BRIAN TOMLINSON
If you wanted to learn a foreign language, how would you do it? Whichever means of learning you chose you would almost certainly need to use materials to help you. These could be artifacts especially designed to facilitate language learning (e.g., textbooks, dictionaries, or computer games) or they could be authentic sources of “information” and experience (e.g., films, manuals, novels, mobile phones). They could be informative in providing information about the language, they could be instructional in teaching language points, they could be experiential in providing experience of the language in use, they could be elicitative in stimulating use of language, or they could be exploratory in encouraging learners to make discoveries about the language for themselves.
Given how important materials are, it is amazing how little attention they had been given by applied linguists until recently. It was not until the mid‐1990s that materials development began to be treated seriously by academics. Before then it tended to be dismissed as something which practitioners did, or treated as a subsection of methodology in which “materials were usually introduced as examples of methods in action rather than as a means to explore the principles and procedures of their development” (Tomlinson, 2001, p. 66). There were some books and articles in the 1980s which focused on such issues as materials evaluation and selection (e.g., Cunningsworth, 1984) but it has been the books of the mid‐1990s onward (e.g., McDonough & Shaw, 1993, 2003; Tomlinson, 1998, 2003, 2008, 2011, 2013a, 2013b, 2016; McGrath, 2002, 2013, 2016; Harwood, 2010, 2014; Tomlinson & Masuhara, 2010, 2017; McDonough, Shaw, & Masuhara, 2013) that have stimulated universities and teacher training institutions to give more time to consideration of how materials could be developed and exploited to facilitate language acquisition. In that period, for example, the University of Luton in England devoted a master's course to materials development, and some universities began to teach materials development as a separate module on their master's courses. At this time, Brian Tomlinson founded the Materials Development Association (MATSDA; www.matsda.org), to run conferences and workshops and to publish the journal Folio, and such associations as the Japan Association for Language Teaching (JALT), the Malaysia International Centre for English Language Teaching (MICELT), and Teaching English to Speakers of Other Languages (TESOL) in the United States set up materials development special‐interest groups.
Today, a number of dedicated materials development master's programs exist, and most universities run materials development modules. Also, there are numerous courses for teachers providing experience of developing materials as a basis for reflection on the principles and procedures which best facilitate language acquisition. In many countries, doctoral students and teachers are researching factors which contribute to the successful development and exploitation of materials, and the findings of some of these studies have been published (see, for example, McGrath, 2013, and Tomlinson & Masuhara, 2010).
Nowadays, materials development is considered to be both a practical undertaking and a field of academic study. As a practical undertaking, it involves the production, evaluation, adaptation, and use of materials. As an academic field, it studies the principles and procedures of the design, writing, implementation, and evaluation of materials. Ideally these two aspects of materials development are interactive in that the theoretical studies inform and are informed by the actual development and use of learning materials. This is one of the aims of MATSDA, and at the MATSDA/SISU Conference at Shanghai International Studies University in June 2018, over 80 presenters from 16 different countries gave presentations on either their research into the effects of the development of materials by teachers or their ideas for helping teachers to gain professional development from applying theory to practice in order to develop effective language‐learning materials. Studying and fostering the interaction between the principles of materials development and their enactment in the classroom is also one of the main objectives of the newly formed materials use research group MUSE (https://museinternational.wordpress.com/).
Many teachers argue that language learners need a dictionary and a textbook—the dictionary to help with new words and the textbook to provide systematic and focused instruction. There is no doubt that a dictionary can be very useful if the learners are taught to use it effectively, but there has been considerable debate about the value of coursebooks (Tomlinson, 2012; Tomlinson & Masuhara, 2017). Proponents argue that coursebooks give learners a sense of system, cohesion, and progress, that they help teachers prepare their lessons, that they help learners revise, and that they provide administrators with a means of ensuring standardization. “Opponents counter that a coursebook is inevitably superficial and reductionist in its coverage of language points and in its provision of language experience . . . it imposes uniformity of syllabus and approach, and it removes initiative and power from teachers” (Tomlinson, 2001, p. 67). Experiments are taking place in which the coursebook has been replaced by teacher‐developed material, by computer software, by sources from the Internet, and by experience of language in use. But most language learners are still using coursebooks.
Forms‐focused materials involve learners in deliberate and explicit learning of discrete target features of the language. Meaning‐focused materials involve learners in such language experiences as reading stories, performing plays, and completing tasks in order to help them acquire language from comprehensible input and motivated use. Proponents of the latter (e.g., Tomlinson & Masuhara, 2017) argue that affective and cognitive engagement while experiencing language in use is the key to effective language acquisition (provided the learners subsequently pay attention to the language they have experienced). Most coursebooks now provide some meaning‐focused tasks but continue to cater mainly to the perceived need for systematic presentation of discrete learning points (Tomlinson & Masuhara, 2017).
Should materials be specially designed for language learners or should they be typical of those used by proficient speakers of the language? The argument for contrived materials is that by focusing on the target feature in easy‐to‐understand examples they make the learning task simpler. The argument for authentic materials (i.e., those not specifically produced for language teaching) is that they expose learners to language as it is typically used and that authentic materials prepare learners for the reality they will encounter as users of the language. Most coursebooks favor the contrived‐materials approach (at least at lower levels) as this is what most teachers and learners seem to think is the best way to learn a language. Some applied linguists, though, consider that exposing learners to language in authentic and comprehensible use can be more motivating and is eventually more likely to help them to become effective communicators outside the classroom (e.g., Tomlinson & Masuhara, 2017).
Language practice involves repeated production of a target item in simplified environments which have been controlled to help the learners to get the item right. Use of language involves producing language in authentic situations which have not been simplified or controlled. In practice, the focus is on getting the form and or function of a target item correct. In use, the focus is on achieving effective communication. Some applied linguists argue that practice is essential in order to eventually achieve effective communication. Others counter that practice is superficial, that it can present an oversimplified and distorted representation of the language, and that it does not help learners when they need to communicate in uncontrolled and unpredictable situations. They argue that materials should focus on providing the learners with opportunities to use the language for communicative purposes. It could be argued that practice helps learners produce correct but limited language quickly whereas use helps them to eventually use language effectively. Most language‐learning materials provide some tasks requiring learners to use language but many of their activities (especially at lower levels) only require them to practice it (Tomlinson & Masuhara, 2017). This was made very evident at the MATSDA/SISU Conference referred to above when presentations from China, Japan, and South Korea in particular revealed how the emphasis of their countries' most popular coursebooks is still on practice rather than use.
While it might seem obvious that language‐learning materials should be based on theories of language acquisition, many publishers and materials developers would argue that they need to provide what teachers and learners want and are used to, and that applied linguists have not conclusively proven what best facilitates language acquisition. Consequently, many coursebooks still do not apply some of the theoretical principles proposed by applied linguists (Tomlinson 2013a, 2013b, 2013c, 2016; Tomlinson & Masuhara, 2017), especially those relating to learners needing opportunities for
Instead, many coursebooks still rely on a presentation–practice–production (PPP) approach in which information is provided about a discrete language feature, which is then practiced in easy exercises before being “produced” in guided activities. This approach has been popular with teachers for many years because it is systematic and focused, and because it provides an impression of progress. It does, however, ignore the fact that language acquisition is delayed and gradual and requires frequent, motivated, and varied encounters with a language feature. See Criado (2016) for a theoretical defense of PPP.
Most commercial language courses are global in that they are developed to be used by any language learner at a specified level anywhere in the world. The advantage is that the potential returns on the investment justify devoting considerable expertise and resources in developing a quality product. The disadvantage is that the materials are inevitably developed for an idealized target learner and cannot be expected to satisfy the needs and wants of any actual learner. Until recently, this was a major problem, as in many countries teachers treated their textbook as a script rather than a resource and imposed materials on their learners which did not connect with the learners' experience of life nor with their learning situation. Nowadays, though, teachers are being encouraged to humanize their global textbooks by, for example, localizing and personalizing the materials for their learners through adaptation and supplementation (Tomlinson, 2013d). Also, many ministries of education and large institutions are publishing their own materials. These have the disadvantage, very often, of not looking as professional and appealing as the global coursebooks but the considerable advantage of being able to explicitly relate to the learning environment of their users (Tomlinson, 2013d; Tomlinson & Masuhara, 2017).
For many years, technology has played an important role in supplementing coursebooks and many language learners nowadays spend more time working at a computer, watching a television screen, listening to a CD, or talking in the target language on a mobile phone than working with a book. The advantages of making use of technology are considerable and include
Many people would also argue, though, that materials which use new technologies
Most of the above complaints seem to be about the abuse of new technologies rather than their potential and it is likely that in the future more technology‐driven materials will be developed which are economical, humanistic, principled, and effective. Grgurović, Chapelle, and Shelley (2013), Kiddle (2013), and Tomlinson and Masuhara (2017) offer discussions about the potential benefits of electronic materials.
Perhaps the most beneficial approaches to the use of technology are blended learning (Mishan, 2013; Tomlinson & Whittaker, 2013), task‐based language teaching (TBLT) (González‐Lloret & Ortega, 2014), and mobile‐assisted learning (MAL) (Pegrum, 2014). Blended learning is an approach in which technology is used to provide the benefits listed above, and face‐to‐face interaction is used to provide experience of supported learning and of communication. In TBLT materials, each unit is driven by a task with intended nonlinguistic outcomes (e.g., organizing a business trip from a number of timetables). MAL makes use of the learners' mobile phones to provide them with experience of listening to the target language in use, of communicating with other learners, and of communicating with proficient users of the target language. One course in Turkey combined blended learning with MAL by providing taxi drivers learning English in the classroom with examples and advice while they were waiting to pick up international passengers.
This entry has outlined some of the issues which make materials development such an interesting field. Many of these issues, and such other issues as those discussed in Tomlinson (2001, 2010, 2012, 2016) and in Tomlinson and Masuhara (2017), may never be resolved. One thing that is certain, however, is that the demand to learn languages and the use of new technologies in developing language‐learning materials will continue to increase. It will be interesting to see if the content of commercial materials will become more authentic and if their activities will become more engaging and communicative.
SEE ALSO: Authenticity in the Language Teaching Curriculum; Computer‐Assisted Language Learning Effectiveness Research; Explicit Knowledge and Grammar Explanation in Second Language Instruction; Form‐Focused Instruction; Task‐Based Learning: Cognitive Underpinnings
DAN KIM
As the use of mixed methods research has noticeably grown in social science in the past decades, applied linguistics research has increasingly adapted to combine quantitative and qualitative research methodologies in a single study as an alternative to single method research. While the movement has been especially vigorous in educational evaluation, the spread of mixed methods research in applied linguistics shows the increasing understanding that “the qualitative exploration of the processes and quantitative measurement of the outcomes in a concurrent design would provide a more complete picture of the phenomenon under study” (Hashemi, 2012, p. 207). Dörnyei (2007), who made the first‐time inclusion of mixed methods in applied linguistics research methods textbook, encouraged considering “applying a mixed method research design in every situation” (p. 313), as mixing methods may bring out richer information than opting for a single methodology. Choi and Richards (2016) proclaimed “the end of the paradigm war,” after the establishment of mixed methods as an essential research methodology in social science including the field of applied linguistics (p. 3).
The mixed methods research methodology involves a process of collecting, analyzing, and mixing quantitative and qualitative methods to answer a research question. The decision to use different methods in a single study largely depends on what is the most appropriate way to approach the research question at hand and how mixing different methods helps better achieve the purpose of the study than using a single method. The underlying belief of adopting a mixed methods approach is that making use of both qualitative and quantitative methods in combination will provide a better understanding of a research problem than choosing either method alone. Tashakkori and Teddlie (2003) go beyond the definition that mixed methods is a combination of different methods, and claim that “mixed methods research has evolved to the point where it is a separate methodological orientation with its own worldview, vocabulary, and techniques” (p. x).
Mixing methods means more than merely using multiple methods: Rather, it also implies mixing different paradigms, or mental models, that are represented in different method epistemologies. The terms “mixed methods” and “multimethods” are sometimes used interchangeably, but multimethod research is used when different methods are employed in parallel or sequence but not integrated until inferences are being made, whereas mixed methods research involves interaction between and within the methods, not limited at the concluding point (Johnson, Onwuegbuzie, & Turner, 2007, p. 119).
There have been six main stances of viewing mixed method paradigms. The purist stance at one extreme claims that important philosophical differences exist between and among different paradigms of inquiry, and therefore methods and paradigms cannot be mixed: Within each paradigm, assumptions form an interconnected whole that must be respected and preserved (Lincoln & Guba, 1985; Smith & Heshusius, 1986). In the pragmatic stance, philosophical differences exist, but they are logically independent of one another; therefore, the differences can be mixed and matched because they are descriptions not prescriptions for practice, and inquiry decisions should be driven by the practical demands of the problem (Reichardt & Cook, 1979; Patton, 2002). From the perspective of the substantive theory stance, philosophical differences are not as important. Instead, what matters more and what should guide methodical decisions are the substantive issues and conceptual theories relevant to the selected context (Cooksy, Gill, & Kelly, 2001).
The complementary strengths stance acknowledges the philosophical differences between methods, but takes the view that paradigms are not fundamentally incompatible: Paradigm differences are important, and they need to be preserved without being altered or corrupted. Hence, the methods in mixed methods inquiry should be kept separate in order to preserve and maximize the integrity and strengths of each method (Brewer & Hunter, 1989; Morse, 2003). The dialectic stance also accepts paradigm differences and their importance, and it especially emphasizes the meaningful engagement and the creative tension between different methods and their mental models in order to achieve dialectical discovery of enhanced or new understandings of the phenomenon (Greene & Caracelli, 1997, 2003). Finally, the alternative paradigm stance claims that historical and philosophical differences among paradigms are reconcilable through emergent paradigms such as contemporary pragmatism, new realism, or emancipation; what should guide mixed methods practice is a new paradigm that actively embraces and promotes the mixing of methods (House & Howe, 1999; Tashakkori & Teddlie, 2003).
Why do researchers choose to mix methods? As the paradigms behind the method choice are diverse, researchers may have different purposes in choosing the mixed methods approach. Greene, Caracelli, and Graham (1989, p. 259) presented five purposes of mixing methods:
Researchers first need to fully understand the strengths and weaknesses of quantitative and qualitative research methods in order to identify the purpose of mixing methods. The purpose should guide the remaining process such as design and analysis strategies.
Several typologies for mixed methods research design have been developed based on various dimensions to consider when planning a mixed methods study. Creswell's mixed methods design typology categorizes research designs based on the purpose and process of using multiple methods. Creswell (2008; Creswell et al., 2003; Creswell & Plano Clark, 2011) claims that the decision to choose which type of mixed methods design to adopt largely depends on the following criteria: Will data be collected concurrently or sequentially? Will you give priority to quantitative or qualitative data? What is your worldview? Who is your audience? Concurrent designs include triangulation, nested, and transformative designs (Table 1).
A triangulation design is used when comparing tendencies with multifaceted perspectives, or when attempting to validate findings from quantitative method with qualitative data. In a triangulation design, the researcher often gives the same priority to both quantitative and qualitative data and compares the results from quantitative and qualitative analysis to determine if the two data sources yield similar or dissimilar results. The nested design is used when giving less priority to one type of data than the other and when using quantitative data and qualitative data separately to answer different research questions. In a nested design, the researcher collects both quantitative and qualitative data simultaneously, asking different questions with each data source. Qualitative data may be implemented within quantitative data, or vice versa. Concurrent transformative design is similar to triangulation design in terms of the stage of integration, implementation, and priority of the methods used, but, unlike other concurrent designs where theoretical perspectives behind the method choice may be present, concurrent transformative design explicitly takes a theoretical stance. Sequential designs can either be explanatory or exploratory.
The explanatory sequential design is used to explain quantitative findings with qualitative data and to identify cases for qualitative data collection from quantitative findings. In an explanatory sequential design, the researcher places priority on quantitative data collection and analysis and collects quantitative data first in the sequence. Qualitative data are used to enhance or explain the results of the quantitative data. The exploratory sequential design can be used when an instrument has not been decided on yet and the researcher wants to explore first in order to identify questions or variables for the instrument. In an exploratory design, the researcher emphasizes qualitative data rather than quantitative data, in order to build upon the initial qualitative findings.
Table 1 Mixed methods design typology (Creswell, Plano Clark, Guttman, & Hanson, 2003, p. 224)
Stage of integration | Implementation | Priority/status | Explicit theoretical perspective | |
Sequential | ||||
|
Interpretation | QUAN→qual | Usually QUAN, can be QUAL or equal | May be present |
|
Interpretation | QUAL→quan | Usually QUAL, can be QUAN or equal | May be present |
|
Interpretation | QUAL→QUAN QUAN→QUAL | Either dominant or both equal | Definitely present |
Concurrent | ||||
|
Interpretation or analysis | QUAL+QUAN | Equal | May be present |
|
Analysis | Qual within QUAN Quan within QUAL | Either dominant | May be present |
|
Usually analysis, can be interpretation | QUAL+QUAN | Either dominant or both equal | Definitely present |
When analyzing mixed data, two types of analysis are commonly used: (a) parallel/sequential (nonintegrated) tracks analysis, and (b) integrated analysis. In the parallel/sequential tracks analysis, the analysis process takes place in a sequence. The researcher may either incorporate quantitative data into a qualitative analysis, or incorporate qualitative data into a quantitative analysis. Figure 1 shows an example process of a parallel tracks analysis. Integrated analysis adopts a concurrent analysis of both qualitative and quantitative data. The researcher moves back and forth between different data sets throughout the whole analysis process. Figure 2 shows an example process of an integrated analysis.
As Figures 1 and 2 demonstrate, the major activities of data analysis are to reduce, transform, and organize the raw data to an organized form that allows for inferences, for assessing interrelationships and patterns within the data set, and for generating and validating inferences, conclusions, or interpretations. First, in the data cleaning, reduction, and transformation stages, the data sets must be compiled, organized, selected, graded, and finally transformed into manageable pieces through thematic categorization, coding, and discovery of patterns. In a parallel tracks analysis, these stages are done separately between different data types, and developing data analysis mechanisms such as a category system and a theme matrix helps organize raw data in different form and volume into a comparable set of data. Data comparison and consolidation are the unique stages in mixed methods analyses. Data comparison is used for exploring interrelationships among different data sets, and data consolidation is done by combining or merging two data sets into one consolidated data set representing a new or enhanced set of constructs. Finally, data integration is for synthesizing different types of data to achieve a coherent and holistic understanding of the research findings.
The quality criteria for mixed methods research are essentially relevant to warranting the mixing process and the result of inferences. This process is not just about the technical aspect of conducting a mixed methods study, but also involves problematization of ethical or relational criteria in the mixed methods research (Caracelli & Riggins, 1994). Ensuring quality in mixed methods research can be especially challenging because of the difficulty with linking integrated methods to credible meta‐inferences (Ivankova, 2014).
Taking a metaphorical analogy can help conceptualize the quality criteria for a mixed methods study. For instance, the research has to have good foundations and understanding of the method choice, as the conductor of an orchestra must have good knowledge and technical basis for orchestrating different instruments, players, and other factors in the course of a performance. The research has to be consistent with the mental models accompanied by the method choice, as orchestrating music also has to relate to the musical theme throughout the whole performance. In using multiple methods, balance and consideration for other instruments is the key leading to harmony. The main player and supporting player(s) must have an understanding and consensus between and among each other as to when to maximize their own techniques and capabilities, and when to allow other players to shine. Finally, the researcher has to take full responsibility in the whole decision making and execution process of conducting a mixed methods study, and this really is an integral process of building trust with the audience as a mixed methods researcher.
