Introduction: the logical problem of evil

In an article in this issue, Kenneth Pearce (2020) suggests a novel solution to the ‘logical problem of evil’. That is, he defends the consistency of the obtaining of some evil with the existence of an omnipotent and perfectly good creator. This article argues that Pearce fails to solve the logical problem of evil.

The logical problem of evil that Pearce addresses is the (alleged) inconsistency between the existence of God – understood, inter alia, as perfectly good and omnipotent – and the existence of evil in the world. The idea is that since God is perfectly good, then God doesn't want there to be evil in the world, and since God is omnipotent, then God gets whatever God wants, and so if God exists there is no evil in the world. In other words, the assumption that God exists clashes with the claim that there is evil in the world. Pearce emphasizes that the logical problem of evil is concerned with the existence of any evil whatsoever, that is, with the (alleged) inconsistency of the existence of God with there being even a minor evil, and he attempts to undermine this alleged inconsistency, that is, to show that God's existence is consistent with there being some evil in the world. Specifically, Pearce points out a kind of evil that, so he argues, is consistent with God's existence.

Pearce, like others, distinguishes the logical problem of evil from the ‘real problem of evil’. In fact, there are a few versions to the real problem of evil, and Pearce refers to that one of them which is concerned with the horrendous evil and suffering that fills the world.Footnote ¹ Since his argument is not concerned with this evil and suffering, Pearce does not take it to address this real problem of evil, but he suggests that his argument does make some progress towards addressing it. I will not discuss the respects in which this is so according to Pearce, but after arguing that he fails to show that the evil in question may be justified, I will point out another difficulty with Pearce's argument – in fact, with Pearce's very approach to the logical problem – that concerns the relation between the two problems of evil.

Pearce's solution to the logical problem of evil

Pearce's general strategy for solving the logical problem of evil is similar to that of Plantinga's (Reference Plantinga1974) free-will solution. This strategy consists in identifying a proposition that entails both that God exists and that evil exists, and showing that this proposition is consistent. The proposition that Pearce identifies is:

Pearce infers the consistency of (M) from his claim that a certain theory of mind is both self-consistent and consistent with God's existence. The theory in question is the naturalistic teleological theory of original intentionality.Footnote ² (Pearce refers to this theory as ‘NFT’ – for ‘naturalistic-functional-teleological’ – but for a reason to be revealed later, I will omit the mention of functionality and refer to this theory as ‘NT’.) According to NT (ibid., 427):

1. 1. A state S of a finite being B is non-derivatively about an object x iff S has the function of indicating x for B.

2. 2. A state S of a finite being B has a function for B only if B's exemplification of S can be given a causal-historical explanation in terms of adaptive advantages that accrue to B when S successfully performs that function.

Pearce's argument does not hinge on the claim that NT is true. What it requires is that it (a) be self-consistent, (b) be consistent with God's existence, and (c) entail that some evil is necessary for the existence of created minds, an evil that is outweighed by the good that is involved in the existence of created minds. The main burden of the argument is to show that NT entails that some evil is necessary for the existence of created minds. Here is how Pearce attempts to do this.

According to NT, original (that is, non-derived) intentionality requires such correlations between representations and their represented objects that are selected for on the grounds of their contribution to the proper functioning of the system in question. Now these correlations must be selected for from among alternatives, and so there must be instances – in the history of the individual organism or in the history of its species – in which the correlations fail, instances that result in malfunctions. Such malfunctions may be, for example, false alarms about the presence of predators, and the system that sustains the correlations is selected for because having a higher rate of false positives is a lesser evil for the organism than even a modest rate of false negatives. The idea, in short, is that original intentionality necessarily involves malfunctions, and malfunctions are evils. It is then impossible for creaturely intentionality, Pearce concludes, to exist in an evil-free world.

