The decline in bargaining power for large groups of workers is at the core of rising inequality. This research aims to provide some of the first causal evidence that contractual language is not merely cheap talk but rather meaningfully shapes the decisions of contracting parties in the labor market. The grant will support an effort to digitize union contracts stored at the Kheel Center at Cornell University. In addition to digitization, the researchers will use language processing tools to extract norms, commitments, and entitlements from the text. The result will be a tool that can be used to understand the role of unions in the 20th century. The dataset will be uniquely detailed, including features of union contracts based on industry sector, union, firm, and year of the contract. The research questions that might be answered with the data range from the fundamental—How are labor contractual terms determined, and how do contractual terms affect workers and firms?—to the more subtle—How and why do contractual terms begin to reflect legal changes and judicial decisions?