Over forty years of research have demonstrated that languages partition the phonetic space differently in order to encode their phonological categories (e.g. Lisker & Abramson, 1964) and that attention to perceptual (including acoustic and visual) cues impact phonological learning and categorization (to cite a few, Abramson & Lisker, 1972; Maye, Werker, & Gerken, 2002; McGuire, 2007, January; McGurk & MacDonald, 1976; Mills, 1987; Narayan, 2006). Likewise, attention to acoustic cues has been shown to vary as an effect of age and language exposure (Nittrouer, Manning, & Meyer, 1993; Nittrouer & Miller, 1997; Nittrouer, Miller, Crowther, & Manhart, 2000; Zhang, Kuhl, Imada, Kotani, & Tokkura, 2005). More recently, it has been proposed that motor experience also plays a key role in phonological category learning (Davidson, 2006; Goffman & Smith,
1999; Smith & Goffman, 2004). However, the interaction between different kinds of cues (acoustic, visual, and articulatory) with respect to phonological categorization is still poorly understood, and diverse phonological primitives have been proposed on the bases of each (Flemming, 2002; Goldstein & Fowler, 2003; Winters, 2000). Furthermore, semantic information may actually interfere with phonological categorization, at least in infants’ word-learning (Pater, Stager, & Werker, 2004).
In a theoretical context where the relationship between phonetics and phonology is still hotly debated, it is important to assess the impact of attention to phonetic cues on the formation of phonological categories, and to begin to understand the way cues in different modalities interact with each other in phonological categorization. For this reason, we have invited papers addressing the following questions:
1. What constitutes a cue in phonology?
2. What evidence is there for acoustic, motoric, and semantic cues impacting phonological acquisition and learning?
3. How does development and language experience affect the weighting of cues used in the formation of phonological categories?
4. What are the implications of this cue weighting for language acquisition and phonological theory?
5. How do cues in different modalities interact in the formation of phonological categories?
2 comments:
First, I want to thank Alex for setting up this blog. I think it's a really good idea for us to be able to talk about these issues before the actual meeting.
I'll just mention a few things I have been thinking about:
1) I have been wondering, whether we haven't underestimated the role of visual information in guiding early speech sound categorization and whether it might not be the case that early classes consist of those with strong visual cues. This is in light of a nice paper that Alex sent me which finds that blind children are later at producing bilabials than sighted ones and also given that my son still won't say [k], but produces many other more articulatorily challenging sounds with strong visual cues. This may have something to do with learning style, but I think if style influences class construction that would be exceedingly interesting as well.
2) I have been hoping that we could try to further specify the distinction between dynamic and static cues. What does it mean to be a dynamic cue --or rather how dynamic does a cue have to be to be useful for children who are more attentive to these sorts of cues than adults. Dutch has a more restricted pitch range than English --does this mean that pitch is not a dynamic cue in this language?
3) When we think about hwo infants go about forming vowel categories like [i] vs [I], we assume, following Kuhl's work, that these vowels in IDS do not have overlapping distributions, however, many recent studies of more naturalistic IDS (not just words in citation measured; e.g., one on Norweigan IDS I just read) seem to suggest that IDS vowels (esp ones in a large inventory system like that in Norweigan) have clearly overlapping distributions. One interesting possibility suggested tome recently by Dan Swingley is that infants may not attend to all instantiations of vowels produced --they may just attend to good ones, e.g., just ones with long durations or something and ignore less clear instances. Some work by Melanie Soderstrom suggests that babies know when someone is producing a well-formed vs. non-well-formed utterance, perhaps they also know when someone is producing a canonical vowel vs. a non-canonical one. I'm curious what you all think of this idea.
-Amanda
Hi Amanda,
I just wanted to mention, re the visual cues, that a new series of studies at UCBerkeley suggest that English speakers, when they have to label place of articulation, rely somewhat on visual cues. When given *no* acoustic information, they'll choose labial in the presence of a labial closure, but about 50% coronal and 50% velar in the absence of a labial closure (in spite of the fact that the stimuli are REALLY clear - you can tell whether the closure is forward/coronal or back/velar).
But, true, these are adults, who don't do much of their listening staring at people's mouths, and perhaps during early language acquisition the impact of visual cues is different.
Post a Comment