Autonomy in second language phonology: Choice vs. limits

Abstract

Learning a new sound system poses challenges of a social, psychological, and cognitive nature, but the learner's decisions are key to ultimate attainment. This presentation focuses on two essential concepts: choice, or how one wants to sound in the target language; and limits, or various challenges to one's goals vis-à-vis accent. Qualitative and quantitative data underscore the relevance of autonomy as a guiding principle from which to explore related constructs such as self-determination, motivation, decision-making and self-concept. I also review several prominent limits on phonological attainment to counterbalance and contextualize the aspect of choice. Suggestions are given for both teaching and research that prioritize autonomy with reference to a complexity perspective.

1. Introduction: The complexity of accent

The field of second language acquisition (SLA) has long focused on universal patterns of language development, attributing the outcomes among adult second language (L2) learners to age of onset – the beginning of exposure to, and experience with, the target language. Within this tradition, the difficulty of acquiring a new system of sounds and articulatory habits is widely cited, perhaps occurring as early as five–six years, if not earlier. In fact, phonology has long been assumed the ‘final frontier’, the realm where limits of a neurocognitive nature most interfere with full learning. This necessarily applies to everyone, across the board, given its presumed biological foundation (Muñoz & Singleton 2011). More than a few scholars have found this theory compelling, given how rarely late language learners end up sounding native-like.

The priorities of the field have changed in the last decade, however. In 2002 Vivian Cook wrote that a focus on the L2 user ‘alters the perspective of SLA research. . . .[it] is no longer about finding excuses why L2 users are failed native speakers, but can explore what makes L2 users what they are’ (2002: 19 – emphasis is mine). Indeed, Dewaele (2005) has encouraged us ‘to abandon the dream of representing a learner's progress in a gently upward line toward native-like status’ (p. 371). There is greater appreciation for individual differences and learners who carve out their own path, gradually building a sense of self in the target language, whether or not they end up sounding native-like. On the negative side, we give up the security of what were previously thought to be ‘hard and fast rules’ about interlanguage and the ‘end-state’ of attainment according to age-based constraints. On the positive side, we gain exciting new directions for research and teaching that incorporate learner priorities and goals. It is in this spirit that I focus on a few essential constructs going forward, to better understand L2 phonological attainment.

With L2 phonology research now pivoting toward context and learner, we are in position to shape the age debate in SLA, to ask not only what makes accent challenging, but to also investigate how it hinges upon what the target language means to the learner, and how that translates into her approach to the task. In so doing, we acknowledge fluidity in the negotiation between the person and her environment, and complexity – the idea that ‘learners interact locally’, and continually undertake a kind of adaptive imitation to suit their communicative needs (Larsen-Freeman 2011: 49). I believe that the notion of a ‘complex, dynamic system’ could be a useful framework for the study of accent, given its emphasis on an adaptive, self-organized system that is continually restructured, as new information and feedback are absorbed and new conditions emerge (Larsen-Freeman 2011: 51).

A constructivist, ‘complex’ approach aligns well with the integrated model for L2 phonological attainment I proposed some time ago (Moyer 2004). See Figure 1.

Figure 1 An integrated view of critical influences in SLA (adapted from Moyer 2004)

What I tried to capture, based on quantitative and qualitative data from a group of immigrants to Berlin, was the balance between age-related influences on accent. These work in tandem to define where on the native/non-native spectrum an individual is located at a given point in time. The idea is that cognitive, social, and psychological processing all affect how the learner uses input strategically, and these realms are interconnected through specific mechanisms.

Larsen-Freeman (2011) describes language acquisition as driven by ‘cobbled together’ resources based on who one's interlocutors are and what activities one is engaged in (p. 54). I would like to expand this idea of resources beyond external factors like opportunities to speak, language instruction, and so on, to include internal processes like motivation, self-regulation, self-concept, and autonomy, all of which are fundamental for getting the L2 user where she is ultimately hoping to go.

While a learner-centered approach is not new to foreign language pedagogy, nor to sociocultural and individual differences research in SLA, it is less widespread in L2 phonology studies (or pedagogy). I believe it is well suited to the complexity of accent, so I will build on that notion today with a few relevant concepts from educational psychology, and the L2 strategies and motivation literature.

