Undermining truthmaker theory

Abstract

Truthmaker theorists hold that there is a metaphysically explanatory relation that holds between true claims and what exists. While some critics (e.g. Merricks 2007) try to provide counterexamples to truthmaker theory, that response quickly leads to a dialectical standoff. The aim of this paper is to move beyond that standoff by attempting to undermine some standard arguments for truthmaker theory. Using realism about truth and a more pragmatic account of explanation, I show how some of those arguments can be undermined.

“[The question “What makes a proposition true or false?”] is a loose way of expressing the question, what are the condition in which $p$ (is true) and what are the conditions in which not-$p$?” Ayer (1946, p. 90)

Elephants roam the African plains. That’s a fact. And, as many know, it is true that elephants roam the African plains. It is almost a banality that elephants roam the African plains if and only if it is true that elephants roam the African plains. Recently, some philosophers—“truthmaker theorists”—have posited a more substantive connection between this fact and this true claim. These philosophers have claimed that ‘elephants roam the African plains’ is made true by the fact that elephants roam the African plains, and not conversely. That is, these philosophers postulate a metaphysically explanatory relation—the “makes true” relation—between what exists—the “truthmakers”—and what is true.

For those skeptical of such a view, a natural somewhat venerable reaction is to provide counterexamples to it—to provide cases of truth without a truthmaker. This is the strategy of Merrick’s (2007)—one of the most prominent criticism of truthmaker theory. But this strategy quickly becomes an unsatisfying dialectical standoff with intuitions getting pitted against one another.

The aim of this paper is to move discussion beyond such a dialectical standoff. Instead of arguing that truthmaker theory is false, I’ll attempt to undermine some standard reasons for accepting truthmaker theory. These reasons are (broadly) abductive in nature, and I intend to show that, given certain plausible theses about truth and a more pragmatic theory of explanation, these reasons can be undermined. The upshot is that for truthmaker theorists to continue to postulate this relation—and use it in doing substantive metaphysics—they ought to shore up the foundations of their theory.

I begin, in Sect. 1, by characterizing truthmaker theory. After distinguishing between different principles various truthmaker theorists accept, I state that the central claim of truthmaker theory is that there is a metaphysically explanatory, asymmetric “makes true” relation. In the next two sections, I consider arguments that there is such a relation. In Sect. 2, I consider the view that without truthmaker theory we cannot account for the truth of certain claims or our knowledge of certain claims. I show how, given what I call “realism about truth,” both of these arguments fail. In Sect. 3, I consider the strongest argument for truthmaker theory: Rodriguez-Pereyra’s argument from an apparent asymmetry in our thinking about truth. After adopting a more pragmatic theory of explanation, I show how this argument fails as well. I conclude by comparing my account with other critics of truthmaker theory.

1 Truthmaker theory

While some philosophers identify as truthmaker theorists, there is much disagreement between them on the nature of truthmaker theory. First, they disagree about what entities are the “truthmakers.” The term ‘truthmaker’ is simply a label for whatever it is that makes claims true. Disagreement about truthmakers is thus disagreement about what sorts of entities “make true,” i.e. stand in the “makes true” relation. Perhaps the most popular suggestion is Armstrong’s states of affairs (1997, pp. 1–11, 113–139). But there are other candidates such as tropes or particulars (Mulligan et al. 1984, Sect. 2). Schaffer (2010) and Cameron (2008b) have argued that the world itself is the truthmaker for (certain) claims (cf. Smith 1999, p. 285). Lowe (2006, pp. 184–188) has suggested that the category of truthmaker is not unified, so that different sorts of things are truthmakers.

Second, truthmaker theorists disagree about what are the truth bearers, i.e. what is it that is made true. Some suggestions are abstract propositions (Merricks 2007), concrete propositions (Armstrong 2004, pp. 12–16), “judgments,” i.e. tropes (Smith 1999, pp. 274–275), and sentences. Lowe (2006, p. 178) has suggested that the category of truth bearers is not unified—so that different sorts of things are truth bearers—but that propositions are the primary truth bearers.

