Patent law’s deep questions and the government’s answers
on Mar 28, 2014 at 3:09 pm
Many lawyers come to the law for its deep philosophical questions—questions of morality, jurisprudence, and the public good. These people usually think that patent law is exclusively for dorks, and that they should avoid it like the plague (in its generic, metaphorical sense—they aren’t really interested in the microbiology). This is ironic, because patent law has a tendency to pose some of the hardest and most interesting philosophical questions the law has to offer: What does it mean to invent a composition of matter or a new thing? Where do we draw the line between the laws of nature and the application of those laws in some product of human ingenuity? Think a computer is obviously a new and patentable machine? (That’s got to be right.) Well how will you deal with the fact that every piece of new software arguably makes it new again, capable of doing something no other machine has done before? That thing the computer is doing may be very close to a patent-ineligible abstract idea or law of nature. Should the software nonetheless be patent-eligible? Always, often, sometimes, never? This year’s intellectual property blockbuster, Alice Corp. v. CLS Bank International, poses some of these questions.
But while these are deep and interesting questions that lawyers should love to puzzle over, like many such ideas in science, they might be dangerous. Intellectual property is hugely important to a number of the Nation’s biggest companies, is a massive economic driver in today’s economy, and poses a critical public policy issue about the balance between encouraging innovation and allowing for competition. There is an abiding concern that questions about patent-eligibility posed at a high degree of generality represent a threat of unforeseen consequences to the very practical world of patent practice and product innovation.
The government’s position—set forth in its brief in Alice—is a study in the balancing act that the patent office and the courts often confront. The government seems to view its take on software patent eligibility as a fairly narrow, middle-ground approach that scales down the deeply philosophical questions inevitably raised in cases like this. For the government, this case is not about software patents so much as it is about business-method patents—which the government loosely defines throughout its brief as patents on “methods of organizing transactions and other human activities” or relationships. These, the government says, are not eligible for patents because they are abstract ideas, not concrete inventions. And furthermore, when you have such an abstract idea, you cannot make it concrete just by telling someone to do it on a computer. Some software should be patent eligible, but at least not that.
This sets up a critical distinction for the government between two kinds of software inventions: those that are patent eligible because they “disclose technological, scientific, or industrial innovations” and otherwise “improve the way computers function” (yay!); versus those that “are directed to organizing abstract concepts and relationships using a computer” (boo!). Put otherwise, the government thinks you can get a patent on an idea for making computers more awesome, but you cannot necessarily patent your idea for making the world more awesome just because it uses a computer.
The government’s position is that this distinction tracks existing skepticism about business-method patents while still preserving a realm of “technical” innovation in computing that might seem intuitively patent-eligible. The government even proposes a two-step approach where one first asks whether the core invention is a patent-ineligible abstract idea, and only considers the software application issues if that is so. Accordingly, even some abstract ideas may be patent eligible as implemented through computer software, and conversely, software never loses its patent eligibility unless it at least embodies an abstract idea. The hope seems to be that this position will preserve the ability of the patent office to work out a distinction between software that meaningfully moves forward the field of computing and software that merely takes a business method and turns it into ones and zeroes.
But without criticizing the government’s position at all—and indeed, while avoiding any comment on the merits of the case before the Supreme Court or what the Justices might do with it—it is worth saying that the government’s approach may turn out to be something of a greased-up watermelon: a thing that looks like it has real heft, but is almost impossible to firmly grasp. Do programs that allow you to ask your computer natural language questions and get accurate answers (you may have met him or her) represent what the government calls “improvements in a computer’s operation as a computer,” or are they just computer-implemented embodiments of the world’s most common human interaction? How about a computer that plays chess—a machine we have pursued since the famous hoax of the Mechanical Turk? This all seems to trade on some deep, ontological account of what it means for a computer to function “as a computer,” which is just the kind of philosophical imponderable that makes these cases so hard, but also so much fun.
The reason this isn’t a criticism is that, as many scholars and judges have pointed out, every proposal for pinning down the distinction between patent-ineligible abstract ideas and patent-eligible innovations seems to slip out of your hands when you try to pick it up and work with it; the distinction between “technical” software innovations and software that simply implements an abstract idea may be the same. The government has a “non-exhaustive list” of at least six factors you might consider in solving such puzzlers, but they do not seem easy to apply in any kind of mechanical way, and may be frustrating to the kinds of lawyers who like to respond to legal questions with answers rather than a new set of considerations.
For similar reasons, the government’s view of software patent eligibility may be more innocuous in name than it is in practice from the standpoint of existing patents (or not, or vice versa). A lot would come to turn on a distinction that seems difficult to rigorously define ex ante. Plain-English categorizations of alleged software inventions do not obviously fit into one or another of the proposed boxes: For example, does a program that gets a computer to do something it has never done before make it function better “as a computer?” (And while you are thinking about that, consider that your own intuitions about what makes a computer “a computer” might be largely a product of the software through which you have experienced its computational power.) It is thus hard to know the exact zigs and zags even the government’s proposed course would cut through the existing software patent portfolios of the many companies watching this case with interest. Perhaps the best that can be predicted is that, whatever comes of Alice, the patent office and the courts will end up drawing lines in practice with a lot of reflective equilibrium, much as they do now.
