In a recent thread I claimed that the Transcendental Argument for God (TAG) was obviously unsound. NotReady challenged me on this claim, so in this thread I’ll try to show that this argument fails (some of our discussion in that thread was over whether my claim that it was obviously unsound was incorrect. Here I’ll just show that it is unsound.

While TAG arguments originated in Immanuel Kant’s philosophy, they are today mostly championed by a followers of a Presbyterian theologian named Cornelius Van Til. Van Til developed a new approach to apologetics called presuppositionalism. He claimed that previous arguments for God’s existence had failed because they falsely supposed that there is a neutral logic or science that can be used to show God exists. Instead, according to van Til, logic and science already presuppose God exists. Thus, when anyone, including atheists, use logic or science to claim that God doesn’t exist, they are implicitly contradicting themselves (similar to how Descartes said it would be a self-contradictory to say that “I don’t exist”). Hence the name, “presuppositionalism.”

This idea is formalized in the TAG argument. This argument is modeled on Kant’s argument for the synthetic a priori—rather than asking what we can know about the world through empirical or rationalistic reasoning, we should ask what are the preconditions of us having experiences in the first place. The TAG argument makes a similar type of claim—whether about morality, logic, science, knowledge, or intelligibility—and identifies God as one of these preconditions.

So here I’ll take logic as an example. According to the presuppositionalist, logic presupposes the existence of the Christian God. Thus, if you think that logic is valid, you should also accept that the Christian God exists. The general argument might go something like this:

1) Logical claims can only be true if the Christian God exists.
2) Logical claims are true.
3) Therefore, the Christian God exists.

Here’s a defense of the first premise. Logical claims are the basic principles on which the rules of logic depend, e.g. the law of non-contradiction, identity, etc. So how do we know that these claims are true? It would seem to be circular to use a logical argument to justify them (e.g. see “What the tortoise said to Achilles”) as you must assume the correctness of these logical claims in order to have an account of validity. You also couldn’t use an empirical or inductive method as these logical claims are absolute and necessary—applying to all claims of truth, whereas empirical and inductive methods cannot ever give us necessary or absolute truths.

So what is the grounding for the truth of logical claims? The Christian God is an orderly god who created the world to be orderly and communicated this order to his creation. Thus, we know that these logical claims are true because God has shared this insight into his character, and hence his creation, with us. Without this knowledge from God, we could not know that logical claims are true.

The second premise is usually assumed by the defender of this argument. He will generally view denying it as self-defeating. And, since the argument is valid, God exists!

I don't get it, are you preparing a follow up post where you give your argument where it is unsound and I have just fast ponied?

That would seem to be a safe assumption when the first paragraph states "Here I’ll just show that it is unsound", yet the last sentence claims "And, since the argument is valid, God exists!"

I'm getting there. Work intervening.

I will start by distinguishing TAG from an argument that has some superficial similarities. So, some theistic philosophers have argued that God is the best explanation of, variously, morality, logic, science, and so on. For instance, Robert Adams argued that theism provides the best explanation for the existence of objective and non-natural moral facts. John Foster argued that theism provide the best explanation for the existence of laws of nature required for the solution of the problem of induction.

Those arguments gain their force from abductive reasoning. Abductive reasoning is where you compare different explanations for some known data and choose the one that is best. A typical way of deciding which is best is done by showing that one of the explanations has to a greater degree the typical characteristics of correct explanations e.g. simpler, testability, consistency, and so on (you might think of these as peripherals that accompany true explanations, thus making the explanation with a greater degree of these characteristics more likely to be true).

This is not the kind of reasoning that underlies the TAG. TAG arguments do not compare explanations for some data, but instead function by showing what is implicitly the case in order for something to be true (or exist). For instance, we would say that in order for sound to exist there must be some medium through which the sound wave travels. Thus, a medium is a necessary precondition of there being any sound at all.

