How the laws of intellectual property have grown–and grown apart
The following is an abridged version of the speech that Professor Fisher delivered upon accepting the Hale and Dorr Chair in Intellectual Property Law.
Over the course of American history, the law of intellectual property has gradually become fragmented into industry-specific subfields. Until now, this trend has been largely inadvertent and uncoordinated. Should it be applauded and pursued deliberately or resisted? I will approach that question circuitously–by first outlining the ways in which the fields of doctrine that together constitute intellectual property law have evolved, and only then considering whether we should strive to disaggregate them.
There are three main zones of intellectual property law: copyright, patent and trademark. Copyright law provides long-term, medium-strength protection for original forms of expression. Copyright originated in a 1790 statute, adopted under the auspices of Article I, Section 8, Clause 8 of the United States Constitution. In the first copyright statute, Congress protected three kinds of works: books, maps and charts.
Since then, Congress and the courts cooperated in expanding the set of protected creations. For example, musical compositions were added in 1831. Photographs became protected in 1865. (Today, not only is xeroxing a landscape photograph clearly illegal, but even taking a photograph substantially similar to a photograph someone has previously taken may give rise to an infringement claim.) Sound recordings–as distinct from the compositions they embody–were added in 1971. Software was formally added in 1978. Architecture became subject to copyright in 1990.
Patent law, by contrast, provides short-term, strong protection for inventions. The subject-matter coverage of patent law has grown, with a few notable exceptions, in ways comparable to copyright.
Patents–more specifically, utility patents–are conventionally divided into two broad categories: products (meaning inventions in the conventional sense) and processes (meaning new ways of doing things). Products are, in turn, subdivided into three subcategories: machines, compositions of matter and a residual category called “articles of manufacture.” Laws of nature, mathematical algorithms and naturally occurring substances were traditionally construed to be outside the zone of patent protection.
During the past 200 years, lawmakers have made a series of additions to the patent field. For example, in the 1930 Plant Protection Act, Congress extended special, limited patent protection to asexually reproduced new plant varieties–like chrysanthemums in new colors, created through radiation. The 1970 Plant Variety Protection Act extended limited patent protection to sexually reproduced new varieties.
Judges have played a part in this expansion as well. In the Parke-Davis case (1911), Judge Learned Hand LL.B. 1896 ruled that adrenaline, a purified form of the suprarenal gland, could be patented. Among the many derivatives of this seemingly innocuous proposition is patent protection for genes. Beginning in the 1990s, the Federal Circuit has consistently upheld patents on purified and isolated full-gene sequences whose physiological functions (in other words, the proteins for which they code) have been identified. Plainly, genes are “naturally occurring substances.” But the courts take the position that, in Arti Rai’s words, an “isolated gene sequence . . . differs from the DNA base pair sequence that is found on the chromosome in nature,” because it has been disentangled from the surrounding “junk” DNA–so called because it has no apparent coding function.
In the 1980 Chakrabarty decision, the Supreme Court held that a genetically engineered bacterium is patentable, announcing broadly that “anything under the sun that is made by man” can be patented. In 2001, the Court adhered to that position, ruling that new plants could be protected, either narrowly under the PPA or PVPA, or broadly using general utility patents.
Since the 1980s, a series of court decisions have removed virtually all impediments to the patenting of software in the United States, resulting in a steady increase in applications and grants. The Patent and Trademark Office, meanwhile, quietly began granting patents on business methods–like Amazon.com’s one-click online checkout system–which had been considered unpatentable on the ground that they consisted of abstract ideas. When challenged, the PTO’s policy (though not Amazon’s specific innovation) held up in court, leading to a vast wave of filings.
In the spirit of a clock striking 13, the PTO has recently issued a few patents on athletic moves–for example, on an unusual way of holding a putter. Patents of this variety are not likely to survive judicial scrutiny. But they are nonetheless suggestive of a worrisome trend.
Trademark provides potentially permanent protection for words and symbols that identify goods and services. In the mid-19th century, when it first emerged from the murk of the common law of unfair competition, trademark law shielded only marks that included the name of the person or company that sold something. Bit by bit, the kinds of things that could be protected against competitors expanded. Originally, just the names of providers of services could be trademarked; later, products were added. Services were initially limited to the names of hotels and newspapers; later, any service could be protected. Arbitrary or fanciful names for products, like EXXON or Kodak, became subject to protection. Descriptive marks–like Fish Fri for frying batter–also became subject to protection against competitors, so long as they had already become associated in the minds of consumers with particular manufacturers.
