When responding to an obviousness rejection, it can be easy to get drawn into debating the secondary reference — what it teaches, what it does not teach, how it fails to solve the problem solved by the invention, and how it differs from the claimed invention. This is especially true when the examiner seems to be using the secondary reference in a way that would make it unfit for its intended purpose. However, do not fall into this trap.

A recent PTAB decision, Ex parte Microsoft Technology Licensing, LLC (Appeal 2025-001222, Application 18/080,407), illustrates how applicants can be led astray by focusing too much on the secondary reference.

MPEP § 2143.01(V) explains that If a proposed modification would render the prior art invention being modified unsatisfactory for its intended purpose, then there is no suggestion or motivation to make the proposed modification. Further, MPEP § 2143.01(VI) adds that if the proposed modification or combination of the prior art would change the principle of operation of a reference, then the teachings of the references are not sufficient to render the claims prima facie obvious.

In application, both rules generally turn on the same premise that the modification is being made to the primary reference (the reference being modified). A secondary reference, by contrast, is typically cited only for what it teaches, not for what it does (at least according to most examiners). Arguments that a secondary reference would be rendered inoperative or changed in principle generally miss the mark, especially before the PTAB.

In the example case noted above, the application on appeal concerned an iterative model-training process. Independent claim 21 required, among other things, selecting trained child models as parent models for subsequent iterations and outputting a final trained model. Claim 21 is reproduced below:

21. A method performed on a computing device, the method
comprising:
outputting a graphical user interface having a first graphical element for designating an operation search space of operations available to be performed by candidate layers in an iterative model-growing process;
receiving first user input via the first graphical element, the first user input identifying a set of multiple operations to include in the operation search space;
adding an initial parent model to a parent model pool;
performing two or more iterations of the iterative model-growing process, the iterative model-growing process comprising:
choosing a selected parent model from the parent model pool;
generating child models by adding individual candidate layers to the selected parent model, the individual candidate layers being configured to perform respective operations selected from the set of multiple operations identified by the first user input received via the first graphical element of the graphical user interface;
training the child models to obtain trained child models;
evaluating the trained child models using one or more criteria; and
based at least on the evaluating, designating an individual trained child model as a new parent model and adding the new parent model to the parent model pool; and
after the two or more iterations, selecting at least one trained child model as a final model and outputting the final model

The examiner relied on a primary reference and two secondary references to fill in the claimed iterative aspects. The Applicant argued that the Examiner erred in rejecting independent claim 21 because neither reference teaches or suggests the limitation “generating child models by adding individual candidate layers to the selected parent model, the individual candidate layers being configured to perform respective operations … identified by the first user input received via the first graphical element of the graphical user interface.” In particular, the Applicant did not argue the examiner’s rationale in combining the references was flawed, but rather that the combination of the primary and secondary references’ teachings, as proposed by the Examiner, would “improperly change the principle of operation of [the secondary reference’s]” graphical user interface.

The Board was not persuaded. It found that the Applicant’s arguments were directed mostly to what the secondary references disclosed — not to whether the proposed modification to the primary would render it unsatisfactory for its intended purpose or change its principle of operation. The Board explained that “we agree with the Examiner that Appellant’s argument is unpersuasive because it is directed to the secondary reference, not to the primary, modified reference. A change in the method of principle of operation of the primary reference can render a modification nonobvious. Here, Appellant proffers no binding legal authority to support its argument based on a purported change in the principle of operation of the secondary reference.” (internal citations omitted).

Because Microsoft never developed that point, the Board affirmed the examiner’s rejections of independent claims 21, 35, and 40, along with the dependent claims.

The “unsatisfactory for its intended purpose” and “change in principle of operation” doctrines can be pursuasive when applied properly. But, they generally apply to the primary reference being modified — the one whose operation would be affected by the examiner’s proposed changes.

Practice Tips

1. Start with the primary reference.
   Before addressing the secondary reference, ask whether the examiner’s proposed modification would leave the primary reference still capable of performing its original function.
2. Frame the issue correctly.
   The doctrine applies to the reference being modified. Minimize effort arguing that the secondary reference would be rendered inoperative.
3. Explain the mechanism of failure.
   If the modification would break or fundamentally alter the primary reference, describe how — for example, by preventing it from performing a key step, removing a necessary structure, or changing its control logic.
4. Don’t take the bait.
   It’s tempting to spend pages on what the secondary reference does or doesn’t teach. Winning 103 arguments often turn on what happens to the primary reference once the examiner’s modification is applied.