I. Introduction

Semantics investigates linguistic entities in their most fundamental dimension; it studies the relations of signs or strings of signs, their meanings, and their eventual references in the world. While linguistic semantics investigates concrete languages, in all their empirical details, it yet presupposes an intuitive understanding of the different kinds of signs and their functions: assertions, commands, and questions; names, predicates, relations; quantifiers, demonstratives, indexicals; and so on. As we all command at least one language, we have a pre-theoretical mastery of the linguistic instruments studied by theoretical analysis.

Philosophical semantics cannot rest content with such intuitions. Its proper task may be defined as the theoretical clarification of different kinds of signs and their functions: how signs in general have meanings and, eventually, references; how they interlock to form meaningful complexes and how their functions differ; what gives them the status of names, predicates, relations; what structurally guarantees that we understand certain strings of signs as assertions, questions, commands; and so on.

Indeed, the fundamental problem of semantics may be to explain how subsentential semantic entities combine to form the basic self-sustained semantic unit, an assertion, and to describe what these subsentential entities must be like. In a sense, this was the main occupation of Frege in his attempt to lay foundations for philosophical semantics. In recent decades, philosophers have at the same time found Frege’s compositional approach seriously wanting and occupied themselves more globally with meaning—all too often with skeptical results. I believe that Frege’s program of clarifying the status of subsentential semantic bits in regard to their function to form full-fledged assertions is a promising one and that it deserves to be refurbished and continued.

The purpose of the present reflections is to show how the analytic-synthetic distinction, as set out by Kant, can be exploited to give answers to some of the Fregean questions in philosophical semantics. Of course, I do not mean we should interpret Kant as an early philosopher of language; if we should list those topics of present-day philosophical concern ignored by Kant, language certainly would be included. However, pioneers of modern philosophy of language, such as Frege, Carnap, and Quine, have occupied themselves with the analytic-synthetic distinction, and invariably have addressed Kant’s influential approach.