Speech segmentation is a language-specific skill: each language provides different cues for optimally segmenting the continuous speech stream into words. When exposed to a novel language, listeners have been shown to use those segmentation cues that they are familiar with from their native language (L1). Learning a second language (L2) entails learning to use the prosodic segmentation cues appropriate for that language. Whereas many studies have addressed the influence of the L1 on segmentation of L2 or novel languages, it is yet unclear how L1 and L2 segmentation strategies might compete in highly proficient L2 learners. In order to assess the prominence of L1 vs. L2 prosodic segmentation strategies, this study investigates which segmentation cues take precedence when listeners are exposed to a novel language. It assesses to what extent proficiency in the L2 and recent exposure to the L2 affect the use of L1 versus L2-specific segmentation cues in segmenting novel-language speech. Participants were (1) native listeners of French, (2) native listeners of English without functional knowledge of French, and (3) native listeners of English with a high proficiency in French as an L2 (assessed with a cloze test) who had lived in a French-speaking country for at least three months. Importantly, all participants were living in the USA at the time of testing. The native listeners of French and the L2 learners of French had been selected to vary in the amount of time elapsed since their last stay in a French-speaking country. In an artificial language learning experiment, participants were tested on their use of a segmentation cue that matches French but not English prosody, namely a high tone on word-final syllables. The word-final high F0 is expected to be a useful segmentation cue for French listeners, because accented syllables in French occur at the right edge of the Accentual Phrase and are word-final; in non-utterance-final positions, they have higher F0 and longer duration than the corresponding unaccented syllables. It is not expected to be a useful segmentation cue for (monolingual) English listeners, because in English, accented syllables tend to be word-initial, with F0 rise, increased amplitude and, to some extent, increased duration signaling this word-initial prominence. During a training phase, 30 participants were exposed to a 20-minute speech stream consisting of six trisyllabic words without any phonetic word-boundary cues except the word-final high F0. In the subsequent test phase, they were tested on their recognition of the trained words. The results showed no effect of French proficiency for the L2 learners. There was, however, a strong effect of recent linguistic exposure: performance increased with decreasing time in the USA since the last stay in a French-speaking country. Surprisingly, the recency of immersion in an L2 environment thus turned out to override L2 proficiency. These results indicate that recent linguistic exposure is a crucial factor influencing listeners’ use of prosodic cues in speech segmentation. This suggests that speech-processing routines are much less rigid and more adaptive than previously assumed. AMLaP 2012 Poster Session 4 September 8, 14:00-15:30
UCLA Working Papers in Phonetics No. 106, pp. 1-33 Effects of initial position versus prominence in English Taehong Cho Hanyang University, Korea tcho@hanyang.ac.kr Patricia Keating keating@humnet.ucla.edu Abstract This study investigates effects of three prosodic factors—prosodic boundary (Utterance-initial vs. Utterance-medial), lexical stress (primary vs. secondary) and phrasal accent (accented vs. unaccented)—on articulatory and acoustic realizations of word-initial CVs (/ne/, /te/) in trisyllabic English words. Articulatory measurements include linguopalatal contact (by electropalatography) for both C and V, and seal duration; acoustic measurements include nasal duration and energy for /n/, VOT, burst energy and spectral center of gravity for /t/, and F1, vowel duration and vowel amplitude for /e/. Several specific points emerge. First, domain-initial articulation is differentiated from stress- or accent-induced articulations in many aspects; for the most part, prominence affects vowel measures while initial position affects consonant measures. Nonetheless, the vowel is also effectively louder domain-initially, suggesting that the boundary effect is not strictly local to the initial consonant. Second, the boundary (domain-initial) effect is not seen across-the-board, but is often constrained by stress and accent factors, revealing that domain-initial strengthening is more effective when a relevant phonetic dimension does not undergo a compelling strengthening coming from stress or accent. Third, some accentual effects can be seen on secondary-stressed syllables, suggesting that accentual influences spread beyond the primary-stressed syllable. But this spread is mainly seen with consonantal measures, showing an asymmetric accentual influence between consonantal and vocalic articulations. 1. Introduction Prosodic structure has been widely recognized as an essential element of speech production, as it conveys a great deal of both structural and discourse information (Selkirk, 1995; Swerts & Geluykens, 1994; Herman, 2000). A large body of phonetic studies in the past two decades has increasingly demonstrated the importance of fine-grained phonetic detail in building up differential prosodic structures of utterances. One of the most conspicuous phonetic hallmarks of prosodic structure is domain-final lengthening (e.g. Klatt, 1975; Wightman et al., 1992; Gussenhoven & Rietveld, 1992, Edwards, Beckman & Fletcher, 1991; Cho, 2002, 2006; Byrd,
No abstract is provided for this article.