One additional (and crucial) factor to consider is that, as different methods and data clash and mesh, researchers may confront unexpected phenomena or issues while engaged in the mixed methods research process. Dealing with the emergent themes and incorporating them into the process of interpretation is another important role of a mixed methods researcher in ensuring the quality of research.
Lazaraton (2003) observed that the two most common qualitative approaches in applied linguistics research are ethnography and conversation analysis. Some statistical analysis incorporated into ethnographical or conversation analysis research may be the most common form of mixed methods approach in applied linguistics research. In terms of research design, concurrent designs seem to be more prevalent than sequential designs (Hashemi & Babaii, 2013). Dörnyei, Durow, and Zahran (2004) conducted an interview study to complement the results of a larger survey of 70 students in investigating language learners' acquisition of formulaic sequences. The study utilized a sequential exploratory design, where the interview data were later accompanied by interpretation of various aptitude and motivation scores of the interviewees. The quantitative findings were used to gain more generalizability of the qualitative findings. Wesely (2010) adopted a modified explanatory design to explore students' language learning motivation as it relates to their attrition from a language immersion program. The primary methods in this study were survey and interview, and qualitative methods were used to expand on or elaborate on quantitative data. Wesely claims that the use of mixed methods helped investigate the complexity and contradiction in the relationship between student motivation and persistence in immersion program.
Riazi and Candlin (2014) reviewed 40 published papers that explicitly used the term “mixed methods approach/method” in their titles or methodology sections, from 30 journals covering one decade from 2002 to 2011. They observed that, although the articles claimed the use of mixed methods, a large number of them could be labeled only as eclectic, in which the researchers seemed to believe that adding either qualitative or quantitative components to the dominant method could work to strengthen the studies, but they seemed to lack a systemic integration of underlying principles and philosophy of mixed methods. Riazi (2018) further asserts that principled mixed methods research studies are only recently emerging in applied linguistics in general (p. 5). Hashemi and Babaii (2013) made a similar observation after their content analysis of 205 research papers published in major applied linguistics journals between 1995 and 2008, concluding that only a limited number of studies achieved high degrees of integration at various research stages as a quality standard for mixed methods study. Recent methodology books on mixed methods for applied linguistics may help to improve the methodological consistency and rigor of current and future research (Brown, 2014; Moeller, Creswell, & Saville, 2016; Riazi, 2017).
SEE ALSO: Quantitative Methods
AGNES KUKULSKA-HULME
Mobile‐assisted language learning (MALL) is the use of smartphones and other mobile technologies in language learning, especially in situations where portability and situated learning offer specific advantages. The main advantages of MALL are immediate access to information, social networks, and situation‐relevant help; flexible use of time and space for learning; continuity of learning between different settings; good alignment with personal needs and preferences; easy creation and sharing of simple content like photos, videos, and audio recordings; and greater opportunity for sustained language practice while carrying out activities such as walking, waiting, or commuting. Hwang, Shih, Ma, Shadiev, and Chen (2016) show the benefits of a mobile learning design that involved students applying their knowledge to real situations and creating meaningful learning material, which led to more frequent language practice.
A key attraction of mobile learning is the ubiquity of mobile phones, both smartphones and less powerful feature phones. The number of mobile subscribers is predicted to reach 5.9 billion by 2025, equivalent to 71% of the world's population (GSMA, 2018). This greatly extends the reach of MALL to groups that have lacked opportunities for language learning or have had limited access to learning materials and resources. At the same time, the increasingly ubiquitous presence of mobile devices in many areas of life encourages a blurring of boundaries between daily life, entertainment, work, and learning. This presents a challenge to conventional, orderly, formal ways of teaching and learning. There is also reason to be concerned about some excessive uses and misuses of mobile phones and associated issues of safety, health, and well‐being.
Besides mobile phones, portable devices used for language learning include tablets, digital audio players, electronic dictionaries, e‐book readers, and handheld game consoles. Access to Wi‐Fi and GPS (global positioning system) can multiply the possibilities for learning on the move, especially when there is free access to the Internet and social media, and the chance to use location‐based services and tools. MALL may also be supported by wearable devices, for example language translation that can be delivered via wearable ear buds that may look like earphones (Gibbs, 2017). Some people will see translation tools as a threat to language learning, while others may find ways to design activities that incorporate it into teaching and learning.
MALL may still be perceived by some as just another form of computer‐assisted language learning (CALL) except that it involves the use of mobile technologies. However, there are considerable differences between CALL and MALL. Since mobile devices provide users with more immediate access to the Internet and to an abundance of apps (applications), many language learners now have access to possibly more attractive alternatives to formal language learning, such as listening to foreign‐language radio on the go, playing language games while queuing, reading blog posts related to personal interests while on holiday, or watching foreign movies while traveling on business. What is more, carrying out conventional language‐learning activities such as vocabulary learning in different settings (on the bus, in a café, in a queue) arguably changes the activities, as they vie for attention in noisy, changeable, stimulating environments that may be at once distracting and potentially enriching. Photos, videos, and audio notes can be used to capture language in use or observations about a situation or setting in which it is used. This captured information can be a lasting memory aid and a tangible link between different learning environments—for example, a captured record of a language issue encountered in a work situation taken into a language class where the issue is discussed. For all these reasons, Kukulska‐Hulme and Shield (2008, p. 273) explain that “MALL differs from computer‐assisted language learning in its use of personal, portable devices that enable new ways of learning emphasizing continuity or spontaneity of access and interaction across different contexts of use.” Jarvis and Achilleos (2013) suggest a new term and acronym, mobile assisted language use (MALU), since learners have many opportunities to “pick up” a language through daily use of mobile devices for a range of social or academic purposes—this is particularly the case for learners of English.
Different types of MALL applications include those that are designed for language study and those that are not explicitly designed for language learners but can be used to support learning, for example, automatic translation or apps that enable image editing. In terms of applications designed for language study, initially vocabulary and grammar learning proved to be popular. Early published studies reported on the use of text messaging and e‐mail for vocabulary learning, including timed interval learning (e.g., Levy & Kennedy, 2005). Stockwell's (2007) intelligent tutor system created a profile of each learner and then delivered vocabulary activities according to the areas they found most difficult. Grammar practice has also received considerable attention. Samuels (2003) described the use of handheld computers for activities such as grammar drills, adding diacritics to Latin texts, and participating in synchronous chat. Castañeda and Cho (2016) reported significant improvements in students' verb conjugation knowledge after they used a game‐like app.
Although resources and applications have tended to focus on the individual learner, there are also reports of collaborative language learning supported by mobile devices (Kukulska‐Hulme & Viberg, 2018). Learners can use their mobile devices to join communities of online learners where language skills may be practiced with other members (Niesner, 2010). Joseph (2009) described a “crowdsourcing” approach, which combined mobile content with materials on language and culture produced by learners and shared via a community site. Learner‐generated and shared language content was also the basis for the Cloudbank and LingoBee apps described by Petersen, Procter‐Legg, and Cacchione (2014).
With the adoption of e‐book readers (devices for reading electronic books) and e‐book software on other portable devices, together with facilities such as integrated dictionaries, parallel texts, and tools for translation, reading in other languages becomes a more attractive possibility, especially for those who have a long daily commute or who spend most of their time away from a fixed computer. The use of mobile devices to access newspapers and other news channels has extended opportunities to read in a second or foreign language on a regular basis. Incidental learning of vocabulary from reading can be supported through e‐books with dictionaries or adaptive software for vocabulary learning; however, unless learners are highly self‐motivated, the effectiveness of this method of learning will depend on good pedagogical design (Fisher et al., 2009). Lin (2014) found positive effects of using tablets in an extensive reading programme encouraging reading outside the classroom and evidence of collaboration among the learners. Access to podcasts and other audio channels has also extended opportunities to listen to languages more frequently, on a casual basis or as part of a routine that may include regular travel. Listening activities on a mobile phone or MP3 player can be carried out successfully while waiting for someone in a car, walking around the house and garden, or traveling (Demouy & Kukulska‐Hulme, 2010). The Audio News Trainer (ANT) app provided news audio recordings to motivate listening comprehension practice on mobile phones, with additional social media‐based interaction to enable sharing and commenting on summaries of news (Read & Kukulska‐Hulme, 2015).
Portable devices make it easier for unconfident learners to find private spaces to practice speaking or pronunciation; even within the home, a computer shared with friends or family members may offer less privacy. The ability to practice speaking and receive private feedback from teachers, while also being able to connect and practice with other learners, were positive factors identified in a successful project with young people who were using mobile phones to learn Irish (Keogh & Ní Mhurchú, 2009). Kirsch (2016) studied children's collaborative storytelling with the iPad, using an app that enabled recording, editing, and playback of oral language; the activity promoted exploratory talk and reflection on language. Writing practice can be more problematic since it is largely constrained by means of input such as small screens and keyboards, which can present a barrier to extensive writing. On the other hand, optical character recognition can enable learners to practice writing unfamiliar scripts such as Japanese kanji characters on phones equipped with a stylus (Koga et al., 2005). Hwang, Chen, Shadiev, Huang, and Chen (2014) have worked on improving elementary school learners' writing skills in English through situated activities involving comments made by peers. The adoption of tools for collaborative annotation of texts, which may be done remotely or by passing round a portable device in class, means that each learner can add a comment. Motivation may be heightened by the possibility of sharing immediately what has been written, including through social media or mobile blogging.
Increasingly, MALL applications relate language learning to a person's physical context when mobile, primarily to provide access to location‐specific language material (e.g., useful vocabulary and phrases) or to enable learners to capture aspects of language use in situ and share it with others. Applications like these make use of technologies that detect a learner's presence and enable media such as photographs, sketches, maps, audio, and video clips to be associated with a physical space for subsequent retrieval. One early system of this type provided learners of Japanese with appropriate polite expressions for their current context (Ogata & Yano, 2004). In a similar vein, Beaudin, Intille, Tapia, Rockinson, and Morris (2007) explored the use of sensors in the home for context‐sensitive learning of vocabulary on a mobile device: Sensors detected learners' interactions with objects in the home, and this triggered the audio presentation of English and Spanish phrases associated with the use of these objects. Ogata, Yin, El‐Bishouty, and Yano (2010) describe a system that detects physical objects around the learner using radio‐frequency identification (RFID) tags, and assigns questions related to the detected object, to improve vocabulary knowledge; this learning environment also allows the participants to share their knowledge. Liu, Chen, and Hwang (2018) designed a system for collaborative listening activities in a fitness center, with language learners watching videos on their phones and then QR codes at the fitness center being used as a mechanism for accessing a quiz, getting information about items of fitness equipment, and enabling collaboration on tasks. Augmented reality (AR) can be used to enrich or gamify a mobile learning experience; when an AR app recognizes an object through the user's camera, an action is triggered on the phone such as displaying text relating to that object, showing social media posts, or playing a sound (Godwin‐Jones, 2016). All these applications and systems focus primarily on learners' mobility and interaction in a designated space, although some activities may be carried out anywhere and at any time.
Mobile learning is poised at the intersection of formal and informal learning, with mobile devices forming a bridge between the two spheres. In a formal education scenario, Shao, Crook, and Koleva (2007) proposed an informal mobile group blog to support students spending time at a foreign university; this enabled the students to share observations about local language use and customs. A similar intent could be discerned in the work of Pemberton, Winter, and Fallahkhair (2010), whose collaborative mobile knowledge‐sharing system for language learners included learner‐generated content and a social network to help international students further their knowledge and understanding of local language and culture in the country where they were studying. Other researchers have also shown interest in informal learning; for example, Song and Fox (2008) reported how some student learners of English used mobile devices to support and extend their learning, driven by a goal to take every opportunity to learn new vocabulary in English. Jones (2015) found that learners of Welsh, whom she interviewed and surveyed, used mobile technologies extensively to access a wide range of resources. Kukulska‐Hulme and Bull (2009) related mobile learning to the issue of “noticing” in second language acquisition, arguing that mobile devices can support noticing; in addition they noted that recording consciously observed features also provides a method of obtaining data on what learners notice, when researching second language learning.
Mobile technology introduces greater flexibility into classroom teaching and it takes learning out of the classroom, often beyond the reach and control of the teacher. This can be perceived as a threat, but it is also clearly an opportunity to revitalize and rethink current approaches to teaching and learning. A framework and guide for teachers to help them rethink their practice and to design mobile learning activities has been provided by Kukulska‐Hulme, Norris, and Donohue (2015). Tracking of learners' actions and behaviors is becoming more common and some mobile learning platforms come with administration programs that track what the students do and how long they do it, reporting it back to the teacher. This can pose a challenge in terms of blurring the boundary between learning and leisure time, if learning is tracked beyond the classroom; learners can become stressed and resentful. It is important to develop mobile learning designs that clearly identify what is best learned in classrooms (real and virtual), what can or should be learned outside (in scheduled time and beyond), and the ways in which connections between these settings can be made. This is best done in consultation with learners.
Mobile learning appeals to a wide range of people, but there are specific groups that have been targeted or have been shown to benefit, and other groups that may have been neglected (Kukulska‐Hulme, 2013). Researchers have cautioned against assuming that the younger generation will automatically like using mobile devices for learning, even if they are avid users of mobile phones and music players in their everyday lives; nevertheless there is growing evidence that if the benefits to the learner of using mobile technology are made evident, uptake increases.
Mobile learners are those who benefit from mobile learning either because they are frequently mobile or because this type of learning is the most appropriate and convenient for them. In the first category are those with mobile lifestyles or whose work involves moving around different locations or substantial travel, for example business professionals needing to improve their language skills quickly on their way to a destination and for specific purposes such as interviews and meetings. In the second category are those who would otherwise find it difficult to access learning opportunities at times and locations to suit them. These categories may overlap, most notably in the case of migrants and refugees. In addition, mobile learning may simply be a preferred way to learn. For example, learners often appreciate the increased social contact and peer support they can draw on when faced with the need to fit language‐learning study and assessment deadlines into increasingly complex lives. When both online and mobile access modes are made available, some learners will opt for one of these while others will make use of both, perhaps changing from one to the other according to their needs and circumstances.
Mobile populations such as migrants often have an urgent need to improve their language skills for work and for social inclusion. Such groups frequently have ready access to mobile phones but perhaps not the latest models, and they may not have the necessary digital skills to make full use of mobile phones for language learning. Danaher, Moriarty, and Danaher (2009), working with mobile communities, drew attention to the fact that access to suitable technologies was often limited to the scope and duration of particular projects (p. 62); how to deal with this issue should still be considered in any new projects.
It has also been argued that mobile device use excludes some learners who have physical difficulties. Older learners are more likely to have more problems with eyesight, hearing, and manual dexterity, which may affect their ability to participate fully in mobile learning, but these problems are not limited to the older age group. The relatively small size of handheld devices presents some challenges in terms of usability (reading text, on‐screen interaction), although this has become less of an issue as new device and user interface designs are continually improving. Speech recognition can make interaction easier for some users and in situations where speech is a more natural medium than writing or typing. Learners with physical conditions meaning they are unable to sit upright for long periods in front of a computer appreciate being able to access reading materials and other activities on a mobile device, since this accommodates a range of positions. Also, students with visual impairments or dyslexia need not be excluded if they are able to listen to digital talking books or podcasts instead of reading (Barton, Penny, & Riordan, 2007). Skiada, Soroniati, Gardeli, and Zissis (2014) developed a game‐based mobile application for children with dyslexia which could help the children improve some fundamental skills such as reading comprehension. It is advisable to involve a range of target users and a variety of physical settings in the design, testing and evaluation of all mobile‐learning developments.
Mobile‐assisted language learning is part of the broader field of mobile, ubiquitous, and contextual learning, also known simply as mobile learning. While, in the early days, mobile learning was often defined in terms of its use of mobile technologies, researchers gradually began to emphasize the mobility of the learner. Winters (2006) noted that due to the rapid uptake of mobile learning in many different communities, the concept had become ill‐defined and needed to be reconceptualized, finally suggesting that “learning is mediated through mobile technologies, which are in themselves interwoven with other learning tools” (p. 11). Indeed, increasingly, mobile‐learning scenarios feature portable devices being used as part of a rich repertoire or blend of technologies and media, used as and when required by learners who may extend their studies beyond the traditional or virtual classroom. Learners who are supported in this way may even be better equipped to become lifelong learners, seeking to continually and opportunistically update and improve their language competences over the course of a lifetime. According to Ally and Prieto‐Blázquez (2014), the next generation of mobile learning will be more ubiquitous and there will be smart systems everywhere that people can learn from. The fields of mobile and ubiquitous learning are converging, and this will be reflected in new applications for MALL.
Mobile‐assisted language learning is one of the key application areas of mobile learning and is likely to continue to grow. Mobile learning is proving its potential to address authentic learner needs at the point at which they arise, and to deliver more flexible models of language learning. It supports skill development in reading, listening, speaking, and writing, and introduces innovations through location‐based learning and by linking formal and informal settings. It requires a good understanding of learner mobility, and highlights issues of accessibility, usability, and inclusion. Learner expectations, skills, and habits, as well as those of teachers, need to evolve to match the potential of mobile and ubiquitous technologies. This should result in new designs for language learning that relate more closely to emerging patterns of technology ownership, time use, mobility, and access.
SEE ALSO: Authenticity in the Language Teaching Curriculum; Collaborative Language Learning; Computer‐Assisted Language Learning Effectiveness Research; Computer‐Mediated Communication and Second Language Development; Innovation in Language Teaching and Learning; Teaching Vocabulary; Vocabulary Learning Strategies
EMA USHIODA
The word motivation derives from the Latin verb movere meaning “to move.” What moves a person to make certain choices, to engage in action, to persist in action—such questions lie at the heart of motivation theory and research. Within the field of second language acquisition, motivation research has predominantly focused on these kinds of what and why questions, such as: What reasons motivate people to learn languages? Why are some people more motivated to learn languages than others? What types of motivation are more effective in language learning and why? Since the turn of the 21st century, however, increasing attention has been paid to how questions, such as how motivation develops and emerges in interaction with the social learning environment. Though informed by theories of learning motivation in educational psychology, motivation research in second language acquisition has evolved somewhat independently, shaped by a concern to address the unique social, psychological, behavioral, and cultural complexities that acquiring a new communication code entails.
Motivation research in second language acquisition owes its origins to two Canadian social psychologists, Gardner and Lambert, who conducted a series of studies investigating language‐learning attitudes and motivation dating back to 1959, and published a report in 1972 that was to prove seminal in shaping the field for the next two decades. They theorized that motivation was a significant cause of variability in second language acquisition, and that it was independent of cognitive characteristics such as ability or language aptitude. They also speculated that learning a second language had important social and psychological dimensions which distinguished it from other types of learning motivation, since learners are expected not simply to acquire knowledge of the language but to identify with the target language community and adopt their distinctive speech behaviors and styles. Individuals' attitudes toward the target language community, as well as their ethnocentric orientation in general, were hypothesized to exert a directive influence on their language‐learning behavior. This hypothesis led Gardner and Lambert (1972, p. 132) to propose two kinds of motivational orientation in language learning: an integrative orientation reflecting a sincere and personal interest in the people and culture represented by the other group and, in its strong form, a desire to integrate into their community; and an instrumental orientation reflecting the practical value and advantages of learning a new language.
