Troubles With Truthmaking: The Truthmaker and Veritas Sequitur Esse Principles

Some recent attempts (by G. Oppy, J. Brower, A. Pruss and perhaps others) at making sense of the doctrine of divine simplicity (DDS) have invoked the truthmaker principle (TMP).  I made heavy use of TMP in my A Paradigm Theory of Existence  (Kluwer 2002), though not in defense of DDS. Being a self-critical sort, I am now re-examining the case for TMP.  Note that acceptance of TMP does not straightaway commit one to acceptance of any particular category of entity as truthmakers such as concrete states of affairs.  One could accept TMP and hold that truthmakers are tropes.  And there are other possibilities. So before we can address the truthmaker defense of DDS we must (i) argue for TMP and then (ii) decide on what can and cannot function as truthmakers.  In this post I consider some of what can be said for and against truthmaking in general. It looks like we might be in for a long series of posts on this fascinating but difficult topic.

1.  The basic idea behind TMP is that for every truth there must be at least one (though there could be more than one) item distinct from the truth that 'makes' it true.  The principle is most intuitive when restricted to contingent non-negative atomic truths monadic or relational, e.g., 'Peter is asleep'  and 'Peter is sleepier than Paul.'  If such a sentence (or the Fregean proposition it expresses) is true, it is not just true; one feels that there must be something in the extralinguistic and extramental world that 'makes' it true.  The making, however, is not causal in nature.  And so philosophers typically reach for the weasel phrase 'in virtue of.'  (I've done it myself.)  Thus the truthmaker of a truth is said to be that 'in virtue of which' it is true.  Or we could say that some or perhaps all truths are in need of an ontological ground of their truth.  'Ontological ground' conveys the idea that outside language and mind there is something that anchors the truth of (some of) our true beliefs, judgments, sentences, propositions, etc.  I myself find TMP to be an exceedingly attractive principle.  But obscurities soon surface as we try to penetrate more deeply into it.

2.  One obscurity has to do with what exactly the truthmaking relation is.  It is first of all clear that the relation is both dyadic and asymmetrical.  It is dyadic in that it connects two items, a truthmaker and a truthbearer. It is asymmetrical in that, if x makes true y, then it is not the case that y makes true x.  A further point about truthmaking is that the relation between truthbearer and truthmaker is many-many, not one-one.  Different truthbearers can have the same truthmaker, and different truthmakers can make true the same truthbearer. 'Peter is asleep' and 'Someone is asleep' are both made true by Peter's being asleep, the concrete fact or state of affairs having Peter himself, warts and all, as constituent.   Peter's being asleep and Paul's being asleep both make true 'Someone is asleep.' It is also clear that truthmakers necessitate the truth of the corresponding truthbearers: in every BL-possible world in which the concrete fact of Peter's being asleep obtains, the proposition *Peter is asleep* is true.  (Asterisks around a declarative sentence form a name of the Fregean proposition expressed by the sentence, and asterisks around a predicate form a name of the predicate's sense.)

But, as I said, truthmaking is not a causal relation, and so the necessitation is not causal necessitation.  If it is true that Peter is tired, then of course there is a causal explanation of this truth in terms, say, of his having stayed up all night talking philosophy with Saul.  What we are seeking, however, is not a 'horizontal' causal explanation of a particular truth's being true, but a 'vertical' philosophical explanation of why any (contingent, atomic, nonnegative) truth is true.  The philosophical explanation is that such truths are true because of the existence of certain entities that make them true, entities that are not themselves truths.

If truthmaking is not causal, then one surmises that it is broadly logical, and that truthmaking necessitation is broadly logical necessitation.  So it looks to be something like entailment.  But entailment is defined over propositions.  Thus, for any propositions p, q, p entails q =_df there is no BL-possible world in which p is true and q false.  That is a standard way of thinking about entailment according to which entailment is the necessitation of material implication.  Now if that is what we mean by the relation of entailment, then it is clear that the truthmaking relation cannot be identical to the entailment relation. For truthmakers are not propositions, but the ontological grounds of true propositions.  It seems that, for the very same reason, truthmaking cannot be a species of entailment.  One cannot take entailment as genus and then add a specific difference so as to arrive at truthmaking.  For again, truthmakers are not propositions or any kind of sentence-like representation.  They are transcendent of mind, language and the representational apparatus of Frege's realm of sense.

