pageok
pageok
pageok
[Adam Mossoff, guest-blogging, May 4, 2009 at 11:17pm] Trackbacks
Patent Thickets, Incremental Invention, and Innovation:

The story of the sewing machine is an important empirical case study of how the American patent system has long dealt with incremental invention and resulting patent thickets. It's also a self-contained case study, which challenges the principal focus of the literature on recent inventions and recent changes in patent law, such as the rise of biotech patenting since the early 1980s. Given the cutting-edge nature of biotech research and its equally innovative commercialization, this new field presents a moving empirical target for patent scholars and economists. This perhaps explains why recently published studies on patent thickets, at best, have found none, or, at worst, have been inconclusive.

As a single case study, it cannot serve as the basis for drawing definitive conclusions, as more empirical studies will have to be done. The Sewing Machine War, however, does point toward some important lessons for the modern policy debates over patent thickets. One lesson is that the incremental invention of complementary elements of new technology seems to be a common feature of the type of cutting-edge discoveries that the patent system has promoted for more than two hundred years. From the sewing machine to automobiles to airplanes to radios, incremental innovation seems to be omnipresent in the historical evolution of science and technology.

There was even incremental innovation in the invention of the incandescent light bulb, which, contrary to popular myth, was not discovered by Thomas Edison. Just as Isaac Singer invented only the final few elements of a practical and successful sewing machine, Edison invented only the first practical incandescent light bulb. (Of course, both of these were tremendous achievements, and thus this is not meant to denigrate their inventive contributions.) In fact, Edison was even sued for patent infringement by one of the earlier inventors of the light bulb. Unlike Singer's hapless luck with Walter Hunt, however, Edison was able to invalidate this earlier patent under one of the statutory requirements for a valid patent grant. Yet, decades later, the inventive cycle repeated itself again, as Edison was again embroiled in controversy, but this time it was with Nikola Tesla, who successfully patented and commercialized follow-on innovation to Edison's own cutting-edge work in electrical power systems.

Professor Michael Heller and other scholars have given passing acknowledgements to a few of these historical examples of incremental innovation and resulting patent thickets. As I noted in my introductory post, much of the chapter on patents in Heller's The Gridlock Economy is spent discussing alleged patent thickets in biotech. In fact, the only historical patent thicket to which Heller devotes anything more than a sentence or two is the airplane patent thicket of the early twentieth century. Coincidentally, this was also the only patent thicket that was solved through a public-ordering solution — a compulsory patent pool imposed on the patent-owners by federal legislation. In fact, Heller devotes more time to discussing this legislatively coerced solution to the airplane patent thicket than to the nature of the patent thicket itself. Again, the underlying assumption is that patent thickets are a property problem to which a public-ordering regulatory model is the best solution.

As commentators have pointed out in prior posts, especially in my first introductory posting (Who Cares About the Invention of the Sewing Machine?), there are important differences between the nineteenth century and our modern digital age. The computer revolution was spawned by the invention of the integrated circuit in 1958 by Jack Kilby and Robert Noyce, and innovation in computer technology and related fields, such as biotech, has grown exponentially. This is technological growth that surely outpaces anything humanity has seen before. This is undeniable.

But there are perhaps just as many myths about the significance of these differences as there are truths. For instance, one commentator claimed that a significant difference in the inventive activity between today and the nineteenth century is that hobbyists can work on computers in their garages today but that inventive work in the nineteenth century required massive capital expenditures. Yet Elias Howe and many other inventors in the nineteenth century, such as Charles Goodyear, Samuel Morse, Samuel Colt, and Eli Whitney, to name just a few, were what we would now call "hobbyists" — people without formal training or full-time employment in the field of technology in which they hit inventive pay dirt.

One could still argue that, compared with the potential success of a new computer technology, nineteenth-century industrial technology required massive upfront investment outlays — what economists call "sunk costs." This is certainly true with respect to railroads, steamboat fleets and, later in the nineteenth century, steel mills and industrial factories. But the investment dollars were there to be had for those with new ideas.

Charles Goodyear, for instance, toiled for years after his invention of vulcanized rubber in 1839 to convince American firms of the value of his invention. Goodyear's troubles were the same as those that affected Howe — both inventors faced skeptical investors and commercial firms who had already lost boatloads of money on prior inventors who had claimed to have solved the technological problem. In fact, the nascent rubber industry experienced in the 1820s and 1830s a bubble that was equivalent to our own "dot com" bubble at the turn of the last century. Hundreds of thousands of dollars in investments went up in smoke when products made from pure rubber lost their cohesion in hot weather or became brittle in cold weather. Goodyear's invention of vulcanized rubber solved these problems, but it was difficult for him to convince firms and the buying public after so many failures — and such spectacular financial wipeouts. Goodyear struggled, but, more important, he succeeded.

The sheer number of patents and the increased amount of inventive activity in the twenty-first century is another important difference between today and the nineteenth century. For instance, some have pointed out that Sewing Machine War was a manageable thicket insofar as it was resolved through an agreement between only four entities: I.M. Singer & Co., Grover & Baker, Wheeler, Wilson & Co., and Howe. Today, a patent thicket may comprise thousands of patents, which are owned by individuals or corporations spread throughout the globe.

