Introduction

As a lay term, fluency is often used to denote general second language (L2) proficiency. In this context, the term typically implies that an L2 user has advanced facility with the grammar, vocabulary, and perhaps even the pronunciation of a second language (Segalowitz 2010). The term fluency might also be used to indicate that a person can comprehend the L2 with ease or that the person has advanced skills in L2 reading and writing. Notably, this lay use of the term necessarily excludes its application to learners who are beginners and even to those with an intermediate knowledge of an L2.

In contrast, applied linguists and language teachers typically use the term fluency to refer to the fluidity or ease with which the second language is spoken (Derwing et al. 2004; Freed 2000; Isaacs and Thomson 2013; Koponen and Riggenbach 2000). Consequently, some lower proficiency L2 learners may be described as fluent, despite the fact that they have only rudimentary grammatical ability, limited vocabulary knowledge, and poor pronunciation. In this context, describing lower proficiency learners as fluent is understood to mean that the language knowledge they do have is easily accessed and that their oral language is produced without undue hesitation (Segalowitz 2010). At the extreme, pidgin languages provide an example of second language varieties that develop into highly fluent systems of communication, despite comprising substantially reduced morphological, grammatical, lexical, and phonological forms (Holm 2000).

Anyone in the field of second language instruction will have encountered learners who produce fluent but structurally simplified L2 speech at the same proficiency level as learners who, despite having similar declarative knowledge, are disfluent. These individual differences in oral performance across speakers with similar knowledge of a language are widely assumed to emerge from a trade-off between accuracy and fluency, whereby L2 learners’ attention to form can adversely affect their fluency (Skehan 1998, 2009; VanPatten 1990). Competition between accuracy and fluency arises when learners access declarative knowledge during online processing, since doing so requires reliance on short-term memory, which is assumed to have a limited capacity (Baddeley 2007). When a speaker’s conscious attention is directed toward one part of the speech production process, such as pronunciation, less attention is available for other processes, such as lexical access, grammatical encoding, etc. This can lead to a breakdown in the speech production system, manifesting itself as a disfluent utterance.

In this chapter, I endeavor to disentangle fluency as a cognitive skill from other constructs of speech production commonly found in the pronunciation literature, and review the few studies that hint at a relationship between fluency and three of these constructs – accentedness, intelligibility, and comprehensibility. I then describe some theoretical models that help illuminate the role of pronunciation in the development of oral fluency. Finally, I discuss some implications for pronunciation instruction.

Defining fluency

In the pronunciation literature, fluency is often considered in combination with other measures of spoken language – especially comprehensibility and accentedness (e.g., Derwing and Munro 1997; Derwing, Munro, and Thomson 2008; Derwing, Thomson, and Munro 2006; Isaacs and Thomson 2013). However, although they are often discussed together, and despite superficial similarities in how they are measured, fluency is quite different from these other constructs.

In accentedness, intelligibility, and comprehensibility research, the interaction between L2 speakers’ production and listeners’ perception forms the locus of attention. For example, accentedness is operationalized using impressionistic judgments of how far L2 speakers’ pronunciation diverges from a native speaker target; intelligibility is operationalized in terms of how accurately listeners are able to identify spoken language relative to an L2 speaker’s intended utterance; and comprehensibility is operationalized as how easy speech is for a listener to understand, referring to how much effort is involved (see Munro and Derwing 1995a; Munro, Derwing, and Morton 2006).

In contrast, investigations of oral fluency typically focus on the state of learners’ L2 speech production systems. Thus, although measures of oral fluency often involve listener judgments, those judgments are typically understood to reflect the underlying cognitive processes involved in planning and producing spoken language, and the degree to which those processes are automatic or controlled. Listeners’ perceptions of L2 fluency in terms of its impact on communication are normally of little interest in this line of research.

Although listener judgments are taken to be indicators of fluency, other features in learner speech undoubtedly influence these judgments (e.g., word choice, grammar, pronunciation). Recognizing this problem, Derwing et al. (2004) blend definitions from Schmidt (1992) and Guillot (1999) to describe fluency as comprising “an automatic procedural skill on the part of the speaker and a perceptual phenomenon in the listener” (2004: 656).

