Rethinking spoken fluency

ELIA 9, 2009, pp. 11-29 RETHINKING SPOKEN FLUENCY Michael McCarthy University of Nottingham, UK [email protected] This article re-examines the notion of spoken fluency. Fluent and fluency are terms commonly used in everyday, lay language, and fluency, or lack of it, has social consequences. The article reviews the main approaches to understanding and measuring spoken fluency and suggest that spoken fluency is best understood as an interactive achievement, and offers the metaphor of ‘confluence’ to replace the term fluency. Many measures of spoken fluency are internal and monologue-based, whereas evidence from a variety of research both within and without linguistics and applied linguistics suggest that speakers fine-tune their performances to one another. I report ongoing research into features of spoken language, such as automaticity and turn-boundary phenomena, which illuminate the interactive processes involved in fluent production. The applications of such research will be directed towards the empirical underpinning of the Common European Framework (CEFR) level descriptors for spoken language through the English Profile project. Key words: fluency, interaction, corpora, spoken language Este artículo revisa el concepto de la fluidez oral. ‘Fluido’ y ‘fluidez’ son términos usados comúnmente en el lenguaje profano y la ‘fluidez’, o la falta de ella, tiene consecuencias sociales para los hablantes. Este artículo revisa los principales enfoques para la comprensión y la medición de la fluidez oral y sugiere que la fluidez oral debe entenderse como un logro de la interacción y no sólo de la expresión individual. También se presenta la metáfora de la ‘confluencia’ para sustituir el término ‘fluidez’. Muchas de Estudios de lingüística inglesa aplicada 12 Michael McCarthy ELIA 9 2009, pp. 11-29 las mediciones de la fluidez oral que se realizan a través de investigaciones están basadas en la producción de un monólogo, mientras que los resultados de las investigaciones llevadas a cabo en el ámbito de la lingüística, la lingüística aplicada, y también en otras disciplinas, muestran que los hablantes sintonizan su discurso oral con el de su interlocutor. En este artículo se presenta una investigación en curso sobre las características propias del lenguaje oral, como son la automaticidad y los fenómenos que aparecen en los intercambios de turnos de habla, que determina los procesos interactivos involucrados en la producción oral fluida. Las aplicaciones de esta investigación se dirigirán a apoyar empíricamente los descriptores del Marco Europeo de Referencia para las Lenguas con el proyecto denominado English Profile. Palabras clave: fluidez, interacción, corpus, lenguaje oral 1. Introduction The terms fluent(ly) and fluency are examples of terminology whose purview goes beyond the specialist language of linguists and applied linguists and which is firmly rooted in everyday lay usage. Most people have at least a rough understanding of what it means to say that someone is ‘fluent in Spanish’ or ‘speaks Japanese fluently’, or to say that one is able to speak a language ‘but my fluency is not what it used to be’. And indeed, in more than just the Romance languages, the equivalent word for fluent is typically based on a metaphor of ‘flow’ or ‘smooth delivery’ (e.g. Norwegian flytende). Few other pieces of linguistic terminology are so effortlessly used in ordinary, everyday language or are so deeply rooted in a common metaphor. We might also note that pieces of music are often said to ‘flow’ (or not). In this paper, I would like to unpack a little what spoken fluency is understood to entail and what it might actually entail the more we look at how speakers interact. Rethinking spoken fluency 13 ELIA 9, 2009 pp. 11-29 As might be expected, a huge amount has been published over the years on the nature of fluency, and the debate entered here is certainly not a novel one. Hieke (1985) spared no words in concluding that ‘... the literature on fluency reveals it to be replete with vacuous definitions ...’ (p.135), a little harsh, perhaps, but nonetheless reflecting a frustration that there seemed to be no happy consensus on what the term meant. It would be fair, however, to sum up the past literature in terms of a number of major preoccupations, albeit themes upon which researchers do not always agree, but which enable us broadly to map the landscape within which fluency is discussed: • Speed of delivery, including number of words per speech unit or per minute, location, distribution and length of pauses, etc. (Fillmore, 1979; Dechert, 1980; Towell, 1987; Towell et al, 1996; Lennon, 1990; Riggenbach, 1991; Freed, 1995; Kormos and Dénes, 2004; Wolf, 2008; Mochizuki and Ortega, 2008; Rossiter, 2009). • Automaticity: the ability to retrieve units of speech (routinised and prefabricated words, phrases, whole clauses) quickly and automatically (Fillmore, 1979; Rehbein, 1987; Gatbonton and Segalowitz, 1988; Towell et al, 1996; Chambers, 1998; Wood, 2001, 2006). • Perceptions and assessments of fluency and their implications by professional practitioners such as teachers and examiners (Derwing et al, 2004; Hasselgreen, 2004; Kormos and Dénes, 2004). • Perceptions of fluency and their implications by non-professionals, for example, the public at large, employers, social peers (Tainer, 1988; Dustmann,1994; Chiswick and Miller, 1998; Dávila and Mora, 2000; Shields and Wheatley Price, 2002; Yeh and Inose, 2003). What typically (but not exclusively) unites the first two preoccupations is a conception of fluency as a monologic achievement, often judged under experimental or quasi-experimental conditions: the speaker either performs fluently or does not, and is more, or less, fluent. The second two preoccupations bring in many more social concerns, especially the fourth, 14 Michael McCarthy ELIA 9 2009, pp. 11-29 and fluency is more typically sited and judged in performance with others (other language learners, interlocutors in social settings, etc.). Some studies naturally fall in between the broad dichotomy, with studies of foreign language learners who have spent considerable periods of time living in the target culture usually being judged by their performance before and after immersion in the target social milieu (e.g. studies by Towell, 1987 and Lennon, 1990). Equally, in Wolf’s (2008) study, the researcher acted as an ‘interlocutor’ of sorts with the informants, while in Lumley and O’Sullivan’s (2005) study, informants interacted with taped stimulus voices to test task effects such as gender of interlocutor. 2. Monologue and Dialogue 2.1. Speed of Delivery Fluency as calculated by temporal measures such as speed of delivery, pausing and so on undoubtedly has some foundation in real perceptions of informants. Freed (1995), in comparing two groups of language students, one stay-at-home group and one which studied abroad in the target culture, used average number of words per second as an instrument of comparison, with the (more fluent) study abroad students managing more words per second. In Kormos and Denés’ (2004) study, various aspects of speech rate (apart from pausing) were said to be ‘the best predictors of fluency scores’ (p.145) (see also Rossiter, 2009), while Foster and Skehan (1999), looking at the reverse of the coin, see longer silences within and between speaker turns in task performances as ‘moments when performance is seriously disrupted and the subject has to engage in regrouping and unexpected on-line planning’ (p.229). In respect of speed of delivery, however, real speech rates vary greatly depending on context and speech genre (Tauroza and Allison, 1990), and so speed of delivery on its own and taken out of context is clearly a Rethinking spoken fluency 15 ELIA 9, 2009 pp. 11-29 blunt instrument for assessing fluency. By the same token, pausing is a complex phenomenon which is sensitive to the cognitive complexity of production and the contextual conditions of speech (Goldman-Eisler, 1972). Pauses may not necessarily be a sign of communicative failure but may indicate complex planning and increased cognitive effort. Moreover, intriguing evidence from divergent sources in communication research, psycholinguistics, sociolinguistics and computer-human-interface science suggests that humans fine-tune their rate of speaking and pausing in accordance with that of their interlocutor(s) (Street et al, 1983; Street and Capella, 1989; Giles et al, 1991; Bosshardt et al, 1997; Street, 2006; Kousidis and Dorran, 2009), a subliminal and delicate harmonising of tempo being achieved in successful interaction, with speech and silence between interlocutors performing a rhythmic dance. Such harmony may be a convincing candidate for an important property of the metaphor of ‘flow’. In sum, with this conception, speed of delivery becomes an interactive, jointly produced phenomenon of discourse, and pauses and hesitations can be investigated with a greater degree of sensitivity to their contexts of occurrence. 2.2. Automaticity It is uncontroversial to suggest that automatic retrieval of language forms plays a part in smooth performance and delivery of spoken language. Dörnyei (2009) sees a distinction between speed of delivery and automaticity, in that ‘fast processing is not necessarily automatic’ (p.287). It is inconceivable that fluent speakers create anew on each occasion of production the patterns that are so common in speech as evidenced in spoken corpora. Once again, context influences the degree of automaticity that is demanded on any given occasion of production (Bialystok, 1982), with dyadic or multi-party conversation placing the highest premium on automaticity because of the need to compete for turns and the imperative to react quickly to one’s opportunities for speaking. The overwhelming 