Translation technology, translanguaging and transknowledging in student academic writing
This article draws attention to the centrality of translation in multilingual pedagogies that include translanguaging and transknowledging with higher education (HE) students in English Medium Instruction (EMI) institutions. At a perplexing time when AI technologies advance, this article provides a background to earlier (southern) research on multilingualisms in education, including the use of human language translation technologies to increase epistemic access to academic texts predominantly written in English. We specifically draw a distinction between more northern perspectives of translanguaging and more southern perspectives that emphasise the central and explicit role of translation in ‘vertical’ translanguaging and ‘transknowledging’. We follow this with a close discussion of a micro-study, the purpose of which is to offer insight into how Chinese-speaking multilingual students engage with academic texts in English through their own human capacity for translation and in interaction with machine translation (MT) and post-editing (PE) of MT to enhance their academic writing (Chang, 2024 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629). In so doing we connect this study to earlier research, pedagogical interventions and theorisation of vertical translanguaging and transknowledging (Heugh, Li & Song, 2017Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. ; Heugh et al., 2022Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35 (1), 89–127. Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35(1), 89–127. ).
Publication history
Table of contents
Introduction
The use of various AI technologies, including human language technology (HLT) or machine translation (MT), has become ubiquitous among students all over the world (Chang, 2022Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. ). Students use this technology when navigating meaningful access to and use of information in a language different from their primary language/s or ones in which they have greatest academic proficiency. Multilingual students have become accustomed to using various translation technologies when attending lectures and participating in tutorials. They also use these to assist their understanding of required academic reading, and in writing academic assessments in higher education (HE) (e.g., Groves & Mundt, 2021Groves, M., & Mundt, K. (2021) A ghostwriter in the machine? Attitudes of academic staff towards machine translation use in internationalised Higher Education. Journal of English for Academic Purposes, 50 , 100957. Groves, M., & Mundt, K. (2021) A ghostwriter in the machine? Attitudes of academic staff towards machine translation use in internationalised Higher Education. Journal of English for Academic Purposes, 50, 100957. ). However, as argued by these authors and in Heugh et al. (2022)Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35 (1), 89–127. Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35(1), 89–127. , academics in the UK and Australia are often resistant to student use of these technologies assuming that they violate university protocols for academic integrity compliance. The latter study shows that students are often penalised when they use MT resources clumsily and/or lack the expertise to navigate between their own human and machine translation capabilities. As both sets of studies argue, students will continue to use these resources regardless, and it is up to educational institutions to discover ways to incorporate them into pedagogies that comply with concerns for academic integrity. We address three questions in the article:
-
How can deliberative pedagogical use of MT and PE advance students’ cognitive and linguistic access to and use of disciplinary knowledge in written academic texts?
-
How can pedagogical use of MT and PE comply with academic integrity?
-
How does understanding student use of MT and PE expand our understanding of the role of translation in (vertical and pedagogical) translanguaging and transknowledging?
In responding to these questions, we revisit earlier research specifically from Southern Africa and Australia. This is research that has been circulating within and in collaboration with scholars in India since 1990. We then link earlier research to Chang’s (2024) (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629 action research micro-study involving six Chinese-speaking, multilingual graduate students at an Australian university. Chang’s micro-study formed part of a larger PhD study that included a parallel but asynchronous, micro-study, conducted in a Taiwanese university, involving four international students who were advanced second language learners of Chinese. Both studies involved pedagogical use of human translation (HT), MT and PE. Each micro-study involved a pedagogical intervention, differentiated and tailored to the needs of two different student cohorts. The Australian-located study was concerned with how students can use HT, MT and PE to advance their academic writing across their disciplines (beyond language), i.e. in an EMI setting. The Taiwan-located study, a Chinese medium instruction (CMI) context, investigated how multilingual students can use HT, MT and PE to advance their second language (L2) learning of Chinese language. We refer mainly to the Australian EMI–located study here.11.PhD fieldwork was conducted in 2020 when COVID and international travel restrictions negatively affected student enrolments in both institutions. Nevertheless, close analysis of a modest number of students’ translation capabilities has rendered valuable data.
This article builds on and follows the first author’s 20 years of teaching international students Chinese and English as additional languages and advanced courses in translation in Taiwan. It draws on the second author’s 35 years of action and field research in multilingual pedagogies in Africa (1990–2025) in collaboration and dialogue with colleagues in South Asia (e.g., Agnihotri, 1995Agnihotri, R. K. (1995) Multilingualism as a classroom resource. In Heugh, K., Siegrühn, A., & Plüddemann, P. (Eds.). Multilingual Education for South Africa, pp. 3–7. Heinemann Educational Books.Agnihotri, R. K. (1995) Multilingualism as a classroom resource. In Heugh, K., Siegrühn, A., & Plüddemann, P. (Eds.). Multilingual Education for South Africa, pp. 3–7. Heinemann Educational Books., 2014 (2014) Multilinguality, education and harmony. International Journal of Multilingualism, 11 (3), 364–379. (2014) Multilinguality, education and harmony. International Journal of Multilingualism, 11(3), 364–379. ; Mohanty, 2018Mohanty, A. K. (2018) The multilingual reality: Living with languages (Vol. 16). Multilingual Matters. Mohanty, A. K. (2018) The multilingual reality: Living with languages (Vol. 16). Multilingual Matters. ). The last 15 years have overlapped with a longitudinal series of studies with multilingual students in an Australian university. This article is furthermore informed by 30 years of research and use of HLT and provision used in South African HE. The purpose in South Africa has been to use translation technologies or MT to redress ongoing historical linguistic and epistemic inequalities in education. These include expectations for majority speakers of African languages to study through EMI alongside minority L1 speakers of English (e.g., Ndebele, 2022Ndebele, H. (2022) Exploring the challenges of information and communication technology localization in South African higher education: a language management approach. International Journal of Multilingualism, 19 (3), 368–382. Ndebele, H. (2022) Exploring the challenges of information and communication technology localization in South African higher education: a language management approach. International Journal of Multilingualism, 19(3), 368–382. ; van Huyssteen et al., 2023van Huyssteen, G. B., Puttkammer, M., McKellar, C., & Griesel, M. (2023) Translation technology in South Africa. In Routledge encyclopedia of translation technology (pp. 373–383). Routledge. van Huyssteen, G. B., Puttkammer, M., McKellar, C., & Griesel, M. (2023) Translation technology in South Africa. In Routledge encyclopedia of translation technology (pp. 373–383). Routledge. ).22.While we draw attention to this research for readers to follow, we do not have the space to offer adequate discussion of this literature here.
Both authors are familiar with prominent applied linguistics’ theory of second language acquisition (SLA), languages education, EMI, and bilingual education that enjoy prominent circulation in Euro-North educational settings. However, our perspectives and first-hand experiences come from contexts identified as postcolonial and ‘Southern’ in decolonial scholarship (e.g., Connell, 2017Connell, R. (2017) Southern theory and world universities. Higher Education Research & Development, 36 (1), 4–15. Connell, R. (2017) Southern theory and world universities. Higher Education Research & Development, 36(1), 4–15. ). In particular, we regard recent discussions of ‘translanguaging’ in Euro-North literature not as an alternative to multilingual education, but rather as a set of practices that fall within a more extensive body of literature and research on multilingualisms, the contexts (histories, linguistic ecologies and locus) of which are more familiar to scholars and practitioners in Africa and the Asia-Pacific.
Elsewhere we have argued that Africa and the Asia-Pacific together comprise 77% of the world’s population and are home to 86% of the world’s languages. Europe and North America together are home to only 3% of the world’s languages. Yet it is literature on multilingualism from the latter that enjoys wide international circulation, while expertise and scholarship in multilingualisms from the former have been largely occluded (e.g., Heugh 2021 (2021) Southern multilingualisms, translanguaging and transknowledging in inclusive and sustainable education. In P. Harding-Esch & H. Coleman (Eds.), Language and the sustainable development goals (pp. 37–47). British Council. (2021) Southern multilingualisms, translanguaging and transknowledging in inclusive and sustainable education. In P. Harding-Esch & H. Coleman (Eds.), Language and the sustainable development goals (pp. 37–47). British Council., 2025a (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. ).
