Buffalo Law Review

First Page


Document Type



From comic conventions to disbanded dioceses, courts continue to struggle with a unique but puzzling question of trademark law. Federal law protects certain terms that refer to a product or service from a specific producer instead of to a product generally. Terms that refer to products are considered generic and cannot receive protection. Courts have also held that a term that was generic at the time the party adopted the mark cannot receive protection, even if the public later views it as being specific to a particular producer. But, many marks were adopted decades or centuries ago. As a result, courts and parties have struggled to find tools that will let them explore what the public understood a term to mean at the time of its adoption. Current methods, such as dictionaries, have fatal flaws. This paper proposes a new and empirical method to make this determination: corpus linguistics.

Corpus linguistics is the study of the meaning of words and phrases using large databases of language use. It allows the user to understand how people used certain words or phrases at a specific point in time instead of being limited to how people understand them now. Using two recent cases as examples, this paper shows how corpus linguistics can provide parties and courts with concrete data on the meaning of a term when the party adopted it, eliminating guess work and other unreliable sources.