Uncategorized

To: mail@americanpublicmedia.org
Subject: Re: MTT Files Program 4: Igor Stravinsky’s Copyright Blues

Regarding “The MTT Files” Program #4 (“Igor Stravinsky’s Copyright Blues”), which aired on San Francisco’s KALW radio tonight:

Michael Tilson Thomas’s intertwined history of music publishing and Stravinsky’s composing life was fascinating and enjoyable — but may I register a complaint? The program came very close to being a probing look at copyright and control, only to shy away and simply reiterate Stravinsky’s grievances without examining them too deeply.

For example:

Stravinsky objected strenuously to music from his ballet “The Firebird” being rearranged as a popular song, “Summer Moon”. In fact, he objected so much that he sued his American publisher for allowing it to happen, on the basis that the rearrangement was devoid of musical merit.

In his sympathic recounting of Stravinsky’s claim that the music had been damaged, Michael Tilson Thomas falls into the trap of treating music as though it were a physical object. The problem with this is easy to explain: if you damage my bicycle, now I cannot ride it, but if you “damage” my song, the world now has both my original version and your new version. Nothing has been lost, something has only been gained. Ultimately, a song cannot be destroyed or damaged. It can be copied, but the original always remains, no matter what is done with a particular copy.

An artist’s attempt to suppress derivative works, as Stravinsky did, is simply censorship by another name. After all, no one was asking Stravinsky to endorse “Summer Moon” artistically. His good name and musical reputation were not on the line: he was always free to disparage “Summer Moon” and dissociate himself from it, if he wanted to.

But by attempting to suppress the song, Stravinsky went from mere disdain to outright hypocrisy. After all, in his own compositions, he had long made liberal use of others’ work — not only of folk tunes, but of composed and attributed music, such as in the Pulcinella Suite. Would Stravinsky say that Giovanni Pergolesi (or Pergolesi’s heirs) should have the right to sue Stravinsky for rearrangements that Pergolesi would probably have found strange? Of course not. And just as Stravinsky was free to build on the work of his predecessors and peers, he should have extended the same rights to others.

My point is not that we should blame Stravinsky — he was behaving the way the copyright system encouraged him to behave, and no differently from many other composers. But an examination of music and copyright, even if it springs from a story about a particular Stravinsky rehearsal, should do more than perpetuate the all-too-common perception that the way copyright enables artists to “control” their work is an uncontroversial and natural right. The reality is far more complex than that. It’s no coincidence that modern copyright law descends directly from a sixteenth-century English censorship law, a history that Michael Tilson Thomas unfortunately didn’t have time to cover.

-Karl Fogel
Editor, QuestionCopyright.org

There’s a famous phenomenon in copyright known as the orphan works problem. It refers to the situation in which the copyright owner of a given work cannot be found. This effectively prevents others from using such a work as part of a new project. For example, if you want to make a movie based on a novel, you must first get permission from the novel’s copyright holder. But if the novel is an orphan work, then you can’t even find the copyright holder. Technically speaking, you could proceed without permission — but you would do so at your own risk. The copyright owner could emerge at any time and demand penalties. You might end up having to pay damages; worse, you might have to abandon or censor your derivative work, no matter how much effort you’d put into it.

Related to the problem of orphan works is another problem, much more serious, yet much less discussed. I call it the ghost works problem. Ghost works are all the works that never get made in the first place, or are made but not released, because copyright concerns prevent them either from being started or from being distributed. Every project that dares not base itself on an orphan work becomes a ghost work, but there are many more ghost works beyond that. Indeed, it would be fair to say that today most works are ghost works. That is, most works either don’t exist or are not accessible, because copyright obstructs them. Whenever you walk into a bookstore, survey the shelves around you and imagine them to be 90% empty, for in a sense they are.

That might sound surprising. After all, the shelves look full, don’t they? To see why they are not, let’s start with an inverse example: a classic work that (fortunately) isn’t a ghost work, but easily could have been, had its authors lived under the modern copyright regime.

