Abstract
Despite the longstanding belief in an inner voice, there is surprisingly little known about the perceptual features of that voice during text processing. This article asked whether readers infer nonlinguistic phonological features, such as speech rate, associated with a character's speech. Previous evidence for this type of auditory imagery has been found, but only when reading was preceded by an explicit auditory presentation of the character's voice (CitationAlexander & Nygaard, 2008; CitationKurby, Magliano, & Rapp, 2009). Three experiments included a fast condition in which the main character was described as speaking quickly and a slow condition in which the main character was described as speaking slowly. When reading aloud, participants in the slow condition read slower. However, when reading silently, participants in the slow condition read slower only when instructed to think about the character's voice. This work concludes that auditory images of speech rate information are not routinely formed during silent reading.
Notes
1Note that only subject analyses were computed for the non-dialogue lines, all of which were null. It would not have been meaningful to compute item analyses given that there were only two early non-dialogue lines and two late non-dialogue lines.
2Additional support for this conclusion comes from unpublished data from our lab in which we essentially replicated Experiment 2, but with the addition of a baseline condition in which the character was described as speaking neither quickly nor slowly. When reading aloud, participants in the baseline condition read the dialogue lines as quickly as participants in the fast condition. In contrast, reading times were slower in the slow condition.