05 / 06

Ducking Friendly Fire: Davison on the Grounding Objection

William Lane Craig


Though Scott Davison is sympathetic to a Molinist account of divine providence, the so-called Grounding Objection still gives him pause. In response to my claim that the Grounding Objection is based upon an untenable view of truth-maker theory, he therefore directs some “friendly fire” at Molinism on behalf of Grounding Objectors. Fortunately for Molinists, his key principle, I hope to show, appears merely to re-state such an untenable view. If construed as a demand for explanations which are not truth-makers, the principle is incompatible with libertarian freedom. Molinists emerge therefore unscathed from his salvo.

Davison first disputes my claim that the Molinist is under no obligation to provide warrant for the assumption that there are true counterfactuals concerning creaturely free choices. [1]  But it seems to me that there are two reasons why it is the so-called Grounding Objector to the doctrine of middle knowledge who bears the burden of proof in this respect. First, Molinism is itself an undercutting defeater of the alleged incompatibility of divine sovereignty and human freedom. Historically, Molinism is the fruit of the Counter-Reformation’s attempt to resolve the antinomy between God’s sovereignty and human free will perceived by Protestant Reformers like Luther and Calvin. In order to defeat the incompatibility claims of the Reformers, Molinists had merely to provide a biblically faithful theory or model exhibiting the compatibility of God’s providence with libertarian free will. To borrow a familiar phrase, the Molinist project is first and foremost one of defense, not theodicy.

Therefore, Davison’s assertion that “true Molinists claim not just that God’s providence might involve knowledge of true counterfactuals of creaturely freedom, but rather that it actually does” [2] is ambiguous. According to the model, God actually does have such knowledge; but the model itself is put forward as an undercutting defeater, as a theory which is epistemically possible. The Molinist will be dialectically successful vis à vis incompatibility claims if he shows that for all we know God’s sovereign decrees are guided by counterfactuals concerning human free decisions which are true and known by God prior to His creative decree. Therefore, the burden rests on the detractor of Molinism to show why this theory is not epistemically an option.

Second, even if the Molinist puts forward his theory, not as a defense, but as a theodicy, it does not follow, as Davison claims, that Molinists “have the initial burden of proof concerning the assumption that there are true counterfactuals of creaturely freedom.”  [3] For that assumption is simply one of the postulates of the theory, which must be assessed as a whole in terms of typical criteria of theory assessment. Compare the case of the Special Theory of Relativity. Crucial to that theory is the so-called Light Postulate, which asserts that light travels in vacuo at a constant speed c. Not only is there no evidence for this postulate, but as numerous commentators have explained, the Light Postulate is inherently unprovable because we can only measure the round-trip speed of light. Nonetheless, Einstein claimed, grant me this assumption along with the theory’s other postulate, the Relativity Postulate (which is also impossible to prove!), and I shall provide you with the best explanation of the electrodynamics of moving bodies. Despite the unprovability of its postulates, the theory as a whole is considered by the majority of physicists to be the best explanation (within its restricted domain of flat spacetime) in view of its theoretical virtues like economy and empirical success. Detractors of the theory do not even dent confidence in the theory (not to speak of showing it to be false!) merely by pointing out that its postulates are unprovable.

In the same way the Molinist theory of providence must be judged as a whole in terms of its philosophical coherence, explanatory power, theological fecundity, and so forth.  It is a non-starter for its detractors to point out that Molinists have not proven the postulates of the theory to be true. Thus, even though there are in this case very good reasons to think that its key postulate is true, [4] the Molinist bears no initial burden to prove this postulate in order to commend his theory as the best account of divine providence available.

So the question is, has the Grounding Objector given us good reason to think that the key postulate of the Molinist theory of providence, viz., that counterfactuals about creaturely free acts are true logically prior to the divine decree, is false? Grounding Objectors, I allege, seem to be presupposing a construal of truth as correspondence known as the theory of truth-makers. Davison agrees that the crude statement of the Grounding Objection in causal terms, so popular in the literature, is a mere rhetorical flourish. But in response to my claim that “the anti-Molinist still seems to be presupposing that in order to be true, counterfactuals of freedom must have truth-makers that either are or imply the existence of concrete objects,” [5] Davison asserts,

The Grounding Objector need not claim that all truths are grounded in a certain way, just that certain truths must be. In particular, the Grounding Objector needs to insist on something like the following principle:

GO: The truth of a true proposition concerning a specific human person must be explained in terms of the actual situation of the person in question.

