Abstract: The individual classifier, a kind of determiner, is a unique grammatical class of Chinese words. There are two theories explaining the lexical access of determiners: direct selection hypothesis and indirect selection hypothesis. Direct selection hypothesis assumes that the selection of determiners is an automatic process, directly retrieved syntactic or semantic information from their own nodes. The indirect selection hypothesis asserts that the selection of determiners is a competitive process, indirectly retrieved syntactic or semantic information from their headwords. This study focused on the lexical access of individual classifiers in language production and comprehension. As the individual classifiers are a kind of the closest determiners of nouns, some of them even have fixed relationships with nouns.
Three experiments investigated the lexical access of individual classifiers during the production of noun phrases, the production of bare nouns, and the comprehension of noun phrases. In Experiment 1 (N = 31), a picture-word interference paradigm was conducted to study the selection of individual classifiers in NP production. On the one hand, utterance onset latencies in NP production were longer when the distractors and the target pictures had different individual classifiers than when they had the same individual classifiers; on the other hand, utterance onset latencies were found longer when the distractors were semantically related to the targets than those were not. It indicated that the picture naming task was inhibited by the semantic relationships between targets and distractors but facilitated by the consistency of their classifiers. In addition, this facilitation can be found even when the SOA was set from 0ms till 100ms. Such a result indicated that the classifiers can be activated in the early stage during NP production and can persist for a short time.
In Experiment 2 (N = 25), the picture-word interference paradigm was again conducted to study the activation of individual classifiers in noun production. It showed the same tendencies as that achieved during bare noun production.
In Experiment 3 (N = 25), the priming paradigm was used with 2(types of noun phrase) × 2(types of classifiers) × 3(SOA) within-subjects design to investigate the lexical access in language comprehension. The results of experiment 3 showed that the reaction time of unmatched classifier noun phrased was longer than the matched ones. It indicated that access of the priming words can activate not only the targets but also the information of classifiers.
According to the achieved results, a theory called “dual-systems and dual-election combination network” of classifier lexical access was proposed. That is, there are two systems in the mental lexicon of Chinese speakers: the system of individual nouns and the system of individual classifiers. Because of the parsimony law in mental process, the indirect route would be chosen under most occasions. This study supplemented the theory of Chinese lexical access theory, and suggested that teaching classifiers with nouns together can expedite learning and help identifying/recalling or using individual classifiers.

Key words: classifiers, lexical access, classifier congruency effect