What is Metaphysical Possibility?

Part 3/3 of Modal Rationalism (From the Archives)

[The final chapter of my undergraduate thesis on Modal Rationalism. The initial sections contrast realist vs conceptualist understandings of “metaphysical possibility”, with an eye to helping skeptical readers to grasp the core concept (e.g. via a “rewind and replay” heuristic to spin out the history of once-open futures). The closing section (paywalled) proposes two candidate principles for determining the scope of metaphysical possibility: what really could have happened.]

§3.1 Metaphysical Realism and Conceptualism

Modal rationalism links metaphysical necessity to a priori knowability. We may wonder what this implies about the metaphysical status of modal discourse: can it still be fully mind-independent? The modal rationalist grants that many modal facts will never be actually known—and perhaps even cannot be known by creatures with our cognitive limitations. So the modal facts are genuinely objective, in that they are completely independent of our minds, and may transcend at least the evidence that is practically available to us. Nevertheless, modal rationalists hold that the sum of all possible rational evidence, including that which is accessible only to more cognitively advanced agents, suffices to settle the modal facts. At the end of the day, all (semantically neutral) necessary truths must be knowable on ideal rational reflection. There are no such necessary truths besides those that are so knowable. Modal reality cannot transcend all possible rational evidence.

Why not? Here modal rationalists may split into two camps. Conceptual modal rationalists, e.g. Chalmers (1999), seek to epistemicize modality by claiming that so-called “metaphysical necessity” is really nothing over and above a priori knowability.¹ On this view, there is an analytic link—perhaps identity—between the concepts of possibility and ideal conceivability that precludes any gap between the two. That way, modal truth just is the ideal limit of a priori inquiry; it does not answer to the sort of independent reality that might sensibly be considered beyond all epistemic reach.² Modal reality is thus a kind of (non-contingent) rational construction. Rather than addressing the metaphysical question of how reality is in itself, modal facts may be considered more fundamentally normative in nature: they tell us what should be concluded at the ideal limit of a priori rational reflection.

Metaphysical Realists about modality, in contrast, wish to uphold the conceptual distinction between necessity and apriority, where the former is taken to be a genuinely metaphysical notion—about the world as it is in itself, rather than our (even idealized) beliefs about it. Realists will be more sympathetic to the previous chapter’s conclusion (§2.3) that we can make sense of a ‘gap’ between the two concepts, according to which even ideally rational agents might be inescapably mistaken about the breadth of modal space. Realist modal rationalists must simply insist that there is no such gap as a matter of fact. Further, this fact about modal space will be necessary if true at all (cf. §2.2), so the modal rationalist is also committed to its apriority. But whereas the conceptualist takes the connection to be analytic, realists will instead propose that it is a substantive, synthetic claim about the metaphysical nature of reality. It is this ‘realistic’ version of modal rationalism that I will seek to elucidate and defend later in this chapter. First, we must depart from the conceptualists by taking seriously the idea of primitive metaphysical modality that underlies the radical challenge from strong necessities.

Note that metaphysical realism puts modal rationalism at risk by opening the door to strong necessities. If—contrary to my arguments—there are indeed some strong necessities, then the inference from ideal conceivability to metaphysical possibility is jeopardized. Even so, this might not leave philosophers quite as hamstrung as typically supposed. It is arguably the fundamentally rational notion, rather than the metaphysical one, that we employ in our philosophical theorizing. That is, for the standard theoretical uses of modality, it may be the conceptualist’s space of coherent scenarios that we really need.³ But I think we have a grip on an independent metaphysical notion in any case, so I will try to bring this out in the sections that follow. My subsequent defense of realist modal rationalism will be of greater significance to those who dispute the theoretical primacy of epistemic space proposed above.

