Home » Library » Modern Library » Job Opening: Creator of the Universe—A Reply to Keith Parsons

Job Opening: Creator of the Universe—A Reply to Keith Parsons


(2009)

1. The Issue
2. Definitions
3. Parsons’ Argument
4. What I Shall Argue
5. Parsons’ Target: Varghese’s “Gap” Argument
6. Stating the Gap Argument
7. Parsons’ First Critique of the Gap Argument: Deflationary Argument #1
8. A Reply to Parsons’ Deflationary Argument #1
     8.1 Why Infinite Regress Does Not Work in This Case
     8.2 Analysis of This
            8.2.1 Preview
     8.3 So What is “Odd” About the Gap Question?
9. Parsons’ Second Critique of the Gap Argument: Deflationary Argument #2
     9.1 The “Principle of Parsony”
10. Why the Principle of Parsony is Unacceptable
11. Another Try: The Physicalist Interpretation of POP
     11.1 Preview
12. Does the Principle of Sufficient Reason Justify asking Varghese’s Question?
     12.1 “No” Argues Parsons
13. A Reply to This Argument
     13.1 Self-Subsuming Explanation
     13.2 Stage One: The Elemental Modal Distinction: Necessary and Contingent Truth
     13.3 Stage Two: From Truth to Being
            13.3.1 Contingent Being
            13.3.2 Necessary Being?
     13.4 Stage Three: Reflection
     13.5 Stage Four: Solution
14. A Critique of This: The Problem of the Infinite Re-Asker
     14.1 How to Stop the Infinite Re-Asker
15. A Powerful Objection to All of This: The Dilemma Argument
     15.1 The Dilemma for the Theist
            15.1.1 Ockham’s Razor
16. A Solution Proposed: The Logic of Personalistic Explanation
     16.1 Freedom and (Lack of a) Sufficient Cause
     16.2 Explaining Free Choices
     16.3 Problem Solved?
17. Back to the Necessary Being Hypothesis: A Closer Look
     17.1 The Philosophical Advantage of a Necessary Being Hypothesis
18. The Argument Continues: Theism vs. Grand Unified Theory
     18.1 Analysis of This: The First Inference
     18.2 The Second Inference
     18.3 Is Theism Explanatory at All?
19. A Further Worry
     19.1 Five Models of Explanation
20. One Last Time: In Defense of Theistic Explanation
21. Concluding Thoughts

1. The Issue

“The danger that science poses for theism,” argues philosopher Keith Parsons in “No Creator Need Apply: A Reply to Roy Abraham Varghese,” “is that as science progresses, God seemingly becomes increasingly irrelevant. A Creator has to have something to do.”

Indeed. For example, continues Parsons, the latest theories in cosmology, such as Stephen Hawking’s “no boundary” quantum cosmology, by theoretically eliminating an absolute beginning point in time and space, explain the universe so completely they leave nothing for a God ever to have done, nothing a God would ever need to do. But if such an all-encompassing scientific theory were to become well confirmed, then it seems that there would no longer be any logical or theoretical need to suppose the universe has a “Creator.” Hence the title of Parsons’ paper.

The problem this raises (for the theist), argues Parsons, is that an otiose God “soon withers into nonexistence.” Belief in God, once it is seen to be logically unnecessary, simply fades away. What the philosophical defender of theism needs, says Parsons, is “a gap for God that cannot be closed by the progress of science, a domain for divine activity in the world that advancing science cannot seal off, marginalize, or supplant with naturalistic explanations.”

Is modern science in the process of rendering belief in God logically unnecessary? Does the success of science at explaining the world mean that it is no longer reasonable to believe in God? This is the issue. But before we go any further, it will be helpful to clarify the meaning of some terms that will play a role in the argument to come.

2. Definitions

In his paper, Parsons uses the terms “God” and “Creator of the universe” interchangeably, and so shall I. By “Creator of the universe” I take it that he means “something that is responsible for the fact that the material universe exists.” By “material universe” I mean the sum total of all that physical science studies or can in principle study, for example: gluons, gravitons, quarks, protons, neutrons, electrons, photons, fields, forces, energy, atoms and molecules, and objects composed of such things. In short, the sum total of all mass-energy in existence.

Here are a couple additional definitions that will clarify other ideas during our discussion. I shall define religion in general as a set of beliefs and practices understood by participants to be directed at an ultimate reality considered worthy of total devotion or unconditional worship. An ultimate reality, if such a thing existed, would be something that all other things depend on for their existence, but that does not itself depend on anything else for its existence. Theism (from “theos,” the Greek word for God) means belief in a God or gods; monotheism is belief in only one God. A religion counts as theistic in a minimum sense of the word if it holds that ultimate reality is a personal and purely spiritual (nonmaterial) being or beings. (A being is personal if it knows, desires, and makes consciously willed choices.) A religion is theistic in the classical sense if it holds that ultimate reality is a supremely perfect, spiritual, personal being, one considered to have unlimited power (“omnipotence”), knowledge (“omniscience”), and goodness (“omnibenevolence”). An atheist is one who denies the existence of a God or gods.

Thus, under these definitions, Buddhism counts as a religion but does not count as a theistic religion, while Hinduism, Judaism, Islam, Christianity, and Native American spiritual practices count as theistic religions. Classical theism encompasses at least Judaism, Islam, and Christianity. Scientific materialism (the view Parsons defends in this paper) does not count as a religion on my definition, since although its advocates consider prime, unoriginate matter to be the ultimate reality, they do not consider prime matter to be worthy of total devotion or unconditional worship.

3. Parsons’ Argument

Now back to the issue at hand. Is modern science making belief in God logically unnecessary? Is the success of science in explaining the world rendering belief in God unreasonable? Parsons answers yes to both questions, and he supports his answers with a powerful philosophical argument. To begin with, I take Parsons to be making a number of claims about explanatory reasoning, which I shall define as follows. Explanatory reasoning occurs when we begin with one or more facts in need of explanation, and then decide what to infer by thinking about what would best explain those facts. In philosophy, an argument based upon explanatory reasoning is usually called an “inference to the best explanation.” The American logician Charles Sanders Peirce (1839-1914), the first to systematically analyze this general form of reasoning, called the arguments that instantiate it “abductive” arguments. I shall call all such arguments explanatory arguments.[1]

An explanatory argument typically takes the following form:

(a) Begin with one or more facts in need of explanation.

(b) Critically examine as many plausible potential explanations as possible (where a potential explanation is one which, if true, would explain the facts in question, and a plausible explanation is one that is consistent with existing, well-established theories and general background information).

(c) Rank one explanation as the “best” explanation on the basis of the standard measures (explanatory power, explanatory scope, prior probability, simplicity, internal and external consistency, fruitfulness, and so on).

(d) Conclude that the explanation ranked as the best explanation is probably true.

Here I am using “true” in its traditional, everyday sense to mean the agreement or correspondence of a declarative statement with reality. More precisely: A declarative statement is true if it corresponds to reality, and it is false if it does not, where “corresponds to reality” means “accurately specifies or describes the intended portion of reality.” For example, the declarative statement “The Earth is perfectly flat” does not correspond to the relevant portion of realty, the shape of the Earth. It does not accurately specify it, which makes the statement false, while the statement “There are mountains on the Moon” does correspond to the relevant patch of reality, for it accurately describes certain geological structures actually present on the surface of the Moon, which makes this statement true in the classical sense of the word “true.”

It is important to keep in mind that explanatory arguments are always inductive and never deductive. That is, in an inductive claim the conclusion is always probably true if the premises are true; it is never the case that the conclusion must be true, or is certain, if the premises are true. (A deductive argument is one in which the conclusion must be true if the premises are all true; that is, an argument in which if the premises are all true, then the conclusion cannot possibly be false. Inductive arguments make a weaker claim, namely, that if the premises are all true, then the conclusion is probably true.)

Since the important arguments in this paper will be explanatory and thus inductive, it is worth keeping in mind that most of what we know about the world is known ultimately on the basis of explanatory reasoning—including most of what we have discovered through the application of the scientific method.[2]

With that said, Parsons’ opening argument can now be put more clearly. I shall suppose that by “a God with nothing to do” Parsons means “a God with no explanatory role to fill in our best theory of the universe.” If science were to reach a point where everything has been explained by a completed and well-confirmed physics, in such a way that nothing is left unaccounted for, then there would indeed be no explanatory need whatsoever to suppose that God exists. That is, belief in God would not make sense of anything at all, and thus would not be logically needed to make sense of the world. In that case, a Creator would indeed “have nothing to do.”

