LECTURE #6: Contemporary Versions: Taylor, Adler

Richard Taylor

Self-Explanatory Existence

In Chapter 10 of his Metaphysics, Richard Taylor presents a contemporary version of the contingency-based cosmological argument. The main difference between Taylor's version and that of al-Farabi, Ibn Sina and Aquinas is a shift from the material to the formal mode. Where the medievals talked about things and their essences, Taylor talks about our ideas and concepts of things. Where the medievals talked about causation, Taylor talks about explanation. For the medievals' principle of causality, Taylor substitutes the principle of sufficient reason. In all these things, Taylor is following a tradition begun in early modern philosophy (Descartes, Locke and Leibniz).

Taylor's Principle of Sufficient Reason states that for every positive truth, there is a reason why it is so, rather than not. In some versions of the PSR, such as those of Leibniz or Samuel Clarke, the principle states that every truth (whether positive or not) has such a reason. For Taylor, there are two kinds of positive truth: self-explanatory, and dependent. A self-explanatory truth contains its own sufficient reason. To ask, of a self-explanatory truth, "Why is this so?" is to ask a question that is meaningless, or in some other way improper. The PSR entails that every proper Why question has a correct answer.

The shift from causation to explanation has two main advantages. First, it is relatively clear what an explanation is -- it is simply an answer to a Why question. According to one popular account of explanation, the explanans (the answer or explanation) logically entails the explanandum (the thing to be explained). At the very least, it would seem that the explanans must necessitate the explanandum. For, suppose it did not. Then in particular case in which both the explanans, X, and the explanandum, Y, were true, we could ask "Why, given X, did Y come to be true?" Suppose the answer to this question is A. Then it is really X+A, and not X alone, that explains why Y is so. Thus, the complete explanation of Y must make it necessary that Y happened. This seems to be Taylor's position.

The second advantage of the shift from causation to explanation is this: it is very easy to see that an infinite regress of explanations is no explanation at all. If I try to explain what a new word "slagu" means, and my explanation includes additional new words, and these are explained in terms of yet more new words, ad infinitum, it is clear that no real explanation of the meaning of the original word has been given. An explanation must terminate in an answer that is self-explanatory.

Taylor begins with the existence of the world as a whole. In this too he differs from the medievals we've studied. Each of the medieval philosophers began with the existence of some finite, contingent things. Taylor begins with the entire universe. He argues that the existence of the universe calls for an explanation. This explanation must be in terms of some self-explanatory truth.

A self-explanatory truth will be a necessary truth. In this case, the self-explanatory truth must explain the existence of the universe. Taylor assumes that this truth will be a true assertion of the existence of some being, God, since any explanation of the existence of things must include the existence of some real thing. Since the self-explanatory truth is necessarily true, God must exist necessarily.

This position raises two questions immediately: how can the existence of God be self-explanatory, and how do we know that the existence of the universe is not self-explanatory? Apparently, it is the necessary existence of God that makes the question, why does God exist, illegitimate. However, unless our concept of God somehow includes His necessary existence, how could the mere fact that God cannot not-exist make our question of why God exists improper? And, how can a concept of a being include or somehow necessitate that being's existence? Surely, it is one thing to ask, What kind of thing is it? and another to ask, Does it exist? This distinction between questions makes sense even if, as al-Farabi contended, God's essence is identical to His existence. God's existence is self-explanatory only if _for us_ there is no distinction between His essence and His existence -- their identity, as a matter of fact, is not enough.

Why does Taylor think that the existence of the universe is not self-explanatory? There seem to be two reasons. First, we are supposed to be able to introspect on our own idea of the universe, and find that the existence of this thing is not self-explanatory to us. However, this exercise would seem to provide the same answer, when we ask whether God's existence is self-explanatory. Second, we are supposed to be able to see that the universe is the kind of thing that can and should be explained by comparing it to other physical objects of more limited size, all of which seem to require some explanation for their existence. This second reason seems to be based on a fairly weak appeal to analogy. Although the universe is in a sense a very large physical objects, it is not typical. The universe has a number of unique features that weaken Taylor's appeal to analogy.

In the end, the shift from causation to explanation, and from the material to the formal mode, seems to have weakened rather than strengthened the cosmological argument. The medieval version of the argument supported the conclusion that there is a necessarily existing object, and this conclusion did not depend on our being able to conceive of such a being. The modern version, in contrast, seeks to convince us of the existence of a being whose existence is self-explanatory. If I cannot conceive of a being whose nature would make its existence self-explanatory, then I have good reason to reject the conclusion of the modern argument. Mere introspection is irrelevant, in contrast, to the medieval version of the argument.

PSR and Free Will

In the section above, I gave some reason for thinking that the PSR guarantees that explanations necessitate their explananda, or, as I will put it, causes necessitate their effects. However, if this is true, then there is really no contingency in the world. Everything that happens had to happen. According to the PSR, every truth has a sufficient reason. The chain of reasons terminates in a self-explanatory truth, that is absolutely necessary. Every link in the chain is also necessary. Hence, it is impossible that any actual truth be false. This is a strong version of fatalism.

There is at least a prima facie difficulty in reconciling fatalism with the reality of choice. If it is necessary that I act as I do, then I would seem to be in error whenever I treat more than one option as real possibilities for me. Hence, all deliberation and choice seems to be grounded in an illusion. Those who see an incompatibility between determinism and free choice are called "incompatibilists". The opposing positiion is that of compatibilism.

