April 15, 2021

The role of extrinsic sources in interpreting patent claims has been a source of debate in the U.S. Court of Appeals for the Federal Circuit. The Court for years encouraged the use of extrinsic sources such as dictionaries, encyclopedias, and treatises. The Court abruptly changed course in 2005, however, largely repudiating its earlier cases extolling the usefulness and reliability of extrinsic sources. The Court justifiably worried that undue reliance on dictionaries and other extrinsic sources subverted the role of intrinsic evidence. The Court also detailed the shortcomings of using extrinsic evidence to shed light on terms in patent claims, noting that reliance on extrinsic evidence would alter the scope of patent claims and “undermin[e] the public notice function of patents.” The Federal Circuit was correct to hesitate before crediting many existing extrinsic sources as universally reliable guides to the meaning of patent claims. A patent’s specification and prosecution history, however, often do not clarify a term sufficiently for courts to jettison extrinsic evidence completely. There is a need for objective sources that shed light on how a person of ordinary skill in the art would interpret a certain term.

This Article provides an overview regarding how an emerging tool—corpus linguistics—could fill the void. Corpus linguistics has the capability to perform language searches in a general or specialized database of words (corpus) to deduce the majority usage of a term. When a patent uses a word in an ordinary manner, litigants may use a general corpus to demonstrate the term’s meaning. When, as is most common, a patent uses a term in a specialized scientific sense, litigants may construct an ad hoc, specialized corpus that could include relevant patents, published patent applications, scientific journals or treatises, or other scientific materials. Such specialized corpora would be specific to each subject area and would assist in determining how a specialized term or phrase is used in publications written by and for skilled artisans. When the intrinsic evidence is not sufficiently clear regarding how a person of ordinary skill in the art would interpret a term, courts may turn to the next best alternative—discovering, quantitatively, how the majority of persons of ordinary skill in the art use that term. In this way, methods of claim construction may more fully support the public notice function of patents.

Authors: Neal Hoopes, Paxton M. Lewis & Amanda Black


Volume 22, Issue 3