PUB - Publikationen an der Universität Bielefeld

We examine how multimodal data collections, resulting mainly from (psycho)linguistic experiments, can be expressed in standardized metadata description formats. We summarize how such data collections differ structurally from the traditional concept of corpora, and we list thoughts, problems, and solutions that occurred when we designed and collected ISOcat data categories and CMDI components for the metadata representation of these data collections. As a result we present plans for an ontology of modalities and related concepts and data units, which we consider a more appropriate enviroment for the kind of multimodal data we are dealing with.