The Very Basis Of Our Patent System... Is A Myth
from the time-to-rethink dept
The theory of patent law is based on the idea that a lone genius can solve problems that stump the experts, and that the lone genius will do so only if properly incented. But the canonical story of the lone genius inventor is largely a myth. Surveys of hundreds of significant new technologies show that almost all of them are invented simultaneously or nearly simultaneously by two or more teams working independently of each other. Invention appears in significant part to be a social, not an individual, phenomenon. The result is a real problem for classic theories of patent law. Our dominant theory of patent law doesn’t seem to explain the way we actually implement that law.He notes that if it's true that most inventions are being simultaneously discovered by multiple parties, and we still believe in the basis for the patent system as put forth today, we should be rejecting nearly all patents filed today:
The result is a real problem for classic theories of patent law. If we are supposed to be encouraging only inventions that others in the field couldn‘t have made, we should be paying a lot more attention than we currently do to simultaneous invention. We should be issuing very few patents—surely not the 200,000 per year we do today. And we should be denying patents on the vast majority of the most important inventions, since most seem to involve near-simultaneous invention. Put simply, our dominant theory of patent law doesn‘t seem to explain the way we actually implement that law.The paper has an extensive discussion on simultaneous invention, and just how prevalent it is -- something we've discussed before. The point raised is that nearly every major innovation has involved simultaneous invention, and appears to be the sort of "natural" progress that would happen no matter what -- contradicting the idea that patents are needed to act as incentive for the invention.
Multiple, independent studies show that what Merton calls -- singletons are extraordinarily rare sorts of inventions. Indeed, Lamb and Easton call multiple, simultaneous invention -- the pattern of scientific progress. Merton's classic work suggests that inventions occur not merely because an individual did something particularly creative or surprising, but because the time and conditions were right. There are two components to this idea. First, invention is not a discontinuity, but an incremental step in an ongoing process. Inventors are working with the tools they are given and trying to improve those tools or use them to make something new. Schoenmakers and Duysters study 157 different inventions and conclude that they are largely based on extensions of existing knowledge.It's not difficult to figure out why this is true. While patent system supporters continue to insist that patents are "necessary" as the incentive to innovate (we hear all the time in our comments how without patents, no one would bother inventing), the truth is that there are much more effective and useful drivers of invention and innovation: need and consumer demand. Other studies have shown that the major driver for invention is one's own need. That same study found that a second big driver of innovation was to get past a certain hurdle that would make an overall market bigger. In such scenarios patents don't make sense, because lots of players in the market are willing to collaborate to get past the initial hurdle. Giving a patent to a single player would hinder much of the innovation (as we've seen).
Second, invention by one and only one person or group is exceedingly rare. Far more common are different groups struggling with the same incremental problem, and achieving the same solution at roughly the same time. Ogburn and Thomas conducted the classic study here. They document 148 instances of simultaneous invention. Only rarely, they find, does an inventor come up with an idea that is not developed in similar form by others working independently. Data on litigation tells a similar story; empirical evidence suggests that between 90 and 98% of modern patent lawsuits are against independent inventors, not copiers.
Lemley goes on to point out that this issue of simultaneous invention isn't just the aggregate of "regular patents" obscuring the big genius breakthroughs. In fact, he finds that the big breakthroughs weren't really big breakthroughs at all:
In fact, however, the evidence does not simply show that most inventions result from simultaneous independent invention. It also shows that the vast majority of the most important inventions of the past two centuries—the pioneering inventions that seem with the passage of history such radical departures from the prior art—were themselves the result of gradual social processes in which multiple inventors developed the idea at about the same time.From there he goes through the history of a variety of such "breakthrough" inventions, and dispels the "myth of the lone genius" in each case. One by one, he goes through the steam engine, steamboats, the cotton gin, the telegraph, the sewing machine, the telephone, the light bulb, the movie projector, the automobile, the airplane, radio, television, the computer, the laser and polymer chemistry -- and in every case shows how lots of people basically came up with the same basic inventions at about the same time, often independently. In fact, sometimes the names most associated with the "invention" were stragglers, who were well aware of the work of others in the field. Lemley does admit to a few true independent inventions... but, in almost every one of those cases, the invention was more of an accident, such as with penicillin, vulcanized rubber, the pacemaker and film -- all of which resulted from accidents or mistakes, rather than any sort of "individual genius" that needed to be rewarded.
