Some years ago, Peter Geach argued that philosophical conceptions of omnipotence are doomed to failure:
When people have tried to read into ‘God can do everything’ a signification not of Pious Intention but of Philosophical Truth, they have only landed themselves in intractable problems and hopeless confusions; no graspable sense has ever been given to this sentence that did not lead to self-contradiction or at least to conclusions manifestly untenable from a Christian point of view. (Geach (Reference Geach1973), 7f.)
Geach recommended that theists conceive of God as almighty, not as omnipotent. Roughly, to be almighty is to have power over all things. Allegedly, this notion is immune to philosophical challenges and it encapsulates everything that faith and scripture impute to God qua creator.
Call ‘almightism’ the view that the divine attribute that expresses God's power is almightiness, not omnipotence. Almightism, as far as I know, has won few adherents in philosophical theology apart from Geach himself, while the doctrine of omnipotence has spawned a whole cottage industry. That industry, however, appears to support some of Geach's contentions. A bewildering number of conceptions compete for the title of the official definition of omnipotence, and there is no consensus about which one of them, if any, is both counterexample-free and theologically proper. Indeed, the confusion seems to have intensified since Geach filed his complaint half a century ago. So what he took to be the chief attraction of almightism might still be a point in its favour.
Unfortunately, the doctrine of almightism suffers from being somewhat unclear. Taking God to have power over all things, which is Geach's first-pass attempt at explicating almightiness, does not distinguish God from other, manifestly ungodlike, beings. An atheist physicalist could very well say that the cosmic quantum field has power over all things, but, presumably, the atheist physicalist is not committed thereby to the claim that one of the necessary attributes of the Judaeo-Christian-Muslim God is instantiated. Geach could reply that the sense in which the cosmic quantum field has power over all things is not the sense in which God should be conceived to have power over all things, but my point is precisely that such clarifications are needed and, as far as the present state of almightism is concerned, are missing.
The goal of this article is twofold. First, I'd like to define almightiness. I'll argue that Geach's heuristic explanations fail to capture the concept, and the definition that seems to be on the right track isn't much simpler than typical definitions of omnipotence. So Geach was wrong to suggest that almightism somehow escapes the toils of analysis. On the other hand, almightiness, properly defined, does turn out to be superior to the concept of omnipotence in the sense that familiar puzzles generated by the latter simply dissolve if God is thought to be almighty. Moreover, the conception of maximal power that underlies the notion of almightiness seems more at home in religious contexts than the conception of maximal power that informs typical definitions of omnipotence. So, all in all, almightism is a viable alternative to omnipotence in the context of analytic theology, and not just (as Geach implied) outside of it.
The puzzle of omnipotence
As a preliminary exercise, it is worth looking at the basic philosophical problem about omnipotence, since it forms the backdrop of the whole dialectic. The following discussion is deliberately sketchy and simplified.
Here are four principles that are individually prima facie plausible but are collectively contradictory:
-
Divine Omnipotence:
(1) God is omnipotent.
-
Hoarding:
(2) If x is omnipotent, then for any proposition P, x has the power to bring it about that P.
-
Impossibility → Inability:
(3) □(x does not bring it about that P) ⊃ x lacks the power to bring it about that P.
-
Impeccability:
(4) □(God does not bring it about that God is culpable.)
By (3) and (4), God lacks the power to sin, so by (2), he is not omnipotent, which contradicts (1). I'll call this ‘the puzzle of omnipotence’, or ‘PO’ for short.
Impeccability is not an indispensable part of the puzzle. Premise (4) can be replaced by any proposition about God's necessary failure to do something. The general form of the fourth premise is the following:
(F) □(God does not bring it about that P.)
Instances of (F) generate a contradiction in conjunction with (1)–(3) the same way Impeccability does. Famous candidate instances of (F) are:
(F1) □(God does not bring it about that 2 + 2 = 5.)
(F2) □(God does not bring it about that an event that was part of the past at t 1 is not part of the past at a later time t 2.)
(F3) □(God does not bring it about that there is a stone that God cannot lift.)
(F4) □(God does not bring it about that Eve freely plucks the apple.)