In this case, what we are explaining is the truth of logical claims. An abductive argument would look at different explanations and choose the one that is best. But that isn’t what the TAG argument does. That looks at what is required for the existence of logical truth. The relevant feature that needs to be explained is the necessity and absoluteness of logical truth. That is, the law of non-contradiction is a tautology, meaning that it is true under all interpretations. This means that logical truths are necessarily true (i.e. in all possible worlds (A = ~A) is false.). It is also supposed to be absolute in that it applies to all true statements. There is, on this view, no way to truly say a contradiction.

God is supposed to ground these claims by locating these claims about logical truth in God’s nature. God’s nature is necessarily what it is and so can provide a ground for the necessity of logical claims. Without God there is nothing that will make logical truths necessary—they would only be contingently true. Similar considerations apply to the absolute nature of logical truths.

Now, the first issue with this argument is that it doesn’t, on its own at least, show that God is a necessary precondition for logical truth. We might say that it is a necessary condition of any logical truth that it is necessary and absolute (a claim which has been challenged). However, the function of God here is just to explain how logical truths have these features. But the existence of one explanation doesn’t preclude their being other explanations for the same data. Using my sound analogy from above, it would be equivalent to saying that air is a necessary precondition for sound. Unless we can show that air is the only possible medium through which sound waves can travel we cannot correctly say that it is a necessary precondition for sound.

So it would be adequate to show that this argument fails to demonstrate that there are other possible explanations for logical truths. The account I’ll defend here as a possible explanation is that of the conventionalist. Logical conventionalists (most typically associated with the logical positivists of the early twentieth-century) claimed that logical truths are true by definition. That is we define, by convention, some words to have certain meanings and from this the logical truths follow. Thus, logical truths are all tautologies—that is empty of any real content. It is necessarily true that a bachelor is a bachelor not because of some special property of the universe, but because of how we’ve defined “bachelor.”

As for the absolute nature of these claims, this ends up being a narrower claim than imagined by the theists. Logical truths are indeed absolute in languages in which they are absolute, but not in languages in which they are not. But any language in which those logical truths do not apply (if there is any such language), just has different definitions for their concept than in our language where the logical truths do apply.

Now, I’m not claiming that this is the correct account of logic (in fact I don’t think it is), but I do think it is adequate as a possible account of logical truth. As such, it is enough to show that presupposing that God exists is not a precondition of logical claims being true, showing that the TAG argument fails.

Now in this response I’ve more or less accepted the account of logical truth put forward by the TAG proponent. However, I should point out that in fact this is in fact very controversial. For instance, in one of the most famous philosophy articles of the last century, “Two Dogmas of Empiricism,” W.V.O. Quine argued that logical truths are not in fact necessary statements. This view of logical truth would be enough on its own to show that God is not a necessary precondition for logical truths, because it shows that logical truths do not have the features claimed by the presuppositionalist.

You might wonder why the presuppositionalist uses the transcendental form of argumentation rather than abduction when doing so places such a greater argumentative burden on him. The reason is theological. Rather than starting by looking for arguments or reasons that show the existence of God, the primary proponents of presuppositionalism start with a theological claim—that the Christian idea of god is of a being so great that every It is as they asked, what would be the most impressive proof possible, and then said that because God is so great the proof of his existence must be one that uses this most impressive proof. This of course skips over the minor step of whether that proof actually exists.

You actually see this in operation in how presuppositionalists argue. They will often start by saying that they will describe what are the necessary preconditions of logic, but when they actually present their argument it often is just them saying that atheists have no explanation for logic and so their explanation wins as it were by default. But the lack of an explanation on the part of the atheist doesn’t show that theism is a necessary precondition for logic. It would only shows that without theism we wouldn’t currently have an explanation for logic.

Euthyphro dilemma except substitute logic for good/the pious?

Sick 1k post imo.

Is the bolded a strawman of canonical defenses of P1? As in, since the underlined is P2, then the implication of the bolded seems to be that P1 follows from (other stuff) plus P2. But if that was the case, the argument would have had a different structure, namely the other stuff plus P2 as assumptions, then P1 would have an intermediary conclusion, not a premise. So it seems that either I am misunderstanding, such a canonical argument is truly bad, or you are misrepresenting the argument.

