Fragmentary thoughts about patents

I'm using this page as a collecting ground for fragments I'd previously left lying around on other pages that I haven't yet tidied up. It sort of keeps them out of the way …

Snippets and links

Early US car manufacture was initially hampered by the Seldon patent, until Henry Ford decided to fight it.

Some quotable snippets from John Paul Stevens' last opinion as a justice of the US supreme court contains some insightful remarks about patents, with excellent references:

Help keep it from getting worse

There has been some effort to extend patents to allow software patents (whence spring all manner of fascinating variations, via the skill of lawyers, including but not limited to process method patents, like the delightful one-click shopping patent – as if cookies hadn't been included in the HTTP spec precisely so as to enable a broad variety of even-before-it-all-got-going obvious ideas like having them carry around enough information to let one's visits to a page contain, for zero observed overhead, information about one's entire (at least) recent interactions with the site).

As the observant reader will have noticed while I descended into a rant (albeit parenthetically), I am not in favour of this idea. Not one bit. I would honestly like to see the software industry become a free market, with all the struggle that implies (and honest humans to deal with, without gratuitously complex motives to entangle their intentions; such being the best guarantee – of security of reward for effort – I can hope for; I'm willing to gamble on this being what folk become, when unencumbered by injustice, save those wholly devoid of virtue – who are, in my experience, rare). In response to various calls for public involvement in the decision-making process, both in Britain and in Europe, I have written what came to mind when I thought about the matter. Mayhap others could have said it better, or I could have done so if I gave it more time, but that's what I wrote, then.

The cauldron of my thinking upon the subject bubbles on, I need somewhere to set down my thoughts; so, on this page, I shall scrawl them, until I get organised and work my gabblings into some coherentish form.

It was never the object of patent laws to grant a monopoly for every trifling device, every shadow of a shade of an idea, which would naturally and spontaneously occur to any skilled mechanic or operator in the ordinary progress of manufactures. Such an indiscriminate creation of exclusive privileges tends rather to obstruct than to stimulate invention. It creates a class of speculative schemers who make it their business to watch the advancing wave of improvement, and gather its foam in the form of patented monopolies, which enable them to lay a heavy tax on the industry of the country, without contributing anything to the real advancement of the arts. It embarrasses the honest pursuit of business with fears and apprehensions of unknown liability lawsuits and vexatious accounting for profits made in good faith.

U.S. Supreme Court, Atlantic Works vs. Brady, 1882 [quoted in Software Patents Tangle the Web]

Imagine

OK, so there is a case for patents which looks like this: if someone comes up with an ingenious way of doing something economically valuable – to a large enough slice of the people in some sense, insofaras and lots of other careful choices of but we need better words for this caveats – it's entirely possible that they'll be able to make more money or achieve more of their goals, whatever those may be, by keeping their idea secret. As soon as anyone else knows how it's done, they'll be able to compete with the innovator far more effectively; by delaying that moment, the innovator gains many strategic advantages in their market. Now it is hard to deny that it'd be a very bad thing indeed if our innovator died leaving no-one able to carry on delivering, so taking with him to the grave, the benefits of the new idea. So, we'd better give our innovator a decisive incentive to tell everyone the idea, and wouldn't it just make sense to have that incentive be a grant of a monopoly – of control, at least – over exploitation of that idea; which amounts to the ability to either tax the uses of that innovation arbitrarily, or simply hold a monopoly.

Well, I may as well deal with the argument as it stands and ask what tests one should apply to an idea before granting a patent for it, ignoring the whole tale of whether a protectionist carte blanche is the right reward to offer for coming forward with an idea. Before granting any special privilege to someone as an incentive to tell other folk about a good idea, be sure you know what price the public must pay and how realistic it is that someone could (quite apart from being so mean-spirited as the hypothesis almost presumes) make any money at all out of their idea without, for instance, first giving some investors concrete guarantees that, in the unfortunate event of the inventor's death, some mutually trusted third party would be beholden to provide (at least) the investors, with the means to recover some reward for their investment.

More quotes, etc.

At the helpless attempt of the author to explain to a befriended computer scientist, why programmes would not be patentable as such, though programme-related theories with an additional technical effect would be, he reacted bluntly: "You are completely nuts!" This clear pronounciation caused the author to think quite a bit.

