I. Introduction

The right to exclude others from using a patent has long been a part of U.S. Patent Law. This right has been criticized by those who think that its justifications are weak. They point out that even if a patent holder has unfettered rights to exploit his technology and to be compensated for doing so, he does not have the right to harm society by refusing to allow his patented technology to be licensed, and thus prevent further technology [p118] from being developed in the field in which the original patent was granted. They find injustice in that all potential improvement inventors are precluded, absent a license, from being able to seek or use an improvement patent on the original patent.

Others counter that society is no worse off when a patent holder exercises his right to exclude. They argue that the technology underlying any patent would not be known to the public had the original inventor not applied for the patent. Since no dependent patent would exist without the original patent, society is no worse off when a patent holder exercises his right to exclude would-be dependent inventors.

The conclusion as to which side has the better argument in the above debate depends largely on an individual’s view of intellectual property policy, as well as theory. The presence and nature of compulsory licensing provisions in the patent code often mark the battleground in the war to decide whether the near absolute right to exclude should continue to be protected. This Article argues for a limited compulsory licensing provision in the area of improvement patents, using policy and theory arguments based in a Lockean rationale of intellectual property. It posits that in the context of improvement patents, labor theory requires that compulsory patent licensing be an integral part of any patent system.

The U.S. Patent System is usually analyzed through the lens of practical results and/or efficiency. This Article explores a different analysis by means of labor theory, but comes to conclusions similar to that of other writers: [1] that the U.S. Patent Code should allow for more compulsory patent licensing than it currently does.

This Article will proceed as follows: Part II will give the doctrinal background about compulsory patent licensing and its role in the intellectual property regimes of other countries. It will then go on to discuss compulsory patent licensing in the United States as it currently stands. Part III will discuss the theory of compulsory patent licensing, arguing that labor theory is applicable to the scenario presented by an improvement patent. Part III analysis continues with an application of labor theory to compulsory patent licensing, ultimately arguing that in the context of improvement patents, labor theory requires that compulsory patent licensing be incorporated into the U.S. Patent Code. Part IV will give a conclusion and some final observations on compulsory patent licensing. [p119]

II. Compulsory Patent Licensing Doctrine

A. Basic Definitions

An improvement patent, also referred to as a dependent patent, is defined as one that cannot be used without infringing an earlier, existing patent (pioneer patent). A compulsory license occurs when the government requires a patent holder to license his patent to another. Many licensing agreements are entered into quite readily between a pioneer patentee and someone who has invented (and possibly patented) an improvement to the pioneer invention. These agreements are struck because they are in the best interests of the parties involved: without a license from the pioneer, the improvement patentee (improver) can be barred from using his improvement patent; without a license from the improver, the pioneer cannot exploit his invention to the fullest by incorporating the improvement to the pioneering invention.

There are cases, however, in which the pioneer refuses to grant a license. Sometimes, a pioneer will refuse to license even if he is not using the invention. The general law in the United States, absent antitrust concerns, allows the pioneer to exclude regardless of whether the patent is being used.[2] This Article deals only with the scenario where a pioneer refuses to grant a license to an improver, and makes arguments that compulsory patent licensing should be allowed, regardless of whether the pioneer is using the patent.

B. International Treatment of Compulsory Patent Licensing

Article 5 of the Paris Convention for the Protection of Industrial Property, an international agreement of which the United States is a member, states that members of the convention may provide for compulsory licensing of patents in order to prevent patent abuse.[3] Specifically mentioned as an example of patent abuse is the failure to work.[4] In addition to other limitations, the Paris Convention gives a time period of non-use before compulsory licensing is allowed, and also requires that a compulsory license be refused if the patentee “justifies his [p120] inaction by legitimate reasons.”[5] The Paris Convention does not have any specific provisions that relate to compulsory patent licensing, although denial of a license to an improver may be considered “patent abuse.”

The Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPs) came about as a result of The Uruguay Round of the General Agreement on Tariffs and Trade (GATT) in 1993. TRIPs places further limitations on compulsory patent licensing under Article 31, which requires the following: each case must be considered on its merits; that the applicant has attempted to obtain a license from the patentee; that the license is non-exclusive and non-assignable; that the use is primarily for the domestic market; and that the patent holder receives adequate remuneration. [6] Most important to this Article is the provision made for improvement patents. The agreement states that a compulsory license may be granted in the context of an improvement patent “where such use is authorized to permit the exploitation of a patent (the second patent) which cannot be exploited without infringing another patent (the first patent).” [7] The following conditions apply:

(i) the invention claimed in the second patent shall involve an important technical advance of considerable economic significance in relation to the invention claimed in the first patent;
(ii) the owner of the first patent shall be entitled to a cross-license on reasonable terms to use the invention claimed in the second patent; and
(iii) the use authorized in respect of the first patent shall be non-assignable except with the assignment of the second patent.[8]

It is noteworthy that while TRIPs does not mandate any of these provisions, many of the industrialized nations have followed the language of Article 31 and made provisions for the grant of compulsory patent licenses, both in the context of nonuse of a patent as well as the instance of an improvement patent. The United Kingdom, for example, allows for a compulsory license (as well as an accompanying cross license) when the improvement patent represents an “important technical advance of [p121] considerable economic significance.”[9] Other countries that do not have specific compulsory licensing provisions for improvement patents may nonetheless provide for it in other broadly written laws. Japan and Germany, for example, allow for compulsory patent licensing when permission to use the patent is in the “public interest.”[10]

Among industrialized nations, the United States allows considerably less compulsory patent licensing than other countries. The commonality of such provisions is underscored by their presence in the TRIPs agreement, which allows, but does not require, countries to provide for compulsory patent licensing of improvement patents. In order to understand the unique nature of the U.S. approach, a brief analysis of U.S. compulsory patent licensing follows.

