Yet, for several reasons, the patent “underdevelopment” problem arguably applies to a large share of potentially valuable inventions. First, patent law encourages inventors to file for patents early in the innovation process. At this stage, especially for modern technologies, an invention is usually not in the form of a finished product ready for sale, and its commercial success is highly uncertain. Instead, the inventor must undertake costly and risky development and testing to transform the invention into a commercially viable product. This uncertainty encourages inventors to delay commercialization in the hopes of reducing risk—for example, by taking advantage of emerging complementary technologies that may lower production costs more than any forgone profits. Indeed, many of the twentieth century’s greatest inventions, including the television, radio, radar, and penicillin, were not commercialized until decades after they were invented. In some instances, the uncertainty is so great that the commercialization of a worthwhile invention never occurs.

Second, the Patent Office and courts tend to countenance expansive “claims”—namely, the inventor’s legal rights—which encompass far more than what the inventor actually discloses in a patent. Although broad claims can reduce commercialization costs by allowing the original patentee to coordinate development efforts among multiple firms, often this coordination fails to occur because of high bargaining costs or strategic behavior, which can stymie the independent commercialization efforts of more efficient firms. Because early patent grants reward the best inventor, but not necessarily the best commercializer, broad claims can impose unwarranted burdens on third-party commercializers. Rampant defects in patent examination, licensing, and litigation often make these undue costs quite large and diminish commercialization.

Third, patent law is primarily designed to induce invention; any protection it provides to commercialization is mostly an afterthought. The dominant “reward” theory of patenting, which undergirds much of today’s law, perceives minimal need to protect risky and costly post-invention development and commercialization efforts. Thus, reward theorists view the patent system as an unfortunate “second-best” compared with one in which all inventions are immediately placed in the public domain. The upshot is that patent law confers direct encouragement to inventors who create and disclose intangible specifications, but not necessarily tangible products.

Fourth, although there has been limited empirical study of the issue, in a 1998 survey of 133 companies worldwide conducted by the British Technology Group, approximately 40% of the patents held by the respondents were uncommercialized. Nonetheless, these companies reported that 32% of these patents were either commercially “very important” or “quite important.” For engineering companies, the figure increased to 40%, and for biosciences and pharmaceutical companies, to 34%. These results are consistent with a European Commission-funded survey that focused on “important” patents, which found that 38% of the patents were never commercialized.

Several scholars have suggested various reforms to improve patent law’s commercialization incentives. One approach, which follows Ed Kitch’s influential “prospect” theory of patents, proposes strengthening patent rights—by, for example, broadening patent scope or lengthening patent terms—so that patentees can “internalize” more of the positive “external” benefits generated by their inventions. Although this view is tempting, reward theorists have rightly criticized it and related proposals as dampening competition and impeding follow-on technological development. Moreover, contrary to one of prospect theory’s core claims, in many cases such a property-rights approach can retard commercialization—particularly when third parties are better commercializers than the inventor. Not only can broad and strong patents enable inventors to exploit defects in the litigation process to inefficiently “hold up” commercializers, but they can also prevent commercializers from “designing around” the original invention and engender “wasted” costs by inducing too many firms to enter the proverbial “patent race.”

Another proposed route to improving commercialization incentives is to modify the reward theory to encourage patenting later in the innovation process, such as by requiring patentees to build a prototype before filing. Although a modified reward theory would improve upon many wanting aspects of today’s patent system by forcing inventors to engage in at least some commercialization in exchange for a patent, it could significantly diminish ex ante incentives to invent and could lead to unnecessary duplicated development costs.

Neither prospect theory nor a modified reward theory can practically achieve an ideal balance because both attempt to “commercialize” traditional patents designed to spur the creation of new and non-obvious knowledge, rather than to prompt the manufacture and sale of new products. Therefore, instead of trying fine-tune patents not originally designed to promote commercial activity, society should adopt a novel policy lever—a “commercialization” patent—granted in exchange for a commitment to commercialize a product not available in the marketplace.

Clearly, the burden of proof for adopting a new type of intellectual property right is high. Patent scholars have generally been opposed to new rights, viewing them as unnecessarily diminishing competition, being too costly and difficult to implement, creating needless complexity, and encouraging legislative rent-seeking. With these hurdles in mind, I sketch a solution and briefly explain why it overcomes these concerns. (Of course, a detailed proposal is ensconced in the Stanford Law Review.)

Commercialization patents could be filed for the same types of product inventions as those within the scope of traditionally patentable subject matter. Only a product that is “substantially novel”—that is, substantially different from a product currently available in the marketplace would qualify for a patent. The commercialization patent would need to be practiced no later than three years after filing. Unlike a traditional patent, which can broadly claim many embodiments, a commercialization patent’s claims would be limited to the product specifically disclosed in the specification and its substantial equivalents.

In contrast to previous proposals for new forms of intellectual property rights, a commercialization patent would not only provide a negative right to exclude others from making and selling the same or equivalent products, but would also include an affirmative equitable and legal right to its holder to make and sell the product. First, the affirmative equitable right would give the commercializer absolute immunity from any injunctive remedies otherwise available in infringement suits by traditional patent holders. Second, any traditional patent holder would be limited to a low, but fairly reasonable, fixed royalty rate it could win at suit, e.g., 1-2%, and would be subject to damages apportionment for multi-component products.

In order to mitigate the potentially harsh consequences of affirmative rights, a commercialization patent could only be filed after a traditional patent covering the product goes uncommercialized for three years after issuance, extended for any regulatory or other unavoidable delays. This window would provide sufficient lead time and a strong incentive to a traditional patent holder to commercialize its invention. Finally, because commercialization cycles tend to be quick, commercialization patents would be of short duration—e.g., five to eight years from filing—though longer terms may be appropriate for a handful of industries.

Such a patent would substantially increase the commercialization of inventions without unduly decreasing competition or imposing dynamic inefficiencies in the system. Because commercialization patents would provide partial immunity from suits by traditional patentees, they would significantly weaken the rights of non-commercializing patentees, reducing transaction costs and pernicious “deadweight losses” from patent licensing and litigation. The administration of a commercialization patent system would not be costly and complex, could reduce the number of traditional patent filings, and could provide significant additional revenue to the Patent Office to improve traditional patent examination. In particular, a commercialization patent would be drafted in the same way as a traditional product patent and include the same kinds of claims. Review for subject matter, utility, enablement, written description, substantial novelty, best mode, and the like would be the same or similar to that for a traditional patent. For similar reasons, judicial oversight of commercialization patents would not be terribly costly. Although the assessment of whether a commercialization patentee sufficiently “worked” the patent may initially be difficult, a “sham sale” doctrine would quickly develop and root out this problem. Finally, because commercialization patents would apply in the same manner to all patentable subject matter, the incentives for industry-specific lobbying of Congress to specially tailor the commercialization patent statute would be minimal.

In sum, commercialization patents of the sort proposed here would fundamentally alter the patent system’s single-minded approach to a multi-faceted problem. Surely, elucidation of the details of commercialization patents requires testing, empirical study, and refinement. Yet, by decoupling the traditional patent into an invention patent, granted in exchange for the disclosure of new and non-obvious knowledge, and a commercialization patent, granted in exchange for the manufacture and sale of a substantially new product, the patent system could offer more optimal incentives for invention and commercialization alike.

Acknowledgments:

Copyright © 2010 New York University Law School.

A lengthy list of colleagues and students who provided invaluable comments and suggestions, as well as copious footnotes for the assertions made herein, can be found in the aforementioned article. This work was funded by a grant from the Ewing Marion Kauffman Foundation.

Ted Sichelman is an Assistant Professor of law at the University of San Diego.

This Legal Workshop Editorial is based on the following Law Review Article: Ted Sichelman, Commercializing Patents, 62 STAN. L. REV. 341 (2010).

Claiming intellectual property is an act of communication, but as with all communication not everything can be spelled out with exactitude all the time—far from it. By drawing out an additional distinction between modes of claiming, Jeanne Fromer advances our understanding of claiming in intellectual property.¹ To the familiar distinction between central and peripheral claiming she adds the contrasting modes of claiming by exemplar and claiming by characteristic. She very usefully applies these two axes not just to patent law, where the issue of central versus peripheral claiming originates, but also to copyright. The possibility of mixing and combining the four different modes of claiming—peripheral by characteristic, peripheral by exemplar, central by characteristic, and central by exemplar—in different areas opens up not only a new lens for looking at the hybrid nature of existing patent and copyright law but also suggests a framework for thinking about improvements to claiming in these two, and other, areas of intellectual property. In this Response I argue that although these two axes can be regarded as distinct, they do share something in common—a differential reliance on formalism versus contextual interpretation—and that the degree of formalism is relevant to the conclusions we can draw from Fromer’s framework.

In peripheral claiming, which is familiar as the main method in current patent law, the would-be patentee claims the invention by describing its outer boundaries, in a process that is sometimes analogized to a metes and bounds description in the law of real property. If something is infringing it must fall within the claim; otherwise there is no infringement. By contrast, in central claiming, which was more characteristic of early nineteenth-century patent law and which lives on in the doctrine of equivalents and means-plus-function claims, the claim sets forth some central or paradigmatic members of the claimed set and then one relies on some notion of similarity to figure out membership of the full set. In Fromer’s fork example, there are four ways to claim.² In peripheral claiming one can either describe the set of forks by characteristic or by pointing out (more on this later). In central claiming, one can either set out a description of a central case of a fork or point to a central exemplar of one, and rely on a notion of relevant similarity to extend the set outward to the full set of claimed forks. This latter determination of similarity in central claiming is typically left for ex post decisionmaking, which allows for more flexibility but at the cost (so it is conventionally thought) of less ex ante certainty.