Dr. Swen Kiesewetter-Kbinger, patent examiner and ex-programmer.

… everything made by people can therefore be understood as technical depending on the spirit of the time. It would only need to be distinguished from the divine creation. About any other delimitation of the term technique is disputable, since by this a painter will understand something else than a physicist, a musician something else than a craftsman, a writer something else than an engineer and a computer scientist finally something else than a lawyer.

Dr. Swen Kiesewetter-Kbinger, patent examiner and ex-programmer.

In the USA, more than six million patents have been issued. Even allowing that a fair proportion of these have expired, how can it be realistic to expect anyone to be able to determine whether their seemingly new invention is, in fact, covered by one of them ? In the year 2000, the U.S.issued 254,329 patents: at that rate, they issue a million every four years and we may expect to see around five million live at any given time. Although patent law only gives rise to civil, not criminal, actions, the practical impossibility of keeping track of which, of the things I might do, are constrained by patents should cause the US to reject its own patent law, for essentially the same reason that a criminal law violates the US constitution: as the Supreme Court found (quoted on page 26 of Groklaw's coverage of U.S. v. Lori Drew) in United States v. Batchelder, 442 U.S. 114, 123 (1979)

[i]t is a fundamental tenet of due process that [n]o one may be required at peril of life, liberty or property to speculate as to the meaning of penal statutes. Lanzetta v. New Jersey, 306 U.S. 451, 453 (1939). A criminal statute is therefore invalid if it fails to give a person of ordinary intelligence fair notice that his contemplated conduct is forbidden. United States v. Harriss, 347 U.S. 612, 617 (1954).

Some quotes from Mayer's dissent in Re Bilski

I wrote at length on Groklaw about Mayer's dissent. In collecting quotes while reading his dissent, I separated them into a few categories for later use in what I wrote; when I was done writing, I found one category that did not fit into what I was attempting to say, so left them out. These do, however, strike me as worthy of repetition. Groklaw's article also provide a link to filing statistics for the US patent office, which strikes me as generally interesting.

[T]he primary purpose of our patent laws is not the creation of private fortunes for the owners of patents but is to promote the progress of science and useful arts. … Although business method patents may do much to enrich their owners, they do little to promote scientific research and technological innovation.

Patents should be granted to those inventions which would not be disclosed or devised but for the inducement of a patent. Id. at 11. Methods of doing business have existed since the earliest days of the Patent Act and have flourished even in the absence of patent protection.

[n]owhere in the substantial literature on innovation is there a statement that the United States economy suffers from a lack of innovation in methods of doing business. Raskind, supra at 92-93. Instead, the long history of U.S. business is one of innovation, emulation, and innovation again. It also is a history of remarkable creativity and success, all without business method patents until the past few years. … (While innovation in business methods is a good thing, it is likely that there would be the same level of innovation even without patents on [such methods].)

Business innovations, by their very nature, provide a competitive advantage and thus generate their own incentives. … The rapid growth of fast food restaurants, self-service gasoline stations, quick oil change facilities … automatic teller devices … and alternatives for long-distance telephone services casts real doubt about the need for the additional incentive of patent protection in the commercial realm.

Business method patents, unlike those granted for pharmaceuticals and other products, offer rewards that are grossly disproportionate to the costs of innovation. In contrast to technological endeavors, business innovations frequently involve little or no investment in research and development. Bilski, for example, likely spent only nominal sums to develop his hedging method. The reward he could reap if his application were allowed — exclusive rights over methods of managing risks in a wide array of commodity transactions — vastly exceeds any costs he might have incurred in devising his invention.

Fragment

Consider any industry in which competing providers of some goods or service come up with (at least moderately) novel and (a.l.m.) non-obvious improvements on their competitors' offerings. Each incorporates their own improvements: once they have inspected their competition, they may well know how to make the other improvements; but patents may restrain them from doing so. This, at least in the first instance, leads to customers being unable to obtain a version of the relevant goods or service in which all the improvements are incorporated.

The competing businesses can cross-licence one another's patents to overcome this. At least in an ideal world, each business would see how well its improvements worked with each of the others and seek to ally itself with the holders of relevant patents; and, at least stochastically, this should lead to the better improvements dominating the market. However, businesses holding stronger positions in the market enjoy a competitive advantage due to various inertial effects (e.g. any overhead a customer must endure to switch from one supplier to another, any exclusivity deals they get their allies to agree to); in the extreme, these effects have been know to be juggled so as to ensure that the benefit they bring out-weighs the advantages a better product enjoys.

