TAG and Epistemic Certainty

A commenter asks why I don’t endorse the claim that the transcendental argument for the existence of God (TAG) gives us epistemic certainty (which I take to mean that the argument delivers a conclusion that has maximal epistemic warrant and could not be rationally doubted). After all, if TAG proves Christian theism “by the impossibility of the contrary”, as many of its advocates have claimed, wouldn’t it follow that TAG’s conclusion is epistemically certain?

As a preliminary matter it should be pointed out that the epistemic status of normal Christian belief doesn’t depend on TAG. We should distinguish between the means by which we know that Christianity is true (in my view, the external testimony of Scripture coupled with the internal testimony of the Holy Spirit, as per chapter 1 of the WCF) and the means by which we show (to others) that Christianity is true. TAG serves the latter purpose, not the former. The “full persuasion and assurance” of our faith has nothing to do with TAG — and it’s a good thing too, since TAG was all but unknown before the 20th century. I make this point only because some Van Tilians have given the impression (usually inadvertently) that the certainty of our knowledge of God and the gospel somehow hangs on the cogency of TAG.

Turning now to the claim that TAG delivers epistemic certainty about Christian theism, I offer three reasons to question that claim. First, as I’ve suggested elsewhere, TAG is best understood not as one unique argument but as a family of arguments with a common theme or goal. So when someone claims that TAG delivers epistemic certainty, my first response is to ask, “Which TAG? Which formulation of TAG?”

Secondly, I’m unaware of any formulation of TAG that proves specifically Christian theism; thus it’s hard to evaluate the claim that TAG proves Christian theism with epistemic certainty.

Thirdly — and this is the most important point — we need to be clear on what it would take for any argument to deliver its conclusion with epistemic certainty. The strongest type of argument is a deductive argument, in which the conclusion follows necessarily from the premises: if the premises are true then the conclusion must be true. But even in a deductive argument, the conclusion cannot enjoy greater epistemic warrant than the premises (at least in the case where the warrant for the conclusion is taken to derive from the argument itself). What this means is that TAG can deliver epistemic certainty only if all of its premises are epistemically certain.

The problem, however, is that the transcendental premise of TAG (that God’s existence is a metaphysically necessary condition of some essential feature of human thought or experience) isn’t self-evidently true and it requires quite sophisticated argumentation in its support. But the more sophisticated an argument, the more room there is for doubt about its cogency — even the doubt that we have adequately understood all of the premises and inferences — and the less plausible it becomes to maintain that all of the premises of the argument are epistemically certain. In the case of TAG, it stretches credibility to claim that every premise of the argument (and of the sub-arguments given in support of the top-level premises) is epistemically certain.

None of this implies that TAG is a failure or a pipe dream. But Van Tilians do themselves no favors by needlessly overstating what TAG can (or could) accomplish. To my mind, the relevant question here isn’t “Does TAG deliver epistemic certainty?” but rather “Why think that TAG should deliver epistemic certainty?”

11 Responses to TAG and Epistemic Certainty

  1. Hi James,
    I am wondering about this statement you made,

    “But even in a deductive argument, the conclusion cannot enjoy greater epistemic warrant than the premises (at least in the case where the warrant for the conclusion is taken to derive from the argument itself). What this means is that TAG can deliver epistemic certainty only if all of its premises are epistemically certain.”

    Are you saying that the conclusion of a deductive argument cannot be certain unless we know all the premises with certainty? If so would this not commit us to the idea that there cannot be any successful deductive arguments? The reason would be that we cannot know any premise with certainty because it is always possible for us to be mistaken, since we are fallible.

    I would really appreciate your clarity on this. I am not as learned as you so forgive me if this sounds like a silly question.
    Blessings,
    Michael

    • Hi Michael,

      Not a silly question at all. But what do you mean by “successful deductive argument”? What counts as successful here? If ‘successful’ simply means sound (i.e., logically valid with true premises) then whether or not an argument’s premises are epistemically certain is neither here nor there; they only need to be true. Philosophers often consider a deductive argument to be a good one if it is logically valid and every premise is more plausible than its denial; again, no need for epistemic certainty on that view of things.