In its most general form, the puzzle of omnipotence concerns necessary limits on God's potential scope of action. There seem to be such limits, and they are prima facie incompatible with the intuitive idea that an omnipotent being can do anything.
I can think of two desiderata for a satisfying solution to PO. The first is that the solution should be unitary in the sense that it should use the same thesis to block the contradiction, regardless of which instance of (F) is in play. This desideratum is motivated by the principle that simple theories are preferable to complex ones, other things being equal. To illustrate, consider a solution to PO on which (F1), (F2), and (F3) are not counterexamples to (1) because impossible states are outside the scope of omnipotence, (F4) is denied on the grounds that there are counterfactuals of freedom that allow an omnipotent being to determine the outcomes of free creaturely decisions,Footnote 1 and, finally, (4) is deflected by claiming that (3) is false and there are necessarily unmanifested powers, God's power to sin being one of them. Arguably, if a solution to PO can achieve the same effect in a less gerrymandered way, it is better off overall.
The second potential desideratum is that the solution should be catholic in the sense that it should avoid controversial or unusual metaphysical principles or posits. This principle is motivated by the thought that a theory that has fewer controversial premises is easier to defend, is less likely to interfere with one's pre-existing metaphysic, and, as a result, can appeal to a wider range of philosophers.
PO is usually solved by constructing an appropriate definition of omnipotence. According to some estimates, there are several dozen such proposals in the literature.Footnote 2 Since Divine Omnipotence is untouchable if one fails to heed Geach's advice, unitary solutions to PO must deny Impossibility → Inability or Hoarding. (Piecemeal denials of candidate instances of (F) seem unlikely to lead to unitary solutions.) Extant definitions of omnipotence fall into two groups depending on which of those premises they target.Footnote 3
I believe, but can't argue in detail here, that no proposed solution to PO is both unitary and catholic, and many are neither. By contrast, if ‘almighty’ replaces ‘omnipotence’ in PO, no contradiction arises under any known variant of (F), and the charge of disunity or lack of catholicity cannot arise. Swapping omnipotence for almightiness solves (or rather dissolves) the puzzle of omnipotence. That's the sales pitch, at any rate.
Geach's criteria
Peter Geach thinks that the philosophical issues about omnipotence can be sidestepped by conceiving of God as almighty. But what is it to be almighty?
‘Almighty’ derives by way of Latin ‘omnipotens’ from the Greek word ‘pantokrātor’; and both this Greek word, like the more classical ‘pankratēs’, and ‘almighty’ itself suggest God's having power over all things. ... God is not just more powerful than any creature; no creature can compete with God in power, even unsuccessfully. For God is also the source of all power; any power a creature has comes from God and is maintained only for such time as God wills. ... Nobody can deceive God or circumvent him or frustrate him; and there is no question of God's trying to do anything and failing. In Heaven and on Earth, God does whatever he will. We shall see that some propositions of the form ‘God cannot do so-and-so’ have to be accepted as true; but what God cannot be said to be able to do he likewise cannot will to do; we cannot drive a logical wedge between his power and his will, which are, as the Scholastics said, really identical, and there is no application to God of the concept of trying but failing. (Geach (Reference Geach1973), 7–8)
I can see four ways to define almightiness, using this passage. I will argue that all four fail and therefore Geach hasn't told us what almightiness consists in.
I assume that almightiness is meant to be a divine attribute: anything that can be conceived to be almighty ought to be either God or a supernatural being of divine prowess. If it is conceivable that one of Geach's criteria is met by a being that is manifestly non-divine in terms of its power, then that criterion cannot define almightiness. And if the conjunction of all of Geach's criteria similarly fails, then Geach's explanation does not supply a definition of almightiness, even though it may identify salient aspects of it.
-
The Way of Comparison
(5) x is almighty =df x is not just more powerful than any creature; no creature can compete with x in power.
This definition fails to entail that an almighty being can create and destroy other beings at will. On the other hand, (5) entails that someone can be almighty simply by being more powerful than any of her worldmates. That sounds wrong.
