Responding to Craig and Hunt (Part 3 – the SFV/OFV distinction)

This is the third installment in a series of posts in which I respond to a 2013 paper entitled “Perils of the Open Road” authored by William Lane Craig and David Hunt (hereafter, C&H). In the paper C&H critique two papers defending open theism: a 2006 paper (hereafter, RBB) that I co-wrote with Greg Boyd and Tom Belt entitled “Open Theism, Omniscience, and the Nature of the Future” and a 2007 paper (hereafter, DT) by Dale Tuggy entitled “Three Roads to Open Theism.”

In part 1 of this series I laid some groundwork by defining key terms like “open theism” and “future contingent” and distinguishing between different types of future contingent proposition (FCP). In part 2 I examined C&H’s dismissive discussion of arguments for the incompatibility of future contingency with an epistemically settled future.

In this post I respond to C&H’s critical discussion of the use of branching diagrams by RBB and DT to depict an indeterministic future. I will argue that their criticisms are extremely uncharitable and very largely miss the point. It seems to me that C&H are much more interested in scoring rhetorical points by making cheap shots than engaging in a fair and irenic discussion of the issues that divide us, contrary to Wolterstorff’s dialogical imperative.

Summary of C&H’s criticisms

On p. 54 C&H say of RBB that our discussion of the future is “so confused that it’s difficult even to know where to begin in one’s commentary.” After several pages of fisking, C&H summarize their criticisms of RBB’s description of how future possibilities can change over time::

Here we see their [RBB’s] confusion of future events with future tense statements about those events, their conflation of two distinct interpretations of tree diagrams of temporal events, one ontic and the other modal/causal, their incoherent blend of tenseless and tensed representations of reality, their idiosyncratic understanding of “might” and “will,” and their confusion of causal closedness and causal determinacy. (p. 59)

That’s five separate confusions they claim to discern! C&H also direct several related criticisms at DT, including the charge of begging the question against Ockhamism (p. 62). In what follows I will reply to C&H’s criticisms against RBB more or less in the order in which they occur. But, first, it will help to set the context by giving a brief overview of the relevant portions of RBB.

RBB – SFV versus OFV

At the beginning of RBB, we spend about five pages introducing and clarifying a distinction between the settled future view (SFV) and the open future view (OFV). We characterize these positions by saying that, according to the SFV, “the future can be completely and truly described in terms of what either will or will not happen” whereas, according to the OFV, “the future cannot be completely and truly described in terms of what either will or will not happen, but must also include reference to what might and might not happen” (RBB, p. 432). Now, these aren’t entirely perspicuous characterizations. If I were to articulate the distinction today I would do so somewhat differently. But the basic idea can be fairly readily gathered from the context. It is clear from RBB’s abstract that the SFV/OFV distinction is to be understood in terms of open theist (OT) and non-open theist (NOT) views of God’s foreknowledge. The purpose of the SFV/OFV distinction is to contrast how the future looks to God given OT with how it looks to God given NOT. (I would add that by “non-open theists” I’m thinking only of non-open broadly classical theists. Process theists would agree with open theists on the matters under discussion here, but I don’t include them within the camp of open theists.)

From the NOT (i.e. SFV) perspective there either eternally is or always has been a unique, complete, linear, and true “story of the future” that God knows to be (or to be going to be) the future. This “story of the future” is one that, in its entirely, will happen. Moreover, God infallibly knows this. Hence, the future, according to the SFV, is epistemically settled for God: “The entire course of history from creation to kingdom come is and always has been settled from God’s epistemic perspective” (RBB, p. 433). And since knowledge entails truth, if the future is epistemically settled then it must also be, as we put it in RBB, “semantically” settled (ibid.). “Semantic”, of course, is the wrong word here. Semantics has to do with meaning, whereas our point is concerned with truth. So we should have said “alethic” rather than “semantic”. Nevertheless, one confusing term aside, the basic point is pretty clear and is reflected in the diagram on p. 434: 

The idea behind the SFV is that there is a complete, true, non-branching story of the future (i.e., of what we temporally situated beings think of as the future). While there may be other ways the future might unfold–other causally possible futures–there is one specific way that it will unfold. Any possible future events that are not included in the complete true story are events that will not happen. This is why we characterized the SFV as the view that “the future can be completely and truly described in terms of what either will or will not happen.”

In contrast, the OFV or “open future view” denies that there is a unique, complete, non-branching story of the future.  As depicted in the diagram (RBB, p. 436), according to the OFV there are multiple possible futures, no one of which is either such that it will happen or that it will not happen. There may be particular future events, like the sun’s coming up tomorrow, that will happen because they occur in all possible futures. Likewise, there may be particular future events, like the temperature’s in Indianapolis reaching -100°F tomorrow, that will not happen because they occur in no possible futures. But there is no complete post-present sequence of events, no complete future, that will happen. This is why we characterized the OFV by saying that “the future cannot be completely and truly described in terms of what either will or will not happen, but must also include reference to what might and might not happen.” For the SFV each causally possible future event either will or will not happen. For the OFV, in contrast, some causally possible future events are such that they neither will nor will not happen, although they might and might not happen.

