“EVIDENTIAL HOLISM,” AS I WILL CALL IT, is the view that individual (“isolated”) hypotheses do not receive evidential support—that only whole systems or groups of hypotheses do.
By contrast, what I shall call “evidential particularism” is the view that individual (“isolated”) hypotheses can receive evidential support, whether or not they are part of an entire system of hypotheses.
These are bare-boned descriptions of the two positions, and each needs spelling out. This can be done in various ways, yielding different versions of the basic views. Here, I will examine one important version of each. The version of evidential particularism I explore is advocated by John Stuart Mill.[CXIII] By contrast, the version of evidential holism I am interested in is not so clearly attributable to anyonein particular, but is suggested in the writings of Whewell,[114] [115] Duhem,[116] and Quine.[117]
The views I present here are bold and brassy. Each has important implications for what is to count as evidence for a scientific theory. Both, I will argue, are (truth-relevant) speculations, in the sense I characterized in chapter 1.[118] Their proponents may believe they have evidence for these speculations, but they do not, at least none sufficient to make them reasonable to believe. Even considered as speculations, however, they have their problems. Those supporting each view make various assumptions for which either no argument is offered or one that is very dubious. In place of each of these positions, I present one that rejects holism completely and defends a type of particularism that is pragmatic and, I think, much richer than that suggested by Mill. Perhaps my version of particularism will still be a speculation. But if it is, I want to claim it is better defended than the versions of holism and particularism I will consider.
1.