1. Introduction
Published in Language Learning in 1995, Munro and Derwing's* investigation of foreign accent, comprehensibility, and intelligibility in second language (L2) speech instigated significant change in L2 pronunciation research (Levis, Reference Levis2020). A key finding was that despite the presence of a foreign accent, listeners could indeed comprehend L2 speech. Within their framework, comprehension of L2 speech was assessed along two dimensions. The first, intelligibility, was an assessment of actual listener comprehension, measured through listener transcriptions of a given utterance. The second, comprehensibility, was a scalar measure of how easy to understand listeners perceived an utterance to be. While these two measures of listener comprehension (i.e., understanding) have been shown to correlate, they have also been shown to measure different forms of understanding (Derwing & Munro, Reference Derwing and Munro2015). This means that while increased intelligibility can be associated with increased comprehensibility, it is still common for listeners to accurately transcribe nonnative speech while simultaneously indicating the speech to be hard to understand (i.e., evidence indicates that intelligibility outpaces comprehensibility in the development of L2 pronunciation). Research published in the 25 years since has repeatedly demonstrated that accentedness, comprehensibility, and intelligibility are partially independent dimensions of L2 speech (Munro & Derwing, Reference Munro and Derwing2020). Whereas Munro and Derwing's (Reference Munro and Derwing2011) research timeline attended to the concepts of accent and broad intelligibility (i.e., inclusive of both actual and ease of understanding), the past decade has seen an increased scholarly emphasis specifically on the global speech dimension of comprehensibility. Given this increased scholarly interest, our timeline is presented with the goal of tracing the post-Munro and Derwing (1995Footnote *) development of L2 speech comprehensibility research.
Comprehensibility is most commonly defined in reference to listeners’ perceived ease or difficulty of understanding a given utterance (Derwing & Munro, Reference Derwing and Munro2015). In this sense, understanding has been associated with the degree of processing difficulty experienced by the listener. For example, Munro and Derwing (1995b*) found that listeners demonstrated longer reaction times in a truth judgment task for less comprehensible speech (see also Ludwig & Mora, 2017*). While our focus on comprehensibility begins with Munro and Derwing (1995a*), we briefly note here several earlier studies that have been influential in our contemporary understanding of comprehensibility.Footnote 1 This earlier influence can most clearly be seen in Varonis and Gass (Reference Varonis and Gass1982), who defined comprehensibility as ‘how easy it is to interpret the message’ (p. 125). Their 5-point scale included end points of ‘I understand the sentence easily’ and ‘I didn't understand this sentence at all’ (p. 126). Interestingly, in a later study, Gass and Varonis (Reference Gass and Varonis1984) would again refer to the dimension of comprehensibility, though their use of listener transcriptions was more characteristic of contemporary measures of intelligibility (see Kang et al., Reference Kang, Thomson and Moran2018). Scalar approaches towards assessing L2 speech remained common prior to Munro and Derwing (1995a*), and indeed had been in use well before for a range of sociolinguistic interests (e.g., Lambert et al., Reference Lambert, Hodgson, Gardner and Fillenbaum1960). Yet, most late-1980/early-1990 studies emphasized listeners’ perceptions of foreign language accent (e.g., Flege, Reference Flege1988; Flege & Fletcher, Reference Flege and Fletcher1992), or in the case of Anderson-Hsieh et al. (Reference Anderson-Hsieh, Johnson and Koehler1992), a construct that combined both accent and intelligibility into a single scale. Utilizing a definition similar to that presented in Varonis and Gass (Reference Varonis and Gass1982), Munro and Derwing (1995a*) firmly established comprehensibility in the sense that it is generally investigated today.
