As many of the other articles in this issue show, recent research in second language acquisition (SLA), corpus linguistics, and psycholinguistics has provided ample evidence of the highly formulaic nature of language. While a plethora of terms and related concepts have been used in discussions of formulaicity, only the terms formulaic or formulas will be used as umbrella termsFootnote 1 in this article to refer to multiword units that follow the idiom principle initially proposed by Sinclair (Reference Sinclair1991) and later defined by Erman and Warren (Reference Erman and Warren2000) as “combinations of at least two words favored by native speakers in preference to an alternative combination which could have been equivalent had there been no conventionalization” (p. 31). This definition suits our purposes well, as it does not imply frozenness of the word combinations and leaves room for some flexibility and variation within the word combinations.
The fact that language is largely formulaic can in some way be considered as essentially bad news for second language (L2) learners, and for many nonnative L2 teachers alike. Knowing that language as a whole is highly conventional, that even when alternative combinations exist only one is usually preferred, and that rules and formalisms are of little help in the acquisition of formulaicity may not necessarily contribute to enhancing learners’ motivation. On a more positive note, however, formulaicity has been shown to play a facilitative role in first (L1) and second language (L2) receptive and productive skills (see articles by N. C. Ellis, Conklin & Schmitt, and Wray in this volume for detailed accounts). Wood (Reference Wood2002) wrote that “the evidence that formulaic language is basic to language development, processing, production, and learning . . . has important implications for how language is dealt with in the classroom, especially for the facilitation of fluent production” (p. 2). As instruction has been shown to make a difference—at least sometimes—in helping learners acquire an L2,Footnote 2 it can be argued that adopting a formulaic approach to L2 teaching seems relevant for three main reasons at least: (a) formulaicity is ubiquitous in language; (b) formulaic language use has been shown to be a marker of proficiency in an L2 (Cortes, Reference Cortes2004; Forsberg, Reference Forsberg2010; Hyland, Reference Hyland2008); and (c) studies have demonstrated that L2 language learners find formulaicity particularly challenging as it is impossible for them to use the innate native intuition usually associated with formulaic language use (see De Cock, Reference De Cock, Campoy, Bellés-Fortuno and Gea-Valor2009, Reference De Cock, Frankenberg-Garcia, Flowerdew and Aston2010; Granger Reference Granger1998; Howarth Reference Howarth1998a, Reference Howarth and Cowie1998b; Nesselhauf, Reference Nesselhauf2003, Reference Nesselhauf2005; Pawley & Syder Reference Pawley, Syder, Richards and Schmidt1983). In sum, if teaching is meant to help learners improve their proficiency levels, it should then—at least in part—be devoted to improving learners’ knowledge and use of formulas.
A decade ago, Wood (Reference Wood2002) said that virtually no classroom materials and programs that capitalize on the vital significance of formulas in production and acquisition were available commercially. One of the aims of the present article is to find out whether this statement still holds in 2012. In order to structure the search for traces of formulaicity in instructed SLA, I will first briefly describe the various stages of instructional intervention where a formulaic approach to L2 is possible. I will then present concrete illustrations for each stage of instruction selected.
Input, intake, interaction, and output are commonly cited as being the key stages of SLA (for more details on each stage, see De Bot, Wander, & Verspoor, Reference De Bot, Wander and Verspoor2005; R. Ellis, Reference Ellis1997; Gass, Reference Gass1997; VanPatten, Reference VanPatten2003, Reference VanPatten2004; VanPatten & Benati, Reference VanPatten and Benati2010). Instructional intervention can take place at several junctures around, between, or within those stages. I have selected three key points in time when instructional intervention is possible and will use them as structuring devices for the rest of the article. The choice of input to be presented to learners will be considered as a first step of instructional intervention. The second step will include pedagogical tasks done in class. The third will deal with the post-output phase and address feedback.
It must be added that no claim of exhaustiveness is being made as far as the examples of teaching methods, tools, or practices presented in each section are concerned. This review is—by nature—incomplete and sometimes even biased, be it intentionally or not. It is unintentionally biased because of space restrictions and also because of my own limitations in expertise. It is sometimes intentionally biased because some issues tackled in other articles of the present volume have not been given lengthier treatment here. The division into three main zones or times of instructional intervention could also be discussed (potential overlap, other divisions possible) but, as explained already, the zones of instructional intervention selected will only serve as structuring devices.
