should be “copyrightable.” Once you grant a sliver of rationale to any copyright, you have undercut your “absurdity” critique, since… all IP law is absurd and unjust.

***

Update: See more recent tweets, e.g. here and here.

Listening again to Riehl’s Ted talk, note that he starts out by saying it’s unfair for an artist to be liable for copyright infringement if the artist didn’t intentionally copy some other song. Notice the implication that intentional copying would have been wrong. But why? This is what’s unjust about copyright. Not that you might accidentally or unknowingly copy, while trying not to; but that you can’t intentionally copy a song. What’s wrong with copying public information, or using it however you like?

Riehl also says that if you’re an artist you have to avoid previous copyrighted melodies. This is not so. If Riehl is right that there is only a finite number of melodies (especially those that sound good), then it’s possible to independently create a melody someone else came up with previously. This is not copyright infringement since it’s not copying.

And as I noted above, this still doesn’t stop people from being liable for copyright infringement if they do more than have a similar melody to an existing song. If they copy it closely, or make a literal duplicate, or have the same lyrics, and so on, then Riehl’s melody stunt is of no avail.

Riehl contrasts the finite number of melodies to paintings, of which there are infinite variations. The implication here is that copyright is not unfair when applied to such art forms, but it is for musical melodies. But a painting can be scanned and digitized to a reasonable resolution which results in a finite digital pattern. A computer could also simply generate every possible painting by cycling through every color and brightness of every pixel of a rectangular field of a sufficient graininess. Sure, the number of paintings generated by this brute force method would be vastly larger than the number of melodies generated from 8 notes, but still, it would be finite. If computers get fast enough and storage media acquire a large enough capacity, we could have a simple program written to generate quadrillions of possible paintings/images, thus exhausting the set of paintings. Store them on a storage medium and release these copyrighted paintings into the public domain. Then when Andy Warhol wants to do a pop art painting of a Prince photograph, he simply uses a search function to find in the Public Domain Image Database one that closely matches the Prince photograph, and now is free to use that public domain image to make his derivative work.

Same with books. You can use the 26-letter alphabet and punctuation marks, and a computer to generate every possible text under, say, 1000 pages. Now every book that’s ever going to be written also exists somewhere in that big thumb drive. So if I want to make a movie version of Harry Potter, I don’t need JK Rowling’s permission; I just find the similar novel in the Public Domain Texts Database and base my derivative work movie on that.

So Riehl says melodies are finite numbers and numbers are facts and facts are not subject to copyright. But every copyrighted work can be represented digitally–every novel, every painting, every movie. They are all “just numbers,” “just facts.” Yet Riehl doens’t extend his analysis to these types of copyrighted works. He talks only about 8-note melodies–“not the lyrics, not the recordings” (here). But why? The universe of possible lyrics for songs is finite, and the lyrics of any given song is just a number, as is the recording itself. The mp3 file for a 4 minute song is just a long sequence of binary digits. It’s just a number. “Just a fact,” as Riehl says of melodies.

So why does Riehl limit his experiment to 8-note melodies? Is it because he’s a copyright lawyer so doesn’t want to abolish all copyright, but he’s a musician so he wants to try to carve out an exception for his buddies and his hobby?

  1. See AI-generated art cannot receive copyrights, US court says; Monkey selfie copyright dispute; see also Ghosts in the Machine; Copyright and AI. []
Share
{ 0 comments… add one }