We therefore include a brief overview of African research and literature on the plurality of multilingualisms that draw attention to the multiple ways in which citizens of southern environs deploy these for convivial and horizontal purposes (for getting along) and for purposes of demonstrating resistance. Multilingualisms are also used for vertical purposes of precision in formal translation and interpreting services, including for access and participation, for example in high-level academic enterprises, legal discourse and documentation, medical and scientific research, international protocols. Vertical use of multilingualisms (keeping distinctions between languages) is also often used for purposes of conflict, exclusion and security. Multilingualisms never appear in an epistemic, ontological or cosmological vacuum. Rather, they occur alongside plural systems of knowledge, ways of being (ontologies) and ways of believing (cosmologies). In most circumstances the purpose or function of multilingual exchanges has to do with the translation and exchange of information (knowledge) rather than language per se. We refer to this process as transknowledging. A ‘southern multilingualisms’ perspective that emerged through decades of research and engagement with multilingual communities in remote, rural and urban settings rests on three key characteristics (Heugh, 2025a (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. ) summarised below:
-
Multidimensional aspects of language use for specific functions or purposes. These include convivial and horizontal language practices (‘mixing of languages’ Nhlapo, 1945Nhlapo, J. M. (1945) Nguni and Sotho: A practical plan for the unification of the South African Bantu languages. African Bookman.Nhlapo, J. M. (1945) Nguni and Sotho: A practical plan for the unification of the South African Bantu languages. African Bookman.), both ‘horizontal and vertical use of African languages’ (Heine, 1977Heine, B. (1977) Vertical and horizontal communication in Africa. Africa Spectrum, 12 (3), 231–238. https://www.africabib.org/rec.php?RID=188823948Heine, B. (1977) Vertical and horizontal communication in Africa. Africa Spectrum, 12 (3), 231–238. https://www.africabib.org/rec.php?RID=188823948) for specific functions, and the notion that ‘multilingualism is the lingua franca in Africa’ (Fardon & Furniss, 1994Fardon, R., & Furniss, G. (Eds.) (1994) African languages, development and the state. Routledge.Fardon, R., & Furniss, G. (Eds.) (1994) African languages, development and the state. Routledge.).33.Multilingualism for these authors is a dynamic, integrative communicative system, not a collection of isolated linguistic entities, and predates northern-facing discussions of translanguaging from 2009 onwards.
-
Linguistic citizenship, where the way people deploy their linguistic repertoires is related to their agency and choice (Stroud, 2001Stroud, C. (2001) African mother-tongue programmes and the politics of language: Linguistic citizenship versus linguistic human rights. Journal of Multilingual and Multicultural development, 22 (4), 339–355. Stroud, C. (2001) African mother-tongue programmes and the politics of language: Linguistic citizenship versus linguistic human rights. Journal of Multilingual and Multicultural development, 22(4), 339–355. ) and dependent upon circumstance.
-
Transknowledging which involves reciprocal (two-or-multiple way) translation and exchange of knowledge (information, epistemology) (Heugh, 2021 (2021) Southern multilingualisms, translanguaging and transknowledging in inclusive and sustainable education. In P. Harding-Esch & H. Coleman (Eds.), Language and the sustainable development goals (pp. 37–47). British Council. (2021) Southern multilingualisms, translanguaging and transknowledging in inclusive and sustainable education. In P. Harding-Esch & H. Coleman (Eds.), Language and the sustainable development goals (pp. 37–47). British Council.).
We bring this background of research and theorisation to Chang’s micro-study discussed later.
Shifting the gaze southwards
The metacognitive and metalinguistic challenges experienced by students who find themselves studying in HE institutions through a second or additional language have generated considerable research, educational and technological response across Africa, the Asia-Pacific and South America during the 20th century. Comparatively, this attention was delayed until the first quarter of the 21st Century in the Euro-North, marked by the popularisation of ‘the multilingual turn’ (May, 2013May, S. (Ed.) (2013) The multilingual turn: Implications for SLA, TESOL, and bilingual education. Routledge. May, S. (Ed.) (2013) The multilingual turn: Implications for SLA, TESOL, and bilingual education. Routledge. ). Pedagogically it was marked by a critique of Canadian and US-based bilingual education as amounting to parallel monolingualism (Heller, 2007Heller, M. (2007) Bilingualism as ideology and practice. In Bilingualism: A social approach (pp. 1–22). Palgrave Macmillan UK. Heller, M. (2007) Bilingualism as ideology and practice. In Bilingualism: A social approach (pp. 1–22). Palgrave Macmillan UK. ). This was followed in the appropriation of the term ‘translanguaging’ used in Welsh bilingual schools (Williams, 1996Williams, C. (1996) Secondary education: Teaching in the bilingual situation. In C. Williams, G. Lewis, & C. Baker (Eds.), The language policy: Taking stock (pp. 39–78). UK:CAI Language Studies Centre.Williams, C. (1996) Secondary education: Teaching in the bilingual situation. In C. Williams, G. Lewis, & C. Baker (Eds.), The language policy: Taking stock (pp. 39–78). UK:CAI Language Studies Centre.) and reconceptualised as distinct from codeswitching, untethered to specific models of bilingual education, and free of ‘named’ languages (García & Li, 2014García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. ).
Understanding theoretically different underpinnings of northern and southern perspectives of multilingualism is important. Northern-located scholars live and work in predominantly monolingual (to a lesser extent bilingual) urban societies. From a predominantly English, and partially French bilingual society in Canada, Heller (2007)Heller, M. (2007) Bilingualism as ideology and practice. In Bilingualism: A social approach (pp. 1–22). Palgrave Macmillan UK. Heller, M. (2007) Bilingualism as ideology and practice. In Bilingualism: A social approach (pp. 1–22). Palgrave Macmillan UK. has argued that French Canadian Immersion programs (and by implication, Dual language immersion in the US) have been designed to retain hermetically sealed separation of languages in ‘parallel monolingual’ isolation rather than programs that employ whole linguistic repertoires. García and Li (2014)García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. and many scholars who follow these authors have reinterpreted translanguaging in Williams (1996)Williams, C. (1996) Secondary education: Teaching in the bilingual situation. In C. Williams, G. Lewis, & C. Baker (Eds.), The language policy: Taking stock (pp. 39–78). UK:CAI Language Studies Centre.Williams, C. (1996) Secondary education: Teaching in the bilingual situation. In C. Williams, G. Lewis, & C. Baker (Eds.), The language policy: Taking stock (pp. 39–78). UK:CAI Language Studies Centre. as linguistic practices that transgress or transcend ‘named’ and separated languages. They further argue that this version of translanguaging, effectively doing away with explicitly bilingual education programs, serves as the linguistic pathway towards social justice in education.
Decades of research from within predominantly multilingual urban and rural societies upon which monolingual education systems have been imposed has been undertaken by many African and South Asian scholars (e.g., Bamgbose, 2000Bamgboṣe, A. (2000) Language and exclusion: The consequences of language policies in Africa (Vol. 12). LIT Verlag Münster.Bamgboṣe, A. (2000) Language and exclusion: The consequences of language policies in Africa (Vol. 12). LIT Verlag Münster.; Obanya, 1999Obanya, P. (1999) Popular fallacies on the use of African languages in education. Social dynamics, 25 (1), 81–100. Obanya, P. (1999) Popular fallacies on the use of African languages in education. Social dynamics, 25(1), 81–100. in Africa; Pattanayak, 1990Pattanayak, D. P. (Ed.) (1990) Multilingualism in India (No. 61). Multilingual Matters.Pattanayak, D. P. (Ed.) (1990) Multilingualism in India (No. 61). Multilingual Matters.; Agnihotri, 1995Agnihotri, R. K. (1995) Multilingualism as a classroom resource. In Heugh, K., Siegrühn, A., & Plüddemann, P. (Eds.). Multilingual Education for South Africa, pp. 3–7. Heinemann Educational Books.Agnihotri, R. K. (1995) Multilingualism as a classroom resource. In Heugh, K., Siegrühn, A., & Plüddemann, P. (Eds.). Multilingual Education for South Africa, pp. 3–7. Heinemann Educational Books., 2014 (2014) Multilinguality, education and harmony. International Journal of Multilingualism, 11 (3), 364–379. (2014) Multilinguality, education and harmony. International Journal of Multilingualism, 11(3), 364–379. ; Mohanty, 2018Mohanty, A. K. (2018) The multilingual reality: Living with languages (Vol. 16). Multilingual Matters. Mohanty, A. K. (2018) The multilingual reality: Living with languages (Vol. 16). Multilingual Matters. , in India). Their theoretical perspectives come from living in and observing complex multilingualisms in situ. In southern settings, despite monolingual regimes in which a former colonial language is imposed as the medium of instruction, bilingualism and multilingualism are the de facto implicit reality. They are the underbelly of entire education systems and the linguistic repertoire of most students and teachers.
To summarise, students are offered rarefied models of education in both southern and northern contexts. These include bilingual programs for a small minority of students in predominantly monolingual Euro-North systems. They are monolingual education systems offered to predominantly multilingual societies in southern locales. Just as theorisation and practice of bilingual education is problematic in the north; so too is theorisation of monolingual education in the multilingual south. Although the contexts are entirely dissimilar, the north has much to learn from the south.