C. Compulsory Patent Licensing in the United States

In Hartford-Empire v. United States,[11] the U.S. Supreme Court stated that a patent owner “has no obligation to use [the patent] or to grant its use to others.” Congress has refused to impose any contrary obligation, and attempts to do so are met with strong criticism. Critics complain that compulsory patent licensing discourages invention[12] and that it amounts to “socialism run rampant.”[13] Because so many view encouraging invention as the main purpose of the patent system, [14] these criticisms have prevented formation of compulsory patent licensing provisions. Only in the areas of atomic energy and air pollution control has Congress made specific provisions for compulsory patent licensing. [15] While these provisions show that Congress is willing to at least consider some form of narrowly tailored compulsory patent licensing, the fact that these provisions deal only with the important areas of public health and nuclear energy underscores United States hesitancy on this issue.

[p122] One area where compulsory patent licensing has taken hold somewhat, albeit not on the legislative side, is in the area of antitrust. Because the patent grant creates a monopoly of sorts, misuse of that grant can lead to a grant of judicially crafted compulsory license to competitors. This concept was visited in Image Technical Services, Inc. v. Eastman Kodak Co., [16] where the Ninth Circuit pointed out the patentee’s right of exclusion is not absolute.[17] The circuit court stated that the right of exclusion cannot include attempts to extend the monopoly beyond the grant of the patent.[18] Defendant Kodak did just that and ran afoul of the Sherman Antitrust Act, according to the circuit court, when it refused to sell or license patented parts for its photocopy machines to independent service organizations.[19] The case was criticized by scholars for being the first time that a court forced a patentee to license a valid patent after a unilateral refusal to sell or license. [20] Most interesting for purposes of this Article is the circuit court’s concern with expansion of compulsory patent licensing. While compulsory patent licensing was granted as a remedy, the circuit court states that the decision was based in part on “Kodak’s failure to strenuously assert its patent rights as a defense from the beginning,” suggesting that better arguments as to the right to refuse to license to others may have swung the case in Kodak’s favor.[21]

The decision in Image Technical Services has been called into question by other courts, reflecting further hostility for compulsory patent licensing. In In Re Independent Service Organizations Antitrust Litigation, [22] on facts similar to those in Image Technical Services, the federal circuit upheld the district court’s grant of summary judgment for the defendant Xerox. The federal court held that, even where an anticompetitive effect is present, an infringer must prove that one of two conditions are met in order for the patentee’s right to exclude to be overcome: that the litigation was a mere sham or that the patentee was obtained fraudulently. [23] The federal court also distinguished Image Technical Services, stating that there was no evidence of illegal tying of patented parts to unpatented parts in this case.[24] [p123] This decision has been characterized as indicative of the general trend in the United States with regards to compulsory patent licensing: that it has fallen out of favor.[25]

Upon comparing the approach of the United States with that of other industrialized nations, one may wonder if any theory of intellectual property justifies the U.S. position on compulsory patent licensing. The next part of this Article is designed to explore that topic and discuss the theoretical underpinnings of compulsory patent licensing.

III. Compulsory Patent Licensing Theory

This part of the Article will discuss the theory behind compulsory patent licensing through the lens of labor theory. It will begin with a brief introduction of labor theory, and then discuss why labor theory, long considered a secondary motivation for U.S. Patent Law, has advantages over other theories. It will then discuss how labor theory applies, beginning with an analysis of the reward that labor theory justifies, then moving onto a discussion of how to frame the commons, then ending with a discussion of how to properly frame appropriation of intellectual property. Throughout this part of the Article, theoretical as well as doctrinal applications of compulsory patent licensing will be discussed in an effort to support the thesis of the Article: that labor theory requires that compulsory patent licensing be incorporated into the U.S. Patent Code.

A. Introduction to Labor Theory

Locke’s analysis of property begins with God’s grant of the commons, which is a state of nature in which goods are held in common.[26] It is up to the individual to convert these goods to private property, and thereby appropriate them, by exerting labor upon them. Because there are enough unclaimed goods, at least in a primitive state, everyone can appropriate the goods upon which their labor is exerted without impinging on goods that others have appropriated.[27]

Allowing appropriation of things with which people have mixed their labor, according to Locke, “increases the common stock of mankind,” [28] an [p124] argument that shows Locke’s belief that his theory, while focused on the individual, is also justifiable from the standpoint of social utility. Locke takes further care to protect those other than the laborer (a form of social utility) with his two provisos, both of which must be present in order for his theory to be justified. The first, referred to as the sufficiency proviso, requires that “enough and as good” be left in the commons from which others can draw.[29] The second, referred to as the spoilage proviso, places the condition that things appropriated not go to waste.[30]

Scholars have debated whether intellectual property is property at all, [31] and the extent to which Locke’s theory should be applied to intellectual property given its original application. [32] Locke’s writings on labor are limited to real property, in the sense that his examples and explanations refer to real property only. Labor theory has been extended to include intellectual property by subsequent scholars, many of whom argue that Locke’s theory applies just as readily to intellectual property as it does to real property.[33] This could be because Locke’s writing concerned the wider philosophy of the nature of government,[34] the administrator of intellectual property regimes throughout the world. Another possible explanation for the application of Locke’s theory to intellectual property may be found in the intuitive appeal of the propositions that intellectual creation requires labor, and that the production and appropriation of ideas can readily satisfy both Lockean provisos. [35] The next section of this Article will incorporate this reasoning in arguing that labor theory is applicable to intellectual property, in addition to extolling some of the advantages that is has over other theories.

