This review focuses on a growing branch of this research effort, that which examines the motivational effects of language teaching. This is pertinent because, despite enhanced mobility and expanding access to foreign languages online, most learners’ early encounters with the second language (L2) still take place in classrooms, and these encounters may shape attitudes and determine students’ willingness to invest further in the L2. Four main types of research are reviewed: first, that which deliberately seeks to identify and evaluate strategies to motivate L2 learners; second, that which has tested the validity of psychological theories of motivation by applying their precepts in L2 classrooms; third, that which assesses the motivational effects of a pedagogical innovation or intervention; fourth, research on what has been too often the unintended outcome of language education, namely learner demotivation. The review highlights the complexity of the relationship between teaching and learner motivation but an attempt is made to articulate some emerging verities and to point towards the most promising avenues for future research.

1. Introduction

This article reviews the growing amount of research evidence about what language teachers can do to motivate their learners. It comes at a time when the language teaching profession faces profound motivational challenges. In many global contexts, the English language is fast becoming established as a basic skill on the national curriculum, increasingly taught at all levels from primary to tertiary, and assessed in high-stakes examinations (Graddol 2006). English teachers find themselves in high demand but also often find that their jobs are increasingly demanding of them; they are under pressure from their institutions to raise standards and respond to curricular innovations, and also from pupils, who are less accepting of a submissive role in class, and expect to be entertained as well as educated. In Anglophone countries, the global spread of English has undermined the raison d’être for foreign language teaching, reflected in a diminished role for languages in national curricula and shrinking enrolments for study at higher levels (Coleman 2009), and forcing institutions and teachers to provide learners with new motives for learning (Gallagher-Brett 2004).

The review aims to address the motivational challenges faced by language teachers in all contexts, though as will be seen there is a preponderance of research on the teaching of English, especially in formal secondary and tertiary education, and much less on the teaching of modern foreign languages (MFL), of heritage languages, of host country languages for migrant adults, or other contexts of language teaching. Its scope will be limited to empirical studies of what is usually within the control of the teacher – for example particular teaching methods or techniques, decisions about learning materials and tasks, forms of classroom assessment, the deployment of new technologies, classroom language – or of their institutions – for example whether a language is taught as a separate subject or as a medium of instruction (e.g. in content and language integrated learning or CLIL). Excluded by this criterion are the motivational impact of national assessment regimes, decisions about when languages are introduced into the curriculum, the motivational results of teacher development initiatives and other policy issues.

The article is warranted because of the surge in publications related to the pedagogical aspects of motivation in the last decade. In the most authoritative recent review of L2 motivation, Dörnyei & Ushioda (2011) comment that the ‘amount of research devoted to the question of motivating learners remains rather meagre relative to the total amount of research on L2 motivation’, but a forthcoming survey article (Boo, Dörnyei & Ryan 2015) finds that approximately one third of 415 papers published in the period 2005–2014 had a main focus on ‘motivating’ as opposed to ‘motivation’. Yet it is only a generation ago that the eminent social psychologist Robert Gardner was able to assert that ‘[t]he learning of a second (or foreign) language in the school situation is often viewed as an educational phenomenon . . . such a perception is categorically wrong’ (1979: 193). Instead the social psychologists’ primary interest was in learners’ attitudes towards the language and its speakers, and their motives for learning it. Perceptions began to change during the 1990s when, with the rise to prominence of social constructivist learning theory, it was recognized that ‘motivational sources closely related to the learners’ immediate classroom environment [had] a stronger impact on the overall L2 motivation complex than had been expected’ (Dörnyei & Ushioda 2011: 47), ushering in what came to be known as the ‘cognitive-situated’ phase of L2 motivation research. Longitudinal studies of L2 learner motivation (e.g. Chambers 1999, Nikolov 1999) reinforced the view that teachers and teaching could play a central role in its development. Utilizing constructs from mainstream educational psychology, book length treatises appeared on how teachers can motivate their pupils to learn an L2 (Williams & Burden 1997; Dörnyei 2001). Partly inspired by these works, and perhaps more recently swayed by an academic climate promoting ‘impact’, motivation researchers have increasingly made aspects of L2 pedagogy their main focus of study.

The first purpose of this review then is to chart the trajectory of this upsurge in research, prioritizing works published in major journals, monographs or edited books and identifying the main lines of inquiry. This provides the organizational structure of the article: the first section deals with research on motivational strategies (MotS), techniques deployed by teachers to deliberately enhance learner motivation; the second section examines research on L2 classrooms which adopts a particular theoretical orientation – two theories dominate here, self-determination theory (SDT) (Deci & Ryan 1985) and the l2 motivational self-system (L2MSS) (Dörnyei 2009); the next group of studies are empirical investigations of L2 pedagogy which have a strong relevance for motivation – applications of new technology forming a significant majority; and then there is a review of research targeting the regrettably common phenomenon of learner demotivation i.e. where classroom processes actually diminish the motivation that learners initially bring with them. Only research papers addressing an aspect of pedagogy are reviewed; those focusing on learner L2 motivation are excluded unless their pedagogical implications are particularly compelling.

The second main purpose of the review is to draw out the main lessons that have been learned so far, to inform the training and education of language teachers, and to point researchers towards promising areas of further inquiry. The final section of the article therefore sets out an agenda for future research on the motivational dimension of language teaching, based on remaining gaps as well as on exciting new theoretical approaches to L2 motivation such as complex dynamic systems theory; it also makes suggestions for methods of investigation that could make the next decade even more productive of insights for motivating language learners.

It could be argued that any good teacher is, by definition, a motivator of learning. Yet there is clearly a motivational dimension to teaching that is distinct from simply ‘good teaching’. Where learners come to class already highly motivated – for example, when preparing for high-stakes exams – it is true that skillful teaching which enables them to achieve their goals as efficiently as possible will be felt to be motivational, even if the teacher makes no conscious effort to motivate. In many other educational contexts, learners enter classrooms with a modicum of motivation which good teachers can work with to produce acceptable results. But teachers who actually target learner motivation could nurture and strengthen it so that it promotes greater learning effort during the course, produces even better results, and perhaps even carries forward to future periods of study. In the most challenging kind of classrooms – this would include many secondary school MFL classes in Anglophone countries – learners may not have chosen to be there at all, and the generation of initial motivation is the teacher's most important role, one that is undeniably complex and difficult. Focus too much on the here-and-now, using traditional carrot and stick techniques, and teachers might preserve classroom discipline at the cost of long-term interest in the subject. Focus too much on the future, and their lessons may neglect learners’ current interests and identities. Such issues demand researchers’ serious attention.

2. Motivational strategies

2.1 Studies based on Dörnyei's (2001) taxonomy

Whichever kind of classroom teachers occupy, most will have at some time or other employed strategies to boost their learners’ motivation. This section focuses on research which directly investigates the prevalence and, in some cases, the effectiveness of these kinds of practical decisions. Dörnyei & Ushioda (2011: 103) define MotS as instructional techniques deployed ‘to consciously generate and enhance student motivation, as well as maintain ongoing motivated behaviour and protect it from distracting and/or competing action tendencies.’ This research has its origins in the The Modern Language Journal debates of the mid-1990s, in which a new generation of academics, concerned to make L2 motivation research more useful to education, were challenged by Gardner & Tremblay (1994) to produce empirical evidence to support their claims that teachers could positively affect their pupils’ motivation. Dörnyei & Csizér's (1998) pioneering study of the views of Hungarian English as a foreign language (EFL) teachers was published just in time for inclusion in the last ‘State-of-the-Art’ review of L2 motivation (Dörnyei 1998). Their ‘10 Commandments for Motivating Language Learners’ represented groups of teaching techniques which Hungarian teachers claimed to use to motivate their pupils. Drawing on this study and on relevant theories of educational psychology, Dörnyei (2001) expanded the list into a full taxonomy of 102 MotS (sometimes termed micro-strategies), grouped into 35 main strategies (or macro-strategies) and organized according to where in the teaching process they would likely be deployed.

The clarity and comprehensiveness of Dörnyei's MotS framework attracted researchers to validate it empirically, first his own doctoral students (Cheng & Dörnyei 2007; Guilloteaux & Dörnyei 2008) and then those beyond his immediate orbit (e.g. Sugita & Takeuchi 2010). Table 1 presents the most important research so far published. It shows how studies have become steadily more ambitious in their aims and more sophisticated in their approach, addressing some of the conceptual and methodological difficulties laid out by Gardner & Tremblay (1994). The most straightforward type of study samples the views of teachers on the relative importance and frequency of use of given MotS (e.g. Cheng & Dörnyei 2007; also see Guilloteaux 2013). Other studies do this but also compare teachers’ views with those of students (e.g. Sugita McEown & Takeuchi 2014; also see Sugita McEown & Takeuchi 2010 and Ruesch, Bown & Dewey 2012). More sophisticated are those studies which actually evaluate the motivational effectiveness of MotS in terms of learners’ behaviour in class and their self-reported motivation (Guilloteaux & Dörnyei 2008; Papi & Abdollahzadeh 2012; Wong 2014). The most ambitious of all are the quasi-experiments (Moskovsky et al. 2012; Alrabai 2016) which identify a set of contextually promising MotS, train teachers in using them, and then compare their motivational effects on classes of students against that of traditional teaching. This last type of study has only been conducted in Saudi Arabia to date, possibly because of the difficulties of gaining ethical approval in many contexts.

