A Complex Eternity

One of the central issues in the philosophy of religion is the relationship between God and time. In the contemporary discussion, the issue is framed between the two opposing positions of eternalism and temporalism. The eternalist claims that God is altogether timeless while the temporalist claims that God is somehow “in” time. We may wonder, however, whether these two positions exhaust the available options. In this essay I hope to show that God may be similar to temporal entities in some respects but not in others. The dichotomy between “inside” and “outside” of time is therefore too simplistic a treatment. I will motivate this view in three ways. First, I will distinguish between a strong and a weak sense of “timeless” and show that it is the strong sense that the eternalist typically needs. On the weak sense, an entity is timeless iff it is not temporal. On the strong sense, however, an entity is timeless iff it possesses none of those characteristics that characterize temporal beings as temporal. Second, I will show that the motivations which lead away from eternalism are not identical to the motivations which lead toward temporalism. One major reason for rejecting strong timelessness is the rejection of a whole philosophical and theological outlook driven by a thoroughgoing conception of divine simplicity. Once one has rejected this picture, however, one is not thereby forced to adopt a full-blown temporalism. Third, I offer an argument from the predication of complex actions to God for the view that God’s eternity must be characterized by complexity and something analogous to our temporal before and after relations without thereby being altogether temporal. I offer this conclusion as a special case of the more general thesis that God is importantly similar to temporal entities in some respects but not in others.


The eternalist–temporalist debate is ill named. From this terminology one would assume that the debate is between those who assert that God is eternal and those that assert that he is not. Most temporalists, however, do assert that God is eternal, they merely wish to dispute what this eternity is like. They maintain that God is “in time” in some way, but typically go on to qualify this claim by saying that God is eternal in the sense that he is without beginning or ending, or that he is not “limited” by the constraints of time. Thus the debate is between those who hold two radically different conceptions of eternity: eternity as timelessness and eternity as everlastingness.1 In order to keep this distinction clear, many have chosen to reserve the word “eternity” for the former position and to use “sempiternity” or the like for the later. This, however, seems inadvisable for the temporalists on rhetorical grounds. The word “eternity” is deeply rooted in most translations of the scripture, the customary religious language, and the theological traditions of most religions—without it being completely clear just what the use of the word means philosophically. If the so-called eternalists claim this word as their own, they have nearly won the debate by this move alone. I propose, therefore, that we define “eternity” as the mode of God’s being and acting whatever this turns out to be. The debate should then be cast as a debate about the nature of eternity, not a debate about whether or not God is eternal.

Defining eternity in this way also helps bring to the fore a further issue which should not be lost in the debate: the positions of eternity as timelessness and eternity as everlastingness do not exhaust the logical space of what eternity could be like. There are two ways that we may understand the term “timeless.” In a weaker sense “timeless” may simply mean “not temporal.” In a stronger sense, however, “timeless” may mean “radically unlike temporal.” As a parallel example, we could wonder about the meaning of the imaginary term “triangleless.” Our ordinary term “triangular” involves as essential characteristics (a) being planar, (b) being bounded, and (c) having three straight sides. Now we may ask whether a square is triangleless. In a weak sense, it would be triangleless because it lacks one of the essential features of being triangular, which is just to say that a square is not a triangle. Given the stronger sense, however, a square is still somewhat similar to a triangle, sharing characteristics (a) and (b). In this sense we need something radically unlike a triangle, e.g. a proposition or the color red, to qualify as triangleless. On a stronger sense of timeless, therefore, if any feature is characteristic of temporal beings qua temporal then we may safely infer that nothing like that feature may characterize a timeless being. This kind of inference is employed frequently in authors like Augustine and Boethius (to name just a few), but would be invalid on the weaker understanding. Thus, it is this stronger sense that eternalists typically want, but this leaves out the middle position that there are some features of temporal beings qua temporal that are analogous to features of eternal beings qua eternal. I therefore propose a terminology which covers these three positions:

Eternal entities possess no feature similar to any feature characteristic of temporal entities.
Eternal entities possess at least one feature similar to a feature characteristic of temporal entities, but in some important respects eternal entities are unlike temporal entities.
Eternal entities possess (nearly) all of the features characteristic of temporal entities.

