LINGUIST List 27.4104

Thu Oct 13 2016

All: Obituary: in Rememberance of Peter Bosch

Editor for this issue: Kenneth Steimel <>

Date: 13-Oct-2016
From: Carla Umbach <>
Subject: Obituary: in Rememberance of Peter Bosch
E-mail this message to a friend

Peter Bosch, Professor of Computational Linguistics and Cognitive Science at the University of Osnabrück, died suddenly on September 24 of a heart attack. He had celebrated his 67th birthday some weeks before and intended to teach for another year and then continue his research projects as a senior in Cologne. He is survived by his wife Prof. Anke Lüdeling and two adult children.

Peter Bosch studied Linguistics and Philosophy in Berlin and Oxford. He completed his MA at the Technical University Berlin in 1974 and his Dr. Phil. at the Ruhr-Universität Bochum in 1980 (supervised by Helmut Schnelle). Between 1975 and 1988 he had teaching and research positions at the University of Nijmegen and the University of Tilburg, as well as visiting appointments at Oxford, Harvard, Brussels and Cologne. From 1988 to 2000 he was a manager in research, development, and services at IBM Germany in Heidelberg and, from 1994 on, the Director of the Institute for Logic and Linguistics at the Scientific Centre of IBM Germany. In 2000, he became Full Professor of Computational Linguistics and Cognitive Science at the University of Osnabrück.

Peter started his studies in Literature but found out quickly that his real interest was in linguistics and, beyond, in Natural Language Processing and Artificial Intelligence. As early as 1971 he participated in an informal seminar discussing Terry Winograd's dissertation and, in passing, learned LISP programming (which required some determination since courses in programming were reserved for electrical engineers). When he moved to Oxford for a year his focus of studies shifted to Philosophy of Language and Logic. He seized the opportunity to attend lectures by renowned philosophers, in particular Peter Strawson, who introduced him to the problem of presupposition. In a later stay, he also got in contact with W.V.O. Quine, which resulted in a translation of From a logical point of view into German plus an epilogue on synonymy in context by Peter, which was published 1979 by Ullstein publishers.

As a result of this experience, already in his MA thesis on contextual prerequisites for text understanding, Peter was able to combine linguistic accounts with logic, and model interpretation processes from a Natural Language Processing perspective. His dissertation on ''Agreement and Anaphora - A Study of the Modes of Pronominal Reference and Their Constraints'' was one of the earliest pieces of evidence that such an approach can lead to results that surpass by far those that can be achieved by taking the standard approaches. It was published 1983 in the then novel Cognitive Science series of Academic Press and is still one of the standard references in this field. The topic of anaphora remained one of Peter's favorite areas of research. Actually, one of the projects left by his sudden death is on reference and binding of demonstrative pronouns in German.

Peter Bosch's professional career began in 1975 at the University of Nijmegen, where he taught philosophy of language sharing the office with Rob van der Sandt. Although occupied with teaching and administrative duties, Peter still found time to get in contact with the newly founded psycholinguistics group at the Max Planck Institute, which led him back to topics in Natural Language Processing and Artificial Intelligence. When in 1985 Leo Noordman was appointed a professor in Tilburg with a decidedly interdisciplinary focus and asked him to join, he readily agreed. That pushed his interests more and more in the direction of computational linguistics and discourse and put him in contact with industrial research. A position at IBM Germany’s Scientific Center then offered him the opportunity to develop systems going beyond the limits of toy implementations. From 1994 on Peter was the Director of the Institute for Logic and Linguistics at the IBM Scientific Center, with considerable influence on the make-up of computational linguistics projects and cooperation with university groups. The most important one was the LILOG project on text understanding and knowledge representation, which spawned many important research results in linguistics and logic, and the number of computational linguists in Germany who were in some or the other way affected by or involved in LILOG activities is hard to overestimate. Towards the end of the 1990s, even though the Center was still highly successful in combining scientific and business-oriented objectives, IBM gradually cut back their financial commitment to it.

In 2000, Peter, who was determined to continue his research agenda, accepted a position as a Full Professor of Computational Linguistics and Cognitive Science at the University of Osnabrück. He played a significant role in the foundation of the Osnabrück Institute of Cognitive Science shortly after his arrival with groups in Computational Linguistics, Cognitive Psychology, and Artificial Intelligence. At the same time the Osnabrück Cognitive Science study program was launched, offering Bachelor’s, Master’s and PhD degrees, which was unique in Germany then (and even now). The program received a number of awards, including grants for teaching and scholarship. The PhD program was Peter's key interest. He made it into a focal point of the institute's interdisciplinary research and a place where new ideas were explored.

Against the cognitive science background Peter developed theories of context-dependence, definiteness and demonstrative pronouns documented in a number of influential publications. One key aspect of these papers is the empirical evidence they provide. Long before it was considered essential in the linguistic community, Peter insisted that linguistic research must be empirically based. And instead of passing the burden of experimental studies to students and research assistants he acquired whatever competence was needed for himself.

Even though he never lost sight of cognitive science, Peter was a genuine semanticist who focused on core linguistic issues, like the productivity and adaptivity of language. He was a reliable constant in the semantic landscape, as a supervisor, as a project partner, as a reviewer and not least as managing editor of the Journal of Semantics for many years. He was a good advisor in matters of research and beyond. He was what one calls in German ein unabhängiger Geist (‘an independent mind') and he was a critically-minded scientist in an old-fashioned way, with no concern for short-term success and with high standards for scientific quality, particularly in his own work. We will miss him.

Linguistic Field(s): Not Applicable

Page Updated: 13-Oct-2016