In such a case, if the bigger business adopts the better improvements the consumer gets a good product: but, otherwise, the consumer loses out, albeit a business winning this way may well suffer some damage, in the process, which may impair its ability to repeat this stunt. Still and all, once the dust has settled, the players in the market hold relevant patents and are parties to cross-licensing deals with one another, regardless of who won and whether the better inventions pulled through or got stomped.

Meanwhile, the barrier to entry into the market has been raised. A new business, seeking to compete in such a climate, not only must raise the capital to set up its premises and equipment, hire suitable staff and so on (I'll call these the fundamental costs of entering the market), but it must also license the patents of the existing players. The existing players can see that the newcomer's sales are going to cost them market share; so will naturally want to license their patents to the newcomer for fees that will at least cover their estimates of these losses; and, having no good cause to do their competitors unnecessary favours, will likely charge more than this. No matter how cheap the fundamental costs, the licensing costs will be on the scale of the profits the existing players are making from being in the market; no matter how cheaply the new business would have been able to sell its goods or service had there been no licensing costs, it will be obliged to sell at a price comparable to that charged by the entrenched competitors simply to break even; but can only make profit on the scale of the variation in price among its competitors, while they (thanks to their cross-licensing) take the whole difference between primary cost and the market price.

Now, it is the intent of the patent system to enable those who take out patents to be able to make money out of their inventions, and this amply shows that it succeeds: but, in the process, the invisible hand of the free market has been disempowered. The way the free market is supposed to arrive at prices for goods and services is based on a balancing act between the price customers will pay and the price suppliers charge; if customers will not pay enough for a business to survive, businesses will leave the market, enabling the remaining businesses to charge enough to survive, safe in the knowledge that customers must either pay up or make do without; if businesses charge unreasonably much, customers will flock to (and potentially be willing to invest in) new businesses entering the market to undercut the unreasonable charges.

One half of this process remains intact, but the other has been disarmed; and, though one might accuse the relevant industry of being a cartel (it clearly needn't have become a monopoly), the businesses in it can honestly enough assert that they are competing. Indeed, when calculating the fair price an anti-trust suit would have to claim they far exceed, each business can claim the license fees their peers charge the new-comer (which each incumbent pays by foregoing, from competing incumbents, the license fees it charges the new-comer) as part of the cost (to them) of producing the product, so that their price will only exceed the fair price by a modest portion (if at all). They will naturally argue that these costs (in the provision of goods or services) are incurred in a separate line of business from (and so be considered separately from the income they make by) licensing their patents, even though the latter, in reality, pays for the former.

Now suppose someone else comes up with a further improvement: they have the usual choice of either letting the existing businesses use it (whether via a patent and licensing, by publishing it or otherwise) or entering the market as a competitor. The patent system is intended to reward inventors, and let us suppose the new improvement is patentable; let us now see whether an inventor can reap those rewards in a market with entrenched cross-licensing.

If the inventor choses to start up her own business, she will need to license at least some of the entrenched patents that cover features that all customers expect the goods or service to have. This is the same barrier to entry endured by any new competitor; but the newcomer with an invention can at least hope to overcome this by cross-licensing. Thus, however the inventor tries to make money out of her invention, she will need to license her improvement to at least some of the businesses already in the market; so let us (regardless of whether she also enters the market) see whether she will be rewarded for her invention.

Each business already in the market, once it hears about the new improvement, will want to be selling a version of the product including it: but will want to do so on terms which allow that business to take as large as possible a share of the increased revenue thus generated. In this matter, it is competing with the inventor and with anyone else to whom she licenses her patent. Each entrenched business will prefer to license the patent on terms which cut out their existing competitors; their ideal would be an exclusive license, giving them an effective monopoly in the new improvement and, thus, the means to make significant gains in market share – hence lots of money, so they can afford to make the inventor a significantly larger offer for such a deal than for a non-exclusive license.

possibly of vastly greater value to the consumer than all the first wave put together

More fragments

In so far as others would have come by the same invention, the alleged benefit is reduced: which the non-obvious requirement pretends to police. Sadly, assessing obviousness is, apart from spectacularly hard, only a partial test – an idea need not be obvious (in the eyes of some outsider) for its time to have come. Several folk may come by the same idea within a few years of one another; several research teams may have embarked on

Now, an invention may be arrived at by several inventors independently: in a monopolistic system, at least all but one inventor is deprived of any benefit from the invention. This is not just to say that it fails to reward all other inventors: it actively deprives them of the rewards they would naturally have obtained by putting the invention to use.

