Broad access to justice under the rule of law means ensuring that similar cases are decided similarly, and for the appropriate reasons. But monitoring and evaluating the reasoning in tens of thousands of fact-specific legal cases require at least partial automation. This presentation discusses the linguistic and logical challenges to developing tools for automatically locating reasoning within legal texts, highlighting it for manual review, and extracting it for further analysis. Examples are drawn from U.S. law on healthcare regulation, veterans’ disability claims, and compensation for vaccine-related injuries. In developing automated semantic analysis of legal texts, we can also gain insights into the philosophy of legal language.