Karthik Durvasula

Neurobiological evidence for voicing underspecification in English

In long-term memory, the phoneme units that make up words are coded for the distinctive features and feature values that are necessary to distinguish between words in the mental lexicon. Underspecification theory says that the phonemes that have unmarked feature values are even more abstract in that the feature is omitted from the representation altogether. This makes phoneme representations in words more sparse than the fully specified phonetic representations of the same words. Eulitz and Lahiri (2004) demonstrated that this theory predicts certain asymmetries in the Mismatch Negativity (MMN) response to phoneme contrasts. We expand on this research by demonstrating underspecification-driven asymmetry in the brain response to laryngeal feature contrasts in English (i.e. what makes /d/ and /t/ different). We add a new test by showing that the asymmetry disappears if the MMN paradigm is modified to encourage the formation of phonetic memory traces instead of phonemic memory traces. This result adds further neurobiological evidence that long-term phonological representations are more sparsely represented than phonetic representations.

Illusory vowels in perceptual epenthesis: the role of phonological alternations

Listeners often perceive illusory vowels when presented with consonant sequences that violate phonotactic constraints in their language. Previous research suggests that the phenomenon motivates speech-perception models that incorporate surface phonotactic information and the acoustics of the speech tokens. In this article, inspired by Bayesian models of speech perception, we claim that the listener attempts to identify target phonemic representations during perception. This predicts that the phenomenon of perceptual illusions will be modulated not only by surface phonotactics and the acoustics of the speech tokens, but also by the phonological alternations of a language. We present the results of three experiments (an AX task, an ABX task and an identification task) with native Korean listeners, and native English listeners as a control group, showing that Korean listeners perceive different sets of illusory vowels in different phonological contexts, in accordance with the phonological processes of vowel deletion and palatalisation in the language.

The role of phrasal phonology in speech perception: What perceptual epenthesis shows us

Abstract Recent research using the phenomenon of illusory vowels has raised our awareness of the extent to which speech perception is modulated by the listeners native-language phonological knowledge. However, most of the focus has been limited to word-level phonological knowledge. In this article, we suggest that the perceptual system recruits segmental phonological knowledge that makes crucial reference to prosodic domains far beyond the word-level. We report the results from three identification experiments on Korean and American English participants. In accordance with their native-language phonotactic constraints at the level of the Intonational Phrase, Korean listeners unlike American English listeners hear more illusory vowels in stimuli containing the sequence of voiced stops followed by nasal consonants (e.g. [eɡma]) than those containing voiceless stops followed by nasal consonants (e.g. [ekma]). The results are interpreted as support for the view that speech perception makes crucial reference to the concept of reverse inference.

Word-internal ambisyllabic consonants are codas

The syllabic affiliation of ambisyllabic consonants (e.g., the word-medial consonants in happy and Danny) is unclear. Research on ambisyllabic consonants has revealed an inconsistent set of phonetic correlates (Krakow, 1989; Turk, 1994; Gick, 2004). While some suggest they behave as onsets or codas (but not both simultaneously), others suggest their gestural durations are intermediate between onsets/codas. At least some of the research is based on comparisons of the ambisyllabic consonants to word-edge onsets/codas. However, comparisons to word-edges are confounded by the fact that such consonants undergo domain-edge related changes (Fougeron, 2001; Keating et al., 2003b). Here, we control for this confound, and compare ambisyllabic consonants to word-medial onsets and codas. We conducted an experiment on 10 native English speakers, who produced 15 repetitions at three different speech rates of 16 English words (8 test, 8 filler) that consisted of the nasal consonants [n or m] in one of four positions: wo...

Voicing, aspiration, and vowel duration in Hindi

There is extensive evidence that consonantal laryngeal features modulate preceding vowel duration (Chen, 1970). However, it is not clear if both consonant voicing and aspiration affect preceding vowel duration. Previous studies produced inconsistent results with respect to the effect of consonant aspiration on vowel duration, while finding a clear positive correlation with consonant voicing (Maddieson & Gandour, 1976; Ohala & Ohala, 1992; Lampp & Reklis, 2004). Furthermore, the locus of the explanation of these effects is unresolved (Kluender et al, 1988; Fowler, 1992). We conducted an experiment on 7 native standard Hindi speakers, who produced 10 repetitions of 12 nonsense words ending in [d, d, t, t] that had 3 different CVCVC contexts. In this article we focus on standard Hindi to show the following: (a) As with other languages, there is a vowel duration difference before voiced and voiceless consonants coda (syllable-final) consonants, (b) Vowel durations preceding aspirated coda consonants are longer than those before unaspirated coda consonants, (c) Closure durations of coda consonants are longer for unaspirated consonants and voiceless consonants, (d) Finally, when crucial confounds are controlled for, there is a slight positive, not negative, correlation between coda consonant duration and preceding vowel length.

Inconsistent Consonantal Effects on F- in Cantonese and Mandarin

Previous research shows that aspiration and sonorancy can have inconsistent consonantal effects on vowel F0 across languages, and even within the same language in different studies: they are reported to either lower F0, raise F0, or have neutral effects. This paper is interested in such inconsistent consonantal effects on vowel F0 in Cantonese and Mandarin. The results of a series of production experiments show that: (1) consonantal effects on vowel F0 are language-specific; (2) the consonantal effects can be conditioned by lexical tones. The findings also provide insights for puzzles in historical tone change.

Acoustic correlates of consonant gesture timing in English

There is extensive research on the organization of syllable-structure as indexed by the relative timing of the articulators (Browman and Goldstein, 1988; Byrd, 1995; Shaw et al., 2011 inter alia). The research suggests consonants in complex onsets (in words such as scream, stream…) are aligned to a single position called the C-center, the mean of the midpoints of the onset consonants. However, such research typically uses very expensive articulatory equipment (X-ray Microbeam, Electromagnetic Articulography…). This restricts the research to a few laboratories across the world with access to such technology. Here, we explore the possibility of using acoustic measurements, which are cheaper and more accessible, for such research. We conducted an experiment on 6 native speakers of English, who produced 12 repetitions of 24 English words (12 test, 12 filler) that varied in the number of onset consonants (C1, C1C2, C1C2C3) in three different vowel contexts. Paralleling previous studies, the results show that o...