Evidence For Selective Adaptation and Recalibration in the Perception of Lexical Stress

Individuals vary in how they produce speech. This variability affects both the segments (vowels and consonants) and the suprasegmental properties of their speech (prosody). Previous literature has demonstrated that listeners can adapt to variability in how different talkers pronounce the segments of speech. This study shows that listeners can also adapt to variability in how talkers produce lexical stress. Experiment 1 demonstrates a selective adaptation effect in lexical stress perception: repeatedly hearing Dutch trochaic words biased perception of a subsequent lexical stress continuum towards more iamb responses. Experiment 2 demonstrates a recalibration effect in lexical stress perception: when ambiguous suprasegmental cues to lexical stress were disambiguated by lexical orthographic context as signaling a trochaic word in an exposure phase, Dutch participants categorized a subsequent test continuum as more trochee-like. Moreover, the selective adaptation and recalibration effects generalized to novel words, not encountered during exposure. Together, the experiments demonstrate that listeners also flexibly adapt to variability in the suprasegmental properties of speech, thus expanding our understanding of the utility of listener adaptation in speech perception. Moreover, the combined outcomes speak for an architecture of spoken word recognition involving abstract prosodic representations at a prelexical level of analysis.

Vowel Quality and Direction of Stress Shift in a Predictive Model Explaining the Varying Impact of Misplaced Word Stress: Evidence From English

The use of suprasegmental cues to word stress occurs across many languages. Nevertheless, L1 English listeners' pay little attention to suprasegmental word stress cues and evidence shows that segmental cues are more important to L1 English listeners in how words are identified in speech. L1 English listeners assume strong syllables with full vowels mark the beginning of a new word, attempting alternative resegmentations only when this heuristic fails to identify a viable word string. English word stress errors have been shown to severely disrupt processing for both L1 and L2 listeners, but not all word stress errors are equally damaging. Vowel quality and direction of stress shift are thought to be predictors of the intelligibility of non-standard stress pronunciations—but most research so far on this topic has been limited to two-syllable words. The current study uses auditory lexical decision and delayed word identification tasks to test a hypothesized English Word Stress Error Gravity Hierarchy for words of two to five syllables. Results indicate that English word stress errors affect intelligibility most when they introduce concomitant vowel errors, an effect that is somewhat mediated by the direction of stress shift. As a consequence, the relative intelligibility impact of any particular lexical stress error can be predicted by the Hierarchy for both L1 and L2 English listeners. These findings have implications for L1 and L2 English pronunciation research and teaching. For research, our results demonstrate that varied findings about loss of intelligibility are connected to vowel quality changes of word stress errors and that these factors must be accounted for in intelligibility research. For teaching, the results indicate that not all word stress errors are equally important, and that only word stress errors that affect vowel quality should be prioritized.

Effects of speech rate on anticipatory eye movements in the visual world paradigm: Evidence from aging, native, and non-native language processing

Research has shown that suprasegmental cues in conjunction with visual context can lead to anticipatory (or predictive) eye movements. However, the impact of speech rate on anticipatory eye movements has received little empirical attention. The purpose of the current study was twofold. From a methodological perspective, we tested the impact of speech rate on anticipatory eye movements by systemically varying speech rate (3.5, 4.5, 5.5, and 6.0 syllables per second) in the processing of filler-gap dependencies. From a theoretical perspective, we examined two groups thought to show fewer anticipatory eye movements, and thus likely to be more impacted by speech rate. Experiment 1 compared anticipatory eye movements across the lifespan with younger (18–24 years old) and older adults (40–75 years old). Experiment 2 compared L1 speakers of English and L2 speakers of English with an L1 of German. Results showed that all groups made anticipatory eye movements. However, L2 speakers only made anticipatory eye movements at 3.5 syllables per second, older adults at 3.5 and 4.5 syllables per second, and younger adults at speech rates up to 5.5 syllables per second. At the fastest speech rate, all groups showed a marked decrease in anticipatory eye movements. This work highlights (1) the importance of speech rate on anticipatory eye movements, and (2) group-level performance differences in filler-gap prediction.

