Our goal in developing this timeline was to trace the empirical bases of current approaches to L2 pronunciation teaching, with particular attention to the concepts of accent and intelligibility. The process of identifying suitable works for inclusion challenged us in several ways. First, the number of empirical studies of pronunciation instruction is far too small to provide an interesting perspective on the issues. In fact, the dearth of such investigations has been noted many times since at least as far back as the 1960s (Strain Reference Strain1963; Sisson Reference Sisson1970). Consequently, tracing the roots of contemporary teaching practices required that we expand our purview to consider theoretically-motivated research, as well as influential non-empirical writing about pronunciation. Here we encountered a second problem: the field of applied phonetics and phonology is so diverse that it was very difficult to decide what to omit. The research follows multifarious threads, some of which can directly inform classroom practices, while others are more concerned with general learning influences and processes. In addition, a large body of speculative and opinionated commentary on pronunciation has been published, much of which has never been submitted to empirical test.

Our approach was to start with an empirical core of classroom pronunciation studies and to situate this within the context of other studies and opinion pieces that reflect some of the major themes in the field. Among the lines of work that we have covered are phonological error prediction and analysis – whether on the basis of interference from the first language, on universals, or on other linguistic and cognitive factors – the relationship between speech perception and production in a second language (L2), and the learner variables that account for individual differences in success in phonetic learning. Still another is the extensively debated question of the role of a critical period as a constraint on L2 pronunciation ability.

In surveying this literature we find that theoretical interests are not necessarily congruent with pedagogical ones. For instance, in a great deal of research, the dependent variable has been strength of foreign accent, frequently assessed through listener ratings or other measures of pronunciation accuracy, assuming that native or near-native production patterns are the acid test for successful learning. However, this focus on accent and on accuracy of production is clearly at odds with a long-standing observation that it is intelligibility – rather than native-like pronunciation – that is most critical for successful communication in an L2. In fact, it surprises us that this fundamental distinction between the nativeness principle and the intelligibility principle (Levis Reference Levis2005) received almost no attention from researchers until the end of the twentieth century. Seen in this light, the voluminous work on error prediction models seems to have limited pedagogical significance because advance knowledge of learner errors is unhelpful in prioritizing phonological structures in an instructional programme that aims at intelligibility. Rather, the mere fact that a phonological structure poses difficulty for a learner says nothing about whether it is actually worth teaching or whether it can even be taught. Nor does much of the theoretically-driven research literature help to identify viable, effective classroom techniques (including the use of technology) that will enhance oral communication in the long term.

With respect to the mismatch between theory-driven research and practice, perhaps one of the most telling comments comes from Pica's (Reference Pica1994) survey of the relevance of research findings to classroom practice. In her coverage of pronunciation teaching, she cites only two papers and concludes as follows:

Identifying native-like production as the central goal in pronunciation teaching inevitably leads to the conclusion that pronunciation is probably not worth teaching because of the limited likelihood of achieving that end. On the other hand, intelligibility has been proposed as a priority in instruction since as early as Sweet's (Reference Sweet1900) handbook, and has been repeatedly identified as a distinct construct throughout the twentieth and twenty-first centuries. But it is not possible to teach intelligibility without a clear understanding of what it is and how it can be achieved. What has been missing until very recently is, first, a conceptualization of intelligibility that assists teachers in setting priorities and second, empirical evidence that identifies effective practices.

Because we have restricted our choices primarily to works that match our own area of expertise, the timeline is heavily oriented toward studies of adult ESL learners that focus on the speech stream and how it is perceived. The following themes, labelled in the last column of the table, are represented:

1. A Empirical teaching studies – the core of our timeline – that examine the performance of real students before and after a pedagogical intervention. We have excluded most laboratory studies designed to evaluate learning under non-classroom conditions.

2. B Empirical studies that employ the constructs of accent, comprehensibility, and intelligibility, and/or demonstrate methodologies for pronunciation assessment. These studies all have applicability to teaching.

3. C Empirical studies of factors that influence success in pronunciation learning.

4. D Theoretical and other non-empirical discussions representing the major historical trends that have characterized the field of applied phonetics and phonology.