By a feature which is “characteristic” of temporal entities I mean a feature that a temporal being possesses qua temporal. This is to rule out the possibility of claiming that both temporal and eternal entities are similar in virtue of something which has nothing to do with being temporal or eternal (e.g. the property of being the subject of predication). This qualification should also rule out gerrymandered or gruesome properties like being temporal- or-eternal. I also mean the word “feature” to be understood in a broad sense so that it includes properties, relations, propositions concerning, predicates etc. The purpose of this paper is to argue that semi-temporalism is a viable position that deserves more attention in the contemporary discussion. I am not at all sure that semi-temporalism is true, only that it is more plausible than the attention given to it would seem to indicate.

To illustrate what it would be like for something to share only some but not all of the features characteristic of temporal entities we may imagine a science-fiction scenario in which there are alien life-forms very different from us. What these aliens call “time” is also very unlike our version. What they call time has a short cyclical topology such that all events repeat on a five “second” period. Or rather, I should not say “repeat” or “period” since each event is numerically identical to the event five “seconds” in the “future” and in the “past.” Further we may imagine that for these beings nothing ever comes into existence or goes out of existence so that their lives are without beginning or ending, yet of finite length. Should we call these aliens temporal or not? It does not seem to me that this question has a straight-forward answer. In many ways they lack certain features that we take to be constitutive of what it means to be temporal, yet in other ways (e.g. having a successive consciousness) they share important features with temporal beings such that it does not seem fair to call them timeless in the way that a proposition is timeless. Clearly these aliens would satisfy no one’s understanding of eternity, but they do illustrate how it is logically possible, or at least imaginable, that an entity could posses some but not all of those features central to being temporal.

As a sampling of the kind of features I have in mind that characterize temporal entities as temporal I offer the following three:

If an entity is temporally extended, then there are at least two times at which it exists one of which is before the other and conversely the other is after the one.
If X causes Y then X cannot be later than Y.
If an entity is temporal then it must have come to be at some time.

Of course all three of these characterizations are disputable, and I do not here intend to work out a carefully articulated philosophy time. Nevertheless, something in the neighborhood of these three must be true. To be sure, there are many other important characteristics of temporal beings, but all I need for my argument is that there is more than one and that each of the characteristics does not entail all the others.

Reject Simplicity not Eternity

It is well known that those who are committed to the doctrine of divine simplicity are also committed to atemoralism because all the features of time depend in some way on complexity. This can be demonstrated by way of a straight-forward argument:

  1. The divine substance is not composed in any way; nor are there entities intrinsic to God distinct from the divine essence.2
  2. Divine action is intrinsic to God.3
  3. Divine action is simple. It is not composed in any way.
  4. Temporal acts are essentially complex rather than simple.4
  5. Divine action is not temporal.

Because eternity is conceived as the mode of God’s being and acting this line of thought leads to the so-called “static” conception of both divine action and eternity. There is a single incomposite and timeless divine act, only the effects of which spread out through all history. For all the wonder the elegance of this conception commands, many philosophers and theologians find it both conceptually mystifying and religiously pernicious. The Christian faith bears witness to a God who interacts with his creatures in complex ways, most fundamentally in the incarnation, death, and resurrection of Jesus Christ, but also in forgiveness, revelation, answered prayer, judgment, miracles, and the like. As the argument goes, a robust faith in God must view all such interaction not merely as the complex effects of a single divine act but as proper acts of God. Led by this conviction, many reject the traditional doctrine of divine simplicity and along with it any substantive distinction between time and eternity. If God’s action is not to be conceived as single and simple, his mode of being and acting should be conceived as being very much like our own, perhaps with the qualification that it is quantitatively infinite without beginning or ending. This latter rejection, however, need not follow from the former, and temporalists may be discarding more of the classical tradition than is necessary in order to satisfy their central motivations.

Once we reject (c) under the conviction that complexity in God’s action is necessary for him to interact in a meaningful way with his creatures, the denial of (d) does not follow directly. Just because God’s action is complex and temporal action is complex we cannot immediately conclude that God’s action is temporal. To do so we would need a modified version of what Stump calls “Hasker’s Principle”:5

  1. to interact directly with temporal beings requires being temporal oneself.

But we should be hesitant before accepting such a categorical principle. Being temporal involves quite a lot besides complexity. Perhaps it would be wise to test out the thesis that a complex but non-temporal being can interact with temporal beings before concluding that eternity is temporal.