Given that listener judgments of cognitive fluency can be influenced by other factors, some researchers have attempted to operationalize fluency using more objectively quantifiable correlates of fluency, which are extracted from the speech signal itself (see Derwing et al. 2004, 2009; Kormos 2006; Kormos and Dénes 2004; Towell, Hawkins, and Bazergui 1996). Among others, these measures often include:

• Speech rate: the average number of syllables spoken per second or minute.
• Phonation time ratio: the percentage of time devoted to speaking relative to the total time taken to produce an utterance.
• Pruned syllables: the average number of syllables spoken per second or minute after any disfluencies have been removed (e.g., syllables attributed to self-repetitions, self-corrections, etc., are not counted).
• Articulation rate: the average number of fluent syllables per second or minute between pauses of a predetermined length (e.g., 400 ms). Like pruned syllables, this measure excludes disfluencies from the total syllable count. Unlike pruned syllables, when calculating the duration of the utterance, any time elapsed during the production of disfluencies (e.g., self-repetitions) and pauses is excluded from the total time.
• Mean length of run: the average number of words or syllables produced between pauses of a length specified by the researcher(s).
• Silent pause ratio: the number and/or time attributed to silent pauses of a particular length per second or minute. The minimum duration for what constitutes a silent pause varies across studies.
• Filled pause ratio: the number and/or duration of pausing attributed to filled pauses (e.g., “um” and “uh”) per second or minute.

In an attempt to overcome some of the limitations associated with using either listener judgments or temporal measurements on their own, Derwing et al. (2009) employed both methods of assessment in a longitudinal study aimed at examining the link between L1 and L2 fluency. Interestingly, while some of the listener judgments closely paralleled the related temporal measures, on other occasions this was not the case. Overall, temporal measures were found to be more sensitive in detecting a relationship between L1 and L2 fluency than were listener judgments of the same speech samples. The authors attributed this difference to the fact that the judges who assessed the L1 samples were not the same judges who assessed the L2 samples. However, one might just as reasonably conclude that temporal measures are simply more accurate than listener judgments, which are influenced by other unrelated factors.

As Segalowitz (2010) points out, notwithstanding the additional insight temporal measures offer relative to listener judgments, operationalizing fluency in these terms is not entirely satisfactory. For example, researchers make subjective judgments in concluding that every self-repetition or self-correction in a given utterance is a sign of cognitive disfluency (Hieke 1981; MacGregor, Corley, and Donaldson 2009). In fact, speakers sometimes use self-repetitions and corrections as a discourse strategy, aimed at clarifying or emphasizing given information for the listener’s benefit.

Similarly, deciding what length of silent pause constitutes a disfluency, and in what context, is also subjective (see Davies 2003). Like self-repetition, pausing can be used as a discourse strategy, especially at clause and sentence boundaries. As a result, silent pauses do not provide fail-safe evidence that there has been a breakdown in cognitive fluency. Conversely, lexical filled pauses (e.g., “like”, “I think”, etc.), which often are a sign of disfluency, are never considered when the filled pause ratio is calculated. This omission is problematic, since like their nonlexical counterparts, lexical filled pauses are sometimes produced in order to buy time for planning and producing utterances that follow.

Defining fluency in terms of temporal phenomena can also be limiting because it reduces the construct to speed of production, which is an oversimplification of the complex cognitive underpinnings of fluency. Taking a cognitive perspective, Schmidt (1992) and Segalowitz (2010) argue that it is the efficiency and automaticity of processing, rather than speech rate, that marks fluency. This means that while, on average, automatic processing is likely to be faster than conscious processing, speech rate cannot serve as a reliable indicator of fluency. Instead, fluency should be viewed as a set of interrelated and overlapping cognitive processes, organized in such a way that they impose the smallest possible demands on working memory, which has limited capacity.

Although a more transdisciplinary approach using experimental techniques borrowed from psychology and neuroscience would unquestionably provide more precision in measuring cognitive fluency (see Segalowitz 2010for a detailed overview), a substantial body of research using less sophisticated approaches has still provided important insights into the development of L2 fluency. It is to this earlier research that we next turn.

Relationships between fluency, accentedness, intelligibility, and comprehensibility

While interrelations between accentedness, intelligibility. and comprehensibility are well attested, their relationship to fluency is less so. To date, no studies appear to have systematically investigated the relationship between oral fluency and these other commonly researched measures of L2 speech production. This is somewhat surprising, since learners’ conscious attention to pronunciation can affect their fluency, which in turn might impact listener perceptions of accent, intelligibility and comprehensibility. Despite the paucity of deliberate research in this area, several studies have examined fluency and these other measures of oral production in tandem. By re-examining several such studies, preliminary evidence emerges that fluency is partially related to these other facets of L2 pronunciation.

