Background: Three experiments were conducted to investigate how character transposition processing of two-character Chinese compound words (canonical and transposed words) at the early N1 stage and whether the patterns of character transposition processing differ between the left and right visual fields. Methods: A dual-target rapid serial visual presentation (RSVP) paradigm was adopted with event-related potentials (ERPs), in which T1 was presented at the center of the screen and T2 was presented randomly either at the center of the screen (0°), left visual field or right visual field across three eccentricities (Exp 1-fovea: 2° to the left or right of center; Exp 2-parafovea: 4° to the left or right of center; Exp 3-periphery: 6° to the left or right of center). Results: Behavioral results confirmed the same quality of lexical processing between the fovea and parafovea, but quality decreased abruptly in the periphery. ERP results showed that eccentricity could be a marginal predictor of the overall N1 amplitudes in regression analysis. More importantly, the early electrophysiological patterns of character transposition effects (a more negative N1 component for transposed words compared with canonical words) were only apparent in the right hemisphere (RH) across the three eccentricities. Conclusions: We concluded that character-order and characterposition coding of compound words occurred at the initial stage of visual processing. This right lateralization pattern further demonstrated a unique fine-grained character-level lexical coding ability of the RH