Sociocultural Theory and Corpus‐Based English Language Teaching

Although Sociocultural Theory (henceforth, SCT) originated nearly a century ago in the writings of Russian psychologist L. S. Vygotsky, historical circumstances meant that the theory began to influence international research communities only in the 1980s. Since its introduction to the second language (L2) field (e.g., Frawley & Lantolf, 1985), central concepts such as mediation, internalization, and zone of proximal development (ZPD) have become widely invoked to understand processes of instructed L2 development. From the early 2000s onward, much L2 SCT scholarship has sought not merely to analyze learner development but to actively guide it through pedagogical practices informed by the theory (see Lantolf & Poehner, 2014). To date, relatively few L2 SCT studies have been carried out in conjunction with corpus linguistics. In this paper, we outline central principles from the theory, with particular attention to how they have been brought into the L2 field. Included here is a line of scholarship that has employed the theory to organize L2 curricula around linguistic concepts and to design activities and resources intended to support learner engagement with them. This work, referred to as concept-based language instruction (C-BLI), provides fertile ground for uses of corpora to help teachers understand learner language abilities and to provide examples of language that illustrate the concept under study. After explaining common procedures associated with C-BLI, we offer examples from recent studies that suggest how the integration of this work with corpus linguistics might enrich L2 education. We then provide suggestions for future research. As mentioned, SCT regards education as having the potential to profoundly impact psychological development. Of course, the premise that our psychological activity is mediated through social (i.e., interactions) and cultural (i.e., artifacts or resources made available to us, including language) means explains development both in and outside of formal educational environments. Aside from rare and tragic cases in which a child is denied access to human interaction, all individuals develop what Vygotsky (2012) termed higher forms of consciousness as the resources available in their culture (e.g., practices, belief systems, forms of knowledge, and, perhaps most important, language) enable them to exert intentional control over innate capacities of attention, memory, and perception. While these latter cognitive processes are part of our natural endowment, it is their transformation through social and cultural means that yields, for instance, our ability to purposefully commit information to memory and to recall it. Just as the introduction of a physical tool into physical activity changes our behaviors, symbolic tools such as language alter our psychological activity. We come to not only act in response to environmental stimuli as other animals do, but instead, we are able to act first on the psychological plane and only later on the physical plane. In other words, through language, we are able to draw upon our knowledge to reflect on situations, devise plans, anticipate possible outcomes, and determine alternatives, all before carrying out an action. As his investigations into the development of higher forms of consciousness continued, Vygotsky realized that with the introduction of schooling come distinct forms of mediation that guide the development of ways of thinking that are not likely to arise in the everyday world. Following Vygotsky (2012), consider the example of fish and mammals that live in the sea. Relying on everyday experience, we may group these together, failing to understand that the latter are warm-blooded, need air to breathe, give live birth, etc., and so should be classified differently. Moving from natural science to economics, we might similarly consider how abstract categories such as social class, capital, and modes of production help us to understand how the social world is structured in ways that would be difficult to apprehend solely on the basis of our own individual experiences. The contrast that Vygotsky drew then was between forms of knowledge that result from direct, often sensory, experience and those that emerge from fields of inquiry such as science and history and that are available to us through education. The former he termed spontaneous conceptual knowledge and the latter scientific. While spontaneous concepts reflect experiences we happen to have had and therefore lack systematicity, scientific concepts are highly systematic, with interrelations and hierarchies existing among them. Schools offer a special environment in which we encounter scientific concepts that have resulted from efforts of a community of scholars, sometimes over generations. Crucially, we do not need to rediscover this knowledge for ourselves but can come to understand it through planned instruction. It is in this regard that Vygotsky (1997, p. 88) considered education to be the "artificial development of the child." To be sure, what Vygotsky envisioned is a model of how education can optimally promote our development through the study of academic subjects, not a description of what always occurs in schools. Within the L2 field, as in any area of education, a range of practices exist, and these are frequently a topic of debate among researchers. As the reader may appreciate, C-BLI advocates the value of explicit knowledge of the L2 as essential mediation with which learners may regulate their language use and interpret its use by others. Importantly, of course, this does not mean a presentation of the language through sets of prescriptive grammar rules but rather as linguistic concepts. While Vygotsky identified the value of a concept-driven curriculum for promoting the development of particular ways of thinking through subject area study, it was P. Y. Gal'perin who devised the model that has most strongly influenced C-BLI. Working within a Vygotskian framework, Gal'perin (1969) became interested in how learners internalize and come to use symbolic mediation within subject areas to regulate their mental functions. Over the course of several decades, Gal'perin