When Expert Testimony Becomes Word Salad: Doctrine of Equivalents and Particularized Linking

Doctrine-of-equivalents testimony has to do limitation-by-limitation work, and the Federal Circuit has shown it will reverse a verdict where the expert glosses past the linking. NexStep v. Comcast (Fed. Cir. 2024) is the cleanest recent statement of the rule. The panel faulted what it called "word salad" expert testimony and held that a DOE finding cannot rest on generalized assertions of equivalence. The opinion is not a doctrinal novelty. It is a reminder of a rule the Federal Circuit has been enforcing for nearly thirty years, and a reminder that retaining counsel can still get it wrong.

The legal standard

Doctrine of equivalents lives in the gap between literal infringement and the public-notice function of patent claims. The Supreme Court framed the modern standard in Warner-Jenkinson v. Hilton Davis Chemical, 520 U.S. 17 (1997), holding that the doctrine must be applied on an element-by-element basis, not to the invention as a whole. Equivalence is judged either under the function-way-result test or under the insubstantial-differences test, and the expert has to walk through whichever framework she is using for each claim limitation that is not literally met.

The particularized-testimony rule is older than Warner-Jenkinson. Texas Instruments v. Cypress Semiconductor (Fed. Cir. 1996) held that "generalized testimony as to the overall similarity between the claims and the accused infringer's product or process" will not support a DOE finding. The expert has to say, for each limitation at issue, what the function is, what the way is, what the result is, and why the accused element performs substantially the same function in substantially the same way to achieve substantially the same result. Or she has to walk through why the difference between the claimed limitation and the accused element is insubstantial.

Either framework is fine. What is not fine is generalized assertions of similarity at the invention level.

The NexStep correction

NexStep, Inc. v. Comcast Cable Communications, LLC, 119 F.4th 1355 (Fed. Cir. 2024), came up on appeal from a Rule 50(b) JMOL of non-infringement. The jury had found DOE infringement; the district court entered JMOL on the ground that the patentee's expert had not supplied particularized testimony tying the accused product to the claim limitations under either DOE framework. The Federal Circuit affirmed. The district court had called the expert's testimony "word salad," and the Federal Circuit endorsed that characterization, treating it as a shorthand for testimony that gestures at equivalence without doing the limitation-by-limitation work.

The doctrinal holding is narrow. The practical holding is broad. A retaining partner who reads NexStep and does not change how she prepares DOE expert reports is leaving a JMOL-and-affirmance issue on the table.

Practical implications

A few points fall out of NexStep and the line of authority behind it.

First, the expert report needs a discrete section for each claim limitation that is asserted under DOE. The function-way-result analysis or insubstantial-differences analysis for that limitation lives in that section. It does not live in a global paragraph at the end.

Second, the trial testimony has to track the report. The Federal Circuit reviews the trial record, not the report alone. An expert who has done the linking work in her report and then drifts into generalized similarity on direct will not be saved by what she wrote.

Third, the linking analysis has to engage with the accused element specifically. "Performs the same function" is not a finding. "Performs the function of converting the analog signal to a digital signal in the same way the claim recites, by sampling at the same rate and applying the same quantization step, to achieve the same digitized output" is a finding. The level of specificity is the rule.

Fourth, the expert should be prepared to identify, on cross, the limitation at which the literal-infringement analysis fails. The DOE testimony only matters for those limitations. Asking the expert "where do you fall short of literal infringement, and what is your DOE theory for each of those limitations" is a productive cross.

How WitnessLens treats DOE expert testimony

Daubert and reliability challenges to DOE expert testimony have moved in step with NexStep. Among the district court Daubert orders tracked on WitnessLens, motions targeting DOE testimony are granted at a higher rate than motions targeting literal-infringement opinions, with the most common ground being failure to do the particularized analysis on a limitation that is not literally met. The pattern is consistent across the major patent venues.

The remedy at trial, if a DOE motion is denied but the underlying testimony is thin, is a Rule 50 motion at the close of the plaintiff's case framed against Texas Instruments and NexStep. The motion is not always granted, but it sets up the appellate record cleanly.

A note on the data

The aggregate observations in this post draw on the district court Daubert orders involving patent expert witnesses tracked on WitnessLens, with the DOE subset filtered for motions that specifically targeted equivalents testimony. Case citations are to published Federal Circuit and Supreme Court decisions. Methodology details are on the methodology page.