Tim Lee

Patents and Tacit Knowledge

I’m currently reading Virginia Postrel’s excellent The Future and Its Enemies. Chapter four gives an excellent exposition of tacit knowledge. It occurs to me that the insights of the chapter bear directly on patent policy:

As Polanyi suggested, much of our most important knowledge is tacit—difficult to articulate, even to ourselves. Contrary to Sale’s imaginings, such knowledge is expensive to share, assuming it can be transferred at all. It is “sticky,” in management scholar Eric von Hippel’s term: “costly to acquire, transfer, and use in a new locus.” Von Hippel notes, for instance, the difficulty of duplicating a scientific apparatus. Subtle information about the lab environment, or procedures that people at the original site take for granted, can make the difference between success and failure. “It’s very difficult to make a carbon copy,” say s a researchers quoted by von Hippel. “You can make a near one, but if it turns out that what’s critical is the way he glued his transducers, and he forgets to tell you that the technician always puts a copy of Physical Review on top of them for weight, well, it could make all the difference.

As a result of this stickiness, tacit knowledge often travels only through apprenticeship, the trial-and-error process of learning from a master. (A form of “apprenticeship” is essentially how as children we learn such complex basic skills as speech. Writing in the 1950s, Polanyi argued that the art of scientific research, as opposed to the scientific information that can be taught in a classroom, had still not passed much beyond the European centers where it had originated centuries earlier: without the opportunity offered to young scientists to serve an apprenticeship in Europe, and without the migration of European scientists to new countries, research centres overseas could hardly have made much headway.”

The application to patent debates should be pretty obvious. Some patent proponents blithely assume that you can copy an invention as easily as you can copy a song or a piece of paper. It’s pretty often, for example, to see an argument that without patent protection, a small software company wouldn’t be able to negotiate on an equal footing with a large one, because the large one will simply listen to the smaller company’s pitch, take careful notes, and then steal the company’s idea without paying a penny.

The problem with this story is that it completely ignores the role of tacit knowledge in duplicating technology. If it’s difficult to duplicate a scientific expermient when the technical details of that experiment are publicly available, how much more difficult is it to duplicate a new technology based on the fragmentary information you get from a technology demo? A company seeking to duplicate a competitor’s technology will typically be forced to go through virtually the same trial-and-error process the original company went through. Which means that in many cases, licensing the smaller company’s technology will be the faster and cheaper than trying to re-invent the wheel.

Obviously, the force of this argument will vary with the degree to which products embody tacit knowledge. For example, it seems like pharmaceutical products would be easier to copy than others because they can be characterized by their chemical formulas. Software seems to be at the opposite extreme—especially if copyright law prevents the verbatim copying of source code. There’s a tremendous amount of tacit knowledge embedded in any software product of non-trivial complexity, so the idea that software companies can duplicate their competitor’s products quickly and easily is unrealistic.

June 2, 2007 | Comments |

Viewing 8 Comments

    • ^
    • v
    Well, Tim, I gotta say this is a turn-around. One day, you write something I completely disagree with, and then you come back with something thats actually pretty good.

    Check out these articles:
    Lemley, Mark A. and Cohen, Julie E., "Patent Scope and Innovation in the Software Industry" . California Law Review, Vol. 89, p. 1, 2001.

    Burk, Dan L. and Lemley, Mark A., "Designing Optimal Software Patents" . INTELLECTUAL PROPERTY RIGHTS IN FRONTIER INDUSTRIES: SOFTWARE AND BIOTECHNOLOGY, Robert Hahn, ed., 2005.
    Lemley and others make several recommendations for patent policy, as it should govern software patents. Relevant items are a more stringent disclosure enforcement for software patent applications. Lemley did not call for lines of code being put into patent applications, but he makes strong arguments on how the USPTO shows little concern with the enablement doctrine due to lax enforcement of the disclosure doctrine. Another key argument by Lemley and others is that there should be a reverse engineering exception in patent doctrine, which is also tied to the USPTO's lax enforcement of the disclosure requirement.