Nhlapo (1945)Nhlapo, J. M. (1945) Nguni and Sotho: A practical plan for the unification of the South African Bantu languages. African Bookman.Nhlapo, J. M. (1945) Nguni and Sotho: A practical plan for the unification of the South African Bantu languages. African Bookman., for example, recognised the everyday horizontal (convivial, mixing) usually spoken multilingualism of urban and mobile citizens as resources for social cohesion seven decades ahead of ‘the multilingual turn’ and translanguaging. He also recognised the vertical (formal separation) of language use in high-stakes circumstances (including academia, justice, safety, security) in postcolonial and multilingual Africa. Heine (1977)Heine, B. (1977) Vertical and horizontal communication in Africa. Africa Spectrum, 12 (3), 231–238. https://www.africabib.org/rec.php?RID=188823948Heine, B. (1977) Vertical and horizontal communication in Africa. Africa Spectrum, 12 (3), 231–238. https://www.africabib.org/rec.php?RID=188823948 argued that African languages are used for different functions or purposes by multilingual citizens. Fardon and Furniss (1994)Fardon, R., & Furniss, G. (Eds.) (1994) African languages, development and the state. Routledge.Fardon, R., & Furniss, G. (Eds.) (1994) African languages, development and the state. Routledge. refer to multilingualism as the lingua franca in Africa — very different from notions of parallel monolingualism. African linguists recognise how de facto multilingualisms are in systemic use despite the mirage of monolingual education regimes.
From a southern perspective, African and South Asian scholars mentioned above do not accept that it is possible for a hermetically sealing off / or separation of languages in any bilingual program. The brain of each human processes considerable metacognitive and metalinguistic activity across linguistic divides, although this may not be audible or visible in classroom observations.
Northern notions of parallel monolingualism and fuzzy borderless repertoires of language in discourses of translanguaging that ignore southern realities cause confusion for many in southern settings. There is pressure to follow scholarship rooted in northern literature even when the contexts are fundamentally different. Makoni (2003)Makoni, S. (Ed.) (2003) Black linguistics: Language, society, and politics in Africa and the Americas. Psychology Press.Makoni, S. (Ed.) (2003) Black linguistics: Language, society, and politics in Africa and the Americas. Psychology Press. for example, incorrectly assumes Heller’s critique of parallel bilingualism in Canadian education is relevant in discussions of bilingualism and multilingualism for majority populations in South Africa. Makalela (2016)Makalela, L. (2016) Ubuntu translanguaging: An alternative framework for complex multilingual encounters. Southern African Linguistics and Applied Language Studies, 34 (3), 187–196. Makalela, L. (2016) Ubuntu translanguaging: An alternative framework for complex multilingual encounters. Southern African Linguistics and Applied Language Studies, 34(3), 187–196. embraces Canadian, UK and US discussions of translanguaging, while simultaneously introducing the horizontal, convivial features of African ‘Ubuntu’ communication with antecedents in earlier work of Nhlapo and others.
Collaborative action research among educators and linguists opposed to apartheid in Southern Africa from the 1980s onwards with scholars from across Africa and India includes action research with multilingual teachers and teacher educators in the National Language Project, an NGO, and English medium university, the University of Cape Town, in South Africa (1983–2005) (e.g., Heugh, Siegrühn & Plüddemann, 1995Heugh, K., Siegrühn, A., & Plüddemann, P. 1995 (Eds.). Multilingual education for South Africa. Heinemann Educational Books.Heugh, K., Siegrühn, A., & Plüddemann, P. 1995 (Eds.). Multilingual education for South Africa. Heinemann Educational Books.; Benson & Plüddemann, 2010Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. ). The purpose was to propose post-apartheid language policy based on multilingualism conceptualised through African and Indian research, including horizontal and vertical features of linguistic repertoires and ecologies. This research was followed by a 25-country study for the Association for the Development of Education in Africa (ADEA) and the UNESCO Institute for Lifelong Learning (Alidou et al., 2006Alidou, H., Boly, A., Brock-Utne, B., Diallo, Y. S., Heugh, K., & Wolff, H. E. (2006) Optimizing learning and education in Africa–the language factor. Paris: ADEA.Alidou, H., Boly, A., Brock-Utne, B., Diallo, Y. S., Heugh, K., & Wolff, H. E. (2006) Optimizing learning and education in Africa–the language factor. Paris: ADEA.). The multi-country study revealed that most students in Africa were unable to complete primary or secondary school education. A critical reason is that students are required to read, write and be assessed through a former colonial language rather than a language in local or wider societal use. Despite de facto (but considered illegitimate) use of informal multilingual spoken practices among students and teachers for purposes of inclusion and meaning making, the advantages afforded by spoken (and/or signed) multilingualism do not carry through monolingual reading, writing and high stakes assessment regimes conducted in silence (see also Reeves et al., 2008Reeves, C., Heugh, K., Prinsloo, C P., Macdonald, C., Netshitangani, T., Alidou, H., & Diedericks, G. (2008) Evaluation of literacy teaching in the primary schools of the Limpopo Province. Pretoria: Human Sciences Research Council.Reeves, C., Heugh, K., Prinsloo, C P., Macdonald, C., Netshitangani, T., Alidou, H., & Diedericks, G. (2008) Evaluation of literacy teaching in the primary schools of the Limpopo Province. Pretoria: Human Sciences Research Council.). It is the case that convivial, horizontal use of spoken language mixing noted in the fieldwork of Reeves et al., resonates with the everyday informal crossing, mixing and switching across or through porous repertoires in much current literature on (horizontal) translanguaging and in Makalela (2016)Makalela, L. (2016) Ubuntu translanguaging: An alternative framework for complex multilingual encounters. Southern African Linguistics and Applied Language Studies, 34 (3), 187–196. Makalela, L. (2016) Ubuntu translanguaging: An alternative framework for complex multilingual encounters. Southern African Linguistics and Applied Language Studies, 34(3), 187–196. . Although these practices are conceived of as a translanguaging pedagogy that attends to social justice as discussed in García and Li (2014)García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. and others, they were found to be insufficient in the African context to secure either retention through secondary school or social justice beyond the classroom door. Instead, they do not prepare the way for academic literacy and EMI which continue to serve as roadblocks in assessment regimes and access to further and higher education.
Indicated earlier, this research has run parallel to three or more decades of innovations in and research on the role of HLT in South Africa, referred to here as MT (e.g., Ndebele, 2022Ndebele, H. (2022) Exploring the challenges of information and communication technology localization in South African higher education: a language management approach. International Journal of Multilingualism, 19 (3), 368–382. Ndebele, H. (2022) Exploring the challenges of information and communication technology localization in South African higher education: a language management approach. International Journal of Multilingualism, 19(3), 368–382. ; van Huyssteen et al., 2023van Huyssteen, G. B., Puttkammer, M., McKellar, C., & Griesel, M. (2023) Translation technology in South Africa. In Routledge encyclopedia of translation technology (pp. 373–383). Routledge. van Huyssteen, G. B., Puttkammer, M., McKellar, C., & Griesel, M. (2023) Translation technology in South Africa. In Routledge encyclopedia of translation technology (pp. 373–383). Routledge. ). Explicit provision of access to and use of MT has been trialled and refined in that setting to diminish unequal linguistic and epistemic access to and use of academic materials in EMI institutions. These are ones in which most students, multilingual speakers of African languages, are disadvantaged by academic teaching and literature in English compared to minority L1 speakers of English.
South-South conversations
Southern-facing research outlined above continued in a longitudinal series of action research studies of students’ engagement with language and epistemic learning in an EMI institution at the University of South Australia between 2010 and 2022. This research was undertaken in eight undergraduate courses in English as an Additional Language and Applied Linguistics, 54 classes; through three surveys involving undergraduate students; collaboration and research participation of nine graduate research students.44.PhD, Masters and Honours students who contributed to the series of studies to 2022, and ongoing follow-up to 2025, are included as student participants in the series. In total, 1072 students participated in the study, the majority of whom participated in two or more semester courses.
| Semester courses | Student linguistic profile | Nr courses | Nr classes | Nr students | Nr academic staff |
|---|---|---|---|---|---|
| 2010–2016 | |||||
| 1st year EAL | Multilingual | 2 | 14 | 350 | 4 |
| 2nd year EAL | 2 | 14 | 210 | ||
| 3rd year EAL | 1 | 7 | 105 | 3 | |
| 3rd year English literature | 1 | 7 | 210 | ||
| Subtotal | 6 | 42 | 875 | 7 | |
| 2017–2022 | |||||
| 3rd year Applied Linguistics | English L1 & multilingual | 2 | 12 | 240 | 3 |
| Survey & interview participants | English L1 & multilingual | Includes students across several academic schools | 148 | 5 | |
| Subtotal | 2 | 12 | 388 | 8 | |
| 2013–2025 | |||||
| BA Honours (research) | Multilingual | n/a | n/a | 2 | 1 |
| Master’s Research | Monolingual | 1 | 4 | ||
| PhD | Multilingual | 6 | 6 | ||
| Supervision & Research analysis | Bi-/multilingual | 12 | |||
| Total | 8 | 54 | 1272 | 38 |
Multilingual international students, typically, have studied English as a foreign or additional language for 10–12 years prior to navigating English language proficiency tests, such as IELTS, and enrolling in Australian universities. Yet most find themselves overwhelmed by the corpus of academic reading and academic writing in English. Observations of student anxiety, high levels of attrition, and limited institutional recognition of multilingual strategies precipitated classroom research which morphed into a longitudinal series of studies. These included a focus on reading and writing academic texts, beginning with the metacognitive and metalinguistic complexities of paraphrasing and summarising, and two-way translation between students’ primary language(s) and English. Explicit pedagogical endorsement of horizontal mixing and switching between languages in both spoken and written form was accompanied by a range of translation practices (reported in Heugh, Li & Song, 2017Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. ).