To this distinction between central and peripheral claiming familiar from patent law, Fromer adds claiming by characteristic versus claiming by exemplar. Patent law typically employs claiming by characteristic, paired with peripheral claiming. This leads to the familiar highly regimented style of claiming in which a single sentence is used to delineate the outer bound of the invention space claimed by the inventor. By contrast, claiming by exemplar involves setting out examples. This fits naturally with central claiming, as in copyright, in which central members of the set are set forth and similarity is judged ex post. But peripheral claiming by exemplar is also possible. Fromer assumes that it would involve listing the members of the claimed set, as is done in so-called Markush claims.⁴ Generally, as we will see, it will not be feasible to literally “point to” every fork that is being claimed (that is, purely by exemplar), and there must be a procedure for inferring from what is pointed out to the entire claimed set. The feasibility of such a procedure and the form it could take would depend in turn on who the relevant interpretive community is; the general public, inventors, and patent attorneys will all bring different background knowledge to bear. ⁵

Even more generally, what Fromer’s framework (implicitly) raises is the question of what types of information should be made explicit and which are best left implicit under different conditions for claiming. Different audiences, different informational resources, and different dynamics of innovation and age of industries can all influence how easy or difficult it is to leave information implicit and conversely how much reliance on formalism is needed.

For many purposes, a simple definition of formalism is helpful. An instance of communication is formal to the extent that its interpretation is invariant to context.⁶ The language of first-order logic, for example, is formal because its semantic rules are highly worked out and thus do not rely much on context, compared with more informal everyday mathematical notation. Natural languages (of which English and Chinese are examples) on this definition are less formal than the language of mathematics, but language in different situations can be more or less formal. Legal language tends to be formal because drafters are wary of reliance on context. For example, legal language in contracts and statutes is notorious for the repetition of noun phrases (as in “the party of the first part agrees with the party of the second part that the party of the first part,” and so on) as opposed to the use of pronouns which do require context for their interpretation. The more that the speaker and hearer are socially close, can rely on common knowledge, and are trying to cooperate, the less formal language needs to be. Thus, saying, “It’s cold in here” is a shorter way of saying, “Because you are closer to the window, please close the window.” But the looser formulation that relies on pragmatic inference relies on a lot of background knowledge and coincidence of communicative interest in order to succeed. In the absence of these factors, the more formal approach would be warranted—possibly including a description of what a window is and how it works. Sociolinguistic studies have shown that individual speakers employ more informal variants with socially closer audiences.⁷

No useful system of communication is perfectly formal in the sense of being absolutely invariant to context, but some communications and systems of communication are more invariant to context than others. I have argued elsewhere that formalism is a matter of degree, and that differential formalism depends largely on the nature of the audience for the communication. Communication is subject to an informational tradeoff. At the same cost one can communicate intensively with a small, socially close, knowledgeable audience or one can communicate in a more stripped-down, formal, and explicit way with a larger, more heterogeneous, and more anonymous audience.⁸ Formalism, which is a matter of degree, should correlate with the breadth and heterogeneity of the audience. The basic distinction between the use of language in contracts and property reflects this tradeoff: communicating an in rem right to “the world” requires a more formal and more regimented style of language and forms, including the standardization supplied by the numerus clausus.¹⁰—but in the examples given here, there is a lot of implicit context-based reasoning going on under the table, which makes claiming by exemplar fall towards the informal end of the spectrum.

To the extent that the costs and benefits of formality and contextuality along different dimensions go together, we might expect to find the formal and informal poles of these spectrums to line up: peripheral claiming will tend to be done by characteristic, and central claiming will tend to rely on exemplars. And the broad contours of patent and copyright do line up this way.¹¹ In this sense the two axes share a lot more in common than Fromer lets on. But Fromer is right to point out that each area uses a hybrid of the types of claiming, and we should ask what factors at this more detailed level push toward or away from the different modes of claiming.

What I suggest here is that the formality versus contextuality of claims—their differential formalism—will emerge as an important lens when trying to apply Fromer’s categories to the problems of claiming in patent, copyright, and other areas of intellectual property. Formalism implicates the costs and benefits of delineation depending on the audience’s background knowledge, the degree of cooperativeness of the communication, and the nature of the informational resource.

The audience for the delineation of rights matters when it comes to the costs and benefits of delineation, as Fromer recognizes. In general, patent law is directed more to expert audiences and much, though not all, of copyright can implicate the activities of nonexperts.¹² Thus, more expert knowledge can generally be presumed in patent law. Delineation in patent law can use technical terms and rely on knowledge in the field that copyright law cannot. But which area of law can rely more on implicit (less formal) means of communication? The role of audiences suggests that different types of implicitness can be employed in patent and copyright. Patent attorneys or agents (including those working for industrial competitors) versed in the relevant art can be expected to fill in certain kinds of information, and potential copyright violators different kinds of information. Possibly the latter kind of background information is closer to everyday knowledge. There are areas of copyright that do involve expertise of an artistic sort, but with a few exceptions expert knowledge (other than possibly legal knowledge) does not seem to feature as much in filling in informal delineation in copyright law as in patent law. Thus the nature of the audience points to the possibility of leaving different information implicit in the two areas. Correspondingly, patent and copyright rely differently on formalism when this reliance on background information is not reasonable.

Claiming intellectual property is an example of noncooperative communication which generally cuts in favor of formal communication. In a cooperative setting, one can assume that the audience will fill in information and draw inferences in a charitable fashion. At the center of the study of the “logic of conversation” is Grice’s Cooperative Principle: “Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.”
¹³ That is, the interpreter will interpret in accordance with the assumption that the speaker is trying to be helpful. Thus, in the Gricean example where Person A says “I’m out of gas” and Person B says there is a gas station around the corner, we and Person A can safely conclude that it is open, because otherwise Person B would be violating the Cooperative Principle and the maxims of conversation.

But legal interpretation is not like this. Disingenuousness of speaker and audience is almost to be presumed. The “logic of legal conversation” diverges from the ordinary conversation studied by Grice. Reliance on central claiming and claiming by exemplar requires context to be filled in and inferences to be drawn. In an adversarial context, it is difficult to converge on the relevant context and these pragmatic inferences, because each party—patentee and alleged infringer, say—has different incentives as to how to fill these items. The noncooperative nature of much of legal communication probably accounts for some of its formality. And legal language in general often relies on a more rigid analogue of pragmatic-style inference. The linguistic study of pragmatics focuses on language meaning in context, as distinct from the study of more formal meaning in semantics. In legal language a lot of what would be loose pragmatics is regimented into the semantics.

Whether to rely on formal or context-dependent modes of communication should depend on the nature of the resource. Do certain kinds of information require greater formality in delineation? One difference between patent and copyright is in how “close” a claimed contribution is to what another might wish to use or develop. In copyright, delineation can be left somewhat unclear when two copyrighted works are not likely to be close. Thus, copyright can allow a defense of independent creation without as much conflict as patent law would have if it provided the same defense (which it does not).¹⁴ Apart from low-authorship works like maps and charts, it is unlikely that two works will be identical without one creator having had access to the other work.¹⁵ But, as Fromer points out, derivative works present a special situation in copyright, where the notion of fortuitous resemblance is a greater possibility.¹⁶ I would say that the right to prepare derivative works is different from the rest of copyright and, for better or worse, comes closer to patent law or even the law of ideas.¹⁷here is a point in this series of abstractions where they are no longer protected, since otherwise the playwright could prevent the use of his ‘ideas,’ to which, apart from their expression, his property is never extended. Nobody has ever been able to fix that boundary, and nobody ever can.”) (citations omitted).] In patent law, particularly in well-established areas of technology, there is a high likelihood of two claims being close or even contiguous in invention space. (Indeed claims can overlap.) Take for example the field of tennis rackets. Many rackets have different “splayed” patterns of strings, with the ends of the strings attached alternately above and below the central plane of the racket frame at different distances from each other, such that in a patent claim, infringement can turn on whether “varies between” requires three or more distances between the strings.¹⁸
In a less crowded field, the potential ambiguity in “varies between” would not matter so much.

Finally, the question of the best degree of formalism in claiming potentially interacts with the psychological evidence that Fromer adduces.¹⁹ This evidence suggests that people use characteristics when learning broad categories and early on in the process of learning categories of all kinds and that they use exemplars for narrower categories and later in the general learning process. This might suggest that claiming by characteristic would be more suited to exclusion strategies and exemplars would be used more in governance strategies, when picking out special uses. As Fromer acknowledges, this evidence is to be taken cautiously, because the conditions in the experimental studies may differ relevantly from the situation of those faced with intellectual property claims. It would be a worthwhile subject for further study whether leaving information implicit (as in central claiming and claiming by exemplar) inhibits learning more or less depending on the level of expertise of the audience and on the nature of the category to be learned. Whether the two kinds of implicitness show the same or different learning effects would be relevant to the question of how formal claiming should be with different audiences in view.

In sum, Fromer’s addition of the distinction between claiming by characteristic and claiming by exemplar to the more familiar central versus peripheral claiming dichotomy is a real advance, which provides a lens for more detailed evaluation of hybrid claiming methods and their variation within areas of intellectual property law. But like central versus peripheral claiming, the exemplar-characteristic claiming distinction also crucially relies on a spectrum of how much information is spelled out formally and how much is left implicitly to be filled in by reference to context. This differential formalism will be important to further development of Fromer’s very fruitful framework.

Acknowledgments:

Copyright © 2010 University of Chicago Law Review.

Henry E. Smith is a Professor of Law at Harvard Law School.
For helpful comments I would like to thank John Golden and Oskar Liivak. All errors are mine.

This hypothetical claiming system looks like that of patent law, under which an invention’s bounds must be demarcated as a prerequisite to patent protection.  But envision for a moment that patent claiming would look more like that of copyright law.  Alexander Graham Bell would receive a patent for his invention of the telephone³ after having fixed (or perhaps commercialized) it in some form.  Assuming the invention complies with the threshold requirements of patent law, the set of protected embodiments would include all substantially similar implementations—a cordless telephone? a fax machine? Internet telephony?—a set to be enumerated on a case-by-case basis in any future infringement litigation, rather than at the time of patenting.  This determination would require courts to ascertain the essential properties of a patented invention.