Fluency and accentedness

Indirect evidence for a link between fluency and accentedness can be found by considering other measures to which these constructs are both related. For example, Derwing et al. (2004) found that speech rate and goodness of prosody (referring to suprasegmental features) are correlated with fluency, while other studies report that speech rate and prosody are correlated with accentedness (e.g., Derwing and Munro 1997; Kang 2010). In another study, Trofimovich and Isaacs (2012) report a moderate negative correlation between mean length of run, a measure of fluency, and number of segmental errors, a measure typically associated with accentedness.

Derwing and Rossiter (2003) investigated the fluency, phonological accuracy, and complexity of L2 speech before and after pronunciation training. They found that the phonological accuracy of learners receiving segmental instruction significantly improved, even on an extemporaneous speaking task, while these learners experienced no improvement in terms of fluency and complexity. This was in marked contrast to a group trained on suprasegmentals, for whom fluency and complexity significantly improved, while their phonological accuracy did not. The researchers concluded that the segmental group had consciously attended to phonological form during their speech production and that this had consumed cognitive resources that could otherwise have been used for more fluent and complex speech. In fact, while there may have been no improvement in the segmental groups’ fluency over time, neither was there a significant decline. If conscious attention to phonological form impacted other processes involved in speech production, as the authors argue, we should expect that the speakers’ fluency and complexity scores would decline as a result. The fact that this did not happen hints at the possibility that real gains in fluency and complexity were masked by the deleterious effect of the learners attending to phonological form. This interpretation suggests that as the segmental groups’ newly learned phonological knowledge becomes automatized, their fluency scores might also improve.

Derwing, Thomson, and Munro (2006) examined Mandarin and Slavic speakers’ development of L2 English fluency and accentedness over a period of eight months. Regrettably, the researchers do not report any correlational analysis between the fluency and accentedness ratings they obtained. Nevertheless, we could surmise that if fluency and accentedness are related, improvement in one should be accompanied by improvement in the other. This prediction is only partially borne out in this study. While both fluency and accent ratings improved for the Slavic group, only accent ratings improved for the Mandarin group. However, since the absolute increase in the Mandarin group’s accent ratings was quite small, this could be taken to indicate that greater improvement in pronunciation is necessary before a relationship with fluency becomes detectable.

A colleague and I recently examined the impact of rater expertise and rating scale length on listener judgments of the fluency, accentedness, and comprehensibility of 38 L2 English learners (Isaacs and Thomson 2013). For the purpose of this chapter, I have revisited that data to explicitly examine the relationship between the fluency and accentedness ratings for the 20 raters who used a 9-point scale. A Pearson’s r coefficient reveals a moderate correlation between these two constructs (r = 0.605, p < 0.001). To probe further, I examined the distribution of accentedness ratings for the five most fluent and five least fluent speakers respectively (see Figure 12.1).

Although skewed toward the low end of the scale (i.e., away from a native-like accent), the distributions are otherwise relatively normal. This indicates that it is sometimes possible for a speaker to be perceived as more fluent, but highly accented, or conversely as less fluent despite having a more native like accent.

Fluency and comprehensibility

Indirect evidence for a relationship between fluency and comprehensibility can be found across several studies. For example, Anderson-Hsieh, Johnson, and Koehler (1992) report that accurate prosody is linked to more comprehensible speech. Similarily, Derwing and Munro (1997) report that goodness of prosody and speech rate were correlated with the comprehensibility ratings of a third of the raters in their study. In addition, 15% of their raters indicated that they were consciously aware that several features normally associated with fluency (e.g., pausing, speech rate, etc.) had affected their comprehensibility judgments.

The pronunciation training studies described previously (i.e., Derwing, Munro and Wiebe 1998; Derwing and Rossiter 2003) also point to an indirect relationship between fluency and comprehensibility. While the group trained on suprasegmentals experienced significant improvement in both dimensions on the extemporaneous speech task, the group trained on segmentals did not. This suggests that fluency and comprehensibility are both closely aligned with suprasegmental features of pronunciation.

Derwing, Munro, and Thomson (2008) compared the development of Mandarin and Slavic learners’ English fluency and comprehensibility over a two-year period. Although the relationship between these two constructs was not a focus of their study, they report moderate to strong correlations between fluency and comprehensibility ratings at three separate data collection points (Pearson’s r coefficients were 0.872, 0.791, and 0.791 respectively).

