In the Fourth of his Meditations on First Philosophy, Descartes confronts an epistemological version of the problem of evil focusing on the evil of error. Descartes knows that he sometimes errs in his search for truth, and this leads him to doubt that he was created by an omnipotent, morally perfect God. The bulk of the Fourth Meditation presents Descartes's theodicy, his response to this ‘problem of error’.
Descartes's theodicy appears at first glance to have two serious flaws. First, it seems very unsystematic. Descartes seems to borrow from a grab-bag of traditional theodicy strategies,Footnote 1 veering wildly from one to another with little concern for the overall coherence of his response: he eschews any knowledge of divine plans, and then immediately proposes a divine purpose for error; he suggests that God has to allow error so that we can have free will, and then claims that we could be free even if it were impossible for us to err. Second, Descartes's theodicy can seem unmotivated: it is not obvious that Descartes needs to construct a theodicy to achieve his philosophical goals. When he proves God's existence in the Third Meditation, doesn't Descartes effectively show that clear and distinct perception is reliable? If so, then his theodicy seems to be (in the words of two recent commentators) ‘a quasi-theological interlude’Footnote 2 in his reasoning ‘rather than an intrinsic part of the argument’.Footnote 3 Little wonder that many scholars devote scant attention to the Fourth Meditation.Footnote 4
This paper will argue that, contrary to initial appearances, Descartes offers a relatively unified, systematic theodicy that contributes substantially to the philosophical progress of the Meditations.Footnote 5 I suggest that Descartes tackles two different arguments from error against the existence of an omniperfect God. Either of these arguments, if successful, would constitute a rational proof of God's nonexistence on a par with Descartes's Third Meditation proof of God's existence, and would therefore lead Descartes into an antimony of reason. To vindicate reason, Descartes must show that these arguments from error fail. He tries to block the atheistic arguments by offering a mixed theodicy: he takes the value of free will to justify an omniperfect God in allowing our actual errors, and he thinks that the overall value of the universe as a whole may justify God in giving us the ability to err. Throughout, I focus on the Fourth Meditation, but use parallel passages from the Principles of Philosophy and other texts to clarify Descartes's position. Most of the paper concentrates on displaying the underappreciated virtues of Descartes's theodicy, but the final section notes a problem with it.
Descartes's problem of error
By the end of the Third Meditation, Descartes takes himself to have proven that God exists. By ‘God’ Descartes means an omniperfect being: ‘the possessor of all the perfections … who is subject to no defects whatsoever’ (AT 7:52/CSM 2:35).Footnote 6 God has all perfections and lacks all imperfections. Therefore, God ‘cannot be a deceiver, since it is manifest by the natural light that all fraud and deception depend on some defect’ (Ibid.). Early in the Fourth Meditation, Descartes notes: ‘since God does not wish to deceive me, he surely did not give me [a faculty of judgement] which would ever enable me to go wrong while using it correctly’ (AT 7:53–54/CSM 2:37–38). This conclusion follows from three assumptions: (1) an omniperfect God exists; (2) God created Descartes's faculty of judgement; and (3) God is veracious.
In what I call the ‘problem passage’, Descartes calls (1) into doubt:
There would be no further doubt on this issue were it not that what I have just said appears to imply that I am incapable of ever going wrong. For if everything that is in me comes from God, and he did not endow me with a faculty for making mistakes, it appears that I can never go wrong. And certainly, so long as I think only of God, and turn my whole attention to him, I can find no cause of error or falsity. But when I turn back to myself, I know by experience that I am prone to countless errors. (AT 7:54/CSM 2:38)
‘Going wrong’ means theoretical error: believing the false when ‘the task now in hand does not involve action but merely the acquisition of knowledge’ (AT 7:22/CSM 2:15). Descartes concerns himself here neither with moral error (sin)Footnote 7 nor with the false beliefs we sometimes acquire when practical needs force us into hasty judgement.Footnote 8 In response to the problem passage, Descartes insists that such error gives us ‘no call to doubt [God's] existence’ (AT 7:55/CSM 2:38), indicating that his worry is about premise (1), not (2) or (3). Descartes defines God (in part) as ‘the creator of all things other than himself’ (AT 7:45/CSM 2:31),Footnote 9 and insists that ‘it is impossible to imagine that [God] is a deceiver’ (AT 7:144/CSM 2:103) because such an idea is ‘self-contradictory’: ‘the form of deception is non-being, towards which the supreme being cannot tend’ (AT 7:428/CSM 2:289).Footnote 10 Descartes conceives of God as a veracious creator. The question is whether any object exemplifies Descartes's concept.
The problem passage raises an analogue of Augustine's problem in On Free Choice of the Will: ‘We believe that everything that exists comes from the one God, and yet we believe that God is not the cause of sins.’Footnote 11 Similarly, in what I will call his ‘causal argument’ from error, Descartes reasons as follows:
(1) If there were an omniperfect God, He would cause everything in me.
(2) An omniperfect God would not cause error.
(3) So, if there were an omniperfect God, there would be no error in me.
(4) There is error in me.
(5) So, there is no omniperfect God.
As we will see, Descartes responds to this argument by ‘looking for the cause of these errors’ (AT 7:38/CSM 2:38) and finding that it is not God.