A series of nonrepresentational marks were later added to the fold. Today, symbols–the Nike swoosh, for instance–may be protected by trademark. Owens Corning is the only manufacturer allowed to make pink insulation, since the color of its product is now protected by trademark. Similarly, the sound of the NBC chimes cannot be used by competitors. “It Just Feels Right” is likely too short a phrase to be protected by copyright law, but Mazda has an exclusive right to use it as a trademark in conjunction with cars. Only Levi’s may place a label on the edge of the back pocket of a pair of jeans, no matter what the label says; other manufacturers must pick different places.
Eventually, companies began to receive trademark protection for things that seemed less like insignia and more like the products or services they were selling. Consider the shape of a Coke bottle or the decor of a Mexican restaurant–the color scheme of the awnings and the waitresses’ uniforms–both of which are shielded against imitation under the rubric of trade dress.
Trademark law now even covers product configurations. The uniforms of the Dallas Cowgirls are shielded, so long as they have acquired secondary meaning and are not, in the eyes of a court, functional. Similarly, no car manufacturer is permitted to make a car that closely resembles a Ferrari Testarossa.
The expansion of the three fields has resulted in increasing amounts of overlap. For example, while all computer software programs are protected by copyright law, a rapidly growing subset (currently over 100,000) are also protected by patent law. Certain industrial designs–objects that are both useful and aesthetically pleasing–are potentially subject to intellectual property protections under offshoots of all three fields: as copyrightable “useful objects,” through design patents and as product configurations recognizable to consumers.
Making Sense of the Growth
No single factor explains why the coverage of intellectual property law has been expanding so steadily. Rather, several intertwined forces–economic, political and ideological–appear to be at work. Here are what seem to be the most important: The transformation of the American economy, from agricultural to industrial to service-based to informational, has resulted in an increase in the number and variety of interest groups clamoring for greater protection of their intellectual products. At the same time, the United States has gone from a net importer to net exporter of intellectual property. There has been a dramatic increase, starting in the early 20th century, in the perceived importance of advertising, which has made strong trademark protection seem even more imperative to the firms that engage in it. Americans’ deep commitment to a labor-desert theory of property, together with an associated equity theory of distributive justice, has helped fuel arguments that creators deserve fair returns for their creativity. Popular suspicion, rooted in classical liberalism, of governmental involvement in the process of identifying and rewarding good works of art and socially valuable inventions has fueled hostility to governmental reward systems as alternatives to intellectual property law. The growth of the romantic notion of authorship and the corresponding celebration of the genius inventor has played a part. The groups favored by most extensions of intellectual property have been highly concentrated, while the groups disfavored have been dispersed. Finally, the term “intellectual property,” with its substantial rhetorical force, has gradually displaced an older vocabulary centered on the phrase “limited monopolies.”
Those are what seem to me the principal causes of the growth. A separate question is whether the expansion has been socially desirable. With respect to the first waves of growth, the answer is probably yes. Most of the extensions of the zones of protection made in the 19th and early 20th centuries were necessary adaptations to the emergence of new kinds of works. The creators of those works had legitimate interests in compensation and would not likely have produced them at optimal rates absent legal shields against competition.
In contrast, most of the most recent expansions have been pernicious. For example, there are plenty of incentives other than the hope of securing a patent for making innovations in athletics. Many aspects of the product configuration doctrine in trademark law are similarly problematic. Consider, for example, the Romm Art case (1992), holding that the Israeli artist Tarkay had established an artistic style–featuring languid women with lowered eyelashes in vaguely European cafés–that was infringed by the posters of Patricia Govezensky, even though no single Govezensky poster was deemed an infringement of Tarkay’s copyright in a given work. Finally, the recognition of business method patents is regrettable. No one suggests that the rate with which new business models were generated before 1998 was suboptimal or that the rate has since increased. The result is that patents on business methods produce all of the well-known disadvantages of patent law–deadweight losses and litigation costs, for instance–without any offsetting social gains. Many other recent extensions of intellectual property law turn out, under scrutiny, to be similarly ill-advised.