None of Craig’s examples is a counterexample to this principle. [6]

We agree that the issue is whether certain truths, viz., counterfactuals of freedom, must be grounded. If “explaining” is a type of truth-making, then (GO) does presuppose that counterfactuals of freedom must have truth-makers which are or imply the existence of concrete objects, viz., persons.  Therefore, my original demand for some explanation and justification of this restriction on what counts as a truth-maker is appropriate and remains unmet by Davison. I went further to allege that the restriction is implausible and offered several counter-examples. Construed as a truth-making principle, (GO) is, in fact, vulnerable to both

5. Napoleon lost the Battle of Waterloo


6. The U.S. President in 2070 will be a woman.

Others of my counter-examples can be converted to statements about specific human persons simply by substituting proper names or definite descriptions where appropriate, e.g.,

2. The ancient Aztecs are extinct today.

4. Torturing the Iraqi prisoners in Abu Ghraib is wrong.

None of these statements can be grounded in the actual situation of the person(s) involved; in some cases the persons do not even exist.

At this point Davison would perhaps protest that “explaining” is not the same as “truth-making.” He says, “The real issue behind the Grounding Objection . . . is this: logically prior to creation, what explains why a given true counterfactual of creaturely freedom is true, and what explains why a given false counterfactual of creaturely freedom is false?” [7]

But then what content is to be given to the notion of “explanation”? Davison complains that counterfacts (the states of affairs revealed by the disquotation principle) are “not really a satisfying explanation.” Why not? Because, he says, “it is customary to regard contingent states of affairs (or facts) as obtaining in virtue of some arrangement of concrete objects." [8] But this answer merely lapses back into the restriction on truth-makers rejected above, for a truth-maker is defined precisely as that in virtue of which a fact is true! Thus, when Davison concludes, “So the truth of a contingent proposition concerning a specific human person . . . is customarily to be explained in terms of the actual situation of the specific human person in question,” [9]he is mistaken, not merely because truth-maker theory neither asserts this nor is customary, but more fundamentally because many such contingently true propositions cannot be explained in these terms.

Davison thinks that “Molinists must simply accept the truth of counterfactuals of creaturely freedom logically prior to creation as a brute, unexplained fact.” [10] Here I think we see the real root of the problem. Grounding Objectors have difficulty accepting the existence of brute facts. These latter-day Leibnizians want everything to be brought into submission to the Principle of Sufficient Reason, including facts concerning human free choices. They want an explanation even for why counterfacts about free creaturely choices obtain. Such a demand bears out what I have suspected all along: Grounding Objectors are implicitly presupposing a non-libertarian view of freedom, for counterfactuals of creaturely freedom are no more inexplicable than certain categorical, present-tense statements about the actual situation of free agents, e.g., “Jones is freely writing a letter to his wife.” We can have all the truth-making we want for such a statement but what we cannot have on a libertarian view is an explanation of why Jones freely so acts [11]

Davison says he is not so sure about this reply. “Libertarians need not be committed to the truth of counterfactuals of freedom logically prior to creation. (And, in any case, the Grounding Objector need not accept Libertarian freedom.)” [12] Both statements are true, but irrelevant. What libertarians are committed to is the brute ultimacy of free choices and so to their final inexplicability. And Grounding Objections not rooted in the want of a sufficient reason for libertarian choices should be seen, so far as I can tell, as objections which are predicated on the assumption that counterfactuals of freedom require truth-makers that either are or imply the existence of concrete objects.

  • [1]

    Scott A. Davison, “Craig on the Grounding Objection to Middle Knowledge,” Faith and Philosophy 21 (2004): 365, in response to William Lane Craig, “Middle Knowledge, Truth Makers, and the ‘Grounding Objection’,” Faith and Philosophy 18 (2001): 337-52.

  • [2]

    Davison, “Grounding Objection,” p. 365.

  • [3]

    Ibid., p. 366.

