Van Inwagen (hereafter: Peter) is dealing here with a domain that over the past three or four years I myself have come to think of as the most profound and perhaps the most profoundly difficult in philosophical theology, viz., the causal dimension of the doctrine of Divine Providence. This is a rich and suggestive paper, as well as a shocking one in some respects. Much of what Peter says I find myself either approving of or wavering back and forth on, so naturally I will just ignore all that and get to the points of disagreement. The target of my reply will be the middle section of the paper, the part having to do with the notions of a divine plan and chance. But I shall proceed by indirection, beginning with comments on what Peter says about (i) God's causal role in the ordinary course of nature, (ii) miracles and (iii) divine decrees. In the end, it seems to me, Peter's more shocking claims about chance and Divine Providence spring in part from an excessively deistic conception of God's causal relation to the world. (Those of you who know Peter's commitment to Christian orthodoxy, a commitment recently documented by the magazine Christianity Today, may think of this as my shocking claim!)

1. God's causal role in the ordinary course of nature. On Peter's simplified model of the universe, God creates and sustains (or conserves) the basic entities and their causal powers. This is the extent of His causal contribution to the ordinary course of nature. Like the power generator in the example involving the two pieces of iron, God is at most a remote or mediate (as opposed to immediate or direct) cause of the changes that the basic entities immediately cause in one another. More simply, God supplies the power, and the created entities then move one another. Their movements are not the immediate effects of God's action.

Anyone familiar with the medieval debates over secondary or creaturely causation will realize that the position Peter propounds here is stigmatized as (in effect) a form of deism by almost every important medieval Christian philosopher. To be sure, this brand of deism, which I will label weak deism is much more benign from a theistic perspective than that strong deism which limits God's causal role in nature to creation alone. Nonetheless, medieval religious thinkers agree almost unanimously that a central element of orthodox theism is the doctrine that God is an immediate cause of every effect brought about in the created universe, that every such effect results directly from an action of God's. Some of these thinkers go so far in the opposite direction as to claim, astonishingly, that God is the only genuine efficient cause (as opposed to merely "occasional" cause) of such effects--this is the position called occasionalism, and it numbers among its advocates such luminaries as al-Ghazali, Gabriel Biel, and, later on, Malebranche and Berkeley. Most of the scholastics, however, endorse what I will call concurrentism, according to which natural effects derive immediately from both God and creatures. That is to say, in addition to conserving natural entities and their causal powers, God must act with or co-operate with those entities in order for them to bring about their characteristic effects. These effects thus result from God's action and from the action of the relevant created things.

However, outside of perhaps a few Latin Averroists, the only medieval Christian thinker I know of who holds the weak deism Peter advocates here is the 14th century Dominican Durandus, whose name came to be the one and only proper name associated with weak deism by later thinkers--among whom I have in mind both concurrentists (e.g., the 16th century Jesuits Luis de Molina and Francisco Suarez), who cite him with mild disapproval, and the occasionalists (e.g., Malebranche and Berkeley), who contemptuously dismiss him. So it is only perhaps by chance, so to speak, that Peter escaped being vilified by Malebranche and Berkeley.

The Christian theological tradition, then, is by and large not sympathetic to Peter's account of God's causal role in the ordinary course of nature. But this is a mere argument from authority, even if the authorities in this instance have no little purchase on us. More telling, though, are the reasons why these authorities reject Durandus' weak deism. For prominent among these reasons is the very one Peter gives in rejecting what he calls the interventionist account of miracles, viz., that Christians ought not to believe that in performing miracles, God has to overpower or struggle with or overcome His creatures. The occasionalists and concurrentists are convinced that there are certain miracles recorded in Scripture that weak deism cannot construe other than as events God was able to bring about only by overpowering certain creatures.