      Besides, I think it’s misleading to speak of a successful deductive argument per se for the same reason it’s misleading to speak of a successful medical procedure per se. A medical procedure can be successful in some cases and not in others; it might even be successful in most cases. But it depends on the circumstances, the people involved, etc. Similar considerations apply to arguments. If D is a valid deductive argument for P, and a person S who has doubts about P recognizes that D provides good support for P, then I’d say D can be considered successful in that case. But none of this requires that the premises of D be indubitable or incorrigible.

      • Hi James,
        Thanks for your reply and for saying it is not a silly question:) It’s been a while since I’ve been in a philosophy class so I am a bit rusty.

        I was sloppy in my language in speaking of a “successful deductive argument.” What I should have said was a “sound deductive argument.” Thanks for forcing me to clarify.

        I guess my confusion is in saying that TAG can render an epistemically certain conclusion only if all of it’s premises are epistemically certain. The reason why I am confused is that it doesn’t seem to be the case with deductive arguments (though I may be confused on this too). Is it not the case that if the premises are true in a deductive argument than the conclusion is certain? And we can know this with epistemic certainty? For example consider the following famous argument

        1. All men are mortal (true but not epistemically certain)

        2. Socrates is a man (true but not epistemically certain)

        C. Socrates is mortal (true and epistemically certain)

        I was taught in my beginning logic class (many moons ago) that if these premises are true, then the conclusion follows with certainty. Now maybe this is wrong, but if it is not wrong, then it would seem to be a case where true but epistemically uncertain premises yield a true and epistemically certain conclusion.

        If this is correct, then it would seem that TAG could have true premises that are not epistemically certain, but yield a true conclusion that is epistemically certain.

        James am I out to lunch on this one? I would appreciate your help in working this out in my head. Thanks for your time,
        Blessings,
        Michael

        • Michael,

          I was taught in my beginning logic class (many moons ago) that if these premises are true, then the conclusion follows with certainty.

          Okay, I see where the confusion has arisen. Your teacher was speaking somewhat loosely here. What is certain is that the conclusion follows from the premises. This is exactly what I meant when I said above that the conclusion follows necessarily from the premises. But the point remains that the certainty that the conclusion is true cannot be greater than the certainty that the premises are true.

          Consider this example:

          (1) There is an even number of books in the seminary library.
          (2) If there is an even number of books in the seminary library then all the books can be arranged into pairs.
          (3) All the books in the seminary library can be arranged into pairs.

          This is a valid deductive argument; it may even be a sound argument. But clearly the conclusion is not certainly true! The reason is obvious: while premise (2) may be certainly true (since it expresses an a priori mathematical truth), premise (1) is not — even if it is indeed true.

          Does that help to clarify matters?

      • Hi James,
        Thanks for your post. Things are coming back to me from my logic class. You are right and it clears things up for me.
        Blessings,
        Michael

  2. James,

    Thanks for clearing the air to help me better understand the nature of the situation. Two things, I see in your response to David Reiter you quote Van Til about the proof of Christianity. Is there a distinction here, do you agree with Van Til we do have proof but also do not believe we have epistemic certainty?

    Second, regarding the impossibility of the contrary. Are you still agreeing that TAG does remove all the other worldviews but even after that we don’t have certainty. Or does it still leave other worldviews intact so other worldviews are possible? Just want some clarification. Thanks

    • It depends what is meant by ‘proof’. I think I understand what Van Til meant by ‘proof’ in that quotation, and if I’m right then no one has (yet) offered a transcendental argument for Christian theism that constitutes a proof in that sense. But I don’t think that’s a problem, because I don’t believe we have to set the bar for successful argumentation that high.

      As for the second point, I’d say that even if there were a version of TAG that successfully refuted every non-Christian worldview, it still wouldn’t furnish us with epistemic certainty for the third reason given in my post (above). But once again, I don’t consider that a big deal, just as I don’t consider it a big deal that the Space Shuttle can’t travel at the speed of light.

  3. Dr. Anderson,

    You asked, “Why think that TAG should deliver epistemic certainty?”

    If there is epistemic certainty with respect to Christianity then one would expect a corresponding apologetic. Perhaps there are candidates other than TAG, but I do not know of any. (Perhaps there is no epistemic certainty with respect to Christianity, but I would question whether or not this claim is faithful to Scripture.)

