|Douglas Biber, Regent’s Professor, English Department, Northern Arizona University|
Corpus analysis of spoken discourse: Research findings, prospects, implications for teaching
This talk introduces corpus-based analyses of spoken discourse, comparing and contrasting this research approach to other empirical studies of speech, illustrating the kinds of research findings that are possible through corpus-based analysis, and discussing implications for pronunciation teaching. The talk begins by surveying the research goals and methods of corpus linguistics. The ‘corpus’ is a collection of natural texts that ‘represents’ a target discourse domain. This requirement of ‘representativeness’ is the first consideration that distinguishes corpus linguistics from research carried out in related sub-disciplines.
But perhaps most importantly, research in corpus linguistics is distinguished by analyzing language variation in natural contexts of use, in ways that can be generalized. For example, many corpus-based descriptions have focused on register and dialect variation, comparing and contrasting the patterns of use found in different spoken varieties. But at a more specific linguistic level, a second important contribution for pronunciation research is the application of corpus-based analysis to study variation in the prosodic realization of particular speech acts.
Historically, most corpora of spoken English contain texts that have been orthographically transcribed, enabling detailed descriptions of grammar and vocabulary but no analysis of phonetic characteristics. More recently, though, there have been a number of projects to build spoken corpora with prosodic annotations, enabling descriptions of linguistic variation that are directly relevant for pronunciation teaching.
Specific case studies are presented that illustrate the strengths of corpus analysis for pronunciation research and teaching. In conclusion, the talk discusses how the results of such research can inform the teaching of pronunciation, and calls for future research efforts to provide the basis for such innovations in pedagogical practice.
|Tracey Derwing, Professor Emeritus of TESL, Dept of Educational Psychology, University of Alberta|
|Murray J. Munro, Professor, Department of Linguistics, Simon Fraser University|
Collecting Data in L2 Pronunciation Research
Obtaining high quality data in L2 pronunciation research requires careful attention to details at multiple levels. Over the years we have conducted a variety of pronunciation studies; we look forward to sharing the benefits of our experience with you. We will limit our focus to the types of research designs and data types that we ourselves are familiar with. In particular, we will explore approaches to measuring the constructs of intelligibility, comprehensibility, accentedness, and fluency, as we have defined these dimensions. We will begin with a consideration of making good speech recordings, followed by the preparation of audio materials for listening tasks. Rating and other judgment tasks will be discussed, as will the effective administration of quasi-experimental listener tasks. We will also touch briefly on data analysis and interpretation. Finally you will engage in some hands-on activities; please bring a laptop or arrange to share one with a colleague. Sample materials will be provided, including Language Background Questionnaires, PRAAT scripts, and consent forms. On completion of the workshop, participants will have developed skills in designing and executing L2 pronunciation studies. This workshop requires advance reading of the article listed below; it is freely available at the link indicated (copy into your browser). Munro, M. J. & Derwing, T. M. (2015). A prospectus for pronunciation research methods in the 21st century: A point of view. Journal of Second Language Pronunciation, 1, 11-42.
|Eric Friginal, Associate Professor, Department of Applied Linguistics and English as a Second Language, Georgia State University|
Compiling and Annotating Spoken Corpora
This workshop overviews the process of designing, compiling, and annotating spoken corpora, in particular, English-based intercultural workplace interactions in settings such as outsourced call centers (business telephone transactions), pilot-air traffic controller communications (Aviation English radio-telephony), and office interactions with workers who use augmentative and alternative communication (AAC) devices and their interlocutors. I will present a summary of traditional corpus-based approaches to text transcription and annotation, advancements in Multimodal Annotation (e.g., Gu, 2008; the ITACorp Project at Penn State University), and current/future approaches and directions. The annotation of spoken corpora for prosody, for example, the Hong Kong Corpus of Spoken English (HKCSE) (Cheng, Greaves, & Warren, 2008) and more detailed contextual transcriptions and annotations of spoken texts paint a promising future for capturing some socio-phonetic features of speech in orthographic transcripts. Although not necessarily considered corpora in the traditional sense, available databases of speech that are designed to be analyzed phonetically, phonologically, or acoustically point to a possible framework for developing a phonetically-annotated corpus. For example, the Speech Accent Archive (http://accent.gmu.edu/) (Weinberger, 2013), currently with over 2,000 speech samples, is an online database of speakers from around the world reading aloud a short paragraph in English. The audio samples are then transcribed phonetically using the International Phonetic Alphabet (IPA), forming a “corpus” of IPA-transcribed texts. By using crowdsourcing techniques, various speakers can also send submissions of their speech patterns and accents digitally. The use of computational tools, dictation and transcription software, qualitative coding programs, and automated sentiment analyzers utilized in customer service will be demoed and discussed.
Language Learning Roundtable Speakers
|Naoko Taguchi, Professor of Japanese and Second Language Acquisition, Department of Modern Languages, Carnegie Mellon University|
The prosody-L2 pragmatics intersection: Can prosody be an indicator of pragmatic competence?