Further, the one beneficiary need not even be an inventor, since the actual inventor may have been coerced into signing away all rights in the invention – for example, many contracts of employment (and, in some jurisdictions, the default implied contract of employment, in so far as not over-ridden by an actual contract) grant the employer all rights in any invention made during the period of employment; this might be just were it limited to inventions pertaining to the employer's business, or made while on their premises or acting on their instructions; but if, while taking a holiday at my own expense to attend a conference of theoretical physicists, I invent a star-ship drive, someone who employs me to write software will be the beneficiary and any benefit I may gain by it will be entirely at their discretion.

A corollary of the overt cost is that – if several patents hold authority over a given invention, e.g. where the invention pulls together several ideas from disparate fields – even the patent holder may be beholden to many patent holders to exercise the patent. One may fairly argue that pulling together ideas from many fields is an activity particularly worty of encouragement, at least in so far as it has a track record of yielding up valuable ideas. Yet in this case the possibility that the inventor is unable to reap their natural benefit is most likely to be realised. Furthermore, any one of the holders of relevant patents is able to obstruct anyone wishing to employ the composite: the benefits they reap are in no way constrained to be proportionate the benefit their contribution delivered. As we live in an era of rapid innovation, this arises frequently: the patent system creates not a monopoly but a cartel.

Inventions are often incremental – the patent on an invention which yields no discernible benefit to the common weal, on its own, may none the less embrace a foundation of some later invention which yields great benefits. In such a case, those who employ the latter are bound to pay not only for their use of it but also for their incidental use of the former. There is no restraint on patents which ensures that the holder of the patent on the less useful invention will benefit less than the holder of the more useful one; indeed, the holder

Finally, note that we inhabit a world giddy with the pace of innovation: even the comparatively modest incentive inventors obtain, from being able to exercise their inventions, may well be sufficient to maintain a fast enough pace of innovation that any increase is of at best marginal benefit and quite possibly, by increasing the giddiness, an actual cost in itself. We no longer inhabit the (comparatively) technologically stagnant renaissance.

Globalisation

Within my lifetime, the great powers of the world have increasingly lined up behind a global agenda under the banner of free trade, which they have consistently adulterated with the various monopolistic provisions of so-called intellectual property law. While there is a case even against the free trade portion of this agenda (the case rests on the issue of huge disparities and inertia; developing nations cannot really compete on an equal footing, and some modicum of protectionism might serve to help them get up to speed and able to survive genuinely free competition), I confess it generally enjoys my support (though I would have no argument against less developed nations insisting on a progressive approach to free trade, as opposed to instantaneous transition thereto). I do believe that free trade is generally good.

However, the patent system in particular (along with monopolies, including intellectual property, in general) actively undermines the good free trade might otherwise be doing for the developing world. A nation may have all the natural resources to participate in some high-profit industry; somehow find the funds to pay for some of that nation's folk to be trained to do the work that industry requires; and yet be unable to reap those high profits because they have agreed to participate in the global patents régime, which grants a monopoly, in some necessary technology, to someone else.

New Scientist (2002/January/12th, issue No. 2325, p43, reviewing Protect or Plunder ? Understanding Intellectual Property Rights, Zed books, by Vandana Shiva) reports:

… a botanical cure for hepatitis traditionally used in India can be patented in the US. The American patent holder may then attempt to stop people using their own remedies. There's another twist: a survey in the US showed that 80 per cent of patents are taken out to block competitors. None of this helps the promised technology transfer from North to South, …

On the authority of governments

Now it cannot be just for government to grant, to others, powers beyond what government itself has any right to exercise. The U.S.A.'s constitution, for instance, forbids government to impede the citizenry's right to free speach: as such, government has not the authority to grant others a privilege which would enable them to impede that right – or any other protected in like manner by the constitution. The grant of a patent provides its holder with a (marketable) authority to, at whim, tax and dictate laws to those who would employ any invention which, even in part, is covered by the patent.