  • [4]

    Davison is much too cavalier in his dismissal of the reasons for thinking that there are true counterfactuals of freedom. (1) We ourselves often appear to know such counterfactuals. Davison’s reiteration of Adams’s claim that all we know is that if someone were in certain circumstances, he would probably perform a certain action fails, in light of Plantinga’s refutation (“Reply to Robert M. Adams,” in Alvin Plantinga, ed. J. E. Tomberlin and P. van Inwagen, Profiles 5 [Dordrecht: D. Reidel, 1985), pp. 379-81), to advance the discussion. Davison’s further claim that we cannot say with confidence that any action is free is remarkable, for it constitutes an objection, not to counterfactual knowledge per se, but to knowledge of libertarian freedom tout court, even our own. This is cutting off one’s nose to spite one’s face. (I’m reminded of William Kneale’s remark apropos Russell’s “no classes” proposal: “The project of getting rid of paradoxes about classes by getting rid of classes is extravagant in the same way as a proposal to get rid of children’s diseases by getting rid of children” [The Development of Logic (Oxford: Clarendon, 1962), p. 659].) Davison’s scepticism would deny to us knowledge of even categorical statements, such as “Curley freely accepted the bribe” or even “I freely performed some action A.” Notice that Davison, on these grounds, should have responded to Plantinga’s example of Curley’s accepting a larger bribe by saying that we do not know that Curley freely accepted the smaller bribe—but Davison instead reverts to the Grounding Objection. The point is that libertarian freedom is a separate issue which is part of the background beliefs assumed by the doctrine of middle knowledge, just as the existence of an external world of physical objects is a background belief assumed by the second law of thermodynamics, and is therefore the subject of independent inquiry. At the least we should agree that anyone who does think that we have good grounds for believing that some of our actions are freely performed also has good grounds for thinking that certain counterfactuals of freedom are true.

    (2) The Law of Conditional Excluded Middle plausibly holds for counterfactuals of creaturely freedom. Davison thinks that this claim assumes that there is a matter of fact about what an agent would do under certain circumstances, which threatens to beg the question. I take it that Davison’s point is that perhaps only “might-counterfactuals” are true of free agents, not “would-counterfactuals.” But reason (2) is meant precisely to foreclose that option. For we are talking in this case about a very special set of counterfactuals involving the choices of some agent in fully specified circumstances. Such restrictive parameters remove the sort of ambiguities that serve to support mere might-counterfactuals. Thus, although there are might-counterfactuals aplenty, it is plausible that LCEM holds for a class of would-counterfactuals as well.

    (3) The Scriptures abound with counterfactuals about creaturely free actions. Since the Scriptures do contain abundant counterfactuals about creaturely actions, Davison, to resist this point, is forced to deny that the Scriptures affirm that these actions are free in a libertarian sense. Again, this seems to me a desperate recourse. The Scriptures breathe libertarian human freedom, even if the Bible makes no explicit mention of it. Take, for example, I Cor. 10.13, which promises that God “will not allow you to be tempted beyond what you are able, but with the temptation will provide the way of escape also, that you may be able to endure it.” It follows that any Christian who does not in some circumstance endure but succumbs to temptation had it within his power to take the way of escape instead, i.e., he had the liberty of opposites in those circumstances. Again, we can at least agree that if the Scriptures do presuppose or affirm libertarian freedom, then there is no basis for denying that sentences like I Cor. 2.8 are true counterfactuals of freedom.

    The question which then remains is whether these counterfactuals are true, not only posterior to, but prior to the divine decree. Since some of them are no more grounded now than prior to the decree, the Grounding Objector owes us an explanation for any restriction he might seek to impose. Moreover, if they are true only posterior to God’s decree, then human freedom and responsibility for sin seem precluded, since God determines which counterfactuals are true and which are false. (The world-type which confronts every human person would include all counterfactuals about his own choices, so that they are outside his control—a conclusion akin to Adams and Hasker’s fatalism.)

    Davison’s pessimistic conclusion that because Hasker has charged that Molinism makes God the author of sin, the disagreement is irresoluble is a rash generalization borne, one suspects, out of impatience rather than a careful weighing of the arguments.

  • [5]

    Craig, “Middle Knowledge.” p. 341.

  • [6]

    Davison, “Grounding Objection,” p. 367.

  • [7]


  • [8]

    Ibid. N.B. that Davison’s attempt to call upon Plantinga in support misreads our Frisian friend. Everyone agrees that if Cantor had not been a mathematician, the proposition expressed by the sentence “Cantor is a mathematician” would have been false. Such a contention is far too weak to constitute an endorsement of any sort of truth-maker theory.

  • [9]

    Davison, “Grounding Objection,” p. 368.

  • [10]


  • [11]

    Another counter-example: “God created Jones,” a statement which cannot be explained in Davison’s sense in terms of the actual situation of Jones.

  • [12]

    Davison, “Grounding Objection,” p. 368.