Founded in the bilingual social context of Canada, second language motivation research thus originated in a social‐psychological framework implicating the social context of second language acquisition and attitudes and relations among different linguistic communities. This social perspective distinguished second language motivation research from the individual‐cognitive perspectives then dominating mainstream motivational psychology. Through the 1990s and the turn of the century, however, efforts were made to bring motivation research in second language acquisition more in line with cognitive theories of motivation, leading to a sharper focus on the classroom setting, and on relationships among individual motivational cognitions (e.g., goals, beliefs, perceptions of competence), learner behaviors (e.g., effort, task engagement or avoidance), learning process and experience (e.g., success, failure), and features of the learning situation (e.g., materials, teaching methods, interpersonal relations, competitive versus collaborative learning environments).
Nevertheless, despite the increased focus on the classroom setting in the study of language‐learning motivation, the complex social and cultural implications of acquiring a new communication code have continued to preoccupy motivation researchers in second language acquisition, and Gardner and Lambert's concepts of integrative and instrumental motivational orientation have maintained their enduring influence on theorizing about why people learn languages or choose particular languages to learn. However, the first decade of the 21st century saw growing discussions in applied linguistics about the global spread of English, which provoked a reconsideration of the integrative motivation concept. In the case of English as a target language, researchers began to question whether the concept of integrative orientation towards English language speakers, their community and culture, made sense, given the status of English as a global language and lingua franca spoken by a global community of which, by definition, the English language learner is also a member, and given the increasingly common status of English as a basic educational skill in many elementary school curricula, alongside mother tongue literacy and mathematical skills. As a consequence, research attention shifted from attitudes to and identification with external reference groups (i.e., target language communities), and turned to focus instead on the internal domain of self and identity and to consider language‐learning motivation as a process of identification within the person's self‐concept.
In particular, Dörnyei (2009a) developed an influential conceptualization of the “L2 motivational self system” centered on people's vision of themselves in the future. Its central concept is the ideal self, signifying the attributes that one would ideally like to possess (i.e., a representation of personal hopes, aspirations, or desires). A complementary concept is the ought‐to self, signifying the attributes that one believes one ought to possess (i.e., a representation of someone else's sense of duty, obligations, or responsibilities). A basic tenet is that if proficiency in the target language is integral to one's ideal or ought‐to self this will serve as a powerful motivator to learn the language because of our psychological desire to reduce the discrepancy between current and desired future self states. In parallel with this theoretical focus on future self states, other researchers have conceptualized language‐learning motivation in terms of identity goals, such as the pursuit of bilingual or global citizen identities (e.g., Lamb & Budiyanto, 2013).
As noted, this reframing of motivation in relation to concepts of self and identity came about largely in the context of English language globalization, and reflects a dominant empirical focus on English‐language‐learning contexts in the study of motivation in second language acquisition in the 21st century. In this respect, the extent to which currently influential self‐and‐identity concepts can be applied to the motivation for learning target languages other than English needs further exploration (Ushioda & Dörnyei, 2017). Much more empirical work is also needed to examine their applicability to motivation for learning multiple languages, and to understand how motivation for learning the dominant global language English may interact or interfere with motivation for learning other languages (e.g., Henry, 2017; Ushioda, 2017).
Whether informed by social‐psychological, individual‐cognitive, or self‐and‐identity perspectives, motivation in second language acquisition has traditionally been conceptualized as an individual difference variable that is implicated in language‐learning success, alongside other individual difference variables such as aptitude, personality, or anxiety. According to this view, motivation is something that is located within the inner psychological life of the person, which may be influenced by a variety of external factors (e.g., teachers, parents, examinations) or experiences (e.g., travel, intercultural encounters), and which affects how the person engages in, persists at, and succeeds in the language‐learning process. In short, motivation has been conceptualized as a component in a cause‐and‐effect relationship, whereby external factors influence motivation, motivation influences learning behavior and outcomes, and these outcomes (e.g., success or failure) may in turn influence subsequent motivation. A major aim of motivation theory and research is to predict and identify what types of external factors (e.g., teacher behaviors or strategies) may have a positive effect on learner motivation, and what forms of motivation may result in optimal learner behaviors and outcomes.
Since the turn of the century, however, researchers have begun to recognize the limitations of cause‐and‐effect models of motivation and learning behavior. Such models conceptualize motivation as measurable cause or product of particular variables, but shed little light on how motivation develops and evolves in dynamic interaction with the complex range of learner‐internal, social, and contextual processes that characterize learning. Nor do they shed light on why the same classroom setting, learning tasks, teacher behaviors, and strategies may affect the motivation of its learners in different ways, leading to different kinds of motivated engagement or nonengagement in learning. In short, cause‐and‐effect models can explain the workings of motivation only in rather general terms of how certain types of learner are likely to feel motivated and behave in certain ways in response to certain kinds of contextual factors. In this sense, in the pursuit of generalizability, such models eliminate much of the messy complexity, unpredictability, and variability of individual motivation.
In the 21st century, alternative theoretical perspectives have thus come to the fore to try to address some of these limitations. There has been a growing concern with the dynamic nature of motivation as process (rather than measurable cause or product), and with how motivation evolves in organic interaction with a host of cognitive, affective, experiential, social, and contextual processes. A key issue now emerging is how to integrate the individual and context in the analysis of motivation. This issue mirrors similar theoretical concerns in contemporary mainstream motivational psychology, and also reflects broader debates within applied linguistics concerning social‐interactive and complexity theory perspectives. In this connection, Dörnyei (2009b) argued against an individual difference approach to motivation (and other learner characteristics) in favor of a dynamic systems theory framework. Dynamic systems approaches concern the behavior of complex systems that contain multiple interconnected components, where development is characterized by nonlinear growth as systems adapt and evolve organically in response to contextual processes, and in ways that contribute to shaping context. As Dörnyei argues, adopting a dynamic systems perspective on second language acquisition processes renders the notion of discrete individual difference variables (such as motivation) rather meaningless, since processes of motivation, cognition, emotion, and their constituent components continuously interact with one another and the developing context, thereby changing and causing change, as the whole system restructures, adapts, and evolves. In a similar vein, in order to capture the mutually constitutive relationship between persons and the contexts in which they act, Ushioda (2009) put forward a “person‐in‐context relational view” of motivation as an organic process that is emergent through the interactions between the reflexive intentional agency of the person, and the fluid and complex systems of social relations, activities, experiences, and multiple micro and macrocontexts in which the person is embedded, moves, and is inherently part of.
The emphasis on sociodynamic interactions also exposes the potential for conflict and tension in these interactions, as highlighted in more critical approaches to integrating individual and context in the analysis of motivation. Such approaches focus in particular on unequal power relations in language learners' struggle to participate in interactional settings in desired social, educational, or professional communities of practice. Strongly influenced by feminist and poststructuralist theory, Norton (2000) developed a concept of language motivation as investment, which she defines as “the socially and historically constructed relationship of learners to the target language, and their often ambivalent desire to learn and practice it” (p. 10). When learners invest in a second language, they do so in the hope that they will acquire a wider range of symbolic and material resources, which will enhance their cultural capital, their conception of themselves, and their desires for the future. A person's investment in a language may be mediated by other investments that may conflict with the desire to speak, such as fear of being marginalized as an immigrant, or resistance when one's professional status or cultural background is not valued. In short, an investment in a new language is also an investment in one's identity, where identity is a site of struggle, being locally constructed, negotiated, and re‐formed each time through a person's interactions with others. In this sense, motivation to learn and use a second language may be socially constrained rather than socially constructed, thus further complicating the issue of how to integrate individual and context in the analysis of motivation.
From a methodological point of view, the current emphasis on relational, sociocontextual, and dynamic systems perspectives on motivation in second language acquisition raises issues and challenges for empirical research (see Dörnyei, MacIntyre, & Henry, 2015). The study of motivation in second language acquisition has been characterized by a solid tradition of empirical investigation. For much of its history, this tradition has been dominated by quantitative methodologies suited to investigating social‐psychological and individual‐cognitive perspectives on motivation as an individual difference characteristic implicated in language‐learning success. Since motivation is not directly observable, such methodologies have typically relied on gathering self‐report measures of motivation, such as how much learners agree or disagree with or rate the importance of certain statements reflecting particular attitudes, intentions, or behaviors. To minimize the inherent subjectivity of such data, considerable attention has been paid to constructing rigorous measurement instruments with good psychometric properties, as exemplified in the standardized attitude/motivation test battery developed by Gardner (1985) and his associates in Canada. These motivation measures are then entered into statistical analyses of relationships with other independent or dependent variables, such as measures of language aptitude, anxiety, or second language proficiency (e.g., Gardner, Tremblay, & Masgoret, 1997).
However, these quantitative methods of inquiry are ill equipped to investigate more complex, dynamic, and contextual perspectives, since such methods typically rely on superficial snapshot measures at a particular point in time and are not sensitive to the complex particularities of evolving motivational experiences or individual–contextual interactions. Since the turn of the century, more qualitative methods of inquiry have gradually begun to complement the dominant quantitative paradigm, in an effort to address the dynamic and situated complexity of language‐learning motivation (see Boo, Dörnyei, & Ryan, 2015, for a comprehensive ten‐year survey of this changing methodological landscape). In particular, unstructured, semi‐structured, or narrative interview techniques are increasingly used to elicit in‐depth self‐report data on motivation and motivational experience, with the transcribed data then subjected to thematic or narrative analysis (e.g., Harvey, 2017; MacIntyre, Baker, & Sparling, 2017). While interview studies are inevitably limited in scale and scope, the data elicited can offer very rich insights into the process and experience of motivation, particularly when a longitudinal research design is adopted with multiple interviews with participants. Moreover, limitations in the scope of the dataset can be offset by complementing it with more large‐scale quantitative data in a mixed methods research design. In particular, with current moves towards more contextually grounded sociodynamic perspectives on motivation, the exploration of individual–contextual interactions is likely to entail triangulation of multiple forms of data from diverse perspectives (e.g., interviews with teachers and students, classroom observations, classroom interaction data, focus group discussions, learner reflective writing), in order to capture a rich holistic analysis of motivation in context, rather than relying (as traditionally) on a single set of self‐report measures (e.g., Henry, 2015; Sampson, 2016).
An interesting concluding observation to make about motivation research in second language acquisition is that it has endured a rather marginalized position within the field, remaining somewhat isolated from the more mainstream linguistic and psycholinguistic traditions of research that have prevailed. This is despite the fact that motivation is widely recognized as a prerequisite for successful second language learning to take place. One explanation for this lack of interaction with mainstream research in the field is that, to date, the analysis of motivation and its role in language learning has largely been at the level of global learning outcomes or measures of proficiency, and research has had rather little to say about how motivational factors and processes relate to the interim processes and intricacies of second language development (Ushioda, 2016). However, once motivation becomes viewed as an integral part of the evolving organic and adaptive system of cognitive, affective, and contextual processes shaping language learning, it is clear that the analysis of motivation can no longer be separated from the central concerns of second language acquisition research. Moreover, it seems likely that the analysis of motivation may now play a major role in any complex or dynamic systems perspective on second language acquisition, given the need to consider the processes of human agency, intentionality, and reflexivity that are fundamental to the dynamic interactions between self and context.
SEE ALSO: Agency in Second Language Acquisition; Attitudes and Motivation in Bilingual Education; Chaos/Complexity Theory for Second Language Acquisition/Development; Identities and Language Teaching in Classrooms; Identity and Second Language Acquisition; Third Language Acquisition
ELENA COTOS
Move analysis emerged in the 1970s and 1980s as an analytic approach within genre analysis. It was first introduced in the study of language for specific purposes (LSP) by John Swales (1981), who proposed it as a pedagogic approach aimed to help both first and second language writers develop academic communicative competence. Grounded in Swales's (1990) view of genre theory, move analysis builds upon the concept of rhetorical action and focuses on the organization of discourse in relation to communicative purposes, which reflect the rationales and constraints underlying the communication of discourse communities.
A discourse community is a social group that shares common public goals. To further their communicative aims, members of a discourse community utilize specific discourse mechanisms, including the conventionalized text structures and language choices of specific genres. Swales (1990) clarifies that discourse communities can be identified within both academic and nonacademic contexts. For example, the intercommunication of academic discourse communities takes place through journal articles, scholarly monographs, conference presentations, and so on. To exemplify a nonacademic discourse community, Swales described a society of stamp collectors who resided in different parts of the world but were united by a common goal to foster interest and knowledge of the stamps of Hong Kong. They communicated through a bimonthly journal and newsletter, for which they developed a set of communicative conventions governing the articles, queries, and replies in this community‐specific newsletter genre (see Swales, 1990, pp. 24–9).
The communicative conventions established by discourse communities shape their genre writing practices and serve as a means of maintaining and extending knowledge, as well as of initiating new members into the community. Swales metaphorically calls the communicative purposes of genres moves, which he sees as discourse segments that can move discourse forward by performing coherent communicative functions. The classic example characterizing genre discourse in terms of moves is Swales's (1981) CARS (create a research space) model for research articles' introduction sections. The CARS model renders the rhetorical structure of introductions as three moves: establishing a territory, establishing a niche, and occupying the niche. The purpose of the first move is to emphasize the importance of the general research topic, making strong claims, and supporting them with domain knowledge and references to previous research. The second move presents a critical stance toward the state‐of‐the‐art knowledge on the topic, asserting the need for new research by highlighting existing gaps. Move three then transitions to the current study, introducing its key aspects to demonstrate how it addressed a given gap.
The moves are realized by steps, or rhetorical strategies that help accomplish communicative purposes. The steps carry more fine‐grained functional meanings expressed with particular language choices. One such strategy used to establish the knowledge territory is the claiming centrality step, which authors use to highlight the prominence, importance, and interest of the discourse community in the topic in order to substantiate the worthiness of investigating it. For instance, in the sentence “This use of the max‐product algorithm for computing approximate MAP assignments on graphs with cycles has been the focus of considerable research in recent years,” the expression “has been the focus of considerable research” indicates the interest of the discourse community in the use of the max‐product algorithm, and “in recent years” suggests how prominent, time wise, their interest is.
CARS was extensively studied and validated across different academic communities and it evolved over the years to reflect how they realize rhetorical intent in conventionalized ways (Swales, 2004, 2011). Research has shown that the moves and their constituent steps frame the structure of genre texts and serve as prototypical criteria for genre identity. They may vary in length, may be recursive, and may be conventionalized or used optionally by discourse communities.
Move analysis is an approach for analysis of the macrostructure and rhetorical composition of instances of text types. Move analysis is essentially a qualitative approach as it requires cognitive judgment to interpret the author's rhetorical intent and to segment a text into discourse units. Traditionally, it was performed on small collections of texts representing a particular text type but is now also applied to large corpora. The book Discourse on the Move by Biber, Connor, and Upton (2007) explains the analytical procedure of corpus‐based top‐down approaches to discourse analysis in general and of move analysis in particular. Unlike bottom‐up approaches, which quantitatively examine variation across text types in terms of linguistic structures characteristic of certain domains of use, top‐down move analysis studies examine meaning‐based units and the language that expresses functional meaning (as exemplified in the sentence of claiming centrality above). Overall, move analysis has evolved into a mixed methods approach to discourse analysis unfolding in several phases: framework development, trialing, validation, and analytic description.
The initial framework development phase entails an inductive analysis, which combines the researchers' knowledge of the communication goal of the target genre and the text data they analyze. Specifically, researchers first identify segments of texts based on their roles in achieving the communication goal of the text. Those segments are then examined more closely to create tentative move categories. The internal composition of each move is further analyzed with a functional‐semantic focus to distinguish the steps, or rhetorical strategies, by which move goals may be accomplished. The researcher conducts this analysis by determining what a text segment is doing functionally to contribute to the move. For example, “Sediment export and turbid runoff from construction sites continue to be a concern” highlights a problem, while “In fact, few studies have documented the effects of highway construction on the water quality of receiving streams” indicates a gap—both these steps contribute to accomplishing the goal of establishing a niche move in Swales's CARS model. Typically, specific language choices signal functional meanings, like “continue to be a concern” and “few studies have documented” in the examples above. Rhetorical intent, however, is not always explicitly articulated, which is why repeated close reading of text segments in context and deliberation with others including content area experts are very important during this phase. Also, because the functional‐semantic themes that emerge may be notionally similar, it is important to address and eliminate conceptual duplication, which may reduce the number of step categories.
Trialing the tentative framework starts with pilot coding. Here, sample texts are segmented into discourse units and classified by applying the tentative move and step categories. Biber et al. (2007) recommend that at least two analysts individually code the texts. This way, the researchers develop a deeper understanding of the global and local purposes of text segments and fine‐tune the moves and steps through extensive reflection and comparative discussion. Additionally, pilot coding enables them to generate detailed descriptors used to formulate definitions for the proposed moves and steps. For example, steps can be defined in terms of function (doing what?), rhetorical intent (why?), content realizations (with what content?), and linguistic realizations (with what language choices?) as shown by Cotos, Link, and Huffman (2016). In their model for discussion/conclusion sections of research articles, they define the step called countering with evidence, which is part of the reshaping the territory move, as follows:
Pilot coding informs the development of a coding protocol, which should address decisions such as how to define the unit of analysis (e.g., proposition, sentence, rhetorically coherent stretch of text), how to identify unit boundaries, how to address multiple functions in the same text segment, how to resolve disagreement between coders, and so on.
Validation of the framework can take different forms. Consulting with expert members of the target discourse community is of utmost importance, although it is not always possible. In corpus‐based studies, collaboration with expert consultants is usually established at the very beginning when the corpus is compiled. To ensure that the experts' input regarding the framework is meaningful, they need to be explained what the moves and steps mean and how they were derived. Then, they can be asked to holistically evaluate each category in view of the proposed definitions and examples and to provide feedback on their clarity and fit with the expectations of their discourse community. Additionally, researchers can code sample texts together with one or more experts. Expert perceptions of whether each move or step is generally common, uncommon, or lacking can also help researchers better understand the genre practices. The insights obtained from experts usually help clarify potentially confusing areas, revise and perhaps add move/step categories, and subsequently refine the protocol for final coding.
Once the moves/steps are comprehensively defined, the framework is used to annotate all the texts in the sample. Annotation can be performed at the level of moves only or at the level of both moves and constituent steps. Because identifying rhetorical functions can naturally be confined by subjective human judgment and interpretation, this phase involves training of multiple coders and requires measures of inter‐coder reliability. Training includes explanation and close examination of each category in the framework, practice using the coding protocol, independent annotation of the same texts by all the coders, and adjudication sessions to discuss agreements and resolve disagreements. The more rigorous the training is, the more likely the coders are to identify the same discourse segments and agree on their classification into moves and steps. Reliability can be determined through simple calculations of percent agreement, or chance‐corrected measures like Cohen's kappa. In larger studies, especially those that are corpus based, the annotation process should be continually informed by reliability checks and iterative calculations of inter‐coder consistency and agreement. In studies where multiple coders are not available, researchers may either have a second coder who annotates a random sample of texts or code the texts twice with some time interval and recode where necessary.