These problems continue to haunt (5) even if it is modally strengthened so that the almighty being is required to be more powerful than any other possible creature. The power to create and destroy is still missing from the picture even after strengthening, along with a number of other powers that God is supposed to have (e.g. the power to work miracles).
-
The Way of Bestowal
(6) x is almighty =df x is the source of all power; any power a creature has comes from x and is maintained only for such time as x wills.
Consider the Solitary Demon, a lonely being who is the source of all the power he has. He is both unable and unwilling to do anything except sustaining himself in existence. By (6), the Solitary Demon is almighty. So (6) is wrong.
Or consider the Puny Demiurge. He is the source of all the power it has, and he is capable of creating. However, he can bestow only relatively useless, minor powers on his creatures, who are his only worldmates. By (6), the Puny Demiurge is almighty, even though he is so unlike God that the comparison is embarrassing. So (6) is wrong.
One might try to strengthen (6), so that an almighty being is required to be the necessary source of all power. But that would rule out contingent almightiness, which is a theologically significant conceptual possibility.
-
The Way of the Unimpedable Will
(7) x is almighty =df nobody can deceive x or circumvent x or frustrate x; and there is no question of x’s trying to do anything and failing
This definiens is vacuously satisfied by the Solitary Demon if he has a relatively sparse motivational structure and he is concerned with nothing but survival. And the definiens is non-vacuously satisfied by the Puny Demiurge if we assume that he never tries to do anything that goes beyond his pre-existing powers.
-
The Way of Aggregation
-
Geach's criteria fail individually as definitions. But perhaps they are meant to be pooled together:
(8) x is almighty =df x satisfies the definientia of (5), (6), and (7).
This suggestion can be rebutted by constructing further counterexamples. (In fact, both the Solitary Demon and the Puny Demiurge are counterexamples to it.) Although this is formally sufficient to refute (8), it is worth pointing out that even if no counterexamples were readily available, (8) would still be unsatisfactory because it amounts to a mere enumeration of features. It lacks internal coherence: instead of displaying the metaphysical ground of the marks of almightiness, it simply conjoins them. It is undeniable that Geach's criteria capture important aspects of God's power, but (8) does not portray them as aspects of a single unified attribute; rather, it portrays the attribute itself as a mere bundle of the aspects in question. If a more principled conception of almightiness is available, it is certainly preferable to (8). As Socrates remarks in Euthyphro:
Bear in mind then that I did not bid you tell me one or two of the many pious actions but that form itself that makes all pious actions pious ... Tell me than what this form itself is, so that I may look upon it and, using it as a model, say that any action of yours or another's that is of that kind is pious, and if it is not that it is not. (Plato (Reference Cooper1997), 6 [6d–e])
Geach's heuristic explanation fails the Euthyphro test, and the individual components of his explanation fail for independent reasons. So Geach hasn't provided a workable definition of almightiness, even though he might have isolated salient marks or aspects of it.
At this point, my dialectic is exposed to an important objection. One theoretical possibility that opens up, if the foregoing criticism of Geach is correct, is that there is no such thing as the concept of almightiness. What we have, instead, is a nebulous intuition of absolute power conveyed by scripture, faith, and religious practice, satisfying Geach's criteria but failing to form a coherent conception.
I'm not sure how to prove that the definition I'm about to present is the definition of a distinctive concept that is different from the idea of omnipotence and is identical to the notion that Geach was driving it. Two observations can support this contention, it seems to me. First, if my proposed definition entails all of Geach's criteria, then it has the inferential potential that Geach attributes to almightiness. If, in addition, the resulting concept is noticeably weaker than extant conceptions of omnipotence, then it falls close to what Geach calls ‘almightiness’ and it falls short of what most people call ‘omnipotence’. As far as I can tell, this is all it takes to be a candidate analysis of ‘almighty’. I'll come back to this issue at the end.
Almightiness defined
-
Almightiness
-
x is almighty with respect to time t =df
For any possible concrete entity y and any power or categorical property F, if it is metaphysically possible but not necessary, given the history of the world before t, that y has F at t, then x can determine whether y has F at t.
