Page 44 - Fall_DTF

Learning to Speak
movements and their associated acoustic consequences. The when Fiona was still 14 months old. Fiona seemed to work on
learning that occurs is the creation of linkages between the speech articulation in parallel (but not simultaneously) with
movements with the sensations: big movements are easy to her work on vocal modulation and prosody. She followed her
feel and make loud, hard sounds and small movements and own, highly individualized path.
whlsPers_make s_imihr,,so“nds' but theyjre qlliett In all Cases’ As a baby progresses through these stages of vocalizations, it
the baby ls leammg to shape the sound’ as Fiona Put ‘L is achieving greater mastery of the coordination of the struc
Frank Gueflther (etg-u T°“fVi]-le and G“eI1the1'- 2010) 1710' tures that articulate speech, with a presumed focus on learn-
P°5ed a t'°h“5t “e“’°C°ml7“tati°“at m°de1 0f 5l7eeCh 5°'1nd ing to independently control these structures. hi the earliest
lea-thing that eXPtath5 that this eXl71°'att°h and Ptactiee hY months of infancy, only the voice is controlled, but the gut-
the baby is an essential 5teP t°Wat'd achieving mathte 5PeeCh- tural sounds of the gooing stage add movement of the lower
Called Directions I.nto Velocities of Articulators (DIVA), jaw to bring the tongue to the roof ofthe mouth, which gener.
the m°de1 eXPtath5 that early listening he1P5 t° hufld midi‘ ates consonant-type sounds. Eventually, babbling coordinates
t°rY targets and that the 'ePetitlVe» CYCHC babbling Stthtgs movements of thejaw, lips, tongue, and voice, with increasing
are what 311°" the baby t° Create mal7Ph‘S5 th theh' 0W“ differentiation of these structures. Finally, in mature speech
brain about how to achieve these targets. These connections Pmdu;_—fi°n_ not just the sueucm,-es but discrete segments of
are the“ “Sed t° hell’ the hahy team What C°mhthati°“5 Of these structures must be controlled independently. Think, for
movements and 5e“5at-tons Pmdhce which 5°““d5 (tatget instance, of the way the tongue moves to produce a “t” versus
maps) and which movements and sensations do not achieve a “1€’ sound. hi the former, the front of the tongue is elevated
the target sound (error maps)- These target and error maps to shape the voice, whereas in the latter, the back of the tongue
are then used throughout the development of speech and is elevated (Edwards, 1992). This independent control, called
1a“S“aSe t° h“i-Id the ’ePet't°he °f Pt'°d“Ctt°“5 and eVe11 in differentiation, is particularly important for mature speech to
adulthood when these skills have fully matured. be Produced accurately ((;i|,|,°n_ 1999),
The small units, or syllables, produced during babble are
eventually layered over the pitch patterns of the native la.n- Doing Nlora with Leas
guage so that the babbling of a baby in the United States Only recently has technology allowed us to capture and
sounds “English,” but the babbling of a baby in Iapan sounds measure the earliest stages of speech motor control. In one
“Iapanese.” This reflects the interface between language and approach, Green and Nip (2010) quantified the movements
speech (Smith, 2006) and the multilevel layering of processes produced in early speech development using line-of-sight
and skills required to achieve mature speech production. For motion capture technology. Tiny reflective markers are
Fiona, the more complex babbling with alternating syllables placed On the lips. face, and jaw Of Children (Figure 3» 13f‘)
did not emerge until after her first birthday. In this video and the 3-dimensional motion of the sensors is captured by
from when she was 14 months old (see Multimedia File 5 special cameras and software. The software assembles the
at acousticstoday.org/mm-Vick), you can clearly hear her wild tracked sensors into a 3-dimensional model, as seen in Fig-
and very loud combinations of many different syllable shapes ure 3, right. This approach allows for analysis of the move-
and structures. Neither her pitch nor the loudness of her voice ments that contribute to the earliest vocalizations and speech
was modulated in this series and she was using these vocaliza- (e.g., Green et al., 2000). For a syllable like “ba,” produced
tions to work on coordination and control for the articulators. commonly in both babbling and early words, the child must
At the same stage, Fiona also produced vocalizations that bring “S bps ntgethen This can be aC_ComPhshed with a W-
. . riety of strategies. For instance, the lips could be held pas-
were communicative but that lacked consonants or the cy- _ _ _ _
. . . . . sive and the Jaw can bring them together. I.n adults, the jaw
clic repetition of babbling. She commonly used single syl- _ _ _ _ _ _
. . does some of the work, but it is handily assisted by bringing
lables that sounded like words but were composed of a sin- _ _ _ _
. . rr » . . the lower lip to the upper lip with independent movements
gle vowel, something like uh. With these productions, she _ _ _
. . . of these structures. The motion capture techniques provide
carefully modulated the pitch of her voice to communicate _ _ _ _
. . . . . . . . clarity about the contribution of each structure to lip closure
questions (rising pitch) and declaratives (rising then falling d h din _ fth
pitch), and she coordinated these vocalizations with facial and an measure t 2 Com anon 0 2 movements‘
hand gestures. This is illustrated in a sequence captured on For instance, the control of the jawand the lip control required
video (see Multimedia File 6 at acousticstoday.org/mm-Vick) to produce the “ba” sound develops gradually over the early
42 1 Acuulclcl Tisuay 1 Fiiizois