It would seem that the PSR entails determinism, which in turn entails compatibilism, on the assumption that there is real freedom in the world. However, some libertarians (incompatibilists who believe in free will) have used a version of the PSR. A libertarian version of the PSR would claim that all truths break down into three categories:

  1. Self-explanatory truths (the existence of God)
  2. Truths for which there is a necessitating causal explanation.
  3. Truths about personal action, for which the agent in question had adequate motivation.
On this view, there are three ways to explain something. First, it might be self-explanatory. Second, it might describe the effect of some determining cause. In this second case, the existence of the cause necessitates the existence of the effect. Third, it might describe a freely chosen action of some personal agent. In this third case, the explanation will consist in specifying that the agent had the power, opportunity, and adequate motive for doing what he did. In this case, the explanans does not necessitate the explanandum. The agent might have done otherwise, even if the prior state of the agent were exactly the same. In many cases, free agents have the power, opportunity and adequate motive for doing several different things. Whatever the agent does do can be "explained" in terms of prior motives, but we cannot explain why the agent chose to act on these motives rather than others, since there is no such explanation.

On this view, then, God's creation of the world could be a contingent matter, since He could have chosen not to create. He had adequate motive to create, and adequate motive not to do so, as well as the power and opportunity for either course of action. What about the truth, "God chose to create the world"? Does this truth have a sufficient reason? It appears not, since if it did, it would either have to be a necessitating cause, in which case God does not act freely, or we would face an infinite regress of divine choices: choosing to create, choosing to choose to create, etc.

One plausible answer is to deny that there is any prior act of choosing on God's part. Perhaps God's choosing to create the world and His actually creating the world are one and the same thing. Wittgenstein has argued that sometimes we do things intentionally without any prior act of decision -- we decide what to do by simply doing something. God's creation might be like this.

This libertarian version of PSR is the modern version of the idea of "determination" developed in the Kalam tradition. Determination is always the result of personal action, and determination introduces real contingency into the world, since motives often "incline without necessitating". We can "explain" what the agent did without explaining why the agent did this rather than something else that equally well accorded with the agent's ends.

The main problem for the libertarian PSR is this: doesn't the possibility of undetermined free action undercut the basic rationale of the PSR? Certainly, the German philosopher Leibniz thought so. Leibniz was a compatibilist, who insisted that all free actions are necessitated by the agent's prior state. Leibniz saw this view as part and parcel of a commitment to PSR. Once we have carved out an exception to the simple PSR as large as that of free actions, do we have good reason to believe that there are no further exceptions?


Mortimer Adler's version of the cosmological argument takes up the synchronic, contingent-existence version developed by al-Farabi. Unlike Taylor, Adler does not shift to the formal mode. Adler talks about causing existences, not about explaining truths. Given the difficulties we have encountered with the PSR, this seems to be a wise strategy on Adler's part. Adler's main contribution is to deal with an objection to the medieval argument that appeals to the modern idea of inertia.

The synchronic argument assumes that the present existence of contingent things must be explained by the contemporaneous activity of some cause. The fact that a thing has existed in the past is not supposed to count as a cause of its present, continued existence. Similarly, the medievals assumed that present motion must be caused by a synchronic mover -- that the fact that a thing has been moving cannot cause its present motion. Contemporary physics, building on the work of 14th century philosophers like Jean Buridan of Paris, includes a principle of inertia: what has been moving continues to move, unless something acts to interfere. Adler extends this idea to that of existence. It seems to reasonable to suppose that there is such a thing as the inertia of existence: what has been existing tends to continue to exist. This would mean that there need be no contemporaneous power that causes things to exist now.

In response, Adler introduces a distinction between radical and supericial contingency. The existence of x is radically contingent if: x might have not existed, in which case nothing would have existed in its place. The existence of y is superficially contingent if: y might not have existed, in which case something else would have existed in its place. Adler admits that where existence is superficially contingent, a principle of inertia applies. We don't need a cause to explain why a superficially contingent thing continues to exist, since it will continue to exist until something causes it to be replaced by something else. In contrast, if a thing is radically contingent, then its existence at every moment requires a cause, since the alternative, nothingness, requires no cause. Only positive existences require causal explanations.

Adler argues that ordinary things are only superficially contingent, since when they cease to exist, they are always succeeded by some sort of remains or residue. In contrast, Adler argues that the universe as a whole is radically contingent, since the only alternative to its existence is sheer nothingness. Thus, the principle of inertia cannot explain the continued existence of the universe as a whole.

Adler's approach is a plausible one. However, I would argue that he misapplies the distinction between radical and superficial contingency. I would argue that the distinction ought to be one between radical existence and superficial existence. The radical existence of a thing x means the existence of x, as opposed to its sheer non-being. The superficial existence of thing x refers to the arrangement or organization of the matter of which x is composed. The radical existence of ordinary things seems clearly to be contingent, since we can clearly conceive of a world in which they are absent, and are replaced by sheer nothingness. There is no need to talk about the contingency of the whole world, as Adler does. It is enough to assume that contingent radical existence must always be caused. This leads us, in the usual way, to the conclusion that there must be an uncaused first cause of all contingent radical existence. The radical existence of this first cause must be necessary.