What have we learned so far? Some macro-strategies do appear to be valued universally by teachers and learners; these include displaying appropriate teacher behaviour (or ‘setting a good example’), fostering good teacher–student relations and promoting learner self-confidence. Instilling these principles in novice language teachers may be time well spent, though how they are realized in particular classrooms is something that teachers will still have to work out ‘on the job’ (see Section 6.2). Furthermore, as Table 2 shows, there is disagreement about the relative importance of different macro-strategies. There is some evidence, for example, that promoting learner autonomy is considered less important by teachers in Asia, but Asian teachers themselves differ in their valuing of a cohesive learner group (much favoured by Taiwanese but not Koreans or Chinese) and ‘offering rewards’ (by far the most popular MotS among Chinese school teachers) (Cheng & Dörnyei 2007; Guilloteaux 2013; Wong 2014). There is also evidence that teachers and students’ views differ; for instance, Ruesch et al. (2012) found that university students see motivational value in comparing themselves to other students – a practice denigrated by their tutors trained in ‘cooperative’ learning principles.

This ‘competition vs cooperation’ dilemma neatly captures a fundamental problem in the MotS approach; Dörnyei (2001) himself seems to come down firmly on the side of ‘cooperation’, arguing that ‘there is nothing “healthy” about even a small dose of competition’ (p. 93), though in other sections he promotes the value of ‘intragroup competition’ (p. 44). Other work suggests that ‘competition within cooperation’ has the highest motivational value (Tauer & Harackiewicz 2004). The point is that it is simply not possible to reduce highly complex issues to pedagogical ‘dos or don'ts’. The successful motivator somehow learns when, where and how to deploy them in particular lessons.

Nevertheless, another major finding of this research is that MotS can work; that is, correlational studies show that they are associated with more motivated behaviour in class, in terms of paying attention and participating in class activities, and with more positive attitudes towards learning and the L2 (Guilloteaux & Dörnyei 2008; Papi & Abdollahzadeh 2012). Admittedly, one can dispute what motivated classroom behaviour actually consists of (Ellis 2009, for example, proposes that ‘alertness’ would be a better criterion of engagement than ‘attention’), and correlation does not indicate a causal effect. But Guilloteaux & Dörnyei (2008) point out that contrasting results for different (mixed-ability) classes in the same schools strongly suggests it is the teacher doing the motivating. Their argument has been strengthened by the subsequent quasi-experimental studies in which teachers specially trained in MotS produce positive changes in their learners’ motivation over the length of a course (Moskovsky et al. 2012) and even generate better L2 achievement, at least when compared to ‘traditional teaching’ as practised in Saudi Arabia (Alrabai 2016).

Clearly, these are exciting results which merit further investigation. In particular, it would be valuable to see whether the deployment of particular MotS by L2 teachers have similar beneficial effects – on learner classroom behaviour, their longer-term motivation and, ultimately, on their achievement – in other global contexts. Education experts may be able to predict which MotS are lacking in, or would be especially valuable in, certain contexts, and these could be targeted by research. For example, we may find that the teaching of MFL in Anglophone countries requires different MotS than the teaching of English as an international language in non-Anglophone countries. Likewise, although cultural and educational traditions will continue to prove important in determining the value of MotS, there are likely to be more significant differences in the motivational needs of learners of different ages within a single country than between countries.

To build a knowledge base of use to educators, future research on MotS also needs to maintain conceptual clarity. Can we be sure that what an Iranian teacher or researcher means by (e.g.) ‘act naturally in front of students’ (Guilloteaux 2013) is the same as what a Korean or American teacher or researcher means? It would seem sensible for research to focus on macro-strategies – general enough to have broad common meanings, and so enabling comparisons across settings – and allowing micro-strategies to be determined locally. So for instance a macro-strategy may relate to ‘Using classroom language sensitively’ – whether this means more use of the L2 (as in Saudi university classrooms, see Moskovsky et al. 2013) or less use of the L2 (as in Chinese elementary schools, see Rui & Chew 2014), will depend on the particular pedagogic context under study.

2.2 Alternative approaches to MotS

There have been other attempts to classify and test out MotS in language education which are not based on the Dörnyei (2001) taxonomy. These contribute to our understanding of MotS, but also cast doubt on the value of trying to build a definitive list of motivational teaching behaviours. Bernaus & Gardner (2008) divided motivational teaching strategies into ‘traditional’ (centred on the teacher's behaviour e.g. ‘I make my students do dictations’) and ‘innovative’ (based on a learner-centred methodology e.g. ‘my students do self-evaluation and co-evaluation’) and investigated their use in Spanish secondary school English classes. The results are significant for suggesting that ‘it is not the actual use of strategies but their perceived use that has an effect on motivation and achievement’ (p. 399). That is, classes of students and teachers do not always agree on whether strategies are being used, and the strategies may only serve to motivate if the students recognize and appreciate them – which they may not do if, for instance, they dislike the teacher's personality or lack any other motives for learning the L2. Meanwhile, in a study of EFL learners in Hungarian schools, Mezei (2014) found that teachers’ use of motivational strategies did not impact on learners’ motivation directly but was mediated by the ideal l2 self and self-regulation. In other words, students with a strong Ideal L2 self (see Section 3.2) and self-regulatory capacity were better able to benefit from the MotS, suggesting that teachers need to work on these other learner qualities too.

Maeng & Lee (2015) based their investigation on a different MotS framework, using Keller's ARCS model (Attention, Relevance, Confidence, Satisfaction) to identify the strategies used by English teachers in micro-teaching sessions during in-service training in Korea. Their most noteworthy finding was that more experienced teachers used fewer strategies than less experienced teachers, again raising questions about how useful it is to categorize behaviours as positive or negative and then count their frequency. The study also draws our attention to the fact there are other potential frameworks for classifying MotS. Among the most significant is Williams & Burden's (1997) 12-point list, both because it is designed specifically for language teachers and because it has been elaborated on the basis of social constructivist learning theory. The general education literature provides other options, notably Anderman & Anderman (2010), Schunk, Meece & Pintrich (2013) and Wentzel & Brophy (2014).

Another approach to understanding MotS was taken by Lamb & Wedell (2015), who asked learners in China and Indonesia to reflect on the pedagogical qualities of inspiring language teachers they had had in the past. In this way they aimed to identify those aspects of L2 pedagogy that had a long-term motivational impact on learners, to complement the shorter-term perspective of MotS research. The most frequently cited qualities related to classroom practices, and again some systematic contextual differences were noted e.g. Indonesian learners valued teachers who made lessons entertaining, while Chinese learners were more appreciative of teachers who made lesson content interesting. However, they also found that about a third of comments related not to what teachers did but what they were like as persons, suggesting that personality can have a major long-term impact on individuals’ motivation to learn.

By focusing on the thinking and decision-making of individual teachers, qualitative researchers have begun to open up new perspectives on MotS. Glas (2015: 2), for example, deliberately eschews the use of a priori lists of MotS in order to take ‘a bottom-up approach towards understanding the role of teacher cognition and decision-making with respect to learner motivation’. What she finds is that Chilean teachers of English feel constrained in how far they can motivate their learners, both by limitations on many learners’ pre-existing motivation and by institutional and political restrictions on their ‘spaces for maneuver’ (e.g. in choice of materials, training opportunities, availability of IT). Nevertheless, some individual teachers do express more agency than others in developing their learners’ motivation, despite working in similar contexts. Investigating the beliefs and behaviours of eight successful ‘motivators’ in Indonesia, Lamb, Astuti & Hadisantosa (2017) locate the source of their agency in their persistent willingness and ability to empathize with their learners – so while macro-strategies like ‘developing a close rapport with students’ and ‘providing enjoyable activities’ are valued by all, their classroom realization will differ according to language level, age, geographical location and other student factors, as well as the teachers’ own training and experience. They suggest that it can take years of reflective practice, as well as a high level of daily commitment, to develop this nuanced response to the needs of individuals and classes.

The weakness of the strategies approach to understanding motivational teaching has been well articulated by Ushioda (2007); it shares with much traditional L2 motivation research, she argues, a positivist, linear model of educational processes, in particular ‘a tendency to view pedagogical interventions in terms of methods or strategies, and to view motivation as the product in a chain of cause and effect’ (p. 23). Adopting a ‘person-in-context relational approach’ (Ushioda 2009), both the studies above, for example, serve to highlight the complexity of MotS and their inseparability from the relations of actual people in contexts of differing constraints and affordances. The categorization and listing of teaching behaviours could give the impression – to novice teachers, educational managers or other stakeholders – that ‘motivating’ is a matter of the teacher deploying the correct strategies, when in fact it is an intensely interactive process, where motivating lessons emerge (sometimes surprisingly) from the coming together and intense mutual engagement from moment to moment of teacher and learners. It is perhaps this capacity for responsiveness, relying on the personal quality of empathy but also built up over years of practice, which defines the successful motivator. This line of argument will be developed further in Section 6.3.

3. Theory-based research on L2 motivation

This section reviews research on L2 pedagogy and motivation which adopts a clearly identifiable theoretical framework. A number of psychological theories have been applied to L2 learning, but only three – SDT, the L2MSS, and social cognitive theory – have inspired the kind of intervention studies which are the main focus of this review. These are described in the following sections; other theories are included in a summary table (Table 3) to help future researchers who wish to develop and test pedagogical approaches that apply their precepts.