Third, and perhaps more importantly, truthmaker theorists disagree about the scope of what is made true; they disagree about whether all propositions are made true, or, if not, which ones are. Armstrong (2004, pp. 5–7) and Cameron (2008a, pp. 107–108, 2008b) advocate “maximalism”—the thesis that every truth has a truthmaker. Rodriguez-Pereyra rejects maximalism, holding instead that there are truthmakers for “an important class of true synthetic propositions, which includes many inessential predications, negative existentials and necessary truths” (2006b, p. 191; cf. Smith 1999, pp. 284–285).

Finally, truthmaker theorists disagree about what principles govern the makes true relation. It is generally accepted that the relation is “cross-categorical,” meaning that it can hold between (i) propositions and (ii) propositions and things that are not propositions (Rodriguez-Pereyra 2006b, p. 189; cf. Bigelow 1988, p. 126). The makes true relation is thus not classical entailment, which holds only between propositions.¹

There is only one uncontroversial principle:

Factive: Necessarily, if S makes $p$ true, then $p$ is true.

Beyond Factive, truthmaker theorists disagree. One widely accepted principle is:

Necessitation: Necessarily, if S makes $p$ true, then necessarily if S exists, then $p$ is true. (Fox 1987, p. 189; Bigelow 1988, p. 126; Restall 1996; Smith 1999, p. 276; Lewis 2001, Armstrong 2004, p. 6; Rodriguez-Pereyra 2005; Jacobs 2010, p. 228; Pruss 2011, p. 18; Cameron 2008a, pp. 109–113)

There is no non-question begging argument for Necessitation.² But it does follow from Factive and:

No Preemption: Necessarily, if S makes $p$ true, then necessarily if S exists, then S makes $p$ true.³

But since No Preemption is a stronger principle than Necessitation, it is not likely to convince those who reject Necessitation.⁴

Beyond Necessitation, there are other principles, each of which has defenders and detractors:

• Entailment: Necessarily, if S makes $p$ true, and $p$ entails $q$, then S makes $q$ true. (See Armstrong 2004, pp. 10–12; Cameron 2008b, p. 411; Mulligan et al. 1984, p. 316; Lopez de Sa 2009; cf. Rodriguez-Pereyra 2006a)

• Disjunction: Necessarily, if S makes $p$ true, then S makes p or q true. (See Mulligan et al. 1984, p. 316; Merricks 2007, p. 168 fn.13)

• Conjunction: Necessarily, if S makes $p \& q$ true, then S makes $p$ true and S makes $q$ true. (See Armstrong 2004, p. 11; Restall 1996, p. 334; and Mulligan et al. 1984, p. 316; Pruss 2011, pp. 71ff; cf. Rodriguez-Pereyra 2006a)

• Aboutness: Necessarily, if S makes $p$ true, then $p$ is about S. (See Smith 1999, pp. 279ff.; Merricks 2007, pp. 32–34; cf. Schaffer 2008, pp. 302–307)

• Fundamental: Necessarily, if S makes $p$ true, then S is a fundamental entity. (See Schaffer 2008, pp. 311–312, 2010)

So truthmaker theorists disagree over (i) what the truthmakers are, (ii) what the truth bearers are, (iii) the scope of truthmaking, and (iv) the principles governing the makes true relation. This brief overview of truthmaker theory thus shows that there are many versions of truthmaker theory. One might think that there is no thesis that ties these theorists together. But that is mistaken. For there are some theses that do tie them together. All truthmaker theorists accept:

• (1) Some true propositions are made true by existing things.

(1) refers to the “makes true” relation and there is some debate regarding principles governing it. Nevertheless, all truthmaker theorists accept:

• (2) S makes true $p$ if and only if $p$ is true in virtue of S’s existence.

From (1) and (2) it follows that:

• (3) Some true propositions are true in virtue of the existence of some entity (or entities).

‘...true in virtue of...’ is to be understood so that:

• (4) There is a metaphysically explanatory, asymmetric relation that holds between truths and entities, namely the makes true relation.⁵

I take (1)–(4) to be the core theses of truthmaker theory. So I will characterize truthmaker theory as holding these theses.

