Recently, much philosophical discussion has centered on the best way to characterize the concepts of random drift and natural selection, and, in particular, on the question of whether selection and drift can be conceptually distinguished (Beatty 1984; Brandon 2005; Hodge 1983, 1987; Millstein 2002, 2005; Pfeifer 2005; Shanahan 1992; Stephens 2004). These authors all contend, to a greater or lesser degree, that their concepts make sense of biological practice. So, it should be instructive to see how the concepts of drift and selection were distinguished by the disputants in a high-profile debate; debates such as these often force biologists to take a more philosophical turn, discussing the concepts at issue in greater detail than usual. A prime candidate for just such a case study is what William Provine (1986) has termed “The Great Snail Debate,” that is, the debate over the highly polymorphic land snails Cepaea nemoralis and Cepaea hortensis in the 1950s and early 1960s. This study will reveal that much of the present-day confusion over the concepts of drift and selection is rooted in confusions of the past. Nonetheless, there are lessons that can be learned about nonadaptiveness, indiscriminate sampling, and causality with respect to these two concepts. In particular, this paper will shed light on the following questions: 1) What is “drift”? Is “drift” a purely mathematical construct, a physical process analogous to the indiscriminate sampling of balls from an urn, or the outcome of a sampling process? 2) What is “nonadaptiveness,” and is a proponent of drift committed to claims that organisms’ traits are nonadaptive? 3) Can disputes concerning selection and drift be settled by statistics alone, or is causal information essential? If causal information is essential, what does that say about the concepts of “drift” and “selection” themselves?
It is a remarkable feature of language, and of human understanding, that the same word refers to various individual objects. This is most evident in the case of generic terms (common nouns). For example, the word 'tree' may apply to one tree in one instance and another in another instance. Because the word applies to different individuals in each case, yet the same cognition "x is a tree" is present, the question arises, "of what object does the word convey knowledge?" In studying their discussions on this topic, the present dissertation translates Patanjali on Panini 1.2.58 and 1.2.64, Vatsyayana on Gautama 2.2.58-69, and Sabara on the Akrtyadhikarana. Patanjali considers the merits, difficulties, and philosophical presuppositions of the extreme views that a word denotes just an individual and that it denotes just a generic property. If a word naturally denotes whatever individual objects it refers to, one cannot account for the occurrence of the same cognition regarding each. The same cognition regarding each of many individuals of a kind is explained as based on a generic property present in each. But if a generic term denotes as a single generic property, one cannot account for linguistic usages reflecting the comprehension of distinction in number, differences in quality, or participation in action, all of which naturally belong to individuals. Because the same cognition and knowledge of distinction, etc. are both present when a generic term is used, Patanjali concludes that the word denotes both the generic property and an individual. In Nyaya, Gautama, as explained by Vatsyayana, favors this solution too. He includes shape as a third element denoted because it is a means to infer the generic property. Sabara affirms that a generic property alone is denoted, demonstrates that cognition of a generic property must derive directly from the word, and explains how cognition of an individual derives from cognition of the generic property. Though his explanation is feasible, his proofs that an individual is not also denoted, depending on a doubtful example and biased application of presumption, are vulnerable.