This study investigates native language effects on phonetic encoding of coda voicing contrast in L2 English by Chinese versus Korean speakers. Results show much smaller phonetic differences in both vowel duration and F0 in marking coda voicing contrast for Chinese speakers than Korean speakers, despite native Chinese speakers' experience with lexical tones. They suggest that producing an F0-related cue in L2 is conditioned by position-specific phonological richness and use of F0 in the speaker's L1. The results are discussed in terms of contrast maximization and effort minimization with reference to the information structure occurring in both L1 and L2.
This study examines the effect of prosodic position on segmental properties of Korean consonants /n, t, th, t*/ along the articulatory parameters peak linguopalatal contact and stop seal duration, and several acoustic parameters. These parameters were compared in initial position in different domains of the Korean prosodic hierarchy. The first result is that consonants initial in higher prosodic domains are articulatorily stronger than those in lower domains, in the sense of having more linguopalatal contact. Second, there is a strong correlation between linguopalatal contact and duration (both articulatory and acoustic), suggesting that “strengthening” and “lengthening” is a single effect in Korean. We interpret this relation as one of undershoot: in weaker positions, consonants are shorter and undershoot contact targets. The different consonant manners of Korean can be characterized as varying in both duration and contact in this way. Third, there is another, less consistent, kind of lengthening and strengthening specific to Korean, namely that tense and aspirated consonant oral articulations can be longer and stronger word-medially than word-initially. Fourth, the acoustic properties VOT, total voiceless interval, %voicing during closure, nasal energy minimum, and to a lesser extent stop burst energy and voicing into closure, were found to vary with prosodic position and, in some cases, to correlate with linguopalatal contact. They could thus potentially provide cues to listeners about prosodic structure.
The current work examines native Korean speakers' perception and production of stop contrasts in their native language (L1, Korean) and second language (L2, English), focusing on three acoustic dimensions that are all used, albeit to different extents, in both languages: voice onset time (VOT), f0 at vowel onset, and closure duration. Participants used all three cues to distinguish the L1 Korean three-way stop distinction in both production and perception. Speakers' productions of the L2 English contrasts were reliably distinguished using both VOT and f0 (even though f0 is only a very weak cue to the English contrast), and, to a lesser extent, closure duration. In contrast to the relative homogeneity of the L2 productions, group patterns on a forced-choice perception task were less clear-cut, due to considerable individual differences in perceptual categorization strategies, with listeners using either primarily VOT duration, primarily f0, or both dimensions equally to distinguish the L2 English contrast. Differences in perception, which were stable across experimental sessions, were not predicted by individual variation in production patterns. This work suggests that reliance on multiple cues in representation of a phonetic contrast can form the basis for distinct individual cue-weighting strategies in phonetic categorization.
This acoustic study explores how Korean learners produce coarticulatory vowel nasalization in English that varies with prosodic structural factors of focus-induced prominence and boundary. N-duration and A1-P0 (degree of V-nasalization) are measured in consonant-vowel-nasal (CVN) and nasal-vowel-consonant (NVC) words in various prosodic structural conditions (phrase-final vs. phrase-medial; focused vs. unfocused). Korean learners show a systematic fine-tuning of the non-contrastive V-nasalization in second language (L2) English in relation to prosodic structure, although it does not pertain to learning new L2 sound categories (i.e., L2 English nasal consonants are directly mapped onto Korean nasal consonants). The prosodic structurally conditioned phonetic detail in English appears to be accessible in most part to Korean learners and was therefore reflected in their production of L2 English. Their L2 production, however, is also found to be constrained by their first language (L1-Korean) to some extent, resulting in some phonetic effects that deviate from both L1 and L2. The results suggest that the seemingly low-level coarticulatory process is indeed under the speaker's control in L2, which reflects interactions of the specificities of the phonetics-prosody interface in L1 and L2. The results are also discussed in terms of their implications for theories of L2 phonetics.