The problem passage also implicitly refers back to what I call the ‘general argument’ from error, which Descartes states in the First Meditation: ‘If it were inconsistent with [God's] goodness to have created me such that I am deceived all the time, it would seem equally foreign to his goodness to allow [permittere] me to be deceived even occasionally; yet this last assertion cannot be made’ (AT 7:21/CSM 2:14). Descartes knows that he sometimes errs. Therefore, either it is not foreign to the goodness of an omniperfect creator to deceive him all the time, or there is no such creator. By the time Descartes refers back to this problem in the Fourth Meditation, he is confident that God's goodness would prevent God from giving Descartes a deceptive nature. Therefore, his experiences of occasional error call God's existence into question:
(i) An omniperfect being would not allow even occasional error.
(ii) There is occasional error.
(iii) So, there is no omniperfect being.
Unlike the causal argument, which assumes merely that God would not cause error, this argument assumes that God would not even allow anyone else to cause error.
Two considerations show that Descartes's arguments from error are not merely evidential arguments aiming to show that God's existence is very improbable on the available evidence, but logical arguments aiming to show that God's existence is impossible (given error).Footnote 12 First, Descartes seems to dismiss probabilistic reasoning in the First Meditation. He resolves to ‘hold back my assent from opinions which are not completely certain and indubitable just as carefully as I do from those which are patently false’ (AT 7:18/CSM 2:12).Footnote 13 Second, Descartes takes himself to have demonstrated (a posteriori) God's existence in the Third Meditation. If the causal and general arguments showed merely that the existence of God is unlikely on the available evidence, they need not call this earlier demonstration into doubt. Taking them together with the earlier demonstration, Descartes could simply conclude that something very unlikely (the coexistence of God and error) was nevertheless true.
But as logical atheistic arguments, Descartes's arguments from error purport to be a posteriori demonstrations that an omniperfect God does not exist – demonstrations resting on premises that are just as certain by Cartesian standards as the premises of his theistic proof. A successful logical argument would create for Descartes an antinomy of pure reason: reason would show that an omniperfect God both does and does not exist. This would be a disaster for Descartes, since one of his goals in the Meditations is to show, in Harry Frankfurt's words, ‘that reason is reliable, in the sense that it does not betray itself by providing reasons for doubting its own reliability’.Footnote 14 Descartes seeks to show that one clear and distinct perception cannot ‘contradict another in the way that one sensory perception may contradict another’.Footnote 15
It should now be clear why, in the Synopsis of the Meditations, Descartes says: ‘In Meditation IV it is proved that everything that we clearly and distinctly perceive is true’ (AT vii 15/CSM 2:11). Until he can eliminate the threat of antimony, he cannot be certain that an omniperfect God truly exists, despite his Third Meditation proof. For why should we trust any rational proof if reason is ‘hopelessly unreliable and inconsistent’?Footnote 16 Only by constructing a theodicy can Descartes prove that reason is reliable, that God truly exists, and hence that all his clear and distinct perceptions are true. The Fourth Meditation theodicy is not an unnecessary detour into theology, but the point at which Descartes tries to overcome his most fundamental epistemological problem.
Descartes's general strategy and eternal truths
Descartes begins his theodicy by noting that premise (4) of the causal argument and premise (ii) of the general argument are not strictly speaking true: ‘error as such is not something real which depends on God, but merely a defect’ (AT 7:54/CSM 2:38). Error is simply the absence of knowledge. Descartes sees, however, that this thought alone will not solve his problem:
For error is not a pure negation, but rather a privation or lack of some knowledge which somehow should be in me. And when I concentrate on the nature of God, it seems impossible that he should have placed in me a faculty which is not perfect of its kind, or which lacks some perfection which it ought to have. (my italics; AT 7:55/CSM 2:38)
As Descartes uses the term here, a negation is the absence of some property in a creature, where that property is not required for the perfection of that kind of creature (e.g. lack of sight in an earthworm). Because God is under no obligation to give any creature all the perfections there are (see AT 7:60/CSM 2:42) – indeed, doing so is logically impossible, since creatures by definition lack the perfection of necessary existence – it is consistent with God's goodness to make creatures with negations. But it is not consistent with God's goodness to make a creature with a privation: the absence of a property that is required to make it perfect of its kind (e.g. lack of sight in a dog).Footnote 17 Because error is a privation and not merely a negation, Descartes's problem remains.
It seems that an omniperfect God would not create a world containing the privation of error, and yet error exists. As Gassendi says in the Fifth Objections: ‘Given that [God] could have made things more perfect but did not do so … he must have lacked either the knowledge or the power or the will to do so’ (AT 7:308/CSM 2:214–15). Descartes hints at his general response to this problem in the following passage: ‘There is … no doubt that God could have given me a nature such that I was never mistaken; again, there is no doubt that he always wills what is best. Is it then better that I should make mistakes than that I should not do so?’ (AT 7:55/CSM 2:38). God allows occasional error because it somehow promotes the good: it is in some way ‘better’ that we make mistakes.