This in itself does not amount to or entail the consistency of (M) – the claim that God decided to create minds although it is impossible that created minds exist in the absence of evil. Indeed, Pearce supports it by making the further claims that (a) NT is self-consistent, (b) NT is consistent with God's existence, (c) the existence of minds is intrinsically valuable or (d) it is a necessary precondition for other goods, and (e) the evil that is necessary for the existence of creaturely intentionality is outweighed by the goods that depend on the existence of minds. Pearce assumes that (a) is true, and provides arguments for (b) and (e). I will not dwell on these arguments, and grant that they are effective. Similarly, I will grant (a) and Pearce's claim that those malfunctions are evils. (I will refer to this latter claim in the fourth section.) As to (c) and (d), Pearce takes them to be extremely plausible. I will argue that they may only be true in a sense in which they cannot serve to establish the consistency of (M) and, thus, the claim that God may be justified to create a world that contains some evil.

Intentionality, evil, and a posteriori necessity

Pearce insists that if NT is true, then it is necessarily true. That is, it purports to be a necessary a posteriori truth. (Of course, taking it to be a priori necessary is implausible.) This point is important. It well accords with the rationale of Pearce's argument, which is that of showing that some evil is metaphysically unavoidable (not per se but) for bringing about some valuable state of affairs whose valence overrides the negative valence of this evil. If this is true, then, according to this idea, God cannot be blamed for creating a world that contains such evil. Of course, this approach cannot serve to defend a notion of an omnipotent God in a sense in which God can also do what is metaphysically impossible, but Pearce, like others who address the problem of evil, is satisfied with a notion of God who is limited by both logic and metaphysics. So let's remain in this framework,Footnote ³ and turn to consider in what sense NT can be said to be necessary and, consequently, in what sense the evils that are implicated by its truth can be justified.

The notion of necessary a posteriori statements (influentially defended by Kripke (Reference Kripke1980) and Putnam (Reference Putnam1975) ) is certainly not uncontroversial, and philosophers who reject it would not be persuaded by Pearce's suggested solution, but for the sake of argument I will grant its legitimacy. The classical example of an a posteriori necessary statement (one that Pearce also mentions) concerns the statement that water is H₂O. Now to take this statement to be necessary a posteriori does not mean or imply that water (or H₂O, for that matter) necessarily exists. Neither does it mean or imply that there can be no watery stuff – that is, stuff with the macro properties of water – that is not H₂0. Rather, what taking this statement to be necessary a posteriori means is that stuff with the macro properties of water that is not H₂0 is not water. For example, twater – that TwinEarth liquid whose chemical structure is XYZ rather than H₂0 but that shares water's macro properties (Putnam's (Reference Putnam1975) ) – isn't water.

How can this point be applied to the case of intentionality and, specifically, to NT? Here is the first point on the matter. Suppose that intentionality is that teleological relation that NT takes it to be, and that the identity between intentionality and that NT relation is necessary a posteriori. So NT's being a necessary a posteriori truth does not mean or imply that the selection history that this thesis takes to constitute intentionality necessarily exists. On the assumption that NT is a necessary a posteriori truth, it necessarily exists only if intentionality exists.

Does this fact pose a problem for Pearce's argument? The claim that intentionality necessarily exists, and, a fortiori, the claim that NTy intentionality – that is, intentionality that is constituted the way NT takes it to be constituted – necessarily exists,Footnote ⁴ do not form part of this argument. What Pearce argues is that the state of affairs in which intentionality exists necessarily involves some evil, and that this state of affairs is both self-consistent and consistent with the existence of (an omnipotent and perfectly good) God. It is consistent with the existence of God since it is necessary for some goods that outweigh this evil. But my claim that taking NT to be a necessary a posteriori truth does not entail the necessity of intentionality is only meant to be the first step in an argument to the effect that Pearce fails to show the consistency of the idea that some evil is necessary for this good. Specifically, realizing what the supposed necessary a posteriori nature of NT is, we can see that relying on it cannot serve to support this consistency claim as well.Footnote ⁵ Here goes.

