Previous research has demonstrated that implicit prosody—the projection of prosody onto text during silent reading—influences parsing choices and agreement computation (Fodor, 2002; Kreiner, 2005). This study examines the effect of manipulating implicit prosody projection in light of several L2 processing models, comparing L1 and late L2 learners of English in comprehension and grammatical processing measures during reading. Participants silently read relative clause sentences such as The reporter who called the senators every once in a while writes awful stories for the newspaper, in one of three text presentation formats: (i) whole sentence, (ii) word-by-word, or (iii) phrase-by-phrase, rating each sentence for grammaticality, and responding to a comprehension probe. Under the assumption that (proficient) readers project appropriate prosodic contours onto text while reading, it was predicted that a word-by-word presentation would be disruptive to this process, while a phrase-by-phrase presentation would be facilitative. In comprehension measures, the word-by-word presentation disrupted performance in the L1 speakers, but facilitated performance in the L2 speakers; in grammaticality ratings, the phrase-by-phrase condition facilitated performance in the L1 speakers, but no effect was found for the L2 speakers. The L1 results support the prediction that readers do project prosodic contours onto text, and this projection can be disrupted or facilitated based on presentation format. The L2 results may best align with predictions from the 'Good Enough' Hypothesis (Ferreira, Bailey, & Ferraro, 2002; Ferreira, 2003), where comprehension is prioritized over grammatical analysis particularly under high task demand. A higher taskload may preclude the use of implicit prosody, requiring more resources to be allocated to the initial step of lexical access. For less proficient readers, this task is most facilitated by a slower word-by-word presentation, which allows better integration of the incoming materials.