16 Michael McCarthy ELIA 9 2009, pp. 11-29 evidence of brief pause-lengths between speaker turns in everyday conversation (typically less than one second) suggests that turn-boundaries are a locus of high demand on automaticity. The work of Stivers et al (2009) shows that across a wide range of languages, pauses between turns are consistently very short (for English around a quarter of a second), suggesting a general principle that humans seek the shortest delay between turns while avoiding too many overlaps. As with other aspects of convergence, there is also evidence that speakers attempt to converge on pause-duration between speaker turns (Kousidis and Dorran, 2009). We return to issues of turntaking below. A major feature of automaticity is the need to retrieve ready-made chunks of language (most typically two to four words in length; see O’Keeffe et al, 2007, for corpus-derived lists of these). The ubiquity in native-speaker and expert-user conversation of such chunks is testimony to their regular, fixed occurrence and to their common, everyday pragmatic functions. Dörnyei (2009: 294-297) gives a useful summary of the notion of chunking from a psycholinguistic viewpoint and mentions the importance of chunking in both production and perception. That such chunks are directly related to automaticity has long been recognised (e.g. Gatbonton and Segalowitz, 1988). The internal syntactic structures of chunks such as on the other hand, you know what I mean, if you like, by and large, that sort of thing, etc. vary considerably but what they have in common is the need to be articulated quickly, automatically, consistently and pre-assembled as intonation units, without internal interruption, thus producing runs of smooth speech that require less frequent pausing (Wood, 2006). Provided chunks are uttered smoothly and without disfluency, the speaker can afford the comparative comfort of taking their time with the more transactional, nonpre-assembled aspects of the utterance. Chunks facilitate reduced processing time (Conklin and Schmitt, 2008) and are therefore more communicatively efficient both for producer and receiver, suggesting a useful underpinning of ‘flow’ and an interactive basis for their use. Rethinking spoken fluency 17 ELIA 9, 2009 pp. 11-29 2.3. Turn-taking and Fluency One aspect of automaticity mentioned above is the ability to respond without long pausing when it is one’s turn to speak. The smooth progression of turntaking in multi-party conversation, with little overlap or interruption, has long been recognised as a basic feature of talk (Sacks et al, 1974), and has recently been reinforced by Stivers et al’s (2009) study, as mentioned above. Two principle aspects of turn-construction would seem to be implicated in the creation and maintenance of flow: turn-opening and turn-closing, each point being the mirror-image of the other, and both points being potential sites for smooth or disfluent transition. What happens at turn-boundaries may tell us a good deal about how flow is constructed interactively, as opposed to the degree of monologic flow achieved by a single speaker. Tao’s (2003) study of turn-openers showed convincingly that items which opened speakers’ turns attended to what the previous speaker had just said, in other words, turn-openers provided links and continuity with what had just occurred and function to create smooth transitions and flow. Evison and McCarthy (forthcoming) give as the 10 most frequent turn-openers in a one-million-word sub-corpus of social conversations which form part of the 5-million word CANCODEi corpus the following: Yeah, Mm, Oh, And, I, No, [laughter], Well, Yes, But. These items show an overwhelming preference for linkage with the preceding utterance, thus avoiding the impression of a butterfly-like hopping from one turn to another. The dominant imperative seems to be to construct one’s turn so that it links with and flows smoothly from the previous speaker’s turn. Such a demand is not present in set-piece monologue performances. The moment we consider real conversation among two or more participants, the issue of flow breaks out of the confines of the single speaker or the single speaking turn and becomes the joint responsibility of all participants. This responsibility axiomatically includes a joint responsibility to fill silences and to avoid uncomfortably long pauses that conversation tries to avoid. For this reason, flow across 18 Michael McCarthy ELIA 9 2009, pp. 11-29 turns may be better captured by the term confluence, reflecting the jointlyproduced phenomenon which constitutes a successful interaction. Turn-closings are complementary to turn-openings. The turn typically transfers to another speaker at predictable points such as the completion of syntactic or intonation units, so-called transitional relevance points (TRPs) (Sacks et al, 1974) and, as we have discussed in relation to the