Revisiting data from Isaacs and Thomson’s (2013) study reveals a strong correlation between fluency and comprehensibility ratings (r = 0.826, p < 0.001). However, comprehensibility ratings for the five most fluent and five least fluent speakers (see Figure 12.2) are quite normally distributed. This indicates that it is possible for some L2 speakers to be perceived as more fluent, but less comprehensible, or conversely as less fluent, but highly comprehensible.

Relevant speech production models

Several theoretical models of speech production help illuminate the complex interactions between fluency and pronunciation. Specifically, these models help point to possible underlying cognitive mechanisms and processes that might explain how improvement in either fluency or pronunciation can promote improvement in the other.

Implications

Although many classroom activities are purported to promote L2 fluency (see Rossiter et al. 2010), there is a dearth of research exploring their long-term impact. Related research has, however, revealed factors that affect oral fluency in the short term. For example, many researchers have examined the impact of task type, planning, and rehearsal (e.g., Ellis 2009; Skehan and Foster 2008), while others have investigated the use of repetition (Gatbonton and Segalowitz 2005), time constraints (Nation 1989), and consciousness-raising (Boers et al. 2006). Unfortunately, the goal of such research is typically limited to validating theoretical models of speech production (e.g., De Bot 1992; Levelt 1999), with little attention to the influence of instructional practice on the development of fluency as a complex cognitive system. Thus, while a variety of factors can clearly be shown to impact fluency during a specific classroom task (state fluency), the extent to which these factors affect permanent changes in fluency (trait fluency) remain uncertain. Nevertheless, some general instructional principles can be inferred from what is known about the relationship between fluency and pronunciation, and through appeal to the theoretical models outlined in the previous section.

One general principle is that for pronunciation instruction to promote fluency, it should aim to stimulate transfer of declarative knowledge to production memory. This means that instructional activities should balance attention to phonological form with activities in which the same forms are represented in communicative contexts. For example, some activities could require learners to consciously attend to pronunciation, which would encourage more accurate but less fluent speech (e.g., Saito 2013). Other activities might use corrective feedback after an utterance has been spoken, since this will not interrupt the speech production process (e.g., Saito and Lyster 2012a, 2012b). Since opportunities to balance fluency and accuracy are rare outside of the classroom, where the demands of communication often prevent conscious attention to form (Lee et al. 1997; Schmidt 2001), this type of instruction is particularly important.

Skehan (2009) provides a useful summary of communicative classroom tasks that promote fluency and accuracy. Personal exchange tasks or other tasks with concrete and familiar topics promote both fluency and accuracy, since they place less of a burden on working memory. Similarly, tasks that provide clear structure also promote both fluency and accuracy. In contrast, tasks introducing new content, such as picture descriptions, or tasks requiring manipulation of information, are ill suited for fluency and pronunciation development, since they impose too many competing demands on working memory.

A second guiding principle for effective pronunciation instruction is that it should include activities that facilitate restructuring of the speech production process to make it more efficient. For example, in Levelt’s (1999) model, a breakdown could first occur during phonological encoding of words. One strategy to reduce the potential for such breakdowns is to improve the overall speed of lexical access (see Skehan 2009). If words are accessed efficiently, more working memory capacity is available to devote to pronunciation. In fact, there is strong evidence that pronunciation accuracy is closely related to lexical frequency and familiarity (e.g., Munro and Derwing 2008; Thomson and Isaacs 2009; Walley and Flege 1999). Thus, vocabulary training and reinforcement should play a central role in fluency and pronunciation instruction.

Another place where restructuring could promote fluency is during the phonetic encoding stage of the speech production process. Connections between mental representations and articulatory gestures can be improved through explicit pronunciation instruction. Training ought to incorporate both perceptual and production practice, since it is widely assumed that in normal L2 development speech perception precedes speech production (Flege 2009). At the same time, in keeping with Complexity Theory, practice in producing sounds might offer another route to improving perception, a claim made by Lowie (2010). Reed and Michaud (2005) also appeal to this view in their argument that speaking helps listening, because learners’ own speech becomes input in their developing L2 speech perception.

