"Learning to Represent Phonological Contrasts: Evidence from Perception and Production"
Suzanne van der Feest
Department of Psychology
University of Pennsylvania
Two distinct traditions have traced the development of phonology in children's
language understanding: a psychological tradition rooted in psychophysics
(with syllable discrimination as the signature task), and a linguistic tradition
based on formal analysis (primarily examining children's speech). Recent
methodological advances have raised the possibility of a truly psycho-linguistic
approach to early phonological development by allowing well-controlled assessment
of even one-year-olds' speech comprehension. Previous studies tested childre's
responses to words in which a sound was replaced with another sound (e.g., Swingley
& Aslin, 2002; Stager & Werker, 1997; Bailey & Plunkett, 2002). Using a
preferential looking paradigm, these studies established that children look for
example shorter to the target "ball" when hearing "dall" than when hearing "ball".
On the basis of these studies, it has been argued that children have detailed
representations of well-known words. A variety of mispronunciations were used,
but different types of mispronunciations were not usually tested in a systematic
way, whereas for example recent work by Nazzi (2004, 2008) shows that consonants
and vowels yield different mispronunciations effects.
In my recent work, I have refined this approach by examining Dutch- versus
English-learning children's interpretation of a single cue at the time, while at
the same time investigating the development of these cues in children's early
productions. In this talk I will discuss the interpretation of Dutch word-initial
voicing and English versus Dutch vowel duration.
Even though previous perception studies have suggested that children have detailed
stored representations of well-known words, studies on child language production
have claimed that children's phonological representations are underspecified:
Fikkert & Levelt (2004) provided evidence for the underspecified nature of
coronal place of articulation and Kager, Van der Feest, Fikkert, Kerkhoff &
Zamuner (2007) for the underspecification of voiceless stops. In both studies the
evidence came from asymmetries in production. If phonological representations of
known words are underspecified, we expect to find evidence for this in perception as well.
In this talk, I will first discuss a study that sets out to test whether the
asymmetries found in children's production of voice and place also show up in
perception. Using the same preferential looking paradigm as Swingley (2003), I
tested the perception of mispronunciations of well-known words by forty-eight
24-month-old Dutch children, keeping the factors voice and place clearly balanced.
These results indicate that children have stored underspecified phonological
representations of words, which they use for both perception and production. I
will further address the development of perceptual phonology and the nature of
early lexical representations by discussing English- versus Dutch-learning
children's interpretation of variation in vowel duration, which provides an
interesting test case because its function varies in each language.
|