We are all keenly aware that phonology relies on a unique combination of skills. L2 users need to perceive fine differences in phonetic categories, apply phonological rules in patterns counter to their native tongue(s), and replicate stress, intonation, juncture, and rhythm features that are essential if they hope to accurately convey meaning and speaker intent. This is a cognitive skill set, but it also involves pragmatic, strategic, and social awareness, not to mention a willingness to project oneself at the basic level of sound and voice, in a context that may be less than familiar or comfortable. Accent, in other words, is uniquely connected to identity and self-concept.

If learning a new sound system poses challenges from many corners – cognitive, psychological, and social – it must take great determination to get to something close to native-like. Let us therefore suspend the traditional view that phonological acquisition follows a pre-determined path and inevitably ends up in a ‘less-than’ place. Instead, let us approach it from the standpoint of autonomy – asking how the learner engages in the process – in order to capture what is, by nature, a very dynamic phenomenon.

I begin by defining autonomy in general. Autonomy centers around volition, or self-determination, defined as the capacity to choose, to determine one's own actions, and to behave in intentional (as opposed to coerced) ways (Deci & Ryan 1985). This idea of ‘choice’ is a foundation of the ‘social turn’ in SLA, but it has too rarely been explicitly incorporated in the study of L2 phonology. Toohey (2007) further characterizes autonomy as socially situated, i.e., what the learner does in certain contexts. The focus is on context and interaction as Toohey emphasizes learning as a product of increasing participation in a community (p. 231f.). This holds for the classroom as well – another kind of community where involving learners in decision-making and reflection is key to fostering their sense of autonomy (Little 2007).

In any context, autonomy centers on the dialogic between person and environment – a cornerstone of complexity theory. So with this in mind, I would like to consider autonomy for L2 phonology in terms of a basic juxtaposition:

• • Choice – the learner's decisions about how she wants to sound

• • Limits – the potential challenges to her goals vis-à-vis accent

I suggest this juxtaposition as a conceptual framework to capture the convergence of intrinsic and extrinsic resources any learner has at hand. This is where the rubber meets the road – where learners really engage with the target language. It is that inevitable tension between choices and limits that holds the key to phonological outcomes.

2. Choice as a critical aspect of phonological development

Let us begin by considering choice: What decisions do learners make about accent, and how deliberate are their choices? This quandary was brought home to me when I studied immigrants to Berlin, L2 users of German, where I found very different approaches to accent among two sets of learners (Moyer 2004). In the late learner group were two immigrants (a woman from Poland and a man from America) with superficial similarities in terms of career goals and types of L2 exposure. In the early exposure group were two Turkish immigrants who had resided in Germany since early childhood (from the age of four). These two had similar access to social networks and schooling in the target language. Now, within each pair – early and late – was one learner whose in-country experience had led to a deeply valued bilingual identity of which accent was a crucial part. And within each pair was another learner who expressed an actual aversion to ‘sounding German’ or being too closely affiliated. Not surprisingly, their strategies for language contact were in no way similar to the strategies adopted by those who sought to assimilate. And their accents, as rated by native listeners, confirmed these differences in approach. So, one of the very early learners sounded identifiably ‘foreign’ and one of the late learners was judged to sound native on most tasks. In short, age of onset was not telling the story of phonological attainment (see similar findings in Moyer 2007). At the time, I thought of this as an identity issue, but looking at it in hindsight, I think it comes down to internally motivated decisions manifested as outward behaviors, or put in broader terms, as affect driving metacognitive responses.

The conundrum of choice is front and center in studies that go beyond situational or episodic passing, i.e., where people really do ‘create personae or identities’ with the help of ‘borrowed linguistic resources’ (Nortier & Dorleijn 2008: 128). Passing is limited, temporary. Much tougher is the more stable phenomenon of styling, seen in Cutler's (2014) story of Isko, a 21-year-old Bosnian Muslim who came to the U.S. as a 16-year-old during the Bosnian civil war:

As a good student with a strong desire to succeed in the US, Isko has mastered English and has few signs of an L2 accent, yet the English he has acquired has a distinctly AAL [African American Language] flavor. . . Isko describes his long-standing interest in hip-hop music and his identification with African American culture. . . he talks about how he feels more comfortable around African Americans, which he says is perhaps partly due to the fact that people in the US often think he's Hispanic rather than White. (p. 155f.)

