In this paper, I discuss the pitfalls of the anthropological notion of culture when used in the field of ethnomathematics, and use Dowling’s myth of reference and myth of emancipation to show how these are linked to what I refer to as the dual dangers – the dangers of decontextualization and knowledge colonization in the field of ethnomathematics. I argue that these two dangers arise from a naïve approach to ethnomathematics research that is informed by the anthropological notion of culture. The paper provides some examples from the Philippines and one from New Zealand on how these dual dangers may manifest while pursuing studies in ethnomathematics and programs for culturally relevant mathematics education.