(This is part two of a series that began here. In this series I am interacting with Tom McCall’s section on theological determinism and compatibilism in his book, An Introduction to Analytic Christian Theology (IVP 2015).)
Key: I will let FW&MR stand for Free Will and Moral Responsibility. I will let TD stand for the Theological Determinism, where God is said to, in some sense, determine all human actions.
In my last post, I made the point that the standard approach to dealing with the claim that FW&MR is compatible with TD is to cite some standard arguments against compatibilism. As I understood it, the standard arguments refer to those arguments that employ the standard definition of determinism as a premise. As I understood it, and as I demonstrated from the definitions McCall enlisted, the standard definition of determinism includes, as an essential component, among other things, the claim that the determining conditions lie in the “remote past.” Call this determinism, D+P. I then argued that many theological determinists—Calvinists, say—would not count as determinists at all on this understanding, since God’s determinative decree is timeless. However, once we allow this feature of classical TD to enter into the picture, the standard arguments are otiose for the purpose of showing that FW&MR is incompatible with TD. One upshot here is that those who argue against Calvinists (of the sort envisioned above) need to restate their argument for it to be relevant.
In the light of this, one might get the impression that I think the standard arguments for the incompatibility of FW&MR with D+P are good (sound and persuasive), while allowing TD to escape out the back door. However, this is false. Even here, I don’t think incompatibilists have made their case. So in this post and the next, I would like to comment on some of the arguments that McCall employed for the incompatibility of FW&MR with D+P.
Before doing so, however, I feel that I need to reiterate a couple of points about McCall’s project, as he sees it. The main point of the book is to illustrate how “analytic theology” can interface with “biblical theology.” In our case, he’s trying to show how work in analytic philosophy can aid and complement biblical theological discussions on sovereignty, providence, free will and moral responsibility. He uses Carson’s writings as a case study. Thus, he does not think that he has offered any sort of knock-down argument for incompatibilism. With this in mind, I would like readers to see these posts as using what McCall wrote as a launching pad for further illustrating how the tools of analytic philosophy (and theology) can aid in these sorts of matters. However, I do not hide the fact that I will be defending theological determinism and compatibilism, much as McCall’s section was skewed more toward defending incompatibilism.
In this post, I will look at McCall’s evaluation of “classical compatibilism.” First, I shall state the standard arguments McCall provides for incompatibilism. Next, I briefly look at what I call “omitted” objections to standard arguments, pointing to some of the most recent work on the standard arguments. Then, I shall look at his evaluation of the first compatibilist strategy for responding to the standard arguments. In the next post, which will be much shorter than this one because I will forgo the lengthy setup, I will evaluate his evaluation of the semi-compatibilist strategy.
McCall and the Standard Arguments
McCall gives two arguments for incompatibilism. The first he calls “the core argument”, and it goes like this:
(1) Free will requires the ability to do otherwise.
(2) If causal determinism is true, then no agent has the ability to do otherwise.
(3) Therefore, free will requires the falsity of causal determinism.
We might wonder what reason there is for thinking that (2) is true. McCall presents another argument—one he got from Keith Yandell—that can be seen as supporting (2), though this is unclear as the terms employed in the second argument wouldn’t establish the consequent of (2). However, the general idea is clear enough, as the second argument could be reformatted to get the consequent of (2). Here’s the second argument:
Let TUDdino = a tensed universal description (a complete statement of every truth at a time) just before the last dinosaur died. Let TUDcup refer to the tensed universal description which includes the proposition that you will decide to have cup of coffee later this afternoon. Let L stand for all the laws (of nature and logic). Then,
(4) TUDdino & L entails TUDcup;
(5) one is not responsible for anything that one has no control over;
(6) one has no control over anything that is entailed by what one has no control over;
(7) one is not responsible for anything that is entailed by what one has no control over (from 5&6);
(8) you have no control over what is true in TUDdino;
(9) you have no control over L;
(10) you have no control over TUDdino & L (from 8-10);
(11) you have no control over what TUDdino & L entails (from 7, 10);
(12) you have no control over TUDcup (from 4, 11);
(13) TUDcup entails that you decided to have a coffee later at 3 o’ clock;
(14) you have no control over whether you decide to have a coffee at 3 o’ click (from 12, 13).
The Standard Arguments and Omitted Objections
For my part, I wish McCall hadn’t used Yandell’s argument since it employs the term ‘control’. One reason is that McCall doesn’t discuss the concept(s) of control after he presents this argument. Rather, he turns to analyses of “can” and “could do otherwise.” He notes these terms are multiply ambiguous, but so is “control,” and perhaps more so! However, I won’t belabor this point, as a discussion of control would take us too far off course (but see Haji’s Incompatibilism’s Allure (pp. 38ff.) for a discussion of the complexities involved in ‘control’, as well as Fischer and Ravizza’s book, Responsibility and Control (1999).)