Yet the same technology that makes inventive activity and patenting more common also reduces the transaction costs in finding information and in coordinating commercial behavior. Word processing programs, email, and web-based searchable databases, such as at the U.S. Patent Office and Google, make it possible to research patents, communicate across the globe, and to reach deals quickly and efficiently in ways that Singer and Howe could have only dreamed about one-hundred-and-fifty years ago. Today, one can search every U.S. patent with a single click of a mouse, send email with proposed licenses as attachments, or FedEx a cease and desist letter in a fraction of the time it took in the nineteenth century. In the mid-nineteenth century, travel between Boston and New York City took a couple days, at best, by horseback (the most efficient method of travel over land). And, even after the telegraph was up and running by the mid-nineteenth country, it facilitated only minimal communication (via the dots and dashes of Morse Code). It is very hard for us today to imagine trying to search patents, negotiate licenses, litigate numerous lawsuits, or operate national commercial enterprises under such conditions — as the sewing machine patentees all had to do in the Sewing Machine War!

In sum, we must beware of anachronisms when we assess technological advances, commercial activities, and legal interactions in a bygone era that lacked, not just the problems of our technology, but also its benefits.

To be clear, it bears emphasizing the empirical merits of the Sewing Machine War and its related features, such as Elias Howe's role as a non-practicing entity in this patent thicket. This is admittedly a single patent thicket involving a single commercial product. I would be committing the logical fallacy of a hasty generalization to draw definitive generalizations from this single data point. At a minimum, though, this case study serves as a cautionary tale against the assumptions that dominate the current discourse concerning patent thickets and closely related policy concerns, such as the impact of incremental innovation and the so-called complicating features of modern inventions. To wit, these are not modern phenomena that are necessarily best resolved with distinctly modern regulatory measures that restrict the property rights secured to patentees. They have long been features of the American patent system, which has long dealt with incremental invention and resulting patent thickets.

In my next post, I'll discuss the issue of "patent trolls" and patent litigation, and whether these have proven to be insurmountable problems to the resolution of patent thickets, past and present.

UPDATE: Some minor spelling errors were corrected.

Fat Man (mail) (www):

To be clear, it bears emphasizing the empirical merits of the Sewing Machine War and its related features, such as Elias Howe’s rollas a non-practicing entity in this patent thicket.


He was on a roll in that role.
5.5.2009 12:06am
Ken Arromdee:
Yet Elias Howe and many other inventors in the nineteenth century, such as Charles Goodyear, Samuel Morse, Samuel Colt, and Eli Whitney, to name just a few, were what we would now call “hobbyists” — people without formal training or full-time employment in the field of technology in which they hit inventive pay dirt.

An occasional physical invention can be made at home. But computer programs are constantly being made at home. Moreover, the problem isn't making an invention that hits pay dirt, the problem is making an invention that steps on someone else's patent. It's impossible to write any substantial computer program without violating a host of patents unknowingly.
5.5.2009 12:57am
Adam Mossoff (mail):
Ken Arromdee: Your argument proves too much. Every person who experimented on a sewing machine was infringing someone's patent somewhere. Yet sewing machines were made, invetions occurred, and improvement patents were issued. Innovation moved forward. The same occurs in the computer industry today.

More important, if someone wants to argue to the contrary, I'm receptive to hearing it. I appreciate evidence-based arguments, which is one of the reasons why I began researching patent history -- everyone was making historical claims without actual historical evidence (or very sketchy evidence).

So, if you're going to make these claims, I'm willing to hear them out, but you can't just make bald assertions. You have to provide empirical evidence to back them up. Thus far, it's just an assertion about "garage inventors infringing others' patents," but I follow patent cases pretty closely and I don't recall hearing about this. I do know of lots of new programs being written all the time, but I haven't heard of a spate of infringement lawsuits against home programmers. Can you identify them? Thanks!
5.5.2009 1:36am
Adam Mossoff (mail):
I should speak more clearly than I did in my last post. What I should have said was that every person who tinkered with a sewing machine was infringing any patents on the sewing machine on which the person was working. Singer, for instance, was infringing Howe's patent the moment he started using the Lerow &Blodgett machine.
5.5.2009 1:44am
Bill Sommerfeld (www):
The software industry has determined that the patent system is useless and that the best strategy is to create impenetrable thickets of patents with fuzzy claims as a form of mutual assured destruction.

My understanding is that most software developers do not allow their engineers to look at patents. The stated reason is the "treble damages for knowing infringement". Instead, engineers are rewarded for filing patent disclosures (which nobody other than patent attorneys ever look at). this does not advance the state of the art. instead, the patent system drains resources from development. in short, software engineers ignore everyone else's patents, file lots of their patents, and hope that anyone else that holds a patent we inadvertently infringe will also have inadvertently infringed some of the patents we hold.

This generally deters patent suits from competitors; you still have the patent troll problem to deal with.

I think Bessen and Meurer are correct in their book "Patent Failure": how can something be property if you can't tell where the infringement boundary is? you can't hire a software patent surveyor to tell if your product does or does not infringe -- you have to litigate to be sure, and that's really expensive.

I have no opinion of the validity of the patent system outside software, but for software it adds no value and drains resources from the industry.
5.5.2009 8:35am
Curt Fischer:
My current understanding of the patent system as practiced by the software industry is similar to Ken Arromdee's and Bill Sommerfeld's.

