R. M. Adams’s Theodicy of Grace (1998)
The following article was originally published in Philo, 1 (1998) pp. 36-44.
Robert M. Adams, in a brilliant, thought-provoking essay, “Must God Create the Best?”, puts forth a theodicy for God’s creating inferior people to those he could have created or, in general, a less perfect world than he could have created, in terms of his bestowing grace upon these created beings. Since grace is recognized as a virtue within traditional theism, God is morally excused for creating less perfect beings than he could have. Adams writes:
For present purposes, grace may be defined as a disposition to love which is not dependent on the merit of the person loved…. A God who is gracious with respect to creating might well choose to create and love less excellent creatures than he could have chosen…. God’s graciousness in creation does not imply that the creatures he has chosen to create must be less excellent than the best possible. It implies, rather, that even if they are the best possible creatures, that is not the ground for his choosing them. And it implies that there is nothing in God’s nature or character which would require him to act on the principle of choosing the best possible creatures to be the object of his creative powers…. The man who worships God does not normally praise him for his moral rectitude and good judgement in creating us. He thanks God for his existence as for an undeserved personal favour.
Exactly what extant evils are addressed by this theodicy of grace? Is it simply that humans are not as beautiful, athletic, or intelligent as they might have been or that there could have existed alternative species to the extant ones which would have exceeded them in respect to such non-moral desiderata? Understood in this restricted way, it might well succeed, and moreover provide a way out of the quandry of what God is supposed to do if, as seems likely, there is no uniquely best of all possible worlds; in such a case, he will actualize some world in which there is a favorable overall balance of good over evil. But one could take Adams’s theodicy of grace to apply to moral evils as well. This reading is suggested by his view of God’s gracious creation of less perfect humans as a bestowal of “an undeserved personal favour,” since questions of merit or desert concern the manner in which creatures employ their free will for good or evil. So interpreted, his theodicy would be an alternative to a free will theodicy (hereafter FWT), such as he himself has so skillfully developed. It rejects the normative premise that is common to every version of a free will theodicy, as well as the free will defense (hereafter FWD)–that God wants (or even is obligated) to create a world containing the best possible overall balance of moral good over moral evil that he can. It makes available to God the following excuse for creating free beings who produce a less favorable balance of moral good over moral evil than that which would have been realized by other free beings he could have created: “Sure I created some rotten apples or, at any rate, people who are morally inferior to others I could have created, but in doing so I was bestowing my grace upon them–creating them without any consideration of their (moral) merit. So don’t bug me about why I permitted there to be moral evil, or at least more moral evil than was required, given what my options were.” Whether or not Adams intended this wide an application of his theodicy of grace, it will be instructive to see how it fares when so interpreted.
There is an initial problem as to which free persons are the recipients of God’s grace qua originating rather than just sustaining creator. They cannot be persons who exist prior to God’s originating creative act, since, according to theism, no contingent beings existed then. When God graciously creates undeserving free beings he is not like a football coach who looks down his bench to see who is available to be “sent into the game” and chooses persons without consideration of how diligently and effectively they have practiced. (Unlike an ordinary coach, God could get away with such arbitrariness, since his job isn’t on the line.) Maybe the recipient is a possible free person–a maximal, personally coinstantiatable set of properties such that it contains, for at least one action A, the property of freely doing (or freely refraining from doing) A. Questions of merit conceptually cannot arise with respect to such set theoretical entities. The property of being bad, pace one who thinks that a property instantiates itself, is not itself bad, and, even if it were, it could not be blamed for this since it is not an agent at all. It also is a category mistake to speak of abstract entities being benefited, since they conceptually cannot have desires and interests.
Obviously, the recipients of God’s grace are the real-life persons who get created by him. They are not possible free persons but the concrete persons who instantiate these abstracta. The problem is that at the time of God’s creative act these persons have not yet established any track record, since they have not yet had the opportunity to exercise their free will for good or ill. Thus questions of their desert or merit cannot yet arise. Adams might respond that this would not preclude God’s creative act from being completely gratuitous. Herein God would be bestowing not an unmerited but a nonmerited benefit on the persons who get created. A nonmerited benefit can be bestowed on an individual who is neither deserving nor undeserving, such as a chair.
