The discussion of the boundary between semantics and pragmatics has also undergone various changes of emphasis and style. In the 1970s, sense-generality and pragmatic inference were brought to the fore (see e.g., Cole 1981; Atlas 1989; Turner 1999; Jaszczolt 1999). Almost two decades later, dynamic perspective in semantics allowed for contextual information to be semanticized (see Kamp & Reyle 1993). Subsequent developments of the idea of underspecification (see e.g., van Deemter & Peters 1996) demonstrated that pragmatics, intentions and intentionality are frequently irreducible and do not yield to formalizations (see Blutner & van der Sandt 1998; van Deemter 1998; Dekker, forthcoming; Jaszczolt & Turner, forthcoming). The predominance of semantic analyses strongly suggests that (i) contrasting meaning in various natural languages requires firm foundations, strict modelling and some degree of formalization; (ii) both (a) cognitive semantics and (b) Tarskian, post-Montagovian semantics supplemented with post-Gricean pragmatics are more productive than the offshoots of the ordinary language philosophy. Finally, to address the empiricism-rationalism dilemma, it can be observed that inferring from quantitative analyses and supporting theories by unquantified data constitute equally successful directions in semantic and pragmatic research.