Against the first premise of the general argument, Descartes insists that God would allow occasional error if there were a morally sufficient reason (MSR) to do so.Footnote 18 An MSR for some error E would be a good (which outweighs E) that God could not instantiate without allowing E, or an evil (as bad or worse than E) that God could not avoid without allowing E; (when discussing morally sufficient reasons below, I will refer to them simply as ‘greater goods’).Footnote 19 To block the general argument's claim that God and error cannot possibly coexist, Descartes needs to show that it is possible for God to have an MSR for allowing error. He does not need to prove an actual MSR, and so is free to draw on theological assumptions without proof.Footnote 20
Descartes's theodicy strategy rests on the assumption that some possible goods are such that not even an omnipotent being could realize them without allowing error. This assumption seems to clash with Descartes's claim that the ‘eternal truths’ of mathematics, logic, and morality ‘have been laid down by God and depend on him entirely no less than the rest of his creatures’ (AT 1:145/CSMK 23). According to this ‘creation doctrine’ of the eternal truths, God ‘was free to make it not true that all the radii of the circle are equal – just as free as he was not to create the world’ (AT 1:152/CSMK 25).Footnote 21 On what E. M. Curley calls ‘the standard interpretation’Footnote 22 of the creation doctrine, God has the power to make contradictions (e.g. ‘the radii of a circle are unequal’) true.Footnote 23 Curley notes this interpretation's startling consequence: ‘Take any contradiction you like. God could have made it true. Hence it could have been true. Hence, it is possible, even if false. Hence, anything is possible.’Footnote 24 But if anything is possible, then God can realize any good without having to allow error, and so Descartes's search for an MSR is doomed.Footnote 25
Does this disprove my claim that Descartes is searching for an MSR? Not at all. For in clashing with the claim that anything is possible, Descartes's theodicy is no different from many other staples of Cartesian philosophy: the argument for mind–body distinctness,Footnote 26 the ontological argument for God's existence, and the use of a priori reasoning in physics, just to name a few.Footnote 27 If the standard interpretation is correct, then Descartes's theodicy is simply one more element of his system that needs to be interpreted in isolation from the creation doctrine.
Furthermore, the standard interpretation of the creation doctrine is probably wrong. For Descartes expressly denies that ‘God could have made p false’ entails ‘p is not necessarily true’. He insists that mathematical truths are eternal, immutable, and necessary despite being freely created by God: ‘[I]t is because [God] willed that the three angles of a triangle should necessarily equal two right angles that this is true and cannot be otherwise; and so on in other cases’ (my emphasis; AT 7:432/CSM 2:291; see also AT 7:380/CSM 2:261). Commentators have proposed various explanations for why Descartes denies this entailment,Footnote 28 but for our purposes the implications of this denial are more important than its motivations.
Because the eternal truths are genuinely necessary, Descartes seems to think, God cannot now make them false. In the Conversation with Burman, Burman asked whether Descartes's creation doctrine implies ‘that God could have commanded a creature to hate him, and thereby made this a good thing to do’. Descartes replied: ‘God could not now do this: but we simply do not know what he could have done. In any case, why should he not have been able to give this command to one of his creatures?’ (AT 5:160/CSMK 343).Footnote 29 Descartes draws a distinction between what God could have done prior to creating the eternal truths, and what God can now do, once the standards of truth and goodness are in place: ‘God could have made p false’ does not entail ‘God can (now) make p false’.
Descartes sometimes suggests that God can make contradictions true even now:
I do not think that we should ever say of anything that it cannot be brought about by God. For since every basis of truth and goodness depends on his omnipotence, I would not dare to say that God cannot make a mountain without a valley, or bring it about that 1 and 2 are not 3. (AT 5:223–224/CSMK 358–359; see also AT 5:272/CSMK 363)
But such passages should be read in light of the letter to Mersenne where Descartes first expounds the creation doctrine. There, he says:
It will be said that if God had established these truths he could change them as a king changes his laws. To this the answer is: Yes he can, if his will can change. ‘But I understand them to be eternal and unchangeable.’ – I make the same judgement about God. (AT 1:145/CSMK 23)
In a move reminiscent of the traditional theological distinction between God's absolute and ordained power, Descartes suggests that we can speak of God's power in two different senses. In the first sense, we consider divine power in abstraction from God's immutability (similar to absolute power). In the second sense, we take God's immutability into account in describing what God can do (similar to ordained power). In the passages suggesting that God can now do anything, Descartes speaks of God's power in the first sense, while in the Burman passage he speaks of it in the second sense. So the two sets of passages do not really conflict. And the passages suggesting that God cannot (now) violate the law of non-contradiction should be taken more seriously than the others, because Descartes thinks that God is in fact immutable: ‘From the metaphysical point of view … it is quite unintelligible that God should be anything but completely unalterable’ (AT 5:166/CSMK 348).Footnote 30
So for Descartes, although God could have created different standards of goodness or logic, God is now bound by the ones God in fact created. Descartes's theodicy aims to show that given the actual standards of goodness, some goods require divine allowance of evils. On the reasonable assumption that God created these standards (explanatorily, if not temporally) before creating contingent beings, then God's options for the creation of contingent beings were restricted: God was not be able to realize certain goods without allowing error. Understood as an exploration of God's goodness toward contingent creatures, then, Descartes's theodicy strategy has a chance after all.
