What is an Exceptional Case such that Court Should Award Attorney Fees?

By Dennis Crouch

Earlier this week, I commented on the two attorney-fees cases pending before the US Supreme Court:

  • Octane Fitness, LLC v. ICON Health & Fitness, Inc., Docket No. 12-1184; and
  • Highmark Inc. v. Allcare Health Management Sys., Docket No. 121163.

Both cases stem from the same perspective that the “exceptional case” requirement for attorney fees is too narrowly construed by the courts and especially by the Federal Circuit. See 35 U.S.C. § 285.

I noted in the prior post that the predecessor of Section 285 was first enacted in a 1946 Patent bill. The 1946 law was substantially similar to the current law (enacted in 1952) with the major difference that the 1946 act expressly gave the court “discretion” to award attorney fees to the prevailing party while the 1952 act removed that “discretion” term and instead indicated that the fee may be awarded “in exceptional cases.”

The Senate Report associated with the 1946 statute indicates that the statute is not intended to make fee awards an “ordinary thing in patent suits” but instead to reserve such awards for “gross injustice.”

It is not contemplated that the recovery of attorney’s fees will become an ordinary thing in patent suits, but the discretion given the court in this respect, in addition to the present discretion to award triple damages, will discourage infringement of a patent by anyone thinking that all he would be required to pay if he loses the suit would be a royalty. The provision is also made general so as to enable the court to prevent a gross injustice to an alleged infringer.

S. Rep. No. 1503, 79th Cong., 2d Sess. (1946), reprinted in 1946 U.S. Code Cong. Serv. 1386, 1387.

When the 1952 act was passed, the House Committee Report briefly mentioned the “exceptional case” amendment to the statute – indicating that the phrase “‘in exceptional cases’ has been added as expressing the intention of the present statute as shown by its legislative history and as interpreted by the courts.”

In several cases, the Federal Circuit has considered the standard and, after some debate, decided in Eltech Systems Corp. v. PPG Industries, Inc., 903 F.2d 805 (Fed. Cir. 1990), that the exceptional case standard should be the same for both prevailing plaintiffs and for prevailing defendants – writing: “we now reach the question and determine that there is and should be no difference in the standards applicable to patentees and infringers who engage in bad faith litigation.” The Eltech Systems case was then cited positively by the Supreme Court in Fogerty v. Fantasy, 510 U.S. 517 (1994) as justification for a party-neutral approach to fee shifting in copyright cases.

We note that the federal fee-shifting statutes in the patent and trademark fields, which are more closely related to that of copyright, support a party-neutral approach. Those statutes contain language similar to that of [17 U.S.C.] § 505, with the added proviso that fees are only to be awarded in “exceptional cases.” 35 U.S.C. § 285 (patent) (“The court in exceptional cases may award reasonable attorney fees to the prevailing party”); 15 U.S.C. § 1117 (trademark) (same). Consistent with the party-neutral language, courts have generally awarded attorney’s fees in an evenhanded manner based on the same criteria. For patent, see, e.g., Eltech Systems Corp. v. PPG Industries, Inc., 903 F.2d 805, 811 (Fed. Cir. 1990) (“[T]here is and should be no difference in the standards applicable to patentees and infringers who engage in bad faith litigation”). For trademark, see, e.g., Motown Productions, Inc. v. Cacomm, Inc., 849 F.2d 781, 786 (2nd Cir. 1988) (exceptional circumstances include cases in which losing party prosecuted or defended action in bad faith); but see Scotch Whisky Assn. v. Majestic Distilling Co., 958 F.2d 594 (4th Cir.) (finding in the legislative history that prevailing defendants are to be treated more favorably than prevailing plaintiffs), cert. denied, 506 U.S. 862 (1992).

Fogerty at note 12.

