2010/09/08

Sumner, M. & Samuel A. G. (2009). The effect of experience on the perception and representation of dialect variants. Journal of Memory and Language, 60, 487–501.

Presentation: Sarah

Summary: Shelly

More and more studies discovered interesting phenomena about dialectal differences in speech perception, such as the structure of vowels in a dialect would influence listeners’ ability in discriminating vowels, and also one’s familiarity toward a dialect would influence his/her percept on a certain sound in the dialect. However, it was still not very clear what the underlying mechanism was for processing different dialects. Therefore, this study aimed itself in examining the processing and representation of dialect variants and the effect that prior experience with a dialect had on spoken word recognition. Two dialects involved in the study were General American (GA) and New York City dialect (NYC). One of the differences between the two dialects was that GA uses r-ful pronunciation in words like baker, but NYC uses r-less variants. Three groups of subjects were recruited: (1) Overt-NYC, who had r-less pronunciation, and were born and raised in the New York city; (2) Covert-NYC, who had the same regional background as overt-NYC, but used r-ful pronunciation; and (3) GA, who were born and raised outside of the New York city, and used r-ful pronunciation. With these groups the author attempted to see how the differences in their experience to the NYC dialect affected their perception. Experiment 1 used a form priming paradigm, aiming to see whether r-less forms were as effective as r-ful forms in priming an identical word. Results showed that Overt-NYC and Covert-NYC groups had similar priming patterns. They were both good when the priming and target were from the same dialects. However, for the GA group, no priming effect was found if the target was in NYC. These results suggested that even for listeners raised in an r-less environment (NYC), they showed more flexibility in perceiving the dialect of GA, which is a strong dialect in American English, and this flexibility was not found in the GA group, because they had little experience with NYC, so little priming effect from NYC can be found. Experiment 2 further examined how such form differences influence lexical activation by using a semantic priming paradigm. Results showed that for all the dialect groups, GA final r-ful primes facilitated responses to semantically-related targets. However, NYC final r-less primes only worked for those who had prior exposure to the NYC dialect. Results of Experiments 1 and 2 both implied that dialect experience played a critical role in participants’ immediate processing of phonetic variants, but it was not clear what the long-term representation was for the sound in listeners’ minds. Therefore, Experiment 3 used a long-term repetition priming paradigm, aiming to see how variants are encoded in the long-term. In this paradigm, primes and targets were put in different blocks, and if the sound in the prime was the mental representation of the participants, they should have a priming effect on the target in the next block, and vice versa. Results revealed that NYC primes only worked for the overt-NYC group, and not for the covert-NYC and GA groups, while GA primes worked for all the three groups, suggesting that overt-NYC had representations of both dialects, whereas covert-NYC and GA had only one, which was the GA r-ful one. It was interesting to see that for the covert-NYC group, even though they appeared to be fluent listeners of r-less NYC in Experiments 1 and 2, they had only one r-ful GA encoded in the long-term. To conclude, in line with previous findings, the present study found that familiarity towards a dialect is very influential on listeners’ ability in processing the sound in the dialect, but it does not ensure that the dialect has the status of long-term representation in listeners’ mind.