Federal Circuit: Construing the Term “A” once Again

by Dennis Crouch

In Salazar v. AT&T Mobility LLC, the Federal Circuit affirmed the district court’s judgment of noninfringement. Once again, the Federal Circuit was called upon to interpret the claim term “a,” this time to determine if Salazar’s claim requiring “a microprocessor” was limited to a single microprocessor. While “a ___” is usually interpreted to include “one or more ___,” the court upheld Judge Gilstrap’s narrower singular construction in this case, finding it limited by later references to “said microprocessor.”

Salazar’s US Patent No. 5,802,467 claims a communication system for interacting with multiple external devices. The system comprises “a microprocessor” that generates reprogrammable communication protocols, a memory device for efficient storage of command code sets retrieved by “said microprocessor,” a user interface for user selections and menu displays rendered by “said microprocessor,” and an infrared frequency transceiver coupled to “said microprocessor” enabling bidirectional communication with external devices.

AT&T systems may have the capability to achieve all these steps, but the defendant argued that each step utilized different processors or multiple processors. Thus, as the District Court explained, the dispute centered on “whether the claims require one microprocessor that is capable of performing the recited ‘generating,’ ‘creating,’ ‘retrieving,’ and ‘generating’ functions.” According to the district court construction, the patentee needed to identify a single processor that performed “all the functional (and relational) limitations recited for ‘said microprocessor.'” As a result, the court found no infringement. The Federal Circuit affirmed this decision on appeal.

While the indefinite article “a” is generally interpreted broadly as not limited to one item, this presumption can shift when necessitated by the patent documents. In this case, the patentee repeatedly used “said microprocessor” to refer back to the already claimed term in a way that “reinvokes [the] non-singular meaning” of the word “a.” Slip Op. (quoting Baldwin Graphic Sys., Inc. v. Siebert, Inc., 512 F.3d 1338 (Fed. Cir. 2008). According to the court, an alternative interpretation would ignore the meaning of “said.”

The key takeaway is that each time the claim uses “said microprocessor,” it refers back to the originally identified microprocessor. This interpretation prevented the argument that the claim covers the use of multiple microprocessors working together to achieve the described functionality.  Non-infringement affirmed.

= = = =

The jury sided with the patentee on anticipation. AT&T also appealed this issue. However, the Federal Circuit declined to hear the question, holding that AT&T failed to move for a Rule 50(a) JMOL before the case was submitted to the jury, thus waiving its right to appeal the sufficiency of the evidence on that issue.

51 thoughts on “Federal Circuit: Construing the Term “A” once Again

  1. 9

    Here’s an idea: claim the actual invention clearly and describe it clearly in the specification as well. If the invention is NOT merely “a thing that can perform a bunch of functions” but rather it’s “one or more things, wherein alone or in combination the one or more things perform a bunch of functions”, then just say so. Is this really hard for a modestly skilled patent attorney or agent? No. On the other hand, if the goal is not to claim what was allegedly invented but to cover some vague universe of scope that is not the prior art, well, reap the whirlwind.

    1. 9.1

      I am more than certain that the entire innovation world would agree with me as I give you a resounding:

      F off

    2. 9.3

      That would require skill beyond “modest skill”. Couldn’t have that. Easier to blame the broken “system”.

  2. 8

    “Claim language itself sets the claim scope. Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582, 39 USPQ2d 1573, 1577 (Fed.Cir.1996). This court has consistently emphasized that the indefinite articles ‘a’ or ‘an,’ when used in a patent claim, mean ‘one or more’ in claims containing open-ended transitional phrases such as ‘comprising.’”? Crystal Semiconductor v. Tritech

    So, the antipatent judicial activists include Gilstrap who does not even mention this holding.

    1. 8.1

      The other thing about this too is that it is again removing the POSITA. Processors are so fluid now. They may have multiple cores, jobs may be pushed out to other processors in a farm, and so forth.

      It would make sense to limit it to one processor if it was used in prosecution or if it was a key element of the invention. But neither are true. This is pure anti-patent judicial activism at work.