Note, first, that it isn't clear why Pearce holds that the evil that creaturely intentionality involves, if it is constituted the way NT asserts it is, is necessary for the existence of created minds. One possible way to defend this idea is to argue that minds are necessarily intentional. Pearce does not argue for this claim, but, shifting from the claim that ‘it is impossible for creaturely intentionality to exist in a world without evil’ to the claim that ‘the existence of created minds requires at least some evil’ (Pearce (2020), 429) presupposes it. Pearce dos not justify this shift. Of course, it would have been justified had minds been intentional by definition, but definitions cannot resolve substantial issues. Specifically, if it is possible for those things called ‘minds’ to exist if their intentionality is subtracted from them, it matters not, for our issue, whether those intentionality-less things are properly called ‘minds’. What seems to be more relevant to Pearce's argument is the idea that it is a posteriori necessary that minds are intentional (and so it is impossible for those things to exist without intentionality).Footnote ⁶ Of course, this idea is similar to the idea that it is necessary a posteriori that intentionality is constituted the way described by NT, and hence that it necessarily involves evil. The conjunction of these two ideas entails that evil is necessary for the existence of created minds. What, then, is the problem with Pearce's reliance on the supposed necessary a posteriori nature of the connection between minds and evil that is mediated by intentionality?

As noted, the claim that water is necessarily H₂O not only does not mean or entail that water necessarily exists; it also does not mean or entail that there cannot be stuff with the macro or identifying properties of water that is not H₂0. Rather, what taking this statement to be necessary a posteriori means is that watery stuff that is not H₂O is not water. Twater, for example, isn't water. In order to apply this point to the relation between intentionality and NT, we have to determine, first, what the identifying properties of intentionality – the analogues of watery properties or, generally, of macro properties – are. The rationale that standardly underlies naturalistic-reductionist accounts of intentionality of the NT kind is that these identifying properties are the functional roles of our intentional states (see Papineau (Reference Papineau1993) and (2001) ). Braddon-Mitchel and Jackson characterize this idea thus: ‘Belief that p is the theoretically interesting state that actually plays the folk functional roles distinctive of belief that p (from reflection of what we master when we master intentional vocabulary)’ (Braddon-Mitchel & Jackson (Reference Braddon-Mitchell and Jackson1997), 481). Determining what this functional role is is supposed to be a matter of conceptual analysis, whereas identifying it with some specific naturalistic world-mind relation (the one that in fact connects intentional states with their objects), such as that historical-teleological relation, is supposed to be an a posteriori matter. An account of intentionality along such lines is not a functional role account of intentionality, since the functional role only plays the role of reference fixer, and is not supposed to be the nature of intentionality.Footnote ⁷

So, given the assumption that the identifying properties of intentionality are the functional roles of intentional states, does taking NT to be a necessary a posteriori truth mean or imply that there cannot be internal states of an organism that have the functional roles of intentional states but that do not involve what, according to NT, intentionality necessarily involves (that is, bearing that adaptive historical relation – call it ‘NTy relation’ – to the environment)? No. taking NT to be a necessary a posteriori truth only means that such states are not intentional states. Indeed, why can't there be internal states with the functional roles of intentional states that do not have the adaptive history that NT takes intentionality to involve? And certainly, the assumption that NT is a necessary a posteriori truth and the claim that our internal states may not bear that adaptive historical relation to worldly items do not jointly entail that they may not bear any (naturalistic) relation to them. So there may be internal states with all properties that are non-intentional and are not constituted by intentional properties, including functional properties, that are related to the environment by means of (naturalistic) relations to the environment different from the NTy relation. This is a metaphysical possibility, and an omnipotent God (even one that is constrained by logical and metaphysical impossibilities) could create human beings with such internal states.

If NT is a necessary truth, then internal states that bear to environmental items only non-NTy naturalistic relations – for example, asymmetric co-variation (see Fodor (Reference Fodor1987) and (Reference Fodor1991) ) – are not intentional, and if mentality is necessarily intentional then these states are not even mental and their possessors do not have minds. This seems to me to be a strange consequence, but I do not rely on its strangeness for rebuffing Pearce's argument. The point I wish to stress is different. Consider entities that are identical to minds in functional role and in any other respect except for bearing to environmental items some naturalistic relations different from NTy relations (e.g. relations of asymmetric co-variation). Call such entities ‘twinminds’, and call their relations to environmental items ‘twintentionality’.Footnote ⁸ The crucial issue is whether God could be justified in creating humans with (NTy) intentionality and (NTy) minds when God could have created creatures with twintentionality and twinminds that do not involve the evils that (NTy) intentionality and (NTy) minds involve. There appears to be no justification for such a choice on the part of God, for twinminds may be as valuable as minds.

