Comprehension of English for‐adverbials: The Nature of Lexical Meanings and the Neurocognitive Architecture of Language

Abstract

AbstractWhat is the nature of lexical meanings such that they can both compose with others and also appear boundless? We investigate this question by examining the compositional properties of for‐time adverbial as in “Ana jumped for an hour.” At issue is the source of the associated iterative reading which lacks overt morphophonological support, yet, the iteration is not disconnected from the lexical meanings in the sentence. This suggests an analysis whereby the iterative reading is the result of the interaction between lexical meanings under a specific compositional configuration. We test the predictions of two competing accounts: Mismatch‐and‐Repair and Partition‐Measure. They differ in their assumptions about lexical meanings: assumptions that have implications for the possible compositional mechanisms that each can invoke. Mismatch‐and‐Repair assumes that lexical meaning representations are discrete, separate from the conceptual system from which they originally emerged and brought into sentence meaning through syntactic composition. Partition‐Measure assumes that lexical meanings are contextually salient conceptual structures substantially indistinguishable from the conceptual system that they inhabit. During comprehension, lexical meanings construe a conceptual representation, in parallel, morphosyntactic and morphophonological composition as determined by the lexical items involved in the sentence. Whereas both hypotheses capture the observed cost in the punctual predicate plus for‐time adverbial composition (e.g., jump (vs. swim) for an hour), their predictions differ regarding iteration with durative predicates; for example, swim for a year (vs. for an hour). Mismatch‐and‐Repair predicts contrasting processing profiles and nonoverlapping activation patterns along punctuality differences. Partition‐Measure predicts overlapping processing and cortical distribution profiles, along the presence of iterativity. Results from a self‐paced reading and an functional Magnetic Resonance Imaging (fMRI) studies bear out the predictions of the Partition‐Measure account, supporting a view of linguistic meaning composition in line with an architecture of language whereby combinatoriality and generativity are distributed, carried out in parallel across linguistic and nonlinguistic subsystems.