Theo Janssen (1948-2018)

         TMVJanssen19761122_480      Janssen_TMV

It is with deep sadness that we inform you that Theo Janssen died on December 6, after a long illness, at the age of 70. A trusted colleague, dedicated teacher and solid researcher passed away. Theo was connected to our university for almost 40 years before retiring in 2013. Interested in formal syntax and semantics since his student days, he worked in this field for all his life. Most of his work centers around the principle of compositionality, of which he was an eloquent and influential defender. Within the ILLC he was both feared and admired for his critical mind in all things, which we will miss greatly.


Obituary

Theo M.V. Janssen (29 July 1948 – 6 December 2018)

 

Theo Janssen started his study in mathematics at the University of Amsterdam in 1966, obtaining a candidate’s degree in 1969 and a master’s degree in 1975. Starting from 1972, he was also a student assistant in the department of mathematics. From 1975 to 1980, he was a junior scientist in the pure mathematics division of the  Mathematical Centre in Amsterdam on a project connecting computation and natural language, and from 1980 to 1985, he held a position in the department of philosophy. Starting from 1985, Theo became an assistant professor in the newly built department of computer science, joining the theory group of Peter van Emde Boas whose main topics were algorithmics and semantics. This group entered the ILLC officially in 1991, where Theo remained active until his retirement in 2013.

Theo’s master’s thesis concerned two-level 'van Wijngaarden grammars’. This well-known formal grammar formalism was known to have universal computing power, but the reverse inclusion was still open. Theo proved that that no supercomputable power arises: two-level grammars compute no more than the standard computable functions.

However, from an early stage, Theo’s interests broadened from the mathematics of formal grammar and computation to connections with the syntactic and semantic structures found in actual natural language. This led to his joining an Amsterdam circle of young linguists, logicians, and philosophers, emerging in the mid 1970s, which included many people who would later become founding members of the ILLC. This group focused its attention on the theory we know today as Montague Grammar which combines logical rigor with serious linguistic content. Theo’s initial interest was a computer implementation of the grammar in Montague’s paper “The proper treatment of quantification in ordinary English”. This implementation was to be flexible and conceptually transparent, in order to cope with the various extensions of the original fragment proposed by many linguists at the time.

Soon this search for connections between mathematics and natural language acquired broader dimensions when Theo discovered further systematic parallels between natural language and computation. In 1976, he managed to capture the concept of a variable in programming languages using Montague’s intensional logic. The resulting model-theoretic semantics, which solved existing problems in understanding crucial computational devices like arrays and pointers, was presented at various prestigious computer science conferences.

To many, this work of Theo Janssen’s has shown that Montague’s program does not just apply to natural and mathematical languages, but also to programming languages, thus considerably expanding its scope, by placing it in an interdisciplinary triangle of logic, language and computation. But he also went on to the foundations of this interface, finding powerful mathematical formulations of the principle of Compositionality, “the meaning of a compound expression is composed from the meanings of its parts”, in terms of many-sorted algebras and homomorphisms.  These were then used as a yardstick for a (sometimes very) critical assessment of proposed linguistic extensions of Montague Grammar. Universal algebraic methods remained close to his heart ever since.

All these ideas came together in Theo’s landmark dissertation “Foundations and Applications of Montague Grammar” (1983), with, in the jargon of the times, Peter van Emde Boas as a promotor, Renate Bartsch as a copromotor, and Johan van Benthem as a coreferent, together representing the various perspectives coming together in this work.

In the decades since, Theo Janssen became an acknowledged international expert on the topic of compositionality, who published widely on a range of topics, from the historical origins, where he discovered that the principle is wrongly attributed to Frege (a result he even presented in the Fregean heartland of the then German Democratic Republic) to the foundations of machine translation. He went on to publish many further insights into the mathematical foundations and linguistic applications of the topic. All these come together in his chapter on compositionality in the “Handbook of Logic and Language” (1997), which shows his mastery of the subject and his international recognition. In later years, one of his main interests were the challenges posed to compositionality by the game-theoretic semantics of Hintikka and others, and the proper understanding of the accompanying ‘information-friendly logics’. His qualities of painstaking analysis, clarity of thought and presentation, and fearless standing up for his critical points, concerning curious variable-binding features of existing independence-friendly systems, were evident throughout.