Pearce claims that plausibly minds are intrinsically valuable (Pearce (2020), 429). We may perhaps suppose that they are, but what is implausible is to suppose that minds have intrinsic values while twinminds lack intrinsic value. To suppose this is to suppose that the intrinsic value of minds is rooted not in their intrinsic nature (minds and twinminds share intrinsic nature – their only difference is external) but in their NTy relations to the environment; that NTy relations but not any other naturalistic relations to the environment are the source of the intrinsic (!) value of the mental. This appears to make no sense.

Claiming that minds are intrinsically valuable is, for Pearce, one possible step in the way to support the view that ‘the existence of created minds is a sufficiently great good that it is consistent with perfect goodness to bring about the existence of created minds even if this implies the existence of at least some evil’ (ibid.). Another step that Pearce undertakes for supporting this view is claiming that ‘the existence of minds is a necessary precondition for an enormous variety of other goods, such as pleasure, virtue, love, and aesthetic appreciation’ (ibid.). He may similarly say that (whether or not intentionality is necessary for mentality) the existence of intentionality is a precondition for such goods. These two claims may seem plausible: isn't intentionality a precondition for love and aesthetic appreciation, for example? However, when understood in the senses that are required for making Pearce's argument effective, these claims are not true. For in the context of this argument twinminds and twintentionality are not minds and intentionality, respectively – only NTy minds and NTy intentionality are minds and intentionality, respectively. So the precondition for these goods according to Pearce is the existence of minds and intentionality as distinguished from twinminds and twinintentionality – as distinguished from any counterparts that differ from NTy mind and NTy intentionality only in some relational-environmental difference, however slight it may be. The former two qualify for the job; the latter two do not. Pearce gives us no reason to believe that only one specific naturalistic relation can enable us to have all these goods. And recall, minds and twinminds share intrinsic natures, so there need not be any intrinsic difference between minds and their non-NTy counterparts. Thus, if Pearce's claim that minds necessarily involve (NTy) intentionality and therefore evil is true, God should not have created minds; rather, God should have created twinminds that do not involve evils.Footnote ⁹ (Of course, one who maintains that minds necessarily involve (NTy) intentionality may argue that love, virtue, aesthetic appreciation, and the like cannot be attributed to twinminds. However, in accordance with what I've just argued, this doesn't matter for our issue, since the crucial point is whether, for example, twinlove is less valuable than love, and there is no reason to believe that it is.)

Thus, there is no reason to suppose that the evils that NT presupposes are necessary for the goods that minds involve. An omnipotent God could bring about such goods without bringing about evils of the NTy kind.

One might think that having (NTy) minds and (NTy) intentionality is necessary for our optimal functioning and is thus instrumental for some goods. (Pearce himself does not argue along such lines.) However, there is no reason to assume that bearing precisely this specific relation – the adaptive historical relation – and thus, on Pearce's assumption, bearing the intentional relation – to environmental items is optimal for our proper functioning. There is nothing inconsistent in the idea of intentionality-less creatures who are behaviorally and functionally identical to us: they lack intentionality because their internal states do not bear the ‘right’ relation to worldly items. Relatedly, there is no reason to suppose that internal states that bear some non-NTy naturalistic relations to the environment that do not presuppose evils such as malfunctions (e.g. naturalistic relations such as asymmetric co-variation) are necessarily less sensitive to the environment and hence less efficient in making us coping with it than (NTy) intentional states. (To claim that adaptive history is necessary for organisms' optimal functioning, so that for ensuring our optimal functioning God cannot but endow us with NTy intentional states, is to put on God limitations beyond metaphysical necessities, and thus to empty the notion of an omnipotent God.)Footnote ¹⁰

Since Pearce has not ruled out the possibility of there being non-NTy analogues of minds the constitution of whose intentionality does not involve evils that are the sources of no lesser good than minds, he has not provided us with a reason to suppose that a perfectly good and omnipotent God can create evil-involving NTy minds. So Pearce has not established the consistency of (M), and has not solved the logical problem of evil.

