Here's my current naturalistic view, much indebted to Post and Weissman:
1. Skepticism can inform, but never derail, (externalist) metaphysical realism. In my view, skepticism, like the quest for certain knowledge or immediate justification (#2 below), looms all too large in epistemology. Skeptical doubts can serve as a check on our most cherished beliefs (not always fun but constructive); at the extreme, however, they amount to little more than a game (fun but not very constructive). But there are limits. Traditional skepticism, like any number of fantastic hypotheses, can easily undermine most of our claims to know the world, but not the claim that the world exists and has a determinate nature, since the latter claim is presupposed by the skeptical thesis itself.
2. Metaphysical realism is not absolutist, impossibly transcendent, or mystical. Metaphysical claims (including realist claims, of course) are, or ought to be, speculative and testable (in terms of the observable differences that would result if they were true). But notice: Very few things outside of logic and other formal systems can be known with certainty, including many, if not all, of our internal states. So, if certainty were the standard of all legitimate knowledge (or justification), realism and internalist constructivism (in fact, all epistemological views, including Merleau-Ponty’s phenomenalist reduction of material objects to qualifications of possible or actual experience) would suffer alike. Often glossing this latter point, the tendency in antirealist or irrealist circles is selectively to condemn realism for a constitutive limitation of all empirical inquiry.
3. Metaphysical realism, as I understand and apply it, is just that: a metaphysical (rather than epistemic) notion about what there is. It amounts to the rather pedestrian-sounding claim that the nonhuman world (alternatively: the external world, nature, reality, the universe, human-independent reality) exists and has a nature independently of what we think, say, or do. It is not reducible to or dependent on any epistemic or semantic thesis. In particular, beyond the bare assertion of its independence from our conceptions, it does not say anything at all about whether and to what degree we can know this world or how that knowledge might come about. Of course, various epistemic theses seem to provide better or worse explanations of our common experience of the world; but none, including correspondence truth, is entailed by realism.
4. The natural world is manifestly impervious to and predates our existence; it certainly takes no interest (pardon the anthropomorphic flourish) in our efforts to conceptualize or structure it. In an exact reversal of the emphases and methods of contemporary linguistic philosophy (which would have us derive antirealist conclusions from some theory of meaning or language), we ought to locate ("naturalize") epistemology within an overarching theory of the world (just as humans, qua terrestrial mammals, are located within nature as one of its creatures). That is, metaphysical speculation, as a collection of abductive inferences, qualifies and extends our understanding of a world that we have always known, at least in general outline. (For an extended defense of this and the previous paragraph, see Devitt, *Realism and Truth*; see also D. Weissman, *Hypothesis and the Spiral of Reflection*.)
5. There can be no "God’s eye," neutral, transcendent, or otherwise non-sensed view of the world (contra Putnam’s and Rorty’s critiques of realism). To perceive the world is always to employ the perceptual resources of some perceiver or other occupying a determinate spatio-temporal location. That is, knowledge is possible in terms first I (employing the chaotic and relatively autonomous cognitive skills of an infant) and then we (as language and sociality enter the scene more completely) supply in our conceptualizations of experience. Furthermore, our choice of conceptual scheme is mostly arbitrary. (I resist saying it is entirely arbitrary given the plausibility of certain nativist, evolutionary, or naturalized accounts of our thinking.) Of course, not all conceptual schemes will make sense, comport with basic logic or rationality, or prove at all useful. These words explain my (trivial) constructivist sympathies and natural preference for certain pedagogical methods: Epistemically speaking, the world never, or very rarely, imposes itself on us (some exceptions can be found in the dictates of elementary logic where, for example, the principle of noncontradiction has a naturally existing counterpart in the refusal of instantiated particulars to possess contradictory properties; see Post); rather, we decide what to look for in the world and how to conceptualize it. Given these qualifications, I can unambiguously (if trivially) assert that all knowledge is a fallibilistic, speculative construction of the knower.