The final, analytic description stage is exploiting the annotated data to derive various descriptions of move features that characterize the target genre. Traditional move analysis can yield rich descriptions of the rhetorical composition and summaries of occurrences of linguistic features of individual texts. These results are somewhat limited because a few randomly selected texts may not be entirely representative of the genre. Corpus‐based studies, on the other hand, afford more comprehensive discovery. Quantifications such as frequency of occurrence and average length of individual move types in the corpus can help researchers determine prototypical and differential distribution of discourse structures, detect typical and alternate cyclical patterns, and infer which moves may be obligatory and which may be optional. Combined top‐down and bottom‐up analyses can reveal the lexicogrammatical characteristics of individual move types and show how linguistic features may co‐occur and interact to accomplish communicative purposes. Additionally, the macrotextual and microtextual distinctiveness of genres can be scrutinized with foci on vocabulary‐based discourse units, lexical bundles, interpersonal features, pragmatic uses of modality, and rhetorical unity defined by spatiotemporal features (see Cotos, 2017). More sophisticated quantitative methods (e.g., multidimensional analysis, k‐means clustering, principal component analysis, chronotopic analysis, linear discriminate analysis) can be employed to identify significant variations and similarities in the rhetorical composition and linguistic realizations in different corpora (Crossley, 2007; Kanoksilapatham, 2007; Kaufer, Cotos, Cai, & Ishizaki, 2017).
Move analysis studies span a wide range of academic and professional genres, including doctoral theses, abstracts, grant proposals, graduate program applications, job application letters, negotiation letters, medical consensus statements, law reports, and corporate press releases. The research article, however, is one of the most extensively investigated genres. Corpora in numerous disciplines and subdisciplines were analyzed to devise move/step frameworks for methods, results, and discussion/conclusions, describing how particular academic fields construct these part‐genres. The move anatomies of articles in multiple fields have also been compared, demonstrating how the moves of the same genre are embedded in field‐specific discursive patterns and how they surface across multiple disciplines. Cotos, Huffman, and Link (2015), for example, show that disciplines form clusters based on the use of steps in each section of the article. While some disciplines stay within the same cluster in all the sections, other disciplines migrate from cluster to cluster, resembling similarity with some disciplines in one section and with other disciplines in other sections. Moreover, the probability of transitioning from one step to another in each section indicates that certain steps are likely to occur sequentially, while some step transitions are almost inexistent.
Offering insights into the textual nature of rhetorical actions, move analysis research reifies Swales's (1981, 1990) pedagogic ideas, which draw on the concepts of genre, discourse community, and language‐learning task. Swales presented his vision for move analysis as a task‐based approach to teaching postsecondary academic writing. In his view, the communicative purposes of genres can serve as core determinants for pedagogic tasks to be shaped as activities in the classroom. Move frameworks are thus “a workable way of making sense of the myriad communicative events that occur in the contemporary English‐speaking academy—a sense‐making directly relevant to those concerned with devising English courses and, by extension, to those participating in such courses” (1990, p. 1).
Applications of move analysis in genre‐based writing pedagogy aim to sensitize students to reoccurring rhetorical effects and structures, enhance their command of functional language and metadiscourse, develop their ability to rationalize their communicative behavior, and potentially enable the transfer of their skills acquired for one genre to another. Introducing and sensitizing learners to genres requires an explicit focus on the textual practices of their discourse community (Hyland, 2003). Teachers thus use specialized corpora as tangible resources for students to explore the structural organization and rhetorical composition of texts and acquire an understanding of how discourse communities achieve their sociodisciplinary purposes. To acquaint students with the contexts inhabited by the genre, teachers often advise students to find and collect texts, frequently seeking direction from their professors. Alternatively, they may choose to compile an in‐house learner corpus and have their students compare it with genre exemplars produced by proficient writers in the target domain. While the directions for text exploration may vary, most proposed guidelines (Cotos, 2017; Flowerdew, 2015; Swales & Feak, 2012) share the following characteristics:
The order of these analytic activities is intentional. In the first activity, students are apprenticed into reading and writing texts through modeling of how to explore and identify discursive patterns that conventionalize the rhetorical identity of texts. Students then delve into detailed analyses of the internal structure of texts, identifying textual boundaries between rhetorical moves based on content and linguistic exponents, also clarifying constraints on move sequences. Annotated corpora can be displayed visually such that the move structure of each text becomes salient, enabling students to notice patterns of distribution, the step composition of moves, the function of a given step in different texts, and different ways in which authors integrate more than one function into the same text segment (see Cotos, Link, & Huffman, 2017). Then, students are geared toward an understanding of how genres systematically connect linguistic actions to communicative functions. For this purpose, teaching methods often include paired top‐down and bottom‐up data‐driven learning tasks, which allow students to detect relations between moves/steps and specific lexicogrammatical features. This becomes a rudimentary prerequisite for rhetorical consciousness raising, which is necessary for students to further decipher how communicative conventions may be indicative of the values, ideologies, and even inbred power relations, and to recognize how social context may influence expectations of genre use.
Thus, apart from explicitness, move analysis in genre‐based pedagogy provides a coherent framework for focusing on both language and contexts, understanding patterns and variation in texts, and challenging valued discourses (Hyland, 2007). While not yet as pervasive as classroom uses of bottom‐up techniques afforded by corpora and concordances, move analysis‐driven pedagogies have made significant strides in instructional settings. Empirical evaluations show that students and teachers alike perceive this approach as aiding the learning of genre conventions and development of advanced literacy (Cortes, 2011).
Over the years, move analysis has advanced to new frontiers yielding more complex theoretical interpretations by applied linguists. Flowerdew (2015) links Swales's approach with educational theories by drawing parallels with the concepts of situated cognition, cognitive apprenticeship, situated learning, and Vygotskyan scaffolding, also recommending a crossover with systemic functional linguistics for an enriched approach to genre‐based pedagogy. Similarly, Bruce (2008) connects communicative moves with cognitive genres, arguing that the communicative purposes of social genres can be realized with a range of cognitive genres (e.g., report, explanation, discussion, recount). Cotos, Huffman, and Link (2017) further this idea, suggesting that different cognitive genres, including those described by Bruce (2008) and Pilegaard and Frandsen (1996), are integrated in the step functions of moves.
Move analysis is also entering unexplored interdisciplinary territories, as moves and steps are mapped onto different dimensions. For example, in rhetoric there is an emerging agenda to connect move analysis with rhetorical priming theory, which views language as an instrument used to prime readers' interactive experiences. Here, the point of convergence is in characterizing the functional meanings of rhetorical moves in terms of so‐called priming strings, that is, socially shared and functionally situated language patterns that lead readers to have particular experiences with texts. The goal is to develop mutual linguistic‐rhetorical tenets for genre theory and writing pedagogy through an integrated linguistic‐rhetorical research approach (Kaufer et al., 2017).
Another extrapolation interfaces move analysis with concept‐matching analysis. The latter is used by researchers in natural language processing (NLP) and machine learning (ML) who are interested in automated identification of discourse elements. In this integrated approach, the moves and steps are modeled as patterns of linked constituent concepts (e.g., novelty, emphasis, trend, contrast), which are represented by syntactically related lexical choices (Sándor, 2007). The two methods are mutually informative, as move analysis can enrich the inventory of concepts, and concept‐based analysis in turn can provide a finer representation of the linguistic instantiations of communicative goals.
The intersections of move analysis with NLP and ML are ushering in innovative technological applications. For example, the Mover software automatically classifies and displays the move structure of research article abstracts (Anthony & Lashkia, 2003). The Academic Word Suggestion Machine autosuggests move‐specific lexical bundles as the student types in a word (Mizumoto, Hamatani, & Imao, 2017). The Research Writing Tutor analyzes student drafts of individual research article sections and generates automated feedback at the level of moves and steps, comparing them with the rhetorical composition of corpora in students' disciplines (Cotos, 2016).
To conclude, move analysis provides a means for examining the constituent parts of a genre, deepening our understanding of how the structure of genre texts is conditioned by the communicative purposes of discourse communities. The growing synergy between move analysis and other epistemologically different traditions is paving the way for enriched theories, blended research methodologies, and more grounded pedagogies and technological applications.
SEE ALSO: Genre and Discourse Analysis in Language for Specific Purposes; Genre‐Based Language Teaching
VIVIAN COOK
The term “multicompetence” was originally defined as “the compound state of a mind with two grammars” (Cook, 1991). The current definition is “the overall system of a mind or a community that uses more than one language” (Cook, 2016a, p. 3), extending the concept to the knowledge and use of two or more languages by the same individual or the same community and avoiding issues with the different interpretations of “grammar.” Multicompetence thus bases second language acquisition (SLA) on the second language (L2) user as a whole person rather than on the monolingual native speaker and on the whole multilingual community rather than separate language communities.
Multicompetence assumes that someone who knows two or more languages is a different person from a monolingual and so needs to be looked at in their own right. Similarly, a multilingual community is not the sum of two monolingual communities. Multicompetence is thus not so much a model or a theory as an overall perspective or framework: it changes the angle from which SLA is viewed. In Grosjean's (2008) terms, it constitutes a bilingual “wholistic” interpretation of bilingualism as opposed to a monolingual “fractional” interpretation.
The multicompetence perspective uses the term “L2 user” for those who know and use a second language at any level, and reserves “L2 learner” for people who are not actually using the second language for authentic purposes, such as children in foreign language classrooms. L2 users may be L2 learners at different times of life or indeed times of day—an L2 learner of English in London who steps out of the classroom immediately becomes an L2 user of English.
Cook (2016a) summed up multicompetence in terms of three premises:
A valid multicompetence research methodology must differ from that used in traditional SLA research (Murahata, Murahata, & Cook, 2016). In terms of the three premises:
Treating all the languages in the user's mind as interconnected means examining not only the conventional transfer from first language to second but also the transfer from second to first: Do you still speak your first language like a monolingual native speaker when you know another language? A negative answer to this could reverberate across SLA research, and indeed much linguistics, as it would be unsafe to regard any person who knows another language as having “normal” monolingual linguistic competence; “the judgments about English of Bloomfield, Halliday or Chomsky are not trustworthy, except where they are supported by evidence from ‘pure’ monolinguals” (Cook, 2002).
L2 users have turned out to be different from monolinguals in many respects. To take some representative studies:
So more than language has changed in the L2 user. Learning a second language also helps with learning to read the first language and with metalinguistic awareness; it develops the areas of the brain responsible for control and delays the onset of Alzheimer's disease (Schweizer, Ware, Fischer, Craik, & Bialystok, 2012). All of these support the proposition that an L2 user is a distinct kind of person from a monolingual.
Recent research has focused on the different ways in which L2 users think. The method is to look at areas where well‐known cognitive differences have been shown to occur between speakers of two languages and then see what happens in L2 users. Typical areas include:
Multicompetence has important implications for language teaching, which has often seen its task as making students as much like native speakers as possible. Multicompetence is now starting to be utilized in books on SLA and language teaching (Cook, 2010).
Overall the development of the multicompetence perspective has been fruitful in suggesting not only new interpretations of existing theories and phenomena but also new research questions to be tackled. It has been linked, inter alia, to the teaching of literacy to Cherokee children (Hirata‐Edds & Peter, 2016), to Hemingway's nickname in Cuba (Herlihy‐Mera, 2013), and to gesture (Brown, 2016). It has gone some way to redressing the bias in favor of the native speaker. But it needs to go further in convincing linguists, psycholinguists, and psychologists of the central importance of the L2 user. The fundamental questions of linguistic competence and language acquisition are changed if most human minds in fact know more than one language.
SEE ALSO: Bilingualism and Cognition; Interlanguage
SIRPA LEPPÄNEN AND SAIJA PEURONEN
The topic of multilingualism and the Internet is important today due to globalization as well as applied linguists' interest in heterogeneous language practices. The Internet is a technosocial system of interconnected networks (Tsatsou, 2014, p. 3) enabling the global dissemination of information and communication between geographically distant individuals and groups. At the same time, for many people— cultural groups, migrants, and political activist groups for example—it functions as a niche medium, providing them with a shared space for participating in transnational and translocal activities (Block, 2004; Leppänen, Pitkänen‐Huhta, Piirainen‐Marsh, Nikula, & Peuronen, 2009; Leppänen, 2012; Kytölä, 2016; Dovchin, Pennycook, & Sultana, 2018). The fact of globalization underlies these processes to contribute to language choice and the use of multilingual and multimodal repertoires, combining ax variety of linguistic and other semiotic resources (Blommaert, 2010, p. 7). At the same time, globalization processes such as the spread of the Internet have also heightened the concerns of some users that the future of the Internet may not, in fact, be multilingual, but monolingual, with English as the dominant language of communication (see, e.g., Crystal, 2001).
This entry refers to multilingualism on the Internet as two interrelated phenomena (see also Androutsopoulos, 2006; Deumert, 2014; Lee, 2017). First, it can mean the choice and diversity of languages as means of communication and analyses of their visibility, accessibility, and status on the Internet (Wright, 2004; Danet & Herring, 2007; Kelly‐Holmes & Atkinson, 2017). In focus here are such language political questions as how much and for what purposes different languages are actually used online, whether or not their distribution reflects linguistic diversity in the world in adequate ways, and whether globally or regionally powerful languages are threatening to replace other (smaller or minority) languages in online settings. Thus, investigations of this type of multilingualism are often concerned with the protection and revitalization of lesser‐used languages on the Internet. In this conceptualization of multilingualism, the Internet is seen as reflecting the language situation of the world offline, as well as a language political agent which has a great deal of power to influence the linguistic preferences and uses of individuals, groups, institutions, and societies. In other words, what happens to languages online is taken to have repercussions on the language situation offline, and vice versa.
Second, multilingualism on the Internet can refer to Internet users' practices of selecting, drawing on, and using more than one language in particular modes and environments of digital communication, whereby they can seek to achieve their situated or sociocultural goals in contextually and communally appropriate and meaningful ways. Importantly, often such uses are not simply an outcome of the Internet users' bilingualism or multilingualism. Rather, they may be a manifestation of their orientation to using all the semiotic resources available to them, including those provided by more than one language—of their metrolingualism (Otsuji & Pennycook, 2010) or polylingualism (Jørgensen, 2008). In many Internet settings this polylingual or metrolingual orientation is also accompanied by intralingual and discursive heteroglossia (Leppänen, 2012), whereby Internet users make use of not only more than one language, but also diverse resources provided by registers, varieties, styles, genres, and intertextuality.
Before launching on a more detailed discussion of these key issues, it is first necessary to define what we mean by the Internet itself. This is because the Internet is, in fact, “an ambiguous term referencing or encompassing innumerable technologies, uses and social spaces” (Markham, 2008, p. 454). It not only refers to the World Wide Web, but also to other information and communication systems outside the Web, such as e‐mail, forum discussions, instant messaging, chat, and Web (phone) call systems.
These different formats of the Internet have made possible a range of synchronous and asynchronous communication modes, some of which are interactive (e.g., discussion forums, different types of chat) while others are more text based (e.g., blogs, fan fiction, microblogging), or image based (e.g., Flickr, Imgur). Furthermore, in many of them (e.g., YouTube, Instagram, Twitter, SnapChat), not only text, but also multimodality—visual effects, pictures, moving image, sound, embodied action, for example—serve as means for sharing information and for engaging in social interaction and cultural production. In particular, digital “social media” is a term emphasizing affordances not just for creating and sharing content but for exchange, participation, and collaboration (Kaplan & Haenlein, 2010; Leppänen, Kytölä, & Westinen, 2017, p. 8). The different formats and technologies establish, in fact, a complex web of media which have become increasingly interconnected and convergent so that one device can, for instance, combine characteristics of several media, for example the computer, television, and telephone. Within this networked mediated space individuals and groups have a great deal of choice as to the kind of communicative space, mode, and style in which they can operate.
At the same time, different Internet settings and communicative modes can constitute quite different environments for language use and require from users the mobilization of sometimes quite distinct register and genre features. Bearing the complexity of the notion of the Internet in mind, it is important that researchers specify what kind of communicative environment they are working on and contextualize their findings in relation to the technical, discursive, social, and cultural parameters of these environments.
As a source of multilingual data, the Internet has meant a range of things to scholars, depending on their respective fields. As in linguistic research on the Internet in general, in the study of multilingualism online a few basic orientations can be identified. First, in some studies the Internet itself is treated as a massive pool of data representative of the language(s) used in it. In such studies, the aim has typically been to discover and describe language facts, for example for the purpose of monitoring language variation, innovation, and change (see, e.g., Crystal, 2001; Davies & Fuchs, 2015). Second, there are studies which have used the Internet as a basis for the compilation of a large and balanced corpus of either particular Internet genres (such as Wikipedia, or blogs, see, e.g., WebCorp, n.d.), or of the whole of the Internet (Kehoe & Gee, 2007; Mair, 2012) for the purpose of exploring general and quantifiable patterns of language(s).
Third, the Internet has also been used as a source of discourse data. In this type of research the data collected is seen to represent particular situated language uses shaped by a number of contextual parameters (such as the participants and goals of the discourse practice). Such studies typically aim for close qualitative analyses ranging from, for example, ethnographic investigations of Internet users' multilingual practices, to analyses of patterns of language use (e.g., code switching, recontextualization, enregisterment), to analyses of the ways in which more than one language is mobilized in the construction of some pragmatic, sociolinguistic, or real world issue (e.g., identity or power), and to studies focusing on the ways in which the Internet itself creates affordances for and constrains particular kinds of language uses (for an overview, see Leppänen & Peuronen, 2012; see also Lee, 2017).
The fact that the Internet is a central and meaningful forum for language use and communication for a growing number of individuals globally poses an important challenge for language studies. More studies are needed to understand and explain the ways in which Internet users select their language of communication, and make use of multilingual resources in their activities, and to discuss the implications of language uses with respect to, for example language policies and language education. Despite the fact that the study of multilingualism on the Internet is a relatively new area of study, it is possible to identify some paradigmatic research orientations within it—these will be briefly discussed next.
Studies of Internet language with a macrosociolinguistic orientation to multilingualism have sought to find out what languages, language varieties, and writing systems are being used online. In this type of research a fundamental premise has been that the quantity, visibility, and accessibility of a language on the Internet are important safeguards of its vitality and importance.
Research of this kind is illustrated by studies which aim to assess the amount and types of linguistic diversity online. For example, a set of studies commissioned by the United Nations Educational, Scientific and Cultural Organization (UNESCO) (Wright, 2004) surveyed attitudes of students of English in 10 countries to find out their language repertoires and which languages they used when engaging in different activities on the Internet (Kelly‐Holmes, 2004, p. 66). In these studies, the role of lesser‐used languages on the Internet was discussed along with considering the possibilities for language revitalization by means of the Internet. In addition, the studies addressed the question whether and in what ways English may pose a threat to other languages used online. The survey indicated that there is a digital and linguistic divide between Internet users in richer and poorer countries, manifesting not only in the different degrees of Internet accessibility (see, e.g., Mafu, 2004), but also in the amount of resources available to provide services in local languages. For example, in developing countries Internet users rely on English‐medium sources whereas users in richer countries are able to access information in their own language (Kelly‐Holmes, 2004).
In another set of studies commissioned by UNESCO, an attempt was made to develop reliable methods for measuring the linguistic diversity on the Internet and, in particular, the actual proportion of the use of English in relation to other languages (Pimienta, 2005; Pimienta, Prado, & Blanco, 2009). On the basis of these studies, it was argued that the linguistic diversity of the world is not adequately represented on the Internet and that the Internet favors large languages (particularly English) and the technical standards associated with these (Paolillo, 2005, pp. 54–5, 79). However, it was also argued (Paolillo, 2005, p. 55) that in reality it is difficult to predict the effect of the Internet on small and minority languages: it may lead to their weakening—as the status of larger languages is consolidated online—or to their strengthening via the visibility, accessibility, and appreciation they may gain when they are used on the Internet (see, e.g., Ferré‐Pavia, Zabaleta, Gutierrez, Fernandez‐Astobiza, & Xamardo, 2018).