Does van Til say more about P1? I assume your defence of P1 is somewhat in line of what van Til is saying, but as it is, this is just metaphysical babble. He needs to show a connection between truth and logical claim. Just saying "God is that connection" doesnt cut it, as it's just an article of faith (as per "God is orderliness" etc. is an article of faith).

I quite like presuppositionalism, as it makes it impossible to know if god exists... it can only be assumed axiomatically.

This is because you end up in a recursive loop. Logic and knowledge presupposes God, but unless this is an axiom it must in itself be knowledge.

This means any knowledge regarding god is presupposed out of a god you can't know. That means personal evidence is impossible. When personal evidence is impossible you can have no revelation. When you have no revelation, you have no revealed religion. When you have no revealed religion Christianity, Islam, Judaism and most forms of Hindusm and religious Buddhism goes out the window (and likely many others).

And yes, you can certainly make a special case for a specific religion (which is what is typically done), which I think is funny... because then if you boil the argument down to the bone this is what is left:
"I am right because I am right."

I'm not sure I'm following your objection here, so let me restate the justification for P1 to see if that addresses it.

To show that P1 is true, presuppositionalists typically try to demonstrate two things: that if the Christian God exists, then he would create a world in which logical claims are true (this roughly tracks the considerations you quoted above) and if the Christian God didn't exist, then logical claims are not true (which typically relies on raising skeptical objections to various philosophical attempts at justifying the foundations of logic.). If you combine those two considerations, you get the claim that logical claims are true if and only if the Christian God exists (sorry, my P1 in the OP is a bit muddled).

A couple points. First, my presentation of presuppositionalism is based more on some Van Til's contemporary defenders (e.g. Greg Bahnsen, John Frame, Matt Slick) than on his own writings, with which I am not very familiar.

That being said, I think the response here would be that the presuppositionalist rejects your framing of the question. Here is a summary of how this might go by Nathan Pitchford:

The idea I think is that any human-based epistemology will always fail to justify the absolute and necessary truths of logic--only a supernatural epistemology based on God could do that. Thus, you'll typically find presuppositionalists essentially make skeptical arguments about the possibility of any (secular) knowledge (their debt to Hume is explicit) and then appeal to faith in grounding their own view.

Meh. This:
is a known smoke bomb: making god an object of rational enquiry doesn't make god an object. 4th lateran council represent.

In case this wasn't clear, Van Til and his followers are hardcore Protestants, so they won't be very impressed with your heretical Roman Catholic councils.

Hm... let's try to make it work:

All the issue is with P1: Logical claims can only be true if the Christian God exists.

Let's take a sub-claim of this and work with that:

P1a: Logical claims about God can only be true if the Christian God exists.

For that, I suppose, one could try arguing:

I don't think he'd object to the 4th lateran principle of any human statement about god always being more wrong than right. However, he'd probably maintain that a logically true claim about god must contain at least some element of (objective) truth as if it didn't, conceiving logical claims as similar fundamental to human reasoning as the Kantian Anschauungen are to having experiences would lead to a deus malignus.
Thus, the primacy of the truth of logical claims would commit him to hold that logically valid/invalid statements about god can never not capture anything true at all about god. Thus, the primacy of truth of logical claims would commit to asserting that it is at least not a fundamentally inadequate means of approaching god, which is something only necessarily true if their "compatibility" (at least in principle) has been established. It seems only god himself can be that sufficient guarantee, which would lead to P1a in the form of "can only be necessarily true if ..."

But I feel very uncomfortable going into waters I usually avoid for a reason.

Sorry if this wasn't clear, but by "logical claim" I mean something like, modus ponens is valid, A equals A is always true, etc. The point, for the presuppositionalist, is that even atheists and other non-Christians use logic, so they are themselves implicitly presupposing that the Christian God exists.

I don't think Van Til would object to someone making a priori claims about God. I think he primarily wants to assert that the tool she is using to make such claims--i.e. logical argumentation--includes as an implicit assumption of its justification that the Christian God grounds it. So any argument based on logic proving that God exists will be circular (although not viciously so). However, presumably logic can still be used to talk about God.