In April 2007, the singer Max Ziff and I gave a concert at the Berkeley Piano Club, in Berkeley, California. We performed one of the great works of nineteenth-century German song: Die Schöne Müllerin, Franz Schubert’s musical setting of twenty poems by Wilhelm Müller. Müller and Schubert were not a team, though. In fact, when Müller wrote the poems, around 1820, it was with the intention that an entirely different composer, Ludwig Berger, would set them to music, which Berger did. Müller and Schubert never met, and Müller apparently never even knew that Schubert too had set his poems to music.

Not that Schubert was trying to hide anything from Müller. It was simply that, at the time, there was no cultural expectation that one must ask permission before making a derivative work from someone else’s original work. Müller’s poems, having been published, were now considered part of the common culture, and if composers wanted to write songs based on them, they were free to do so. To our eternal benefit, Franz Schubert had this freedom: Die Schöne Müllerin is a truly inspired piece of music, one that has influenced generations of singers and composers.

Our concert thus depended on the public domain in two ways. One way is obvious: Die Schöne Müllerin is, legally, in the public domain today, so we are free to perform it without arranging royalty payments to anyone. But there is a deeper dependency, too: this music would not exist in the first place had there not been a healthy public domain at the time the poems were written.

Imagine if Müller and Schubert had lived in the present day, instead of the early nineteenth century. Müller writes his poems, intending for Ludwig Berger to set them to music; Berger does so. Then a mostly unknown composer, Franz Schubert, appears out of the blue, wishing to set them to his own music and asking Müller’s permission. But Müller can’t give permission — he doesn’t own the copyright anymore, his publisher does, and the publisher, not wishing to encourage competition with the Berger settings, is inclined to refuse. Perhaps Schubert could pay for the privilege? But no, he doesn’t have those kinds of resources. Or perhaps he’d like to negotiate a royalty-sharing arrangement? But Schubert has no lawyer, and no head for haggling over contracts. He’s a composer, not a negotiator. Well then, he is free to set the songs for his private enjoyment (that’s “fair use”) but he certainly may not distribute them!

In all likelihood, things wouldn’t even get that far, of course. Instead, Schubert would know in advance that he cannot always follow where his inspiration leads, when where it leads him is into someone else’s copyrighted territory. Instead, he would just accept that the work of most of his contemporaries is out-of-bounds for someone like him, an unknown with no resources. And so we would not have Die Schöne Müllerin… nor Die Winterreise (another of his song cycles), nor many of his individual songs, which often set the work of living poets.

And that’s just one composer.

This exercise in imagination highlights one of the most insidious aspects of the ghost works problem: that it cannot be easily measured, no matter how great its magnitude. We can point to an existing album, a movie, or a book and say “There! There is that thing, that physical object, whose existence is beyond doubt.” But how can we point to something that is not there? How can we know what we do not have? We can only measure the loss indirectly; nevertheless, there is compelling anecdotal evidence that it is large.

Some of this evidence comes from the world of free and open source software, where it is traditional not only that the software itself be released under open copyright licenses, but that the software’s documentation be similarly licensed. This means, among other things, that derivative works such as translations can be made by anyone. In theory, this could be done without permission or cooperation from the original authors, but in practice translations are almost always authorized and approved, because cooperation is easier than territorialism.

The result of this freedom is that the documentation for virtually all major open source programs, and many minor ones, has been translated into several languages, usually enough languages to cover the vast majority of the software’s user base. Furthermore, the translations are usually kept up-to-date as the software and its documentation evolve.

This phenomenon is not limited to technical documentation. In late 2005, I published a book entitled Producing Open Source Software: How to Run a Successful Free Software Project. The publisher, O’Reilly Media, although marketing the book through traditional trade and bookstore channels, agreed to release it under a permissive (open source) copyright. Accordingly, I put the book’s full text online at producingoss.com — and pretty soon people showed up to translate it! I did nothing to seek out translators, except release the book under a liberal license; only after the first translators showed up did I put a notice on the front page soliciting more. Now we’ve got a German translation under very active development (with two separate translators who only met through cooperating on this project), a Hebrew one happening somewhat more slowly, and some recent arrivals looking at doing a Chinese version.