Evidence provided by measurements of the linguistic situation on the Internet have often led to attempts to document and advocate the use of the Internet to archive, promote, and teach small, minority, and endangered languages (Warschauer, 1998; Littell, Pine, & Davis, 2017). Such initiatives can be very important in raising awareness and increasing the appreciation of small or endangered languages, but their success often critically depends on whether or not speakers of the lesser‐used languages themselves have access to the Internet and, if they do, whether they are committed to actually using and circulating its resources to maintain and revitalize their language (Mafu, 2004; Sperlich, 2005; Androutsopoulos, 2006, p. 429; Kelly‐Holmes & Atkinson, 2017).
For many people the Internet has become a translocal linguistic contact zone in which multilingual resources and repertoires can turn out to be crucial capital for communication, action, and interaction (Leppänen, 2012; Kytölä, 2016). Along with the “translingual turn” in studies of multilingualism, many scholars investigate, instead of individual languages as separate entities, the variety of resources people draw on and mobilize in different contexts of work, education, and popular culture (Dovchin, Pennycook, & Sultana, 2018, pp. 29–30). This has also resulted in new terminologies which “share a desire to move away from the language of bi‐ or multilingualism” (Pennycook, 2016, p. 201). In particular, by using these new terms, such as transglossia and translanguaging (García & Wei, 2014), translingual practices (Canagarajah, 2018), polylingual languaging (Jørgensen, 2008), metrolingualism (Otsuji & Pennycook, 2010) and superdiversity (Androutsopoulos & Juffermans, 2014), researchers of multilingualism emphasize the need to focus on social practices communicated through repertoires of semiotic resources.
For Internet researchers, the rise of such a conceptualization of multilingualism in the early 2000s has meant a shift from measuring the online presence of global and other languages to the investigation of the Internet as a space for language practices (Androutsopoulos, 2008) which are simultaneously rooted in and voicing the local, as well as responding to and engaging with the global (Leppänen et al., 2009). For example, in many international and local social media environments which focus on topics or interests relating to some aspect of popular culture as a globally shared frame of reference (e.g., football, pop music, films, and TV series) such a translocal orientation is very evident. In this area of research the focus is on the specific multilingual practices, preferences, and policies of Internet users, the motivations behind their language choices, and the functions and meanings their language choices and uses have for them in the specific Internet contexts in which they operate. What such language practices call for is research which can help us understand and explain them as situated and socioculturally meaningful action; research that stems from, for example, the discourse analytic, ethnographic, pragmatic, and microsociolinguistic traditions.
In such investigations of digital discourse, identity and sociocultural practice have often proved to be important explanatory perspectives. Multilingual resources have been shown to be part of the indexical means through which identities and a sense of shared social world can be constructed and negotiated, contested and policed (Leppänen, 2009; Kytölä, 2012; Deumert, 2014). For instance, it has been shown (Androutsopoulos, 2013) how participants can signal their interpersonal relations and alignments through code switching and style shifting between social varieties.
The ways in which Internet users' language practices contribute to the establishment and negotiation of affinity spaces or communities of practice with shared cultural and social conventions have also been investigated in ethnographic studies. In these studies, online activities have been analyzed either alone or complemented by fieldwork in offline settings in which also the participants' Internet uses are examined with the help of, for example, surveys, interviews, and observations (Androutsopoulos, 2008). What is more, because of the omnipresent nature of digital communication media, online and offline spaces are often best viewed as intertwined and not separate realities as such (Dovchin, Pennycook, & Sultana, 2018). Such multidimensional ethnographic investigations can prove particularly useful in helping to explain the ways in which Internet users themselves make sense of and account for their engagement with the Internet and the affordances it offers them for meaningful social and cultural action (e.g., Peuronen, 2017).
When analyzing multilingualism on the Internet, it is important to bear in mind that there is not just one, but several possible language settings in which more than one language may be used. One particular language setting is the kind in which participants have no shared language. Consequently, they need to choose a vehicular language which they can relatively safely assume that the other participants also know and are able to use in communication. For this purpose, English is often a self‐evident choice. For example, it has been shown (e.g., Wodak & Wright, 2006) how in multilingual Internet networks the participants favor English as their shared language because it ensures the widest possible understanding among the participants.
Another typical language setting in which participants in digital communication often operate multilingually is one in which they choose some other language than their L1 as either the primary communicative code or as an additional resource alternating and mixing with their primary language. However, in such a setting, the motivations for their choice of language can vary a great deal. Sometimes the language choice online is an extension of the participants' offline language preferences and uses, but it may also be the outcome of an official language policy or grassroots, group‐specific language policing online (Leppänen & Piirainen‐Marsh, 2009; Karrebæk, Stæhr, Juffermans, & Muhonen, 2017). Such language policing is often evident in not only the participants' particular language choice but also in the ways in which they orient to and monitor choices by others. For example, there is evidence from different countries (see, e.g., Warschauer, Said, & Zohry, 2002; Androutsopoulos, 2006; Siebenhaar, 2006) that in more spoken‐like and dialogic Internet environments, in discussion forums and different types of chat applications, for example, when participants choose to use a language other than their L1 language, vernacular varieties tend to be preferred over standard or formal varieties. In addition, such a language choice can be a sociocultural index; for instance, in Web discussion by Finnish football fans (Kytölä, 2012) their choice of Finnish, English, or Swedish at different points of their discussion can effectively cue their varied stances toward the topic under discussion and their co‐conversationalists.
Sometimes participants in digital communication can draw on resources provided by several languages to such an extent that the outcome is thoroughly heterogeneous discourse. For example, it has been shown how interactional digital discourse (e.g., forum discussions) can be eclectic, creative, and dynamic in the way they adopt, adapt, and appropriate available forms and patterns from more than one language to serve particular communicative purposes and contextualize meanings in their interactional context (Peuronen, 2011; Androutsopoulos, 2013). In a similar vein, studies of writing in Internet genres such as blogs or fan fiction have shown how drawing on resources from several languages is part of a more general heteroglossic repertoire of stylistic, intertextual, and narrative resources which can be used by writers to craft contextually meaningful and socially appropriate texts (Leppänen, 2012).
In the investigation of many informal multilingual Internet genres which often are not only linguistically but also discursively heterogeneous, the concept of heteroglossia can serve as a heuristically useful tool with which to unravel and explain their complexity (Leppänen et al., 2009; Deumert, 2014; Peuronen, 2017). As originally defined by Bakhtin (1981, p. 291; see also Bailey, 2007; Androutsopoulos, 2011), heteroglossia refers to discourse which combines and mixes forms and content that represent “the co‐existence of socio‐ideological contradictions between the present and the past, between differing epochs of the past, between different socio‐ideological groups in the present, between tendencies, schools, circles and so forth.” In focus in heteroglossia are thus not only the co‐occurrence of and shifts between languages or language varieties, but also how language uses index ideological meanings and points of view, including links and tensions between these, and engage with larger social, historical, and ideological processes.
More specifically, the notion of heteroglossia makes it possible to identify different aspects of linguistic heterogeneity. First, it helps to identify and distinguish between the selection and combination of resources from more than one language and intralingualism, the selection and combination of features from more than one variety, register, or style of a language. Second, heteroglossia allows the identification of two types of multilingual heterogeneity. One of these is a mixed in‐group style in which Internet users orient to and make use of the linguistic resources available to them, no matter which language they may technically be associated with (see also Jørgensen, 2008; Otsuji & Pennycook, 2010). Another aspect of multilingual heterogeneity is code switching—one of the semiotic resources available to Internet users to contextualize, highlight, or to make relevant possibilities for the interpretation of discourse and reality within a discourse entity. Finally, heteroglossia can also entail a discursive dimension: the selection, combination, and juxtaposition of discursive resources, for example, aspects of genres, modalities, and intertextuality (Leppänen, 2009; Leppänen & Piirainen‐Marsh, 2009) within a discourse entity. For example, in blogging or fan fiction discursive heteroglossia can play a key role in writers' efforts to create the kind of generically and intertextually complex and multilayered discourse that is prized by their respective communities of practice.
Despite the doubts that many parents and educators have about the usefulness and value of young people's online activities, the fact that many of them operate on the Internet multilingually can have a great deal of potential in terms of first, second, and foreign‐language learning. One reason for this is that for many young people Internet‐based peer‐focused communication can provide stimulating opportunities for informal learning—possibilities for the recognition, application, and practice of patterns of language and discourse, aspects of new literacy, and forms of interaction for purposes of authentic communication, social life, and cultural production (e.g., Yaman Ntelioglou, Fannin, Montanera, & Cummins, 2014). In addition, in many Internet activities which involve planning and design—such as fan fiction writing—participants have developed resources and services which are in principle instructional, teaching novice participants, through feedback, commentary, and explicit sets of instructional materials, appropriate ways of language use in the particular Internet genres their writing inhabits (Leppänen, Kytölä, & Westinen, 2017).
In recent years, a great deal of research interest has been shown in the pedagogic potential of young people's Internet‐based activities (see, e.g., Lam & Rosario‐Ramos, 2009; Prinsloo & Stroud, 2014) and pedagogic experimentation relating to the uses of technology in second and foreign‐language classrooms (for an overview see, e.g., Thorne, Black, & Sykes, 2009; Guth & Helm, 2010). These studies have argued strongly for the necessity of a renewed vision of educational practice that would harness the potential of communicative technologies in language instruction and suggested ways in which language education can be both enhanced and reformed by taking seriously the significance and role of multilingual and multimodal digital communication in the contemporary world (Barton & Lee, 2013).
However, while the Internet, or more generally, digital information and communication technologies, can offer new incentives, opportunities, and activities for language learning, it may be that for many young people the integration of their out‐of‐school digital activities within school‐based language instruction can be problematic. One reason for this is that their Internet activities are part of their own lifeworld, and a forum of social life outside the control and agendas of the adult world, parents, and teachers. As a result, teachers wishing to make use of young people's online interests and activities in language teaching may sometimes face the resentment and reluctance on the part of their students. On the other hand, the recognition of the linguistic and discursive expertise that students may have and its sensitive use as a means to scaffold opportunities for further language learning in a way that does not intrude too much into young people's own cultures on the Internet, can be an effective strategy to empower students and enhance their language learning.
This entry aimed at outlining how multilingualism on the Internet can mean a variety of things, depending on how the notion of the Internet is conceptualized, which genres and modes of communication (e.g., edited webpage contents vs. interactive modes) or language settings are focused on, and the disciplinary perspective from which their analysis is approached. It also aimed at illustrating the emergent paradigms of current research on multilingualism on the Internet. One of these was labeled the macrosociolinguistic approach, and it was shown to be concerned with assessing the degree of linguistic diversity online, evaluating the threat posed by English as a popular vehicular language to other (smaller, minority) languages in the world, and using the Internet as an instrument and forum in language maintenance and revitalization. Another central research paradigm was shown to focus on the study of the multilingual discourse practices of Internet users in different modes of communication, drawing on concepts and microanalytic methods provided by, for example, discourse studies, pragmatics, interactional sociolinguistics, linguistic anthropology, and ethnography. The entry also described some of the crucial features of multilingual Internet discourse, and suggested that often it can usefully be anatomized with the help of the notion of heteroglossia which allows the identification of not only of different types of linguistic heterogeneity, but also of discursive heterogeneity.
The present overview of this emergent and multifaceted field of the study of multilingualism on the Internet is necessarily cursory, and many crucial issues could only be touched upon here. However, it needs to be emphasized that the multidimensionality of Internet multilingualism, which is now beginning to be recognized in research, profits from investigations which approach it from a variety of perspectives and methodologies. These can help shed light on not only the specificities of but also the interconnections between language uses online and offline, and on the functions and meanings of these language uses in the varied, but increasingly convergent linguistic, communicative, and discursive spaces of the Internet. Understanding the complex ways in which multilingualism works on the Internet, how and why Internet users make use of more than one language, is particularly important for applied language study. Research on Internet users' language practices can help us understand how language choice and linguistic (and discursive) heterogeneity are drawn on as semiotic and sociocultural resources. Understanding such motivations, goals, forms, and meanings of multilingualism on the Internet can thus help us define and formulate, for example, language policies when such policies are needed to protect and revitalize languages under threat and educational visions that can profitably build on and develop people's multilingual practices and that in turn can lead to learning, empowerment, and agency.
SEE ALSO: Lingua Franca and Language of Wider Communication; Multicompetence
CRISTINA SANZ
Metalinguistic awareness is only one of the cognitive advantages resulting from living with two languages. The study of the effects of bilingualism on cognition has also included nonverbal tasks, such as mathematical and numerical (Bialystok & Codd, 1997), creativity (Kessler & Quinn, 1980), recognition memory (Koster & Cadierno, 2018), and problem‐solving tasks (Bialystok & Majumder, 1998; Bialystok, 1999; Bialystok & Martin, 2004), with findings supporting those from verbal tasks (Kessler & Quinn, 1980, 1987) showing that language experience may play an important role in the development of abstract thought (Rosenblum & Pinker, 1983) and semantic creativity. All these studies compared monolinguals and bilinguals completing tasks that varied in their degree of difficulty, with the most demanding tasks requiring that participants ignore irrelevant information that would interfere with the actual task if given attention. These results support previous research (Bialystok, 1988) showing that bilinguals are better than monolinguals at solving problems that are based on conflict and attention. A few recent studies looked at whether the bilingual cognitive benefits could extend to bidialectal populations. When comparing bidialectal children's performance in cognitive control tasks with monolingual (monodialectal) controls, no bidialectal advantage was found (Ross & Melinger, 2016), while, in another study, Swabian‐dominant bidialectal speakers outperformed balanced German–Swabian bidialectal speakers in inhibitory control tasks (Poarch, Vanhove, & Berthele, 2018).
Although the majority of studies have looked at child bilingualism, research has also indicated that some of the results found in children can be replicated with adults, and that, in fact, bilingualism attenuates the effect of aging on cognitive control (Bialystok, Craik, Klein, & Viswanathan, 2004; Kefi et al., 2004; Bialystok, Craik, & Luk, 2012). In particular, Zirnstein, Hell, and Kroll (2018) found that, compared with their monolingual counterparts, older adult bilinguals are more capable of generating predictions during online reading comprehension, which is a resource‐demanding processing behavior that appears to rely heavily on executive functions that are known to decline with age (Brickman et al., 2005). However, some studies showed that the positive effects associated with bilingualism may decrease in young adults, who are at the peak of their attentional abilities, when compared to children or older adults (Bialystok, Martin, & Viswanathan, 2005; Bialystok, Craik, & Ruocco, 2006; Bialystok, Craik, & Ryan, 2006).
More accurate descriptions of how bilingualism affects cognitive performance, according to Bialystok (2007b), have led to the interpretation that bilingualism must be considered in terms of both advantages and disadvantages. This has been investigated in studies of cognitive and linguistic processing in adults and in children (see the section on “Metalinguistic Awareness: Learning to Read, Learning other Languages”). For adult bilinguals, research has shown that they are at a disadvantage on tasks measuring lexical retrieval and fluency (Michael & Gollan, 2005) but that they are better at controlling attention when performing a challenging task (Bialystok et al., 2004). For bilingual children, different mechanisms are involved in linguistic and cognitive outcomes of bilingualism: While level of proficiency is related to performance on metalinguistic tasks, executive control is predicted by length of time in the immersion program (Bialystok & Barac, 2012). It was also revealed that changes in children's metalinguistic concepts associated with bilingualism emerge gradually, over a period of several years (Bialystok, Peets, & Moreno, 2014). However, the lack of conclusive results in some of the areas investigated suggests that further research is still needed to understand the complex nature of the relationship between bilingualism and cognition, including metalinguistic awareness. Given that the majority of the research has been carried out with bilingual children as subjects, research must further expand on the recent findings that have examined the consequences of adult bilingualism, especially in young adults, an age group where the bilingual advantage seems to be less obvious.
The language‐teaching profession considers that the body of empirical studies showing that second language learning fosters cognitive development and contributes to stronger native language skills, especially in writing, is significant enough that they can use it as a convincing argument in favor of the strengthening of foreign languages in school and college curricula (Kecskes & Papp, 2000). Foreign‐language learning, even at an intermediate level, improves critical thinking and context‐dependent processing of language material, enhances the development of interpersonal relations, and teaches self‐reflection on language behavior (Sanders, 2006).
Research investigating cognitive advantages for bilingual children has revealed that bilingualism helps develop higher levels of mental flexibility and metalinguistic awareness at an earlier age. Children exhibit bilingual awareness after their second birthday through self‐repair, other‐repair, translation, and labeling of languages (Mishina‐Mori, 2004). Two‐year‐olds can make corrections, criticisms, or initiate language games (Moreno‐Zazo, 1998). Bialystok (2001) summarizes several studies that investigated the effects of bilingualism on word awareness, syntactic awareness, and phonological awareness. In word awareness, a number of studies agree on the advantage of bilinguals over monolinguals in their ability to separate word and meaning (Ben‐Zeev, 1977; Bialystok, 1988; Ricciardelli, 1992). However, counterevidence also exists, as shown in a study by Rosenblum and Pinker (1983), whose monolingual and bilingual preschoolers were equally capable of associating objects with new arbitrary names. Nevertheless, monolinguals and bilinguals varied in the way they referred to the properties of an object. Whereas monolinguals were more likely to talk about the physical attributes, bilinguals emphasized the social context in which objects could be used, thus revealing bilinguals as more capable of extrapolating the meaning of a word and connecting it to the real world.
In the field of syntactic awareness, research has observed an advantage for bilingual children over monolinguals when performing tasks that require noting and correcting errors (Galambos & Goldin‐Meadow, 1990) or avoiding misleading information (Bialystok, 1986, 1988). Galambos and Goldin‐Meadow argue that bilingual children have a greater awareness of linguistic forms than monolingual children, and that metalinguistic awareness improves with intellectual development. However, studies dealing with morphosyntactic structures have also shown that there are areas in which bilinguals do not outperform monolinguals, such as the acquisition of the mass/count distinction in English (Gathercole, 1997). This study examined whether bilingual children differ from monolingual in the way they draw inferences about the referents of new nouns based on mass/count structures such as an X and some X. The results showed that young monolinguals (seven years old) outperformed bilinguals of the same age. However, at nine years old, bilinguals seemed to have achieved the same levels of performance as their monolingual counterparts. Gathercole (1997) provided a number of possible explanations for these results; the one that seemed more plausible is that bilinguals may just need more time to acquire this specific feature. The differences in results between this and other studies in the area may also lie in the nature of the task. In this case, either an elicitation task (Gathercole, 1997) or a grammaticality judgment task that requires the inhibition of misleading information (Bialystok, 1986, 1988; Galambos & Goldin‐Meadow, 1990) were used. In addition, this study suggests that the difficulty of the task proposed may play an important role on the effects of bilingualism on cognition depending upon the age and developmental stage of the participants.
In phonological awareness, positive findings for bilingual children are revealed in aspects such as the ability to judge whether objects in pictures had long or short names (Yelland, Pollard, & Mercuri, 1993). However, Bialystok, Majumder, and Martin (2003) did not find any difference in performance between bilingual and monolingual children who were requested to substitute a sound in the target word with another sound from another word.
According to Ianco‐Worrall (1972), bilingual children appear to develop semantic understanding before monolingual children, although this difference diminishes over time. Ben‐Zeev (1977) concludes that bilingual children are better prepared to reorganize their perceptions than their monolingual counterparts, and that bilingual children are prepared to treat words as units separate from their semantic content within a greater syntactic system.
