Missing and Superfluous Prosodic Accents in L2 English Focus Processing: Behavioral and ERP Evidence

Background/Objectives: Second-language (L2) listeners interpret prosodic prominence in relation to contextual-focus expectations. This study examined behavioral and ERP responses to English prosody–focus mismatch across two recruitment-defined proficiency strata, two focus positions, and two linked word-locked prominence configurations. Methods: Thirty participants completed an auditory question–answer appropriateness-judgment task while electroencephalography (EEG) was recorded. Prosody–focus matching and focus position were manipulated. Event-related potentials (ERPs) were analyzed separately at the contextually focused word and the background word. Behavioral responses were modeled with mixed-effects models and time-resolved ERP trajectories were analyzed with generalized additive mixed models. Results: Mismatch reduced accuracy at both focus positions, with a numerically larger reduction for sentence-final focus. Reaction-time mismatch effects were not reliable within either position and did not differ reliably across positions. At focused words, missing prominence elicited predominantly negative-going effects, with nominal Group × Matching interactions at both focus positions in the 300–550-ms range. At background words, superfluous prominence elicited predominantly positive-going effects, with a nominal Group × Matching interaction under sentence-final focus in the 550–1000-ms range. Direct window-level comparisons further showed that mismatch responses differed between focused and background words, with nominal Group × Word Role interactions, whereas no reliable overall modulation of this contrast by focus position was observed. Conclusions: Missing and superfluous prominence were associated with different ERP response patterns, whose interpretation requires consideration of both local accentuation and contextual focus. Evidence for proficiency-related modulation remained uncertain, and the observed sentence-final accuracy difference was sensitive to random-effects structure.