Of course, the problem with a party-neutral approach is that it can only be applied at a certain high level of granularity or abstractness. Since patentees and accused infringers make systematically different argument and have systematically different strategies, the particular causes of exceptional case findings tends to be different. Thus, losing infringers most often pay fees based upon ongoing willful and reckless infringement while losing patentees most often pay fees for suing on patent obtained through inequitable conduct or for bringing baseless lawsuits. At that low level of granularity, the differences are such that it is difficult to compare whether a party-neutral approach is being applied.


Thanks to John Pinkerton at Thompson & Knight for providing me with this legislative history from the 1946 Act.

38 thoughts on “What is an Exceptional Case such that Court Should Award Attorney Fees?

  1. 2

    Exceptional:1. Willful infringement;2. Asserting an invalidity defense based on art that is the same or cumulative of the art before the PTO, i.e, where there would be no substantial new question of patentability. Having to litigate that again in court is vexatious.3. Filing a DJ action that a patent claim is invalid without an opinion of counsel that there is a substantial new question of patentability.3. Filing an infringement lawsuit without obtaining an opinion of counsel that there is infringement or that information makes infringement probable.4. Filing a DJ action of non infringement where one does not have a prior opinion of counsel that there is no infringement.4. Filing a lawsuit where one knows an asserted claim is invalid.Regarding so called i.e., I strongly disagree that a case should be deemed exceptional unless the claim involved with the fraud is found to be invalid. If the claim is valid, there was no harm done to the public.Litigation misconduct.Calling the patent owner a troll in front of the jury ala Microsoft in i4i.

    1. 2.1

      Funny, Ned, how you can be fairly reasonable when you don’t have an axe to grind. Do you ever consider that by advocating not applying the law or understanding information processing patents that you denigrate our entire system? That your behavior negates anything positive that you could ever do?

      1. 2.1.1

        Night, We agree on so many things up to a point. The point is where the law is on patenting mathematics… You call it information processing, but it is the same thing.We all know that math is useful. But the Supreme Court requires more than just the math. I requires an application. Now, it really doesn’t take much to leap that barrier from the abstract to the concrete. Just talk about signals. That just might be sufficient. But we have no clear authority on point.


          Ned, a machine is sitting there performing useful work. How can it not be patentable based on how you decide to characterize it?


                102 and 103 is fine. 101 is not.Oh, so it’s all about adhering to some super strict interpretation of 101 whereby any claim that recites a “machine” is eligible, period, no matter how that “machine” is distinguished from prior art machines. But when we get to 102 and 103, you tell us that you have no problem finding all these functionally claimed computers obvious. Isn’t that what you meant by “fine”? If not, then, please let us know what you meant and explain to us how we determine what makes a new functionality on an old computer “non-obvious.” What exactly is “non-obvious” about a specific type of information being stored or transmitted by a computer that renders the computer patentable over the prior art? Is it like a non-obvious book being carried in the back of an old car that makes that car patentable over the prior art? Or is there some profound difference? You know, cars are pretty popular and if you don’t think that they should be patentable just by reciting some new functionality without describing any new structure then maybe you want to just live in the stone age. Is that the 103 argument you want to make? Or will it be that you can show that a lot of people bought the non-obvious book and put it in their car?When you make these statements (102 and 103 is fine, 101 is not”) I’m reminded of the patent txxbxgger arguments made in the run-up to Prometheus. We were all told that those claims just had to be eligible under 101 and that any “issues” could be handled by 102/103. But when pressed about the analysis under 102/103, suddenly those cheerleaders for a toothless 101 became very very quiet. The reason they became quiet, of course, is that the only way those analyses succeed in invalidating patents is to apply the same reasoning used to tank the claims under 101 (i.e., the claims effectively protect the ineligible subject matter).That’s exactly what’s happening with the typical computer-implemented j*nk. Computers exist to process information. They are proxies for the human brain. Computers are also old. By allowing old computers to become “new” for patent purposes merely be reciting information processing (or storage or transmission) functionality, then we are effectively granting patents on those functions. “Functionality” is not eligible subject matter, however.Therein lies the rub, friends. I can hardly wait for the Supreme Court to dive in. You think they are going to buy Randy “I can touch the computer so it’s eligible” Rader’s arguments? Or the arguments about the “essence of electronic structure”? I doubt that. And that’s doubly true when (1) the shriliest and shriekiest defenders of computer-implemented j*nk are invariably the least persuasive and credible; and (2) many many “industry” participants agree that such claims promote nothing except waste and the lining of attorney/grifter pockets.