Truthmaking, then, is neither causation nor entailment. And yet it combines elements of both.  For a truthmaker is a pre-propositional ground of propositional truth the existence of which somehow BL-necessitates the truth of the truthbearer.  And this is obscure.  Truthmaking is an asymmetrical relation that connects a nonproposition to a proposition.  We understand what it means to say that one proposition entails, and by entailing BL-necessitates, the truth of another -- see the definition above -- but we don't quite understand what it means to say that a truthmaker, a nonproposition, BL-necessitates the truth of a proposition.

And so some of us are in the following predicament.  We have the strong realist intuition that at least some truthbearers are in need of ontological grounds of their truth, but we are unable to make quite clear just what this relation of truthmaking is.  It is neither causation nor entailment nor a species of entailment.  What exactly is it then?

3.  TMP is sometimes formulated as follows:

TMP-A  For a proposition to be true, there must exist at least one item, distinct from the proposition, whose existence entails that the proposition is true. (J. Dodd, "Farewell to States of Affairs," AJP, June 1999, 147)

For a more precise statement we turn to Beebee & Dodd (2005):

TMP-A' Necessarily, if *p* is true, then there exists at least one entity t such that *t exists* entails **p* is true.*

Note the difference between TMP-A' and

TMP-B Necessarily, if *p* is true, then there exists at least one entity t such that t entails **p* is true.*

In effect, I rejected TMP-B in #2 above on the ground that entailment is a relation all the relata of which are propositions which implies that the truthmaker t cannot stand in an entailment relation. But a proposition about t can, and this yields TMP-A'.

What TMP-A' implies is that the mere existence of a truthmaker logically suffices for the being true of the corresponding truthbearer, and therefore that the internal structure of the truthmaker is irrelevant.  For on this formulation it is not t that stands in an entailment relation, but the proposition *t exists.* This formulation has the advantage of not implying that truthmakers make true their truthbearers by entailing them, a notion that literally makes no sense given that truthmakers are not propositions.  Unfortunately, if the mere existence of a truthmaker BL-necessitates the truth of the corresponding truthbearer, then a concrete individual such as Peter himself could be a truthmaker.  Of course, Peter himself cannot serve as truthmaker of *Peter is asleep* for the reason that there are possible worlds in which Peter exists but the proposition in question is not true.  There is also the point that Peter himself does not 'select' between *Peter is asleep* and the true *Peter is drunk* as the truthmaker of an atomic predication must. TMP-A and TMP-A' do however allow that Peter himself is the truthmaker of  such essential predications as 'Peter is human,' 'Peter is an animal,' 'Peter is self-identical,' and 'Peter exists.'

TMP-A and -A' thus allow two sorts of truthmakers, those with a proposition-like structure (e.g. the fact of Peter's being tired) and those with no such structure (e.g., Peter). And note that in both cases it is the mere existence of the truthmaker, or a proposition to that effect,  that is said to necessitate the truth of the corresponding truthbearer.  But surely it cannot be the mere existence of the fact of Peter's being tired that makes true *Peter is tired.*  It is the nature or internal proposition-like structure of this fact that makes *tired* true of Peter.  Clearly, *Peter is tired* is true iff *tired* is true of Peter, which implies that the truthmaker of *Peter is tired* must be the ontological ground of predicate's (or the corresponding property's) applying to (or being instantiated by) Peter.  So it cannot be the mere existence of the fact of Peter's being tired that is the truthmaker of *Peter is tired.*

Therefore, in allowing two sorts of truthmakers, TMP-A and -A' give aid and comfort to an equivocation on 'truthmaking' and 'truthmaker.'  On the one hand, a truthmaker is the ontological ground (in the simple atomic monadic case) of a predicate's applying to a subject.  On the other hand, a truthmaker is an entity whose mere existence is such that, in every world in which it exists, the corresponding truthbearer is true.  This is tantamount to a confusion of TMP in its original sense -- TMP-B -- with a distinct principle which I shall now explain.