This response confounds God’s preparatory or prevenient grace with his salvific or sanctifying grace. Adams’s notion of God’s gracious creation is similar to St. Thomas’s view of God as acting mercifully in creating things (see footnote 7). Questions of merit pertain only to his salvific grace in which he freely bestows some benefit, some aid toward sanctification, upon persons who have already proven themselves by their misuse of free will not to be meritorious. Because the theodicy of grace under consideration is directed toward reconciling God’s existence with moral evil, it must be the salvific sense of grace that is the relevant one. Since God’s creating people who are undeserving is an instance of salvific grace, it must be based upon what they will freely do after they are created. But can God know about this at the time of his creation? Does God, in other words, possess “middle knowledge,” knowledge of what would result if various possible free persons were to be actualized?
A more precise account of middle knowledge is needed. In order to explain this notion let us begin with the above concept of a possible free person. For purposes of illustration, let us take as our specimen of such a person a maximal and compossible set of personal properties that contains the property of freely doing action A, this being the only property in the set that involves freely doing (or freely refraining from doing) some act. Call this possible person “P“. We can form P‘s diminished possible free person by deleting from P every property that alone, or conjunction with other members of the set, entails freely doing A or entails freely refraining from doing A. Call this proper subset of P “DP.” Notice that DP also is the diminished possible free person of another possible free person, named P‘, who contains all the same properties as P save for containing freely refraining from doing action A in place of freely doing action A. They will form a Dr. Jekyll and Mr. Hyde pair, P being the Mr. Hyde member if the doing of A is immoral, as we will assume it to be. Next, consider this specimen free will subjunctive conditional proposition, called an “F-conditional,” which says what would ensue if a diminished possible free person were to be actualized or instantiated. F. If DP were instantiated, its instantiator would freely do action A. In order for God to have middle knowledge it must be the case that every F-conditional has a contingent truth-value and, in answer to the question, “Just what did God know and when did he know it?”, he knows what they are, even prior to his creative decision.
With this concept of middle knowledge in mind, a dilemma argument can be constructed to show that that God cannot bestow salvific grace in advance.
- Either God has middle knowledge or he does not.
- If God does not have middle knowledge, then God cannot grant salvific grace in advance. first horn
- If God does have middle knowledge, then God cannot bestow salvific grace in advance. second horn
- God cannot bestow salvific grace in advance. From 1-3
On either horn, Adams’s theodicy of grace does not work. Each horn will now be argued for separately.
Argument for the First Horn
In “Middle Knowledge and the Problem of Evil,” Adams argued at length that it is impossible that God create free persons and also have middle knowledge of what they will freely do. Thus, Adams is personally committed to attacking the first horn of the dilemma argument. This, of course, is only of ad hominem interest since it concerns internal consistency within Adams’s philosophy. Unfortunately, it looks very much like God must have middle knowledge in order to be able to bestow salvific grace in advance upon the free people he creates. For God cannot bestow salvific grace in advance unless he knows at the time of his creative act that he is doing so. But he cannot know that he is doing so unless he knows that the free persons he creates will prove themselves morally unmeritorius by their subsequent free actions. This requires that God have middle knowledge prior to his act of actualizing a possible free person, that is, he must know that if he were to actualize this possible free person’s diminished possible free person, the instantiator would freely perform at least one morally wrong action. In other words, he must know the relevant F-conditionals predicting how the instantiator of this diminished possible free person would freely act if it were to be created. Without the requisite middle knowledge, God’s creation of free persons is a gamble, since he cannot have any prior assurance that these persons will come through for him and freely do the morally right thing. This makes availble (sic) to God the morally exonerating excuse for permitting moral evil of unavoidable ignorance. But whatever merit this might have in providing a theodicy for God’s permitting moral evil, it precludes his granting salvific grace in his very act of creating free persons.
