Scott Seyfarth

Word informativity influences acoustic duration: effects of contextual predictability on lexical representation.

Language-users reduce words in predictable contexts. Previous research indicates that reduction may be stored in lexical representation if a word is often reduced. Because representation influences production regardless of context, production should be biased by how often each word has been reduced in the speakers prior experience. This study investigates whether speakers have a context-independent bias to reduce low-informativity words, which are usually predictable and therefore usually reduced. Content word durations were extracted from the Buckeye and Switchboard speech corpora, and analyzed for probabilistic reduction effects using a language model based on spontaneous speech in the Fisher corpus. The analysis supported the hypothesis: low-informativity words have shorter durations, even when the effects of local contextual predictability, frequency, speech rate, and several other variables are controlled for. Additional models that compared word types against only other words of the same segmental length further supported this conclusion. Words that usually appear in predictable contexts are reduced in all contexts, even those in which they are unpredictable. The result supports representational models in which reduction is stored, and where sufficiently frequent reduction biases later production. The finding provides new evidence that probabilistic reduction interacts with lexical representation.

Dynamic hyperarticulation of coda voicing contrasts

This study investigates the capacity for targeted hyperarticulation of contextually-relevant contrasts. Participants communicated target words with final /s/ or /z/ when a voicing minimal-pair (e.g., target dose, minimal-pair doze) either was or was not available as an alternative in the context. The results indicate that talkers enhance the durational cues associated with the word-final voicing contrast based on whether the context requires it, and that this can involve both elongation as well as shortening, depending on what enhances the contextually-relevant contrast. This suggests that talkers are capable of targeted, context-sensitive temporal enhancements.

Implicative organization facilitates morphological learning

One dimension of this task is segmentation. For example, how can learners separate the stems from the affixes that signal a particular morphosyntactic property? If this information was all that learners had, they might hypothesize that zavod has the lexical meaning ‘factory’, and the –ov suffix signals genitive plural. A large body of experimental research addresses this syntagmatic, structural challenge of identifying recurrent partial forms (e.g., Saffran et al. 1996; Finley and Newport 2011; Aslin and Newport 2012). However, there is also a paradigmatic aspect to the problem. For example, the table below shows some alternative possibilities for how a plural form might be realized with different cases in Russian (Baerman et al. 2009).

Acoustic Differences Between English /t/ Glottalization and Phrasal Creak.

In American English, the presence of creaky voice can derive from distinct linguistic processes, including phrasal creak (prolonged irregular voicing, often at edges of prosodic phrases) and coda /t/ glottalization (when the alveolar closure for syllablefinal /t/ is replaced by or produced simultaneously with glottal constriction). Previous work has shown that listeners can differentiate words in phrasal creak from those with /t/ glottalization, which suggests that there are acoustic differences between the creaky voice derived from phrasal creak and /t/ glottalization. In this study, we analyzed vowels preceding syllable-final /t/ in the Buckeye Corpus, which includes audio recordings of spontaneous speech from 40 speakers of American English. Tokens were coded for presence of phrasal creak (prolonged irregular voicing extending beyond the target syllable) and /t/ glottalization (whether the /t/ was produced only with glottal constriction). Eleven spectral measures of voice quality, including both harmonic and noise measures, were extracted automatically and discriminant analyses were performed. The results indicate that the discriminant functions can classify these sources of creaky voice above chance, and that Cepstral Peak Prominence, a measure of harmonics-to-noise ratio, is important for distinguishing phrasal creak from glottalization.

Acoustic differences in morphologically-distinct homophones

ABSTRACT Previous work demonstrates that a words status as morphologically-simple or complex may be reflected in its phonetic realisation. One possible source for these effects is phonetic paradigm uniformity, in which an intended words phonetic realisation is influenced by its morphological relatives. For example, the realisation of the inflected word frees should be influenced by the phonological plan for free, and thus be non-homophonous with the morphologically-simple word freeze. We test this prediction by analysing productions of forty such inflected/simple word pairs, embedded in pseudo-conversational speech structured to avoid metalinguistic task effects, and balanced for frequency, orthography, as well as segmental and prosodic context. We find that stem and suffix durations are significantly longer by about 4–7% in fricative-final inflected words (frees, laps) compared to their simple counterparts (freeze, lapse), while we find a null effect for stop-final words. The result suggests that wordforms influence production of their relatives.

Does Implicit Voice Learning Improve Spoken Language Processing? Implications for Clinical Practice

PurposenIn typical interactions with other speakers, including a clinical environment, listeners become familiar with voices through implicit learning. Previous studies have found evidence for a Familiar Talker Advantage (better speech perception and spoken language processing for familiar voices) following explicit voice learning. The current study examined whether a Familiar Talker Advantage would result from implicit voice learning.nnnMethodnThirty-three adults and 16 second graders were familiarized with 1 of 2 talkers voices over 2 days through live interactions as 1 of 2 experimenters administered standardized tests and interacted with the listeners. To assess whether this implicit voice learning would generate a Familiar Talker Advantage, listeners completed a baseline sentence recognition task and a post-learning sentence recognition task with both the familiar talker and the unfamiliar talker.nnnResultsnNo significant effect of voice familiarity was found for either the children or the adults following implicit voice learning. Effect size estimates suggest that familiarity with the voice may benefit some listeners, despite the lack of an overall effect of familiarity.nnnDiscussionnWe discuss possible clinical implications of this finding and directions for future research.

Acoustics of the tense-lax stop contrast in Semarang Javanese

Javanese has a contrast between tense and lax stops. While both tense and lax stops are voiceless and unaspirated, the contrast at least in word-initial position is realized through acoustic differences in the following vowel, including lower f0, breathier voice quality, and higher F1 for the lax stops relative to their tense counterparts. However, previous reports have indicated substantial cross-speaker variation, and in some cases involve differing characterizations of the acoustic contrast, possibly due to small sample sizes. Moreover, it is still unclear whether (and how) this contrast is maintained in word-final position. In this study, we investigate the tense-lax contrast based on audio recordings of 27 speakers of Central Javanese from Semarang, Indonesia who each read 30 or more items with a word-initial or word-final stop in a carrier phrase. Stops and their adjacent vowels were hand-annotated, and measurements were taken including voice onset and offset time, stop closure and release duration,...

Acoustic comparison of /t/ glottalization and phrasal creak

In American English, the presence of creaky voice can derive from distinct linguistic processes, including phrasal creak (prolonged irregular voicing, often at edges of prosodic phrases) and coda /t/ glottalization (when the alveolar closure for syllable-final /t/ is replaced by or produced simultaneously with glottal constriction). Garellek (2015) showed that listeners can differentiate words in phrasal creak from those with /t/ glottalization, which suggests that the creaky voice derived from phrasal creak and /t/ glottalization differ acoustically. To test this, we analyzed vowels preceding syllable-final /t/ in the Buckeye Corpus, which is comprised of audio recordings of spontaneous speech from 40 speakers of American English. Tokens were coded for presence of phrasal creak (prolonged irregular voicing extending beyond the target syllable) and /t/ glottalization (whether the /t/ was produced only with glottal constriction). Spectral measures of voice quality, including both harmonic and noise measure...