## Video Presentation ## Check out our **3-min video**: https://youtu.be/lVcssZw5FvY ---------- Questions? Comments? Let's chat on Zoom! ---------------------------------------- **Zoom:** We will be available over Zoom during **poster session 2**: https://harvard.zoom.us/j/98466004667 **Email:** Feel free to email me (Yingtong Liu) for any quesions: y_liu@g.harvard.edu ---------- Abstract -------- Comprehenders must recover a speaker’s intended meaning from linguistic input that may be corrupted by noise. Previous studies have demonstrated that comprehenders do so by considering the prior probability of the intended *meaning* (Gibson et al., 2013; Ryskin et al., 2018) , but the noisy-channel framework also predicts that the ***form*** of the sentence matters. Low-frequency structures should bias the listener to infer more plausible meanings/structures. To test this hypothesis, we examined how comprehenders interpret events of varying plausibility formed in word orders of varying frequencies in English and Mandarin. Our results showed that speakers of both languages drew more inference for less frequent word orders. Thus, structural frequency does affect comprehenders’ noisy-channel inference.