TYPE OF INPUT PRESENTED TO LEARNERS
Carroll (Reference Carroll2001) wrote, “one point on which there is consensus is that SLA requires exposure to second language,” but “there is no agreement on what kind or how much exposure a learner needs” (p. 2). With regard to the kind of input recommended, although Krashen (Reference Krashen1985) pleaded for the exclusive use of comprehensible input, Schachter (Reference Schachter and Rutherford1984) commented on the potential of erroneous input. Swain (Reference Swain, Gass and Madden1985) discussed the benefits of interaction, which can include both correct and incorrect output. In sum, the literature has shown that the development of learners’ communicative competence does not depend exclusively on exposure to comprehensible input, and that interaction—even if it includes non-target-like forms potentially leading to misunderstandings or communication breakdowns—is central in SLA (Ohta, Reference Ohta and Lantolf2000). This said, none of the paradigms listed above rejects the fact that exposure to a large amount of natural native-like input (be it oral or written) remains one of the prerequisites to promoting the acquisition of formulaic language.
In instructed contexts, authentic documents, textbooks, and teacher-talk are the key sources of native(-like)Footnote 3 input. The proportion of authentic documents used in the classroom is difficult to assess: Some curricula recommend exclusive use of authentic documents, while in some other countries, textbooks are compulsory. Figures are, however, available for the proportion of teacher-talk. Teacher talking time (usually abbreviated as TTT) takes up an impressive 70% of classroom time on average (for a review of TTT studies, see Chaudron, Reference Chaudron1988), which makes it a rather massive type of input for learners. More recent research confirms this high proportion of TTT. In a study of advanced Spanish classes, Polio and Zyzik (Reference Polio and Zyzik2009) also showed that teachers themselves are aware of the fact they do “most of the talking in class” (p. 561). As for textbooks, they still represent one of the most commonly used sources of input in instructed SLA contexts. The two coming paragraphs will be thus be devoted to these two frequent sources of input in instructed SLA and to the link between those input types and formulaic language.
Textbooks
As explained here earlier, textbooks are still widely used by teachers as a main source of target language input that learners get. Yet, research has pointed discrepancies between corpus-based language descriptions and written and oral input included in textbooks. In an overview of textbook research over the last two decades, Meunier and Gouverneur (Reference Meunier, Gouverneur and Aijmer2009) showed that such discrepancies are found in several areas. Regarding lexis and phraseology, some textbooks, for example, include very infrequent words and phrases that are sometimes not worth teaching, be it for receptive purposes or productive purposes (Meunier & Gouverneur, Reference Meunier, Gouverneur, Encarnación, Quereda and Santana2007). Regarding grammar, the typical typology of first, second, and third conditionals often found in textbooks only covers 15% of a sample of 1,000 conditional sentences retrieved from an English native speaker corpus (Gabrielatos, Reference Gabrielatos2003). Other findings include that in syntax, striking discrepancies exist between the spoken language included in the textbooks and real spoken data (Römer, Reference Römer, Aston, Bernardini and Stewart2004), and in pragmatics, Hyland (Reference Hyland1994) found that despite the interest devoted to hedging in the research literature, a number of widely used textbooks display an ignorance of empirical usage, and pedagogic writing materials would benefit from revisions based on authentic data.
