Abstract: Age of acquisition (AoA) is an important variable that has recently drawn considerable attention as a determinant of lexical processing. Several hypotheses have been advanced to explain the origin of the AoA effects. One of the hypotheses is the phonological completeness hypothesis, which assumes that AoA effects arise at the level of phonological representation (Brown & Watson, 1987; Gerhand & Barry, 1998). However, it had failed to explain the AoA effects in a phonological segment experiment (Monaghan & Ellis, 2002) and some experiments which phonological output was not necessary (Brysbaert et al., 2000). Another important hypothesis is semantic hypothesis, which assumes that at least parts of the AoA effect originate from the semantic system (Van Loon-Vervoon, 1989; Brysbaert et al., 2000). According to this hypothesis, the order of acquisition has a lasting effect on the time needed to activate the meanings of words. However, some studies have failed to support the claims of the semantic hypothesis (Morrison, Ellis, & Quinlan, 1992). In the present study, three experiments were conducted to examine the locus of the age of acquisition effects in the processing of Chinese and especially tested the semantic hypothesis of AoA effects.
Method
In Experiment 1, participants performed word naming and picture naming tasks. The stimuli comprised 36 single characters and 36 line drawings obtained from the Snodgrass and Vanderwart (1980) pictures. The 36 characters included 18 early-acquired and 18 late-acquired items. The 36 characters were also the corresponding Chinese names of the Snodgrass and Vanderwart (1980) .We aimed to compare the AoA effects between picture naming tasks and character naming tasks. To reduce the influence of the phonological output on the AoA effects in the picture naming task of Experiment 1, semantic category judgment task was used in Experiment 2, which required the subjects to judge whether the character belongs to the semantic category of “action”. There were 40 early-acquired and 40 late-acquired characters. In Experiment 3, picture classification task was used，which required the subjects to decide whether the object in the picture belongs to living things or non-living things. There were 28 pictures in each category, with 14 early-acquired and 14 late-acquired picture names. In all three experiments, the age-of-acquisition of words was rated by 30 subjects as Gilhooly and Logie (1980) did.
Results
Reaction times for correct responses and error rates were analyzed by subject and by item variance. In Experiment 1, there was an interaction between AoA and task, with substantial AoA effect in picture naming task (119ms), but no AoA effect in character naming task. In Experiment 2, the semantic category judgment task showed significant AoA effects (17ms). In Experiment 3, the significant AoA effects were also found in the task of picture classification of non-living things (20ms).
Conclusion
The results of the present study suggested that the AoA effects could be partially produced within the semantic processing system, rather than only within the speech output system

Key words: Age of acquisition, Chinese, Semantic hypothesis, Phonological completeness hypothesis