-
x can determine whether P =df
x has the power to bring it about, through intentional action, that P is true and x has the power to bring it about, through intentional action, that P is false.
According to this definition, almightiness is the power to determine what concrete beings are like and what powers they have. It is a ‘second-order power’ that allows one to grant and revoke contingent powers and categorical properties.
By ‘powers’, I mean intrinsic properties that make their bearers capable of bringing about, or liable to undergo, changes. Powers can be individuated in a fine-grained fashion, so that the power to pluck a green apple is different from the power to pluck a red one, or in a coarse-grained fashion, so that both of those powers are identical to the power to move one's hand. I intend to remain neutral on the graininess issue, and I'm not even assuming that powers are fundamental properties (if they are reducible to subjunctive conditionals, so be it). By ‘categorical properties’ I mean fundamental monadic or relational properties that (i) do not entail being capable of bringing about, or being liable to undergo, some change, and (ii) can only be instantiated by concrete beings. Being a prime number is not a categorical property on the present terminology, nor is being such that Big Ben is in London. Typical candidate categorical properties include colour, shape, and spatial position. Metaphysicians sometimes argue whether these (or, indeed, any) properties are categorical, but I'd like to steer clear of those debates. If there are no categorical properties, then reference to them can be removed from the definition.
Assuming that no concrete entity can exist without having some power or categorical property, the definition entails that an almighty being can create and destroy contingent entities at will.Footnote 4 It also entails that he can do all the things that God is traditionally assumed to be able to do, such as causing miracles.
The restriction to contingent powers and categorical properties is justified by the thought that necessary properties cannot be granted or revoked. If it is metaphysically necessary that Alice has the power to play chequers at t, then nobody, not even God, can take that power away from her. Such cases, however, are unlikely to arise in classical theism. If it is necessary that Alice is capable of playing chequers at t, then it is necessary that Alice exists at t. But given that God is the only necessary concrete being in classical theism, Alice is then identical to God. Generally, only divine powers and categorical properties are non-contingent in classical theism.Footnote 5 Restricting the scope of almightiness, a candidate divine attribute, to contingent powers and categorical properties amounts to assuming that God cannot alter his own essential powers, but given that it is impossible to alter the essential powers of a necessary being, this isn't much of a concession.
Almightiness has very little to do with the idea of a being who can do anything, even if ‘anything’ is qualified to avoid logical contradiction. God can be almighty despite being unable to sin, because sinning has nothing to do with the ability to grant or revoke contingent powers and categorical features. God can also be almighty if some powers belong exclusively to individuals other than God, for example, if Eve's power to bring it about that she freely plucks the apple belongs exclusively to her. Almightiness only requires that God be able to grant and revoke such powers at will, and this requirement does not create a theological conundrum. As a result, the present proposal dissolves the Puzzle of Omnipotence. Consider the version of PO where tokens of ‘omnipotent’ are replaced by tokens of ‘almighty’ (time indices are suppressed):
(1*) God is almighty.
(2*) If x is almighty, then for any proposition P, x has the power to bring it about that P.
(3) □(x does not bring it about that P) ⊃ x does not have the power bring it about that P.
(4) □(God does not bring it about that God is culpable.)
By Almightiness, (2*) is false if there are propositions of the form ‘S is F’ where F is not a power or a categorical property. Here's such a proposition:
(9) Eve chooses to pluck the apple.
The property of choosing to pluck the apple is not a categorical property, because it entails being able to bring about some change. And it is not a power either, since exercising a power is not itself a power. One can have a power without exercising it but this distinction is inapplicable to the property of exercising a power. So God qua almighty is not required to bring it about that (9) and therefore (2*) is false.Footnote 6
To show that PO is dissolved as a result, one must also show that the propositions that make (2*) false include the propositions that generate instances of (F). That is, one must show that instances of (F) do not concern powers and categorical properties that are contingently instantiated by concrete beings. And that much seems true. So PO is a non-issue if God is almighty.
