A review study on discourse-analytical approaches to language testing policy in the South Korean context

Abstract

Research on language planning and policy (LPP) has transcended beyond enacting government-initiated language policy and progressed towards uncovering inherent power relations. The discourse-analytical approach (DA) effectively captures the relation between texts and other elements of social practices. Therefore, this approach has been adopted to analyze the meaning-making process in ideologies, practices, and different social contexts. The current study expands on previous research and argues that language testing policies (LTP) be explored using the DA because the former is also produced within the discursive practices of intertextual and interdiscursive links among texts (e.g., policy documents and media texts) and social structures. Three empirical contexts are analyzed to demonstrate the need for introducing tests: (1) introducing the National English Ability Test to replace the preexisting English subject test of the College Scholastic Test in Korea to measure speaking, writing, reading, and listening skills; (2) adopting cut-off scores for graduation based on the Test of English for International Communication; and (3) implementing the Test of Proficiency in Korean scores for screening and regulating the flow of immigrants. The results facilitate an understanding of the variability in LTP contexts, thereby contributing to democratic pluralism in the LPP field.

Introduction

In recent decades, increasing awareness of the social impact of language tests has been noted (McNamara & Roever, 2006; Messick, 1989; Shohamy, 2001). Shohamy (2008) raised concerns over the power of high-stakes testing; she reported that tests act “as powerful educational, societal, and political devices that have strong effects on de facto language policy” (p. 371). When language testing is adopted to enable policy enactment, the power of language tests becomes notably significant. In recent years, research on language testing has explored the enactment of language-related policies fueled by high-stakes language testing and focused on the discourse-analytic approaches (DA) to language planning and policy (LPP). Similar to other education policies (Ball, 1993) and language education policies (Johnson, 2011; Ricento & Hornberger, 1996), language testing policies (LTP) must be conceptualized and probed as multilayered phenomena and processes.

In this study, LTP is considered a complex, multilayered social practice. The consideration is based on three example contexts wherein test development and administration of English and Korean proficiency were adopted to enact the related policies in South Korea (henceforth, Korea). This study aimed to illustrate that LTP is planned and implemented through “discourse.” The term “discourse” in this study does not mean a mere level of language “beyond the sentence.” Discourse reflects social practices. It is not neutral but is motivated by sociopolitical interests. Discursive events in language testing and policy are understood as the sites of motivated meaning and actions, in which power structures and actors affect each other dialectically. Policy meaning is discursively produced, reinforced, contested, and resisted, as policy discourse is both constitutive and constructive of policy (Barakos & Unger, 2016).

By focusing on three empirical research contexts in Korea, wherein we participated in LTP data collection and analysis as primary investigators, the relevance of the DA to LTP (hereafter, DALTP) was revisited and elaborated. The following government-led educational policies are discussed in this study: (1) reform policy for English language education by the National English Ability Test (NEAT), (2) English benchmark policy for graduation, initiated by the government and operationalized by Korean universities to use the Test of English for International Communication (TOEIC), and (3) educational policy for screening international students and foreign-born adolescents of an immigrant spouse by using the Test of Proficiency in Korean (TOPIK) scores.

Literature review

Development of language planning and policy

Barakos and Unger (2016) categorized the LPP research into four “waves” of history, building upon the framework established by Ricento (2000) and expanded upon by Hult (2010) and Johnson (2013). It is essential to underscore that the “waves” should not be perceived as rigid categories; rather, they overlap, existing along a continuum. They serve as a useful framework for future researchers to retroactively analyze trends in the field.

In the first wave, the focus of LPP research was on identifying, describing, and defining official and indigenous languages. Standard languages were polished with the intention of ensuring state modernization and with the aim to resolve language-related problems in postcolonial contexts following World War II. Societies with diverse linguistic communities often saw LPP as a means for the government to address these challenges (Rubin, 1971). Governments often adopted problem-solving and cost–benefit models to optimize economic growth (Ricento, 2000), and LPP research was considered ideologically neutral and scientific.

The second wave of LPP research began to challenge the positivist linguistic approach of first wave, giving greater consideration to policy impact. Reseachers shifted their attention to the multifaceted contexts in which language planning occurred, moving beyond the idea that it was solely imposed by government bodies. Scholars began to question the feasibility of earlier language planning models and put the growing focus on the sociopolitical and ideological dimensions inherent in LPP. The shift paved the way for the emergence of the third wave, known as “critical language policy” (Tollefson, 1991).