Students responded to invitations to bring information available in their home / preferred languages into tutorials and oral presentations, demonstrating their linguistic agency and ownership of multilingual repertoires in classroom discussions. Increasingly, written assessment tasks included opportunities for students to use materials in two or more languages, including use of translation of key information. This brought about invigorated tutorial discussions in which different perspectives and kinds of knowledge available through their languages were compared. By 2014 changes to criteria for both oral presentations and written assessment in courses designed to bridge English Additional Language and Applied Linguistics included up to 30% of reading material to support assignments in the language/s of students’ choice, the remainder in English. Students were encouraged to (i) incorporate key quotations in the source language, (ii) translate and/or paraphrase these, (iii) disclose how the translation was provided (human translation, MT, or a combination of both), (iv) cite and reference in the source language, and translate in English. By 2017, the practice was included for monolingual English-speaking students of Applied Linguistics, who were encouraged to develop capabilities in the use of MT, and to attempt PE through collaboration with their multilingual peers. Multilingual students were invited to offer oral presentations and demonstrations to monolingual peers on how to navigate various MT platforms and resources, including how to make best judgment calls on accuracy and appropriacy. This process served to change power dynamics between L1 speakers of English and multilingual students, equalising respect for knowledge and language brokers across several languages and English in tutorials, and collaboration on assignments. It also served to open the door for monolingual students to begin to develop multilingual capabilities. It introduced monolingual students to the potential of MT, making visible multilingual students’ use of translation technologies (Heugh et al., 2022Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35 (1), 89–127. Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35(1), 89–127. ; Heugh 2025a (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. , b (2025b) Heugh, K. (2025) Southern multilingualisms and transknowledging: towards restoring plurality in applied linguistics and education. Applied Linguistics Review. (2025b) Heugh, K. (2025) Southern multilingualisms and transknowledging: towards restoring plurality in applied linguistics and education. Applied Linguistics Review. ). By 2019 the teaching staff developed a set of pedagogical resources for use by students and teachers on the use of MT and human interaction with MT, circulated within the university.
Setting the scene
This discussion above has laid the groundwork for explicit or deliberative use of translation to support language learning and meaningful access to and advance educational equality for students. We regard explicit pedagogical use of translation in academic reading and writing as an example of vertical translanguaging (Heugh, 2015Heugh, K. (2015) Epistemologies in multilingual education: Translanguaging and genre — companions in conversation with policy and practice. Special Issue, Language and Education. 29 (3): 280–285. Heugh, K. (2015) Epistemologies in multilingual education: Translanguaging and genre — companions in conversation with policy and practice. Special Issue, Language and Education. 29 (3): 280–285. ) or pedagogical translanguaging (Cenoz & Gorter, 2021Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. ). Through the interlinked studies using multilingual pedagogies with students who exercise their linguistic agency and citizenship, we have come to understand that translation extends beyond language teaching and learning. It serves to facilitate reciprocal translation and exchange of disciplinary knowledge beyond language, what we refer to as ‘transknowledging’. Thus, together we regard translation, translanguaging and transknowledging as core elements of multilingualisms in education, certainly for many students we encounter from southern settings (see also Heugh, 2025a (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. (2025a) EMI, Cognitive capture and decoloniality. In Mirhosseini, S-A., & De Costa, P. I. (Eds). Critical English medium instruction in higher education, 215–231. Cambridge University Press. , b (2025b) Heugh, K. (2025) Southern multilingualisms and transknowledging: towards restoring plurality in applied linguistics and education. Applied Linguistics Review. (2025b) Heugh, K. (2025) Southern multilingualisms and transknowledging: towards restoring plurality in applied linguistics and education. Applied Linguistics Review. ).
Figure 1 below illustrates the schematic relationship among the three core elements of southern multilingualisms in education:
-
Horizontal translanguaging can be adapted and expanded from García, 2009García, O. (2009) Bilingual education in the 21st century: A global perspective. Wiley-Blackwell.García, O. (2009) Bilingual education in the 21st century: A global perspective. Wiley-Blackwell.; García & Li, 2014García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. García, O., & Li, W. (2014) Translanguaging: Language, bilingualism and education. Palgrave Macmillan. ; Li & García, 2022Li, W., & García, O. (2022) Not a first language but one repertoire: Translanguaging as a decolonizing project. RELC journal, 53 (2), 313–324. Li, W., & García, O. (2022) Not a first language but one repertoire: Translanguaging as a decolonizing project. RELC journal, 53(2), 313–324. (including spoken translation, i.e. interpreting).
-
Vertical translanguaging (including written translation).
-
Transknowledging (reciprocal or two-way exchange and translation of knowledge).
Translation technology in source-based academic writing
We complement the research discussed thus far with further evidence of difficulties that multilingual students encounter with texts in English (e.g., Singh, 2019Singh, M. K. M. (2019) International graduate students’ academic writing practices in Malaysia: Challenges and solutions. Journal of International Students, 5 (1), 12–22. http://jistudents.org/Singh, M. K. M. (2019) International graduate students’ academic writing practices in Malaysia: Challenges and solutions. Journal of International Students, 5(1), 12–22. http://jistudents.org/; Tajik et al., 2022Tajik, M. A., Abdul Manan, S., Arvatu, A.-C., & Shegebayev, M. (2022) Growing pains: Graduate students grappling with English medium instruction in Kazakhstan. Asian Englishes, 1–19. Tajik, M. A., Abdul Manan, S., Arvatu, A.-C., & Shegebayev, M. (2022) Growing pains: Graduate students grappling with English medium instruction in Kazakhstan. Asian Englishes, 1–19. ). We refer to academic sources used in writing as source-based academic writing, with particular attention to paraphrasing, summarising, and synthesising information from multiple texts (e.g., Safari & Ahmadi, 2025Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53 (1), 53–70. Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53(1), 53–70. ). Source-based writing requires understanding of texts and selection of relevant information from reading materials to paraphrase and then integrate reworded information into coherent writing. While this complex metacognitive and metalinguistic process is often difficult for L1 writers, it is more challenging for L2 writers (Hirvela & Du, 2013Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12 (2), 87–98. Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12(2), 87–98. ; Safari & Ahmadi, 2025Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53 (1), 53–70. Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53(1), 53–70. ; Wette, 2017Wette, R. (2017) L2 undergraduate students’ learning to write using sources: A trajectory of skill development. In J. Bitchener, N. Storch, & R. Wette (Eds.), Teaching writing for academic purposes to multilingual students: Instruction approaches, (pp. 99–112). Routledge. Wette, R. (2017) L2 undergraduate students’ learning to write using sources: A trajectory of skill development. In J. Bitchener, N. Storch, & R. Wette (Eds.), Teaching writing for academic purposes to multilingual students: Instruction approaches, (pp. 99–112). Routledge. ) who undertake complex academic language input and output processes in their L2.
