Working Draft: The Case Against Facts

Comments appreciated if you are en rapport with the subject matter.

 

The Case Against Facts

 

Arianna Betti, Against Facts, The MIT Press, 2015, pp. 296 + xxvii

 
If Buridan's contribution to the bestiarum philosophorum was the ass, and David Armstrong's the ostrich, Arianna Betti's is the hedgehog bristling with spines. The hedgehog is an appropriate totemic animal inasmuch as her book too bristles with sharp distinctions and prickly arguments designed to inflict pain upon the friends of facts. In this penetrating and beautifully organized volume Professor Betti deploys her distinctions and arguments against two sorts of facts, compositional and propositional, as she calls them. The states of affairs of David Malet Armstrong's middle period (Armstrong 2007) are examples of the first kind of fact. These items are the main target of Betti's animadversions in the first part of her two-part book. She does not go so far as to claim that Armstrongian facts do not exist; her claim is the rather more modest one that we have no reason to posit them, since the work they do, if it needs doing at all, can be done just as well by a certain sort of mereological sum. (101) Betti ignores, however, Armstrong's very different later conception of states of affairs or facts. (Armstrong 2009; Armstrong 2010, 26-34; Vallicella 2016) This later conception also counts as compositional in her sense and ought to have been discussed for the sake of completeness, especially since it in some ways approximates to Betti's mereological  position.

 

One might wonder how a fact could fail to be compositional. Facts are complex or composite items, after all, not simples. So they must all have some internal composition or other, whether they be truthmaking facts or facts of the Chisholmian-Plantingian sort. At a bare minimum, a's being F is composed of a and F-ness. Thus I find less than felicitous Betti's talk of propositional facts in contrast to compositional facts as “noncompositional objects at the level of reference.” (24) She makes it clear, however, that she is using 'compositional' in a narrow sense that implies that compositional facts and their constituents are “part[s] of the furniture of the world.” (37) We shall soon see that being in the world involves being real as opposed to being ideal. An example of a compositional fact is the fact of Guido's being hungry. This fact has Guido himself, all 200 lbs of him, as a constituent. An example of a propositional fact is the putative referent of the that-clause in a sentence like 'Guido sees that Francesca is serving spaghetti puttanesca.' This putative referent is the fact that Francesca is serving spaghetti puttanesca. This propositional fact is like a (Fregean) proposition, though it is not a proposition, in that it does not have Francesca herself as a constituent, but rather an abstract surrogate that represents her. (170) (This fact-of vs. fact-that terminology is mine, not Betti's. I got it from Milton Fisk.)

 

Betti describes in marvellous detail seven features of compositional facts (18) and five of propositional facts (170). I will speak of C-facts and P-facts. Here are some salient differences. C-facts are in the world, and thus suited to play the truthmaking role whereas P-facts are not in the world and hence not fit for truthmaking. To be in the world is to be real where to be real is to exist “through time and in time as causes or effects in a causal chain.” (22) So C-facts are real while P-facts are ideal. The ideality of P-facts, however, is not that of propositions since P-facts are not propositions. Betti is greatly and rightly exercised by the curious in-between status of these “ghostly critters” (114) that are neither truthbearers nor truthmakers and yet are championed by such distinguished philosophers as Roderick Chisholm, Alvin Plantinga, and Kit Fine. These “ghostly critters” are not truthbearers because they are neither true nor false. But while they are not bivalent in terms of truthvalue, they are 'bipolar' (my term): while all exist, some of them obtain while some do not. They are not truthmakers since truthmakers are real and 'monopolar': if they don't exist they are nothing. Thus the fact of Guido's being hungry does not exist at all if Guido is not hungry. Propositional facts are neither fish nor fowl. The conclusion Betti arrives at strikes me as correct: “Propositional facts collapse into true propositions.” (179) Propositional facts are thus not a distinctive category of entity. We need them, she thinks, as little as we need compositional facts. Actually, her position is far more radical than this since she denies that that-clauses are referential parts of speech. So her position is best expressed conditionally by the following quotation: “If there were nominal reference to facts, facts would be true propositions . . . . (113) Her view, if I understand it, is eliminativist not identitarian: she is not saying that there are propositional facts and that what they are are true propositions; she is saying that that there are no propositional facts.

 

Leaving propositional facts to languish in their ghostly realm, the rest of this article will take issue with Betti's critique of compositional facts, the ones dear to my heart, the facts involved in the flux and shove of the real order. On a personal note, I want to thank Professor Betti for her very close attention to my articles on the topic.