Nevertheless, some publishers have recently made considerable efforts to use corpus data (be it native or learner corpora) to inform textbook writing. To take one example, Cambridge University Press now offers a list of English language teaching (ELT) corpus-informed materials.Footnote 4 The list contains the following categories: adults (three textbook series), dictionaries (nine dictionaries), exams (containing all the Cambridge exams’ series), grammar (10 titles), methodology and applied linguistics (five titles), photocopiable resources (including three titles, two of which directly related to formulaic language, i.e., collocations and metaphors), professional English (eight references), pronunciation (three titles), teens (two textbooks of the English in Mind series, with one specifically dedicated to Italian learners of English), and vocabulary (23 references, several of which are devoted to collocations, idioms, or phrasal verbs). In the second edition of the English in Mind textbook, for instance, two explicit references to the use of corpora are being made. The editors explain that “a Get it right section allows students to focus on common areas of difficulty (as informed by the Cambridge Learner Corpus)” and that “there is a large number of ‘Everyday English’ expressions, selected on the basis of the CANCODE [Cambridge and Nottingham Corpus of Discourse in English] Corpus of spoken language, ensuring that your students learn lexical chunks which are of proven frequency in conversational English.”Footnote 5 Despite such efforts, however, Boulton (Reference Boulton, Harris and Moreno Jaén2010) showed that while the use of corpora tends to be given a high profile in the accompanying publicity (book covers and introductions), this visibility is often lost in the materials themselves. One other study that concretely shows the discrepancy between publicity and reality is Gouverneur (Reference Gouverneur, Meunier and Granger2008), who wrote that textbooks that are said to be corpus-based should ideally focus in part on formulaic language. In contrast, she showed not only that textbooks offer a limited range of vocabulary exercises devoted to formulaicity, but also that there is very little overlap between the textbooks (of the same level) in terms of the patterns included. Even if major publishers employ corpus linguists, the concrete impact of corpus studies in course books for general purposes is not clearly visible yet. In another recent review of one of the volumes of the new English Unlimited textbook series, Posser (Reference Posser2011) argued that the book has a present-day and authentic feel, that the learning of grammar and vocabulary is done in a natural context, and that attention is paid to typical features of English such as collocations and vague language. He added, however, that whether or not this feel is a good mirror of authentic language is difficult to assess precisely. McCarten (Reference McCarten, O'Keeffe and McCarthy2010) also claimed that course books have generally been slow to exploit corpora as a resource. This statement is supported by Burton (Reference Burton2012), who provided the results of a questionnaire completed by 13 current course book authors on their use of and attitudes toward corpora. The results of the survey suggest that while some course book authors already make use of corpora in their writing, many do not, primarily due to lack of access to corpora and a lack of time in writing schedules. Burton's conclusion is rather pessimistic, as he stated that “the nature of ELT publishing means that the influence of corpora on course books may remain limited in the future, regardless of the outcome of debates in academic literature on the subject” (p. 106).
Teacher-Talk
As for the second type of input analyzed here, teacher-talk, studies have shown that if using native-like formulaic language constitutes a major challenge for learners, it is also the case for many of their nonnative speaker teachers (García Merino, Reference García Merino1997; Maum, Reference Maum2002; Milambiling, Reference Milambiling2000). Given the high proportion of nonnative teachers in ELT, it seems unlikely that teacher-talk might be a key factor in promoting the acquisition of formulaic language. Another reason why native and nonnative teacher-talk alike might not prove very helpful in promoting incidental acquisition of formulaic language is given by Horst (Reference Horst2010), which examined the lexical characteristics of a corpus of teacher-talk and explained that “in terms of its potential for supporting incidental vocabulary acquisition, the teacher speech proved to be long on richness—hundreds of words likely to be unfamiliar were used—but short on redundancy, with few words recycled often enough to be remembered. Given the known rates of uptake (defined below) in listening contexts, incidental L2 word gains achieved in the course are likely to have been minimal” (pp. 176–177). While Horst's study did not specifically tackle formulaicity, it can nonetheless be inferred that the benefits for the uptake of formulaic language might be equally low.
As learners have been shown to be sensitive to input frequency (N. C. Ellis, Reference Ellis2002a, Reference Ellis2002b, Reference Ellis, Doughty and Long2003), the quantity and quality of the input that learners get in instructed SLA is thus of paramount importance. Despite solid scientific evidence, however, some teachers still reject the helpfulness of using authentic data and of explicitly focusing on aspects of fluency and formulaicity. Andrewes (Reference Andrewes2011) said, for instance, that he does not believe in the value of using spoken corpora. He claimed that even if some features of spoken language have been shown to be highly frequent devices used to maintain spoken conversational fluency, they are not teachable. He saw no point “in actively teaching such authentic aspects of fluency even to our advanced class in the heart of London,” and he even adds that such aspects of fluency “are likely to be irrelevant in most teaching/learning contexts” (p. 7). However, the dominant view in SLA posits that the lack of authenticity in teaching materials (more specifically in course books) will not help promote the acquisition of formulaic language. This lack of authenticity may also have circular and detrimental consequences on teacher-talk, especially in countries where L2 teachers are mostly nonnative speakers of the target language. If future teachers get poor input as learners, they might understandably find it difficult to become formulaically fluent teachers.
