Investigating factors that influence English L2+ perceptual vowel acquisition
Loading...
Date
2025
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Alternative Title(s)
modality, speaker, and feedback
Abstract
This thesis is based upon a selected 5 papers and articles themed around perceptual phonology acquisition and listening for English as a second/additional language (L2+). The articles are focused upon the use of Global Englishes and prestige varieties as a source of vowel acquisition (Jones & Blume, 2022) and as potential factors in listening difficulty (Jones 2024d); visual salience as a factor in vowel acquisition (Jones, 2024c), and the effect of feedback timing on vowel acquisition (Jones, 2024e). Additionally, a conference paper discusses potential methods for small sample analysis, particularly Bayesian methods (Jones 2024a), which uses Jones and Blume (2022) as an example, but which is also a rationale for the analysis methods in all three of the quantitative articles (Jones & Blume, 2022; Jones 2024c and 2024e).
The rationale behind the studies is that L2+ listening is a difficult skill and partly this is due to many learners having incomplete English phonology development. This is particularly the case in the author's context, English for Academic Purposes (EAP) for English Medium Instruction (EMI) in Japanese higher education. Many L1 Japanese, and other L1 Asian language learners of English, experience problems due to the larger vowel inventory of English. Such a difference in vowel inventory means that vowels are likely categorised as L1 equivalents, which may cause difficulties when two different English vowels are categorised as examples of the same L1 vowel.
An overview is given of the different models that inform the theoretical scaffold of the dissertation. The Perceptual assimilation Model (PAM; Best 1995) and PAM-L2 (Best & Tyler, 2007) are discussed, with their relevance for L2+ learners of English, as well as being a model which assumes all language categorisation occurs within the same template, i.e. all language categories are stored centrally and used across all languages. The Speech Learning Model (SLM; Flege, 1995) and its revision, SLM-r (Flege & Bohn, 2021) is discussed as an example of what occurs with regard to high achieving learners; additionally, the SLM and SLM-r are of interest because they are models of phonetic rather than phonemic based learning. A further point of interest is that the SLM claims that first and second languages affect one another over time, with phonetic categories of L1 and L2 blurring. The Second Language Linguistic Perception model (L2LP; Escudero, 2005) examines the L2 state as it develops over time with exposure to L2 learning activities, ultimately culminating in an end state, and being able to be compared to the optimal state of language learning the L2(+). A further model, the Natural Referent Vowel framework (NRV; Polka and Bohn, 2011) explains why some vowels, in particular vowels with
extremes of openness, closedness, and tongue protrusion/extrusion, are more readily learned than others.
A description of the methodologies of the studies used in producing the included articles is given: essentially following a pretest-post-test structure, although one study used a pretest-post-test-delayed post-test structure. As mentioned above, Bayesian analysis methods were used to analyse the data obtained. The reason for Bayesian methods, as opposed to more commonly used frequentist statistical analyses, is that Bayesian methods are more suited to the smaller sample sizes that are common in applied linguistics, and particularly classroom-based research. It is noted that the nature of all of this research is exploratory, and the case is made that this should be the default state of applied linguistics research, particularly classroom-based research, and that null hypothesis statistical testing is an ill-advised process with novel research (see also Trafimow, 2024).
General findings are that the use of Global Englishes as opposed to prestige varieties has no effect on phonology acquisition (Jones & Blume, 2022) or listening difficulty (Jones, 2024d). Additionally, preliminary evidence is found that suggests a link between visual salience and perceptual vowel acquisition (Jones, 2024c). However, the Bayes Factors in the statistical analysis do not suggest strong evidence of a connection, therefore it is suggested that further research is required. Feedback timing is found to have a role in acquisition of L2+ vowels (Jones, 2024e), with delayed feedback being more beneficial for learning. Furthermore, a coincidental finding across the studies is that crammed learning appears to affect vowel acquisition, and has an interaction with learning of all target vowels
and certain vowels depending upon the learning interventions.
The findings above are used to construct and/or support theories, such as a Bayesian model of phonology learning, where prior input is used as a reference when L2+ listening takes place. Such prior input is constantly revised, and thus parsing and listening skills develop. The dissertation concludes with the contributions to research and language teaching pedagogy and avenues for further research based upon the findings within the dissertation, some of which the author is undertaking as the dissertation goes to press.
Description
Table of contents
Keywords
Computer assisted language learning, English, Listening, Second language
Subjects based on RSWK
Phonologie, Fremdsprachenlernen, Modalität (Linguistik), Hören