There is a growing consensus in psycholinguistics that humans actively predict during language comprehension. There is evidence that we use a variety of different kinds of linguistic information to anticipate things like what kinds of words are likely to come next in a sentence, what kind of sentence structures are likely, and even fine-grained details about upcoming words themselves. What is not clear, however, is when prediction does and does not occur. Do we only do it in certain circumstances? What kinds of linguistic information do we use to generate expectations? Does everyone predict to the same degree in the same circumstances? I am conducting research geared toward answering some of these questions. In particular, I am examining when, if at all, various types of linguistic cues are mapped on to a specific prediction; I am also exploring what factors predict the extent to which adult second language learners anticipate upcoming words in sentences.