Maybe things will change after I read the Gridlock Economy, but at this point, I have to wonder why "patent reform" needs to be systemic and legislated. Can't the USPTO simply adopt a much more stringent threshold for "non-obvious" when examining software patents?
5.5.2009 9:06am
Ted Sichelman (mail):
Thanks for the insightful post, Adam. The web of cross-licenses within the hard disk drive industry is a modern-day version of the sewing machine combine that protects -- and shares -- incremental invention across a host of manufacturers. Of course, the downside of protecting incremental invention through keiretsu-style cross-licenses is that by creating an effective R &D monopolist, it can dampen incentives for radical innovation, particularly by shutting out market entrants with disruptive technologies -- who generally are unwelcome in these cross-licensing combines. Although new industries may sprout up by avoiding the combine altogether (compare the rise of flash memory as a soon-to-be replacement for hard-disk drive technology), private ordering solutions to patent thickets may entail significant dynamic costs in long-run innovation. Perhaps you have discussed as much on other posts -- and I know it's covered briefly in your paper -- but I think it's a point worth raising in the context of the discussion above.
5.5.2009 9:29am
Ken Arromdee:
Your argument proves too much. Every person who experimented on a sewing machine was infringing someone's patent somewhere. Yet sewing machines were made, invetions occurred, and improvement patents were issued.

This assumes that experimenting on sewing machines is an activity of similar scope and popularity to writing computer programs, and that the harms of restricting such experimentation are similar to the harms caused by restricting the writing of computer programs. That's not true.

Innovation moved forward. The same occurs in the computer industry today.

Software patents have nothing to do with innovation in any non-legal sense. What actually happens is that everyone has to pretend the patents don't exist and be prepared for the possibility of a lawsuit from a patent troll at any time. I do not believe that this is what happened in the sewing machine industry.
5.5.2009 9:44am
einhverfr (mail) (www):
First of all, any engineer will tell you that incremental innovation beats wholesale innovation every time because one can use as many tried-and-true components as possible. A constrained vision regarding innovation in technology (Linux, Windows NT/2000/XP) beats an unconstrained one (Microsoft Bob) in delivering quality products every time.

However, what Massoff misses in these posts is that the scope and complexity of patent thickets in the IT world are new. Far worse, a very large number of software patents do not seem to my mind to meet statutory requirements but whether courts would limit or invalidate the patents can only be proved through costly litigation. The only solution to the problem from a software engineering point of view is to ignore patents and support organizations which are active in challenging and invalidating some patents in the industry and provide legal defence for small software developers. For Free/Open Source software, this means the Software Freedom Law Center and a few others.

The scope of Singer's problems, as bad as it was, was far smaller than what any of us in the IT world have to deal with today.
5.5.2009 10:23am
David Drake:
An heretical and perhaps off-point view from a lawyer who is not an IP practictioner: I've always thought patent protection was out of place for computer software except in the most extreme cases of invention. If copyright law and its much weaker protection were the standard here it seems like it would cut down the controversies and yield better societal results. Sorry--this is one of those "If only the law was x" comments that I generally dislike.

BTW--enjoyed this series very much. I've learned a lot of history, law and economics. Best of all, the comments have been on-point, free from acrimony, and helpful to the discussion.
5.5.2009 10:29am
Adam Mossoff (mail):
I’ll be responding to some of the comments in my next couple posts, in which I talk further about such things as patent litigation and patent trolls, but here are several quick points:

First, I agree that software patents are a complicated issue today, and I think James Bessen and Mike Meurer’s book identifies some of the problems with computer patents today. Their economic analysis is outside of my professional bailiwick, but I would like to point out that their claims have not gone unchallenged within by those who do have the economic wherewithal to assess such things. So, for those who have read Patent Failure by Bessen and Meurer, I’d also recommend taking a look at some of the work by Professor Rosemarie Ziedonis of the University of Michigan, who has critiqued their metrics. I haven’t yet worked through her analysis in its entirety, but it seems evenhanded, because Professor Ziedonis also acknowledges some of the strengths of their empirical study (and so she’s not just providing a knee-jerk critical reaction).

Second, another problem with Bessen’s and Meurer’s book is that it reflects a certain time bias in its assessment of software patents vis-à-vis patents for “old technology.” Their analysis claims to show that the patent system promoted economic efficiency in the nineteenth century, but this conclusion has the benefit of a time horizon of over hundred years of working out the details of the patent system. Software patents are relatively new -- the began in the 1980s but really did not take off until the early 1990s -- and so what we might be witnessing in the software industry is the inevitable institutional problems in the implementation of a legal system to protect rights that were previously undefined and unprotected. When the automobile industry first took off at the turn of the twentieth century, there were all sorts of patent problems and several hundred competing car companies. It took a couple decades, but the industry eventually moved beyond these initial growth pangs and settled into a market of several dominant companies. There was lots of legal and commercial tumult at that time, and, in retrospect, it would have been destructive to the entire industry if the response to this tumult was to throw out all legal protection of automobiles (patents or otherwise). So, before we condemn software patents as completely unworkable, let’s give the system time to work out the kinks. Otherwise, we seriously risk throwing the baby out with the bathwater.

Third, and last, Ted Sichelman is right about making sure that we always keep in mind the concern about dynamic efficiency. But the problem with such complaints about the potential efficiency-choking aspects of private ordering is that such concerns are almost always counter-factual, especially when proffered on the systemic level. I am deeply suspicious of any quantitative claims about lost innovation, because, frankly, we just don’t know how much or what will lead to new inventions and innovation -- beyond knowing that political and economic freedom, rule of law, and enforcement of property rights (both tangible and intangible) strongly correlate with innovation and economic prosperity. For example, see the United States versus the Soviet Union. (I would further argue that there’s a necessary philosophic connection here, but that is beyond the purview of my comments here.) So, for both philosophic and economic reasons, my default rule is that, absent specific empirical knowledge about the harmful impact of private ordering on dynamic efficiency, I prefer private-ordering to systemic changes in the legal system that mandate public ordering.
5.5.2009 11:18am
Soronel Haetir (mail):
David Drake,

The problem you run into is how do you define extreme.