As far as I can tell, there are three prevalent motivations for adopting temporalism. One is the difficulty of fitting the biblical witness of a God who forgives and answers prayers into the picture of a single, partless, divine act. A second motivation is the conceptual difficulty of understanding how two things of such ontologically distinct categories as the temporal and the atemporal can interact with one another at all. We should not conflate these two. My proposal is that the first motivation is the dominant one and can be satisfied without adopting a full-blown temporalism. A third motivation comes from the difficulty of reconciling human free will with a timeless, omniscient God. My intuition is that if the atemporalist can successfully deal with the first and second motivations, they stand a good chance of dealing with this third. Further, although the temporalist has some quick answers for the typical questions in this area, temporalism raises difficult questions of its own. I will, therefore, leave an investigation of the link between eternity and theological determinism as the topic for another paper.

Argument From Forgiveness

So far I have attempted to render plausible the claim that eternity is similar to time in some respects but not in others. I now hope to offer a positive argument with the conclusion that the eternal and the temporal are similar in one particular respect: sequential ordering. The truth of this conclusion implies the falsity of the doctrine of divine simplicity. In framing this argument, however, I hope to show that I can say most of what the temporalist wants to say against the atemporal conception of God’s action without committing to temporalism. To complete this argument I must first define two terms: ‘complex action’ and ‘stage.’ An action is complex just in case it is essentially characterized by a division into stages. By ‘stage’ in this context I just mean a part of an action rather anything having to do with four dimensionalism. An interesting and plausible thesis is that all temporal actions are complex, but all I need for my argument is that there is at least one complex temporal action that finds a relevantly analogous corollary in God. I will take forgiving as a placeholder for such an action, but presumably there are many more.

When one person forgives another there is an important sense of before and after. Suppose that Anna forgives Benjamin for stealing money from her wallet. For this to be a genuine case of forgiving, there must be some first stage (s1) in which Anna holds this theft against him. Then, there must be some second stage (s2) in which Anna releases Benjamin from this hold. The structure of forgiving is more complex than this and likely involves more essential stages than two, but what matters for our purposes is that at least two distinct parts of the action can be identified and that these are essential to its being a case of forgiving. If there were no initial holding-against—perhaps because Anna was ignorant of the crime or simply did not care—then she would not be in a position to forgive, and if there were no second stage, there would be no forgiveness. In the case of Anna and Bob, it is likely (and psychologically necessary) that s1 and s2 occur at distinct instants or over distinct periods, but it seems at least imaginable that some being could accomplish both stages of forgiving in a single instant. What is not conceivable is that s2 takes place at a time before the time at which s1 takes place.

An important tradition in religious thought maintains that (nearly) all predication of terms to God must be understood as merely analogous to the ordinary usage of these terms. On this view, it is false to say “God forgives” if by “forgives” we mean an action of exactly the same kind that Anna preforms toward Benjamin. After all, it is reasonable to suppose that God knows precisely how wrong Benjamin’s action is and also knows that he is always just in his judgments. Surely such differences are going to make whatever God does toward Benjamin quite distinct from what Anna does. To stress this distinction we can term God’s version “forgivingG,” but before we go too far in this direction we must note that there must be some resemblance between vanilla forgiving and forgivingG. If this were not so, the things we say about God would loose their grip on any meaningful content. We would have no idea what we were saying when we claim that God forgives. Surely one respect in which forgivingG and forgiving are similar is their complexity. If forgivingG were not divisible into something like Anna’s s1 and s2 it would be too unlike forgiving to reasonably be called an analogous predicate.6 What remains an open question at this stage, however, is whether forgivingG needs to be temporal for it to be relevantly analogous to forgiving.

For temporal beings we may safely assume that the stages which compose their actions must be themselves temporal. Every stage of Anna’s forgiving must occur at some instant or period of time. There is an inheritance of the property being temporal that runs from the agent through the action to the action’s parts. Thus the structure of Anna’s action is characterized by (i). There is an asymmetric relationship between s1 and s2 such that s1 must occur at some time before the time at which s2 occurs. Indeed this is so essentially. If Anna’s holding-against came after her release of Benjamin’s debt this would not be a case of forgiveness at all. Further, (ii) and (iii) seem to hold for Anna’s action. There must be some causal chain that runs from s1 to s2 and her forgiving must begin at some point. These latter features, however, do not seem to be features of Anna’s action qua a case of forgiving, but rather qua a case of temporal action. They may be essential to forgiving in the sense that they are features of every case of forgiving in every possible world where forgiving occurs, but they do not seem to be essential to forgiving in the sense of being central to our understanding of what it means to forgive.

