William Lane Craig's Middle Knowledge

William Lane Craig's

"Middle Knowledge, Truth-Makers, and the 'Grounding Objection'"

A Reformed Critique

The present author is in the process of responding to William Lane Craig's (WLC's) book, "The Only Wise God." In the process, however, it seems as though it would be useful to focus on the topic of the final chapter of that book, Middle Knowledge, and particularly on another of WLC's writings on the subject, specifically his article, "Middle knowledge, Truth Makers, and the 'Grounding Objection.'" This topic is doubly timely due to the present author's on-going discussion in the comments section of a treatment of another of WLC's related articles, on Newcomb's Paradox, in which the topic of truth-makers came up.

Presently, a copy of WLC's article is available here (link).

WLC's article purports to address a critical problem for any theory of Middle Knowledge, which he phrases as there being no true counterfactuals of creaturely freedom, because there is nothing to make them true, but which might more aptly be described as there being no way for God to have "free" knowledge (including knowledge based on the (actual or hypothetical) exercise of creaturely freedom) prior to the divine decree.

As WLC admits, the objection is not an undercutting, but a rebutting rejection of Molinism, because if God does not have "middle knowledge," there can be no Molinism, and one must accept either Thomism/Calvinism (which states that God has "free knowledge" posterior to the divine decree) or Open Theism (which denies that God has "free knowledge" at least with respect to creatures).

It is worthwhile to point out, before we begin a detailed analysis of WLC's article, what middle knowledge is alleged to be. It is alleged to be God's knowledge of what "would" happen if a free creature were in a particular circumstance. Thus, it is God's knowledge of what state of affairs "would" result from the exercise of man's (or other free creature's) "free will" in a particular circumstance. Accordingly, this knowledge is a special kind of "free" knowledge, namely knowledge that depends on the creature's exercise of "free will," as constrasted with what Molinists (and most philosophers) refer to as God's "free knowledge," namely the knowledge that God has as a result of the exercise of his own free will.

WLC begins by setting forth the alleged warrant for the Molinist assumption (WLC himself calls it an "assumption").

The first alleged warrant is that "we ourselves often appear to know such counterfactuals." This warrant, however, can be readily dismissed. Our knowledge of "counterfactuals" generally falls into one of two categories, neither of which is consistent with the Molinistic assumption. The first category are counterfactuals that are true by virtue of the law of the excluded middle (i.e. we know that it is true that if President Fox were held by terrorists, the US government either would or would not intervene. The second category are counterfactuals for which our ground is causal inference, which negates freedom. An example in this category would include your knowledge that if you showed up late for work for a week and then insulted your boss to his face, your boss would react by firing you. There is not some third category of counterfactuals that we know to be true by some other mechansim. Thus, this alleged warrant falls flat on its face.

The second alleged warrant is that "it is plausable that the Law of the Conditional Excluded Middle (LCEM) holds for counterfactuals of a certain special form." As noted above, however, this is simply an example of the first category of counterfactuals that are true because they are just restatements of the law of the excluded middle. They do not assist Molinism in any way, because they do not depend at all on the exercise of a free will, and do not correspond to anything more than the application of a logical rule to a hypothetical scenario. Thus, this alleged warrant is also insufficient.

The third alleged warrant is that Scriptures are "replete with counterfactual statements." Scriptures, however, do not identify any creaturely acts as "free" in the Molinist (Libertarian) sense, and Scriptures also include counterfactual statements describing God's own actions. Finally, all such Scriptural statements are made posterior to the divine decree. Thus, such statements cannot possibly be a warrant for the Molinist assumption.

WLC implicitly notes at least some of these objections and suggests that "anti-Molinists" retreat to asserting that counterfactuals have no ground prior to the divine decree. WLC's only two responses are that "there seems to be no more ground now for many counterfactuals about creaturely free acts than there is logically prior to God's decree," and if "it is God who decrees which counterfactuals about creaturely free acts are true" then this seems to "make God the author of sin and obliterate human freedom."