3.1 SDT

Originally formulated by Deci &Ryan (1985), SDT is one of the most enduring theories of motivation, actively applied in a wide range of human activities from education through healthcare to management, supported by a flourishing website (www.selfdeterminationtheory.org) and popularized in bestselling paperbacks (Pink 2009). The theory is perhaps best known for elaborating the binary distinction between intrinsic motivation, where people learn because they enjoy the process of learning, extrinsic motivation, where they are doing it not for its own sake but for some alterior motive, which can be more or less ‘external’ i.e. from very externalized (e.g. to avoid punishment) to very internalized (e.g. doing it because it is part of ‘who you are’). In educational settings, the ideal scenario is where all pupils are intrinsically motivated to learn the subject, because then they will be naturally inclined to study hard, act responsibly and be receptive to new knowledge (Ryan & Deci 2000). In reality, this is unlikely to be the case, and SDT's key insight for teachers is that with skillful instruction, they can help to make their pupils’ motivation more intrinsic and more internalized. They can do this by trying to ensure that their classrooms satisfy three conditions necessary for human growth and learning: the need for autonomy (a feeling of being able to choose personally meaningful activities), for competence (a sense of gaining mastery of a subject area or skill) and for relatedness (feeling connected to and valued by others engaged in the activity).

As Stroet, Opdenakker & Minnaert's (2013) review article shows, there is now a wealth of research in general education showing that ‘need-supportive teaching’ does promote school age learners’ academic motivation, at least when students perceive the teaching in this way (similar to Bernaus & Gardner's 2008, previously mentioned finding about MotS). Attempts to apply SDT to L2 pedagogy were led by Noels and colleagues (Noels, Clément & Pelletier 1999; Noels 2001), who found that, in line with the theory, Spanish language teachers with a communicative style which supported their North American students’ sense of autonomy and competence tended to have learners with more intrinsic and internalized forms of motivation. In other words, when students perceive their teachers as giving them choice in what and how to learn, and as providing helpful friendly feedback on progress, they will be encouraged to put more effort into study. Noels (2001) speculated that such effects might be even stronger for school pupils, who are at a more impressionable age and usually interact more intensively with their teachers than university students.

Since those pioneering studies, a number of other researchers working in a variety of global contexts have confirmed the validity of this and other aspects of the theory. Carreira, Ozaki & Maeda (2014) investigated elementary school English classes in Japan and found that the children's perceptions of teacher autonomy support (e.g. the amount of praise and encouragement they give) contributed to their own sense of autonomy, competence and relatedness, and in turn their intrinsic motivation, especially in the middle grades (years 3–4). Pae & Shin (2011) compared the motivational impact of Korean EFL teachers using communicative methodology – presumed to be more supportive of learners’ autonomy, competence and relatedness – with those having a more traditional teaching style. They found that the extrinsic motivation of students was not affected by which methodology they were exposed to but, as expected, those in communicative classrooms tended to have higher intrinsic motivation and self-confidence. Two other studies had the common aim of enhancing students’ intrinsic motivation, but designed quite different teaching interventions suited to their contrasting higher education (HE) contexts. In the USA, Jones, Llacer-Arrastia & Newbill (2009) created a series of language learning tasks that enabled their foreign language (FL) students to feel renewed enjoyment for their work; in Japan, Fukuda, Sakata & Takeuchi (2011) introduced a ‘guided-autonomy’ syllabus deliberately to boost their students’ sense of autonomy for learning, which they believed was lacking, though they found that this worked only when student–teacher relatedness was also strong.

More recent studies by Noels and colleagues have countered the criticism that SDT has more relevance to individualist ‘western’ societies than to collectivist ‘eastern’ societies. Noels et al. (2014), for example, compared the beliefs of Euro-Canadian and Asian-Canadian undergraduates and found no difference in their valuing of autonomy in the language classroom, though interestingly both groups preferred ‘reactive’ to ‘proactive’ autonomy – in other words, both appreciated the active guidance of their teachers throughout their course. Sugita McEown, Noels & Saumure (2014) found that Canadian students’ perceptions of their teachers’ support for relatedness and competence did correlate with their level of self-determination (i.e. autonomous motivation) for FL learning, yet perception of autonomy support did not. They suggest that in environments where high levels of autonomy already exist, such as at North American colleges, this aspect of pedagogy may not be as salient as in contexts where regular teaching and institutional curricula are more controlling. Other recent research has thrown doubt on the view that more choice is always better – Mozgalina (2015) found that German learners of Russian did not necessarily engage more fully on language learning tasks when they had chosen them themselves.

Although not always referencing SDT, a distinct strand of L2 motivation research has addressed the principle of ‘relatedness’. Assisted by authoritative works on positive group dynamics (Dörnyei & Murphey 2003; Dörnyei 2007), researchers have recently focused on whether and how teachers can motivate learners through deliberately fostering positive relationships in class. For example, both Chang (2010) and Koga (2010) found correlations between the cohesiveness of language classes and the motivation of individual class members. In a continuing programme of research at Japanese universities, Murphey and colleagues have demonstrated the reciprocal nature of the relationship between learner motivation and group dynamics; for example Murphey et al. (2014) report an action research cycle where lecturers deliberately tried to engineer more positive group dynamics in English major classes through an adaptation of possible selves theory (see Section 3.2); students were asked to generate idealizations of helpful classmates, and these visualizations in turn helped them to behave in cooperative ways with others, evidence – they argue – of positive emotional contagion.

3.2 The L2MSS

3.2.1 Future selves

Mainstream educational psychology has attributed an increasing role to the concepts of ‘self’ and ‘identity’ in motivating learners to put effort into academic study (Oyserman, Bybee & Terry 2006; Kaplan & Flum 2009). In L2 education, this development has been reflected in the currently dominant theoretical framework, the ‘L2 motivational self-system’ (Dörnyei 2009; for more on how this scheme built on earlier L2 motivation theory and on ‘self psychology’, see Dörnyei & Ushioda 2011). This proposes that L2 learners’ motivation can in most contexts be best predicted and explained through three main constructs:

• The Ideal L2 self – the learner's personally valued vision of themselves as a competent user of the L2 in the future (in SDT terms, a strongly internalized motive, encouraging an active aspiration towards the desired goal)
• The Ought-to L2 self – the learner's conception of what significant others believe they ‘ought to’ be like in the future (in SDT terms, a more extrinsic motive, where avoidance of undesirable end states is prioritized)
• The L2 learning experience – the learner's attitudes to, and experiences of, the learning process, inside and outside of classrooms.

When investigated empirically in diverse global contexts, the Ideal L2 self and the L2 learning experience are consistently found to be more closely related to actual or intended learning effort than any other motivational constructs (such as integrative or instrumental orientations, cultural interest, parental encouragement – see Taguchi, Magid & Papi 2009; Kormos, Kiddle & Csizér 2011).

From a pedagogical point of view, this framework has two potential sources of insight. First, the fact that the ‘L2 learning experience’ is found to be predictive of motivation to learn, and with younger learners may even be the most important factor (e.g. You & Dörnyei 2016), foregrounds the role of the teacher and their methodology as a decisive influence on motivation. To date, however, there has been little research exploring the interaction between this component and the future selves, and it is an issue that Ushioda (2014: 134) claims is ‘undertheorized’. Second, and in contrast, promising work has already been done, both conceptually and empirically, to explore the pedagogical implications of the ‘Ideal L2 self’. Future selves have their origins in the social role models and media images available to young people in childhood and early adolescence; in many global societies, the valorization of English language skills and the association of the language with desirable cosmopolitan lifestyles means that by the time they enter secondary school, many pupils do already have incipient ideal English-speaking selves (Ryan 2006; Lamb 2012).

Two recent books (Hadfield & Dörnyei 2013; Dörnyei & Kubanyiova 2014) offer teachers practical ideas for developing learners’ L2-related visions, organized in a sequence of six stages:

1. Initially constructing the desired future self (where it is entirely lacking)

2. Enhancing the image e.g. through classroom visualization exercises

3. Making the image plausible e.g. through ensuring learner goals are realistic

4. Developing an action plan e.g. helping learners set proximal sub-goals

5. Activating the vision to keep it alive e.g. through regularly engaging with learners’ transportable identities

6. Counterbalancing the vision e.g. sensitively reminding learners of the dangers of failure.

Some empirical studies already offer evidence that these strategies can work. Magid & Chan (2012) describe simultaneous interventions with Chinese undergraduate students learning English (as a ‘minor’) in the UK and Hong Kong. The 3–4-month treatment involved guided imagery sessions where students were helped to visualize themselves in future using English, workshops and language counselling sessions where they developed action plans to achieve them, an ‘Ideal L2 self Tree’ activity to elaborate their L2 future selves (see Chan 2014). The results were s a strengthening of participants’ ideal L2 selves (see Magid 2014), and the strong qualitative endorsement of the sessions by students in interviews. Two semester-long studies with Japanese Majors in English, one an action research cycle involving various tasks designed to elaborate and strengthen their ideal L2 selves (Sampson 2012), the other involving extended role-playing as international ‘experts’ using English (Munezane 2013) also provided positive results in terms of students’ affective response to the class activities themselves and the effects on their future L2 self-guides. Mackay (2014) conducted a quasi-experiment with upper-intermediate level Catalan students of English, exposing the treatment group to a series of sessions based on Hadfield & Dörnyei (2013); quantitative evidence of success was limited but again qualitative data indicated that the activities were valued and did enrich students’ ideal L2 selves.