Repetition enhancement to voice identities in the dog brain

In the human speech signal, cues of speech sounds and voice identities are conflated, but they are processed separately in the human brain. The processing of speech sounds and voice identities is typically performed by non-primary auditory regions in humans and non-human primates. Additionally, these processes exhibit functional asymmetry in humans, indicating the involvement of distinct mechanisms. Behavioural studies indicate analogue side biases in dogs, but neural evidence for this functional dissociation is missing. In two experiments, using an fMRI adaptation paradigm, we presented awake dogs with natural human speech that either varied in segmental (change in speech sound) or suprasegmental (change in voice identity) content. In auditory regions, we found a repetition enhancement effect for voice identity processing in a secondary auditory region – the caudal ectosylvian gyrus. The same region did not show repetition effects for speech sounds, nor did the primary auditory cortex exhibit sensitivity to changes either in the segmental or in the suprasegmental content. Furthermore, we did not find evidence for functional asymmetry neither in the processing of speech sounds or voice identities. Our results in dogs corroborate former human and non-human primate evidence on the role of secondary auditory regions in the processing of suprasegmental cues, suggesting similar neural sensitivity to the identity of the vocalizer across the mammalian order.

The role of lexical stress in spoken English word recognition by listeners of English and Taiwan Mandarin

Two perceptual experiments investigated how the suprasegmental information of monosyllables is perceived and exploited in spoken English word recognition by listeners of English and Taiwan Mandarin (TM). Using an auditory lexical decision task in which correctly stressed English words and mis-stressed nonwords (e.g. camPAIGN vs. *CAMpaign) were presented for lexical decisions, Experiment I demonstrated that TM listeners could perceive the differences between stressed and unstressed syllables with native-like accuracy and rapidity. To examine how the perceived suprasegmental contrast would constrain English lexical access, Experiment II was conducted. It used a cross-modal fragment priming task in which a lexical decision had to be made for a visually presented English word or nonword following an auditory prime, which was a spoken word-initial syllable. The results showed that English and TM listeners recognized the displayed word (e.g. campus) faster both after a stress-matching (e.g. CAM-) prime and a stress-mismatching (e.g. cam-) prime than after a control prime (e.g. MOUN-, with mismatching segments). This indicates that suprasegmental information does not inhibit a segmentally matching but suprasegmentally mismatching word candidate for both the two groups, although TM is a language where lexical prosody is expressed syllabically and its listeners tend to interpret lexical stress tonally. Yet, the two groups’ responses were slower after the stressed primes than after the unstressed ones, presumably because the former generally had more possible continuations than the latter do. It is therefore concluded that when recognizing spoken English words, both the native and non-native (TM-speaking) listeners can exploit the suprasegmental cues of monosyllables, which, however, are not so effective that they will outweigh the segmental cues.

Discrepancy between heritage speakers' use of suprasegmental cues in the perception and production of Spanish lexical stress

This study investigates Spanish heritage speakers' perception and production of Spanish lexical stress. Stress minimal pairs in various prosodic contexts were used to examine whether heritage speakers successfully identify the stress location despite varying suprasegmental cues (Experiment 1) and whether they use these cues in their production (Experiment 2). Heritage speakers' performance was compared to that of Spanish monolinguals and English L2 learners. In Experiment 1, the heritage speakers showed a clear advantage over the L2 learners and their performance was comparable to that of the monolinguals. In Experiment 2, both the heritage speakers and the L2 learners showed deviating patterns from the monolinguals; they produced a large overlap between paroxytones and oxytones, especially in duration. The discrepancy between heritage speakers' perception and production suggests that, while early exposure to heritage language is beneficial for the perception of heritage language speech sounds, this factor alone does not guarantee target-like production.

Explaining Cross-Language Asymmetries in Prosodic Processing: The Cue-Driven Window Length Hypothesis

Cross-language studies have shown that English speakers use suprasegmental cues to lexical stress less consistently than speakers of Spanish and other Germanic languages ; accordingly, these studies have attributed this asymmetry to a possible trade-off between the use of vowel reduction and suprasegmental cues in lexical access. We put forward the hypothesis that this “cue trade-off” modulates intonation processing as well, so that English speakers make less use of suprasegmental cues in comparison to Spanish speakers when processing intonation in utterances causing processing asymmetries between these two languages. In three cross-language experiments comparing English and Spanish speakers’ prediction of hypo-articulated utterances in focal sentences and reporting speech, we have provided evidence for our hypothesis and proposed a mechanism, the Cue-Driven Window Length model, which accounts for the observed cross-language processing asymmetries between English and Spanish at both lexical and utterance levels. Altogether, results from these experiments illustrated in detail how different types of low-level acoustic information (e.g., vowel reduction versus duration) interacted with higher-level expectations based on the speakers’ knowledge of intonation providing support for our hypothesis. These interactions were coherent with an active model of speech perception that entailed real-time adjusting to feedback and to information from the context, challenging more traditional models that consider speech perception as a passive, bottom-up pattern-matching process.