Invited Historical Essay

—In The Descent of Man (1871), Darwin observed “curious parallels” between the processes of biological and linguistic evolution. These parallels mean that evolutionary biologists and historical linguists seek answers to similar questions and face similar problems. As a result, the theory and methodology of the two disciplines have evolved in remarkably similar ways. In addition to Darwin’s curious parallels of process, there are a number of equally curious parallels and connections between the development of methods in biology and historical linguistics. Here we briefly review the parallels between biological and linguistic evolution and contrast the historical development of phylogenetic methods in the two disciplines. We then look at a number of recent studies that have applied phylogenetic methods to language data and outline some current problems shared by the two fields. [Comparative method; Darwin; evolution; historical linguistics; phylogeny; Schleicher.] CURIOUS PARALLELS IN THE DOCUMENTS OF EVOLUTIONARY HISTORY In The Descent of Man, Darwin (1871) noted that the process of evolution is not limited to just the biological realm. “The formation of different languages and of distinct species, and the proofs that both have been developed through a gradual process, are curiously parallel . . . . We find in distinct languages striking homologies due to community of descent, and analogies due to a similar process of formation.” (pp. 89–90) Many of the fundamental features of biological and linguistic evolution are demonstrably analogous (see Table 1 and Croft, 2000). Just as DNA sequences contain discrete heritable units, so too do languages in their grammatical and phonological structures and their vocabularies (lexicons). These may differ from language to language and can be inherited as the languages are learned by subsequent generations. With knowledge of the processes of linguistic change, it is possible to identify homologous linguistic characters that, like homologous biological structures, indicate inheritance from a common ancestor. For example, homologous words, or cognates, meaning “water” exist in English (water), German (wasser), Swedish (vatten), and Gothic (wato), reflecting descent from proto-Germanic (*water [in historical linguistics, inferred proto forms are denoted with a ‘*’]). Cognates are words of similar meaning with systematic sound correspondences indicating they were related due to common ancestry. Processes of mutation and random drift can operate on linguistic characters, just as they do on genes. An example of lexical mutation, or innovation as it is known in linguistics, is the word boy, which arose at some point after English split off from the other west Germanic languages (Campbell, 2004). A phonological example is the unconditional sound change of /t/ to /k/ in Hawaiian. So the ancestral Polynesian word *tapu, “forbidden,” changed to kapu, and *tolu, “three,” changed to kolu, and so on (Crowley, 1992). As well as these “point mutations,” words, like gene sequences, can show insertions (e.g., Old Swedish *bökr, “books,” to böker; Campbell, 2004), deletions (e.g., Proto-Oceanic *tupa, “derris root,” to Selau tua; Blust, 2003) and reversals, or metathesis (e.g., Old English brid to modern English bird). Linguistic changes, like changes in biological form, are also sometimes structurally and/or functionally linked. Terms for “five,” for instance, tend to be correlated with terms for “hand” for obvious reasons. As in biology, mutation and drift create variation that may be subject to selection. For example, Pawley and Syder (1983) found evidence that Darwinian selection pressures have acted on English syntax. Specifically, they identified differences between vernacular and literary English grammar that they argue were “adaptive to the particular conditions imposed by the mode of language use” (p. 577). The fundamental process of language formation involves cladogenesis, where a single lineage splits to form two new languages. Often, as in biology, this is due to geographic separation or migration events. Horizontal gene transfer and hybridization also have a linguistic equivalent in borrowing between languages. For example, the English word mountain is borrowed from French, montagne. Borrowing between languages can produce reticulated evolution in a similar fashion to horizontal gene transfer in plants or bacteria. Extreme cases of contact between languages can produce a form of language “hybrid,” as in the case of some Creoles. For example, Sranan, a Creole spoken in Surinam,