Uniformitarianism

For language, uniformitarianism is the idea that arguments about linguistics generalize. Claims about how language interacts with social organization, culture, and the brain must be generalizable if they are to be useful. Uniformitarianism therefore gives us a benchmark against which we can evaluate general claims about language and language change. Where it appears to fail, it encourages us to look deeper. The comparative method in historical linguistics is based on this principle. The uniformitarian principle also underlies the methods in computational phylogenetics. Researchers in the humanities and social sciences investigating the past face an acute data problem. The vast majority of past human behavior, material and intellectual culture, and human ecologies is undocumented and unpreserved in the historical and prehistorical record. What is there is sparse, unevenly sampled, and often removed from its context. Reconstruction methods (that is, methods which compare systematic variation among contemporary and recorded items) allow us to recover some of this information, but these methods are contingent on contemporary data availability and on the accuracy of the assumptions used to infer change, and are subject to various biases because of ways data are lost. Uniformitarianism is therefore one way to project insights about the present back into the past. Assumptions about uniformity pervade science, including linguistics. However, the extent to which uniformitarianism is useful as a distinct principle, in and of itself, is quite unclear. In this chapter, I begin with a brief review of work on uniformity and the history of uniformitarian thought in linguistics. In Section 2 I discuss some of the ways that uniformitarian concepts have been defined. In Section 3 I raise some consequences for uniformitarian thinking in historical linguistics, while in Section 4 I briefly discuss two case studies.