Second language (L2) pragmatics, a field within SLA, studies L2 learners’ ability to perform a communicative act in a socio-culturally appropriate manner and how that ability develops over time (Taguchi & Roever, 2017). While this definition is relatively uniform, there are different understandings as to what counts as evidence for pragmatic development and how to assess such development. Traditionally, pragmatic competence has been assessed at pragma-linguistic and discourse-level. Researchers have focused on what linguistic forms learners use to convey their intention; whether their linguistic choice is appropriate in a given context; and how they can adapt their linguistic resources to changing discourse and co-construct meaning in interaction. Going beyond the linguistic and discourse-level analysis, I will discuss the potential of prosody as an indicator of L2 pragmatic competence. I will present a survey of a small amount of existing research on L2 learners’ use and understanding of prosody in speech acts and irony. Following this, I will present a portion of a cross-sectional study (Kermad, Kang, & Taguchi, 2017), which compares two ESL groups of different proficiency and a native English speaker group for their use of rising tone in spoken speech acts (requests and opinions). I will illustrate how tone and prominence can differentiate the groups and how the differences compare with those from other indicators of pragmatic competence analyzed in the study (i.e., native speaker ratings of speech acts; types of linguistic forms in speech acts).
|Shawn Loewen, Professor, Second Language Studies, Michigan State University|
Instructed Second Language Acquisition and Pronunciation
Traditionally, pronunciation has not received as much attention in instructed second language acquisition (ISLA) research as have other areas of language such as grammar and vocabulary (Derwing & Munro, 2005). However, that oversight is quickly being addressed as numerous recent studies have investigated classroom pronunciation learning (e.g., Lee, Jang & Plonsky, 2014). Questions that ISLA researchers are asking include:
- Is instruction effective for pronunciation development?
- If so, are some types of instruction more effective than others?
- Are some elements of pronunciation more amenable to instruction than others?
In general, ISLA researchers respond affirmatively to the first question; instruction can make a difference for second language (L2) learners’ pronunciation. Types of instruction that have been investigated range from explicit instruction, including information about points of articulation and crosslinguistic influences, to implicit instruction through recasts or other types phonological information provided during meaning-focused interaction. In terms of the target of instruction, both segmental (i.e., individual consonants and vowels) and suprasegmental features (e.g., intonation, stress, pitch) have been investigated. Additionally, research has investigated the effects of targeting one phonological feature intensively versus multiple phonological features. One issue that has yet to be explored in detail is the relationship between pronunciation and other aspects of language. For example, when morphosyntactic and phonological features overlap, such as with the pronunciation of the past tense English morpheme –ed, should errors be treated as a phonological or grammatical? What are the implications of each interpretation? Finally, although some research has investigated L2s such as Spanish and Mandarin, the majority of research has examined L2 English.
|Donald Rubin, Professor Emeritus, Departments of Speech Communication and Language and Literacy Education, University of Georgia|
UPDATE: 8/21/19 We regret to announce that Dr. Rubin will not be able to join us at PSLLT this year. Below please find a message from Dr. Rubin.
Don Rubin deeply regrets that he is unable to attend PSLLT 2019. Were he able to attend, Don would be asking the group to consider the consequentiality of our teaching and research about pronunciation. The history of language and attitude research—which examines both linguistic stereotyping based on accent and now also reverse linguistic stereotyping—is long. While not all these studies reach identical conclusions (that’s a good thing!), there remains little doubt about the big picture: Listeners’ perceptions of a speaker’s accent shape all kinds of social judgments about that speaker. Conversely, listeners’ prior beliefs or knowledge about a speaker’s social identity shape the way they perceive those accents. What are some real world consequences for teaching pronunciation? At very least, teachers need to complement technical pronunciation instruction with instruction about inoculating listeners against negative stereotyping related to non-mainstream accents. What are the consequences for research on pronunciation? Let’s extend more research beyond the walls of the classroom into cross-cultural settings such as immigration hearings, patient-provider healthcare interactions, and customer service encounters to assay the impact of language attitudes on conversations that make a material difference in language learners’ lives.
|Lucy Pickering, Professor, Department of Literature and Languages, Texas A&M Commerce|
The role of prosody in International Communication in English: Data from a call center corpus
Corpus linguistics is a powerful tool for investigating issues in pronunciation teaching and learning. This talk will focus on the context of call center interactions to illustrate the ways in which corpus methods can be used for analysis of and training for pronunciation. Public opinion in the U.S. and the U.K. regarding the perceived “pronunciation problems” of agents based in call centers in Outer Circle English-speaking countries (e.g., India and the Philippines) is typically negative (e.g., Forey & Lockwood, 2007). However, it is often difficult for researchers to pinpoint the specific issues involved as access to sound files of authentic calls is scarce due to their proprietary nature. This paper examines the role of the understanding of prosodic conventions (e.g., the use of intonation and rhythm) in call center interactions recorded in the Philippines as part of a corpus collected by Eric Friginal in 2006 (Friginal, 2009). Analysis suggests that where conflict occurs, it is mirrored in these linguistic features, and this has important implications for models of intelligibility and training for high-stakes interactions.
|Talia Isaacs, Associate Professor, UCL Institute of Education, University College London|
Pronunciation Assessment and Testing
This presentation centers on trends in second language pronunciation research, teaching, and assessment by highlighting the ways in which pronunciation instructional priorities and assessment targets have shifted over time, social dimensions that, although presented in a different guise, appear to have remained static, and principles in need of clearer conceptualization. The reorientation of the goal in pronunciation teaching and assessment from the traditional focus on accent reduction to the more suitable goal of intelligibility will feed into a discussion of major global constructs often subsumed under the umbrella term of “pronunciation.” The talk will overview research and developments related to human judgments of pronunciation, lingua franca communication, and the consequences of machine-driven scoring. Finally, recommendations for advancing an ambitious research agenda will be proposed to disassociate pronunciation assessment from the neglect of the past, bring it in line with the more voluminous work on pronunciation teaching and learning, and propel it to the forefront of developments in our field.