The influence of speaker information on the processing of the ING variable
The focus of this study is on the sociolinguistic variable ING, and how its canonical variant, -ing (jumping) and non-canonical variant, -in’ (jumpin’) are processed with different prior information on the speaker. The aim of this study is to find out what effect certain regional and socioeconomic background information on a speaker has on listeners’ reaction time and accuracy of words ending with -ing and -in’. Across three experiments, participants took part in the same lexical decision task with audio files of the same speaker. These audio files consisted of single words ending with -ing and -in’, alongside fillers of real words and nonwords. In each experiment, participants were asked to identify the stimuli they heard as real words or nonwords. Participants of Experiment 1 formed a control group, and received no prior information on the speaker. Participants of Experiment 2 were told that the speaker they were about to hear was an upper-middle class man from Connecticut, a state associated with the -ing form. Participants from Experiment 3 were told that the speaker was a working-class man from Alabama, a state associated with the -in’ form. It was hypothesised that words ending with -ing would be processed faster and more accurately identified as existing words across all three experiments, due to the canonicality advantage. This turned out to be the case. The second hypothesis was that words ending with -in’ would be processed faster and more accurately identified as existing words in Experiment 3 (Alabama guise) than in the other two experiments, due to their association with the area in question. In addition, the accuracy of identifying words ending with -in’ as existing words was expected to be higher in Experiment 3 compared to Experiment 1 and 2. The results show that this second hypothesis is not the case, and that both variants were processed in a similar speed and identified as accurate in a similar manner across all three experiments. These results are in support of the canonicality advantage, and show that speaker information turned out to have little to no effect on the processing of the ING variable.
Faculteit der Letteren