Isko's ‘styling’ – his adoption of culturally and linguistically symbolic features in a generalized (not just situational) way – shows how he adopts a specific accent to project aspirational belonging, even crossing racial boundaries (even if the acceptability and legitimacy of doing so is questioned, as Rampton (1999) points out). Based on insights gleaned from Isko and interviews with others of various first language (L1) backgrounds, Cutler concludes that L2 users are conscious of their choices and the implications of using certain accents. They also purposefully adopt attitudes, stance, dress, mannerisms, etc. to express belonging, although not necessarily conformity to the ‘native speaker/standard ideal’ – a construct that has been problematized in SLA research generally, and certainly within the phonology literature as well (see Levis 2005). Even in limited ways, learners can choose to avoid correct pronunciation of some canonical features, e.g., English th– (to pronounce that as dat) to preserve their cultural or ethnic identity (Gatbonton 1975, cited in Hansen-Edwards 2014), or they may vary pronunciation of –ing based on the gender of their interlocutor (Adamson & Regan 1991; see also Gatbonton, Trofimovich & Magid 2005; Rampton 2013).

So accent is malleable at a conscious level, but like any good story, it's complicated. The boundaries of learner choice can get blurry. Some contexts impose constraints not of our choosing. I may experience negative reception in a host community as a result of my L1 identity, and this could affect my chances to establish solid social networks, thus affecting my long-term fluency. Another example is the interpersonal dynamics of a given classroom, and the social pressures imposed by my classmates and teacher vis-à-vis accent (Lefkowitz & Hedgcock 2002). This can impinge on my willingness to ‘try on’ new sounds and ways of speaking if I do not want to sound ‘different’ or be held to an unrealistic standard. In short, even autonomy is subject to larger ideologies and contextual constraints, no matter how otherwise willing or able we are.

With that noted, let us look in more detail at the concept of autonomy: How is it influential for accent, and what are its related dimensions?

3. Autonomy as a framework for understanding phonological outcomes

I have referenced self-determination as key to the concept of autonomy. Self-determination guides our actions toward a future vision, as Deci & Ryan put it (1985), and requires psychological ‘flexibility’ to negotiate the interaction of ourselves in the environment (p. 38). Required to sustain that future vision is motivation, long confirmed as significant for phonology (see Moyer 2013). Nowadays, motivation is no longer seen as an achievement-oriented construct, but as an integral part of identity and personal learning trajectories (Ushioda 2011: 222). So this individual difference construct has evolved beyond the dichotomy of instrumental vs. integrative, taking on a much more learner-centered aspect, and furthermore, a much more individualized aspect.

3.1 Motivation and self-regulation

Motivation influences cognitive and metacognitive strategies directed at pronunciation regardless of environment, i.e., for both immersion and foreign language (FL) classroom learners, as confirmed by Baker-Smemoe & Haslam (2012). I also know that highly motivated learners adopt more, and varied, learning strategies, which correlates with comprehensibility and/or closer-to-native ratings (Baker-Smemoe & Haslam 2012; see also Moyer 2004). Motivation likely works its magic as a function of intensity and consistency rather than type (see findings in Moyer 1999, 2004, 2007) and is most significant for accent when it includes more than one of the traditional categories (instrumental or integrative) (Moyer 2007). This suggests a complex investment in the target language, which is, arguably, necessary to overcome many of the age-related challenges we’ll speak of momentarily.

Several key dimensions likely keep motivation alive, according to Dörnyei (2005):

• • commitment control to preserve one's focus on goals

• • metacognitive control to monitor and control concentration

• • emotional control to generate emotions that support one's intentions

• • environmental control to reinforce a positive outlook in difficult circumstances

All of these dimensions can be found in self-report data from various phonology studies, so I include a few quotes here from several of those Berlin immigrants mentioned who were rated within the native speaker range for most tasks (Moyer 2004):

1. 1. Commitment control: (The prompt from the semi-structured interview was: If you find your motivation waning, what specific behaviors do you undertake?)