This thought experiment seems to indicate that claiming the set of protected embodiments under patent law looks very different than copyright law.  And in a sense, it does.  Patent law has adopted a system of peripheral claiming, requiring patentees to articulate by the time of the patent grant their invention’s bounds,⁴ usually by listing its necessary and sufficient characteristics.  And copyright law has implicitly adopted a system of central claiming⁵ by exemplar, requiring the articulation only of a prototypical member of the set of protected works—namely, the copyrightable work itself fixed in a tangible form.⁶ Copyright protection then extends beyond the exemplar to substantially similar works,⁷ a set of works to be enumerated only down the road in case-by-case infringement litigation.

Investigating the claiming practices of patent and copyright law side by side thus illuminates two salient axes for claiming the set of works protected by an intellectual-property right: peripheral versus central and characteristic versus exemplar.  Peripheral claimants must delineate to the public the set’s bounds so that a third party could determine whether any particular embodiment is a set member, thus protected by the right.  By contrast, central claimants must describe only some central members of the set, which are clearly protected under the right and are used to determine whether other items are similar enough to the enumerated members to also fall within the same right.  Central and peripheral claiming might be seen as two points on a spectrum of how many members of the set must be described by the claim, with peripheral claims describing more members than central ones.  Either peripheral or central claiming can be done by exemplar or by characteristic.  Claiming by exemplar entails enumerating particular members of the set.⁸ In the case of peripheral claiming by exemplar, one would enumerate each set member, while for central claiming by exemplar, one would catalog only some set members.  Claiming by characteristic, on the other hand, requires a description of the essential properties of the set’s members.  For peripheral claiming by characteristic, one would describe the necessary and sufficient features common to all members of the set of protected embodiments.  And for central claiming by characteristic, one would express the features common to at least some central members of the set.  Claiming by exemplar and by characteristic can be seen as opposing points on a spectrum of how much distillation of the actual works’ characteristics is necessary.  The table illustrates how the fork might be claimed under each of the four claiming types.  Though scholarship mentions patent law’s peripheral claims⁹ and Clarisa Long and Henry Smith discuss patent law’s claiming requirements and copyright law’s lack thereof,¹⁰ until now these dual claiming dimensions have not been expressly appreciated.

¹¹¹²

The method of ascertaining the extent of the set of protected embodiments, for evaluating validity or infringement, depends on the claiming type.  For peripheral claiming by exemplar, one can determine whether a particular embodiment is protected by observing whether the embodiment matches any claimed exemplar.  Does the fork in question match a fork in the pictured set?  For peripheral claiming by characteristic, one must decide whether the embodiment possesses the claimed features.  Central claiming by exemplar requires divination of the essential features of each claimed exemplar, followed by a determination whether the embodiment is sufficiently similar in those features to a claimed exemplar.  And with central claiming by characteristic, one must decide whether the embodiment is sufficiently similar in its features to those claimed.

Despite patent law’s typical peripheral claims by characteristic, in practice, patent claiming is heterogeneous.  Patent law retains some vestiges of central claiming under which it used to operate,¹³ through the doctrine of equivalents, means-plus-function limitations, and dependent claims.  With some exceptions, the doctrine of equivalents allows a patentee to “claim those insubstantial alterations that were not captured in drafting the original patent claim but which could be created through trivial changes.”¹⁴ Means-plus-function limitations in claims permit a patentee to describe his invention in terms of the functions it typically performs rather than the parts of the invention that carry out those functions.¹⁵ With these claims, protection is statutorily extended beyond the inventions described in the patent’s specification that carry out the claimed function to any equivalents of those inventions.¹⁶ And a dependent claim—which incorporates an independent claim in the patent further limiting the independent claim¹⁷—has no “central claim”-like legal effect of extending protection beyond the described characteristics, but nonetheless has a “central claim”-like communicative effect of highlighting prototypes of the associated independent claim.  Moreover, patent law encourages some claiming by exemplar through its best-mode requirement and Markush claims.  Markush claims allow the grouping of exemplars in otherwise characteristic claims so long as they possess a “community of . . . characteristics which justify their inclusion in a common group.”¹⁸ And to fulfill the best-mode requirement, the inventor must disclose the best embodiment subjectively conceived by the time the patent application was filed,¹⁹ thereby effectively claiming one superior exemplar of the invention.

Just as patent claiming practice can be varied, so can copyright claiming.  By design, copyright law institutes central claims by exemplar; but through the approved use of licenses to permit others to make substantially similar works, copyright in operation involves expression of the bounds of works permissibly created under such licenses and delineation of characteristic features of the set of protected works.  In one striking example, NBC Universal recently licensed a French television company the right to create a French version of the criminal procedural television series, Law & Order: Criminal Intent, and this license incorporated the creator’s thousand-page manual detailing essential characteristics of a Law & Order production that must be followed, such as how many of the show’s characteristic “ching-ching” sounds could be used per act.²⁰ These expressions in legally binding contracts bring forms of peripheral claiming and claiming by characteristic into copyright law.

This description of patent and copyright claiming raises the question of which sorts of claiming work best for each.  The Constitution sets forth the goal of intellectual property laws: “To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.”²¹ Each of the four types of claiming can affect the costs of drafting claims, efficacy of content notice to the public of the set of protected embodiments,²² ascertainment of protectability, breadth of the set of protected works, and the protection of works grounded in after-developed technologies (technologies postdating claiming).

As to costs of claim drafting, keeping them low is important when the law seeks to encourage easy protection of intellectual property and high otherwise.  Drafting is costlier for peripheral than for central claims and for peripheral claims by exemplar than by characteristic because of the foresight needed. It is also costlier for central claims by characteristic than by exemplar if an embodiment, which could otherwise serve as a central claim by exemplar, already needs to be produced.

Clear content notice to the public of the set of protected embodiments is valuable so that the public can avoid improper use of the set’s members without permission and can, concomitantly, understand what is free for the taking, thereby furthering innovation.²³ The established view is that peripheral claiming provides better content notice than central claiming.²⁴ This view has rightfully come under attack because, among other things, ambiguities inherent in language can render claim construction unpredictable²⁵ and peripheral claims tend to contain highly abstract formulations, rather than colloquial terms, to ensure that a broader set of embodiments is protected.  As Doug Lichtman observes, a reader “might find it easier to read simple, concrete claim language (‘shoelace’) rather than more abstract expressions (‘mechanism by which to bind tightly around the foot’) that are in fact technically superior.”²⁶

Research in cognitive science also suggests that the established view might fail to account for how people process, comprehend, and construct categories and neglect the other dimension of claiming (exemplar or characteristic).  According to this research, people’s categories are formed and comprehended not with a list of necessary and sufficient criteria to test for membership, as the classical Aristotelian view would suggest,²⁷ but rather with prototypes against which to compare potential category members for sufficient similarity.²⁸ Though these studies were done in other contexts, they are valuable in suggesting how to communicate categorical information successfully.  To the extent this research generalizes to categories of inventions or artistic creations in intellectual property law, the necessary and sufficient features set forth by peripheral claiming by characteristic do not accord with the way people learn categories.  Content notice might be better—or just as good—with (well-constructed) central claims conveying prototypes or exemplars, or even peripheral claims by exemplar.  Even if central claims do not provide perfect content notice of categorical boundaries, the evidence indicates that they might do a good job at communicating the crux of the protected embodiments.  There might also be differences in the effectiveness of content notice depending on whether claiming by exemplar or characteristic is used.  Recent experiments suggest that when it comes to learning larger, more differentiated categories, learning by characteristic is prevalent early in the learning process but slowly gives way to exemplar learning.²⁹ By contrast, in learning smaller, less differentiated categories, learning by exemplar is dominant.³⁰ To the extent this research applies to intellectual-property categories—as intuition would suggest—when intellectual property law is protecting a small, poorly differentiated category, claiming by exemplar would be appropriate.  But claiming by characteristic would be more suitable for larger, better differentiated categories, and claiming by exemplar might also play a role in teaching categorical boundaries.

In addition to the varying degrees of difficulty and cost associated with claim drafting and content notice, the claiming systems differ in the ease of assessing whether the claimed set of embodiments is protectable.  This factor is closely related to content notice because the assessor—typically a government actor—needs to understand the set’s scope to ascertain protectability by the relevant intellectual property laws.  Notwithstanding their content-notice problems, peripheral claims are thought to enable the examiner to assess the novel or original features of the claimed set because the full extent of the set is expressly set out by the creator.  For central claims by characteristic, protectability of those set members not literally described by the central claim might not be assessed during examination.  And for central claims by exemplar, assessment of protectability of the full set is more difficult, as the examiner must deduce the essential features of the exemplar.  Of course, embodiments not literally encompassed by the central claim but falling within the set of protected embodiments will be adjudged separately down the line for protectability, if at all.

Another aspect one might think affected by the choice of claiming system is the breadth of the set of protected works.  Though one asks very different questions under each type of claiming to enumerate the corresponding set of protected embodiments, the interpreter does not necessarily arrive at significantly different breadths.  In construing peripheral claims, one can assign claim terms either narrow or broad meaning.³¹ In construing central claims, one can limit or broaden the set of embodiments that are sufficiently similar to the centrally claimed prototype.³² And just as with central and peripheral claims, there is elasticity in construing both characteristic and exemplar claims.  Of course, concern lies not only with the true breadth of the set, as determined by government officials, but with its operational breadth.  Given that there is uncertainty of the extent of the set of protected embodiments under peripheral or central claiming and characteristic or exemplar claiming, aversion to the risk of liability attracts licensees.³³ Licenses will be taken sometimes even when unnecessary—when a patent or copyright is invalid or the valid claims do not cover the licensed activity.  Operationally, such licensing leads to a set of protected embodiments broader than the law would otherwise allow.  Therefore, the better the content notice provided by the claims, as discussed above, the less broad the set of protected embodiments will tend to be operationally relative to actual breadth.