      1. 8.1.1

        “Processors are so fluid now. They may have multiple cores, jobs may be pushed out to other processors in a farm, and so forth.”

        That’s been true for a while. Any historians out there want to “guess” how long the above technology has been a reality?

  3. 7

    I wonder if the claim had recited “the” instead of “said” if the outcome would have been any different, because “said” is more specific than “the.” I suspect that the outcome of the case would have been the same, but that is not clear to me.

  4. 6

    “Yesterday you were assigned to the Unconscious Bias: Finding and Mitigating Blind Spots class. The purpose of this message is to clarify the training requirement and PPA code:
    • This training is mandatory and applies to all employees
    • The PPA code for all Patents employees is XYZ
    • One hour of non-production/Other Time is authorized to attend this training
    Please direct any additional questions to the program manager.”

    1. 6.1

      Yay “equity!”

      What are the chances that this “blind spot” training will “properly” account for the “pre-Biden and Liberal Left ascent to power“ forced change to “ISMs” that put in an element of “transgressor need be in power” (which the Liberal Left so quickly forgot that had ‘engineered’ when Biden won….

      That’s right folks – with the Liberal Left in power, they fell out of “vict1m class.”

      1. 6.1.1

        Most of the people subject to bias are not liberals. 40% of women vote republican and the majority of african americans identify as moderate rather than liberal, and those two groups numerically make up most of the people subject to bias.

        Also I thought the liberal left ascended with Obama. Obamacare was not particularly liberal (as it was only the republican position from Clinton’s term) but surely that was more liberal than anything Biden has done. Its weird how the current democrat is always dangerously liberal, amirite?

          1. 6.1.1.1.1

            Obviously not. Who is the puppetmaster and what have they achieved in the past two years?

            1. 6.1.1.1.1.1

              The “who” is the universal deep state “they.”

              What they have achieved in the past two years is the natural fruition of Liberal Left machinations.

              Ask yourself, are you (both personally and collectively) better off now than you were before Biden?

    2. 6.2

      Don’t worry. You can be like Intel. They went woke about 10 years ago and have completely destroyed their core business. White people don’t even want to work there anymore.

      But, it’s OK. Biden is giving them $50 billion dollars to make up for it.

      1. 6.2.1

        “White people don’t even want to work there anymore.”

        This is just sad. You should probably try to get a real friend.

  5. 5

    With all due respect to the Court, I think their formulation (“a” usually means “one or more”) is silly and unnecessarily legalistic. The underlying rule is that claims are open-ended, reciting what they include, not (usually) what they exclude. So when a claim says “a microprocessor,” it doesn’t mean ONE AND ONLY ONE microprocessor. That’s the real rule. The infringing article must have one, but it could have many. If you look at it in this way, which I think is the way that is most sensible and true to ordinary English, then the part about “said” simply follows. Yes, if I say “a microprocessor” and then later say “said microprocessor,” which simply means “THAT microprocessor,” then of course there must exist a single processor that has all the characteristics I attribute to “a processor” and “said processor.” There might be more, but I don’t really care about them.
    The contrary idea, that those characteristics might be met by a group of processors acting collectively, runs contrary to the normal English construction. If that’s what you mean, say it.

    1. 5.1

      Nobody disagrees with you. the point is, if i claim “a microprocessor configured to do function 1, 2, and 3”, then can you get by infringement by using one microprocessor configured to do function 1, and a second microprocessor to do function 2. Especially these days, as almost all computers use multi-core processors (essentially different microprocessors), so all the old patents with “a microprocessor configured to ….” would run into problems.

      But yeah, for at least the past 5 years or so, i have made sure to claim “one or more microprocessors configured to…” or similar language. thats one of the easy things to look for to tell good drafters from bad, these days.