In addition to his solid record in research and interdisciplinary bridge-building, Theo was also a dedicated and well-respected teacher for many generations of students, an active supervisor of master’s and PhD students, an advocate and implementer of curriculum reform whenever the time was ripe, and, and in recent years also, of outreach to high schools.

The ILLC has lost one of its founding members who was there right from the initial period in the 1970s, helping build the interface of logic, language and computation, and exemplifying a harmonious combination of interdisciplinary spirit and mathematical standards. Moreover, Theo Janssen exhibited personal qualities that were plain for all to see: he thought deeply, he could be critical without social sugaring to anyone whenever called for – but at the same time, he was a trusted team player, deeply loyal to his colleagues and his broader environment.

Theo will be missed by his colleagues and students. Our thoughts go out to his daughters Marlieke and Jorien.

 

Sonja Smets, director ILLC

Johan van Benthem

Peter van Emde Boas

Frank Veltman

(11 December 2018)


 

Grafrede

Beste Marlieke, Miriam, Jorien, familie en vrienden van Theo

Dit is de tweede keer dat ik afscheid moet nemen van een van mijn
wetenschappelijke kinderen.

Ik leerde Theo kennen toen hij in 1969 op trad als studentenvertegenwoordiger bij de vakgroep wiskunde van de UvA, enkele jaren voor de invoering van de radendemocratie aan de Universiteit. Daarbij was hij zowel innovatief als effectief.

Hij was degene die reeds in 1974 het initiatief nam om met een groep
wiskundigen de studie van de natuurlijke taal serieus te nemen, en nam ons mee naar de Montague gemeenschap waaruit zo veel moois is voortgekomen. Zijn rol bij het tot stand komen van de samenwerking tussen wiskundigen, logici, filosofen, taalkundigen en informatici waaruit uiteindelijk ons Institute for Logica Language and Computation is ontstaan kan niet genoeg worden benadrukt.

In 1985 trad Theo bij mij in dienst bij de sectie Theoretische Informatica met de bedoeling een programma grondslagen der semantiek op te starten. Echter, omdat ik voor 8 maanden vertrok naar het buitenland moest hij eerst, samen met mijn andere medewerker Leen Torenvliet, de sectie bestieren, en onder meer een nieuw curriculum Informatica opstellen, en dat alles als part time medewerker.

Met Theo heb ik ook de nodige reizen gemaakt, zowel door de wetenschap als door de landen waar wij onze resultaten mochten presenteren. Met programma semantiek naar Turku in Finland en naar de hoge Tatras in Slowakije, en met het compositionaliteitsprincipe naar het Volksdemocratische Jena en
Schwerin, en meer recent, met de geschiedenis van de Humaniora naar Rome.

Illustratief voor zijn respect voor het publiek is dat Theo vermoedelijk de eerste promovendus aan de UvA is geweest die voor de verdediging van zijn proefschrift een lekenpraatje heeft gehouden.

Theo had een groot hart voor het onderwijs, de studenten en zijn
toehoorders. In mijn notities zie ik dat wij reeds in 1971 delibereerden over geschikte thema’s voor een seminarium Automatentheorie. Later toen hij het onderwijs in dat vak van mij had overgenomen koste het hem geen moeite om, op grond van didactische motieven, mijn ordening geheel overhoop te
halen. Theo was immers altijd zeer kritisch op wat zijn collega’s wilden presenteren, onderwijzen of opschrijven. Zijn opmerkingen waren meestal correct en werden gewaardeerd.

Wij zullen het in de toekomst zonder zijn constructieve bijdragen moeten stellen. Dat geldt niet alleen voor de leden van ons ILLC maar ook voor de wiskundigen, informatici, taalkundigen en filosofen in onze omgeving, want hij werd alom gewaardeerd.