§3.2 Content-Based Modalities vs. Metaphysical Modality

I wish to distinguish two very different ways of specifying a modal space. In the first case, philosophers may isolate and identify the particular modal space they wish to work with by offering a (more or less) formal specification of the contents they wish to include or exclude. That is, they begin with some framework F of rules or limitations, and then define the space of F-possibilities as simply a matter of what is not ruled out by F.⁴ I will say that such a specification is “content-based”, to highlight the point that its delimiting rules are directly and exclusively concerned with the internal contents of possible worlds, so that one may determine whether or not to allow a world-candidate solely on the basis of descriptions of what that world contains. For example, nomological possibility is sometimes understood as simply consisting in the non-violation of the actual laws of nature.⁵ This specification is content-based insofar as it can be applied simply by examining a complete description of the internal workings of a candidate world, and determining whether any of the described events contravene our laws of nature. A claim is nomologically “possible”, in this sense, so long as the laws of nature don’t rule it out. Conceptual possibility can similarly be settled simply by determining whether a candidate world-description contains any overt or implicit self-contradictions. This might not be purely formal: if rational insight cannot be captured algorithmically, there will be no finite set of rules that can determine a priori coherence. So the latter may need to be taken as a primitive in its own right. Nevertheless, this modality is “content-based” as I use the term, for it serves to directly fix the breadth of the modal space, by means of a particular method.

In this chapter, I wish to explore the proposal that metaphysical possibility is not to be understood in such content-based terms. Its breadth is not defined in terms of epistemic procedures. Instead, this modal space is, in a sense, “world-oriented”. It is to be characterized first and foremost in terms of its metaphysical nature, thus leaving its breadth of content to be fixed by reality rather than built explicitly into the concept. This aims to connect with our intuitive notion of ‘metaphysical possibility’ as reflecting ways the world really could have been—a concept for which our primitive grasp leaves open, at least initially, what breadth of content this modal space contains. The answer is fixed by reality, not our concepts alone. The question of what really could have been is here assumed to be a question fundamentally about the world—or reality in itself—that admits of an objective and exclusive answer. Though philosophers might propose whatever content-based restrictions suit their purposes, the world itself provides just one space of real possibilities.

Of course, this space of metaphysically possible worlds must have some or other breadth, and so be specifiable in terms of restrictions on content. Perhaps it includes all the conceptually possible worlds. Or perhaps it includes only the actual world, as would be the case if things never really had any genuine opportunity to be different. If all else fails, one could simply give an exhaustive specification of each and every world that it includes, and construct their disjunction as the content restriction. Any of these “spatial breadths” are prima facie consistent with the concept I’m trying to point to, because its fundamental character lies in a different dimension. Unlike the other modal concepts, we don’t immediately characterize it in terms of breadth or restrictions on content. The criterion for a world’s inclusion in this space is instead its brute modal nature. We don’t ask: “Does this world contain anything which violates such-and-such content restrictions?” Instead we ask the irreducibly modal question: “Is this a world that had the opportunity to be actualized?” Or, equivalently: “could it really have come about?”

Once we have a grasped metaphysical modal space by way of the above questions, we can go on to inquire into the space’s breadth of content—as below. But for now I emphasize that the concept must be initially grasped in these primitive modal terms. You cannot begin by characterizing metaphysical modal space in terms of its contents, because those are not included in the concept as it initially presents itself to us. If you begin with them, you are really grasping a different concept altogether. After all, for any space of worlds characterized in terms of their content, one can still coherently ask: “but might they really have been actualized?” It wrongly remains an open question, unless one builds this modal requirement right into the fundamental conceptual character of metaphysical modality.

Note that this conception makes no explicit demands on what content must be found within candidate worlds. Metaphysical modality, thus understood, is not to be analyzed in terms of any collection of formal rules or laws that must be satisfied, nor even a primitive content restriction such as “rational coherence”. Rather, what matters is simply whether the candidate world is one that really could have been actualized. So long as this external modal property is satisfied, we need not worry about what is in the candidate world. This makes the specification of metaphysical modal space significantly different in kind from the content-based spaces mentioned earlier.

§3.3 Grasping Metaphysical Modality

It might be wondered what, exactly, this notion of “really could have been actualized” involves. (Merely emphasizing the “really” will do little to help one who lacks an antecedent grasp of the concept.) Since it is presented as a primitive or bedrock concept, no reductive analysis can be offered to explicate it. But some general remarks may help bring the intuitive notion to light.