A few metaphysical details may be worth clearing up before we move on. Almightiness appears to quantify over merely possible individuals, and some philosophers may find an ontological commitment to such posits implausible or overly costly. I reply that apparent quantification over merely possible individuals is a mere notational convenience here. One can switch to an ontologically more perspicuous idiom if needed. Detailed discussion is in the endnotes,Footnote 7 together with a remark on conjunctive properties and collectives of individuals.Footnote 8
Almightiness, as stated, fails to make it clear when an almighty being is able to determine which powers and categorical properties are present at t. The baseline assumption is that the power to affect time t is possessed by the almighty being at t, but this assumption must be modified if one has an eternalist conception of the divinity or if changing the past is metaphysically possible. Again, the details are relegated to the endnotes.Footnote 9
The definition is immune to the counterexamples that render Geach's criteria inadequate. The Solitary Demon and the Puny Demiurge are classified as non-almighty as long as they are contingent. Take a time t in the history of one of those beings. Assuming that the Demon and the Demiurge are contingent beings, it is possible at t, given history up to t, that they cease to exist and a cosmos with a wholly different character comes into being. But neither the Demon nor the Demiurge is capable of bringing such a change about. So they are not almighty at t.
In response, an interlocutor could strengthen the modal profile of these beings. Suppose that the Solitary Demon is essentially solitary and that the Puny Demiurge is essentially the source of all power in the world yet he is essentially puny. At any time t during their existence, the Demon and the Demiurge have the power to determine, within the modal confines set by prior history, which contingent powers and categorical properties are instantiated at t. So they qualify as almighty, which sounds wrong.
One way to resist these counterexamples is to argue that arbitrary essences cannot be made up at will (the ‘anti-McEar’ objection).Footnote 10 This response is not quite satisfactory, in my view, because the inadequacy of a concept can be demonstrated by appealing to scenarios that are conceivable but metaphysically impossible. In any case, I don't want to rule out that this is the case.
A different way to deflect the challenge is to admit that the Essentially Solitary Demon and the Essentially Puny Demiurge are, in a qualified sense, almighty. They have their little pocket of modal space where everything happens as they wish. They are small gods, as it were. Indeed, if such beings were metaphysically possible under theism, then they could overrule even God's will because God would be unable to create anything in a world where the Demiurge or the Demon exists.Footnote 11 So these modal beasts do not refute the present proposal; rather, they illustrate that one can imagine a modal space that is fragmented into separate pockets each of which has its own almighty ruler.
Almightiness is correct by Geachean lights in the sense that it entails, with the help of a few intuitive supplementary principles, all the Geachean marks of almightiness.
The Way of Comparison
‘God is not just more powerful than any creature; no creature can compete with God in power, even unsuccessfully.’
As we saw, Almightiness entails that God Almighty can freely create and destroy other concrete beings. He can also alter their powers at will, within the limits of contingency. You can't compete with such a person, even unsuccessfully, because he can prevent you from doing anything at any time. So Geach's Way of Comparison is part of the present conception of almightiness.
The Way of the Unimpedable Will
‘Nobody can deceive God or circumvent him or frustrate him; and there is no question of God's trying to do anything and failing.’
That nobody can prevent God from achieving his goals follows readily from Almightiness if God is the only necessary concrete being. For then God can destroy or incapacitate anyone and anything that might stand in his way. That it is impossible for God to try but fail to do something follows by the same route, provided that x’s trying to ϕ entails that it is metaphysically possible for x to ϕ.Footnote 12 If nobody can prevent God from achieving his goals and if God will ϕ whenever he wants to ϕ, then it is impossible for God to try but fail to do something.
The Way of Bestowal
‘God is the source of all power; any power a creature has comes from God and is maintained only for such time as God wills.’
This follows from Almightiness under the presupposition that contingent beings cannot create themselves or generate spontaneously out of nothing. Granted that assumption, Almightiness entails that lesser beings have their powers in virtue of God's having bestowed those powers on them, and have them only as long as God wills.