In one of the few studies that have looked at bilingual pragmatic awareness, Safont Jordà, (2003) compared monolingual (Castilian) and bilingual (Catalan and Castilian) female learners of English in a discourse completion test, a role‐play task, and a discourse evaluation test. Results showed advantage of bilinguals over monolinguals in justifying their evaluation on the appropriateness of certain requests to particular contexts, as well as on their use of request realizations.
To resolve some of the contradictory findings from studies investigating the relationship between child bilingualism and cognition and specifically to explain some of the negative findings that had been obtained in research with minority and migrant bilingual children in subtractive contexts (i.e., where two languages are competing), Cummins developed a new framework (1976), the threshold hypothesis. Cummins related cognitive development in children, internal in nature, to the attainment of proficiency thresholds in their L1 and L2, and this to a combination of external (i.e., social, attitudinal, educational) factors. Under the model, bilinguals are expected to develop cognitive advantages, including metalinguistic awareness, when compared to monolinguals, only after a certain threshold is reached in their two languages.
Research has in fact provided evidence that higher levels of bilingualism are associated with cognitive advantages when compared to lower levels of bilingualism not only in children (Bialystok, 1988; Ricciardelli, 1992; Bialystok & Majumder, 1998) but also in adults (Thomas, 1988; Kefi et al., 2004). For children, studies such as Yelland et al. (1993) found that being partially proficient in a second language confers participants with a cognitive advantage when compared to monolinguals in tasks that involved metalinguistic awareness. Yelland et al.'s children receiving partial instruction in Italian as second language were better than monolinguals at judging whether a number of pictures contained long or short names. Bialystok's (1988) bilingual groups outperformed the monolingual group on an arbitrariness of language task, while the fully bilingual group performed significantly better than the partial‐bilingual and the monolingual group on the concept of word task. Demont (2001) shows an advantage for children in German/French bilingual classes at grammatical judgment and correction tasks and at word recognition, while Dillon's comparison (2009) of children in both early Irish immersion schools and English‐medium schools where Irish is taught as L2, concluded that more balanced bilinguals are more likely to independently display higher levels of metalinguistic awareness and evidence of positive crosslinguistic transfer. Parallel conclusions were drawn by Kenner, Gregory, Ruby, and Al‐Azami (2008) for second and third generation Bengali heritage language learners in the UK. In conclusion then, evidence suggests that, for children and adults, the higher the level of bilingualism the more cognitive advantages, but when a partial level of bilingualism is reached benefits also appear when compared to monolinguals.
Studies with adult participants, like that of Thomas (1988), show comparable results, in that with more linguistic experience comes greater metalinguistic awareness. According to Thomas, instruction in a second language aids in the development of sensitivity to grammar. More recent research by Ransdell, Barbier, and Niit (2006) compares working memory capacity and metalinguistic awareness as it applies to reading in monolingual, bilingual, and multilingual university students within three cultural contexts, America, Estonia, and France. Their results show that bilingual and multilingual students have better metalinguistic awareness of their language skills in reading and greater working memory capacity than do students who are monolingual, but who have comparable native language skills. Gibson and Hufeisen (2006) also concluded that multilingual learners of English are highly metalinguistically aware and adept at sifting formal grammatical information from semantic information, both in listening and written tasks. More research is needed that investigates where to place the threshold for benefits to show, especially since these thresholds do not appear to be constant and seem to vary with the nature of the task; that is, different levels of bilingualism will lead to different levels of metalinguistc awareness and as task complexity varies, they require different levels of metalinguistc awareness.
Some scholars attribute the cognitive benefits of bilingualism not as much to superior oral proficiency but to literacy in two languages (Cummins, 1981; Swain, Lapkin, Rowen, & Hart, 1990; Cook, 1997; Bialystok, 2001). Sanz (2007) found no effects for balance in oral skills on the acquisition of an L3, showing instead that level of biliteracy is a predictor of success. Furthermore, her study identifies a specific level of development of literacy in both languages that needs to be reached before we can observe enhanced ability to learn the L3; this level is possible in socioeducational situations that promote additive bilingualism. More recently, Rauch, Naumann, and Jude (2012) found that students who were fully biliterate in German and Turkish outperformed monolingual and partially biliterate students in both L3 English reading proficiency and metalinguistic awareness. Taken together, results of these studies suggest that the crucial factor in successful L3 acquisition is the development of heritage language literacy skills, rather than exclusively oral skills.
A second explanation to account for contradicting behavior elicited by metalinguistic tasks was proposed by Bialystok (1999, 2001). Her framework describes to what degree task characteristics determine a bilingual's performance depending on the involvement of two processing components: analysis and control. Whereas analysis is defined as the ability to represent explicit and abstract rules, control is described as the ability to attend to specific aspects of the input, especially in situations that include ambiguous information. In order to test how bilingualism interacts with these external conditions, the task is manipulated such that demands of analysis and control are increased. Reexamining the results in this way reveals that bilingual advantages occur reliably on tasks that make high demands on control but are not evident in tasks that make high demands on analysis.
Bilinguals' greater attention to language form to determine constantly which language to use and how to switch between languages develops the component of control which serves as a guide during this linguistic process. Also, reading decontextualized language in comparison to oral exercises demand a greater understanding of the grammatical structure, making easier the development of the analysis component. Biliteracy is key in developing the analysis component.
A number of factors involved in language acquisition may be enhanced as a result of the language‐learning experience of bilinguals, including cognitive flexibility, automatization of skills, and metalinguistic awareness (Cummins, 1976; Cromdal, 1999; Bialystok, 2001; Gibson & Hufeisen, 2006). Such factors have been posited to explain the advantage shown by bilinguals over monolinguals in the acquisition of a third language (L3) both in classroom (Swain et al., 1990; Cenoz & Valencia, 1994; Sanz, 2000) and in laboratory studies (Nation & McLaughlin, 1986; McLaughlin & Nayak, 1989; McLaughlin, 1990; Nayak, Hansen, Krueger, & McLaughlin, 1990). Classroom studies conclude that instruction in the minority language is associated with improved competence in an L3 at no cost either to competence in the majority language or to academic development. In Sanz (2000), Spanish‐Catalan biliterates were shown to have an advantage over Spanish monolinguals as learners of English as an L3. Sanz concludes that, in general, bilinguals are more efficient language learners because they are faster and consequently achieve a higher level of attainment in the L3. But what gives multilinguals the edge?
From a cognitive psychology perspective, bilinguals are experienced language learners and users and experience allows them to approach the task of learning a new language in a different, more efficient way. This cognitive explanation is exemplified by Klein (1995), whose multilinguals showed heightened ability to identify and retain key verbs necessary to trigger development of the key grammatical structure to be acquired. From a second language research perspective, metalinguistic awareness works as an advanced organizer for incoming linguistic stimuli, giving learners the capacity to focus on form and to pay attention to the relevant features in the input, thereby increasing the amount of processed linguistic information that feeds into their internal grammar.
In an effort to determine whether bilinguals at different stages of L2 development make different use of information about how the language works (i.e., explicit information) during the acquisition of a third language, Sanz, Stafford, and Bowden developed the Latin Project, which has produced a number of publications. Lado, Bowden, Stafford, and Sanz (2017) compares performance by emerging bilinguals at different levels of Spanish proficiency during the initial stages of development of an L3; their two groups differ on the type of feedback they received (with or without grammar). Results show that level of bilingualism is closely related to L3 proficiency level (e.g., Cenoz, 2003) and that even partial bilingualism enhances subsequent language learning (e.g., Thomas, 1988). Importantly, as hypothesized, the study identifies two different thresholds for the positive effects on cognition of adult bilingualism, with their positions changing depending on external conditions: When there is no grammatical information, only one threshold emerges and is delimited by near‐native proficiency.
To sum up, experience with more than one language—multilingualism—promotes the automation of the basic necessary abilities in the processing of information and language. As they become automatic, these basic necessary abilities depend less on working memory and allow room for other processes. Also, the development of input restructuring strategies is a result of metalinguistic analysis. Language learners restructure input in categories that allow for quicker access and more transparent relationships with other units of information, making the processing of new input definitively easier. In this way, multilingualism, metalinguistic awareness, and attentional control are interconnected and explain multilinguals' advantage in the acquisition of new languages.
The research on literacy agrees there are three prerequisite skills for the acquisition of literacy: establishment of metalinguistic awareness, competence with the oral language, and understanding of symbolic concepts of print (Bialystok, 2007a). Metalinguistic awareness, especially morphological awareness, has been reported to play a key role in the development of L1 reading comprehension among Chinese monolingual children (Li & Wu, 2015). Research on bilingualism and literacy explores the extent to which these skills differ for monolingual and bilingual children. Its findings suggest that bilingualism is clearly a factor in children's development of literacy, but the effect of that factor is neither simple nor unitary. Sometimes bilinguals are at an advantage (concepts of print), sometimes at a disadvantage (oral language competence). Likewise, Francis (1999) concludes that metalinguistic awareness is related to difficult aspects of literacy development in different ways, the key variables being the degree of decontextualization and expressive versus receptive language tasks. More recent theories of L2 literacy argue that reading subskills, like metalinguistic awareness, can be transferred across languages, as predicted by Koda's (2005, 2008) transfer facilitation model, confirmed by a small but growing body of research. Ramirez, Chen, Geva, and Kiefer (2010) found that young Spanish‐speaking English learners' knowledge of derivational affixes and morphological production ability in Spanish uniquely predicted their English word reading. Similarly, Zhang, Koda, and Sun (2014) observed compound awareness in L1 Chinese had a facilitating factor for reading comprehension in both Chinese and English among fifth and sixth graders in China. Results from these studies lend credit to transfer of morphological awareness (knowledge of word structure) in literacy development.
SEE ALSO: African Union; Aptitude in Second Language Acquisition; Early Bilingualism; Explicit Knowledge and Grammar Explanation in Second Language Instruction; Instructed Second Language Acquisition; Multicompetence; Teaching Reading: Foundations and Practices; Third Language Acquisition
RODNEY H. JONES
Multimodal discourse analysis is an approach to discourse which focuses on how meaning is made through the use of multiple modes of communication as opposed to just language. Although there is a long tradition in semiotics, anthropology, sociology, and cultural studies of engaging analytically with visual texts and other modes, it is only relatively recently that this challenge has been taken up by applied linguists, who, not surprisingly, have traditionally privileged the modes of speech and writing. One reason for this recent surge of interest in multimodality is the increasing importance of multimodal representation in everyday life with the rise of digital technologies. Multimodal texts, however, are far from new. In fact, before the development of the printing press, most texts were “illuminated” in some way, with considerable attention paid to things like illustrations and calligraphy, and face‐to‐face interaction has always depended on a rich array of modes and materialities. Another reason for the development of multimodal discourse analysis is the concurrent development of technologies—including digital video cameras and computer editing and analysis software—which have made this kind of analysis easier.
It should be stressed at the outset that multimodal discourse analysis is not the study of a special kind of discourse, that is, “multimodal” discourse. Rather, it is the engagement in the multimodal dimension of all discourse. As Scollon and Levine (2004, pp. 1–2) point out,
language in use, whether this is in the form of spoken language or text, is always and inevitably constructed across multiple modes of communication, including speech and gesture not just in spoken language but through such “contextual” phenomena as the use of the physical spaces in which we carry out our discursive actions or the design, paper, and typography of the documents within which our texts are presented.
The concept of “mode” in multimodal discourse analysis should not be confused with the concept of “modality” in grammar (which refers to the expression of possibility and obligation in language), or with the concept of “mode” in Halliday's tripartite model of context (consisting of field, tenor, and mode), in which mode refers to the “channel” of linguistic communication (i.e., speech or writing). When multimodal discourse analysts speak of mode they usually mean a “regularized, organized set of resources for meaning‐making, including, image, gaze, gesture, movement, music, speech and sound‐effect” (Jewitt & Kress, 2003, p. 1). Linguists who take different approaches to multimodal discourse analysis (see below), however, emphasize different aspects of this definition. Social semioticians, for example, tend to regard modes primarily as resources for meaning making, while those taking a more systemic approach see them as organized systems of “meaning potential” (Halliday, 1994), and conversation analysts and interactional sociolinguists see them chiefly as equipment for managing interactions. Some multimodal discourse analysts have been critical of viewing modes as reified, bounded semiotic systems with stable characteristics, especially in situations where the development of new media results in the blending of modes (as computer technology has done in the case of speech and writing; see Goddard, 2004). Norris (2004, p. 11) argues that modes are best thought of as heuristics rather than as stable categories or systems. While it is unavoidable for analysts to construe modes as distinct entities, she notes, they are in reality a loose “grouping of signs that have acquired meaning in our historical development.”
Modes are also clearly distinguished from media, which are the physical or material carriers of modes (e.g., paper, telephones, televisions, computer screens, and the like). For example, while dress would be considered a mode (a system governing the combination of various semiotic elements involving color, shape, texture, and their placement on a human body), clothing is a medium, the physical means through which this system is expressed. At the same time, there is an intimate and complex relationship between modes and media: the functions and significance of various modes evolving in relation to the development of technologies and the social practices governing their use in particular sociohistorical contexts (Kress & van Leeuwen, 1996). As Bateman, Wildfeuer, and Hiippala (2017) argue, the materiality of different media are shaped for different communicative purposes, and so act as incubators for the development of new semiotic modes and new ways of combining them.
In its early years in the late 1990s and early 2000s, multimodal discourse analysis focused on two broad areas of research—the study of multimodality in texts like magazines, comic books, films, works of art, and Internet Web pages, and the study of multimodality in interaction—and each of these foci drew on rather different traditions in linguistic and social science research. In more recent research, however, especially involving digital media where Web sites and social media interaction involve the exchange of things like images and videos, and in classroom research in which the manipulation of multimodal textual objects plays an important role in teaching and learning, this neat distinction has become blurred.
Early work on the analysis of multimodality in texts tended to draw from either the semiotic traditions of scholars like Barthes (1978) or from methods developed in film and design studies. A turning point came in the late 1970s, when the critical linguistics group at the University of East Anglia developed what they called social semiotics, drawing on the work of the British linguist M. A. K. Halliday (1978, 1994). Halliday's systemic functional linguistics (SFL) represented a significant departure from previous approaches to language which focused mostly on syntax and semantics, viewing it instead as a resource for making meaning based on the needs that had evolved in the social world. Disciples of Halliday reasoned that if language could be analyzed as a “social semiotic,” other forms of communication such as images could also be subjected to a similar kind of analysis. Perhaps the most influential proponents of this approach were Gunther Kress and Theo van Leeuwen, who, in their 1996 classic, Reading Images: The Grammar of Visual Design, proposed that images can be analyzed by applying conceptual tools from functional grammar. This project was later taken up by others and applied to a range of other types of texts including paintings, sculpture, and architecture (O'Toole, 1994), mathematical discourse (O'Halloran, 2005), school textbooks (Guo, 2004), and Internet Web pages (Martinec, 2005).
The influence of Halliday on multimodal discourse analysis gave rise to a range of different approaches to the analysis of multimodal texts and interactions. Some scholars, such as Kress, Jewitt, and Bezemer (Jewitt & Kress, 2003; Bezemer & Kress, 2015), focused more on the social semiotic aspect of Halliday's work, exploring how different modal resources are used by particular groups of people in particular contexts, such as education (Jewitt, 2006) and medicine (Bezemer, Murtagh, Cope, Kress, & Kneebone, 2011). Others, such as O'Halloran (2004a), Royce and Bowcher (2006), and Unsworth (2008), focused more on the systems through which multimodal meaning is produced, applying Halliday's lexicogrammatical ranks and categories to the analysis of multimodal texts such as museum exhibitions (Meng, 2004), films (O'Halloran, 2004b; Tseng, 2008), and the use of space in classrooms (Royce 2002; Jones, 2008). Still others attempted to develop ways of combining concepts from systemic functional grammar with corpus tools, the most notable work in this area being that of Baldry and Thibault (Baldry & Thibault, 2005; Baldry, 2007). Finally, there are scholars such as Machin, Mayr, and van Leeuwen (Machin & van Leeuwen, 2007; Machin & Mayr, 2012) who have championed a multimodal critical discourse analysis, which seeks to apply perspectives from critical discourse analysis, itself heavily influenced by Halliday's view of language (Fairclough, 1992), to multimodal texts with the aim of uncovering various ideological and political agendas.
Some scholars, while influenced by Halliday's work, developed methods of analysis that sharply diverged from both social semiotics and systemic functional analysis. Among the most notable has been the Genre and Multimodality framework (sometimes abbreviated GeM), developed chiefly by John Bateman (2008). Bateman, Delin, and Henschel (2002) argue that the application of analytical frameworks such as SFL, developed for the analysis of temporally organized modes like writing, to the analysis of visual layout is problematic, especially without sufficient empirical evidence regarding how people actually process visual information. The GeM framework makes use of the notion of genre to account for variation in document structure, using an annotation schema with multiple layers of description designed to be applied to large corpora of documents so that significant patterns can be empirically identified.
Another important tradition in multimodal discourse analysis has focused more on the analysis of face‐to‐face interaction, drawing its inspiration from work in sociology, anthropology, and psychiatry. The foundation for much of this work can be traced back to the “ecological perspective” on communication developed by Scheflen (1974), Birdwhistle (1970), Bateson (Ruesch & Bateson, 1951/1968), and others working in the Palo Alto research group in the 1950s; the work of sociologists such as Goffman and Garfinkel; and development of discourse analytical approaches to social interaction such as conversation analysis and interactional sociolinguistics. The focus of approaches in this tradition is not just on how modes work together at particular moments to create meaning, but also on how modes are deployed sequentially to manage the ongoing flow of interaction.
Among the most influential proponents of this approach is Adam Kendon (1990), whose studies of the interactional functions of gaze and facial expression and of the spatial organization of naturally occurring interaction combine Bateson's “natural history” approach with Goffman's microsociological perspective, and Charles Goodwin (2000), whose extension of principles from conversation analysis (CA) to the study of multimodal interaction has heavily influenced scholars such as Mondada (2007, 2011) and Nevile (2004), who have applied CA frameworks to the analysis of interaction in settings as diverse as operating theaters and airplane cockpits. Another strand in this tradition draws on principles from interactional sociolinguistics and mediated discourse analysis, analyzing the use of modes as tools to perform social actions and enact social identities. Perhaps the best exemplar of this approach is Sigrid Norris, whose book Analyzing Multimodal Interaction (2004) lays out a systematic framework for analyzing how, in face‐to‐face interaction, people negotiate their simultaneous participation in multiple actions and enactment of multiple identities though the strategic deployment of different modes.
There has been other work that combines an interest in multimodal texts with a concern with how those texts are shaped and reshaped in the context of ongoing interaction. Of particular note is the work of Rick Iedema (2001), who takes the position that multimodal analysis must be complemented by a more dynamic perspective on semiosis which examines how social interaction helps to shape the ongoing emergence of multimodal meaning. Iedema's main contribution to the field of multimodal discourse analysis is his concept of resemiotization, which he defines as the dynamic process through which meaning changes as it becomes instantiated in various configurations of modes and media along the trajectories of various social practices.