      1. 5.1.1

        I understand the point of the case, and its result, I think. I was only commenting on the unnatural formalism of saying that “the article ‘a’ usually means ‘one or more’.” It’s simply not necessary, and in this case results in the Court having to say that this case presents an exception. It’s silly, and I think my comment shows how you get to the correct answers with a much simpler, common-sense approach to the language.

        On the “one or more microprocessors” as good practice – I hear you. My rule is to say what you mean, and understand what you say. If you say that “a microprocessor” can do X, Y, and Z, you better mean it.

        I’ve taken to avoiding the term “microprocessor” in a claim, in any case, unless I’m really sure I mean it. Who needs the drama of arguing about whether that term includes, or is equivalent, to a microcontroller, or to an FPGA or ASIC that is specially designed to carry out the functions?

        1. 5.1.1.1

          Actually, the rule that “a” means “one or more” adds something that is not present just because the claim is open-ended. If I say “a processor performs a, b, and c,” and a means only one, then there needs to be one processor that performs a, b, and c regardless of whether the claim is otherwise open-ended. The fact that the claim is open-ended only means that there could be another processor that does the same, does less, does more, or does something completely different – it does not free the claim from the requirement of there being having all of the functions performed by one processor (assuming that “a” means “a single”. Such is the case in the decision in question. the court did not say that the claim is not open-ended. They only said that although there may be many processors present doing all sorts of things, there needs to be one that performs all the functions recited as being performed by “said processor.”

          1. 5.1.1.1.1

            I agree that the Court didn’t say the claim wasn’t open-ended, and I agree that they said there has to be at least one processor that performs all the functions attributed to “a processor” and “said processor.” I think that’s the right answer. But because of silly things they’ve said before, they had to get there by saying “usually ‘a’ means one or more, but not this time.” This Court didn’t mess up -earlier ones screwed up with the whole “a” means “one or more.” That was not necessary. Instead, what they should have been saying all along is that “‘a’ means ‘one’ but doesn’t mean ‘one and only one’.” That gets you to the same results you would expect, as well as to the result in this case, without the silliness of redefining ordinary words.

      2. 5.1.2

        Ocho,

        You mention “last five years,” while the patent at point (US 5,802,467) was filed in 1995 (a bit more than five years ago).

        1. 5.1.2.1

          1995! Wowee zowee that was YEARS before anyone thought of using multiple computers to perform a combination of tasks. These guys were ADVANCED, like some amazing brainiac hybrids of Einstein, Gates, and of course the immortal Fred Hyatt.

    2. 5.2

      This is the best comment so far, and gets to the same point I was making below in 3.1.1.

  6. 4

    This is where you recite a controller and in the specification define that the controller can include one or more microprocessors, memories, etc.

      1. 4.1.1

        Two questions:

        Is that a bug or a feature)

        Was that intended to be a bug or a feature?

        And one more question:

        Congress wrote this (in answers to the above) FOR applicants, but who re-wrote it AGAINST applicants?

  7. 3

    With respect, I don’t really agree this case was about construing “a” one way or the other, i.e., singular or plural. It was in fact about construing “a microprocessor” / “said microprocessor”.

    A sentence on p. 11 sums it all up quite nicely: “We agree with the district court that while the claim term ‘a microprocessor’ does not require there be only one microprocessor, the subsequent limitations referring back to ‘said microprocessor’ require that at least one microprocessor be capable of performing each of the claimed functions.” (emphasis mine)

    1. 3.1

      Certainly it is not noteworthy to state that “said refers to what was previously introduced.”

      I think it’s accurate to say that some think that “a = one or more” and “one or more” creates three important embodiments: One, multiple, and group. This case confirms that “one or more” does not create the group embodiment. So I think it is most accurate to say that it is construing “a.”

      1. 3.1.1

        I think I generally agree. But I also think it’s good to be precise on this kind of thing.

        For situations like these, you basically have three possible outcomes.

        1. “a/said microprocessor” = only 1 microprocessor
        In that case, of course the sole microprocessor has to perform all the listed functions.

