From Semantic Underdetermination to Contextual Resolution: A Layered Framework for the Semantics–Pragmatics Interface in English

Abstract

The relation between linguistically encoded meaning and contextually interpreted content remains a central problem at the semantics–pragmatics interface. English utterances routinely communicate contents that are more determinate, contextually restricted, or conceptually adjusted than their overt lexical and syntactic forms alone appear to specify. The difficulty is not simply that context matters, but that competing theories assign context different representational roles: value assignment to encoded parameters, completion of underspecified structure, optional enrichment, lexical modulation, implicature, or selection among conventional senses. This article develops a critical conceptual analysis, supported by a concept-led review of foundational and recent scholarship, to clarify these operations and their evidential consequences. It proposes a Layered Contextual Resolution Framework (LCRF) comprising six analytical layers: encoded semantic content, linguistic under specification, contextual parameter resolution, pragmatic enrichment or modulation, inferential stabilization, and occasion meaning. The framework is applied to English cases involving predicative incompleteness, weather predicates, gradable adjectives, possessives, temporal interpretation, scalar strengthening, loose use, and polysemy. Its contribution is diagnostic rather than merely taxonomic: each layer is associated with distinct evidential questions concerning linguistic mandate, conventionality, contextual value assignment, conceptual adjustment, alternative structure, and processing. The analysis argues that context dependence is heterogeneous and that no single semantics–pragmatics boundary is likely to explain all cases. A constrained layered account preserves a substantive role for semantic structure while allowing pragmatically developed explicit content where independent evidence supports it. The article concludes with testable predictions for experimental, corpus, and cross-linguistic work and identifies conditions under which a phenomenon should be reclassified across layers as evidence changes.