B. Why Labor Theory Should Apply

The main focus of labor theory is on the individual rather than on general societal welfare or practical considerations. Any application of [p125] theory to intellectual property rights should focus on individuals because they are essential to the process of invention. Without the inventor, there is no invention. Therefore, intellectual property regimes should be concerned with rewarding inventors in a manner that is fair. Any system that does not focus on the individual, but instead focuses on general societal welfare, runs a greater risk of treating inventors less fairly.

Fairness to inventors, in the form of carefully determined intellectual property rights, will not harm society and may even help it. Other applications of theory might place greater emphasis in societal welfare or practical consequence, but in placing their focus elsewhere, these theories run the risk that the inventor may receive an award that is less than what he contributes, thereby discouraging people from laboring. Further, the two goals of rewarding the inventor and ensuring the benefit to society are not mutually exclusive. They may even be complementary, as Locke seemed to believe and as the U.S. Supreme Court pointed out in the landmark case Mazer v. Stein:

The economic philosophy behind the clause empowering Congress to grant patents and copyrights is the conviction that encouragement of individual effort by personal gain is the best way to advance public welfare through the talents of authors and inventors . . . Sacrificial days devoted to such creative activities deserve rewards commensurate with the services rendered.[36]

Stated differently, if inventor talents are neglected through an intellectual property system that is less than fair in its rewards to the inventor, a sub-optimal system for the creation of works results. Labor theory, with its eye on fairness to the inventor, is best suited to avoid this problem.

One possible counterargument to the above line of reasoning is that many creators of intellectual works do not create because of a desire to receive a just reward, but rather to benefit society or gain fulfillment through self-expression. In fact, it is likely that any given inventor has a host of different motivations for laboring over his intellectual creation. This fact is used to question the validity of incentive theory, which is often used to justify intellectual property rights in the United States. Labor theory, however, may provide an answer, in that any inventor who is not motivated by the desire to receive what is deemed by society to be a “fair” property right is free to express that motivation, by simply choosing not to appropriate that information for which he labored.

[p126] There may be a question as to whether appropriation constitutes an independent act that the laborer chooses to perform as opposed to it being an unavoidable consequence of labor itself. But just as property owners are free to abandon their land, intellectual property holders may exercise the right of abandonment with respect to their intellectual creations. Overall, labor theory better deals with the conundrum presented by varying incentive levels by focusing on the individual’s right, of which no inventor is required to take advantage, rather than trying to generalize the motivations of any particular inventor. That said, this Article will assume that most inventors will act in their own best interest in seeking a just reward for their intellectual labor.

Another assumption commonly made in the explanation as to why labor theory applies to intellectual property is that the creation of patentable subject matter involves Lockean labor. Justin Hughes points out that this might not be so, that Locke envisioned labor becoming property because the energy, consciousness, and control that fuel our labor are also our property.[37] Ideas might not be capable of being property, especially when they do not require our energy, consciousness, or control, and even more so when they come about as a result of “huge scales of research,” which pushes idea-making closer to mechanical labor and thus further from what Locke envisioned.[38]

While Hughes’s point is well taken, his other examples of ideas coming about as a result of great mental labor, such as the complete plans to a suspension bridge and a scholar’s research-enriched dissertation, are compelling enough to justify the assumption made.[39] The same assumption was expressed by Eaton Drone, who argued that ownership is created by production. Because ideas are produced by none other than our own person, which is surely our property, they also are our property.[40] While Locke did not apply his theory outside the context of the individual laborer, that does not necessarily entail that he meant to limit rewards only to individual laborers. The effects of group labor on the intellectual commons are just as capable of increasing the “common stock of man” as any other type of labor. Hughes concedes that Locke’s theory, which states that if the provisos are met there is no good reason not to grant property rights, is largely a justification by negation.[41] The provisos can be satisfied when a group appropriates adequately as individual laborers. Therefore, [p127] even if group intellectual labor of a more mechanical nature is not exactly what Locke envisioned, it is still capable of an application to labor theory.

Labor theory, which is used to justify receipt of a government monopoly for mixing one’s labor with information, is applicable in the area of intellectual property because of its concern that inventors be rewarded fairly. What one receives ought to be fair, meaning that one should only get what has become, through one’s labor, a part of one’s self. It is not fair, therefore, to give a monopoly in information with which an inventor has not mixed his labor, any more than it is fair to allow someone to appropriate private property for which no labor has been performed. To do otherwise would be to inflict harm to other’s claim to the commons, a situation that would run counter to Locke’s first law of nature that no harm be done to others.[42] Harm is caused because others who have not labored have an equal claim to the property appropriated by the original non-laborer, and by denying them that claim, their ability to labor is impaired.

The above arguments advocate the applicability of labor theory to intellectual property. Many of the arguments are based on a concept of fairness that runs throughout Locke’s work. Whether or not labor theory is applicable depends in part on how it is employed to define fairness to inventors. Crucial to this inquiry is deciding the reward that labor theory justifies. The next section of the Article explores this topic in greater detail, and goes into further detail on the application of labor theory itself to intellectual property generally, and specifically considers the case of compulsory patent licensing.

C. How Labor Theory Applies

1. Deciding the Reward that Labor Theory Justifies

This Article considers two options when deciding the reward that labor theory justifies. The first is that labor theory may justify a complete monopoly over the information patented. The necessary and logical extension of this reward is that no subsequent inventor can reap rewards of improvement labor should the pioneer refuse to grant a license. The labor of the improver can in no way be rewarded by the grant of monopoly (however limited). The only reward given to the improver is his bargaining chip to convince the pioneer to grant a license that could be beneficial to [p128] both parties. Labor theory’s concern with the rights of others is discussed in greater detail in the section on framing the intellectual commons.

