Human speech shows an unparalleled richness in geographic variation. However, few attempts have been made to understand this linguistic diversity from an evolutionary and comparative framework. Here, we a) review extensively what is known about geographic variation of acoustic signals in terrestrial mammals, using common terminology adopted from linguistics to define different forms of variation (i.e. accents and dialects), and b) examine which factors may determine this variation (i.e. genetic, environmental and/or social). Heretofore, terminology has been used inconsistently within and across taxa, and geographic variation among terrestrial mammals has never been defined as in human speech. Our results show that accents, phonologically different varieties, occur widely in terrestrial mammals. Conversely, dialects, lexically and phonologically different varieties, have only been documented thus far in great white-lined bats, red deer, chimpanzees and orangutans. Although relatively rare among terrestrial mammals, dialects are thus not unique to humans. This finding also implies that such species possess the capacity for acoustic learning. Within primates, the two great apes showing dialects are those who also show extensive cultures in the wild, suggesting that, in hominoids, intricacy of acoustic geographic variation is potentially associated with cultural complexity; namely, both have derived from selection increasingly favoring social learning across varied contexts, including the acoustic domain.