When deciding on the content of instruction, suprasegmental features should be given priority, since they are likely to impact fluency more than segmental features. When segmental features are taught, those that occur the most frequently in contrast with other sounds are likely to provide the greatest long-term benefit to fluency. The relative contribution of individual sounds to communication is known as their functional load (Brown 1991; Munro and Derwing 2006). Spending time on sounds with a low functional load can cause learners to unnecessarily divert attention toward features of pronunciation that do not have a major impact on the intelligibility of speech. Furthermore, when L1 sounds can be used in place of L2 sounds without a loss of intelligibility (e.g., a trilled /r/ instead of a standard English /r/), for reasons of efficiency there is merit in allowing learners to continue using the L1 sound rather than expecting them to acquire the more native-like form, if fluency is a goal.

As more accurate perceptual representations emerge, pronunciation instruction should provide learners with substantial practice in speech articulation at the level of segments, words, and phrases. This will promote fluency at the articulation stage of Levelt’s (1999) model. As with the previous stage, it is advisable to allow learners to rely on L1 speech sounds, whenever doing so does not adversely affect their intelligibility. Unrealistically attempting to achieve accent-free production can lead learners to overmonitor their speech, causing at least temporary and unnecessary destabilization of an already efficient L1 system.

Conclusion

In this chapter I have related fluency to some common constructs from the pronunciation literature. This relationship can be further understood through reference to cognitive mechanisms that are known to impact fluency and pronunciation. Given the many questions that surround the validity of fluency measures used in existing L2 research, the precise nature of the relationship between fluency and pronunciation remains uncertain. Future research is needed that is more methodical in relating fluency to pronunciation. There is also a clear demand for longitudinal research in this area. This can lead to evidence-based pedagogical interventions, which will encourage both more fluent and more comprehensible L2 speech.