One reason this is such an important issue today is that many people would find the following inference extremely attractive: If science explains literally everything, so that belief in God is no longer needed in order to explain or make sense of anything, such as why the universe exists or why anything within the universe exists, then belief in God does not make sense of anything, and therefore belief in God just does not make sense. Parsons’ argument may now be put as follows:

  1. As science progresses, it explains more and more of the material universe, leaving less and less unexplained. It does so in three ways: by expanding its explanations out to larger and larger scales, by extending scientific explanation down to the smallest dimensions, and by tracing cause and effect further and further back in time.

  2. Over time, the progress of science leaves us with fewer and fewer unexplained aspects of our world—aspects of reality that we might feel we can only explain or make sense of by supposing that God exists and is their cause. In other words, science leaves fewer aspects of our world that we might believe can only be accounted for on the basis of an explanatory inference to God’s existence.

  3. Thus, the progress of science leaves us with less and less explanatory need to suppose or hypothesize the existence of a Creator of the material universe, and thus with fewer and fewer logical grounds for any theistic explanatory argument (i.e., for any explanatory argument for God’s existence).

  4. In contemporary scientific cosmology, eternal universe theories do away with a singularity or beginning point in time, and contain the explanatory resources to eventually completely account for the existence of the material universe as a whole so as to leave “nothing for a Creator to do or to have done.”

  5. These theories are either well confirmed, or are rapidly on their way to becoming well confirmed.

  6. Thus, it is the case that either (a) science has now explained the universe so completely that there is no longer any explanatory reason to suppose that God exists, or (b) science is on the verge of doing so and can be expected to complete the task in the near future.

  7. Consequently, thanks to the progress of science, belief in the existence of God (or a Creator of the universe) is either epistemically unjustified right now, or is fast becoming so.

Underlying this argument is an unstated assumption:

A: It is not reasonable to suppose that God exists unless there is an adequate explanatory reason to do so.

In other words, the only adequate philosophical justification for theism would be an explanatory argument. Now I happen to believe that a very attractive version of Anselm’s ontological argument for God’s existence can be formulated and defended within a quantified system of S-5 modal logic, thus offering a nonexplanatory theistic argument that is entirely a priori in nature. But I shall not challenge assumption A here, for, with the exception of the ontological argument, all of the classic philosophical arguments for God’s existence are indeed explanatory arguments, and theism would be in deep philosophical trouble if the highly controversial (and difficult to understand) ontological argument were to be the only philosophical reason to believe God exists!

This valuable and intriguing paper by Keith Parsons raises extremely important questions about the ultimate limits of scientific explanation, and about whether or not belief in God is logically vulnerable to the progress of science. Would anyone have any good reason to continue believing in God if cosmologists one day achieve a well-confirmed grand unified theory of physics, one that combines in a logically consistent way general relativity theory and quantum theory, and eliminates an initial singularity or an absolute beginning point of space and time, thereby accounting for all known particles and forces within one comprehensive explanatory framework? In light of the rapid progress physics is making toward this goal, this is certainly one of the greatest questions of our day. Is belief in God destined to go the way of other once widespread beliefs overthrown by the advance of natural science, for instance, belief in the existence of elves, fairies, leprechauns, and witches’ spells? Parsons puts his conclusion this way:

I [shall] argue that there is no intellectual difficulty in postulating an initial state of the universe as an ultimate brute fact…. [By this] I mean a primordial or original state of affairs, one which, though its existence is logically contingent, is not caused by, or dependent on, conditioned by, or reducible to, or supervenient upon any other prior or more basic entity or state of affairs.

4. What I Shall Argue

But is Parsons right about this? Setting aside nonexplanatory arguments (like various ontological arguments for God), if his argument is sound, then atheism is indeed rationally justified, or is fast becoming so while theism is destined for the trash heap of intellectual history (along with all of the other claims that have been falsified by the progress of natural science). However, I shall argue, Parsons’ argument does not hold up when subjected to close critical examination.

In this paper I shall defend two claims. First, Parsons’ crucial fourth premise is false. Second, philosophical theism, far from being vulnerable to the continued progress of science, rests on a rationally satisfying and philosophically attractive logical basis that cannot, in principle, be overturned by the continued progress of natural science. Theism cannot even be overturned by the “Holy Grail” of contemporary physics: a completed, well-confirmed grand unified “theory of everything” that harmoniously unites general relativity and quantum theory so as to account for all known particles and fields (i.e., for all quantities of mass-energy). The philosophical theist must reply to Parsons’ argument, however, because if there is no good reason at all to believe in God, then a philosopher who is also a theist ought to just give it up.

5. Parsons’ Target: Varghese’s “Gap” Argument

Parsons rightly notes that one way to answer his argument would be to establish the existence of an explanatory gap that has the following two features: (a) advancing science cannot in principle be expected to cross the gap; (b) the gap can be closed by a suitably formulated theistic hypothesis. In other words, an answer would be an aspect of the universe that science cannot be expected to explain in principle, but that is explained by theism. This would be “a gap for God that cannot be closed by the progress of science, a domain for divine activity in the world that advancing science cannot seal off, marginalize, or supplant with naturalistic explanations.” Exactly: what is needed is an aspect of the universe that can only be explained by supposing, on the basis of an inference to the best explanation, the existence of a Creator.

Thus, in the rest of his paper Parsons supports his fourth premise by defending it against exactly this type of challenge. More specifically, he defends it against “one of the clearest statements of the [explanatory] case for a Creator, even given the progress and promise of physical cosmology”—namely, the explanatory argument for God’s existence presented by Roy Abraham Varghese in his Introduction to the volume Cosmos, Bios, and Theos.[3] If Varghese’s argument is successful, then the crucial fourth premise of Parsons’ argument is false, and we have a well-grounded reason to suppose that God exists—a reason that even the continued progress of science cannot undermine.

Varghese’s argument is that there is at least one explanatory gap which theism closes (in a rationally compelling way) that atheism cannot possibly close, even given the continued progress of science or a completed and comprehensive grand unified theory of physics. And that is the explanatory gulf between (a) absolute nothingness and (b) the existence of a contingent, material universe. How do you get from (a) to (b)?

When thinking about nothing, the following is worth keeping in mind: Parsons warns that “it is essential that we not hypostasize [the concept of nothing] and turn absolute Nothing into Something…. [for example by conceiving] of absolute Nothing as a sort of ghostly, empty, precosmic matrix—like the primordial chaos of Hesiod’s Theogony—that existed prior to the Big Bang.” He continues:

Language bewitches us here. When we say things like “before the Big Bang there was nothing,” we seem to be naming something we call “Nothing” and asserting that it existed before the universe did. This way of speaking is pernicious. It creates the misleading picture that there was this empty Something—which we call by the name of “Nothing”—which somehow mysteriously gave birth to a universe. But this is wrong. The universe did not come from “Nothing.”

In short, by “nothing” Parsons really and seriously means nothing.[4]

If I may amplify the point: Cosmologists have explained, in scientific terms and in precise step-by-step detail, the physical reactions and transformations that took already existing matter from a uniform soup of elementary particles dominated by radiation (during the Quark-Lepton era) to the present state of the universe containing electrons “orbiting” atomic nuclei composed of protons and neutrons (both of which are in turn composed of quarks). However, nobody has any idea how to get (through scientifically explainable steps) from a state of pure, absolute, brute nothingness to matter existing.[5]

6. Stating the Gap Argument

In the Introduction to Cosmos, Bios, Theos, Varghese goes straight to the heart of the matter: How much explanatory power would be possessed by a completed and well-confirmed grand unified model of the cosmos that in theory eliminates an initial singularity and posits an eternal universe with an endless past and an endless future? Would anything be left unexplained? Would there remain any explanatory reason to suppose that the universe has a Creator? Would there be any reason to suppose that there is something beyond the material universe that is responsible for the fact that a past-eternal material universe exists? Would there be anything for God to have done? Or, would we finally have a scientific explanation that accounts for everything, leaving absolutely nothing whatsoever unexplained (and thus no explanatory reason at all to believe God exists)? Varghese replies:

[P]hilosophically, [an eternal universe] hypothesis begs the bottom line question: how did matter and the universe come into being? (Introduction, page 4, emphasis in the original)

To which Parsons responds:

Actually, the question “How did matter and the universe come into being?” cannot be addressed on the assumption that the universe is eternal. That assumption precludes that the universe came into being.

Right; if we suppose that the universe has always existed and never began to exist, then the question, “How did it all come into existence?” is not to be asked. But Parsons goes on the offensive too quickly here, for in the very next sentence Varghese observes that this “bottom-line” question needs to be reformulated for an eternal universe, and accordingly immediately reframes his bottom-line question as thus:

Such a question can be addressed even with regard to an eternally existing universe…. [W]e want to know how it is that there is a universe with the property of eternal existence. (Introduction, p. 4, my emphasis)

In other words, why does an eternal universe exist? Why an eternally existing anything, rather than just eternal nothingness? Varghese thus raises the ultimate existential question without presupposing a “creation event,” an initial singularity, or any beginning point at all.