In the third wave, scholars delved deeper into the inherent power dynamics in language policy, critically examining how policies often served the interests of dominant social groups and perpetuated various forms of social inequalities (Tollefson, 2006). They aimed at mitigating inequality and developing more democratic language policies, where minority languages could coexist. The third wave of critical language policies highlighted the interactions between language ideologies and discourses, particularly in the media.

Fourth-wave researchers have re-evaluated the conceptualizations of macro–micro dichotomy (Johnson, 2009) and have proposed a new understanding of the dialectic relationships between the policy as text and discourse within the power structure (Ball, 1993; Hult, 2010; Johnson & Johnson, 2015). Starting from this fourth wave, often referred to as “discursive approaches to language policy” (DALP) (Barakos & Unger, 2016), LPP was perceived as a multilayered phenomenon shaped through the dialectical process within historical contexts. Numerous theoretical and methodological considerations aimed to move beyond the analysis of policy as mere text. For instance, Ricento and Hornberger (1996) applied the concept of situated practices of different actors to language educational policy, using the metaphor of an onion to illustrate the multiple layers of policy development.

Discursive approaches to test-driven language policy enactment

The contemporary understanding of LPP has evolved to be complex and dynamic (Lo Bianco, 2009), transitioning from a view of policy texts as objective knowledge to seeing them as subjective constructs shaped by individual or societal experiences and interactions (Fairclough, 1992). This shift also underscores the need to view formulation and implementation of LPP not as a linear policy formulation process but as a discursive construct and a battleground of competing interests. Among various facets of LPP, issues related to language education policy have been recognized as highly dynamic, value laden, and context bound. Therefore, it is essential to examine the analytic techniques in texts, such as intertextuality observed in policy documents and social media, as well as the interdiscursivity among different genres, discourses, and styles. This assertion finds support in the empirical research conducted by Bubikova-Moan (2017) in the Netherlands, Edwards (2020) in Norway, and Harding et al. (2020) in Iran.

The emphasis in the studies lies in comprehending LPP as a discursive construct that is entrenched within text and practice, eventually manifesting in empirical evidence. For instance, Rasti’s (2023) investigation of Iran’s foreign language education policy illustrates how the demand for global communication skills was justified, concurrently portraying social subjects as passive recipients of the policy. The transformation towards communicative competence was depicted as an achievement of policymakers, thereby legitimizing their actions. While Rasti’s study did not delve into the impact of this educational policy on testing cultures in Iran, it did provide crucial insights into how the policy implementation process establishes its credibility.

The research conducted by Edwards (2020) and Bubikova-Moan (2017) showcased how policy discourses influence the actions of social agents. Edwards (2020) investigated the implementation of English as the medium of instruction in higher education, where institutions had to navigate conflicting interests. They had to comply with state regulations that aimed to preserve Dutch while simultaneously promoting internationalization through English-medium instruction. By adhering to a language code of conduct, institutions strategically positioned themselves as responsible actors amidst these divergent interests. Similarly, Bubikova-Moan’s (2017) study analyzed two policy documents concerning children of non-Norwegian ethnolinguistic heritage in Norway, published 15 years apart. This analysis revealed a shift in discourse from promoting a group-based pluralist framework to emphasizing the importance of a proficient command of the Norwegian language for achieving educational and social equity. These studies emphasize the need to perceive language education policy as a discursive construct that extends beyond mere text and significantly impacts society.

In the following paragraphs, we will narrow our focus further, transitioning from language education policy to language testing policy. We will explore why the discursive approach is equally crucial for analyzing language testing policy. Language testing issues often intersect due to the power dynamics that affect various stakeholders. Language tests, often designed for specific purposes or used to legitimize the intentions through standardized scores, serve to concretize policies and set objective thresholds. Considering the interconnection between tests and policies, it is imperative for further research to integrate language policy and testing to comprehend tests beyond their apparent objectivity, delving into their broader historical and societal dimensions. In this study, policy enactment driven by language tests is deemed as LTP. Moreover, this study posits that LTP enactment is a discursive process that must also be examined from the discursive approach, as advocated by Barakos and Unger (2016).