One might object that this characterization is not sufficiently discriminating. For example, Trenton Merricks is a critic of truthmaker theorists such as Armstrong (see his 2007). Nevertheless, he would accept (1)–(4). For instance, he writes: “...I think that truths like that Merricks exists, truths that really are entirely about what exists, must have truthmakers” (Merricks 2007, p. 168; cf. Merricks 2007, p. xvii, 2011, pp. 215, 230). But this is primarily a verbal issue. Merricks uses ‘truthmaker theory’ to refer to what I’ve called above ‘maximalism’—the view that every truth has a truthmaker (cf. (Merricks 2007, p. xiii)). In this more narrow sense, Merricks is not a truthmaker theorist. But I’m here concerned with the wider—and arguably more important—sense of truthmaker theory as accepting (1)–(4). In this wider sense, Merricks is a truthmaker theorist.⁶

2 Arguments from truth and knowledge

In the remainder of this paper, I consider some arguments for truthmaker theory. I will not consider all arguments, as I have nothing to add to some standard responses to some arguments.⁷ Instead, I’ll focus on what I take to be some of the more powerful, and to some degree, less-discussed arguments. In this section, I’ll focus on two arguments that are more abductive in nature: without truthmaker theory we would be unable to explain the relation between truth and reality and our knowledge of reality.

Key to my defense will be what I call realism about truth.⁸ Realism about truth is the thesis that, necessarily, a proposition is true just in case what the proposition says to be the case actually is the case. In other words, a proposition is true just in case the world is as the proposition says it is.⁹ According to realism about truth, the proposition that ‘there is a coffee mug on my desk’ is true just in case the world is such that there is a coffee mug on my desk, and not (e.g.) whether or not the proposition coheres with other beliefs or would be adopted by an ideal scientific community. With realism about truth in hand, we can now turn to objections to the view that there is no metaphysically explanatory asymmetric relation that holds between truths and entities.

2.1 Truths that “Float Free”?

The first objection is that this view does not allow for an adequate connection between truth and the world. Thus, Bennett (2011, p. 187) writes: “Truth—or, better, truth-value—depends on being. I find this intuition very plausible. After all, what’s the alternative? That truth floats free of being? Surely that’s the kind of thought that leads to berets, and a job in a bad Comp Lit department.” Robin Stenwall (2010, p. 212) writes: “Whenever a proposition is true, it is not just true. There is not the Earth, and by an outstanding coincidence it happens to be true that the Earth exists. We should rather say that the proposition that the Earth exists is true in virtue of, or because of the Earth...” And Rodriguez-Pereyra (2005, p. 22) writes that the view that rejects truthmaker theory “is not good. Surely, the rose’s being red and the truth of the proposition that the rose is red are connected in some way. The proposition that the rose is red is about the color of the rose, and so if it is true, it must have to do with the rose’s being red.”

But what exactly is the problem here? First, given realism about truth, there is a connection between the rose’s being red and the truth of the proposition that the rose is red—a necessary connection! Necessarily, the proposition that the rose is red is true if and only if the rose is red. Necessarily, the proposition that the Earth exists is true if and only if the Earth exists. Thus, it is not an outstanding coincidence between what is true and how the world is. Second, it is not clear what Rodriguez-Pereyra means by “have to do with” in the above passage. The proposition that the rose is red is about the rose’s being red, and so in that fairly obvious sense it “has to do with” the rose’s being red.¹⁰

Thus, we do not need the makes true relation to account for how it is that truths do not float free. Realism about truth is sufficient. Now the truthmaker theorist might respond that realism about truth does not explain how truths do not float free of the world—it merely has a principle that entails that they do not. (There’s a more sophisticated version of this objection I’ll consider in Sect. 3. I focus here on a cruder one that I’ve heard expressed before.) But the truthmaker theorist should be dissuaded of this response. For this puts such a high bar for explanation that not even truthmaker theory will clear it. For truthmaker theory “merely has” principles—like Factive and Necessitation—that entail that truths will not float free. (The truthmaker theorist might press that these are “explanatory” principles, but as David Lewis reminds us (1983: 40), just because you give a title to something does not mean it deserves it.) So if realism about truth does not explain how truths do not float free, then neither will truthmaker theory and the end result will be the same: this is no reason for accepting truthmaker theory.