If I am correct that (i) is central to forgiving, then forgivingG cannot do without something analogous to it without the analogy between forgiving and forgivingG loosing its integrity. ForgivingG need not, however, be characterized by anything like the non-central features of (ii) and (iii). But how should we spell out this “something” which is analogous to (i)? (i) was specified by explicit reference to individual times, and the before and after relations between the stages of Anna’s forgiving are only understood by mapping them onto the before and after relations that hold between instants of time. As with other talk about God, it seems the most advisable way forward is to claim that there is an analogous rather than identical structure characterizing God’s action. Hence, the stages of God’s forgiving may be describe by an asymetrical, transitive relation beforeG (and its corollary afterG) such that one part of God’s forgivingG is beforeG another. This relation is merely analogous to the ordinary before relation because we cannot infer from the proposition “one stage of God’s forgiving is beforeG another stage of his forgiving” such things as that the first stage caused the second stage or that God’s action has a beginning in any sense analogous to the ordinary sense of “caused” or “beginning.”

This may give some sense to such inevitable ways of speaking as “before God created time…,” or “God first…and then…” Most atemporalists will maintain that these are simply mistaken ways of speaking, but it seems to me that when the layman wonders what God was like before creation he knows full well that “before” in this context does not refer to the ordinary relation of before in time, and is not asking something meaningless or obviously absurd. In this way there is a certain structural similarity between time and eternity which is not total. Although there are many ways in which eternity is very different from time, we should not say with the atemporalists that it is completely unlike time. This similarity may also shed some light on how it is that an eternal entity can interact meaningfully with a temporal one. To fully work this out is a subject for another paper, but it seems initially plausible that there is more hope for a semi-temporalist account of time- eternity interaction than an atemporalist one.


In this essay I have hoped to open up for discussion a third option in the debate between temporalism and atemporalism and motivate this middle position in three ways. First, by suggesting a more helpful terminological scheme we may begin to talk precisely about what we mean when we claim that God is timeless. Second, I have cautioned against conflating the motivations which lead away from atemporalism and the motivations which lead toward temporalism. If a middle position between the two is tenable, it may be sufficient to reject atemporalism without bringing on board all the problems associated with temporalism. Third, a positive argument for this middle position arises from a consideration of some of God’s actions which are analogous to our own. For this analogy to hold, God’s actions must be characterized by something like our before and after, yet the relata of God’s version need not be times and therefore need not be temporal. Eternity must therefore be complex and structured in a sequential way but need not have all the other features of time.

  1. There is also a breakdown in the temporalist camp between those who hold that “without beginning” means that God’s own existence antecedent to any act of creation is characterized by a temporal structure stretching back infinitely into the past, and those, like William Lane Craig who hold that God is without beginning in the sense of being the cause of the first moment of time and only being contingently temporal now. Craig’s is an interesting hybrid view that deserves a careful treatment I cannot offer here.

  2. I have pulled this definition directly from W. Matthews Grant, “Divine Simplicity, Contingent Truths, and Extrinsic Models of Divine Knowledge,” Faith and Philosophy 29, no. 3 (July 2012), 254.

  3. One may try to escape the conclusion of this argument by adopting a model of divine action on which God’s actions are extrinsic to him. This move, however, seems to jeopardize the the immediacy of God–creature interaction even more than divine simplicity.

  4. It may be objected to (d) that at least some temporal actions are mereologically simple. It seems to me that all temporal actions are composed of distinct stages and the stages which compose an action are not themselves actions. If we “freeze” an action at a single instant we do not have an action properly speaking. Nevertheless, supposing that there can be such a thing as an instantaneous action, this kind of simplicity will not work for divine action, as though God acted once at a single instant of time and no others. The non-temporality of God’s action still follows, therefore, even if we allow for instantaneous temporal action.

  5. Eleonore Stump, “Eternity, Simplicity, and Presence,” in God, Eternity, and Time, ed. Christian Tapp and Edmund Runggaldier (Ashgate, 2011), 29–45 34.

  6. The advocate of divine simplicity will, of course, reject this claim, and this is just the point on which the debate for and against the doctrine turns. Alternatively, the defender of divine simplicity could concede that some of God’s actions are complex, but hold that they are extrinsic to his essence. I do not intend to venture too far into this debate, merely to suggest that if one rejects divine simplicity on something like these grounds one need not reject a substantive notion of eternity with it.

Source: md