There is one likely exception to the notion that claiming choice does not per se affect the actual breadth of the set of protected embodiments, namely the extent to which works grounded in after-developed technologies are protected.  A peripheral claimant must think well beyond the manifestation of his invention to the future set of manifestations likely to be valuable enough to prevent others from using, even if not presently practicable.³⁴ A central claimant, on the other hand, can defer delineation of the full set of protected embodiments at the outset until the future when, by virtue of time’s passage, the unfolding of related innovation has become clearer.³⁵ Central claims, then, more easily allow for an expansion of protection beyond the claimed core in light of future innovation.

That the choice of claiming systems implicates the foregoing factors—factors essential to calibrating intellectual-property law to stimulate innovation—in different ways underscores the importance of choosing a claiming system with care.

Though previous scholarship principally defends the typical claiming forms for each patent and copyright law,³⁶ I undertake a thought experiment to analyze whether they are ideal for either type of intellectual property.  This task, in fact, is suggested by patent law’s incorporation of not insignificant elements of central and exemplar claiming and copyright practice’s use of peripheral and characteristic claiming.  I explore whether claiming in copyright and patent law can learn from one another.

Though patent law admirably incorporates all four types of claiming flexibly, it can be tweaked to stimulate innovation by adding claiming elements more reminiscent of copyright law.  I suggest that patent law’s typical peripheral claims by characteristic, adopted principally to provide content notice of the set of embodiments protected by a patent, do not provide sufficient content notice, which negatively affects the public’s ability to know what is in and out of the public domain, assessments of protectability (particularly when an issued patent is accorded a presumption of validity³⁷), and the operational breadth of the set of protected works.  To ameliorate these and other concerns, I propose—contrary to conventional wisdom³⁸—that ex ante patent-claim drafting be modified to include central claiming by characteristic.  And claiming by exemplar ought to serve a role in patent law.  Claims by characteristic can be supplemented by requiring the registration of certain exemplars—all commercial implementations, if any, by the patentee or licensee—claimed to be within the set of protected embodiments.  Exemplar registration, which would be available to the public and linked to the associated patent, would help sharpen the understanding of the bounds of the set of protected embodiments.  And it would occur in the situations in which exemplars are most useful, when the patented invention is commercialized and is therefore likely to be valuable—when content notice is important.  These modifications to patent claiming would better serve patent law’s purpose to stimulate innovation by making it easier for the public to identify that which must be licensed to be used and that which can be used freely for follow-up innovation.

Claiming in copyright law is more complicated.  Current claiming practice in copyright law is rigid, and it suffers from various defects, all tied to the poor content notice effected by the central claims by exemplar.  Those defects might suggest that copyright claiming ought to co-opt claiming by characteristic from patent law. But the structure and theory of copyright accentuate significant—perhaps insurmountable—barriers to making changes to the current claiming system.

As it stands, copyright’s central claims by exemplar provide little content notice to the public, leading risk-averse third parties either to take licenses even as to works not protected by copyright or avoid them completely, a situation that grants too heavy a copyright reward at the expense of generating further creativity.  From this vantage point, it would seem far more productive to require or at least provide significant incentive to copyright claimants ex ante to claim their works centrally by characteristic.  This claiming would merely entail, at insubstantial cost, a succinctly expressed pattern of the work at issue.  Such claims might provide significantly better ex ante notice in two ways.  First, the enablement of feature-by-feature comparisons would help indicate those works that would be considered substantially similar to the created work and thus protected under the copyright.  Second, such claims would help explicate which substantially similar works would nonetheless be permissible uses under the doctrine of fair use³⁹ by encouraging determinations of works that borrow from the copyrighted work in ways that do not implicate too many of the claimed features or transform it significantly.  These effects, on the view of one concerned with copyright’s poor notice, would increase overall production of creative works by providing the incentive to create copyrighted works and by encouraging creation by third parties beyond the copyright.

But that is far from the complete analysis.  Viewing the copyright system as a whole through a wider lens identifies significant, and perhaps insurmountable, theoretical and practical concerns with central claiming by characteristic.  First, there is a concern that requiring copyright holders to claim their works by characteristic, even if centrally, might cause copyright law to contravene the First Amendment by granting protection to ideas, rather than the expression copyright seeks to protect.⁴⁰ A second concern lies in whether artistic creators will feel at ease demarcating the essential pattern of their creation.  A related third concern is whether the described characteristics could capture the artistic essence of the work.  Consider the likely claim for a Jackson Pollock painting: a painting comprising varied colors of dripped, flung, or spattered paint.⁴¹ That description does not capture—and arguably cheapens—the artistic essence and effect of Pollock’s paintings.  Fourth, there is apprehension about the cost and viability of the administrative and legal support necessary to institute central claims by characteristic, due to the ex ante examination claims by characteristic might require, the number of claims to be written, and the legal support necessary to assist copyright holders.

All in all, I suggest that there has been a severe underinvestigation of claiming practice in intellectual property and that the handful of explorations of the topic have overlooked some of the important aspects of the taxonomical, descriptive, and normative features of intellectual property claims. To maximize innovation, it is imperative that claiming practice, a key factor of intellectual property systems, be explored and optimized.

Acknowledgments:

Copyright © 2010 University of Chicago Law Review.

Jeanne C. Fromer is an Associate Professor at Fordham Law School.

In this Article, I suggest that trade secrets can be justified as a form, not of traditional property, but of intellectual property (IP). The incentive justification for encouraging new inventions is straightforward. Granting legal protection for those new inventions not only encourages their creation, but enables an inventor to sell her idea. And while we have other laws that encourage inventions, notably patent law, trade secrecy offers some significant advantages for inventors over patent protection. It is cheaper and quicker to obtain, since it doesn’t require government approval, and it extends to protection of types of business and process information that likely would not be patentable.

It seems odd, though, for the law to encourage secrets, or to encourage only those inventions that are kept secret. I argue that, paradoxically, trade secret law actually encourages disclosure, not secrecy. Without legal protection, companies in certain industries would invest too much in keeping secrets. Trade secret law develops as a substitute for the physical and contractual restrictions those companies would otherwise impose in an effort to prevent competitors from acquiring their information.

The puzzle then becomes why the law would require secrecy as an element of the cause of action if its goal is to reduce secrecy. I argue that the secrecy requirement serves a channeling function. Only the developers of some kinds of inventions have the option to overinvest in physical secrecy in the absence of legal protection. For products that are inherently self-disclosing (the wheel, say, or the paper clip), trying to keep the idea secret is a lost cause. We don’t need trade secret law to encourage disclosure of inherently self-disclosing products—inventors of such products will get patent protection or nothing. But if trade secret law prevented the use of ideas whether or not they were secret, the result would be less, not more, diffusion of valuable information. The secrecy requirement therefore serves a gatekeeper function, ensuring that the law encourages disclosure of information that would otherwise be kept secret, while channeling inventors of self-disclosing products to the patent system.

My argument has a number of implications for trade secret policy. First, the theory works only if we treat trade secrets as IP rights, requiring proof of secrecy as an element of protection. If we give the protection to things that are public, we defeat the purpose and give windfalls to people who may not be inventors (what we might call “trade secret trolls”). Courts that think of trade secret law as a common law tort rather than an IP right are apt to overlook the secrecy requirement in their zeal to reach “bad actors.” But it is the courts that emphasize secrecy, not appropriation, as the key element of the cause of action that have it right. Second, an IP theory of trade secrets also encourages preemption of “unjust enrichment” theories and other common law ways courts are tempted to give private parties legal control over information in the public domain. Thus, an IP theory of trade secrets is in part a “negative” one: the value of trade secret law lies in part in defining the boundaries of the cause of action and preempting others that might reach too far. Analyzing trade secret claims as IP claims rather than common law contract or tort claims requires courts to focus on what the law is protecting, how, and why—something the common law did not do. As a result, the unified trade secret approach does not expand, but rather cabins, the overbroad reach of the common law.

I.
Constructing an IP Theory of Trade Secrets

Trade secrets are best understood not as applications or extensions of existing common law principles (warranted or unwarranted), but as IP rights. In this Part, I explain the two critical features trade secrets share with other IP rights—they promote inventive activity and they promote disclosure of those inventions. I then seek to explain the most significant anomaly—the requirement of secrecy.

A.     Incentives to Invent

Trade secret law confers an exclusive right on the possessor of valuable information not generally known to or readily ascertainable by competitors. Exclusivity is the hallmark of an IP right. Both patents and copyrights confer similar rights to prevent use by others on the developers of new and valuable information. In so doing, patents and copyrights are generally acknowledged to serve a utilitarian purpose—the grant of that legal control encourages the development of new and valuable information by offering the prospect of supracompetitive returns, returns possible only if the developer does not face competition by others who use the same idea. In this way, patents and copyrights avoid the risk of underinvestment inherent with public goods, which are more costly to invent than to imitate once invented.

Trade secrecy has the same effect. It gives the developer of new and valuable information the right to restrict others from using it, and therefore the prospect of deriving supracompetitive profits from the information. This may be true of business as well as technical secrets, since some protection for business ideas helps ensure a first-mover advantage for those who take risks on untested business models. True, the right of exclusion in trade secret law is not absolute. The trade secret owner cannot sue someone who develops the idea independently, or who reverse engineers a product on the open market to learn the secret. But the same is true of copyright law. A right to exclude does not have to be absolute to be effective in rewarding and therefore encouraging innovation. It need merely provide sufficient advantage in terms of lead time or relative costs to minimize or eliminate the public goods problem.

The additional incentive provided by trade secret law is important for innovation. Trade secret law reaches into a number of corners patent law cannot. The definition of trade secret (valuable information) is broader than the definition of patentable subject matter, for example, protecting business plans, customer lists, and so-called “negative know-how” against use by others. Patent law cannot protect valuable information of that sort. Further, inventors must apply for patents, publish their applications after eighteen months, and then wait perhaps four years for the Patent and Trademark Office to decide whether to grant protection. That significant delay renders patents unavailable as a practical matter in fast-moving industries. Trade secrets, by contrast, are automatically protected upon creation provided the requirements of the statute are met. Finally, patent litigation is as much as three times as expensive as trade secret litigation, with a price tag—a median of $5 million per side in legal fees for large cases—that puts it out of reach of many small firms. Small wonder, then, that economic literature suggests that some firms, particularly start-ups, rely heavily on the incentive to invent provided by trade secret law. In many cases patents are simply not an adequate substitute.