The Way of Aggregation
Since the previous Geachean marks all follow from Almightiness, it is trivial that their conjunction also follows from it. Almightiness therefore falls close to what Geach calls ‘almightiness’. Although Almightiness does not entail Geach's criteria without auxiliary premises, those premises are either innocuous or sound entirely natural in a theistic context. They include the axiom that God is the only necessary concrete being, the claim that contingent beings need some external metaphysical ground to exist, and the principle that someone's trying to ϕ entails that it is possible from them to ϕ. Using these presuppositions to connect Almightiness to Geach's criteria does not rig the game in favour of a definition that is otherwise on the wrong track.
In the next section, I argue that Almightiness falls short of what most people call ‘omnipotence’, therefore it can justifiably be seen as a candidate definition of the concept that Geach was driving at.
Two concepts of maximal power
In conclusion, I'd like to argue that the conception of maximal power that underlies Almightiness is very different from the conception of maximal power that underlies typical definitions of omnipotence. Moreover, or so I'll claim, the conception of maximal power that underlies Almightiness is more at home in theological contexts than its distant cousin that haunts the omnipotence industry. I intend this thesis to support Geach's claim that almightiness is better suited to describe God's power.
To recapitulate, almightiness is defined as the power to grant or revoke powers and categorical properties, within the modal confines set by prior history. An almighty being has a sort of ultimate ‘second-order’ power, the power to determine the distribution of contingently instantiated powers and categorical properties.
As a result, the claim that God is almighty is different from the claim that God can do anything, which is the guiding intuition behind the doctrine of omnipotence. That doctrine, as it is typically laid out in analytic theology, conceives of maximal power along the following lines:
-
Maximal Power as Maximal Versatility
(10) x is maximally powerful =df
for any state S, x is able to bring about S.
I believe that extant definitions of omnipotence are variations on (10).Footnote 13 Establishing this thesis would require a long and somewhat tedious paper, so I'll simply assume that the thesis is correct. If you are dubious, please take the following to be conditional on a convincing survey.
It is clear, in any case, that Almightiness is based on a different conception of maximal power than the one expressed by (10). Purely formally speaking, the definiens of Almightiness can be portrayed as a variation on the definiens of (10) if one restricts the scope of ‘S’ to states that only concern the possession of contingent powers or categorical properties. But substantively, Almightiness has very little to do with the idea of maximal versatility. Nothing about Almightiness suggests that an almighty being should be able to bring about anything (even if ‘anything’ is qualified to avoid contradiction). This is clear if one recovers the respective notions of comparative power:
-
First-Order Comparative Power
x is more powerful1 than y =df x can bring about a wider range of states than y.
-
Second-Order Comparative Power
x is more powerful2 than y =df x can grant and revoke powers and categorical properties more freely than y.
(10) defines the concept of being maximally powerful1, whereas Almightiness defines the concept of being maximally powerful2.
To round off the article, I'd like to argue that the second-order conception of comparative power, which lies at the heart of Almightiness, is more at home in religious thought and behaviour than the first-order conception of comparative power, which appears to inform the doctrine of omnipotence.
Note, first of all, that the property of being powerful1 (that is, the property of being more powerful1 than the members of an appropriate reference class) is different from the property of being powerful2 (defined in terms of the same reference class). To see why, consider two imaginary humans, Versatile Vilma and Controller Carl. Versatile Vilma is an exceptionally gifted person who has cultivated an amazing array of talents, both mental and physical. She can speak dozens of languages, she is an expert in many fields of science, she is a great poet, musician, dancer, and painter, she excels at sports, and she is an international celebrity who has the ear of world leaders and can bring about all sorts of changes in the world. Controller Carl, in contrast, is the quintessential couch potato. He can't do much apart from sitting in front of his TV, moving from the couch to the kitchen and back, and working the remote. However, Carl owns a very special remote. It can send a signal into people's brains, turning them comatose for an arbitrary period of time. There is no known defence against Carl's device.
Who is more powerful, Vilma or Carl? It all depends on subscripting. Clearly, Vilma is much more powerful1 than Carl, because she can bring about a much wider range of states than Carl can. On the other hand, Carl is much more powerful2 than Vilma. He can freely meddle with people's powers in ways she cannot. Being powerful 1 and being powerful 2 are therefore different properties.