All of this is pretty straightforward. If some version of NOT (non-open theism) is true, then something like the SFV must be correct, whereas if OT (open theism) is true, and if God is unqualifiedly omniscient, as we maintain, then something like the OFV must be true.

C&H: the SFV/OFV distinction is poorly drawn

C&H claim to be “puzzled” by our respective characterizations of the SFV and the OFV because “these two ways of describing the future do not seem to be mutually exclusive” (p. 55). Why they say this is not entirely clear, but their argument seems to be that our characterization of the OFV could apply just as well to the SFV when the latter is properly understood to include moral and modal truths (p. 56). The future, for example, presumably includes moral truths to the effect that this or that ought or ought not to happen. And it presumably includes modal truths about what could or might and might not happen. If this is right, and I concede that it is, then it is not true according to the SFV strictly and properly understood that the future is “exhaustively” or “completely” describable in terms of what either will or will not happen. Moveover, for non-open freewill theists it is also true that a “complete” description of “the future” must include reference to what might and might not happen. Hence, C&H conclude, our distinction between the SFV and the OFV collapses.

In response I would say that this criticism, while technically correct, completely misses the larger point. The status of moral and modal truths is completely tangential to the fundamental purpose of the SFV/OFV distinction, which is to contrast open theistic and non-open theistic views about the future. No one on either side of the OT/NOT debate would dispute that moral and modal truths are important or that a “full” characterization of “the future” would have to include both. The dispute centers on how the future appears epistemically to God and more specifically on whether there is a unique possible future that will happen, as the diagrams above depict. If there is a unique possible future that will happen, then either God knows that it will happen (which entails non-open theism), or God is not unqualifiedly omniscient. But if there is no such future and there are truths about what might and might not happen, then God’s being unqualifiedly omniscient entails open theism. That’s the point of the SFV/OFV distinction.

C&H: the future cannot “change” (except trivially)

In the course of explaining the differences between the OFV and the SFV we note in RBB that there is a sense in which the future can and does change on the OFV but doesn’t change on the SFV. We argue (RBB, p. 436) that it changes by attrition of possibilities. As future contingencies are resolved one way or the other, branches representing “unchosen” possible futures fall out of the realm of causal possibility.

While there is nothing particularly complex about this idea, C&H find it “unclear” (p. 55). First, they say that “it is a familiar analytic truth that the future cannot be changed” because “to change the future would be to bring it about that an event which will occur will not occur, which is logically absurd” (p. 55). This charge, however, is question-begging. It is only analytic that future cannot be changed if one first defines “the future” as “what will occur”. This is how SFV proponents think of “the future”. But the central argument of RBB is that there is an alternative way to think about the future. On the OFV “the future” is not equivalent to “what will occur” but rather to “what either will occur or might-and-might-not occur”.

C&H consider another sense in which the future might be thought to “change” on the OFV:

The idea, we take it, is that on the “open” view, since the only future-tense statements that are true are those describing events which are causally determined to occur, then as causally determinative conditions for some event become actual, statements about that event switch truth values, in Rhoda et al.’s view, from false to true. … [W]hat changes is [sic] the truth values presently possessed by future-tense statements. (p. 56)

This is, in fact, a passable description of what occurs on the OFV—provided that we do not restrict “future-tense statements” to will and will not statements as C&H do. For will and will not statements are not the only kinds of future-tense statements. There are also will probably statements and might and might not statements. These are future-oriented just as much as the will and will not variety.

But C&H charge that this is only a change in a “Pickwickian” sense (i.e., that it is not meaningful change in any usual sense). First, they note that we talk of the “set” of truths about the future changing whereas, strictly speaking, sets have their membership fixed and thus cannot change. Rather, “at different times different sets of future-tense statements are true” (p. 56), something that is also true on the SFV given a tensed (or A-theory) view of time. Second, C&H charge that change in the future on the OFV “consists wholly in [change in] the truth-values presently possessed by future-tense statements” (p. 56) which is also the case on the SFV, the only difference lying “in which statements are presently true” (ibid.).

This last charge, however, is incorrect. There is an important difference in what sorts of changes are possible on the two views. Setting aside changes due merely to the passage of time, on the OFV a true might and might not statement becomes false as its corresponding will or will not statement becomes true after having been either false (according to RBB) or neither true nor false (according to DT). But on the SFV a might and might not statement’s becoming false is never accompanied by its corresponding will or will not statement’s becoming true. This is why, on the SFV, there is no change in “the actual future” (apart from the trivial and Pickwickian “change” that results simply from the passage of time), whereas on the OFV there is a change in “the actual future” because that phrase now refers to a branching array of possibilities, which, as already noted, changes by branch attrition.

Now, I no longer think this point about the “changeability” of the future is as important as I did when writing the RBB paper. Indeed, it may do little to rise above the level of trivial or “Pickwickian”, as C&H allege. Certainly even for non-open freewill theists like C&H there is a sense in which the branching array of possibility changes by branch attrition. The salient point is not really how “the actual future” changes but whether there is a unique possible future for “the actual future” to refer to.

