An Investigation into the Obligatory Decomposition of Opaque Morphemes and Pseudomorphemes in Chinese

Abstract

The obligatory decomposition model of morphological processing is a robustly attested hypothesis that proposes that the language processing system automatically decomposes all input into its morphological constituents. However, it is unclear if and how this decomposition occurs for input that is semantically opaque: that is, input for which the constituent morphemes do not contribute to the meaning of the whole. A further point of uncertainty is if the system distinguishes between semantically opaque lexical items, and pseudo-derived/pseudocompound items, which on the surface seem to have an opaque morphological structure but have an underlyingly monomorphemic lexical representation. The present study addressed these questions by conducting a priming experiment on Chinese archaic idioms (which are semantically opaque, as well contain archaic grammar structures) and loan words (which are pseudocompounds). The experiment used a lexical decision task paradigm to investigate if the morphological constituents of the archaic idiom and loan word items are decomposed and semantically activated during processing. Archaic idiom primes that were semantically related to a target did yield priming effects, which is consistent with the hypothesis that Obligatory Decomposition occurs in spite of both semantic opacity and archaic morphosyntactic structure. Loan words also yielded priming effects, but this could not be connected to the question of semantic opacity because the priming effect was found regardless of if the loan word prime was semantically related to the target. From this we conclude that semantically opaque lexical items are obligatorily decomposed during processing, but more research is required to understand if the same is true for pseudocompounds.