Foreign language (FL) aptitude generally refers to a specific talent for learning a foreign or second language (L2). After experiencing a long period of marginalized interest, FL aptitude research in recent years has witnessed renewed enthusiasm across the disciplines of educational psychology, second language acquisition (SLA) and cognitive neuroscience. This paper sets out to offer a historical and an updated account of this recent progress in FL aptitude theory development and research. As its subtitle indicates, the paper centres on three major issues: following the introduction and clarification of basic concepts, Section 1 traces the early conceptions of FL aptitude dominated by John Carroll's pioneering work. Section 2 summarizes and examines more recent theoretical perspectives and FL aptitude models proposed by researchers from multiple disciplines that have significantly broadened the conventional research traditions associated with Carroll's original conception. Based on the research synthesis of current FL aptitude models, Section 3 suggests the directions FL aptitude theory and research might take in coming years. We conclude that a working memory perspective on FL aptitude presents one promising avenue for advance, as does the development of new aptitude tests to predict speed of automatization, implicit learning and greater control over an emerging language system. In addition, it is argued that issues of domain-specificity versus domain-generality for aptitude tests may lead to aptitude theory and research becoming more central in applied linguistics.

Experience tells us that some people learn a second or foreign language with greater ease, more quickly, or with apparently better results than others. One perspective on this phenomenon is the concept of Foreign Language Aptitude (FLA). Originally, the notion of FLA presumed a relatively stable talent for learning a foreign language that differs between individuals (Dörnyei and Skehan 2003: 590). However, whether FLA is fixed/innate or amenable to training has become the departure point for most research in this area.

Working memory (WM) generally refers to the human ability to temporarily maintain and manipulate a limited amount of information in immediate consciousness when carrying out complex cognitive tasks such as problem-solving and language comprehension. Though much controversy has surrounded the WM concept since its inception by Baddeley & Hitch (1974), an increasing number of cognitive psychologists have accepted WM as a multi-component system comprising both domain-specific storage mechanisms and domain-general executive functions (Miyake & Shah 1999; Baddeley 2012; Williams 2012). Such a fractionated view of this cognitive construct manifests itself clearly in distinct strands of WM-language research, where two contrasting research paradigms have emerged (Wen 2012).

The present study sets out to explore the distinctive roles played by two working memory (WM) components in various aspects of L2 task-based speech planning and performance. A group of 40 post-intermediate proficiency level Chinese EFL learners took part in the empirical study. Following the tenets and basic principles of the Phonological/Executive (P/E) Hypothesis [i.e. the P/E model, Wen, Z. 2015. Working memory in second language acquisition and processing: the phonological/executive model. In Working Memory in Second Language Acquisition and Processing, ed. Z. Wen, M. Mota and A. Mcneill, 41–62. Bristol, UK: Multilingual Matters; Wen, Z. 2016. Working Memory and Second Language Learning: Towards an Integrated Approach. Bristol, UK: Multilingual Matters], WM was conceptualised as consisting of phonological working memory (PWM) and executive working memory (EWM), measured respectively by a non-word repetition span task and speaking span tasks in participants’ L1 Chinese and L2 English. Their L2 speech performance on video-clip narrative tasks was analysed and coded in terms of a series of fluency, accuracy, complexity and formality measures. Statistical analyses indicate that although participants’ PWM was not related to any of the speech performance measures, their EWM was significantly correlated with some lexical, syntactic and stylistic features of L2 speech performance. These results and findings are discussed in relation to the hypotheses laid out in the P/E model as well as their implications for future L2 task planning research.