Sociogeographical differences in the pronunciation of Anglicisms
in Spanish
Acoustic vowel analysis using multilevel regression
models
In this paper we propose a possible statistical
analysis of sociogeographical differences in two Spanish varieties
in the context of loanword phonology. For this purpose, we extracted
formant frequencies of Anglicisms from corpus data from Mexico and
Spain, with a particular focus on the realisation of the English
vowel /ʌ/ (e.g., [bɾʌnt͡ʃ]). We used a Multilevel Bayesian
regression model to infer various fixed effects on the phonic
realisation. The analysis shows that geographical proximity to the
U.S. does not impact the realisation in our data. Speakers’ exposure
and affinity towards the English language, on the other hand, favour
imitation of the English vowel. This seems to be especially relevant
in our globalised world, where English media are ubiquitously
available.
Article outline
- 1.Introduction
- 2.Methodology
- 2.1Choice of the phoneme /ʌ/
- 2.2Corpus
- 2.2.1Participants
- 2.2.2Tasks
- 2.2.3Internal variables
- 2.2.4External variables
- 2.3Acoustic analysis
- 2.4Statistical analysis
- 3.Results
- 3.1Distribution of the frequency values in the vowel
space
- 3.2Analysis of predictors
- 3.3Posterior varying effects
- 3.4Analysis of the latent variable
- 4.Discussion
- 4.1Adaptation of the phoneme
- 4.1.1Realisation of grapheme-phoneme correspondences vs. imitation of the English vowel
- 4.1.2Differences in pronunciation between words
- 4.2Factors influencing the realisation of the phoneme
- 4.2.1Language exposure and affinity
- 4.2.2Speakers’ provenance
- 4.2.3Internal variables
- 5.Conclusion
-
Acknowledgements
-
Notes
-
References