On Monday, the Supreme Court granted certiorari in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc. (No. 13-854), which asks whether a district court’s factual findings in interpreting a patent claim are reviewed de novo or for clear error on appeal. Stated that way, the answer would probably seem obvious: Federal Rule of Civil Procedure 52(a)(6) states that findings of fact must not be set aside on appeal unless clearly erroneous, and that provision would seem to answer the question presented.
The complication is that a “factual finding in interpreting a patent claim” is, in an important sense, an oxymoron. In its prior decision in Markman v. Westview Instruments, Inc., 517 U.S. 370 (1996), the Supreme Court held that claim construction was a “purely legal” question “under the authority of the single appeals court.” In this view, there are simply no “facts” to be found in claim construction, and the entire purely legal question would be reviewed de novo as the interpretation of a statute would be. To add to the confusion, however, the Markman Court also stated that claim construction often involves receiving evidence regarding how particular words are used and understood within a certain scientific or technological field, and in this manner it seemed to endorse a view of claim construction as a fact-bound inquiry.
One key to resolving this apparent paradox is to understand that appellate courts routinely find “facts” using de novo review. To take the most salient example, today the Supreme Court (in McCutcheon v. Federal Election Commission) struck down aggregate contribution limits—which prohibit a donor from donating more than $57,500 overall to multiple candidates—in campaign financing. Key to the decision was the Court’s assertion that high aggregate contributions do not lead to quid pro quo corruption or the appearance of corruption. Whether this is true is ultimately an empirical question, yet the plurality opinion does require a district court to find it. Similarly, on a great range of legal issues that involve contested empirical predicates—e.g. whether capital markets are efficient, whether affirmative action results in mismatch, whether the death penalty deters crime—the Court regularly takes positions with no regard to district court fact finding.
More to the point, discerning the meaning of English words is itself a factual exercise. Unlike languages that have a prescriptive authority—such as the Académie Française that prescribes official definitions of French words—the meaning of English words is entirely a matter of social fact: a word means whatever the relevant community of users thinks it means. The word “apple” means a particular type of fruit only because everyone else who speaks English mentally connects the word with the definition.
Understood in this way, there is nothing in principle that is different between patent claim interpretation and other types of textual interpretation such as statutory or constitutional interpretation. Statutory interpretation generally asks what a statutory phrase means to the broad community of English users; patent claim interpretation asks what a claim term means to a smaller community of scientists or engineers. The only difference is that the “facts” required to interpret statutes are usually so second-nature to a judge (judges are themselves members of the broad community of English users, and therefore they learned the meanings of common statutory terms in high school) that the fact-finding process is done subconsciously and summarily; whereas more specialized terms appearing in patent claims may require formal fact-finding procedures such as an evidentiary hearing with expert testimony. But all textual interpretation involves “fact-finding” in the most basic sense of the term.
This does not mean that patent claim interpretation must be treated just like statutory interpretation. One can argue, for example, that appellate judges are just as qualified as trial judges to interpret generalist statutes (since the fact-finding simply entails relying on pre-existing knowledge of vocabulary) but should give deference when formal evidentiary procedures as invoked, as they are during patent claim construction. But understanding the pervasiveness of appellate fact-finding in our legal system does refute the simplistic notion that, merely because patent claim interpretation involves “factual findings” in some sense of the term, de novo appellate review is automatically precluded by FRCP 52(a)(6). There are facts in the philosophical sense of the term, and then there are facts for purposes of FRCP 52(a)(6). The latter is a much narrower set than the former.