This paper examines the effects of morpheme boundaries on intergestural timing, and demonstrates that low-level phonetic realization is influenced by morphological structure, i.e. compounding and affixation. It reports two experiments, one using electromagnetic midsagittal articulography (EMA) and one electropalatography (EPG), examining Korean data. The results of the EMA study show that intergestural timing is less variable for adjacent gestures across the word boundary inside a lexicalized compound than inside a nonlexicalized compound, and inside a monomorphemic word than across a morpheme boundary. The EPG study (which examined the timing in palatalization of a coronal) shows that both [ti] and [ni] have more variability in gestural timing when heteromorphemic than when tautomorphemic. Furthermore, the phonetic details of gestural overlap shed light on the asymmetry on palatalization between tautomorphemic and heteromorphemic gestural sequences (e.g. ni vs. n-i), presumably driven by paradigmatic contrast and preference of overlap. In short, what emerges from two experiments is that gestures are coordinated more stably within a single lexical item (a morpheme or a lexicalized compound) than across a boundary between lexical items. In accounting for the stability of intergestural timing within a lexical entry, several hypotheses were discussed including the Phase Window, Bonding Strength, Phonological Timing and Extended Phase Window model newly proposed here. The implication is that the morphological structure may be encoded in the phonetic realization, as is the case with other linguistic structure (e.g. prosodic structure).
This study compares prosodic structural effects on nasal (N) duration and coarticulatory vowel (V) nasalization in NV (Nasal-Vowel) and CVN (Consonant-Vowel-Nasal) sequences in Mandarin Chinese with those found in English and Korean. Focus-induced prominence effects show cross-linguistically applicable coarticulatory resistance that enhances the vowel's phonological features. Boundary effects on the initial NV reduced N's nasality without having a robust effect on V-nasalization, whose direction is comparable to that in English and Korean. Boundary effects on the final CVN showed language specificity of V-nasalization, which could be partly attributable to the ongoing sound change of coda nasal lenition in Mandarin.
This study investigated effects of three prosodic factors—prosodic boundary, lexical stress, and accent—on articulatory and acoustic realizations of two CV syllables, /nE/ and /tE/. These syllables occurred at the beginning of trisyllabic English nonwords; their position in the larger phrase (prosodic boundary conditions), and whether they were lexically stressed and/or accented (prominence conditions) were varied. Articulatory measurements included linguopalatal contact (by electropalatography) for both C and V, stop consonant seal duration, and C-to-V contact difference; acoustic measurements include nasal duration and energy for /n/; VOT, burst energy and spectral center of gravity for /t/; and F1, vowel duration and vowel amplitude for /E/. We tested whether domain-initial strengthening occurs in the C and/or the V segments independently of stress or accent conditions. We found that the effects of position and of stress/accent can be distinguished in the production and the acoustics of these syllables. One domain-initial effect (greater consonant contact domain-initially) was complementary to one stress/accent effect (greater vowel opening with stress/accent); in other cases the effects overlapped (greater vowel energy and tendency to longer consonant both domain-initially and with stress/accent); in one case they conflicted (less consonant energy domain-initially, more consonant energy with stress/accent).
This paper reports on the articulation of two aspects of speech prosody: word and sentence stress. Word stress is examined in reiterant renditions of minimal pairs differing in the location of main stress (e.g., DIScharge–disCHARGE), with the contrastive stress difference manifested as differences in the location of the nuclear pitch accent. Sentence stress is examined in sets of minimal sentences differing in the presence or location of the nuclear pitch accent (e.g., so TOMMY gave Debby a song from Timmy). Three talkers read these materials while movements of the jaw and tongue (using the Carstens Articulograph) or of 20 reflective infrared markers on the face (using the Qualysis Motion Capture system) were recorded. Syllables with and without contrastive pitch accents are compared along several articulatory dimensions, including: jaw displacement and duration of movement, jaw peak opening and closing velocities, jaw peak acceleration, whole head displacement, and eyebrow displacement. These comparisons allow us to determine whether the talkers articulate the focused syllables/words differently. The three talkers were selected to differ in their overall visual speech intelligibility, as rated by expert deaf speechreaders, and therefore interspeaker differences in articulation are expected. [Work supported by NSF KDI Grant 9996088.]
