Chapter · January 2017 citations reads 8 14,254 1 author

Figure 18.1 Cognitive Demands on Learners During Speech Processing (Based on Levelt’s Model of Speech Processing)

15031-0341-Fullbook.indd 245 8/23/2016 9:27:12 PM

246. 
     • Anne Burns
     

the theorisation of second language speech production (e.g. McLaughlin, 1987; Poulisse, 1993; John-son, 1996; Bygate, 1998). To produce speech, the speaker must engage in processes that express both form, or structure, and meaning, or content. The model presents three components that are directly involved in the production of speech: (i) conceptual preparation, (ii) formulation, and (iii) articula-tion. A fourth component of this model is self-monitoring, which operates at a superordinate level of the other three components (see Figure 18.1). While presented separately in the model, these components are likely to recur and overlap throughout speech production, as interconnected neural networks in the brain are activated at about the same time (Bechtel & Abrahamsen, 1991).

Conceptual preparation is where speech production begins, as speakers select topics or contribute to preceding speech content, using their background or ‘encyclopedic’ world knowledge. In the case of language learners, conceptual preparation relates to how much learners already know and com-prehend, to the linguistic repertoire at their disposal, and to their sociocultural awareness. Formula-tion involves mapping the messages speakers have in mind on to their available lexico-grammatical system and stringing them together (Garman, 1990). In order to do this, speakers must make lexico-grammatical choices relevant to the intended message and flesh them out syntactically, selecting appropriate forms of ‘bounded markers’ (of tense, mood, number, and so on), as well as pay attention to overall discourse coherence and register variance. Language learners may resort to translation or mental representations of how messages are structured and produced from their first language to assist in this process. Articulation brings in physiological skills, which are closely linked to memory and information processing. The articulatory system must be activated to produce streams of sound through control of muscle groups (vocal tract, larynx, and lungs). For fluent speakers, phonological encodings have become automated, enabling them to manipulate and draw attention to meaning through assigning stress and intonation at various points in the speech stream. For learners, espe-cially beginners, articulation can be very challenging and may provoke anxiety as they struggle to reach greater automisation and thus greater comprehensibility. The fourth dimension at work in speech processing, according to Levelt’s model, is self-monitoring. This is a metacognitive process that works at a higher level than the other components, involving checking speech for accuracy and acceptability, and managing disfluencies, or errors in pronunciation or grammar. Self-monitoring of speech in first language speakers has been shown to begin at an early age (Foster-Cohen, 1999), but there is also evidence that language learners make productive use of oral communication strategies to monitor speech (Nakatani & Goh, 2007).
Speaking and Affective Factors
Language teachers frequently point to their learners’ lack of motivation and reluctance to speak in class (Wiltse, 2006). Such comments touch on emotional and cognitive factors that can have a sub-stantial impact on learners’ willingness to speak. Arnold (1999, p. 8) notes that “anxiety is possibly the affective factor that most pervasively obstructs the learning process”; it creates feelings of tension, apprehension, nervousness, embarrassment, and worry, and it is interwoven with factors of social pressure, personality, self-esteem, inhibition, and risk-taking. In some individuals it may substan-tially impair cognitive processes and memory during speech production. Horwitz and her colleagues (Horwitz et al., 1986) elaborated the concept of language anxiety, describing it as “a composite of self-perceptions, beliefs, feelings, and behaviors related to the learning process and arising from the unique situation in the classroom (p. 31). They argue that because “complex and non-spontaneous mental operations are required in order to communicate at all, any performance in the L2 is likely to challenge an individual’s self-concept as a competent communicator and lead to reticence, self-consciousness, fear, or even panic” (p. 128). Several studies have built on the concept of language anxiety. Gregerson (2003) found that anxious learners exhibit a high degree of frustration and may

