The Nordic Network for research on pronunciation and pronunciation teaching in a second language perspective, NNL2P, connects researchers and PhD candidates with a special interest in pronunciation and pronuciation teaching.

Two people talking. Photo credit Priscilla Du Preez/Unsplash.

The conducted research is set in a Nordic context where second language learning/teaching of the Nordic languages and English are included.

Today, about 50 researchers are connected to the network, which intends to meet once a year in one of the Nordic countries. On June 14-15 2021 a digital gathering hosted by Linköping University was organised.

Conference 2021

This year we had the pleasure to listen to Alice Henderson, Université Grenoble Alpes, Switzerland, John Levis, Iowa State University, USA, Jacques Koreman, Norwegian University of Science and Technology, Norway and Ocke-Schwen Bohn, Aarhus University, Denmark.

Conference 2021

Programme for the conference held on 14-15 June 2021

Read or download the conference programme (PDF)

Read more about the main speakers and their presentations.

Ocke-Schwen Bohn: Core aspects of the revised Speech learning Model (SLM-r)

Portrait on Ocke-Schwen Bohn.Ocke-Schwen Bohn is the professor of English Linguistics at Aarhus University, Denmark. He received his PhD from Kiel University (Germany) and spent his time as a post-doc working with James Flege at the University of Alabama in Birmingham. With funding from German and Danish research agencies, and in collaboration with American, Canadian, and Australian colleagues, Bohn’s research focuses on the causes and characteristics of foreign accented speech, speech perception (in infants, cross-language, and second language acquisition), and bilingual memory.

About the presentation

Over the past 25 years, much of the research on L2 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.

Like its predecessor, the revised SLM (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 because the mechanisms and processes that function “perfectly” in L1 acquisition do not yield the same results in L2 acquisition.

The SLM-r models 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 (e.g., child vs adult 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.

Alice Henderson: Teaching English pronunciation in a communicative framework: Principles, tools and techniques

Alice Henderson laying in the green grass with a book. Mountins are towering behind her.Alice Henderson is an Associate Professor at Université Grenoble - Alpes, France, where she now teaches English for Specific Purposes, after teaching English phonetics and phonology for 24 years at the nearby Université Savoie-Mont Blanc. In 2009 she initiated the international conference English Pronunciation: Issues & Practices. She has published internationally on English pronunciation learning and teaching, the perception of foreign-accented speech, and English Medium Instruction (EMI), and has been involved in teacher training in France, Norway, Poland, and Spain.

About the presentation

Language teachers often mention needing more time or knowledge in order to truly integrate pronunciation teaching in their classrooms. The major objective of this ‘virtually hands-on’ workshop is for participants to feel more comfortable selecting and adjusting exercises, to meet their learners’ needs. To this end, we’ll actively explore how Celce-Murcia et al.’s (2010) framework can help teachers to organize a communicative approach to pronunciation teaching. Although most of the examples will be from English, it is applicable to any language. 

This framework is implicitly aligned with Levis’ (2005) Intelligibility Principle, whereby pronunciation exercises should aim to achieve intelligible pronunciation, but it is also useful for classes based on the Nativeness Principle, where achieving nativelike pronunciation is the goal, e.g., in certain countries’ examinations for future language teachers. 
The workshop includes time for discussion and for partner activities (in breakout rooms) on both segmental and suprasegmental features of English. Participants will work through exercises which they can then use or adapt for their own classrooms, and will receive an annotated “Further Reading” list.

  • Celce-Murcia, M., Brinton, D., Goodwin, J. M., & Griner, B. (2010). Teaching pronunciation: A course book and reference guide (2nd ed). Cambridge University Press.
  • Levis, J. M. (2005). Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly, 39(3), 369–377.

Jacques Koreman: L2 pronunciation training from a multilingual perspective

Portrait on Jacques Koreman.Jacques Koreman is a professor of Phonetics at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway. He is interested in experimental phonetics and speech technology, but has in the past ten years mainly focused on applied research. Together with his team, he has developed the Computer-Assisted Listening and Speaking Tutor (CALST). CALST is an online, multilingual platform for pronunciation training.

About the presentation

The Computer-Assisted Listening and Speaking Tutor (CALST) is a platform for multilingual pronunciation training which covers several Norwegian dialects as well as British English. It is presently being extended with Greek, Spanish, and Italian.

The organization into exercises is based on phonemes, since these are important to distinguish words in the target language. Using L1-L2map to compare the target language with the native language (based on an extended version of UPSID), exercises are tailored to the learner. The advantages and disadvantages of this process will be discussed. Exercises contrast a sound to other sounds in the target language which differ in one phonetic dimension (as represented in the IPA tables). The exercises in CALST capture the allophonic variation in the target language. Since the correspondence between writing and pronunciation is important in literate societies, we also ensure that this is reflected in the exercise material.