Perhaps this situation can be justified by the fact that were it not for the pioneer, the improver would not have had the ability to labor in any way to improve the pioneer patent. The broad monopoly reward, therefore, is perhaps justified because it does not cause any harm to society or to the improver. But this justification ignores the reality of patents in a modern society, where a patent is often granted after a race between various entities. The U.S. Patent Office regularly adjudicates interference proceedings when two similar patents have been applied for at nearly the same time to decide which party has the superior right to the invention claimed. [43] The existence of such a proceeding underscores that many inventions, if not patented by one party, will shortly thereafter be patented by another. The losing entity in this patent race is most likely in the best position to be the improver, and those improvements exist independent of the winner of the patent race. Therefore, harm is done to the independent inventor/improver in that he receives no reward at all if the patent holder refuses to license. Harm is also caused to society because it has to bear the burden of a long-term monopoly, which was designed to approximate the amount of time that it would have taken someone else to invent the patented information,[44] when a much shorter term would have been the actual time required for independent invention.

Another possible argument that the patent grant may be overbroad comes from an analogy to tort law and is posited by Wendy Gordon: “Once a creator exposes her intellectual product to the public, and that product influences the stream of culture and events, excluding the public from access to it can harm.” [45] Gordon’s argument may be weaker in the patent context, however, because patented information may be less reflective of culture than a copyrighted play or novel. Hughes speaks differently of the effect of exposing an intellectual product to society: “New idea X may be the key to a whole new range of ideas which would not have been thought of without X.” [46] If Hughes is correct, Gordon’s argument loses its muster unless she can articulate the harm in such a way that it outweighs the great benefit of having access to new lines of ideas. It may be that a definitive answer as to the reward that labor theory justifies is elusive by these arguments alone, and must be found elsewhere.

[p129] Giving a broad monopoly over the information of one’s intellectual product seems especially odd in the face of the requirement of disclosure of the invention. Disclosure is required as a condition of the patent grant, and is pointless from the standpoint of an improver who is denied a license to use the patented invention. There must be some other reason for requiring disclosure. There is a notice function that serves as a warning to independent inventors, who have no defense to patent infringement, although scholars have long argued for such a defense. [47] By giving notice of what is claimed, the disclosure benefits other inventors because it gives them the opportunity to “design-around” the pioneer patent. Other inventors may use the patent as a springboard for finding other solutions to the problem, as the Hughes quote above notes.

In all of the preceding scenarios, information in the patent is indeed used, just not in the particular way disclosed by the patent. The monopoly granted, therefore, is limited to the use of information in the particular methods claimed in the patent, and does not include the information itself. The distinction between real and intellectual property is apparent here, as real property that has been granted to one person is less likely to be used by another. Perhaps this suggests the limitations in applying labor theory to intellectual property. Because information is different from real property in so many ways, a theme that is developed in more detail later in this Article, complete monopoly over the information patented is not warranted. The disclosure requirement and subsequent “use” of the information suggests as much.

Labor theory does not seem to suggest that disclosure is necessary; neither does it suggest that disclosure runs afoul of its provisions. Disclosure is, however, quite practical, in addition to being universally accepted. Since disclosure is such an integral part of patent systems, this Article assumes it is here to stay and does not engage in arguments about whether or not it is justified or required by labor theory. Disclosure under the patent system suggests some understanding of a limitation on the monopoly granted to the inventor. But this is only true if the other possible reasons for disclosure discussed above are not a sufficient justification for the disclosure requirement. Whether they are or not is beyond the scope of this Article. The point here is to raise the possibility that the disclosure requirement itself, whether justified by labor theory or not, may lend itself to justify limiting the monopoly granted to the inventor.

[p130] The second option for the reward that labor theory justifies is a monopoly over the value that the patent adds to society. This is called the “value-added” theory, which was stated by Lawrence Becker as follows: “when labor produces something of value to others — something beyond what morality requires the laborer to produce — then the laborer deserves some benefit for it.” [48] Becker suggests that property rights be justified only to the extent that value has been added to society. Justin Hughes puts a further refinement on this theory by stating that “Labor often creates social value, and it is this production of social value that ‘deserves’ reward, not the labor that produced it.”[49] It is important to recognize that all labor cannot possibly be rewarded with a property right. Duplicative labor is one example. Containing the reward for labor to its value added to society is the best option, especially given Locke’s sufficiency and spoilage provisos, which are arguably designed to ensure such containment.

The provisos are often framed as means of avoiding harm that potentially ensues with granting property rights.[50] The harm flowing from an overly broad property right that effectively allows a pioneer to exclude improvers has already been discussed, and will be discussed further in this Article. Therefore, the provisos, in addition to serving as a threshold to whether any property right is granted at all, can also be viewed as designed to avoid harm by limiting the property right itself. This is especially so with something so difficult to define as the property right in information, which cannot be acquired in the same sense that real property can be acquired. Different intellectual property regimes such as patent, trademark, and copyright, with their varying levels of property right, are potentially one way in which this difficulty is addressed. These different levels of property right all involve labor that has added value to society, but are constrained in different ways by the provisos in an effort to avoid harm.

