Epstein on Technology Patents

by on December 22, 2007 · 10 comments

One of my favorite podcasts is David Levine’s Hearsay Culture, which I stumbled across this summer. I noticed recently that back in March he did a podcast with Richard Epstein, a giant of classical liberal legal thought, back in March, so I’ve been listening to that episode.

Epstein has long been one of my favorite libertarian thinkers. On most subjects, I find myself nodding along in agreement. But when he gets to the application of patent and copyright law to the technology industry he has a tendency to go off the rails pretty quickly. This is apparent in his Hearsay Culture interview as well. For example, about 25 minutes into the interview, he talks about the problem of patent thickets in the tech industry thus:

The question is, how do we know when there’s a blockade? Well, a lot of it depends on the topology of the landscape if you’re looking at physical resources, and the same thing will happen with respect to intellectual property. So just to take the general question, let’s start with a background on the tech side before we get to the documentary film. Do we think that the addition of any new patent in any particular area is going to increase or reduce the blockade effect of other patents? And the answer to that question is frankly my dear we don’t know in the abstract but the betting would be that the more patents you have, the fewer the blockades. Now why is that? If you imagine these things as being in an array. Suppose you have to go through six steps in order to get a process, and at every one of these steps you’ve got four alternatives. Well if that’s what’s happening, you’ve got a lot of choices at each stage and you’ll be able to bargain one off against another and presumably find a path through this thicket. Somebody comes up with a new invention, now instead of having four alternatives at the first stage you may have five. Or you’ll get a new invention which means that you don’t have to bargain with anybody at stages 6 or 7. And if that guy comes in with a blockbuster patent, he will not be able to charge more than the old 6 and 7 combination could have been able to do, and if he’s really good somebody else is going to come into that same field because the patent doesn’t give you a monopoly over the functionality as such, only the device or the invention that allows you to actualize that functionality. I mean, that’s not 100% correct, but it’s 95% correct for these particular points. Samuel Morse could patent the telegraph, he could not get control over all uses of the electromagnetic spectrum. So new inventions in the tech area generallly seems to me to expand possibilities by giving you alternative stepping stones through the thicket.


Epstein then launches into a very sensible discussion of the difficulties faced by documentary filmmakers, and concludes that documentary filmmakers should have more latitude to include excerpts of copyrighted materials in their films without getting the permission of the relevant copyright holders. He continues:

Why not negotiation [between song owners and documentary filmmakers]? Well, we know what the answer to that is. It’s that you’ve gotta negotiate with 80 different people, each of whom thinks that his song is worth 4 percent of the value of the total package. Well, you do the multiplication, and it turns out that the royalties exceed the yield, and the whole thing gets scrapped.

Now, as anyone who’s looked at actual software patents will tell you, this is exactly the situation faced by a prospective software developer. There are now so patents on the books that it’s basically impossible to even identify all the people who have a plausible claim on a given product, to say nothing of negotiating licenses with all of them. The idea that software patents don’t “give you a monopoly over functionality” is just false, as the Vonage and Research in Motion cases, among many others, plainly demonstrate.

Epstein has a mental model of the way the patent system works that, at least in the high-tech industry, has no connection to the real world. Software patents don’t cover clearly-defined steps in a manufacturing process. They tend to cover broad and vaguely defined categories of functionality like “one click shopping,” (Amazon) “wireless email,” (NTP) and “faster-than-real-time downloading” (Burst). These patents aren’t freak exceptions that happened to sneak by the patent office, they’re the bed and butter of the patent system, at least when it comes to software.

How can Epstein be so sensible with documentary filmmakers and so confused about software patents? I think an interest clue is the fact that apparently, Epstein’s son is a documentary filmmakers. That means that documentary filmmakers aren’t just abstractions in his theoretical model of the world. He has actually talked to some of them and found out about the problems the copyright system is creating for them. I suspect he’s never sat down for an in-depth discussion with a computer programmer about the effects of software patents (or the DMCA, for that matter) on software development.

The ironic thing is that Epstein should know better. As he says around 4 minutes into the podcast:

I always think that Aaron Director, bless his memory, put it very nicely when he said “If you’re dealing with political institutions, examine all new laws under a presumption of error.” And what he meant by that, in effect, was to say that any legal system is going to have a certain degree of drag by virtue of the fact that all legals systems when they introduce new laws have administrative costs that are capable of making errors. So if you know that those two things are around, you want to be reasonably confident that you know what you’re trying to do when you want to stop certain activities from taking place.

The DMCA and software patents, of course, are two new laws that were introduced in the last decade or two over the loud objections of many of the people who would be most directly affected by them. I would think that Director’s admonition would apply here as much as anyone else. Yet Epstein seems to have fallen into the same trap as his progressive intellectual opponents, overestimating his own ability to use the law to re-shape an industry he doesn’t understand very well for the public benefit.

Update: Close to the end of the podcast, he mentions the eBay v. MercExchange case as an example of patent thicket problems, but he seems to regard this is merely a matter of occasional, inadvertant patent infringement, not a systematic problem with patent thickets.

  • Timon

    There is a major confirmation bias in the thinking of (normally) clasical liberal thinkers that Epstein put unusually clearly in the podcast when he said something like, “You can’t understand intellectual property unless you know about these 18th and 19th century liberals who describe why property is a good idea.” He doesn’t pause to wonder if it was in fact those very ideas (in Lawyerstan, DC, or just in his imagination) that were shoehorned into service in an area outside the environment where they evolved. That is to say, a Marxist historian will see class struggle in everything from amoebas to microprocessors, and there is a property vector in classical liberal ideology that can be just as totalizing. The problem – any problem – becomes “How do I apply my proven formula to this new thing?” rather than, “How does this new thing force me to revise what I though I knew?” And of course it is much easier to simply misunderstand or ignore key parts* of the new thing than to allow that your philosophy might not have perfectly anticipated centuries of chaotic human change.