Was the invention of RSA extreme enough? How about DES )although I believe DES was published as an open standard and not patented). At the other end is something like the XOR operation. How do you draw the line?

This brings to mind an example of incremental invention I saw awhile back, the home ice cream maker. Was adding a motor to replace the hand crank really such a leap as to warrent the patent protection that it received?
5.5.2009 11:20am
einhverfr (mail) (www):
Bill Sommerfeld wrote:

The software industry has determined that the patent system is useless and that the best strategy is to create impenetrable thickets of patents with fuzzy claims as a form of mutual assured destruction.


That is the exact problem. And for an engineer, even if you DO read the patent, it is almost impossible to tell if you are infringing on it because either the patent is so vaguely worded it could cover any software of a specific class, or it could impact dependencies of yours that you are not intimately familiar with. When you have a few new physical parts which are new enough to be patent-worthy, that is one thing. When you have potentially hundreds of thousands of dependencies, each of which could infringe on any number of patents of uncertain scope.

In computer hardware you have a subset of these problems. Unlike software, the patents are usually of certain scope, but you have tremendous patent proliferation problems.

Also the typical argument against worrying about patent thickets as relate to physical inventions like computer hardware and sewing machines doesn't apply to software. In normal hardware, patents expire creating a fairly large safe harbor area so patent thickets are generally self-correcting. In software (and business process patents as well), it is not clear that this is the case because the patent process doesn't work well for what amount to process inventions not tied to physical equipment. This leads to uncertain scope, which means that costly litigation is an ever-present possibility.

Once again, if you find a patent thicket in biotech areas, it will be on a scale similar to what we see in The Sewing Machine Wars. A few key components will have possibly overlapping patent claims, and these aren't common. In software, the problem is fundamentally different.

Hint:

A search at uspto.gov for "Microsoft and communication" yields over 35000 results! A search for "IBM and communication" yields almost 42000 patents! And given the low bar for colorable claims of patent infringement, it seems that the only solution is to ignore patents as a whole.
5.5.2009 11:27am
Ken Arromdee:
I am deeply suspicious of any quantitative claims about lost innovation, because, frankly, we just don’t know how much or what will lead to new inventions and innovation -- beyond knowing that political and economic freedom, rule of law, and enforcement of property rights (both tangible and intangible) strongly correlate with innovation and economic prosperity.

Unfortunately for this theory, a patent is a government-granted monopoly. We have them anyway because we think that a certain amount of government interference in the market is beneficial, but government interference is not without its downsides and we should be deeply suspicious of it.

You could just as well say that a tax is enforcement of property rights because it gives someone a property right to your tax money.
5.5.2009 11:34am
einhverfr (mail) (www):
Adam Massof wrote:

When the automobile industry first took off at the turn of the twentieth century, there were all sorts of patent problems and several hundred competing car companies. It took a couple decades, but the industry eventually moved beyond these initial growth pangs and settled into a market of several dominant companies.


However, what you didn't see in the automobile industry was wholesale disregard for patents.

However..... Let me explain the problem.......

Microsoft hasn't sued (outside of countersuits) anyone that I can find for patent infringement. Same with IBM. The big companies get patents for countersuit reasons. They don't go out looking for competitors to sue. Consequently, if I infringe on Microsoft's patents, they might issue a press release but that will probably be the end of it. Suing me, OTOH, would be bad press and somewhat legally risky. So they don't.

Now, instead what we see is that the big vendors are the ones asking for the legal protections to be thrown out. This did not happen with the automobile. The reason is, it is a losing business strategy to sue lots of folks for patent infringement (see Rambus's story for example). To make a business of it, you have to sue a few wealthy companies. This means that Microsoft and IBM are patent lawsuit TARGETS but unlikely to initiate lawsuits on their own accord. Microsoft his thus been sued (and lost) many times and have not initiated a new lawsuit against folks for patent infringement.

The good thing about Free/Open Source Software is that there aren't many targets big enough to be worth suing. The bad thing is that there aren't many targets big enough to defend themselves.

The way the industry is going is towards de facto non-protection. This is done through:

1) Collective defence (vendor-neutral organizations which provide free assistance in invalidating patents)
2) Target elimination by decentralization. Basically it isn't worth it to sue a large number of people one at a time for pennies.

It is now feasible for most software shops to ignore software patents. This was never the case with the automobile or the sewing machine. Eliminating patent protections for software would not fundamentally be disruptive tot he current state of the industry, though the bigger companies naturally have more to gain from such.

(Now, I have issues with the way patents work in biotech but these are separate issues. For example, I think that standard plant patents are OK, but when you have patents one can infringe through gathering seeds from a crop and replanting, which plant patents do NOT protect, you have real problems.)
5.5.2009 11:47am
Adam Mossoff (mail):
Unfortunately for this theory, a patent is a government-granted monopoly.

Actually, this is wrong -- theoretically and historically. A patent is a property right, secured by a government-granted title that is no different in conceptual or normative content than the title that secures property in land, chattels, or water.

Thankfully, I don't have to rest on my laurels with this broad assertion, as I have published extensively on why the "monopoly" conventional wisdom that infects the modern patent policy debates is predicated on bad history and a fallacious understanding of property theory.

On the history of patents as property rights -- defined, secured and justified under natural rights philosophy in both theory and doctrine -- see my two articles:

http://ssrn.com/abstract=892062

http://ssrn.com/abstract=924226

On property theory and the relationship between patents and property theory, see:

http://ssrn.com/abstract=438780

http://ssrn.com/abstract=1239182

These articles contain extensive analysis of primary sources in both patent law, property law, and property rights theory, especially the natural rights theory of property, which dominated nineteenth-century patent law doctrine.

