Poster #68 - Non-autistic versus autistic children: Why should I say “um” if I’m not talking to anybody?

• In Event: Poster Session 17

Abstract

When speakers use “um,” they may be conveying that they are pausing to plan an utterance (Clark & Fox Tree, 2002), and/or intending to take a turn in the exchange (Fischer, 2000). Research finds autistic individuals use “um” less often than non-autistic (NonAu) individuals when they answer questions during diagnostic testing (Gorman et al., 2016), describe how to play a sport (McGregor & Hadden, 2020), and describe pictures (Irvine et al., 2016). These authors argue that differences in “um” frequency are due to autistic individuals’ general pragmatic challenges. However, a recent study reported no difference in “um” usage between NonAu and autistic children during dyadic conversation (Boo et al., 2022). This suggests that differences in “um” use may not persist in all contexts. The current investigation compares “um” use by NonAu and autistic speakers across datasets that utilize different discourse elicitation tasks.

Data came from two sites (A and B): At Site A, 15 NonAu (Mage = 13.4, SD = 2.1) and 18 autistic (Mage = 13.5, SD = 2.3) children participated in two tasks – first, they completed the Trier Social Stress Test (TSST; Kirschbaum et al., 1993), where they had three-uninterrupted minutes to narrate a story in front of a panel of judges. Then, they conversed with an adult researcher, where they answered questions about their family and hobbies. At Site B, 22 NonAu children (Mage = 12.5, SD = 2.3) and 21 autistic children (Mage = 11.6, SD = 2.2) also engaged in conversation, but they interacted with virtual avatars that mimicked peers instead of an adult. In all three language samples, “um” tokens were tallied.

“Um” tokens did not differ between NonAu and autistic children in conversational tasks from Sites A or B, ps > 0.05. At Site A, NonAu children produced more “um” tokens than autistic children during the TSST task, t(38) = -2.097, p = 0.043. Significant cross-task differences also emerged (F(5, 110) = 3.886, p = 0.003, ηp2 = 0.150): both groups from Site A produced fewer “um” tokens in the TSST task than children from both sites who engaged in conversation, ps < 0.05 (Figure 1). No significant differences between the conversational tasks from Sites A and B emerged.

Findings suggest that “um” use differs depending on the nature of the discourse context (i.e., monologic vs. conversational). Interestingly, “um” use was statistically equivalent regardless of whom participants were speaking to, suggesting that discourse type (dyadic conversation vs. monologic storytelling), and not discourse partner, most affects “um” usage for both autistic and NonAu speakers. Further, the fact that autistic participants use “um” significantly more often during conversation suggests that they recognize its usefulness as a turn-taking device. This function of “um” is unnecessary in a context like TSST, where speakers are asked to talk for a specified amount of time without interruption. Overall, findings reveal pragmatic strengths for autistic speakers, including not only their use of “um” in the first place, but also their ability to vary their frequency of “um” use across contexts.

Authors

• Cynthia Boo, University of Connecticut
  Presenting Author

• Aaron Shield

• Nancy McIntyre, University of Central Florida

• Ruth B Grossman, Emerson College

• Peter Clive Mundy, University of California - Davis

• Letitia R Naigles, University of Connecticut

• Emily Zane, James Madison University