Of course, if they exercise their privilege too starkly, they may suppress all exercise of affected inventions: hence depriving themselves of any revenues from those inventions' deployers. Yet this may, indeed, prove to their advantage – consider two competing businesses, one of which owns the patent on a technology obsoleted, but built on, by the other: the former can deprive the latter of any revenue from the more valuable invention, under the guise of protecting their own reward for the obsolete invention. The temporary nature of patents would ideally prevent this from mattering: however, in an era of rapid (and often incremental) innovation, such overlaps are by no means scarce. Thus, while the dynamics of markets will affect each patent-holder's exercise of their privilege, this need not restrain the patent-holder's authority.

The parable of the carpenter

If a carpenter makes a chair which is indistinguishable from one I own, the question of whether he has seen my chair or not is immaterial – he owns the chair he made – but if I have patented some innovative design feature in my chair I can, none the less (and still regardless of whether he has seen my chair) sue him for the theft of my property.

He can only escape this bind by over-turning my patent. To do that he must show that the idea was one that an unexceptional carpenter could reasonably have been expected to come up with entirely naturally; which he can, in practice, only do by pointing to such carpenters who have done so. Yet it is practically impossible for him to prove that he, or any other carpenter, has neither seen my chair nor read my patent, nor heard my idea described by someone who has; so I shall be able to repudiate any number of example carpenters by claiming they stole my idea, they did not come up with the idea for themselves entirely naturally, that is just the lie they would naturally tell and have an unmistakable (and undeniable) vested interest in telling. Furthermore, to make this case, he needs carpenters who are prepared to help him prove in court that they are unexceptional. Just what their reputations need and their egos are begging for – plus they also have to admit in court to making chairs which infringe my patent, thereby ensuring I know who to sue next. They need to be sure I'll lose before they'll agree to that.

Not that I actually have to worry about that – I shall bring my charge against the first carpenter I discover infringing my patent, so he shall only have himself as an example of someone else who came up with the idea and I shall be able to charge that he is not such an unexceptional carpenter in any case – either he stole my idea or he must have been a pretty exceptional carpenter to have come up with my oh so clever idea idependently, after all, it's not like any other carpenters have done so. As long as I thus prevent my patent being over-turned, we're back to it not mattering whether he came up with the idea independently or by stealing it, all that matters is that his chair infringes my patent. Once I've successfully defended my patent, no-one else is going to have a chance of over-turning it, so I'll be safely assured of royalties from anyone who wants to make chairs incorporating my idea.

At the heart of this is the problem that an idea can spread and its spread is impossible to trace, hence to disprove. Over-turning a patent thus suffers from an impossible burden of proof – how on earth does one prove that an idea is obvious ? Indeed, once exposed to an idea, how does one determine whether it is obvious ?

Im/Materiality

The primary problem with ideas as property is their immateriality: two folk can produce what both will acknowledge is the same idea; whereas two folk making chairs, even be these indistinguishable, will declare them distinct chairs; I cannot tell two sacks of wheat apart, yet the two farmers who grew them are not at all inconvenienced by this, whereas the corresponding coincidence between two inventors' ideas is a dire complication.

No amount of similarity between my chair and yours will suffice to prove you stole mine, if I still have mine. You cannot come by my chair while honestly believing that you made it yourself after it came into my posession. If you acquire a chair and mine goes missing, I might accuse you of stealing mine: but if you can find witnesses who saw mine before it went missing and attest that yours is distinguishable from it, I have no case (likewise if you can find witnesses who will attest that you were in posession of yours before mine went missing). Even if mine has gone missing and yours is indistinguishable from mine, if a witness comes forward who saw you make your chair from its raw materials, my case evaporates. All of this owes to the materiality of chairs.

Because ideas are immaterial, you can come by my idea without my having it any the less; without, even, my being able to detect that you've come by it, or your being aware that it existed before you had it; even if you have it from me, what you get may appear drastically different from what I have; even if you come on it for yourself, what you get may be indistinguishable from what I have. No witness can, seeing your researches, attest with certainty that you were not simply acting so as to appear to come up, independently, with an idea you actually got from me.

