Lee, On-Soon2021-01-132021-01-132011-03-01Lee, On-Soon. 2011. Forward and Backward Feature Agreement: Evidence from Korean Numeral Classifier Constructions. University of Hawai‘i at Mānoa Working Papers in Linguistics 42(1).http://hdl.handle.net/10125/73238This paper investigates the processing of semantic feature agreement between classifiers (specific vs. general) and associated nouns in two constructions, prenominal vs. postnominal. To examine this, two experiments were conducted. Results from an off-line judgment experiment show that phrases containing specific classifiers and associated nouns sound more natural than phrases containing a general classifier and associated nouns. This supports the hypothesis that the expectation of semantic information at specific classifiers helps to relatively facilitate the processing of associated NPs, compared to that at a general classifier. Results from the self-paced reading experiment show that semantically related information helps to reactivate preceding associated nouns in postnominal constructions, whereas it fails to facilitate processing of following associated nouns in prenominal constructions. Findings indicate that postnominal constructions are easier to comprehend because the expectation triggered by specific classifiers helps to reactivate the preceding nouns, compared to the condition with specific classifiers in prenominal constructions. This study supports the predictions of expectation-based theories.Creative Commons Attribution Non-Commercial Share Alike Licenselinguistics