As the three fields of intellectual property have expanded, they have begun to fragment. Sometimes, this disaggregation has been deliberate. Congress has, from time to time, identified a particular industry or a particular type of intellectual product, concluded that it warranted special treatment and created a customized set of rules to fit it. The first example was the design patent statute, which created an entirely new patent regime for ornamental inventions. Similarly, Congress, in an effort to comply with the Berne Convention, deliberately differentiated architecture from other types of copyrightable materials.
Sometimes this disaggregation has been deliberate in a narrower sense. Confronted with a problem or plea peculiar to one industry, Congress or the courts have adopted rules limited to that industry. For example, copyright law now permits performers to make “covers” of musical compositions, provided they pay a governmentally determined fee. Another example is the judge-developed requirement of conceptual separability, which applies only to three-dimensional useful objects. Applications for business-method patents, unlike most other patent applications, must be reviewed by two examiners. The requirement, first developed by the courts, that descriptive marks, colors and, most recently, product configurations must have acquired secondary meaning before being shielded by trademark law is yet another example.
Sometimes, by contrast, disaggregation has been a largely inadvertent by-product of judicial decisions attempting to respond to the distinctive features of certain industries. For example, as Professors Dan Burk and Mark Lemley have shown, the Federal Circuit has recently semiconsciously differentiated the software and biotechnology fields while applying the requirements for patentability. In the former, the nonobviousness doctrine has been applied strictly, while disclosure requirements have been applied leniently. In the latter, the reverse bias has been applied. Why? In the court’s view, software is a mature field. Persons having “ordinary skill” in programming are thought to be especially knowledgeable. Thus, you need a big “inventive step” to reach higher than they could reach, and you don’t need to disclose much, because they can fill in the gaps in a skimpy patent application. By contrast, the court (oddly) persists in treating biotechnology as an immature and unpredictable field. Thus, a smaller inventive step will suffice to render an invention nonobvious. But to obtain a patent, the inventor must reveal more about it to enable others to replicate it.
In sum, the field of intellectual property has already become, to a significant degree, fragmented. Is this trend good or bad? Should we celebrate and generalize it? Or should we resist it?
Two substantial factors create a strong initial justification for disaggregation. But some other, institutional factors suggest that disaggregation is risky. On balance, my suggestion is that lawmakers should disaggregate more often and certainly more consciously but should be watchful for and seek to avoid certain hazards.
The first benefit of self-conscious disaggregation is that it would help to offset the dangerous rhetorical force of the term “intellectual property.” If lawmakers could be persuaded to ask not: “What is necessary adequately to protect creators’ intellectual property?” but instead: “What should be the scope of rights to colors–or genes–or advertisements?” they would be less likely constantly to expand the zone of protection. Lawmakers might then concentrate on the economic and social benefits and costs associated with different types and amounts of legal control over ideas.
The second, more complex factor favoring disaggregation is that industries differ–and differ in ways that cannot be accommodated by the general doctrines of copyright, patent and trademark law. They vary most obviously and perhaps importantly in terms of the amount of legal incentives necessary to spur innovation. In some fields, the lure of a copyright or a patent is crucial to induce people to invent; in others, it’s not. A few examples: Poets are less price-sensitive than screenwriters. Studies have shown that biochemists are primarily driven not by thirst for high incomes, but by a cluster of motivations we might call a commitment to science.
But to warrant reducing levels of intellectual property protection in a given field, it is insufficient to observe that inventors in that area would continue to invent for nonpecuniary reasons. Firms must also be willing to invest in and market their inventions, despite the absence of intellectual property rights. It turns out that investment patterns also vary by industry. In some fields, custom or first-mover advantages enable firms to reap adequate returns without the need for additional incentives, as Justice Stephen Breyer ’64 long ago suggested in his study of the trade-book industry. In other fields, intellectual property protection is more important in attracting funding.