The former of these objections ignores the obvious: posterior to the divine decree we have reality - the world that exists and operates according to the laws of Ordinary Providence. Thus, counterfactuals become grounded in reality.

The latter of these objections is wrong for enough reasons to occupy a major treatise (and Jonathan Edwards has written one that fully rebuts it). To briefly summarize, God is not the "author" of sin because he does not do sin: he is not a voluntary agent that performs a sinful act. Furthermore, God is not the "author" of sin because no one who sins is coerced into sin contrary to their will by God. If WLC means something else by "author of sin" we can respond with Vincent Cheung that we will simply grant the objection, and ask why that should be a problem. Used outside the first two items listed above, it is simply pejorative rhetoric, not a meaningful objection. Additionally, it is "libertarian" (not compatible) freedom that obliterates human responsibility, because it necessarily denies that man determines what man does. Or, to take a hint from WLC, we could simply deny the charge of obliteration of human responsibility and ask WLC to provide evidence. Nevertheless, as noted above Jonathan Edwards' "Freedom of the Will," has already thoroughly disabused any serious reader of the truth of either of the charges above, though (of course) WLC neither recognizes nor responds to Edwards' detailed and famous rebuttals.

Although we have seen that the Molinist assumption is philosophically warrantless, let us proceed to see what kind of response WLC is able to provide to the grounding objection.

WLC begins by explaining an alleged relationship between the grounding objection and a theory of "truth-makers." Theory, in essence, states that in addition to truth bearers (sentences, thoughts, or propositions) there are entities by virtue of which the truth bearers are true. Someimtes these truth-makers are labelled as "facts" or "states of affairs."

WLC recognizes what is intuitively obvious that the term "truth maker" suggests a causal relationship between the truth-maker and the truth of the truth bearer. (This same problem for the Molinist position had been observed by Dan in the comments section of the post linked-to above.) WLC's response is bafflingly off-base:

First, WLC responds that the truth-maker is usually conceived to be "such abstract entities as facts or states of affairs" and asserts that consequently a causal relation is not at issue here. This objection is bizarre, to say the least. While the terms "facts" and "states of affairs" are abstract, "facts" and "states of affairs" are categories of things that (mostly) have (or had or will have) tangible existence. Individual facts are not "abstract."

Second, WLC attempts to show that "negative existential statements" do not have truth-makers. If this were not presented as an attempted rebuttal, we would assume it were a poor play on words, for "do not have" with respect to statements of "negative existence" is fully consistent with a theory of truth-makers. The fact is that negative existential statements can be rephrased quasi-positively. Thus, "Baal does not exist" can be rephrased as "Everything that exists is a non-Baal," or more simply "All gods are non-Baals," since it is alleged that Baal is a god. Negative existential statements, therefore, can be converted to quasi-positive statements, for which there is an obvious truth-maker.

WLC then asserts that the various phrasing of the grounding objection are "crude construal[s]" of the problem. We have demonstrated, however, that WLC does not have an answer to the question variously phrased, but which amounts to:
Counterfactuals are supposedly true but not caused to be true by God, who then causes them to be true?
(A similar general question remains unanswered in the discussion thread identified above.)
The result of WLC's failure to answer this question is that WLC cannot provide an explanation for the truth of counterfactuals, and consequently can provide only an incomplete epistonomy.

Instead of identifying a cause, WLC simply calls the call for a cause (the request for grounding) "inept" and states that the objector must set forth some kind of "very special causal theory of truth-makers" in order to counter the "customary truth-maker theories." As we noted above, however, WLC has failed properly to analyze the customary truth-maker theories. Furthermore, no special theory of causality is required: everything external to God cannot be self-existent, and therefore requires a cause. The truth of counterfactuals is alleged to be external to God. Therefore, the Molinist should identify the cause, or acknowledge that the truth of counterfactuals is simply special pleading for Molinists.

