Alice Faber

Phonation differences and the phonetic content of the tense-lax contrast in Utah English

This article presents data bearing on the question of what happens at the phonetic level during a sound change of the type which Labov, Yaeger, and Steiner (1972) labeled an “apparent merger.” Our production data come from three generations of native Utahns who participated in the Intermountain Language Survey (ImLS) and four New Yorkers who served as control subjects. The phonetic subject of our study is the ongoing change in the tense-lax pairs /i-I, e-e, u-υ/ before tautosyllabic dark [†] in Utah English. Previous studies reported that the resultant vowels are usually, but not always, perceived by both transcribers and speakers as lax. Acoustic analysis, self-categorization data, and perception data demonstrate that, after the usual F1/F2 contrast has been lost, contrasts between these and lax vowels may persist in phonation differences and that these phonation differences may be available to hearers.

Perceptual assimilation and discrimination of non-native vowel contrasts.

Research on language-specific tuning in speech perception has focused mainly on consonants, while that on non-native vowel perception has failed to address whether the same principles apply. Therefore, non-native vowel perception was investigated here in light of relevant theoretical models: the Perceptual Assimilation Model (PAM) and the Natural Referent Vowel (NRV) framework. American-English speakers completed discrimination and native language assimilation (categorization and goodness rating) tests on six nonnative vowel contrasts. Discrimination was consistent with PAM assimilation types, but asymmetries predicted by NRV were only observed for single-category assimilations, suggesting that perceptual assimilation might modulate the effects of vowel peripherality on non-native vowel perception.

Assimilation of non‐native vowel contrasts to the American English vowel system.

The perceptual assimilation model (PAM) [Best et al., JEP:HPP 14, 345–360 (1988)] predicts that two non‐native sounds that are assimilated to the same nature category will be harder for listeners to discriminate between than sounds that are assimilated to two different native categories (TC contrasts); how difficult will depend on whether they are equally good (or bad) exemplars of the single native category (SC contrasts) or not (CG contrasts). PAM was based on studies of consonant perception, as were subsequent tests of the model. This study extends the model to non‐native vowels. American listeners performed keyword identification [W. Strange and T. L. Gottfried, J. Acoust. Soc. Am. 68, 1622–1625 (1979)] and categorical AXB discrimination tasks using six non‐native vowel contrasts, Norwegian /i–y/, /i–■/, French /o–o/, /œ–y/, /œ–■/, and Thai /■–■/. Assimilation patterns for a particular vowel contrast, inferred from keyword results, were more variable than in consonant studies but nonetheless strongly ...

The Discriminability of Nearly Merged Sounds.

In a near merger, speakers produce two contrasting words differently without being able to reliably discern the contrast in their own speech or in the speech of others. Acoustic measurements typically reveal small differences between the elements of near merged minimal pairs, along several acoustic dimensions. This paper argues that statistical evaluation of the potential distinctiveness of these near merged elements must take simultaneous account of all these dimensions. For that reason, discriminant analysis was used to assess the differences between near merged lil-dl, leI-eli, and IUI-ul! for five Utah speakers. In contrast with independent univariate Analyses of Variance of Fl, F2, fo, and spectral slope, the multivariate discriminant analyses suggest that all three contrasts are preserved by all five speakers. However, homophones like heel and heal were not distinguished by the discriminant analyses. Discriminant analysis is thus a powerful technique for assessing whether a reliable basis exists for the claim that two potentially contrastive items are in fact distinctive.

Tongue-jaw coordination in vowel production: Isolated words versus connected speech

Abstract This study investigates the positions of the tongue body and the jaw in the production of Italian vowels /i/ and /a/ in different phonetic, prosodic and utterance contexts, with the aim of assessing the role of and the coordination between the two articulators in the processes of coarticulation, reduction and compensation. The data indicate that, within the same phonetic context, a change in utterance type (from isolated words to words in connected speech) or in lexical stress position (from stressed to unstressed vowels) induces decreased displacement of the jaw and the tongue from their rest position along the high/low dimension for both vowels. Thus prosodic and utterance contexts induce vowel reduction through a decrease in displacement of both jaw and tongue. Variation in vowels as a function of the consonantal context (/t,d,z,f,l/) was observed in jaw displacements only in the front/back dimension: vowels were more fronted when adjacent to fricatives. All the other coarticulatory effects concern tongue body movements and tend to increase, as does reduction, from isolated words to connected speech. In symmetric VCV sequences extensive compensatory tongue displacements in the back direction were observed during the production of reduced /a/ vowels: thus, if vowel reduction causes a decrease in the articulatory distance between /i/ and /a/ along the high/low dimension, this compensatory tongue movement appears to counteract such effect by increasing the articulatory distance along the front/back dimension. In asymmetric sequences, the V-to-V effects seem to overrule the compensatory movements, and, adding to the reduction effects, cause a further decrease in the articulatory distance between the two vowel types.