The enactment of language policies driven by high-stakes language tests is influential because of the power of testing. Language tests should not only be regarded as pedagogical tools but also as sociopolitical instruments or decision-making tools, with substantial impacts on different stakeholders (Shohamy, 2006). The widely referred to example is the “No Child Left Behind” policy, which was implemented in the USA in 2001 and led to explicit sanctions, including the closure of underperforming schools, student transfers to other schools, and reductions in school funding. The implementation of language tests for immigration policies is another example that is frequently mentioned. These tests serve not only to regulate immigration but also to assess eligibility for residency, naturalization, and citizenship across the globe, as noted by Extra et al. (2009), Hogan-Brun et al. (2009), and Milani (2008).

Numerous researchers have focused on discursive approaches to test development and validation due to the social impact of language tests. Building on Messick’s (1989) concept of test validity, for example, McNamara (2001) highlighted that sociopolitical values are embodied in the development and administration of language proficiency tests. The practice of testing should be seen in the arena of contestable social and cultural values, as well as “in terms of the discourses within which language testing have their meaning” (McNamara & Roever, 2006, p. 199). Furthermore, language testing requires cultural and critical policy analysis, which frequently involves exploring questions about history and ideology and utilizing qualitative and discursive interpretations (McNamara & Ryan, 2011).

While expanding on discursive approaches to the test development and validation processes, LTP must be conceptualized and examined as a sociopolitical process. The DALTP emphasizes the critical role of discourse in shaping language testing policy, which is discursively created, implemented, interpreted, and re-situated (Barakos, 2012), as a form of social practice. As discourses in LTP dialectically interact with the material world, educational stakeholders, even the public, see policy meaning as, and through, discourses. The double functions of discourse as (1) the objective of LPP and (2) a vehicle for (re)constructing discourse should be highlighted more. “Discourse” in DALTP should be understood as both a medium that facilitates social practices affecting individuals and as an outcome with the power to influence those individuals.

Critical discourse analysis as a discourse-analytic framework

From a methodological perspective, various DAs have been posited in the literature. For example, the ethnographical approach to policy enactment has been welcomed because it is expected to address the gap by clarifying the agencies’ voices (Hornberger & Johnson, 2007; Johnson & Johnson, 2015). Critical discourse analysis (CDA) (Barakos, 2016; Johnson, 2011, 2015) offers a comprehensive structure for examining how language policy has evolved over time and across different locations (Savski, 2016).

CDA researchers can analyze how policy activities at the local level relate to macro-level texts (such as government policy documents) and historically constituted discourses (Wodak & Meyer, 2009). Wodak and Meyer (2009) categorized the CDA into various theoretical orientations, such as Fairclough’s (1989, 1992) “dialectical–relational approach,” van Dijk’s (1988) “sociocognitive approach,” van Leeuwen’s (2008) “social actors approach,” and Reisigl and Wodak’s (2016) “discourse-historical approach.” Among these approaches, Fairclough’s (2016) CDA takes the stance that semiosis/language, discourse, and social structure are in a dialectical relationship.

Defining discourse as a social practice, Fairclough (1989) argued that the language choice of a speaker or writer is primarily related to the social context of interpretation and production. Fairclough visualized three levels of social organizations, as depicted in Fig. 1. These levels include the immediate social context, social institutions that provide a larger framework for the production and interpretation of discourse, and society as a whole. Fairclough’s intention was to emphasize the social reality in which texts are embedded. Text is an integral part of a socially conditioned process. Discourse and social structure as a social practice are in dialectical relations that contribute to producing events and actions and lead to social change.

Fig. 1

Discourse as text, interaction, and context (as cited in Fairclough, 1989, p. 25)

The text in the innermost rectangle relates to formal features, such as vocabulary and grammar. Texts are linguistically construed within social organizations; therefore, they function as traces and cues for understanding how they are produced and interpreted in the discourse practice. Textual practice traces how texts are represented for meaning-making resources. This emphasis on linguistic cues underlines the importance of the discursive formation of language tests into entities that ultimately exert practical power. The outer square in the middle is the interaction between text and context that captures discourse practice. The meaning-making process is captured by focusing on the intertextuality among genres, styles, and discourses of texts. Finally, the outermost rectangle is a context that aims to explain social practices for the schematic framework, or ideology, which in turn enables textual and discourse practice.

