Page 30 - Winter 2020

BABY TALK
adult and infant, presumably to ensure that infant offspring survive and thrive. Kalashnikova et al. (2017) claim that in early infancy, any benefits of IDS related to clarifying speech units are secondary to this basic social/emotional bonding goal, and, as in evolution, these linguistically motivated patterns likely emerge later in development and piggy-back on to this social bonding function.
Overall, what is happening to the resonance component of the speech signal when caregivers use IDS is not fully resolved. Caregivers may be modifying their speech to clarify speech units and boost language development, to convey positive emotions, to sound smaller and build social bonds, or some combination of these effects. Although the details remain unclear, understanding how these modifications impact infant development contin- ues to ignite and steer ongoing research.
Infant-Directed Speech and Infant Speech: An Important Connection?
As outlined in Vocal Resonance Properties of Infant- Directed Speech, there are different viewpoints regarding what motivates the use of IDS. One idea to emerge recently is that when mothers use IDS, they are altering their speech to sound smaller and more like an infant. Although this is a new perspective on IDS, the act of unconsciously adapting your speech to mirror or imi- tate features of your conversational partner is not a new observation. This has been noted and studied extensively in adult speech communication and is often referred to as phonetic convergence. Moreover, in adult-to-adult
interaction, speech convergence is typically associated with liking or holding a positive attitude toward your conversational partner (Pardo, 2013).
Other findings point to an important connection between IDS and infant speech. First, there are indeed clear paral- lels between IDS and infant speech. With respect to vocal source properties, infant speech and IDS have similar voice pitch values, particularly when IDS is produced by a female adult/mother. Figure 3 shows voice pitch values across the life span, including voice pitch values for IDS produced by female adults. Figure 3, pink box, highlights the range in which voice pitch values overlap across infant speech and speech produced by an adult female using IDS.
Although voice pitch values can overlap across IDS and infant speech, the vocal filter properties of infant speech and IDS are more distinct. When an adult female raises her larynx and spreads her lips to shorten her vocal tract length, she will sound like a smaller person. Neverthe- less, a mother cannot shorten her vocal tract enough to match the vocal tract length of her infant. Infant speech has much higher vocal resonances, reflected in the for- mant frequencies uniquely associated with a talker with a very short vocal tract. This results in higher formant frequency values for infant speech compared with adult speech which are shown by a spectrogram of the vowel "ee" produced by an infant and a female adult (Figure 4). These differences are also apparent in the vowel space shown in Figure 2, where you can see that the corner
Figure 3. Typical average voice pitch (f0) values for speakers across the life span. Blue lines, observed range of values observed within each group. Pink box, voice pitch range where infant and adult female IDS values overlap. Data from Masapollo et al., 2016, Table 1.
30 Acoustics Today • Spring 2021