        2. “a/said microprocessor” = >1 microprocessor, but at least one must perform the functions

        3. same as #2 above, but any combination/permutation of the plural microprocessors may perform the functions

        So yes, if you construe “a” as singular, then you inevitably end up with outcome #1 above. But CAFC didn’t do that. They construed “a” as one or more, which follows the default rule. However, as shown above, that’s not enough by itself to distinguish between outcomes #2 and #3.

        Accordingly, CAFC still had to do some further reasoning to end up where it did, at outcome #2. That to me is the interesting part of the decision, not the starting point of just following the default rule when construing “a”.

  8. 2

    None of the individual functions of this alleged “invention” were new and (expectedly) there is no teaching away in the art that these functions could not be combined to achieve … their combined functionality. There are also no unexpected results. This is typical of a lot of “do it on a computer” claims out there and also true of just as many “do it on multiple computer” claims. At least with regards to patenting, huge swaths of this “art” have been propped up by a patent office dedicated to the churning out of meritless patents because computers are shiny and the fees have to come from somewhere. The irony, of course, is watching the beneficiaries complain that they are being treated unfairly which has been a nonstop whine for decades.

    1. 2.1

      Adding: also true of reams upon teams of “do it remotely” claims, “do it on a mobile device” claims, “do it on a wrist-bound device” claims, “do it encryptedly” claims, and (more recently) “do it with AI/machine learning” claims. Just pick your shiny new trend/buzzword and someone in the compooter arts at the PTO is sure to get the vapors.

      1. 2.2.1

        He disappeared for almost fifteen months right after he celebrated 15 consecutive years.

        He tried a number of names (s0ckies) upon his return, and quickly settled to “The Prophet.”

  9. 1

    This opinion took it a step too far. Fed Cir now requires that each processor be configured to perform the entire method. That is not the purpose of that type of language. When a claim says “a”, when interpreted to mean “one or more” then, the one or more has to do the functions as a whole. Meaning, it’s a black box. The claims should merely require that the combination of processors together perform the function, not that each and every processor perform every process step in parallel. That’s not how distributed computing works. Same as saying a system configured to do x, y,z, doesn’t mean that every component of the system performs all of it, just that the combination performs the function. I think bad arguments by the attorneys led to an overextended opinion.

    1. 1.1

      With respect, I am not sure why you think this is an “overextended” opinion. If you look at cases like In re Varma, 816 F.3d 1352 (Fed. Cir. 2016), relied upon heavily in this new case, the court has repeatedly made clear that even when “a” encompasses more than one thing, each of those things still has to satisfy all applicable claim limitations. The court in Varma offered the following analogy which is again quoted in this decision: “For a dog owner to have ‘a dog that rolls over and fetches sticks,’ it does not suffice that he have two dogs, each able to perform just one of the tasks.” If you have more than one dog, therefore, each dog must be able to both roll over and fetch sticks.

      This type of problem could be avoided through claim drafting. To use your example, if the claim had recited a “system for doing x, y, z,” instead of a “microprocessor for doing x, y, z,” the plaintiff could likely have crafted an infringement theory in which the claimed “system” compassed multiple microprocessors. But reciting something very specific like a microprocessor often limits the plaintiff’s freedom in crafting an infringement theory.

      1. 1.1.2

        How about if I say I have businesses (plural) for selling cars and selling medical devices? Does that mean every business has to do both? The term is being interpreted to mean “one or more” so “the” one or more ( not each of the one or more, but the one or more) has the claimed characteristics. How about I have one or more dogs, and those one or more dogs help with fetching sticks and herding sheep. Is it clearly wrong to say one of “the one or more” fetches and the other herds? The combination clearly performs the claimed feature. How about, I have 100 dogs that help with fetching, hearding, emotional support, and detecting cancer. So all of them need to do that or the combination? If I claim a system, wherein “the” system is configured to do x y or z, do we care what sub component or the claimed structure performs it? In the instance where there the “one or more” is only a single processor, then yes it has to do it all, but when the “one or more”processor is two then that combination (singular) has to do it. Again, it’s just a black box. If I have one or more severs for housing email, doing payroll, housing customer data, etc, does that mean that each of the one or more severs (when I have more than one sever) need to do that or merely that combination (the one or more severs) need to do that?

