Prof. Robert P. Merges, University of California at Berkeley School of Law
September 18, 2013

Two news items caught my eye in recent days, and I have been thinking about how they connect.  Ray Dolby, the great sound engineer and entrepreneur, died.1  And I saw the latest in a string of ominous-sounding reports on the building wave of patent litigation.  To the world at large, I do not suppose the two stories have much in common.  But to someone like me, a student of all things related to patents and inventions, there was something about the confluence of these items that hit the mark.

The company that carried the Dolby name, Dolby Labs, was a byword in the world of innovation.  Recording engineers, consumer electronics companies, movie theatre operators, and eventually consumers themselves – they all associated Dolby Labs with high-quality sound systems.  Yet in the terms that frame the contemporary debate over patents and patent litigation, at least parts of Dolby Labs’ operation might be characterized by some as a patent troll.  They rely on technology licensing rather than actual product sales to make money.2  It is precisely the sort of organization featured in the recent report on patent litigation, which tells us that 66 percent of patent lawsuits are now being initiated by trolls (or “patent assertion entities” [PAEs]).3  To some, who would defend current trends in patent litigation, and wrap themselves in the positive aura of Dolby Labs, equating Dolby Labs with the trolls shows definitively the poverty of the troll label.  They would have us accept this syllogism: Dolby Labs was innovative; Dolby Labs is a “patent troll”: therefore the “patent troll” label is wrong-headed and useless.

Is Dolby Labs a Troll?

My simple point here is to argue – with as much vigor as I can – the falsity of this syllogism.  For my money, Dolby Labs is not a patent troll.  But at the same time, I think the troll label is valuable and useful.  It is shorthand for a set of practices that have grown up around patents; that rely heavily on litigation or its threat; and that undermine the purpose of the patent system as enshrined at the founding in our Constitution: promoting real (valuable) innovation.  There really are trolls, in other words; Dolby Labs is not one of them.

To place Dolby outside the troll category is to reject a simplistic approach to constructing the category.  The most common simplistic approach is this: to define a troll as any patent owner that does not manufacture products.  There are two camps that would advocate for this.  The first might be called the virulent anti-trolls, many of whom are tied closely to an era when patents were typically held by manufacturing companies.  While there is something to be said – maybe a lot to be said – about the connection among patents, the manufacturing base, and economic prosperity,4 the criticisms of this camp are just not selective enough.  They would sweep Dolby Labs into the disfavored category of troll.  And that would both dishonor what Ray Dolby achieved, and disfavor future companies trying to establish themselves as pure “idea factories,” dedicated to marketing technologies (covered by patents) rather than physical products.

The other camp pushing for a simplistic definition of trolls is, paradoxically, trolls themselves.  Their thinking is simple: if they can help establish a definition of the troll category that is essentially absurd, they will bring the category itself into disrepute.  Define trolls to include obvious innovators and you discredit the very idea of the troll company.  Define a supposed vice to include practices of undoubted virtue and you eliminate that definition from serious consideration.

It’s a clever strategy, but it should not be allowed to work.  There is too much at stake.  To simply accept on faith that all issued patents are presumptively valid, therefore represent real innovation, and that therefore patent litigation in any volume is acceptable is to capitulate to the forces of troll-dom.  This would be a big mistake, because out-of-control patent litigation is a real drain on the economy.  Excessive patent litigation demonstrates that the patent system has become decoupled from its true and original purpose – fostering real innovation.  The question is how to correct course, without crushing the life work of Ray Dolby and other innovators who use patents as the foundation for productive “idea factories.”

Patrolling the Troll Line: Two Sets of Tools

I’ve said some things before about the importance of patrolling the troll line.  Just above I was defending the legitimacy of this line.  Here I take a crack at how the line should be drawn.  Conceptually this is easy: Reward only real innovations.  Do not allow patents to become rent-seeking instruments that permit their owners to recover economic benefits created entirely by other people and other companies.