So Varghese claims that even if we had a well-confirmed grand unified model of an eternal universe—a model that unifies the field equations of general relativity and the mathematics of quantum theory so as to account for all known particles and interactions without positing an initial singularity or beginning—one live question would nevertheless remain: Why does the whole (past-eternal) thing exist? (One might add: including the pure mathematical structure it instantiates?) Or, to use some interesting wording suggested by Varghese later in his Introduction: Why is an eternal universe in being?

Is this what cosmologist Stephen Hawking meant when he poignantly asked (at the end of A Brief History of Time) whether finding a grand unified theory of the whole material universe would still leave us asking: “What is it that breathes fire into the equations and makes there a universe for them to describe?”[6] Mathematician Roger Penrose, Hawking’s colleague, seems to ask the same question: “Once you have put more and more of your physical world into a mathematical structure, you realize how profound and mysterious this mathematical structure is. How you can get all these things out of it is very mysterious.”[7] Varghese thus seems to be in good company when he continues to press the ultimate existential question even in the face of a completed eternal universe theory.[8]

But so far the argument has rested on an intuitive claim: that a completed physics, even one postulating an eternal universe, would not resolve the ultimate existential question, “Why does a universe—past eternal or not—exist?” Perhaps (educated) intuitions are unavoidable at some point, but can it be demonstrated by reasoned argument that the ultimate question of existence remains unanswered even in the face of a completed physical theory of an eternal universe? It would be nice to have an explicit argument here, one that backs up the intuitions of philosophers, mathematicians, and leading cosmologists, and that goes deeper than the sense of puzzlement we feel when we come face to face with an eternal universe model.

On this point Varghese’s argument disappoints. After examining each of a quantum gravity, vacuum fluctuation, and inflationary model of the universe, Varghese asserts that none offers an answer to the ultimate question of existence: Why is there a universe in a state of physical reality for the theoretical model to describe? The reader is thus left to ponder the question in light of the various cosmological models.

An explicit argument that eternal universe models do not resolve the question of existence would be better; and that is what I shall offer in a few moments. But first I examine and reject Parsons’ argument on this matter.

7. Parsons’ First Critique of the Gap Argument: Deflationary Argument #1

As we have seen, Varghese asserts that even if we hypothesize that the universe is eternal, the question of its existence remains: Why does an eternal universe exist? Parsons’ first response is to argue that we ought to reject the logical legitimacy of the very question at hand. First, says Parsons, Varghese’s question, “Why does it all exist?”, is odd when asked of an eternal universe given that:

the universe is eternal, and given, as Varghese never denies, that each state of the universe is scientifically explicable in terms of the laws of nature and the previous states of the universe…. For instance, the reason that the universe at any given time contains just so much matter or energy and no more or less is explained in terms of the preceding states of the universe and the applicable conservation laws. In such an eternally existing universe, as opposed to one with a temporal origin, there is, by definition, no inexplicable initial state, but a seamless web of explicable states.

I shall call this Parsons’ first “deflationary” argument—it purports to let the air out of the question entirely, leaving it flat and out of the game. Fleshed out, Parsons’ reasoning would seem to be as follows:

  1. If the material universe has an endless (infinite) past with no beginning, then the present state of the universe can be sufficiently explained in terms of the laws of nature and the previous state of the universe, which in turn can in principle be sufficiently explained by reference to the laws of nature and the preceding state of the universe, and so on back in time, stage by stage, explanation by explanation, without end.

  2. If so, then there is no initial state of the universe left unexplained.

  3. Since there is no first state of the universe lacking an explanation, and since in principle no intermediate state of the universe lacks an explanation, it follows that no state of the universe is in principle unexplained: no state of the universe lacks an explanation.

  4. Therefore, Varghese’s question, “Why does the eternal universe as a whole exist?” is an odd one, since nothing in the universe, no individual part of the universe, lacks an explanation.

In short: If each individual state of the universe at each moment in time has an explanation—known or unknown—in terms of previous states of the universe, it would be odd (or logically inappropriate) to ask why the whole sum total exists. It would be like asking, after someone asserts a mathematical equation, “What color is that?”

Of course, the Scottish philosopher David Hume (1711-1776) gave this argument its classic expression when he wrote the following in his Dialogues Concerning Natural Religion (published posthumously in 1778):

In tracing an eternal succession of objects, it seems absurd to inquire for a general cause or first author. How can anything, that exists from eternity, have a cause, since that relation implies a priority in time, and a beginning of existence? In such a chain too, or succession of objects, each part is caused by that which preceded it, and causes that which succeeds it. Where then is the difficulty? But the WHOLE, you say, wants a cause. I answer, that the uniting of these parts into a whole, like the uniting of…several distinct members into one body, is performed merely by an arbitrary act of the mind, and has no influence on the nature of things. Did I show you the particular causes of each individual in a collection of twenty particles of matter, I should think it very unreasonable, should you afterwards ask me, what was the cause of the whole twenty. This is sufficiently explained in explaining the cause of the parts.[9]

8. A Reply to Parsons’ Deflationary Argument #1

I was once convinced by Hume’s argument. However, I became unconvinced after reflecting for a long time on the following counterexample proposed by philosopher William Wainwright.[10] Suppose that you want an explanation for the fact that human beings exist, and imagine that you’re offered the following account. The history of human beings forms an ordered series stretching back in time, with each human being in the series caused to exist by a pair of parents who came into existence at an earlier time in the series, with each parent in turn caused to exist by a pair of parents who came into existence at an earlier time in the series, and so on back in time forever. So the series is just one person after another, brought into existence by two previous persons, with this process going back in time forever, with no beginning at all. In other words, the history of humanity is an infinite regression: the existence of each human being A in the series is explained by reference to two prior human beings B and C in the series, and the actions of B and C brought A into existence, with B and C each explained by reference to two prior human beings, D and E—and so on down the line.)

Now, such a series is certainly logically possible—there is no logical contradiction in the idea. And it is certainly true that, on an infinite regression hypothesis, the existence of each and every individual human being in the series in principle has an explanation, known or unknown. That is, each human being in the series can in principle be explained by reference to two or more previously existing human beings in the series, so that no individual human being in the series lacks an explanation (of his or her existence). It is also the case that on the infinite regression hypothesis, there is no first human being in the series, so no first step in the series lacks an explanation. Thus, to paraphrase Parsons, each human being’s existence is explicable in terms of the laws of nature and the actions of previous persons, such that there is, by hypothesis, no inexplicable first person, and there are no inexplicable intermediate persons—just a “seamless web of explicable persons.”

Nevertheless, there are two good reasons to maintain that the existence of such an infinite series as a whole would lack an explanation even if each individual member of the series in principle has one. First, although each human being (by hypothesis) has an explanation of his or her existence in terms of previous human beings in the series, so that no individual is lacking an explanation, it does not follow, by any recognized rule of deductive or inductive logic, that the existence of the series as a whole is thereby explained. To reason from the claim that each part of the series has an explanation to the conclusion that the whole series is thereby explained commits a composition fallacy. (In general, a composition fallacy occurs when someone reasons that since each part of a whole has a property p, it follows that the whole must have property p as well. For instance, “Since each individual member of the school basketball team is a good player, the team as a whole must be a good team.”) That is the first reason.

Second, and more importantly, it seems crystal clear, upon sustained reflection, that even if no one individual member in the series lacks an explanation, the following very general question about the series as a whole remains totally unanswered: “Why do human beings exist?” In other words, “Why does a past-eternal-series of humans as a whole exist?” Why is it not the case that no human beings ever existed? Or, to put it still another way: “Why does the inventory of all that actually exists include a (past-eternal) series of human beings?” Why not no human beings ever at all? (Note: I intend the question in a causal sense rather than a teleological one: I am asking for the cause of the series as a whole, not its purpose, end result, or telos.)

One could also ask the more specific question: “Why does this past-eternal-series of humans as a whole exist?” Why not a different series of human beings, one with different members? Why does the inventory of all that actually exists include this (past-eternal) series of human beings, rather than some other (past-eternal) series of human beings? Now, don’t these questions remain live even if we suppose that the series is an unending infinite regression, with each member explained in terms of previous members of the series?

