Humans imitate each other during social interaction. This imitative behavior streamlines social interaction and aids in learning to replicate actions. However, the effect of imitation on action comprehension is unclear. This study investigated whether vocal imitation of an unfamiliar accent improved spoken-language comprehension. Following a pretraining accent comprehension test, participants were assigned to one of six groups. The baseline group received no training, but participants in the other five groups listened to accented sentences, listened to and repeated accented sentences in their own accent, listened to and transcribed accented sentences, listened to and imitated accented sentences, or listened to and imitated accented sentences without being able to hear their own vocalizations. Posttraining measures showed that accent comprehension was most improved for participants who imitated the speaker’s accent. These results show that imitation may aid in streamlining interaction by improving spoken-language comprehension under adverse listening condition
Humans often imitate each other in social interaction (Chen, Chartrand, Lee-Chai, & Bargh, 1998). These imitative actions cover a wide range of behaviors, including manual gestures, body postures, facial expressions, mannerisms, and speech patterns (Chartrand & Bargh, 1999). Imitative behavior in humans streamlines social interaction by increasing affiliation and empathy between interaction partners (La France, 1979). In addition, imitative behavior aids in vicarious learning (Bandura, 1977; Mattar & Gribble, 2005), which occurs as a function of observing, processing, and replicating the actions of other people. The present study was designed to investigate whether vocal imitation can improve comprehension of lan-guage spoken in an unfamiliar accent.
The motivation for this study comes from theoretical approaches that propose that action comprehension activates internal cognitive mechanisms also used in action execution (Bandura, 1986; Brass, Wohlsläger, Bekkering, & Prinz, 2000; Meltzoff & Moore, 1977; Meltzoff & Prinz, 2002; Wilson & Knoblich, 2005), and that action execution can, in turn, improve action understanding (Hecht, Vogt, & Prinz, 2001; Prinz, 1992, 1997). Moreover, it has been suggested that imi-tating other people’s actions may make it easier to predict their subsequent actions, especially when the meaning they are trying to convey is ambiguous or distorted (Kappes, Baumgaertner, Peschke, & Ziegler, 2009; Pickering & Garrod, 2007; Wilson & Knoblich, 2005).
The present study used a novel speech accent to test whether imitation of unfamiliar actions indeed improves understanding of these actions. Imitation of accented speech is especially suited for testing this hypothesis, for two reasons. First, accented speech contains phonetic and phonological variations (Adank, van Hout, & Van de Velde, 2007; Best, McRoberts, & Goodell, 2001) that may lead to ambiguities and other distor-tions that listeners must resolve during spoken-language understanding. For instance, Japanese learners of English have difficulty in producing a distinction between the vowels in slip and sleep, as these vowels are not contrastive in Japanese. This ambiguity in pronunciation may make it difficult for native English listeners to determine which word is being spoken. Consequently, comprehension of an unfamiliar accent is reflected in slower and less efficient processing compared with comprehension of a familiar accent (Adank, Evans, Stuart-Smith, & Scott, 2009; Floccia, Goslin, Girard, & Konopczynski, 2006).
The second reason that speech with an unfamiliar accent is suited for testing our hypothesis is that vocal imitation of pho-netic and phonological variation is already commonplace in everyday life. For instance, people engaged in dialogue spon-taneously imitate each others’ intonation patterns (Goldinger, 1998), clarity of speech (Lakin & Chartrand, 2003), speech rate (Giles, Coupland, & Coupland, 1992), regional accent (Delvaux & Soquet, 2007), and style of speech (Kappes et al., 2009). If imitation of an ambiguous action improves subsequent understanding of that action (Kappes et al., 2009; Pickering & Garrod, 2007; Wilson & Knoblich, 2005), then imitating an unfamiliar accent may subsequently improve comprehension of utterances spoken in that accent; that is, imitation may allow listeners to better anticipate the phonetic and phonological variation in accented speech.
Our study evaluated the effect of several forms of training on the comprehension of accented speech. Listeners heard sentences spoken in an unfamiliar accent of Dutch. This unfa-miliar accent was obtained by systematically altering the pro-nunciation of vowels in stressed lexical positions, with the aim of creating a nonexistent accent of Dutch. Using such a novel accent ensured that all listeners were equally unfamiliar with the speech we used in the study. This approach was necessary because relative familiarity with an accent affects the process-ing of accented speech (Adank et al., 2009; Floccia et al., 2006). Comprehension of this novel accent, which the first author has used in a previous study, has been found to resem-ble comprehension of existing regional and foreign accents (Adank & Janse, 2010).
Following a pretest that measured their accent comprehen-sion, listeners were split into six groups. The baseline group received no training. Participants in the other groups heard one-hundred sentences spoken in the unfamiliar accent. In the listening group, participants were instructed to merely think about the sentences they heard. In the repeating group, participants repeated the sentences in their own accent (i.e., without imitat-ing the unfamiliar accent). The transcription group wrote a semiphonetic transcription of the pronunciation aspects of the sentences. The imitation group verbally imitated the exact pro-nunciation of the sentences, and the imitation-plus-noise group listened to the sentences without background noise and then imitated the sentences while their own vocalizations were masked with noise. Finally, all listeners were tested again on their comprehension of the unfamiliar accent, using the same procedure as in the pretest. The duration of the whole experi-ment was approximately 25 min for the baseline group and 35 min for the other groups.
The baseline group was included in the study to ascertain whether training in itself would lead to posttest improvement. The listening group was included to ensure that mere addi-tional auditory exposure to the unfamiliar accent could not explain posttest improvement. The repeating group was included to test whether the motor act of speaking influenced posttest performance. The transcription group was included to ensure that paying attention to phonetic and phonological aspects of the accented sentence could not explain posttest improvement. The imitation group was included to test whether imitation of the unfamiliar accent resulted in a posttest improvement in comprehension of the unfamiliar accent. The imitation-plus-noise group was included to investigate whether improvement in the imitation group was due to imitation per se or to auditory feedback from the participants’ own voices while imitating the recordings. We reasoned that if comprehension improved more in the imitation and imitation-plus-noise groups than in the other groups, this would support the hypothesis that imitation improves understanding of ambiguous or distorted actions.