ABSTRACT
The conclusion that should be drawn from these reflections is not that Hume's subjectless solipsism is incoherent or refutable on the basis of the very language he used. Quite the contrary, the language Hume used (as ordinarily understood) is based on assumptions that Hume, on the basis of plausible epistemic principles, has shown to be baseless and questionable. If his principles are not amended, revised, or superseded by better ones, the appropriate outcome is perhaps Wittgensteinian silence: "W ovon man nicht sprechen kann, dariiber muss man schweigen." 35 To ask, as Wittgenstein did in the private language argument he included in his Investigations, whether sentences (one could equally say opinions) about certain sensory objects are "correct" is to assume distinctions whose credentials Hume implicitly challenged. I believe and shall argue in the next chapter that such distinctions are, in fact, defensible and that Wittgensteinian silence is not the appropriate outcome to Hume's skeptical reflections. But to defend those distinctions one must solve the problem about observation that I posed earlier in this section - the problem of how the principles or general considerations needed to assess observational beliefs can be rationally defended in a critical, anti-Cartesian way. The key to solving this problem is a satisfactory theory of experimental inference.