2.2 Epistemological problems?

One might think that there is an epistemological problem with rejecting truthmaker theory. Sorensen (2001, p. 171) expresses this problem as follows:

A violation of the truthmaker principle¹¹ is serious business. In addition to grating against metaphysical principles, the violations are epistemologically grave. The phenomenalist claims to know that a chair continues to exist while unobserved. If there is no truthmaker for the proposition, how can the knower discriminate between the proposition being true and being false? ...I know contingent truths in virtue of connections between me and bits of reality that make those statements true. Absolute skepticism follows when there is no connection between a contingent statement and a state of affairs upon which it is contingent.

There are at least two ways to understand this objection. First, the objection might be that without truthmaker theory one is unable to discriminate between the sorts of circumstances under which something will be true or false. (This is suggested by Sorensen’s “...how can the knower discriminate between the proposition being true and being false?”) But this is clearly wrong. I can discriminate between the sorts of circumstances in which it is true that there is a book on my desk and when it is false—it is true just in case there is a book on my desk, and false otherwise. One does not need to appeal to truthmaker theory.

There is a second, more promising, way of understanding this objection. On this way, in order for me to know some contingent proposition—e.g. that there is a book on my desk—$I$ must be able to discriminate whether or not that proposition actually is true or false. (This is suggested by Sorensen’s “$I$ know contingent truths in virtue of connections between me and bits of reality that make those statements true.”) But if truthmaker theory is false, then there won’t be a truthmaker for the proposition that there is a book on my desk, and so—it might seem—I will not be able to discriminate when that proposition is true or false. So $I$ do not know some contingent proposition, e.g., that there is a book on my desk.

But this is unpersuasive as well. First, note that in denying that there is a truthmaker for ‘there is a book on my desk,’ I need not deny that there is a book on my desk. (I am not denying that there are “bits of reality.”) I am merely denying that there exists an entity that stands in a metaphysically explanatory asymmetric relation to the true proposition ‘there is a book on my desk.’ But denying that is consistent with holding that there is a book on my desk.

Once this is recognized, it seems obvious that I am able to discriminate when it is true or false that there is a book on my desk. Since my view of the desk is unobstructed, if there weren’t a book on my desk, I would be able to tell that it is false that there’s a book on the desk; and if there were a book on my desk, I would be able to tell that it is true that there’s a book on my desk. Of course, there are certain propositions for which I am unable to discriminate whether it is true or not, for instance, there will be Moon prisons in 2500 or Columbus had over five hundred ounces of ale on his first voyage in 1492. But this has nothing to do with truthmaker theory. Truthmaker theorists have the same problems as I for determining whether these propositions are true or not.

As I see it, the mistake behind both of these objections is the same: if one rejects truthmaker theory, then there will be no connection between the truth of a contingent proposition $p$ and the fact that $p$. But, as we’ve seen, this is not the case. Given realism about truth, there is a very important connection—a necessary connection. Thus, I do not think that worries concerning “free floating” truth or knowledge give us reason for accepting truthmaker theory.

But those sympathetic to truthmaker theory might feel that there’s more to be said. I’ve undercut these arguments by urging a modal connection between truth and the world. But surely, the thought might go, we want some of sort of explanatory connection between truth and the world. The shift to a modal connection will not accommodate this. The idea that there is some sort of explanatory connection is, I think, the strongest reason for adopting truthmaker theory. I consider it at length in the next section.

3 Truthmaker theory and explanation

One might think there needs to be more than a modal connection between truth and ontology, but also an explanatory connection. One might express this thought by claiming that truth is “grounded” and this gives us reason for accepting truthmaker theory. The chief proponent of this kind of argument is Rodriguez-Pereyra (see his 2005, pp. 22–31; 2006b, p. 191; 2009). In this section, I criticize Rodriguez-Pereyra’s argument. As I see it, the idea that truth is “grounded” may mean either a plausible and widely held thesis (what I call below the ‘Asymmetry Thesis’) or a much more controversial thesis that needs argumentation to support it (what I call below the ‘Grounding Thesis’). The argument for truthmaker theory begins with the Asymmetry Thesis, then infers the Grounding Thesis, and then concludes that truthmaker theory is true. Here I resist the first inference. I argue that one can account for why the Asymmetry Thesis is plausible without appealing to the Grounding Thesis. Consequently, truthmaker theorists need to provide independent reason for showing the Grounding Thesis true or show why my account is less plausible than an appeal to the Grounding Thesis.

