Participants in psycholinguistic experiments are typically asked to read or listen to dozens of individual, uncontextualized, disconnected sentences, ranging from perfectly normal to extremely odd in one or more respects. In this chapter, we consider the possibility that a nontrivial number of participants are not terribly motivated to process such sentences in a “typical” manner. Furthermore, this lack of motivation may well result in data that are not reflective of normal human language processing because it is not reflective of everyday language use. Contrary to the idea that language processing is automatic, we speculate that the depth and accuracy of language processing and comprehension can vary as a function of task and comprehenders' motivation to perform the task, and therefore that “typical” processing is variable. We argue that both online and offline measures of processing and comprehension are susceptible to this variability, and we offer some thoughts and suggestions about experimental methods that may help to reduce or manipulate it in order to collect data that are more faithful to both the linguistic input and language processing outside of the laboratory.