Human beings have an extraordinary capacity for language. Until recently, this capacity was thought to rely on the cerebral cortex, with the emergence of language in humans linked to the evolutionary expansion of higher-order association cortices such as the prefrontal cortex in phylogenetically newer species. However, the cerebellum has also increased in size and cell number at a rate equal to or greater than the expansion of the prefrontal cortex [1], contains 80% of the neurons in the brain, and has almost 80% of the surface area of the cerebral cortex in humans [2], suggesting it, too, might be important for the emergence of higher-order human cognition. Indeed, the expansion of the cerebro-cerebellar system is the largest driving factor of increased brain size over evolution [3]. Relevant for language specifically, the emergence of supratentorial brain regions that support human-specific language coincides with the expansion of the posterolateral cerebellum, including Crus I and Crus II. These cerebellar regions form reciprocal circuits with language regions of the cerebral cortex, including the prefrontal cortex and temporoparietal areas 4, 5•. Human diffusion tensor imaging, tractography, and resting-state functional connectivity has echoed animal studies, finding connections between several posterolateral cerebellar subregions (e.g. Crus I, II, and VIII) and supratentorial language-sensitive areas 6, 7••, 8, 9••. Critically, primary cerebellar damage impacts the development of supratentorial regions, including the prefrontal cortex, providing further evidence for an intimate relationship between the cerebellum and cerebral cortex [10]. Together, evolutionary expansion of the cerebellum and connections between cerebellum and supratentorial language regions provide the neurobiological substrates for cerebellar involvement in language processing, and the enormous expansion of language abilities in humans.

Although historically considered a motor structure, the idea that the cerebellum may contribute to higher-order language emerged from patient studies which found that damage or disruption of the cerebellum during development and in adulthood can result in a range cognitive disruptions. This led to a theoretical framework suggesting that given the perceived uniformity of cerebellar cytoarchitecture, the cerebellum might play a similar computational role in both motor and nonmotor tasks such as language (‘Universal Cerebellar Transform’ [11]). Examples of nonmotor deficits associated with the cerebellum include phonological, grammar, and semantic-pragmatic challenges 12, 13, 14, 15. Crucially, language deficits following cerebellar damage can occur in the absence of motor disturbances, and in the absence of any cerebral cortical damage. More recently, the advent of neuroimaging has enabled a more precise view of the cerebellum’s role in language. Neuroimaging studies find that the cerebellum emerges as a consistent site of language processing starting quite early in development and across the lifespan, and is co-active with language-selective regions [16] and for a variety of motor and nonmotor language tasks 17, 18, 19. Despite this, the cerebellum is rarely incorporated into neurobiological frameworks of language, and frameworks of language that do incorporate the cerebellum often suggest that it plays a low-order role [20]. Here, we review existing empirical evidence for cerebellar contributions to the language processing hierarchy, evidence linking cerebellar language regions to language development and disorders, and discuss unique cerebellar contributions to language.