While these studies are encouraging, especially in suggesting that a treatment of about three months is sufficient to make a qualitative difference in L2 learner motivation, it should be noted that they are all relatively small-scale and were limited to HE students; we need trials with younger learners, first to find out if visualization and action-planning activities are popular with younger age groups, at what age learners can form realistic images of their future selves, and also to see whether enhanced motivational visions are sustainable over time. Teachers also need to be aware of research suggesting that individual learners may have different capacities for L2 self-imagery (Al-Shehri 2009; Dörnyei & Chan 2013).

3.2.2 Present selves and identities

One of the criticisms of the L2MSS is that it has directed too much attention towards future-oriented aspects of the self, and ignores the motivational consequences of other self-concepts and more immediately relevant identities. Taylor (2013) has argued on the basis of her large-scale study of Romanian teenagers, for example, that many school pupils display ‘public selves’ in their English language classroom which are at odds with how they really see themselves as language learners (e.g. some trying to appear coolly detached to peers while actually ambitious, others demonstrating false ambition to the teacher). The result is dysfunctional school classrooms full of dissimulation. Instead, she argues, teenagers are desperate for teachers to treat them as real people with complex personalities and varied extra-curricular interests. Support for this perspective comes from other studies of classrooms where teachers try to do just that – for example Richards (2006) demonstrates how student class participation is enhanced when teachers engage with learners’ ‘transportable identities’, that is their out-of-class personas (e.g. as big sister, Manchester United supporter, player of Minecraft). As Ushioda (2011a) points out, this mirrors to some extent the theme of ‘personalization’ within communicative language teaching, but Luk's (2005) study shows how ‘off-the-shelf’ communicative tasks which might work in a western context fall flat in a Hong Kong school; instead, learner motivation is enhanced by more traditional teacher-fronted classes where the teacher makes a concerted effort to invoke and perhaps challenge learner identities in a humorous way – for similar examples see Sullivan (2000) in Vietnam, Forman (2011) in Thailand, Lamb & Budiyanto (2013) in Indonesia.

While the imagining of ideal future selves can provide a motivational impetus for study, the more immediate goal for teachers is perhaps the development of ‘L2-mediated identities’ (Block 2007) – helping learners to feel comfortable using the L2 and convincing them that it can become a vehicle for expressing their own voice. The Asian studies cited above show that nurturing individual identities can be a group enterprise. Programmes of research in Japanese university classrooms offer further evidence of this. Yashima & Zenuk-Nishide (2008) show how elaborate class role plays can promote the ‘international posture’ of their students and their readiness for authentic communication in English. Falout and colleagues (Falout et al. 2013) report on how their ‘present communities of imagining’ framework enables class members to collectively reprocess past negative educational experiences, share present learning activities and exchange and elaborate visions of their possible English-speaking futures, with quantifiable benefits for their learning motivation.

Researchers taking a more sociological and specifically poststructuralist approach to second language acquisition (SLA) offer other perspectives on learner identity, with different implications for teachers and institutions (Norton & Toohey 2011). In this view, identity is far more complex and dynamic than is captured in the neat categories of psychological theories, and Norton's term ‘investment’ is offered as a more contextually sensitive, politically responsive complement to the concept of motivation – for example, a learner may be motivated to learn a language yet not invested in a particular class, perhaps because they do not like the teacher's methodology or feel rejected by peers (Norton 2001). Indeed, much of the research in this tradition has highlighted the problems that occur when teachers or their institutions do not recognize or respect the identities of classroom language learners, as in Toohey (2000), Duff (2002) and Talmy (2008), all case studies of English as a second language (ESL) learners in North American mainstream schools who reject the stigmatized ‘ESL’ identity imposed upon them and feel disempowered in relation to their ‘native’ peers. Examples from other educational settings are Cambodian women rejecting ESL classes because they felt their domestic and professional identities were not taken into account (Skilton-Sylvester 2002), and Japanese university students being side-lined in Canadian academic seminars (Morita 2004). The research is also valuable in bringing to light systematic inequalities related to identity categories like race, gender and sexual orientation that potentially undermine L2 learning opportunities.

As Norton & Toohey (2011) acknowledge, this research has so far been successful in revealing structural constraints on L2 learner motivation/investment in western educational contexts but more needs to be done to uncover the ‘problems and triumphs of language teaching’ in postcolonial and multilingual global settings (p. 437), not least the way social class is implicated in the distribution of opportunities to learn and use English in the developing world (Block 2013). Available evidence suggests, however, that to promote learner investment teachers need to view language not as a static linguistic system but ‘as a social practice in which experiences are organized and identities negotiated’, to offer learners ‘multiple identity positions from which to engage in the language practices of the classroom’ and to be constantly vigilant about ways in which they may be reinforcing subordinate student identities (Norton & Toohey 2011: 430).

3.3 Social cognitive theory

3.3.1 Self-efficacy

Bandura's (1997) social cognitive theory emphasized the important role that self-evaluation plays in shaping a learner's approach to learning tasks, and introduced the concept of self-efficacy: ‘one of the most heavily researched and viable psychological constructs in different areas of human functioning’ (Schunk & Pajares 2004: 115). A person's self-efficacy beliefs are formed mainly through ‘mastery experiences’ (i.e. successful performance of an activity in the past) as well as from comparison with other people, other people's judgements of their ability, and positive emotions. In general education, learners’ beliefs about their own capability to carry out particular tasks have been found to affect their choice of activities, how much effort they put into them, and what they ultimately achieve (Mills 2014). Moreover, there is plenty of evidence that classroom experiences are an important source of self-efficacy beliefs (Wang & Pape 2007) and so what the teacher does (e.g. the type of questions they ask, how they group individuals with other students, the kind of feedback they give) may contribute to shaping each learner's self-efficacy, with potentially long-term impact on their academic success.

Self-efficacy has been found to be a significant component of language learner motivation in a variety of L2 contexts, often mediating the impact of L2-related attitudes on motivated behaviour (e.g. as in Tremblay & Gardner's model, 1995). Recently, studies by Kormos et al. (2011) and Iwaniec (2013) in Chile and Poland respectively have suggested close links between the Ideal L2 self and self-efficacy beliefs i.e. one has to believe in one's capacity to learn and use the L2 in order to visualize a future L2-using self. Self-efficacy has also been shown to predict L2 achievement in school (Hsieh & Kang 2010) and university settings (Mills, Pajares & Herron 2007). It is also notable that promoting students’ self-confidence is consistently found to be a teaching strategy valued by teachers in a variety of global contexts (see Table 2).

Research describing actual attempts to foster self-efficacy in the L2 classroom, and to monitor its effects on motivation and achievement, is limited but persuasive. In the UK, Graham & Macaro (2008) hypothesized that learners’ low self-efficacy with French stemmed partly from difficulty with understanding French speech, and this could be ameliorated by training in a cluster of listening comprehension strategies (e.g. making predictions, identifying key words, recognizing word boundaries). A six-month intervention with two groups of low-intermediate students of French demonstrated lasting gains in L2 listening proficiency and in learner self-efficacy. In the USA, Mills & Peron (2009) used a ‘global simulation course’ to improve the writing self-efficacy of 148 college students of French – role-playing French characters, they hypothesize, helped the students ‘validate’ their French-speaking identities. In Japan, Matsumoto, Hiromori & Nakayama (2013) tested the impact of instruction in key reading strategies (e.g. identifying the main idea, adjusting reading speed to test difficulty) on the motivation and beliefs of 360 Japanese undergraduate students of English (as a ‘minor’), producing positive results for extrinsic and intrinsic motivation and reading self-efficacy.

3.3.2 L2 self-confidence and self-esteem

Two related constructs have also been associated with L2 motivation and achievement. ‘Linguistic self-confidence’ has its origins in social psychological approaches to L2 education (Clément 1980) when it was hypothesized that students who experience close contact with the L2 community would have stronger confidence in their capacity to learn. More recent research in FL contexts suggests the relationship can be reversed, i.e. that more confident learners seek out intercultural contact, which in turn can improve L2 attitudes and lead to motivated learning behaviour (Csizér & Kormos 2008). As Sampasivam & Clément argue (2014) the relationship is likely to be reciprocal, and offers a reason for educators to promote intercultural contact as a means of improving L2 learner motivation, whether directly through short-term study abroad projects (e.g. Hernandez 2010; Reynolds-Case 2013) or through computer-mediated communication (see Section 4.1).

The more general construct of ‘self-esteem’ (the overall estimation of one's worth as an individual) has a quite different history in L2 education (Rubio 2014). The threat that L2 learning can bring to a person's self-esteem (e.g. through depriving one of the capacity for self-expression) has fueled the popularity of humanistic approaches to language teaching, embodied in ‘alternative’ methods like Suggestopedia (Lozanov 1978) as well as in teachers’ handbooks deliberately aimed at protecting or enhancing learner self-esteem (e.g. Rinvolucri 2002; Andres & Arnold 2009). However, empirical evidence of the motivational effectiveness of these teaching techniques is lacking.

3.4 Further theories of motivation

A number of other theories, prominent in mainstream education, have attracted the attention of L2 researchers and been found to have explanatory power for L2 motivation. However, with one or two exceptions, they have not yet been systematically applied and empirically evaluated in L2 education. These are included in Table 3, along with references both to the original authors of the theories and to L2 educational publications which explain them.

