Introduction

Pronunciation teaching and phonetic training in second language development: What do they have to offer?

The learning process of a new sound system may be challenging not only to students, but also to their teachers. When facing this challenge, L2 learners need to develop new strategies to perceive as well as to produce those new sounds. In turn, when trying to help their students in this task, teachers may find it difficult to set the goals to be reached in their pronunciation classes, as well as to decide on which aspects have to be taught and how these aspects should be addressed in their classrooms.

In order to help both learners and teachers overcome these challenges, research on L2 pronunciation (be it in the classroom or in the language laboratory) plays a fundamental role. Considering this scenario, as we go through the pages of the most consolidated journals on L2 learning and teaching, we may easily notice that there has been a significant increase in the number of studies focusing on L2 pronunciation instruction and perceptual/production training in the last two decades. This growth accompanies the rising number of studies on L2 acquisition in general, being the result of new developments in both the fields of L2 speech and L2 teaching.

As for the developments in the field of L2 speech, the last twenty years have witnessed a significant growth in the propositions of new L2 perceptual models, such as the Native Language Magnet Model (NML - Kuhl 2000), the Perceptual Assimilation Model-L2 (PAM-L2 - Best and Tyler 2007), The Second Language Linguistic Perception model (L2LP - Escudero 2005) and the recent Revised Speech Learning Model (SLM-r - Flege and Bohn 2021),1 among others. Even though all these models are related in what regards their empirical object of investigation, each one of them reflects different views of language and phonetic primitives, ranging from a psychoacoustic account, such as the SLM-r, to a direct-realist, articulatory basis, as claimed in the PAM-L2. These different accounts and the discussions proposed in each of them have contributed to different fields of Linguistics, such as Phonetics and Phonology, Psycholinguistics and Language Acquisition, helping us understand how L2 sounds are learned. Moreover, as these models are based on the recent developments of Formal Linguistics, they also have a lot to contribute to these core research areas, as L2 data constitute a rich source of evidence to these fields.

Along with this better understanding of the L2 learning process, new theoretical constructs have also been proposed in the L2 teaching field, and new classroom methodologies and goals for L2 pronunciation teaching have been set. Munro and Derwing's (1995) definitions of 'intelligibility', 'comprehensibility' and 'accentedness' have made it clear that these are independent (though related) constructs, as an L2 learner may be very intelligible and yet show traces of their L1 accent. In the last 35 years, these constructs have been revisited and new research methodologies have also been developed (Albuquerque 2019; Derwing and Munro 2015; Kang, Thomson, and Moran 2018; Munro and Derwing 2015; Nagle, Trofimovich, and Bergeron 2019; Thomson 2018; Trovimovich et al. 2020, among many others), having thus reshaped the goals of pronunciation teaching. In this scenario, Levis' (2005, 2018, 2020) intelligibility principle sustains that the goal of pronunciation teaching is to promote intelligible L2 speech, instead of aiming at accent-free productions. This has led teachers to rethink their practices, also helping researchers to find new methodologies on how to measure the efficacy of pronunciation instruction or training, be it in the classroom or in the laboratory.

In view of the academic progress in both the fields of L2 speech perception/production and L2 teaching, research on L2 pronunciation has progressed quantitatively and qualitatively, becoming even more heterogeneous and complex. This is clear in the literature reviews and meta-analyses carried out in the last decade by Saito (2012), Lee, Jang, and Plonsky (2015), Thomson and Derwing (2015) and Saito and Plonsky (2019). In their meta-analysis of 77 studies of L2 pronunciation teaching published between 1982 and 2017, Saito and Plonsky (2019) propose a framework for studies on instructed second language pronunciation, according to three main criteria: (i) whether studies adhere to a 'global' (such as intelligibility and comprehensibility judgements) or 'specific' (such as investigations on specific segments) character of pronunciation performance; (ii) whether L2 speech is assessed by human judges or with acoustic analyses; (iii) whether data are obtained spontaneously or in a more controlled setting/instrument. This study not only shows the growth of the pronunciation field in the last years, but also sets important methodological challenges that are still to be faced by researchers in the field, paving the way for future investigations. There remains no doubt, therefore, that L2 pronunciation instruction/training constitutes an effervescent theme of investigation in the current L2 scenario, bridging the gap between studies on L2 speech and L2 teaching.

The present volume congregates these different approaches to L2 pronunciation research in the classroom or in the laboratory. The chapters consist of top-ranked proposals selected for oral presentation at the symposium on "L2 Pronunciation Teaching and Training: Different Approaches", which took place at the AILA online Convention at the University of Groningen in August 2021. This book presents 13 chapters (besides this 'Introduction' and the 'Conclusion'), all of them written by well-known researchers from different universities in the world, who focus their investigations on a variety of first and target languages. The chapters address L2 pronunciation teaching or training in their diversity of approaches, goals, methods and background theories, aiming to strengthen the (imperative) connection between studies on L2 acquisition and L2 teaching.

The 13 chapters of this book are organized in four main themes, departing from L2 speech studies and moving towards the role of pronunciation teaching and classroom and laboratory approaches to pronunciation training. The first part ('Pronunciation development and intelligibility: implications for teaching and training studies') consists of four chapters that focus on the development of L2 pronunciation and its impact on speech intelligibility and comprehensibility. In all of its chapters, the implications of these studies for the pronunciation instruction/training scenario are highlighted, reinforcing the expected connection between Formal and Applied research. The second set of chapters addresses L2 pronunciation teaching, focusing on different approaches and language perspectives to the discussion on the role of classroom intervention studies in L2 development. The last two modules of the book address pronunciation training. The three chapters of the third part focus on the use of technology to foster pronunciation learning in the classroom. These studies deal with the use of speech technologies in order to understand learners' difficulties and how to address them. Finally, the last section of the book presents four studies on High Variability Phonetic Training (HVPT), which consists of a laboratory approach in which learners are trained with stimuli from a variety of talkers and in different phonetic environments (Logan, Lively, and Pisoni 1993; Logan and Pruitt 1995; Thomson and Derwing 2016). Besides their explicit focus on laboratory approaches to L2 speech development, the implications of their results for L2 pronunciation teaching are also highlighted, once again aiming to connect the classroom and laboratory environments, as well as the fields of L2 speech and L2 teaching. In what follows, we provide a brief description of each chapter.

In Chapter 1, Thaïs Cristófaro Silva and Wellington Mendes focus on the role of orthography in the production of plural formation in English by Brazilian learners (L1: Brazilian Portuguese). The authors investigate two orthographic patterns in stop + sibilant sequences: one presenting two consonants in word-final position ('cups', 'cats'), and the other presenting a vowel letter between two consonants ('grapes', 'plates'). Their data show significantly higher rates of vowel productions when the orthographic form contains the letter <e>, thus suggesting that orthographic information is part of phonological representations. The data is analyzed in light of the Exemplar Model (Bybee 2001, 2008; Johnson 1997) combined with the Revised Speech Learning Model (SLM-r) (Flege and Bohn 2021). In view of this combined theoretical approach, we consider this chapter to be representative of our claim that L2 studies can also contribute to revisit (and also reshape) models of Phonetics/Phonology. As suggestions for pronunciation teaching are...