Aizawa et al.’s (2023)Aizawa, I., Rose, H., Thompson, G., & Curle, S. (2023) Beyond the threshold: Exploring English language proficiency, linguistic challenges, and academic language skills of Japanese students in an English medium instruction programme. Language Teaching Research, 27 (4), 837–861. Aizawa, I., Rose, H., Thompson, G., & Curle, S. (2023) Beyond the threshold: Exploring English language proficiency, linguistic challenges, and academic language skills of Japanese students in an English medium instruction programme. Language Teaching Research, 27(4), 837–861. study with 264 multilingual students in an EMI program at a Japanese university showed that even advanced learners (CEFR B2-C1) found reading English academic articles highly challenging. Similarly, Safari and Ahmadi (2025)Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53 (1), 53–70. Safari, F., & Ahmadi, A. (2025) L2 learners’ challenges in integrated writing tasks: Implications for writing teachers and developers of diagnostic assessments. The Language Learning Journal, 53(1), 53–70. reported that language challenges (limited English vocabulary and grammar) prevented multilingual students from clearly expressing ideas. Students often have no alternative other than to retain words and sentence structures from source texts when paraphrasing. Barkaoui (2015)Barkaoui, K. (2015) Test takers’ writing activities during the TOEFL iBT® writing tasks: A stimulated recall study. ETS Research Report Series, 2015 (1), 1–42. Barkaoui, K. (2015) Test takers’ writing activities during the TOEFL iBT® writing tasks: A stimulated recall study. ETS Research Report Series, 2015(1), 1–42. also highlighted challenges in choosing words to convey appropriate meaning. Neumann et al.’s (2019)Neumann, H., Leu, S., & McDonough, K. (2019) L2 writers’ use of outside sources and the related challenges. Journal of English for Academic Purposes, 38 , 106–120. Neumann, H., Leu, S., & McDonough, K. (2019) L2 writers’ use of outside sources and the related challenges. Journal of English for Academic Purposes, 38, 106–120. research in Canada similarly found that multilingual students in that EMI context struggle with academic reading and writing from source-based texts. Students in their study found difficulty in identifying key ideas and selecting relevant information to support their arguments. Overall, paraphrasing and summarising in source-based writing remains especially challenging for multilingual learners (Hirvela & Du, 2013Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12 (2), 87–98. Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12(2), 87–98. ; Shi & Dong, 2018Shi, L., & Dong, Y. (2018) Chinese graduate students paraphrasing in English and Chinese contexts. Journal of English for Academic Purposes, 34 , 46–56. Shi, L., & Dong, Y. (2018) Chinese graduate students paraphrasing in English and Chinese contexts. Journal of English for Academic Purposes, 34, 46–56. ), even when students meet English proficiency requirements for university admission (Singh, 2019Singh, M. K. M. (2019) International graduate students’ academic writing practices in Malaysia: Challenges and solutions. Journal of International Students, 5 (1), 12–22. http://jistudents.org/Singh, M. K. M. (2019) International graduate students’ academic writing practices in Malaysia: Challenges and solutions. Journal of International Students, 5(1), 12–22. http://jistudents.org/; Tajik et al., 2022Tajik, M. A., Abdul Manan, S., Arvatu, A.-C., & Shegebayev, M. (2022) Growing pains: Graduate students grappling with English medium instruction in Kazakhstan. Asian Englishes, 1–19. Tajik, M. A., Abdul Manan, S., Arvatu, A.-C., & Shegebayev, M. (2022) Growing pains: Graduate students grappling with English medium instruction in Kazakhstan. Asian Englishes, 1–19. ). Trenkic and Warmington (2019)Trenkic, D., & Warmington, M. (2019) Language and literacy skills of home and international university students: How different are they, and does it matter? Bilingualism: Language and Cognition, 22 (2), 349–365. Trenkic, D., & Warmington, M. (2019) Language and literacy skills of home and international university students: How different are they, and does it matter? Bilingualism: Language and Cognition, 22(2), 349–365. found that Chinese international students with CEFR B2/C1 English proficiency levels are disadvantaged compared with L1 English students. They have smaller vocabularies in English, restricted comprehension and summarising capabilities, and require twice as long to read academic texts.
From previous research we know that multilingual learners use MT to support L2 writing. MT is used for translating lexical items, sentences, and paragraphs; assisting comprehension of L2 texts and generating drafts from L1 into L2 (Bowker & Ciro, 2019Bowker, L., & Ciro, J. B. (2019) Machine translation and global research: Towards improved machine translation literacy in the scholarly community. Bingley: Emerald Publishing. Bowker, L., & Ciro, J. B. (2019) Machine translation and global research: Towards improved machine translation literacy in the scholarly community. Bingley: Emerald Publishing. ; Zhou et al., 2022Zhou, S., Zhao, S., & Groves, M. (2022) Towards a digital bilingualism? Students’ use of machine translation in international higher education. Journal of English for Academic Purposes, 60 , 101193. Zhou, S., Zhao, S., & Groves, M. (2022) Towards a digital bilingualism? Students’ use of machine translation in international higher education. Journal of English for Academic Purposes, 60, 101193. ). Learners also use MT to translate their L2 writing into L1 to check meaning and then back translate to L2 for comparison and revision. MT is thus a tool for revision (Chang, 2024 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629; Tsai, 2022Tsai, S.-C. (2022) Interactive academic EFL writing assisted by GT for Chinese non-English major students. In J. Qin & P. Stapleton (Eds.), Technology in Second Language Writing (pp. 10–27). Routledge. Tsai, S.-C. (2022) Interactive academic EFL writing assisted by GT for Chinese non-English major students. In J. Qin & P. Stapleton (Eds.), Technology in Second Language Writing (pp. 10–27). Routledge. ). Patterns of students’ MT use include searching for equivalent words, correcting errors, comparing translations, and making lexical choices. MT is therefore a multifunctional resource (Yu & Jiang, 2025Yu, R., & Jiang, L. (2025) Exploring patterns of AI-powered machine translation use in second language writing: a translanguaging perspective. Computer Assisted Language Learning, 1–27. Yu, R., & Jiang, L. (2025) Exploring patterns of AI-powered machine translation use in second language writing: a translanguaging perspective. Computer Assisted Language Learning, 1–27. ).
However, despite its usefulness, MT remains imprecise. Post-editing (PE) involves modifying MT-generated texts through a process of correcting and editing machine-produced text to achieve an acceptable translation (Allen, 2003Allen, J. (2003) Post-editing. In H. Somers (Ed.), Computers and Translation (Vol. 35, pp. 297–318). John Benjamins. http://digital.casalini.it/9789027296696. Allen, J. (2003) Post-editing. In H. Somers (Ed.), Computers and Translation (Vol. 35, pp. 297–318). John Benjamins. http://digital.casalini.it/9789027296696. ). In a pedagogical intervention Chang (2022)Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. incorporated PE into an English-Chinese translation class for advanced learners of Chinese. She found that adopting more than one MT tool may assist with the judgment of MT accuracy. That study also shows that multilingual learners draw on their repositories of linguistic expertise (repertoires) and technological capabilities to help evaluate the appropriateness of translation produced by MT before making PE decisions. Consistent with findings in Heugh et al. (2017)Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. , Chang notes that students with advanced bilingual capability (strong capability in two or more languages), are best placed to judge MT accuracy and make best use of PE.
Our earlier discussions indicate that multilinguals often use translation to move back and forth between different languages when writing in L2 English (Heugh et al, 2017Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. Heugh, K., Li, X., & Song, Y. (2017) Multilingualism and translanguaging in the teaching of and through English: rethinking linguistic boundaries in an Australian University. In B. Fenton-Smith, P. Humphreys, & I. Walkinshaw (Eds.), English medium instruction in higher education in Asia-Pacific (pp. 259–279). Springer. ), and that MT, has been used as a tool in their L2 writing processes (Chang, 2022Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. ; Groves & Mundt, 2021Groves, M., & Mundt, K. (2021) A ghostwriter in the machine? Attitudes of academic staff towards machine translation use in internationalised Higher Education. Journal of English for Academic Purposes, 50 , 100957. Groves, M., & Mundt, K. (2021) A ghostwriter in the machine? Attitudes of academic staff towards machine translation use in internationalised Higher Education. Journal of English for Academic Purposes, 50, 100957. ). However, until recently it has remained unclear how multilingual students use MT to support their HT when writing with source texts. Although PE has been considered essential for accuracy and appropriateness (Bowker & Ciro, 2019Bowker, L., & Ciro, J. B. (2019) Machine translation and global research: Towards improved machine translation literacy in the scholarly community. Bingley: Emerald Publishing. Bowker, L., & Ciro, J. B. (2019) Machine translation and global research: Towards improved machine translation literacy in the scholarly community. Bingley: Emerald Publishing. ; Chang, 2022Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. ), little attention has been paid to how or whether an explicit pedagogy involving both MT and PE can facilitate source-based academic writing while students maintain academic integrity. We address this below, offering insights on the potential cognitive, linguistic and pedagogical advantages of MT and PE in EMI multilingual education as well as the implications for vertical translanguaging and transknowledging.