Trade secret law also reaches where contract alone cannot. Trade secret law precludes acquisition of information by strangers using improper means—computer hacking and other forms of corporate espionage. Further, it extends the reach of the law beyond privity of contract to anyone who comes into contact with a secret knowing that they have acquired it by accident, mistake, or by another’s malfeasance.

B.     Incentives to Disclose

Patent and copyright law do not exist solely to encourage invention, however. A second purpose—some argue the main one—is to ensure that the public receives the benefit of those inventions. There is decent evidence to support the idea that at least one function of an IP right is not just to encourage new invention, but to encourage the dissemination of those new ideas.

At first blush, trade secret law seems to push in the opposite direction. After all, protection under trade secret laws is conditioned on secrecy, and so it seems to encourage secrecy, or at least the development of inventions that can be kept secret. Paradoxically, however, trade secret law actually encourages broader disclosure and use of information, not secrecy. It does so in two ways. First, the legal protection trade secret law provides serves as a substitute for investments in physical secrecy that companies might otherwise make. The facts of E. I. du Pont de Nemours & Co. v. Christopher² provide an example. There, the plaintiff was constructing a chemical plant, and during construction it was apparently possible to see the layout of the plant from the air and so to discern the secret process du Pont was using. The court noted that du Pont could have built a temporary roof over the plant during construction, but only at “enormous expense.” It didn’t need to build that roof because the law protected its interest in avoiding (aerial) prying eyes. Had the law not done so, however, it is reasonable to suppose that du Pont might have built the roof rather than risk loss of its trade secrets. That investment in secrecy would have been inefficient; it is cheaper (both for du Pont and for society) for the law to provide that protection. And even if the investment in secrecy were efficient for du Pont in the absence of the law, it would still impose a social cost by restricting the flow of information—a cost du Pont would have no reason to take into account.

There is empirical evidence that overinvestment in secrecy is a real problem in the absence of trade secret protection. Examples can be found as far back as the guild system that pervaded Western economies in the Middle Ages. Guilds were places that could and did develop technical knowledge, but in the absence of legal means to protect that knowledge they went to great lengths to prevent others from learning of it, imposing draconian limits on the mobility of employees and the development of competing firms. The same problem remains today in countries that do not provide legal protection for secrets. Robert Sherwood studied business practices in Mexico and Brazil, two countries that do not have strong legal protection for trade secrets and in which resort to the courts may not be viable for a variety of reasons.³ He found that companies in those countries make business decisions that inefficiently limit the disclosure of information because they fear that they cannot rely on the courts to prevent the use of information they do disclose. For example, they may be less willing to contract production out to third parties if it means giving out information about secret processes, even where the third party could use the process more efficiently. They may take elaborate security measures, building walls and fences and hiring armed guards. And they may hire employees whom they expect to be loyal—such as family members—rather than strangers who would do a better job.

The problem also remains for products or industries that do not qualify for IP protection. Michael Pollan explains that the developers of new breakfast cereals, for example, engage in enormous efforts to protect the secrecy of their new ideas in order to gain a few months’ first-mover advantage. For the same reason, they operate their own machine shops to design the cereals, rather than outsourcing that work to those presumably more specialized in it.⁴

None of this evidence is perfect. One of the problems with social science is that it is hard to run clean tests in the real world. Nonetheless, there are both logical and evidentiary reasons to believe that, without legal protection, companies in certain industries would invest too much in keeping secrets. These investments are inefficient, in several senses. In many cases, the problem they address could be avoided by the courts at lower cost than the building of walls and fences. Second, physical investments must be made for each secret, while legal investments need be made only if there is misappropriation. That means that even if a physical investment in secrecy is individually cheap, in the aggregate the cost of having to make that investment for every secret may outweigh the cost of resort to law, which will be necessary only in those few cases in which the secret is actually misappropriated. Finally, and most importantly, restrictions on the flow of information between business partners or to new employees slow the process of commercialization and improvement of the secret inventions, and therefore interfere with both the invention and disclosure functions of IP law. Trade secret law developed as a substitute for the physical and contractual restrictions those companies would otherwise impose in an effort to prevent a competitor from acquiring their information. In so doing, it encourages disclosure of information that companies might otherwise be reluctant to share for fear of losing the competitive advantage it provides.

To be sure, trade secret law still encourages some secrecy. So if the alternative were a world in which companies freely disclosed their inventions, that world might be preferable. But the empirical evidence suggests that is unlikely to be the alternative. Rather, a world without trade secret protection is likely to have more, not less, secrecy.

Trade secret laws can encourage disclosure in a second way as well: they serve as a partial solution to Arrow’s Information Paradox. The paradox is this: In the absence of any legal protection, the developer of a potentially valuable but secret idea will have a difficult time selling that idea to someone who could make more efficient use of it. In order to sell the idea he will have to disclose it to allow the buyer to evaluate it, but disclosing it destroys the value inherent in its secrecy. To see this, imagine that I tell you I have a great idea, and I’ll share it with you for $1 million. Should you take the deal? You can’t know the answer to that question unless I tell you what the idea is. But in the absence of legal protection, if I tell you what my idea is, you no longer need to pay me $1 million.

Now add trade secret law (or any IP right) to the picture. The existence of a legal right to prevent others from using or disclosing my idea in breach of a confidential relationship allows me to disclose the idea in precontractual negotiations, secure in the knowledge that the other side is not free to take the idea without compensating me. The law, by giving certain rights to the holder of the secret, allows him to disclose information he would otherwise have been unwilling to share, and therefore permits business negotiations that can lead to commercialization of the invention or sale of the idea, serving both the disclosure and incentive functions of IP law. True, the parties could have entered into a contract limiting what could be done with the information, but the putative buyer may be reluctant to sign such a contract without knowing what they might be limiting themselves from using. Both venture capitalists and Hollywood executives, for example, are notoriously unwilling to sign nondisclosure agreements before reading business plans or movie scripts. Trade secret law reaches beyond contract law by allowing courts to infer the existence of a confidential relationship from circumstances in which transactions might be difficult or impossible without that assumption.

C.     Channeling Protection Between Patents and Trade Secrets

So far, so good. But at this point the reader might object that, if the goal of trade secret law is to give legal rights over an invention while encouraging its disclosure, we don’t really need the secrecy requirement at all. In this vein, a number of scholars have suggested that any investment in protecting trade secrecy is wasted, since the law is requiring companies to spend money in ways that reduce, not increase, the dissemination of ideas. If the goal of trade secret law is to encourage dissemination by giving the security of a legal right, this argument runs, why not just grant that right to any information, regardless of whether it is secret?

The problem with this argument is that without some basis for defining the legal right, it will sweep too broadly. If I can get ownership rights in any information, no matter how public, the result will be to deter, not promote, the dissemination of that information. Broad legal rights may restrict employee mobility, with negative consequences for the economy. If any idea, no matter how public, is subject to a claim of legal rights, individuals and companies will reasonably worry about using any information they do not themselves develop. If I could sue you for repeating my explanation of trade secret law, the result is not likely to be wide discussion of that explanation, even if I have no intention of actually suing you for discussing my idea.⁵ And while we could theoretically substitute a defendant’s conduct for proof of secrecy as the basis for entitlement to a legal right, such conduct-based definitions are circular—competition is unfair if it is likely to be defined by courts as unfair—and ultimately empty.

Granted that we need some definition of the entitlement, why secrecy? The answer, I believe, is that the secrecy requirement serves to channel inventors into the appropriate form of IP protection. Consider three different types of inventions: one that is impossible to conceal once it is in widespread use (think of the wheel or the paper clip), one that is impossible to discern by evaluating the product (think of the formula for Coca-Cola), and one that can be discerned by evaluating the product, but only with difficulty (think of software source code, which is not evident from the object code sold to customers but which might be reverse engineered). In a world with patent law but no trade secret law, companies with inventions in the first category—those who have developed inherently self-disclosing inventions—will turn to patent law if they can. If not, they will be out of luck. If the paper clip were not patentable, companies wouldn’t be able to keep it secret and still make much profit from it. Their best option would likely be to sell the paper clip and hope to make some profit from brand recognition or first-mover advantages.

Companies with inventions in the second category, by contrast—those who develop inventions that are not transparent to the world, such as chemical processes and some formulas—might well decide to keep an invention secret in the absence of legal protection. They may reason that secrecy may give them a greater advantage than patent law, since patents may be held invalid, may be easy to design around, and in any event will expire within twenty years. Indeed, there is some empirical evidence that they do so—that where secrecy is possible, inventors choose it over patent protection. Without trade secret law, the efforts those companies take to protect their secrets may be excessive. That overinvestment may be specific—protection of a particular idea—or general—imposing too many restrictions on employees and business partners. Either way, the result is both inefficiency from overinvestment in secrecy and the loss of the benefits of public disclosure of information.

A secrecy requirement provides protection to companies in the second category, not in the first. Thus, it ensures that trade secret law provides legal protection in circumstances in which inventors might otherwise choose excessive secrecy, but denies protection to inventions that companies would not keep secret in the absence of patent protection. By drawing this line, even the secrecy requirement of trade secret law has the surprising effect of reducing, not increasing, the secrecy of inventions.

What, then, of inventions in the intermediate category? Companies with inventions in this third category might or might not rely on secrecy rather than patent law. Both approaches have risks. As noted above, patents might be invalid, or easy to evade, and in any event will expire in a set period of time. On the other hand, reliance on secrecy provides only tenuous protection, since the secret could be discerned by reverse engineering or independent development or disclosed by an employee or business partner in the absence of trade secret law. In this case, the effects of introducing trade secret law are ambiguous. If companies in this third category would have opted for secrecy, then the introduction of trade secret law reduces the negative effects of that secrecy for the same reasons it did in the second category. But if they would have opted for patent protection rather than secrecy without law, adding trade secret law might encourage them to keep secret information they would otherwise have patented (and therefore disclosed).

