Students who immigrate often lose their social networks, face discrimination, and deal with the challenges of adjusting to a new culture (Da Silva & Ravindran 2016). African immigrants enrolled in South African HEIs encounter several challenges such as financial and systemic barriers (Mabaso & Bhengu 2021), academic, language and social barriers (Gumede & Ndlovu 2023; Van Wyk & Davids 2025), as well as experiences of prejudice and discrimination (Omoniyi 2020). Some of the challenges that migrant African students face are similar to those that black South African students—especially from rural contexts—face, such as language barriers in accessing academic discourses (Booyse & Ngwenya 2020; Ndimande-Hlongwa & Khumalo 2022), code-switching, pedagogical, and institutional barriers (Diko 2024; Mabuso & Mhengu 2021; Mkhize 2023), as well as financial, psychological, and emotional stressors (Omoniyi 2020; Van Wyk & Davids 2025). While some of these challenges often originate from a lack of social connections and financial limitations, they lead to stress and psychological issues that are worsened by limited social support (Da Silva & Ravindran 2016). According to Ma (2022), the hurdles these students encounter could be narrowed down to language, culture, classroom discussions, academic expectations, and interpersonal relationship issues. Ma emphasises the importance of teaching methods that consider cultural and language differences to provide adequate support. Their findings also underscore the need for targeted resources and support to address the unique challenges that immigrant students face in finance courses, thereby enabling them to succeed.

According to Gumperz (1982:60), ‘code-switching is the juxtaposition within the same speech exchange of passages of speech belonging to two different grammatical systems or subsystems’. Myers-Scotton (1993:8) sees code-switching as ‘the selection by bilingual speakers of forms from an embedded language in utterances primarily structured by a matrix language’. The above two definitions are important in laying the ground for a critical discussion on the application of code-switching in classroom settings.

In Gumperz’s (1982) seminal work, Discourse Strategies, he highlights four grounding principles in code-switching: a discursive approach to code-switching, the role of grammatical integrity, the intensity of code-switching analysis, and how the careful use of ethnography offers a sound empirical ground for a robust code-switching understanding. Meanwhile, Myers-Scotton (1993, 2002) offers the Matrix Language Frame (MLF) as a model for the grammatical understanding of code-switching. For her, the MLF ensures a more nuanced understanding of code-switching in that it highlights the embeddedness of one or more languages in a dominant language, thus showing how the switching occurs structurally (Myers-Scotton 1993). Myers-Scotton (2002) further suggests that code-switching has a predictive power through the way words are organised and the distribution of morpheme, as well as that code-switching has cross-linguistic applicability.

Code-switching provides multiple benefits in educational contexts for students acquiring English as a second language (L2). Several studies have highlighted the significance of code-switching in enhancing learner engagement and comprehension, particularly in contexts where English is not the primary language (Mokgwathi & Webb 2013; Nguyen, Grainger & Carey 2016; Shafi, Kazmi & Asif 2020). Positive results have been recorded with code-switching in Mathematics courses noted to promote clarity in difficult concepts and enhance student understanding (Ali & Ihsan 2023). These studies collectively highlight the importance of code-switching as an instructional strategy that can improve the learning outcomes for L2 English learners.

To O’Leary (2010), AT is a psychological framework stressing the interaction of tools, results, and the community in forming human activity. The main ideas of the theory are (Engeström 2009):

Just as in other fields, AT has been used in project-based language learning (Gibbes & Carson 2014), qualitative analysis (Hashim & Jones 2007), and enterprise ontology (O’Leary 2010). A model activity system is shown in Figure 1.

Lam (2021) and concurs that AT facilitates understanding of the complex relationships and mediations involved in learning activities. In support, Lam (2021) and Çakıroğlu et al. (2016) found it relevant when studying the mediating effect of technology, such as peer feedback systems, in online learning contexts. Emphasising the social and cognitive factors that impact its implementation, Yarzebinski et al. (2015) broaden AT’s scope to include code-switching in classrooms with students who speak more than one language. An excellent foundation for learning by means of tutorials, AT permits in-depth examination of code-switching’s many components and mediations.

