Databases for Linguistic Purposes: a case study of being always too early and too late

At the MPI for Psycholinguistics “databases” were in the center of interest as long as it exists, since experimental and observational data is the basis of all its empirically driven research and “databases” are necessary to organize and administer research. This usage of the word “databases” already may indicate its different usage and indicate the large variety of data sets that is associated with it. In this paper we do not want to give a definition, but first describe and analyze four concrete examples of data sets that are used in the MPI that required modeling and implementation decisions. Then we will summarize our experience and come to some general statements.