Philosophy has confidently resumed debating the nature of ultimate reality, treating metaphysics like just another science and dismissing old positivistic fears that its claims are meaningless. But this confidence is unearned, argues philosopher Nicholas Stang, since, as Kant showed, we lack any good explanation of how language can refer to such an ultimate reality. It’s time metaphysics stopped trying to describe a world that lies somewhere out there, utterly other than and outside of our minds. Instead, a more promising approach lies in the German Idealist idea that mind and reality are in some way co-constitutive.

Metaphysics talks about the ultimate nature of reality—not how we know about the world, nor how we ought to act or live in it, but how the world itself is. Metaphysics aims to peer behind the way things appear in ordinary experience, or even our best scientific theories, and ask: is that how things really are? The world seems to contain a bunch of different, separate entities—rocks and rivers, tables and chairs, bushes and trees, particles and fields—but is this true, or is there in reality only one single thing? Ordinarily we assume that complex material objects, our houses, cars, and coffee tables, are real—but is this true, or are only their simplest parts real? On the other side, is the whole universe of which they are parts the only “real” thing? Physics tells us that the universe is a vast spacetime that is home to physical forces—but is this true, or is spacetime itself founded on something non-physical, like minds? Or a single mind—or God? These questions are questions of metaphysics.

For several decades, analytic philosophers, under the influence of logical positivists like Rudolf Carnap, rejected these questions as meaningless. The positivists held that every meaningful question could be answered either empirically, or else through logical analysis of the meaning of our terms. But the questions of metaphysics do not fit into either category. They are not empirical in any straightforward sense; the latest results from CERN are not going to settle whether reality is ultimately physical, or whether it depends on something mental, or even on something that is neither mental nor physical. But these questions are not conceptual or logical, either. We can define our terms as rigorously as we please; this alone will not settle whether reality is ultimately mental or physical (or neither or both), or whether the way things actually are is the only way they possibly could have been. Logical analysis of concepts will not settle the questions of metaphysics.

Having rejected the logical positivist criterion of meaningfulness (that every meaningful claim is answerable empirically or through logical analysis), analytic philosophers have returned in recent decades to metaphysics, and metaphysics has enjoyed a splendid resurgence. Questions and debates that would have been familiar to any eighteenth-century metaphysician—whether reality is ultimately physical or mental, whether there are many things or only one “big” thing, whether complex material objects are real, whether other “possible worlds” exist, whether everything in reality has a complete explanation—are back on the table,  and pursued with the characteristic rigor and precision of analytic philosophy. We appear to be living in a golden age of metaphysics.

___

The mere fact that the methodology of contemporary metaphysics is modeled on the methodology of natural science should give us no confidence that we are obtaining, or even approaching, knowledge of the truth about ultimate reality.

___

But in rejecting the Carnapian criticism of metaphysics, philosophy ignored an earlier and potentially even more powerful critique of metaphysics (one that influenced the logical positivists themselves) by Immanuel Kant. In the Critique of Pure Reason, Kant posed a deceptively simple question: how is metaphysics possible?

This question has many dimensions. One of these is: how is it possible for us to so much as refer to the ultimate structure of reality? Metaphysics aims, after all, not merely to state truths about reality, but to get at its ultimate structure. A botanist who describes a tree as made up of an upper half and a lower half, or a right and a left half, is not a very good botanist, for the structure of a tree is quite different than this: it has roots, bark, branches, and leaves. Likewise, the metaphysician who just gave you a list of what exists would not be a very good metaphysician; she owes you an account of how the items in the list are structured, both internally and in relation to one another. Are they all independent of one another, or do they all depend upon one another (e.g., as effects depend upon causes)? Do they all depend upon one single, fundamental thing? Do they have parts, or are they simple? To use a metaphor that goes back to Plato, the aim of metaphysics is to “carve” reality at its “joints,” much as a butcher carves a carcass at (or at least near to) its joints.