A second source of interindustry variation involves the manner in which technological advances characteristically occur. Some fields, such as the development of drugs from plants, conform reasonably well to a model of discrete innovation, rendering discoveries relatively autonomous. In other fields, such as telecommunications, an initial, pioneering invention commonly triggers successive waves of secondary inventions. In still other fields, like software, the typical member of each generation of innovators draws on many forerunners and then spawns many more. A patent regime optimal for one type of industry is likely to be ill-suited to the others.
There are several other sources of industry variation that argue in favor of disaggregation. Within trademark law, for instance, the strength of the public interests served by protecting names and insignia varies sharply by subfield. Within copyright law, some kinds of works are primarily informational in character, while some are primarily creative or expressive in character. Rules appropriate to the latter are not necessarily appropriate to the former.
Industries also vary, increasingly, as to whether intellectual property protection is adequate to stimulate innovation and to reward participants properly. The clearest examples of outliers are music and film. Traditional copyright protections are no longer capable of preventing massive, nonpermissive reproduction and distribution of digital audio and video recordings over the Internet. Partly as a result, the revenues of the music industry are dropping fast, and the fortunes of the film industry are in peril. The search is currently on for a new, better system. My own view–presented in a forthcoming book–is that the most promising candidate would be an alternative compensation system, under which creators would be compensated (in proportion to the relative popularity of their creations) with funds raised by the government through taxes. Many benefits–cost savings, convenience, cultural diversity and semiotic democracy–would be reaped if, with respect to online distribution of recordings, we replaced the crumbling copyright regime with such an alternative source of funding. But that’s a subject for another day. The point for the purposes of the present subject is that some system sensitive to the special features of the entertainment industries would be better than continued reliance upon the generic rules of copyright.
There are four offsetting concerns that suggest, at a minimum, that disaggregation should not be employed casually. The first and most parochial is that it makes the field harder to teach and understand.
Second, the process of drawing the boundaries between the newly subdivided categories will sometimes be difficult. It may be easy to tell the difference between a smell and a sound, a movie and a song, or a plant and an animal. But other boundaries will be harder to police. For example, it’s notoriously tricky to differentiate business methods from other kinds of processes. And the boundary between hardware and software is not as clear as one might think. Whenever a frontier is less than crisp, disaggregation will incur unwanted costs.
Third, it’s not clear which, if any, institution is well-positioned to formulate and revise industry-specific rules. Congress has a decent track record in this respect. But Congress is vulnerable to rent seeking, which would almost certainly increase if Congress began more often to set rules specific to each industry. Equally important, industry-specific rules can become obsolete quickly (the Audio Home Recording Act leaps to mind), and Congress may not be sufficiently nimble to adjust. The courts are more agile. Judges sometimes do a good job of fashioning industry-specific rules, as when they used a combination of antitrust and copyright law to empower, but also to discipline, the performing-rights societies in the music industry, ASCAP and BMI. But the Supreme Court is very busy, and leaving the job in the hands of lower-court judges is bound to create circuit conflicts and forum-shopping. And the single, specialized tribunal that oversees patent disputes, the Court of Appeals for the Federal Circuit, has a mixed record on this front. That leaves administrative agencies. We might restructure the Copyright Office and the Patent and Trademark Office to equip them for lawmaking, and then give them the job. The difficulty–sadly familiar to administrative-law scholars–would be to devise a way of either designing or supervising the agencies so that they acted wisely and independently. In short, no simple solution to this crucial challenge is yet apparent.
Fourth and finally, lawmakers attempting to disaggregate the field would need either to abide by or to alter the tight constraints imposed by the several intellectual property treaties to which the United States is now a party. The TRIPS Agreement, for example, would require amendment. Even where treaty amendments were not necessary or could be obtained, industry-specific reforms not paralleled by similar reforms in other countries would create problems.
The disaggregation of the law of intellectual property is happening. Two considerations suggest that this trend is desirable. Disaggregation would help neutralize the rhetoric of property rights that currently infuses and distorts the field. And it would accommodate the multiplying and increasingly important respects in which fields of innovative activity differ. When pursuing this strategy, however, we should try to locate reasonably definite boundaries between subfields. We should pay close attention to the design of the institution to which we entrust this task. And we would frequently be obliged to harmonize our adjustments to those made in other countries. Last, we’ll have to rethink the way we teach the subject. P
A video recording of the full speech can be obtained at www.cyber.law.harvard.edu/fisherlecture.