WLC then identifies seven proposition that allegedly do not have "truth-makers." These include:

1) No physical objects exist.
(Which is not true, but if it were true the truth maker would be readily seen from the fact that "all things that exist are non-phsical things" has the truth-maker "all things that exist."
2) Dinosaurs are exitinct today.
(Which may be true, and if it is true today then the truth maker is a combination of the dinosaurs that once were (as to the implication of previous non-extinct status for the class of dinosaurs) and the status of all things as being members of the class "not-dinosaurs."
3) All ravens are black.
(If this were true, the truth-maker would be the color of all ravens.)
4) Torturing a child is wrong.
(If this is true, then the truth-maker is the standard of wrong, the moral law.)
5) Napolean lost the battle of Waterloo.
(The truth maker here could variously be ascribed to historians, Napolean's blunder, British genius, or God's providence.)
6) The president of the U.S. in 2070 will be a woman.
(If this is true, the truth-maker is inter alia God's decree.)
7) If a rigid rod were placed in uniform motion through the aether, it would suffer a FitzGerald-Lorentz contraction.
(Which has as its truth-maker the definitions of certain theories of 19th century mechanics and/or causal inference therefrom.)

In short, WLC presents seven propositions that he believes there is no truth-maker to be found for, and implicitly asserts that consequently there is no problem if no truth-maker can be found for counterfactual statements. But WLC, fortunately, does not rest on that idea (which has been demonstrated wrong above by provision of relevant truthmakers).

Instead, WLC continues with a section entitled "Do Counterfactuals of Creaturely Freedom Need Truth-Makers?" Of course, if WLC can answer this question negatively, then the problem would appear to be solved - regardless of the merit of any of the seven examples identified above.

WLC's contention appears to revolve around first, asserting that truth-maker theory is a minority philosophical position; second, some truth-maker philosophers permit some true statements to lack a truth-maker; and third, WLC has not found an explanation as to why counterfactuals should not be permitted to be without a truth-maker. None of these, of course, provide an excuse for counterfactuals to avoid the need for a truth-maker.

WLC's next question is whether such counterfactuals have truth-makers. WLC notes that Fredodoso has attempted to provide a truth-maker, namely the state of affairs that the propisition will have a truth-maker or that the proposition would have a truth-maker under the relevant condition. WLC recognizes that this alleged truth-maker creates the obvious, immediate, and insoluable problem of infinite regress, but seems to think that it is not so. WLC asserts that facts do not have to have truth-makers. That assertion may be true generally, to wit that ordinarily facts do not require truth-makers. This particular alleged fact, however, makes reference to another truth-maker, and thus creates the self-referential problem.

The problem is that WLC is taking shelter in the use of a similar technique to that used by philosphers who deny present grounding of future tense statements. Thus, WLC takes comfort in the fact that some explanatiosn of the grounding of future tense statements also suffer from infinite regress. WLC, however, seems to have overlooked the fact that many future tense statements can have other forms of grounding.

For example, the statement "tomorrow will be another day" is grounded in causal inference from the succession of time (and/or the definition of the word "tomorrow"), and "tomorrow will be rainy or not" is grounded in the law of the excluded middle. Finally, with respect to human perspective future tense statements that are not based on laws of logic or causal inference, such statements very rarely have grounding, and consequently are not "true" if, for a statement to be true, it must have grounds of truth. Note that what is being implied (but not stated) is that the statement must have grounds of truth for the speaker. In other words, the speaker must have a basis for the statement. The statement "a woman will be president of the U.S. in 2070" may turn out to be a correct statement of the state of affairs that transpires in 2070, but an ordinary human cannot have grounds for that statement - and, thus, while we might label the statement "true" after the fact, at the time the statement was made it was simply a guess.

This is not so with respect to divine omniprescience. God knows the future with the positive grounding of the divine decree. Thus, God's beliefs about the future are not guesses, but instead have the ground of the divine decree. Furthermore, God's beliefs about the hypothetical future are most readily grounded in causal inference, just as are ours - but God knows all the factors and thus has perfect knowledge of all hypotheticals posterior to the divine decree.