To avoid inadvertently encouraging secrecy rather than disclosure, trade secret law incorporates limits on the scope of the right, notably the defenses of independent development and reverse engineering. As the Supreme Court suggested in Kewanee Oil Co. v. Bicron Corp.,⁶ and as commentators have suggested, these defenses weaken the trade secret right sufficiently that it does not entice inventors to choose secrecy over patent protection. That isn’t always true, of course; the inventors of Coca-Cola could have chosen to patent it but didn’t. But importantly, weakening trade secrets means that those in the intermediate category are unlikely to choose secrecy over patenting. Taken together, the secrecy requirement and the relative weakness of the trade secret law help ensure that the law protects those who would otherwise rely on secrecy without law, and encourages disclosure in those cases, while not displacing patent law as the means of protection for self-disclosing inventions. Put another way, the secrecy requirement channels particular inventors to the form of IP protection that best achieves the goals of society.

II.
Implications for Trade Secret Law

The IP theory of trade secret rights has several implications for the development of trade secret doctrine.

A.     The Centrality of Secrecy

One implication of the theory I articulated in Part I is that the requirement of secrecy is not an accident or a mistake. It is a central part of what makes trade secret law work. A significant benefit of thinking of trade secrets as IP rights rather than as unfair competition torts is that it puts the focus of the legal inquiry first and foremost on whether the plaintiff has an IP right at all. The UTSA, for example, defines the legal rights of trade secret owners by requiring the existence of a secret and defining what constitutes a secret. Doing so prevents plaintiffs from ignoring or glossing over proof of the existence of a trade secret in their effort to prevent what they see as improper use of their information.

This point may seem obvious—of course winning a trade secret case requires the plaintiff to prove the existence of a trade secret. But in fact a number of cases and commentators that have applied the tort theory of trade secrecy have minimized or even ignored that requirement. The Supreme Court itself led courts astray in E.I. du Pont de Nemours Powder Co. v. Masland, where it said that “[w]hether the plaintiffs have any valuable secret or not, the defendant knows the facts, whatever they are, through a special confidence that he accepted. The property may be denied, but the confidence cannot be.”⁷ A number of courts applying the Restatement of Torts have followed the lead of the Masland dictum, holding that defendants misappropriated trade secrets by acquiring or using a secret by improper means or in breach of a confidential relationship without determining that the information was itself a secret at all. An example is Smith v. Dravo Corp., in which the defendant had clearly made use of information obtained from the plaintiff during acquisition negotiations in later entering the market in competition with the plaintiff.⁸ The court found liability on the basis of the defendant’s admittedly troubling business behavior. But in doing so, the court elided the distinction between the use of information that was truly secret, such as the plaintiff’s confidential patent applications, and information that was readily accessible to the public, such as the dimensions of plaintiff’s shipping containers that were already on the market.

There are a number of other examples. In United States Sporting Products, Inc. v. Johnny Stewart Game Calls, Inc.,⁹ for example, the court held that publicly sold, uncopyrightable recordings of bird calls were protectable. The court focused on the labor the plaintiff had put into collecting them, but ignored the fact that they were not secret. In Rohm & Haas Co. v. Adco Chemical Co.,¹⁰ the court ignored the fact that the defendant’s alleged secret process was in fact disclosed in a number of industry publications because it found that the defendant did not in fact learn the information from those publications, but instead from the plaintiff. And in Franke v. Wiltschek, the Second Circuit elevated this idea to a general rule based on Masland:

It matters not that defendants could have gained their knowledge from a study of the expired patent and plaintiffs’ publicly marketed product. The fact is that they did not. Instead they gained it from plaintiffs via their confidential relationship, and in so doing incurred a duty not to use it to plaintiffs’ detriment. This duty they have breached.¹¹

These courts have departed from the principle of trade secrets as IP rights. Perhaps they are blinded by the defendant’s suspicious conduct, or perhaps they view employee mobility itself as suspect. Whatever the reason, they ignore the critical limit on the scope of that IP right. Doing so risks turning trade secrets from a well-defined legal right that serves the broader purposes of IP law into a standardless, free-roaming right to sue competitors for business conduct that courts or juries might be persuaded to deem objectionable. Secrecy is critical to ensuring that trade secret law does not interfere with robust competition or with the dissemination of new ideas. Courts that ignore that requirement undermine the purpose of trade secret law. The dictum of Masland should not only be disregarded but reversed: “The starting point in every case of this sort is not whether there was a confidential relationship, but whether, in fact, there was a trade secret to be misappropriated.” Understanding trade secrets as IP rights, and therefore as premised first and foremost on the existence of such a legal right, will help restore the centrality of the secrecy inquiry. And as a corollary, it may help ensure that the plaintiff clearly defines what it claims to own, rather than (as happens all too often in practice) falling back on vague hand waving.

B.     The Relationship Between Trade Secret Law and Other Torts

The importance of secrecy in channeling inventors between patent and non-patent IP protection has a second implication as well. Requiring trade secret plaintiffs to prove that they own real secrets will do little good if those same plaintiffs can turn to other legal doctrines to provide equivalent protection without the requirement of secrecy. Unfortunately, there are a number of state common law doctrines that offer just that prospect. The common law doctrine of breach of confidence, for example, required only proof that something was offered to the defendant in confidence, and that the defendant disclosed that information. Other common law doctrines, including misappropriation, unfair competition, and unjust enrichment (at least in those states in which it is an independent cause of action), similarly have no elements other than a loose definition of improper conduct. And still other torts, such as interference with contract or “idea submission,” may well overlap almost completely with trade secret claims in particular cases.

Trade secret law should preempt these torts when they are applied to protect information that would, if secret, have been protected by trade secret law. That is, a plaintiff who complains of the defendant’s use of its information, but who cannot prove that the information is secret, should not be able to rely on one of these torts (or any other common law variants) to bypass the requirement that it prove secrecy. If trade secret law does not preempt these torts, the point of the secrecy requirement will be lost, and with it the benefits of dissemination of new inventions. Companies will be unable to rely on the presence of ideas in the public domain; any information might potentially be subject to one of these torts. As a result, companies will be less willing to compete vigorously on the merits. Departing employees will be less willing to rely on information in the public domain to start new companies, and as a result more reluctant at the margins to start those companies. As Jim Pooley notes, “there is arguably little social utility” in allowing state claims based on misappropriation of trade secrets to go forward if the plaintiff cannot prove the elements of a trade secret claim.

Trade secret law should not, however, preempt state laws that have as an object something other than the protection of information. A defendant who breaks into an office to steal information has committed a tort (and indeed a crime) regardless of whether the information in question was secret. Trade secret law should preempt laws within the same general scope as trade secrecy, but not laws that serve fundamentally different purposes.

Once again, conceiving of trade secrets as IP rights helps achieve the goal of preemption of conflicting common law torts. If trade secret law is one tort among many common law torts, there is no reason to privilege it over other torts when the two conflict. But we have a well-established set of principles by which IP rights preempt state common law rules that interfere with those rights. We have those preemption principles because we recognize IP rights as utilitarian rules created by government to address public-goods problems, and the policy decisions implicit in those rules will at a minimum be complicated and may even be overridden by layering on additional causes of action not designed with public goods problems in mind. Treating trade secret law as an IP right dependent on proof of secrecy highlights the policy stakes, and will encourage courts to preempt common law claims that threaten to undermine the balance trade secret law strikes. In so doing, it may further advance the trade secret policy of disclosure by removing state laws that block the flow of non-trade secret information.

C.     Reasonable Efforts to Protect Secrecy

While proof that the plaintiff’s information is secret serves a critical role in channeling towards trade secret protection only those inventions that are best served by trade secret law, the same is not necessarily true of the parallel requirement that trade secret owners take reasonable efforts to protect their secrets. That requirement seems to stem from traditional tort notions of contributory negligence, under which plaintiffs were barred from relief if they themselves contributed to the tort. The explanation I have offered for trade secret law (and for the secrecy requirement) is not one that values secrecy as an end in itself; far from it. The benefit of trade secret law is that it reduces investment in secrecy compared to what would happen absent that law. So there is no reason we should want to establish a minimum investment level as an end in itself. And it may have negative consequences in particular circumstances.

The question then becomes whether reasonable efforts serve some other end. For example, some courts suggest that efforts to protect information as a secret are a sufficiently strong proxy for the secrecy of the invention that we should rely on them as evidence in support of the existence of a secret. But they are surely not perfect evidence; any litigator will tell you that companies regularly label as secret lots of things that clearly are not secret. Even assuming that reasonable efforts at secrecy do offer such evidence, that doesn’t justify the imposition of reasonable efforts as a separate requirement, just the consideration of that evidence in the overall secrecy inquiry. Alternatively, it may be that efforts to protect secrecy serve to put potential defendants on notice of the claim of secrecy, and therefore prevent inadvertent misappropriation. This may be true of some, but not all, efforts at secrecy, so again, it seems to justify reasonable efforts only as evidence, not as a separate requirement. More to the point, it will be true only as to some defendants; others may be aware of the secrecy of the information they take whether or not those secrets were reasonably protected. It seems more logical to cabin the risk of liability for inadvertent misappropriation by imposing some kind of scienter requirement than through this kind of constructive notice through enforcement efforts.

Reasonable efforts to protect secrecy, then, may make sense as evidence of secrecy or even as evidence of scienter, but they probably don’t make sense as a separate requirement.

Conclusion

Trade secrets are IP rights. They serve the same purposes as patent and copyright law—they encourage innovation and the disclosure and dissemination of that innovation, though they sometimes serve those purposes in surprising ways. Trade secret law reduces investments in secrecy and encourages the dissemination of the secret to more people who can make productive use of it. Indeed, trade secret rights may serve the purposes of IP law better than more traditional IP rights, at least for certain classes of inventions. The public disclosure function of the patent system doesn’t work very well in most industries, and doesn’t work at all if inventors opt out of the patent system.