Besides exercises for individual sounds, CALST also offers exercises for consonant clusters, word stress and lexical accent. If permitted by the learner, user data are logged. Logged data form the basis for a further tailoring of the exercises in CALST. In the presentation, the above principles will be explained using Greek, Spanish and Italian data.

John Levis: L2 pronunciation research and teaching: The importance of many languages

John Levis is Angela B. Pavitt Professor of English at Iowa State University. He is the author of Intelligibility, Oral Communication, and the Teaching of Pronunciation (Cambridge University Press). He initiated the Pronunciation in Second Language Learning and Teaching Conference and is founding editor of the Journal of Second Language Pronunciation.

About the presentation

Over the past several decades, the research and teaching of second language (L2) pronunciation has frequently been described as a neglected area, but it is now quickly developing as an interdisciplinary field. However, despite large changes in the visibility of L2 pronunciation, it continues to be dominated by research into English with far less research on other languages.

Questions related to issues of intelligibility and comprehensibility, longitudinal acquisition, individual differences among learners, functional load and many other topics rely on research and theorizing largely based on English as an L2. Likewise, journal articles, professional books, and even conferences are overly filled with varied approaches to pedagogy that are directly relevant to English, and researchers and teachers interested in other L2s are often forced to apply findings from English to their own contexts. We thankfully see increasing types of L2 research in other languages such as Chinese, Spanish, Swedish, and Japanese, but for the field of L2 pronunciation to continue progressing, we need to take a wider view of the questions we ask.

This talk argues that L2 pronunciation research must not only broaden its focus to varied L2 learning contexts but that is also needs to explicitly connect L2 pronunciation research to theories related to general areas of interest such as phonological acquisition and learning, factors related to the effectiveness of instruction, and speech intelligibility, so that research on multiple L2s can inform how pronunciation learning occurs more generally. This also means that the training of teachers should incorporate information about these basic questions in the field into pedagogical decision making.

Read more about the speakers and their presentations.

Typological rarity and apparent similarity to L1 as sources of difficulty for the development of pronunciation in Swedish as a second language

Memet Aktürk-Drake & Federica Raschellà, Uppsala University

In this cross-sectional study, we investigated the pronunciation development of 26 learners with Italian as their L1 and varying experience with Swedish as a second language (SSL). We divided the participants into three experience-based groups: beginners, speakers on an intermediate level and experienced speakers. We focused on accuracy of pronunciation in 8 vowels and 7 consonants. We investigated how different sources of difficulty such as typological rarity (PHOIBLE) as well as apparent phonetic (Flege 1995) and orthographic similarity to L1 interacted with experience. The participants pronounced single words in a picture-naming task. The pronunciation data were analysed auditively by two judges with Swedish and Italian as their respective mother tongues. The results suggest that increasing experience leads to more accurate segment pronunciation overall, but that different segments benefit from experience to different degrees depending on the source of difficulty. While orthographic similarity and typological rarity seem to complicate the acquisition of pronunciation for beginners, it is, instead, phonetic similarity that can account for the experienced speakers’ remaining difficulties. Since previous studies have shown that segment pronunciation is important for perceived nativelikeness in SSL (Kuronen & Zetterholm 2017), we conclude that pronunciation teaching needs to focus more on phonetic similarity.

Comprehensibility and accentedness of L2 English spoken by Finns: Key findings

Elina Tergujeff, University of Jyväskylä

This presentation is focussed on the key findings of a three-year project on L1 Finnish and L1 Finland-Swedish speakers’ L2 English that aimed to shed light on, e.g. (1) how English-speaking listeners ratethe speakers’ L2 English for comprehensibility and accentedness, (2) which speech features are linkedwith comprehensibility and/or accentedness, (3) the relationship between overall oral proficiency andcomprehensibility/accentedness, and (4) possibility of the shared language benefit applying to L2comprehensibility and accentedness.The project made use of speech materials that were gathered for the purposes of a national languageassessment. A selection of the materials were subjected to comprehensibility ratings with Englishspeakinglisteners and listeners who share the speakers’ L1. In addition, the speech samples wereanalysed for various speech features to seek connections between the features and the ratings. Theresults support earlier findings on the role of fluency and segmental accuracy for L2 comprehensibilityand accentedness, and the shared language benefit being language dependent. In addition, newknowledge was obtained about the relationship between overall oral proficiency andcomprehensibility/accentedness. The results imply that fluency features might deserve a strongerfoothold within language teaching and assessment.

