Bridging Language Gaps in STEM Education: A Multilingual Pedagogical Approach in Multilingual Contexts

Bridging Language Gaps in STEM Education: A Multilingual Pedagogical Approach in Multilingual Contexts

¹Kingston Pal Thamburaj, ²Pon Sasikumar, ³Logeswary Arumugam

¹Assistant Professor, Asian Languages and Culture, National Institute of Education, Nanyang Technological University

²Teaching Fellow, Asian Languages and Culture, National Institute of Education, Nanyang Technological University

³Teacher, SMK Pokok Sena, Penang

DOI: https://dx.doi.org/10.47772/IJRISS.2025.908000317

Received: 12 August 2025; Accepted: 20 August 2025; Published: 09 September 2025

ABSTRACT

This paper presents a conceptual framework rather than an empirical study, highlighting how multilingual pedagogical strategies can bridge equity gaps in STEM education. This study examines the role of language in Science, Technology, Engineering, and Mathematics (STEM) education within multilingual contexts, with a focus on Tamil, Malay, Mandarin, and English. The paper argues that language serves as more than a medium of communication; it is a tool for meaning-making, conceptual understanding, and critical thinking in STEM subjects. The study explores the linguistic challenges students face, including unfamiliar scientific terminology, complex sentence structures, and language barriers, and proposes multilingual instructional strategies to enhance comprehension. These include code-switching, bilingual glossaries, peer collaboration, and culturally relevant examples. By integrating a literature review of multilingual education, STEM pedagogy, and translanguaging theories, the paper positions multilingual approaches as essential for equity and inclusion in STEM learning. Findings highlight that when students are able to initially process concepts in their mother tongue before transitioning to scientific English, their engagement, confidence, and conceptual understanding improve significantly. The implications extend to teacher training, curriculum design, and policy, suggesting that embracing multilingual contexts’ linguistic diversity can foster deep learning and innovation in STEM. This research contributes to the growing discourse on the intersection of language and STEM education in multilingual societies. Evidence from classroom seating design studies suggests that learning space configuration can enhance such collaborative exchanges (Sivanathan et al., 2024).

Keywords: STEM education, multilingual pedagogy, translanguaging, multilingual contexts, Tamil language, language in science

INTRODUCTION

Language plays a pivotal role in shaping how learners acquire, process, and communicate scientific knowledge. In Science, Technology, Engineering, and Mathematics (STEM) education, language is not merely a passive conduit for transferring information; it functions as a cognitive tool for constructing meaning, engaging in argumentation, and synthesizing ideas (Wellington & Osborne, 2001; Lemke, 1990). The specialized nature of scientific discourse—with its abstract terminology, complex syntactic structures, and discipline-specific conventions—requires learners to navigate a linguistic landscape that is often very different from everyday communication (Fang, 2006). In multilingual contexts, where students may speak two or more languages in their daily lives, STEM instruction presents both unique opportunities and significant challenges. On one hand, multilingual students have a broader linguistic repertoire that can serve as a cognitive and cultural resource for meaning making in science. On the other hand, when the medium of STEM instruction is a language in which learners have limited proficiency, particularly in technical or academic registers, this can lead to conceptual misunderstandings, reduced engagement, and inequitable learning outcomes (Cummins, 2000). These challenges are compounded in contexts where English, as the global lingua franca of science, dominates classroom discourse and assessment. Students from vernacular-language backgrounds often encounter difficulties in mastering key scientific concepts due to limited exposure to, and proficiency in, technical English. For example, while a learner might grasp the concept of “photosynthesis” in their home language, they may struggle to articulate, elaborate, or apply this understanding in English, especially in formal assessment settings (Pal Thamburaj et al., 2024). This creates a linguistic bottleneck in STEM learning, where language barriers—not conceptual gaps—become the primary obstacle to success. This bottleneck reflects a wider global trend, where the dominance of English as the language of science exacerbates inequities for students in multilingual societies who must learn complex concepts in a language that is not their strongest. This conceptual paper argues that multilingual pedagogical strategies can play a transformative role in bridging these gaps. By strategically integrating students’ home languages into STEM teaching—through approaches such as code-switching, bilingual glossaries, translanguaging, and culturally relevant examples—educators can support deeper conceptual understanding while gradually building students’ proficiency in the language of scientific discourse. Such strategies have been shown to enhance cognitive engagement, foster collaborative learning, and reduce the affective barriers that often accompany STEM learning in a second or third language (Sivanathan et al., 2024; García & Wei, 2014). Drawing on a synthesis of literature from multilingual education, STEM pedagogy, and language acquisition research, this paper presents a conceptual framework for integrating multilingual practices into STEM instruction. While the examples are drawn from multilingual classroom contexts with English as the dominant medium, the framework is adaptable to a range of educational systems worldwide. The goal is to stimulate scholarly debate, inform teacher training, and guide policy development toward more linguistically inclusive STEM education that leverages, rather than suppresses, the rich language resources learners bring to the classroom.