15031-0341-Fullbook.indd 246 8/23/2016 9:27:12 PM

Teaching in the Second Language Classroom • 247

fail to monitor and respond to their own mistakes. Ohata (2005) reports that for Japanese learn-ers, anxiety was related to negative evaluation by others, leading to a sense of loss of face in oral presentation situations. Various other studies show that reticence to speak should not necessarily be interpreted as a lack of motivation; rather, it may reflect deep-seated beliefs and fears when having to perform in class in front of peers (Tsui, 1996; Lui & Jackson, 2008). It follows that teachers need to create a supportive and collaborate classroom environment (Wiltse, 2006), where they are attuned to learners’ verbal and non-verbal behaviour, where learning is carefully scaffolded, and where the notion of error and disfluencies in performance are portrayed as a natural part of language learning. Teachers can also assist learners to manage their language anxiety through strategies such as keeping a portfolio (Öztürk & Çeçen, 2007).
Fluency, Accuracy, and Complexity
The interaction of linguistic knowledge and skills, cognitive processing, and affective factors place heavy demands on speakers, and on language learners in particular, and can have a direct impact on the quality, characterised as fluency, accuracy, and complexity (e.g. Bygate, 2009), of spoken perfor-mance. The main priority of speakers is to exchange meaning (Skehan, 1998), for which fluency is vital so that the message is conveyed coherently with few pauses and hesitations, and with maximum comprehensibility for listeners. However, there is limited consensus on what constitutes fluency, and Koponen and Riggenbach (2000) argue that fluency and pronunciation overlap and can be difficult to differentiate. Language learner fluency in learners may, however, be restricted by the pressure to process speech rapidly, so that accuracy, the ability to create syntactically accurate messages and to articulate them appropriately, is impaired. Limited working memory capacity may mean that accu-racy, or form, is sacrificed when formulating meaning. With growing automisation of processes such as lexical retrieval, greater complexity is possible. Complexity combines both form and meaning, leading to greater precision and use of more advanced grammatical forms such as clause embedding, subordination, and superior awareness of social roles and relationships with interlocutors.
Fluency is likely to be of greatest concern to learners of English at all levels. Studies measuring what constructs underlie the notion of fluency, and therefore contribute to its measurability, are still relatively scarce. A study by Kormos and Dénes (2004), conducted in Hungary on native speaker lis-teners’ perceptions of fluency, concluded that “fluency is best conceived of as fast, smooth and accu-rate performance”, which can also be characterised temporally by pace (the number of stressed words per minute), which includes a specific feature, stress, that it is possible to calculate. They found that mean length of runs and speech rate were also found to be good indicators of fluency. Nation (2011) argues that fluency practice needs to be skill specific, but that there is likely to be transfer of fluency between the four skills of listening, speaking, reading, and writing, with greatest transfer between productive and receptive skills in the same mode (e.g. speaking and listening). He suggests that fluency development requires the following conditions: familiarity of content and language (lexico-grammar); communicative tasks where meaning can be conveyed; pressure to perform at faster than usual speed (e.g. 4/3/2 task where learners change partners and time in minutes to perform the same task is reduced); and quantity of practice where fluency activities form the majority of the course.
Discourses of Speech
A major criticism of Levelt’s model of speech processing is that it “appears not to take too much account of the interpersonal collaborative nature of talk” (Bygate, 2009, p. 408). Speakers do not formulate utterances based on individualized conceptualisations of what they want to say. Rather, speech is produced in cultural and social contexts where speakers must take account of the topics

15031-0341-Fullbook.indd 247 8/23/2016 9:27:13 PM

248. 
     • Anne Burns
     

they wish to raise, their relationships with other speakers, and the ways in which they can express their meanings within particular cultural and social conditions. They build interaction in the light of what their interlocutors can already be expected to know and how they expect these interlocutors to respond pragmatically (Wilkes-Gibbs, 1997). The growing trend over the last decades in researching and analyzing naturally occurring interaction has provided considerable insights into pedagogical knowledge for the teaching of speaking. Some of these areas are now briefly surveyed.