 

The Case Against Compositional Facts

A compositional fact, as opposed to a propositional fact, is an entity fit to play the role of truthmaker. The truthmaker role may be introduced as follows. Consider the assertive utterance of some such contingent sentence as 'Tom is sad.' If true, this assertively uttered sentence cannot just be true: if true, it is true because or in virtue of something external to it. This use of 'because' is not causal which is why philosophers reach for the weasel phrase 'in virtue of,' which, despite its slipperiness, may well be indispensable for metaphysics. I say it is indispensable. (Or do hedgehogs eat weasels?) Roughly, there has to be something that 'makes' the sentence true. This external something cannot be another declarative sentence, even if true. More generally, a truth is a true truthbearer (a Fregean proposition, say, or perhaps an Aristotelian proposition, see pp. 31-32 for Betti's helpful explanation of the difference) and no true truthbearer is made true by another such item in the specific sense of 'makes true' in play in truthmaker theory. Nor can someone's say-so be what makes true a true truthbearer. The truthmaker has to be something 'in the world,' something extralinguistic and mind-independent in the realm of reference as opposed to the realm of sense. The friends of truthmakers are realists about truth: they are convinced that at least some truths are in need of an ontological ground of their being true.¹

 

Truthmaker maximalists hold that all truths need such grounds, but one needn't be a maximalist to be a truthmaker theorist. As for 'makes true,' this is neither entailment nor causation. Not entailment, because entailment is a relation between propositions, assuming that truthbearers are propositions, whereas truthmaking is a relation between extra-propositional reality and propositions. So if x makes true y, then y is a truthbearer, but x is not. If someone says that the proposition expressed by 'Snow is white' makes true the proposition expressed by 'Something is white,' then that person, while talking sense, is not using 'makes true' in the specific way in which the phrase is used in truthmaker theory. Truthmaking is not causation for a similar reason: causation does not connect the extra-propositional to the propositional whereas truthmaking does. As Armstrong says, truthmaking is “cross-categorial.” (Armstrong 2004b, 5) It links the extra-propositional to the propositional.

 

It is important to note, however, that while truthmakers cannot be Fregean or Aristotelian propositions, and thus must be extra-propositional, they must also be proposition-like on Armstrong's approach. This is a point I think Betti misses. Speaking of compositional facts, she tells us that “facts are neither linguistic nor languagelike entities at the lowest level of reference. (28, emphasis in original) But this is certainly not Armstrong's view, the view that is supposed to be the target of Betti's critique of compositional facts. His view is that the world is a world of states of affairs, a “totality of facts not of things” (Wittgenstein) and “sentence-like rather than list-like.” (Armstrong 2010, 34) If the world is sentence-like, then, pace Betti, it is language-like. Armstrong was profoundly influenced by his teacher in Sydney, the Scots philosopher John Anderson, who held that “reality, while independent of the mind that knows it, has a 'propositional' structure.” (Armstrong 1997, 3) Armstrong goes on to say that “the propositional view of reality which he [Anderson] championed is the facts or states of affairs view of reality.” (Armstrong 1997, 3-4) That Armstrongian facts are proposition-like and thus language-like is fairly obvious when we consider the truthmakers of contingent predications of the form 'a is F.' The truthmaker cannot be a by itself, or F-ness by itself, or the mereological sum a + F-ness. It must be a-instantiating F-ness, which has a proposition-like structure. Armstrongian facts have a logos-like and thus logical articulation contrary to what Betti says in opposition to Kit Fine. (28) But now I am getting ahead of myself.

 

Suppose you accept the legitimacy of the truthmaker role and the need for some type of entity to play it. It doesn't follow straightaway that the entities needed to play the role must be what Betti calls compositional facts or what David Armstrong calls states of affairs. This is so even if we confine ourselves to the really clear examples of truthbearers in need of truthmakers, namely, synthetic, contingent predications such as 'Guido is hungry' or the propositions expressed by assertive utterances of such sentences. Nevertheless, a powerful argument can be mounted for compositional facts as truthmakers. The argument Armstrong and I consider powerful, however, Betti calls “unsound.” (106) Surprise!

 

Although she is skeptical of the need for truthmakers, she is willing to grant the need arguendo, insisting only that if we need truthmakers, a certain type of mereological complex can do the job thus rendering Armstrong's facts, as unmereological complexes, unnecessary. (102) This is why she thinks the truthmaker argument for Armstrongian facts is unsound. As she sees it, compositional facts are not givens, but theoretical posits, and unnecessary ones at that. They were invented to solve a problem, the unity problem, that arises only because of certain optional assumptions about relations and properties that one is not bound to make. (94-95) Compositional facts are an ad hoc, indeed a “maximally ad hoc,” solution to a pseudo-problem. (64)

 

Now let me say something in exposition of Armstrong's argument for facts or states of affairs as truthmakers on the assumption that the truthmaker role is legitimate and needs to be filled by some category of entity or other. I will then consider Betti's counter-proposal.

