One approach to comparing the neural bases of language and music is through the use of song, which is a unique and ecological combination of these two cognitive domains. In song, language and music are merged into one acoustic signal with two salient dimensions. By manipulating either the linguistic or musical dimensions (or both) of song and studying their relationships, it is possible to gain important information about the neural networks underlying language and music cognition. We will present a brief review followed by recent behavioral, electrophysiological, and neuroimaging studies concerned with the functional and structural relationships of music and language. These results, together with the previous studies in the field, help understanding whether the different levels of music and language processing are independent or interactive.