This is happening with a book that has, let’s face it, a fairly limited audience. Not only that, it’s the second time this has happened to a book I’ve published (see cvsbook.red-bean.com). Once could be coincidence; twice is starting to look like a pattern. And I’m only using my own books as examples because they’re the first thing that came to mind. There are hundreds of open source projects that could tell a similar tale about their documentation. The lesson to draw here is that, were it not for copyright restrictions, most books in the world would be translated into several languages. After all, the better the book, the more some multilingual reader will be motivated to translate it. The translation doesn’t have to be perfect, because there will also be people who show up to edit it. These projects tend to self-organize in exactly the same way that open source software projects do.

But under the current copyright regime, if you want to release a translation of a book that was published under traditional restrictions, you don’t just sit down and start translating. Instead, you start by negotiating the right to translate — a process which is completely unrelated to actually translating, and is also daunting, time-consuming, and likely to fail. It’s hard to imagine a more potent gumption sink than “rights negotiation”. The mere prospect is enough to shut down most translation projects — which is why I look at those bookstore shelves and see them as mostly empty.

Now sensitized to the presence of ghost works all around us, I usually don’t need to go out of my way to find examples. Instead, I can depend on them finding me with some regularity. Such was the case here. While writing this article, I opened my New Yorker and saw the above piece. I could have chosen from several other examples that crossed my path in the last week — and those are just the stories that someone bothers to tell. This fact alone is a clue to the size of the problem.

If one person can gather a few examples of ghost works without even trying, many people working together can really start to catalogue the problem. Maybe after we have enough we’ll start to notice some patterns. So please keep your eyes open, and if a ghost work crosses your path, let us know. Our contact page is http://www.questioncopright.org/contact.

The conference “Copyright, DRM Technologies, and Consumer Protection”, on March 9 and 10 at UC Berkeley, was a great example of competing assumptions jockeying for mental floor space.

(DRM is the set of software and hardware handicaps that prevents computers and media players from sharing files freely with each other. It’s why when you download a song to your iPod, you can’t copy it to another iPod or upload it to somewhere else, for example; or why when you burn a CD with a standalone CD burner, you often have trouble making copies from the new CD.)

The panelists at the conference were varied: lawyers and professors of law, some economists and other academics, executives from content-owning companies and content-carrying companies, officials from governmental and quasi-governmental bodies (e.g., WIPO), someone from the British Library, from the Electronic Frontier Foundation, from Public Knowledge, etc. Each panel held five or six people, facing the audience. A moderator introduced each one, and then that panelist spoke for fifteen to twenty minutes on the panel’s topic. After all had spoken, the panel as a whole took questions from the audience.

Given the disagreements in the room, the conference understandably didn’t come to any conclusions about DRM. But it did two other useful things: it gave an opportunity for every possible analysis of DRM to be heard and debated (there were some non-obvious ones), and, perhaps more importantly, it revealed the rhetorical terms in which different parties want the public to think about DRM.

One good point a few panelists made is that successful DRM is likely to weaken the user’s privacy. All DRM prevents computers and media devices from sharing files freely with each other. But in order to merely curb freedom, rather than end it entirely, DRM must identify which files can be shared and which can’t, and which methods of sharing are permissible. The more sophisticated this process of determination becomes, the more it is necessary for devices to analyze information about the files in complex ways. The burden of this analysis will often be too great to implement in typical consumer electronics — so instead the data will be sent to an online server, which will figure out your rights and tell the client device what to do. But step back and consider where this is going: devices all over your house, sending information about your viewing and listening habits to a central server. Is this data certain to be subpoena-able someday? You bet. It probably already is.