 

If it is true that Tom is sad, could the truthmaker of this truth be the item that Betti calls (8) the sentence-subject of 'Tom is sad,' namely, Tom? No, since Tom needn't be sad. So Tom by himself cannot be what makes true 'Tom is sad.' The same goes for the property of being sad. By itself the property cannot be the truthmaker of the sentence in question. (I am assuming, with Armstrong, that properties are immanent universals. Immanent, in that they cannot exist uninstantiated; universal, in that they are repeatable.) Now if Tom exists and sadness exists, then so does the mereological sum Tom + sadness. But this sum cannot be the truthmaker either. For the sum exists whether or not Tom is sad. How so?

 

Suppose that Tom is not sad, but Shlomo is. If properties are immanent universals, then sadness cannot exist uninstantiated; suppose it exists in virtue of being instantiated by Shlomo. So Tom exists, sadness exists, and their sum exists. But this does not suffice for Tom's being sad. There is a missing ontological ingredient: something to connect sadness to Tom. You might think that the missing ingredient would have to be the worldly correlate of the 'is' of predication. But if you take this correlate to be an exemplification/instantiation relation then you ignite Bradley's relation regress which is unfortunately vicious. Other moves invoking Strawsonian nonrelational ties, Bergmannian nexus, Fregean unsaturated concepts, and benign fact-internal infinite regresses (see Vallicella 2010), are equally unavailing. The unifier of a fact's constituents cannot be a further constituent or anything internal to the fact. This leaves two possibilities: (i) the unifier is external to the fact, which Betti rejects, and (ii) Armstrong's middle-period suggestion that facts are entities in addition to their consituents and it is they who hold fact-appropriate constituents together so that they can exercise the truthmaking function. Betti has mastered the dialectic and considers the least bad solution to be Armstrong's: facts hold their constituents together. Although she doesn't say so, she considers my solution in terms of an external unifier to be the worst. The extant putative solutions to the unity problem of course presuppose that it is a genuine problem. Betti thinks it isn't.

 

Betti's Dissolution of the Unity Problem

 

After rejecting the extant putative solutions to the unity problem, Betti proposes to dissolve it by collapsing the distinction between “relations that relate relata and relations that do not: all relations relate relata and carry out their own unifying work.” (95) She means this to apply to properties as well. All properties qualify their bearers and carry out their own qualifying work. Thus there needn't be anything to hold the constituents of a relational or as monadic fact together: nothing internal to the fact, nothing external, and not the fact itself. Betti's point is that there is no need for Armstrongian facts, facts as entities in addition to their constituents. (Cf. Armstrong 1997, 117) Her point is not that there are no facts. There may well be facts; it is just that if there are, they are a special sort of mereological sum. Perhaps we can say that she is an identitarian about compositional facts, not an eliminativist, whereas she is an eliminativist about propositional facts, not an identitarian. More on this in a moment.

 

What Betti has to do is block a possibility like the following. In the actual world, call it Charley, Tim loves Tina. In a merely possible world w in which Tim and Tina both exist, Tim does not love Tina, but Tim loves Toni. In Charley we have both the relational fact of Tim's loving Tina and the mereological sum Tim + loves + Tina. In w, we have the sum Tim + loves + Tina but not the corresponding fact. This implies that there is more to the fact than the sum of its constituents: the sum can exist without constituting a fact. The something more is that which makes of the constituents a real truthmaking unity. Call it the unifier. Betti thinks that the least bad of the extant proposals as to what the unifier is is Armstrong's: facts hold their constituents together; facts are unmereological complexes over and above their constituents. In short, what Betti needs to do is counter the seductive thought that in an actual relational situation such as that of Tom's loving Tina, the constituents can exist without forming a real truthmaking unity. What she needs to maintain is that, necessarily, if all the constituents exist, then the relatedness exists. If the mere existence of the constituents ensures their connectedness, then there is no need for Armstrongian facts. You would then have real unity on the cheap, real truthmaking unity from mereology alone, or rather from mereology operating upon the right sorts of constituents. The mereological principle of the extensionality of parthood would hold for all complexes. Nice work if you can get it!

 

Betti can achieve her end if she holds that relations are relata-specific where “A relation is relata-specific if and only if it is in its nature to relate specific relata.” (89) Suppose that the relation loves as it figures in the sum Tom + loves + Tina is necessarily such that, if it exists, then it relates Tom and Tina. Then there would be no distinction in reality between loves as a relating relation and loves as an inert relation that is merely a constituent but not also a unifier of the complex into which it enters.

