Sonya Bird

Techniques for field application of lingual ultrasound imaging.

Techniques are discussed for using ultrasound for lingual imaging in field‐related applications. The greatest challenges we have faced distinguishing the field setting from the laboratory setting are the lack of controlled head/transducer movement, and the related issue of tissue compression. Two experiments are reported. First, a pilot study identifies important factors in controlling head/transducer movement in field settings. Second, an Optotrak/ultrasound study reports the range of head movement in an optimal field‐like setting within and across varying phonetic contexts, as well as the effect of tongue tissue compression on tongue image data. Results suggest that with a simple arrangement involving a head rest or surface, a fixed transducer, and careful design and presentation of stimuli, reliable lingual ultrasound data can be collected in the field.

Oral–laryngeal timing in glottalised resonants

Abstract Although previous studies of intergestural timing in multi-gesture segments have identified some consistent patterns, fundamental questions remain about the underlying causes of these patterns. Hypotheses based on universal perceptual or biomechanical restrictions have proven difficult to test because of confounding factors, e.g. anatomical coupling and aerodynamic necessity. The cross-linguistically rare class of glottalised resonants (GRs) involves oral and laryngeal gestures, which are neither anatomically nor aerodynamically interdependent, thereby providing a revealing test case for these hypotheses. If intergestural timing is determined by universal perceptual factors, GR timing patterns should be consistent across languages. This comparative study of GRs in three endangered British Columbian languages reveals distinct patterns: the timing of GRs is consistently pre-glottalised in Nuu-chah-nulth, post-glottalised in Nɬeʡkepmxcin, and dependent upon syllable position in St’at’imcets. These findings indicate that a strong hypothesis based on perceptual recovery must be rejected, and suggest instead that intergestural timing must be specified on a language-specific basis.

Lheidli intervocalic consonants: phonetic and morphological effects

Several researchers have noted that intervocalic consonants are unusually long in Navajo (Sapir & Hoijer 1967, Young & Morgan 1987, McDonough & Ladefoged 1993). This paper explores intervocalic consonant duration in the Lheidli dialect of Dakelh (Carrier) in order to determine whether the long intervocalic consonants found in Navajo are characteristic of other Athabaskan languages as well. It is shown that Lheidli intervocalic consonants are substantially longer overall than (a) consonants in other positions within Lheidli, (b) vowels within Lheidli, and (c) singletons and geminates in other non-Athabaskan languages. Furthermore, intervocalic consonant duration is at least in part a function of morphological structure, with intervocalic consonants lengthening to signal a stem morpheme boundary. The data presented here provide new evidence for the existence of a language-specific phonetic component of grammar as well as for the interaction between phonetics and morphology.

The phonetic rhythm/syntax headedness connection: Evidence from Tagalog

Ramus, Nespor, and Mehler [Cognition (1999)] show that the rhythm of a language (broadly: stress‐ versus syllable‐ versus mora‐timing) results from the proportion of vocalic material in an utterance (%V) and the standard deviation of consonantal intervals (delta‐C). Based on 14 languages, Shukla, Nespor, and Mehler [submitted] further argue that rhythm is correlated with syntactic headedness: low %V is correlated with head‐first languages (e.g., English); high %V is correlated with head‐final languages (e.g., Japanese). Together, these proposals have important implications for language acquisition: infants can discriminate across rhythm classes [Nazzi, Bertoncini, and Mehler, J. Exp. Psych: Human Perception and Performance (1998)]. If rhythm, as defined by %V and delta‐C, can predict headedness, then infants can potentially use rhythm information to bootstrap into their languages syntactic structure. This paper reports on a study analyzing rhythm in a language not yet considered: Tagalog. Results support the Shukla et al. proposal in an interesting way: based on its %V and delta‐C, Tagalog falls between head‐first and head‐last languages, slighty closer to the head‐first group. This placement correlates well with the fact that, although Tagalog is said to be primarily head‐first syntactically, head‐last phrases are permitted and common in the language.

Pronunciation change in SENĆOŦEN: A acoustic study of /k, kw, kw’, q, q’, qw qw’/ across generations of speakers

This paper presents an acoustic study of /k, kw, kw’, q, q’, qw qw’/ in SENCOŦEN, a dialect of North Straits Salish. The “K series”—as these sounds are collectively called—are of particular concern among speakers, the perception being that both the uvular ~ velar and the plain ~ ejective contrasts may be disappearing. To understand how these contrasts are currently being realized, 12 speakers (3 speakers x 4 generations) were recorded pronouncing 20 isolated words (7 consonants x 2–3 words). Based on auditory impression and preliminary acoustic analysis (PLACE: spectral composition of bursts/frication; formant transitions into adjacent vowels; VOICING: VOT; jitter; shimmer; amplitude rise time), the velar ~ uvular contrast is indeed less consistent among younger speakers than among their elders; there is also a lot of variability in the plain ~ ejective contrast, among all speakers. In addition, the plain ~ ejective contrast interacts with the velar ~ uvular contrast for some speakers: /k kw/ and /q qw/ a...

