Measuring Semantic Relatedness: A Proposal for a New Textual Tool
Judicial decisions, statutes, constitutions, sentencing guidelines, and ERISA-related documents have at least one thing in common: at a molecular level, the laws are all composed of words. The scientific study of linguistics, particularly the field of semantics, analyzes what words mean and how they are connected with each other. And yet, thus far, the legal field has taken little notice of academic and technological breakthroughs in the field of linguistic semantics. This Note seeks to highlight the potential utility of linguistic semantic tools in interpreting legal texts. Specifically, applying algorithms to a free online lexical database allows anyone with a computer to measure the level of relatedness between two nouns. Like more classical and widely accepted textual tools, these algorithms shed light on the plain meaning and semantic nuances of different words. Applying them to two prominent federal circuit splits regarding federal sentencing guidelines and ERISA benefits further underscores their usefulness across the legal discipline. The legal field stands to benefit from employing semantic linguistic algorithms in the law to help resolve semantic ambiguity in legal texts and arrive at more consistent, quantifiable conclusions.
Katherine A. Cohen