Foggy connections, cloudy frontiers: On the (non-)adaptation of lexical structures.

While research on possible adaptive processes in language history has recently centered mostly on phonological variables, here, I return the focus on the lexicon in two different ways. First, I take up the familiar theme of the responsiveness of language structure to the local conditions at different elevations of the earth's surface by exploring further the idea that language communities at high altitudes may tend not to distinguish lexically, as, e.g., English does, between "cloud" and "fog." Analyses of a global dataset of languages as well as in-depth study of the languages of the Central Andes are consistent in showing a wide spread of colexification of "cloud" and "fog" across elevations, whereas distinguishing languages tend more to be spoken at lower elevations. Statistically, there is global support for the idea that colexification is triggered by high elevation, but a closer look, in particular at the Andean dataset, paints a more nuanced picture. Concretely, it shows that in some language families, there are consistent preferences for either colexifying or distinguishing between "cloud" and "fog." In particular, the behavior of the large Quechuan family, which ranges across high- and low-elevation environments but still is consistently colexifying, shows no evidence for adaptive processes within language families. This result is open to various interpretations and explanations, for they suggest lineage-specific preferences for or against colexification that run counter to global trends. It is also at odds with the notions of "efficient communication" and "communicative need" as far as they relate to lexical categories and bars mechanistic or deterministic views on the processes in which the categories of languages are molded.