CLASSROOM TASKS
This section addresses activities that can be done in class to maximize learning. Nunan (Reference Nunan2004) reviewed many definitions of the notion of task in language teaching and proposed the following definition of a pedagogical task: “a piece of classroom work that involves learners in comprehending, manipulating, producing or interacting in the target language whilst their attention is focused on mobilizing their grammatical knowledge in order to express meaning, and with the intention to convey meaning rather than to manipulate form” (p. 3). He added a few lines later that his own definition “refers to the deployment of grammatical knowledge to express meaning, highlighting the fact that meaning and form are highly interrelated, and that grammar exists to enable the language user to express different communicative meanings” (p. 3). The author divided pedagogical tasks into rehearsal tasks and activation tasks. In rehearsal tasks, learners rehearse, in class, a communicative act they will carry out outside of the class. As for activation tasks, they involve communicative interaction, but not one in which learners will be rehearsing for some out-of-class communication. Rather they are designed to activate the acquisition process. Bygate, Skehan, and Swain (Reference Bygate, Skehan and Swain2001) also stressed the importance of oral tasks in introducing students to aspects of authentic spoken language use (e.g., discourse patterns, lexico-grammatical features). To the best of my knowledge, however, very few studies provide results of experiments carried out to foster formulaicity within a pedagogical task. One exception is Wray and Fitzpatrick (Reference Wray, Fitzpatrick, Meunier and Granger2008). Their study can be classified as a rehearsal task, as they tested the efficacy of memorizing and rehearsing conversational turns for L2 learners (to prepare a visit to the vet). Adopting a psycholinguistic approach, the authors analyzed what can be learned from learners’ deviations from memorized language. Their students described the fact of memorizing sentences as a liberating experience that promotes fluency and confidence; the authors also explain that memorization can be used as a means of establishing the strengths and weaknesses of learners in relation to morphology, lexis, and phraseology. Given the paucity of information that can be found on the link between pedagogical tasks and formulaicity, we will focus on other types of classroom tasks considered as essential preparatory steps or activities paving the way to further successful pedagogical tasks. In Nunan's typology (Reference Nunan2004), they are called enabling skills, language exercises, and communication activities.Footnote 6
Enabling Skills and Language Exercises
Enabling skills and language exercises have been grouped in one subsection, as the distinction between the two is not always easy to make. Nunan (Reference Nunan2004) defined enabling skills as related to the mastery of language systems (e.g., grammar, pronunciation, vocabulary) that enable learners to take part in communicative tasks. As for language exercises, they are defined as pieces of classroom work focusing learners on, and involving learners in, manipulating some aspect of the linguistic system.
The use of reference tools (in the form of dictionary searches, corpus consultation procedures, or data-driven learning [DDL]) will be included in this section. The formulaic features found in many (paper and CD-ROM versions of) current learners’ dictionariesFootnote 7 include access to collocations, idiomatic expressions, phrases, usage notes, comparisons of native and nonnative usages, natural example sentences, and sometimes extra corpus-based examples. All the main ELT publishers also offer specific stand-alone dictionaries of collocations, idioms, and phrasal verbs. Such paper and online dictionaries are also available for languages other than English (e.g., see http://www.tonitraduction.net/ for French). New types of reference tools have also been created. Such tools include web-based dictionaries cum writing aids. Granger and Paquot (Reference Granger, Paquot, Granger and Paquot2010a, Reference Granger and Paquot2010b) and Paquot (in press) have designed a web-based English for academic purposes (EAP) dictionary for nonnative writers that includes a rich corpus-based description of about 900 nontechnical words and phrases that express key functions in academic discourse (e.g., contrast, exemplification, or cause and effect), with particular focus on their phraseology (collocations and recurrent phrases). The web interface is customizable, as the content is automatically adapted to the user's discipline and L1 background. It is designed as an integrated tool where the dictionary is linked up to other language resources, in particular, a corpus-handling tool and discipline-specific corpora and exercises. A similar tool is available for learners of Italian, The Dictionary of Italian Collocations (Spina, Reference Spina and Calzolari2010). This dictionary is also integrated with an online language-learning platform. Another recent tool is StringNet (Wible & Tsao, Reference Wible, Tsao, Meunier, De Cock, Gilquin and Paquot2011). StringNet is defined as a lexical knowledge base that includes corpus-derived hybrid n-grams in which part-of-speech categories can occur alongside lexemes or word forms. This technique makes it possible to uncover the larger patterns in which collocations often tend to be embedded. Wible and Tsao suggested that access to such a knowledge base could help learners discover patterns they had not yet thought of looking for. Yet another example of such new tools is presented by Chen (Reference Chen2011). This web-based collocation retrieval tool, called WebCollocate, is based on a large section of the part-of-speech tagged Gutenberg corpusFootnote 8. Chen's study shows that students who worked on translation tasks with WebCollocate found significantly more proper English collocations than students who used other tools.