4. Empirical studies of teaching innovations

This section reviews empirical studies in which the rationale for the research derives primarily from the motivational potential of an innovative practice or teaching approach, rather than from theory (though theory may well be invoked in support). Topics are presented in a rough order of frequency of publication in major journals over the past decade. From this it will quickly be noted that computer assisted language learning (CALL)-related research is a dominant theme, though this might be slightly exaggerated by the existence of several specialist journals dedicated to this aspect of L2 pedagogy.

The majority of studies share a common basic research methodology – a pedagogical innovation or intervention is conducted with a non-generalizable sample of students, followed by a survey measuring learner attitudes towards or feelings about the specific L2 learning experience and/or L2 learning generally. In some studies this is triangulated with other kinds of data, usually interviews; observational data, such as evaluating learners’ classroom behaviour (as in some of the MotS research detailed here), is rare. Even rarer are intervention studies with pre- and post-tests of motivation. Randomized control trials are almost unheard of in L2 motivation research. Another general characteristic of the research is that it is short-term – usually reporting on an innovation lasting one semester or less – and the results might therefore be vulnerable to a novelty effect (as acknowledged in many studies). In the descriptions below, research methodology is only mentioned if it departs from the norm, or if the evidence it produces is particularly strong or weak.

4.1 CALL/digital technology

Developments in digital technology are probably the most prolific source of innovation in L2 teaching methodology in contemporary times, at least in western or developed world contexts, and the motivational properties of each innovation are usually considered an important aspect of its instructional qualities – after all, if learners do not like the innovation, then it will not be taken up, however great its capacity for language learning or acquisition.

Indeed, in their systematic review Macaro, Handley & Walter (2012) comment that the evidence for CALL promoting positive learner attitudes towards L2 learning is stronger than the evidence for beneficial linguistic outcomes.

For reasons of space it is not possible to review all studies reporting motivational impact from the use of CALL in L2 courses – readers are referred to recent authoritative review articles (Macaro et al. 2012; Stockwell 2013; Golonka et al. 2014; Bodnar et al. 2016). Instead, what follows is a synthesis of the main motivational benefits identified in studies of CALL innovations, which should be read in cognizance of Stockwell's (2013) point that students’ readiness to engage with new learning technologies will always be a product of a complex range of interacting factors, such as their familiarity with IT at home, the teacher's enthusiasm for and skills in IT, and the nature of the learners’ pre-existing motivation for learning a particular language.

4.1.1 Greater autonomy and individualization

One of the original rationales for CALL was its capacity for enhancing learners’ control over the learning process and hence for tailoring courses of instruction to learners’ individual needs (Warschauer 1996). There is evidence that this potential is being realized. For example, Bhattacharya & Chauhan (2010) report how a blogging project enhanced their Indian postgraduate students’ autonomous motivation to write, as well as their metacognitive skills. In a controlled experiment, Sanprasert (2010) shows how the introduction of an online course management system, with space for online communication and collaboration, improved her Thai students’ ability and willingness to learn independently. The individual convenience and personalization made available through mobile learning (Kim et al. 2013) and ‘ubiquitous’ learning (Jung 2014) have also been shown to increase learner satisfaction, though Kukulska-Hulme, Norris & Donohue (2015) warn that teachers need to ensure all learners in a class have access to similar devices and are willing for them to be used for educational purposes. On the basis of their own literature review, Golonka et al. (2014) argue that advances in automatic speech recognition software – offering finely tuned individual feedback for learning pronunciation – is one of the most promising uses of CALL, with evidence of learners showing increased confidence in their oral skills, when used regularly.

4.1.2 Enhanced opportunities for communication

IT can greatly expand the opportunities available for learners to try out their L2. This can be especially valuable in educational contexts where traditional classroom methodology does not offer much scope for communicative practice and few opportunities exist outside the classroom. Freiermuth & Huang (2012) for example describe how an ‘intercultural online synchronic chat task’ involving undergraduate learners of English in Japan and Taiwan was very positively evaluated by the participants, though they stress that the intense L2 practice it engendered was a product not just of the technology but of careful task design. Even in a multilingual European context, Jauregi et al. (2012) show the motivational benefits of synchronous computer-mediated communication as the Czech learners of Dutch in their quasi-experiment engaged with enthusiasm in native-speaker–non-native-speaker (NS-NNS) video-conferencing and reported deep satisfaction with the intensive speaking practice this provided (see also Wu, Yen & Marek 2011).

4.1.3 Identity development

Block (2007), Kramsch (2009), Norton & Toohey (2011) and others have stressed the potential of digital technologies for ‘trying out new and alternative identities and modes of self-presentation . . . without posing a threat to students’ real-world identities and private selves’ (Ushioda 2011b: 207). The empirical evidence for this phenomenon remains thin, however. Klimanova & Dembovskaya (2013) show through discourse analysis how American learners of Russian as a heritage language asserted their identities as authentic Russian speakers in a two-month-long tele-collaboration project with Russian learners of English, though there is no information on whether this impacted their longer-term investment in the language. Gleason & Suvorov (2012) found that the use of Wimba Voice for asynchronous oral language learning tasks, while popular with students, did not appear in the short-term to contribute to their future L2 selves.

4.1.4 Recognizing and utilizing learners’ existing IT skills

Henry (2013) offers a distinct motivational rationale for the use of CALL in L2 classrooms. In many contemporary societies, L2 learners are at least as ‘tech-savvy’ as their teachers, often rather more so, and therefore ‘teachers of English would profit from a better understanding of young people's leisure time activities’ (Henry 2013: 151) in order to create classroom activities that feel more authentic for learners i.e. more like what they do out-of-class, and with scope for creativity and meaning-making in the L2. Examples in the literature include Freiermuth's (in press) smartphone ‘treasure hunt’ using geocaching software with Japanese undergraduates, and Wehner, Gump & Downey's (2011) venture into the virtual world of Second Life with American students of Spanish, which both report highly positive affective responses from participants. There is likely to be a surge in publications on the use of digital games and ‘gamification’ in language learning in coming years.

To round off this section, the literature makes clear that CALL has the capacity to demotivate too, and such cases are probably under-reported for the obvious reason that innovators may prefer not to broadcast them. Stockwell (2013) cites two studies (Chen & Cheng 2008; Castellano, Mynard & Rubesch 2011) which had adverse effects, and ‘give us insights into what is necessary to avoid inadvertently damaging learners’ motivation to study using technology’ (p. 165); these include making sure that learners have the necessary IT skills or L2 proficiency to use the technology, and providing sufficient guidance and support throughout the learning process.

4.2 Content-based instruction

CLIL, in its various forms, is an increasingly common feature of national language (and especially English language) curricula worldwide, particularly popular with curriculum designers in Spain and in elite private education systems in the Far East. In addition to its efficacy in promoting language acquisition, its motivational impact has also been subject to empirical investigation in recent years. Doiz, Lasagabaster & Sierra (2014) built on an earlier study by Lasagabaster (2011) to compare the L2 motivation of 393 Spanish secondary school students of English divided into CLIL and non-CLIL groups in the first and third school years; they found the former to have significantly higher intrinsic motivation, instrumental orientation and interest in the L2 (though also higher anxiety levels). Mearns (2012) reports motivational gains from an experiment with English secondary school learners of German in the UK. Huang (2011) used Spada & Fröhlich's (1995) Communicative Orientation of Language Teaching Observation scheme to compare the behaviour patterns of the same 26 Taiwanese six-year-olds in two content-based language instruction classes and two direct language instruction classes, and found they participated more actively in the former. Seikkula-Leino (2007) found a slight motivational advantage among CLIL vs non-CLIL primary school children in Finland, while Lasagabaster & Beloqui (2015) found stronger intrinsic and integrative motives for learning English among CLIL-taught pupils in a Spanish primary school.

Although these results are encouraging, it should be noted that the studies conducted so far are all of modest scale and do not control for intervening variables (except gender). What is more, there are some less encouraging reports of practice in other global contexts. Coleman (2011) argued on the basis of a survey of English ‘bilingual education’ in three Asian countries that it was being implemented at the behest of ambitious school administrators and parents, with little concern for quality control, and could have damaging effects on young people's academic motivation and achievement. Kuchah (2013) raises similar questions about English-medium education in Francophone West Africa. The negative motivational impact of being educated in a second or third language is well documented in some multilingual developing world settings (e.g. Mohanty et al. 2009), and more substantial research is warranted on its effect in privileged urban educational settings too.

4.3 Intercultural content

Calls for a greater integration of culture and language teaching have increased in recent years, partly from an enhanced awareness of the interconnectedness of language and culture but also from a belief that such an approach would, by enriching syllabus content, be more motivating for learners (e.g. Kramsch 2009). Two recent studies have tested this hypothesis in relation to MFL teaching in Anglophone countries. Acheson, Nelson & Luna (2015) report a two-year study with 391 students of Spanish in two American secondary schools in which an experimental group were given direct instruction and practice activities in intercultural competence during their Spanish language lessons – post-tests revealed significant gains in their attitudes to Hispanic languages and people and greater motivation for L2 learning. However, Peiser & Jones (2013) suggest the issue is complex. In a mixed-method study of over 700 UK secondary school pupils, they found that their attitudes towards having more ‘intercultural understanding’ (IU) in the MFL curriculum were influenced by a range of societal, social, cognitive and institutional factors (e.g. more academic groups showed less interest in IU) and wrote ‘we cannot conclude from our study that a greater focus on IU is a panacea to motivation issues in the MFL curriculum’ (p. 354).

