Science Research UvA-VU: Logic and Language
Two entirely different disciplines are being brought together. Professor and philosopher Arianna Betti and natural sciences researcher Stefan Schlobach are performing research together. How does this unique cooperative venture work? Stefan Schlobach explains.
05/26/2015 | 4:12 PM
The cooperation between Arianna Betti (Institute for Logic, Language and Computation, University of Amsterdam) and Stefan Schlobach (Artificial Intelligence, VU University Amsterdam) goes back eight years. 'Back then, the Netherlands Organisation for Scientific Research (NWO) called for the submission of research proposals involving a cultural institution,' recalls Schlobach. "The National Library of the Netherlands course come up with terms with which to describe and identify books, to help people find them. We came up with the idea to study how those book descriptions change over time, and whether the older descriptions are still useable today.'
According to Schlobach, the concept of feminism is a good example of a concept that has changed over time and is continuing to change. 'The meaning of that word a hundred years ago was totally different from its meaning in the 1920s - when women had just gained the right to vote - and in the 1960s, when the contraceptive pill became widely available and campaigning began for the legalization of abortion. In formal terms, the same word is used, but it means something different,' Schlobach explains. 'How can you measure ideas that change over time? And how can you study what happens to them? These questions formed the basis for the research proposal that we wrote.'
According to Schlobach, the two disciplines differ in their approaches. 'In contrast with our methods, the methodology of the philosophers is hundreds of years old, and very detailed. They think about and question everything relating to concepts such as time and word meaning, and they are also very critical of assumptions. Our working methods are more pragmatic: What is the problem, and how can we resolve it, by building a system or developing computer languages.' The AI researcher considers it informative and useful to examine in more detail questions such as: What are words/concepts? What is the relationship between concepts and the world? We want to investigate how words can be described, so that ultimately a computer can use logic to reason it out, to bring data together, and to develop tools.
The practical impact of the research is great, says Schlobach. "Let's say that historians want to analyse how many civil servants there were between 1790 and 2010. First you have to know what a civil servant is before you can perform a valid analysis. However, the definition of a civil servant in 1880 was probably different from today's definition. For doctors, the question 'What is a tumour?' is important. What has changed with regard to that concept, the meaning, or the name over the years?" The objective of the researchers is to formulate criteria with which to indicate changing word meanings, so that we can develop a generic theory. 'And, naturally, we aim to create smart databases that can overcome those kinds of changes over time,' concludes Schlobach.