Corpus consultation procedures are gaining ground in instructed SLA contexts. Farr (Reference Farr2008) claimed that the language-teaching profession has realized that the teaching of lexical and phraseological patterns needs higher priority and that these patterns are best accessed through corpora. Farr studied the actual impact of corpus consultation procedures and outlined the findings from a survey on student teachers’ experiences of using corpora. While the results show an overall positive predisposition toward the use of corpora, they also demonstrate that the real teaching scenario often does not permit the ideal application. Several other studies have also been carried out to assess the pros and cons of corpus use. Varley (Reference Varley2009) noted that corpus consultation is gaining in prominence as a language-learning tool. Varley analyzed the response of second-year undergraduate students to a course assignment including the use of a popular concordancing software program to investigate language features characteristic of a range of genres. The results showed that students have a positive response to corpus consultation. This study also found that interest was strongest among students who had clear language-learning goals, and that the assignments produced by the students demonstrated an increased awareness of lexico-grammatical usage, particularly with regard to vocabulary use, phrases, and colligational patterns. The results of other studies on the impact of corpus consultation practices in the classrooms (Belz & Vyatkina, Reference Belz and Vyatkina2008; Boulton, Reference Boulton2009a, Reference Boulton2009b; Vannestal & Lindquist, Reference Vannestal and Lindquist2007) show that using corpora with students may require substantial support, can be time-consuming, may not appeal to all the students, and may prove beneficial for some skills and tasks but not for others. Breyer (Reference Breyer2009) highlighted the difficulty in popularizing the use of corpus tools, explaining that despite the enthusiasm in the research community, the application of corpus tools and resources in the classroom remains limited. Breyer (Reference Breyer2009, Reference Breyer2011) also insisted, as Römer (Reference Römer2006) did, that corpus literacy be part of the initial training of teachers. In countries or places where corpus literacy is more widespread in teacher populations (e.g., Hong Kong or Taiwan) studies analyzing the impact of corpus tools in foreign and second language teaching have shown overall positive results. Wible (Reference Wible, Meunier and Granger2008) argued that L2 learners have been increasingly exposed to digital language input, and that this digital turn should have implications for the learning of multiword expressions.
Dictionary consultation, corpus consultation, and DDL activities can be done in class under the guidance of a teacher but also outside the classroom (when learners use reference tools to work on assignments). Such consultation practices are more commonly used during reading and writing activities and are usually warmly recommended in the literature for their capacity of revealing the formulaic aspects of language (Cheng, Reference Cheng, O'Keeffe and McCarthy2010; Gilquin & Granger, Reference Gilquin, Granger, O'Keeffe and McCarthy2010; Greaves & Warren, Reference Greaves, Warren, O'Keeffe and McCarthy2010). Here again, some discrepancy can be found between what is suggested in academic literature and what actually happens in the classrooms. While dictionary consultation should not be too problematic, access to corpora coupled with powerful corpus query tools is not always that easy. This is due not so much to the lack of availability of corpora (see, e.g., the free access provided to the 425-million word Contemporary Corpus of American English and its powerful online corpus query tool, http://corpus.byu.edu/coca/) but rather to the lack of training of some teachers (corpus literacy not being part of their initial training), the lack of access to well-equipped computer rooms in some learning environments, the types of corpora available (many more corpora are available for general or specific academic purposes than for more general English), and also to the learners’ needs and levels (why consult corpora if you simply need to write a postcard to a friend or describe your school?). Boulton (Reference Boulton, Harris and Moreno Jaén2010) nicely summarized the issues: “Given the wealth of resources available, there would seem to be every reason to rejoice as we enjoy the prospect of ever-increasing corpus use in the classroom. . . . Despite the considerable research interest and the multiplicity of resources available, public awareness is low: corpus consultation remains rare even in university and research environments . . . and it has had virtually no impact on ‘ordinary’ learning practices elsewhere” (p. 18).