It seems clear that the Hume-Parsons argument does not give us a good reason to accept the general principle that an infinite series as a whole is fully explained once we see that each individual member of the series has an individual explanation in terms of previous members of the series. Moreover, the principle is actually undermined by Wainwright’s compelling counterexample of an infinite regression of human beings. I conclude that Parsons does not show in his paper that a well-confirmed eternal universe theory would leave no question of existence unanswered.

8.1 Why Infinite Regress Does Not Work in This Case

Now, although I believe that Wainwright’s thought experiment provides compelling support for Varghese’s claim (that a well-confirmed eternal universe hypothesis would not by itself explain why the material universe as a whole exists), I think we need to press the issue further. We must ask: Why is a past-eternal regression of contingent beings or states of affairs, in which each member of the series is explained in terms of previous members, not a rationally satisfying answer to the question at issue—why a contingent universe exists? (Keep in mind that the logical contingency of the universe is one of Parsons‘ working assumptions throughout his paper.) Can an account be given for the explanatory failure of this case? I believe so. The account I propose is inspired by Lawrence BonJour’s argument for the existence of epistemically basic beliefs. In his latest book, Epistemology: Classic Arguments and Contemporary Responses, BonJour writes:

[I]t is clear on reflection that merely having an infinite chain of beliefs related in the right way is not in fact sufficient for [epistemic] justification. Suppose that instead of believing that there are no armadillos on my desk, I am crazy enough to have the infinite set of beliefs to the effect that for each natural number n, there are at least n armadillos on my desk…. [T]hen I could construct an infinite justificatory chain: that there are at least two armadillos [on my desk] is a conclusive reason for believing that there is at least one, that there are at least three is a conclusive reason for believing that there are at least two, and so on. But it still seems clear that none of these beliefs would really be justified. The reason is that in such a justificatory chain, the justification conferred at each step is only provisional, dependent on whether the beliefs further along in the chain are justified. But then if the regression continues infinitely, all of the alleged justification remains merely provisional: we can never say more than that the beliefs up to a particular stage would be justified if all of the others that come further back in the sequence are justified. And if this is all that we can ever say in such a case, and if all chains of inferential justification were infinite in this way, and if there were no other account of how beliefs are justified that does not rely on inference from other beliefs, then we again would have the unpalatable skeptical result that no belief is ever genuinely justified.[11]

This seems right to me, and so does a qualified application of this general form of reasoning to an infinite explanatory regression of contingent beings, such as the one posited by an eternal universe hypothesis. When the contingent existence of A is offered as the explanatory ground for the contingently existing object B, then the existence of B is accounted for by saying that A caused B; but the contingently existing object A itself has not been given an explanation. In that case, the existence of B is not conclusively accounted for, and further explanation is called for—the existence of A needs an account.

For example, suppose that one morning I find that my car won’t start. When I discover that my newly installed battery is now dead, I only partly understand why my car isn’t starting, although for some practical purposes that might be all the explanatory depth I need. (For instance, that would be all the explanation that I need to know in order to justifiably head back to the local auto parts store, battery in hand, to request a replacement.) But the dead battery, by itself, would not be a complete explanation for my car not starting, for a deeper question would remain: Why is the new battery dead? Wouldn’t that further contingent fact stand in need of explanation as well? And until that fact is accounted for, the fact that my car won’t start lacks a conclusive explanation. (Notice that any good mechanic would not stop with the dead battery; he would search for the underlying reason for the dead battery, in case the cause lies inside the car itself.) Now if each explanation in an explanatory series is similarly inconclusive, then at each stage in the series we lack a rationally satisfying, completed explanation. Therefore, if a series of contingent things goes back like this forever—as it must if it only contains contingent things—then there simply is never going to be a completed, rationally satisfying explanation, no matter how far back we go.

To paraphrase BonJour’s argument: In such an explanatory chain, the explanation conferred at each step is inconclusive and provisional only, passing the explanatory buck back to previous steps in the chain on the presumption that a conclusive explanation exists somewhere else down the line. But then if the regression continues infinitely, all of the alleged explanations remain merely inconclusive and provisional: we can never say more than that the steps up to a particular stage would be completely explained if all of the others further back in the sequence are completely explained. If this is all that we can ever say in such a case—that all chains of explanation are infinite in this way, and that the only available account of contingent existence relies on explanation in terms of other contingent things—then we have the unpalatable result that contingent existence has no conclusive explanation.[12]

I submit that examination of our ordinary explanatory practices shows that explanation is transmitted from one state of affairs to another in a manner resembling the transmission of epistemic justification from one belief to another. Just as belief A does not transmit conclusive epistemic justification to belief B when A is itself unjustified, so potential explanans A does not transmit conclusive explanation to potential explanandum B when A itself is not fully explained. This is simply the way our explanatory practices work, the logic of ordinary explanation. (In an explanatory context, the explanandum is the object that is to be explained, and the explanans is that which does the explaining.)

Perhaps this is the point of the famous (but probably apocryphal) story of Bertrand Russell and the turtle lady. According to one version of the legend, shortly after publication of Russell’s book on the subject, Russell was giving a public lecture on Einstein’s general theory of relativity. The question came up: What holds the Earth in place in space? A woman in the audience offered: “The Earth rests on the back of a giant turtle—that’s what holds the planet up!” When Russell asked, “But what holds the turtle in place?” she is said to have answered: “The turtle rests on an even bigger turtle under it.” Humoring her, Russell is said to have replied: “But what holds that turtle in place?” She answered: “An even bigger turtle under it.” After several iterations, when Russell asked her again, “But what holds up that turtle?” she is said to have replied, “Sonny, it’s turtles all the way down.”[13]

8.2 Analysis of This

When the first turtle was proposed to account for what holds the Earth in place, why did Russell ask, “But what holds that up?” Because the first turtle was not by itself a conclusive explanation. For essentially the same question that started the explanatory regress arose again with respect to the turtle: a turtle sitting there in empty space is as much in need of explanation as is the Earth. Why did Russell keep repeating the question? (“What holds that one up?”) Because each new turtle added to the lineup (“It rests on an even bigger turtle”) was also inconclusive in itself. You might say that each new explanatory step made zero explanatory progress.

So, on reflection, it seems pretty clear that if a series of explanations of contingent things regresses to infinity—whether the contingent things be turtles, planets, or what have you, with each contingent thing’s existence explained only in terms of some previously existing contingent thing—then no conclusive explanation of existence is given, no matter how far back one goes. This is the reason, I believe, why the eternal universe hypothesis advanced by Parsons and Hume fails, and why the atheist’s explanatory enterprise, at least as presented by Parsons and Hume, is explanatorily unsatisfactory. So, I think Parsons is wrong on this one. An eternal universe hypothesis would not solve the problem of ultimate existence. And I believe that the account I have presented explains why the theist, even in the face of an eternal universe hypothesis, is on philosophically solid ground to keep asking why. I conclude from this that, contrary to Parsons, there is nothing philosophically odd about the theist’s question at all.

I submit that the account provided above explains why an infinite regression of nothing but contingent beings is, in itself and on its own, intellectually unsatisfying and rationally unsatisfactory.[14]

8.2.1 Preview

What if there is a way to provide a rationally conclusive explanation for the existence of a contingent universe? On purely rational grounds, wouldn’t this be preferable to an infinite regression explanation? It would seem so; isn’t it always reasonable to prefer a complete over a partial explanation? The history of science makes no sense otherwise. Later in this paper I will argue that classic philosophical theism provides a conclusive explanation, which in turn renders belief in God, as defined within classical theism, rationally preferable to an atheistic eternal universe hypothesis of the type Parsons champions in his paper.

8.3 So What is “Odd” About the Gap Question?

The upshot: Theists persist in asking “Why?” because an infinite regression of contingent beings alone just does not resolve the question of why the whole exists. What is “odd” about asking why the whole exists? Isn’t that standard philosophical practice, to keep asking why if the question is not resolved? But is there some sense in which the theist’s question is odd? Well, it is certainly an unusual question to ask; it is not like the questions we typically ask in the course of everyday life (“Who ate the last piece of pie?” “What’s on TV tonight?” “Does the car need gas?”). But most philosophical questions are unusual in this sense. And like all great philosophical questions, one only feels the force of the theist’s question once one has reached a certain level of intellectual maturity, and only when in an extremely reflective mood.

I conclude here that the question of existence is a distinctively philosophical question, one that has been asked by important philosophers from the beginning of the discipline, and as such does not seem odd in some pejorative sense of the term. However, it is definitely most at home in a philosophical context where we are seeking answers to the most fundamental questions that a human being can possibly ask.

Parsons’ deflationary argument therefore gets us nowhere. Even if we have a well-confirmed grand unified theory of an eternal universe, a philosophical question remains: Why does it all exist? The question remains because a regression of individually inconclusive explanations, no matter how far back it stretches, will never add up to anything more than an inconclusive explanation.