Or does it? It might be objected that on Descartes's picture, God could have created the standards of goodness in such a way that He would not have to allow evils to get goods (at the later stage of creation). The fact that God did not do this suggests that God is not perfectly good.Footnote 31 I think that Descartes (rightly or wrongly) would consider this objection confused. For according to Descartes, it is not true ‘that God's idea of the good impelled him to choose one thing rather than another’:
For example, God did not will the creation of the world in time because he saw that it would be better this way than if he had created it from eternity … . On the contrary, it is because he willed to create the world in time that it is better this way than if he had created it from eternity. (AT 7:432/CSM 2:291)
The objection supposes that the act by which God creates moral standards could itself be subject to moral evaluation. But how could any of God's alternatives for establishing moral standards (if God even had alternatives)Footnote 32 be right or wrong, better or worse, if no moral standards were in place yet? It would seem that for Descartes, God's creation of the eternal truths is simply a brute fact that is beyond good and evil precisely because it establishes the meanings of ‘good’ and ‘evil’.
The big picture
After hinting at his general theodicy strategy, Descartes says:
… it is no cause for surprise if I do not understand the reasons for some of God's actions; and there is no call to doubt his existence if I happen to find that there are other instances where I do not grasp why or how certain things were made by him. For since I now know that my own nature is very weak and limited, whereas the nature of God is immense, incomprehensible and infinite, I also know without more ado that he is capable of countless things whose causes are beyond my knowledge. … there is considerable rashness in thinking myself capable of investigating the purposes of God. (AT 7:55/CSM 2:38–39).
Descartes here adopts what is often called a ‘sceptical theist’ position: he assumes that even if he cannot understand why God allows error, it is nevertheless possible that God has an MSR for doing so. God may allow error for the sake of a good beyond our ken.Footnote 33
Descartes's next sentence creates a strong impression that he is not thinking systematically. For immediately after claiming that it is rash to investigate God's purposes, he undertakes just such an investigation, suggesting a possible MSR for God to allow error.
Whenever we are inquiring whether the works of God are perfect, we ought to look at the whole universe, not just at one created thing on its own. For what would not without merit appear very imperfect if it existed on its own is quite perfect when its condition as a part of the universe is considered. It is true that, since my decision to doubt everything, it is so far only myself and God whose existence I have been able to know with certainty; but after considering the immense power of God, I cannot deny that many other things have been made by him, or at least could have been made, so that I would acquire the condition of a part in the universe of things. (AT 7:55–56/CSM 2:39; my translation)
For all we know, Descartes insists, our imperfection somehow increases the perfection of the whole universe. I call this Descartes's ‘big-picture’ theodicy strategy.Footnote 34
This strategy seems to appeal to the principle of organic unities, which says: ‘the intrinsic value of a whole is neither identical with nor proportional to the sum of the values of its parts’.Footnote 35 Descartes seems to have believed this principle is actually true, for he told Mersenne: ‘God leads everything to perfection, in one sense, i.e. collectively, but not in another, i.e. in particular. The very fact that particular things perish and that others appear in their place is one of the principal perfections of the universe’ (AT 1:154/CSMK 26). However, in the Fourth Meditation, Descartes does not yet know for sure that he is a part in a universe of other things, and so offers up the overall value of the universe as a possible MSR for God to allow error.
Though we can conceive in the abstract that the universe as a whole has some value, we cannot know concretely what that value is; it is inaccessible to finite humans. Therefore, if God's reason for allowing error were to maximize the universe's global perfection, that reason would be in an important sense beyond our ken. So Descartes's big-picture strategy, far from conflicting with his sceptical-theist response, actually fits with it like a hand in a glove.Footnote 36
The big-picture theodicy may block the general argument from error, but it does nothing to show that God is not the cause of error. Indeed, the chance to increase the perfection of the universe might seem to provide God with a sufficient reason directly to cause us to err. In order to address the causal argument, Descartes goes on to argue that our mistakes result from a misuse of free will.
Free will
According to Descartes, judgement depends ‘on two concurrent causes’ (AT 7:56/CSM 2:39): the intellect – which puts forward various propositions for affirmation or denial – and the will or ‘faculty of choice’ (AT 7:56/CSM 2:39) which ‘simply consists in our ability to do or not so something (that is, to affirm or deny, to pursue or avoid)’ (AT 7:57/CSM 2:40). Both the intellect and will are perfect in their kind.
