An acoustic analysis of rhoticity in Nigerian English

Although Nigerian English (NE) has been described as a non-rhotic English variety (Jowitt-2018), I present the first systematic acoustic analysis of rhoticity (non-prevocalic-/r/s in ca r , fa r me r ) in NE. The dataset comprises spontaneous and elicited speech of 30 Yoruba native speakers born and raised in Lagos, Nigeria. Interviews were transcribed and force-aligned in hand-checked FAVE. A modified Praat script measured the first three formants in each V + /r/ sequence at nine equally spaced time points. For a feature intrinsic normalizing measure of rhoticity, this study focuses on the static measurement of the F3–F2 difference. The formant measurements are subjected to mixed-effects linear regression analyses in R, using lme4 package to model F2–F3 at the F3 minimum. Based on 11 823 tokens of non-prevocalic-/r/ and with higher values of F3–F2 (Hertz) indicating weaker rhoticity, findings show the following results: gender (female = 941; male = 1972), age (18–29 = 901; 30–49 = 1634; 50–63 = 2082), social class (upper-class = 703; middle-class = 1598; lower-class = 2097), and positional contexts (word-final = 964; word-initial = 1673). As there is overt prestige associated with American English (AE), I hypothesize that NE is becoming rhotic with influence from AE. Younger upper-class females lead the change, especially in elicited speech. Low F3 close to F2 serves as an accurate correlate of rhoticity in NE. Importantly, there was hyper-rhoticity (i.e.-“Africaɹ”).