The LLT Lab creates semantic data through its various projects. The law finds physical expression in documents, but we must understand the meaning of the words in those documents before we understand what the law is. Semantic interpretation of the textual contents of legal documents will enable computers to assist humans to mine information from those documents and to use that information to perform useful work – for example, to suggest new legal arguments, or to retrieve legal rules or evidence for use in such arguments. In order to write automatic annotation programs or to submit annotated material to machine learning programs, raw legal texts must be interpreted for their meaning. The Lab annotates spans of text within legal documents for the semantic roles played in legal reasoning and argumentation. Current areas where we create semantic data are vaccine-injury compensation decisions, medical malpractice judicial decisions, and veterans claims decisions.