Think of Shadrach sitting in the fiery furnace. Here we have real human flesh exposed unprotected to real fire, and yet Shadrach survives unscathed--even though the fire is so hot that it consumes the soldiers who usher him into the furnace. How, on the weak deist view, can God save Shadrach? Only, it seems, by either (i) taking from the fire its power to consume Shadrach, which is inconsistent with the soldiers' being incinerated but in any case amounts (or so the anti-deists all claim) to destroying the fire and in that sense overpowering it; or (ii) endowing Shadrach's clothing and flesh with a special power of resistance, in which case God is opposing His creature, the fire; or (iii) placing some impediment (say, an invisible heat-resistant shield) between Shadrach and the flames, in which case God is yet again resisting the power of the fire. By contrast, on the occasionalist and concurrentist models, God accomplishes this miracle simply by withholding His own action. The (real) fire is, as it were, beholden to God's word; He does not have to struggle with it or overcome it or oppose it. The fire's natural effect cannot occur without God's action, and in this case God chooses not to act in the way required. An elegant account, and one that does not in any way give any creature a power that God must oppose.

In summary, then, I have some worries about the way in which Peter conceives of the relation of God to the world. What's more, I believe it is this weak deism that allows and perhaps even induces Peter to adopt such a restricted notion of a divine decree. More on this below.

2. Miracles. My remark here is a brief one. I am very surprised by the metaphysical characterization Peter gives of miracles, viz., that God accomplishes miracles by supplying created entities with unusual causal powers, powers that they themselves then exercise ("they rearrange themselves" (p. 5)) to bring about abnormal (i.e., miraculous) effects. On this view, then, created things bring about miraculous effects in the same way they bring about their ordinary effects, viz., by exercising their own powers.

I would think, to the contrary, (i) that a miraculous effect is brought about either, as above, by an omission on God's part or by God's acting by Himself directly on some created entity, and, further, (ii) that a miraculous effect is such that the created entities present in the circumstances cannot bring it about, at least not at that very time. So, for instance, God acts directly to accomplish in Mary's womb what would normally be accomplished by the sperm, viz., the fertilization of an ovum. I'm not sure how Peter would think of this miracle. Does God give the ovum special powers? That doesn't seem quite right--after all, the miracle is something that happens to the ovum, not something that it does.

3. Decrees. According to Peter's account, divine decrees are divine actions. Fine and good. Further, God's decreeing a state of affairs S entails that S obtains. Once again, fine and good. But Peter also seems to think that if God decrees S, then the nature of God's causal contribution to S's obtaining rules out its being the case that any creature, and a fortiori any free creature, should also causally contribute to S's obtaining. Now perhaps my disagreement with Peter here is nothing more than a verbal disagreement. But the term 'decree' is one worth fighting for in the metaphysics of divine causation. And I find Peter's use of the term much too restrictive and, further, restrictive in a way that both reflects his weak deism on the one hand and leads directly to some strikingly implausible claims about God's plan on the other.

The weak deism is reflected as follows: Only if some form of deism is true can created effects be divided neatly into those that God directly causes and those that creatures directly cause. (By contrast, according to occasionalism and concurrentism, God is a direct or immediate cause of all natural effects.) But this seems to be exactly the distinction Peter uses to separate off what is decreed by God from what is not decreed by God. That is, God decrees just those states of affairs that He directly brings about. Moreover, since on Peter's view God is an immediate cause of an effect E just in case He is also the sole cause of E, it will be "obvious" (p. 13) that God cannot decree, say, the free actions of His rational creatures. But all this seems excessively restrictive to me, and leads directly to Peter's very limited, indeed shockingly limited, conception of God's plan.

Before we turn to what Peter says about chance and the divine plan, let me sketch an alternative account of divine decrees.

Following Peter, let's distinguish between God's non-reactive decrees and His reactive decrees. Unlike Peter, however, I do not restrict the objects of God's decrees just to states of affairs that God brings about by Himself. Rather, the objects of God's decrees are non-evil states of affairs that God brings about directly either by Himself or in cooperation with (free and non-free) created or secondary causes.

There are also divine decrees of a third sort, though these are called decrees only in an improper and extended sense. For they are not, strictly speaking, actions of God's, though, like decrees proper, they do entail that their objects obtain. I will call them permissive decrees. Their objects are evil states of affairs that God knowingly permits to obtain.