Another point (made by Peter Swire among others) was the computer security implications of running DRM. The code in a DRM system must be a black box: it cannot be open source, because if the user could understand and change it, she could disable it and copy her files without restriction. But if the code is opaque, it cannot be examined for security flaws — and in fact, the Digital Millennium Copyright Act makes it illegal to even attempt such an examination in most circumstances. Basically, you have to run this code, for even if you are technically capable of modifying it, doing so would be illegal. (In response to this situation, Jim Blandy proposed a new slogan: “It’s my computer, damn it!”)

There was also some discussion of DRM in terms of consumer law and contract law, rather than copyright law. Consumer law takes into account the “reasonable expectations” of the consumer (for example, that having obtained a copy of a movie, you might reasonably expect your television or computer to actually play it when asked). But this solid-sounding phrase slowly disintegrated as people pointed out that the expectations of consumers are not static: as people experience digital restrictions more and more often, they begin to accept them — their “reasonable expectations” begin to incorporate DRM behaviors.

It was fascinating to see how determinedly the representatives from content-owning companies used the words “balance” and “choice”. Over and over again, we heard that the best DRM systems are those that strike an appropriate “balance” between the rights of content owners and the rights of consumers (also sometimes called “users”). The invocation of “choice” as a guiding principle often came as part of a self-addressed call-and-response formula, as in “What do consumers really want? They want choices.” I don’t see any way to understand that other than as a fake question designed to make DRM deployment appear to be a response to some market need — which it isn’t, because users have been indicating pretty clearly what they want: bits that flow freely.

These sorts of assertions often came in the context of the broader claim that DRM has the potential to enable a wide variety of new business models (see “An Economic Explanation For Why DRM Cannot Open Up New Business Model Opportunities” for a rebuttal of this point of view). The business-model argument was repeated by several panelists, and it’s worth some attention for the assumption underlying it: that enabling any particular business model is a positive good, a prima facie justification for whatever DRM mechanism might be required to enable it. My friend Ben Gross has an intensely practical answer to this kind of thinking. He objects to draconian copyright laws on the grounds that it’s simply not the government’s job to prop up failing business models, and he applies the same reasoning to DRM. Panelist Andrew Bridges (of Winston & Strawn, LLP) said essentially the same thing, in a memorable comment on DRM’s essential role in the marketplace: “There are two ways to make money by connecting supply and demand: by making it easy, or making it hard.”

Some panelists made reference to DRM protecting “integrity” (e.g., Victoria Bassetti of EMI: “DRM preserves our products’ integrity”), but we never got a concrete explanation of how it does so, or even what precisely it would mean. “Integrity” is a loaded word here, because whenever it is used in a conversation about filesharing and copyright, one can easily imagine that it refers at least partly to plagiarism. I don’t know whether that’s how these people meant it, but the inference is hard to avoid, and it’s completely backwards: DRM works against the detection of plagiarism, because it impedes digital technologies’ ability to arbitrarily examine and compare files, and prevents people from uploading files to locations where they can be viewed and downloaded publicly. Plagiarism cannot flourish where there is transparency, but DRM prevents transparent behavior at a technical level, and thus drives people toward non-transparent methods of sharing.

There were also various attempts to talk of DRM-restricted products as being essentially the same as physical products or limited-resource services. Thus, Thomas Rubin of Microsoft said that DRM has been accepted for years as a means of controlling access to satellite TV (and now satellite radio), to websites that require login accounts, to cell phone networks, and even to traditional libraries! At some point during the reading of that list, he mentioned that he was being deliberately provocative and tossing up some examples as fodder for thought. I’m glad he included that caveat, because his list didn’t contribute anything constructive to the debate, except to outrage more than one person in the audience (as I learned chatting in the hallway afterwards). Web sites and cell phone networks have limited bandwidth and computational resources, so their products really can be used up, if too many people log on. And libraries deal with physical objects, so access control is as understandable for them as for jewelry stores.