If anyone is interested in understanding the substantial degree to which American courts repeatedl;y relied on common law real property cases as precedent for deciding patent law cases, used common law property concepts (title, tenants in common, licenses, assignments, etc.), and used property rhetoric (called infringers “pirates”), check out these articles. In fact, the courts went even further and directly imported into patent law common law real property doctrines, such as the canon of liberal construction of title deeds. This is the historical source of the interpretative rule that patents are to be construed in favor of patentees in infringement actions (what we now refer to as “presumptive validity”). Notably, this was done under the 1793 Patent Act, which did not provide for an examination system in applying for patents, and thus it was not originally justified as an deference to the administrative work of patent examiners. For further explanation, see my Cornell Law Review article.
5.5.2009 11:59am
einhverfr (mail) (www):

Actually, this is wrong -- theoretically and historically. A patent is a property right, secured by a government-granted title that is no different in conceptual or normative content than the title that secures property in land, chattels, or water.


Sorry, I don't see things this way.

A patent is a government granted right to form a temporary monopoly in exchange for disclosure of an invention. The idea is that we are served by limited terms of these monopolies by the fact that they reward things that help us progress.

The danger in seeing a patent as a property right is that, while this isn't entirely inaccurate, it is a TEMPORARY right, much more akin to a lease contract than a title which secures property in land, chattels or water.
5.5.2009 12:41pm
einhverfr (mail) (www):
The problem with seeing patents as property is that it tends to push for ever-longer patent terms. After all, it is unlikely that ownership in the title of my house will ever expire through statute and my house will revert to the public domain. I favor bringing the terms back to what they have been through most of our history. However.....

The way I see the quid quo pro in this is:

The inventor gives everyone a disclosure on the patent and in return is granted a temporary from the public domain back to him/her. Then the patent expires back into the public domain.
5.5.2009 12:49pm
Adam Mossoff (mail):
einhverfr: I appreciate your responses, but how about reading the evidence before dismissing the claim out-of-hand? In fact, I'm more than willing to engage you in a debate on this issue, but first I need some indication that you're at least willing to hear the opposing arguments (as opposed to responding to one-sentence summaries of them).

I'll confess that my comments don't objective reflect what I've done in terms of my consideration of your arguments. In fact, I freely admit that I'm in the minority among academic IP scholars on the connection between IP and proeprty.

So, on my part, I've read over the years probably hundreds of articles in which the arguments you're presenting in your comments have been stated and restated. Reading the majority of academic scholarship and related literature is actually what prompted me to investigate the historical case law and other documents, as history is often used to buttress arguments in the IP field (see, e.g., the arguments made by the petitioner in Eldred v. Ashcroft). In my historical research, I was surprised to come upon substantial original historical sources -- court opinions, treatises, press accounts, speeches, etc. -- that contradicted the claims made today about the monopoly status of IP (including patents).

So, here's my proposal: Take a look at the research I have uncovered and the respond to that, and I'll gladly join you in intellectual discussion. I enjoy very much all of the comments I've received, and I love even more intellectual discussion, but there has to be a quid pro quo.
5.5.2009 1:26pm
Ken Arromdee:
If anyone is interested in understanding the substantial degree to which American courts repeatedl;y relied on common law real property cases as precedent for deciding patent law cases, used common law property concepts (title, tenants in common, licenses, assignments, etc.), and used property rhetoric (called infringers “pirates”), check out these articles.

I'm sure that courts use precedents involving natural persons to handle cases dealing with corporations all the time. That doesn't mean that corporate personhood is anything other than a government construct.
5.5.2009 1:57pm
Javert:

First of all, any engineer will tell you that incremental innovation beats wholesale innovation every time because one can use as many tried-and-true components as possible. A constrained vision regarding innovation in technology (Linux, Windows NT/2000/XP) beats an unconstrained one (Microsoft Bob) in delivering quality products every time.
Maybe. But thanks to the "unconstrained" visionaries, mankind progressed beyond the wheel.


Unfortunately for this theory, a patent is a government-granted monopoly [that allows] a certain amount of government interference in the market . . ."
This is like claiming that there is no difference between the government using its police power to exclude competitors from a market (e.g., for cable service) and the government using its police power to exclude trespassers from my property. With the former, the government injects force into the free market; with the latter, it uses force in retaliation.

Patents are not a grant from the government. They merely define the proper boundaries of a creator's right to property -- in the same way that a title does for one's land. Patents need the backing of the state's police power so that we do not degenerate into anarchy and the bloody feuds of private armies.

Some have alleged that certain software patents are hopelessly thicketed or are vague. As one who have been through numerous real estate transactions, I can tell you that these problems also exist with titles to property. Sometimes, land thickets can take a year and tons of money to untangle -- e.g., an individual on the title is somewhere in the Australian outback or has long-since died (with no publicly recorded will.) Or the title's boundaries, easements et al. amount to: four steps from the tallest oak tree, and farmer John can use my property to haul his hay.

However, in no case do we say: let's just use the property and deal with the legalities later. Nor, in my view, are such problems justifications for abolishing titles or for blithely ignoring another person's right to property.