Chang’s research places the spotlight on how multilingual students use MT and PE to support paraphrasing and summarising in academic writing, paying attention to a sequence of procedures that students adopt. She has found more nuanced data specifically about the use of MT and PE in L2 students’ translation of academic text in English and their L1. She has also found evidence of how and why they engage in cognitive and linguistic processes relating to PE. Her research sheds light on how practices of translation, translanguaging and transknowledging co-occur as students use both MT and PE to make meaning from and into texts in L2
Research methods
The qualitative action research highlighted here involved online collaborative MT and PE teaching interventions. Six Chinese-speaking postgraduate students in different disciplinary fields, interested in using MT and PE to further both their expertise in English and disciplinary knowledge participated. Their IELTs scores of 6.5 (or equivalent, e.g., CEFR B2) had secured enrolment in graduate programs. Table 2 shows baseline information of students’ linguistic repertoires and study programs.
| Name (Pseudonym) | Qing | Yun | Lian | Ying | Peng | Xiang |
|---|---|---|---|---|---|---|
| L1 | Mandarin | Mandarin | Mandarin | Mandarin | Mandarin | Mandarin |
| L2 or addition language(s) | English | English, Taiwanese | English | English, | English | English |
| Study program | Master of Teaching (Early Childhood) | Master of Social Work | Master of Research (Business) | Master of Teaching (Early Childhood) | PhD in Architecture | Master of Education (TESOL) |
The intervention, therefore, was designed to guide students towards achieving their aims to advance their disciplinary knowledge and English through translation. It was also designed to guide students towards meeting academic integrity expectations. It was also to consider how translation modes, MT and PE, might support pedagogical use of translation, translanguaging and transknowledging as multilingual pedagogies. Human translation (HT) involves students producing translations on their own. MT involves students using machine-translated texts. PE involves students editing or modifying MT outputs. The pedagogical intervention included the use of all three translation modes in five one-hour collaborative sessions. Following findings from Chang (2022)Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. , students were advised to use more than one MT tool to cross-check the accuracy of MT-generated texts, such as Google Translate and Youdao Translator. Since the intervention included collaborative teaching and learning methodology, in addition to researcher-led guidance, student participants were invited to demonstrate and share how they use(d) MT.
Table 3 shows the teacher-student collaborative tutoring activities and the data collection process.
| Tutorial sessions 1hr/ session | Tutoring activities | Data collection |
|---|---|---|
| Session 1 |
|
N/A |
| Session 2 |
|
N/A |
| Session 3 |
|
N/A |
| Session 4 |
|
|
| Session 5 |
|
|
| Post Intervention |
|
|
While the teacher researcher focused on the use of HT, MT and PE as complementary strategies to support academic reading and writing, three students (Yun, Lian and Xiang) volunteered to demonstrate how they use MT for paraphrasing in tutorial sessions 4 and 5. Their demonstrations and explanations of MT use and subsequent tutorial discussions were screen-recorded via Zoom for transcription. Screen-recorded data of translation materials from the discipline-specific working texts of these three students are illustrated in Tables 4 and 5 below.
After the completion of the intervention, a 90-minute semi-structured focus group interview with five participants and a 50-minute semi-structured individual interview with one participant who could not join the focus group interview were conducted. The interviews were used to further understand students’ thoughts about their use of MT and PE.
The collected screen recording and interview data were transcribed for coding and analysis. The textual data were coloured and highlighted to generate codes while reading through the transcripts, followed by classifying codes into categories. Prior to finalising categories, the initial categories were reviewed and revised by re-reading the whole body of textual data along with each category. Some initial categories were merged and deleted until final categories were decided. Screenshots with commentaries were incorporated to supplement textual descriptions, providing visual evidence of the processes discussed and making the analysis more accessible.
What the data reveal
Students’ rationales, strategies, and sequences in using MT and PE to support paraphrasing and summarising include:
-
use of MT followed by PE (MT and PE) to help with comprehension and paraphrasing;
-
use of MT and PE to improve their paraphrased and summarised texts and to reduce similarity between the source text and their own writing.
Using MT and PE to help with comprehension and paraphrasing
When students found English source texts difficult to understand, let alone paraphrase, they reported using MT and PE to assist comprehension and paraphrasing. For example, Ying explained:
If I come across some complicated sentences, I would use the machine translation tool to translate into Chinese first. Then I would modify some characters if I think the translation is not completely accurate. Next, I would translate the Chinese translation into English [using MT] …then still check the English translation accuracy and do some modifications.
Ying’s account indicated a six-step cognitive and linguistic process. She first used MT to translate complex sentences from English to Chinese. Second, she evaluated MT accuracy. Third, she modified MT output in Chinese. Fourth, she used MT to translate the modified MT Chinese output back into English. Fifth, she evaluated MT accuracy. Sixth, she modified the MT output in English (see Figure 2).
As shown in Figure 2, translation from English to Chinese seemed to help Ying understand the challenging text in English, while back translation from Chinese to English appeared to help her produce an adequately paraphrased text in English.
Lian also used MT and PE to understand difficult texts and to complete paraphrasing. In the fourth tutorial, Lian used the sentence in Table 4 as an example to demonstrate her paraphrasing strategy with MT (Google Translate) when encountering difficult academic texts.
As shown in Table 4, Lian’s process also involved a six-step cognitive and linguistic process in search of high quality and accurate paraphrasing. At first, Lian used MT to translate the English source text into Chinese to get MT output in Chinese (Step 1). She then evaluated the Chinese translation (Step 2). This MT Chinese output sounded unnatural because of the phrasing and word order problem. For instance, the English phrase ‘buying…at unprecedented levels’ was translated into Chinese ‘以前所未有的水平购买… (yi qiansuoweiyou de shuiping goumai…, using unprecedented level buying…)’ as marked in yellow. Therefore, Lian post-edited the Chinese translation and paraphrased the text in Chinese (Step 3). For example, she modified and reworded the underlined text into Chinese ‘中国投资者对澳大利亚葡萄园的投资热情剧增 (zhongguo touzizhe dui aodaliya putaoyuan de touzi reqing juzeng, Chinese investors’ enthusiasm for investing in Australian vineyards has surged)’. Subsequently, she translated the post-edited and paraphrased Chinese text back to English using MT (Step 4). She then evaluated the English translation from MT (Step 5). Finally, she post-edited the MT English output to produce her perception of an appropriate translation (Step 6). Figure 3 shows Lian’s six-step process of using MT and PE for comprehension and paraphrasing.
Analysis of Lian’s paraphrasing process shows that PE and paraphrasing in Chinese at Step 3 seems essential to obtain a more appropriate translation at Step 4 and to avoid high similarity between the source text and paraphrased text. Also, PE at Step 6 appears important to produce a more accurate translation, even though the English translation produced by MT in this example was adequate in terms of source text meaning and the use of vocabulary and syntax. Although Lian indicated in the demonstration that this paraphrasing process is time-consuming as it involves engaging in translation activities twice, from English to Chinese and from Chinese to English, she considers it a viable strategy for paraphrasing difficult academic texts:
…if you have no ideas on how to do that [paraphrasing], this is another option… because for us some English sentences are really hard to understand or even to paraphrase.
This shows Lian’s recognition that paraphrasing involves both reading comprehension and writing production, and that a better understanding of the text precedes paraphrasing. After seeing Lian’s demonstration, Xiang agreed with this strategy, arguing:
…it’s easier for me to understand Chinese terminology than English terminology. So, this applies to a lot of students who are learning professional [knowledge/subjects] through EMI. So maybe it’s easier for them to use this method.
Xiang believes the use of MT in this way to be beneficial for EMI students as it can overcome their difficulties in understanding and paraphrasing difficult academic texts. Prior to participating in this study, she had herself shown other Chinese international students in the fields of Science, Technology, Engineering and Mathematics (STEM) how to paraphrase using MT. In the fifth tutorial, Xiang demonstrated using MT for paraphrasing using the material she had used in her previous presentation prior to the intervention (see Table 5).
As shown in Table 5, Xiang first used three MT tools (Google, Baidu and Youdao) to obtain a translation from English to Chinese which was then directly translated back to English to provide three versions of MT English output (no Chinese translations were shown in this demonstration). Second, Xiang marked terminology that she wanted to keep in yellow and marked other words in red to distinguish the difference in word choice between the source text and MT output. Third, she produced a paraphrased text by selecting words, phrases, and sentence structures from the three MT outputs. According to Xiang, the text highlighted in grey in the rephrased version is a combination of translations from the three MT tools. Figure 4 shows Xiang’s three-step process of using different MT tools for paraphrasing.
This example shows that Xiang obtained different translations from the three MT tools as points of reference, including the translations of words (e.g., major/significant/main), phrases (e.g., concentrated on/focused on), and sentence structures (e.g., …make a major contribution to…/…are the main sources of…/…are the main causes of…). She subsequently rewrote the original text by selecting her perception of suitable MT output and combining the translations together. Her paraphrased text, however, seems to have rather high similarity of sentence structure between the source text and rephrased text, as shown in Table 5.
At first sight, Xiang appears to navigate complex cognitive and linguistic processes. However, she admits that understanding the source texts is not essential when using this three-step process. As Xiang stated in her demonstration, ‘using this kind of method you don’t even need to know what you are reading’. This means Xiang can still produce a paraphrase that has relatively high similarity to the original text through using MT, even if she does not really understand the source text material. Xiang’s use of MT for paraphrasing, therefore, may be interpreted both as a problem in comprehension and one of the risks compromising academic integrity.