Thus, if one is to make property of an idea, either one practices the injustice of depriving folk of the natural fruits (which would have existed without the law) of their labours – i.e. the use of ideas they had independently (on grounds that someone else filed first) – or one has to grant the idea to all who claim to have come by it independently and can't be shown to be lying – at which point there is little sense in which it is property. Either way, ideas as property are qualitatively so different from material property as to annihilate any sense in which what is just or sensible for one may reasonably be expected to be so for the other. Any attempt to describe ownership of an idea by analogy with ownership of a chair or plot of land is bound to mis-lead; calling it property encourage the folly of using inappropriate analogies.

Further, the im/material distinction is crucial to the origin of property: absent the state, my viking forebear could – and would – defend his property against whoever came to take it from him. Any institutions which failed to recognise this were doomed to be for ever at war with the whole citizenry. As such, the phenomenon of material property exists regardless of laws and governments. Now consider an idea: if my viking forebear invented a better chair, he could not even know whether someone else was making similar chairs elsewhere, nor is there anything intrinsic to his idea that empowers him to guard it as he does his material property.

Similarly, if we trade (purely) material property, each goes away wholly posessed of what the other gave him and disposessed of what he traded for it. If we trade purely immaterial property, even though you go away with my patent and I with yours, each retains his understanding of the idea he patented. If we trade material property which is an expression of some immaterial property, however, without trading the immaterial aspects, we put into the world some goods my viking forebear could not have understood. When his chair broke, he made another like it: do not tell him he thereby wronged the owner of the patent on some clever design feature he thus copied !

In effect, the system of patents does this:

While the intent may be that this power is proportionate to the extent to which the invention is used, in practice it has more to do with how much power the patent-holder wields – which is directly correlated with how little the patent-holder needed any further reward than arises naturally from having made the discovery. Thus those who most need the allegedly intended help are least apt to benefit by it; and the monopoly power overtly bestowed can only be fully realized by those who posess such great power already that we should sooner concern ourselves with the threat to our liberty posed by that power – let alone any expansion of it – than with any worry that they may lack an incentive to innovate.

On the significance of artificial intelligence

Governments are throwing money at Artificial Intelligence (AI) research, the ostensible function of which is to ultimately produce computer systems capable of doing all the mental things people do. If that dream is ever attained, it will be necessary to accept that things produced by computer programs may contain the necessary creative step that separates automated processing from creativity. Given that one branch of government (the executive) presumably considers that goal attainable, it would be perverse for another (the judiciary) to deny that the action of a program may constitute such a creative step.

Ultimately, even before we reach full AI, we can expect to have programs which do make contributions to fields of art and science that, had they been done by a human, would be entitled to the monopolistic privileges of copyright and patent. Arguably this has been done; some of the researchers in neural networks and evolutionary programming have successfully produced optimal and novel solutions to real-world problems (e.g. sorting a list of numbers).

Such software is apt to re-invent something that's already been patented: in such a case, presumably, the law will hold that the patent has been infringed; in principal, the person operating the machine on which the software ran will be liable for this infringement. Yet they shall have had no practical way to know even that the program was trying to solve the problem whose solution the patent covered – let alone that a (likely utterly unintelligible) patent covered it.

One might hope to turn this into a cure for the problem of obvious patents; if a widely-used automated problem-solver comes up with the same solution to a problem, it can be counted as obvious. Over time, this might even reduce the number of patents to a non-insane level. However, it'll definitely require some rethinking of the law's stance on automated processing (e.g. of copyrightable works) and creativity.

However, that would only be reasonable if applied to an isolated problem solver; one connected to the net might appear to re-invent an idea when actually it found it via the internet. Yet it would likely be very useful to have automatic processes that can connect to the internet and take inspiration from what they find there; indeed, if we're to have much chance of coping with the era of information overload we'll need such processes to serve as our agents. When these independently re-invent a patented idea – or indeed borrow it from a source which neglected to mention that it is covered by a patent – we are back to the same problem.

Effectively, patents (particularly on software, but ultimately the same argument will apply to anything that can be made by a machine controlled by software – i.e. pretty much anything patentable) are incompatible with the deployment of artificial intelligence if it gets far enough to be useful; and, if it never does, we're wasting a lot of tax dollars funding futile research.


Valid CSSValid HTML 4.01 Written by Eddy.