 

Betti's contention, then, is that all relations, just in virtue of existing, are relating relations, active ontological ingredients if you will, and none are inert ingredients. A relation cannot exist without actually relating its relata. If so, there cannot be a difference between the mereological sum a + R+ b and the fact of a's standing in R to b. Given the constituents, the fact is given: it is not an ontological extra, something over and above the constituents. There is no possibility of the constituents existing without the fact existing. It follows that there is no need for facts as unmereological compositions, facts as “additions to being,” in a phrase from Armstrong. If a fact just is a mereological complex, then it is an “ontological free lunch,” to employ yet another signature phrase of the late Australian. Of course, not just any old mereological sum is a fact; only those with the right constituents.

 

And the same goes for properties: all properties, just in virtue of existing, qualify their bearers. There is no need for a tertium quid such as an instantiation relation to tie a property to its bearer. Nor is there any need for monadic facts as entities in addition to their constituents to do this unifying work. There is no difference between the sum a + F-ness and the fact of a's being F. For this to work, all properties have to be “bearer-specific.” “A property is bearer-specific if and only if it is in its nature to be had by specific bearers.” (90) Suppose it is true that Hargle is happy, and that being happy is “bearer-specific.” We can display the property as follows: __(H) being happy. '__' indicates that the property is unsaturated or incomplete or gappy in something like Frege's sense: if it is had by an individual it is had directly without the need of a connector such as an instantiation relation or Strawsonian nonrelational tie or a Bergmannian nexus. '(H)' indicates that the property is bearer-specific or rather bearer-individuated: if the property is had, it is had by Hargle and nothing else. That the property is had follows from its existence: necessarily, if the property exists, then it is had, had by Hargle and nothing else, and had directly without the service of a tertium quid. What this all implies is that the mereological sum Hargle + __(H) being happy suffices as truthmaker of 'Hargle is happy.' There is no need for a fact over and above this sum. Indeed, as Betti points out, the property alone suffices as truthmaker since it cannot exist unless Hargle exists. (101)

 

Questions and Objections

 

1. Why is Betti's proposal superior to Armstrong's?

 

Betti presents us with an alternative way of thinking about truthmaking facts, namely, as mereological sums whose parts include relata-specific relations and bearer-specific properties. Betti's main point is that “mereological complexes are viable as truthmakers; facts are not needed for the role.” (101) When she says that facts are not needed, she means Armstrongian, middle-period facts. She is not denying that there are truthmakers. Nor is she is denying the existence of facts as long as they are assayed as mereological complexes. If a fact is a complex entity that functions as a truthmaker, then her mereological complexes containing relata-specific relations and bearer-dependent properties are facts, though not in Armstrong 's robust sense. She is denying, or rather refusing to countenance on grounds of theoretical economy, facts as unmereological complexes. Her claim is that there is no explanatory need for facts as the middle-period Armstrong conceives of them, namely, as “additions to being.” Betti may bristle at my use of 'facts' in describing her position but surely there is an innocuous and nearly datanic, as opposed to theoretical, use of 'fact' according to which an individual's having a property, or two or more things standing in a relation, is a fact. Indeed, she needs this use of 'fact' just to state her theory, according to which the fact aRb is identical to the sum a + R + b, when R is relata-specific. On her view facts are a proper subset of mereological sums. That is not a denial of facts, but an acceptance of them. Unfortunately, Betti sometimes expresses herself in a misleading way. She tells us, for example, that “the thought that the world is a world without facts – one in which there is no difference between facts and sums – is shown to be perfectly sensible.” (88) This formulation equivocates on 'fact.' What she wants to say is that the world is without Armstrongian facts, not that the world is without truthmaking facts. It is the latter that are no different from sums, namely those sums whose constituents include relata-specific relations and object-dependent properties.

 

Betti thinks her theory is preferable to Armstrong's. I question whether she is justified in this preference. We face a tough choice. Armstrong's theory violates the extensionality of parthood and countenances unmereological complexes. This is a strike against it. Betti's theory avoids unmereological complexes, thereby upholding the extensionality of parthood, but accepts relata-specific relations and bearer-dependent properties. How plausible is it that all relations are relata- specific and all properties bearer-dependent? Are these notions even coherent? Let's consider the coherence question.