But this raises a natural question, a version of the Kantian question for contemporary metaphysics: why do our concepts carve ultimate reality at its joints? In our metaphysical theorizing about ultimate reality, why are we not like the botanist who divides the tree into the upper and the lower halves? Or the butcher who swings wildly at the side of beef, cutting it into all kinds of unnatural shapes?

Contemporary philosophers have offered answers to these questions. What I think is the most promising answer can be found in the work of Ted Sider, who draws on earlier work of the late great David Lewis. The core idea of the Lewis/Sider response is to make “joint carving” (getting the structure of reality right) part of what it is to refer to the world at all. On the Lewis/Sider view, of all the things we could be talking about (all of the interpretations of our language that are compatible with what we say), the one we are actually talking about is the one that maximizes structuralness or, to use Lewis’s term, naturalness. For instance, the “good” botanist earlier uses terms like “roots,” “bark,” “branches,” and “leaves.” In doing so, why is he referring to roots, bark, branches and leaves, rather than some unnatural division of the tree into non-structural parts?

Here, Lewis and Sider are responding to a (disturbing) result of philosophical logic: as long as our language is consistent, we can generate arbitrarily many interpretations of it on which it comes out true. This makes reference easy, but it also makes it cheap: there is always an interpretation of what your words are referring to which makes what you are saying come out true. But which of those interpretations is the “correct” one? In virtue of what do we refer to the things we refer to? The Sider/Lewis idea is that the correct interpretation of our words is the one that maximizes structuralness. So to the Kantian question “why do we refer to the structure?” Lewis and Sider answer, in effect: that is part of what reference is. Structure partly determines reference, so there is no possibility that we are systematically carving reality wrongly. We might get it wrong in some circumstances, but it is not possible for us to be like the bad botanist with respect to the ultimate structure of reality all the time.

But this assumes that there is a unique interpretation that is more structural than all the others, and, even worse, it assumes that this interpretation includes maximally structural items at the ultimate level of reality.  We might not be like the “bad botanist” who gets it totally wrong, but we might be like a decent botanist, who gets the surface structure of the tree basically right, but cannot talk about its deep structure: its pith (the spongy heart of the tree), cambium (a thin layer of growing cells), phloem (the liquid that transports its nutrients), etc.  For all Lewis and Sider can show, we might be cut off from even talking about the ultimate structure of reality. It would be impossible for us to even talk about what we intend to talk about in metaphysics.

But even if we grant that we can refer to the ultimate structure of reality, another dimension of the Kantian question arises: how can we know about that structure? Why is metaphysics anything more than idle speculation? Strange though it may sound, the dominant answer in contemporary work is that metaphysics is, broadly speaking, empirical: in metaphysics we infer to the best explanation of the phenomena we are studying. Just as we do not directly observe sub-atomic particles, but infer their existence because they are the best explanation of the data we uncover in large particle accelerators, goes the argument, in metaphysics we cannot directly observe the answer to any interesting metaphysical questions; we infer our metaphysical theories because they are the best explanations of the phenomena.

But I don’t think this explanation is much better than the Lewis/Sider explanation of reference. For what makes an explanation a “good” explanation, and what makes the best explanation “best”? An explanation has to be internally coherent and consistent with the data, but those virtues are easy to come by. For one, a mere list of the data fits that criterion; likewise, for any coherent theory that is consistent with the data we can easily generate a spurious alternative theory that has those features too. In natural science we infer the theory that is most unified or most elegant, and so too, the metaphysicians say, in metaphysics we should infer the theory that maximizes simplicity (it posits fewer basic kinds of things, and has fewer underived axioms) with explanatory power (it explains more than its rivals). Hence what is known as Occam’s razor: do not multiply entities without necessity.