So what then is the source of my mistakes? It must be simply this: the scope of the will is wider than that of the intellect; but instead of restricting it within the same limits, I extend its use to matters which I do not understand. Since the will is indifferent in such cases, it easily turns aside from what is true and good, and this is the source of my error and sin. (AT 7:58/CSM 2:40–41)
Descartes says that the will's scope ‘extends to anything that can possibly be an object of any other will – even the immeasurable will of God’ (Principles I.35; AT 8a:18/CSM 1:204). The scope of the will is infinite, but the scope of the understanding – by which Descartes means clear and distinct understanding – is finite.Footnote 37 Therefore, free creatures can pass judgement on any proposition they entertain, even if they do not perceive it clearly. We have ‘the freedom to assent or not to assent’ (AT 7:61/CSM 2:42) to what we perceive obscurely, and we misuse this freedom by making judgements when we should not:
If … I simply refrain from making a judgement in cases where I do not perceive the truth with sufficient clarity and distinctness, then it is clear that I am behaving correctly and avoiding error. But if in such cases I either affirm or deny, then I am not using my free will correctly … . In this incorrect use of free will may be found the privation which constitutes the essence of error. The privation, I say, lies in the operation of the will in so far as it proceeds from me, but not in the faculty of will which I received from God, nor even in its operation, in so far as it depends on him. (my emphases; AT 7:59–60/CSM 2:41)
In the Second Replies, Descartes says: ‘They go astray [peccare] who make a judgement when ignorant of the grounds for making it. Whenever we call a conception obscure or confused this is because it contains some element of which we are ignorant’ (AT 7:147/CSM 2:105). By using the term peccare, Descartes suggests that judging without clear and distinct evidence is at least an epistemological analogue of sin, and perhaps even a type of sin.Footnote 38 Descartes's search for the origin of error, like Augustine's search for the origin of evil, terminates in human misuse of freedom.Footnote 39
We cause ourselves to err, and so God is not ‘in the strict and positive sense the cause of the errors to which … we are prone’ (Principles I.29; AT 8a:16/CSM 1:203). Strictly speaking, premise (1) of the causal argument is false: some things in us are not caused by God. Nevertheless, everything in us ‘comes from God’ (AT 7:54/CSM 2:38) in the sense that it depends on God's concurrence. When Descartes causes a free choice of his will, God concurs with him by bringing the choice into existence. Descartes and God working together produce the volition. If we want to know why that volition exists, the answer is first and foremost, God. However, if we want to know why Descartes makes that particular choice (rather than some other choice, or no choice at all), the explanation terminates with Descartes. Descartes, not God, is the cause of error strictly speaking (see The World, AT 11:46–47/CSM 1:97).
Descartes puts his earlier privation/negation distinction to work in explaining divine concurrence with our bad decisions:
… the privation involved, which is all that the essential definition of falsity and wrong consists in … does not in any way require the concurrence of God, since it is not a thing; indeed, when it is referred to God as its cause, it should be called not a privation but simply a negation. (AT 7:60–61/CSM 2:42)
Since God is under no obligation to make us use our wills in the correct way – and in fact may be unable to make us do so if we are truly free – errors are simply a negation relative to God: they are an example of God's producing a kind of thing with less reality or perfection than some other thing God might have produced, without thereby violating any obligation. We are the ones who violate our obligations when we pass judgement on matters that we do not perceive clearly and distinctly.
Still, we might wonder whether God's concurrence with our bad decisions compromises God's goodness. Those who co-operate with criminals (when they can resist) are themselves guilty of crime. Descartes responds:
I must not complain that the forming of those acts of will or judgements in which I go wrong happens with God's concurrence. For in so far as these acts depend on God, they are wholly true and good; and there is in a way more perfection in me because I can elicit them than there would be if I could not … [God] has given me the freedom to assent or not to assent in those cases where he did not endow my intellect with a clear and distinct perception. (AT 7:60/CSM 2:42; my emphasis and translation)
Whether my choice is good or ill, it is still a free act of will, and as such is something good. God has an MSR for concurring even with our bad choices: namely, that in doing so, God gives us freedom, which is ‘a supreme perfection in man’ (AT 8a:18/CSM 1:205) because it makes us ‘bear in some way the image and likeness of God’ (AT 7:57/CSM 2:40).
At this point, Descartes invokes the core reasoning of traditional free-will theodicy. God wants us to be free because freedom makes us more perfect, but God cannot make us free without agreeing to concur with our bad choices. For if God's policy were to concur only with good decisions – such as the decision to suspend judgement about obscure matters – then we would not really have the ability (in which freedom consists) to suspend judgement or not. Though the existence of human freedom does not entail that we will make bad choices (it is logically possible that we might always use freedom correctly), it does entail that God will concur with bad choices if we make them.Footnote 40
Some will object that Descartes cannot really embrace the sort of free-will theodicy I have just attributed to him. For such a theodicy depends on the assumption that free will necessarily involves the ability to choose otherwise, and hence is incompatible with determinism. But there is considerable evidence that for Descartes, we can be free even when we cannot choose otherwise, as when ‘the will of a thinking thing is drawn voluntarily and freely … but nevertheless infallibly, toward a clearly known good’ (AT 7:166/CSM 2:117).Footnote 41
Elsewhere, I have argued that, for Descartes, the will enjoys the power to choose otherwise (in the sense necessary for freedom) even in cases of clear perception.Footnote 42 But even if Cartesian freedom does not consist in the ability to ‘do or not do’ (AT 7:57/CSM 2:40) with respect to clear perceptions, it surely does consist in such an ability with respect to obscure perceptions. Most commentators agree that for Descartes, the will enjoys the power either to pass judgement (by affirming or denying) or to suspend judgement concerning what we perceive obscurely.Footnote 43
And it seems clear that, for Descartes, this freedom during obscure perception is not compatible with divine determinism, either direct or indirect (through the operations of the intellect). For suppose that in cases of obscure perception, God determined us to suspend judgement, so that we lacked the ability to err. If such determinism were compatible with freedom, then it would be possible to fully enjoy the perfection of freedom without being able to err. But then it would be inconsistent for Descartes to suggest – as he does in the passage above – that we would be less perfect if we lacked the ability to err. Furthermore, in responding to the causal argument, Descartes suggests that because error results from human free choice, it follows that God is not responsible for it. But if God can efficiently cause humans to make particular free choices, then God might be responsible for our errors even if we make them freely. So in the passage above, Descartes implies that during obscure perception, we enjoy a kind of freedom that is incompatible with determinism. Therefore, he can construct a free-will theodicy of error, even if he does not think that freedom always involves the ability to do otherwise.