This study examines how young speakers of Seoul Korean produce tri-consonantal clusters /1kt/ and /1pt/ as in palk-ta ('to be bright') and palp-ta ('to step on'). Production data were collected from 20 speakers of Seoul Korean. The results of narrow transcription of the data showed that simplification is not obligatory as some speakers often preserve all three consonants. When simplified, there was a clear asymmetry between /1kt/ and /1pt/. Speakers showed no clear preference for either C1 preservation (C1=/1/) or C2 preservation (C2=/k/ in /1kt/ and /p/ in /1pt/) in production of /1kt/, but in production of /1pt/, strong preference was found for C1-preserved to C2-preserved variant. When compared with production data in Cho (1999), simplification patterns appear to have changed over the past 10 years, in a direction to preserve the first member of the cluster (/1/) more often, especially with /1kt/. There was no substantial between-item variation, indicating that simplification patterns are not lexically specified. Finally, the results suggest that the process of tri-consonantal simplification has not been fully phonologized in the grammar of the language as evident in substantial inter- and intra-speaker variation.
An articulatory study was conducted to explore effects of prosodic boundary and syllable structure on temporal realizations of /ma/ in C♯V vs. ♯CV in Korean (where ‘♯’ denotes an Intonational Phrase or a Word boundary). The vocalic gesture underwent boundary-induced lengthening more in C♯V than in ♯CV, implying that the boundary effect is largely localized to the initial element whether consonantal or vocalic. CV coordination patterns were temporally neutralized between ♯CV and C♯V in the phrase-internal Word boundary condition, showing a possible ‘resyllabifiation’ of ‘C’ with the following vowel in C♯V in the articulatory temporal measures taken in the present study. It was suggested that CV gestures in C♯V, whose phasing relationship has to be determined postlexically, reorganize temporally in an in-phase coupling mode just like the way CV gestures are phased in ♯CV. Finally, while there was leftward shifting of the consonantal gesture in C♯V with some temporal variability across an IP vs. a Word boundary, intergestural timing in ♯CV remained invariant regardless of boundary strength. But the most stable temporal pattern was observed with an IP boundary in ♯CV, interpretable as an important temporal characteristic of domain-initial strengthening. Some of these results were further discussed in terms of their implications for the theory of π-gesture and the gestural coupling model of syllable structure.
We explore the role of the acoustic consequences of domain-initial strengthening in spoken-word recognition. In two cross-modal identity-priming experiments, listeners heard sentences and made lexical decisions to visual targets, presented at the onset of the second word in two-word sequences containing lexical ambiguities (e.g., bus tickets, with the competitor bust). These sequences contained Intonational Phrase (IP) or Prosodic Word (Wd) boundaries, and the second word's initial Consonant and Vowel (CV, e.g., [tɪ]) was spliced from another token of the sequence in IP- or Wd-initial position. Acoustic analyses showed that IP-initial consonants were articulated more strongly than Wd-initial consonants. In Experiment 1, related targets were post-boundary words (e.g., tickets). No strengthening effect was observed (i.e., identity priming effects did not vary across splicing conditions). In Experiment 2, related targets were pre-boundary words (e.g., bus). There was a strengthening effect (stronger priming when the post-boundary CVs were spliced from IP-initial than from Wd-initial position), but only in Wd-boundary contexts. These were the conditions where phonetic detail associated with domain-initial strengthening could assist listeners most in lexical disambiguation. We discuss how speakers may strengthen domain-initial segments during production and how listeners may use the resulting acoustic correlates of prosodic strengthening during word recognition.
Manifestation of prosodic structure in articulatory variation: Evidence from lip kinematics in English was published in Laboratory Phonology 8 on page 519.
No abstract is provided for this article.