Understanding trade secrets as IP rights allows them to take their proper place in the pantheon of social policy designed to encourage innovation. It also gives us a way to think about how those rights are designed, a way that has significant implications for how trade secret law looks and how it interacts with other laws. Most surprisingly, those implications are ones that offer greater, not lesser, latitude for competitors and departing employees than the unfair competition rationale most commonly articulated as an alternative.

Acknowledgments:

Copyright © 2009 Stanford Law Review.

Thanks to Chuck Adams, John Barton, Bob Bone, Dick Craswell, Zohar Efroni, Paul Goldstein, Wendy Gordon, Tait Graves, Joe Grundfest, Rose Hagan, Eran Kahana, Larry Lessig, David Levine, Jacqueline Lipton, Rob Merges, Mike Meurer, Roger Milgrim, Michael Risch, Sharon Sandeen, Peter Swire, Rebecca Tushnet, and participants in workshops at Stanford Law School and the IP Scholars’ Conference for discussions of these issues or comments on a prior draft.

Mark Lemley is William H. Neukom Professor at Stanford Law School and Of Counsel at Durie Tangri Lemley Roberts & Kent LLP.

This Legal Workshop Editorial is based on the following Article:   Mark A. Lemley, The Surprising Virtues of Treating Trade Secrets as IP Rights, 61 STAN. L. REV. 311 (2008).

The answers to these questions all lie in the scope and limits of intellectual property law’s relationship to market experimentation.  To those familiar with intellectual property law, the phrase “intellectual property for market experimentation” may sound like an oxymoron.  Intellectual property law has long been understood principally as a means of encouraging the production of different forms of technological information and creative works.  Patent law and trade secret law protect inventions, while copyright law protects books, art, and music.  Trademark law, rather than protecting innovations, is justified by minimizing consumer search costs and confusion about products.  However, once these various intellectual property laws have protected the relevant information or symbols, the forces of competition are assumed to operate without further interference.  Intellectual property, on the traditional account, is not needed to promote market experimentation.  The invisible hand itself does that.

I.
The Connection Between Technology and Market Experimentation

But there is a parallel between market and technological experimentation that suggests that intellectual property law could encourage market experimentation in much the same way that it encourages technological experimentation.

Consider two similarly situated companies.  Company A is contemplating testing thousands of materials to see if it can find one that could make standard light bulb filaments cheaper.  No one knows for sure which, if any, will work.  The testing process is expected to cost $100,000,000 and holds about a 50% chance of yielding a successful new technology.  If testing is successful and Company A obtains exclusive rights to the new material for a period of years, the firm would gain over $200,000,000.

Company B, meanwhile, is considering promoting a more efficient type of light bulb that was discovered long ago but was never marketed effectively.  A marketing campaign costing $100,000,000 is expected to hold about a 50% chance of convincing consumers to accept the bulb, which produces a different tint of light from conventional bulbs and which, though cheaper in the long run, requires higher up-front costs.  If the campaign is successful and the firm obtains exclusive rights to the newly commercialized bulbs for a period of years, the firm would reap over $200,000,000.

The social benefits and costs of providing exclusive property rights are about equal in these two cases.  It makes no economic difference whether the source of risk arises from scientific or market uncertainty.  The applicable legal rules, however, present a paradox of differential treatment:  Company A likely can secure exclusive rights, while Company B likely cannot.  This distinction may help explain why some technologies exist for years before being effectively commercialized.

Indeed, the hypothetical Company B may help explain the historical experience with compact fluorescent light bulbs.  The basic technology associated with these bulbs is now more than a half-century old and thus in the public domain.  By the 1980s, several companies were engaged in limited manufacturing of this type of bulb.  Yet, despite the availability of the technology, the bulbs remained for decades only a small fraction of the market.  The new bulbs faced numerous marketing and consumer-information hurdles, many of which remain to some extent unsolved to this day.  Perhaps existing market barriers could be overcome with extensive marketing campaigns, product demonstrations, and warranties, but with hundreds of competing manufacturers, no one firm would reap any significant rewards from such efforts if they made the new product successful.

This absence of protection suggests that the existing level of market experimentation may be inefficiently low.  Empirical assessment of this is difficult; one cannot easily measure what does not exist.  Instead, the empirical evidence must come from businesses that have been created.

The Netflix example advances the case because it is an example of a business that easily might not have existed but for the possibility of sufficient legal and nonlegal first-mover advantages.  Also essential was the persistence of its multimillionaire founder, who could personally afford to capitalize the initial business and who had a sufficient track record of success that he could attract venture capital (albeit not without difficulty) as the business was developing.  We cannot identify with certainty business ideas that would have been successful if only they had been implemented, but we can show that even businesses like Netflix that proved to be phenomenally successful may at one time have appeared to be marginal projects or likely losers.  Indeed, even after Netflix took off against the odds, there were many analysts in 2002 and beyond who doubted that it would be able to survive competition from Blockbuster and Wal-Mart.  Netflix’s own disclosures of market data during its 2002 initial public offering have been blamed for attracting Blockbuster and Amazon.com as competitors, further showing that competitors do try to free ride on the market experiments of other firms.

We can also see inadequate protection for market experimentation at work in explaining why some seemingly obvious products and product features take so long to emerge.  A possible example of this phenomenon is luggage with wheels, a feature that came into common use only in the late twentieth century.  The idea that adding wheels to luggage might be useful is old in the art.  A 1914 patent, for example, protected a device that secured wheels to a suitcase.  Even such a patent would provide little market exclusivity, given the myriad of other ways that one might attach wheels to a suitcase.

The chief impediment to wheeled suitcases, as in the case of the light bulb example discussed above, appears to have been not technological uncertainty but market uncertainty:  here, uncertainty about consumer demand.  If a widespread market experiment with wheeled luggage proved successful, established luggage companies would surely copy the innovation.  This was, of course, precisely what happened once a suitcase with wheels was finally marketed successfully.  While we cannot eliminate the possibility that other factors, such as technological complications, may have contributed to the delayed widespread introduction of luggage wheels, the inadequacy of incentives to engage in market experiments seems likely to have played some role.

II.
Intellectual Property Protection for Market Innovation

Whatever the empirical evidence, there are strong theoretical arguments that incentives for market innovation are insufficient.  Industrial organization scholars have long recognized that in many contexts, free entry will not lead to socially optimal entry.  Similarly, the literature on business management has recognized that, while first movers enjoy significant advantages, second movers have significant advantages too.  A market experiment can produce information of a type generally ignored by the literature—information about whether consumer demand and other market conditions will permit commercial success.  Straightforward economic models can show that a system of free competition with no exclusive rights for market experiments may provide inadequate incentives to induce socially valuable experiments such as the introduction of a new (but technologically uninnovative) product or service or even the marketing of an old product or service to a new geographic market.  Intellectual property protection for market experimentation could raise social welfare, even if the exclusive rights impose substantial losses.

Modern intellectual property theory posits that exclusivity imposes a static cost, but that the dynamic benefit of encouraging information production and dissemination may make this cost worth bearing.  This same logic applies to market experimentation.  The effective launch of a new product or service may require substantial investments.  Commercial success or failure produces information about market demand and supply upon which competitors can often free ride.  To be sure, early experimenters will gain some first-mover advantages, as they also do with technological innovations, but late-entering competitors obtain two important second-mover advantages against early market experimenters.

First, they do not have to bear the cost of investing in market development.  Second, they can copy the first experimenter’s market successes and avoid repeating its failures.  Once such market information is created, consumers would benefit from competition, but without a sufficient guarantee of exclusivity there may be no one with enough incentive to undertake the risky initial investment in developing and testing the market.  Just as patents encourage risky but ultimately beneficial technological experimentation, some form of intellectual property protection could result in a socially beneficial increase in market experimentation and entrepreneurial activity.

Economists, ranging from leading twentieth-century champions of free markets to modern analysts of matters as seemingly diverse as high-tech entrepreneurship and economic growth in developing economies, have long recognized this parallel between market experimentation and technological experimentation but, along with legal theorists, have failed to draw a connection to intellectual property law.  Leading economist Frank Knight, for example, viewed trade secret law and the patent system as important means of addressing the problem that, “owing to the low cost of indefinitely multiplying an idea, it is usually difficult to capitalize an increase in productive power.”¹ Knight recognized that existing intellectual property law would make it difficult for an innovator to “secur[e] any permanent gain” from an “improvement of business organization and methods” because such an improvement was “usually neither patentable nor capable of being kept secret.”²

Yet, after thoughtfully identifying this problem, Knight largely dodged solving it.  He asserted that “there is no evidence of any unwillingness to make expenditures in this form of improvement,” even though that “fact” was puzzling to him.  Perhaps Knight simply assumed that the large amount of market innovation precluded the possibility that there might be inefficiently little of it.

Professors Scott Shane and S. Venkataraman, two leading scholars in the study of entrepreneurship, also recognize the parallel between market and technological experimentation without recognizing how intellectual property law might help.  Shane and Venkataraman view entrepreneurship as “the crucial engine” driving change in a capitalist society.  They also recognize that entrepreneurship is tied to the production and exploitation of new information, and they expressly note that the opportunities for entrepreneurship “need not be restricted to . . . technological developments.”³ They acknowledge the appropriability problem of information generated by entrepreneurship, noting that the new “information diffuses to other members of society who can imitate the innovator and appropriate some of the innovator’s entrepreneurial profit.”⁴

Yet they accept the legal structure addressing the appropriability problem as a given; Shane and Venkataraman view entrepreneurial activity as depending upon factors such as “[t]he provision of monopoly rights, as occurs with patent protection or an exclusive contract,” and “the slowness of information diffusion or the lags in the timeliness with which others recognize information.”⁵

III.
Objections to Intellectual Property Rights for Market Innovations

There are at least two general objections to the argument for intellectual property rights for market experimentation that may help explain why it has received almost no attention in the past.  The first was voiced long ago by Friedrich Hayek as a general objection to any form of a government-sanctioned exclusive right designed to promote innovation:  The free market already provides abundant incentives to experiment and innovate.⁶ This Hayekian position draws no distinction between technological and market experimentation; thus, it would resolve the paradox of differential treatment present in the law by abolishing intellectual property rights for innovation generally. The Hayekian position has remained a polar position, and, while free competition remains an important alternative to intellectual property in some circumstances, all developed nations now recognize intellectual property rights as a means of fostering technological experimentation.