Fairclough (1989) outlined three dimensions of discourse, each requiring a different analytical approach: description, interpretation, and explanation. In the “description” stage, formal and semantic features of the texts, such as vocabulary choice and grammatical structures, are analyzed. Halliday’s (2014) systemic functional grammar is a useful tool for understanding the semantic functions of texts. During the “interpretation” stage of iterative interpretation, the formal features of descriptive texts are re-examined based on the members’ resources in the process of social interaction. In the “explanation” stage, the social effects of the texts on production and interpretation are discussed within the specific social context. Discourse is viewed as both constitutive of and constituted by social practice, and CDA researchers explain it as a form of social practice using relevant social theories.

Fairclough (2003, 2016) highlighted three semiotic (or discourse-analytical) methods, wherein texts are embedded in the elements of social practices and social events. Genres, discourses, and styles mediate between social practices and social events and reflect the semiotic networks of social practices or intertextuality. Genres are linked to the “methods of acting and interacting.” Genre conventions determine how text is arranged and presented and influence the semiotic features of interaction. For example, different genres, such as newspaper articles, policy documents, academic reports, or interviews as speech data, will entail different rules or conventions for production and interpretation. Discourses are a “semiotic way of construing aspects of the world (physical, social, or mental) which can generally be identified with different positions or perspectives of different groups of social actors” (Fairclough, 2016, p. 11). Discourses are represented through the textual formation and offer a means to interpret the process of meaning-making, as they are materialized from their stances and identities. Styles are “ways of being,” and an individual’s stylistic repertoire typically reflects their epistemological stance among interlocutors. For instance, depending on how writers identify themselves, the overall stance of editorials can be didactic or affiliating to the readers.

The overt focus on text analysis, however, has led to this approach becoming vulnerable to criticism that the voices of stakeholders are disregarded, thereby undermining the ideological motivation. As Fairclough (2016) was also aware of the limitations, he also noted that dialectical-relational perspectives allow for economic and sociopolitical context to be included in the semiotic analysis, and the analyses enable researchers to understand the dialectic relations between structure (social practices) and events (actions). Moreover, while the term “critical” might suggest that the methodology falls into the third wave of LPP (refer to “Development of language planning and policy” section), Fairclough’s (1989, 2003, 2016) work has consistently acknowledged the dialectical relationship between texts and social practices, highlighting the essential role of semiotic networks in shaping social practices. This could situate Fairclough’s CDA as a DA approach within the fourth wave of LPP.

Only a limited number of studies have employed critical discourse analysis as a method for investigating the multilayers of language testing policies. The subsequent discursive analysis of language tests, along with their respective findings, demonstrates the adaptability of this approach in various global contexts. For instance, Milani’s (2008, 2009) research in Sweden focused on the proposal by the Swedish Liberal Party to implement a language test for naturalization. Employing critical discourse analysis on policy documents and newspaper articles, Milani (2008) scrutinized the overt and covert ideologies behind the language test for citizenship. Despite its stated aim of promoting social cohesion, the policy texts discursively associated language proficiency with cultural knowledge, treating it as a prerequisite for citizenship. This discursive practice effectively heightened social differentiation, thereby legitimizing the exclusion of certain groups from the civic and symbolic domains in Sweden.

In another study, Millar (2013) conducted a critical discourse analysis of the Canadian policy discourse concerning language and immigrant integration to intensify the motivation for the increased emphasis on language in the policy, Through an interdiscursivity analysis (Fairclough, 2003), the study revealed how elements from neoliberal skills discourse, an academic discourse of “language as communicative competency,” and a research discourse of “language as a factor of immigrant incorporation” were recontextualized within the Canadian immigration regime. The results indicated that changes in language requirements and language testing within Canadian immigration policy were driven by and interconnected with a neoliberal transformation of Canadian immigration and integration policy.

Additionally, Harding et al. (2020) examined the emergence of the “Secure English Language Test,” officially recognized by the UK Home Office for immigration and visa purposes in 2010. Through critical discourse analysis of the official tender documents, the researchers explored the discursive construction of the test and different social actors in relation to the concept of “securitization” (Buzan et al., 1998). Immigrants were represented as customers, bidders, and providers rather than identified as test-takers, testing agencies, or government bodies. Furthermore, immigrants were often portrayed as potential threats to national peace, particularly those engaging in unethical testing behaviors, as they were discursively framed as disruptors of national tranquility. The study revealed that security was a prominent theme in the tender, with prospective bidders required to adhere to detailed security requirements and supervise subcontractors. Social actors, spaces, objects, policies, and procedures were consistently portrayed in terms associated with securitization.

