Counter-Arguments and Concessions: Reply to Andrew Harland-Smith’s Grounding Objection

By John A. Limanto


August 26, 2018

1. Preliminaries

In avoiding unintended off-tracking, it is necessary for me to state the points I concede in Andrew’s essay. First, I concede that there may be some ways in construing (DT) that should not bother Molinists (Andrew’s formulation of [(DT)] does not necessarily entail his conclusion as I will argue later). Second, I concede that in the Frankfurtian example he gives, William is not morally culpable for the voting of Hillary if the chip activates. Although, I will argue that the example he gives is unparallel to the case of CCFs and free actions. I concede that in order for an action to be free, agents must be the first causes for their actions. I concede that backward causation is impossible. Lastly, most importantly, I concede that the agent does not cause his own CCFs to be true as there may be agents who do not exist and yet still have true CCFs about them. However, I will argue that this should not be problematic to libertarians.

Peter Van Inwagen once quipped that the “horrible word” is the word “can.”[1] Along similar veins, my thesis in this essay is that “made it true,” as Andrew formulates in (DT), is ambiguous and (DT) may be revised so as to not exclude moral accountability. Therefore, this essay rejects (DT), not on the grounds that it is explicitly contradictory with Molinism, but on the grounds that it is implausible.

2. Scrutinizing (DT)

(DT) For any person S and CCF C, if C is true, S has made it true [emphasis added] that C

In justifying (DT)’s relation with moral culpability, Andrew posits the Frankfurtian examples. Yet, it seems that Frankfurtian examples are ambiguous in illustrating the situation between (DT) and the agent in question. For example, consider the following quote:

So, what, we may ask, grounds his accountability in voting for the wrong candidate? The answer, presumably, is that notwithstanding his inability to realise the alternate possibility, he is still the source of his actions. Simply put, he, and the powers and dispositions within him, are the first cause of the salient outcome.

Thus, the grounds for agent S in bringing about an action is S as the first cause. Andrew further gives the following concluding thoughts:

Where (P) previously depended on the chip implanted in William’s brain, the same cannot be said here. Here, the truth of (P) depends only upon William himself. And it is for this reason that he (William) is accountable for his voting for the wrong candidate. Thus, (‘DT’).

One wonders: in what sense is (P) dependent on the chip or even William himself? To clarify this principle, let us consider certain possibilities:

MT: ‘Made True’ thesis

(MT) For any person S and CCF C, if C is true, S causes it [emphasis added] that C

(MT) is, by far, the closest it seems to get to what Andrew had meant. If (MT) is what Andrew had meant, the Molinists can reject the thesis by saying that the relation between a proposition with its corresponding state-of-affairs (in this case being the state-of-affairs of S causing the consequent when put in the circumstances in the antecedent) is in terms of correspondence and not causation.

To prove that the relation between a proposition and the state-of-affairs is in terms of correspondence, it should be noted that statements about future and past events likewise are not caused by us in the present, and yet, we could claim that agents mentioned in those propositions have properly performed their free actions during the aforementioned time. Thus, (MT) would not be applicable to the Frankfurtian example given that (P) would have been true long before the action takes place (given that the (P) is in its tenseless form). Take, for example, a tensed form of Andrew’s proposition:

(4) Andrew will engage John in a written exchange.

As it is in tensed form, prior to the event occurring, (4) should already be true long before Andrew engages John in a written exchange. Our intuition, at this point, is not to say that Andrew is not free in engaging John in a written exchange, but that Andrew is free because when t comes, Andrew acts as a first cause to engaging John in a written exchange. If this may be the ground (in the sense of providing moral culpability) for (4), then it follows that CCFs may likewise be grounded in terms of the agent’s moral culpability the same way. Therefore, a CCF C is grounded morally in the sense that if agent p is put in c, then p would act as the first cause for X.

Now, what revision may a Molinist make to (MT)? As we have seen, (MT) fails because it cannot account for truths of propositions about the future or the past. The following is a possibility:

(MT2) For any person S and CCF C, if C is true, C corresponds to a state-of-affairs of what S would do in circumstances C1

What is evident is that (MT2) is readily palatable to a Molinist.[2] Here, however, we have defused the notion of ‘made true’ that should exclude Molinism. If God ‘middle knows’ CCF C, it is because C corresponds to a counterfactual state-of-affairs. For example, given:

(5) In circumstance C1, P would freely do X,

then, the corresponding state-of-affairs is the state-of-affairs of its being the case that in C1, P would freely do X. In other words, these states-of-affairs are the truthmakers of the propositions. It may be objected at this moment that such a counterfactual state-of-affairs does not rely on the agent as the agent may not exist! However, this is a false move as it mirrors the mistake that (MT) has proposed earlier: namely, we recognize many states-of-affairs are true apart from there being a concrete (i.e. causally potent) realization of the states-of-affairs at a moment. Therefore, take, for example, a tenseless form of (4):

(6) At t, Andrew engages John in a written exchange,

Prior to t, (6) would be grounded in the state-of-affairs of its being the case that (6). Similarly, past propositions, too, would be grounded in non-concrete state-of-affairs. Take, for example, the time t1 after t. It follows that in such a time, the ground for (6) is still the state-of-affairs that (6). Therefore, as we have settled that states-of-affairs need not consist of concrete objects, it follows that Molinist CCFs may be grounded in a similar states-of-affairs requiring no concrete objects.

