1. Jamin Asay’s A Theory of Truthmaking is one of the most important books on truthmaking, full of important ideas from beginning to end. There is much in the book I agree with. There are two main things in the book I disagree about. One is what Asay says about nominalism and truthmakers; the other is what he says about what he calls explanation-first truthmaking. In this brief note, I shall concentrate on the latter, and I shall explain why I disagree with him about it.

So, what is explanation-first truthmaking? It is the explanatory project of explaining why truths are true. This is in contrast with ontology-first truthmaking, the project defended by Asay, and which is the project of answering the fundamental ontological question “What is there?”. Asay thinks explanation-first truthmaking should be abandoned because (a) it is problematic and (b) it offers no benefits over those of ontology-first truthmaking. I shall argue that he has not established the problematicity of explanation-first truthmaking, and I shall also argue that explanation-first truthmaking offers benefits over ontology-first truthmaking.

2. What is Asay’s argument against explanation-first truthmaking? It takes the form of a dilemma: the goal of explanation-first truthmaking is either to provide some explanation or other of truth, or to provide a particular kind of explanation of truth. If the former, there are non-truthmaking explanations that do a better job; if the latter, the view takes on a theoretical task that is insufficiently motivated (p. 36).Footnote 1

I think the first horn of the dilemma does not represent a genuine alternative. The goal of explanation-first truthmaking is to provide a particular kind of explanation of truths. Nevertheless, I am not persuaded by the reasons Asay gives to conclude that there are non-truthmaking explanations that do a better job. Thus, I shall discuss the first horn of the dilemma before I discuss its second horn.

Asay notes that there are deductive-nomological explanations and since, according to him, “any explanation (…) is an explanation of truth” (p. 36), it follows that there are non-truthmaking explanations of truth. He gives the following example: suppose we want to explain the truth of <The liquid in the Yangtze boils at 100 degrees Celsius>. A deductive-nomological explanation would proceed as follows:

  1. (1)

    All samples of water at standard pressure boil at 100 degrees Celsius.

  2. (2)

    The liquid in the Yangtze is a sample of water at standard pressure.

    Therefore,

  3. (3)

    The liquid in the Yangtze boils at 100 degrees Celsius.

As Asay notes, this explanation invokes no truthmakers. And it explains a truth, namely the truth that the liquid in the Yangtze boils at 100 degrees Celsius. But it does not explain why that truth is a truth. Indeed, that explanation invokes truth no more than it invokes truthmakers. In other words, one needs to distinguish between explaining why p – explaining a truth – and explaining why <p> is true – explaining the truth of a truth.

There might be two replies to this point. One would be that explaining the truth of a truth is explaining semantically ascended truths, that is, truths like <<p> is true>. If so, explaining the truth of a truth is simply explaining a particular kind of truth, but then what is the explanation of the truth of non-semantically ascended truths? The answer is that explaining the truth of <p> and explaining <<p> is true> is the same. And this shows that if explaining truths like <<p> is true> requires giving truthmakers, as I think it does, not only semantically ascended truths have truthmakers. For when one explains <<p> is true> one gives the truthmaker of <p>.

Another reply might be that given the necessary equivalence between p and <p> is true, there is no distinction between explaining why p and explaining why <p> is true. But explanation is hyperintensional, as a reflection on explanation in mathematics and logic should show. Therefore, the necessary equivalence between p and <p> is true does not guarantee that explaining p is explaining the truth of p.

Thus, Asay’s assertion that “any explanation is an explanation of truth” is ambiguous. It can mean either that any explanation is an explanation of a truth, or that any explanation is an explanation of the truth of a truth. The former is correct: any explanation is an explanation of a truth—explanation is factive. The latter is incorrect: not every explanation is an explanation of the truth of a truth—some explanations are explanations of the boiling temperature of a certain liquid, for instance.

3. Asay knows that there is another way of attempting to explain truth without recourse to truthmakers. This is to appeal to instances of the following schema:

(B) For every true proposition: <p> is true because p.Footnote 2

Asay points out that explanations of truth in terms of truthmakers are less ontologically parsimonious than explanations of truth in terms of schema (B), for explanations of truth in terms of truthmakers typically invoke controversial or contentious entities, such as states of affairs, absences, tropes, totalities, etc., while explanations in terms of instances of schema (B) do not invoke any entities (p. 38).