Three research contexts where DALTP were adopted

NEAT-driven reform policy in secondary school contexts

The first example where LTP was implemented in Korea is the NEAT, an English proficiency test developed and executed by the Korean government as a domestic test to measure English speaking, writing, reading, and listening skills. A public need for communicative competence in English language abilities were noted, and the government responded by creating a new English test. The NEAT, especially levels 2 and 3, could exert influential power on individuals in determining their performance on college admissions. Therefore, the introduction of the NEAT was one of the educational reform policies driven by high-stakes testing in secondary school contexts. The English subject curriculum primarily focused on reading and listening and necessitated a shift towards incorporating communicative competences. This curriculum had transformed the culture of English learning and teaching in public schools. However, a new test led to discursive conflicts (Shin & Cho, 2020) and was then abolished rather abruptly.

Drawing upon Fairclough’s (2003) critical discourse analysis on the relevant policy documents and newspaper articles, the following studies have demonstrated that test-driven policy enactment was planned within the discursive relationship among policy documents and the production and interpretation of stakeholders within the social structure. The discursive analytic approach to the introduction and abolishment of the NEAT (Shin, 2012, 2019, 2023; Shin & Cho, 2020, 2021) revealed that the process was created, legitimized, and contextualized in and through discourses and mediated by the newspapers and policy document.

The use of the DALTP indicates that the test was not developed in a vacuum but rather materialized from the discursive approach to language policy. The social practice of NEAT-driven policies led to the related texts being considered socially constitutive and the discourses being shaped by the sociohistorical contexts. The textual analysis of relevant policy documents and newspaper articles has shown that the test-driven policy enactment was planned within the dialectic relationship between stakeholders and policy documents (Shin, 2012, 2019; Shin & Cho, 2020, 2021). The process of test development, implementation, and abolishment was the product of negotiation.

The underlying rationale for the introduction of the NEAT by the Korean government showed “bureaucratic ideologies” (Weber, 2015). For instance, the test was introduced under the guise of technology-oriented practice, where the test was represented as a technology to resolve the current educational problem (Madaus, 2001). The test was legitimized by highlighting the cost and benefit of developing a new test, thereby appealing to economic discourses. The domestic test was considered cost-effective for the nation, particularly when compared with the foreign tests, as well as for the private expenditure of each household. The utilitarian approach also claimed that the domestic test was more convenient and easier to access than the imported test. These discourses were accepted within the underlying social structure, where technology is regarded as an objective, apolitical, value-free panacea to resolve the current shortcomings in English education. Reckless belief in the test and naïve perspectives on the positive washback effect of the test also justified the teach-to-the-test culture.

These analyses were made available by examining the policy documents drawing upon analyses related to genre, discourse, and style. Policy documents as a genre followed the conventional practice of reductive-positivist methods for resolving social problems, which resulted in simplifying the multilayered sociopolitical issues in English education into individualized and manageable problems. To improve the communicative competence of the public, the state initiated a test that would lead to washback effects on public schools. Two discourses on policy documents revealed their attitudes and stances towards languages and tests. First was that English language competence was a reifiable entity that can be objectified with constructs, such as “practical English ability.” Second was the test implementation procedure as technology for achieving the goals efficiently. Styles are the means through which producers of policy documents identify themselves. The non-animate, agentless sentence structure erased the process (e.g., the introduction of the NEAT) hiding explicit agents. Frequent use of measurement registers evoked credibility of the style, and obligatory modalities showed their stance of having the upper hand. These styles contributed to legitimizing the government’s voices.

Although the intertextuality between media and policy discourse has not been scrutinized in the literature, Shin (2019) and Shin and Cho (2020) have revealed that the media contributed to policy enactment and functioned as one of the most important sites for public discourses to be propagated and contested. The media managed a site for discourses to be mediated by the newspaper’s orientation. Shin and Cho (2020) analyzed two newspapers (Chosun Ilbo and Hankyoreh) considered to be on opposite ideological sides. Chosun Ilbo arranged texts from market principles, thereby highlighting the importance of fulfilling supply and demand, for which the state ought to intervene in the test market and for which a test should replace another. Conversely, Hankyoreh focused on the effectiveness of tests, thereby casting doubt on the household expenditure for test preparation and the loss of national funds from NEAT. Hankyoreh also doubted the effectiveness of tests in normalizing public education (e.g., making the classes more interactive), thereby leaning towards the abolition of the test.