Communication Activities
Communication activities are pieces of classroom work that involve both a focus on a particular linguistic feature and a genuine exchange of meaning. Such activities include a limited focus on form during an activity that has initially been designed to be more communicative. Examples are reading a text and discussing it; listening to an audio file in order to search for specific information to be transmitted to others or discussed with others; or using peer or teacher assistance while writing a note to somebody. Activities carried out to promote intake will be addressed in this section.
In his work on the noticing hypothesis, Schmidt (Reference Schmidt1995) stated that “what learners notice in input is what becomes intake for learning” (p. 20). Another often-debated issue related to intake is the relationship between intake and uptake. Lyster and Ranta (Reference Lyster and Ranta1997) defined typical uptake as “a student's utterance that immediately follows the teacher's feedback, and which constitutes a reaction in some way to the teacher's intention to draw attention to some aspect of the student's initial utterance” (p. 49). The authors, however, distinguish between two types of student uptake: “(a) uptake that results in ‘repair’ of the error on which the feedback focused and (b) uptake that results in an utterance that still needs repair (coded as ‘needs-repair’)” (p. 49). As can be seen from the definition, in some cases intake (i.e., what is consciously noticed by the learner) might be a first step to uptake (e.g., in the form of repairs), but in some other cases intake and uptake cannot be clearly distinguished (as is the case in needs-repairs, when uptake might be delayed). I will therefore use the term intake in the present section as an umbrella term.
Several techniques can be used by teachers to promote noticing and hence facilitate intake that constitutes an essential, though not necessarily sufficient, step for language uptake and subsequent use in communicative activities. Examples include training students to notice useful chunks, furnishing typographical enhancement (such as italics, bolding, underlining, highlighting, or color-coding), or providing explicit metalinguistic comments. Many textbooks include awareness-raising activities related to the formulaic nature of language. English Unlimited: Pre-intermediate Coursebook (Tilbury, Clementson, Hendra, & Rea, Reference Tilbury, Clementson, Hendra and Rea2010) contains a section on “noticing and recording collocations” (p. 23). Another example is found in Cutting Edge (Cunningham & Moor, Reference Cunningham and Moor2005), which contains a useful language section devoted to a functional classification of useful expressions to express opinions, agreement, disagreement, and so on. Teachers’ books also regularly contain teaching tips, which include the importance of drawing learners’ attention to the formulaic nature of language. The types of sequences that receive most attention in textbooks are usually collocations (particularly verb + noun and adjective + noun combinations) and pragmatically or functionally useful formulas (e.g., introducing yourself, giving advice, [dis]agreeing with somebody, asking for help). The presence of such language boxes in textbooks helps highlight the formulaic nature of language and also makes it possible for learners to learn useful chunks as wholes (as is the case for “I'd like to + verb,” “could you please + verb,” “you could maybe + verb,” or “if I were you, I would . . .”) without focusing on the underlying syntax.
The use of corpora and natural-language-processing techniques can also facilitate the typographical enhancement of formulas. Wu, Franken, and Witten (Reference Wu, Franken and Witten2010), for instance, described a system that exploits the facilities offered by digital libraries to provide a rich collocation-learning environment. Collocations are automatically identified and highlighted in input documents with the help of a natural-language-processing system using a database of 1.3 billion short phrases drawn from the web and from which 29 million collocations have been automatically identified. This automatic identification and highlighting of collocations could be used more widely to draw learners’ attention to such sequences.