Practically speaking, when dealing with such issues, we have to assess whether the risks -- including time, money and the degree of uncertainty -- are worth pursuing the project. When we do proceed with due diligence, sometimes we successfully resolve the thickets and vagueness, other times we fail. Sometimes, we regard the risks as too high and just walk away.
5.5.2009 2:07pm
Phil R:
If chattels and IP are on equal footing, why would the latter be singled out specifically in the constitution to be given a utilitarian, quid-pro-quo justification?
5.5.2009 2:10pm
einhverfr (mail) (www):
Adam Massoff:

I think my disagreement is one of degree rather than substance.

A patent is analogous to some extent to a title granting property rights to chattels. In fact, if you are doing financial accounting, you treat these in closely approximate ways (grants of land, and most water rights grants, are somewhat different in both these cases for the simple reason that they have an unlimited usable life and generally don't depreciate).

Thus my primary disagreement is that I don't think that permanent property rights grants can be simply lumped together with patents and other temporary grants. I suppose a patent is no different from a land title in the same sense that a gift of land is no different from a gift of a machine capable of producing a certain number of 500000 units of a product before it needs to be replaced. Certainly these three cases (title to a machine, title to a patent, and title to land) are fundamentally different on certain levels despite the fact that they are somewhat analogous. Each poses unique concerns and hence is subject to unique limitations.

The second issue involves the question of natural rights analysis, as you mention in your first paper. In this area, I think your points have some merit. Certainly natural rights analysis is applicable here at least by analogy, and perhaps also by substance. One thing you will not find in my posts is the idea that patents are grants of privilege as opposed to rights. Rather they are grants to a privileged position in the marketplace-- this is a government grant of rights and rights theory is appropriate for approaching this. In short, I don't think that natural rights theory is entirely divorced from the question of patents.

However, I would make an argument that even if we disregard the Jeffersonian version of the history of patent law (or at least supplement it with a fruits-of-labor/social contract approach), natural rights theory breaks down at some point regarding how patents and copyrights are implemented in our Constitution. Perpetual grants for goods, chattels, water, and land are permitted, and we would take issue from a natural rights perspective if these were abridged in the same way that patents are, via arbitrary time limits. In short, patents and copyrights in the US both represent what are fundamentally commercial as opposed to moral rights.

For example, suppose the Federal Government passed a law stating that "no car more than 17 years old may be driven on public highways." Would this run amok with the 4th Amendment's property guarantee more than the 17-year patent term? Suppose the Federal Government ordered all buildings more than 17 years old to be demolished in order to stimulate new construction. Would that be an illegal taking any more than a patent expiring after 17 years would be?

I think there is some middle ground here, which is to the idea that patent is a social contract involving temporary rights analogous to property rights in exchange for disclosure. These rights grant a privileged position in the marketplace.

I am not saying that the rights are not analogous to property rights. I am saying that a title to a patent is fundamentally different from the title to a perpetual resource, such as a perpetual water rights claim, or a title to a piece of land.

Going back to my lease analogy.... A lease provides a set of property rights to the one who obtains it. In fact, the specific analogies to trespassing seem applicable someone (perhaps other than the owner) entering a leased property without the renter's permission. Here is a good example: If someone is prosecuted for trespassing by parking a car in a spot intended for customers of a business (but who is not doing so as a customer), I don't think that the question of whether the property was leased or owned would be particularly relevant (unless it was the owner who allegedly committed the trespass).

So I agree with you that it is analogous to property rights. I disagree with you that a patent is fully analogous to a title to a piece of land, water rights, or any other resource with an indefinite usable life. So I suspect the disagreement is more minor than it appears.
5.5.2009 2:39pm
Adam Mossoff (mail):

If chattels and IP are on equal footing, why would the latter be singled out specifically in the constitution to be given a utilitarian, quid-pro-quo justification?


This is a great question, Phil R.! Luckily, I've already answered it at length in my 2007 article, Who Cares What Thomas Jefferson Thought About Patents? Reevaluating the Patent 'Privilege' in Historical Context, which you can download as a pdf by clicking on the title.

To phrase my favorite economists, I won't be inefficient by repeating what I said there, and I think the explanation is pretty clear and well presented in that article.
5.5.2009 2:47pm
einhverfr (mail) (www):
Javert:

Maybe. But thanks to the "unconstrained" visionaries, mankind progressed beyond the wheel.


Then you misunderstand me. Nearly every major invention I can think of that has changed our lives has been an incremental development on something which came before it. There are a few true breakthroughs, but even here those don't usually become commercially viable until they have been around for a while. An integrated circuit is nothing more than solid state electronics manufactured together on the same semiconductor wafer, for example.

An interesting way to look at this is to see how many of these important breakthroughs came out of Bell Labs, and how long they were around before they were commercialized. Even where they are commercialized, they are done so where it is one component in a larger system. For example, lasers and fiberoptics would replace with light what was transmitted using electrical current, but the fundamental way that voice data would be transmitted over a phone system didn't change.

Another good example is the development of steel sword-making in Europe between the La Tiene Celts and the early Vikings. We see a number of changes which by themselves don't seem terribly impressive despite the fact that they were quite disruptive. First we see simple iron blades (early La Tiene). Then we see iron with steel edges welded (early La Tiene). Then we see coil-forged iron with steel edges (middle La Tiene). THen we see pattern welded steel (Goths, other Germanic tribes 3rd century or so), then we see homogeneous steel swords (7th century, Scandinavia). So we can see incremental innovation, based on a conservative tradition, providing solid advances.