I understand that. The idea was to take a more narrow claim, show that it is valid and then argue that the broader claim cannot be wrong if the more narrow claim is true.

I think the problem here is that, at least as usually conceived, logical truths are true by virtue of their form, not their content. For example, "God is God" is true in standard first-order logic. But it is not true because of the semantics of "God," but because of the form of the statement. So why would it matter if the logical claims (i.e. the tautologies) are about God or about my socks?

True. Sorry, I got mixed up in two different thoughts there.

Basically, my initial feeling was that it's somewhat weird to put the argument in the form given by you above:

1) Logical claims can only be true if the Christian God exists.
2) Logical claims are true.
3) Therefore, the Christian God exists.

... when basically the entire music of it lies in 1)

So I think his reasoning - as well as its problems - could be presented more clearly if he presented more clearly an argument like this first.

1)
2)
3) Logical claims can only be true if the Christian God exists.

The second thought was that my initial criticism of

was somewhat misguided and I tried to make more sense of it - which I then implanted in the original argument for some reason.

I guess what he's trying to say here is that if we assume an independent logic, yet use this logic in our arguments and discussions about god, we are implicitly presupposing that that the use of arguments governed by logic is able to grasp something "real" or "true" about god. In that sense we would indeed take logical claims as being more fundamental than god and the only way to ensure that this does not end in a merely contingent adequacy between reasoning and whatever-truth-about-god-we-can-grasp would be to again pre-order god as the reason of a necessary adequacy. However, I don't see at all how this is to make sense to an atheist.

Right. Incidentally, you made comments at the start about what was "obvious". I think in what you have posted you have demonstrated that it is "obvious" that P1 (in particular the "if not god then not logic is true" direction) is not supported and is just effectively skepticism at the possibility of alternate explanations. And you have given a nonobvious alternate explanation.

Now I am still slightly confused about the canonical defense you presented of the former side (if god then logic claims are true). Previously when I have heard people talk about this, this is side is effectively trivial. Namely, God is effectively being defined such that this property falls either immediately or almost immediately. Namely, one of the properties of god is "that which imbues the universe with logical truths" in which case it falls immediately, or something like "god has properties of being logical and wanting his creation to be in his image" in which case it falls almost immediately by concluding that "therefore he would have created the universe with logical truths". Is there more to it than this? (In my original quote it seems like you are giving a similar defense until the bolded which seems out of place but no matter)

This is wrong because

This premise is wrong. Here is the same using knowledge.

1. If there is no god (most often the entity God, defined as the god of the Christian Bible, Yahweh), knowledge is not possible.
2. Knowledge is possible (or some other statement pertaining to logic or morality).
3. Therefore a god exists.

Notice how your argument does not hold water any more because it relied on your premise 2 not some sensible premise like "logic exists".

Matt Slick's reaches that point a little differently, his method is somewhat more satisfying, but still flawed. His way goes something like.

-Logical absolutes (A=A, etc) exist
-things that exist are either physical (matter, energy, etc) or conceptual (dependant on a mind)
-logical absolutes are not physical, so therefore conceptual
-if they were dependant on human minds, we would be able to change them, but we can't
-So, these concepts must have been the result of a greater mind, which is called god.

this path is more satisfying since it presents a reason to believe that it may be true, not just an assertion. However, one obvious flow is a possible false dichotomy in #2, physical/conceptual, which is not necessarily an exhaustive list of all existence.

I can't make sense of what your objection is here, so I'll just ignore it for now.

I don't disagree with any of this. Most of the argument (though not all of it), lies in justifying (1). This is what the rest of my posts were about.
None of it is supposed to make sense to the atheist as he is explicitly not putting forward a "neutral" (with respect to theism) justification of logic. The argument is that what goes on in apologetics is not proving the existence of god through natural human reason, but a comparison of atheist, Islamic, Buddhist, Christian and other worldviews. Van Til then argues for the thesis that when we compare them only the Christian one ends up being coherent as it is the only one that can justify logic (and morality and knowledge).