Linking verbal transformations to their causes.

Abstract

The verbal transformation effect (VTE) is a perceptual phenomenon in which listeners report hearing illusory utterances when a spoken word is rapidly repeated for an extended period of time. The cause of the illusion was investigated by identifying regularities across the transformations that listeners reported and then testing hypotheses about the cause of those regularities. Variants of the standard transformation paradigm were used across 3 experiments to demonstrate that perceptual regrouping of the elements in the repeating utterance is 1 cause of the VTE. Findings also suggest that regrouping is influenced by whether the stimulus is perceived as speech or as nonspeech.Within the information-processing approach to perception, one method of inquiry is to investigate the causes of illusions, which are viewed as windows into the processes that operate during veridical perception. That is, illusions are temporary lesions (Warren, 1968) of the perceptual system that reveal its inner workings.

Stream segregation of concurrent speech and the verbal transformation effect: Influence of fundamental frequency and lateralization cues

“…The results support the idea that perceptual segregation and regrouping contribute to the VTE (Pitt and Shoaf, 2002;Stachurski et al, 2015). Specifically, when two sequences are present in both ears, the greater opportunity this affords for the regrouping of acoustic-phonetic elements typically increases the number of forms heard, speeds up the time to the first VT, and reduces the dwell time of the initial form.…”

“…The delay for the ITD condition was accommodated within the terminal silence, so that the tokens used in all three conditions were precisely 551 ms long, corresponding to 327 repetitions per sequence in 3 min. In line with previous studies (e.g., Warren and Ackroff, 1976;Pitt and Shoaf, 2002), the interval between sequentially presented tokens was kept small to facilitate maximum resegmentation and perceptual regrouping of acoustic-phonetic elements within the stimuli presented. An additional word, "train", was recorded and transformed in the same way as described above for the experimental stimuli.…”

“…The two versions were played with a halfcycle offset (i.e., half the duration of the token), in order to prevent across-sequence fusion of the identical parts of the tokens (i.e., the aperiodic segments). Each sequence was 3 min long; this choice was based on the observations of Pitt and Shoaf (2002), who found that listeners tended to stop reporting changes after that time owing to fatigue. On each trial, listeners were asked to monitor both sequences and to indicate throughout how they perceived each one of them.…”

“…However, it is only since the millennium that the relationship between the VTE and cues for auditory stream segregation (Bregman and Campbell, 1971;Bregman, 1990) has been explored in any detail. Pitt and Shoaf (2002) showed that the verbal transformations experienced by listeners are related to the acoustic cues that help bind together the rapidly changing and diverse sounds of speech (see, e.g., Darwin, 2008). In particular, acoustic-phonetic elements that are periodic and have a low-frequency centroid, such as nasals, cohere better with neighbouring voiced vowels than do acoustic-phonetic elements that are aperiodic and have a high-frequency centroid, such as unvoiced fricatives, affricates, and plosives.…”

“…Our current understanding of the relationship between the VTE and auditory stream segregation, which is based on studies using single sequences (e.g., Pitt and Shoaf, 2002;Stachurski et al, 2015), suggests that stimulus configurations that increase the possibilities for perceptual regrouping of acoustic-phonetic elements should facilitate the VTE. Furthermore, it also suggests an alternative or additional explanation for the independent verbal transformations for two concurrent sequences of identical stimuli observed by Warren and Ackroff (1976).…”