The trick is and always has been how to draw this line.  A number of people argue, implicitly at any rate, that the cost of drawing the line is not worth whatever benefits it would bring.  The central idea is that the market for patents represents the spearhead of an advancing “information economy,” and any attempt to discredit trolls moves us in a backward direction.  At heart, this criticism of the “troll line” seeks to contrast contemporary patents, patent litigation, and the market for patents with a now-superseded 19th-century view of patents as an adjunct to manufacturing.  A less-sweeping condemnation of the troll line idea concedes that perhaps some aspects of litigation-driven patent practices are not beneficial, but that drawing the line between good and bad uses of patents is too difficult and too costly to be worthwhile.

I disagree with both versions of the troll-line critique.  The first boils down to the simplistic notion that if patents help create a market, often with court enforcement forming the backdrop to consensual transactions, then there is no need to worry.  Voluntary exchange, based on enforceable property rights, is always good.  A moment’s reflection shows how wrong this is; society often chooses to shut down or at least not support consensual markets because they produce no net social benefits.  The classic case is rules prohibiting blackmail, but there are many other examples as well.

What about the second argument – that the troll line is too difficult to draw with precision, that policing it would eat up more resources than society gains by having it in the first place?  This I find wrong as well.  And my reasoning here leads me right to one crucial set of tools for policing the troll line.  To me, the existence of numerous and rigorous requirements for patentability demonstrates convincingly that society does think it worthwhile to weed out economically useful inventions from those that are of little or no value.  The first set of tools we can and should use to draw a firm line, in other words, is ready to hand: They are the conventional tools of patentability, laid out in Sections 102, 103, and 112 of the Patent Act.  For various reasons, I would lay particular emphasis on two of these: nonobviousness and the disclosure requirements in Section 112.

It is hardly revolutionary to say that all patents ought to meet the requirements of the statute.  In fact, it sounds very close to the troll argument that all we need to do is let the Patent Office do its job, and then allow full enforcement of all duly issued patents.  But I am saying something a bit different.  I want the various actors in the patent system – particularly judges presiding over patent litigation – to reorient their approach to patentability.  These judges need to understand that patent standards reflect crucial policy choices.  These technical-sounding requirements embody and give meaning to the deep underlying purpose of the patent system: to weed out non-innovative patents, so that only true innovations are rewarded.  Arguably, the Supreme Court’s KSR opinion5 pushed the courts in this direction, at least with respect to nonobviousness.  We need to encourage the post-KSR trend toward toughened patent standards.  And we need judges to extend it, by toughening disclosure standards under Section 112.  This could be particularly helpful in cutting back on the number of flabby, overbroad, and overreaching software patents that are often the basis for troll litigation.  Requiring real disclosure of operating software, and hemming in software claims, would go a long way toward weeding out the weak and overbroad patents that abound in the software field.

In an ideal world, that would be enough.  Consistent enforcement of tough patent standards would separate the innovative wheat from the rent-seeking chaff.  But given the imbalance between patent applicants and examiners, and the inevitability of administrative caprice and mistake, we can assume that even with toughened standards some bad patents will still issue.  That is where the second set of tools comes into play.  Various proposals aimed specifically at litigation abuse have been proposed – rightly, in my view.  A modified British Rule (loser pays) for lawyers’ fees may make sense.6  Rules requiring trolls to identify themselves clearly, rather than hide behind corporate shells, and even to post a bond from which lawyers’ fees may be later paid, could make sense too.7  And even more aggressive remedies, such as the development of a doctrine of predatory patent litigation, may become necessary to stem the tide of patent litigation.8


Ray Dolby built a successful licensing empire by developing innovative sound technologies.  His business model was based on the patent system.  In order to preserve that system, we need to take aggressive action against excessive patent litigation.  It is difficult to say in advance what is the optimal rate and mix of patent litigation.  But that should not scare us away from the conclusion that litigation-based troll models, dependent on mistake patents, should not be encouraged or even tolerated.  The simple policy that should animate our efforts is this: Patents should be about real innovation.  Bending our collective will toward this policy is the best guarantee we have that the patent system will be in good shape, ready for the future Ray Dolbys, and eager to encourage the ideas that they will hatch.