Humans are peculiarly badly placed to understand why language is as it is. An ornithologist studying the nest-building habits of a particular bird species, for example, has ample evidence from other species to determine which habits of the bird in question are unique, which are unusual but shared with other species that may be closely related to it, and which are general, shared with many species. The unusual-but-shared habits will be particularly useful in establishing the species’ prehistory and genetic relationships, especially if these habits serve no obvious contemporary function and are therefore unlikely to have been diffused independently through more than one bird population owing to similar selection pressures. But scientists studying the vocalisation and communication habits of humans have no such comparisons to assist them. No other species’ communicative repertoire is anything like as elaborate as ours — or at least, any superficially comparable elaboration seems to serve much narrower ends, such as self-advertisement for mating purposes.The purpose of this article is to draw attention to three features of language that (I suggest) can be understood properly only by recourse to the organismas- document approach. I hope thereby to stimulate more linguists who are interested in the history of language to look for other such features.