Before transitioning into the specifics of our timeline, we feel it is important to make clear the importance of previous and ongoing research into different dimensions of L2 speech. The partial independence of comprehensibility, intelligibility, and accentedness has served as the basis for proposed revisions for both how L2 pronunciation is taught and assessed. Levis (Reference Levis2005) described both a Nativeness Principle and an Intelligibility Principle as present in language curriculums and discussed the need for pedagogical practices and materials that better reflect both researchers’ and teachers’ increasing recognition of intelligible over nativelike speech as the target of L2 pronunciation learning (see Levis, Reference Levis2020, for a more elaborative discussion). Levis's proposal has served to motivate much of the research into comprehensibility over the past decade. Echoing this pedagogical shift from nativelike speech to intelligible speech, recent scholarship aiming to bridge second language acquisition (SLA) and assessment-focused research has extensively discussed the importance of emphasizing intelligible over nativelike speech in pronunciation assessment (e.g., Isaacs & Trofimovich, Reference Issacs and Trofimovich2017; Kang & Ginther, Reference Kang and Ginther2018; see also Isaacs & Harding, Reference Isaacs and Harding2017, whose research timeline on pronunciation assessment includes multiple references to comprehensibility). As we continue to push intelligible speech as a target of pronunciation teaching and assessment, it is necessary to understand where the field currently stands in reference to understanding relevant dimensions of intelligible speech. In the current timeline, this relevant dimension is comprehensibility.
2. Timeline methodology
The current timeline uses Munro and Derwing (1995a*) as a starting point, following up on the Journal of Second Language Pronunciation's 2020 special issue entitled ‘25 years of intelligibility, comprehensibility and accentedness’. We present 41 empirical studies drawn from 14 peer-reviewed applied linguistics journals,Footnote 2 as a means to show how research into comprehensibility has both grown and transformed our understanding of the dimension. In line with common practices in meta-analytic research (e.g., Marsden et al., Reference Marsden, Morgan-Short, Thompson and Abugaber2018; Plonsky et al., Reference Plonsky, Marsden, Crowther, Gass and Spinner2019), we here focus on empirical journal publications rather than book chapters, dissertations, theses, and conference proceedings to ensure included research has been through the journal peer review system. An initial review conducted by the first author identified 99 empirical studies published between 1995 and 2020 that included a measure of comprehensibility (spread across 19 journals). Only 21 studies were published prior to 2011, showing the significant growth in scholarly interest over the last decade (N = 78). We additionally considered post-2020 publications published during the drafting of our timeline.
Given space constraints, and in keeping with the essence of this strand in Language Teaching, we selectively identified 41 studies that demonstrate how contemporary research builds on, and is enriched by, that which preceded it, or which provide a novel approach towards investigating and understanding comprehensibility. Though we have strived to include a range of scholars to the extent possible, a review of the timeline will clearly reveal several prevalent names post-2010 (in addition, of course, to numerous studies published by Murray Munro, Tracey Derwing, and colleagues). Such inclusions are necessary to best represent the growth in comprehensibility research over the past 25 years.
3. Existing trends in L2 speech comprehensibility research
As previously stated, the 41 studies in our timeline either build on previous comprehensibility research or take a novel approach to investigate comprehensibility. We specifically highlight several key themes we see as prevalent across 25 years of comprehensibility research, each of which we summarize below.
As articulated in Levis (Reference Levis2005), both speaker and listener are essential elements in the production of intelligible speech, and thus it is no surprise that both have received focus in L2 comprehensibility research. Whereas Munro and Derwing (1995a*) featured one set of listeners (native English) assessing the speech of one set of speakers (first language (L1) Mandarin learners of English), subsequent research has considered how varying both speaker L1 (e.g., Crowther et al., 2015a*; Derwing & Munro, 1997*) and listener L1 (e.g., Foote & Trofimovich, 2018*; Munro et al., 2006*) may inform comprehensibility ratings. Beyond potential L1 effects, speakers’ degree of comprehensibility appears tied to a range of speaker-based variables, including length of residence (Saito, 2015*), motivation (Nagle, 2018*; Saito et al., 2018*), and aptitude (Saito et al., 2019*).
Extending directly out of listener-based considerations, and given the strong support that comprehensibility was a partially independent dimension from accentedness (e.g., Derwing & Munro, 1997*; Munro & Derwing, 1995a*, b*), attention turned to understanding what listeners attended to when rating comprehensibility. While initial research focused on segmental and suprasegmental influences (e.g., Munro & Derwing, 2006*; Kang, 2010*; Kang et al., 2010*), subsequent research revealed that listeners attended to not only phonological but also lexicogrammatical characteristics of speech (e.g., O'Brien, 2014*; Trofimovich & Isaacs, 2012*), with a strong emphasis on lexical use (Saito et al., 2016b*). However, the linguistic profile that informs listeners’ perception of L2 comprehensibility appears to vary depending on a range of variables, including speaker L1 (Crowther et al., 2015a*), listener status (e.g., Foote & Trofimovich, 2018*; O'Brien, 2014*), task (e.g., Bergeron & Trofimovich, 2017*; Crowther et al., 2015b*), and various individual differences (e.g., Saito et al., 2019*).