REFERENCES

1. Anderson J.R. 1983. The Architecture of Cognition, Cambridge, MA: Harvard University Press.
2. Anderson, J.R. 1993. Rules of the Mind, NJ: Lawrence Erlbaum Associates, Inc.
3. Anderson-Hsieh, J., Johnson, R., and Koehler, K. 1992. The relationship between native speaker judgments of nonnative pronunciation and deviance in segmentals, prosody, and syllable structure. Language Learning 42(4): 529–555.
4. Baddeley, A. 2007. Working Memory, Thought, and Action, Oxford: Oxford University Press.
5. Boers, F., Eyckmans, J., Kappel, J., Stengers, H., and Demecheleer, M. 2006. Formulaic sequences and perceived oral proficiency: putting a lexical approach to the test. Language Teaching Research 10(3): 245261. doi: 10.1191/1362168806lr195oa.
6. Brown, A. 1991. Functional load and the teaching of pronunciation. In: Teaching English Pronunciation: A Book of Readings, A. Brown (ed.), 221–224, London: Routledge.
7. Davies, A. 2003. The Native Speaker: Myth and Reality, 2nd edition, Clevedon: Multilingual Matters.
8. De Bot, K. 1992. A bilingual production model: Levelt’s “Speaking” model adapted. Applied Linguistics 13: 1–24.
9. Derwing, T.M. and Munro, M.J. 1997. Accent, comprehensibility and intelligibility: evidence from four L1s. Studies in Second Language Acquisition 19: 1–16. doi: 10.1111/0023-8333.49.s1.8.
10. Derwing, T.M., Munro, M.J., and Thomson, R.I. 2008. A longitudinal study of ESL learners’ fluency and comprehensibility development. Applied Linguistics 29: 359–380. doi: 10.1093/applin/amm041.
11. Derwing, T.M., Munro, M.J., and Wiebe, G.E. 1997. Pronunciation instruction for “fossilized” learners: Can it help? Applied Language Learning 8: 217–235.
12. Derwing, T.M., Munro, M.J., and Wiebe, G. 1998. Evidence in favor of a broad framework for pronunciation instruction. Language Learning 48: 393–410. doi: 10.1111/0023-8333.00047.
13. Derwing, T.M. and Rossiter, M.J. 2003. The effects of pronunciation instruction on the accuracy, fluency and complexity of L2 accented speech. Applied Language Learning 13: 1–18.
14. Derwing, T.M., Thomson, R.I., and Munro, M.J. 2006. English pronunciation and fluency development in Mandarin and Slavic speakers. System 34: 183–193. doi:10.1016/j.system.2006.01.005.
15. Derwing, T.M., Rossiter, M.J., Munro, M.J., and Thomson, R.I. 2004. Second language fluency: judgments on different tasks. Language Learning 54: 655–679. doi: 10.1017/S0272263109990015.
16. Derwing, T.M., Munro, M.J., Thomson, R.I., and Rossiter, M.J. (2009). The relationship between L1 fluency and L2 fluency development. Studies in Second Language Acquisition 31: 533–557. doi: 10.1017/S0272263109990015.
17. Ellis, R. 2009. The differential effects of three types of task planningon the fluency, Ccomplexity, and accuracy in L2 oral production. AppliedLinguistics 30: 474–509. doi:10.1093/applin/amp042
18. Flege, J. E. 2009. Give input a chance! In T. Piske and M. Young- Scholten (Eds.) Input Matters in SLA, pp. 175–190, Bristol: Multilingual Matters.
19. Freed, B. 2000. Is fluency, like beauty, in the eyes (and ears) of the beholder? In: Perspectives on Fluency, H. Riggenbach (ed.), 243–265, Ann Arbor: MI: University of Michigan Press.
20. Gatbonton, E., and Segalowitz, N. 2005. Rethinking communicative language teaching: a focus on access to fluency. The Canadian Modern Language Review 61: 325–353. doi:.10.2307/3587026.
21. Guillot, M.-N. 1999. Fluency and Its Teaching, Clevedon: Multilingual Matters.
22. Hieke, A.E. 1981. A content-processing view of hesitation phenomena. Language and Speech 24(2): 147–160.
23. Holm, J. 2000. An Introduction to Pidgins and Creoles, Cambridge: Cambridge University Press.
24. Isaacs, T. and Thomson, R.I. (2013). Rater experience, rating scale length, and judgments of L2 pronunciation: revisiting research conventions. Language Assessment Quarterly 10(2): 135–159.
25. Kang, O. 2010. Relative salience of suprasegmental features on judgments of L2 comprehensibility and accentedness. System 38(2): 301–315. doi: 10.1016/j.system.2010.01.005.
26. Kormos, J. 2006. Speech Production and Second Language Acquisition, Mahwah, MJ: Lawrence Erlbaum Associates.
27. Kormos, J. and Dénes, M. 2004. Exploring measures and perceptions of fluency in the speech of second language learners. System 32(2): 145–164. doi: 10.1016/j.system.2004.01.001.
28. Larsen-Freeman, D. 1997. Chaos/complexity science and second language acquisition. Applied Linguistics 18(2): 141–165.
29. Larsen-Freeman, D. 2012. Complex, dynamic systems: a new transdisciplinary theme for applied linguistics? Language Teaching 46: 202–214 doi: 10.1017/S026144811000061.
30. Larsen-Freeman, D. and Cameron, L. 2008. Complex Systems and Applied Linguistics, Oxford: Oxford University Press.
31. Lee, J.G., Cadierno, T., Glass, W.R., and VanPatten, B. 1997. The effects of lexical and grammatical cues on processing past temporal reference in second language input. Applied Language Learning 8: 1–23.
32. Levelt, W. 1989. Speaking: From Intention to Articulation, Cambridge, MA: MIT Press.
