The computational neuro-architecture of language has been difficult to characterize. A long-standing excuse for the field’s slow progress as compared to, say, the visual domain, is the lack of an animal model for such a unique human trait. However, over the last 20 years of neuroscience research, we have made enough progress in mapping language functions to reveal close neuro-architectural homologies, not only between human language and systems in non-human primates, but between linguistic and nonlinguistic systems within the human brain. This realization has guided the development of new models of the neurocomputational architecture of language. I will discuss two architectural models covering phonological and syntactic domains. I will highlight homologies between linguistic levels of representation as well as to non-linguistic sensorimotor and sensory-conceptual systems in human and non-human primates. These advances clarify the neural organization of language and its possible evolutionary history.