Oxford University Press (forthcoming; co-edited with Annalisa Coliva)
Philosophical Studies (2021) 178: 493–513
Analysis (2021) 81: 207–215 (with Evan Welchance)
Synthese (2022) 200: 1–14 (with Annalisa Coliva)
In M. Baghramian, J. A. Carter, and R. Rowland (eds.), The Routledge Handbook of Philosophy of Disagreement, Routledge (2022) (with Annalisa Coliva)
The overwhelming consensus among commentators is that G.E. Moore’s proof of an external world is a circular failure: whether its intended target is the skeptic or the idealist, Moore’s belief that he has hands fails to be independent of his belief that there is an external world. Hence, Moore’s proof fails to be a good proof.
Moore’s proof is puzzling in large part because nowhere in his writings does he seem to address or anticipate this worry. Indeed, that Moore seems largely oblivious to such a worry is one of the reasons the proof has struck so many as naïve. By drawing on archival evidence and overlooked passages in Moore’s posthumously published lectures, my dissertation, Proof and Circularity, Reconsidered, challenges this consensus in several unprecedented ways.
Conspicuously absent from discussions of Moore’s proof is an account of how Moore conceived of circularity. No such account exists. I provide one, showing that Moore was not only deeply engaged with the problem of circular proof but, as early as 1927/28, building on the work of logicians W.E. Johnson and J.N. Keynes, Moore formulated an anti-circularity condition that guards against the kind of epistemic circularity glossed above. This leads to a puzzle. For Moore’s “Proof of an External World,” (1939) contains no mention of this condition; its absence naturally suggests an oversight of Moore’s.
Archival evidence tells a different story. Building on this evidence, I claim that by 1939 Moore had reached a kind of aporia with respect to circularity prompted by his struggles with the contextual nature of proof giving—the idea that circularity is context-relative and dependent on an arguer’s and audience’s beliefs. Moore, that is, struggled to formulate an anti-circularity condition capable of capturing this contextuality: that his proof is circular relative to an audience of skeptics, but cogent relative to an audience of idealists, his intended target. Yet, despite Moore’s struggles to formulate such a condition, I urge that he never abandoned it and that a close and careful reading of “Proof” supports this conjecture.
First, I show that Moore’s long analysis prefacing his two-handed proof is an attempt at providing non-circular justification for its key anti-idealist premise: not the proof’s first premise, but its second, conditional premise, i.e., that hands are externally constituted. Here, I argue, Moore implicitly satisfies this anti-circularity condition even if this condition isn’t explicitly included among the three conditions Moore enumerates in “Proof.” The upshot is that Moore's proof isn't obviously circular against certain idealist views.
Second, I show that the concluding passages of “Proof” can be read as a confessional of sorts, one wherein Moore acknowledges the circularity of his proof relative to an audience of skeptics. The unexpected upshot is that Moore and some of his most vehement critics are in agreement: his proof is circular. This, I suggest, provides further evidence against received readings of Moore’s proof, i.e., “epistemological” readings that characterize the proof as an anti-skeptical proof. On my reading, then, Moore only considered his proof to be circular (and thereby ineffective) against the skeptic but not his intended target, the idealist.
What results is a new, deeper, and more contextually faithful reading of Moore’s proof, one that resolves various longstanding exegetical issues that have puzzled commentators since the proof’s publication in 1939.