C&H: equivocal interpretations of branching diagrams

C&H next charge that RBB “conflate two quite different interpretations of such diagrams: as representations of the ontology of space-time events and as representations of the causal modality of such events” (pp. 56-57, emphasis in original).

Referring to the above diagram depicting the SFV, C&H state that “confusion is evident”:

If the diagram exhibits real events, then the diagram is a futile attempt to represent at once both a tenseless and a tensed perspective on the world. If the diagram is a tenseless representation of temporal moments or events, then there should be no present moment designated and so no branches. If the diagram is a tensed representation, then there should be no solid line representing the sequence of future events, since on presentism there are no such events. (p. 57)

This charge bewilders me. After all, the whole point of the diagrams was to contrast the SFV and OFV with respect to the “semantic” or, more accurately, alethic settledness of the future. They were never intended to depict a temporal ontology, and where C&H get this idea from is not clear to me. We do note later in the RBB paper that ontology significantly bears upon the alethic question (and who disputes this?), but ontology isn’t even primarily in view when we introduce and describe the diagram in question (cf. RBB, p. 433). When we describe it we do so exclusively in terms of what is now true about the future.

Turning to the OFV diagram, C&H charge that

the absence in the diagram of a solid line later than the present in no way illustrates either the failure of Bivalence for future contingent propositions or the falsity of such propositions, for the diagram does not depict the truth status of propositions at a time, but the ontology of events in time (p. 57).

But, again, this is wholly false. The diagram was never intended to depict “the ontology of events in time” but to depict what is now true about the future according to the OFV (cf. RBB, p. 436). It seems to me that C&H are being very uncharitable here in their criticisms. If only they had bothered to contact either myself, or my coauthors, or Dale Tuggy (author of the DT paper), we could have easily set them straight about how to read these diagrams.

At any rate, C&H continue their criticisms:

Rhoda et al. then begin to construe the diagram as a representation of the causal modal status of events. If an event occurs on all branches, it is causally determined; if it occurs on none, it is causally impossible; if it occurs on some but not all, it is causally contingent. But if the diagram represents the modal status of events, why is there a single line from the past through the present? If there have been free decisions in the past or if quantum indeterminacy is ontic, then the past should be branching, too. The diagram should be a thicket, not a tree. (p. 58)

Again, the charge is seriously confused, indeed, I’m strongly tempted to think that C&H are being willfully obtuse here. First, we don’t after introducing the diagrams then “begin” to construe them in causal/modal terms after having construed them ontologically. Rather, the causal/modal interpretation is there right when the diagrams are introduced. (I invite the reader to read the paragraphs in RBB that immediately precede each of the diagrams to confirm that this is so.) Second, as to why we don’t depict a branching past, the answer should be obvious. Our concern in the paper was with the alethic status of the future. That there has been a unique actual past has never been up for question in the OT/NOT debate. For C&H to raise this as a problem for us, much less as a point of confusion on our part, is just bonkers.

C&H: Tuggy’s use of branching diagrams begs the question

Turning from RBB to DT, C&H charge that Tuggy, who also uses a branching diagram to represent the openness of the future, similarly confuses ontology and modality (p. 59), despite the fact that in a quote from DT that C&H provide (pp. 59-60) Tuggy explicitly rejects an ontological interpretation of the diagram! If this isn’t a clear case of uncharitable reading on the part of C&H I don’t know what is.

C&H further charge Tuggy with begging the question against Ockhamism: “by depicting the past as a single line segment rather than as branching, it precludes there being any accidentally contingent propositions about the past” (p. 61). This criticism is correct, but it’s really not a big deal because it’s easily fixable and doesn’t undermine Tuggy’s central project, which is to articulate and defend a non-bivalentist version of open theism. To fix it all one has to do is switch from asking what futures are possible given the entire actual past to asking what futures are possible given the explanatorily relevant actual past.

In an amusing case of irony, C&H immediately go on to beg a question of their own when they claim that “all libertarians think that some past events are ‘soft’ facts and that therefore propositions about their occurrence are not at every later time accidentally necessary” (p. 61). Since the authors of RBB and DT are clear counterexamples to this claim, it is simply false that all libertarians think that some past events are soft facts. To assume, in the dialectical context of their paper, that libertarians as such must believe in soft facts is as clear a case of question-begging as one could wish for.

Closing thoughts

I thank C&H for pointing out some infelicities in our presentation of the SFV/OFV distinction. In particular, we don’t clearly distinguish our characterizations of the SFV/OFV from definitions, we make too much of the changeability of the future on the OFV, and we use the term “set” inappropriately. In addition, though C&H don’t specifically call attention to this, instead of our talking of the future’s being “semantically” open or settled we should have used the term “alethic”.

Understood as definitions our characterizations are inadequate because they leave out, for example, moral and modal truths. But when understood more loosely in terms of the purpose for which the distinction was introduced C&H’s criticisms of the distinction largely miss the point. And they completely miss the point when proposing that we offer an ontological interpretation of the SFV and OFV diagrams.