Real Consumer Language: A Corpus-Based Methodology for Genericide Claims and Trademark Protection
The Teflon and Thermos genericness surveys have been dispositive in numerous trademark infringement and dilution cases pertaining to genericness, helping to prove trademark significance, create reasonable inferences, or establish material issues of fact as to the genericness of marks such as “PILATES,” “SWISS ARMY,” and “SPORTS CHANNEL.” Despite the influence genericness surveys have exhibited, they are prone to a host of limitations, including inefficient sampling, expensiveness, sampling errors, inadequate sampling size, and demand characteristics. Further, surveys are prohibitively costly and indiscreet for trademark holders to use for non- litigious monitoring of their trademarks. This paper proposes the use of corpora for measuring consumers’ understanding of a mark as generic. Corpora are large databases of text that allow for the statistical approximation of how a given word is used. In this case, capitalization rates, the frequencies of alternative common names, and other linguistic indicators can be used to ascertain whether a mark is being used as a generic term. With large quantities of textual data from blogs, websites, and other sources readily available, expert witnesses can construct or use already existent corpora to analyze disputed marks. This paper exhaustively delineates the methods by which corpora can be used to measure genericness, informing practitioners on using corpora in genericness disputes and trademark holders on using corpora to monitor the use of their trademarks. Ultimately, the paper proposes that corpora are more cost-effective, empirical, and efficient than surveys in doing so.