"This proposal describes a highly interdisciplinary approach to the empirical study of cultural language evolution. It draws on ideas and methods from *historical linguistics and typology*, *natural language processing*, *biology*, *bioinformatics*, *computer science*, and *statistics*.The computer aided study of cultural language evolution has seen a tremendous upturn over the past fifteen years. This comprises both model-driven approaches - studying the consequences of design assumptions regarding language production, comprehension, and learning for their long-term population-wide consequences - and data-driven approaches that employ algorithmic techniques from bioinformatics to recover otherwise inaccessible information about language history. At the current junction, the field faces two challenges:- The specifics of language evolution - which includes parallels with but also key differences to biological evolution - require central attention.- Model-driven and data-driven approaches need to inform each other to achieve explanatory power and to assess the statistical significance of the findings.The project will establish a radically data-oriented framework for the study of language evolution. This includes three aspects:- replacing the off-the-shelf tools from bioinformatics that are currently in use in computational language classification by linguistically informed algorithms, esp.\ *multiple sequence alignment techniques*,- identifying characteristic traits of language evolution via *exploratory data analysis*, guided by the theory of *complex systems* and employing cutting-edge methods from *machine learning* such as *kernel methods* and *causal inference*, and- developing, implementing and testing models of language evolution that correctly predict the *statistical fingerprints of language evolution*, i.e. pay sufficient attention to the domain specific features of language evolution that have no counterpart in biological evolution."