3.1 An argument for Truthmaker theory

Aristotle wrote:

...if there is a man, the statement whereby we say that there is a man is true, and reciprocally—since if the statement whereby we say that there is a man is true, there is a man. And whereas the true statement is in no way the cause of the actual thing’s existence, the actual thing does seem in some way the cause of the statement’s being true: it is because the actual thing exists or does not that the statement is called true or false...¹²

Perhaps we can restate Aristotle’s point like this. It is plausible to think:

• (5) It is true that there is man because there is a man,

but not at all plausible to think:

• (6) There is a man because it is true that there is a man.

This holds for many truths, making it plausible to generalize from this case (and others) to say that plausibly there are some¹³ true instances of the schema:

• (7) $p$ is true because $p$,

and plausibly no true instances of the schema,

• (8) $p$ because it is true that $p$,

where ‘$p$’ ranges over propositions.¹⁴ ${}^{,}$ ¹⁵ Let’s call the ‘Asymmetry Thesis’ the thesis that there are some plausible instances of (7) and no plausible instances of (8).

The Asymmetry Thesis is plausible. But its plausibility raises a question: why are we inclined to think there is such an asymmetry? Truthmaker theorists are likely to think that an account of the plausibility of the Asymmetry Thesis should be metaphysical in some important sense. Exactly what important sense is unclear, and may be left at least initially at an intuitive level. Nevertheless, they are likely to claim that the Asymmetry Thesis supports the idea that truth is “grounded” or “determined by reality” or “holds in virtue of reality,” where these phrases are understood in some metaphysical sense and not in a semantic, epistemic, or logical sense (cf. Rodriguez-Pereyra 2005, pp. 20–21; 2006b, p. 191). So, they will claim that the reason why it is plausible to think that

• (5) It is true that there is a man because there is a man,

but not

• (6) There is a man because it is true that there is man

is because it is plausible to think that the truth of ‘there is a man’ is grounded in there being a man, and not vice versa. Again, exactly what it means to say that truth is “grounded” or “determined by reality” may be left at an intuitive level. Keeping this in mind, let us call the ‘Grounding Thesis’ the thesis that truth is “grounded” in reality, where ‘grounded’ is understood in some metaphysical sense that accounts for the plausibility of the Asymmetry Thesis.¹⁶

From the Grounding Thesis, one can argue to truthmaker theory. The most promising way to do this is to follow Rodriguez-Pereyra, who argues that the best way to understand grounding is as a relation (2005, pp. 27–31). But relations link entities. If a proposition is true, then it stands in the grounding relation. If it stands in the grounding relation, then there must be at least one thing to which that true proposition stands in the grounding relation. So, true propositions are grounded in entities (cf. Rodriguez-Pereyra 2005, p. 25). If this grounding relation is just the makes true relation, then this argument establishes (or comes very close to establishing) truthmaker theory (as defined above).¹⁷

So we can reconstruct the following argument for truthmaker theory:

• (9) The Asymmetry Thesis is true.

• (10) So, the Grounding Thesis is true.

• (11) So, truthmaker theory is true.

In the next section, I challenge the move from (9) to (10), that is, I challenge whether the Asymmetry Thesis supports the Grounding Thesis. I argue that there is an equally plausible account of the Asymmetry Thesis that does not require the truth of the Grounding Thesis. I further argue that it is realism about truth and not the makes true relation that explains the plausibility of the Asymmetry Thesis.

3.2 Does the asymmetry thesis support the grounding thesis?

In this section, I argue that the Asymmetry Thesis does not support the Grounding Thesis because there is an alternative explanation for this apparent asymmetry in our thinking about truth. To see this, I begin with a brief discussion of ‘because’ statements before turning to the issue at hand.¹⁸

We can usefully think of ‘because’ statements as answers to questions. The statement ‘P because Q’ can be thought of as an answer to the question ‘why P’ with ‘because Q’. Further, ‘why’ questions are often contrastive in nature. In asking why it is that P, we are often implicitly asking why it is not the case that Q or R or S, etc. Of course, as is well-known, an utterance of a single sentence (‘Why did Victoria eat the sandwich?’) might be used to express different questions, and thus different contrast classes will be relevant depending upon the question being asked. But oftentimes which question is intended is either clear from the context or can become clear by emphasis or elaboration (e.g. ‘why did Victoria eat the sandwich?’/‘why did Victoria, and not some other person, eat the sandwich?’ or ‘why did Victoria eat the sandwich?’/ ‘why did Victoria eat the sandwich, and not the piece of cake right next to it?’).