The second objection maintains that even if free competition is suboptimal in encouraging market experimentation, the remedy of expanding intellectual property protection would be worse than the original disease.  This argument is more formidable than the Hayekian objection.  Even a limited grant of exclusive rights may impose significant costs on society.  Because society generally relies on exclusive rights to encourage technological experimentation, however, it would be highly surprising, given the fundamental similarities between technological and market experimentation, if the optimal policy choice for encouraging market experimentation were always to rely upon whatever natural first-mover advantages exist in a particular market and never to deploy some form of exclusive rights.

IV.
Current Protections for Market Innovations

It is thus not surprising that exclusive rights have been used in some circumstances to encourage market experimentation.  British “patents of importation,” which provided exclusive rights for the importer of a technology already existing outside of Britain, were permissible for hundreds of years.  Likewise, exclusive franchise agreements are a widely employed means by which franchisors encourage new franchisees to risk developing a business in new geographic locations.

Although there is no general public law analogue to franchise agreements, and although the United States’s legal system was the first to reject patents of importation, the goal of promoting market experimentation is not as alien to modern American intellectual property doctrine as it may first seem.  Perhaps as much by accident as by design, our existing systems of intellectual property already include several doctrines that are difficult to explain unless the relevant intellectual property rights are recognized as partially advancing the goal of encouraging market experimentation.

For example, some observers have argued that productive American industry is being overrun by “patent trolls,” companies that produce no actual products but merely obtain and enforce patents.  The conventional theory of the patent system maintains that the basic quid pro quo for obtaining exclusive patent rights is the disclosure of the technology set forth in the patent document itself.  Under that theory, the concern about patent trolls seems inexplicable:  Someone who makes a sufficient disclosure and obtains a valid patent cannot be gaming the system.  But if the patent system is recognized as having mixed goals—both spurring the disclosure of technological information and fostering actual investment in real-world market experiments—then the concern over patent trolls makes sense.  The law should be more generous to firms that have both made technological disclosure in patent documents and risked assets in launching new businesses based on the technology.

A.     Market Innovation and Trademark Law

Trademark law can also be understood as advancing the goal of market experimentation.  The generally accepted purpose of trademark theory is to economize on consumer search costs. But trademarks are also central to allowing an entrant into a new market to maintain its market share in the face of competition.  If, for example, any competitor were permitted to use the label “Netflix” to describe services similar to Netflix’s, then Netflix likely would lose its market share much more rapidly, since the Netflix product would seem less distinctive and attractive.  The goal of encouraging market experiments like Netflix therefore provides an additional justification for trademark protection.

In their classic analysis of the economics of trademark law, William Landes and Richard Posner note that prices for brand-named goods have “seemed to some economists and more lawyers an example of the power of brand advertising to bamboozle the public and thereby promote monopoly.”⁷ Landes and Posner point out that the presence of differentiated prices between brand and generic products may not imply deadweight costs if the basis for the difference is that consumers are paying extra for guarantees of high-quality manufacturing or to avoid the expense of determining whether alternatives are in fact of equal quality.  In the pharmaceuticals market, though, the absence of empirical evidence showing that generic drugs have dramatically inferior quality as compared to brand drugs makes Landes and Posner’s empirical claim difficult to verify.

The market experimentation theory provides a more satisfactory response to the argument that there are costs associated with trademark protection.  Trademarks can serve a useful function even if many consumers, acting solely in their own private interests, are irrationally loyal to a brand.  If there is some static inefficiency to consumers’ preferences for brand names, early market entrants will expect a greater market share and, therefore, engage in more market experimentation.

Market experimentation theory thus may help explain our courts’ general reluctance to commit “genericide” by concluding that a trademark has become generic.  Everyday language suggests that many consumers use “Band-Aid” and “Rollerblades” generically, and yet they persist as trademarks.  Some scholars have argued that this aspect of the law cannot be reconciled with trademark’s goal of reducing consumer search costs.  The market experimentation theory suggests that even so, the first-mover advantages provided by such trademarks encourage market entry for future potential products.  Similarly, protection for branding, such as with Major League Baseball caps, does not seem explained by the theory that consumers associate the New York Mets logo with high quality millinery.  But it may make sense to protect these hats to encourage creation of other businesses that might furnish branding opportunities later.

B.     Market Innovation and Trade Secret Law

Market experimentation theory also helps make sense of trade secret law.  A puzzle of trade secret law is why it extends to nontechnological information, such as customer lists and sales figures.  This aspect of trade secret law is difficult to explain through its two traditional justifications:  that it provides incentives to produce technological information or that it reduces social costs associated with “self-help” remedies.  Indeed, conventional theories of competition might suggest that there should be an affirmative obligation to disclose such information, perhaps justifying a Freedom of Information Act for the private sector.

Explaining these aspects of trade secret law as an appropriate social subsidy to encourage market experimentation makes for a more solid foundation.  The law protects whatever business data can be hidden, thus discouraging subsequent entry, increasing a first entrant’s expected share of rents, and creating stronger incentives for the market experiments that produce the data.  Trade secret law may be overinclusive—it protects copycat businesses too—but, in general, innovators are the businesses that have the most information worth protecting.

In short, existing trademark and trade secret law can be more fully explained by consideration of market experimentation.

C.     Market Innovation and Patent Law

Patent law, on the other hand, does not provide well-tailored incentives for market experimentation.  The problem is that the granting of patents is not dependent upon the extent to which an innovation will promote market experimentation.  This holds true even with—perhaps especially with—the modern advent of so-called business method patents.  Such patents might be granted for innovations that are based on new technologies but that would have been created and marketed even without patent protection.  On the other hand, such patents could be refused for being technologically obvious, non-novel, or outside the scope of patentable subject matter, even though some grant of economic exclusivity might be needed to encourage the risky testing of such ideas in the marketplace.

A more reasonable system would be open to granting exclusive rights based on market, rather than technological innovation, but only in those cases where the market success is truly doubtful—in other words, where the market innovation is nonobvious.  This general problem has been recognized only in narrow areas, such as in protection for “orphan drugs.”⁸ Part of the concern with such drugs is that no firm may have the incentive to invest in the Food and Drug Administration’s approval process if they won’t receive exclusive rights, and so the Orphan Drug Act provides a limited term of exclusivity even for drugs that otherwise have fallen into the public domain.⁹ Legal scholars have not previously recognized that the logic of the Orphan Drug Act could be more broadly applicable to the patent system.¹⁰

V.
Conclusion

Why don’t intellectual property theory and property rights theory more generally contemplate property protection for market experimentation?  It is not because information about market success is a type of information that inherently needs no protection.  Rather, it may be because our property rights institutions, as currently designed, are poorly suited to afford such protection.  Institutional limitations have become theoretical limitations that inhibit a clear conceptualization of the entire field of intellectual property.  With the theoretical limitations exposed, perhaps intellectual property scholars can begin to consider whether it might be possible to imagine new institutions or smaller institutional changes that might improve protection for market experimentation.
 

Acknowledgments:

Copyright © 2009 New York University Law Review.

Michael Abramowicz is Professor of Law at George Washington University Law School.

John F. Duffy is Oswald Symister Colclough Research Professor of Law at George Washington University Law School.

The authors thank Martin Adelman, Shamnad Basheer, Robert Brauneis, Richard Pierce, David Post, Michael Risch, Roger Schechter, and Alex Tabborak, as well as participants in workshops at the George Mason University School of Law, the Boston University School of Law, the University of Pennsylvania Law School, and the University of Michigan Law School for helpful comments and suggestions. We also thank Megan Keane for excellent research assistance.

This Editorial is based on the following full-length Article:   Michael Abromowicz & John F. Duffy, Intellectual Property For Market Experimentation, 83 N.Y.U. L. REV. 337 (2008).

Mission:

The Legal Workshop website provides a single online forum for cutting-edge legal scholarship from the top law journals in the country.

The Legal Workshop features “op-ed” versions of the articles published by the member journals. These concise and lively pieces are written for a generalist audience, combining the best elements of print and online publication.

Each Legal Workshop Editorial undergoes the same rigorous editorial treatment and quality screening as the journals’ print content, but readers are able to offer comments and esteemed academics have the option of submitting response pieces, which are checked for citations and substance.

By aggregating the work of multiple law reviews, The Legal Workshop is able to provide frequently updated content. New article-based content is posted every Monday and most Wednesdays and Fridays. The Legal Workshop provides a one-stop forum for readers wishing to stay abreast of contemporary legal scholarship.
 

Founding Members:

New York University Law Review
Stanford Law Review
Cornell Law Review
Duke Law Journal
Georgetown Law Journal
Northwestern University Law Review
University of Chicago Law Review
 

Acknowledgments:

The generous pro bono work of David Sando, an attorney with Skadden, Arps, Slate, Meagher & Flom LLP, has been and continues to be essential to The Legal Workshop.

The idea for The Legal Workshop was originally conceived by Joe Ross, Volume 59 President of the Stanford Law Review. He and Erin Delaney, then Editor-in-Chief of the New York University Law Review, first solicited potential members for The Legal Workshop in the spring of 2007.

Editors at NYU and Stanford who have played a special role in carrying the torch include Thomas Haymore, Ben Kingsley, Matt Lawrence, Lincoln Mayer, Michael Montaño, Sam Nitze, Sean Nutall, and William Rawson.
 

Recommended Browsers:

For Mac:  Safari 3.2.1. and Firefox 3.0.8.
For PC:  Safari 3.2.2., Firefox 3.0.8., and Internet Explorer 8

Thanks for visiting our new website. We hope you enjoy it.