6. Every view of the world is a view of someone and from somewhere (#5 above). Nevertheless, from that fact alone nothing of (especially antirealist) interest logically follows about what can or cannot be known. In particular, it does not follow that we cannot know the world as it is independent of our experience. What does follow is the related, second-order and skeptical worry that we might not know in any particular case if we fail or succeed to know the world as it is independent of our experience (#1 above). But this (limited skepticism) is no threat to externalist realism, for two reasons. First, realism supplies the very condition of intelligibility for all varieties of skepticism – if there were no world, there would be nothing to be skeptical about. And second, many of our speculative, referential claims about the world receive experiential confirmation as we infer abductively from successful or unsuccessful action to the external conditions for the truth of these representations. At any rate, these are epistemic worries unrelated to the basic assumption of metaphysical realism (#3 above), that the world exists and has a nature independent of all that we think, say, or do.
7. Constructivism's internalist radical empiricism (alternatively: radical phenomenalism, relativistic neo-Kantianism, postmodern idealism, irrealism, antirealism) supposes that knowledge must consist of the interanimation of these two things: The indubitable contents of immediate experience and the rules of deductive inference. These restrictions on knowledge guarantee that the world remains unknowable or unthinkable, since none of its features would be directly inspectable in this way. Further, it restricts constructivists to the flat, descriptive plane of sensory data (so-called "experiential worlds"), disallowing reference either to extra-mental causes of the differentiations and orders present to experience or to the (uninspectable) properties of knowers capable of organizing and conceptualizing the data received (see also Weissman on "prescriptivist intuitionism"). Despite the well advertised utility of their view, constructivists/pragmatists of a radical stripe have embraced what amounts to, when compared to realism, an explanatorily useless doctrine. Relaxing the logical empiricist demand for verification, realism proposes this alternative, hypothetical method: we infer abductively from the inspectable differences and samenesses to their extra-linguistic, extra-conceptual conditions and causes. The existence and nature of the external world then serves as the conclusion of an (rather global and often unconscious) inference to the best explanation of successful practice (pragmatism). Contrary to radical empiricist scruples, experience is not the object of knowledge, but that activity by which we come to know the world.
8. Analysis suggests that coherence truth collapses in every case to correspondence truth, which is apparently ingredient in the very notion of truth. Combining our best theory of truth with realism we say, in a neo-Aristotelian spirit, that our thoughts are true when and if they correspond to the way things are.
9. Given the arbitrariness and diversity of our conceptualizations, correspondence is rarely picture-like or isomorphic. Rather, our thoughts and sentences signify possible configurations of properties and relations obtaining in the world. When the possibilities signified are actual, the sentences are true. Though correspondence is the apparent meaning of truth, what it means in any particular instance for some element of a conceptual scheme to correspond to the world is determined internally to that conceptual scheme. For example, however I might chose to conceptualize an individual chipmunk, it is the world, not my scheme, that determines the truthfulness of the claim "the world contains more than one chipmunk." That is, from my naturalistic perspective, truth is radically non-epistemic.
10. Furthermore, given the arbitrariness and diversity of our conceptualizations, the world determines but never guarantees our access to truth. We may rarely, if ever, be right about the world (but even when, as traditional skepticism would have it) we are entirely or mostly wrong, we are wrong about the way the world really is. That is, contemporary realism combined with correspondence truth is always fallibilistic and never "naïve."
11. Since I deny the sense of deriving realism from the immediate contents of experience, my view does not attempt to do what is clearly impossible: To know the world extraconceptually (or without the assistance of any concepts or input from experience). Rather, I claim only to infer abductively (from the noticeable effects to its conditions) within a particular conceptual scheme to the extraconceptual and determinate nature of the world (just as a mapmaker makes a map of the world that contains his or her own position; see Weissman). While to know the world without (employing) thoughts is quite nonsensical, to know the world (as it is) without thoughts is commonplace. Appeals to science and commonsense, rather than signaling a naïve faith in the world as guarantor of truth, are used fallibilistically and hypothetically to provide (billions of daily bits of) evidence in favor of a more robust realism that posits the mind-independent existence and nature of the many objects and relations of everyday experience (trees, cats, rocks, other people, etc.). The (abductively) confirmed hypothesis of the existence of the objects and relations of this extra-conceptual world, its successful application to science, practice, and everyday life, its sheer popularity and persistence, all conspire to make this the default position against which alternatives inherit the burden of proof. There is, I suppose, no non-question begging, deductive proof of realism (but see Stove, *The Plato Cult,* for a compelling defense of the view that metaphysical realism is a necessary truth).