Because the justified reward flows from the value added, it follows that the reward cannot exceed the value added. Hughes expresses this in his normative statement of value-added theory: “People should be rewarded for how much value they add to other people’s lives, regardless of whether they are motivated by such rewards.” [51] This in part depends upon the presumption that no more value can be added to society than labor is expended. Patents granted on certain discoveries may counter that presumption, but the norm is still that much work needs to be done in [p131] to add value to society. Considering the amount of time and resources invested by today’s research and development companies, it is safe to say that labor expended will be excessively greater than value added.

The patent grant, as it is currently constituted, is designed to represent the value added to society. The patent code’s requirements of novelty, non-obviousness, and operability highlight the validity of the value-added theory. If any requirement is absent, the invention will lack value to society because it does not advance the state of the art. Each requirement raises the threshold of the additional value requirement in its own unique way. U.S. patent holders who refuse to license to an improver are effectively allowed to extend their monopoly beyond the patent grant and bar other new, non-obvious, and operable (and therefore, by and large, valuable) inventions from coming forth. In doing so, they reap value that they have not added to society.

This problem is recognized in countries with more liberal compulsory licensing regimes. They seek to ensure the scope of patent rights by defining them in terms of value. TRIPs, for example, allows member countries to provide for the granting of a compulsory license only if, in addition to other requirements, “the invention claimed in the second patent shall involve an important technical advance of considerable economic significance in relation to the invention claimed in the first patent.” [52] By requiring that considerable value be added before an improver can seek a compulsory license, the TRIPs framework acknowledges the role that value plays.

Harm to society is perhaps most stark in the case where only the improver’s patent is commercially feasible. The public would be deprived of an invention should the U.S. pioneer patentee refuse to license, despite the fact that the pioneer patentee did not add the required value to bring forth the invention to a marketable product. The TRIPs framework, if implemented, would possibly avoid this harm, because improvement patents that allow reduction of the patent to a commercialized product will certainly qualify as having the requisite economic significance. There is a question as to whether an improvement patent that prescribes reduction of the patent to a commercialized product will always satisfy the “important technical advance” threshold found in TRIPs Article 31, and further question as to whether that threshold is greater than that of patentability under the patent law of that member state. In any event, the [p132] combination of “important technical advance” and “considerable economic significance” highlights the validity of the value-added theory.

Labor theory stresses that rights should be granted commensurate with value added to society. One may question why so many patents are granted on seemingly worthless inventions. Peter Rosenberg makes an interesting observation on this point:

To satisfy the operability standard, an inventor need not establish that his invention is better than, or that it is even as good as, existing means for accomplishing the same result . . . [T]he law does not ask how useful is the invention. A device that may not operate well may nonetheless be operative.[53]

Rosenberg is correct to point this out, but only seems to focus on one aspect of value that can be derived from a patent: that of a patented product. There is other value to be found in a patent; the advancement of the state of the art and the use of the patent as a springboard to other ideas are but two examples. The U.S. approach can therefore be said to be concerned with value, and uses the imperfect but best way of ensuring it: novelty, non-obviousness, and enablement. Nonetheless, conflict is present between the U.S. requirement for patentability and the TRIPs “value” requirement.

One may question why the requirement of TRIPs is so stringent in requiring, for those member countries that elect to provide compulsory patent licensing to improvement patents, that the value threshold for the improvement be so high. The TRIPs language is probably more in line with Locke, but arguably requires that the improvement patent generate too much value. Perhaps the TRIPs language reflects some sympathy towards the U.S. position that property rights in a patent should only be disturbed in the most rare of circumstances. The TRIPs framework, if implemented, would ameliorate some harm, but still leave many improvers without the benefit of being able to obtain a compulsory license.

The analysis so far leads us to the conclusion that the monopoly, limited by the value added to society, would not extend to allow a pioneer to prevent an improver from gaining protection for improvements to the patent. This result is contemplated under labor theory, which states that an individual should be entitled to no more than that value which he has brought about by his labor, represented by the patent grant, which arguably represents the value given. The Lockean provisos serve as the major [p133] constraint on property rights in information. A determination of compliance with the provisos is dependent on the way in which the intellectual commons is framed. The next section of this Article explains why this is so, and seeks to frame the intellectual commons. It also will analyze whether the Lockean provisos are violated by a regime that refuses to issue compulsory licenses to improvement patents.

2. Framing of the Intellectual Commons

Labor theory dictates that the commons contain that with which man mixes his labor in order to gain a property right, which is manifested in the patent grant. The commons is a grant from God to humanity, and is drawn upon to convert things from their natural state, in which they cannot be enjoyed, to a different state where they can be enjoyed by human beings.[54] This framework is quite applicable to intellectual creations, where information is taken from somewhere, worked upon, and then presented to society as part of the patent grant. This Article considers two options in deciding how the intellectual commons of labor theory are properly framed. The first is to define it as everything in the public domain, meaning all information that is known and can be used freely. The second option is to frame the commons as what lies just beyond the state of the art, with the state of the art including information disclosed in patents, even though this information is not available for use in the particular way claimed in the patent.

Patentable inventions come from two sources. The first source is what lies beyond the state of the art. The second comes from new and non-obvious combinations of information that is entirely or in part in the public domain. This second source of patentable inventions can properly be viewed as being just beyond the state of the art as well, and therefore is probably a sub-category of the first source of patentable inventions. While a patent may be made up entirely of public domain information, the way in which the public domain elements are combined can qualify an invention as patentable. The idea that the commons are composed of what lies beyond the state of the art is reflected in the novelty and non-obviousness requirements in the U.S. Patent System. Any information that is not novel or that is obvious is part of the current state of the art.

