The aim of this study is to investigate Swedish upper secondary school students ́ views on their use of English and Swedish in different situations during English lessons and their perceptions of the optimal balance between the two languages to achieve the learning outcomes of English as an upper secondary school subject. The Willingness to communicate (WTC) model (Macintyre et al., 1998) and the concept of codeswitching (CS) (e.g. Carson, 2014; Simon 2001) were adapted as the theoretical framework. A qualitative interview method was used for purposeful data elicitation, and for examining the first-person perspectives of four students at different ages and different levels of proficiency.The findings of this study were discussed in view of previous research, through which it wasindicated that the optimal balance between the target language (TL) and the first language (L1) might be generalizable to some degree. A finding congruent with several previous studies was that the target language was considered to be a desirable and purposeful codeduring communicative activities. However, it emerged that students' own willingness to use TL alone was often not a sufficient driving force for actual TL use, whereas successful TL use was likely to take place when the teacher's clear, firm instructions and management for encouraging students to use TL were accompanied. In addition, the first language was thought to be the most fulfilling code in form-focused tasks such as grammar exercises as well as in information intensive tasks, whether students worked individually or in pairs or groups. Although students felt comfortable with using TL with interlocutors who they know well andin small groups, they from time to time spoke TL beyond their comfort zone, adapting to different situations. Within this general tendency, there were yet individual differences in defining the optimal TL/L1 balance. Proficiency and self-perceived competence in L1 also emerged as important for code choice between TL and L1, although not the most decisive factor. Based on the findings, this study suggests that pedagogical interventions, such as constructive, informed dialogues about TL/L1 use, can help students plan and deploy strategies for using the two languages to benefit their TL learning.