My task here is to write a response to Larry Lessig’s meditation on the free-software movement and its relationship to the general law of copyright (see “The People Own Ideas!”). But as the dry tone of my first sentence suggests, we have very different approaches to our common topic. Lessig is a master at weaving personal vignettes with structural arguments. The vignettes are intended to introduce an intimate personal dimension to the arcane world of intellectual property. His readers receive a gut-level education about the immense impact that legal rules have on ordinary people, whose voices, he tells us, can only be heard above the din if they speak in unison.
I demur. The selective imagery of eager students in Porto Alegre, Brazil, “remixing culture” using free software does nothing to address the central policy issues around intellectual property. The complex trade-offs needed to govern software and copyright aren’t illuminated by the artful juxtaposition of real users of free software with the nameless stick figures stuck with proprietary alternatives. One could as easily paint a picture of high-spirited inner-city youths mastering Microsoft Office under the benevolent gaze of the Bill and Melinda Gates Foundation. Neither helps.
Private and Common Property
My qualified defense of proprietary software rests on my general approach to property rights. It may seem odd that I see land law as a place to begin thinking about copyright law in the digital age, but in the law, continuity counts for more than novelty. While we always have to tend to the differences among different forms of property, we are likely to make fewer mistakes by proceeding carefully from established understandings.
Every legal system in history has blended two separate property regimes: the private and common. Both are important to software and copyright. Private property confers on individual owners exclusive rights to the possession, use, and disposition (sale, lease, mortgage, gift) of some given tangible resource. Virtually all civilizations start with a decentralized system in which the person who first takes an unowned thing is entitled to keep it against the rest of the world. Providing a plot of land or individual object with a single, determinate owner facilitates its effective use. The farmer who sows today knows that she can reap tomorrow, without fearing the incursions of others. The ability to sell, lease, or mortgage property allows for everything from a simple transfer of land from person A to person B to the formation of complex cooperative ventures among multiple parties. The GNU General Public License (GPL) that Lessig so admires offers a shining example of how this last, iterative process works.
Any system of private ownership requires state enforcement, first, to protect private property from forced occupation, misappropriation, and invasion, and second, to enforce voluntary deals. But any theory of property rights that includes a key role for the state should also emphatically reject the use of centralized state power to determine who shall own what resource or why. Governments, for instance, should not pick technologies.
In all legal systems, however, a system of private property rests on an infrastructure of common property. The air we breathe, the roads we travel, and the language we speak cannot easily be reduced to private possession. They remain part of the commons because their separation impedes respiration, transportation, and communication. At the edges, we recognize useful exceptions. Although everyone may use the word “monopoly” to describe a market with a single seller, only Hasbro may market a board game with hotels and a jovial top-hatted mascot under that trade name. The private creation of a trade name pulls that name out of the linguistic commons for the limited purpose of identification.
Owners and Nonowners
These simple observations can be generalized. Property rights are organized to minimize the obstacles to human prosperity and well-being by maximizing the public benefit that emerges from the self-interested actions of many individuals. Common property works when we want, say, to travel freely on a river. Private property works when the development and trade of separable assets creates enormous gains. But the justification for private rights in everything from sponge cake to software has to be social. Private property provides the right incentives for innovation, from which nonowners benefit through voluntary exchange.
In general, private property is a great bargain for society. Let’s assume that Bill Gates’s net worth is $45 billion. That’s a lot, but it’s no big deal compared to the gains his customers have received from purchasing Microsoft products. My copy of Microsoft Office may have cost me $500, but that is a tiny fraction of my gains in productivity. The most important gains from all forms of property, whether tangible or intellectual, accrue to the nonowner who buys the products of the owner. The price one actually pays for a thing is almost always less than the amount one would pay if necessary. The difference, called consumer surplus, is a pure gain for the buyer, and it exists because private ownership gave the seller the incentive to create or maintain the thing. In other words, granting a temporary patent or copyright monopoly to get the benefit of a new product now – rather than having to wait for some free product later on – is usually a good deal for both the producer and the consumer. This system of property rights isn’t antithetical to free software or “free culture.” Indeed, it is their very foundation. Let’s start with software.
Freeware versus Payware
Lessig’s defense of free software reads more like a disquisition on good and evil than a measured assessment of its benefits compared to proprietary alternatives. But the “four freedoms” contained in the GNU GPL, which governs the way much free and open-source software is distributed, weren’t inscribed on tablets brought down from Mount Sinai. They were created in the 1980s by Richard Stallman, an MIT computer scientist with a particular social agenda. Lessig describes these four freedoms and their functions cogently enough in his essay. The bottom line: free software means free and open access to a program’s source code.
