Niculae-Gabriel Sandu - Research#

About myself

I arrived in Helsinki in the late 70s after taking a degree in economics at the Academy of Economic Studies, Bucharest, Romania. I studied philosophy at the University of Helsinki. The department had a strong profile in logic. The combination of classical game-theory and logic led to my main research interests: game-theoretical semantics and IF logic (Independence-friendly logic).










Game-theoretical semantics (GTS)

GTS was developed by the Finnish philosopher Jaakko Hintikka in the 70s. His idea is to interpret logical languages by two person games, played by a verifier and a falsifier. The first one tries to show that a given sentence is true (in a particular situation or model) whereas the second tries to show that it is false. In this new setting, classical logical systems like first-order logic turn out to correspond to what in classical game theory is known as win-lose games of perfect information. I applied, with other philosophers and semanticists, game-theoretical ideas to the analysis of natural language phenomena: quantifiers and pronouns.

IF logic

I started to work on IF logic in the late 80s (with Jaakko Hintikka). The starting observation was that the linear structure of our familiar logical systems makes them inadequate to express certain mathematical and natural language statements which are not only meaningful, but much in use. In other words, I observed that there are certain relation of dependence and independence between quantifiers and connectives which cannot be expressed by first-order languages, but require stronger ones. The result of these investigations is IF logic, an extension of first-order logic which is interpreted by win-lose games, as its first-order relative, but these games are now games of imperfect information. Apart from their greater expressive power, IF languages also allow an interesting combination between truth-conditional and probabilistic interpretations.

Truth

Our classical logical languages cannot express their own truth-predicate, that is, they cannot, on pain of inconsistency, speak about a statement of a given language being true or false. This is known among philosophers as the thesis of the ineffability of semantics: semantical properties like truth and falsity are undefinable in one’s language. A well known result in model theory, due to Alfred Tarski, shows that in order to express truth for the sentences of a given language, one has to resort to a richer metalanguage. I showed that IF languages can define their truth-predicate.