9. Parsons’ Second Critique of the Gap Argument: Deflationary Argument #2

Parsons’ second deflationary argument attempts to support a stronger claim, that the theist’s question (“Why does the series of things as a whole exist?”) is worse than odd: it is logically illegitimate, and therefore should not even be asked in the first place. First, argues Parsons, the theist’s question with respect to an eternal universe model breaks down into two distinct subquestions:

(a) Why is there something (instead of nothing)?

(b) Why this eternally existing universe (rather than some other)?

The problem is that:

[I]t is hard to see what motivates these questions. What is the further mystery we are trying to address when we keep asking why at this point? Why should it surprise us that there is a universe? Why should it surprise us that there is this universe?

Parsons is claiming here that there is no reason at all why the existence of the universe should surprise or puzzle us, and consequently there is no need to look for a cause of the material universe as a whole, an explanation that would be deeper than anything a well-confirmed and completed physics could provide.

Of course, as Parsons recognizes, the question of existence remains open for the theist, even in the face of an eternal universe hypothesis, because “there … conceivably could … have been nothing at all.” For the philosophical theist, it is the conceptual possibility of absolute nothingness that logically grounds the question, “Why is there anything at all?” Likewise, says Parsons, the question, “Why this universe rather than another?” remains for the theist because “there … conceivably could … have been … all sorts of other universes … instead of ours.” In other words, since we can coherently conceive of alternative possible worlds with different configurations of matter, different laws of nature, and even different types of matter, we are justified in asking the question: “Why does this universe (rather than another) exist?”

Parsons’ response to the theist on these two arguments is interesting and, as far as I know, original:

Of course, we can imagine that there might (i.e., conceivably could) have been nothing at all or that all sorts of other universes might have existed instead of ours, but this need not create any mystery. There are always innumerable imaginable possibilities whose failure to be realized creates no mystery at all. The moon could conceivably have been made of cheese, but it is no mystery that it isn’t. In general, it is no mystery why something does not exist unless, given our background knowledge, its existence was expected, or at least no more unexpected than what does exist. Nothing in our knowledge base supports the slightest expectation that the moon would be made of cheese. Nor do we have any basis for thinking that some other (ex hypothesi eternal) universe should have existed all along instead of ours. Therefore, it is hard to see how asking “Why doesn’t some other universe exist?” is very different from asking “Why don’t we have a moon made out of cheese?”

9.1 The “Principle of Parsony”

Parsons’ wording suggests the following principle of explanation, which I shall name the principle of Parsony:

In general, the nonexistence of X is no mystery unless, given general background knowledge, its existence is either expected or is no more unexpected than what does exist (i.e., its existence is at least as expected as that which does exist).

Now, since we have no rational basis for expecting the moon to have been made of green cheese, the principle of Parsony accounts nicely for the fact that the nonexistence of a cheese-moon is no mystery at all and needs no explanation.

The application of the principle of Parsony to the material universe apparently seems straightforward to Parsons: In scientific contexts, he claims, when just one specific outcome out of a set of possible alternatives occurs,

we naturally and rightly assume that there is some reason why this happened rather than that. We rightly assume that there were physical antecedents and relevant physical laws that determined, at least probabilistically, the occurrence of one event out of a contrast class. But when the “occurrence” we are interested in is the existence of the universe as a whole, by definition there are no physical antecedents or applicable natural laws that justify the expectation of an answer to our question “Why this rather than that?” What could make the universe as a whole probable or improbable, expected or unexpected?

Presumably nothing. From which Parsons concludes that we have no rational basis for expecting that some other universe might have existed instead, or that there might have been nothing at all, just as we have no basis for expecting that the moon might have been made of green cheese. The existence of an ex-hypothesi eternal universe, therefore, is no more of a mystery than the nonexistence of a cheese-moon. It follows that an eternal universe is no mystery at all and calls for no explanation. The theist’s two cosmological questions are thus exposed as logically illegitimate: there is no good reason why they should even be asked in the first place. Says Parsons:

In the absence of grounds for thinking this universe, given that it has existed eternally, is unexpected, surprising, or improbable … the correct riposte to the questions “Why this universe?” or “Why any universe at all?” is “Why not?”

10. Why the Principle of Parsony is Unacceptable

But is the principle of Parsony true, or even plausible? Why should we accept it? What is the reason for thinking it true? Is there a noncircular, independent argument for it? Does the principle reflect our actual explanatory practices? First off, the principle of Parsony is certainly not necessarily true: its denial is not a contradiction or an impossibility. Neither is it self-evident. So it stands in need of substantiation or justification. Thus it is curious that Parsons offers no philosophical argument for it at all. But aside from this, the principle of Parsony simply does not match standard philosophical or scientific explanatory practices. Consider the four following counterexamples.

One day in the physics lab, Susan, an undergraduate, overhears some grad students talking about the charm quark, c, and the fact that it has a 2/3 charge. Knowing nothing about quarks except that they are particles inside protons and neutrons, she asks: “Why does the charm quark have a 2/3 charge (rather than no charge at all, or rather than some other charge)?” Common sense suggests that Susan’s question is a perfectly reasonable one, even if (to paraphrase Parsons) nothing in her knowledge base supports the slightest quantitative expectation that the charge of the charm quark would be some other number, or that it would have no charge at all—and even if she has no background reason to expect that some other charge is “no more unexpected than” the actual 2/3 charge. In other words, it seems to me that Susan’s question is legitimate without the basis of expectations required by the principle of Parsony.

A philosopher is pondering the free will problem. “Why,” she asks, “do we have this strong intuitive feeling that the future is open to us, when we act as if it actually is not?” Now, nothing in her knowledge base leads her to expect that we should or should not have this feeling. And the persistence of the feeling is not more unexpected than its opposite, the intuitive feeling that the future is not open. Yet even in the absence of the rational, probability-based expectations required by Parsons’ principle, this philosopher still wonders and still seeks an answer—in accord with standard philosophical practice. Again, asking the question seems rational even in the absence of the basis of expectations required by the principle of Parsony.

Imagine Thales sitting on the dock at Miletus pondering the problem of the one and the many, the first distinctively philosophical question on record. Nothing in his experience justified a quantitative expectation of the sort required by Parsons that there would, or would not, exist one ultimate unity behind the apparent plurality of things. Yet he asked the question, and he gave an argument for the conclusion that he reached—the first distinctively philosophical argument ever recorded. Good question, but no Parsonian basis of expectations.

In a class on medieval philosophy, a student learns that in approximately 1248 Thomas Aquinas walked from Italy to Paris to study under Albert the Great at the University of Paris. (When Aquinas arrived, Albert had already moved to a university in Germany, so Thomas walked all the way to Cologne to start his graduate work.) Without knowing how often people walked in those days, rather than rode a horse or donkey, and thus without any basis for making the required Parsonian estimates of probability, the student nevertheless asks: Why did he walk to Paris (rather than ride an animal, and rather than stay home)?

In these and other realistic cases of explanatory practice, the underlying explanatory principle is not the principle of Parsony (POP), but instead seems to be what I shall name the daring inquiry principle (DIP):

When confronted with the existence of some unexplained phenomenon X, it is reasonable to seek an explanation for X if we can coherently conceive of a state of affairs in which it would not be the case that X exists.

Reflection on my proposed counterexamples supports, I submit, DIP over POP. In addition (and I believe this is a significant point), I submit that if anyone will reflect on the nature of intellectual history, it is DIP rather than POP that has driven a great deal of human intellectual inquiry over the past several thousand years, not just in philosophy, but in every field of academic thought. When you look back at the historic leaps of intellectual inquiry, DIP has been the driver, not POP.

So why suppose the principle of Parsony is true? At least in Parsons’ paper, the principle is nothing more than an assertion.

11. Another Try: The Physicalist Interpretation of POP

Parsons next supplements POP with the claim that, in general, the only legitimate basis for expecting one thing or another to be or not to be are “physical antecedents and laws of nature.” (Hence these are the only grounds for expecting that some other universe might have existed, or that it might have been that nothing at all exists.) I shall call this the “physicalist interpretation” of POP. Parsons does not give an argument in support of this claim, although he does tell us that in a separate paper he has criticized one well-known attempt at providing nonempirical grounds for holding that the existence of the universe as a whole is surprising, unexpected, and stands in need of explanation. That attempt was Richard Swinburne’s argument that, even without nontautological background information, we can judge on a priori grounds alone that the existence of our material universe is highly improbable, and therefore surprising and in need of external explanation.