    You wrote, “We should distinguish between the means by which we know that Christianity is true (in my view, the external testimony of Scripture coupled with the internal testimony of the Holy Spirit, as per chapter 1 of the WCF) and the means by which we show (to others) that Christianity is true.” You then explained, “TAG serves the latter purpose, not the former.” However, you then wrote, “I’m unaware of any formulation of TAG that proves specifically Christian theism; thus it’s hard to evaluate the claim that TAG proves Christian theism with epistemic certainty.”

    I do not think that you mean to imply that TAG is “*the* means” by which we show to others that Christianity is true, but one of “the *means*.” Here TAG is understood not as that argument which is pertinent to Christian belief but rather that argument which is pertinent to theistic belief as an element of Christian belief (though not *necessarily* an element of Christian belief). Have I understood you correctly?

    Though we may do well in distinguishing between the means by which we know and the means by which we show we have not thereby excluded TAG or any other argument from the former category. Here you excluded TAG from the former category without argument. You may have good reasons for doing so that were not stated. Additionally I would question in what respect the unbeliever being shown that Christianity is true does not at the same time or subsequently contribute to or even constitute the unbeliever’s having known that Christianity is true especially in those cases where the unbeliever comes to agree with the conclusion of the argument(s) or converts to Christianity.

    When you state that you do not endorse the claim that TAG gives us epistemic certainty, I would follow you in asking, “Which TAG? Which formulation of TAG?” Perhaps you mean to say that *no* formulation of TAG gives us epistemic certainty by virtue of the other considerations you have brought out. I wonder though if epistemic certainty is really possible by virtue of the external testimony of Scripture coupled with the internal testimony of the Holy Spirit, and I do so by virtue of the same kinds of considerations you bring to bear upon TAG.

    I have some problems with thinking that we do not have certainty about our faith (I am not attributing this view to you), and I have some problems with thinking that even given such certainty we are unable to show others this feature of our faith through apologetic argument (I am not even necessarily attributing this view to you). It would be wrong for me to conclude from what I am questioning here that TAG must therefore provide certainty, but I think there are real problems with denying that it does. Hopefully this serves to begin to answer your question about why we should think TAG provides epistemic certainty, though I would dare say that most would simply appeal to some of TAG’s major proponents like Van Til and Bahnsen, who certainly held that TAG provides epistemic certainty. This is not to say that they were correct to hold this, but people tend to turn to them for knowledge concerning the claims of TAG.

    What is skepticism with respect to Christianity if not the affirmation of the possibility that God does not exist and has not spoken?

    • Thanks for dropping by, Chris. You’ll forgive me if I’m selective in my response to your comments.

      If there is epistemic certainty with respect to Christianity then one would expect a corresponding apologetic. Perhaps there are candidates other than TAG, but I do not know of any. (Perhaps there is no epistemic certainty with respect to Christianity, but I would question whether or not this claim is faithful to Scripture.)

      Well, I question both premises of your argument here. A case could be made, perhaps from Romans 1, that the existence of God is (or can be) epistemically certain. But I don’t see any biblical or theological basis for the claim that Christianity (however you wish to delineate that) is (or can be) epistemically certain. Given the confessional Reformed model for knowledge of the Christian gospel, it’s hard to see how it could be — and apparently you agree with me on this point. Neither do I see any good reason to accept the conditional claim above (regarding our expectations for apologetics). Perhaps you can spell out the basis for accepting this claim.

      You say you question whether the claim that Christianity is not epistemically certain is “faithful to Scripture”. Fine — just show me the money. What is the exegetical argument here? If you’ve written on this before, feel free to post a link.

      As for your question about TAG: I don’t rule out the feasibility of a transcendental argument for specifically Christian theism. Indeed, that’s an ongoing project of mine. But we shouldn’t understate the challenges of that project, nor should we issue checks that we can’t yet fund.

      I will gladly concede that in principle TAG could be a means by which Christianity is known to be true. But clearly it could not be the normal means by which it is known. I would go further and say that a knowledge of Christianity based solely on TAG would be quite dysfunctional. Far more plausibly, TAG could be a means by which a properly basic belief in Christianity is maintained and fortified in the face of objections and alleged counterevidence.