Some answers are correct answers to questions; some are not. Answers can be evaluated along three dimensions (cf. van Fraaseen (1980: pp. 146–147)). First, the plausibility or probability of the answer itself, that is, on its own. (For instance, an answer that is utterly ad-hoc is frequently not a good answer because it does poorly along this line.) Second, the relation between a particular answer and the topic of the question. (For instance, if the topic of the question is not very expectable given the answer, then it is not going to be a good answer.) Third, the relation between the given answer and other possible answers that one might give. Some answers might be deemed irrelevant, and thus incorrect, because they are eclipsed by other possible answers. Let me dwell on this point, by providing an extended example, because it will be relevant below.

Suppose I know Lu took 17th street instead of the bypass or 3rd to get to the movie theater. I ask why she did, since I know that 17th street is, of the three ways, the longest and know that Lu knows this as well. The answer ‘because you can get to the theater by 17th street’ might be incorrect, but not necessarily because it does poorly along the first two dimensions. Taking 17th might be a good way to get to the theater, and thus not an implausible way to go; further, maybe Lu only takes 17th when she is going to the theater. Nevertheless, the reason it is incorrect is because there is a better answer—Lu thought 3rd was under construction, and the bypass would be swamped with traffic after the football game. What makes that initial answer incorrect is not that it is false that you can get to the theater by 17th nor that Lu doesn’t know that. It is true, and she knows it. But that answer is not correct because there is a better one, which is easily revealed by how a conversation would unfold if a passenger of Lu’s car was asked about the route. “Why did Lu take 17th street?” “Because you can get to the theater by 17th street.” “Yes, I know that, and so does Lu. But it is the longest of the routes. Given that, why did she take it?” “O, because she thought that 3rd was under construction, and that the bypass would be swamped with traffic.”

We are now able to turn to the issue at hand. The truthmaker theorist is attempting to support truthmaker theory by pointing out that there are some plausible instances of:

• (7) $p$ is true because $p$,

and no plausible instance of:

• (8) $p$ because $p$ is true.

The first thing to note is that the instances of these schemas will be ‘because’ statements, and so on the analysis above, can be treated on the question/answer model.

Now the truthmaker theorist is right that in some cases—though perhaps not all—when we ask ‘why is it true that P?’ the correct answer is ‘because P.’¹⁹ Why is this? For the following two-part reason. First, often when asking ‘why is it true that P?’ we are implicitly contrasting truth and falsity, so that we are really asking ‘why is it that $p$ is true and not false?’ Second, the correct answer to that question is ‘because P.’

Why is ‘because P’ the correct answer? I suggest the reason is an implicit acceptance of realism about truth and not any makes true relation. Given realism about truth, the answer ‘because P’ does well along the first two dimensions. Depending upon the case, P will not itself be implausible and there is a connection between P and it being true that P, namely a modal connection! Further, given an implicit commitment to realism about truth, other answers—e.g. answers that might be given by antirealist theorists about truth—are not correct answers because they are simply wrong. Thus, ‘because P’ does well along the third dimension as well. Realism about truth can thus explain why ‘because P’ is the correct answer.

Further, a makes true relation is not sufficient for answering that P. For instance, one can concede that there is a makes true relation that makes P true, but deny that it is true that P because P because one accepts some version of antirealism about truth. On such a view, it is true that P because (e.g.) an ideal scientific community would assent to P.²⁰ Thus, ‘because P’ can be the correct answer but not because of truthmaker theory. Rather, it is some tacit acceptance of realism about truth (or a similar view) that accounts for why ‘because P’ can be the correct answer.

At this point, one might immediately object—the truth of a ‘because’ statements require an asymmetric relation and I’ve only pointed to a bi-conditional (albeit a necessary bi-conditional) that is not asymmetric. But this is wrong. The truth of certain ‘because’ statements rests upon the kinds of questions one is asking, and different questions can exploit different sides of a bi-conditional. As van Fraassen illustrates,²¹ one can explain the length of the shadow by the particular height of the tower, but one can also explain the height of the tower by the length of the shadow—though one cannot do both in the same breath. It all depends upon the question being asked and what counts as a relevant answer.

