Neurocognitive studies can approach gene-based developmental language impairments from two angles, which are complementary and ideally combined in a research program. One approach aims at an optimal phenotypic description of a disorder and from there proceeds to a biological and developmental understanding. Complementary to such a top-down approach, a bottom-up perspective will primarily focus on potential etiological pathways and attempt to explain complex outcome phenotypes in terms of elementary developmental disturbances. My paper is dedicated to this latter approach. I argue that in behaviorally defined disorders (such as specific language impairment or autism) shared genetic risk and common etiology can at best be expected for specific aspects of language deficit and that such shared etiology will only apply to subtypes of these disorders. One reason for this skepticism is that the emerging language system in children can be affected in many different ways via more elementary sensory, perceptual, cognitive, and motor impairments. Neurocognitive research on developmental language disorders relies on an understanding of such potential elementary disturbances before it can confidently proceed to the study of complex linguistic impairments.