Compared to the paraphrasing processes of Ying (Figure 2) and Lian (Figure 3), who both evaluated and modified MT Chinese output at Steps 2 and 3 prior to back translating to English, Xiang omitted these two steps. More specifically, Lian not only post-edited MT Chinese output but also paraphrased in her own words in Chinese to reduce similarity between the source text and her paraphrased text. Analyses of Ying and Lian’s processes suggest that the PE step and paraphrasing in Chinese may push students to better understand source texts. Conversely, without this step, it may be difficult to ensure students’ understanding of the reading materials and compliance with academic integrity.
Despite these three different uses of MT for academic writing, interview data show these students have come to understand how to use this tool more critically and judiciously after the intervention:
I used [MT] a lot before, I was very bold in terms of using machine translation; now I have to be more cautious. Because sometimes when you [teacher researcher] show the translations, there are maybe a few mistakes [in the MT output]. I need to be very careful [Xiang].
I try to use it [MT] just as a tool and not rely on it…if you want to improve English reading and writing, you need to find the balance [Lian].
Following the intervention, Xiang recognises that MT might produce incorrect translations that she might have overlooked previously, while Lian emphasises the value of finding a ‘balance’ between using her own HT and MT. The above comments indicate that the intervention helped these students to understand that MT should be used cautiously, critically, and judiciously. In addition, some students turned a negative attitude towards MT into a positive and recognised the importance of PE when using MT for academic writing after the intervention:
Before, I really disagreed with using translation machine in writing, but now I begin to use it in my writing. I think it’s helpful [Qing]
I wouldn’t know MT has progressed as a good tool for students if it wasn’t this group [Yun].
Post-editing really matters…to avoid plagiarism… they [MT outputs] are just materials… you have to paraphrase… based on the materials we have [Peng].
On the one hand, the intervention reduces some students’ over-dependence on MT; on the other hand, it increases other students’ level of trust in MT as a useful tool. Taken together, learning and developing capabilities in MT and PE may increase students’ awareness of using the tool critically and judiciously.
Using MT and PE to improve paraphrased and summarised texts while reducing similarity
Some students mentioned that they adopt MT and PE to improve their paraphrased texts, as Ying reported:
I would write a paragraph by myself first, then if the similarity [to the source text] is high, I use machine translation tool to reduce the similarity.
This means Ying adopts MT to reduce similarity between her paraphrased text and the source text. This serves to avoid potential breaches of academic integrity. When asked how she managed this process, Ying explained:
[I] put English sentences [paraphrased] in the MT to get Chinese translations and then slightly adjust the Chinese translations and then translate back to English.
This comment shows that Ying starts with HT to produce her paraphrased text, then adopts MT and PE to make modifications to MT Chinese translations, and then translates back to English, through which she reduces similarity between the source text and her paraphrased text. Likewise, Yun described a similar sequence of using HT, MT and PE to support summarising:
I think one of the difficulties when I wrote my assignment is I want to paraphrase all these [papers] I have, but I don’t know how to put them all together. And then I will try to do that in Chinese and then translate back [to English]. But I will definitely make some mistakes in grammar so [MT and PE] will help this, and I have found it really helpful in that scenario.
Although Yun used the word ‘paraphrase’, what she actually does is to summarise several papers for her assignment. The student finds it challenging to summarise or synthesise the main ideas from different articles in English, so she uses HT, MT and PE. She demonstrated how she used MT (Google Translate) to improve her summarised texts in the fourth tutorial (see Table 6).
As shown in Table 6, Yun went through a four-step process. At Step 1, she put her own English text into the MT to obtain MT Chinese output. The Chinese translation from MT allowed Yun to check whether her English text was equivalent to what she wanted to express. This MT Chinese output was not error-free as the word ‘mindset’ was translated as ‘ 最認真的人 (zui renzhen de ren, the most diligent people)’ as indicated in the orange underlined text. At Step 2, therefore, Yun post-edited unsuitable MT output and conveyed her intended meaning. For example, she changed the Chinese translation ‘ 涉及不同的因素 (sheji butongde yinsu, involving different factors)’ to ‘ 牽涉到不同的層面與原因 (qianshe dao butongde cengmian yu yuanyin, involving different aspects and causes)’ to fully express her intended meaning. At Step 3, Yun translated the post-edited Chinese text back to English using MT to obtain MT English output. At Step 4, Yun compared her own written text with the MT English output produced at Step 3 for making revisions. By comparing the two texts, Yun identified grammatical inaccuracies in her own writing as indicated by the yellow underlined text in Table 6. Yun reported that she replaced the inaccuracies in her writing with the underlined MT produced text at Step 3, although she did not show her final revised text in the demonstration. Figure 5 shows Yun’s four-step process of using MT and PE to assist with revisions to improve her summarised texts.
The four-step process seems to help Yun to convey her intended meaning, first through the PE step in Chinese, and then to identify her grammatical inaccuracies in English through the final step of comparing her own writing with the MT text. Yun claimed these steps as beneficial for making revisions and improving grammar in her English written text.
When including HT into the process Yun indicates above, the sequence of her summarising process is to (1) summarise in L1 Chinese; (2) translate the summary from L1 Chinese to L2 English by herself through HT; and (3) start the four-step procedure of using MT and PE to improve her summarised texts (see Figure 6).
Together, the above data show that this student uses MT and PE to reduce similarity to the source text (as Ying explained) and to improve the paraphrased or summarised texts (as Yun demonstrated). This way of using MT requires linguistic and cognitive effort to first evaluate and modify MT output and then make revisions of her own text. Yun’s example shows that the summary of articles was generated by the student because MT (at the time that this research was undertaken)55.This study was conducted prior to the arrival of generative AI such as ChatGPT in late 2022. was unable to produce a summary for her. Analysis of Yun’s strategy and sequence of using MT reveals her ethical use of this tool, in which MT is merely to make their paraphrased and summarised texts more appropriate while maintaining academic integrity.
Discussion
Findings from this modest study help us understand multilingual students using HT, MT, and PE as a combined strategy to support paraphrasing and summarising. Students utilise MT and PE to help with understanding and paraphrasing complex academic source texts. Additionally, they adopt MT and PE to improve their own (HT) paraphrased and summarised texts while reducing similarity between the source text and their own texts. Except for Xiang’s example, use of MT and PE by Ying, Lian and Yun displays complex metacognitive and metalinguistic processes.
Analysis of Yun’s summarising process (Figure 6) shows that she has drawn on her repertoire of languages in Chinese and English to read academic articles in L2 English, to summarise the information in L1 Chinese, and then to translate this into L2 English. This sequence offers insight into the complexity of her translanguaging practices, i.e., alternating use of two languages for input and output activities (see also Cenoz & Gorter, 2021Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. Cenoz, J., & Gorter, D. (2021) Pedagogical translanguaging. Cambridge University Press. ). L1 served as a cognitive and linguistic resource for thinking, which helped Yun engage in the first cognitive (disciplinary knowledge) and linguistic activity, summarising. She then continued with the second cognitive (epistemic) and linguistic process, translating. MT and PE were involved in the translation practice from the third to the sixth step to assist with revision and improving the summarised text in L2 English. This demonstrates that Yun employed her linguistic repertoire, prior knowledge, MT, and cognitive skills to move back and forth between two language and knowledge systems, for checking the accuracy of meaning and syntax in her own English writing, conveying her intended meaning in L1 Chinese, and comparing HT and MT outputs for making modifications. Yun thus drew on multilingualism, translanguaging and transknowledging to complete summarising activities.
