Sunday, August 23, 2009

Don't forget to address the combination

In dealing with obviousness rejections, my preference is to always argue that the combination of references doesn't teach all the claimed elements. [As long as the facts support this position, of course. If the combination really does teach the claim, then you're left with arguing that the combination isn't obvious – and that's usually an uphill battle.]

In the vast majority of my cases with a claim including elements X and Y, the Examiner relies on one reference for teaching element X and a different reference for teaching Y. That is, the Examiner doesn't usually rely on the combination as teaching a particular element. In such a case, I can make the following logical argument that the claim isn't obvious because X isn't taught:
  • Reference #1 doesn't teach X.
  • Neither does Reference #2. 
  • Therefore the combination doesn't teach X.
This simplistic approach is not appropriate when the Examiner instead relies on the combination for teaching a particular element. For example, suppose my claim includes "a green widget." If the Examiner relies on Reference #1 for teaching a widget, and on Reference #2 for teaching a green wiki, then the allegation is really that the combination teaches "green widget." In such cases, arguing that the Examiner didn't make a prima facie case of obviousness requires you to go farther than the conclusory statement "therefore the combination doesn't teach," and instead to actually explain why the combination doesn't teach.

The BPAI made this point recently in Ex parte Parmelee (Appeal 2008-2907, Application 09/683,943, Technology Center 2400). Rejected claim 1 included a "data store including digital safe deposit accounts stored therein," where "each of the digital safe deposit accounts is associated with at least one private key." The Board noted that the Examiner relied on Wheeler for multiple stored digital accounts with public/private keys, and on Cohen for electronic safe deposit accounts. (Decision, p. 10.) The Board also found that the Examiner provided a compelling motivation for modifying Wheeler's method to use an electronic safe deposit account instead of a digital account. (Decision, p. 10.)

In addition to upholding the rejection, the Board criticized the Applicant for not addressing "the combined teachings as set forth by the Examiner." (Decision, p. 12, emphasis added.) "Repeatedly, Appellants' arguments include a portion of the subject matter relied upon by the secondary reference [Wheeler] in each limitation which is argued to not be taught or suggested by each of the individual references." (Decision, p.12, emphasis added.)

After reviewing the appeal brief, I find I agree with the Board. The closest thing I see to an argument addressing the combination is a statement that neither reference contained a suggestion to combine. [And as I see it, "no suggestion" goes to overcoming prima facie obviousness, not to arguing there is no prima facie obviousness.] There are instead several statements that amount to "neither reference teaches digital safe deposit accounts, where each is associated with at least one private key." (See, e.g, Brief, 2nd para. on p. 17 and last para. on p. 19.)

Ex parte Parmelee is a good reminder to make sure that your arguments truly address the combined teachings when appropriate.


  1. Thank you for addressing this confusing issue! Great job!! I have a few questions.

    With regard to your green widget example, are you saying that the possible arguments are?:
    1) Reference 1 does not teach a widget;
    2) Reference 2 does not teach green -- maybe argue Reference 2 teaches teal (ignoring my inappropriate addition/change of facts);
    3) The combination of Reference 1 and 2 does not teach a green widget or combining green with a widget.

    Also, what exactly is a combination? Is it the glue that holds "green" and "widget" together or is it the "green widget" itself? ("Glue" perhaps being the reason for combining green and widget.)

    Further, isn't a green widget just a specific kind of widget rather than a combination? (Not critizing your example -- most adjectives further narrow a noun.) Plus, you discussed a green wiki, too... so my statement of the combination could be viewed as incorrect.

  2. Thanks for your comments. I'll try to clarify my reasoning.

    >Further, isn't a green widget just a specific
    >kind of widget rather than a combination?

    I guess it's both. It's a combination in the sense that the Examiner chose to parse my element into its components and to find the components in different references.

    Your list of 3 arguments points out -- rightly so -- that there are other options here. I agree that in my hypothetical, arguments #1 and #2 are also available: you could argue that the individual components of the combination are not present.

    In fact, arguing that individual components are not present is probably a stronger argument, when available. My post was focused on the need to sometimes argue the combination itself, i.e., if the components "widget" and "green" are in fact present, you need to address "green widget".

    >Also, what exactly is a combination?

    Great question, and one that causes me to think about this more. The kind of argument I'm thinking about here is along the lines of: "although Ref1 teaches a widget and Ref2 teaches a green wiki, putting them together does not actually result in a green widget because [wikis have nothing to do with widgets]."

    That sounds like what you refer to as "glue". I still view this as a does-not-teach argument rather than a motivation-to-combine argument. Because my basic position is that simply sticking two individual components (green and widget) next to each other does not in fact result in what I claimed, even though the Examiner characterized it that way. Simply calling it a green widget does not make it so.

    Now that you've brought this up, I can see how one would instead characterize "wikis have nothing to do with widgets" as a motivation-to-combine argument.

    Personally, I reserve the term motivation-to-combine-argument for the one I make when substituting a widget for a wiki DOES make sense. In that case, my factual starting point is that the combination does teach a green widget, so what I'm left with is the argument that a POSITA would not choose to make this particular combination. [Maybe something like "although substituting a widget for a wiki makes sense generically because they're the same class of object, this particular set of references teaches away from doing so."]

    A fine distinction, and one that's hard to explain, especially in a hypothetical context. I think the most important thing is that your response addresses the combination, regardless of how you characterize your argument.

  3. Just found out about this blog. I do have a question regarding the simple argument you talk about. Your claim contains X and Y. Reference 1 contains X and reference 2 contains Y. Then why do you say in your argument that reference 1 doesn't contain X? I understand that you are talking about taking the combination attack route, so I understand that by combination your claim wouldn't be obvious, but if the examiner has already made a rejection based on single references, how could you talk about combination because you are not addressing the examiner's rejection. Sorry for the ramble, but I'm majorly confused. Thank you.


  4. >Ref 1 contains X and Ref 2 contains Y.

    Is this a new hypo? Because my hypo was different. In my hypo, Ex. *alleged* 1=X and 2=Y but as Applicant, I disagreed about X.

    >if the examiner has already made a rejection
    >based on single references how could you talk
    >about combination because you are not
    >addressing the examiner's rejection.

    Nah, I *don't* advocate arguing the combination when Ex. relies on individual refs. In such a case, it's acceptable to argue individual refs.

    The point of my original post was how to handle those rare cases where the Examiner really does rely on the combination to teach a particular element ("green widget").

  5. Oh I see! You are *disagreeing* with the examiner that reference 1 contains X. I understand now. Sorry, sometimes I need to know what is not being stated in order for me to understand what is. Thank you.

  6. Thank you for this very useful post. This is an argument I often make. However, I have been struggling to find case law that supports the simple proposition that if the combination of references does not teach all the elements of the claim, then the obviousness rejection fails. There used to be language to this effect in the MPEP that I would cite to support this simple proposition. But since KSR, the language was dropped from the MPEP and I am left trying to find some solid case law to back it up.

  7. >references does not teach all the elements of
    >the claim, then the obviousness rejection

    The Patentably Defined blog has a) noted that the MPEP doesn't include the all elements test and b) addressed your question about how to handle that.



    But I don't think it's that simple. I think Fed Cir decisions such as PerfectWeb, MasterLock and Tokai show that the refs don't have to teach each element if a POSITA could use common sense to add an element which is completely missing or to modify an existing element.