Noam Chomsky’s theories and insights have had a huge impact on linguistics, philosophy and other branches of cognitive science. The idea of the human faculty of language as a specialized mental organ and the insistence on treating linguistics as a part of natural science have in many ways revolutionized our understanding of language and the mind. In New Horizons in the Study of Language and Mind (2000), he makes two claims with which this thesis is concerned. First, that the study of language should be approached in accordance with what he coins methodological naturalism. Second, that referential semantics is deeply flawed. This thesis unpacks and assesses the arguments for these claims, and then argues that they do not succeed in threatening referential semantics. I begin by providing a map of methodological naturalism and the Chomskyan view of language. I hold that Chomsky is right that a purely naturalistic approach is valuable to linguistics. On the other hand, I argue that the demand that philosophical accounts of language must also accept and work within the doctrines of methodological naturalism is unwarranted. Against this background, I assess three specific Chomskyan arguments against referential semantics. These arguments touch on ontological concerns which go over and beyond the methodological issues. According to Chomsky, the things humans talk about have such intricate properties that it cannot plausibly be the case that words refer to things in the world. I argue that these ontological arguments presuppose too much on behalf of referential semantics. Therefore, I conclude that Chomsky’s strongest case against referential semantics lies in his methodological arguments, but that the ramifications of these arguments are limited to naturalistic theories.