Page 42 - 2017Spring

Communication Across the Life Span
would be affected by the difficulty in interacting described above for older adults, for example. Indeed, in everyday life, speech is typically produced while communicating with an- other speaker, and our key aim is to ensure that the message that we are imparting to our interlocutor is understood so that communication can continue efficiently. We typically do this by adapting our speech dynamically throughout our interactions, producing more clearly articulated or “hyper- speech” when communicating in adverse conditions but resulting in less clearly articulated “hypospeech,” requiring less effort to produce when the message we are imparting is highly predictable (Lindblom, 1990). Most communication occurs at some point along this “hyper” to “hypo” continu- um, and the degree of effort used to produce speech changes dynamically according to the ongoing level of understand- ing of our interlocutor. We assess this level of understanding via the appropriateness of their responses, the frequency of requests for clarification, pauses, and hesitations. Recently, there has been a move toward investigating how talkers of different age groups adapt their speech in different commu- nicative conditions using problem-solving tasks involving dialogues between two individuals (for a review, see Cooke et al., 2014). These dialogues may still be far from natural communication because they are recorded in the laboratory and may involve talkers carrying out a specific problem- solving task to maintain some control over the content of the interaction, but they do provide information about speaker adaptations inherent to speech communication that cannot reliably be gleaned from read speech or spontaneous speech monologues.
One of the challenges of studies carried out with very dif- ferent age ranges is to find a task that is useable across a broad age range and that imposes a similar degree of cog- nitive load, as far as this can be ascertained. Some studies have used a “spot-the-difference” picture task, “diapix,” (Van Engen et al., 2010; Baker and Hazan, 2011) that involves pairs of talkers conversing to find differences between their pictures without seeing their partner’s picture. Others have used different interactive tasks such as Sudoku, the match- ing of complex shapes (tangrams), or tasks that involve one talker describing a trajectory on a map to another (map task). To investigate how individuals of different ages adapt their speech when communicating in adverse listening con- ditions, controlled disruptions to communication between two talkers, such as adding noise or spectrally distorting the speech of one or both talkers during their interactions, can be intro- duced (for a review of this type of work, see Cooke et al., 2014).
Figure 1. Conversational articulation rate based on data collected during “spot-the-difference” tasks (diapix) carried out between pairs of talkers. These data have been accumulated across studies carried out with children (reported in Hazan et al., 2016) and with young and older adults (reported in Tuomainen and Hazan, 2016). Circles denote outliers.
In a series of related studies with children aged 9 to 14 years, young adults, and older adults aged 65 to 85 years using a diapix task, trends for articulation rate (syllables produced per second) in conversational speech showed an inverted U shape: children up to the age of 11 years (Hazan et al., 2016) and older adults in the 65-85 year age range (Tuomainen and Hazan, 2016; Figure 1) spoke with a lower articulation rate than young adults. Normalized pitch range showed a simi- lar picture: 13-14 year olds and adults used a narrower pitch range in their conversational speech than both 9-12 year olds and 65-85 year olds. The change in mean fundamental frequency followed the expected trends in terms of talker sex and age (Figure 2; http://acousticstoday.org/hazan).
In those same diapix studies, when communication was made more difficult between conversational partners, the ad- aptations made by 9-14 year olds, and especially the younger group of 9-10 year olds, were quite consistent, with a general increase in vocal effort (shouting), while young adults varied more in the strategies that they used to make their speech clearer (Hazan et al., 2016). This suggests that 9-14 year olds were still developing a full range of these strategies that are essential for efficient communication. Ongoing work with older adults appears to be showing a similar trend, with changes also consistent with an increase in vocal effort and less evidence of reducing their articulation rate, which is a strategy typically used by younger adults (Tuomainen and Hazan, 2016).
40 | Acoustics Today | Spring 2017