The Wrong Embodiment Problem: When Examiners Improperly Broaden Claims

One of the most familiar rules of patent examination is that claims are given their broadest reasonable interpretation consistent with the specification. MPEP § 2111. That standard is often shortened to “BRI,” but the shorthand sometimes obscures the real constraint: the interpretation must be reasonable, and it must be consistent with what the applicant actually discloses in the specification.

The specification plays a central role in this analysis. Claims are not read in a vacuum, and the MPEP instructs examiners to interpret claim language in light of the specification. MPEP § 2111.01. At the same time, the MPEP also warns against importing limitations from preferred embodiments into the claims. That balance—reading claims broadly, but not rewriting them based on selected examples—creates a recurring source of dispute in prosecution.

A recent PTAB decision illustrates a subtle but important version of this problem: the examiner uses the wrong embodiment from the specification (one one covered by the claims) to justify an overbroad interpretation of the claim language, especially where the claims are drafted to track a different embodiment (note that this post does not address other issues that may occur by drafting claims that leave out particular embodiments, such as dedication to the public).

In Ex parte Takahashi et al., Appeal No. 2025-001403, aspects of the appeal turned on claim interpretation. The claims recited a mathematical relationship involving area ratios of different regions in a semiconductor device. The language was explicit: the claim required a sum of two area ratios to be smaller than a third area ratio.

The examiner adopted a different construction. Instead of summing ratios, the examiner interpreted the claim as requiring a ratio formed from sums of areas. To support that reading, the examiner cited to a particular embodiment in the specification that described that alternative formulation.

The Board summarized the examiner’s approach succinctly: “As shown above, the Examiner selected a particular embodiment from the Specification as a lens through which to view the disputed claim feature … but claim 1 refers to a different embodiment.”

That framing is critical. The problem was not that the examiner consulted the specification. The problem was that the examiner selected the wrong part of the specification and used it to override the claim language.

The Board went on to explain that the claim language—“a sum of the first area ratio and the second area ratio”—was supported by a specific embodiment in the specification that expressly described summing ratios. Nevertheless, the examiner relied on a different embodiment that discussed sums of areas, and then used that disclosure to reinterpret what the claim meant. The Board rejected that move: “The Examiner has improperly read the claim 1 language that requires summing the ‘area ratio of the first p-type contact region per unit cell’ and the ‘area ratio of the first p-type region per unit cell’ out of the claim, in favor of a different embodiment that relates to ‘the sum of the areas’ of these regions.”

This is not a minor technical error. It goes to the heart of what BRI permits. The broadest reasonable interpretation standard should not allow an examiner to replace the claim’s actual words with different words taken from a different embodiment. Consistency with the specification should not mean that any disclosed embodiment can be used to redefine the claim. Rather, it means that the claim must be read in a way that makes sense in view of the disclosure the applicant actually chose to claim.

Because the examiner’s construction was foundational to the obviousness rejection, the Board treated the error as dispositive: “As the Examiner’s incorrect interpretation is relied upon to reject claim 1 … the rejection is fatally flawed and we need not reach Appellant’s other arguments.”

That is a strong statement, and practitioners should take note. Claim interpretation errors of this type are more common that one would thin. When an examiner anchors an obviousness analysis to an improper construction, the entire rejection can collapse.

This case also illustrates a pattern that many prosecutors have seen in practice. An application discloses multiple embodiments. The applicant narrows the claims to one of them (or a subset). The examiner nevertheless continues to interpret the claims through the lens of a broader or different embodiment. The narrower claims are then treated as if they still cover that broader disclosure. This then leads to improperly maintained rejections, forcing the applicant to either take a different approach (through drawn out prosecution), or appeal.

However, as this case illustrates, an examiner should not treat the specification as a menu from which to select whichever embodiment yields the broadest possible interpretation most convenient to the rejection. The claims, not the embodiment labels, control.

There are several practical takeaways for practitioners.

First, when amending claims to track a specific embodiment, consider explicitly tying the claim language to that embodiment in your remarks. This can make it harder for the examiner to later pivot to a different embodiment as an interpretive anchor. However, consideration should be given as to the implications of abandoning coverage of the unclaimed embodiment.

Second, when an examiner relies on a different embodiment to justify an interpretation, frame the problem precisely. This is not just a disagreement over meaning; it is an embodiment mismatch. Make that explicit.

Third, quote the claim language relentlessly. The Board focused on the fact that the examiner “read the claim language … out of the claim.” That framing can be persuasive. If the examiner’s interpretation requires deleting or replacing words, say so.

Finally, do not let “consistency with the specification” become a one-way ratchet. Consistency cuts both ways. It does not mean “as broad as any embodiment.” It means faithful to what was actually claimed.

The broadest reasonable interpretation standard is not necessarily hostile to applicants. But when it is applied through the wrong embodiment, it stops being reasonable. This decision is a useful reminder that the specification is meant to illuminate the claims—not rewrite them.


Comments

Leave a comment