While the discursive approach to NEAT effectively maps intertextual relationships among various discourses, it exhibits two notable shortcomings: it does not thoroughly delve into the dialectic impact of policy changes on individuals affected by them, nor does it investigate the ways in which people navigate language requirements within power dynamics. An ethnographic investigation could enhance our understanding by revealing how a government announcement regarding a new test initiates a chain reactions among teachers, students, and parents, for instance. Fieldwork could shed light on how students and teachers exercise their agency by opting for alternative English tests over NEAT for college admissions. The necessity for such ethnographic insights will be further elaborated upon.

TOEIC for English graduation benchmark policy in higher education

The second example is TOEIC for English graduation benchmark policy in higher education. During the mid-1990s, the Korean government’s declaration of “Segyehwa” (literally, “globalization”) had begun to influence the education system and society. The government-led discourse on globalization went beyond slogans and created actual policies that increased Korea’s national competitiveness in the global market and encouraged citizens to be more competitive.

Large corporations, such as Samsung, LG, and Hyundai, overhauled their hiring practices and required applicants to demonstrate their work-related English language skills. TOEIC has been one of the popular instruments designed for evaluating practical English-language communication skills required in the workplace. During the late 1990s, many universities adopted TOEIC as a graduation requirement for their students, arguing that it would enhance the global competitiveness of individuals and universities in the job market. According to the survey conducted by the Korea TOEIC Committee (2021), 112 out of 219 universities nationwide have continued to use TOEIC scores as a graduation requirement, as of January 2021.

As there has been limited research on the policy of English graduation benchmarks, Park and Shin (2016) investigated the policy in relation to policy power and interpretative agency, recognizing it as a multilayered phenomenon that is shaped and implemented in discourse (Barakos & Unger, 2016). Drawing on Fairclough’s (1989) CDA as a research methodology, they examined how the related discourses (re)produced a belief about English language requirements for graduation in Korean universities. Therefore, they focused on the media discourse because it reflected and engendered ideologies about English language learning and testing that, in turn, encouraged specific perceptions of testing-led policies as common sense—or as “just the way things are” (Johnson, 2013, p. 140)—and proliferated dominant discourses.

Park and Shin (2016) collected 86 newspaper articles from three major Korean newspapers, namely Chosun-ilbo, JoonAng ilbo, and Dong-a ilbo, starting with the government’s official declaration of “Segyehwa” in November 1994 to June 2016, which marked the end of the study period. They adopted Fairclough’s relational perspective of texts and analyzed relations between three levels: internal, external, and interdiscursive relations (Fairclough, 2003). The analysis of internal relations showed that different lexical and rhetorical strategies, such as rewording or overwording, synonyms and antonyms, collocation, and metaphors, contributed to creating coherent vocabulary relations. Modal verbs of obligations (e.g., “must,” “should,” “ought to”) and legitimation strategies for authorization and rationalization (van Leeuwen & Wodak, 1999) also contributed to forming consistent semantic relations to the texts.

An analysis of interdiscursive relations was conducted using the elements (genres, styles, discourses) of “orders of discourse” in order to identify the link between the text per se and the social context. The genres of report and interview dominated the newspaper articles. Regarding the styles, however, the authoritarian voice of the presidents was overwhelming in the genre of the interview, whereas the voices of different news sources, such as presidents, professors, students, academic affairs deans, teaching staff, and employment managers, appeared in the genre of the report. The texts of the graduation benchmark policy were embedded in the following themes (discourses): “English language as (controllable) entities,” “language incompetence as a problem,” and “English proficiency as an object (to be conquered).” After the English language competence of controllable entities was identified to be deficient and problematic, it was then noted that English language competence should be managed and conquered through testing instruments, such as TOEIC.

