Abstract: Researchers have conducted lots of studies on the masked onset priming (MOP) effect in different languages such as French, Dutch and English. However, there are rare studies on this phenomenon in the field of Chinese at present. Two theories have been proposed to explain this phenomenon including a response competition hypothesis within a dual route framework and the speech-planning account. One of the main discrepancies of the two theories is the locus of the MOP effect: whether it arises during grapheme-to-phoneme conversion or at the level of phonological encoding of the speech response. Once pure phonological and orthographic effect can be separated from MOP effect, the locus where the effect arises would be clear: if only phonology was involved in MOP effect, MOP effect may arise at the level of phonological encoding; by contrast, MOP effect may arise during grapheme-to-phoneme conversion, if there was a orthographic effect, as well.
In experiment 1, types of the primes (phonemic, syllabled, or unrelated primes) and position of the (Shengmu or Yunmu) were used to test the pure phonological effect. The picture naming was facilitated by Shengmu much more than Yunmu, which indicated a pure phonological effect in MOP effect. What’s more, the different priming effects between different positions indicated that (1) phoneme instead of syllable is the least unit in phonological encoding; (2) a serial process was involved during phonological encoding; (3) the syllable is provisionally formed just before the speech production.
In experiment 2, the pure orthographic effect was tested by using left-component prime and right-component prime. It showed that right-component has more facilitation effect than left-component prime which is just contradict to the prediction of MOP effect, which indicated there was no orthographic factor in this effect.
Experiment 3 adopted picture-word interference paradigm to prove the finding of orthographic position effect in experiment 2. Result showed that right–components of interference words produced stronger prime effect than left-components. The superiority of right-component priming could be attributed to the fluency Chinese readers’ knowledge that left-components of Chinese characters indicate the meaning of while right- components indicate the pronunciation.
In short, MOP effect is a pure phonological effect. Thus the effect arises at the level of phonological encoding and the final results support the speech - planning account for MOP effect.

Key words: the masked onset priming effect, phonology, orthographic &, phonological encoding, orthographic effect