UC San Diego

In this dissertation, I describe a series of experiments designed to answer a number of questions about parafoveal pre-processing : (1) How many words can readers pre- process parafoveally at the same time? (2) How do readers parafoveally pre-process unspaced compound words? (3) What role does parafoveal pre-processing play in word skipping? In the first set of studies, Experiments 2.1 and 2.2 (Angele & Rayner, 2011), the availability of both the first word (n+1) and the second word (n+2) to the right of the pre-boundary target word (word n) were manipulated orthogonally. Additionally, in Experiment 2.2, the word frequency of word n was manipulated. The results show that availability of n+2 preview had no effect on fixation times on either word n, n+1, or n+2. In the second set of studies, Experiments 3.1 and 3.2 (Angele & Rayner, 2013a), both the order of morphemes in the preview for a compound target word (cowboy vs. boycow) and, in Experiment 3.2, the identity of the letters comprising the constituent morphemes were manipulated (e.g. cowboy vs. boyenz). The results show that (1) subjects obtained less preview benefit from a reversed order preview (e.g. boycow) than from an identical preview (e.g. cowboy) and that (2) a correct preview for one of the morphemes led to a slight preview benefit even when the morpheme order was reversed. This suggests that, at least during early processing, compound words are processed as a whole. In the third study, Experiment 4 (Angele & Rayner, 2013b), the preview that readers received for a three-letter target verb (e.g. eat) was manipulated, so that the upcoming word appeared to be (1) the definite article the (although in an infelicitous position), (2) a random letter string (cxf ), or (3) identical to the target verb. The results show that readers are likely to skip the even when it does not fit into the sentence context. In summary, the results of the experiments performed as part of this dissertation show that parafoveal pre-processing is, in general, rather shallow, involves only one parafoveal word, and is restricted to the orthographic level