Arguments for the Second Horn
It has just been shown that God cannot bestow salvific grace by creating a person unless he has middle knowledge. God’s gracious creation of persons in advance of their freely proving themselves to be unmeritorious might be compared with ordinary cases in which a parent, in advance of foreseen misdeeds, bestows a future unmerited benefit upon a child. To use an example that was given to me by Donald Turner, parents might leave cookies out for their children to eat upon their return from school even though they foreknow that the children will freely perform some bad acts at school, such as telling the teacher that they missed school the previous day to attend their parent’s funeral–an excuse my children often used, even in a single term and with the same teacher, which explains why I got such strange looks from their teachers when I met them at PTA meetings. We can now spell out God’s morally exonerating excuse in the theodicy of grace-in-advance for permitting moral evil. By bestowing the benefit of existence upon persons whom he foreknows to be morally inferior to others he could have created, he is bestowing grace, an unmerited benefit, upon them. More accurately, by electing to actualize diminished possible free persons whose instantiators he knows will create a less favorable balance of moral good over moral evil than other diminished possible free persons he could have instantiated, he is bestowing grace upon these instantiators. And, since grace is a virtue, God is not to be blamed for the moral evils that result from his creative decision.
Two arguments will now be advanced to show that if God has middle knowledge, and he must have it, he cannot bestow salvific grace in advance upon the free persons he creates. The first argument shows that God’s possession of middle knowledge is freedom-canceling for those whom he creates, thereby precluding their being unmeritorious and thus being fit targets of God’s grace. The second argument attempts to show that if God were to have middle knowledge, he would have to know what he is going to do in advance of his deciding to do it, which is a conceptual impossibility.
The First Argument. That God’s possession of middle knowledge is freedom-canceling for the beings he creates is argued for at length in Chapter 5 of my book, On The Nature and Existence of God (Cambridge: Cambridge University Press, 1991), and only a brief summary of it will be presented here. Let us assume that it is true that F. If DP were instantiated, its instantiator would freely do action A. in which doing A is morally wrong. Furthermore, God, in virtue of possessing knowledge, knows that F is true. He then elects to exercise grace by creating a person who will instantiate DP, foreknowing at the time he does this that this person will freely go wrong by doing A. According to the libertarian intuitions of the theist, the events and circumstances reported in the antecedent of F, which really is a complete description of the world up until the time at which the instantiator freely does A, do not sufficiently cause A. Rather, A is agent-determined in the Aristotelian sense by the instantiator. It will be argued that under these circumstances the instantiator of DP, pace F, does not freely do action A.
It would be a mistake to think that under these libertarian conditions God’s initial creative act cannot sufficiently cause the doing of A, for a sufficient cause can reach through the interposition of an indeterministic or stochastic process, such as that which leads up to A. Consider in this connection a stochastic machine which, when its button is pressed, triggers a stochastic process, the decay of a radioactive element or the spinning of a wheel of fortune, the outcome of which determines whether a gun is fired resulting in an innocent person’s death. Imagine that I chance on the scene and inadvertently press the button, resulting in the horrendous outcome. Given that I did not have “middle knowledge” of the relevant subjunctive conditional reporting what would ensue if its button were pressed and did not intend to bring about or even risk this outcome, I am blameless for the resulting evil. Furthermore, I do not even cause this evil. Let us change the circumstances so that I now have “middle knowledge” via some ESP faculty of the truth of this subjunctive conditional and press the button so as to bring about this person’s death. In this case my action is a sufficient cause of the death, and is so in spite of the interposition of a stochastic process. If it seems odd to you that in the former case my pressing the button did not cause the death but did in the second, given that the only difference between the cases concerned my possession of knowledge of the subjunctive conditional, remember that the relevant sense of causation is the forensic one having to do with moral and legal responsibility and blame. Herein, one’s psychological state is relevant to whether one causes something.
