Learning Phonology With Substantive Bias: An Experimental and Computational Study of Velar Palatalization

Abstract

There is an active debate within the field of phonology concerning the cognitive status of substantive phonetic factors such as ease of articulation and perceptual distinctiveness. A new framework is proposed in which substance acts as a bias, or prior, on phonological learning. Two experiments tested this framework with a method in which participants are first provided highly impoverished evidence of a new phonological pattern, and then tested on how they extend this pattern to novel contexts and novel sounds. Participants were found to generalize velar palatalization (e.g., the change from [k] as in keep to [t ʃ] as in cheap) in a way that accords with linguistic typology, and that is predicted by a cognitive bias in favor of changes that relate perceptually similar sounds. Velar palatalization was extended from the mid front vowel context (i.e., before [e] as in cape) to the high front vowel context (i.e., before [i] as in keep), but not vice versa. The key explanatory notion of perceptual similarity is quantified with a psychological model of categorization, and the substantively biased framework is formalized as a conditional random field. Implications of these results for the debate on substance, theories of phonological generalization, and the formalization of similarity are discussed.

Evidence for a learning bias against saltatory phonological alternations

“…A soft bias is typically implemented in models using a prior, which makes certain outcomes, within the space of possible outcomes, have higher a priori likelihoods relative to others. The idea that such soft biases have a role in phonological learning has been growing in the literature (e.g., see Wilson, 2006;Zuraw, 2007;Finley & Badecker, 2008;Moreton, 2008;Hayes et al, 2009;Hayes & White, 2013). Implementing a fully functioning learning model to account for these results is beyond the scope of this paper.…”

“…(Saffran & Thiessen, 2003;Wilson, 2006;Zuraw, 2007;Finley & Badecker, 2008;Moreton, 2008;Hayes et al, 2009;Becker et al, 2011;Skoruppa, Lambrechts, & Peperkamp, 2011;Baer-Henney & van de Vijver, 2012;Becker et al, 2012;Finley & Badecker, 2012;Hayes & White, 2013).…”

“…These results suggest that learners find alternations between highly similar sounds easier to learn, but the effect of similarity beyond a single feature remains unclear. Wilson (2006) used a different artificial language paradigm. Participants were trained on novel alternations involving palatalization, with one group learning that it occurred before high vowels (e.g., [ki] à [tʃi]) and a different group learning that it occurred before mid vowels (e.g.,…”

“…4 Another way of defining similarity is in terms of perception. Similarity is likely evaluated by listeners on the basis of multiple sources of information, and forming a complete understanding of how people judge the similarity of two speech sounds is not trivial (see Steriade, 2001Steriade, /2008Wilson, 2006;Mielke, 2012;Cristia et al, in press). For simplicity, I take the relatively straightforward notion of mutual confusability as an approximation of perceptual similarity.…”

“…Steriade's Pmap was originally conceived within the phonological framework of Optimality Theory (Prince & Smolensky, 1993/2004, but its basic principles can be extrapolated to other models of phonological learning. Indeed, Wilson (2006) implemented a computational version of the Pmap as a way to explain why learners extended novel alternations involving palatalization more often to contexts where it would be less perceptibly noticeable (before high vowels) compared to contexts where it would be more noticeable (before mid vowels). Using maximum entropy grammar models (see also Goldwater & Johnson, 2003;Hayes & Wilson, 2008), Wilson implemented the similarity bias as a prior.…”