But how, then, to account for our intuition that ‘it is true that P’ could almost never be an answer to the question ‘why P?’ if one can exploit either side of a bi-conditional in providing explanations? My answer to this is that there is an important difference between our asking the question ‘why is it true that P?’ and ‘why is it the case that P?’ The former is to ask the question ‘why is P true, not false?’ By contrast, the latter is to ask the question ‘why is it the case that P and not not-P (e.g. Q, R, or S instead)?’ And here an important difference opens up. For in the first question, the contrast class semantically ascended—why is it true, not false, that P? But in the second question, we have not. This results in different answers being relevant to the two questions. Importantly, because we have not semantically ascended in the second question, there is often an implicit causal constraint at play when we ask the latter question that need not be there when we ask the former. For instance, if you ask of my dog Ava, ‘why is she overweight?’ the appropriate set of answers do not semantically ascend, but instead tell a causal story—I feed her too much so that she’ll leave me alone while I type; she has a thyroid problem; she does not like play, etc. This causal constraint is in play for the simple reason that the majority of times that human persons ask these kinds of why questions, we are asking about a causal nexus and the causal relations that constitute it.

The result of this is that, if one asks ‘why is it the case that P?’ the answer ‘because it is true that P’ is superseded by other answers, in particular, answer that put together a causal story of how it is that P came about. Of course, if Ava is overweight, then it is true that she is overweight. But in asking for what causal antecedents led to her being (or continuing to be) overweight, ‘it is true that she is overweight’ will play little to no role, instead answers like ‘she has a thyroid problem’ or ‘I feed her too much so she’ll leave me alone to type’ are better answers.

Summing up, then, I agree with the truthmaker theorist that the Asymmetry Thesis is plausible—that there are some plausible instances of:

• (7) $p$ is true because $p$,

and no plausible instances of:

• (8) $p$ because $p$ is true.

But I do not account for this asymmetry by postulating a metaphysical explanatory relation. Rather, on my account, the difference has to do with the kinds of questions that we ask, when we ask instances of these schemas. When we ask ‘why is it true that $p$’ vs. ‘why is it that $p$’ we are asking different kinds of questions, with different constraints on what is to be a relevant answer. It is this that accounts for the asymmetry, not the makes true relation.

Before moving on, I think it is important to get the dialectic straight at this point. The truthmaker theorist is providing an argument from truthmaker theory. I’m resisting the first step of the argument—the step from The Asymmetry Thesis to the Grounding Thesis—by offering a non-metaphysical way of understanding the asymmetry in our thinking about truth. I’ve further suggested that the makes true relation by itself is not sufficient to account for the asymmetry in our thinking about truth. If the truthmaker theorist responds by claiming that I have yet to “account for how truths are true in virtue of reality” or something along those lines, then she is merely begging the question and asserting the truth of the Grounding Thesis. She needs to either provide a different argument for the Grounding Thesis or show how my account is not sufficient to explain the Asymmetry Thesis.

3.3 Comparisons to other accounts

I am not alone in being skeptical that the Asymmetry Thesis (or something similar) provides reason for accepting truthmaker theory and trying to account for it otherwise. Here I briefly compare my account to three others—Jennifer Hornsby’s, Julian Dodd’s, and Benjamin Schnieder’s.

Jennifer Hornsby believes there is “a certain asymmetry in our thinking about truth [that] needs to be explained” (2005, p. 42). Her explanation is this:

For the rose to be red, it isn’t required that there should be anything meaningful to say or to think; in particular, it isn’t required that the proposition that the rose is red should be propounded. Again, for the rose to be red, it isn’t required that one should be able to speak or to think of a set whose sole member is the rose. Thus when we say that the proposition that the rose is red is true, or that the set whose sole member is the rose has a member that is red, we say more than what is required for the rose to be red. If we simply say that the rose is red, by contrast, then we say no more than that which is required for the rose to be red. (2005, pp. 43–44).

