Neural correlates of semantic integration in Spanish verb-noun compounds: an ERP study

Abstract

It is assumed that semantic transparency in compounds depends on integration of constituents and head transparency. Spanish verb-noun compounds are semantically and morphologically exocentric, lacking a head. We studied if verb argument structure determines differences in semantic integration and morphological decomposition. During a lexical decision task, Agentives (referent takes the role of agent), and Metaphoricals (involving methaphorical patterns to construct meaning), were displayed as whole words or separated into constituents. Agentives were responded better and faster than Metaphoricals, with no interaction between compound type and presentation mode, suggesting constituent decomposition for both compounds. Early electrophysiological effects related to visual recognition, and an N400 component for Metaphoricals, suggest constituent semantic integration. Differences seemed related to combinatorial procedures: in Agentives the meaning of constituents is easily integrated to form compound meaning; in Metaphoricals, this procedure is inhibited, and compound meaning is accessed through the metaphoric relation that links both constituents.