21 August 2003. Thanks to Mike Godwin.
Remarks for the Progress and Freedom Foundation conference on “The Future of Online Music, Movies & Games”
By Mike Godwin
Senior Technology Counsel
Public Knowledge
<mnemonic{at}well.com>
Based on remarks originally delivered June 10, 2003
In his comments on the first panel at this conference, our master of ceremonies Jim DeLong expressed his distress at anti-property rhetoric. And it’s true that current debates about copyright in the Internet Age have led to some commentators to argue for the distinction between copyright interests and the traditional interests associated with real and tangible property. Those commentators have a point, insofar as the bundle of rights associated with “intellectual property” (which is itself only a few decades old as a term of art) is different in many ways from the rights associated with real or tangible property (as well as from the rights associated with “non-intellectual but intangible property,” such as accounts receivable).
But it is a point that doesn’t advance the current debate at the intersection of copyright and technology policy, because it doesn’t point to a resolution of that debate that will generate much of a consensus among policymakers or among the rest of us.
Furthermore, we should also consider that one source of anti-property rhetoric is the rhetoric of those who want to stress the connection between (or even identity of) intellectual property and traditional property interests. That’s because there is a regrettable tendency to invoke property doctrine as if, standing alone, it were dispositive of any the questions raised by computers, file-trading, the Internet, and the current struggle at the intersection of copyright and technology policy. There’s a tendency to invoke property doctrine loudly and then sit down, satisfied that all questions have been answered.
Jeff Eisenach suggested in his remarks from the audience that the pro-property/anti-property rhetorical divide maps to the traditional left-right polical spectrum, but that characterization is clearly incorrect; Richard Epstein noted, for example, that even within the American libertarian movement, itself generally sympathetic to pro-property arguments, there is division these issues, and when it comes to libertarian views on copyright and technology policy, I can assure you that there’s even more division.
Assume copyrighted works are property -- then what?
To judge from what I’ve heard here at this conference, there’s apparently a huge debate over whether copyright interests map to property. I’m not sure I agree with any argument that this mapping is a clear or obvious one -- in particular I note that term limitations on protection of copyrighted works don’t have any easy parallel in the worlds of traditional property. Nevertheless, I want us all to assume for the sake of this argument that copyrighted works are property. At the same time, I also want us all to assume for the sake of this argument that my computer is my property. Finally, I want to assume also that there is such a thing as a commons, as we understand it from traditional property theory, which is property whose ownership is shared by all members of the community. But the commons I’m going to defend at the end of these comments is probably not the one you’re thinking of.
The content industries currently are pushing a number of legislative and regulatory proposals that aim to resolve the real and imagined problems posed by computer copying and Internet file-trading. The main thing to understand about these proposals is that neither support of nor opposition to them reduces to any division between those who favor property interests, and those who oppose them. What we’re seeing here is not a conflict between those who value (or understand) property rights and those whose don’t. Instead, if you accept my ad argumentum assumptions in the preceding paragraph as true, and look at the content industries’ proposals in the light of those assumptions, it becomes apparent that the conflict is between two sets of property rights.
To understand how pro-copyright regulations might put other property rights at risk, I remind you of the current (and ever stronger) political consensus that broad regulation should be approached with fear and trembling. We embark on such regulation only when it is clear that we know what we’re doing, and when it is clear there is no alternative. The lesson of the past of much broad government regulation can be reduced to two words: “unintended consequences.” That is a lesson that many of us, on the left as well as on the right, have taken to heart.
Generally speaking, the difference between a broad regulation that works well (or mostly works well) and one that is seriously marred by unintended consequences is that the first is governed by some sort of narrowing principle. Thus, a broad prohibition of copyright infringement, such as that of Title 17, is made tolerable because infringement itself is narrowly defined, and because the prohibition has other limits and well-considered exceptions. A broad prohibition of tools which might be used for infringement is more problematic, because the class of actions involving things that might be used for infringement is far broader than the class of infringements itself. A still-broader regulation that touches every digital device that might receive, directly or indirectly, content without authorization is likely to be even more problematic, to raise greater risks of unintended consequences, and to put other kinds of property interests into jeopardy. This is true even if such a regulation does succeed in protecting copyrighted works from unauthorized distributions. (And of course it’s even worse if the regulation happens to be ineffective at protecting copyrighted works from such distribution.)
Richard Epstein, Solveig Singleton, and others have noted that the costs of rights enforcement are higher in the IP context. I think this is a sensible and necessary observation, but it raises the question of whether we believe there should be any limits on the measures of enforcement, based on the cost of those measures or on those measures’ harm to other property interests. This is, I think, the fundamental question that remains to be addressed at the intersection of copyright and technology policy.
