Process mining focuses on the analysis of recorded event data in order to\ngain insights about the true execution of business processes. While\nfoundational process mining techniques treat such data as sequences of abstract\nevents, more advanced techniques depend on the availability of specific kinds\nof information, such as resources in organizational mining and business objects\nin artifact-centric analysis. However, this information is generally not\nreadily available, but rather associated with events in an ad hoc manner, often\neven as part of unstructured textual attributes. Given the size and complexity\nof event logs, this calls for automated support to extract such process\ninformation and, thereby, enable advanced process mining techniques. In this\npaper, we present an approach that achieves this through so-called semantic\nrole labeling of event data. We combine the analysis of textual attribute\nvalues, based on a state-of-the-art language model, with a novel attribute\nclassification technique. In this manner, our approach extracts information\nabout up to eight semantic roles per event. We demonstrate the approach's\nefficacy through a quantitative evaluation using a broad range of event logs\nand demonstrate the usefulness of the extracted information in a case study.\n