The cultural content of ‘authentic’ materials has long been recognized as having motivational potential (Peacock 1997). In the case of English as a global language, though, the question arises as to which culture is relevant, given that the language is used across a multitude of cultural and national borders, and what kind of communication is authentic, given that native speakers make up a minority of actual users (Pinner 2014). On the one hand, it can be argued that learners may be motivated by their (and their teachers) being convinced that local varieties of English are acceptable, even desirable, models to aspire to (Jenkins 2007) and that NNS-NNS communication in international settings is just as authentic as NS-NS communication (Hall 2013). On the other hand, there is empirical evidence that many learners still put a high value on prestige, NS varieties of English (Pan & Block 2011; Sung 2014) and can be motivated by exposure to NS teachers (Lamb & Budiyanto 2013), a perception that still drives employment practices in some education systems (see Hu & Mackay 2012).

4.4 Designing motivating tasks

Just as tasks had become a central unit of analysis in the study of SLA, Dörnyei (2002) argued that they could become a logical focus of study for a more situated, education-friendly approach to L2 motivation. As it turns out, there has been a rather limited amount of research conducted on task motivation in subsequent years. Broadly speaking, a distinction can be made between studies which have examined the interaction between task design features and learners’ actual engagement on the task, and those looking at the potential longer-term motivational effects of repeated use of a certain type of task. In the latter category can be included research studies which have empirically tested the assumption made by so many practising teachers in the ‘communicative’ era, viz. that learners of all ages can be motivated through the use of intrinsically stimulating and more personalized activities in class. Thus, small-scale studies by Ajibade & Ndububa (2008) in Nigerian secondary schools and Chou (2014) in Taiwanese primary schools provide some evidence for the positive motivational impact of using culturally appropriate word games, songs and stories in English classes. In an innovative research article, Kao & Oxford (2014) argue for the motivational properties of hip hop music in the English language classroom based on the first author's personal experience as father and teacher. Shaaban (2006) describes how cooperative (‘jigsaw’) reading tasks improved the motivation to read of Year 5 students of English in Lebanon, while Lo & Hyland (2007) report an action research study in a ‘well-established’ Hong Kong primary school which produced gains in young learners’ motivation to write English by (a) introducing more personally relevant topics and (b) providing new audiences beyond the teacher.

The more micro-level analysis of learner engagement on language learning tasks was pioneered by Julkunen (2001), Dörnyei (2002) and Dörnyei & Kormos (2000), whose work brought to light the formidably complex range of situational factors, learner characteristics and dynamic processes which shape any individual's motivation on task. Indeed, this very complexity may have deterred further empirical inquiry (though see Dörnyei & Tseng 2009 and Yanguas 2011) until recently, when complex dynamic systems theory has offered a promising new frame of reference in which the learner's task engagement could be viewed as a complex system within an ‘attractor state’ subject to various ‘control parameters’ like task topic, learner proficiency level, relations with peers, degree of teacher support, and intended outcomes. Some of the most interesting work is on oral tasks and the learner's ‘willingness to communicate’ (e.g. MacIntyre & Serroul 2015), which is beyond the remit of this review, but recent analyses of the dynamics of task (Poupore 2013) and lesson motivation (Waninge, Dörnyei & de Bot 2014) suggest this approach may eventually produce insights of value to teachers.

4.5 Increasing the relevance of the L2

A number of studies have looked at whether enhancing learners’ perception of the L2’s relevance for their lives can increase their motivation to learn it. In a rare example of a randomized control trial, Donitsa-Schmidt, Inbar & Shohamy (2004) report how teaching the oral form of Arabic, rather than the literary form, increased the motivation of Israeli Jewish primary school learners, because it was deemed to be more useful for pragmatic and cultural reasons. Fryer et al. (2014) argue that utility value also plays an important motivational role among the large numbers of tertiary level students worldwide who are forced to study English on non-elective courses, and Johnson's (2013) longitudinal study showed that awareness of its instrumental value for work was the single most important motivator among second year Japanese university students of Engineering, suggesting that teaching should try to reinforce the relevance by, for example, focusing on vocationally relevant genres. Meanwhile in Anglophone countries, perceptions of low relevance are claimed to be a major reason for learner demotivation (Coleman 2009). Taylor & Marsden (2014) describe an experiment in UK secondary schools in which 13–14 year-olds were exposed to one of two different ‘advocacy’-based interventions: a panel of external speakers who had all benefited from their knowledge of an L2, or a lesson in the language with an external tutor who had a strong message to convey about their experience with that language. Results showed that participation in the intervention did have a modest beneficial effect on uptake of the L2 in later years, and this was related to more positive perception of its personal relevance.

4.6 Alternative forms of assessment

The relationship between motivation and assessment is under-researched in mainstream education as in L2 education; a popular view in the teaching profession is that traditional summative pen and paper tests can provide some short-term motivation for more confident pupils (see Huang 2012, for evidence of this effect in Taiwanese English classes) but can also undermine long-term intrinsic motivation to study by encouraging all to value grades more than new knowledge and skills. Indeed, high-stakes testing regimes feature as a prominent demotivating factor in some research (see Section 5). Some empirical evidence is emerging on the motivational impact of more learner-centred forms of assessment. The introduction of self-assessment has been shown to boost learner motivation in Japanese university settings through a structured project (Kato 2009) and with intermediate Iranian EFL learners through written journals (Birjandi & Tamjid 2010). Lam (2014) presents testimony and a theoretical model of how portfolio assessment of L2 writing can promote student motivation and self-regulated learning. Other research is examining how formative feedback can be optimized to promote learners’ short- and long-term motivation to write in the L2. Busse (2013) for example found that UK university students of German had a strong preference for informational feedback expressed in an encouraging tone, though in a controlled experiment with Dutch HE students, Duijnhouwer, Prins & Stokking (2012) found no effect of such feedback on learners’ L2 writing self-efficacy.

4.7 Motivating boys

Gender has long featured as a significant individual difference in L2 motivation research, with females generally scoring higher for integrativeness and self-efficacy (e.g. Williams, Burden & Lanvers 2002), for having ideal L2 selves (e.g. Henry & Cliffordson 2013) for intended learning effort (e.g. Ryan 2009), for participation in foreign exchange programmes (Taylor 2000) and so on. In fact, there is a common perception globally that language is a ‘feminine subject’, taught by female teachers. Debate continues over whether the underlying source of these differences is biological, social or educational (Carr & Pauwels 2006) but there have been several noteworthy attempts to overcome the male deficit. One approach has been to separate the sexes for language classes, in the belief that L2 use invokes teenagers’ gendered identities in ways that other subjects do not. Results of actual experiments are mixed. Chambers (2005), for example, found in a UK comprehensive school pilot project that pupils’ overall enjoyment and confidence increased but teachers complained of difficulties teaching mixed-ability classes (they were normally streamed by ability) and of dealing with bad behaviour in all-male classes. By contrast, in Canada, Kissau, Quach & Wang (2009) found that boys flourished in all-male classes but girls did not enjoy their all-female classes. A different approach was taken in the Score in French project in a UK secondary school (McCall 2011), which unashamedly aimed to make the language classroom more ‘masculine’ by creating a French curriculum based on football. Through sensitive task design, exploiting the appeal not just of the game but of certain French players, the pilot project was successful in increasing the L2 motivation of both boys and girls.

5. Demotivation

Such is the global allure of English that, in urban settings at least, few young people enter secondary school without a modicum of desire to learn it, and for many the desire is deeply felt and persuasive (Motha & Lin 2014). So, as Littlejohn (2008) has argued, in some ways the English teacher's main task is not to motivate learners but to prevent their demotivation, that is, the gradual loss of a pre-existing motivation to learn. Moreover, it is a well-established truth that more intrinsic forms of motivation tend to decline during the early years of secondary school (see Yeung, Lau & Nie 2011, for confirmation of this trend for English language learners in Singapore) and learner motivation is particularly vulnerable at transition points like primary to secondary (Burns et al. 2013) or foundation course to HE (Woodrow 2013). It is not surprising then that during the 2000s a research literature began to appear on demotivation in L2, ably summarized in Dörnyei & Ushioda (2011).

As they make clear, some sources of learner demotivation are beyond the immediate control of teachers. For example, societal discourses may negate the value of learning an L2 (e.g. see Lanvers & Coleman (2013) on UK media attitudes towards L2 learning); nationally imposed curricula can curtail the freedom of teachers and pupils to pursue local interests in class (e.g. Wedell 2009); high-stakes assessment regimes can raise anxiety levels and displace intrinsic motives for learning with a pursuit of grades and qualifications (e.g. Cheng, Watanabe & Curtis 2004); low investment in language education by governments results in poorly resourced classrooms and poorly rewarded teachers (e.g. in many developing countries, see Bennell & Akyeampong 2007). However, there is also plenty of evidence to suggest that learners are sometimes demotivated by what happens inside the language classroom. In fact, in some, studies teachers have been clearly identified as the main source of demotivation (e.g. Falout & Maruyama 2004). A non-exhaustive list of negative practices identified in recent research is presented in Table 4.