Now, my experience has been that content owners are uncomfortable when you move the discussion to the costs of enforcement, especially when they are advancing legislation or regulation that is likely to be extremely costly. This is typically the point when a certain familiar moral-rights rhetoric kicks in. Content owners would like you to understand these issues as centering on what we (the rest of us) get to do with their property (if anything). But I’m far more strongly concerned with the distinct moral-rights question of what they, and the government, are claiming to be able to do with my property, and with our collectively owned property, in the interests of protecting copyrighted works.
Let’s begin by considering what some believe to be the sine qua non of property -- the right to exclude. With regard to the right to exclude, Richard Epstein commented sagely during his panel that in certain contexts this property right trumps my (otherwise constitutionally protected) freedom of movement -- for example, I can’t invoke it to invade his house. My freedom of speech is also constrained -- I can’t invoke it to break into the Marriott and organize a political meeting. But what are the constraints in the other direction?
Consider that Richard is a prolific writer whose works I admire. Suppose that I had many of you from this audience come over to my house, and suppose I decided to perform one of his works in a dramatic reading. Do we think it would be appropriate if Richard, learning of an unauthorized commercial performance, engaged in self-help by bursting into my house to personally prevent me from reading anything further? (I note in passing that the Copyright Act does allow something like this, but it’s not self-help -- there are significant procedural safeguards, including the requirement to obtain a court order and to have a law officer present.) What if Richard merely had a good suspicion that I might be reading -- do we think it would be okay in that instance if he burst in to prevent the suspected or anticipated performance from occurring? Or suppose the government simply prohibited me even from taking a copy of Richard’s latest article into my house, or from having you over for a visit, or from putting locks on my door -- all because it is possible that I might engage in an unauthorized dramatic reading, or, worse, an unauthorized copying. To what extent do we believe that his right to exclude me from use of his copyrighted work entails a government regulatory regime that constrains the design of my property, or my use of it, assuming that my use of my property is otherwise lawful?
Or let’s take a different intellectual-property scenario. Suppose it were determined that it is too easy for individuals to inspect the interior of new hydrogen-hybrid automobiles and figure out how to build one on their own, out of off-the-shelf materials, in their backyards. And suppose that auto manufacturers started sealing the hoods of cars or padlocking them so that only authorized mechanics could access them. This could be implemented in part through the sales contracts for the new cars. Still, the patent holders might entreat policymakers for greater protections -- of course there already are protections against patent infringement, but perhaps we’d need a new law: The Don’t Manipulate Car Parts Act (DMCPA). This Act would make it unlawful for me to peek under the hood of my own car, regardless of whether I’m planning to engage in patent infringement, because of the grave risk that I’d use the information obtained by such peeking to infringe an economically very important patent.
Clearly, there would be social costs associated with this policy choice embodied in the DMCPA. We wouldn’t be able to fix our own cars anymore. We wouldn’t be able to learn how cars work by inspecting them (a traditional pastime of young American males when I was growing up). Inventers wouldn’t be able to develop inventions that interoperate with the engine, based on inspection and reverse engineering. Indeed, it would be harder to become an inventor in this arena, because hands-on experience would be much harder to come by.
Now, of course we could modify the DMCPA to allow certain exceptions, like reverse engineering. We could make the exceptions exceedingly narrow and not entirely consistent, but, what the heck, this sort of legislation is never pretty, as we learned with the Digital Millennium Copyright Act (DMCA). Also, as an economically minded rational actor, one could choose to buy another car that didn’t have this restriction -- unless of course one couldn’t, because of inter-industry agreements and/or government regulation forbidding other kinds of design.
Shifting the Balance Away from Infringement to Potential Infringement.
My hypothetical DMCPA, like the actual DMCA, and like the technology mandates that copyright holders are currently seeking to impose on computers and other digital devices, shifts the focus of enforcement away from actual violations, and more toward possible or potential violations. The problem with this shift in focus is that, definitionally speaking, we’re all potential violators, and the burdens and sanctions of enforcement will increasingly fall on all of us.
This is a key point: Originally the law of copyright was directed toward preventing infringement itself. The DMCA moved the ball a bit -- under the DMCA, enforcement is directed against tools that could be used for infringement, even if their actual use is noninfringing. (Adam Mossoff in today’s first panel unfortunately compared circumvention tools to burglary tools, but of course the law of burglary tools is quite nuanced compared to the DMCA, since burglary-tools offenses characteristically include the element of criminal intent.)