There is a difference between the stochastic machine case and God’s creation of free persons in that the intervening stochastic process in the latter involves a free choice by a created being. This difference, however, concerns not whether God sufficiently causes A, since there can be cases of causal overdetermination (e.g. the actions of the hirer and hired assassin are each sufficient causes of the ensuing death), but whether God shares blame along with the creaturely middleman. In my On the Nature and Existence of God it is argued that since God sufficiently causes every action of a created person, the person is not free; and, because persons are not free, this defeats the underlying strategy of the FWD and FWT to deflect blame, though not responsibility, for moral evil from God onto the created middlemen. Since, pace the versions of the FWD and FWT that are based on God having middle knowledge, they are not free since God exercises a freedom-canceling control over them, they cannot take “the fall” for the evils that result from their intentional actions. And, furthermore, since they are not free, this also defeats Adams’s theodicy of graceful creation; for only a free being can qualify as morally unmeritorious and thereby be a suitable object of God’s salvific grace, be it in being initially created or being given subsequent help by God in finding sanctification.
The Second Argument. It would be good to have a backup argument for the second horn, since the first argument is complex and contains some disputable premises. Let us suppose that God graciously creates a morally undeserving person, Jones, who is the instantiator of DP. God’s decision to graciously create Jones requires that God foresee that Jones will freely perform at least one morally wrong action, which in this case involve knowing F. If DP were instantiated, its instantiator would freely do action A. Thus, God’s decision to graciously create Jones has as one of its reasons or decision-determining factors that Jones will freely do the morally wrong action A. But God’s decision to create Jones is a necessary cause of Jones’s freely doing the morally wrong action A, and, furthermore, God knows that it is. And this violates the conceptual truth that C. It is not possible that an agent know what she will decide to do under the description under which she decides to do it before she makes her decision. Thus God must know before he decides to graciously create Jones of an event that he knows to be causally dependent on his decision to create Jones, and this clearly violates conceptual truth C.
Notice that it does not matter whether or not God’s knowing that Jones freely does A is temporally before his decision to graciously create Jones. The “before” can mean either temporally before, if God is omnitemporaly eternal, or before in the order of determination or explanation, if God is timelessly eternal. In the latter case, What God’s knows is a reason that helps to explain or determine his decision, even though there is no temporal relation of precedence between his knowing and his decision.
Alexander Pruss has presented me with a potentially devastating objection to this argument. In order to show that God’s gracious creation of Jones violates C, I had to have God know prior to his creative decision the future indicative proposition that Jones will freely do the morally wrong action A. But, his objection goes, all God needs to know is the subjunctive conditional, “If Jones were to be created, Jones would freely do the morally wrong action A.” And his having this knowledge prior to his creative decision does not run afoul of C. If Pruss’s objection holds up, then the value of my presenting my second argument for the second horn is only as a warning to others who, like myself, might be mislead into giving this argument.
Richard M. Gale University of Pittsburgh
 Reprinted in his The Virtue of Faith and Other Essays in Philosophical Theology (New York: 1987).
 Ibid., pp. 56-7.
 It could conceivably be extended to certain natural evils, such as disease, corruption and death, by seeing these natural evils as resulting from non-moral imperfections of created beings.
 In “Middle Knowledge and the Problem of Evil,” American Philosophical Quarterly 14 (1977).
 Adams considers the objection to his theodicy of grace based on an analogy with human parents who deliberately create a seriously defective offspring by having their genes altered, and thereby act immorally. To counter this analogical argument he finds a theistically-based reason for the immorality of their creative act that does not apply to God’s creative act, namely, that by creating the defective child they are subverting God’s purpose in creating human beings, that being to have creatures who could freely come to know and love him. There is a more straightforward non-theistic response which Adams ignores. The analogy misfires because God, being omnipotent, can assure that his imperfect created beings have a life that is over-all worthwhile whereas the finite parents, given the contingency of the world and their severe limitations in power, cannot do so for their defective child. And when we try to imagine counter-factual circumstances in which this isn’t so, we are really imagining a possible world in which our ordinary moral precepts and rules would have no application. The thought-experiment misfires, because the presupposition for our playing the morality language-game is not realized in the imagined world. For more on this see my “On Some Pernicious Thought Experiments,” in Thought Experiments, eds. T. Horowitz and G. Massey (Lanham: 1991). Donald Turner has rightly suggested to me that a far more powerful version of the analogical objection is to imagine the human parents intentionally conceiving a morally defective child, when it was possible for them to conceive a morally nondefective child instead. This analogy is not so easily subverted.
