ABSTRACT
Self-paced reading (SPR), or the moving window technique (MWT) (Just, Carpenter & Woolley, 1982), is one of a number of psycholinguistic methods that has been developed in order to tap into the moment-by-moment (or on-line) processing of language input. During a SPR experiment, words (or phrases) appear on a computer screen, usually in the center, one at a time, and once the participant has read the input, s/he pushes a button to bring up the next constituent, which most often replaces the last, until the sentence has been read to the end. The researchers thus obtain the participant’s reading time (RT) profile for each experimental item, with a task that allows them to isolate the particular word or phrase that may be of theoretical interest: RTs are compared across conditions, with processing slowdowns being of most interest. Readers may slow down for a number of reasons, of course, but the assumption is that some part of the input has caused processing difficulty, for example, if the word does not fit semantically with the current analysis or if it is unexpected or ungrammatical: Researchers must therefore infer the underlying cognitive processes from careful manipulation of the experimental materials (this is discussed in more detail in the following sections). It is common to have a secondary task with SPR. To keep participants ‘on track,’ a comprehension question or a truth value judgment might follow all or some of the items, or perhaps readers are asked to make metalinguistic judgments, assessing the sentence’s acceptability or grammaticality at the end of each item, for instance. It should be cautioned that each task type may influence the way readers process the input in different ways (more details on this point will be raised in the sections below).