There are places for unconstrained visionaries. We call them "hobbyists." They built the first (mechanical!) programmable computing machines. However, none of these great breakthroughs could be commercially helpful until decades or centuries had passed after they were invented. Presumably the first person to try to ride a horse (then raised as food animals) was just playing around. Big breakthroughs even today happen more often by accident than by intent.
5.5.2009 3:03pm
einhverfr (mail) (www):
One more point to Javert:

However, in no case do we say: let's just use the property and deal with the legalities later. Nor, in my view, are such problems justifications for abolishing titles or for blithely ignoring another person's right to property.


First, such patent thickets are the norm, rather than the exception, in the software industry. They are the exception in industries like real estate.

Secondly, if you blithely ignore someone else's right to property provided that they don't put fences in, no trespassing signs, etc, and don't act fundamentally unreasonable about it (entering someone's house, threatening someone with physical injury, etc), you might end up with an additional partial claim to the property out of it (historical use, adverse possession, etc).

So, actually, you are wrong on both counts.
5.5.2009 3:33pm
Javert:

First, such patent thickets are the norm, rather than the exception, in the software industry.
If you know that, then you know that others have erected fences and no trespassing signs. In which case, you seem to want to use their property without their permission, and to profit from their efforts without compensating them.
5.5.2009 3:53pm
Ken Arromdee:
This is like claiming that there is no difference between the government using its police power to exclude competitors from a market (e.g., for cable service) and the government using its police power to exclude trespassers from my property. With the former, the government injects force into the free market; with the latter, it uses force in retaliation.

All you have to do is define a "cable television entertainment property right", which we can call a "catent" for short. Now the government is merely enforcing the cable company's catent right when it prevents other people from being competitors--if you compete, you're infringing on the cable company's property right in their "catent".
5.5.2009 4:26pm
einhverfr (mail) (www):
Javert:

Not really. As I have pointed out, there is no feasible way to guarantee that one is not infringing on software patents. And as a one-person software shop, I could spend all my time working, or I could spend all my time reading patents and not working. Guess which one will put food on my table.
5.5.2009 4:27pm
Ken Arromdee:
If you know that, then you know that others have erected fences and no trespassing signs. In which case, you seem to want to use their property without their permission, and to profit from their efforts without compensating them.

No, because:
1) Knowing that every program is covered by a ton of patents is different from knowing *which* patents it's covered by. In the analogy, instead of seeing a no trespassing sign at a particular location on a piece of property, there's a box in the local courthouse full of no trespassing signs, each of which says "this sign applies to some piece of property out there". (And this doesn't even consider the issue of submarine patents.)
2) The idea that violating a software patent means profiting from someone's efforts is ludicrous given the way software patents actually work in practice. If there's one thing which software patents don't involve, it's creative effort.
5.5.2009 4:31pm
einhverfr (mail) (www):
Also here is a limitation to natural rights analysis at least in theory (though not in practice regarding American law). Natural rights analysis tends to hold that there are specific unalienable rights which we hold by virtue of being human. These come out of 17th century philosophy and are a well developed train of Western thought.

However, the very premise of natural rights is flawed and based on our CULTURAL ideas of the interplay between individuality and society. "Natural rights" are cultural/social constructs.

This doesn't really have any bearing on whether natural rights analysis applies to patent law. However, it suggests that the distinction between a legal right and a moral right is not as fundamental as one might think.

Natural rights analysis, as it defines how we view ourselves in relation to others in our society, is perfectly applicable to patent law questions where treaty considerations are absent.
5.5.2009 4:34pm
Bill Sommerfeld (www):

If you know that, then you know that others have erected fences and no trespassing signs. In which case, you seem to want to use their property without their permission, and to profit from their efforts without compensating them.

Software patents are not like fences. You can see a fence as you approach it and route around it.

There is no fence. There are no marked property lines. There are just cranky owners that occasionally yell at you and guard dogs which bark at you.

But you might have been walking along the same path once a week for 30 years before someone noticed your trespass and called the cops.

Oh, and because there's a bonus penalty for knowing trespass, you're not allowed to carry a map showing property lines -- it would be inaccurate anyway so it wouldn't effectively keep you out of trouble.
5.5.2009 4:59pm
Javert:

All you have to do is define a "cable television entertainment property right", which we can call a "catent" for short. Now the government is merely enforcing the cable company's catent right when it prevents other people from being competitors--if you compete, you're infringing on the cable company's property right in their "catent".
Sure, and all we have to do is to define "bananas" as "weapons consisting of a metal tube from which a projectile is fired at high velocity into a relatively flat trajectory" -- and we can use them to shoot intruders.
5.5.2009 5:31pm
Ken Arromdee:
Sure, and all we have to do is to define "bananas" as "weapons consisting of a metal tube from which a projectile is fired at high velocity into a relatively flat trajectory" -- and we can use them to shoot intruders.

If you're suggesting that banana guns, "catents" as property, and patents as property are all equally ludicrous redefinitions, then we agree.
5.5.2009 5:44pm
Halbert:
The position of some commenters is that the quantity and vagueness of patents, and their overlapping claims, makes it necessary for a small developer of software to disregard patents altogether, even though it's impossible to write software that does not infringe a patent somewhere. A large company can build an offensive patent portfolio to cross license in case of a dispute, but it puts a small developer at a real disadvantage. Let me know if I got it wrong (the above mashes together several comments and might not be entirely found in any one), but that is the argument for a harm caused by a "patent thicket" in software.
So how does a small software developer function? Maybe a programmer can shrug off the liability, but it's unlikely that investors across the board would be so brazen. If the customer is sophisticated user, he (again, generally) would not knowingly accept liability. I'm not sure whether the UCC indemnity for the sale of goods applies to software, but do you address indemnity in your software contracts or licenses?

