In last week’s post, I introduced a theory of possibility called Leibnizian Theism (LT), the view that the causal powers of a necessarily existing, independent, omnipotent being are the ultimate basis for anything that is possible. In this week’s reflection, we will evaluate this theory by exploring its major benefits and problems.
First, in addition to having all of the benefits of the Aristotelian View, LT has the virtue of grounding alien possibilities. It accounts for why possible worlds which are radically different from our own might have existed. Alien possibilities are taken seriously in disciplines such as physics, ethics, theology, metaphysics, and philosophy of mind, so a theory which explains them is more likely that it would be otherwise.
Second, LT unifies the broadly logical and natural categories of modality by reducing them to a single class – causal modality. It does this by placing or removing restrictions on whose causal powers count in determining what is possible. Natural modality defines possibility in terms of what entities governed by the laws of nature can do, and broadly logical modality defines possibility in terms of what any or all entities can do. Reducing these two categories to a single class is beneficial because it is simpler than explaining them separately.
Third, LT gains support from a modest explanatory principle which underlies the success of abductive reasoning. According to this principle, it is more plausible than not to posit an explanation for some phenomenon if doing so yields gains in terms of explanatory power or prior probability. In combination with other premises, this principle indicates that the existence of all contingent things is best explained by a necessarily existing personal agent whose causal powers raise the probability that contingents things would exist. If this argument is sound, then there are independent reasons for thinking that a being similar to Leibniz’s God exists.
Fourth, LT avoids some of the pitfalls of Extreme Modal Realism (EMR), the theory defended by David Lewis. Recall that the problems of ethical indifference and inductive skepticism plague Lewis’ view because he thinks that all possibilities are equally real. Consequently, there are infinitely many real worlds which (1) contain counterparts of me who perform evil acts that offset the good acts I perform in this world, and which (2) have histories that are identical to this world, but futures that are not friendly to induction. LT avoids these ethical and inductive problems because it does not require that all possibilities be real; nor does it imply the existence of other-worldly counterparts or other-worldly futures which are unfriendly to induction. But despite its benefits, there are several problems that Leibnizian Theism needs to address.
First, LT provides a helpful causal analysis of broad/natural modality, but not of narrow logical modality. In the narrow sense, something is impossible if it entails a contradiction but necessary if its denial is contradictory. So, for example, a round square is impossible because it is inconsistent to assert that a squared object is round. Likewise, a round circle is necessary because denying the roundness of a circle is also inconsistent. But how can a causal analysis of modality be applied to these narrow cases?
One approach is to analyze the concept of impossibility in terms of power by asking, “Why can’t God perform contradictions?” But this is where the Leibnizian runs into trouble: if he answers by saying that contradictions are inconsistent, then he is illicitly helping himself to the modal notion of consistency to define what God can do, whereas God’s power is supposed to define what consistency is. If he answers that contradictions cannot be performed, then the question of why some acts can or cannot be performed arises all over again. For these reasons, the philosopher Eric Hiddleston (2012) concludes that a causal account of narrow modality fails.
Second, LT trivializes the concept of omnipotence. If omnipotence is defined as the ability to do anything possible, and possibility is defined by what God can do, then the concept becomes trivial: God can do anything that God can do! This is hardly illuminating. However, maybe there are other ways of grasping what omnipotence means besides defining it in terms of God abilities. One approach is to define the concept by extrapolating from what finite beings can do. Allow me to explain:
The most basic understanding of causal power comes from the body and its success or failure at “trying.” Through various attempts at pushing, pulling, moving and being moved by other objects, we develop an understanding of what we can, can’t, and must do – of how our embodied powers are limited and liberated in various circumstances. Soon, this understanding expands to include objects that interact apart from our bodies. Whether close by or far away, these objects (like us) exhibit limited abilities to interact in various circumstances. Now, it doesn’t take long before we begin to ask ourselves, “could there be something like unlimited power? Could a being have powers which are not limited by the circumstances that everyday objects face?”
At face value, the idea of unlimited power is far removed from everyday experience, but the coherence of the idea as an extrapolation from our experience should not be ruled out a priori. Of course, we do not know what it is like to be omnipotent, but we do know what it is like to act within limits and to remove obstacles so that our actions are increasingly less limited. If so, then the idea may very well be coherent and (at least partially) understandable through such a process of extrapolation.
How does this relate to the triviality objection to LT? Admittedly, it is trivially true that God can do anything God can do, but our experience of limited power along with our ability to extrapolate from it adds non-trivial content to “what God can do.” This content does not come from the concept of God. It comes from the material we use to construct a concept of God, namely our grasp of what is possible vis-à-vis our own powers and those of other objects.  Of course, this point does not prove that such a being exists or that the idea of such a being will not turn out to be incoherent upon further inspection, but that is another matter.
Third, LT forces us to revise many of our modal intuitions. If an independently existing personal agent exists in all possible worlds, then many of the propositions that seem to us to be possible (epistemically) turn out not to be (ontologically). For example, there seem to be possible worlds in which the following are true: (1) once upon a time there were no rational beings; (2) the only substances are four mutually repelling steel balls; (3) there are no states of consciousness; and (4) only contingent beings exist. 
But if LT is true, our intuitions on this matter are incorrect. In fact, there are many more examples besides (1)-(4) which turn out (surprisingly!) to be false in all possible worlds. This is a problem because a theory that forces us to revise too many of our modal intuitions is probably untrue.
With that said, our intuitions can (and often should) be revised because they depend on an interpretive process that is fallible. This process involves three main stages: we start with an initial understanding of what is possible, impossible, or necessary; then we test that understanding against various thought experiments, conceptual analyses, and other background information that is relevant; and finally, we arrive at a new understanding which may or may not be similar to the first.
Clearly, if modal knowledge depends on a process of interpretation, then our intuitions are fallible and subject to revision. Thus, we should not automatically reject a modal theory like LT just because it mandates such revisions. In fact, LT might provide overriding reasons for thinking that (1) – (4) are necessarily false, such as the four benefits discussed above. So the critic of LT must judge for herself whether the reasons in favour of the theory outweigh the reasons she has for believing that (1)-(4) are possible.
In addition to the three problems I have just outlined, there are many other objections which could have been developed if space had permitted. I could have mentioned David Hume’s objection to attributing “necessity” to any concrete being – not merely to a being like God! Or I might have explored naturalistic objections to the probability of an unembodied Mind, given that we have only ever encountered embodied minds that depend on physical mechanisms in order to function. This list of objections could go on, but my point is that LT faces serious challenges that cannot be ignored if it is to provide an adequate basis for modality.
 Eric Hiddleston (2012) Notre Dame Philosophical Reviews. Retrieved on March 2nd, 2014, from: http://ndpr.nd.edu/news/30418-actuality-possibility-and-worlds/. In my view, Hiddleston’s conclusion prevents the Leibnizian from reducing all modalities to a single class – namely, causal modality – and thereby weakens LT’s simplicity as a theory. However, it remains to be seen whether other modal theories can fair any better at explaining narrow modalities. Perhaps the best option for the Leibnizian is to regard contradictions as meaningless non-acts which get ruled out for merely conceptual reasons, not because of causal restrictions on God’s power.
 Perhaps this is what Alex Pruss (2011) meant when he said, “even if the claim that God is omnipotent is tautologous on the [Leibnizian] theory, the claim that God is omnipotent is still contentful” (p.267).
 Some of these examples of possible propositions come from Richard Swinburne (2012) “What Kind of Necessary Being Could God Be?” European Journal for Philosophy of Religion Vol. 4, no. 2, 1-18.