Spoken and Written Grammar
Significant insights have arisen from the analysis of speech compared with writing over several decades. Bygate (1998, citing Hatch, 1992) notes the major characteristics relate to three areas: planning, contextualization, and formality. Unlike writing, speech is more commonly unplanned, embedded in contexts of immediate use, and must respond to reciprocity conditions. Prototypically, speech is less formal than writing and characterized by features of clausal and phrasal organization (in comparison with sentence organization); lower levels of grammatical formality; left dislocation (occurrence of a noun/prepositional phrase, or pronoun to the immediate left of an already com-plete clause); involvement versus detachment; parataxis (or nextness); parallelism; repetition; repair; and conjunction (in contrast to subordination). To these could be minimally added grammatical intricacy (in contrast to lexical density), grammatical congruence (unmarked typical realisations of meaning), ellipsis, relexicalisation, and exophoric reference (out into the context). McCarthy and Carter argue that awareness of differences in spoken and written grammars is pedagogically impor-tant since “descriptions that rest on the written mode or on restricted genres and registers of spo-ken language are likely to omit many common features of everyday informal grammar and usage” (McCarthy & Carter, 1995, p. 154).
In recent years, major (if not definitive) contributions to descriptive grammars of spoken gram-mars have come from such compilations as Biber et al. (1999) and Carter and McCarthy (2006). Each of these works is based on naturally occurring language in contrast to invented examples; the data are drawn from large corpora (40 million words British and American English in 37,000 texts in Biber et al. and up to 900 million words in Carter and McCarthy), and strong emphasis is placed on linguistic function, rather than form, the importance of register in grammatical choice, and explicit emphasis on relationships between discourse factors and syntactic and lexical choices. While acknowledging that spoken and written English are based on the same foundational grammar (cf. Halliday, 1989), they each clearly reflect “the increasing attention that has been paid to register variation of language in use, ‘the lexical end of grammar’ and phraseology, in linguistic studies of English over the past two decades” (Rowley-Jolivet, 2002, p. 91). McCarthy and O’Keeffe (2004; see also O’Keeffe et al., 2009; Reppen, 2010) point to the gradually growing impact that spoken corpora analysis is having on the theory and practice of speaking pedagogy, including those for English (e.g. British National Corpus, British/Irish CANCODE spoken corpus, Michigan Corpus of Academic Spoken English, Longman Spoken American Corpus) as well as other language varieties (e.g. International Corpus of English with data from the Englishes of the Caribbean, Hong Kong, Nigeria, and Singapore). Other corpora, such as the Vienna-Oxford International Corpus of English (VOICE) corpus, have emerged to con-test the predominance of so-called native speaker models that have so long underpinned English language teaching. Given the expansion of English globally, they argue for the need to capture the use of English as a lingua franca which can represent more realistically how the language is being used by speakers from a wide variety of first language backgrounds, whose upbringing and education did not involve English (Seidlhofer, 2004).
VOICE data are unscripted, largely face-to-face, and cover a range of settings, both formal and informal, functions, and participant roles and relationships. According to Seidlhofer, the corpus