It is important to note another point emphasized by Dörnyei & Ushioda (2011); demotivation is as complex as motivation, in that learners react differently to the same classroom procedures, and the same learner may react differently to the same stimulus depending on their stage of learning, their mood, the disposition of peers and so on. One-off studies have already shown that different factors will tend to demotivate learners of high versus low proficiency (Falout & Maruyama 2004) and high versus low pre-existing motivation (Sakai & Kikuchi 2009). Longitudinal studies suggest that for some learners a spiral of decline can develop, when frustrating classroom experiences engender negative attitudes which discourage extra-curricular practice, lowering confidence further, and so on (e.g. Lamb 2011; Busse & Walter 2013). Other learners seem able to have negative classroom experiences without letting them affect their overall longer-term motivation (Campbell & Storch 2011). As Dörnyei & Ushioda (2011: 156) write, ‘the current shift towards socio-dynamic perspectives on L2 motivation research . . . seems well suited to investigating the “dark side” of motivation’, because of its capacity to identify patterns of interaction over time between different elements in the classroom environment and beyond. Very recent studies which have adopted this approach do indeed present vivid pictures of fluctuation in learner motivation, from the scale of one academic year (Kikuchi in press) to single language lessons (Waninge et al. 2014); it is perhaps too soon to draw strong pedagogical implications, beyond a heightened awareness of complexity and dynamism, and of the importance of ‘initial conditions’ i.e. getting off to a good start.

One further issue raised by research on demotivation is that of teacher motivation, for evidence suggests the two are closely connected. The teacher factors which repeatedly appear in lists of potential demotivators (cf. Dörnyei & Ushioda 2011; Kikuchi 2013) – monotonous methodology, low enthusiasm for the subject, unfriendly demeanour, lack of attention to individual needs – are often a symptom of an underlying lack of work motivation. Conversely, inspiring teachers are often those who are highly motivated themselves, since it takes great energy and commitment to offer constant variety, and show unwavering enthusiasm, approachability and adaptability (Lamb & Wedell 2015). In a rare attempt to link learner motivation with teacher motivation, Bernaus, Wilson & Gardner (2009) found that the relationship was indeed mediated by teachers’ use (or not) of MotS, concluding: ‘Our results suggest that both students and teachers need to be motivated. If teachers are motivated, students are more actively involved in class activities and feel more motivated’ (p. 33). The few published studies that have looked directly at L2 teacher motivation (e.g. Karavas 2010, in Greece; Aydin 2012, in Turkey; Wyatt 2013, in Oman) portray it as an area of concern, both for its own sake and for the knock-on effects of learner motivation.

6. Summary

This review has identified well over 200 pieces of research published in academically respectable outlets in the last 15 years that have addressed the motivational dimension of language teaching. The research can be classified into four broad types: investigations of the strategies that teachers deliberately use to motivate classes of students, a line of research that is gradually growing in sophistication (Section 2); studies that apply theoretical insights about human motivation to L2 educational contexts, and here SDT and the L2 motivational self-system dominate (Section 3); studies that have described the motivational impact of various pedagogical innovations, the majority of which have involved the introduction of learning technology (Section 4); and research into the phenomenon of learner demotivation (Section 5), often located in Anglophone countries or in Japan. It is time to pause for reflection: what can we state with confidence about the effects of teaching on L2 learner motivation? What should teacher educators do with this knowledge? And what areas of new knowledge should future researchers prioritize?

6.1 What do we know?

6.1.1 Teachers can motivate

Danziger (1997) points out that ‘motivation’ as a field of psychology has its origins in the expansion of mass education in the early decades of the twentieth century, and the recognition that people, whether children in classrooms or consumers on the high street, could not be manipulated by force; instead ‘[o]ne had to play upon what individuals wanted, what they were interested in, what they privately wished for’ (p. 113). We now have sufficient research evidence to suggest that language teachers are able to influence their learners’ motivation, both for better and worse. This probably accords with the experiences of most teachers, who will have registered the impact of their work on their learners’ views and feelings about the subject. The bulk of the research evidence, however, is indirect, deriving from non-controlled studies of learner attitudes in response to some kind of short-term innovation in teaching; there have been very few controlled experiments measuring the impact of teachers, and few that have included the direct observation of learner behaviour, in class or outside. What is more, the focus has almost always been on immediate outcomes. There is as yet very little evidence about the long-term impact of teachers on learner motivation, apart from in retrospective accounts by learners recalling favoured teachers in the past (e.g. Shoaib & Dörnyei 2005; Lamb & Wedell 2015).

6.1.2 The personal is paramount

The aspects of teacher behaviour which appear to have the most motivational impact on language learners are those that relate to the human side of teaching. In the MotS research, the strategies most consistently valued by teachers as well as learners emphasize the connections between people: learners have to respect the teacher as a professional, and ideally like them as a person; the teacher has to understand the needs, goals and desires that learners bring to class from their lives outside and from prior experiences of learning; they have to create a group dynamic which accommodates and excites all of them; they need the professional dedication to continually respond to, and sometimes try to change, those needs, goals and desires. Teacher–student and peer relations have been found to be important to learner motivation in general education (Wentzel 2009), but it is possible that they are even more central to language teaching, which necessarily involves more frequent interpersonal communication.

6.1.3 Methods matter, but so does context

To maintain good relations while also ensuring that sufficient language learning takes place to give learners a sense of progress, takes considerable technical skill. Much of the evidence reviewed here relates to creative classroom practice, techniques for providing learners with stimulating and satisfying tasks. Novelty and variety may themselves be motivating, but also presenting classroom activities so that learners know why they are doing them, setting just the right level of challenge, and providing informative feedback afterwards are all key motivational skills. These actions must always be made to fit the precise context too – appropriate to the local educational culture, meeting broader curriculum goals, matching the preferences and interests of particular learner groups, and wherever possible adapted to individual learner needs, wants and identities (present and future).

6.1.4 Learner control

As if all this was not difficult enough, the paradox of good teaching is that it must be done while allowing learners to feel in control of events. It is not a coincidence that the common theme running through all the motivational theories reviewed here is that people need to feel they are learning for their own reasons, in a self-determined way, autonomously, efficaciously, in pursuit of an ideal self perhaps, or at least an internalized goal. The research evidence in favour of ‘autonomy-supportive’ teaching in general education is increasingly persuasive (Reeve, Deci & Ryan 2004). In L2 education, the research evidence is mostly negative in nature; that is, experienced teachers and motivated learners do not mention reward schemes or other overt forms of manipulation as being likely to motivate learners; entertaining classroom activities might be valued, but only when offered strategically as part of a course of constructive learning. This is the reason why some commentators have suggested that the key question teachers need to ask themselves is not ‘how can I motivate my students?’ but ‘How can I create the conditions under which students will be able to motivate themselves?’ (Reeve et al. 2004: 53).

6.2 The role of teacher educators

Indeed, this might be one of the key questions for novice teachers to answer while in training. If success in language learning is partly determined by the learner's motivation – and research suggests that it is associated with 18–33% of the variation in achievement (Masgoret & Gardner 2003) – and the teacher is an important influence on that motivation, then it is reasonable to argue that the motivational dimension should be an essential component in any course of training for language teachers. In fact, Wigfield, Cambria & Eccles (2012: 474) argue that ‘with the current press for more assessments and evaluation of student performance, and teacher and principal accountability for student performance on those tests’ it becomes even more imperative that motivation research makes a contribution to teachers’ professional development.

However, the existing research also tells us that there are many cultural and contextual factors which mediate the motivational impact of teacher behaviour. What works in one educational context may not work in another. The moderate results usually obtained in L2 motivation research also reminds us of the significance of individual learner differences, since pedagogical innovations rarely gain universal approval – what works for one learner may not work for another. Therefore, language teachers-in-training need to become aware of general principles of L2 motivational practice – the four earlier points (6.1.1 – 6.1.4) are a place to start – and of sets of available strategies, as so clearly presented in Dörnyei (2001); yet they also need to be made keenly aware that the effective operationalization of principles and strategies is contingent on local conditions and the needs, goals and desires of the actual persons involved.

This matches what we now know about the complexities of teacher learning. It is recognized that training novice teachers or persuading experienced teachers to use a pre-ordained set of instructional strategies is a recipe for failure if the strategies do not align with teachers’ underlying beliefs and values (Borg 2003; Kubanyiova 2012). Even the four broad ‘truths’ about motivation elucidated here may present a challenge to teachers’ existing beliefs. For example, it is possible that many language teachers have what might be termed a ‘fixed mindset’ (Dweck 1999), believing that learners either are or are not motivated for languages, just as they either have or do not have L2 aptitude, and that this is not open to change. Other teachers may resist the notion that warm personal relations are important for motivation, preferring instead to maintain a disciplinary distance and authority. They may well have firm views on what practices ‘work’ in the classroom and be resistant to methodological innovation. And even when convinced that learner motivation is their own responsibility, there is the danger that they will assume too dominant a role, undermining learners’ internalized motivation for the subject. In their motivational intervention study in the USA, Turner, Warzon & Christensen (2011) found that maths teachers’ ability to reflect on their own beliefs and understand the rationale behind motivational practice was crucial to their ability to change.

In addition to motivational awareness-raising, teachers might benefit from training in recognizing learner motivation and demotivation. As argued above, successful practice involves constantly responding to learners’ needs, goals and desires, over the timescale of lessons and courses. This, in turn, requires the ability to infer what learners dislike and predict what they might like. This can be developed through training in classroom observation, the use of devices like needs analysis questionnaires, lesson reaction slips and course evaluation surveys, journal writing and other techniques of reflective practice (Farrell 2008).