Is it so much to ask, at this late date, that everyone debating issues of copyright and DRM agree to stop talking about digital data as though it were a limited resource? You can’t “steal” songs and movies, you can only copy them. They’re not like library books, or cell phone bandwidth, or an artist’s reputation (all of which are, in one way or another, diminishable resources). Talking about data in that way is a disservice to logic. I think, deep down, Thomas Rubin knew this, which is why he inserted his disclaimer.

If I were to take away two lessons from the conference, they’d be that language matters, and assumptions matter. The rhetorical advantage gained by being in favor of “balance” is nearly unbeatable. I think the only way to deal with it is to redefine “balance”, to start using that word to talk about balancing new things, for example, the right of the public to copy and make derivative works versus the right of the artist to have (very) temporary control over the initial distribution of her work.

As for assumptions: I heard Victoria Bassetti of EMI respond to a questioner by asking (paraphrasing, as I don’t have the transcript) if he cared whether artists earned any money or not. Since artists mostly don’t make money from copyright royalties anyway, her response was a non-sequitur, but it effectively placed her on the artists’ side and the questioner on the side of those lazy, freeloading filesharers. Because she knows that most people share a certain assumption — one which she may even sincerely believe herself — about artists earning their livings from copyright royalties, she’s able to use this kind of response to deflect attention from the problems DRM creates. It would be a bad outcome indeed for these to be the terms under which the public considers DRM.

I’ve concentrated mostly on the remarks of unreservedly pro-DRM panelists here, but I don’t want to give the impression that they set the tone of the conference. There were impressive critical presentations and questions from the aforementioned Andrew Bridges, from Gigi Sohn (of Public Knowledge), Cindy Cohn (of the EFF), Ian Kerr (University of Ottowa), Deirdre Mulligan (Berkeley Center for Law and Technology, Samuelson Clinic, and Boalt Hall School of Law), Peter Swire (Ohio State University), and others. I went to the conference partly to see how people were talking about DRM, from all points of view, in preparation for being on a similar panel in Montréal next month. I was not disappointed. If there remains any major point about DRM not raised at this conference, I’d be very surprised; kudos to the organizers for that.

[References: my notes from the conference are here.]

On April 18th, 2007, I’ll be a panelist at a session with the provocative (and somewhat enigmatic) title of “Interoperability: computer industry giants versus music?”, at Les Rencontres québécoises de l’industrie de la musique, at the Bonsecours Market in Montréal, Quebec. The other panelists will be from the music distribution industry, plus at least one from the Electronic Frontier Foundation. I’m looking forward to a lively discussion! Full report here afterwards…

Update:

Well, this isn’t really a “full report”, but the panel discussion was terrific, and not at all the slugfest one might expect — genuine discussion took place. But the most interesting thing about the conference as a whole was that although it was mainly composed of people from the recording and radio industries, many were very receptive to the message that copyright is not always good, and many also showed signs of giving up on DRM as a strategy for controlling copying. It may be that the industry is starting to see the light, at least in Québec, Canada.

By now, the whole classical music world has heard of the Joyce Hatto scandal (Wikipedia’s article is excellent).

Joyce Hatto was a pianist who died in June, 2006. She didn’t play many concerts, but she recorded prolifically — or so everyone thought, until it was discovered, in early 2007, that most of her recordings were plagiarized from the records of other pianists. She never knew about it, apparently: the plagiarism was the work of her recording engineer and husband, William Barrington-Coupe.

The best part is how the deception was uncovered: when someone put her recordings onto a computer, automated comparison routines kept stubbornly identifying them as other pianists’ tracks!

It’s a great example of what we’ve been saying about artists putting their work online: sharing files widely prevents plagiarism, by making it much easier to detect. Forget Hatto herself for a moment — think instead of all those other pianists, whose recordings were passed off as her work: the reason the hoax was detected at all was because their track information was available online. And if the recordings themselves had been available online, the problem would only have been detected more quickly, probably years ago.