Some other types of noticing and awareness-raising activities have also been shown to favor intake and potential subsequent retention and use of specific type of formulas. Boers, Demecheleer, and Eyckmans (Reference Boers, Demecheleer, Eyckmans, Bogaards and Laufer2004) suggested that etymological elaboration can be used to retrace the meaning of figurative idioms (e.g., in jump the gun, the gun refers to the staring pistol in track sports, and therefore jump the gun literally means start running before the signal is given). A second technique is sound repetition, that is, drawing learners’ attention to the fact that alliteration (repetition of initial consonants, e.g. day dream, good as gold, blind as a bat) and assonance (repetition of identical or similar vowel sounds in neighboring words, e.g., my kind of guy) have considerable mnemonic potential for facilitating the learning of lexical chunks. As learners are not very likely to notice these sound patterns spontaneously, it is up to teachers to encourage them to recognize and exploit them (see Boers & Lindstromberg, this volume; Lindstromberg & Boers, Reference Lindstromberg and Boers2008a, Reference Lindstromberg and Boers2008b). The experiments carried out by these authors have proved beneficial for long-term retention. The use of visual media is another technique that has been put forward to help reveal fundamental characteristics of figurative language. Naciscione's (Reference Naciscione2010) work provides many examples of how visual representations of what she calls phraseological units are of stylistic and cognitive interest in bringing out the creative aspects of the verbal and the visual (e.g., hanging bank notes on a washing line to illustrate the “money laundering” expression).
Despite the existence (in some of the studies already listed here) of statistically validated results for long-term retention benefits, the actual uptake of formulaic sequences in learner language is not always easy to assess. Jones and Haywood (Reference Jones, Haywood and Schmitt2004), in a study of the acquisition of formulaic sequences in academic settings, have shown that various types of awareness-raising activities led to a considerable improvement of the recognition of formulaic language, but that the effects were more limited in terms of production. Another problem is finding an agreement on the type of formulas that merit teaching time. Some argue that because high-frequency chunks stand the best chance of being picked up by learners autonomously, teaching time should therefore be devoted to medium-frequency chunks (see Boers, Deconinck, & Lindstromberg, Reference Boers, Deconinck, Lindstromberg, De Knop, Boers and De Rycker2010). Some other authors argue, in contrast, that as even at advanced stages of proficiency frequent chunks are not mastered, teaching time should be devoted mainly to those frequent chunks (for frequent verb + noun collocations, see Nesselhauf, Reference Nesselhauf2005).
PROVIDING FEEDBACK
This final section focuses on the provision of feedback to learners, that is, the provision of positive or negative comments on learners’ output. There are two main types of feedback: implicit and explicit. R. Ellis, Loewen, and Erlam (Reference Ellis, Loewen and Erlam2006) stated that implicit feedback often takes the form of recasts, that is, positive reformulations of all or part of a learner's non-target-like utterance. It must be noted, however, that such recasts are not always noticed (and hence not subsequently taken up, as already mentioned) by learners as being corrective attempts from the part of the teacher.
As for explicit feedback, it can take two forms: explicit correction, in which the response clearly indicates that what the learner said (during teacher-learner or peer-interactions for instance) or wrote (during proofreading or written correction phases) was incorrect; or metalinguistic feedback, in which teachers provide comments, information, or questions about the learner's utterance. As was the case for some of the teaching practices listed above, neither quantitative nor qualitative evidence on the use of a formulaic type of feedback in classroom can easily be found. In the present section the focus will thus be laid on the use of more recent tools and approaches documented in the literature, namely, the use of computer tools for self-correction, whether or not initial feedback has been provided by teachers.
Chang and Sun (Reference Chang and Sun2009) and Gao (Reference Gao2011) provided the results of proofreading experiments involving the use of corpora. Chang and Sun (Reference Chang and Sun2009) analyzed how the gradual introduction to corpus literacy (and more specifically to concordance searches) enhances students’ proofreading performance. They also tried to find out whether or not scaffoldedFootnote 9 activities have a lasting effect on students’ concordance-assisted proofreading performance. Their results show significant improvement in students’ proofreading performance with the support of concordance searches. Gao (Reference Gao2011) even went one step further and focused on fully independent use of concordances, but this time with parallel corpora. He explained that studies on self-correction using corpora usually involve monolingual concordances and intervention from instructors (such as the initial marking of errors); he then wondered whether L2 learners can independently refine their output by using a parallel concordancer without any teacher-initiated feedback. He reported on a pilot study on the effects and use of a parallel concordancer in a Chinese–English translation task. The results of the pretest/posttest design study show that learners can independently make corrections by consulting parallel corpora without any guidelines from their instructors, and that the improvements made in the posttest centered on formulaicity, involving word choices and word combinations.