From there, Lemley shows how this basically destroys the very basis on which patent theory rests. It's supposed to be about creating the incentive for what wouldn't be created otherwise. And he notes that patents are "costly" to society, but that we as society agree to bear those costs under the (false) belief that we get these unique inventions because of it. As Lemley notes: "If we are patenting things we would have obtained without the cost of a patent, on this theory, we‘re wasting our money and probably harming rather than helping innovation downstream."
He also debunks the "ex post" theory for the patent system, which argues that even if patents aren't needed to create the incentives for the initial invention, they are then needed afterwards to incentivize the development of the invention into a product for the market. This theory falls apart quickly for anyone who understands basic free market economics: you don't need a monopoly to bring something to market. In fact, monopolies tend to hinder needed innovations. As Lemley points out, existing incentives such as brand reputation and first-mover advantage seem to suffice. And limiting a market to just a single provider, such as the one who invented it, often seems to lead to the invention being mistargeted, away from the most useful application:
Marconi thought the use of wireless radio technology would be to permit ships at sea to communicate with each other; while that is in fact a use, it is hardly the most important one. Armstrong, the inventor of FM radio, thought he had invented a way of extending the reach of AM radio and broadening its bandwidth; he missed the things (like the absence of static) that actually made FM a success. Bell described his telephone as an improvement in telegraphy, and Western Union turned down an opportunity to buy the patent for $100,000, rejecting the telephone as inherently of no value to us. IBM didn‘t foresee the market for personal computers. The transistor was originally conceived primarily as useful in hearing aids. The steam engine was developed to pump water out of flooded mines. Railroads were originally envisioned as a way of getting goods to canals, which would be the dominant form of overland transportation. The VCR was initially marketed to TV stations as a means of airing reruns. And so on.Next up, Lemley debunks the "disclosure theory" as a basis for patents. We've gone through a variety of reasons why disclosure theory is a myth. Lemley's critique is more direct: patents don't disclose anything useful, for the most part. This was highlighted in the recent This American Life episode on patents, in which a software developer with a patent admitted that he didn't even understand what his own patent said. Lemley puts it simply:
The problem is in part one of law; the Federal Circuit has permitted a number of vague general disclosures that don‘t in fact communicate very much to anyone, and patent lawyers often have incentives to write those vague disclosures. So even those who read patents hoping to learn the state of the art would often be disappointed today.He also discusses a few other reasons why disclosure doesn't work, such as the fact that patent applications are kept secret for 18 months, meaning any disclosure is already quite out of date, and the sheer impossibility of finding relevant patents thanks to the terrible classification system and vague wording. Instead of disclosure, reading patents is almost only done for lawsuit avoidance. As Lemley says: "If they read patents at all, it is to know what is owned, not what is known."
The final section of the paper is where Lemley suggests an alternate theory. That is, he isn't convinced that all of this means we should dump the patent system, but that we might need to rethink what patents are for and how to set them up... though even he concludes that his theory isn't entirely convincing and probably needs additional thought. His idea is that "patent races," whereby the opportunity to gain such a monopoly makes inventors rush their work to get the patent, may be a better way of viewing the patent system, with people often "racing" to beat others to the patent. He admits that this setup likely has negative consequences (many of which we see in the market today -- such as blocking out those who independently invent, but come in second or later). However, as an explanation for how the patent system works, it probably presents a much better theory than the standard theories, which he showed didn't work, earlier in the paper.
As always with Lemley, the paper is a great read. While it runs 104 pages, it's quite easy to get through. It's thorough, well-argued and well-documented. I'd be curious, though. to hear any critiques from patent system supporters. Too often when we write about these kinds of papers, we just get angry insults in the comments from patent holders, who insist that it's all just an "excuse to steal." This paper goes to great lengths to back up its arguments, and it would be nice if those who disagree with the position at least were able to do the same.