Our research has a very broad orientation and covers a multitude of topics from computational linguistics, general linguistics and comparative linguistics. Additionally, we are interested in scientific methods and scientific methodology and try to address these topics from the viewpoint of linguistic research questions.

Language comparison – be it typologically or historically orientied – has a long tradition in linguistics, reaching back to the beginnins of the 19th century. Thanks to new computer-assisted methods and thanks to the availability of digital language data it is now finally possible to evaluate the traditional methodology again and to try to enhance it by means of computational methods. In the context of our research we develop new computer-assisted methods, that we use to address concrete problems in historical and typological language comparison.

Chinese is a historically extremely well documented language. This enables us to investigate the development of hte language in great detail. Many factors, however, make the concrete investigation difficult. The Chinese writing system, for example, does not tell us anything about the way the words were pronounced in ancient times and ancient readings have to be reconstructed. The writing system also masks the huge dialectal variety of Chinese, with numerous mutually unintelligible varieties, whose diversity in ancient times can as well only be uncovered through linguistic reconstruction. In the context of Multilingual Computational Linguistics we try to successively automatize the multiple methods and procedures which were developed in the discipline of Traditional Chinese Phonology. Our work covers many different topics in this regard, ranging from network analyses, via phonetic alignments, up to the phylogenetic reconstructioon of Chinese dialect history

All languages change as long as they exist. Already August Schleicher mentioned this in 1863, and this view has not changed until today. Language change proceeds in interesting turns, touching upon all core areas of language, ranging from phonology (the sound system) via the pronunciation of individual words and their meaning, up to morphology and syntax. Languages change independently of whether they come into contact with other languages, but in the case of language contact, the processes underlying language change become much more complex but also much more fascinating. This is specifically also due to the fact that words do not need to be borrowed directly from one language to the other. They can also be transferred indirectly (as can be seen from the German word Wolkenkratzer, lit. “cloud scraper”, that was coined after English skyscraper). We investigate language change and language contact in multiple ways and try specifically to develop new method that allow us to investicate language contact and language change in a computer-assisted form.

Our work does not remain in the abstract but is always based on concrete examples from concrete language families. Due to past and ongoing collaborations and due to our individual experience, our focus lies on language families in South-East Asia and South America. In South-East Asia, languages from very different language families have developed in very close contact with each other over thousands of years. In South America, we can also find language families that developed in close contact with each other. Additionally, we find dominant varieties in both regions, such as Chinese or Spanish, which often have a strong influence on the development of less dominant language varieties with a smaller amount of speakers. Our work consists in the collection of data on individual language families and their analysis with the help of the computer-assisted tools we develop.

Lexical semantics deals with the meaning of words in individual languages in specific and in human languages in general. Lexical typology tries to compare the structures of the lexicon in the languages of the world in order to identify commonalities, differences, and basic tendencies of development and structure. Multilingual Computational Linguistics contributes through the application of computational methods in a very specific form by applying, extending, and developing procedures that allow us to compare lexical structures in the world’s languages. These procedures are all computer-assisted and make use of lexical data that have to be standardized and curated in various forms. For the visualization of our findings we develop interactive applications (such as, for example, the CLICS database) that allow us to investigate semantic structures in a transparent way.

Due to the important role which data and their analysis play in Multilingual Computational Linguistics, we are very interested to make the curation of data and the sustainable analysis of research data an object of our research itself. Based on the multitude of topics that we investigate in a data-driven way, we aim to provide examples for good scientific practice in the treatment and curation of research data and offer them to the research community. Central topics of our work in this context are the standardization, the computer-assisted curation, and the interactive analysis of research data. Here, annotation and visualization play a very important role.

Since most of the problems that we have to deal with in Multilingual Computational Linguistics do not have ready-made solutions that we could use, we often have to search actively for solutions. For this reason, we have gained quite some experience in the field of scientific problem solving and we try to share our insights with the research community. Our central strategy for the development of consistent solutions to scientific problems consists in the targeted modeling of research data, accompanied by an interdisciplinary orientation that helps us to adapt existing solutions in neighboring disciplines to our specific needs. Algorithmic procedures play an important role when searching for concrete solutions to individual problems.