 

2. Against Relata-Specific Relations and Bearer-Dependent Properties

 

Suppose Argle is two feet from Bargle. There is nothing in the nature of either relatum to necessitate their standing in this external relation. Each can exist apart from the relation. And as I see it, there cannot be anything in the nature of the relation itself to necessitate that it be precisely these two critters that the relation relates. So on my view a relational situation such as Argle's being two feet from Bargle involves a double externality: there is nothing in the nature of the terms to dictate their standing in the external relation in question, and there is nothing in the nature of the external relation to dictate the terms. But as Betti sees it, it is the nature of this relation to relate Argle and Bargle and nothing else: the relation cannot exist/be instantiated without relating precisely these two. This implies that “as soon as” (105) the relation exists, it relates Argle and Bargle. If this conception is coherent, it has the desired consequence of undercutting Bertrand Russell's distinction between actually relating relations and those same relations as inert, and with it the distinction between a fact as a real unity of fact-appropriate constituents and the 'mere' mereological sum of those very same constituents. If this works, it puts paid to Armstrong's commitment to unmereological complexes: mereology suffices for truthmakers provided the parts of the sums include relata-specific relations or bearer-dependent properties.

 

It seems to me, however, that the notion of relata-specificity reduces to absurdity by way of the following argument in which R is any relata-specific dyadic external relation, and a and b are its individual relata. (See also my critique of D. W. Mertz in Vallicella 2004.) Generalization beyond the dyadic case is straightforward but unnecessary. Betti's definition of 'external relation' is standard and perfectly serviceable: “A relation is external if and only if it is not grounded in corresponding properties of its relata, that is, is an entity over and above its relata.” (89) An internal relation is then one that is grounded in corresponding properties and is not an entity in addition to its relata. Now to the argument:

 

P1. R is entirely dependent for its existence on both a and b. (Betti's theory of relata-specificity)

 

This is because (i) R cannot exist without being instantiated and thus cannot exist without actually relating some pair of individuals or other, and (ii) R cannot, as relata-specific, relate any pair of individuals other than a, b. If dyadic R were an immanent universal, then it could not exist without relating some pair or other; but it would not necessarily have to relate the precise pair, a, b. R's existence would then not depend on its relating a and b. But as it is, R is a particular (an unrepeatable), not a universal (a repeatable); it is a non-transferable relational trope. It is as particular as the particulars it relates. Its being or existence is exhausted by its particular occurrence, unlike an immanent universal the being or existence of which is not exhausted by its instantiation in a particular case. So R, as a relational trope, is entirely dependent for its existence on the exact relata it has: its being or existence is exhausted by its relating of those exact relata, the individuals a and b. Therefore,

 

C1. R is not distinct in reality from the particular relatedness aRb: R = aRb.

 

Of course, R can be thought of in abstraction from aRb. But R in reality is identical to aRb. You cannot say that they are different because aRb has constituents a, b while R does not. For R exists when and only when it is relating a and b. Apart from them it is nothing at all.

 

P2. The particular relatedness or relational fact aRb is identical to the mereological sum a + R + b, given that R is relata-specific. (Betti's theory) Therefore,

 

C2. R is identical to the sum a + R + b. (from C1 and P2 by Transitivity of Identity)

 

P3. No proper part of a mereological sum having two or more members is identical to the sum of which it is a proper part. (Principle of mereology) Therefore,

 

C3. R is not identical to the sum a + R + b. (from P3) Therefore,

 

C4. R is and is not identical to the sum a + R + b. (from C2, C3) Contradiction! Therefore,

 

C5. Either P1 or P2 is false; either way, Betti's theory fails.

 

Betti will presumably reject (C1). But how? She tells us that it is the nature of R to relate exactly a and b. Now if it is the nature of R to relate exactly these relata, then it is intrinsic to R that it do so. But then R is intrinsically relational, relational in and of itself. If this is neither contradictory nor magical, then it involves importing mind (intentionality) into the bowels of R. For if it is intrinsic to R that it relate exactly a and b, then R, quite apart from actually relating a and b, 'pre-selects' a and b as its relata. But this is what mind in its intentional states does. Such states are intrinsically relational: it is their nature to be of or about items that need not exist for the states to be of or about them. But surely there is no intentionality within the non-transferable relational trope R!

 

But what is the alternative? Will we be told that a and b are constituents of R? But then R is identical to aRb, when it cannot be given that aRb is a + R + b.

 

Now let's consider bearer-dependent properties. Suppose we grant, along with Armstrong (2004, 49), that some mereological complexes are truthmakers. Is it not also the case that some are not? Suppose that Gargle is lachrymose but Hargle is not. Then the following sum exists: Hargle + __(G)being lachrymose. The sum exists because its two parts exist. But the parts are not connected to form a truthmaker. This implies that on Betti's account there are two sorts of mereological sum: those that are truthmakers and those that are not. It also implies that what makes a mereological sum a truthmaker is not its being a mereological sum. What makes a sum a truthmaker is the nature of its members. Thus what makes Hargle + __(H)being happy a truthmaking sum is its second member.