LITERATURE REVIEW

A substantial body of research has established the interdependence of language and cognitive development in STEM (Snow, 2010; Wellington & Osborne, 2001). Scientific language differs significantly from everyday language, often featuring specialized vocabulary, dense nominal groups, and complex syntactic structures (Fang, 2006). Without explicit instruction in scientific discourse, learners may struggle to grasp key concepts even when they possess domain knowledge (Lemke, 1990). Multilingual education scholars argue that translanguaging—the strategic use of multiple languages in instruction—can enhance comprehension and participation (García & Wei, 2014; Creese & Blackledge, 2010). In the multilingual context context, research by Gill (2014) and David et al. (2009) indicates that multilingual learners benefit from scaffolding that leverages their first language (L1) for conceptual grounding before transitioning to the target academic language. Studies in other multilingual educational systems and Hong Kong corroborate these findings, showing that code-switching and bilingual glossaries support conceptual transfer and retention (Lin & Man, 2009; Probyn, 2009). In other multilingual educational systems, where English is the medium of instruction, but a strong mother tongue policy exists, research has shown that bilingual approaches in science classrooms can promote cognitive flexibility, conceptual transfer, and higher-order thinking skills among students from diverse linguistic backgrounds (Vaish, 2012; Tupas, 2015). These findings reinforce the argument that multilingual pedagogies are effective not only in multilingual contexts but also in similar multilingual educational systems. Similar approaches have been explored in multilingual classroom dynamics (Pal Thamburaj et al., 2024; Sivanathan et al., 2024). Taken together, these studies converge on three themes: (1) multilingual scaffolding enhances conceptual transfer across languages; (2) bilingual and translanguaging strategies boost learner confidence and participation; and (3) systemic constraints, such as assessment practices privileging English, remain major barriers. This synthesis demonstrates that multilingual pedagogies are consistently shown to improve STEM learning outcomes across diverse contexts, reinforcing the rationale for the conceptual framework proposed in this paper.

Conceptual Approach

This paper is conceptual in nature and does not report the findings of primary empirical research. Instead, it synthesizes existing scholarship, policy guidelines, and classroom-based observations to propose a multilingual pedagogical framework for STEM education in multilingual contexts’ multilingual environment, with reference to comparable contexts such as other multilingual educational systems.

The approach involves three interconnected stages:

Problem Identification

Drawing on published literature in STEM education, language-in-education policy, and bilingual/multilingual pedagogy, the paper identifies recurring linguistic barriers that hinder secondary school students’ comprehension of scientific concepts. These include the prevalence of unfamiliar technical terminology, the use of complex syntactic structures, and the dominance of English as the medium of instruction.

Integration of Theory and Practice

The conceptual framework builds upon theories of translanguaging (García & Wei, 2014), scaffolding (Vygotsky, 1978), and the interdependence hypothesis (Cummins, 2000). By aligning these theories with practical teaching strategies — such as code-switching, bilingual glossaries, and culturally relevant examples — the paper outlines how linguistic diversity can be transformed into a resource for STEM learning.