The main contribution of Engeström (2009, 2014) to AT is his focus on extensive learning and growth in particular situations. He emphasises how collective activity is shaped by mediation and conversation (eds. Sannino, Daniels & Gutiérrez 2009). Engeström’s work has also had an impact on organisational studies by expanding the understanding of human activity to embrace formal organisations (Blackler 2009). These contributions have expanded the scope and relevance of AT, bringing it new dynamism in accounting for human behaviour and learning.

Code-switching can be studied in educational settings using tools provided by AT. At its core, AT offers a framework for examining human activities that are mediated by tools and artefacts. Applied to code-switching, it allows the researchers to view this sometimes-charged phenomenon through the perspective of the dense social setting in which it occurs. Who are the various actors involved in code-switching, and what interrelationships and possible tensions exist among them? As students and teachers, we are all part of the learning environment that sometimes engulfs us in educational code-switching (Desyana, Murni & Agust 2020). Activity Theory could be a valuable means of studying the types of code-switching that sometimes occur in English-language medium classes, along with the contextual variables that affect its usage (Hendra 2018). More generally, it could illuminate the sometimes-contradictory tensions across various educational settings in which code-switching is used.

The qualitative case study method is particularly suitable for this investigation, given its small and concentrated sample size. This study focused on five African immigrant tutees and three tutors, all of whom were engaged in finance tutorials. Enhanced by the unique features of the context, the case study method gains strength by narrowing its focus to tutors working with students whose first language is not English.

The data were acquired using two primary methods: focus group interviews and observations of random instructional sessions. This approach was deemed appropriate for this study because it had previously proven beneficial in investigating the use of code-switching in first-year finance tutorials. For instance, Irawan (2022) and Mauliddiyah, Murni and Mustofa (2020) coded their conversations after engaging in the focus group method. They, too, were searching for answers concerning code-switching, albeit in English as a Foreign Language (EFL) context. Harmilawati (2018) and Subon and Tarmim (2021) also employed focus group methods, which reinforced our decision to use this technique, particularly in our campus context. Using conversations collected from focus groups, these latter two studies found that code-switching served beneficial functions for both EFL teachers and their students.

The unobtrusive observations of tutorials marked the start of the data collection process. Deliberate choices were made to select the tutorials that involved three tutors and seven tutees from African countries, none of which are English-speaking. In these settings, the likelihood of code-switching between English and other African languages is presumably highest. Observing these tutorial sessions provided the authors with the first glimpse of the interactions between the participants and tutors. If this premise were incorrect and the instances of code-switching the researchers were looking for didn’t actually occur in any significant number, then the research project could not have gotten off the ground. But the project did get off the ground, as the first layer of conversation analysis was recorded during the observations.

Data analysis began with coding, followed by the application of relevant theoretical frameworks. A two-tiered coding approach (Bryman, Bell & Hirscsohn 2021) was employed. First, thematic coding was applied based on key themes guiding the interviews and observations. These themes included participants’ demographic backgrounds (e.g., nationality, language spoken at home), tutors’ level of study and linguistic competencies, tutees’ linguistic backgrounds, and dispositions towards code-switching held by both tutors and tutees. The second level of coding involved applying relevant variables derived from AT, including the subject, tools, object, rules, community, division of labour, and outcome. Table 1 illustrates the coding process using these variables.

The chosen approach of thematic and theoretical coding is justified by its ability to capture the multifaceted nature of code-switching in the context of finance tutorials for African immigrant students. Thematic coding enables the identification of key themes related to code-switching, such as interactional functions (Candilas et al. 2023; Galegane 2020) and students’ attitudes towards this practice (Ma 2014). Theoretical coding, informed by AT, facilitates the analysis of these themes through the lens of social mediation (see Table 1). Furthermore, research on the types and functions of code-switching in educational settings (Purnama & Bali 2020) provides a foundation for interpreting the coded data within the specific context of finance tutorials for first-year students.

This article is one of the outcomes of a department-wide longitudinal study examining teaching and learning in entrepreneurship at the host institution. Permission was granted by the host faculty and ethics clearance was awarded by the institution ethics board.

From the above, it was possible to capture, code and analyse the data as presented in the findings and discussion sections of this article.