I will make two further points before engaging with what McCall says explicitly. First, both of the above arguments rely on some unstated rules of inference that McCall omits from the discussion, probably because such nuances are beyond the scope of an introductory text intended for nonspecialists. The first rule is called Rule α, it states no one has a choice about (or control over) necessary truths. Most philosophers have thought this rule unassailable, but recent work has cast doubt on this (see Kearns’ “Responsibility for Necessities” (Philosophical Studies 2011, 155(2)). Peter van Inwagen says that if Rule α is false, then someone could have a choice about necessary truths—and at best, only God could have that choice. But see Jack Spencer’s paper “Able to do the Impossible” (forthcoming in Mind) for counterexamples. There’s also a gaggle of theological counterexamples to Rule α, it seems to me, but I shall forgo discussion of these here. The other rule is Rule B, which is a transfer principle that, in the case of argument 2, transfers ‘no responsibility’ over the proposition stated in the antecedent to the proposition stated in the consequent. So it transfers nonresponsibility (a version of Rule ß seems to be involved as well, but this isn’t clear and I’m not going to address it). McCall uses Rules α and B together to derive some of his (sub) conclusion(s). I’ll address Rule B in a future post, as I believe it has counterexamples.
The second point is that a recent objection to consequence-style arguments by Joe Campbell threatens to render all of them unable to secure incompatibilism as a conclusion. I think any discussion of the consequence argument must take Campbell’s objection into account. Here’s why: if the standard approach to theological determinism and compatibilism use the standard arguments for incompatibilism, and those arguments simply cannot secure incompatibilism, then the standard approach doesn’t do what the typical proponents think it does—bracketing that I’ve shown that they don’t do this in the light of God’s timelessness. Let’s see why Campbell’s argument is relevant. Let “incompatibilism” name the thesis that it is impossible for (nomological) determinism to be true at a world and someone do a free action at that world. This is the conclusion of the consequence argument. Since the conclusion is supposed to be a necessary truth (if true at all), the premises must be necessary truths (or derived from necessary truths). Campbell’s objection (first stated in his paper “Incompatibilism and the Necessity of the Past” (2007) points out that the existence of a remote past is a contingent feature of determinism. He illustrates this with a possible case of someone—Adam—doing a free action at the first instant of time. Thus, Adam had no remote past. If determinism were true at this world, how would we show that Adam’s action wasn’t free? Not with the consequence argument, for it includes the existence of a remote past as an essential feature of (nomological) determinism. Moreover, Campbell’s objection seems to affect almost all arguments for incompatibilism (see Bailey’s (2012) “Incompatibilism and The No Past Objection,” and for a response to Bailey, see Zhang’s (2013) “Can the Incompatibilist Get Past the No Past Objection?”).
I don’t bring up the above points as a criticism of McCall. He states that he’s not trying to be exhaustive. However, I wanted to get these on the table for those who read McCall’s section and want to explore the matter further.
Compatibilisms and The Standard Arguments: Classical Compatibilism
In this section I want to explore the explicit points McCall does raise in relation to his two arguments and the ways compatibilists have responded to it. He cites three compatibilist strategies for taking on these arguments: (i) the classical compatibilist strategy, (ii) the Frankfurt strategy, and (iii) the semi-compatibilist strategy. I address (i) in this post and (ii) and (iii) in the next post.
The Classical Compatibilist Strategy
One strategy McCall discusses (p.65ff.) is the classical compatibilist strategy. He notes that they will reject (2), urging instead that “even though determinism is true, agents may still have the ability to do otherwise (in some relevant sense.” McCall notes that they claim that there are different senses of ‘can’, landing on the “conditional” or “hypothetical” sense of can. According to this sense, to we say that an agent can (is able to) do otherwise just in case, if they had willed (decided, chosen, etc.) otherwise, then they would have done otherwise. Call this the Simple Conditional Analysis (SCA) of ability.
McCall’s response to the SCA rests on the infinite regress objection, specifically Richard Taylor’s version. Essentially, the infinite regress objection allows that statements of the form ‘If S had chosen or tried to do X, S would have done X’ are equivalent to statements of the form ‘S could have done X,’ but this is because we are normally presupposing that ‘S could have chosen or tried to do X’. But when we get there, we must admit that S could not have chosen or tried to do X. Any attempt to offer a compatibilist-friendly analysis of ‘S could have chosen or tried to do X’ will be met with the same objection, and so on ad infinitum.
McCall is on firm ground in choosing Taylor’s objection to SCA. I grant McCall’s response to SCA and would urge theological compatibilists to refrain from endorsing classical compatibilism of the SCA variety. However, classical compatibilism has richer resources than those reading McCall’s criticism of it would pick up on. While McCall does admit that some classical compatibilists still defend a SCA, he doesn’t note that most classical compatibilists have abandoned it in favor of more sophisticated conditional analyses.