It sounds innocent enough. But as with every contract or license, there’s a catch. No legal system ever creates unlimited rights, and every freedom has its correlative duties. In the case of the GPL, the kicker is that anyone who incorporates open-source software into his work must release any derivative work under the same license. The precise language is as follows: “You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.” The word “must” says it all. Content protected by a general public license is just that, licensed. Software distributed under the GPL should not be confused with ideas, writings, and inventions lodged happily in the public domain, available for all to use as they see fit, without any restrictions.
The GPL, in short, is vintage capitalism. Stallman created his own software from scratch, and the ordinary rules of property and contract let him license that software on whatever terms he chooses, with no questions asked. Those who don’t want to play by the rules of the free-software community are free to do business with Microsoft. Likewise, Microsoft is free to say, “Forget that free-access stuff: you can license our software, but you cannot see all our source code. If you don’t like our conditions, then you may switch to some open-source product.” Take-it-or-leave-it works both ways.
Why, then, prefer software bound to Stallman’s particular mix of freedoms and restraints to software that is proprietary or in the public domain? The defenders of free software often claim that their GPL inspires production and creativity, while proprietary software encourages secrecy. But this is a false opposition. The law of trade secrets can also inspire creativity; it recognizes the premise that some people will invest in a new invention only if they can retain the exclusive right to control its use. They can keep the code dark and sell the products made with it; or they can license the use of the code under a confidentiality agreement. Protecting trade secrets ensures that original creators are rewarded for their work. By contrast, open-source projects reward those who contribute to the code later.
So which plan is better? Perhaps the choice is not so stark. IBM makes millions on its database and server software yet actively encourages its customers to use the open-source Linux operating system. Sun Microsystems is relicensing its Solaris operating system under open-source terms in order to let its own software developers tap into the thriving open-source community. Even Microsoft shares its code, on a limited basis, with outside developers of Windows programs. If the copyleft movement (as it sometimes likes to call itself) requires that all derivative work be governed by the GPL, that’s fair enough; developers know the terms of the deal. Yet proprietary firms can profit from similar networks of license agreements, albeit agreements that take a different form. The state’s job is to enforce both sets of arrangements as written (with the caveat that private contracts are invalid if they create monopolies in restraint of trade).
We can now see why Lessig’s homage to free software is at odds with the principles of a free society where people can choose whatever business arrangements they prefer. We shouldn’t praise the Brazilian government for “pushing itself and the nation to substitute free software for proprietary software.” We would be equally wrong to urge the Brazilian government to promote proprietary software. In free-market societies, it is wholly illiberal for governments to take any side in controversies involving varying business models. A government’s role as a neutral arbiter is compromised whenever it engages in propaganda to persuade folks to prefer one type of contract to another. By putting its thumb on the scale, it makes true competition impossible. Arbiters cannot be cheerleaders.
The same principle, I believe, applies to the state’s own procurement decisions. Governments have fiduciary duties to their citizens similar to those that boards of directors owe to shareholders. Their job is not to satisfy their own ideological predilections; they should buy the software that offers the best combination of price and quality. The great threat to free culture is not proprietary software. It is the dogmatic insistence that one form of industrial organization is a priori better than its rivals.
That same overconfidence about software licensing pervades Lessig’s treatment of copyright. No matter one’s political beliefs, it is critical to remember the strong economic imperatives that drive modern societies to legislate some form of copyright protection. Just as we protect private rights in land for the benefit of the community, not solely for a property’s owner, so too we have a social reason to protect writings and other intellectual creations.
As John Locke would have it, a just society recognizes the natural rights of its citizens, including the right to protection of their productive labor. But copyright has an additional justification: it fosters huge positive contributions to culture, in the form of novels, movies, manuals, music, and other works. Some creators are motivated solely by the desire to create and would be happy to distribute their works under simple terms such as a Creative Commons license requiring attribution only. But for most authors, compensation matters, and we increase their production by limiting the rights of others to copy their work. Of course, authors who claim copyright protection today necessarily build on the efforts of prior writers. But Lessig’s rhapsodic praise of free culture ignores the necessary trade-offs between producers and users that any mature system of copyright must take into account.
In the end, all is a search for balance. Here are the key trade-offs:
Copyright duration. The ownership of land is normally indefinite. But the U.S. Constitution allows for Congress to issue copyrights for limited periods only. Why the difference? Because there is no sensible way to return privately owned land to the commons. Forcing the current owner out after 50 years would create a free-for-all, because only one owner can possess a plot of land. But writings (including software) are different, because many people can use them without depriving others of their use. The only question is how long to wait before returning them to the public domain. Lessig and I agree that the current rules are too generous. In the 19th century and for most of the 20th, U.S. copyrights initially expired after 28 years, a rational length of time. But the 1976 Copyright Act lengthened that period to 75 years, and the Sonny Bono Copyright Term Extension Act (CTEA) of 1998 gave producers another windfall, adding 20 more years – even for works whose 75-year copyrights were about to expire. Disney and the Gershwin estate didn’t do anything to deserve such extensions on their expiring copyrights. (Remember, someone with one year left to run on a copyright gets a lot more out of a 20-year extension than a new author whose 20 extra years start in 2080.)