3. Is (DT) Essential to Molinism?

Recall to our previous joint statement that Andrew cites in his essay:

“…if the CCFs are a feature of God’s free knowledge, they are no longer truths about what libertarian free agents would do. To the extent that the ultimate source of their truth lies in God’s creative decree, their truth has been fixed independently the relevant agent’s causal powers”.

However, Andrew gives the following complaints:

Of interest here, is not so much the validity of the argument (though that may be challenged). Rather, the interest here, is with the essence of the complaint. It would, we are supposed to agree, be a problem if the truth of the (‘CCFs’) had been set prior to, and hence independently of, the relevant agents’ causal powers. Specifically, we are told, it would undermine our status as free agents.

But what, we may ask, is the general principle here? Well, to some extent, that question has already been answered. Ostensibly, it would be a problem if the truth of the (‘CCFs’) had been fixed independently of the relevant agents’ causal powers. The obverse of this, is, necessarily, that the truth of the (‘CCFs’) are dependent on the causal powers of the relevant agent. Thus, (‘DT’).

The essence of Andrew’s complaint is that if Molinists would not allow for God to cause the truth of the CCFs, then should not it entail logically that agents be the ‘cause’ for their CCFs? Here, it almost sounds as though Andrew resorts to the refuted (MT). Yet, entertaining this thought for a second, consider the following simple syllogism to expose the logical fallacy here:

Given CCF C, person P, and ‘→’ for ‘cause’:

(7) ¬ (¬ PC)

(8) ∴ PC

Yet, (8) is fallacious. Just because nothing outside of P causes C does not entail logically that P causes C. It may be the case that C is an uncaused, brute fact. In fact, this is precisely the strategy that I am going to bring in this exchange: it may be the case that CCFs are true as uncaused brute facts. Andrew must be able to prove a certain principle that excludes brute facts from being realizable (which he has not done!).

4. (DT), Therefore, Molinism is Not True?

Andrew claims that there arises an issue of timing. Simply, truths ((the (‘CCFs’)) about how we would act were true long before we made them so.” Let us use the concrete example that Andrew himself uses:

[C]onsider a situation where God elects to create, not this world, but w2. That is, the world in which I never reject Molinism, and hence never engage in this exchange. In that case, the following proposition is true:

(C’) Andrew never engages John in a written exchange.

As from the very first moment of creation (call it t), God knew (C’). And yet, as from t, there was still some 13.8 billion years to traverse before I would make it true that (C’). To the extent then, that (‘DT’) is true, it follows that my actions caused God to be in the state of knowing that (C’) before I performed them. Hence, backwards causation.

First of all, it should be pointed out that this argument is not an argument against middle knowledge, but an argument against foreknowledge. Even if granted that the argument can easily be applied to middle knowledge (by shifting [(C’)] to a CCF as Andrew himself has hinted on in his paper), this argument falls to the refuted major assumption that CCFs are true because the agent causes his CCFs to be true. As we have demonstrated, this should not be the case. The agent may still be a first-cause in the sense that when the antecedent of the CCF is realized, the agent acts as a first cause in bringing about the consequent in the state-of-affairs. However, the agent himself does not cause his own CCFs to be true in the same way that an agent does not cause the propositions about his future or past actions to be true or false actively in the present.

Andrew himself has raised a good question in the following:

But, to what extent can Molinism agree that the (‘CCFs’) have their genesis in us. Prima facie, at least, it is not altogether obvious that it can.

However, as it is not clear at all what Andrew stipulates for CCFs to have their genesis in us, Molinists may get away by proclaiming (MT2) boldly and saying that these CCFs are true because they correspond to a counterfactual state-of-affairs about what we would do in different circumstances. I shall bring up that moral culpability concerns whether or not we are a first-cause only when the concrete state-of-affairs is actualized and not before or after.


As we have pointed out within this essay, Molinism may maintain that CCFs have their genesis in us in the sense of (MT2). Andrew’s concerns may be alleviated easily through such a principle. Thereby, I’d like to proclaim that (DT) may even be innocuous for the Molinists (granted that there may be other interpretations of [(MT)] that renders Molinism self-contradicting).  Even further, Andrew’s concerns that (DT) entails ‘backward causation’ is a non-sequitur. (DT) combined with (MT2) only entails that propositions correspond to a counterfactual states-of-affairs in the same way that tenseless propositions may be true even before the aforementioned time. Therefore, I accept (1), but reject both (2) and (3).

[1] Inwagen, Peter van. 1983. An Essay on Free Will. New York: Oxford University Press. 65.

[2] I am not here espousing that (MT2) is necessary for Molinism, I think it is a possible route to take.


About the Author

By John A. Limanto

John A. Limanto is a fellow leader of Philosophy group in Pelita Harapan School and an aspiring Christian apologist in his local community. After living for 7 years in Borneo, he now resides with his family in Jakarta, Indonesia where he is pursuing his research on Molinism and free will.