/ #SLA #Phonetics 

Seminar 5 (11th, 13th & 18th November, 6PM): Ocke-Schwen Bohn (Aarhus Univeristy)

Characteristics, consequences, and causes of foreign accented speech in perception and in production


The fifth seminar held at the Department of Linguistics, SNU in Fall 2020 is “Characteristics, consequences, and causes of foreign accented speech in perception and in production” by Ocke-Schwen Bohn.


The lecturer currently works at Aarhus University, Denmark. He is current associate editor at the Journal of Phonetics.

Description:


  1. 11/11 (Wed.): Social, psychological, and communicative consequences of accentedness

Foreign accented speech can have a range of consequences, including often negative social evaluations of nonnative speakers, and potential intelligibility problems which may lead to miscommunication. This lecture attempts to provide a taxonomy of the consequences of foreign accented speech. Examples from published and unpublished work on listener reactions to nonnative speech are used to exemplify the overall classification of its effects into linguistic-communicative processing cost on the one hand, and social and psychological cost on the other. Both types of cost can be examined in a number of ways. For processing cost, these are comprehensibility, acceptability, and measures of intelligibility, and this contribution highlights both well-documented and possible relations between these aspects of processing cost. The nonlinguistic cost of (foreign) accented speech can be roughly classified as resulting in biases regarding the personality or the social characteristics of the speaker. This lecture will provide examples of how these consequences have been examined, which will result in a discussion of the relation between (aspects of) the two types of cost and the usefulness and appropriateness of different methods used to study the consequences of accentedness.

Link to the slides

  1. 11/13 (Fri.): Predicting difficulty in L2 speech learning

Non-native speech sounds are hardly ever perceived as easily and accurately as native speech sounds. A variety of approaches has been used to predict difficulty in cross-language and non-native speech perception, among them the comparison of phonetic symbols used to transcribe native (L1) and nonnative (L2) speech sounds, acoustic comparisons of the sounds of the L1 and the L2, and various measures of perceived similarity including perceptual assimilation of the sounds of the L2 to L1 categories. This presentation evaluates the predictive power of these approaches through a series of experiments which examined a) L1 Danish listeners’ perception of English consonants, b) L1 English listeners’ perception of L2 Danish unrounded front vowels. This evaluation confirms that the comparison of phonetic symbols can, at best, serve as an initial heuristic. However, this method can be quite misleading. Acoustic comparisons and three measures of perceived similarity each account for some of the perceptual problems, but none of these methods does so in a fully satisfactory manner. The presentation concludes with a discussion of possible reasons for the relative successes and failures of different methods used to predict difficulty in cross-language and non-native speech perception.

Link to the slides

  1. 11/18 (Wed.): Core aspects of the revised Speech learning Model (SLM-r)

Over the past 25 years, much of the research on second language speech has been inspired by the Speech Learning Model (SLM, Flege 1995). Despite its success in generating testable hypotheses and initiating research which has provided considerable support for the SLM, we (Flege & Bohn, 2021) felt the need to revise the model because several of its assumptions needed clarification, and because part of the focus of the original SLM had become unproductive. This talk presents some of the core aspects of the revised Speech Learning Model (SLM-r). Like its predecessor, the SLM-r assumes that L2 learners of any age make use of the same mechanisms and processes to learn L2 speech that children exploit when learning their L1. Native vs. nonnative differences in L2 production and perception are ubiquitous not because humans lose the capacity to learn speech at a certain stage of neuro-cognitive development but because applying the mechanisms and processes that functioned “perfectly” in L1 acquisition to the sounds of an L2 does not yield the same results. The SLM-r attempts to model age effects by referring to age-related differences in the quality and the quantity of L2 input, and to age-related differences in perceived cross-language similarity. Whereas the SLM focused on between-group differences (child vs adult learners, experienced vs inexperienced L2 learners), the SLM-r is an individual differences model which focuses on how individuals learn L2 sounds and how L2 learning influences their production and perception of L1 sounds.

Link to the slides