Content uploaded by Kalim Gonzales
Author content
All content in this area was uploaded by Kalim Gonzales on Sep 11, 2019
Content may be subject to copyright.
12-month-olds use voice and temporal
cues to extract grammatical structure
Kalim Gonzales, Rebecca Gómez, & LouAnn Gerken
University of Arizona
Aggregating
Children’s tolerance for
inconsistency
•Children sometimes show sensitivity to
grammar patterns in highly inconsistent
input
–Hudson Kam & Newport (2005; 2009)
•But not always
–Singleton & Newport (2004); Ross & Newport
(in prep)
–Gomez & Lakusta (2004)
Gomez & Lakusta (2004)
Structure 1 Structure 2
S { aX
bY }
S { bX
aY }
|a, b| = 2; |X, Y| = 6
Gomez & Lakusta (2004)
Aggregating
Segregating
Segregation by voice?
•newborns can distinguish between their own and
another mother’s voice (de Casper & Fifer,
1980).
•infants’ word representations retain speaker-
specific information (Jusczyk, Pisoni, &
Mullennix, 1992; Jusczyk, Hohne, Jusczyk, &
Redanz, 1993; Houston & Jusczyk, 2000), just
as do adults’ (e.g., Goldinger et al, 1991).
Weiss, Mitchel, & Gerfen (2009)
Gebhart, Aslin, & Newport (2009)
*
NS *
NS
When is a voice cue necessary?
Present Study
•Do 12-month-olds use voice and blocking
cues to segregate more and less
consistent speakers?
–Exp 1
•both voice and blocking cues
– Exp 2
•voice cue only
•blocking cue only
Artificial Language
•We adapted the artificial language paradigm from our
lab’s previous work on infants’ tolerance for grammatical
inconsistency (Gómez & Lakusta, 2004).
Structure 1 Structure 2
S { aX
bY }
S { bX
aY }
|a, b| = 2; |X, Y| = 6
Familiarization stimuli
In Structure 1, Xs are 2-syllable words, Ys are 1-syllable words. The opposite holds for Structure 2.
Test stimuli
Syllable number is the same for grammatical test items, but the X- and Y-words are new
Voice and blocking cue
manipulations
Exp 1 Exp 2
both cues neither cues voice cue blocking
only cue only
FAMILIARIZATION PROCEDURE
Head-turn preference procedure
Results: Experiment 1
Results: Experiment 2
Conclusion
•Together, the present results provide the first
direct evidence that segregation of statistical
structure can be achieved:
1. not only for separate languages but also for more and less
consistent versions of the same language,
2. by infants as well as adults, and
1. with a temporal cue (blocking).
•Remaining question: Can infants segregate by
voice?