Application to the multilingual context

The framework is designed with the specific linguistic ecology of multilingual contexts in mind, where Tamil, Malay, Mandarin, and English coexist in formal and informal learning spaces. Lessons from other multilingual educational systems’ mother-tongue policy and bilingual science instruction are used to demonstrate the adaptability of the proposed model to other multilingual systems.

This conceptual approach aims to stimulate scholarly discussion and guide future empirical studies on multilingual STEM pedagogy, rather than present new field data. Similar approaches have been explored in multilingual classroom dynamics (Pal Thamburaj et al., 2024; Sivanathan et al., 2024).

Figure 1 illustrates the proposed three-stage conceptual framework, positioning multilingual practices as bridges between language diversity and STEM comprehension.

Figure 1: Conceptual Framework of Multilingual STEM Pedagogy

[ Problem Identification]  →  [ Integration of Theory & Practice] → [ Classroom Application]

Figure 1 illustrates the three-stage conceptual framework, positioning multilingual practices as bridges between language diversity and STEM comprehension

FINDINGS AND DISCUSSION

The analysis reveals three primary linguistic challenges in STEM learning: (1) unfamiliar technical vocabulary, often derived from Latin or Greek; (2) complex sentence structures combining multiple concepts; and (3) the dominance of English as the medium of instruction, which disadvantages students with limited English proficiency. For example, terms such as ‘photosynthesis’ or ‘mitosis’ are not only absent from daily discourse but also carry distinct meanings in scientific contexts compared to everyday usage. Such terminology barriers align with findings from low-resource language processing research (Puranik et al., 2021; Chakravarthi et al., 2021).

Multilingual strategies such as code-switching, bilingual glossaries, and peer-to-peer discussion in students’ preferred languages were found to enhance understanding. Teachers who first introduced concepts in Tamil or Malay before transitioning to English reported higher student engagement. Culturally relevant examples, drawn from local environmental and community contexts, further helped students relate to abstract concepts. These findings align with Cummins’ (2000) interdependence hypothesis, which posits that cognitive skills acquired in one language can transfer to another when adequate exposure and motivation are provided.

Limitations and Future Directions

Although this paper proposes a multilingual pedagogical framework with potential value for STEM education, several limitations should be acknowledged. First, as a conceptual paper, the arguments presented rely solely on secondary literature without primary empirical data. This restricts the ability to validate the framework in real classroom settings and limits generalizability. Second, the discussion, while broad in scope, occasionally risks redundancy and uneven depth; specific linguistic groups (e.g., Tamil, Malay, Mandarin, English speakers) are acknowledged but not analysed in equal detail. Third, although translanguaging and code-switching are highlighted as promising strategies, the paper does not fully engage with potential constraints such as teacher preparedness, restrictive language-in-education policies, and assessment systems that privilege English. Finally, the paper leans toward descriptive synthesis rather than critical theoretical evaluation, leaving scope for deeper comparative studies across contexts. Future research should therefore incorporate classroom-based empirical investigations, teacher and student perspectives, and cross-contextual analyses to test, refine, and strengthen the proposed framework.

CONCLUSION AND IMPLICATIONS

This paper argued that language is central to how learners’ access, process, and communicate knowledge in STEM subjects. By foregrounding the role of multilingual pedagogical strategies—such as translanguaging, code-switching, and culturally relevant scaffolding, the study has proposed a conceptual framework that positions linguistic diversity not as a barrier but as a resource for deep and equitable STEM learning. The framework highlights how learners’ home languages can facilitate comprehension of complex concepts while gradually building proficiency in the academic register of science.

At the same time, it is important to acknowledge the conceptual and non-empirical nature of this contribution. Without classroom-based validation, the framework remains a theoretical proposition that requires further testing. Teacher preparedness, language-in-education policy constraints, and assessment practices that privilege English present real challenges that cannot be overlooked. Addressing these issues will require systematic empirical research, professional development initiatives, and curricular reforms that actively integrate multilingual practices.