    Despite all that, we should not concede these historical points. First of all, because these hoary traditions like tax-loophole patents, software patents, and business process patents are generally younger than Britney Spears’s singing career. That has to be the beginning of any discussion of these things, especially when patent lawyers and trolls have the chutzpah to invoke the founding revolutionaries in reference to the mischief they have been making for a couple decades, and mainly in the last 8-10 years.

    * Epstein talks at length about open source licensing while betraying no knowledge of the subject, specifically the existence of the Apache, Mozilla/LGPL, and Berkeley Unix licenses, which account for enormous chunks of open source, and many of which are in fact in the public domain.

  • Timon

    There is a major confirmation bias in the thinking of (normally) clasical liberal thinkers that Epstein put unusually clearly in the podcast when he said something like, “You can’t understand intellectual property unless you know about these 18th and 19th century liberals who describe why property is a good idea.” He doesn’t pause to wonder if it was in fact those very ideas (in Lawyerstan, DC, or just in his imagination) that were shoehorned into service in an area outside the environment where they evolved. That is to say, a Marxist historian will see class struggle in everything from amoebas to microprocessors, and there is a property vector in classical liberal ideology that can be just as totalizing. The problem – any problem – becomes “How do I apply my proven formula to this new thing?” rather than, “How does this new thing force me to revise what I though I knew?” And of course it is much easier to simply misunderstand or ignore key parts* of the new thing than to allow that your philosophy might not have perfectly anticipated centuries of chaotic human change.

    Despite all that, we should not concede these historical points. First of all, because these hoary traditions like tax-loophole patents, software patents, and business process patents are generally younger than Britney Spears’s singing career. That has to be the beginning of any discussion of these things, especially when patent lawyers and trolls have the chutzpah to invoke the founding revolutionaries in reference to the mischief they have been making for a couple decades, and mainly in the last 8-10 years.

    * Epstein talks at length about open source licensing while betraying no knowledge of the subject, specifically the existence of the Apache, Mozilla/LGPL, and Berkeley Unix licenses, which account for enormous chunks of open source, and many of which are in fact in the public domain.

  • http://www.techliberation.com/ Tim Lee

    Timon, that’s an excellent point that I hadn’t noticed from listening to his talk. He did seem to be oblivious to the fact that there are non-copyleft open source licenses. Which isn’t a serious problem in and of itself, but does suggest that he’s theorizing about things he doesn’t understand very well.

  • http://www.techliberation.com/ Tim Lee

    Timon, that’s an excellent point that I hadn’t noticed from listening to his talk. He did seem to be oblivious to the fact that there are non-copyleft open source licenses. Which isn’t a serious problem in and of itself, but does suggest that he’s theorizing about things he doesn’t understand very well.

  • http://techdirt.com/ Mike Masnick

    Tim,

    One really really really minor point. You wrote:

    “The DMCA and software patents, of course, are two new laws that were introduced in the last decade or two over the loud objections of many of the people who would be most directly affected by them.”

    It’s worth pointing out that the change to include software patents wasn’t a “new law” but a new interpretation of the law thanks to CAFC. In fact, if you listen to some patent attorneys talk about it, they’ll claim that software has ALWAYS been patentable — it was just that people didn’t take advantage of it, which is a bit of revisionist history of course… but…

  • http://techdirt.com/ Mike Masnick

    Tim,

    One really really really minor point. You wrote:

    “The DMCA and software patents, of course, are two new laws that were introduced in the last decade or two over the loud objections of many of the people who would be most directly affected by them.”

    It’s worth pointing out that the change to include software patents wasn’t a “new law” but a new interpretation of the law thanks to CAFC. In fact, if you listen to some patent attorneys talk about it, they’ll claim that software has ALWAYS been patentable — it was just that people didn’t take advantage of it, which is a bit of revisionist history of course… but…

  • http://www.techliberation.com/ Tim Lee

    Mike: quite right. But I think Epstein’s policy point applies just as well: when the courts announce a dramatic re-interpretation of an existing statute–especially one that expands government power over the private sector–it deserves the same sort of skepticism as when Congress does the same thing.

  • http://www.techliberation.com/ Tim Lee

    Mike: quite right. But I think Epstein’s policy point applies just as well: when the courts announce a dramatic re-interpretation of an existing statute–especially one that expands government power over the private sector–it deserves the same sort of skepticism as when Congress does the same thing.

  • http://www.ikeelliott.typepad.com Ike Elliott

    Tim, this is an excellent post, and thanks for drawing attention to the topic. Thanks, also, for your comments on my blog regarding the question of whether software should be patentable.

    I have posted on the blog in response to your comments there, comparing software to music. Music is not patentable, of course, while software is patentable. Imagine if music were patentable, though. Would we have been deprived of many great musical works due to the temporary monopoly granted for certain musical forms? I fear that would have been the case, and it supports the idea that we should not have allowed software to be patentable.

    More on my blog at http://ikeelliott.typepad.com/telecosm/2007/12/licensing-the-b.html

  • http://www.ikeelliott.typepad.com Ike Elliott

    Tim, this is an excellent post, and thanks for drawing attention to the topic. Thanks, also, for your comments on my blog regarding the question of whether software should be patentable.

    I have posted on the blog in response to your comments there, comparing software to music. Music is not patentable, of course, while software is patentable. Imagine if music were patentable, though. Would we have been deprived of many great musical works due to the temporary monopoly granted for certain musical forms? I fear that would have been the case, and it supports the idea that we should not have allowed software to be patentable.

    More on my blog at http://ikeelliott.typepad.com/telecosm/2007/12/

Previous post:

Next post: