Our long-term goal is to contribute to the integration of formal and lexical semantics. Our more immediate theoretical starting point is the idea of “text as theory”, within a model-theoretic semantic framework. We describe a set of empirical problems in the domain of genitive modifiers that offers a challenge to theories of the integration of lexical, compositional, and contextual information. After sketching a solution, we raise the issue of metonymy in the interpretation of genitives, and examine the role of sortal information in the specification of underspecified meanings and in processes of type-shifting and sort-shifting, including metonymy.

This paper investigates the status of the lexicon and the possibilities for expressing lexical generalizations in the paradigm of Head-Driven Phrase Structure Grammar (HPSG). We illustrate that the architecture readily supports the use of implicational principles to express generalizations over a class of word objects. A second kind of lexical generalizations expressing relations between classes of words is often expressed in terms of lexical rules. We show how lexical rules can be integrated into the formal setup for HPSG developed by King (1989, 1994), investigate a lexical rule specification language allowing the linguist to only specify those properties which are supposed to differ between the related classes, and define how this lexical rule specification language is interpreted. We thereby provide a formalization of lexical rules as used in HPSG.

Following the spirit of Relativized Extreme Functionalism, I argue that a set of grammatical features, traditionally thought of as devoid of semantics, lexical class (declension and conjugation), in fact has semantic content. Taking Josefsson (1995. 1997. 1998) as a point of departure, I suggest that the lexical class determines the word class of a word, hence relating the word to a major ontological category such as THING and EVENT. A certain lexical class may correspond to a semantic subclass of a major ontological category, but this does not need to be the case. The approach taken explains certain morphological phenomena in Swedish, such as stem vowels, and explains why the non-head part of a compound is undetermined for major ontological category.

Based on a functional approach, the article proposes a role for lexical knowledge in human languages in relation to syntactic and encyclopaedic knowledge. A lexicon presupposes encyclopaedic knowledge in terms of which the semantic domain of lexical items can be defined – but this does not mean that there is no distinction between lexicon and encyclopaedia, only that one stands on the shoulders of the other. Syntax similarly presupposes a lexicon: there can be no combinations without items to be combined, whereas you can have (holophrastic) languages consisting solely of items. However, inside the domain of human, i.e. syntactically organized languages, syntax and lexicon presuppose each other: lexical items below full utterance size make no sense except in relation to a combinatory syntax, and a combinatory syntax presupposes elements that can enter into combinations.