This paper reexamines the question of whether logic is analytic or synthetic, arguing that the issue has transformed rather than vanished. I propose that model-theoretic logic, when interpreted through interpretational semantics, can be regarded as analytic, and that logical necessity can be derived from this analyticity. Revisiting Etchemendy’s distinction between interpretational and representational semantics, I argue that it mirrors the analytic–synthetic divide: the former aligns with truths grounded in meaning, while the latter corresponds to truths dependent on the structure of possible worlds. Etchemendy’s claim that interpretational semantics cannot secure necessity thus challenges the traditional view that logical necessity is explainable through analyticity. In contrast, Gila Sher’s Formality Thesis grounds logic in formal features of the world, presenting logic as synthetic yet necessary. I interpret her view as a reformulation of the analytic/synthetic problem within a formal-structural framework. The paper concludes by defending interpretational semantics and suggesting that the analyticity of logic, properly understood through truth-conditions, can indeed account for logical necessity.
This paper addresses Quine’s critique of the analytic-synthetic distinction from Two Dogmas of Empiricism, where he argues that analyticity—truth by meaning alone—cannot be coherently defined without circular reasoning. Quine contends that defining analyticity in terms of necessity leads to circularity because necessity itself presupposes analyticity. In response, this paper uses Kripkean semantics to define analyticity as “global necessity”, or truth in all points of evaluation, thus countering Quine’s circularity objection. The proposed approach relies on Kripke’s modal logic framework, where analyticity can be anchored in the relational structure of possible worlds. By defining analytic statements as those true in all worlds, this proposal allows analyticity to be coherently applied across a family of logical systems, namely the family of normal modal logics. Comparing this approach with Rudolf Carnap’s definition of analyticity from Meaning and Necessity, the paper highlights that Kripkean semantics offers greater flexibility by allowing the same definition of analyticity to be applied across a whole range of frameworks. Thus, while analytic truths become framework-dependent, they preserve their value within formal systems, as well as their conventional character—which was first recognized by Carnap. After discussing a possible application of this work to current debates on topic-neutrality of logic and logical pluralism, the paper concludes that the proposed account of analyticity provides a counterpoint to Quine’s dismissal, supporting the analytic- synthetic distinction as both meaningful and practically useful in logical analyses.
This article proposes a reconstruction of the reason why the Marburg School of Neo-Kantianism seemingly disregarded the analytic-/synthetic-distinction altogether: Members of this school regarded incessant conceptual change as a necessary feature of human knowledge. When concepts change, the analytic-/synthetic-distinction shifts accordingly. Since Hermann Cohen (and his students) found no use for such an unstable distinction, they left it out of consideration entirely. As a paradigm example I discuss Hermann Cohen's view in detail, then compare it to Natorp and Cassirer. Finally I point out that there is a striking contrast on this point between the Marburg and the South-Western School of Neo-Kantianism, which correlates with their embrace of realism and idealism respectively.
This paper aims to present a largely overlooked chapter in the debate concerning the analytic-synthetic distinction (ASD). London’s intellectual trajectory was distinctive: despite his enduring ambition to become a physicist, he began with a philosophical dissertation exploring the development of deductive theory, drawing inspiration from Husserl’s logical investigations. This framework also informed his first scientific paper, Die Bedingungen der Möglichkeit einer Maßbestimmung in einer physikalischen Mannigfaltigkeit und das Prinzip der Ähnlichkeit, which will be the primary focus of our analysis. In summary, we argue that London’s reflections offer valuable insights into a potential approach that removes the ASD from its usual context, which entails a firm distinction between analytic and synthetic judgements. We shall demonstrate that, while rejecting a strictly Kantian conception of the synthetic a priori, London’s defence of the “autonomy of ideality” does not result in a purely analytic theory of judgments. Rather, echoing Husserl and Cassirer, his proposal aligns more closely with a category that we may term “infra-conceptuality,” which integrates both the conceptual containment of the analytic and the fecundity and productivity of the synthetic. However, we suggest that London’s emphasis on “conventions” and the relative nature of deductive theory would necessitate a reassessment of his relationship with neo-empiricist thought.
Strongly analytic assertions (or deductions) are known to be true (or valid) solely by virtue of knowledge of their meaning and are therefore unrevisable. Modestly analytic assertions (or deductions) are recognized as correct in a language solely by virtue of a knowledge of their meaning. Despite Quine’s effective critique of analyticity, philosophers still debate whether there are analytic assertions or deductions. The answer depends on the conception of meaning adopted. According to an inferentialist conception of meaning (inferentialism), to understand a sentence in a language is to apprehend a set of possible meaning-constitutive inferences and assertions in which the sentence occurs. Paul Boghossian proposed a brand of inferentialism in his attempt to salvage strong analyticity. Boghossian’s version of inferentialism faces serious objections, some raised by Timothy Williamson. The main thesis of this paper is that in epistemic contexts there are modestly analytic assertions and inferences. To support this thesis I outline a non-holistic, fallibilist inferentialism in which there are no strongly analytic assertions, but the meaning-constitutive assertions and deductions are modestly analytic.
