Recent evidence suggests that individuals generate written words based on both spelling and sound. The present study used event-related potentials (ERPs) to examine the relative time course of orthographic and phonological activation. We adopted Chinese as a target language in which spelling and sound are largely dissociated. Native speakers of Chinese Mandarin were presented with coloured pictures and wrote down colour and picture names as adjective-noun phrases. Colour and picture names were either phonologically related, orthographically related, or unrelated. EEG revealed phonological effects in the 200-500 ms time window, starting at 206 ms after picture onset, and orthographic effects in the 300-400 ms time window, starting at 298 ms. The results of our study suggest that activation of phonological codes takes place approximately 100 ms earlier than access to orthographic codes, which provides evidence for phonological encoding as early sources of constraint in written word production.