   Maria: I try to make more friends because personal contact with native speakers is more motivating than just studying.

2. 2. Metacognitive control:

   Maria: I try to address difficulties through social interactions with native speakers and imitate their pronunciation as much as I can.

3. 3. Emotional control:

   Blanka: If you aren't really proficient with the language or you have a strong accent it's probably true that it's harder to make contacts, but for me, I don't differentiate between Germans and others. I’m open to making all kinds of friends.

4. 4. Environmental control:

   Dorota: What doesn't please me I have long since learned to live with . . . I guess I’ve become half-socialized in that way.

   Maria: Sometimes I make a conscious effort to remind myself what I have in common with Germans, when I come across a situation where I might react negatively as a Russian.

(Noteworthy is the element of control over self-concept revealed in these comments.) These mechanisms are likely interconnected, meaning that a breakdown in one impacts the others, according to Rose (2012).

So motivation directs and sustains behaviors aimed at improving fluency, and this is important in a classroom setting as well, where Wu (2003) has tied motivation to autonomy statistically through students’ choices about content and language learning tasks, and opportunities to be creative, which increase one's sense of competence and control. Bernaus (2010) summarizes the potential of the classroom this way:

If learners feel that they are their own masters in their learning they will be motivated to learn and of course to attain higher levels of achievement. . . . Classroom activities and materials in language learning should therefore utilize meaningful contexts of genuine communication with persons together engaged in the process of becoming persons. (p.185) (emphasis is mine).

3.2 The dynamics of self-concept

Let us explore that link between autonomy and self-concept (see Ushioda 2011). The self is a ‘coherently organized dynamic system encompassing all the beliefs, cognitions, emotions, motives and processes related to and concerning oneself’ (Mercer 2014: 163). The self continually evolves (Mercer 2014) and perhaps even ‘oscillates’ due to our affiliation with various communities (Sade 2011). What does that mean from the learner's perspective? All of us have some idea of the attributes we would most like to see in ourselves as L2 users. If I desire a certain level of mastery, and believe that this goal is attainable, I develop a ‘roadmap’ for reaching that ‘ideal L2 self’ (Dörnyei 2005).

Ideal selves, or possible selves are goal-oriented and action-oriented (Frazier & Hooker 2006). A mountain climber first envisions herself on the top of a mountain, as they put it. She knows she needs great strength and willpower to achieve that aim, so she goes to the gym today (and next week, and so on) (p. 42f.). Gregerson & MacIntyre (2014) write that language learners must use the ‘power of their imagination’ as they invest in the target language, develop a social identity through ‘imagined communities’, and visualize possible selves. ‘These processes evoke positive and negative emotion . . . intimately tied to the ongoing motivation process’ (p. 118). So it starts as a fantasy; we will ourselves to speak in ways that are counterfactual to our status and identity. Ryan & Irie (2014) describe the interaction with our environment as such: We ‘anticipate the actions or reactions of others and guide our own behavior accordingly’ (p. 109). If we anticipate a positive reaction, we are encouraged to continue, and our sense of self is positively reinforced. When we anticipate a negative reaction, we may retreat before establishing a solid foundation in speaking ability.

The phonology research has very few up-close, personal narratives that describe this arduous, often uncomfortable process (e.g., Lybeck 2002; Rindal 2010). In Marx’ (2002) first-person account of her residence in Germany, she describes appropriating a French accent at first (French being the other foreign language she had studied at a younger age) believing that this would yield a more positive reception than English-accented German. (Actually, her greatest fear as a proud Canadian was being mistaken for an American – better to sound French since she could not yet pull off a German accent.) Within a year, she changed course, aiming for a German accent in order to better assimilate into her community (p. 272). By the time she returned to Canada three years later, even her native English was tinged with a German accent; she had successfully constructed a new identity as ‘a foreigner’, in her words (p. 275).

Marx’ narrative highlights a very non-linear path toward her ideal L2 self, a path that shifts in step with her evolving attitudes toward her community, and an increasing sense of security and acceptance. By her own account this was a performance at first; this new identity was constructed in relation to her environment, affirmed through a series of conscious decisions and a habit of self-reflection on what accent meant to her personally, as a signal of affiliation. It speaks to all the elements of volition, self-regulation, motivation, and conscious decision-making to achieve a specific goal. It also shines a light on Marx’ flexibility, her responsiveness and willingness to reassess, revise, restart, reinvent.