    I have to find this article by Robert Merges, in which he argues that the scope of enforcement for a patent should correlate with the extent of its disclosure. Can't recall the article offhand, since Merges has released a number of articles recently.

    Having referred you to the argument above, I do take issue with this aspect of the post:
    It’s pretty often... to see an argument that without patent protection, a small software company wouldn’t be able to negotiate on an equal footing with a large one, because the large one will simply listen to the smaller company’s pitch, take careful notes, and then steal the company’s idea without paying a penny.
    You make two mistakes Tim: 1) confusing less leverage and no leverage, 2) not acknowing practices in the industry. For #1, lets try this: without patent protection, a small firm would have less leverage to negotiate with larger firms and would have to rely on trade secrets, which inherently do not invite collaboration and exchange between innovators. For #2, imagine that a small firm works with a larger one to commercialize an invention. The small firm provides a component part of an invention, the larger one wraps complementary technologies around it and manufactures. Without patent protection for the smaller firm, once the commercial invention is completed, the larger firm would not bother to pay licensing fees. They'd say, thanks sucka, but you don't have a patent on it. With patent protection, small firms hold leverage from the initial negotiation phase to the commercial phase.
    • ^
    • v
    Noel - for your #2 can't the smaller company sue the larger for breach of contract just as easily as for patent infringement ? (And it costs much lesser to get a good contract drafted, negotiated and signed than to get a patent awarded).
    • ^
    • v
    Chris, excellent point, but without a patent, how exactly do you define the smaller firm's contribution? There is little to prevent the larger firm from reworking the technology to the point where it can say its no longer using the smaller firm's contribution.
    • ^
    • v
    Noel, did you read my post? The whole point was that, thanks to local knowledge, the large company will be able to "rework the technology" much more cheaply with the small company's help than without. So the small company says "we won't help you adopt our technology unless you give us a contract agreeing to give us 20 percent of the revenues for 10 years (or whatever). If my point about tacit knowledge is true, the patent is completely superfluous.
    • ^
    • v
    Tim, again you use this logic where you only compare extremes: "If my point about tacit knowledge is true, the patent is completely superfluous." No, no, no...

    If a small company says: "we won't help you adopt our technology unless you give us a contract agreeing to give us X percent of the revenues for X years" the larger company would be constrained in how it commercializes technology, since it could adopt a better solution yet still owe the small company royalties. This is terrible for the innovation process in a sector where firms must adapt and be flexible. Such a scenario, as you suggest Tim, would also probably curb entry of new firms, who can't just license their patents but must essentially go rent-seeking with unproven technology.

    You make some good points in this post Tim- suggesting that licensing agreements can also be seen as collaboration agreements, but then you over-generalize, and arrive at a very bizarre scenario.
    • ^
    • v
    Good post, Tim. Tacit knowledge also plays a very large part in design. Have you read, Donald Schon's Theory in Practice: Increasing Professional Effectiveness or Displacement of Concepts? They'd both be relevant.

    I'd also suggest:

    http://enigmafoundry.wordpress.com/2006/12/17/t...
    • ^
    • v
    I have not read that. Thanks for the pointer!
    • ^
    • v
    This is also relevant to org theory. Tacit knowledge is another reason to be dubious of Mises' ideas on the "entrepreneurial corporation" and the division of labor between entrepreneurs and tech types. Mises argued that rational calculation in the choice of alternative technical means was impossible without recourse to entrepreneurial knowledge of the relative costs of alternative inputs. Although Mises didn't acknowledge it, the obverse is also true: it is impossible to make rational entrepreneurial decisions about which black box to channel money into without a knowledge of the production process. And if much of either technical or entrepreneurial knowledge is tacit, it may not be feasible to delegate them to different people. The smaller the organization, the less stovepiping of functions, and the more ownership and control are combined with direct involvement in production, the more rationally the organization will be run.

Trackbacks

blog comments powered by Disqus