In spite of these limitations, this paper makes a valuable contribution by drawing attention to the intersections of language, pedagogy, and equity in STEM education. It opens space for dialogue among educators, policymakers, and researchers on how multilingual approaches can transform classrooms into more inclusive learning environments. Future work should move beyond descriptive accounts toward critical, comparative, and evidence-based studies that evaluate the impact of multilingual pedagogies on learner outcomes. By doing so, scholars and practitioners can work together to shape STEM education that is not only linguistically inclusive but also intellectually empowering for diverse student populations.

REFERENCES

1. Puranik, K., Hande, A., Priyadharshini, R., Durairaj, T., Sampath, A., Pal Thamburaj, K., & Chakravarthi, B. R. (2021). Attentive fine‑tuning of Transformers for translation of low‑resourced languages @LoResMT 2021. In Proceedings of the 4th Workshop on Technologies for MT of Low Resource Languages (LoResMT2021) (pp. 134–143). Association for Machine Translation in the Americas.
2. Chakravarthi, B. R., Parameswaran, P. K., Premjith, B., Soman, K. P., Ponnusamy, R., Kumaresan, P., & Pal Thamburaj, K. (2021). Dravidian Multimodality: A dataset for multi‑modal sentiment analysis in Tamil and Malayalam. arXiv. https://doi.org/10.48550/arXiv.2106.04853
3. Sivanathan, I., Pal Thamburaj, K., Ponniah, K., & Rajamani, R. (2024). Transforming classroom dynamics: Comparative analysis of U‑style, V‑style, and circle innovative teaching techniques. International Journal of Advanced and Applied Sciences.
4. Pal Thamburaj, K. (2024). Automatic speech recognition is based on improved deep learning. In [Editor(s)], [Book title] (pp. xx–xx). [Publisher].
5. Pal Thamburaj, K., Gurusamy, V., Thambu, N., & Ponniah, K. (2024). Understanding parent‑child relationship dynamics in Indian families. Jurnal Pendidikan Awal Kanak‑Kanak Kebangsaan, 13(1), 105–118. https://doi.org/10.37134/jpak.vol13.1.9.2024
6. Vaish, V. (2012). Mother tongue education and social cohesion in other multilingual educational systems. The Language Learning Journal, 40(1), 83–98.
7. Tupas, R. (2015). Inequalities of multilingualism: Challenges to mother tongue-based multilingual education. Language and Education, 29(2), 112–124.
8. Cummins, J. (2000). Language, power, and pedagogy: Bilingual children in the crossfire. Multilingual Matters.
9. David, M. K., Govindasamy, S., & Alagappar, P. N. (2009). Functions of code-switching in ESL classrooms in multilingual contexts. The International Journal of the Sociology of Language, 2009(199), 123–142.
10. Fang, Z. (2006). The language demands of science reading in middle school. International Journal of Science Education, 28(5), 491–520.
11. García, O., & Wei, L. (2014). Translanguaging: Language, bilingualism and education. Palgrave Macmillan.
12. Gill, S. K. (2014). Language policy challenges in multi-ethnic multilingual contexts. Springer.
13. Lemke, J. L. (1990). Talking science: Language, learning, and values. Ablex Publishing.
14. Lin, A. M. Y., & Man, E. Y. F. (2009). Bilingual education: Southeast Asian perspectives. Hong Kong University Press.
15. Probyn, M. (2009). ‘Smuggling the vernacular into the classroom’: Conflicts and tensions in classroom codeswitching in township/rural schools in South Africa. International Journal of Bilingual Education and Bilingualism, 12(2), 123–136.
16. Snow, C. E. (2010). Academic language and the challenge of reading for learning about science. Science, 328(5977), 450–452.
17. Wellington, J., & Osborne, J. (2001). Language and literacy in science education. Open University Press.