This study focuses on South Korean university students' perceptions of the English-Medium Instruction (EMI) courses that they have experienced. Research into EMI in Asian contexts suggests that, there is not only a rapid implementation of EMI policy linked to internationalisation, but the policy is also often top-down driven, and hence, is not sensitive to the linguistic challenges of individual university contexts; i.e. it is not context appropriate (D.-W. Cho, 2012; Dearden & Macaro, 2016; Kang & Park, 2005). This thesis explores an EMI context at a research-intensive university in South Korea through semi-structured interviews with eight engineering and business students. The interview data are analysed using Constructivist Grounded Theory (CGT) (Charmaz, 2006, 2014), enabling a bottom-up, empirically, grounded understanding of the context. A constructivist approach enabled my a priori thinking to play a role in the selection, sifting, and organisation of data. Through the grounded theory analysis I developed four distinct micro-CGTs which are labelled as: 'Past Experiences Shape Trust'; 'Support Shapes Trust'; 'Trust Affects Access'; and 'Trust Affects Interactions'. The results indicate that trust plays a fundamental role in the students' EMI experiences. Trust, is both, being shaped by students' past experiences and support, and is affecting their access to subject content and interactions. For example, experiences of preparing, and support for EMI, form a degree of trust in situated linguistic capital. Furthermore, a degree of trust affects affordances that flow from situated linguistic capital, and these affordances promote future EMI access to subject content and interactions. From the above four micro-CGTs, a single macro-CGT was developed, which I refer to as 'Situated Linguistic Capital Theory'. This theorises an ecological situated version of linguistic capital, thus extending existing versions theorised in terms of social capital and power. The situated version describes how linguistic codes present linguistic capital in different situations. Thus, this thesis highlights the perceptions of trust that students have in their first language and second language linguistic codes in different situations. I conclude the thesis by considering if EMI in South Korean higher education might move towards a Contextualised EMI (CEMI), which could cater to the particularity of the context.