In contrast, for the Molinist, there is nothing that could ground God's knowledge of "counterfactuals" prior to the divine decree. Thus, God could not "know" such counterfactuals, and if God could not know such counterfactuals, then Molinism is dead.

WLC suggests that the ground of truth for a conditional "if A were in C, A would do S" is the hypothetical state of affairs described in the statement. This is somewhat similar to the non-predestinarian alleged grounding of future tense statements. The problem with that assertion, however, is that the hypothetical state of affairs does not exist, and (if it is really a counterfactual) will never exist. In other words, unlike future tense statements, there will never have been a corresponding state of affairs in which A is in C and does S, and the fact there will never have been a corresponding state of affairs in which A is in C and does S would "falsify" the future tense statement. Thus, WLC is relying for truth grounding on the very thing that would falsify a future tense statement, and a conditional statement does not purport to convey negative, but positive, information. Accordingly, even under a non-predestinarian analysis, WLC's basis for grounding is all wet.

Elsewhere, such as in "The Only Wise God," WLC has proposed an alternative basis for grounding. That basis is the fact that God knows all true counterfactuals. The problem with this attempted grounding should be immediately obvious. First, there is no reason to suppose that God knows all true counterfactuals prior to the divine decree, and perhaps even more plainly, it itself cannot have adequate grounding, unless we are to suggest that it is in God's nature to know all true counterfactuals. But if so, then true counterfactuals are part of God's natural knowledge, which both creates a paradox with respect to God himself, imposes necessity on counterfactuals, and takes out the "middle" from "middle knowledge." In short, an assertion that God's natural knowledge encompasses counterfactuals as much destroys Molinism as an assertion that God's free knowledge encompasses counterfactuals. Both scenarios destroy Molinism utterly.

WLC recognizes that the way that we treat normal hypotheticals (if I fire a shotgun at a pane of ordinary glass from a distance of 5 feet, the glass will break) is by identifying the causal inference. WLC, however, does not want to apply this to creaturely free acts, because he recognizes that doing so would destroy their libertarian freedom.

WLC even goes so far as to state that asking the question, "Why is F a fact?" or "What makes F a fact?" is to deny libertarian freedom, if F is a libertarian free act. The result is that WLC is unable to provide an account for the existence of F, and WLC is unable to handle Scriptural assertions that there are reasons why F is a fact (for example, there is a reason why Thomas believed). In short, WLC's philosophy falls prey to all the objections raised at great length and depth by Jonathan Edwards centuries ago.

Thus, when we arrive at WLC's conclusion that "anti-Molinists have not even begun to do the necessary homework," we are left with precisely the opposite impression. Specifically, we left with the impression that there is no warrant for accepting the Molinist assumptions, we are left with the impression that Molinism's assumption ends up contradicting Molinism, and we are left with the impression that Molinism fails to provide a complete epistemology: Molinism asserts that there are true counterfactuals of creaturely freedom and that they were known by God prior to the divine decree, but cannot give any account of them.

Finally, we may simply point out that the primary criticism raised against the Reformed view is essentially transitive to the Molinist viewpoint (in debate terms it is a "non-unique" criticism). That is to say, that while Molinists may deny that God makes the true counterfactuals of creaturely freedom true, nevertheless, God's selection of a particular world from all possible worlds (in Molinism) makes God in essence as much the cause of the evil that will be, as does God's selection of a particular world from all possible worlds in Calvinism or Thomism, because God recognizes that the consequence "If A is in C, A will do S," attaches just as certainly to C whether the link between C and S is causal (Calvinism/Thomism) or mysterious/unexplainable/impenetrable (Molinism/Arminianism). Thus, either way, C => S, and God's selection of C implies God's selection of S. In short, it is only the obviously mistaken position of Open Theism that an escape the charge that WLC attempts to level.

Thus, we may safely conclude that WLC has been unable to withstand the grounding objection proposed against Molinism, and the reader has no reason to believe in such a philosophical position. Instead, the reader should believe in the God who knows all hypotheticals on the basis of His infinite wisdom in combination with the Divine decree.

-Turretinfan