Consider what has been referred to as “the new dispositionalism.” On this view, the ‘can’ relevant to freedom of action and moral responsibility is the ‘can’ of capacity or disposition or power, specifically rational abilities.Thus Vihvelin, “To have free will is to have the ability to make choices on the basis of reasons and to have this ability is to have a bundle of dispositions” (“Free Will Demystified: A Dispositional Account”). This is taken to be compatibilist-friendly since most philosophers don’t reject that things can have dispositions if determinism is true (e.g., causal determinism may be true and sugar cubes still be soluble). This view rejects the simple conditional analysis of ‘can’ that is expressed in terms of a subjunctive conditional. It does so in light of recent work in the nature of dispositions, capacities, or powers. Let’s look at the SCA more closely:
(SCA) x is disposed to X at time t just in case condition C obtained at t, then S would X at t.
Here, condition C might be “choosing (wanting or willing) for reasons.” But due largely to the work of David Lewis, it doesn’t seem like we should analyze dispositions in terms of such a conditional analysis. Most philosophers now see that something that has the disposition to X can X even if it is not X-ing and even if it never in fact X’s. Consider a crystal vase wrapped in bubble wrap. The vase can break (it’s breakable) if struck even if it would not break if struck (because it’s wrapped in bubble wrap).
In the light of recent work on dispositions, new dispositionalists have offered complicated (not simple) conditional analyses (CCAs) of ‘can’. Vihvelin (Causes, Laws, and Free Will: Why Determinism Doesn’t Matter (2013)) expends a tremendous amount of effort offering a CCA of ‘can’. Stating the general gist, she writes that “A person has the narrow ability to R just in case she has some intrinsic property or set of properties B that makes her well suited for (good enough at) doing R in circumstances in which her surroundings provide her with the opportunity (extrinsic enablers, no extrinsic masks) to do R and she tries to do R” (p.187). The reader can consult her book for her painstaking work explaining each of the key terms in the quote. One upshot of this approach is that the regress argument McCall employed against the SCA is entirely impotent (cf. Vihvelin 200), in part because it would be assuming a false view of dispositions, viz., even if S would not X (do R) if S chose to (where we hold all the laws and past up to the moment of choice fixed), that doesn’t imply that S can’t (is unable) do X.
As with McCall, I myself have not intended to offer a knock-down defense of classical compatibilism via the new dispositionalism. I have merely intended to show that the main argument McCall employed against classical compatibilism only applies to a classical compatibilism that arose before recent work on dispositions, one many classical compatibilists now reject. To be sure, there are objections to the new dispositionalism. One is that it hasn’t given us sufficient conditions for doing otherwise, even if it has given us correct necessary conditions. One reply, however, is that no analysis of ‘could do otherwise’, compatibilist or incompatibilist, can meet this ‘sufficiency objection.’ For the argument here see Joe Campbell, “Two Problems for Classical Incompatibilism” (in Haji and Caouette‘s (eds.) Free Will and Moral Responsibility (2013)). If Campbell is right, then classical compatibilists and classical incompatibilists are in the same boat.
Finally, I shall close by making the interesting point that it appears that classical Reformed theologians may have been ahead of their time here. First, I note that the Westminster Confession of Faith (9.2) seems to view free will as a power (ability, capacity, or disposition). But more interesting is the fact that Calvin seems to have held to the view of dispositions (or abilities) that accords with recent work on dispositions. Remarking on how Y’s being determined to X does not make Y’s nature absolutely necessary such that it lacks the capacity (ability, power) to not X, Calvin offers this example from the bones of our Lord,
At the same time, that which God has determined, though it must come to pass, is not, however, precisely, or in its own nature, necessary. We have a familiar example in the case of our Saviour’s bones. As he assumed a body similar to ours, no sane man will deny that his bones were capable of being broken, and yet it was impossible that they should be broken (John 19.33, 36). Hence, again, we see that there was good ground for the distinction which the Schoolmen made between necessity, secundum quid, and necessity absolute, also between the necessity of consequent and of consequence. God made the bones of his Son frangible, though he exempted them from actual fracture; and thus, in reference to the necessity of his counsel, made that impossible which might have naturally taken place.
Here we see Calvin saying that Jesus’ bones are breakable even though they are determined not to break. Calvin tied this in with our ability to do otherwise than X even though given God’s decree it’s necessary that we will X. In the light of these facts, it would seem better for McCall to have addressed the new dispositionalism rather than classical compatibilism of the SCA variety for the purpose of addressing theological compatibilism. However, it is quite understandable why he did address the SCA variety, for many (contemporary) Calvinist theologians have assumed the more simple analysis of ‘can do otherwise’. So I appreciate the opportunity McCall’s section provided for some theological compatibilists to try to get clearer about their classical compatibilism. In the next post I turn to semi-compatibilism.