      I wonder though if epistemic certainty is really possible by virtue of the external testimony of Scripture coupled with the internal testimony of the Holy Spirit, and I do so by virtue of the same kinds of considerations you bring to bear upon TAG.

      Good — it seems we have some substantial agreement here. :) But then what’s your point? If anything, you’ve conceded that epistemic certainty isn’t so important after all.

      I have some problems with thinking that we do not have certainty about our faith (I am not attributing this view to you), and I have some problems with thinking that even given such certainty we are unable to show others this feature of our faith through apologetic argument (I am not even necessarily attributing this view to you).

      I agree that we have (or can have) certainty about our faith. But the question is: what kind of certainty? We can’t just assume that when the Bible speaks about ‘certainty’ or ‘assurance’ that such concepts map neatly onto a particular philosophical understanding of certainty.

      You say you have ‘problems’ with certain (!) claims made about our faith and our apologetic. Okay — but why not tell us what those problems are exactly? Otherwise you’ve given me no reason to change my current views; you’ve only given me some relatively uninteresting autobiographical facts. Again, if you’ve explained this elsewhere, just post the links.

      What is skepticism with respect to Christianity if not the affirmation of the possibility that God does not exist and has not spoken?

      Skepticism with respect to Christianity is doubt about the claims of Christianity, including the claim that the Christian faith can be known with confidence to be true. But repudiating such skepticism is quite compatible with the idea that some of the claims of Christianity are not epistemically certain (in the technical sense). Only one who bought into a discredited Cartesianism would think otherwise. Descartes is not the lens through we are to view the epistemological claims of Scripture.

  4. Hegel,

    An argument can be sound, and so the conclusion impossible to be false, even if you don’t have epistemic certainty regarding the premises or the conclusion. Consider:

    [1] All totalitarians have an I.Q. of over 150.
    [2] Hitler was a totalitarian.
    [3] Ergo, Hitler had an I.Q. over 150.

    No, *if* the premises are true, then the conclusion cannot be false. But, we may not be said to be warranted in having epistemic certainty regarding [1], and thus we cannot be epistemically certain of the conclusion, *even though* [1] may be true and so the conclusion must be true.

    Now, you may *believe* [1], and you may even be said to *know* that [1] is true, and so you could be said to *know* the conclusion, but could you be said to know it with epistemic certainty (i.e., that no rational person could possibly doubt the premises)? I don’t think so.

    Consider the what the argument for [1] would look like? You’d have to know all the supporting premises with epistemic certainty too, and that seems doubtful. Likewise with TAG. A fully developed TAG (i.e., not one that says, “If logic then God, not God, therefore not logic, FAIL!) is going to require a lot of supporting arguments. And, *if* those premises are all true, and if the form is valid, then the conclusion cannot be false. But whether they’re true or not depends on the way things are. It’s a metaphysical relationship between premises and the world. But to *know* the premises with epistemic certainty is an *epistemological* thesis, and it seems doubtful that you could know each and every individual premise of the fully-spelled-out instance of TAG. Suppose it requires 30 premises? Suppose more than a few of those premises touch on matters you know next to nothing about (say, some highly complex, sophisticated point in the philosophy of logic), how could you reasonably take yourself to know the conclusion with complete epistemic certainty?

  5. Another problem, I guess, is that those who claim TAG gives them epistemic certainty also claim it is non-deductive. But they have not produced the “transcendental logic” by which the claim “God exists” supposedly follows from the premises. They have produced no rules of inference, they have not proven their system at the meta-level, neither have they shown whether it is a decidable system of logic. They could not take a “bad” TAG argument and a “good” TAG argument and show us how, where, and why one went wrong when the other didn’t. Given this situation, it would seem atheists and other non-Christians can easily resist the argument, they “have an excuse,” and the Christians who give the non-deductive TAG cannot be epistemically certain that the conclusion follows from the premises. It would seem, then, that they *begin with* the idea that Christianity is and can be shown to be epistemically certain, and they have *faith* that the “argument” shows it even though they have no idea of the rules and system of the logic they are using to derive the conclusion.