Many types of human behaviour, from scientific research to political decision-making, are based on implicit assumptions, considered to be so self-evident that they do not need any further justification. Such assumptions are particularly powerful in topics related to language: one of the most universal and fundamental human abilities and a prerequisite for social life, civilisation and culture. They become a driving force in the current debates about multilingualism.We identify three central assumptions underlying key controversies related to language: (a) the "limited resources model" assuming that learning languages has a detrimental effect on learning other subjects, (b) the notion that the "normal" state of human brain, mind and society is either monolingualism, or a strong dominance of a "mother tongue", accompanied by less relevant "additional" languages, (c) the belief that the aim of language learning is a "native-like" proficiency and anything that fails to reach it has only limited value.Combining radically different academic backgrounds (education and cognitive neuroscience) and methodologies (qualitative and quantitative) we examine how these assumptions influence attitudes towards multilingualism. We evaluate the available empirical evidence and explore conceptual common ground, from the design of school curricula to the promotion of healthy ageing. We conclude that our perspectives complement each other, providing a valuable tool to inform language policy.