The logical problem and the real problem: a difficulty with Pearce's very approach

Though solving the logical problem of evil falls short of solving the real problem of evil, Pearce claims that his solution to the logical problem puts us in a better position vis-à-vis the real problem. I will not discuss the respects in which this is so according to Pearce, but rather point out another difficulty with Pearce's argument – in fact with Pearce's very approach – that concerns the relations between the two problems of evil.

Pearce does not attempt to reconcile the existence of horrendous evil and suffering – which he takes to underlie the real problem of evil – with God's existence. His solution to the logical problem of evil concerns evils of another kind – the malfunctions of our representational systems, which he claims to be evils. Now recall that Pearce's strategy for dealing with the logical problem involves the idea that some evils are necessary for bringing about greater goods. Pearce's claim that our representational systems are subject to evils may worsen the dialectical situation of one who wishes to solve the real problem of evil by employing the same strategy. For this claim implies that one has to take into account not only the horrendous evils but also those ‘representational’ or NTy evils. So in the best case (from the theist perspective), everything remains the same: Pearce points out a feature that appears to be a difficulty for theism, but that (if Pearce successfully shows that it is necessary for a greater good) isn't.

More importantly for our present concern, a similar consideration exposes another difficulty for Pearce's solution of the logical problem itself. We may wonder whether the comparison between the ‘representational’ or NTY evils and the good that is involved in the existence of created minds is the relevant comparison. Why focus on this good? Evil that brings about good but also greater evil cannot be justified. So if, in light of the existence of the evil that Pearce points out (those representational malfunctions), we wish to justify the creation of minds, we should not compare this evil with this good. Rather, we should examine the overall balance of good and evil, an examination which should take into account, on the side of evil, not only the evil that NTy intentionality brings about, but also, for example, the familiar horrendous evil and suffering that give rise to the real problem of evil. In other words, for Pearce, minds (qua intentional) necessarily involve some evils, but these evils are outweighed by the greater good that minds involve, hence the creation of minds that involves those representational evils is justified; but if minds also involve, at least sometimes, other evils – for example, such that do not arise from people's own errors and malfunctions but, say, from natural disasters that make them suffer for the rest of their lives, then the creation of those minds that involve both those representational evils and those horrendous evils would only be justified by showing that the sum of those two kinds of evils is outweighed by the goods that minds involve. And of course the task of showing this is a task of a different order of magnitude from that of showing that those representational evils themselves are outweighed by those goods.

Some clarification of the framework in which the issue arises is in order. The real problem of evil in the version under discussion concerns the justification of severe evil. The problem is especially challenging for theism if the evil is so great as to make life not worth living – if the balance between good and evil in life is negative. (On this issue see Adams (Reference Adams1999).) It is difficult to think of a plausible explanation for how such a severe situation could be part of any divine plan. But even if the balance is positive, theism must account for the existence of evil. Theism must justify the existence of any evil, however minor it may be. This forms the logical problem of evil, and it appears to involve no quantitative dimension. However, Pearce's justification for the existence of some specific evil by appealing to its role in enabling some good does involve a quantitative dimension, for, of course, that good should be greater than the evil. Still, what I have just argued is that its being greater than the specific evil in question – the one that intentionality is supposed to involve necessarily – is not enough for the theist justification; a total positive balance of good and evil is necessary for justifying it. If the overall balance for minds is not positive, it would not fit with any divine plan – a perfectly good and omnipotent God would not create minds whose balance of good and evil is negative.Footnote ¹¹

Thus, it is not only that, even on the assumption that Pearce's argument for the consistency of NTy evils with God's existence is effective, the theist should still address all other instances of evil and suffering. An effective argument for this consistency itself must adequately address the existence of these other instances. Pearce treats the logical problem of evil in isolation from the real problem, and believes that his treatment of it puts us in a better position vis-à-vis the real problem. But if what I argue in this section is along the right lines, such an isolation of the logical problem within the framework of Pearce's reasoning is impossible: addressing the logical problem in Pearce's way requires us to address the real problem of evil – to reconcile God's existence with the existence of those horrendous evils as well. And of course, if the theist has at her disposal an effective solution to the real problem of evil, she does not have to bother about the logical problem.Footnote ¹²