Although Parsons does not repeat his criticism of Swinburne’s argument in his paper, he does provide the conclusion of that criticism: that the a priori grounds employed by Swinburne (to motivate his claim that the existence of the material universe is unexpected and therefore needs explanation) “lead only to a maze of metaphysical imponderables from which there is no escape.” (A proposition is said to be known a priori if its justification does not depend on our sense experience of the world, i.e., on the information that we receive from our physical senses—sight, touch, hearing, taste, and smell.)

But Parsons must be wrong about this; physical antecedents and laws of nature cannot be the only legitimate grounds for the types of expectations allegedly governed by POP. Here is why:

(a) We obviously do have knowledge of the physical antecedents and laws of nature required by POP, and this knowledge is certainly empirical in nature (empirical or a posteriori knowledge is knowledge justified on the basis of our sense experience of the world, in other words on the basis of data provided by our physical senses—sight, touch, taste, etc.).

(b) But if the physicalist interpretation of POP were strictly true, we could never attain knowledge of physical antecedents and laws of nature in the first place. Consider the exact moment just before the first instance of coming to know a physical antecedent or a law of nature. At that moment, we would not know any physical antecedents or laws of nature. But then (according to POP) we could not ask questions such as “Why does this rather than that happen?” and “Why does this rather than that exist?”

Yet questions of this type are surely necessary for any scientific investigation of the world, including the discovery of the very physical laws of nature and physical antecedents required by POP. In other words, at the beginning of all empirical investigation, lacking the basis of expectations required by the physicalist interpretation of POP—knowledge of physical antecedents and laws of nature—we could not legitimately ask the sorts of questions that must be asked if we are to pursue and attain the empirical knowledge presupposed by the physicalist interpretation of POP. Thus, the physicalist interpretation of POP is self-referentially incoherent: it presupposes that we possess a particular type of knowledge, while at the same time making it impossible that we ever attain that type of knowledge in the first place. The physicalist interpretation of POP cannot be right if we have any empirical knowledge at all.

So much for POP: the theist’s question of existence survives. When judged in the context of our normal explanatory practices, the theist is asking legitimate (albeit philosophical) questions when he or she asks, “Why does something, rather than nothing at all, exist?” and “Why this material universe and not a different one?” This is true even on the assumption of an eternal universe, for as Parsons himself admits, “we can imagine that there might (i.e., conceivably could) have been nothing at all,” and we can coherently conceive that “all sorts of other universes might have existed instead of ours.” But, as I have argued, that is logically sufficient to ground the theist’s question concerning the existence of the universe as a whole.

11.1 Preview

Shortly I will argue that: (a) philosophical theism provides us with a (philosophical) hypothesis that answers the ultimate question of existence, while atheism does not and cannot; and (b) this counts as a pretty good explanatory reason to believe that God exists. After all, if one hypothesis H1 accounts for the data, while a second hypothesis H2 does not, that is a good reason to prefer H1 over H2. But before I get to this part of my argument, back to Parsons.

12. Does the Principle of Sufficient Reason Justify asking Varghese’s Question?

Seeking to put an end, once and for all, to the theist’s existential question about the universe as a whole, Parsons next takes on what may be the most historically important reason theists have given for seeking an answer to the question of existence—the principle of sufficient reason (PSR). Though this was first given philosophical expression by the great German philosopher and mathematician Gottfried Leibniz (1646-1716), Parsons renders it as “Nothing can be so unless there is a sufficient reason why it is so.” Many theistic philosophers have claimed that this principle is a necessary truth, and it has served as a key premise in many different philosophical arguments for God’s existence.[15]

The PSR is indeed one reason why some theists believe that there must be an answer to the questions: “What explains why there is something rather than nothing?” and “Why does this universe exist rather than some other?” If the PSR is true then there certainly must be an answer, in which case the question of existence is a legitimate one. So the PSR, if true, would legitimate the existential question.

12.1 “No” Argues Parsons

Parsons considers PSR, but only to repudiate it on the grounds already given by J. L. Mackie in his own critique of theism, The Miracle of Theism. In Mackie’s words, “[t]he principle of sufficient reason expresses a demand that things should be intelligible through and through…. [but] nothing justifies this demand and nothing supports the belief that it is satisfiable even in principle.”[16]

But there is more. In a passage quoted by Parsons, Mackie argues for a much more radical (and potentially devastating) conclusion, namely: not only have theists provided no adequate justification for PSR, and not only have they not shown that it can, at least in principle, be satisfied, but PSR necessarily cannot be satisfied. If Mackie is right about this, then PSR is not merely unjustified, and it is not merely false; it is necessarily false. Mackie’s famous argument goes approximately as follows:

  1. It is a conceptual truth about explanation that every explanation contains an explanans considered brute for the purposes of the explanation, which is then employed to explain or account for the explanandum.

  2. Even if the explanans of one explanation is in turn explained by a deeper explanation (such that the explanans of the first explanation becomes the explanandum of the deeper explanation), the deeper explanation in turn will also start with an explanans considered brute for present purposes, which will in turn become the explanandum of a still deeper explanation, and so on back and back.

  3. It follows that no matter how far back our actual working explanations take us, the explanans furthest back must always be brutely factual (since being furthest back it is not explained in terms of a previous explanation).

  4. Thus, it is simply not possible that all things ever be explained through and through, leaving no brute facts whatsoever.

  5. Hence PSR cannot be satisfied.

  6. Therefore PSR cannot be true, it is necessarily false.

Mackie quickly adds: “But there is no need to see this as unsatisfactory.”

If Mackie is right, it is a conceptual truth about explanation that every explanation contains a part considered brute for the purposes of the explanation. This entails that every explanation must leave brute facts in its wake. Parsons concludes from this the ultimately inevitable bruteness of existence: brute facts must collect and remain forever unresolved at the base of any explanatory chain, like coffee grounds permanently caught in the drain trap of a sink. No escape from brute contingency will ever be possible, and the theistic hope for a conclusive explanation of the whole of reality is necessarily a vain hope.

After quoting (and endorsing) Mackie’s argument, Parsons makes a telling comment:

If we repudiate the PSR, then we will no longer automatically infer that the fact that something lacks an explanation means that it requires one. It might well be entirely reasonable for some logically contingent things just to be and to have no reason for their existence (my emphasis).

The upshot: If Mackie’s argument is sound, brute (unexplained) facts will always be with us; they are an unavoidable facet of our contingent existence, an unconquerable mountain standing in the way of our explanatory powers. I consider Mackie’s argument to be a very important one: it is, in a sense, his Gödel’s Incompleteness Proof for explanation. But is Mackie right?

13. A Reply to This Argument

Every explanation in science clearly contains an explanans that is brute for the purpose of the explanation. (And that means that science alone will never answer all questions of existence.) But it is not at all obvious or self-evident that this is true of all explanations, such as philosophical ones. Thus an argument is needed for the universal claim.[17]

Consider one way in which the universal claim that all explanations must leave something completely brute in their wake might be false. Suppose there were an existentially self-subsuming explainer—an explainer of existences whose own explanatory field curves around to account for its own existence (and other existents explained in the explanandum). In that case, the hypothesized existence of the existentially self-subsuming explainer could serve as the explanation for the existence of a whole realm of things without leaving its own existence brute (and thus without leaving any irreducibly brute facts of existence in its wake). This idea may be summed up by saying that such an explainer would be a “self-explaining explainer.”

We are considering a self-explaining explainer of existence only as a theoretical possibility at this point. I am not claiming that there is such a thing. My point is that such an explainer, although unlike anything ever encountered in physical or natural science, would be a counterexample to Mackie’s assertion—unbacked by argument—that all explanation leaves behind something brute in its wake. Wouldn’t it?

Since nothing Mackie says rules out the theoretical possibility of an existentially self-explaining explainer, and since the concept is not on its face self-contradictory, it would seem to behoove the honest seeker of ultimate explanation to ask: Is such an explanation possible—an explanation of existence in terms of a self-explaining explainer, an explainer whose explanatory field leaves no purely brute existential facts in its wake? Let’s see.

13.1 Self-Subsuming Explanation

There is a rather obvious candidate for self-subsuming explanation, namely, the type of explanation actually referred to in passing by Parsons himself:

Clearly, the PSR entails that if the chain of sufficient reasons is not to extend ad infinitum, then all contingencies must ultimately be grounded in something that is, in some sense, necessary.

I appreciate that Parsons at least mentions in passing the explanatory strategy of grounding the contingent in that which is necessary. Such “necessity grounded explanation” actually has a long and venerable provenance in the history of both philosophy and mathematics. I only wish he would have pursued the idea a little, for the trail does not grow cold. Rather, I shall argue, reasoning about it naturally leads to a fascinating and rationally satisfying solution to the ultimate problem of existence—a self-explaining explainer that logically entails philosophical theism. The investigation of self-subsuming explanation thus ends naturally in an explanatory argument for God’s existence.