Back to the big picture
Even if we grant that Descartes's free-will theodicy blocks the causal argument, it does not fully address the general argument. For Descartes thinks that our freedom makes us liable to error only because it outstrips our faculty of clear and distinct understanding; if the understanding were larger in scope, we would not err, despite being free. So why didn't God create ‘enlightened’ humans who have a more expansive faculty of clear and distinct perception? As Gassendi put it in the Fifth Objections:
… is it not still an imperfection not to perceive clearly matters which you need to decide upon, and hence to be perpetually liable to the risk of error? … although error does not immediately reside in the faculty God gave you, it does indirectly attach to it, since it was created with the kind of imperfection which makes error possible. (AT 7:313/CSM 2:217–218; my emphasis)
To block the general argument, Descartes must provide a (possible) MSR for God to make creatures like us (with such limited intellects) when God could have made us with ‘more perfect’ intellects.
In some places, Descartes suggests that the human intellect is necessarily finite in scope, the will necessarily infinite: ‘it is in the nature of a created intellect to be finite’ but the will's nature ‘rules out the possibility of anything being taken away from it’ (AT 7:60/CSM 2:42; see also AT 8a:18/CSM 1:205). Since it is logically impossible to make a human whose will does not outstrip her (clear and distinct) intellect, perhaps not even God could make enlightened humans. But in fact, Descartes sees that to overcome our liability to error, God would not need to make the intellect infinite:
… God could easily have brought it about that without losing my freedom, and despite the limitations in my knowledge, I should nonetheless never make a mistake. He could, for example, have endowed my intellect with a clear and distinct perception of everything about which I was ever likely to deliberate. (my emphasis; AT 7:61/CSM 2:42)
God could have avoided our liability to error simply by making the faculty of clear and distinct perception much larger than it actually is – large enough to cover any issue we might ever deliberate about (Descartes assumes a finite number of such issues).
What then is God's MSR for creating creatures liable to error? Descartes answers with a return to the big-picture strategy:
Had God made me this way [i.e. enlightened], then I can easily understand that, considered as a totality, I would have been more perfect than I am now. But I cannot therefore deny that there may in some way be more perfection in the universe as a whole because some of its parts are not immune from error, while others are immune, than there would be if all the parts were exactly alike. And I have no right to complain that the role God wished me to undertake in the world is not the principal one or the most perfect of all. (AT 7:61/CSM 2:42–43)
It may be that a universe containing creatures like us is more perfect on the whole than a world containing enlightened humans instead.
Though Descartes's appeal to the principle of organic unity here strongly resembles Leibniz's theodicy, there is an important difference: Leibniz appeals to organic unity to explain God's MSR for allowing the occurrence of evil, but Descartes appeals to organic unity to provide God with an MSR for allowing the possibility of evil. Descartes claims that our ability to err, not our actual erring, may perfect the universe as a whole (just as Augustine insisted that ‘what is necessary to the perfection of the universe is not our sins … but the existence of souls that … sin if they so will’).Footnote 44
Some interpreters suggest that by returning to the big-picture strategy, Descartes revokes his earlier free-will theodicy.Footnote 45 Since our possession of freedom does not necessarily require God to allow error, freedom cannot be God's MSR for allowing error. While initially plausible, this reasoning ignores the difference between actual unenlightened humans, who suffer from obscure perception in the intellect, and enlightened humans. Though enlightened humans could enjoy freedom without being liable to error, Descartes thinks that the freedom of unenlightened humans does necessarily require God to allow error, and so he accepts the free-will theodicy as an explanation of why God allows the errors of actual human beings. Descartes intends the big-picture theodicy to supplement, rather than supplant, the free-will theodicy.