There are no objective measures of Marx’ accent to confirm how successful this transformation was, but in a similar portrait of a learner's personal trajectory, Major (1993) describes an adult American learner of Portuguese whose deep affiliations with Brazil motivated her to develop an accent indistinguishable from native speakers. This was confirmed by phonetic analyses and native listener ratings. As with Marx, at the height of her affiliation with Brazil, her mother tongue had traces of a Portuguese accent. She later disavowed her connections to Brazil – decisively – moved back home, and one year later, lost all traces of Portuguese influence in her English accent (even her Portuguese was less native-sounding than it had been before).

If chaos and non-linearity are hallmarks of a dynamic system, then here is evidence of what this looks like up close. Both of these stories speak to the fluidity and complexity of accent, and the key role that internal resources play at a conscious level. Let us also note an important bidirectional relationship: for both learners, L2 actually influenced their L1 phonology at the height of their affiliation with the second language.

4. Limits on learner autonomy: Challenges to accent goals

While championing autonomy, I do not claim that anyone can achieve anything in terms of pronunciation. There are surely a number of developmental phenomena that take time to get past (see Hansen-Edwards 2014), and I acknowledge that age-related factors (of a still-uncertain nature) can work against even the most ardent intentions and efforts. However, I believe the age issue for phonology should be understood as a question of limits on both external and internal resources – the kind I am talking about today. So I turn now to several such limits that represent newer areas in phonology research which deserve greater focus.

4.1 Social limits

LeVelle & Levis (2014) maintain the following:

. . .[G]aining access to real speakers . . . opens up opportunities to notice how people talk, how they interact, the ways in which they package their words and gestures, and the sociolinguistically marked variants that evoke comfort in interactions. We are social beings, and communication and pronunciation live and develop within social contexts. (p. 103)

First, however, we must be willing to communicate; we must be ready to ‘enter into discourse at a particular time with a specific person or persons, using a L2’ (MacIntyre et al. 1998: 547). Yet adult learners are famously self-conscious about having an accent, even if it does not legitimately interfere with actual understanding (see Derwing & Munro 1997).

It is precisely the linking of these constructs to the long-term picture that we should do more of. We know, for example, that self-consciousness can take a toll on attainment. For Gluszek & Dovidio's (2010) 77 learners of English, perceived stigmatization of accent and a poor sense of belonging correlated with self-reported communication difficulties. When non-native speakers (NNS) are expected to carry the ‘communicative burden’ and adhere to accent standards (Lippi-Green 1997), this can exacerbate insecurities and lead them to communicate in circumscribed ways (Subtirelu 2014). Over the long-term, the effect on phonological attainment is surely detrimental (see Hansen 1995). Stigma can lead to emotional and social isolation, as pointed out by LeVelle & Levis (2014: 105), and a weak sense of belonging plays ‘a pivotal role in the social initiatives . . . of nonnative speakers’, according to Gluszek, Newheiser & Dovidio (2011: 37). Their study (as well as Moyer 2007) shows that if I identify strongly with the target community and feel comfortable with assimilation, I will likely have a more native-like accent. On the flipside, if my identity differs from L2-majority speakers (see Gatbonton, Trofimovich & Segalowitz 2011), or the identity of my classmates (Cao 2011), I may have little desire to try to change my noticeably foreign pronunciation. (The problem of native speakers’ willingness to listen and communicate with me as an accented speaker is another matter altogether (see Derwing & Munro 2014; Kang & Rubin 2014)).

As Wenger has described the construction of identity, it depends on engagement, imagination, and alignment with the group around us (2010, cited in Levelle & Levis 2014). I believe this is what exceptional learners model for us. In a recent analysis (Moyer 2014a), I reported that those who end up sounding native-like are not immune to social awkwardness or discomfort, yet they describe themselves as outgoing and motivated to connect socially with native speakers. Perhaps they are extraverts, with (potentially) lower anxiety, superior short-term memory, and greater willingness to communicate (Dewaele & Furnham 1999; van Daele et al. 2006). So this is the ‘social’ challenge for late learners: to prioritize the ‘possible L2 self’, and to align that vision with the surrounding community (see Moyer 2014b).