    2. 1.2

      Fed Cir now requires that each processor be configured to perform the entire method.

      The claims should merely require that the combination of processors together perform the function, not that each and every processor perform every process step in parallel.

      That’s not what the decision says, it specifically says that there may be one or more processors, but that at least one of the processors perform all of the functions. It does not require all processors in the system to each perform all the functions, nor does it require it in parallel.

      How about if I say I have businesses (plural) for selling cars and selling medical devices? Does that mean every business has to do both?

      It means of all the businesses you have, at least two of them must sell cars and medical devices. Assuming it’s a comprising claim and not intended use, you have two limitations directed to a business – sell cars, sell medical devices, and a limitation that you have plural of that. Because it comprises a plurality, you may have other nonconforming businesses so long as you you meet the two businesses selling cars and medical devices standard.

      The term is being interpreted to mean “one or more” so “the” one or more ( not each of the one or more, but the one or more) has the claimed characteristics.

      You introduce X, then THE X must have Y, and THE X must have Z. Therefore, there is only infringement if X has Y and Z. But X is referring to “one or more” microprocessors that have Y and Z, i.e. any nonzero number of things that all have Y and Z. Pluralizing something does not spread limitations out amongst the group because the subject of the limitation is still on “something” and not “group.” There’s a difference between a group of one or more processors where said processors do Y and Z and a group of one or more processors where said group does Y and Z.

      How about I have one or more dogs, and those one or more dogs help with fetching sticks and herding sheep. Is it clearly wrong to say one of “the one or more” fetches and the other herds?

      One fetches, another herds is noninfringing, because you said one or more dogs and THAT PRIOR ITEM (“those one or more dogs”) has two qualities. That places two requirements upon (embodiment A) one dog, or upon (embodiment B) greater than one dogs. Contrast with “one or more dogs, wherein at least one dog of the one or more dogs fetches and at least one dog of the one or more dogs herds” which are serial singular limitations placed upon one dog or upon multiple dogs. Contrast again with “a group of one or more dogs, wherein the group of dogs help with fetching and herding” which limits the sum characteristics of the group, and not individual characters within it.

      If I claim a system, wherein “the” system is configured to do x y or z, do we care what sub component or the claimed structure performs it?

      No, because this is a limitation on the group and perfectly explains why “A system comprising one or more processors where the system does X, Y and Z” is completely different than “A system comprising one or more processors where the processors do X, Y and Z.” The former places four limitations on a box (inside are some number of processors, and the box can perform three functions) and the latter places one limitation on the box (inside are some number of processors) and three limitations on the processors (and each of those processors does X Y and Z).

      Note that this does not require every processor in the system to do X, Y and Z, because it is a comprising claim with a singular microprocessor embodiment. That means there is infringement when one processor does X, Y and Z, and the open ended system may have other processors (that don’t do X Y and Z) as it wishes.

    3. 1.3

      I disagree with you, Mike. A “microprocessor” is not a black box.
      If you want to cover the possibility of implementing it with multiple processors, then it’s easy enough to say so. Recite a “processing circuit” and clearly define that in the specification.

      1. 1.3.1

        Absolutely agree with you there. My standard practice was to have something like “a system comprising (memory + processor etc), wherein the system is configured to” x,y,z.

        1. 1.3.1.1

          Joke is on you then, because my system has multiple memory chips and multiple memory layers 😉

        2. 1.3.1.2

          hope that “was” your practice, because you are probably invoking 112(f) when you say “system configured to”.

          1. 1.3.1.2.1

            Great point. If I merely said a system configured to, then yes I agree. But I most often recited something like “a system comprising a processor and a memory, the memory having instructions that when executed by the processor, cause the system to” do xyz. I don’t often see 112f in those circumstances.

Comments are closed.