Freedom (of actual, unenlightened humans) explains why God allows actual errors to occur. The principle of organic unities explains why God allows the possibility of error: it shows why God created actual humans (whose freedom entails the possibility of error) rather than enlightened humans (whose freedom would not entail the possibility of error). Descartes's theodicy appears internally incoherent or unsystematic only if we fail to see that these two different MSR's (freedom and organic unity) are operating at two different levels to resolve two different problems.Footnote 46
Descartes's appeal to organic unities might seem to compromise God's role as guarantor of clear and distinct perception. As Michael Della Rocca says, Descartes's big-picture theodicy ‘would seem to prove too much’:
If God's goodness is, in some mysterious way, compatible with errors with regard to non-clear and distinct ideas, why can it not equally and equally mysteriously be compatible with erroneous c&d judgments? Similarly, if the big picture may – in a way unknown to us – justify erroneous non-clear and distinct ideas, may it not – in a way equally unknown to us – justify erroneous clear and distinct ideas?Footnote 47
In fact, Descartes takes the big picture to justify only the fact that we are able (due to our limited intellects) to err, not the fact that we actually do err. But Descartes still seems to have a problem. For if the big picture can justify the possibility of error with respect to obscure matters, then could it not also justify the possibility of errors with respect to clear and distinct ideas? If God's purposes are inscrutable, then for all Descartes knows, he may fulfil those purposes by erring, even about clear and distinct matters. But this possibility would undermine Descartes's ability to be certain that clear and distinct perceptions are always true.Footnote 48
I believe that Descartes has an answer to this problem. For Descartes, though we do not know what God's purposes are, we know that some purposes cannot belong to God: we know, for example, that it impossible for God to have any malicious purpose. God cannot have ‘the internal malice which is involved in deception’ (AT 7:143/CSM 2:102). Nevertheless, in the Second Replies, Descartes says: ‘I would not want to criticize those who allow that through the mouths of the prophets God can produce verbal untruths which, like the lies of doctors who deceive their patients in order to cure them, are free of any malicious intent to deceive’ (AT 7:143/CSM 2:102). Similarly, victims of dropsy illustrate that in some instances, it ‘is not inconsistent with the goodness or veracity of God’ for us to be ‘deceived by natural instinct’ (Ibid.). In the prophesy and dropsy cases, our errors need not indicate divine malice because we can avoid them, or at least recognize and correct them in the future by using some of our other faculties.
However, Descartes seems to think that it would necessarily be malicious of God to build us in such a way that we formed false beliefs that ‘could not be corrected by any clearer judgements or by means of any other natural faculty’ (AT 7:144/CSM 2:103; see also AT 7:80/CSM 2:55–56). Since ‘there cannot be another faculty both as trustworthy as the natural light and also capable of showing me that [things revealed by the natural light] are not true’ (AT 7:38–39/CSM 2:27), if God designed us to have false clear and distinct perceptions, God would be acting out of malice, which ‘implies a conceptual contradiction – that is, it cannot be conceived’ (AT 8b:60/CSMK 222).
Thus, for Descartes, God is not only what Marilyn McCord Adams calls a ‘producer of global goods’, but also is bound to show a certain amount of ‘goodness to’ individual creatures.Footnote 49 Even if giving us false clear and distinct perceptions would somehow make the universe more perfect overall, Descartes's God would not do it. Though Descartes seems to think that God's actual MSR is beyond our ken, he is certain that it must be consistent with God's goodness to us, a goodness that requires the reliability of clear and distinct perception.
But why assume that it would necessarily be malicious of God to build us with an uncorrectable tendency toward error? If the truth were horrible enough, might not God shield us from it out of benevolence?Footnote 50 Descartes could respond to this worry by suggesting that the imagined situation is actually incoherent. It is morally permissible for doctors to deceive their patients only if they lack the power to cure the patients without deception. As Descartes says, if ‘the will to deceive’ is not ‘evidence of malice’, then it must be evidence of weakness (AT 7:53/CSM 2:37). Similarly, God could resort to deception only if God lacked the power to change reality for the better. Since God is omnipotent, this could never occur. Indeed, it is hard to imagine that an omniperfect God could create a world so bad that we would need to be perpetually shielded from it.
A problem for Descartes's theodicy
Descartes's theodicy is not only important for his epistemological project, but also systematic and internally coherent. It is a mixed theodicy, appealing to two different morally sufficient reasons. First, God's MSR for allowing our actual errors is freedom: God cannot grant creatures like us (with such a limited faculty of clear and distinct perception) freedom without agreeing to concur with our errors if we choose to make them. This appeal to free will refutes the causal argument for error by insisting that God is not the sole explanation for our actions: the will explains why we err. Second, God's MSR for making us capable of error could be the perfection of the universe as a whole. This second MSR, when taken together with the first, seems to show that it is possible for an omniperfect God to allow error. This implication blocks the general argument from error. In appealing to these two MSRs, Descartes also makes use of the ideas that error is merely a negation with respect to God, and that God's reasons are beyond our ken. Descartes unites the four seemingly disparate themes in his theodicy into a consistent whole, showing that occasional error forces him to deny neither God's existence nor God's ability to underwrite the reliability of clear and distinct perception.
But though Descartes's theodicy coheres with itself, it has some serious defects. Rather than going into them all, here I will concentrate on one: the free-will component of Descartes's theodicy conflicts with his belief that God providentially preordains all of history.
Descartes's traditional view of divine providence incorporates two main ideas.Footnote 51 First, God, before creating contingent beings, has a particular plan for how history should unfold. This plan includes which choices creatures should make. Second, God's causal contribution to the world is sufficient to ensure that God's plan is realized.Footnote 52 These two assumptions seem to entail that God is the sufficient cause of all human actions. As Descartes's himself says: ‘the least thought cannot enter the mind of man if God has not wished and willed from all eternity that it enter therein’ (AT 4:313–314/CSMK 272), and ‘nothing can possibly happen other than as it has been determined from all eternity by Providence; so that Providence is like a fate or immutable necessity’ (AT 11:438/CSM 1:380; my emphasis). This understanding of providence appears to rule out the existence of the incompatibilist freedom upon which Descartes's theodicy is (partially) built.