Having said that, the burden is now on me to show that a self-subsuming explanation is available, that it can provide a conclusive, logically sufficient explanation for the existence of the material universe, and that it entails philosophical theism (as defined in this paper). Certainly if there is such an explanation, it would be rationally preferable to the solution offered by Parsons—that the existence of the universe is absolutely brute—for then atheism offers no explanation for why a contingent universe exists. I shall back up my claim in distinct stages.

13.2 Stage One: The Elemental Modal Distinction: Necessary and Contingent Truth

A proper understanding of explanatory self-subsumption must begin with the distinction, central to logical theory, between that which is necessary and that which is merely contingent.[18]

The ancient Greek philosophers were the first to give philosophical expression to this distinction—probably first drawn in an alethic context (with respect to truth and falsity). After philosophical reflection on the nature of mathematical proof, it seems likely that someone noticed a radical difference between the way in which a pure mathematical truth (such as the Pythagorean theorem) is true, and the way in which something mundane like “Thales has a beard” is true.

The Pythagorean theorem is not only true, it is so rock-solidly true that it cannot even possibly be false; its denial is not even consistently conceivable or describable. Its truth is intrinsic, unloseable, essential. In ordinary language, if something cannot possibly be otherwise, we say that it is “necessary.” The word necessary means “cannot be otherwise.” Thus it was natural to call the Pythagorean theorem a “necessary” truth.

Pythagorean theorem is a theorem of Euclidean (noncurved) geometry. It states that on a right triangle in a noncurved space, the square of the hypotenuse is always equal to the sum of the squares of the two other sides, where the hypotenuse is the side opposite the right angle.

By contrast, the proposition that Thales has a beard, though true, could have been false. There are consistently conceivable or describable circumstances in which it would have been false that he had a beard (for instance, if he had previously shaved it off). Thus, there are possible circumstances in which the proposition is true, and possible circumstances in which it is false; the truth or falsity of “Thales has a beard” is therefore dependent on which external circumstances happen to obtain. The proposition that Thales has a beard is thus neither intrinsically true to its core, nor intrinsically false to its core; it is neither true nor false essentially, or in its own right. You might say that it is a logical “flip-flopper,” capable of being either true or false depending on circumstances. It is therefore appropriate to call such a proposition contingently true, since in ordinary language “contingent” means “dependent.”

Thus, in logical theory we say that a proposition is necessarily true if and only if it is true in all possible circumstances, and thus false in absolutely none. A necessary truth like Pythagorean theorem is therefore steadfast across all possible circumstances. One might say that it is so rock-bottomly true that it cannot even possibly not be true. It is intrinsically true, essentially true, if you will; true in its own right and not on account of something external that made it true. This is truth independent of all circumstance.

A proposition is contingently true if and only if it is true in some possible circumstances and false in others; thus its truth is dependent on which circumstances obtain and which do not. It is made true (or false) by external circumstances. A contingent truth like the proposition that Thales has a beard is not so steadfast as a necessary truth; it can be either true or false, depending on the circumstances. Such a truth is dependent on circumstance.

13.3 Stage Two: From Truth to Being

13.3.1 Contingent Being

It is not that much of a leap from a recognition of two modes of truth—the contingent and the necessary—to the thought that there might be two corresponding modes of existence or being as well. Here is one way into the distinction. Begin with the existence of any ordinary thing around you, for instance: a rock, a tree, a lake, a mountain, a planet, a bird, a fellow human being, or even a molecule, an atom, or a subatomic particle such as a quark. Next, think about the many conditions that are required if that thing is to remain in existence. For example, suppose you are thinking of a human being. Certainly oxygen, food, water, and temperatures within a certain range are required if a human being is to remain in existence. Suppose that you picked a lake. Surely a lake depends for its existence on numerous geological preconditions, and so on.

Now think about the many conditions that would end the existence of the item you picked. For example, all human life would end if oxygen ceased to exist, all lakes would dry up if the temperature did not stay within certain limits, and so on. Even an ordinary atom or molecule will go out of existence under certain circumstances, and the typical subatomic particle is toast if it encounters its antiparticle. The existence of each of the ordinary material things around us is constantly threatened at every moment by various external circumstances, any one of which could end its existence at any time.

Next consider beginnings of existence. Conditions must be just right if any one of these ordinary things is to even come into existence. A lake won’t even form unless the temperature is above a certain minimum. A human baby won’t be born if food, water, and oxygen are not first available for the mother. An island won’t pop up unless certain necessary geologic preconditions are just right. And so on.

So you could say that, for ourselves and for the ordinary material things around us, existence hangs by a thread. Existence is for us a precarious, insecure affair, constantly under the threat of nonbeing. This is truly “existence on the edge.” For this reason, the adjective contingent (signifying dependency) is certainly an appropriate one for the type of existence we share with the ordinary material or physical things around us—a type of existence that is radically dependent on external conditions from start to finish, and at every turn.

Thus we get the strict definition of a contingent entity or being: there are possible circumstances in which it would exist, and possible circumstances in which it would not exist. To which it is seems logical to add: If certain conditions had not first existed, then it would never have come into existence; and if certain conditions fail to obtain, its existence is toast, snuffed out like a candle in the rain. This is the miserable contingency, the “might-not-have-been-ness,” that is our lot, and that is the lot of all the mundane or material things around us, from atoms to molecules, chemicals, rocks, lakes, mountains, planets, stars, galaxies, and so on.

13.3.2 Necessary Being?

But reflection cannot stop here. To stop with contingency is to stop in the logical middle of a process, like stopping at right without admitting left, or stopping at top without admitting bottom, or admitting half without admitting whole. Could there, at least in principle, be something that is noncontingent? If so, what would it be like? A completely noncontingent being, if such were to exist, would logically have to be a thing whose existence is not dependent on any external circumstances (nor ever could have been). It would seem to follow that its existence could not possibly ever have been threatened by any possible circumstance or entity, nor ever could be. Moreover, no external circumstance or entity could possibly have caused its existence (or prevented or aborted it). That is the only way that such a being could exist in a manner completely independent of all possible circumstance, that is, independently of all contingencies. This is all a matter of pure logic.

In theory, such a being would have existed no matter what, and would continue to exist no matter what. That is, it would exist in all possible circumstances, and be nonexistent in absolutely none. This entails that its nonexistence is an impossibility: if there are no possible circumstances in which it does not exist, then its nonexistence is not possible. Thus it has always existed and always will, as there never was a time when it did not exist, nor will there ever be a time when it does not exist (or else its existence would be dependent on circumstances beyond itself, namely, temporal circumstances).

Thus we see that the easily conceivable concept of contingent being has an opposite, a logical counterpart that reflection is bound to discover if pushed far enough. Abstracting from the concept of contingent being, we naturally reach the concept of noncontingent being. Since “necessary” means “cannot be otherwise,” the term “necessary existence” or “necessary being” is appropriate for this theoretical category or mode of being. Just as “top” cannot ultimately be admitted without “bottom,” and just as “fast” cannot be fully grasped without reference to “slow,” the concept of contingency cannot be fully grasped until it is understood as the immediate logical counterpart of necessity. Necessity is the logical opposite of something that might not have existed had circumstances been sufficiently different; it is something that could not possibly have not existed. And so the philosophical mind, upon sustained reflection, reaches the refined and exalted concept of necessary being.

So, whereas a contingent being typically owes its existence to some external circumstance, entity, or process that preceded it in existence and caused it to be (for example, you received existence from your parents), a necessary being would owe its existence to absolutely nothing external to itself.[19] But let us go just a little further.

The concept of necessary being is not logically primitive; it can be and has been analyzed, and the analyses are fascinating, though this is not the place to enter them in any depth. Medieval philosophers argued that a necessary being, if one existed, would be “self-existent” (in contrast to contingent beings, which owe their existence to factors outside themselves). As a first approximation, by self-existent they meant that a necessary being would exist “from within itself,” or “of itself rather than of (from) another being or circumstance.” Certainly this is on the right track, for since a necessary being would not depend on external circumstances for its existence in any way, it would exist wholly and absolutely on its own steam, so to speak. Some philosophers put the idea this way: A necessary being, if one existed, would contain within its own essence the fundamental ground of its existence. (This would be one sense in which a necessary being is a self-explanatory being.)