But this does not mean that explanations of truth in terms of schema (B) are better than explanations of truth in terms of truthmakers. Take any instance of schema (B), say: <John is tall> is true because John is tall. Why is it the case that <John is tall> is true because John is tall? In other words, why are the instances of schema (B) explanatory instances? In yet other words, what grounds explanations based on schema (B)?

In general, explanations track relations between entities and it is such relations that ground explanations. In some cases, the explanans metaphysically determines the explanandum, in other cases the explanans causes the explanandum, and in other cases the relation between explanans and explanandum might be a different one, but explanation is always based on some relations between some entities.

Truthmaker explanations clearly satisfy this constraint on explanation: there are, say, two facts, one consisting of <John is tall>’s being true and another consisting of John’s being tall, and the latter fact determines the former fact.

Those wishing to use no more than schema (B) to explain truth, however, cannot account for what grounds those explanations. Indeed, one cannot say, for instance, that what grounds the explanatory character of the relevant instance of schema (B) is that John is tall determines the proposition that John is tall is true. This is ungrammatical, since “John is tall” cannot occupy a nominal position, and so it cannot be the subject of a sentence. And it is ungrammatical because it is non-sense: indeed, determination is a relation and so it needs entities to relate, but “John is tall” does not stand for any entity.

Some will think that one can ground explanation by iterating because clauses. For instance:

(C) <John is tall> is true because John is tall, because q.

“q” here is a sentence variable. Some might propose “q” to be replaced by “John is tall,” others by “<John is tall> is true,” others by “<John is tall> is true and John is tall,” and perhaps others might propose it to be replaced by yet other sentences. But it does not really matter for the current purposes what sentences might or might not replace it. For whatever sentence we replace “q” with, that sentence will not stand for an entity any more than “John is tall” does.

Thus, the proponents of schema (B) must take explanation of truth to be brute, ungrounded. But no explanation is brute (though grounding might be brute). Thus, although explanations of truth based on no more than schema (B) fare better than truthmaker explanations in terms of ontological costs, they fare worse in terms of accommodating the constraints on the concept of explanation, namely the constraint that it must be grounded in a relation between entities. And this is not a mere trade-off between two independent and equally important constraints. The constraint that explanation must be grounded in a relation between entities implies that any explanation of the truth of a truth must allow for some entities grounding that explanation. The idea that one should avoid explanations that require contentious entities serves only to select explanations with non-contentious ontological commitments over explanations with contentious ontological commitments, but not to select explanations with no ontological commitments over explanations with contentious ontological commitments.

4. Asay also argues that explanations in terms of truthmakers fail in cases of contrastive explanation. For instance, suppose that <Victor hiked up Victoria Peak during the Chung Yeung festival> is true (p. 38). Given that explanation can be contrastive, Asay argues that there are four candidates for explanation here, each one with its one distinctive explanation, and that the explanation of none of them involves a truthmaker for <Victor hiked up Victoria Peak during the Chung Yeung festival>:

Victor hiked up Victoria Peak during the Chung Yeung festival

Victor hiked up Victoria Peak during the Chung Yeung festival

Victor hiked up Victoria Peak during the Chung Yeung festival

Victor hiked up Victoria Peak during the Chung Yeung festival

I agree that explaining those four facts need not involve a truthmaker for <Victor hiked up Victoria Peak during the Chung Yeung festival>. But there is a fifth explanandum:

<Victor hiked up Victoria Peak during the Chung Yeung festival> is true.

Here the only possible contrast is with the proposition in question being false, rather than true. And the explanation of this explanandum requires truthmakers. To further elaborate on a point I made in Section 2, if the explanandum is the truth of a truth, the explanans is the truthmaker of the truth, and such a truth need not be semantically ascended: in this case the explanans of why <Victor hiked up Victoria Peak during the Chung Yeung festival> is true is the truthmaker of the truth that Victor hiked up Victoria Peak during the Chung Yeung festival.