Approaches which combine attention to the composition of multimodal texts and the sequential deployment of modes in interaction have become increasingly important with the development of digital media platforms through which users are able both to interact with other people in real time, and to interact with texts. Adami (2015), for example, applies a social semiotic approach to the analysis of interactivity in digital texts, considering how features such as hypertext and drop‐down menus introduce both aesthetic and functional potentials. Jovanovic and van Leeuwen (2018) analyze the exchange structure and generic features of interactions on multimodal social media sites such as Instagram and WhatsApp, noting how users work within the constraints imposed by the apps and exploit the predesigned visual templates they provide to engage in evaluative and emotive communication. And Albawardi (2017) applies the Scollons' theory of “geosemiotics” (Scollon & Scollon, 2003) to explore how users of Snapchat use space and place as a resource for making meaning.
More recent trends in multimodal discourse analysis are pushing scholars to consider more fully the complex ways different modes interact in communication beyond conventional understandings of composition and the organization of interaction, including how meanings change as they move across modes, what Kress (2003) calls “transduction,” and the way the deployment of one mode can sometimes result in sensations associated with other modes (sounds, for instance, can be perceived to have color, or smells can be perceived to have shape), a phenomenon known as “synaesthesia” (Nelson, 2006; van Leeuwen, 2016). Scholars are also being pushed to consider a range of previously neglected modes of communication such as touch and smell, both in the context of face‐to‐face interaction (Norris, 2013; Bezemer & Kress, 2014) and mediated through new technologies like VR (virtual reality) glasses and wearable computers, moving multimodal analysis beyond the audiovisual to the realm of multisensory experiences (Ensslin, 2018).
SEE ALSO: Multimodal Interaction Analysis; Multimodality and Systemic Functional Analysis
SIGRID NORRIS
Multimodal interaction analysis (Norris, 2004, 2011) is a holistic methodological framework that allows the analyst to integrate the verbal with the nonverbal, and to integrate these with material objects and the environment as they are being used by individuals acting and interacting in the world. In short, multimodal interaction analysis allows a researcher to study real people interacting with others, with technology, and with the environment. Multimodal interaction analysis, to give a few and diverse examples, has been utilized to study topics such as identity production during video‐conferencing interactions and acquisition of interactive aptitudes (Geenen, 2017, 2018), ethnic identity (Matelau, 2014), agency and intersubjectivity (Pirini, 2016, 2017), car talk (Norris, 2017), identity production through objects (Norris & Makboon, 2015), communicating knowledge (Norris & Pirini, 2017), or academic lectures (Bernad‐Mechó, 2017).
The point of view taken in multimodal interaction analysis is that all actions in fact are interactions and that all of these (inter)actions are linked to people (referred to as “social actors” in multimodal interaction analysis)—no matter whether the researcher is investigating the real or the virtual world, someone buying ice cream, or interaction with a computer program. When thinking about buying ice cream, it is obvious that a social actor is buying and another social actor is selling the ice cream. It is also obvious that a social actor (or more likely a whole number of social actors) made the ice cream using ingredients, tools, and machines. When looking at software use, we look at it in the same way: a social actor is using it; but a social actor (or rather, again, a number of social actors) produced the software, embedded fonts, color schemes, and the like. Social actors and cultural tools are the key elements in multimodal interaction analysis. Multimodal interaction analysis emphasizes that all social actors act by using cultural tools—from language and gestures to coffee makers and computers. On the one hand, multimodal interaction analysis (following mediated discourse analysis) argues that social actors cannot act without using some cultural tools; and on the other hand, multimodal interaction analysis claims that cultural tools do not exist without social actors producing, using, and re‐producing them. Because of these two interlinked claims, there is a constant tension in (inter)action between social actors and the cultural tools that they are using.
The framework primarily grew out of mediated discourse analysis (Scollon, 1998, 2001) in connection with interactional sociolinguistics (Goffman, 1959, 1963, 1974; Gumperz, 1982; Tannen, 1984) and multimodal analysis (Van Leeuwen, 1999; Kress & Van Leeuwen, 2001). Multimodal interaction analysis allows the integration of all communicative modes, where communicative modes are defined as systems of mediated actions (Norris, 2019). In this view, all modes of communication together build one system of communication, which comes about through actions and interactions that people produce. Strongly building upon Scollon's (1998, 2001) mediated discourse theory, multimodal mediated theory has been developed to underpin multimodal interaction analysis (Norris, 2019).
In multimodal interaction analysis, as in mediated discourse analysis, the mediated action is the unit of analysis. A mediated action is defined as a social actor acting with or through mediational means or cultural tools (Scollon, 1998; Wertsch, 1998). Within the mediated action we find the constant tension alluded to above between social actor and cultural tools. Certainly, not all mediated actions are of the same kind; the three general demarcations are as follows.
First are the lower‐level mediated actions, which are the smallest pragmatic meaning units of communicative modes, such as an utterance for spoken language, a gesture unit for the mode of gesture, or a postural shift for the mode of posture. These lower‐level meaning units are of interest to multimodal interaction analysis as they allow for microanalyses.
Second, we investigate higher‐level mediated actions, which are made up of an abundance of chains of lower‐level mediated actions; for example, a conversation is made up of chains of utterances, chains of gestures, chains of postural shifts, and so on. While a higher‐level mediated action such as a conversation is, on the one hand, produced through these chains of lower‐level mediated actions of utterances, gestures, or postural shifts, the lower‐level mediated actions, on the other hand, only come about because the higher‐level mediated action is being produced. Thus lower‐level and higher‐level mediated actions cocreate each other. While there is a bottom‐up and simultaneously a top‐down development in (inter)action between the various levels of action, it is, however, usually the higher‐level mediated action (such as a conversation) that social actors actively produce.
Third, we investigate frozen mediated actions, which are actions entailed in objects and the environment. For example, the mediated action of building is frozen in a house, or the mediated action of making coffee is entailed in the cup of coffee itself, just as the mediated action of producing a software program is embedded in the actual software program.
While there are three units of mediated action, and any or all of them can be taken as the starting point of an analysis, there are, of course, many different levels of mediated action. Some are smaller than the lower‐level mediated actions and some are larger than the higher‐level mediated actions discussed above. A lowest‐level mediated action might be an in‐breath, and a larger scale, higher‐level mediated action might be a marriage. What is important is the fact that each mediated action, no matter how small or how large, has a beginning and an ending point. However, neither the beginning nor the ending points of the larger scale higher‐level mediated actions may in fact be part of the data collected for an individual study. Thus, we know theoretically that there is a beginning and an end when investigating larger scale higher‐level mediated actions, while the actual starting and ending points may be unclear. Certainly, the lower the levels of action under investigation are, the clearer the beginning and ending points become. For example, an in‐breath begins at some point and ends at another, a gesture begins at some point and ends at another, and a dinner begins at some point and ends at another.
Mediated actions are phenomenological and visible or audible. But, rather than dismissing psychology, multimodal interaction analysis embeds the notion of attention/awareness into the framework. Attention and awareness are viewed as two sides of (almost) the same thing: When a social actor pays attention to an action, the social actor is also aware of it. However, attention and awareness are not always perceivable in the same way. Just think of a social actor cutting a Christmas ham for the first time. The social actor necessarily has to pay close attention and is also highly aware of the action. But now think of a social actor (we will call her Petra) sitting in a café, holding a sleeping baby, and chatting with a friend about a problem. Petra pays close attention to the friend and the problem that they discuss, and, at the same time, she is holding the sleeping baby in her arms. She pays no attention to the baby (or so it appears), but when you think about it more, you will find that she has to be (and necessarily is) aware of the sleeping child and has to (and does)—in turn— pay some attention to it, even if only a little. Thus, paying attention to and being aware of an action constitute one state, even though sometimes we will be able to perceive this state by investigating the paying of attention, while at other times we are able to perceive it by investigating the state of awareness necessary to perform the action.
When looking at interaction in this way, we discover that a social actor is often involved in more than one higher‐level action at a time: The social actor in the example above is interacting with her friend and at the very same time is interacting with the baby. She thus (co‐)produces two higher‐level actions with two different social actors simultaneously, but on different levels of attention/awareness. This notion of social actors (co‐)producing simultaneous actions was groundbreaking and only possible (a) because of the multimodal lens that multimodal interaction analysis utilizes, and (b) because multimodal interaction analysis is interested in how actions are produced in social actors' attention/awareness levels. While a social actor can always only focus on one higher‐level mediated action at a time, social actors can pay decreasing attention/awareness to many more higher‐level mediated actions. The modal density foreground–background continuum in Figure 1 is used as a heuristic tool to illustrate such simultaneous (co‐) production of higher‐level mediated actions by one social actor.
While the simultaneity of the (co‐)production of higher‐level mediated actions and the attention/awareness levels appears plausible by itself, multimodal interaction analysis uses a scientific notion, called modal density, to delineate the higher‐level mediated actions and to place them relationally on a heuristic social actor's foreground–background continuum of attention/awareness. Modal density is the complexity or the intensity that the social actor utilizes to produce a higher‐level mediated action. Modal density in simultaneous higher‐level mediated actions can be high, medium, or low, or anything in between. When thinking once more of Petra sitting with her friend and the baby in the café, we can see that she is utilizing many modes highly intertwined when speaking with her friend: She uses the modes of spoken language (talking about the problem), gesture (moving her hands/arms every so often), gaze (looking at her friend and looking away at certain intervals), posture (sitting in such a way that she can easily talk with and understand her friend), proxemics (sitting close to her friend), and objects and object handling (sitting in a chair, in front of a table in a café, and drinking coffee out of a cup). There are many modes at play in this interaction which indicate to the onlooker and to the social actors themselves that Petra pays close attention and is highly aware of the higher‐level mediated action that she (co‐)produces with her friend.
When we now think about the modal density in her interaction with the baby, an interaction that she appears to pay no attention to at first sight, we find that she uses the modes of proxemics (holding the baby close to her), posture (sitting in such a way as to easily hold the baby), and touch (holding the baby close to her torso and in her arm) in this interaction. While these modes are closely interwoven, they do not build the same modal density as the many interwoven modes used in the conversation. This example is straightforward and easy to analyze, but there are many more complex simultaneous interactions which can only be analyzed through this very notion of modal density.
Multimodal interaction analysis also investigates the psychological levels of attention/awareness besides the phenomenological ones (Norris, 2009). The idea behind this is the following: A social actor has feelings when performing an action in their everyday life, and, while (co‐)producing one mediated action with other social actors or the environment, the social actor has not forgotten what they did before, and what is even more important, the feelings often linger. Because of this lingering of feelings, prior higher‐level mediated actions often have an impact upon current ones. It is the lingering of feelings that indeed becomes phenomenological when social actors are studied over time and it is this phenomenological appearance of the psychological states that is of interest to multimodal interaction analysis.
Just as social actors are involved in more than one mediated action at a time horizontally, so social actors' mediated actions are embedded in more than one layer of discourse at a time vertically. Generally, we distinguish between three layers of discourse: the central, the intermediary, and the outer layer of discourse (Norris, 2011).
When taking a closer look at Petra in the example above, we find Petra's lower‐level mediated actions, and with them the higher‐level mediated action of speaking to her friend as well as the higher‐level mediated action of holding the sleeping baby, produce the central layer of discourse. This is the immediate layer of discourse, in which Petra as well as her friend and the sleeping baby are involved. These are the immediate, one‐time mediated actions performed.
At the very same time, Petra's mediated action of speaking with her friend and also the mediated action of holding the sleeping baby produce an intermediary layer of discourse. This is formed by Petra in connection with and through her networks, enacting certain long‐term mediated actions. In other words, perhaps Petra speaks with this friend much in the same way she speaks with other friends (illustrating certain knowledge in Petra's network on how she acts), or perhaps Petra speaks with this friend in a café as she never has spoken before (thus negating the network expectation with regard to Petra's usual mediated actions). However, only close research on Petra within her network can give a researcher the intermediary layer of discourse that Petra produces with this mediated action. Similarly, of course, we can see that the mediated action of holding the baby is also embedded in an intermediary layer of discourse. Maybe Petra is a babysitter and takes the babies she watches to meet her friends; or maybe she is a mother, who takes her baby everywhere. In either case, Petra enacts this mediated action of holding the sleeping baby on an intermediary layer of discourse as well as on the central layer explained above. Thus, on the intermediary layer of discourse, these one‐time mediated actions performed locate Petra in her network as a social actor who performs these kinds of mediated actions on a regular basis (or not).
Besides the central and the intermediary layers of discourse, Petra also enacts her mediated action of speaking with her friend and the mediated action of holding the sleeping baby producing and reproducing an outer layer of discourse. This is formed by the larger society, and is enforced by the extended networks that a social actor is part of and by institutions that the social actor engages with. Let us assume that Petra is in fact the mother of the baby and frequently meets this and other friends at cafés, taking her baby along. Petra thus may act in accordance with the institution called “family.” Similarly, when thinking of institutionalized layers of discourse, we may find that Petra speaks and acts in accordance with society's expectation of how a woman of her age is supposed to speak and act. An examination of the outer layer of discourse (institutionalization of actions and societal forces upon the actions performed) in connection with the intermediary layer of discourse (network forces upon the actions performed) and the central layers of discourse (micro‐interactional forces upon the actions performed) as well as the agentive nature of these actions on each of the levels can shed light upon Petra's possibly diverse higher‐level mediated actions that she is combining at this and other times.
With its systematic phases and steps for a multimodal discourse analysis, multimodal interaction analysis is a strong method that is clear in its practical applicability. In multimodal interaction analysis, philosophy, theory, and analytical tools build a coherent framework (Norris, 2019). The multimodal transcription conventions move beyond traditional discourse approaches that add other modes by selecting particular images to demonstrate nonverbal actions. In multimodal interaction analysis, language is transcribed by including intonational curves overlaid over images that show the nonverbal actions and the physical environment as well as objects within, demonstrating the actual ongoing mediated actions as concretely as possible (Norris, 2004). In multimodal interaction analysis, researchers work in a systematic way through analytical phases and steps. They utilize multimodal transcription conventions and concrete analytical tools, all of which are theoretically founded, allowing for replicability of their analyses and reliability of their findings. Thus, in multimodal interaction analysis, we step away from interpretation and focus upon theoretically founded analysis. This coherent framework and way of working can easily be taught and learned and is proving useful for any kind of research where we wish to examine what people do, how they communicate, and how they interact (Norris, 2019).
SEE ALSO: Conversation Analysis and Interactional Linguistics; Multimodality and Systemic Functional Analysis; Multimodal Discourse Analysis
JOHN S. KNOX AND JING HAO
Multimodality is the study of how multiple communication modes (e.g., language, gesture, and proxemics; language, image, and layout) combine to make meaning. But multimodality is not a discipline or a theory; it is “a field of application” (Jewitt, 2009, p. 2) which transcends and draws on different disciplines and theories. The recognition that all communication is multimodal has profound implications for applied linguistics because it questions long‐standing assumptions about the nature of language, and how it should be learned, taught, researched, and theorized. One approach to theorizing multimodality draws on the principles of systemic functional (SF) theory, and this distinguishes systemic functional multimodal discourse analysis (SF‐MDA) from other approaches to multimodality.
Systemic functional theory has a number of defining characteristics which are derived from systemic functional linguistics (SFL). SFL is a theory of language as a resource for making meaning in social contexts. The theory is generated through the study of written and spoken texts, and the contexts in which these texts make meaning. SFL is a functional theory. Its focus is not on the forms of language per se (though these, too, are important), but on the function of linguistic forms in context. SFL is also a systemic theory, meaning that it accounts for meaning by positing meaningful systems of choices available for language users.
SF‐MDA shares fundamental tenets with SFL, perhaps most centrally that it is also founded on systems of choice. To illustrate, in a children's picture book, written text can be integrated in an image, or text and image can appear in distinct spaces, as complementary aspects of the page layout (Painter, Martin, & Unsworth, 2013). There is a functional difference between these layout structures, and in SF theory this difference in structure realizes a systemic opposition.
This relationship between system and structure (called the system–structure axis) is the basis of SF theory. It is fundamental to explanations and hypotheses about language and other semiotic systems, about the relations between text and context, and about other “dimensions” of analysis in the theory (see Martin, 2010).
One key aspect of SF theory, which has been widely applied in studies of multimodality, is the observation that there are three broad kinds of meaning (called metafunctions) in language. These three kinds of meaning (which are derived theoretically from systemic oppositions) are labeled in Kress and van Leeuwen's (1990, 1996) seminal “grammar of visual design” as:
The terminology used to label the metafunctions in SF‐MDA varies across studies and approaches (cf. O'Toole, 1994; Lemke, 2002), but the theoretical debt to the metafunctions of SFL is clear in each approach.
Another key element of SF theory which is important in SF‐MDA is the identification of genres: consistent patterns of structure in groups of texts sharing similar social purposes (see Bateman, 2008).
In SF terms, multimodality can be understood as co‐instantiation, or the simultaneous selection of choices from system networks across different semiotic systems (e.g., from language, from gesture, and from movement through space). Such co‐selection results not in one meaning “added” to another, but in an ensemble which provides us with a high degree of choice and flexibility in meaning making, in what Lemke (2002) refers to as “multiplicative meaning” (see Painter et al., 2013).
SFL focuses on meaning in context, and has long understood language as one semiotic resource among many (e.g., Halliday, 1978). Thus, SF theory has always been amenable to the study of semiotic systems other than language, and of the interaction between such systems in texts.
An important early step in the progression of SF theory from language to other semiotic systems is Kress and van Leeuwen's (1990) Reading Images, motivated in part by “the need for a better understanding of all the things that go with the verbal: facial expressions, gestures, images, music and so on” (Kress & van Leeuwen, 1990, p. 1). This work was soon followed by the introduction of the journal Social Semiotics (in 1991), O'Toole's (1994) study of displayed art and architecture, and the extended version of Kress and van Leeuwen's earlier work (Kress & van Leeuwen, 1996). These last two works (by O'Toole and by Kress & van Leeuwen), in particular, have informed much of the subsequent work in SF‐MDA.
SF‐MDA began with the application of SF theory to the description of semiotic systems other than language. Kress and van Leeuwen's (1996) “visual grammar” illustrates this, with explicit systems, and the structures realizing them, developed and described for visual communication.
Other semiotic systems such as music (e.g., van Leeuwen, 1999) and three‐dimensional space (e.g., O'Toole, 1994; Stenglin, 2009; Ravelli & McMurtrie, 2016) have also been described, as discussed below. SF‐MDA has also been concerned with how different semiotic systems are integrated in acts of meaning: one notable example of this is O'Halloran's (e.g., 2005) description of mathematical discourse as a combination of language, image, and mathematical symbolism. These descriptions of semiotic systems and investigations of multimodality “in action” vary in the extent to which they draw on the system–structure axis developed in SFL, and/or on other aspects of SF theory which have derived from the system–structure axis. They also vary in the extent to which they problematize the “theoretical borrowing” from one semiotic system (i.e., language) to others.
SF‐MDA continues to explore a range of fundamental questions. These include but are not limited to whether (the same) three metafunctions operate across semiotic systems (e.g., van Leeuwen, 1999; McDonald, 2011); the nature of relations between semiotic systems in texts (e.g., Royce, 1998; Bateman, 2008; Matthiessen, 2009; Painter et al., 2013; Zappavigna & Martin, 2018); and whether “levels” of meaning proposed for different semiotic systems have the same theoretical basis (e.g., Martinec, 2005; Baldry & Thibault, 2006; Matthiessen, 2009; Zhao, 2010b).
In the subsections that follow, work that explores these and other questions is grouped according to the kinds of texts explored: two‐dimensional, three‐dimensional, and those including the fourth dimension of time.
Multimodal SF studies of (primarily) static, two‐dimensional texts have looked at a wide range of text types produced in print and hypertext, including instructional materials and textbooks, illustrated books and picture books, print advertisements, newspapers, and texts on social media platforms.