Peter van Emde Boas


 

Theo Janssen was right there when the Amsterdam environment in logic, language and computation formed in the 1970s, and he stood for mathematical rigor and constancy of purpose, sometimes also single-mindedness -- though I always found him open to reasonable argumentation. Our personalities and working styles could not have been more different, but we matched.

I felt honored to be involved with his thesis, and I still keep our extensive correspondence. It ranges from a shared love for universal algebra (I wrote a paper on homomorphism extensions triggered by questions of Theo's) to sometimes heated discussions on what what natural language really is, and what semantics should be about. Naturally, Alice ter Meulen and I asked Theo to contribute the central compositionality chapter to the "Handbook of Logic and Language". In later years, we shared an interest in game-theoretic semantics and the wonderworld of IF logic with its beguiling mixture of insights and mirages, a poetic maze that Theo assiduously tried to sort out.

In real life, Theo was the opposite of a game-theoretic strategist. He said what he thought at all times, instead of saying what is useful for gentle social manoevering, something he considered (as he once told me) dishonest. I am afraid that Theo's working style is not what makes for successful institutes or universities, but I always felt as a director of the ILLC how valuable it was to have him close by as a moral compass.

It is hard to imagine not seeing Theo any more in our corridors and offices. The world of the ILLC founding generation has suddenly become poorer.

Johan van Benthem


 

My most vivid memories from Theo go back to the days in the early 1980s when Theo and Jeroen and I were working on our PhD theses. At the time Theo was stationed at the Department of Philosophy and had an office right across the hall from ours. We interacted daily, sharing lunch in our (slightly bigger) office, often with Fred Landman and  sometimes Renate Bartsch present as well. The conversation was about semantics and logic almost all of the time. Often Theo would eat his sandwich standing at the blackboard writing something, in many cases to show that we got something wrong. His focus on getting things right, and his single-mindedness about what getting things right amounts to, no matter what other people said, was inspiring. And so was his willingness to share his insights and critical views without any reservations. At times that led to heated debates, and prolonged differences of opinion. But it never resulted in any hard feelings. Not just because Theo was often right (he was), but also because whatever the difference of opinion, it was for Theo never something  personal. That is a rare quality, and I remember it fondly and miss it dearly.

Martin Stokhof


 

In 1981 I edited together with Theo Janssen and Martin Stokhof the Proceedings of the Third Amsterdam Colloquium. One of the papers was Hans Kamp’s A Theory of Truth and Semantic Representation. As editors we did not fail to notice that the theory faced a serious problem: it did not adhere to the semantic Principle of Compositionality, prominently present in Theo’s academic work. So, we urgently tried to get Kamp to do something about this. But he didn’t care too much. We decided to be pragmatic rather than principled, and accepted the paper. Non-compositional as it was, it became one of the most influential papers in semantics of the previous century.

Jeroen Groenendijk


Although Theo and I both had our roles in the pre-IILC cooperation in logic, language and computation we remained rather distant colleagues until from 2001-2003 we joined forces in the logic course for medical informatics students at the AMC. On my side with a somewhat anxious start given Theo’s strong opinions on teaching. It turned out to be a smooth cooperation. His presence clearly improved the course I had given there some years before and he was a very reliable and considerate colleague. Our resulting friendship was further solidified when from 2006-2009 we had to share a room in the Euclid building. He was a tolerant and pleasant roommate. As a colleague and friend he will be missed by many of us.

Dick de Jongh


Theo was a man of principles. He professed and promoted pure intellectual integrity, and thus educated and shaped my generation. Theo was also the one and only person who presented at all of the famous first ten of the Amsterdam Colloquia, and not (only) because of his name or fame, but because of the quality of his submissions. He continued to contribute until the fifteenth Amsterdam Colloquium. Here are the abstracts of 15 x Theo in 30YAC. Must reads!

Paul Dekker