I believe that the theologically relevant intuitions about divine power concern the property of being powerful2 and not the property of being powerful1. When theists think about God, their paradigm concern is the influence that God has over them and the rest of reality. The notion of being powerful2 heavily informs religious ritual and worship: sacrifice, expiation, and petitionary prayer are all motivated by the intuition that God can grant and revoke powers and categorical properties at will, having absolute control over the fate of humans, both living and dead. Faith in general seems to be a response to a hypothetical cosmic imbalance in second-order power.
Given that the second-order notion of comparative power seems to inform religious thought and behaviour in ways that the first-order notion, which expresses a wholly different property, does not, and given that almightiness is defined as the property of being maximally powerful2, the doctrine that God is almighty is better suited to capture the theologically relevant intuitions about divine power. Religious thought and behaviour, to the extent that they are motivated by concerns about divine power, are motivated by concerns about God's ability to influence our lives, not by concerns about God's relative versatility.
Finally, let me address the worry that my proposal is a definition of omnipotence in disguise. An interlocutor could argue that what I'm offering is in fact a new conception of omnipotence that pretends to be an interpretation of Geach. The only reason I have provided for thinking that Almightiness is an interpretation of Geach is that it entails all the Geachean marks of almightiness while failing to entail that God can do anything. But, the interlocutor could point out, all extant definitions of omnipotence fit this description. And so there is no reason not to take Almightiness to be a candidate definition of omnipotence.
I wouldn't mind selling Almightiness as a new definition of omnipotence, but I do not think that it is one. One way to distinguish the concept that Geach was driving at from traditional notions of omnipotence is to note that the former is considerably weaker than the latter. A concept of divine power is considerably weaker than another if it excludes some salient power that the latter tends to attribute to God. Almightiness satisfies this criterion with respect to extant definitions of omnipotence, it seems to me. As far as I know, it is consistent with all extant definitions of omnipotence that God singlehandedly activates the powers of created entities – for example, that he makes a radioactive atom decay at a specific moment in time. But God is unable to do such things if Almightiness exhaustively describes his potential scope of action. Decaying is not a categorical property, because it entails having the power to decay. And decaying is not a power either; it is the manifestation of a power.Footnote 14 So if God is almighty, then God cannot cause radioactive decay, he can only create its possibility. Generally, God cannot activate the powers of created entities directly. Extant conceptions of omnipotence do not have this consequence, so Almightiness defines a notion that is considerably weaker than traditional conceptions of omnipotence.Footnote 15
The interlocutor could suggest that Almightiness is a non-traditional conception of omnipotence, one that happens to be considerably weaker than its rivals. If ‘omnipotence’ is just a name for divine power and it has no further default semantic content, then this is certainly correct. In that case, Geach misunderstood his own project: he argued, unwittingly, that the scope of omnipotence should be taken to be narrower than orthodoxy takes it to be. And on such an understanding of ‘omnipotence’, the present article argues that the best way to narrow the scope omnipotence along Geachean lines is to restrict it to the power of granting or revoking contingent powers and categorical properties.
On the other hand, if ‘omnipotence’ is not just a name for whatever attribute turns out to express God's power, but a term that has some semantic implications on its own, then Almightiness may not be a conception of omnipotence. I don't want to legislate on this issue. I can only register my own semantic intuitions. For me, ‘omnipotence’ connotes being able to do absolutely anything, to the extent that this is metaphysically possible. And, as I argued above, it seems to me that Almightiness is quite distinct from this notion. On the present proposal, God cannot make creatures do things, not even in situations where free will is absent. God Almighty brings creatures into existence and he endows them with powers, but all the activity in the created world comes from creaturely sources (save, perhaps, for God's miraculously granting or revoking powers or categorical features from time to time). Apart from ruling out certain forms of occasionalism and some theories of divine action,Footnote 16 this conception of divine sovereignty seems to clash with extant definitions of omnipotence, and it represents a step back from the idea that God can do anything, even if ‘anything’ is suitably qualified to avoid contradiction. Given these limitations, I find it hard to advertise the present proposal as a new definition of omnipotence, and I find it closer to the notion that Geach was driving at.Footnote 17
