Although it has generally thought that the duration of saccades should be subtracted from the reading time in eye movement research, Irwin (1998) has demonstrated that lexical processing such as word recognition is not suppressed during saccades, and has thus called that conventional wisdom into question (Murray, 2000; Radach & Kennedy, 2004; Rayner, 1998; Vonk & Cozijn, 2003). In Irwin’s experiments, however, the subjects were merely reading isolated words, not sentences. It is therefore not clear what we can legitimately infer from the results of these experiments about what happens during saccades when people are reading sentences. In this thesis, I report the result of an eye-tracking experiment which tests whether cognitive processing for reading continues during saccades when people are reading sentences. The subjects in the experiment read the first half of a sentence, and then made either a short or a long saccade, before going on to read the rest of the sentence; in half of the experimental sentences, a low-frequency word was used at the location immediately preceding the saccade, so that some cognitive processing regarding that word would “spill over,” that is, continue even after the eyes have left the word. The result of the experiment showed that the spill-over effect, that is, the difference between the effect that a high-frequency word has on the processing of the ensuing material and the effect that low-frequency word has on the processing of the ensuing material, was smaller after a longer saccade. This result indicates that linguistic processing is not completely suppressed during saccades when people are reading sentences.