5. Asay thinks the best response for the explanation-first theorist is to embrace the second horn of the dilemma and maintain that the job of truthmaker theory is to offer a particular kind of explanation that only truthmaker theory can offer (p. 39). I agree that this is the best response.

But this response is better than Asay thinks. Indeed, Asay thinks this response is problematic because this amounts to building into the conception of truthmaker theory “the task of explaining truth itself” (p. 41). But, Asay correctly points out, explaining what it is to be true and explaining what makes something true are two different things.

Now, it is precisely this distinction that allows the explanation-first theorist to maintain that there is a distinctive kind of explanation that only truthmaker theory can offer while at the same time avoiding Asay’s criticisms.

Indeed, explanation-first theorists do not need to compete with deflationists and they do not need to espouse any version of the correspondence theory of truth. All the explanation-first theorist needs to maintain is that since the truth of most truths depends on (extra-linguistic or extra-propositional) reality, there must be an ontological explanation of why (those) true propositions are true (where such ontological explanation makes reference to extra-linguistic or extra-propositional entities).

Thus, what the explanation-first theorist needs to say is simply that there is a particular form of explanation of truths, namely ontological explanation, which consists in providing the entities in virtue of which certain truths are true. This is a sort of explanation that only truthmaking theory can provide, since the entities in virtue of which any propositions are true are the truthmakers of those propositions. Such an explanation has nothing to do with explaining what it is to be true, and it is therefore as available to the deflationist as the ontology-first version of truthmaker theory is. Asay is right that ontology-first truthmaking is available to deflationists. I am claiming that explanation-first truthmaking is also available to them.

6. Asay has a final challenge for explanation-first truthmaking. This is that this view of truthmaking needs what he calls “Goldilocks” truthmakers—entities that are neither too big, nor too small, but just right (p. 43). Take the truth <Pandas exist> and a particular panda, Jia Jia. Most truthmaker theorists will say that Jia Jia is a truthmaker of <Pandas exist>. Asay objects that <Pandas exist> does not depend on Jia Jia, since it would have been true even if she had not existed. Yes, that is right. But <Pandas exist> depends on there being some panda or another, and if Jia Jia is one of them, Jia Jia makes that truth true (assuming pandas are essentially pandas and so Jia Jia necessitates the truth of <Pandas exist>).Footnote 3

But can Jia Jia really be a truthmaker of <Pandas exist>? Asay points out, correctly, that I and other explanation-first theorists would say that Jia Jia plus the piece of bamboo she is chewing is too big to make <Pandas exist> true. But then he notes that Jia Jia could lose a few parts (some teeth, say) without ceasing to be a truthmaker for <Pandas exist>, and he suggests that the fact that Jia Jia can lose such parts without effect on the truth of <Pandas exist> suggests that such parts are extraneous to the truthmaking of <Pandas exist>. Asay also points out that once a demand is made for truthmakers that are perfectly matched to their truths, a full account of the requisite matching is called for (p. 43).

Fair enough, such a general account of the requisite matching is indeed a challenge, but it is not a challenge I intend to meet here. But I can explain here in what sense the teeth or other parts of Jia Jia are extraneous to the truthmaking of <Pandas exist>. This is because any truthmaker for <Pandas exist> must be a panda, and Jia Jia would still be a panda without her teeth—thus they are extraneous to the truthmaking of <Pandas exist> in the sense that they are not necessary for Jia Jia to make true that proposition. But the teeth and other parts of Jia Jia are not extraneous to the actual truthmaking of <Pandas exist>, since the truthmaker is Jia Jia, and Jia Jia includes all her parts and therefore all her teeth.

7. Thus explanation-first truthmaking purports to give a particular kind of explanation of the truth of most true propositions, namely an ontological explanation. And such a kind of explanation is not problematic, at least not in the sense in which Asay thinks it is—or so I have argued. Thus, I conclude that no compelling reasons have been given to abandon explanation-first truthmaking. Furthermore, an ontological explanation of the truth of true propositions is a benefit offered by explanation-first truthmaking but not by ontology-first truthmaking. Thus, there are reasons to adopt explanation-first truthmaking.