Chalmers expresses his skepticism as follows:

It seems to me that we do not even have a distinct concept of metaphysical necessity to which the second primitive [besides rational coherence] can answer. The momentary impression of such a concept may be a residue from initial impressions of the Kripkean distinction between epistemic and metaphysical modality. But once we recognize that this distinction can be explained with one modal primitive, and that there are constitutive ties between the Kripkean modalities, the grounds for this impression disappears. The only concept of a “metaphysically possible world” that we have is that of a logically possible world. If someone thinks they have a distinct concept here, there is no reason to believe that anything answers to it.⁶

I think both challenges may be answered by pointing to a familiar—and metaphysically ‘realist’—concept of possibility that is sometimes relegated to the merely “nomological”. The concept I have in mind naturally relates to commonplace ideas about objective chance, indeterminism, and the open future. Many people think that the future is metaphysically open, in that it really could turn out in any one of a number of different ways—the truth of the matter hasn’t been decided yet. Each open possibility has some non-zero objective chance of eventuating. Note that this isn’t just a claim about our epistemic situation, or even the rational ideal: it’s about how reality is in itself. This has nothing to do with any actual or possible minds. As a rough heuristic: if God were to rewind time and play it back again, things would unfold differently. Admittedly, this commonsense belief assumes indeterminism. If, instead, the future is already determined, then—given the present state of affairs—there is only one way that things can really turn out. No matter how many times God “replays” history from this point, he’ll never get a different result.

Of course, metaphysical possibility cannot simply be identified with non-zero objective chance. The past is presumably now fixed, so there’s no chance it will suddenly change on us. But even though the past certainly won’t be different, nevertheless we might still think that it could have differed.⁷ Perhaps there were open alternatives at a time in the even more distant past. In extending our intuitive notion of the open future back into the past, we will find various (now closed) branches that really were, at one point, dynamically open possibilities. Our concept of metaphysical possibility should at least include the entire history of such dynamic possibilities.⁸ They are all ways that the world really could have turned out. Hopefully it is now intuitively clear what I mean by this.

It’s worth pausing here for a moment to clarify what we have established. In exploring the metaphysical specification of modality, we have thus far reached a space of worlds that could be given the content-based specification of “nomological possibilities given the initial starting conditions of the universe”. But such content-based descriptions fail to capture the metaphysical significance—the idea that the world really could have turned out in any of those ways. (Though readers might implicitly project significance on to it, in light of their background knowledge that anything satisfying this content restriction would in fact have had an objective chance of eventuating.) I wish to draw attention to this modal primitive—the one we invoke when thinking about objective chances, physical indeterminism, and open futures—and how natural it is for us to be Metaphysical Realists about it. This seems to be a species of modality that is no mere rational idealization, but rather is truly in the world, as a basic component of reality.

It might be objected that what I’ve pointed to here just is “nomological modality” in some sense. Non-Humeans, at least, could be expected to imbue some form of natural necessity with the primitive metaphysical significance proposed above—in which case the end result will be much the same.⁹ Whatever you want to call it, once we have a grip on this ‘realist’ modal primitive, we may ask: could the laws of nature themselves have been different, in this primitive sense? It seems a reasonable question, though of course the nomological impossibility of it is trivial. This suggests that the primitive notion in this vicinity is not, strictly speaking, nomological possibility after all—at least, it doesn’t seem built into the modal concept that the laws of nature must be necessary in this primitive sense.¹⁰ We can always stipulate such a content-based restriction later, if we need it for other purposes. But the core concept here is—prima facie—potentially broader than that. So I think ‘metaphysical possibility’ is the more fitting term.

We have here a metaphysically ‘realist’ modal concept that has worldly application: at the very least, it spans the entire history of dynamic possibilities or “open futures”. This undermines the skeptical basis for conceptualism—we should be realists about metaphysical possibility instead. But how far does it extend? We are now faced with the awesome questions of why our universe exists at all, and whether a wholly different universe—say with alternative laws of nature or initial conditions—could have existed in its place. This provides the focus for the next section. Could absolutely any coherent scenario really have been actualized, as the modal rationalist proposes? Or are some rationally apparent possibilities necessarily excluded by the nature of reality, creating “strong necessities” that fall outside the 2-D framework and falsify modal rationalism?