33. Levelt, W. 1999. Producing spoken language: a blueprint of the speaker. In: The Neurocognition of Language, C. Brown and P. Hagoort (eds.), 83–122, Oxford: Oxford University Press.
34. Lowie, W. 2010. The development of early L2 phonology: a dynamic approach. In: Proceedings of the International Symposium on the Acquisition of Second Language Speech, New Sounds, vol. 6, Poznán, Poland. Retrieved November 15, 2013 from http://ifa.amu.edu.pl/newsounds/Proceedings_guidelines.
35. MacGregor, L.J., Corley, M., and Donaldson, D.I. 2009. Not all disfluencies are equal: the effects of disfluent repetitions on language comprehension. Brain and Language 111(1): 36–45.
36. Miller, G. 1956. The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological Review 63: 81–97. doi: 10.1037/h0043158.
37. Munro, M. J., and Derwing, T.M. 1995a. Foreign accent, comprehensibility and intelligibility in the speech of second language learners. Language Learning 45: 73–97. doi: 10.1111/j.1467-1770.1995.tb00963.x.
38. Munro, M.J., and Derwing, T. 1995b. Processing time, accent, and comprehensibility in the perception of native and foreign-accented speech. Language and Speech 38,: 289–306.
39. Munro, M. J. and Derwing, T.M. 2006. The Functional Load principlein ESL pronunciation instruction: an exploratory study. System 34(4): 520–531. doi: 10.1016/j.system.2006.09.004.
40. Munro, M. J., and Derwing, T.M. 2008. Segmental acquisition inadult ESL learners: a longitudinal study of vowel production. Language Learning 58: 479–502. doi: 10.1111/j.1467-9922.2008.00448.x.
41. Munro, M.J., Derwing, T.M., and Morton, S. 2006. The mutual intelligibility of L2 speech. Studies in Second Language Acquisition 28: 111–131. doi: 10.1017/S0272263106060049.
42. Nakuma, C. 1998. A new theoretical account of “fossilization”: implications for L2 attrition research. IRAL 36(3): 247–257.
43. Nation, I.S.P. 1989. Improving speaking fluency. System 17: 377–384. doi: 10.1016/0346-251X(89)90010-9.
44. Reed, M., and Michaud, C. 2005. Sound Concepts: An Integrated Pronunciation Course, New York: McGraw-Hill.
45. Rossiter, M.J., Derwing, T.M., Manimtim, L., and Thomson, R.I. 2010. Fluency: the missing component in the “communicative” classroom. Canadian Modern Language Review 66: 583–606.
46. Saito, K. 2013. Re-examining effects of form-focused instruction onL2 pronunciation development: the role of explicit phonetic information.Studies in Second Language Acquisition 3: 1–29.
47. Saito, K., and Lyster, R. 2012a. Effects of form-focused instructionand corrective feedback on L2 pronunciation development of /r/ by Japaneselearners of English. Language Learning 62(2): 595–633. doi: 10.1111/j.1467-9922.2011.00639.x.
48. Saito, K., and Lyster, R. 2012b. Investigating pedagogical potentialof recasts for L2 vowel acquisition. TESOL Quarterly 46(2): 387–398. doi: 10.1002/tesq.25.
49. Schmidt, R. 1992. Psychological mechanisms underlying second language fluency. Studies in Second Language Acquisition 14: 357–385. doi: 10.1017/S0272263100011189.
50. Schmidt, R. 2001. Attention. In: Cognition and Second Language Instruction, P. Robinson (ed.), 3–32, Cambridge: Cambridge University Press.
51. Segalowitz, N. 2010. Cognitive Basis of Second Language Fluency, New York: Routledge.
52. Selinker, L. 1972. Interlanguage. IRAL 10(3): 209–231.
53. Skehan, P. 1998. A Cognitive Approach to Language Learning, Oxford University Press.
54. Skehan, P. 2009. Modelling second language performance: integrating complexity, accuracy, fluency, and lexis. Applied Linguistics 30: 510–532. doi: 10.1093/applin/amp047.
55. Skehan, P. and Foster, P. 2008. Complexity, accuracy, fluency and lexis in task-based performance: a meta-analysis of the Ealing research. In: Complexity, Accuracy, and Fluency in Second Language Use, Learning, and Teaching, S. Van Daele, A. Housen, F. Kuiken, M. Pierrard, and I. Vedder (eds.), Brussels: University of Brussels Press.
56. Thomson, R.I. 2011. Computer assisted pronunciation training: targeting second language vowel perception improves pronunciation. CALICO Journal 28: 744–765.
57. Thomson, R.I. 2012. Improving L2 listeners’ perception of English vowels: a computer-mediated approach. Language Learning 62: 1231–1258. doi: 10.1111/j.1467-9922.2012.00724.x.
58. Thomson, R.I. and Isaacs, T. 2009. Within-category variation in L2English vowel learning. Canadian Acoustics 37(3): 138–139. Proceedings ofthe Annual Conference of the Canadian Acoustics Association, Niagara-on-the-Lake, ON.
59. Towell, R., Hawkins, R., and Bazergui, N. 1996. The development of fluency in advanced learners of French. Applied Linguistics 17(1): 84–119. doi: 10.1093/applin/17.1.84.
60. Trofimovich, P., and Isaacs, T. (2012). Disentangling accent from comprehensibility. Bilinguialism: Language and Cognition 15: 905–916. doi: 10.1017/S1366728912000168.
61. VanPatten, W. 1990. Attending to form and content in the input: an experiment in consciousness. Studies in Second Language Acquisition 12– 287–301. doi: 10.1017/S0272263100009177.
62. Walley, A.C. and Flege, J.E. 1999. Effect of lexical status on children’s and adults’ perception of native and non-native vowels. Journal of Phonetics 27(3): 307–332. doi: 10.1006/jpho.1999.0098.