Still, the DMCA’s prohibitions are comparatively narrow -- many of us will never have need for circumvention tools. The truly troubling aspect of the shift in focus brought about by the DMCA is that it moved the prohibition from mere infringement, and from vicarious infringement or planned infringement, to the mere possibility of infringement and vicarious infringement.
And now the burdens are being increased on all of us, yet there is no cost/benefit analysis taking place, even though there should be. The theory here is that, since we’re all potential infringers, we should bear that burden. But this burden is a costly one. To get an idea of what the cost/benefit analysis will look like, consider the case of the broadcast flag scheme.
The Price
The broadcast flag itself is nothing more than the addition of (the equivalent of) a one-bit label to broadcast content; for the broadcast-flag scheme to work, there must also be a comprehensive law that requires some range of devices to monitor for the label at the point of demodulation of the broadcast signal. This is an interesting enforcement model. It needs a regulation that requires that all devices monitor for this “flag,” that they govern copying accordingly, that competitors can’t create noncompliant devices, and that compliant devices be nontamperable. The enforcement model as described by its proponents, and as offered to the Federal Communications Commission, reaches not just to TVs but to any device that can receive TV content, or be that can be a “downstream” device from any device that can receive content. These latter categories turn out to include computers and just about any other digital device you can think of.
As I noted a few paragraphs ago, relatively few of us have a need for circumvention tools. But pretty much all of us will have a need for computers. And the latest legislative/regulatory push from content owners is for regulations that will affect all computers. The call by the movie studios for a “broadcast flag” regulation has been only the most recent and obvious manifestation of this push. The studios assert, in what most charitably can be characterized as technological ignorance, that high-definition television content is particularly prone to Internet piracy -- an assertion that is trivially easy to disprove. (But don’t take my word for it -- try downloading high-definition content yourself -- you’ll find that copying a 20-gigabyte program via a file-sharing network is excruciatingly difficult, regardless of how good your broadband Internet connection is.)
On this point, however, I think the content owners are not merely misinformed, but actively disingenuous. As it happens, comparatively low-resolution television (the analog television we grew up with and that we see today on our non-digital TVs) is far easier to pirate, because the files that result from digitizing analog content are smaller -- typically by an order of magnitude. Nevertheless, the studios have asked the FCC to adopt the broadcast-flag regulations on the theory that such regulations would protect, not current analog content (easier to pirate) but digital television content (as I’ve noted, difficult to pirate, at least over the Internet) in order to promote the transition to digital television. They cite the “Napster” phenomenon as a harbinger of woes soon to face television producing companies without noting that songs are far smaller and easier to copy than television is or likely to be. It is increasingly clear, especially considered in the light of MPAA proposals in a number of other forums, that the real goal of the broadcast flag and similar proposals is to restructure the architecture of computers generally, making them “closed” rather than “open.” And the studios want to do the same thing to the Internet as a whole.
When you actually read these proposed broadcast-flag regulations, what you see is that, despite what is being said about them, their scope is far larger than simple television reception -- the MPAA calls for regulation of every device that could receive, directly or indirectly, television content. As I’ve noted above, it is difficult to conceive of what kind of computing device this category does not include. Essentially, it calls for regulations that would require a broad range of digital devices to monitor for copyright information. In addition, it will require that of computers that “the hood be padlocked” -- that technologies in your computer must be rendered inaccessible and tamper-proof.
The Computer Revolution and Counter-Revolution
Now, apart from the question of whether this sort of initiative truly protects copyright interests, it is clear there are aspects of the computer revolution that it directly attacks. Consider, first of all, that computers make copies. In a fundamental sense, that is how computers work -- they copy information from storage to RAM to video and back again, countless times a day. Therefore, efforts to prevent unlicensed copying by changing the design of computers will run the risk of affecting their fundamental ability to function.
Second, the increasingly lowering costs (and greater efficiencies) of computers are attributable to modular, open architectures -- standard parts that can be plugged in, and standard interfaces that anybody can build to and tinker with and add to and improve.
Third, the aspects of computer functionality that can’t be modified through hardware can be modified through software -- computers are general-purpose tools that can be altered by programming.
The current designs of computers, including openness and flexibility have been responsible for the great explosion of economic growth in the information-technology sector. (And, arguably, growth in the information-technology sector has triggered economic growth in other sectors, including the content/entertainment sectors.) Attack this openness and flexibility, and you kill the computer revolution. Computers will become mere “players” of software and content and just another wing of the consumer electronics industry, and won’t be useable as open testbeds for individual inventiveness.