The best approach, therefore, is to frame the commons as what lies just beyond the state of the art. But why does it matter how the commons are framed? Because labor theory is thought to be constrained by the Lockean proviso that “enough and as good” (sufficiency) be left in the commons [p134] with which others can mix their labor to receive their own reward. Even advocates of using labor theory to justify broad intellectual property rights have acknowledged that the grant of rights is constrained by the Lockean provisos of sufficiency and spoilage. [55] If the commons are defined as what is in the public domain, a patent on information in the public domain (excluding a patentable combination of public domain information) would not allow for the first Lockean proviso to be met. This is because there is a finite amount of information that is in the public domain. Therefore, any removal of information leaves the rest with less, which is certainly not “as good.” If the commons are defined as what is beyond the state of the art, the sufficiency proviso will likely be met. In order for this to be so, there needs to be an infinite amount of information (and thus, potential inventions) to be created such that any removal of information will by necessity leave “enough and as good” for those who follow. This Article will assume that an infinite amount of information lies beyond the state of the art.

Some scholars feel that patent strays from copyright, which seems to satisfy the sufficiency proviso, in that when a patent is granted, an inventive step is removed from the commons. Thus, enough and as good remains after removal of the inventor’s originality from the commons, but removal of the inventive step does not leave enough and as good. [56] This position is weak, however, because it assumes a finite level of inventive steps in the commons. If inventive steps are indeed a subset of information, which is infinite, then it follows that there must be an infinite amount of inventive steps. The ingenuity of inventors, from which inventive steps are derived, is not likely to run out. Steady increases in the number of patent applications in the United States seem to suggest as much.

Peter Drahos develops this theme further: “If the world is more or less an infinite set of equivalents then it follows . . . that any given appropriation by an individual of an abstract object would be allowed.”[57] Drahos later says that the extension of this would lead to the “radical” implication that a much larger scale of appropriation is justified by Locke. The example Drahos gives is that someone who labors to discover the second law of thermodynamics should be allowed to appropriate it.[58] This depends on exactly what appropriation entails, a theme developed later in [p135] this Article. In any event, Drahos’ example would violate the sufficiency proviso by the nature of what is taken, a law of nature, of which there may not be a set of equivalents, as opposed to an inventive step, of which there seems to be no limit.

Once the commons are properly framed, an analysis of compliance with the provisos can be properly considered. There are various arguments that the U.S. rejection of compulsory patent licensing violates the sufficiency proviso. Some of these arguments depend on what happens to the information that is removed from the commons. It is not part of the public domain in the purest sense because it cannot be used freely. It can, however, be used, as pointed previously in this Article. Thus, it follows that an improvement inventor who has the information of the patent at his disposal, is only better off by virtue of the patent grant. Acceptance of others being better off by disclosing the information in the patent assumes that all known information is held as part of the commons. But, as the previous section of this Article points out this is not so. The sufficiency proviso only applies to the commons, and it is there that “enough and as good” must be left.

The first argument is that there may not be “as good” for an improvement inventor that seeks to invent with the information that lies just beyond the state of the art. The state of the art is surely changed when the information underlying a patent is disclosed. The information underlying a patent is disclosed presumably for the benefit of those to whom it is disclosed — the public. This benefit is only realized to its fullest extent if the public can combine their labor with the information to make improvements. In this sense, there may not be “as good” because, even if the commons is not affected by removing from it the information found in the patent, the state of the art that lies just beyond the patent itself cannot be explored without fear that the pioneer will refuse to grant a license. Thus, “as good” is not left for others who wish to improve upon a patented invention.

This argument assumes that when considering what is “as good,” the reference point is the state of the commons after the patent is granted. The commons includes both what is beyond the state of the art as well as new, non-obvious, and useful combinations of information found in the public domain. Such a definition may include improvement patents, which are nothing more than the patent itself combined with information either beyond the state of the art or within the state of the art in a new, useful, and non-obvious way. Because information or combination found in improvement patents is removed from the commons, the sufficiency proviso, it can be argued, is not met. One counter to this is that if there really are an infinite number of inventions in the commons waiting to be [p136] patented, then no matter how many potential improvement patents the pioneer inventor takes with him when he is granted a property right, there is “enough and as good” left behind. While the sufficiency proviso is not much help in this sense, the spoilage provision might be. Since the pioneer cannot possibly explore all the improvements that are to be made, if he is given a monopoly over them, he may have taken more than he can put to use. Some property would, in a sense, be destroyed without being used.[59] Regardless of whether or not the provisos are violated, a monopoly over things not patented is not justified by labor theory.

The above arguments are dependent upon what is referred to in the literature as the baseline problem. The baseline problem explores what reference points to use when deciding whether or not the property right granted causes the harm that the sufficiency proviso is designed to prevent. Adam Moore argues that the baseline should remain static, and should be defined as the moment after acquisition of the property versus the moment before the property is taken.[60] Moore acknowledges the temptation to use a sliding baseline, stating that “since an individual’s level of well being changes over time the baseline of comparison must also change.”[61] This forces one to accept that “to determine moral bettering and worsening we are to compare how individuals are before the creation of some value to how they would be if they possessed or consumed that value.” [62] This is error, according to Moore, because it lends credibility to the overly broad argument that we are all worse off when value is held exclusively by anyone.[63]

Moore may not be so correct in his analysis when it is applied in the context of compulsory patent licensing. Moore constrains his analysis by the creation of “some value,” and as discussed previously in this Article, he faces an uphill battle in arguing that improvement patents are part of the value created by a patent. They arguably only are a part of that value when the improvement would not have been created were it not for the pioneer patent. This assertion is dealt with in more depth in other sections of this Article, so it will be left for now. Further, the argument that the sufficiency provision is not satisfied by the rejection of compulsory patent licensing may hold true regardless of whether the baseline is static or dynamic. A static baseline that leaves an improvement inventor unable to explore what lies beyond the state of the art, whereas before the patent grant he could [p137] explore it, certainly includes the “morally relevant changes in well being” that violate labor theory principles.

