Current projects
What we do
The members of KUPPL conduct research that addresses questions within a wide range of topics in phonetics, phonology, and psycholinguistics, along with their interfaces with fields such as second language acquisition. To get a sense of the kind of work we do, please see our recent publications and recent conference presentations, and some descriptions of a few recent and current projects below.
Acoustics of clear speech
Allard Jongman, Morgan Robertson
In our research, we study clear (hyper-articulated) speech as a way to gain insight into which acoustic properties contribute most to increased intelligibility. We use an elicitation paradigm to prompt two different types of clear speech: overall and contrast-specific clear speech. Overall clear speech is the speech produced when a speaker is led to believe that their production was incomprehensible (e.g., Maniwa et al., The Journal of the Acoustical Society of America, 2009). Contrast-specific clear speech is the speech produced when a speaker is led to believe that their production was confused with a different word or specific phoneme. For example, when a speaker produced “first” and is led to believe that it was perceived as “thirst”. Our results show that speakers make fine-grained and refined acoustic modifications depending on the confusion that occurred. That is, instead of employing one generic way of speaking more clearly, speakers make distinct, phonetically sophisticated adjustments for a confusion based on place of articulation versus manner of articulation versus voicing. We are currently extending this research to second-language learners.
Does form similarity matter for long-lag morphological priming?
Steve Politzer-Ahles, Xuan Wang, Münir Özturhan
It has long been assumed that people can access a word from their lexicon more easily if they have recently been exposed to another word that shares a morpheme with it, even if the form of that morpheme is different. (This phenomenon is known as long-lag morphological priming.) But previous studies mainly used words in which the difference in form was very small – e.g., pairs like give…gave. This study takes advantage of English irregular verbs like teach…taught and go…went to see if long-lag morphological priming can really occur between words with very different forms.
Enhancing the effectiveness of high variability phonetic training (HVPT)
We are investigating the impact of incorporating cue-attention switching training into traditional HVPT. This modification aims to reallocate learners' attention to L2-relevant acoustic cues, thereby reducing L2 learning gaps derived from their L1 cue-weighting strategies and fostering more native-like perceptual patterns.
Fricatives
Fricatives are a wonderful class of sounds to test theories of speech perception. Some fricatives have robust intrinsic cues while others are more context-dependent. Thus, fricatives represent an ideal platform for examining the informational assumptions of models of speech categorization. However, fricatives are also difficult to categorize. Consequently, listeners utilize a large number of cues to do so.
While fricatives have received considerable attention, there is current debate about what the best method is to characterize them and to what extent high-frequency information (above 10-15 kHz) contributes to their distinction. We plan to conduct a comprehensive study to determine the utility of various metrics, including spectral moments with and without taper analysis, and DCT coefficients (Jongman, Oxford Research Encyclopedia of Linguistics, 2024).
In understanding how fricatives as a class of consonants are perceived, both invariant and cue-combination approaches may be appropriate, but the context-dependence of many, if not all, cues raises the possibility that compensation (the interpretation of a cue relative to others) is necessary. Modeling the relationship between the acoustics and perception of fricatives first gave rise to our C-CuRE model (McMurray&Jongman, Psychological Review, 2011), the core tenet of which is that listeners do not make decisions based on raw cues but instead build up expectations about what a segment produced by a specific talker and in a specific context should sound like, and that what drives perception is the difference between those expectations and the actual signal. The weighting of signal information and expectation gives rise to perception. Since then, we have used C-CuRE to understand categorization of other sound classes, modalities (audio and visual), and speech styles (clear and conversational) (e.g., Redmon et al., Journal of Phonetics, 2020).
Individual differences in learning novel phonological contrasts
In this project, we are investigating individual differences in within-category acoustic cue sensitivity in native language perception and domain-general inhibitory control ability as potential sources of variability in L2 learning success.
Perceptual adaptation to lexical stress in different intonations
Tzu-Hsuan Yang, Jeff Holliday, Jieun Lee
We are investigating listeners’ ability to rapidly adjust to changes in the phonetic realization of a contrast, focusing on lexical stress in English. Specifically, we're looking at whether listeners can adjust their perceptual strategies by relying more on a secondary acoustic dimension (duration) when the primary cue (vowel quality) is absent and another second diagnostic cue (f0) varies with intonation.
Phonetic drift in Mandarin sibilants
Tzu-Hsuan Yang, Jeff Holliday, Allard Jongman
This project investigates how L2 experience influences the phonetic realization of L1 categories. We are focusing on L2 English learners who speak Taiwan Mandarin as their L1, where there are large individual differences in the L1 phonetic space due to sibilant merger. The goal is to explore whether the amount of drift will be modulated by the sparseness/density of the L1 phonetic space.
Updated November, 2024