Lingenio GmbH cares for active interchange with numerous research institutions, is involved in various research projects and continuously makes own contributions to research in the areas of machine translation, lexicography, text and corpus analysis.


Interest and competence in research results from the company’s philosophy and credo according to which it is only possibly to fill a leading position in language technology over a long time and to develop corresponding top products if a permanent dialog is cultivated between development and research so that research can sensitively react on demands from customer oriented development and that linguistic development can quickly turn research results into improvements and new applications and make them available to the customer.


This philosophy has its roots in the genesis of Lingenio as spin off from the former scientific center of IBM in Heidelberg and in the fact that the core of Lingenio’s software systems result from IBM’s Logic Based Machine Translation project (LMT) which was one of the most ambitious and influential machine translation projects throughout the nineties.

Current Work

Today Lingenio’s research has a main focus on integrating statistical methods into so-called rule based architectures and, generally, in accelerating the development and extension of reliable language information with the help of automatic corpus analysis.

Research Versions of the Lingenio Software

We think that a good text analysis infrastructure helps a lot with respect to many research areas. Sometimes it is even a necessity in order to derive qualitatively new or more precise knowledge from corpus data. If for example the words of a corpus are annotated morphologically and related to their basic forms (lemmata) and maybe even classified in addition according to their semantic contribution, it may be possible to cluster individual cases, which as such may be statistically insignificant, to statistically ‘recognizable’ and significant classes of events. In this way, it may be possible for example to recognize collocations or selectional restrictions of complements or other relations basing on linguistic abstraction which otherwise would be unrecognizable because of the diverseness of cases due to morphological differentiation and/or size of semantic classes etc. Keep on reading.


In order to ease scientific work research in such areas, Lingenio provides a number of research prototypes. Keep on reading.