Scope of protection: derivative works and DRM. Opposition to the CTEA shouldn’t translate, however, into reluctance to extend copyright protection to derivative works like the French translation of my latest novel. The extra income stream is an added incentive for the original author, while rivals must compete by producing novel works rather than derivative ones.
We should also welcome the expanded options that digital rights management (DRM) provides for marketing new works. Forcing people to pay for films and music on a per-use basis is a sensible response to the technologies that allow protected works to be copied infinitely at close to zero cost. With old-fashioned books, a work’s value to a second reader is built into the cover price. But there’s no way to price an initial sale to cover anywhere from one to a million performances of a song. Charging by use allows for price discrimination between heavy and light users, which neatly brings into the marketplace those low-intensity users who are unwilling to pay the flat fee for records or tapes. DRM is no more threatening to free culture than metered phone calls.
Nor will DRM impede the “remixing” of bits and pieces of shared experience into new creative works. I can be inspired by Hemingway or Bellow to write my own masterpiece, so long as it is not a derivative work. Indeed, often the copyright law does not give sufficient protection to original creators. The old rules work well; the only problem is that an artist or author who wants to assemble snippets from previous works into something new can find it prohibitively expensive to acquire the rights to those snippets. What’s needed is some fine-tuning around the edges.
Fair use. Section 107 of Title 17 of the U.S. Code contains a turgid account of the factors that determine whether any particular use of a copyrighted work is protected as a fair use – the limited use of the work of someone else in your own work. Fair use lets a critic quote from an author she hopes to savage: her article will be suspect if it does not show the basis for her judgment. Asking the author for permission won’t work because the author will deny access to his enemies and allow it to his friends. So weakening the property right makes sense as a way to bolster the market. Similar arguments can be made for allowing use of copyrighted works in other cases, such as news reporting, teaching, and research, as Section 107 now dutifully provides.
The harder problems are those like the great 1984 case of Sony v. Universal Studios, in which the U.S. Supreme Court held that fair use allowed people to use the Sony Betamax VCR to record television shows. The Court ruled that Sony did not illegally aid copyright infringement because its equipment had “substantial noninfringing uses”; in fact, the justices reasoned that VCRs expanded the numbers of TV viewers. But the choice involved hard trade-offs. To hold Sony liable might have retarded the use of valuable new technology, but letting it off ran the risk of undermining the TV and movie industry’s ability to protect copyrights. The Supreme Court’s decision was proved right when VCRs opened up a new income stream for copyright holders.
These same trade-offs are at issue in this year’s cause celebre, a Supreme Court case pitting MGM against the operators of Grokster, a peer-to-peer file-sharing program that some consumers use to download pirated copies of songs or movies. The friend-of-the-court brief Lessig submitted in support of Grokster shows the same habits of mind that dominate his piece for Technology Review. The late Fred Rogers of Mr. Rogers’ Neighborhood, who testified in Sony v. Universal, wanted his works to be freely available for noncommercial use: Lessig applauds this virtuous impulse and worries that suits against Grokster will frustrate the wishes of people like Rogers. So he wants Grokster to be free of liability even if individual file sharers should be punished. Fair enough; but once again he pleads his case through anecdote rather than solid legal reasoning.
My fundamental objection to Lessig’s essay is that he argues by appealing to attractive examples of free spirits and not from legal principle, as a good jurist should. My son Benjamin is a young, gifted filmmaker in New York who depends on copyright protection for his livelihood. I would never argue, however, that we should have strong copyright protection just to help Benjamin’s career. But Lessig’s arguments from anecdote do the equivalent for his own cause. In spite of his fervor, he has not explained why the standard view, which offers sensible if limited protection of intellectual-property rights, is wrong. We have yet to learn why free culture depends on free software. For me, at least, the opposite is closer to the truth: free society also rests on the strong protection of proprietary software.
Richard A. Epstein is the James Parker Hall Distinguished Service Professor of Law at the University of Chicago. He was the editor of the Journal of Legal Studies from 1981 to 1991 and of the Journal of Law and Economics from 1991 to 2001. His books include Skepticism and Freedom: A Modern Case for Classical Liberalism. Epstein taught a seminar on the intellectual origins of private property taken by none other than Lawrence Lessig.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Everything you need to know about artificial wombs
Artificial wombs are nearing human trials. But the goal is to save the littlest preemies, not replace the uterus.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.