There are also various arguments that the U.S. rejection of compulsory licensing does not violate the sufficiency proviso. The public is left with “enough and as good” because removing the information from what is simply beyond the state of the art does not cause the commons to be diminished. The commons are only changed as the state of the art advances. Moreover, “enough and as good” can be said to be left to other laborers because the improvement inventor is no worse off than he would have been had the pioneer not released the information.

This line of argument, however, ignores the possibility of subsequent independent invention, where neither “enough” nor “as good” is left to the subsequent independent inventor. This is achieved once he, through his labor, much of which was mixed with the commons prior to the grant of the pioneer patent, advances the now redundant technology to the point of patentability. This scenario has been used to argue for a lesser patent term for inventions in some fields where patent races are common. [64] It is exasperated when a subsequent independent inventor makes an improvement, and even more so in the rare case where the improvement is made prior to the grant of the pioneer patent.

Wendy Gordon’s argument that once action has begun, inaction can result in harm, is also applicable here. This is what Moore refers to as a “dynamic baseline.” “That an intellectual product is new, would not have otherwise existed, and may initially bring benefit to the public, does not guarantee that later exclusions from it will be harmless.”[65] The harm is that a potential improvement inventor finds himself with neither “enough” nor “as good” because knowing of the patent affects his thinking in a way that limits him to finding other technological solutions. This is, as Gordon admits, essentially a reliance argument: “[H]aving changed people’s position, the inventor cannot then refuse them the tools they need for surviving under their new condition.”[66] While Gordon did not apply her argument to compulsory patent licensing, it is a good fit. The “change in position” is a result of the change in the state of the art of the information underlying the patent. Refusal of a license is an example of the tools that one would need in order to acquire property, and thus survive. The harm of creation of a reliance and subsequent denial, as well as violation of the proviso, can therefore be avoided by allowing compulsory patent [p138] licensing. The U.S. system, as seen above in Part II of this Article, falls short in this regard.

The analysis so far leads to the conclusion that compulsory patent licensing ought to be accepted by Lockean theory. There are, however, other reasons for compulsory patent licensing’s acceptance. They are discussed in the following section of this Article. They are somewhat more policy based, and explore both the point at which appropriation of intellectual property is justified by labor theory, as well as how appropriation is limited by what is being appropriated.

3. Framing of Appropriation

Appropriation occurs where someone mixes their labor with information in the commons and takes that information for himself. This can be done in various ways, one of which is to receive legal protection in the form of a patent for that information which is appropriated. Many labor theorists view such legal protection as the just reward for the laborer’s efforts in mixing his labor with information. This section of the Article deals with two principal issues relating to appropriation. The first involves the circumstances under which appropriation is appropriate. The second explores how information, by its nature, limits what an intellectual laborer ought to be able to appropriate. The analysis on both counts will be based in compulsory patent licensing.

Some scholars believe that appropriation is permitted only to allow property to reach its full use.[67] Locke may have had full use in mind when explaining his spoilage proviso. Many items by their very nature cannot be put to their full use if they are appropriated. One example is found in the laws of nature, which are so broad and are used as a basis for so many inventions that limiting access to them would keep them from their full use, which may include serving as the basis for other inventions. Drahos’ argument that Lockean theory, if taken to the limits that are often suggested, might require granting intellectual property rights to the laborer who discovers the second law of thermodynamics meets an end here. While the laborer has surely added value to society through his labor, the provisos are not satisfied when a law of nature is appropriated. Sufficiency is not left behind for others, who may not be able to labor in areas beyond the state of the art, and the waste associated with the limitations of being the only person with the exclusive right to develop the law of nature is too great.

[p139] The policy goal of ensuring that property reaches its full use is, according to Seanna Shiffrin, buoyed by the fact that the “world is initially held in common by all people.”[68] In order to reach its full use, some information must be held in common. According to Shiffrin, allowing such information to be held by many highlights the importance of the common grant of the world,[69] an idea held by Locke as an important part of his theory on labor. Regardless of whether or not Shiffrin’s belief that Locke envisioned the world to be held in common is correct, her argument that property ought to be put to full use is compelling. While this seems to be more of a practical consideration than anything Lockean, it is not inconsistent with labor theory, especially because the provisos, in particular that of spoilage, seem designed to encourage property’s full use. Labor theory may not require full use as a condition to appropriate, but it does dictate that full use will be made of whatever is appropriated. It is important to distinguish between non-wasteful possession, which is what the spoilage proviso is designed to prevent, and full use. It is quite possible that property being held exclusively by one person might not reach its full use, but spoilage will also fail to occur. The exact level of use that labor theory requires is not considered here. It is only pointed out that the spoilage proviso is concerned with the level of use, and plausible arguments can be made that full use is contemplated by Locke. An argument that labor theory requires full use in the area of information might be stronger than in the realm of real property given the unique nature of information. This is because information by its nature is more prone to waste when held exclusively.

Appropriation of information is different than appropriation of tangible things, because information is different. Even when appropriated, it can be discovered by another or secretly used by another without any legal implications. Further, information, unlike land, need not be owned exclusively in order to be put to its full use. Full use may occur independent of whether or not the information is appropriated. These characteristics of information limit what is appropriated, because if they can reach their full use in other ways, or even if they are less likely to spoil if not appropriated, then the justification for allowing private ownership is not as strong. This is what Shiffrin is getting at when she states that the first condition necessary to justify a person’s appropriation of something is that “things of that sort must be susceptible to justified private [p140] ownership.” [70] Information, for the aforementioned reasons, is an example of something that is less susceptible to private ownership.