15031-0341-Fullbook.indd 248 8/23/2016 9:27:13 PM

Teaching in the Second Language Classroom • 249

demonstrates that errors most English teachers spend much time and effort correcting can be consid-ered as unproblematic, since in the data they do not appear to impede comprehension. These include omission of third-person present tense -s, confusion of relative pronouns who and which, omission or insertion of definite and indefinite articles, failure to use grammatically correct question tags, insertion of redundant prepositions (study about), overuse of verbs of high semantic generality (e.g. have, make, put, take), replacement of infinitive constructions with that-clauses, and over-explicitness (black colour). Communication problems are more noticeably caused by vocabulary unfamiliarity, lack of paraphrasing skills, and incomprehension of idiomatic expressions.
Speaking and Discourse Analysis
Mohan (2011) notes that linguistic research on language and learning has largely been dominated by 20th-century structuralist models analysing items below the sentence. Advances in discourse analy-sis (as well as the advent of sophisticated recording technologies) have significantly expanded the detailed examination of naturally occurring speech and have illuminated typical patternings that characterise spoken text. At its heart, contemporary discourse analysis unites form and function to scrutinise the work that language does (what it accomplishes) in naturalistic settings, rather than through invented examples. Various approaches to discourse analysis, including (but not exclusively) systemic-functional linguistics (Halliday & Matthiessen, 2014), conversation analysis (Garfinkel, 1967; Sacks et al., 1974), pragmatics (Austin, 1962; Searle, 1969; Grice, 1975), and critical discourse analysis (Fairclough & Wodak, 1997), have constituted a social and critical “turn” in approaches to language analysis and have been instrumental in these advances (Silberstein, 2011). While there are differentiations in epistemology and the primary locus of analysis, all of these approaches unite in a common interest in language above the level of the sentence and in interrelationships between texts and the contexts in which they arise. Systemic functional linguistics (SFL) is a linguistic theory:
oriented to descriptions of language as a resource for meaning, rather than as a system of rules. It is oriented, in other words, to speakers’ meaning potential (what they can mean). . . . SFL is concerned with texts, rather than sentences, as the basic unit through which meaning is nego-tiated. It treats grammar, in other words, as the realisation of discourse.
(Halliday & Martin, 1993, p. 22)
The contribution of SFL to the teaching of speaking is that it can illuminate the primary social and functional motivation of the text (transactional or interactional), prototypical generic structures in spoken discourse (e.g. narrative, recount), and how ideational, interpersonal, and textual register variables are played out through particular choices of language. Burns et al. (1996), Paltridge (2001), and Martinez (2011) outline ways in which spoken texts can be analysed generically and provide suggestions for practical classroom activities. While text is the analytical unit of meaning in systemic-functional linguistics, in conversation analysis it is primarily the analysis of the turn (Atkinson & Heritage, 1984). Conversation analysis (CA) is concerned with ethnomethodological questions of how people interpret everyday life and develop a sense of social order. Garfinkel (1967, p. 1) was interested in “paying attention to the most common activities of daily life the attention usually accorded extraordinary events”. As the term implies, CA is concerned mainly with informal spoken discourse and employs fine-grained analysis to illuminate the “detailed organization of everyday language” (McCarthy et al., 2010, p. 58). Key questions are related to how people take and keep turns in conversation, how they commence and shut down conversation, how they initiate, close, and shift topic, and how they collaborate to negotiate communication from utterance to utterance. Major contributions from conversation analysis are the analysis of turn-taking and the underlying localized

15031-0341-Fullbook.indd 249 8/23/2016 9:27:13 PM

250. 
     • Anne Burns
     

‘rules’ for taking turns, with the basic pattern being the adjacency pairs, which highlights the related-ness of one type of turn to another (e.g. offer-acceptance).

Based on classroom research, Barraja-Rohan (2011) shows how CA initially raised awareness in lower and intermediate second language learners of the mechanisms and norms of spoken interac-tion, and also eventually helped them to become analysts of conversation and more effective con-versationalists. Bowles and Seedhouse (2007) contains a collection of studies considering how CA can be applied in language for specific purpose contexts. Nicholas (2015), wishing to combine socio-cultural, concept-based ideas with findings from CA research. employed a three-stage process for speech act instruction: (1) orientation (instruction in fundamental concepts relevant to interactional competence; (2) execution (strategic interaction tasks carried out with a partner); and (3) control (reflection on videotaped performances in the light of understanding of the relevant concepts). He concludes that enabling EFL learners to gain more nuanced conceptual understandings of pragmatic norms should equip them to apply this knowledge in other contexts.
Pragmatics has its roots in philosophy. Grice (1975) proposed the notion of conversational maxims, based on what he termed ‘the cooperative principle’. He argued that the maxims of quality (be true), quantity (be brief), relevance (be relevant), and manner (be clear) enable conversational partners to relate to each other but also trace implicatures, or hidden or indirect meanings. Speakers must infer from implicatures what a speaker intends by drawing on both linguistic and contextual knowledge, including the cooperative principle (Schiffrin, 1990). In pragmatic theory, a primary unit of analysis is the speech act, the realisation of particular purposes, such as greeting, apologising, requesting, comple-menting, threatening, disagreeing, and so on, through language. Speech act theory focuses on the illo-cutionary act, or the performance of a particular piece of language, given form and meaning in relation to the surrounding situation. Illocutionary acts (Austin, 1962) can be both direct and indirect (as when a question is used as a command, or a statement of identification can also be an offer of service).
Brock and Nagasaka (2005) suggest a pedagogical heuristic that teachers can use to raise prag-matic awareness when teaching speaking: See, Use, Review, and Experience (SURE). See involves helping students observe language in context and raising awareness of pragmatic factors; Use intro-duces activities where students can choose how to interact based on their understanding of real or simulated contexts; Review allows for revision and recycling of areas of pragmatic competence focused on in previous activities; and Experience exposes students to video, television shows, or other media, where they observe the role pragmatics plays in communication. Tatsuki and Houck (2010) provide a range of studies that demonstrate practical examples of the use of pragmatics in language classrooms, while Wigglesworth and Yates (2007) report on a study designed to provide specific information to help teachers develop classroom activities on how native English speakers (NESs) mitigate difficult requests in the workplace context in English-speaking countries.
Critical discourse analysis (CDA) has its base in the poststructuralist theories of social philosophers such as Foucault, Bourdieu, and Habermas and in the work of critical linguists such as Fowler at el. (1979), Fairclough (e.g. 1989), and Gee (e.g. 1990). CDA interrogates how discourses are used in the construction of power relationships and in the maintenance of institutional and social dominance. CDA theorists are concerned with issues of how language operates to construct and maintain ideolo-gies of hegemony, race, gender, discrimination, class, politics, immigration, and crime. Van Dijk (2003, p. 352) notes that while other approaches in discourse analysis may to a greater or lesser extent include critical perspectives, CDA “aims to offer a different ‘mode’ or ‘perspective’ of theorizing, analysis, and application throughout the whole field.” Fairclough and Wodak (1997, pp. 271–280) set out the main principles of CDA:

1. 
   CDA addresses social problems.
   

2. 
   Power relations are discursive.
   

15031-0341-Fullbook.indd 250 8/23/2016 9:27:13 PM

Teaching in the Second Language Classroom • 251

3. 
   Discourse constitutes society and culture.
   

3. 
   Discourse does ideological work.
   

3. 
   Discourse is historical.
   

3. 
   The link between text and society is mediated.
   

3. 
   Discourse analysis is interpretative and explanatory.
   

3. 
   Discourse is a form of social action.
   

Applied to the teaching of speaking, a critical discourse perspective enables teachers and learners to consider areas such as who or what dominates or is omitted from talk, what underlying social norms are at play and how they are maintained or contested, and what gatekeeping roles are incurred. Recent studies applying a CDA perspective to spoken language are Eggins and Slade (2005), who analyse genres of casual conversation such as story-telling, gossiping, and joke-telling, and Harrington et al. (2008), focusing on language and gender research. Considering applications to language teaching, Cots (2006) argues that the development of learners’ capacities to examine and judge critical world issues is too often absent from language programs and proposes how CDA can be deployed through specific kinds of language activities. He suggests a three-part framework that raises questions teach-ers can discuss with learners when introducing a text: (1) social practice (e.g. In what kind of social situation is the text produced? What social identities are involved? What are the social consequences of the text?); (2) discourse practice (e.g. Does it require us to read between the lines? How conven-tional is the text in its context of use?); (3) textual practice (e.g. Is it obvious in any way that one of the participants is more in control of the construction than the others? How does syntactic structure as well as lexical choice affect the meaning? Are there alternatives?).

Despite the considerable advances in discourse analysis research, and calls from Firth and Wagner (1997), Rampton (1997), and Boxer and Cohen (2004) for closer interconnectedness between second language acquisition research and discourse analysis, it can still be argued that such approaches are only spasmodically exploited in teacher education. Where would-be teachers and currently serving teachers are exposed to courses in discourse analysis, it is rarely the case, I would argue, that discourse analytical insights are readily transposed into language teaching classroom practice, and perhaps even less readily in the case of teaching speaking (see, for example, findings from the studies edited by Bartels, 2005). As Borg (2005, p. 339) in Bartel’s collection argues:
Even in teacher education contexts where [knowledge about language] and methodology are addressed directly, it is not productive for what and how teachers learn about language to be disconnected from the roles their KAL [knowledge about language] will play in the classroom.
The interconnection of teacher education courses in second language acquisition, discourse analysis, and methodology is an area for research that should attract much wider attention if the findings from research on speaking are to be productively applied (see also Andrews, 2001).