4.  There is more to TMP in its original intuitive sense as explained in #1 above than the idea that true sentences (propositions, etc.) must be about something that exists.  One could grant that 'Peter is asleep' cannot be true unless there exists something to which 'Peter' refers without granting that 'Peter is asleep' has a truthmaker.  One could maintain that veritas sequitur esse rerum (truth follows upon, is logically subsequent to, the being of things) without subscribing to TPM.  And so we ought to distinguish VSE from TMP-B.

VSE:  Every truth is about one or more items that exist or else presupposes the existence of one or more items; there are no truths about nonexistents.

VSE is not trivial: Meinongians deny it.  One who accepts VSE must say that 'Peter is asleep' cannot be true unless Peter exists.  And such a person must say about 'Someone is asleep' that it cannot be true unless there exists someone who is asleep.  But such a person cannot say that Peter is the truthmaker of 'Peter is asleep.'  Otherwise he would have to say that Peter is also the truthmaker of the true 'Peter is drunk' when it is clear that Peter cannot be the truthmaker of both.  Truthmakers must 'select' between different affirmative atomic contingent predications even in cases in which the subject terms have the same referent. Peter's existence is a necessary condition of the truth of both sentences; but that is not to say that either Peter or Peter's existence (which we ought to distinguish) is the truthmaker of either sentence. In the atomic monadic case, a truthmaker is an ontological ground of a predicate's being true of a concrete individual, which implies that neither the referent of 'Peter' nor the referent of 'asleep'  nor the ordered pair <Peter, being asleep> can be the truthmaker of 'Peter is asleep.'  The truthmaker either must be or be relevantly similar to the concrete fact or state of affairs, Peter's being asleep.

We can put this by saying that the truthmaker of a contingent atomic truth must be a proposition-like entity, and that this proposition-like structure must play a role in truthmaking if TMP-B is not to collapse into the rather less interesting (but also less controversial) VSE.   It must be proposition-like, but it cannot be a (Fregean) proposition, for no such proposition is a truthmaker. Why not?  Truthmakers are not (Fregean) senses; they are not part of the apparatus by means of which we refer to objects in the realm of primary reference; they are denizens of the realm of primary reference.  To put it another way, truthmakers are external to mind, language and the abstract semantic apparatus by means of which we secure access to what is external to mind and language. 

5.  So what am I arguing?  If there is to be a distinctive truthmaker principle worth discussing and worth putting to work in different areas of philosophy, then it cannot collapse into the uncontroversial veritas sequitur esse rerum principle.  But it does so collapse if spelled out in terms of TMP-A or TMP-A'.  For what these formulations do is construe truthmaking as a relation of entailment that connects a proposition, one that states that a nonpropositional entity exists, to a proposition.  But then what is distinctive to truthmaking is lost.  In the affirmative atomic monadic case, a truthmaker is a nonpropositional entity that grounds or accounts for a predicate's being true of its subject.  But it cannot do this unless it has an analogous, a proposition-like, structure that somehow determines the truth value of the truthbearer.  It is not in virtue of its mere existence that anything is a truthmaker, but in virtue of its internal proposition-like structure.  This is why Peter himself, that very concrete individual, cannot be a truthmaker of any truth.

To call Peter himself the truthmaker of 'Peter is human,' 'Peter is self-identical,' and 'Peter exists' is to show a lack of understanding by confusing VSE with TMP-B.  None of these essential predications can be true unless Peter exists; but that is not to say that Peter makes them true in the distinctive sense of 'makes true' explained in #1 above.  In the case of 'Peter is self-identical,' there is no need for a truthmaker since necessarily nothing can exist that is not self-identical. Similalrly for 'Peter exists':  there is no need for a truthmaker since necessarily nothing can exist that does not exist.  Nor does 'Peter is human' require a truthmaker.  For if 'Peter' is being used as a name for a human being, then it is automatically the case that Peter is human.

I conclude aporetically.  My realist intuitions tell me that there must be truthmakers, but so far I have been unable to come up with a clear and precise explanation of just what truthmaking is.

Now if the nature of truthmaking is in the dark, then of course we cannot use it to clarify the doctrine of divine simplicity.