Appropriation is thus limited by what is being appropriated. Any intellectual property system that grants to the appropriator the right to exclude others from use under every circumstance goes too far. Information must be more fully exploited in order to reach its full use, and overly broad rights have the potential to hamper that full use. “Wasteful appropriation would frustrate the charge to make the common grant work to humankind’s benefit.”[71] Although something less than full use might have been tolerated by Locke in order to ensure the just reward for value-adding labor, waste is certainly to be avoided. Shiffrin argues that information is more susceptible to waste when it is under exclusive control: “Attempts to control, suppress, manipulate, or monopolize ideas and information run counter to the intellectual spirit of open public discussions that promote learning and appreciation for the truth.”[72]

Compulsory patent licensing in the context of improvement patents should be allowed because it facilitates full use, or at the very least more use, especially when there is nonuse of the patent. Allowing complete control over the information in the pioneer patent as part of appropriation can prevent information from reaching its full use. Information that cannot be used to bring forth other ideas and inventions is not being put to its full use. One of the unique characteristics of knowledge, and something else that sets it apart from real property, is its ability to beget more of itself. The difference between real and intellectual property is again evident. The full use of real property is often a unique characteristic of information that is opposed to real property, the full use of which is often reached by exclusion of others. Some pioneer patents may even have some of the characteristics of the laws of nature in that they can form the basis for many other types of inventions, and it would be impossible for the patent holder to explore all the possible areas of application that the patent might have.

In conclusion, Locke’s concern with avoiding spoilage, and his possible desire to ensure full use, cuts in favor of Congress placing a provision for compulsory patent licensing in the U.S. Patent Code. Compulsory licenses allow inventors to ensure that their labor is not wasted by a pioneer’s refusal to license. The unique nature of information only serves to strengthen the claim that the United States should take action in this regard. Improvement patents should be viewed as a [p141] consequence of information’s ability to give rise to more knowledge. Allowing improvement patents to come forth more readily by this means is required by labor theory, and only serves to strengthen the common stock of mankind.

IV. Conclusion

Various issues need to be taken into consideration in deciding whether labor theory requires compulsory licensing of improvement patents. The first is an inquiry into the validity of labor theory itself as it relates to the creation and appropriation of intellectual works. The analysis above, in addition to arguing that labor theory is indeed applicable, shows that labor theory has some particular advantages over other theories. Once it is determined that labor theory is applicable, it remains to be seen how that theory should be applied. The fact that labor theory has been thoroughly discussed by many different scholars in many fields suggests that any conclusion reached might be easily refutable. Even so, this Article concludes that labor theory is applicable to compulsory patent licensing theory in that, if properly applied, labor theory requires a substantive change to U.S. law.

This thesis is supported by analyzing labor theory in three main areas. The first is the reward that labor theory justifies. Because a property right is the reward for labor, the assumption is that if labor theory is applied, expansive property rights are granted. Locke, however, intended that the property right granted be attuned to the value added to society by virtue of the labor exerted. This value serves as a constraint on the property right granted in the patent. It prevents a pioneer from being able to deny a license to an improver by virtue of the fact that the value added only extends as far as the pioneer patent itself, and not into the value of the improvement patent.

The second area of analysis deals with the intellectual commons and how they should be defined in applying labor theory to intellectual works. The commons are best defined as all information that lies just beyond the state of the art. By removing information from the commons, an inventor changes the state of the art and causes the appropriated information to become the state of the art. In refusing to license to an improver, the pioneer effectually bars the improver from developing information beyond the state of the art. This is in violation of Locke’s sufficiency proviso. Furthermore, the spoilage proviso may be violated if the pioneer is given the right to exclude would be inventors because the pioneer may not be [p142] able to make sufficient use of the property. Only by requiring licenses for holders of improvement patents can the provisos be met.

The third area of analysis is that of the appropriation of property itself. The nature of information warrants that it be put to its full use, and this is more likely to occur when there are many people with access to that information who intend to work with it. The unique nature of information limits what is appropriated because of information’s tendency to go to waste when held exclusively. Compulsory patent licensing avoids the waste associated with exclusive control by requiring that a pioneer patentee not appropriate for himself the information labored on by improvement inventors.

The United States, by granting pioneer patentees an expansive right to exclude, is allowing the monopoly of the patent to extend beyond the rights that the patent grant justifiably bestows upon the pioneer. Courts should not shy away from using compulsory patent licensing as a remedy for antitrust violations. However, the best solution to the problem of compulsory patent licensing in the context of improvement patents would involve legislative reform. It would be one thing to say only whether the lack of compulsory patent licensing in the context of improvement patents is inconsistent with labor theory. However, to affirmatively argue that the current system needs to be changed is somewhat more bold.

If it is true that labor theory calls for compulsory patent licensing in the face of improvement patents, then one might inquire as to the limitations that labor theory places on compulsory patent licensing of improvement patents. Although labor theory does not require as much value be added by the improvement patent as does the permissive TRIPs compulsory license provisions, the author is sensitive that some concessions will need to be made if any legislative reform is to come forward. The tendency of the law to shy away from compulsory patent license, even in the antitrust arena, proves as much. Whatever the outcome, should any legislation that provides for more compulsory patent licensing in the area of improvement patents be enacted in the United States, it will help ensure the just reward to inventors that Locke envisioned.