In the Principles of Philosophy, Descartes addresses the problem as follows:
We can easily get ourselves into great difficulties if we attempt to reconcile … divine preordination with the freedom of our will, or attempt to grasp both these things at once. … We shall get out of these difficulties if we remember that our mind is finite, while the power of God is infinite … we cannot get a sufficient grasp of [divine power] to see how it leaves the free actions of men undetermined. Nonetheless, we have such close awareness of the freedom and indifference which is in us, that there is nothing we can grasp more evidently or more perfectly. (AT 8:20/CSM 1:206)
Descartes insists that despite God's providential control over them, free human choices remain somehow undetermined, though we will never understand how this could be.Footnote 53
It is tempting read the passage above as follows: though divine providence entails the absence of incompatibilist freedom, God is nevertheless able to make such freedom be present in the world – in other words, God can do the logically impossible. If this is what Descartes means, then he opens himself up once again to two objections considered earlier in this paper. First, if God can do the impossible, then there is no possible MSR for God to allow error, and Descartes's general theodicy strategy is defunct. Second, if God can do the impossible, we cannot be sure that God is not a deceiver. If incompatibilist freedom can exist despite contradicting divine providential control, then why can't systematic divine deception exist despite contradicting perfect goodness?
But in the passage above, Descartes does not really intend to suggest that God can do the logically impossible. For he thinks that divine providence does not logically entail the absence of libertarian freedom. It is enough for such freedom, he seems to think, that our free actions be causally undetermined. And in his correspondence with Princess Elizabeth (see especially AT 4:352–354/CSMK 282), he develops a roughly Molinist account of divine providence, according to which God's providential control over human choices does not require God to causally influence those choices. Instead, God uses God's knowledge of counterfactuals of freedom (statements such as ‘if Adam were tempted in a certain situation S, he would freely succumb to temptation and eat the fruit’) to arrange for creatures to fulfil God’s plan through their own free choices (e.g. God could arrange for Adam to eat the fruit by placing Adam in S). Descartes could maintain (not without controversy, but with some plausibility) that this understanding of providence does not entail causal determinism.
Unfortunately for Descartes, his creation doctrine of the eternal truths impinges in a problematic way on his Molinist account of providence. Ordinary Molinism trades on the assumption that although God controls which choice situations we face, God does not control which counterfactuals of freedom are true of us (God may put Adam in S, but God does not make it true that Adam would eat the fruit in S). But for Descartes, God must create the counterfactuals of freedom, just as God creates the eternal truths of logic and mathematics (see AT 7:431–432/CSM 2:291). Therefore, Descartes's God not only determines what situations we face, but what we would do in those situations. Our actions are entailed entirely by factors (situations and counterfactuals of freedom) over which we have no control, and so it seems that on Descartes's view, our actions cannot be free in an incompatibilist sense, even if they remain causally undetermined.
Descartes could say that God can simultaneously create the counterfactuals of freedom and give us control over them. This response looks self-contradictory (I have control over whether a certain counterfactual is true only if I determine its truth; so if God makes it true, I lose control over it), but Descartes might insist that it only seems contradictory to us because our concepts are not adequate to God's power. Given the concept of causation that we derive from our experience of creation (see AT 5:347/CSMK 375), if another creature were to cause a counterfactual of freedom be true of me, it would indeed follow that I had no control over the truth of that counterfactual. But according to Descartes, ‘no essence can belong univocally to both God and his creatures’ (AT 7:433/CSM 2:292). Therefore, statements about God's activities do not necessarily have the implications we would normally expect.Footnote 54 In particular, it may be that although God causes a certain counterfactual to be true of me, it is nevertheless also true that I have control over the truth or falsehood of that counterfactual.
I do not think that this way of trying to reconcile providence and freedom is successful. But even if it is, Descartes still has a very big problem. For his reconciliation strategy depends ultimately on the thought that our concepts, derived as they are from creation, are not adequate to God. Therefore, not all apparent limitations on divine power are real limitations. But how are we to tell the difference? And why suppose that we grasp the true implications of divine goodness any more firmly than we grasp the implications of divine creative power? How can we be sure that the contradiction between divine goodness and deceptiveness is not also merely apparent, born of an attempt to apply the mundane concept of goodness univocally to God? If our concepts are inadequate to God, it seems we can never be sure that our clear and distinct ideas really are true.
Descartes could of course appeal to divine revelation as the criterion for distinguishing real from merely apparent constraints on God's power; he might say, for example, that scripture (and/or the magisterium of the church) affirms both providence and freedom, but consistently denies that God could be a deceiver. But this would be to ground reason on faith. Such an appeal to theological authority would undermine Descartes's attempt to vindicate reason through philosophy alone, and to construct an argument that could persuade unbelievers (see his dedicatory letter to the Sorbonne, AT 7:1–2/CSM 2:3–4). So it seems that Descartes's attempt to reconcile the free-will component of his theodicy with his view of divine providence must (given his creation doctrine of the eternal truths) lead either to scepticism or fideism. Either way, it vitiates his philosophical programme.Footnote 55