 For ease of exposition God is taken to enjoy an omnitemporal sort of eternality rather than the more fancy timeless form, but everything that is said about what is prior in the order of time can be understood in terms of priority in the order of explanation or determination.
 St. Thomas’s claim that God acts mercifully infringing a thing out of non-being into being” (Summa Theologiae, 1a, 21,4) suggests this likening of God to a football coach.
 As Tom Berry pointed out to me, there is a way of viewing God’s originating creation of morally free creatures as a bestowal of salvific rather than just prevenient grace, the reason being that their possession of free will permits them to do what is morally wrong, thereby rendering them morally undeserving ab initio. That all free persons are constitutionally flawed in a way in which brutes or automata are not, however, clashes with the normative premise of the FWD and FWT, according to which there is such a great value to the possession of free will that God is exonerated for creating free persons who wrought moral evil, provided there is a favorable balance of moral good over moral evil.
 The New Catholic Encyclopedia (New York: 1967) says: “In the context of grace considered as an absolutely free and personal act through which God communicates Himself to men…the doctrine of grace supposes a creature already constituted in its own being in such wise that it has the possibility of entering into a free and personal relationship with the Divine Persons or of rejecting that relationship.” (p. 666. My italics.) The New English Dictionary defines “grace” as “the free unmerited favour of God manifested in the salvation of sinners” and also as “the divine influence which operates in men to regenerate and sanctify, to inspire virtuous impulses, and to impart strength to endure trial and resist temptation.” Paul saw Christianity as a religion of grace based on “the conception of God bestowing undeserved favour and fellowship upon men” that deals with “men in their mortal weakness and estrangement, reconciling them to Himself.” (James Moffat, Grace in the New Testament (New York: 1932), pp. 8-9.) Grace is a redeeming action. For Saint Augustine, grace “heals the effects of original sin and personal sin and so frees men to live a genuinely Christian life.” (New Catholic Encyclopedia, p. 660.) According to another author, “grace is God’s way of meeting man whom He came in search of and found lost in the solitude of an earthly sinful nature,” again presupposing a prior track record for created persons who are targets of grace. (Peter Fransen, Divine Grace and Man, revised edition (New York: 1965), p. xiv). A good overview of grace is N. P. Williams, The Grace of God (New York: 1930).
 It won’t do for Adams to object that God could bestow grace upon deserving persons on the ground that his bestowal of grace is based on the counterfactual, “Were I to know that the person is undeserving, I would still give my gift.” For, if God could be gracious to a deserving person, then Adams’s theodicy is undercut, because the idea that it is especially valuable to give grace to an undeserving person would be undercut. I owe this point to Alexander Pruss.
 Alexander Pruss has pointed out to me the need for the addition of “under the description under which she decides to do it.” Without this qualification C would face the following type of counterexamples. At the time Oedipus decided to kill the insulting stranger, who he did not then know to be his father, he knew (because of a prophecy) that he would be the necessary cause of his father’s death. Thus, at the time of his decision he did know of an event for which his decision is a necessary cause only he did not conceive of his decision to kill the stranger as the decision to kill his father. Another counter-example to an unqualified version of C is my freely deciding the numbers in a lottery when I have been informed in advance by an angel that I will win the lottery. The Pruss-qualified version of C takes care of this example, since my decision is made not under the description “pick the winning numbers” but under the description, say, “pick 13 12 63 13 23 13.”