But to apply this methodology in metaphysics, the study of ultimate reality, is to assume that ultimate reality is more likely to be simple than complex, more likely to have fewer basic kinds of things in it than more. To put it bluntly, what entitles us to assume that ultimate reality itself does not multiply entities without necessity? Perhaps ultimate reality is deeply uneconomical. Admittedly, inference to the best explanation, guided by maxims like Occam’s razor, has been very successful in the natural sciences. We have reason to think that this works in the natural sciences because they progress over time, different scientists converge (given enough time) in their theories (even if some pockets of disagreement persist), and they make empirical predictions that turn out to be true. We should have no such confidence that inference to the best explanation works in metaphysics, because the history of metaphysics is not a history of convergence; the questions that are debated today have been debated for centuries (in some cases millennia), and agreement seems as far off as ever. Metaphysics does not make surprising empirical predictions that can then be tested and verified. A typical metaphysical theory might “predict” that tables exist (duh) or that although your table exists, it might not have existed (again, not very surprising). As Kant puts it, “metaphysics has not yet entered upon the secure path of a science.” So the mere fact that the methodology of contemporary metaphysics is modeled on the methodology of natural science should give us no confidence that we are obtaining, or even approaching, knowledge of the truth about ultimate reality in this central part of philosophy.

SUGGESTED VIEWING The return of metaphysics With Hilary Lawson, Galen Strawson, Patricia Churchland, Maria Balaska

Far more promising than the Lewis/Sider attempt to save metaphysics as the study of ultimate reality, I think, is Kant’s approach to metaphysics: we give up on any claim to be studying ultimate reality, and study instead reality insofar as it can appear to us in experience and natural science. Metaphysics, in this transformed sense, remains an armchair discipline because, Kant argues, we can study the structure of any possible experience and the structure of any possible natural science based on that experience (natural science, after all, is always answerable to experience, that is, experiment and observation). But we do not give up on claims about ultimate reality altogether: for in a different philosophical discipline, ethics, we have a distinctive yet no less certain way of grounding claims about our ultimate nature—we are free rational beings capable of determining ourselves to act according to our moral duty, or contrary to our moral duty. These claims do not compete with any results of the natural sciences, and so cannot be contradicted or supported by them, because they are about ultimate reality that lies behind the spatiotemporal reality disclosed in natural science and ordinary experience.

There are, however, reasons to be skeptical about Kant’s solution to the problem of metaphysics. Kant’s own claims about the formal structure of any possible experience and thus any possible natural science appear to be simply contradicted by later discoveries: the non-Euclidean structure of physical spacetime and non-deterministic physical laws in quantum mechanics, to name only two of the most pressing. What is more, Kant’s claim that our awareness of ourselves as bound by an unconditional moral law discloses the ultimate nature of reality more certainly than any possible natural science will strike many as special pleading on behalf of an implausibly strong and culturally idiosyncratic (essentially, pietist Lutheran) conception of morality.

Although I think these problems can be overcome, I agree that Kant’s conception of metaphysics faces grave challenges, in particular those concerning the “interface” between ultimate reality and our experience. For ultimate reality appears to us in some way in experience, even if we cannot infer back from the content of that experience (and theories we construct on its basis) to the structure of that ultimate reality. After all, the structure of our minds alone does not determine the specific ways that reality shows up to us in science and experience; that is determined, in some way we cannot know, by ultimate reality. It seems that Kant needs our concepts to be able to refer to that interface between ultimate reality and experience in order to so much as state his theory, but also cannot explain how that reference is possible. For that interface is not purely within our experience: it concerns how that experience is determined by, or grounded in, ultimate reality.

But I do not think that these problems with Kant’s answer to his own questions should lead us back to Lewis and Sider, or to Kant’s early-modern metaphysical predecessors like Spinoza or Leibniz. They should lead us forward to post-Kantian German idealists like Hegel, who take the Kantian question of the possibility of metaphysics seriously and would reject answers like those of Lewis and Sider—but who also attempt to provide a positive explanation of the possibility of metaphysics that avoids the Kantian problem of the reality/experience interface. But that is a story for a future occasion.