Introduction to papers on speech recognition and perception from an articulatory point of view

The following group of papers resulted from a special session entitled Speech Recognition and Perception from an Articulatory Point of View that was held during the spring 1994 meeting of the Acoustical Society of America in Cambridge, Massachusetts. Organization of the session began when Richard McGowan, Terry Nearey, and Juergen Schroeter invited speakers to give papers and critiques on the role of articulation in human perception and machine recognition. Presentations were invited from three speakers or groups of speakers who were representative of the three speech areas in the Society: production, perception, and processing. One talk was given by Bjorn Lindblom, another was given by John Ohala, and the third by Rick Rose, Juergen Schroeter, Mohan Sondhi, and Oded Ghitza. The invited critiquers for the Lindblom paper were Ken Stevens, Robert Remez, and Bishnu Atal; for the Rose et al. paper, Roger Moore, Joe Perkell, and Terry Nearey; and for the Ohala paper, Mary Beckman, Douglas O’Shaughnessy, and Ca...

Linguo‐mandibular coordination in consonant production

The coordination between the jaw and tongue in regulating tongue position for consonant articulation was explored by simultaneous use of electropalatography and an alternating magnetic field device (Movetrack: Branderud, 1985). Linguo‐palatal constriction was monitored via palatography, and tongue body and jaw height and frontness were monitored by small receiver coils adhering to the articulators. Speech materials were real and nonsense words of the form (C)V1CV2, where C ranged over {t d ∫ z I} and V1 and V2 over {a i}. The “real” words were produced in isolation and in sentences. Each token was repeated three times by one female native speaker of Italian. Regression analysis was used both to distinguish between rotational and translational movement of the jaw and to separate between the tongues independent movement and positional effects caused by jaw movement. Examination of the inherent differences among consonants in tongue position reveals both synergistic and antagonistic patterns of linguo‐mandi...

Relationship of recorded EMG signals to within‐ and cross‐utterance acoustic variation

Analysis of simultaneous acoustic and EMG recordings of one speaker of New York English producing /əpVp/ utterances under several conditions shows differences between inter‐ and intrautterance interactions. For three anterior genioglossus insertions (♯1–♯3) and one posterior insertion (♯4), the peak activation amplitude relevant to articulation of the vowel was measured. Amplitude differences reliably differentiated target vowels. For vowel/muscle pairs with significant amplitudes, activity duration and phasing of the peak within the activity were also measured. Despite strong correlations among the peak amplitudes for the anterior insertions, these measures differed in their correlations with measures of F0, F1, and F2, both within and across utterances. Overall, F2 correlated best with ♯1/duration, F1 with ♯2/amplitude, and F0 with ♯2/amplitude. For /ɪ/, intertoken F2 variation correlated best with ♯3/phasing, F1 with ♯2/amplitude, and F0 with ♯3/phasing. The posterior ♯4/duration correlated best with /...

The effect of consonant context on vowel goodness rating

The effect on within‐category discrimination of supposed phonetic prototypes for the sounds of a given language depends on the assumption that the number of prototypes used by speakers of that language is of the same order of magnitude as the number of phonemes in that language. In the present experiment, listeners provided goodness ratings for three sets of synthetic tokens varying in F2 and F3. One set consisted of isolated /i/ tokens while the other two contained appropriate transitions and release bursts for BEEP and GEEK, respectively. Goodness ratings depended not only on a token’s position in the grid but also on its phonetic context (zero vs b——p vs g——k), reflecting the well‐known coarticulatory effects of consonant context on vowel production. The listener judgments thus reflect the relative appropriateness of a given set of formant values for a consonantal context and not an abstract phonological target underlying all three contexts. [Work supported by NIH Grant No. HD‐01994.]

Dialect differences in vowel production and perception

Earlier work [Faber et al., J. Acoust. Soc. Am. 94, 1865 (1994)] reported differences among American‐English‐speaking listeners from Utah and Connecticut/NY in perception of the HEEL–HILL and POOL–PULL contrasts (pairs that are nearly merged in Utah but distinct in the northeast US), measured by three tasks, labeling, AXB discrimination, and keyword identification. On these tasks, a few CT/NY listeners (those with parents from the southern US) performed differently from the other subjects. Their vowel spaces also were qualitatively different from those of the other listeners, based on acoustic analysis of three readings of the keywords. The CT/southern listeners had more high back crowding and did better on Utah POOL/PULL than the other listeners, while the Utah listeners had more high front crowding and did better on Utah HEEL/HILL. These results accord with the literature reviewed by Bradlow [Cross‐Linguistic Study of Vowel Inventories, Cornell (1993)] relating listeners’ ability to discern small vowel ...