4.2 Limits in experience

SLA scholars know that simply living in-country is no guarantee of a better accent. Children typically acquire L2 as the language of instruction in school, and easily establish social networks with peers due to frequency of contact and the availability of meaningful interaction across various contexts and functions. But opportunities to use the target language in authentic and meaningful ways, i.e., beyond perfunctory exchanges and formulaic language, may be infrequent for older learners. In a classroom setting, these limits are inherent, but even in-country, limited access to meaningful interactions hinder adults’ ability to form social networks with native speakers which, we know, fosters greater fluency. Lower input frequency can diminish the salience of specific features and thus their complete acquisition (Hansen-Edwards 2014: 57). Overall, the quality of L2 experience appears to be more significant than the quantity (Moyer 2011), so the social network aspect is critical.

Two relevant factors correlate significantly to a more authentic accent and highlight the link between context, and a strategic approach: (a) the extent to which one interacts with native speaker friends, adopting the target language in one's innermost circles (Moyer 2004, 2011); (b) the extent to which one relies on L2 relative to L1 in daily life more generally (Flege, MacKay & Piske 2002; Moyer 2011). Knowing this, we should look at the whole trajectory of attainment, in particular, the ways that advanced fluency moves gradually toward L2 dominance, a potential predictor of accent (see Moyer 2004, 2014a).

4.3 Limits of personal orientation

I mentioned earlier that very advanced students deploy more sophisticated, complex strategies (Macaro 2010), and this is significant for accent outcomes (Moyer 2004, 2007). In addition, older learners utilize more strategies overall than younger ones do (Pawlak 2011) – an age advantage that is metacognitive in nature. And here is something that is perhaps not surprising: those who already sound native-like are not complacent, but have a self-critical bent and continually try new techniques and seek new venues for practice (Moyer 2014a).

Of course, accent-oriented strategic behaviors depend on individual goals and personality. Desire to sound native, risk-taking, extraversion and empathy have all been linked to phonological fluency (Moyer 2007; Hu & Reiterer 2009; Ockey 2011), as has the ability to adapt to challenging circumstances (Moyer 2014a). In fact, these appear to have independent effects on accent, apart from age of onset (e.g., Moyer 2004). It is at the level of the individual that we find out how these relate to the possible self and manifest as specific strategic behaviors (Gregersen & MacIntyre 2014: 118).

4.4 Limits in ability

A construct that is front and center in the individual differences literature is ability, or aptitude in cognitive circles. Defining it has been a challenge. In our field, ability is typically described as: read-aloud or mimicry abilities; the perception of categorical distinctions; or some kind of implicit talent, aka ‘having an ear for languages’ (e.g., Ioup et al. 1994), the mechanisms for which could be working memory, rhythm and pitch perception, or phonetic coding ability – the ability to identify distinct sounds and to retain sound-symbol mappings (famously measured by the Modern Languages Admissions Test (MLAT), but not applied to phonological learning, per se – see Moyer 2013, for discussion). Empathy could be another aspect of phonological ability, recently confirmed by Rota & Reiterer (2009), measured as sensitivity and concern over real-life vs. fictitious situations) and correlated with suprasegmental perception, phonetic coding ability, and the enjoyment of mimicry (see also Ibrahim, Eviatar & Leikin 2008).

A serious caveat to this fascination with aptitude is that, first of all, decontextualized test batteries cannot account for real mastery, much less real-life language use. Viewed in such an isolated way, the construct leaves no room for learner choice and self-determination, nor how the person and the environment interact. Secondly, aptitude is ostensibly about a priori ‘talents’ or skills; however, it is not strictly innate; new phonetic and suprasegmental categories can be learned (see Saito 2011), improvement is arguably within anyone's reach, and ‘talent’ may be less influential than one's overall orientation (see Moyer 2013, 2014a). Let us therefore prioritize what a learner actually can do, i.e., how she marshals specific resources targeted to accent.