Also, to add my two cents to a couple of comments. Based on public information, Microsoft has been listed as a plaintiff in at least 14 patent infringement lawsuits in the last 5 years. Sometimes multiple litigations have been filed against the same party, but there have been suits against seven unique parties over that time span. Finally, its well settled that a patent has the attributes of personal property, even though it's nature is different from tangible property.
5.5.2009 5:57pm
ASlyJD (mail):
Here's my problem with the natural rights view of patents: they fundamentally undercut the infringer's right in their property.

Example: Starbucks has a patent on their little coffee sleeves. I can buy my own cardboard, but I cannot shape it into the shape that Starbucks thinks is most efficient because that would violate their patents. From whence does Starbucks get the right to tell me what I can or cannot do with my cardboard? Most vexingly, this alienation of my rights is present regardless of my mens rea. Thus, I may be liable to Starbucks for my infringing cardboard sleeve even if I don't even know about their work.

Now, I'm willing to accept this restraint on my property rights as part of a utilitarian arrangement, just as I do schools, roads, and all those other government provided things that aren't strictly "life, liberty, and property." But to call a patent a natural right is to shrink all other property rights.
5.5.2009 6:07pm
einhverfr (mail) (www):
Halbert:

A large company can build an offensive patent portfolio to cross license in case of a dispute, but it puts a small developer at a real disadvantage.


That is the fear. In practice, small developers are not worth suing. In practice, they would spend everything they have defending themselves and the the result would be nothing for the patent owner. This a major factor that keeps things from becoming a huge mess.

What is worth noting is that it is expensive to build a factory capable of building automobiles or sewing machines. Consequently if you put a couple competitors out of business with patent infringement claims, few will follow. However with software, the barrier to entry is so low (I haven't paid a dime for any of my current software development tools), you will go broke just from court costs before you secure a practical monopoly.
5.5.2009 6:14pm
Ted Sichelman (mail):
Thanks for you response to my comment, Adam. A few responses. First, there is good (though, not exhaustive) evidence that directly runs counter to your assertion that my sort of "complaint[] about the potential efficiency-choking aspects of private ordering ... [is] almost always counter-factual, especially when proffered on the systemic level." In particular, numerous studies show that startups are more innovative per R &D dollar than incumbents by a variety of measures. Our 2008 Berkeley Patent Survey -- and common sense -- indicates that startups are much less likely to be part of cross-licensing cohorts than incumbents. Moreover, case studies -- such as those in Merges &Nelson's 1990 "Patent Scope" article -- tend to show that competition leads to greater and faster inventive activity. In this regard, cross-licensing among incumbents reduces competition, at least along the technology axis. Finally, incumbents in the hard disk drive and ultrasound industries -- both of which I know fairly well from practicing -- have mainly produced incremental, not radical innovations.

Second, and perhaps more importantly, why should the quantitative burden fall on those in favor of greater competition? You state that we know "that political and economic freedom, rule of law, and enforcement of property rights (both tangible and intangible) strongly correlate with innovation and economic prosperity. For example, see the United States versus the Soviet Union." Although there is a good evidence that strong real property rights have led to economic prosperity, I don't know of good evidence that strong intellectual property rights have led to such prosperity (except perhaps in the pharma field).

Rather, we mainly rely on theoretical arguments to justify the patent system. If we were to follow your "absent empirical knowledge" approach, it seems that based on the available empirical evidence, we should only have patents for pharmaceuticals. Of course, I'm happy for you to point me to empirical evidence otherwise. Assuming you cannot, I don't think the "absent empirical knowledge" approach should be accorded much "weight". So, returning to theory, I think there are a host of strong arguments as to why cross-licensing "combines" -- the result of private ordering generally among incumbents -- present serious concerns of dynamic inefficiency.
5.5.2009 9:57pm
Javert:

I don't know of good evidence that strong intellectual property rights have led to such prosperity (except perhaps in the pharma field).
And except in the fields of music, movies, painting, sculpture, novels, plays, and the entire publishing industry.
5.6.2009 12:15am
Ted Sichelman (mail):
Worthwhile point, Javert. I should have limited my statement to "strong patent rights," which was my intent, since the above discussion is focused on patents.

On the other hand, there are convincing arguments that strong copyrights -- at least along the lines of current U.S. copyright protection -- aren't necessary to spur today's creative industries.
5.6.2009 10:32am
statfan (mail):
"I.M. Singer &Co., Grover &Baker, Wheeler, Wilson &Co., and Howe." is unparsable without knowledge of company naming conventions. You can use semicolons to separate the entities here.
5.7.2009 2:07pm

Post as: [Register] [Log In]

Account:
Password:
Remember info?

If you have a comment about spelling, typos, or format errors, please e-mail the poster directly rather than posting a comment.

Comment Policy: We reserve the right to edit or delete comments, and in extreme cases to ban commenters, at our discretion. Comments must be relevant and civil (and, especially, free of name-calling). We think of comment threads like dinner parties at our homes. If you make the party unpleasant for us or for others, we'd rather you went elsewhere. We're happy to see a wide range of viewpoints, but we want all of them to be expressed as politely as possible.

We realize that such a comment policy can never be evenly enforced, because we can't possibly monitor every comment equally well. Hundreds of comments are posted every day here, and we don't read them all. Those we read, we read with different degrees of attention, and in different moods. We try to be fair, but we make no promises.

And remember, it's a big Internet. If you think we were mistaken in removing your post (or, in extreme cases, in removing you) -- or if you prefer a more free-for-all approach -- there are surely plenty of ways you can still get your views out.