Comments

[Alexander R Pruss]

1. The truthmaker-based defenses of divine simplicity do not require TMP. They only require that the notion of truthmaking make sense, and that some propositions have truthmakers.

2. I think that few philosophically interesting notions have definitions. But we can introduce truthmakers by way of example, and then by delineating some formal properties. That Obama exists has Obama as a truthmaker. That Obama exists or Bush is not president has Obama as a truthmaker. That Obama exists or Bush exists has Obama as a truthmaker; it also has Bush as a truthmaker. On some trope theories, that Fred is a triangle is made true by Fred's triangularity (which is a particular that couldn't exist without being the triangularity of Fred). Etc. We can delineate some formal properties. Only true propositions have truthmakers. Any truthmaker for a disjunct is a truthmaker for a disjunction. If F(x) is a formula open in x, and then any truthmaker for the proposition that F(a), where a is a name, is a truthmaker for (Exists x)F(x). If x is a truthmaker for p, then that x exists entails that p. Moreover, if there is such a thing as relevant entailment, then this entailment is relevant. If x is a truthmaker for p, then p is at least partly about x. If x is a truthmaker for p, and p is not just a proposition asserting the existence of x, then the proposition that x exists is a constitutive explanation of p.

3. I also think the truthmaker relation can be reduced to a more general relation, the true-in-virtue-of relation that holds between propositions. Specifically, x is a truthmaker for p iff either p is the proposition that x exists or p is true in virtue of the proposition that x exists. The true-in-virtue-of relation is, in turn, something that, as far as I know, can't be defined. But we can give examples and formal properties, once again. That the knife is hot is true in virtue of the fact (=true proposition; this is stipulative, and not a substantive theory of facts) that the knife's molecules have high molecular kinetic energy. If divine command theory holds, then that murder is wrong is true in virtue of the fact that God has forbidden murder; if Kantianism is true, then that murder is wrong is true in virtue of the fact that murder violates the first categorical imperative; it is also wrong in virtue of the fact that it violates the second. If Platonism is true, than that George is F, where Fness is a basic property, is true in virtue of the fact that George instantiates Fness. If A is a mereological sum of distinct items B and C, then, I think, that A exists is true in virtue of the fact that B exists and that C exists. (I don't think there are any mereological sums, actually.) If one of Hume's accounts of causation is right, then that A causes B is true in virtue of the fact that A-type events are followed by B-type events. Much of the task of the philosopher, in every major field of philosophy, is, given an interesting truth p, to find an illuminating truth q in virtue of which p holds. If p is true in virtue of q, then q explains and entails p, and if there is such a thing as relevant entailment, it entails relevantly. If a disjunct is true in virtue of q, so is the disjunction. Only true propositions stand in the true-in-virtue-of relation.

[Bill Vallicella]

Thanks for the comments, Alex.

Ad (1). You've lost me on this one. If every version of TMP is false, then it cannot be the case that "some propositions have truthmakers."

Ad (2) and (3). "That Obama exists has Obama as a truthmaker." No. Did you read my post? If you want to introduce truthmakers, then you need to give a clear example, e.g. 'Obama occupies the White House.' The clear examples suggest that if there are truthmakers, then they are either concrete states of affairs or something like them.

[Alexander R Pruss]

Bill:

TMP says, basically, that all true propositions have truthmakers. At least that's how you introduce TMP. But it is quite coherent to think that some, but not all, true propositions have truthmakers. For instance, take Lewis's paper on truthmakers. Lewis rejects TMP, but thinks that some propositions are true in virtue of having a truthmaker; some other propositions are true in virtue of lacking a falsemaker (cf. Aristotle's remark that to speak truly is to say of what is that it is or of what is not that it is not--Aristotle seems to be committed here to the theory that every true proposition is true by having a truthmaker or lacking a falsemaker); some other propositions are true in virtue of an entity having a certain property; and some propositions are combinations of these.

I am simply unconvinced that truthmakers have to have propositional structure. You write: "In the affirmative atomic monadic case, a truthmaker is a nonpropositional entity that grounds or accounts for a predicate's being true of its subject." Maybe that's true in the affirmative atomic monadic case. But it's not true in the case of a singular existential proposition.