Looking closely at exceptional learners, they seem to have several things in common, including: strong identification with the target language; comfort with assimilation; pride in their L2 abilities; and consistent use of L2 across multiple domains, especially on the home front (Moyer 2007, 2014a). This seems to hold true regardless of age of onset. According to my 2007 study, those with an early start and years-long residence in-country do not reach the same level of phonological fluency if they are missing that deep investment (Moyer 2007). Equally striking is the outward direction of their approach; they take risks to engage with, to be a part of, their community. Perhaps a healthy dose of assertiveness is necessary to achieve L2 oral fluency, as Ockey has indicated (2011). I would add that for a motor-based skill like accent, a strong metacognitive orientation is required. Here is what exceptional learners do, based on my recent analysis (Moyer 2014a; see also Biedron & Szczepaniak 2009):

• • seek and incorporate feedback directed at accuracy of form

• • practice problem sounds aloud and enjoy mimicry

• • strategize ways to improve oral fluency, and accent specifically

• • accept difficult experiences as part of the process

Their habits embody learner autonomy on multiple fronts. Along with a metacognitive approach, they stay flexible and versatile, and ‘check in’ regularly on whether their behaviors adequately match their goals. This supports Dörnyei's description of self-regulation as multi-dimensional; it activates cognitive, metacognitive, motivational, and behavioral resources to order to enhance achievement (2005: 191).

5. Conclusions

Why prioritize learner autonomy now? There are several things we know about very advanced learners that compel us to consider autonomy in order to move the research and the pedagogy forward:

• • Those who achieve a native-like accent are keenly aware of the relevance of accent, both for their own lives and in society more broadly (Moyer 2007).

• • They go to great lengths and social risks, even avoiding other L1 speakers to maximize L2 use, in order to stick to their goals (see Muñoz & Singleton 2007).

• • They re-double their efforts after setbacks and failures, continually redrawing their ‘roadmap’ in order to achieve their ‘ideal L2 self’ (Moyer 2004, 2014a).

Their remarkable attainment is supported by the belief that they are capable of improving their accent (Moyer 2007), whereas the opposite seems to obtain for those considered ‘poor’ or ‘very poor’ in pronunciation; they tend to blame their difficulties on factors beyond their control (Baran-Łucarz 2012.) Self-efficacy, the belief that one's accomplishments are due to one's own efforts and resources, may be critical to moving beyond the initial enthusiasm of beginning a new language, when every step forward feels like leaps and bounds (Ushioda 2014). With time, motivation will need to be consciously directed to take on ‘increasingly complex tasks and challenges’ (p. 136; see also Mercer 2014).

Throughout their new volume on ‘the self’ in SLA, Mercer & Williams emphasize that self-worth, identity, motivation and self-regulation are central to successful language learning in both formal and informal settings (2014: 178). Based on available evidence, it seems that the willingness and ability to establish a new self-concept in the target language confers significant advantages for pronunciation attainment. From that foundation, learners can consciously construct ways to assess their progress and make adjustments in approach, as needed.

5.1 Pedagogical applications

How can teachers integrate the principle of learner autonomy for accent? Even in the FL classroom, learners are still essentially in charge of their own learning, and we can foster their sense of competence, autonomy and self-efficacy (Mills 2014) in several ways. First, deepen awareness of accent through the following:

• • Introduce various tools and techniques to accommodate different learning styles and modes, e.g., include visual feedback along with auditory stimuli and feedback (see Chun 2007 for review).

• • Expose students to a range of different ‘voices’ in the target language to enhance their perception of new categories and sound patterns.

• • Discuss regional, generational and gendered variation and its social consequences, even overtly discussing bias (see Lindemann, Litzenberg & Subtirelu 2014).

• • Describe how accent is used situationally to express emotion and to control the flow of interaction. This raises awareness of pronunciation as communicatively meaningful. Role-play and read-aloud could be very helpful and enjoyable here as well.

Second, teach self-regulation and boost self-efficacy in these ways:

• • Ask students to cite the types of practice they most enjoy and find helpful, and incorporate these into activities and assessments.

• • Get learners invested in their own progress: ask them to specify goals for pronunciation based on how they will use the target language in the short- and long-term. How will accent matter to them?

• • Incorporate self-reflection and self-assessment. These are especially effective learning tools since they require a deeper level of reflection and noticing.