Despite some differences in the results, with some studies suggesting the need for guidance and others promoting exclusive self-correction, all the studies stress the potential of corpora and corpus-based tools to help L2 learners improve formulaicity. The learners’ (and sometimes also the teachers’) lack of native intuition on formulaic language can thus be partly compensated thanks to the use of such tools.
CONCLUDING REMARKS
This article has shown that what Wood wrote in 2002—that virtually no classroom materials and programs that capitalize on the vital significance of formulas in production and acquisition were available commercially—can safely be invalidated in 2012. Our increasingly refined understanding and knowledge of the formulaic nature of language has clearly impacted L2 teaching. A recent volume edited by Wood (Reference Wood2010) on perspectives on formulaic language shows that work continues to be done, including in languages other than English (such as French, Chilean Spanish, Japanese, Arabic, and Italian).
It has also been demonstrated that digital tools, corpora, and natural-language-processing techniques not only offer great potential for extracting the formulaicity of language, but also—and perhaps more importantly—now make the world of formulaicity accessible to learners and teachers. Digital literacy offers a welcome compensation for the lack of authentic input inherent to instructed SLA contexts, and also for the nonnative learners’ and teachers’ lack of intuition about formulaicity.
This review has also shown that despite considerable efforts made by some publishers to include more authenticity (and hence formulaicity) in teaching materials, more work will be needed on that front to make such efforts more visible. Although the formulaic move is more easily perceptible in dictionaries and EAP and ESP (English for specific purposes) course books, language textbooks for general purposes seem to lag behind in this respect. Two issues that might be worth addressing in the future include the selection of the type of formulas that should be exploited more fully in textbooks (high-frequency vs. low-frequency chunks?) and the way in which they should be exploited (simple typographical enhancement for some, receptive focus for others, and more productively oriented approaches for some others).
Another conclusion that can be drawn from the review is that the promotion of communicative competence, fluency, and accuracy in L2 can be achieved through the use of activities that are not communicative in essence. Examples include corpus consultation activities, DDL, and cognitively motivated activities on phonological similarities and visual imagery. N. C. Ellis (Reference Ellis, Meunier and Granger2008) explained that two particular types of tasks promoting awareness of the nature and role of formulaic sequences are shadowing and dictogloss. Shadowing is used when learners read aloud along with a transcript while listening to an audio or video file. Dictogloss is used when learners listen to a short text read by the teacher or played on a tape, and then have to work in teams to reconstruct the entire text. Such focus-on-form activities provide a chance for learners to attend to formulaic sequences by having them discover their constituent parts.
The review has also illustrated the digital gap that exists between countries or areas where, to use Wible's expression (Reference Wible, Meunier and Granger2008), the digital turn has taken place and those where learners’ and teachers’ access to and knowledge of information and communication technologies are lagging behind. Other variables explaining the discrepancy between what is advocated in the literature and what is actually done are the influence of learning strategies, learners’ styles, levels of proficiency, and needs.
Another issue that will need to be addressed in the future is the paucity of solid empirical studies reporting on formulaic-enhanced teaching practices. We desperately need studies that focus on questions such as: How many times do teachers draw learners’ attention to the formulaic nature of language? What words or expressions do they use to refer to formulaicity? Which concrete exercises do they do in class with their students to promote the productive use of formulas? How is formulaicity actually valued and assessed by teachers? In that respect, I believe that more (multimodal) corpora of authentic classroom interactions should be collected, transcribed, annotated, and analyzed.
To conclude, though L2 teaching no longer ignores the formulaic nature of language, the exact paths to follow to better teach it are still insufficiently lit. It is to be hoped that the work carried out by SLA specialists, (corpus) linguists, psycholinguists, natural-language-processing experts, and teachers will help improve our knowledge of, access to, and pedagogical treatment of formulaicity.
ACKNOWLEDGMENTS
I would like to thank the anonymous reviewers for their insightful comments on the first version of this article.