 

But this second member has a rather intricate and puzzling structure. It is a bearer-individuated property, a property that exists only if instantiated by Hargle. Hargle can exist without being happy, but the property in question cannot exist unless Hargle exists. It is in the nature of the property to qualify precisely Hargle “as soon as it exists,” (105) i.e., as soon as the property exists. But when does it exist? When Hargle instantiates it. So it is not as if the property has its individuated nature apart from its being instantiated; rather, it receives its individuated nature by being instantiated by Hargle. It is only the existing Hargle that can make the property individuative of precisely Hargle and nothing else. So Hargle supplies the nature that makes the property Hargle-specific, or rather Hargle-individuated.

 

Does this not smack of absurdity? The nature of an entity is intrinsic to it; it cannot consist in a relation to an item external to it. So it cannot be instantiation by Hargle that gives the property its nature. If, on the other hand, Hargle were a constituent of the property in question, namely, __(H)being happy, then it would make sense to say that it is the nature of the property to be instantiated by Hargle. But Hargle is not a constituent of the property; otherwise the property would not be a property but the fact of Hargle's being happy.

 

Betti seems to face a dilemma. Either Hargle is not a constituent of the property or he is. If Hargle is not a constituent of the property, then the property has no nature that makes it dependent on precisely Hargle and nothing else. But if Hargle is a constituent of the property, then the property is a fact.

 

If Betti's account is incoherent, as I have just argued that it is, then it cannot be superior to Armstrong's even if Armstrong's is also incoherent. I should make it clear that I am not defending Armstrong; I admit that his view of facts is problematic. In fact, I argue that it is incoherent in Vallicella 2016. My point is that Betti's theory is not an acceptable replacement for it. Even if her theory is not incoherent, it is problematic as I will now further demonstrate.

 

3. Digging Deeper: Further Questions about Betti's Theory of Relations

 

Betti faults me (92-93) for failing to distinguish between externality and relata-unspecificity. A relation is external just in case it is not “grounded in corresponding properties of its relata . . . .” (89) “A rela tion is relata-unspecific if and only if it is not in its nature to relate specific relata.” (90) I fail to distinguish externality from relata-unspecificity in that I hold that, in Betti's words, “A relation is external if and only if it could have related another pair (or triple, quadruple, etc.) of relata.” (93, citing Vallicella 2002, 14-15, 31; 2004, 164). As I see it, no external relation has a nature that dictates that it relate only a particular pair, triple, quadruple, etc. of relata. As against this, Betti envisages the following possibility: an external relation such as being two feet from that holds, if it holds at all, between Argle and Bargle but cannot hold between any other pair of relata. The relation is external in that there is nothing in the natures of the relata that dictates that they stand in the relation in question; the relation is relata-specific in that there is something in the nature of the relation to dictate that, if it holds, it holds only between Argle and Bargle.

 

Now if Betti's scenario is possible, then I have blundered by conflating externality and relata-unspecificity. But while I grant that Betti's 'possibility' is combinatorially possible given her definitions, it is not metaphysically possible. I gave an argument above. So my conflation of externality and relata-unspecificity strikes me as justified.

 

I found Betti's theory of relata-specific relations (which draws on the work of her student Jan Willem Wieland) obscure and in need of further development. One intriguing suggestion is that “relata-specific relations can still be universals.” (91) Now there is a wholly uncontroversial sense of 'relata-specific universal' which Betti does not intend. Consider the universal taller than. This is a dyadic relation that is instantiated by ordered pairs of objects, but not just by any old pair. The pairs must be pairs of things having height. Taller than is thus specific to all and only such pairs and not to pairs of numbers or pairs of sets or pairs of propositions or pairs of angels or pairs of acts of thinking. But Betti means something different. She is apparently envisaging the possibility of a relation that is universal but that, say, relates only Guido, Francesca, Giacomo, and Maria in respect of height. Unfortunately, she gives no exemples and I am not sure what she is driving at. She brings this up because she thinks that her solution to the unity problem works whether or not one assays properties as universals or as tropes. (91) But this is all very obscure and here is a lacuna that needs filling.

 

Conclusion

 

My interim verdict with respect to compositional facts is that Betti has not provided a viable mereological alternative to the admittedly untenable facts or states of affairs of Armstrong's middle period.

 

 

REFERENCES

 

Armstrong, D. M. 1978. Nominalism and Realism: Universals and Scientific Realism, vol. I. Cambridge: Cambridge University Press.

Armstrong, D. M. 1983. What is a Law of Nature? Cambridge: Cambridge University Press.

Armstrong, D. M. 1989a. Universals: An Opinionated Introduction. Boulder: Westview Press.

Armstrong, D. M. 1989b. A Combinatorial Theory of Possibility. Cambridge: Cambridge University Press.

Armstrong, D. M. 1993. “A World of States of Affairs”. Philosophical Perspectives, vol.7, 429-440.