Each type of decree is associated with God's will in a distinctive way. God's non-reactive decrees have as their objects states of affairs that God wills antecedently, where what God wills antecedently constitutes, as it were, His preferred plan for the created realm. However, for His own good reasons God permits (or "wills permissively") deviations from this plan as a concession to creaturely weakness and defectiveness. (God's so-called permissive will is God's will only in an extended sense--just as in the case of permissive decrees.) Finally, when God re-channels, as it were, such deviations from his preferred plan toward further goods, He is said to will such goods consequently. So God's reactive decrees are directed at states of affairs that He wills consequently. Notice, this picture yields at least three different but wholly understandable senses in which a (metaphysically contingent) state of affairs might be a part of God's plan:

    Sense 1: S is part of God's plan1 iff S obtains and is antecedently willed by God.

    Sense 2: S is part of God's plan2 iff S obtains and is either antecedently or consequently willed by God.

    Sense 3: S is part of God's plan3 iff S obtains and is either antecedently or consequently or permissively willed by God.

With this background in hand we can now turn to the central section of Peter's paper.

4. Chance and God's plan. The most striking feature of Peter's account of God's plan is how limited and narrow God's plan turns out to be. According to Peter, S is a part of God's plan only if S obtains and is non-reactively decreed by God. What's more, on Peter's view no free creaturely action is non-reactively decreed by God; nor, as far as I can tell, is any non-free action of any creature non-reactively decreed by God. God sustains the basic entities and their powers, but He is not an immediate and sole cause of their motions. So these motions are not part of God's plan. (I may be misinterpreting Peter here, but I don't think I am.)

Well, given this construal of God's plan, it is no surprise that Alice's death turns out not to be part of God's plan. After all, neither do other events which, with Holy Week just behind us, we might have thought part of God's plan. In fact, if Peter is correct, then no significant event in the history of salvation is part of God's plan. For each such event involves free human choices and/or results from decrees issued by God in response or reaction to the sin of Adam. So, for instance, what Peter says entails that the call of Abraham is not part of God's plan; neither is Abraham's positive response to that call; neither is the commissioning of Moses nor his positive response. The same holds for Mary's freely consenting to be the mother of God, the virginal conception (which, though miraculous, would not have happened had Adam not sinned), Christ's life, death and resurrection, etc. I can speak only for myself, but it seems to me too obvious for words that Peter's construal of God's plan is too limited. At the very least, it is misleading to attribute all these events to chance and leave it at that. (Notice, by the way, that the threefold account of God's plan I suggested above does indeed give us a sense in which the events of salvation history were not part of God's preferred plan, but my account does not forbid us to say that there is yet another wholly proper sense in which (O felix culpa!) those events are most assuredly part of God's plan. It seems to me important for Christians to have both such senses among our conceptual resources.)

5. Caveats and retractions. However, it does not follow that I reject all of the most important theses of Peter's paper. To be sure, if a chance event is defined as one that occurs but is not part of God's plan even in sense 3 laid out above, then there are no chance events--indeed, it is a necessary truth that there are no chance events. That is to say, I hold the doctrine of Divine Providence to entail that every event E occurring in the created universe is such that E is willed either antecedently or consequently or permissively by God. That is, each such event is either specifically and knowingly intended by God (providentia approbationis) or specifically and knowingly permitted by God (providentia concessionis). Nevertheless, even such a strong conception of Providence as this is able, I believe, to accommodate many of the most important claims Peter wants to make.

For instance, this conception of Providence is perfectly consistent with the presence of genuine causal indeterminism in the world, as long as God is in a position to know what will result from the action of genuinely indeterministic causes--be they free or natural causes. (The theory of middle knowledge provides one way, but not the only way, to demonstrate this consistency claim.) So my strong view of Providence is arguably compatible with there being many events that "might very well not have happened."