Hornsby’s thought seems to be that the asymmetry should be explained in terms of the existence of propositions. It can be the case that the rose is red without the proposition that the rose is red existing and being true. This, I take it, is the main drawback of Hornsby’s account in comparison to mine. As Dodd (2007, pp. 398–399) points out, her explanation requires one to reject the following principle: necessarily, if it is the case that $p$, then there is the proposition that $p$, and it is true. This is a plausible principle, and while I won’t defend it here, it is worth pointing out that my account above does not require rejecting it. This is a mark in favor of my account.

Julian Dodd (2007, p. 389ff.) explains the Asymmetry Thesis as follows. Unlike Hornsby, Dodd thinks that propositions are necessary existents, and so every world in which it is the case that $p$ there is a proposition that $p$ and it is true, and vice versa. His explanation of the Asymmetry Thesis thus could not trade on such a difference. Consequently, he uses a hyperintensional notion: Kit Fine’s “ontological dependence” (1995). On Dodd’s account (roughly put), the reason why it is plausible that

• (5) It is true that there is a man because there is a man,

is because the proposition is ontologically dependent upon there being a man, and not vice versa.

While it is beyond the scope of this paper to fully discuss this point, as I see it, the chief problem with Dodd’s explanation is that it requires the use of Fine’s notion of ontological dependence—and that notion is another metaphysically explanatory primitive notion. To be sure, the makes true relation and the relation of ontological dependence might be importantly different according to their respective proponents. But, for many skeptical of the makes true relation, to trade the makes true relation for the relation of ontological dependence is just to jump from the frying pain into the fire. My account allows us to stay out of both.

Finally, Schnieder (2006) attempts to account for the asymmetry of truth by using the notion of a “conceptual explanation.” To provide a conceptual explanation is to provide an explanation that illuminates certain conceptual relations. He gives the following example: “Thorsten is my brother-in-law because he is married to my sister” (2006, p. 32). In this case, the concept ‘brother-in-law’ is explained in terms of more basic concepts such as ‘married’ and ‘sister.’ While this is a case where one concept is explained it terms of others because the others are part of its analysis, Schnieder allows that this need not always be the case.

He then considers the following principle, which is closely related to my Asymmetry Thesis:

(T) For all $p$: if it is true that $p$ at all, then it is true that $p$ because $p$.

This principle, he says, “exploits a crucial part of our understanding of the concept of truth.” This “crucial part” is the Tarski schema—$p$ is true just in case $p$. This schema “lies at the heart of the concept of truth” (Schnieder (2006), p. 36). Principle (T), then, provides a conceptual explanation—since it helps illuminate the concept of truth—and instances of it are consequently also cases of conceptual explanation (Schnieder (2006), pp. 36–37). So goes Schnieder’s positive account.²²

Schnieder’s notion of conceptual explanation is an interesting and useful one. I find what he has to say for certain concepts—such as brother-in-law or widow—quite promising. But I’m dubious of his extension of this concept to truth. I do not think that (T) “exploits a crucial part of our understanding of the concept of truth” or generally understand how (T) is meant to illuminate that concept. Importantly, the case of (T) is different from the case of ‘brother-in-law.’ In the latter case, the conceptual explanation holds because one is providing more basic concepts that might be part of an analysis of the more complex one. But this is not the case for (T)—generally speaking, P will not contain any concepts that will be part of an analysis of the ‘it is true that’ in “it is true that P.”

Schnieder recognizes this but does not think that it is problematic (2006, p. 36). For, plausibly, one can provide a conceptual explanation by providing insight into a concept and one can provide insight into a concept by showing how it interacts with other already possessed concepts: “mastery of the concept is constituted by the ability to relate statements involving it to statements involving only conceptual resources already at hand” (2006, p. 36).

But it is not clear that this helps Schnieder in the case of (T) and truth. For, at best, (T) illuminates how we are to use the concept of truth by indicating that one can use a truth-predicate to apply to propositions. But so would other principles that Schnieder would presumably reject, such as:

(T*) For all $p$: if it is true that $p$ at all, then $p$ because it is true that $p$.

Further, (T) contains a certain asymmetry, whereas the Tarski schema that is to be illuminated does not. It is thus not clear why he thinks that a principle like (T) illuminates the concept of truth any more so than does realism about truth, with lacks such an asymmetry.