Speakers are usually quite accurate and fluent when speaking, but we all make errors. While these speech errors have been a rich source of data on the mechanisms of language production, less is known about the consequences of these errors for our listeners. This is largely because most research on comprehension focuses on "ideal delivery" utterances; that is, comprehension of language that is error- and disfluency-free. I will discuss some work-in-progress investigating what happens to the syntactic parse of material that was erroneously produced and then corrected. In particular, I rely on the phenomenon of structural priming (where people tend to choose syntactic structures that are influenced by what they have previously comprehended) to see if listeners' subsequent production is influenced by errors that they have heard.