Effects of the distribution of acoustic cues on infants' perception of sibilants

Abstract

a b s t r a c tA current theoretical view proposes that infants converge on the speech categories of their native language by attending to frequency distributions that occur in the acoustic input. To date, the only empirical support for this statistical learning hypothesis comes from studies where a single, salient dimension was manipulated. Additional evidence is sought here, by introducing a less salient pair of categories supported by multiple cues. We exposed English-learning infants to a multi-cue bidimensional grid ranging between retroflex and alveolopalatal sibilants in prevocalic position. This contrast is substantially more difficult according to previous cross-linguistic and perceptual research, and its perception is driven by cues in both the consonantal and the following vowel portions. Infants heard one of two distributions (flat, or with two peaks), and were tested with sounds varying along only one dimension. Infants' responses differed depending on the familiarization distribution, and their performance was equally good for the vocalic and the frication dimension, lending some support to the statistical hypothesis even in this harder learning situation. However, learning was restricted to the retroflex category, and a control experiment showed that lack of learning for the alveolopalatal category was not due to the presence of a competing category. Thus, these results contribute fundamental evidence on the extent and limitations of the statistical hypothesis as an explanation for infants' perceptual tuning.

The Distribution of Talker Variability Impacts Infants’ Word Learning

“…In previous work, variability in talker's voice, for word learning (Rost & McMurray, 2009) or in intervening elements, for nonadjacent-dependency learning (Gómez, 2002) has been unstructured relative to the categories or pattern to be learned. When the distribution of variation has been bimodal, it has been on a phonological dimension (e.g., voicing; Maye et al, 2002;Maye et al, 2008; or visual cues to phoneme identity; Teinonen et al, 2008; see also Cristia et al, 2011; though see Zhao et al, 2013, with adults).…”

Semantics guide infants’ vowel learning: Computational and experimental evidence

“…Experimental evidence shows that real infants can indeed learn a novel phonetic contrast from only auditory input, even within several minutes (Cristia, McGuire, Seidl, & Francis, 2011;Maye et al, 2002;Maye, Weiss, & Aslin, 2008;Yoshida, Pons, Maye, & Werker, 2010;Wanrooij, Boersma, & van Zuijen, 2014). For example, Maye et al (2002Maye et al ( , 2008 presented infants with a continuum of a phonetic contrast.…”

Distributional training of speech sounds can be done with continuous distributions

“…We can therefore erase the fear that earlier results demonstrating an effect of training with discontinuous distributions (e.g., Gerken, 2000, 2001;Maye et al, 2002Maye et al, , 2008Hayes-Harb, 2007;Gulian et al, 2007;Yoshida et al, 2010;Escudero et al, 2011;Cristi a et al, 2011;Wanrooij et al, 2013) could have been artifacts of the discontinuous sampling method; after all, these results have now been replicated with the arguably more natural continuous distributions, so it has become more likely that the observed perceptual improvements are a realistic result of bimodal training. However, as both types of sampling have now been shown to exhibit distributional learning effects and continuous distributions can be considered more ecologically valid than discontinuous distributions, we recommend for future distributional learning experiments not to artificially reduce the variation in the stimuli to 8 or 10 auditory values but to solely employ continuous distributions.…”

“…In all previous studies on distributional learning, bimodal distributions were based on stimuli with 8 or 10 different values for voice onset time (e.g., Gerken, 2000, 2001;Hayes-Harb, 2007;Maye et al, 2002;Maye et al, 2008;Yoshida et al, 2010), vowel formants (e.g., Gulian et al, 2007;Escudero et al, 2011;Wanrooij et al, 2013), or fricative frequencies and formant transitions (Cristi a et al, 2011), and these stimuli were repeated in certain proportions. In Fig.…”