              How very insightful–not. The machines are projected to replace 50% of the current jobs in the US, so I think the burden should be on you to tell me how the machines are not useful.And, again, this bizarre and disturbing notion that somehow you can characterize an invention in some way to invalidate the invention (which is a useful machine) is simply against every principle of our laws.

                1. So what? So a machine that replaces a person isn’t eligible for patentability? How is that possible Ned? How!!!!!!!!!!Reality refutes your 101 arguments.

                2. But the computer was invented in WWII. Talking about the utility of a computer is not a convincing argument.

                3. Clearly Ned, you are engaging in the vacuous ‘House’ argument and you are forgetting the simple words that Congress chose in 35 USC 101:”or any new and useful improvement thereof”The computer that was invented in WWII is not the Morse like claim that would prevent all future inventions.


              Useful – not technical.Do you recall the conversations about the difference in US jurisprudence?


                Anon, fortunately, the Supreme Court defines useful, not Judge Rich as in State Street Bank. Sent from Windows Mail

                1. Ned,Why do you keep on trying to pin me to State Street?My position is simply not tied to State Street. The Useful Arts being more extensive than technical is not tied to State Street.I suggest that you stop your crusade against Judge Rich – you clearly go too far in your dislike of someone who knew patent law far better than the Supreme Court. Shall we get on the merry-go-round yet again about what was holding in Benson (the quote I repeatedly show you) and teach you (yet again) what was dicta in Benson (and why Judge Rich could toss that dicta aside) – using the very reference that you supplied in our conversation about BIlski? Please stop trying arguments that I have defeated.

                2. Anon and State Street Bank. What do they have in common?Think sirs.Does a programmed computer that produces an improved price come to mind?

                3. Again, Ned, I reiterate my point that my position is not tied to State Street Bank.Please refrain from dust-kicking.

                4. anon, tied to and is are related concepts.How do you distinguish your position from State Street Bank?


              Ned,Can you obtain copyright protection for math?(and no, you are completely incorrect that software is math).

                1. Ned,That does not answer my question. I asked about math, not about function.Please answer my question.

                2. One can obtain copyright protection for math in the form of computer software, but is protects only the expression and not its function.

                3. Ned,In three different sub-threads you are (once again) playing games and being less than forthright.Here: You cannot copyright math (pure math) – you fail the function/expression dichotomy. The logical conclusion that you CAN copyright aspects of software is that software is not math. Yet, you torture the plain law in order to avoid making a straightforward admission.Below: you continue to insist on trying to pigeonhole me to State Street Bank. I do not need to distinguish my position because I do not derive my position from State Street Bank.Finally, you continue to ignore the multiple points about the fallacious ‘House’ argument, the Grand Hall experiment (must configure before you can ‘use’), Morse analogy and improvements to a computer (and your statement of NWPA not addressing a modern computer is – to be blunt – complete B$ as it is, well, obvious, that it is improvements to a machine under discussion.We all deserve better from you Ned.

                4. Anon, I don’t know where you are going with this copyright discussion, but Congress concluded in ’76 that software could be protected by copyright, but such did not extend to software’s function. How this relates to math is the question I have for you. Math is typically pure function and should not be protectable by copyright. But even if this were true, what does this have to do with the patentability of software? Software is neither a machine, process, article or composition. It is not patentable on any of these grounds.