6.3 Directions for future research

This final section will consider how research on the motivational dimension of language teaching can be improved in quality, so we can have more confidence in its results, and can be made more useful to the language teaching profession so that as educators we can more readily provide the conditions for students to motivate themselves.

6.3.1 Teachers’ beliefs, motivation and practice

As discussed above, teachers’ beliefs and values profoundly influence their practice, and would mediate any attempts to make their teaching more motivating. It is true that the MotS research has brought important insights into the specific strategies that teachers value, but because it has largely been done through surveys based on researchers’ own conceptions of motivational practice, we cannot be sure that they represent teachers’ own core beliefs about how to motivate learners. We also lack evidence about how teachers’ conceptions of motivation are reflected in their practice. A recent study by Muñoz & Ramirez (2015), for instance, using an SDT theoretical framework, found that Colombian university language centre teachers appeared to recognize the motivational value of giving learners choice and introducing meaningful language tasks, yet rarely did either during observed classes. Kubanyiova's (2006) study of a training course designed to help teachers move towards a more ‘motivational teaching practice’ found that, although the participants were enthusiastic about the course and recognized the value of the proposed MotS, they did not adopt them in their teaching; she ascribed this ‘failure’ to both contextual constraints and to the nature of the participants’ own motivation to teach. Yet Feryok & Oranje (2015) show that even when teachers’ main focus is on other aspects of pedagogy, such as an intercultural class project, a desire to interest and motivate learners is a recurring motif in their lesson planning and moment-to-moment classroom decisions,

If we want to understand why and how teachers adopt and adapt MotS, researchers need to engage with the literature on teacher cognition, and conduct in-depth qualitative studies of individual teachers as ‘persons-in-context’ (Ushioda 2009). They need to recognize the complexity of teachers’ mental lives by enquiring into teachers’ prior educational experiences, as learners, as teacher trainees, and as novice teachers, to see how their thinking has evolved over time, how they orient to the profession, and how they perceive the affordances and constraints of their particular context (Kubanyiova & Feryok 2015). They need to observe classroom events closely (for example using stimulated recall techniques) to understand teachers’ ‘thinking-in-action’ and the way they interactively create engaging experiences in class, or alternatively fail to do so. Such methods would also help to overcome another flaw noted in the MotS approach, namely uncertainty over the meaning of particular strategies.

6.3.2 Motivating young language learners

L2 education is beginning at ever younger ages in many global contexts, and one of the main rationales for its introduction into the primary school curriculum is motivational – the belief that teachers can foster positive attitudes towards L2 learning which will pay dividends later (Enever 2011; Heinzmann 2013). But most of the research reviewed here has been concerned with motivating older learners. The future self- and identity-based theories which have recently dominated thinking about L2 motivation are of questionable value in informing young learner teaching, for psychological evidence suggests that ideal selves only take shape in early adolescence (Zentner 2007). Children's learning behaviour is likely to be much more contingent on immediate classroom processes, emphasizing the key role of the teacher, yet the MotS taxonomy was built up on the views of experienced teachers working mainly at secondary and tertiary levels. A priority for future L2 motivation research is therefore to analyse the classroom experiences of young L2 learners and teachers, and this is likely to require innovative methods such as metaphor analysis (Jin et al. 2014) and the use of system logs to track learners’ online game playing (Bodnar et al. 2014). There are suggestions that the initial positive effect generated in early years classrooms may be undermined as L2 learning becomes more challenging (Enever 2011), so longitudinal studies would be especially valuable.

6.3.3 Intervention studies

Understanding what teachers and learners currently do is important, but motivation researchers should also be pushing the field forward through theoretically informed, empirically robust interventions, working with groups of teachers to help make their practice more motivating for their learners. The practical and ethical challenges involved in such research no doubt explain why quasi-experimental interventions – involving attempts to control intervening variables and gaining permission form diverse stakeholders (e.g. as in Moskovsky et al. 2012; Alrabai 2016) – make up such a tiny proportion of the research reviewed in this paper, yet such studies are to be encouraged because they offer the most persuasive evidence of motivational impact. They do not need to be large-scale, as long as the intervention itself is theoretically sound and implemented consistently, and appropriate statistical procedures are used (see Rosenzweig & Wigfield (2016) for an authoritative review of motivation interventions in science, technology, engineering and maths subjects).

The new theoretical concept of directed motivational currents (DMCs) (Dörnyei, Henry & Muir 2016) – periods of intense activity in pursuit of a short-term goal – may prove to be a fertile source of interventions. The rise of digital technology, as well as aspirations towards more radically learner-centred teaching, has led to renewed interest in project work across the curriculum (e.g. see Patton 2012). Dörnyei et al. (2016) present a series of frameworks for class projects in language learning based on their understanding of the essential features of a group DMC:

• It should start with a clear collective goal, for which everyone feels a sense of ownership
• There are sub-goals and progress checks built into the process
• It generates positive emotionality in the group
• It has a demonstrable outcome, in the form of a performance/exhibition/production

An alternative form of intervention is exemplified in Turner et al. (2014), who worked with a number of US high school teachers over a three-year period to help them apply a set of four motivational principles (partly based on SDT) and found that repeated class observations and post-class discussions did help to make the teaching style of some (but not all) teachers more supportive of their pupils’ engagement. The longitudinal nature of this study is worth remarking on: recognizing the complexity of classrooms, the researchers were not seeking immediate or linear cause and effect relations between teacher behaviour and learner engagement, but working towards the gradual emergence of new patterns of teacher–student relations.

6.3.4 Teachers (and learners) as (co-)researchers

Any successful intervention study requires the sustained cooperation of teachers, not just to grant access to their classrooms but because only a teacher will know exactly how motivational principles or strategies should be operationalized in any particular classroom. There are other good reasons why future research on L2 motivation should be done with rather than on teachers. First, experienced teachers build huge funds of knowledge over their professional lifetime which can and should inform theory. In a chapter on the motivational power of having learners ‘speak as themselves’ in class, Ushioda (2011a) acknowledges that ‘communicative’ and ‘humanistic’ language teachers had discovered this for themselves at least two decades ago. In this case, academics were adding a psychological rationale for a practice that was more usually justified in psycholinguistic terms, for producing ‘enhanced output’. In other areas motivational theorizing is guiding pedagogy, for instance in developing L2 future self-imagery (Hadfield & Dörnyei 2013), or has the capacity to do so. The point is that greater understanding of the motivational dimension of language teaching can only come through close collaboration between researchers and teachers.

A further reason why practising teachers need to be involved in L2 motivation research is that the process of researching can itself be motivating, for teachers and learners. Proponents of action research and exploratory practice (e.g. Farrell 2008; Allwright & Hanks 2009) claim that systematic investigation of their own classes helps teachers develop feelings of control and competence, a stance supported by some of the practitioner researchers cited in this review (e.g. Mearns 2012; Sampson 2012); learners too can be motivated by involvement in classroom research as they gain insight into aspects of their own practices and thought processes (e.g. Murphey & Falout 2010; Coyle 2014).

6.3.5 Topics of controversy

Debates have raged in mainstream educational psychology that have barely caused a flicker of interest in L2 motivation circles, possibly because we are still profoundly influenced by Gardnerian views (1979) on the distinctiveness of desire to learn language. For teachers working in language classes in schools or colleges, however, there are fundamental motivational issues that they share with teachers of other subjects, and these deserve more empirical investigation. One example is the use of rewards, particularly prevalent in primary/elementary schools where language teaching is increasingly based, but also common in the certification systems of private language schools, and most recently in the gamification of digital learning tasks. The success of such methods would appear to contravene the tenets of SDT, in that the regular offer of extrinsic rewards for school work may undermine the development of long-term intrinsic motivation to study, but this view has also been challenged (e.g. Cameron, Banko & Pierce 2001). Likewise, the way teachers commonly praise pupils, focusing on the person rather than performance of the task-in-hand, has been criticized for substituting short-term emotional rewards for more enduring forms of motivation (Dweck 1999). Summative assessment is another ubiquitous extrinsic motivator, undeniably stimulating in certain contexts, as all experienced teachers know, yet also perceived as controlling by students and frequently a source of anxiety (Pulfrey, Darnon & Butera 2013). Notwithstanding their complexity, these are surely issues that should interest L2 motivation researchers because they are deeply embedded in the practices of language teachers.

7. Conclusion

The surge of interest among academics in the motivational dimension of language teaching is to be broadly welcomed. It is serving to mediate between L2 motivation theory and classroom practice, enhancing the real-world impact of the former, expanding the motivational awareness of pedagogical innovators, and potentially providing novice teachers with useful guidance on this important aspect of their job. It is helping us to understand the nature of real-world problems too: why so many state school pupils are demotivated in their study of English when it could be so valuable in their lives; how English mother tongue pupils can be persuaded to learn FL when their value is not immediately obvious; how people who do have L2 goals can be pushed to invest even more effort into learning than they currently do; why teachers need to find a balance between short-term stimulation, through entertainment, rewards or the threat of sanctions, and inspiring a long-term dedication to learn the L2 through the nurturing of internal motives.

However, at this stage in its evolution, quality needs prioritizing over quantity. This could be served by following the directions outlined above, as well as by upscaling sample sizes, implementing bolder interventions, deliberately targeting areas of pedagogical or theoretical challenge. The groundwork has been done; it is time for more ambitious projects to build our understanding of how teaching motivates and demotivates language learners.