Exploiting random intercepts: Two case studies in sociophonetics

Drager, Katie; Hay, Jennifer

doi:10.1017/s0954394512000014

Cited by 57 publications

(47 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A larger coefficient means that a speaker was more likely to produce the /k/ in the quotative than in the discourse particle; the more negative the coefficient, the more likely a speaker was to exhibit a strong CR trend in production. These results are also presented in Drager & Hay (2012).…”

Section: Variation At the Individual Levelsupporting

confidence: 56%

Linguistic variation, identity construction and cognition

K.¹

2015

View full text Add to dashboard Cite

Section: Variation At the Individual Levelsupporting

confidence: 56%

Linguistic variation, identity construction and cognition

K.¹

2015

View full text Add to dashboard Cite

“…SPEAKER and PROMPT were included as random factors to prevent extremes of variation in the behavior of particular speakers or extremes of variation in the production of particular prompts having an undue effect on statistical results, see Drager and Hay (2012) and Hay (2011). The inclusion of SPEAKER as a random factor was particularly important in the construction of models for the formant data, as it helped to remove formant variation that might be attributable to variation in vocal-tract length.…”

Section: H Statistical Analysismentioning

confidence: 99%

The role of gesture delay in coda /r/ weakening: An articulatory, auditory and acoustic study

Lawson

Stuart‐Smith

Scobbie

2018

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

The cross-linguistic tendency of coda consonants to weaken, vocalize, or be deleted is shown to have a phonetic basis, resulting from gesture reduction, or variation in gesture timing. This study investigates the effects of the timing of the anterior tongue gesture for coda /r/ on acoustics and perceived strength of rhoticity, making use of two sociolects of Central Scotland (working- and middle-class) where coda /r/ is weakening and strengthening, respectively. Previous articulatory analysis revealed a strong tendency for these sociolects to use different coda /r/ tongue configurations—working- and middle-class speakers tend to use tip/front raised and bunched variants, respectively; however, this finding does not explain working-class /r/ weakening. A correlational analysis in the current study showed a robust relationship between anterior lingual gesture timing, F3, and percept of rhoticity. A linear mixed effects regression analysis showed that both speaker social class and linguistic factors (word structure and the checked/unchecked status of the prerhotic vowel) had significant effects on tongue gesture timing and formant values. This study provides further evidence that gesture delay can be a phonetic mechanism for coda rhotic weakening and apparent loss, but social class emerges as the dominant factor driving lingual gesture timing variation.

show abstract

“…The idea is to assess the effects of vocal tract length in a different data set, and then statistically remove these effects in our data set of interest. This statistical technique is discussed in d etail, and with reference to the dataset presented here, in Drager and Hay (2012).…”

Section: F3 Of Linking /R/mentioning

confidence: 99%

/r/-sandhi in early 20th century New Zealand English

Hay

Maclagan²

2012

Linguistics

Self Cite

View full text Add to dashboard Cite

Most work on linking and intrusive /r/ is conducted within the context of theoretical model building. This paper contributes some empirical data on the phenomenon by conducting the first large-scale acoustic analysis on the production of linking and intrusive /r/ in spontaneous speech. We analyze linking and intrusive /r/ in the speech of New Zealanders born between 1900 and 1935 -a period during which New Zealand English had recently become reliably nonrhotic, and the system of linking and intrusive /r/ was newly emergent.The acoustic analysis shows that the more often a speaker tends to produce linking /r/, the lower the /r/ F3 tends to be. That is, across speakers, there is a robust correlation between the frequency of use of /r/, and the degree of constriction used. Consideration of the distributional patterns of words reveals that words which more often tend to occur before vowels are more likely to be produced with /r/ in linking environments. This suggests that the pattern of contexts in which we experience a word is stored in a way that affects the word's subsequent production. Moreover, the F3 of the /r/s appearing in these words which tend to be pre-vocalic are significantly lower. There is thus a r elationship between frequency of use and degree, both when considered across speakers and across words.

show abstract

Exploiting random intercepts: Two case studies in sociophonetics

Cited by 57 publications

References 17 publications

Linguistic variation, identity construction and cognition

Linguistic variation, identity construction and cognition

The role of gesture delay in coda /r/ weakening: An articulatory, auditory and acoustic study

/r/-sandhi in early 20th century New Zealand English

Contact Info

Product

Resources

About