Bimodal schwa: Evidence from acoustic measurements

The question of whether schwa is targeted or targetless has been the subject of much debate (Browman et al., 1992; Browman and Goldstein, 1995; Gick, 1999, 2002). Gick (2002) found that there is a pharyngeal constriction during schwa relative to rest position, and concluded that schwa is not targetless. This experiment further showed a ‘‘bimodal’’ pattern in schwa in a nonrhotic speaker, indicating that the subject has distinct schwas in lexical words and function words. The present study examines the existence of the ‘‘bimodal’’ pattern in schwas in nonrhotic dialects through an acoustic experiment. It is predicted that there is a significant difference in formant values between lexical schwas and function schwas. Results to date indicate a significant difference in them between schwas in lexical versus function words, both between historical schwas and those derived from final /r/ reduction. Data from several additional nonrhotic subjects will be presented. Implications for intrusive r as well as for th...

The battle for F0: Glottalization versus stress in Statimcets

The conflicts that arise in speech production give us insight into the interaction between faithfulness to mental representations and articulatory and acoustic limitations on speech. This paper presents an example of a perceptually based conflict [Kochetov, LabPhon8 (to appear)] in Statimcets, a Salish language spoken in the Interior of British Columbia. In a pilot study on glottalized resonants in Statimcets, Bird [ICSNL 38 (2003)] found that the primary cue to glottalization was creaky voicing, and that glottalization was reduced in stressed syllables. This paper expands on the previous study, incorporating new acoustic data on stress, and a more detailed acoustic analysis of glottalization. It is shown that one of the cues to stress, raised fundamental frequency (F0), conflicts with the lowered F0 associated with creaky voicing. In Statimcets, this conflict is resolved by preserving the perceptual salience of stress cues at the expense of cues to glottalization. This results in loss of glottalization i...

Laryngeal contrasts in first and second language speakers of Hul’q’umi’num’

This paper investigates laryngeal contrasts in first (L1) and second (L2) language speakers of Hul’q’umi’num’, a dialect of Halkomelem (Salish). The language, which lacks a thorough acoustic description of its consonants, is highly endangered with 50-100 native speakers, but has a growing number of L2 speakers. Since the learners come from an English background and are therefore unfamiliar with glottalized consonants, the sounds can prove challenging. This study addresses this issue by examining how L1 and L2 speakers pronounce plain and ejective stops in terms of what acoustic correlates they’re using. Tokens of words read in isolation will be analyzed from L1 and L2 speakers participating in a language course in Duncan, BC. Tokens of both speaker groups will be classified by phoneme and acoustic measures of duration (e.g., voice onset time, closure duration), centre of gravity, and vowel coarticulation (e.g., spectral tilt, F0, rise time) will be made. The results of the two groups will be compared with statistical analysis, and the findings will be used to create guidelines to assess future learners’ and teach pronunciation. The findings will also contribute to a broader understanding of how Hul’q’umi’num’ consonants fit into voicing and ejective typology.This paper investigates laryngeal contrasts in first (L1) and second (L2) language speakers of Hul’q’umi’num’, a dialect of Halkomelem (Salish). The language, which lacks a thorough acoustic description of its consonants, is highly endangered with 50-100 native speakers, but has a growing number of L2 speakers. Since the learners come from an English background and are therefore unfamiliar with glottalized consonants, the sounds can prove challenging. This study addresses this issue by examining how L1 and L2 speakers pronounce plain and ejective stops in terms of what acoustic correlates they’re using. Tokens of words read in isolation will be analyzed from L1 and L2 speakers participating in a language course in Duncan, BC. Tokens of both speaker groups will be classified by phoneme and acoustic measures of duration (e.g., voice onset time, closure duration), centre of gravity, and vowel coarticulation (e.g., spectral tilt, F0, rise time) will be made. The results of the two groups will be compared with...

Developing ultrasound overlay videos for SENĆOŦEN language learners

This paper reports on an initiative to develop a pronunciation training tool for learners of SENCOŦEN (Salish: Canada). In acquiring new speech sounds, second language learners rely on both acoustic and articulatory information. Ultrasound has been demonstrated to be effective for pronunciation instruction, as it allows learners to visualize tongue movements during speech. However, ultrasound-based instruction can be difficult to implement with large groups or with learners wanting to learn independently. To address these limitations, we developed a technique for creating ultrasound overlay videos which combine ultrasound images of tongue movements with external profile views of a speaker’s head. Ultrasound is particularly useful in SENCOŦEN, as its inventory includes many lingual articulations that are difficult to distinguish without visual cues, such as velar/uvular contrasts, and a complex coronal series (e.g., ť, ťθ, ʧ, ʃ, ƛ, ɫ, l, ľ). SENCOŦEN is critically endangered, but efforts are underway to re...

An Exemplar Dynamic Approach to Language Shift

Phonetic variability refers to the variability with which we speak. This variability may be a function of socio-linguistic factors such as dialect, for example the intervocalic /t/ in city is often pronounced as flap [R] in Canadian English but as glottal stop [P] in British English. It may also be a function of linguistic factors. For example in Canadian English, the pronunciation of /t/ is a function of prosodic environment: it is aspirated [t] in top, flapped [R] in city and glottalised [t ∼ P] in nightlife. In Quebecois French, /t/ varies as a function of segmental environment: it is affricated [t] preceding high front vowels /i/ and /y/, as in [ty] tu ‘you sg.’ but unaffricated elsewhere, as in [to] tot ‘early’. Finally, a certain amount of variability is intrinsic to speech, and results from the fact that our articulatory mechanisms are not characterized by robotic levels of precision. Phonetic variability has recently become the focus of much research because of its implications for how sound structure is represented in a language’s grammar. Generative models of phonology represent speech sounds lexically as single, abstract, invariant underlying forms. For example, [t] [R], and [P] in the English