Of additional interest has been, of course, understanding to what extent speakers are able to improve the comprehensibility of their speech, and what effect pronunciation instruction may have on this development. Longitudinal research, ranging from a few semesters of study (e.g., Saito et al., 2019*) through to seven years of immersive contact (see Derwing et al., Reference Derwing, Thomson and Munro2006, Reference Derwing, Munro and Thomson2008, 2013*) has shown that comprehensibility can indeed continue to develop, though such development seems highly tied to the extent to which the language is used (Derwing et al., 2013*; Zielinski & Prior, 2020*). Extending back to Derwing et al. (1998*), pronunciation instruction has fairly consistently been shown to benefit the development of comprehensible speech, with positive benefits found for instruction provided in both second and foreign language contexts (Zhang & Yuan, 2020*), as well as for both English and non-English languages (e.g., Derwing et al., 1998*; Isbell et al., 2019*).
Drawing from the above discussion, we highlight the following themes throughout our timeline:
• SPE – Speaker-based variables
• LIST – Listener-based variables
• LC – Linguistic correlates of comprehensibility
• LONG – Longitudinal development
• PED – Pedagogical inquiry
• MTD – Methodological practices
Studies thematically labelled as MTD explicitly identified as investigating/validating some aspect of methodological practice in comprehensibility research. In this sense, the types of scales used to rate comprehensibility have been the focus of methodological inquiry, whether as a means to validate current rating procedures (e.g., Isaacs & Thomson, 2013*; O'Brien, 2014*), or to introduce alternative rating options (Saito et al., 2017*). In recent years, there has been a push to investigate listeners’ perception of comprehensibility as dynamic, rather than static (e.g., Nagle et al., 2019*), as well as in interactive, as opposed to monologic, performance (Trofimovich et al., 2020*). The use of MTD for specific studies does not indicate that additional methodological innovation does not exist elsewhere in our timeline, but only that such studies did not forefront their innovation. To help with framing comprehensibility within the larger body of global speech research (see Munro & Derwing, Reference Munro and Derwing2011), we make note of additional global dimensions of speech considered in each study:
• ACC – Accentedness
• INT – Intelligibility
• FLU – Fluency
Though not included in Munro and Derwing (1995a*), interest in perceived fluency, or the ease of flow of L2 speech, typically in reference to the presence/absence of pauses and other dysfluency markers (Derwing & Munro, Reference Derwing and Munro2015), has since often been considered in relationship to comprehensibility (e.g., Derwing et al., 1998*; Suzuki & Kormos, 2020*), and is thus considered in this timeline similarly to accentedness and intelligibility. Finally, given that ~80% of identified studies featured English as the target language, we have explicitly identified studies in which a non-English language was the target of analysis.
Dustin Crowther is an Assistant Professor in the Department of Second Language Studies at the University of Hawaiʻi at Mānoa. His research agenda emphasizes the attainment of intelligible speech for additional language speakers, inclusive of speaker- and listener-based variables. Specifically, he takes into account linguistic and intercultural considerations that define native-nonnative and nonnative-nonnative interaction. Given the increased spread of English, much of his research is informed by scholarship derived from Global Englishes. As an experienced language instructor, his long-term scholarly objective is to link research to pedagogy. Dr. Crowther additionally emphasizes the promotion of methodological rigor within applied linguistics research.
Daniel Holden is currently a Ph.D. candidate in the Department of Second Language Studies at the University of Hawaiʻi at Mānoa. His main research interests are critical administration, second language teaching and learning, language assessment, materials development, and the impact of edu-business and globalization of international English programs. His long-term research goal is to foster a stronger connection between administrators and the members of their school communities through critical dialogue. He has most recently been doing assessment-related research for a short-term intensive English program, focusing on the revision of its writing and speaking placement exams.
Kristen Urada is a Ph.D. student in the Department of Second Language Studies at the University of Hawaiʻi at Mānoa. Her main research interests are instructed second language acquisition, task-based language teaching, Hawaiʻi Creole, and speech perception. As a teacher-researcher, she aims to provide research-informed pedagogy, thus supporting the linkage of research to pedagogy, as well as making research accessible to the multilingual community through the Multiʻōlelo initiative (https://multiolelo.com).