Armstrong, D. M. 2004a. “How Do Particulars Stand to Universals?” In D. W. Zimmerman, (ed.), Oxford Studies in Metaphysics, vol. 1. Oxford: Oxford University Press, 139-154.

Armstrong, D. M. 2004b. Truth and Truthmakers. Cambridge: Cambridge University Press.

Armstrong, D. M. 2009. “Questions about States of Affairs”. In M. E. Reicher (ed.), States of Affairs. Frankfurt: Ontos Verlag, 39-50.

Armstrong, D. M. 2010a. Sketch for a Systematic Metaphysics. Oxford: Oxford University Press.

Armstrong, D. M. 2010b. “Reinhardt Grossmann's Ontology”. In Cumpa, J. (ed.), Studies in the Ontology of Reinhardt Grossmann. Frankfurt: Ontos Verlag, 29-43.

Baxter, D. 2001. “Instantiation as Partial Identity”. Australasian Journal of Philosophy, 1979, 449-64.

Bergmann, G. 1967. Realism: A Critique of Brentano and Meinong. Madison: The University of Wisconsin Press.

Butchvarov, P. 1979. Being Qua Being: A Theory of Identity, Existence, and Predication. Bloomington: Indiana University Press.

Butchvarov, P. 1986. “States of Affairs”. In Bogdan, R. (ed.), Roderick M. Chisholm. Dordrecht: D. Reidel, 113-133.

Butchvarov, P. 2010. “Facts”. In Cumpa, J. (ed.), Studies in the Ontology of Reinhardt Grossmann. Frankfurt: Ontos Verlag, 71-93.

Chisholm, R 1976. Person and Object: A Metaphysical Study. La Salle: Open Court.

Cumpa, J. and Tegtmeier, E. (eds.), 2009. Phenomenological Realism Versus Scientific Realism: Reinhardt Grossmann-David M. Armstrong Metaphysical Correspondence. Frankfurt: Ontos Verlag.

Frege, G. 1960.”On Concept and Object”. In P. Geach and M. Black (eds.), Translations from the Philosophical Writings of Gottlob Frege. Oxford: Basil Blackwell, 42-55.

Frege, G. 1976. “Der Gedanke”. In G. Patzig (ed.), Logische Untersuchungen. Goettingen:Vandenhoeck and Ruprecht, 30-53.

Grossmann, R. 1974. “Bergmann's Ontology and the Principle of Acquintance”. In Gram, M. S. and Klemke, E. D. (eds.), The Ontological Turn: Studies in the Philosophy of Gustav Bergmann. Iowa City: University of Iowa Press, 89-113.

Grossmann, R. 1983. The Categorial Structure of the World. Bloomington: Indiana University Press.

Grossmann, R. 1984. Phenomenology and Existentialism: An Introduction. London: Routledge and Kegan Paul.

Grossmann, R. 1990. The Fourth Way: A Theory of Knowledge. Bloomington: Indiana University Press.

Grossmann, R. 1992. The Existence of the World: An Introduction to Ontology. London: Routledge.

Mulligan, K., Simons, P. and Smith, B. 2009. “Truth-makers”. In Lowe, E. J. and Rami, A., Truth and Truth-Making. Montreal: McGill-Queen's University Press, 59-86.

Mumford, S. 2007. David Armstrong. Montreal: McGill-Queen's University Press.

Plantinga, A. 1974. The Nature of Necessity. Oxford: Oxford University Press.

Strawson, P. F. 1950. “Truth”. In Aristotelian Society Suplementary Volume 24, 136-137.

Vallicella, W. F. 2000. “From Facts to God: An Onto-Cosmological Argument”. International Journal for the Philosophy of Religion 48, 157-181.

Vallicella, W. F. 2002. A Paradigm Theory of Existence: Onto-Theology Vindicated. Dordrecht: Kluwer Academic Publishers.

Vallicella, W. F. 2004. “Bradley's Regress and Relation-Instances”. The Modern Schoolman, vol. LXXXI, no. 3, 159-183.

Vallicella, W. F. 2010. “Gaskin on the Unity of the Proposition”. Dialectica 64, 265-277.

Vallicella, W. F. 2016. “Facts: An Essay in Aporetics”. In Calemi, Francesco F. ed, Metaphysics and Scientific Realism: Essays in Honour of David Malet Armstrong. Boston/Berlin: Walter de Gruyter, 105-131, esp. 115-126.

 

1It is an interesting question whether one could be an idealist and also a truthmaker theorist. Consider a Kantian who holds that phenomenal objects and events are “empirically real but transcendentally ideal” to employ a signature Kantian phrase. It seems to me that such a philosopher could maintain a need for truthmakers for some truthbearers, namely those synthetic aposteriori, and thus contingent, judgments about empirical objects and events. It seems one could combine realism about empirical truth with transcendental idealism.