Again, this strong conception of Providence does not entail that every event has an explanation of the sort Alice's husband is looking for. I say this even though I also believe that for any evil effect E, God permits E only if He then channels or orders E toward some good. This last thesis does not, I want to argue, entail that God has a specific reason for permitting each particular evil, though it does entail that He permits each evil knowingly. So, for instance, it may well be, as Peter suggests, that God has adopted, for whatever reasons, a general policy in keeping with which He allows various moral and natural evils to occur. In that case, there may well be evils, like Alice's untimely death, for which there is no (antecedent) providential reason other than that God has adopted a certain general policy. And this is so, even if there is some specific future good toward which God orders Alice's death. So even if, for every existent evil E, God knowingly permits E and then orders E toward some good, it does not follow that God has some special reason for permitting E over and beyond his reason for adopting the general policy in question. To make my negative point bluntly, from the fact that God orders an existent evil E toward some further specific good G, it doesn't follow that G provides a specific reason for God's permitting E in the first place. Aiming for G may well be God's 'response' to E; it doesn't follow that attaining G is God's specific reason for allowing E. So even on my very strong conception of Divine Providence, there may well be no very specific answer to the question "Why did Alice have to die that way?". In several places St. Thomas uses the following example: God wills that every human being be saved, but permits, say, Jones to sin mortally and be punished by eternal damnation; Jones' sin is thus ordered to a manifestation of divine justice, viz., the punishment of the wicked. "So even though he does not fulfill God's will, God's will is fulfilled in him." However, it does not seem to follow that God allows Jones to sin mortally in order that He might manifest His justice by punishing Jones. The permission to sin comes first, perhaps as the result of a general policy; the ordering of this particular sin, despite itself, to some good comes later.

So I deny that there is chance or fate in an ultimate sense if God is provident. That is, I deny that anything can happen which is not in any sense at all--even sense (3)--part of God's freely and knowingly chosen plan for the created world. But this does not mean that every evil has an explanation of the sort Alice's husband is after, or that there are no events that might very well not have happened, or that there is no such thing as causal indeterminism among either free or natural causes. (In effect, then, I am attempting to put asunder what Peter has joined together, viz., the various characteristics of "chance" listed on p. 9 and again on p. 10.)

6. A note on disjunctive decrees. I really don't see why God shouldn't be able to choose arbitrarily among equally attractive options. There seems to me nothing demeaning about this. In fact, such an ability may very well be built into the notion of free choice (I'm not sure it is, but I'm not sure it isn't, either.)

But grant for the sake of argument that God can issue disjunctive decrees without explicitly decreeing any of the disjuncts by itself. Now suppose that X was in fact the initial state of the universe, that X and Y were the only two choices consistent with God's purposes, and that X and Y were equally preferrable. Then it seems there are only three possibilities:

    (1) In the beginning God decreed X-or-Y and then "looked to see" which of X or Y had been actualized, thus learning that it was X that had been actualized.

    (2) In the beginning God decreed X-or-Y while foreknowing that X would occur were He to decree X-or-Y.

    (3) In the beginning God decreed X.

(1) seems demeaning to a provident God, but it does not appear to be what Peter has in mind, since Peter affirms in at least three places (p. 12, p. 15, p. 16) that God has foreknowledge of future contingents. Now while (2) and (3) are admittedly distinct from one another, my own theory of decrees has as a consequence that (2) describes a way in which God might (antecedently) decree X as an initial state of the universe. That is, as long as X and Y are non-evil states of affairs, (2) just entails (3)--though not vice versa. But even if Peter balks at this, he has to admit that it seems a bit hyperbolic to call (2) an instance of "chance" in the sense of "not being a part of anyone's plan".

Still, the idea of irreducibly disjunctive decrees does strike me as a bit bizarre--especially if God is acknowledged to have infallible foreknowledge of future contingents. As I see it, God foreknows future contingents because He decrees them--and, I believe, this claim can be shown to be compatible with freedom and causal indeterminism, generally. (The doctrine of middle knowledge is one way, but not the only way, to do this.)

Alfred J. Freddoso
Philosophy Department
University of Notre Dame
Notre Dame, IN 46556
April, 1987