MultiFluency: Multilingual speakers’ fluency across Swedish, Finnish, and English

Pauliina Peltonen & Pekka Lintunen, University of Turku

Speech fluency has been widely studied as a component of L2 oral proficiency. However, studiescomparing L1 and L2 fluency from the same speakers are relatively rare, and multilingual learners’fluency has been studied even less. In the present study, we examined multilingual speakers’fluency across Swedish, Finnish, and English among Swedish-speaking (Group 1) and Finnishspeakinguniversity students (Group 2) in Finland. We focused on the differences in fluencybetween the groups and the connections across the participants’ L1 and L2 productions fromquantitative and qualitative perspectives.The study is part of a project “MultiFluency”. From the project data set, 90 monologue speechsamples in Finnish (L1/L2), Swedish (L1/L2), and English (L2) from the two groups (N=30) wereanalyzed for temporal fluency (e.g., articulation rate; silent pause frequency, duration, andlocation) and stalling mechanisms (e.g., filler words, drawls). The results demonstrated minordifferences between the groups regarding Finnish (L1/L2) and English (L2) fluency, but cleardifferences in Swedish (L1/L2) fluency. L1 speech fluency was also found to correlate with L2fluency, but the connections varied across languages and depending on the participants’ L1.

Beliefs about Orthographic Diacritics: Effects on L2 vowel acquisition

Mara Haslam, Stockholm University

 In Swedish the letters “å” “ä” and “ö” are considered separate letters of the alphabet from “a” and “o”, but English speakers often seem to have trouble distinguishing the vowel sounds that these letters represent in perception and production. This study hypothesizes that difficulty in vowel acquisition is connected to the fact that English speakers do not see “å”, “ä”, and “ö” as separate letters of the alphabet from “a” and “o”. Polish has a similar vowel inventory to English in relation to Swedish, but written Polish uses letters with diacritics, while written English uses unadorned letters of the Latin alphabet. In this study native Polish speakers and native English speakers are taught Swedish words containing the vowel sounds in question along with their orthographic representations, then tested on their perception and production of the target words. In addition, their beliefs about whether certain orthographic characters are the same or different are tested. Their results will also be compared to the performance of native Swedish controls. Previous results show that Polish speakers outperformed English speakers on vowel acquisition and that they were also more likely to indicate that they believed “å”, “ä” and “ö” were distinct letters from “a” and “o”. This new set of data will include speakers at different proficiency levels so that the effect of proficiency may also be explored.

Towards a targetlike L2 Swedish intonation: Pitch accents and boundary tones of L1 Spanish speakers

Johanna Einarsson, Stockholm University

A number of studies have stressed the important role of prosody for a well-functioning communication in the L2 (e.g. Anderson-Hsieh et al., 1992; Hahn, 2004). L2 prosody is also the subject of my ongoing PhD project, in which I look at the intonational system of L1 (Peninsular) Spanish speakers of Swedish. I examine the speakers’ production of pitch accents and boundary tones in three sentential contexts: declaratives, polar questions, and pronominal questions. In-depth profiles on the speakers’ L2 intonation are made in a qualitative study. The materials consist of read and (semi)spontaneous speech produced by nine experienced L2 speakers. Read elicited sentences produced by three of them (FP1–FP3) are focus materials. Their L2 Swedish phonological performance is at various levels: FP1 (non-nativelike), FP2 (partially nativelike), and FP3 (near nativelike).
Results show that the L2 Swedish intonation differs considerably between the subjects. For example, FP1’s speech is characterized by a tangible influence from the L1. FP3, on the other hand, has acquired the distinction between the two prominence levels in Swedish as well as the tone accent distinction, yielding strikingly targetlike forms.

Acceptability in Finnish and Danish

Lisa Tulaja, Leena Maria Heikkola & Antti Saloranta, Kiel University, Åbo Akademi University & University of Turku

Immigration has increased sharply in Europe over the past decade, also in the Nordiccountries (UN, 2017). At the heart of a successful integration of an immigrant arelanguage skills. Research has shown that pronunciation in particular can affect how asecond language (L2) speaker is perceived (Toivola, 2011). The article presents aninternational study on the acceptability, namely the communicative success of speech(Tulaja, 2020), of L2 speakers of Finnish and Danish. The aim of the study is toinvestigate the acceptability dimension of pronunciation in two Nordic languages,namely Finnish and Danish. It is of interest to examine how acceptability can be used asa dimension in the assessment of L2 pronunciation. Knowledge of what native speakersconsider to be acceptable second language speech can be utilized in the design andimplementation of new avenues for research into the social aspects of pronunciation,and new guidelines as to how the social aspects of pronunciation should be included insecond language teaching.