. Wstęp

In the thesis computational semantics for monadic quantifiers in natural language in the spirit of J. van Benthem’s works [van Benthem 1986] is considered. The possible correlations between quantifiers and procedures which can be used to compute quantifiers’ denotations in finite models are discussed. We prove M. Mostowski’s theorem that monadic quantifiers definable in the divisibility logic are exactly quantifiers recognized by finite state automata [M. Mostowski 1992]. We also discuss the case of more complex quantifiers recognized by push-down automata, like ”most”, ”the same number as”. Our results let us make hypothesis concerning human ability to interpret quantifiers. Referring to the works of P. Tichy [Tichy 1969] and Y. Moschovakis [Moschovakis 1990] we also consider the explication of the meaning (intension) of sentences in terms of algorithms computing their truth values in finite interpretations.