duration of pauses, above, transition is typically quick and automatic. Evison and McCarthy (forthcoming) further demonstrate that there is a strong tendency for the occurrence of certain lexical items and longer chunks to trigger speaker-change. These include vague language markers such as or something, and stuff (like that), which invite the listener to fill in unstated or fuzzy members of categories from shared knowledge (Evison, McCarthy and O’Keeffe, 2007). An immediate convergent response, however brief, confirms the projection of assumed shared knowledge for the speaker. Evison and McCarthy (ibid) show that such responses frequently occur, even if they are only acknowledgements such as I know or Yeah. Likewise, common evaluative adjectives such as lovely, awful, wonderful, funny regularly trigger listener verbal reaction, with a strong preference for convergent reactions. The reaction is not necessarily (and is very often not at all) floor-grabbing; it can be minimal (e.g. backchannel noises such as Mm), or non-minimal (e.g. Really?) (McCarthy, 2003), while still leaving the floor in the hands of another speaker. The importance of these types of items is that they invite reciprocity and convergence and project seamlessly to the following turn-opener by the next speaker. Turn closing also includes the possibility of turn-completion by a listener, or the joint production of syntactic formations such as main clause (speaker 1) followed by subordinate clause (speaker 2). Examples of such fluent joint production may be seen in Tao and McCarthy’s (2001) study of non-restrictive whichclauses in conversation. One might sum up the contribution of turn-taking to the notion of fluency by the simple and seemingly banal observation that all utterances in a conversation, apart from the opening utterance, may be seen as responsive, Rethinking spoken fluency 19 ELIA 9, 2009 pp. 11-29 with the primary motivation of turn-construction being the creation of a responsive unit of speech, quickly and automatically. For the purposes of the present discussion, turn-opening and closing may be viewed as two sides of the coin of confluence, in other words, the co-creation of fluency in a conversation rather than the fluency of an individual speaker. Judgements of fluency which lack such an interactive dimension may therefore be considered as providing only a partial picture of the speech event, lacking as they do the attention which is merited by the efforts of conversational participants to create confluence on so many different levels. 3. Assessing Fluency in the Pedagogic Context Assessing fluency has long preoccupied language practitioners and many language-proficiency measures and scales of achievement explicitly acknowledge fluency as a component of proficiency measures. The Common European Framework of Reference (CEFR) refers to fluency as a descriptive element at the higher levels. In the description of the B2 level, for example, the successful B2 language learner should be able to ‘interact with a degree of fluency and spontaneity that makes regular interaction with native speakers quite possible without strain for either party’ (Council of Europe, 2001: 24). Not only does this description link fluency with spontaneity, echoing the importance of quick and automatic production, the implications for interaction with another speaker are at least acknowledged, even if only vaguely stated. The more specific description of spoken language in the CEFR describes the C2 user as being able to speak ‘so smoothly that the interlocutor is hardly aware of it.’ (ibid., p.28), highlighting the importance of smooth performance and, once again, acknowledging the interactive dimension of fluency. Even at a lower level of achievement (B2), the speaker should be able to produce language ‘with a fairly even tempo’ and ‘few noticeably long pauses’ (ibid.). Tempo and pausing, as we have argued, 26 Michael McCarthy ELIA 9 2009, pp. 11-29 Developments in Applied Sociolinguistics. Cambridge: Cambridge University Press. Goldman-Eisler, F. (1972). Pauses, clauses, sentences. Language and Speech, 15: 103-113. Guillot, M.-N. (1999). Fluency and its Teaching. Clevedon: Multilingual Matters. Hasselgreen, A. (2004). Testing the spoken English of young Norwegians: A study of test validity and the role of ‘smallwords’ in contributing to pupils’ fluency. Cambridge: Cambridge University Press. Hieke, A. E. (1985). A componential approach to oral fluency evaluation. The Modern Language Journal, 69 (2): 135-142. Kormos, J. & Dénes, M. (2004). Exploring measures and perceptions of fluency in the speech of second language learners. System, 32 (2): 145-164. Kousidis, S., & Dorran, D. (2009). Monitoring convergence of temporal features in spontaneous dialogue speech. Dublin Institute of Technology: Digital media Centre Conference Papers. Accessed November 2009 at: http://arrow.dit.ie/cgi/viewcontent.cgi?article=1003&context=dmccon . Lennon, P. (1990). Investigating fluency in EFL: a quantitative approach. Language Learning, 40 (3): 387-417. Lumley, T. & O’Sullivan, B. (2005). The effect of test-taker gender, audience and topic on task performance in tape-mediated assessment of speaking. Language Testing , 22 (4): 415–437. McCarthy, M. J. (1998). Spoken Language and Applied Linguistics. Cambridge: Cambridge University Press. Rethinking spoken fluency 27 ELIA 9, 2009 pp. 11-29 McCarthy, M. J. (2003). Talking back: ‘small’ interactional response tokens in everyday conversation. Research on Language in Social Interaction, 36 (1): 33-63. Mochizuki, N. & Ortega, L. (2008). Balancing communication and grammar in beginning-level foreign language classrooms: A study of guided planning and relativization. Language Teaching Research, 12 (1):11– 37. O’Sullivan, B. (2002). Learner acquaintanceship and oral proficiency test pair-task performance. Language Testing, 19 (3): 277–295. O’Sullivan, B., Weir, C., & Saville, N. (2002). Using observation checklists to validate speaking-test tasks. Language Testing, 19 (1) 33–56. Prodromou, L. (2008). English as a Lingua Franca. London: Contimuum. Rehbein, J. (1987). On fluency in second language speech. In: H. W. Dechert & M. Raupach (Eds.), Psycholinguistic Models of Production (pp.97-105). Norwood, NJ: Ablex. Riggenbach, H. (1991). Towards an understanding of fluency: A microanalysis of nonnative speaker conversations. Discourse Processes, 14: 423–441. Rossiter, M. J. (2009). Perceptions of L2 fluency by native and non-native speakers of English. Canadian Modern Language Review, 65 (3): 395-412. Sacks, H., Schegloff, E. A. & Jefferson, G. (1974). A simplest systematics for the organisation of turn-taking for conversation. Language, 50 (4): 696-735. Shields, M & Wheatley Price, S. (2002). The English language fluency and occupational success of ethnic minority immigrant men living in English metropolitan areas. Journal of Population Economics, 15 (1): 137-160. 28 Michael McCarthy ELIA 9 2009, pp. 11-29 Stivers, T., Enfield, N. J., Brown, P., Englert, C., Hayashi, M., Heinemann, T., Hoymann, G., Rossano, F., de Ruiter, J. P., Yoon, K.-E., & Levinson, S. C. (2009). Universals and cultural variation in turntaking in conversation. PNAS (Proceedings of the National Academy of Sciences), 106 (26): 10587-10592. Street, R. (2006). Speech convergence and speech evaluation in fact-finding interviews. Human Communication Research, 11(2): 139 – 169. Street, R., Brady, R. M., & Putman, W.B. (1983). The influence of speech rate stereotypes and rate similarity on listeners' evaluations of speakers. Journal of Language and Social Psychology , 2 (1): 37-56. Street, R., & Capella, J. (1989). Social and linguistic factors influencing adaptation in children's speech. Journal of Psycholinguistic Research, 18 (5): 497-519 Tainer, E. (1988). English language proficiency and the determination of earnings among foreign-born men. Journal of Human Resources, 23 (1):108-122. Tao, H. (2003). Turn initiators in spoken English: A corpus-based approach to interaction and grammar. In P. Leistyna & C. F. Meyer (Eds.). Corpus Analysis: Language Structure and Language Use (pp.187- 207). Amsterdam: Rodopi. Tao, H. & McCarthy, M. J. (2001). Understanding non-restrictive whichclauses in spoken English, which is not an easy thing. Language Sciences, 23: 651-677. Tauroza, S & Allison, D. (1990). Speech rates in British English. Applied Linguistics, 11(1): 90-105. Towell, R. (1987). Variability and progress in the language development of advanced learners of a foreign language. In R. Ellis (Ed.) Second Language Acquisition in Context (pp.113-127). Toronto: Prentice- Hall. Rethinking spoken fluency 29 ELIA 9, 2009 pp. 11-29 Towell, R., Hawkins, R., & Bazergui, N. (1996). The development of fluency in advanced learners of French. Applied Linguistics, 17 (1): 84-119. Wood, D. (2001). In search of fluency: what is it and how can we teach it? Canadian Modern Language Review, 57 (4): 573-589. Wood, D. (2006). Uses and functions of formulaic sequences in second language speech: An exploration of the foundations of fluency. Canadian Modern Language Review, 63 (1): 13–33. Wolf, J. P. (2008). The effects of backchannels on fluency in L2 oral task production. System, 36 (2): 279-294. Yeh, C. J. & Inose, M. (2003). International students’ reported English fluency, social support satisfaction, and social connectedness as predictors of acculturative stress. Counselling Psychology Quarterly, 16 (1): 15-28. First version received: October2009. Final version accepted: December 2009. i CANCODE (Cambridge and Nottingham Corpus of Discourse in English) was a collaborative project between the University of Nottingham, UK and Cambridge University Press. Cambridge University Press is the sole copyright holder. Details of the corpus and its construction may be found in McCarthy (1998).