The unmasking had nothing to do with DRM, by the way. DRM is the set of software and hardware handicaps that prevents computers and music players from sharing files freely with each other. It’s true that some of the programs that detected the similarities between Hatto’s recordings and other pianists’ also have built-in DRM, but the DRM is utterly irrelevant to the comparison techniques that spotted the correlations. In fact, if DRM were as effective as the record companies wish it were, it would only have hindered the comparisons, since then the other pianists’ track information might not have been readily available for examination.

But we’ve still got a long way to go. The Wikipedia article on Hatto had the following sentence, as of early February 28th:

Meanwhile the British Phonographic Industry (BPI) has begun an investigation. If the allegations are true, it would be one of the most extraordinary cases of copyright infringement the record industry had ever seen, according to a BPI spokesman.

Notice how the hoax is identified as “copyright infringement“, not “plagiarism“. I checked the reference: the BPI spokesman apparently referred to “piracy”, so in the interests of accurate quoting, I’ve changed the Wikipedia article to say “piracy”. But that’s not really satisfactory: the word “piracy” is often used to refer to both unauthorized printing and plagiarism, as though the two are the same offense. The word thus provided a semantic pivot, around which some Wikipedian was able turn from one of the word’s meanings to the other, making it into a case of “copyright infringement”.

Who was this mysterious misquoter?

We’ll never know, because they did it anonymously, though clearly on purpose. For when the sentence was originally added, it quoted the BPI representative correctly. Later, someone came along and changed just one thing: “piracy” to “copyright infringement”. You can see the edit here. Probably they felt that “piracy” was too loaded a term, and that “copyright infringement” would be more accurate. Unfortunately, this is exactly the conflation — equating unauthorized copying with stealing credit — that the record industry promotes; the pity is that their effort has been so successful.

I’ve fixed the text to say “piracy” again. But the BPI spokesman should have talked about “plagiarism” in the first place, because that’s what we’re dealing with here, and the more we let digital files circulate freely, the less plagiarism there will be.

Back in late 2006, Matthew Gertner (of AllPeers) and I did a mutual interview about copyright reform. It was a fascinating and wide-ranging conversation, and he’s posted it on PeerPressure, the AllPeers blog.

From Matthew’s introduction:

“…Rather than assuming that some copyright is necessary and trying to decide exactly how much is optimal, [Fogel] suggests that we imagine a world without copyright and take it from there.

He contends at the beginning of the podcast that, not only does he not know personally what the right level of copyright is, but that it isn’t possible to know this based on current evidence, a view that I find eminently reasonable. I also agree wholeheartedly with the way he concludes our discussion:

I think that there is some built-in exclusivity there but I also think… whatever change is going to happen is going to happen essentially through a market process. It’s not going to be that Congress suddenly wakes up and drastically reconsiders copyright law. Instead, some number of artists, just as some number of software developers did a couple of decades ago, will by choice release their stuff under these liberal copyrights, And they will create this little fertile safe space for sharing that will grow, and basically we’ll have two parallel streams: there’s the old stream and the free stream. And people will just start choosing stuff based on what they like, not based on ideological concerns about how it was produced. And we’ll just see what happens.

At the end of the day, we need to create an environment where individuals can test their own approaches to copyright and let the market decide what works best. I don’t necessarily see as strong a connection as Karl between liberal copyright terms and free content, however, and I hope that this makes our discussion more dynamic and thought-provoking.”

There are both download and streaming links available — listen to it here.

Click the image above to watch QuestionCopyright.Org Executive Director Karl Fogel delivering a talk at the Stanford University Library’s Technology Chalk Talk Series on October 19, 2006. The video is available to view and download on the Internet Archive.

The talk is about 90 minutes long, including the question-and-answer session after the presentation. The audience members’ backgrounds were in library science, computer science, publishing, and law, so the Q&A was particularly good in this talk.