Analysis of Lian’s paraphrasing process (Figure 3) reveals the complexity of paraphrasing in L2, which requires two rounds of the translation process, i.e., translation and back-translation. This is because paraphrasing is the rewriting of sentences in a new structure but retaining the original text meaning within the same language, as explained by Widdowson (2014)Widdowson, H. G. (2014) The role of translation in language learning and teaching. In J. House (Ed.), Translation: A multidisciplinary approach (pp. 222–240). Palgrave Macmillan. Widdowson, H. G. (2014) The role of translation in language learning and teaching. In J. House (Ed.), Translation: A multidisciplinary approach (pp. 222–240). Palgrave Macmillan. as intralingual textual reformulation and by Jakobson (2021)Jakobson, R. (2021) On linguistic aspects of translation. In L. Venuti (Ed.), The translation studies reader (4 ed., pp. 156–161). Routledge. Jakobson, R. (2021) On linguistic aspects of translation. In L. Venuti (Ed.), The translation studies reader (4 ed., pp. 156–161). Routledge. as intralingual translation or rewording. Thus, when the source text in L2 is difficult to understand and paraphrase, the students need to carry out translation from L2 English to L1 Chinese followed by back translation from L1 Chinese to L2 English. This finding further confirms that paraphrasing is complex and challenging for L2 writers in both metacognitive and metalinguistic processes (Gao & Zheng, 2019Gao, X., & Zheng, Y. (2019) Multilingualism and higher education in Greater China. Journal of Multilingual and Multicultural Development, 40 (7), 555–561. Gao, X., & Zheng, Y. (2019) Multilingualism and higher education in Greater China. Journal of Multilingual and Multicultural Development, 40(7), 555–561. ; Hirvela & Du, 2013Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12 (2), 87–98. Hirvela, A., & Du, Q. (2013) “Why am I paraphrasing?”: Undergraduate ESL writers’ engagement with source-based academic writing and reading. Journal of English for Academic Purposes, 12(2), 87–98. ; Shi & Dong, 2018Shi, L., & Dong, Y. (2018) Chinese graduate students paraphrasing in English and Chinese contexts. Journal of English for Academic Purposes, 34 , 46–56. Shi, L., & Dong, Y. (2018) Chinese graduate students paraphrasing in English and Chinese contexts. Journal of English for Academic Purposes, 34, 46–56. ). This explains why Lian employed MT and PE twice: first to understand and rewrite the source text; and second to translate it back to English. This sequence suggests why Xiang’s use of MT might be problematic. She may not have fully understood the source text because she omitted steps involving post-editing and paraphrasing in her own words. Students who do not ensure their own human translation interaction with the subject matter (information, knowledge, episteme) with MT through PE, not only risk misunderstanding academic texts, but they may also risk breaching academic integrity.
There is early evidence from this study that knowledge of and capability in MT and PE can increase Chinese students’ understanding of source texts in English. Also, using MT to back translate post-edited texts from Chinese to English followed by a further round of PE is necessary to increase accuracy and appropriateness. This sequence of steps seemed to help students in this study to avoid potential plagiarism. Thus, the sequence of steps suggests the ethical potential of MT use together with PE.
Moreover, the findings indicate that translation with MT and PE, namely the complementary use of HT, MT and PE, can facilitate students’ translanguaging and transknowledging processes. This chimes with Ortega’s (2017)Ortega, L. (2017) New CALL-SLA research interfaces for the 21st century: Towards equitable multilingualism. CALICO Journal, 34 (3), 283–316. https://www.jstor.org/stable/90014697. Ortega, L. (2017) New CALL-SLA research interfaces for the 21st century: Towards equitable multilingualism. CALICO Journal, 34(3), 283–316. https://www.jstor.org/stable/90014697. proposition that communication with digital technology ‘rarely results in one-language-only or one-language-at-a-time texts’ (p. 294). The students in this study ultimately work with at least two languages to produce sophisticated one-target-language texts through a complex translingual process with the use of MT and PE. The data illustrates how explicit use of translation involving HT, MT and PE, involves vertical translanguaging and transknowledging. Together the data suggest that the three modes of translation allow students to connect their prior language(s) and knowledge(s) and to expand their repertoires of language(s) and knowledge(s). In so doing, translation, (vertical) translanguaging and transknowledging appear to advance students’ linguistic (bilingual or multilingual) and epistemic capabilities (see also Heugh et al., 2022Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35 (1), 89–127. Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35(1), 89–127. ) and academic writing.
Concluding remarks
This article addresses the potential and role of translation as a key feature of both translanguaging (particularly vertical or pedagogical translanguaging) and transknowledging within a broader set of (southern) perspective of multilingualisms in education. This broader set of multilingualisms arises from decades of research in highly multilingual societies in postcolonial Africa, the Asia-Pacific and more recently in Australia. The article further explores the potential of the interaction between HT and human language translation technologies (MT technologies) with a long association with AI. Chang’s study builds on early experimental engagement with MT in undergraduate Applied Linguistics and English Additional Language courses and a study on human language translation technology (Heugh et al., 2022Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35 (1), 89–127. Heugh, K., French, M., Arya, V., Pham, M., Tudini, V., Billinghurst, N., Tippett, N., Chang, L. C., Nichols, J., & Viljoen, J.-M. (2022) Multilingualism, translanguaging and transknowledging: Translation technology in EMI higher education. AILA Review, 35(1), 89–127. ). Chang’s study was designed to explore whether an intervention providing explicit guidance to students on how to use MT (and by extension other AI tools) can increase meaningful access to and engagement with academic readings, such that students’ paraphrasing and summarising capabilities are enhanced. Secondly, the intervention was designed to explore whether such guidance could facilitate strategies that would assist students to maintain academic integrity through deliberative use of human interaction (cognitive engagement) with MT.
Earlier substantive system-wide and longitudinal studies in the African context argued that the horizontal and spontaneous practices of translanguaging can support positive affective conditions for learning, student confidence and conviviality. We also argue based on 15 years of exploration of multilingual and translanguaging pedagogies in an Australian university and Chang (2022Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18 (1), 110–136. Chang, L. C. (2022) Chinese language learners evaluating machine translation accuracy. JALT CALL Journal, 18(1), 110–136. , 2024 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629 (2024) An investigation of the potential of machine translation and post-editing to advance students’ learning of Chinese and English languages in higher education [Doctoral dissertation, University of South Australia]. https://hdl.handle.net/11541.2/39629) that systematic use of vertical translanguaging is necessary in high-stakes academic contexts of higher education. We further argue that multilingual education involves more than a focus on language and translanguaging. It involves reciprocal or two-way exchange and translation of knowledge, that is, transknowledging (see also Windle et al., 2023Windle, J., Heugh, K., French, M., Armitage, J., & Chang, L. C. (2023) Reciprocal multilingual awareness for linguistic citizenship. Language Awareness, 32 (4), 582–599. Windle, J., Heugh, K., French, M., Armitage, J., & Chang, L. C. (2023) Reciprocal multilingual awareness for linguistic citizenship. Language Awareness, 32(4), 582–599. ; Song & Lin, 2025Song, Y., & Lin, A. M. (2025) From translanguaging to transknowledging: decolonizing knowledge production in applied linguistics. Applied Linguistics Review. Song, Y., & Lin, A. M. (2025) From translanguaging to transknowledging: decolonizing knowledge production in applied linguistics. Applied Linguistics Review. ).
Chang’s micro-study discussed here, while focussing specifically on two areas of academic writing, paraphrasing and summarising, along with student use of MT and PE, expands previous research and literature. Whereas much attention is paid to the relationship between language and translation, far less attention has been paid to the relationship between translation as conduit for epistemic exchange across linguistic and disciplinary divides, especially in EMI. The micro-study furthermore offers insight into how students navigate both the metacognitive and metalinguistic aspects of knowledge translation. These insights respond in part to concerns that student use of MT may result in academic integrity violations. Students in this study demonstrate that when they are guided towards human engagement with MT, their cognitive and linguistic processes go beyond simple ‘plagiarism’.
There are limitations to this study. Multilingual students need explicit pedagogies to support high-level academic capability in both the home / preferred language and English (in EMI contexts). These pedagogies can include horizontal (spoken or spontaneous) translanguaging, but they also require vertical or pedagogical translanguaging through translation in reading and writing academic texts across the disciplines. These in turn require explicit attention to translation (HT, MT and PE), and this brings about transknowledging capability.
Further implications of this study include the potential for a larger, more comprehensive and longitudinal, study of HT, MT and PE. Second, while AI developments with ChatGPT and other technologies may have overtaken MT systems, their outputs are not yet reliable or error-free (Moneus & Sahari, 2024Moneus, A. M. & Sahari, Y. (2024) Artificial intelligence and human translation: A contrastive study based on legal texts. Heliyon, 10 (6), e28106. Moneus, A. M. & Sahari, Y. (2024) Artificial intelligence and human translation: A contrastive study based on legal texts. Heliyon, 10(6), e28106. ; Tsou et al., 2024Tsou, W., Lin, A. M., & Chen, F. (2024) Co-journeying with ChatGPT in tertiary education: Identity transformation of EMI teachers in Taiwan. Language, Culture and Curriculum, 37 (4), 529–543. Tsou, W., Lin, A. M., & Chen, F. (2024) Co-journeying with ChatGPT in tertiary education: Identity transformation of EMI teachers in Taiwan. Language, Culture and Curriculum, 37(4), 529–543. ). Human cognition and interaction with the technology continues to be necessary in high-stakes settings. For this reason, we argue that the trial pedagogical intervention here alerts HEI students and academic staff to the potential value of human interaction with MT and AI more widely. We suggest therefore that the findings of this study are portable to wider discussions of multilingualisms in education and AI tools and their use beyond applied linguistics.
Funding
Open Access publication of this article was funded through a Transformative Agreement with Adelaide University.
Acknowledgements
We sincerely appreciate the insightful feedback provided by two reviewers of this paper.