One could adopt something akin to the European Language Portfolio, ² which asks students to describe their linguistic identity and language learning experiences, to write reflections and assess their progress periodically perhaps noting intercultural experiences that enrich their language learning (Little 2007). Any such activities could help stimulate interest and motivation in pronunciation learning. (Gregersen & MacIntyre 2014 outline motivation-oriented activities around interactive role-play, imagination of self-concept, and self-reflection which could be adapted for pronunciation.)

As teachers, let us also engage in a bit of self-reflection to support our learners’ autonomy (Kang & Rubin 2014). What biases do we have about accent, even among our own students? Are we open to thinking about why a ‘native speaker ideal’ might not be an appropriate goal for everyone, or do we ‘implicitly support a particular [standard] model as prestigious’? (Levis & Moyer 2014: 276). Finding out about our students’ target language goals will help us to set realistic and fair expectations.

5.2 Research implications

How can researchers best incorporate the idea of learner autonomy? The phonology research today is rich with many questions – about age effects, about the efficacy of instruction, about comprehensibility and intelligibility, and so on. To a large extent these questions revolve around a basic one: What constitutes ‘success’, and by extension, what hinders or fosters, it?

In my own work, I have focused on what Macaro calls ‘achievement against the odds’ (2010: 277), which in this context is native-likeness in ultimate attainment. Because I consistently come across learners who defy the odds and stump the judges, so to speak, I think it behooves us to consider success from their perspective, finding out about their goals, their intentions, and the ways that accent is intertwined with their self-concept. How do they manage to do what they do? Learners themselves can best describe their process toward greater fluency (Levis & Moyer 2014: 276).

This requires complex methodological approaches, going beyond self-report and survey, for example, into grounded theory analyses for situated studies of self working from the data up (e.g., Aveni 2005; Hemmi 2014), first-person narratives such as the Marx (2002) study, and even brain-imaging to track the less conscious aspects of self or the ways that memory and emotion are enhanced when concepts and events are connected to one's sense of self (see Northoff 2014). Most of all, abilities should be measured at different points in time in order to demonstrate a consistent link between what a learner intends, what she does to support those intentions, and what she achieves (Macaro 2010: 280).

Dewaele reminds us: ‘The learner is not only an object of scientific curiosity, but also a crucial witness of his or her own learning process’ (2005: 369). So looking forward, let us ask What does accent mean? – not just to external evaluators like ourselves, but to the language learner. With this in mind, I offer a few specific questions:

• • At what stage do learners become conscious of specific aspects of their own L2 accent, and how does this affect how they actually ‘hear’ and use targeted feedback? Is this different in an instructed, as opposed to an immersion-type, setting? More experimental and self-report data could help us answer these questions.

• • How do learners draw (and redraw) ‘roadmaps’ corresponding to their ‘possible selves’, or, alternatively, to resist that sort of transformation? What moves them to incorporate non-linguistic features like dress and body language? More longitudinal case studies are needed to observe how accent is constructed in relation to an emerging sense of self in context.

• • How does learning context influence the motivational control dimensions mentioned, and the specific behaviors that correspond to them? For example, do classrooms evoke more metacognitive and commitment control strategies while immersion learning relies more on emotional and environmental control? More self-reflective, introspective data are needed to answer such questions.

The bottom line is that individual differences should not be viewed as staid, decontextualized personality or a priori cognitive traits. Individual differences need to be afforded the dynamic quality that is their true nature. The field of L2 phonology is uniquely poised to do this kind of work, to show how ‘open and adaptive’ the underlying system driving attainment really is. I think this can be done without becoming overly subjective or relativist. We will need to operationalize some of these constructs (learner autonomy, self-concept, etc.) for our field while also finding ways to draw the links between affective and cognitive mechanisms as mutually responsive to the environment. In so doing, I think we will find commonalities, even as we continue to discover how socially embedded accent is.

After all, some late learners are willing to take risks, imagine new possibilities for themselves, and engage with others in order to acquire the fluency they desire. Perhaps by incorporating some of the concepts discussed here today we can better understand the ways that ‘mind, body and world work together’ in learning additional languages (Atkinson 2011).