The external analysis results indicated that the discourse of “deficiency” and “problem” was associated with monolingualism or linguistic determinism. Considering monolingualism as a norm and allowing this norm to prevail over the media texts led to a social system that perceives language as the primary cause of problems. According to Harrison (2007), the discourse on “lack of language skills as a barrier” (p. 75) leads to another discourse that language deficiencies reduce work efficiency and impede social integration. Consequently, the target language should be taught, learned, tested, and corrected. This leads to silencing, or even disparaging, other language resources, life experiences, or other social variables that an individual already possesses and presumes the need for monolingualism as a solution.

Although English was not specified as the language of a majority or that of the official language in the media discourse, the policy texts highlighted that native-like proficiency in English prepared college students for improved employment opportunities in the era of globalization. The consistent arrangement of texts validated the policy, thereby normalizing the power of English (language testing). At least in the space where the English benchmark policy was implemented, the English proficiency requirements can be considered a result of the discursive intervention that English should be efficiently utilized, educated, and tested to resolve related problems. If such discourses are reinforced, “monolingual habitus” will be legitimized, thereby recognizing only one linguistic norm as valid in society or school (Gogolin, 1997, p. 41).

This study holds significance as it delves into the media discourse regarding the English graduation benchmark policy within Korean universities. However, the creation and implementation of a language policy open up to a range of interpretations, not solely confined to its creators but also extended to those expected to implement and use it (Johnson, 2013). Consequently, researchers should take into account the diverse array of actors involved in language policy, including school administrators and students, and their potential to exert influence or establish dominance within these contexts. Furthermore, empirical findings derived from ethnographic investigations of this policy will provide insight into the process of language policy development and explore the interconnections across multiple layers, spanning from policy formulation to practical implementation (Hornberger & Johnson, 2007; Johnson & Ricento, 2013).

TOPIK as a high-stakes decision-making tool for foreign residents

The third example is TOPIK as a high-stakes decision-making instrument for immigrants in Korea. Similar to the way in which national language tests have played key roles in collecting information and decision-making for controlling foreign residents (Extra et al., 2009; Harding et al., 2020; Hogan-Brun et al., 2009; Milani, 2008, 2009; Millar, 2013), the TOPIK has wielded significant influence among foreigners living in Korea. The number of immigrants has increased due to government-led immigration policies since the 2000s. According to the Korea Immigration Service (2020), 2,524,656 foreigners were residing in the country, accounting for approximately 5% of the total population. Over the past 5 years, the number of foreign residents has grown at a rate of 7.4% on average each year. Korean language proficiency tests play an important role in gatekeeping to determine the rights to residency, naturalization, and citizenship. Among them, TOPIK is the most influential test for foreigners in educational and social settings.

TOPIK is a written test designed to test Korean language proficiency in reading, listening, and writing comprehension. The National Institute for International Education (NIIED), which is under the direct control of the Ministry of Education, organizes and supervises the test. According to the NIIED (2020), since the TOPIK’s first implementation in 1997, the cumulative number of applicants exceeded 2 million in 2018, and 375,871 people took the test in 2019 alone. TOPIK results are commonly used to aid in high-stakes decisions regarding foreign residents for college admission, employment, visa issuance, permanent residency, or naturalization application.

Drawing on Fairclough’s CDA, Park (2020) examined how the TOPIK has been discursively constructed as a high-stakes decision-making instrument for immigrants in Korea. She collected 377 articles from five major newspapers since 1995. The collected data was analyzed using Fairclough’s (1989) model and was discussed in five stages in accordance with the changes in Korean language testing policy: test development, policy introduction, implementation, expansion, and enrichment.

In the description of policy texts, different linguistic features, such as rewording, metaphor, modes, modality, and foregrounding, were presented at each stage. Different discursive practices were interpreted in this way. The TOPIK development was justified by the “Korean-style globalization” discourse. In the introductory period of TOPIK-led policies, the “languages as problem” discourse that problematized the lack of Korean proficiency among immigrants, and the “test as a prescription” discourse wherein social problems can be resolved through testing policies, appeared. In the implementation stage, the “language as a requirement” discourse emerged, as standard Korean language proficiency was assumed to be a prerequisite for participating in Korean society. In the expansion stage, the “national language” discourse legitimized the urgent need for the development of TOPIK speaking, academic TOPIK, and junior TOPIK. In the enrichment stage, the TOPIK had been used as a requirement for permanent residency and naturalization. The “test as a precaution” discourse, in which language testing policies could prevent immigrants from causing problems, was introduced in the media. The neoconservative ideology (Apple, 2004), which can play a crucial role in teaching “correct” norms, knowledge, and values in a national and official language, was deeply ingrained in the policy discourses. The TOPIK-led policies were discursively used to regain the “correct” form of Korean language knowledge, as well as to resolve various social problems, such as exclusion, alienation, underachievement, mental disorders, and domestic violence.