and colleagues conducted numerous studies in a variety of school subjects and across age and grade levels (see Haenen, 1996). This work led to an instructional model in which each step is devoted to supporting learner internalization of scientific concepts within a domain. It was this model that was adapted for the L2 classroom by Negueruela (2003) in the first C-BLI project. Conducted with university learners of L2 Spanish at a U.S. university, Negueruela's (2003) study focused on verbal tense and aspect, specifically the preterit-imperfect distinction. Traditionally, these are presented to learners as two contrasting "tenses," with selection based on a range of conditions (e.g., "duration" of an action), often resulting in confusion over which rule to apply in a given circumstance. Negueruela presented verbal tense and aspect as a coherent linguistic concept with preterito and imperfecto as options for realizing it in Spanish. Rather than determining a rule to apply, instruction emphasized a speaker's perspective on an action, specifically whether it is brought into discourse as completed at a fixed point in time. Along with a verbal explanation and contrasting examples, the concept was also represented through visual models. As we detail below, the C-BLI approach included leading learners through a sequence of activities designed to promote their use of the concept as a symbolic tool, that is, to mediate their internalization of the concept so that eventually they could think with the tool even when supporting materials were no longer present (see Lantolf & Poehner, 2014). Learner performance on oral and written tasks requiring the use of the preterito and imperfecto, along with verbalizations of the reasons underlying their choices provided evidence of their conceptual development. Following Neguereula's (2003) initial project, C-BLI programs have been implemented with learners of different languages, across proficiency levels, and focused on a range of linguistic concepts (for review, see Lantolf, Xi, & Minakova, 2021). While there is some variability, C-BLI programs generally proceed through six phases, the first of which concerns identifying learner current understanding of the language features in question. This means determining how learners interpret the language features and their potential for creating and expressing meaning. In C-BLI, this is referred to as the learner's current orienting basis for action, or OBA, as it is upon this understanding that learners select among available options in the language to communicate (i.e., it orients learners to act with/through the language). In this phase, corpora of learner language could serve as a useful resource to help teachers identify an individual learner's or a group of learners' current knowledge of the target language features, assuming the tasks used to elicit learner production entail abundant use of such features. As demonstrated in many learner corpus studies, learner corpora can be analyzed to assess the use of target language features (e.g., in terms of the frequency, accuracy, appropriateness, diversity, sophistication of the target feature use) at a given time point, across multiple time points, and/or in reference to other learners or learner groups (see Lu, 2023 for a review of such studies). In line with our observations concerning L2 instructional approaches, this phase often reveals that learners have only a partial understanding of the concepts either because previous teaching provided prescriptive rules to follow or, in the case of immersion experiences, learners have not been able to infer their own how a language feature works. The second and third phases in C-BLI involve the presentation of the concept. In the second phase, the concept is presented verbally as a coherent whole and accompanied by examples. This supports comparisons between learner OBA and the full concept. Especially important here is that learners understand the concept rather than simply memorizing its definition. To support this goal, in the third phase, the concept appears in a material (e.g., a physical model learners can manipulate) or materialized (i.e., a symbolic representation such as an image, graph, or chart) form. In this way, learner attention shifts from the words of the definition to the meaning represented by its material/materialized form. Much C-BLI work has drawn upon explanations of language from research in Systemic Functional Linguistics or Cognitive Linguistics. Accounts of language framed in these approaches eschew simple rules in favor of meaning-based analyses, and in this way, they often provide an appropriate "content" around which a C-BLI program may be organized. The full concept and its representations form what is referred to as the Schema for Orienting Basis for Action (SCOBA). The SCOBA is the symbolic tool learners may use to regulate their language use. In these phases, corpora could be drawn upon to inform the formulation, presentation, and illustration of the target concept. A good example of how corpus analysis could help formulate and present a concept is Casal's (2020) analysis of various formal linguistic features (i.e., reporting verbs, phrase-frames, shell nouns, and syntactically complex structures) used to realize different rhetorical move-steps in a corpus of published research article introductions. The findings on the distribution of such formal features across different rhetorical move-steps formed the basis of his subsequent concept-based pedagogical intervention aimed at helping graduate student writers develop their genre knowledge. Additionally, corpus analysis can also inform the selection of relevant focal items for pedagogical intervention. For example, in a study of the effects of C-BLI on L2 English learners' acquisition of verb–noun collocations, Tsai (2020) selected 12 focal items (i.e., specific verb–noun collocations) based on their frequency of occurrence in the Corpus of Contemporary American English in addition to insight into their degree of difficulty for L2 learners from previous learner corpus studies of collocational errors. Phase four of C-BLI begins the process of learner internalization of the SCOBA as activities require them to