Even if you’re not a computer programmer or tinkerer, this will affect you. Your digital tools will be less efficient. They will be less flexible. Your own property will assume you’re a would-be copyright infringer. (As with nearly all mechanical detection mechanisms, there will be false positives.) It will be difficult for inventors to develop new tools for you to use. It will be difficult for people to learn enough to become inventors.
I think the costs associated with these changes will be huge, and that they involve not only a cost as to our individual property (the computers and other digital tools that we own as individuals, which will become more encumbered and less functional), but also to a kind of collective property -- the “commons” that is composed of open, easy-to-study-and-manipulate computer standards and architectures. It is these open standards and architectures that have enabled whole new companies and industries to arise in the computer and Internet sectors. I think that before we blithely go forward with adopting the broadcast-flag proposal, which could permanently close the “commons” of open computer and Internet standards, we should ask whether we will be imposing costs so great that they may exceed the value of what we are supposedly protecting.
I raise these questions not because I think it is somehow illegitimate for content owners to want to protect their property -- it is no more illegitimate than my desire to protect my home from invasion or my computers from a destructive legislatively mandated redesign. But I think there are ways of protecting content, or of allowing content owners to protect content, that are more efficient, less costly, and more equitable.
Encryption, Costs, and Equity
Encryption of the broadcast signal at the source is one way of doing this. Now, this would, it has been pointed out, put the existing digital TV receivers in the field -- from 250,000 to 500,000, depending on how you count -- at a disadvantage. Existing DTVs would have to be retrofitted, perhaps with a modified set-top box, to receive encrypted content. And I’ll stipulate that encryption at the source might require changes in government regulation to allow broadcasting of content that is encrypted rather than “in the clear.” But the costs associated with retrofitting existing DTV sets to handle encrypted content are likely far lower than the costs of rearchitecting the entire information-technology industry.
I think the major factor that has been missing in the broadcast flag debate has been this assessment of relative costs. I’ve given you a rough assessment here, but we ought to demand a more precise calculation of the value to be gained in dollars by imposing such a regulation, and of the costs? This calculation should include an analysis of the costs in efficiency -- what will it mean to our society when computers do less, are less flexible, are less interoperable with each other and with other devices? What will it mean when we must always replace rather than repair? (Indeed, repairing digital tools will become impossible or nearly so because repairing them is a kind of tampering.) We’ll be less enthused about new product offerings in the computer arena, because, while they’ll offer more speed and more storage, they won’t offer the same degree of open-ended potential. I can predict that if the computer industry “matures” in this way as a response to regulation, we can expect sales of computers to flatten and stay flat. And, as I’ve suggested, this will be a day when the computer industry becomes just another wing of the consumer-electronics industry.
The equitable aspect of protecting television content via encryption at the broadcast source is that it puts the bulk of the costs of protecting the content on the content holders. (It also works better.) The broadcast flag scheme, in contrast, puts none of the costs of protecting the content on the content holders -- instead, it puts those costs on everybody else, including you and me. I think there is a serious inequity issue here.
Another Commons That Matters
In closing, though, I want to come back to the issue of the “commons,” as I have defined it in these remarks. Remember, I’m not talking here about the commons of the public domain, created and renewed by copyrighted works slipping out of from under statutory protection, although that commons is also important. Instead, I’m talking about a commons of intellectual inquiry, of shared understanding of how things work, of the ability to tinker with one’s own property, of the prerogative to inspect and learn how the things you own work, and propose improvements to them, and improve your understanding of them. (I assume that it’s possible to do all of these things today without infringing copyright.) The Internet protocol is part of that commons. The standards shared by computer motherboards are part of that commons. The writings about how computers work are part of that commons. The social consensus that it’s okay to use our own property to learn how things work is also a part of that commons. It is a commons that rests on what computer scientist Ed Felten calls “the freedom to tinker.” The commons of the “freedom to tinker” with one’s own property is similar to the commons of the English language itself -- it not only involves the use of a nonrivalrous resource, but it also gives rise to the further growth and enrichment of that resource
There are two types of technological copyright protection -- those that don’t require us to close systems up (such as encryption at the source of broadcast television) and those that do (such as the broadcast flag scheme, and any other system that relies solely on “marking” copyrighted content). We collectively have a stake in open, understandable systems, based on our uses of our own property, and our posterity has a stake in that as well.
We are at a crossroads and one road ahead of us entails throttling back or even killing the computer revolution while the other allows us to continue riding that accelerating revolution into the future. It’s up to all of us to make sure that, even as we continue to develop innovative ways of protecting content in the new digital world, we also insist that our society takes the right road.