 

Related articles

The Ramsey Problem and the Problem of the Intrinsically Unpropertied Particular

Van Inwagen, Properties, and Bare Particulars

Bare Particulars versus Aristotelian Substances

Comments

[Javier Cumpa]

Hello Bill,

I find your review of Betti's terrific book just excellent. It will be great to have it in the next issue of Metaphysica!

Here you can find some comments:

1. With respect to Betti's assumption that Armstrong's facts are designed to solve the unity problem. As I understand Armstrong's factualism (from the 70ties to 90ties), it is intended to solve two related problems: (a) the problem of universals and (b) truthmaking. Betti seems to be focused only on truthmaking to criticize Armstrong's view on non-mereological composition. But the big problem of non-mereological composition in Armstrong's ontology arises in the context of the problem of universals. Armstrong needs to defend that facts have a non-mereological mode of composition to defend his identity view by which properties are capable of being constituents of more than one fact, i.e, multiple location. So, Armstrong would always resist Betti's general conclusion of reducing facts to sums, because he needs non-mereological composition to solve the problem of multiple location of property universals, and the mereological composition of mereological sums, according to Armstrong, cannot solve it.

2. As to sums with Betti's relata-specific relations, it seems to me that if the entities of a mereological sum are related by a certain relation, then the resulting entity will be a mereological whole with parts related to each other, but not a mereological sum where the parts of the whole are not related to each other. In this regard, I wonder: to what extent Betti's approach to facts really avoids non-mereological composition? Is not she really offering a new kind of it? For she seems to be defending that there is a kind of whole without mereological composition, i.e., a kind of whole governed by special bearers and special relations, a sort of whole that is not governed by the laws of mereological calculus.

3. You write: "An internal relation is then one that is not grounded in corresponding properties and is not an entity in addition to its relata." Is this definition of an internal relation right?

Many thanks for this amazing review!

All the best,
Javi

[BV]

Thanks for the comments, Javi. I'm glad you liked the review. I will get a revised electronic copy to you in a couple of days.

As for #3, You are right: there is a typographical error. It should read: "An internal relation is then one that IS grounded in corresponding properties and is not an entity in addition to its relata."

For example, suppose A is red and B is (the same shade) of red. Then A and B stand in the *same color as* relation. This relation is grounded in the intrinsic property of being red shared by A and B. It is not an entity over and above its relata and their properties.

[BV]

As for #2, you are definitely on to something. Suppose a is a particular and __F is a bearer-dependent property. This property exists iff it is attached to a. The particular can exist without the property, but the property cannot exist without that very particular. So there cannot exist the sum a + __F "where the parts of the whole are not related to each other."

Your point seems to be that the composition in Fa is unmereological.

Part of the puzzle here is that we tend to think of mereological sums as collections in which abstraction is made from connections among the parts. So if there is the sum a + F, this cannot be identical to Fa -- in which case something more is required to get to Fa. And this will be an unmereological form of composition.

Is this a fair summary of your point?

[Javier Cumpa]

You are more than welcome, Bill!

Thanks also for the clarification of the definition of an internal relation.

As to #2, yes, that is a fair summary of my point.

[Astute]

I have simplified the aporia we discussed by email, and which I believe is related to the points you raise in the review above.

1. Language signifies reality (C facts?)
2. What language signifies has a grammatical structure (P facts?)
3. Reality does not have a grammatical structure
All three have strong demands on acceptance, but all three cannot be true.

(1) If language did not signify reality, how could we use language to communicate our thoughts beliefs about reality?
(2) If meaning is compositional, and the component parts correspond to the words, then what language signifies has a grammatical structure.
But (3) it is implausible that reality, i.e. the world as it actually is, independent of any representation of it, has a grammatical structure.

[BV]

I don't like the word 'signifies.' How many analytic philosophers use it?

Here is my preferred aporetic triad:

A. Some sentences are true in virtue of their correspondence with extralinguistic reality.

B. If so, then reality must have a sentence-like structure.

C. Reality does not have a sentence-like structure.

This is a nasty little puzzle since this trio is inconsistent. And yet one can make a plausible case for each member of the trio.

Do you want to discuss the problem in this form?

[Astute correspondent]

>>Do you want to discuss the problem in this form?

That will do.

Wittgenstein: "a name ought really to signify a simple".

[Astute]

So which of your A B C above do you reject? I am tempted to reject B. If there is a correspondence as asserted by A, then language and reality have to correspond in virtue of something. Perhaps that 'something' is not the sentence structure, but something else?