Under this theme the researchers discuss how the tutors and tutees perceive code-switching as a pedagogical tool and the analysis thereof, using AT.

For tutors, the object of the activity is to enhance understanding of lecture material by adapting teaching methods and providing explanations in languages familiar to the students (French or Lingala). Conversely, tutees aim to grasp challenging lecture content more effectively through explanations delivered in their preferred language. A key pedagogical tool employed by tutors is a conscious linguistic adaptation, using various languages (English, French, Lingala) to explain complex concepts and bridge the language gap. This is evident in other research that explores peer learning and tutoring in multilingual settings (Diko 2024; Mthethwa 2021). Another evident tool is pacing and simplification, where tutors adjust the speed and complexity of explanations to cater to tutees with varying levels of English proficiency. This was evident in the narration from Tutee C and corroborates other research such as Mawela and Mahlambi (2021) and Mkhize (2023).

The application of these tools differs slightly between tutors. For instance, Tutor 2 frequently interjects with English phrases to check for understanding or repeat explanations in English after using kiSwahili. In contrast, Tutors 1 and 3 predominantly use French and Lingala, respectively, with minimal code-switching back to English. This variation might reflect individual tutors’ comfort levels using multiple languages during tutorials and their awareness of the pedagogical value of code-switching. This is similar to the findings in the research by Shinga and Pillay (2021) who explored teachers’ levels of confidence in using code-switching as a pedagogical tool, as well as Mthethwa’s (2021) research, which examines the uncomfortable position of English as a language of instruction in multilingual settings.

The tutorial interactions involve the Department of Entrepreneurship and potentially benefit other students from diverse backgrounds who might find multilingual explanations helpful. The department and faculty operate within a context that necessitates such interactions due to the multicultural composition of the student body. In terms of the rules, academic norms dictate that lectures are delivered in English, posing a challenge for non-native speakers. Tutors’ approaches promote inclusivity by allowing the use of multiple languages in tutorials despite the dominance of English in lectures. However, tutors’ training does not explicitly address multilingualism, creating a minor inconsistency within the activity system.

In terms of division of labour, tutors assume a dual role as content facilitators and linguistic mediators. This was evident in Tutor 1 serving as both a subject and a tool in the activity system. Tutor 1 served as the subject in the activity system when the object of the activity is to improve comprehension of finance concepts by foreign African students, and as a tool when Tutee C served as the subject of the activity and the object is to enhance tutee’s satisfaction in the tutorial. Activity Theory enables the identification of dynamic mediation (Engeström & Sannino 2025) wherein there can be simultaneous shifts in subject – role performance in the activity system (Lekganya 2021). While this may appear to extend beyond conventional teaching responsibilities, research suggests that employing students’ native languages during tutorials in multilingual settings such as South Africa is a recommended practice (Charamba 2019; Mathole 2016). Tutees actively participate by communicating their language preferences and engaging with content in their preferred language. Engaging actively in tutorials helps African immigrant students understand the concept taught and demonstrates the benefits of linguistic flexibility and tailored teaching strategies.

The findings of this study have far-reaching consequences for the pedagogy of finance education in South Africa. They could lead to profound changes in both the performance of African immigrant students in finance classes and the way educators in those classes teach. To begin with, the use of both English and African languages in instruction would represent a significant shift toward a more inclusive teaching environment. That alone could lead to better comprehension among students of complicated financial concepts. In this way, the study’s findings could lead to improved financial education for African immigrants and, by extension, for finance students in South Africa more broadly.

The study was done at one department of a South African university with a relatively small sample. This could make it more challenging for the results to be applied at other universities or in different areas. The tutorial observation focused on one specific finance concept, the TMV, which may not be relevant to all areas of finance education, making this restriction even more problematic. Using only observations and interviews to collect data may not give a complete picture of the experiences and opinions of both tutors and students.

Research on a larger scale and scope should be undertaken across multiple universities in various regions of South Africa to enhance the generalisability of the results. Employing a mixed-methods research framework in a future study would significantly improve our understanding of university students’ and tutors’ experiences with code-switching in the South African context, particularly among those who use it in a finance course. Finally, stretching the research to include other academic disciplines and the languages of the student populations would serve to validate and generalise the findings.