verbalize their understanding of it, its meaning, and its relevance for communication. Two subphases of verbalization include language production for others, as when learners explain their understanding to their peers or the teacher, and language production for the self, that is, talking one's way through use of the SCOBA as a mediating tool. Phases five and six then extend this to use of the SCOBA during communicative activities. At first, the SCOBA may be physically present as a reference, but over time learner verbalizations of the SCOBA's meaning take on greater importance, and ultimately the concept is internalized. At phase six, internalization, external representations of the concept are no longer required by the learner, whose use of the language is now regulated by new conceptual understanding. In these phases, corpora could form part of the context and/or materials to support learner efforts to use the concept under study (and the forms/structures through which it is realized) in interpreting and producing language. For example, Casal (2020) reported how integrating corpus and genre analysis in his C-BLI-informed pedagogy helped his graduate student participants develop genre knowledge and conceptual resources for engaging in goal-driven genre practices. To date, C-BLI studies that take advantage of corpus resources and methods remain limited. Future research can explore and evaluate effective ways to integrate corpus research findings, corpus data, corpus analysis, and corpus activities in various phases of C-BLI. In the first phase, the following question can be addressed with learner corpus analysis in cases where the target concept has already been determined: What is the learners' current understanding of the concept as evidenced in their language production? The corpus analysis results can complement teacher judgments or results from other assessments, including those that do not rely upon language production. As explained, this can help to identify learner OBA at the outset of a C-BLI program. Beyond analyses of specified concepts, it is further possible to adopt a data-driven approach to learner corpus analysis (e.g., Ivaska, 2015) for the purpose of determining which concepts learners do not understand or only partially understand and that could be prioritized in a subsequent C-BLI intervention. In the second and third phrases of C-BLI, systematic corpus-based analysis can be performed to address a number of questions surrounding concept formulation, presentation, and illustration. What is the range of forms, meanings, and uses of the target concept evident in a representative corpus that needs to be captured? Does an existing SCOBA adequately capture this range? Is a new SCOBA—or multiple ones—required to represent the full range of the concept's forms, meanings, and uses? How might the relative frequency of usage of different forms, meanings, and uses of the target concept be reflected in its presentation to learners? Would using authentic instances from a representative corpus to illustrate different meanings and uses of the target concept facilitate learner understanding of the concept? In the last three phases of C-BLI, it may be productive to engage learners in interacting with corpus data as they endeavor to use the concept in communicative activities. An especially useful research question that can be investigated in these phases is: What form of data-driven learning and/or what types of corpus activities may be effectively integrated with C-BLI to facilitate learners' practice with and internalization of the target concept? Contingent on positive findings for such research questions, it will then be important to consider which types of teacher training and/or teacher-researcher collaboration may be necessary to enable effective integration of corpus linguistic techniques in C-BLI. Recent research has also called for the synergy of SCT and intelligent computer-assisted language learning (Ai & Lu, 2018). Looking ahead, a promising area for future research lies in the exploration of ways to exploit large language model-based generative artificial intelligence systems such as ChatGPT in C-BLI. This research area is relevant to our discussion as large language models are trained on large-scale corpus data. Many questions regarding the instructional implementation of such systems remain to be addressed. For example, how can generative AI systems be leveraged to profile the learners' understanding of a target concept? This could be approached through the delivery of a series of evaluation questions or an automated analysis of the learners' use of the target concept during production tasks. Another question concerns how such systems might be trained to formulate, present, and illustrate linguistic concepts automatically. Can the systems elicit verbalizations from the learners and assess their verbalized conceptual understanding? Can they provide contextualized communicative activities to help learners internalize a target concept? Iterations of rigorous prompt engineering research can provide answers to these important questions. Subsequent research can then investigate the types of teacher and learner training necessary to facilitate effective interaction with such systems for C-BLI, the learning outcomes obtained from and learner conceptions of AI-powered C-BLI, and the degree of learner autonomy in conceptual learning developed through AI-powered C-BLI as well as the longer-term sustainability of such autonomy. Matthew E. Poehner is Professor of World Languages Education and Applied Linguistics at The Pennsylvania State University. Dr. Poehner's research engages Vygotskian Theory to understand processes of instructed L2 development and to organize educational activities to promote learner language abilities. He is co-editor of the journal Language and Sociocultural Theory. Xiaofei Lu is Professor of Applied Linguistics and Asian Studies at The Pennsylvania State University. His research interests are in corpus linguistics, English for Academic Purposes, second language writing, and computer-assisted language learning. He is the author of Corpus linguistics and second language acquisition: Perspectives, issues, and findings (Routledge, 2023).