                5. Ned,You ask me where I am going and you should already know the answer.You (quite incorrectly) state that software is math and you know that such simply cannot be true as you know that you cannot obtain copyright on math.Yet, you cannot seem to come to the correct (and oh so simple) conclusion.Further, you (should) know that aspects of software (the expressive ones) are what garners copyright protection – just as other aspects – the functional and manufacture ones, are what garner patent protection.Then you (quite incorrectly) state that software is not covered under a patent eligible category.And yet, just before the blog change, you admitted that software is covered under 101.Quite simply – and undeniably – software is a manufacture, a machine component.Can we have you stop your games now?

                6. Anon:”You (quite incorrectly) state that software is math and you know that such simplycannot be true as you know that you cannot obtain copyright on math.” Anon, you seem to not be able to understand the differences between genus and species. Software is a form of math because when executed software does nothingelse besides execute arithmetic/logical instructions and/or execute instructionssuch as loads and stores. However, allmatter is not software even though all software might be math. This is same thing as saying that all men areanimals but not all animals are men.

                7. Anon, “Then you (quite incorrectly) state that software is notcovered under a patent eligible category.And yet, just before the blogchange, you admitted that software is covered under 101.”I believe I said that a programmed general purpose digitalcomputer is not categorically excluded under section 101. The real problem with such claims is lack ofnovelty because the computer is not new even though the program he might benew. The new use of an old machine ispatentable as a process by statute.

                8. Anon, “Quite simply – and undeniably – software is amanufacture, a machine component.”Then, quite simply and undeniably all software should be claimedas a component of a machine. But when Iask you whether the program has to be part of the machine in order for themachine to be new in any patentable sense, you decline to answer thequestion. And yet, through thisbackhanded admission, you seem to suggest that a programmed computer is newbecause it has new software as a structural component. But you will not say this, or agree tothis. Why?

                9. Ned, You are not listening. You continue to pursue lines of argument that have been offered before while ignoring the counterpoints that have been offered before to those arguments.The merry-go-round continues.Point one: components and patentability on the component itself. You are absolutely wrong Ned about some need to claim the larger item for an invention of a component.I do not need to claim a joined structure to claim an invention to a new rivet.I do not need to claim a total vehicle to claim an invention to a new tire.I do not need to claim a loaded gun to claim an invention to a new bullet.Stop the fallacies.Point two: Quite simply, Alappat – your position fails.As to your admission, your backtracking here (what you admitted as to 101) is not what you explicitly said (and in fact, is a self-contradicting error, as you cannot have a general purpose machine that has been programmed) – again, your inability to accept the plain meaning of the Alappat case interferes with your appreciation of what the law is now. Reread Alappat – once a genreal prupose machine is programmed, it NO LONGER IS A GENERAL MACHINE. You cannot ‘use’ a machine unless FIRST you change the machine into a new machine by configuring that machine with the software in question. This is not a new point, and we ride the merry-go-round yet again because you refuse to acknowledge the real world around you.Your comments about backhandedness and what I won’t say are nothing more than dust-kicking. I have fully explained (many,many,many times now) what Alappat means.Point three: Math and software – simple set theory. It is not a genus/species issue. I suggest that you try something even simpler: basic set theory. Math, as a set, does not intersect the set of things for which copyright attains. Period.Software does intersect. Thus, software cannot be a subset of math. You are simply mistaken in your attempt to portray math as a genus – simple logical error.

                10. Anon, State Street Bank decided a programmed computer that produced an improved price was eligible as a machine. Now you may or may not associate yourself with this case, but I as you to provide your view on the result, if not the rationale.For example, do you agree with the result?

    1. 1.1

      In UK, unsuccessful parties have to pay court-assessed (taxed) costs. So why not in the US?Restoration of the language “may” would be a step in the tight direction. But in order to know whether that step would suffice it would be necessary to establish what the relevant pre-1952 case law was – a suitable task for an academic researcher, I suggest. Could we see another posting on thsi blog?

Comments are closed.