Software manuals in English and Japanese are analyzed by Martinec (2003), who found the Japanese manuals more explicit in their staging and their visual coding of instructions. English‐language instructions for setting up a computer are analyzed by Stenglin and Iedema (2001), who found image and language, and their relations, crucial to understanding the text.
In textbooks and teaching materials, the visual design of information has been shown to have an important role in the construction and reproduction of knowledge (Bateman, 2008). Images and diagrams play critical roles in representing content knowledge across disciplinary areas, such as mathematics (O'Halloran, 2005), history (Derewianka & Coffin, 2008), biology (Guo, 2004), and economics (Royce, 2006). Doran (2018) demonstrates that building knowledge of physics relies on both images and mathematics. In addition to the representation of epistemological knowledge, studies have also revealed that images are a critical resource for cultivating students' attitude toward knowledge. Akashi's (2017) comparative study of history textbooks in Japan and Hong Kong demonstrates how images are used differently to enact opposing attitudes about the same historical events. Chen's (2010) study of English textbooks used in China finds that positive attitudes are inscribed in images for inviting students' positive attitude toward English learning.
Text–image relations in children's picture books have attracted attention from SF‐MDA researchers. Painter et al. (2013) develop a comprehensive set of image systems in children's picture books and model the co‐instantiation of the independent choices of image and verbiage in the narratives. The use of CD‐ROMs and online materials in educational contexts has also been examined (e.g., Unsworth, 2001; van Leeuwen, 2005; Jewitt, 2006; Djonov, 2008; Zhao, 2008).
Many SF‐MDA studies have explored legacy media (e.g., print and online newspapers). Caple (e.g., 2013) explores press photography and the interactions between texts and images, both in print and digital news media. Economou (2014) finds that visual images can be used to position readers with very different attitudinal meanings in respect to the same “factual” content. Much work on legacy media has also examined composition and multimodal design, including Kress and van Leeuwen's (1996, 1998) description of print home pages; Knox's (2007) description of home pages of online newspapers; Caple's description of image‐nuclear stories; Zhang's (2017) analysis of science news Web sites; and Caple and Knox's (e.g., 2017) description of emerging story genres in online news galleries. In addition, Bednarek and Caple (2017) propose an approach that incorporates corpus linguistic methods to assist multimodal discourse analysis. In addition to legacy media in English, studies have also explored other linguistic contexts, including Greek (e.g., Economou, 2008) and Thai (Knox, Patpong, & Piriyasilpa, 2010).
Image–verbiage relations have also been analyzed in magazines. Zhao's study of a popular Australian independent magazine (Zhao, 2014) reveals the multimodal construal of “indie taste.” Other SF‐MDA influenced studies on magazines include Stöckl's (2017) examination of text–image relations in a popular science magazine, and Pflaeging's (2017) analysis of narrativity in National Geographic feature articles, both of which adopt a diachronic genre perspective to trace the development of text–image interactions through time.
In recent years, SF analysis has informed a series of studies on social media platforms including Twitter (e.g., Zappavigna, 2018) and Instagram (e.g., Zappavigna, 2016; Caple, 2018). Zappavigna's work has demonstrated that hashtags are a significant metafunctional social semiotic resource that is critically used by microbloggers to perform their identities and position themselves in “ambient affiliation.” SF‐MDA studies have also revealed some unique features of social media photographs. Zappavigna (2016) identifies a system of subjectification which accounts for a novel interaction between the viewer and the photographer in “selfie” posts. Zhao and Zappavigna (2018) argue that the selfie is a distinctive visual genre as it can articulate and generate perspectives in different ways. Studies of social media have also demonstrated its significant function in public engagement with political events (e.g., Caple 2018; Zappavigna 2018).
The most well‐known early work on developing a “grammar of space” is that of O'Toole (1994), who poses four ranks for analyzing architecture (Building, Floor, Room, Element), mapped against the three metafunctions. This approach to SF‐MDA has been applied across a range of semiotic systems and multimodal contexts by O'Toole and others (see O'Toole, 2004; cf. Zhao, 2010b), and specifically to three‐dimensional space in museums.
Ravelli's (2006) study of museum texts builds on an analysis of language in museums to describe exhibitions, and the institutions that house them. Her conceptualization of the ranks of Institution, Exhibition/gallery, and Exhibit/display, is shared theoretically by Pang, who identifies the ranks of Museum, Gallery, Area, Surface/item (Pang, 2004). Taking an alternative approach, Stenglin (e.g., 2009) also theorizes the construction of space in museums, developing a metafunctional grammar of three‐dimensional space more broadly. Lim, O'Halloran, and Podlasov (2012) analyze the functional spaces in classrooms as a dynamic interplay of the design and arrangement of the classroom, and the use of the space. McMurtrie's (2017) grammar of movement in space construes our movement as a semiotic resource that enables us to transform the meanings of the semiotic systems that have been co‐deployed in a space. Drawing on the emerging works on spatial texts, Ravelli and McMurtrie (2016) establish spatial discourse analysis as a field of exploration, and model how spatial texts can be analyzed in a range of examples including the foyers of high‐rise apartment complexes, a university library, a shopping center, and an art museum.
A complementary approach to analyzing space in the field of education can be found in Kress et al. (2005), who illustrate how English classrooms can be “read” (and “written”) as multimodal signs which reflect the teachers' philosophies of teaching and learning, and their institutional practices (see also Jones, 2008).
An emerging number of SF‐MDA studies have explored the relationship between gesture and language. Drawing on Cleírigh's unpublished work on linguistic body language, and following Matthiessen's (2009) identification of sound quality and body language as paralanguage, Martin and Zappavigna (2018) make a distinction between body language in sync with the prosodic phonology of spoken language (sonovergent resources) and body language that is oriented to the multifunctional content plane of spoken language (semovergent resources). From a metafunctional perspective, Hood and Forey (2005) reveal the co‐patterning in conference presentations of gesture with interpersonal meanings in language. Similarly, Jewitt (2006) describes the role of gesture and embodied meanings in expressing the interpersonal relations between teacher and students, and also their attitude toward the reading of a poem, and Amundrud (2018) analyzes the use of gaze in EFL (English as a foreign language) classrooms. Hao and Hood (2019) illustrate the convergent interaction between spoken language and paralanguage in cultivating a set of scientific values in an undergraduate lecture. Zappavigna and Martin (2018) describe a youth justice conference where young offenders are brought together with a range of people from the community. They demonstrate how embodied language and arrangement in a space are used critically to negotiate shared values and create bonds in a community. Ideationally, Martinec (2004) identifies detailed and systematic consistencies between experiential meanings in language and the gestures that co‐occur with speech. Hood and Hao (in press) demonstrate the ways in which body language interacts with verbiage to build complex knowledge of scientific taxonomies and activities.
Studies of body language have also explored the adaptation of gestures in animated films (Unsworth, 2014; Ngo, 2018a). Ngo's study of a children's animated film reveals that gesture and co‐occurring speech by the character typically do not realize the same kinds of attitudinal meanings. Unsworth's (2014) comparative analyses of selected segments in book and movie form indicate the different possibilities of employing semiotic resources in two different modes. The findings of children's multimodal literacy provide implications for teaching comprehension and representation of gestural meanings in multimodal literacy (Ngo, 2018b).
Body language as a critical communicative resource has also been demonstrated in Dreyfus's study on nonverbal communication of a child with an intellectual disability (Dreyfus, 2011). She illustrates that a variety of semiotic behavior needs to be instantiated to enable the success of nonverbal communication. This study also poses the challenges of determining the boundaries between semiotic and non‐semiotic behavior.
Some multimodal texts, such as music and film, are “time‐based”: the unfolding of the text and its speed is a property of the text, rather than of the reader's ability to process it.
Music has been studied from an SF perspective most notably by Theo van Leeuwen. By systematizing meaningful sound qualities of speech and music such as perspective, timing, interaction, and pitch, van Leeuwen (1999) provides the beginnings of a map of the meaning potential of sound as used in speech, in music, and as experienced in human contexts beyond speech and music (see Machin, 2010). Building on van Leeuwen's work and work on phonology by Halliday and Matthiessen, Caldwell (2014) accounts for the differences between the sung and the rap voice, providing a comparative set of features of embodied meanings. Weekes (2016) also builds on the work of van Leeuwen and explores how senior secondary students of music use diagrams to construe musical meanings visually in examinations. McDonald (2011) problematizes the approach taken by van Leeuwen, and by music scholarship more broadly, arguing that music should be approached as embodied meaning.
SF‐informed studies have also examined film and television texts, or filmic texts. A key issue in the analysis of such texts is their segmentation and transcription (Baldry & Thibault, 2006), and different analysts use different transcription models in order to illustrate different textual features: van Leeuwen (2005) illustrates rhythm and links between the information provided in/with shots; Pun (2008) explores the use of sound and silence in films; and Tseng (2013) describes coherence across semiotic resources and modes in filmic texts. Bateman and Schmidt (2012) employ the metafunctions of SF theory and established non‐SF approaches to analyzing film to outline an approach to analyzing and understanding film discourse from both paradigmatic and syntactic perspectives. Likewise, Wildfeuer (2014) employs metafunctions and stratification from SF theory in describing “the logic of film discourse interpretation.” Chen and Wang (2016) also employ the metafunctions in their categorization of the types and functions of semiotic interplay between dynamic images and subtitles in a film.
In addition to time‐based texts, some online texts have been studied from the perspective of readers' trajectories through them. Djonov's (2005) study of children's Web sites documents different readers' paths through the same site, demonstrating that different paths through a Web site, and even different amounts of time spent on the same page, can change the meaning of the text experienced by the reader. Zhao incorporates the dimension of time into her modeling of readers' paths through “educational interactives,” theorizing the “critical path,” or “traversal,” as the unit of analysis (Zhao, 2010a). The notion of “resemiotization,” is developed by Iedema (2003), who traces the process of meanings being “resemiotized” as part of institutional processes with, for example, meetings resemiotized into minutes and briefs, which are resemiotized into building plans, which are eventually resemiotized into a building. Complexity in multimodal texts has also been explored in the software design, slideshow composition, and presentation of PowerPoint slide shows (e.g., Zhao, Djonov, & van Leeuwen, 2014). These studies illustrate some of the ways in which multimodality has challenged traditional notions of what constitutes a text, and how we approach text analysis.
Numerous fundamental questions in multimodality still require (further) empirical investigation (Bateman, Wildfeuer, & Hiippala, 2017). Yet, SF‐MDA offers a consistent theoretical basis for describing multimodal texts, studying semiotic systems, and for understanding linguistic texts not as isolated entities, but as sets of choices from linguistic systems that are always in interaction with sets of choices from other, related systems.
SEE ALSO: Multimodal Interaction Analysis; Systemic Functional Linguistics
SCOTT E. GRAPIN
Multimodal perspectives on teaching and learning build on the basic assumption that meanings are made (as well as distributed, interpreted, and remade) through many forms and resources—image, gesture, gaze, body posture, sound, writing, music, speech, and so on— of which language is but one (Kress & van Leeuwen, 2001; Jewitt, 2014). That communication is multimodal is not, however, new—people have always used image and nonverbal forms to communicate. The multimodal character of communication, teaching, and learning is increasingly commonplace across a range of disciplines (e.g., anthropology, education, design, linguistics, media and culture studies, musicology, sociology). My focus here is on multimodal teaching and learning from a social semiotic perspective, which is informed by linguistic theories, in particular the work of Halliday's (1978) social semiotic theory of communication and developments of that theory by Hodge and Kress (1988).
Detailed studies on specific modes, such as Kress and van Leeuwen's seminal work on images (1996), and other key works that contribute to an evolving “inventory” of semiotic modal resources, have helped begin to describe nonlinguistic semiotic resources and their material affordances and organizing principles. The concept of modal affordance refers to what it is possible to express and represent easily. How a mode has been used, what it has been repeatedly used to mean and do, and the social conventions that inform its use in context, shape its affordance. Multimodal research shows that people draw upon their available modal resources to make meaning in specific contexts. Further, these resources come to display regularities through everyday patterns of use. The more a set of resources has been used in the social life of a particular community, the more fully and finely articulated its regularities and patterns become.
Alongside the assumption that all modes in a communicative event or text contribute to meaning, multimodality asserts that all modes are partial. That is, all modes, including the linguistic modes of writing and speech, contribute to the construction of meaning in different ways. Therefore, no one mode stands alone in the process of making meaning; rather, each plays a discrete role in the whole. This has significant implications in terms of epistemology and research methodology: Multimodal understandings of teaching and learning require the investigation of the full multimodal ensemble used in any communicative event. The imperative, then, is to incorporate nonlinguistic representation into understandings of what teachers and students say and write in the classroom.
Given emergent local foci of multimodal practices, research on teaching and learning from a multimodal perspective is generally small scale, ethnographic, and case based—with limited analysis of the impact on teaching and learning. Much of this work is descriptive, offering detailed inventories of the resources used by students and teachers and how these resources are developed into multimodal ensembles.
Multimodality has been taken up to explore teaching practices in a variety of contexts and to inform models of teaching (New London Group, 1996; Cope & Kalantzis, 2015). Multimodal research has examined the ways in which teachers orchestrate a range of modal resources, including gesture, gaze, position, posture, action with books and boards, and talk in the classroom. It has examined the ways in which language policy, student identities, official curriculum, and school knowledge are mediated through multimodal communication in the classroom (Bourne & Jewitt, 2003; Kress et al., 2005; Kress, Jewitt, Ogborn, & Tsatsarelis, 2014). Such work has shown that significant aspects of teaching work are realized through a range of modes.
Studies of multimodal practices of science and English classrooms in the UK have shown that this holds true even in a curriculum context such as English where talk and writing dominate the classroom (Kress et al., 2005). The Multimodal Production of School English project involved detailed video recording and observation of nine English teachers in three inner London schools, as well as interviews with teachers and students and the collation of texts made and used in the classroom. The project shows the complex ways in which image, gesture, gaze, interaction with objects, body posture and writing, and speech interact in the classroom production of school subject knowledge. The School English project highlighted how students and teachers coproduce notions of ability, resistance, and identity in the classroom through their nonverbal interaction. The way in which classroom displays, space, furniture, and artifacts were designed to realize versions of English as a school subject was also documented. Overall, this research showed that the work of interpreting school English is beyond language and requires the ability to make sense of a range of modes and the relationships between them.
Within the multimodal environment of the classroom, modes come to take on specific situated roles in the construction of curriculum knowledge. A teacher's choice of mode can therefore be viewed as a process of pedagogic design, matching the target knowledge with the particular modal affordance in a given situation. For example, representation of a cell in the science classroom as an image or through writing, in color or black and white, as a three‐dimensional model or an animated sequence, makes available and foregrounds different aspects of the concept of cell (Kress et al., 2014). As different modes have differential potential effects for learning, the shaping of learner identities, and how learners find pathways through texts, the choice of mode is central to the epistemological shaping of knowledge and ideological design. In this regard, the long‐standing focus on language as the principal, if not sole, medium of instruction can at best offer a very partial view of the work of communicating in the classroom. The relationship between modes in learning materials raises important questions related to the gains and losses realized through the representation of curriculum knowledge in one mode rather than another (Bezemer & Kress, 2008).
Understanding the communication landscape of the classroom through a multimodal lens has significant implications for learning. Thinking about learning as a process of design and choice (of representational modes) gives a new focus to the role of the learner. From this perspective, the multimodal texts and artifacts that students produce can be viewed as one kind of sign of learning, a material trace of the process of meaning making (Kress et al., 2014). The interpretative work of students is also reshaped through their engagement with a range of modes. In such a view, students need to learn how to recognize what is salient in a complex multimodal text, how to read across the modal elements in a textbook or interactive whiteboard, and how to move from the representation of a phenomenon in an animation to a static image or written paragraph. Learning increasingly involves students in working across different sites of expression, negotiating and creating flexible spaces for planning, thinking, hypothesizing, testing, designing, and realizing ideas (Jewitt, 2008).
New skills for reading, and for finding, authenticating, manipulating, linking, and recontextualizing information, go beyond traditional taxonomies of print skills (Kress, 2003; Leander & Wells Rowe, 2006). The structure of many digital texts opens up options about where to start reading a text—what reading path to take. The multimodal design of texts in the contemporary digital landscape offers students different points of entry into a text and alternative possible paths through a text, highlighting the potential for readers to remake a text via their reading of it. The reader is involved in the task of finding and creating reading paths through the multimodal, multidirectional texts on the screen—an openness and fluidity of structure that seeps onto the page of printed books (Kress, 2003; Moss, 2003). Writing, image, and other modes combine to convey multiple meanings and encourage the reader to reject a single interpretation and to hold multiple possible readings of a text (Coles & Hall, 2001). In addition, new technologies make different kinds of cultural forms available, such as computer games and Web sites, offering new spaces for publication and dissemination and building new relationships between readers and writers in a way that challenges and breaks down some of the traditional distinctions between them.
Although teaching and learning have long been recognized as multimodal, relatively less attention has been paid to the implications of this for second language (L2) students, who are learning content in and through a language that is not their first. This lack of attention may be due, in part, to the different ways in which multimodality has been interpreted in the fields of L2 education and content area education, respectively. Traditionally, L2 education has viewed nonlinguistic modes (e.g., image) as scaffolds or supports for language development. For example, L2 students may be asked to draw their understanding of science concepts until they develop proficiency in the more privileged linguistic modes of speech and writing. In contrast, content area education has viewed nonlinguistic modes as essential semiotic resources of academic disciplines. From this perspective, drawing may be beneficial to L2 students, but it may also offer affordances for all students to represent concepts more effectively and to engage in the type of visual representation that is valued in disciplines such as science (Kress, 2000; Lemke, 1998). These divergent perspectives have been referred to as the “weak and strong versions of multimodality,” respectively (Grapin, 2019). Interrogating these divergent perspectives and their consequences for teaching and learning with L2 students is particularly urgent in the US context, as new K‐12 content standards expect all students, including a growing population of L2 students, to engage in multimodal disciplinary practices (e.g., constructing digital texts in language arts; analyzing visual representations in mathematics; developing models in science). The multimodal nature of these new standards sends a clear message that multimodality is not compensatory for L2 students (i.e., the weak version) but essential for all students to do the work of the content areas (i.e., the strong version).
Despite the tendency to privilege linguistic modes in L2 education, recent theoretical developments have expanded what “counts” as meaning making with L2 students. In particular, socially oriented theories in L2 education have blurred traditional boundaries between meaning‐making systems and argued for a view of language as fully integrated with other modes (van Lier, 2004; García & Li, 2014). In conjunction with these theoretical developments, research has begun to explore how L2 students across various contexts put all of their semiotic resources into the service of meaning making. This line of inquiry has shed light on the benefits of multimodality for promoting identity expression and empowerment of L2 students, particularly in out‐of‐school learning contexts (Lam, 2009). Moreover, a small but emerging body of literature has also begun to document the ways in which L2 students draw on multiple modes to access and engage with rigorous content as part of the core curriculum (Smith, Pacheco, & de Almeida, 2017). By transcending traditional disciplinary boundaries between L2 education and content area education, this research holds great promise for turning a deficit view of L2 students into an asset view that leverages the diverse semiotic resources they bring to the content classroom.
SEE ALSO: Multimodal Discourse Analysis; Multimodal Interaction Analysis; Multimodality and Systemic Functional Analysis; New Literacies of Online Research and Comprehension
3.147.80.3