The significance of this study lies in its examination of the media discourse concerning the Korean language testing policy for immigrants. It is meaningful to undertake additional research that examines intertextual and interdiscursive connections between media and other related documents within each specific domain. TOPIK holds a prominent position, serving as an important instrument in high-stakes decisions affecting migrants, such as admissions, employment, residency status, permanent residence, and naturalization. DALTP research can engage in further investigatioins into the interplay between macro-level discourses present in government policy reports and press releases, and micro-level discourses observed in interviews with key policy stakeholders, including test takers and administrators.

We have now examined exemplary research works in the Korean context, affirming its relevance in language testing policy studies. DALTP is highly anticipated to enrich the field by highlighting the constructive power of discourse in facilitating political communication among local actors. It is essential to collect and interpret the research texts from diverse stakeholders for a better understanding of how much DALTP has the potential to apply various analytic forms into the development and validation research of language testing policy. As its significance needs to be enhanced by the critiques from various approaches in different contexts of test-driven policy implementation, the examination should not be restricted to just one particular framework, like Fairclough’s model mainly discussed in this study.

Conclusion

The growing discussions on critical and discursive approaches to LPP can be attributed to the global migration in the era of postmodern globalization. The movement of people for various reasons, such as employment, immigration, naturalization, education, marriage, and tourism, leads to the convergence of diverse races, ethnicities, capital, knowledge, and technology. In this interconnected world, where various media channels play a crucial role, there has been a need to re-evaluate the underestimated area of research, “language”-related requirements in testing, and policies. In this context, the current study has explored the meaningfulness of adopting discursive approach to “language” testing policy. By using DA, the understanding of LTP, often implemented through high-stakes language testing, can be enhanced.

The empirical studies (the English education reform through the adoption of speaking and writing sections in the NEAT, the implementation of TOEIC as a graduation benchmark requirement in higher education, and the use of TOPIK as a high-stakes decision-making tool for foreign residents) discussed in this paper exemplify that LTP should be viewed as both critical and discursive practices. Adopting the DALTP can delve into the constitutive power of discourse, thereby uncovering dominant social structures and their impact on individual discourses, perpetuating these structures. Furthermore, DALTP can provide a valuable framework for analyzing intertextual discourses in interdisciplinary research, investigating how certain discourses prevail or fail in ideological battles.

The implications of adopting DALTP are significant. It allows a deeper understanding of the variability in LTP contexts by acknowledging the dialectic relationships between policy as text and discourse within a power network. LTP, in this case, is no longer seen as a top-down, deterministic process where structural approaches tend to prevail. Instead, it is viewed as a dialectic relationship where policy text and discourses on the policy interact and influence each other (Barakos & Unger, 2016; Johnson, 2004, 2016). Moreover, by examining unfolding discourses on different LTPs, DALTP contributes to democratic pluralism (Tollefson, 2013) by aiding government agencies in understanding standard (official) language policies, helping local organizations preserve dialects, assisting schools in reviewing languages to teach, guiding enterprises to focus on multilingual resources, and supporting civic groups in advocating for the language rights of minorities.

Despite being in the nascent stage, DALTP, however, is in need of extending its scope beyond considering media and policy text as the sole source of discourse. While earlier studies on DALTP have made strides in understanding the multilayered and sociopolitical aspects that influence the discourse formulation, future studies could benefit from engaging with these members involved. Integrating ethnographic studies into DALTP could enhance the analytic insights by offering a more nuanced framework, capturing the interplay between discourses and social actors in shaping language education policy. By doing so, researchers would be able to trace the impact of policies directly on those affected, deeply exploring members’ perceptions, attitudes, and values in the context of discourses on test-driven language policies. Incorporating ethnographic studies could offer profound insights into how individuals affected by language testing policies interpret, apply, and even exercise their agency to shape situated practices (Pérez-Milans and Tollefson, 2018; Johnson, 2011; Johnson & Ricento, 2013). This approach would further enrich our understanding of the complex, dialectical relationship that exists among various levels of policy texts, discourses, and practices.