Other philosophers have explained “self-existent” by theorizing that in the case of a necessary being, and only in the case of such a being, essence and existence must be identical. In other words, being and essence must be one and the same thing, not two different things (as is the case with contingent beings). Thus, whereas when a contingent thing exists this is always the result of a union of being with essence, in the case of a necessary being, its very essence is identical with its to be. (In logical theory, when we say that A “is identical with” B, we mean that A and B are one and the same thing, not two different things; as, for example, Bob Dylan and Robert Zimmerman are one and the same person, not two different persons, for the two names refer to one and the same person.)

This theory of necessity as an identity of being and essence was first given systematic philosophical expression by Thomas Aquinas, the greatest philosopher of the Middle Ages, in his early essay On Being and Essence. The best contemporary treatment and defense of the idea, in my opinion, occurs in a new book by Steven Duncan, where the idea is developed systematically in the context of working out a solution to the philosophical problem of the external world.[20]

13.4 Stage Three: Reflection

A certain realization naturally emerges after sustained and mature philosophical reflection on the existence of the material universe. It is not the mere existence of the many things around us (atoms, molecules, dirt, trees, rocks, lakes, stars, galaxies, etc.) that is puzzling and cries out for explanation; it is the contingency of their existence, their sheer might not have been-ness. Jean-Paul Sartre was drawing our attention to this when he referred to the existential gratuity of all that surrounds us:

[C]ontingency is not an illusion, an appearance which can be dissipated; it is absolute; and consequently perfectly gratuitous. Everything is gratuitous, that park, this town, and myself. When you realize that, it turns your stomach over and everything starts floating about.[21]

But doesn’t this naturally raise the question that Varghese asks: Why does it all exist?

13.5 Stage Four: Solution

Now imagine a necessary being with certain features: sufficient power, knowledge, and will to create a universe; free will in the normal, everyday-language sense of the term “free” (i.e., the power to initiate a course of action, to make a choice that is not necessitated, caused, or predetermined by the sum total of all antecedent conditions and external circumstances or forces); and the will to freely give the gift of existence to a world of creatures, either directly by divine fiat, or through an independent evolutionary process. It seems crystal clear that, hypothetically, that sum total—a free act of creation by a sufficiently knowledgeable and powerful necessary being—would all by itself constitute a logically sufficient causal condition for the existence of a contingent universe. (Roughly, S is a sufficient causal condition for some phenomenon P if and only if the fact that S obtains logically implies that P is the case. In other words, P must be the case once S is the case.) Of course, such a being would logically have to be personal in nature, given the definition of a person as a being capable of (among other things) making consciously willed choices.[22]

Suppose that a necessary being with the requisite powers described above exists and freely wills that a contingent universe exists. Then it would logically have to be the case that a contingent universe exists: a necessary being with the requisite powers (for instance, omnipotence) choosing to create a universe is all you need to get a universe. That is, a universe follows if such a choice by such a being is made. It is a matter of strict deductive logic: That the stated condition obtains—a sufficiently knowledgeable and powerful necessary being chooses to create a contingent world—logically implies that a contingent universe exists.

Thus, the necessary being hypothesis provides a sufficient cause, a sufficient reason, or sufficient explanation (depending on your preferred terminology) for the contingent existence of the material universe as a whole.

This is a very significant result. A sufficient condition would produce or result in the existence of the item in question. Citing a sufficient causal condition for the existence of something X that stands in need of explanation (a “sufficient reason” or “sufficient cause” for X) is one important way we explain things in science, philosophy, and many other explanatory contexts. (Of course, we also need to rule out other possible explanations for the existence of X before we finally conclude that we have found the actual cause of X.) There are exceptions (which I will explain below), but in many explanatory contexts, to find the sufficient casual condition or sufficient reason for something X simply is to adequately explain the existence of X. And note: The sufficient cause or reason for something is normally considered to be a rationally satisfying explanation for its existence.

Intuitively this explanatory move—appealing to something noncontingent in order to account for that which is contingent—seems to be on the right track. For in the normal course of explanation, we appeal to something outside of a given category of things in order to adequately explain the existence of the things within that category as a whole. For instance, to explain why there are cars in general, we find that we must refer not to more cars, but to those who design and build cars—to engineers, factory workers, and so on. These are all things that do not themselves belong to the category of cars. Likewise, to explain why there are houses, we find that we must appeal not to more houses, but to something outside the category of houses altogether—to carpenters, electricians, wood, etc. So, right off the bat, the theist’s explanatory move—appealing to something noncontingent in order to account for the contingent as a whole—seems in harmony with our normal explanatory practices.

So let’s take stock. Philosophical theism provides us with a sufficient cause, sufficient reason, and sufficient explanation for why a contingent universe exists. On the other hand, atheism, on Parsons’ own admission, does not; rather, it leaves the existence of the material universe an unexplained, purely brute fact. Our contingent universe has no explanation, claims Parsons; it just is. But in science and philosophy, and in many explanatory contexts, providing a logically sufficient reason for something is normally considered a very respectable form of explanation. The philosophical theist adds: It sure beats the only alternative—leaving it all just unexplained.

On the basis of an inference to the best explanation, one might reason from the contingency of the material universe to the existence of a necessary being responsible for the existence of the universe. This is an explanatory argument for God’s existence.

It looks like we have come full circle. If true, the necessary being hypothesis would explanatorily “ground” all contingencies, in Parsons’ words, “in something that is, in some sense, necessary.” Thus it seems that we have arrived at the explanatory possibility mentioned in passing by Parsons, which I named “necessity grounded explanation.” But would this be a self-subsumptive explanation, one that ends the regress of questions of existence at a self-explanatory stopping point? Let’s see.

14. A Critique of This: The Problem of the Infinite Re-Asker

At this point an obvious problem arises. If we explain the existence of contingent beings by reference to a noncontingent being (more precisely, the free creative act of a necessary being), then a seemingly devastating reply suggests itself right away:

You maintain that the existence of contingent beings needs a sufficient explanation, and so you hypothesize a free act of creation by a sufficiently powerful and knowledgeable necessary being. But doesn’t essentially the same question arise all over again with respect to the necessary being? What explains the existence of it? What accounts for its existence? What is its explanatory ground? And once we ask this, aren’t we back where we started, needing another explanation of existence, in terms of a more fundamental level of being, and then another, and another, with no real explanatory progress made?

This is a powerful reply, a potentially devastating objection. What can the defender of the necessary being hypothesis (“NB” for short) say in response? Suppose the defender of NB replies:

What explains or accounts for the existence of the necessary being? Why absolutely nothing at all, it has no explanation!

That certainly won’t do, for if that is the theistic reply, then an atheist such as Parsons would have every right to say the very same thing with respect to the existence of the contingent universe:

What accounts for the existence of the contingent universe? Why nothing at all, it has no explanation whatsoever!

If the theist has just finished saying that the existence of a necessary being has no explanation whatsoever, then he or she certainly cannot complain if the atheist makes the statement that the universe has no explanation. If the theist were to complain that atheism leaves something unexplained, the atheist would be within his rights to respond: Tu quoque! (“You do it too”!). In that case, atheism and theism would be on a par, with both leaving things unexplained in the end.

Suppose instead that the theist answers:

What explains the existence of the necessary being—call it “N”? Why, a more fundamental necessary being N1, an even “necessarier” being than N.

Wouldn’t essentially the same question arise with respect to the existence of N1? And wouldn’t this require basically the same answer again, an appeal to an even “necessarier” necessary being N2? And wouldn’t the question repeat again, giving us a prior necessary being N3, and again, giving us N4, and so on to infinity? The inevitable result would be an infinite regression of necessary beings, wouldn’t it? Why would this be an improvement over an infinite regression of contingent beings? It’s hard to see how it would.

The astronomer Carl Sagan, author of Cosmos, put it this way:

If we say ‘God made the universe,’ then surely the next question is ‘Who made God?’ If we say, ‘God was always here,’ why not say ‘the universe was always here’? If we say that the question, ‘Where did God come from?’ is too tough for us poor mortals to understand, then why not say that the question of ‘Where did the universe come from?’ is too tough for us mortals? In what way, exactly, does the God hypothesis advance our knowledge of cosmology? (cited in Varghese’s Introduction, p. 4).

So, the defender of the necessary being explanation seems to be in a bind, for essentially the same problem of existence seems to arise all over again if we hypothesize a necessary being as the ground of existence for the (contingent) material universe. For once we posit a necessary being, it seems we then need a further explanation for its existence, an explanation that refers to something more fundamental still, which leads us to “re-ask” the question again and again: “But what explains its existence?” “Now what accounts for that?” And it seems that we are back to where we started, trapped in an infinite loop, an infinite regression of “What explains the existence of that?”

But this line of questioning is wonderful because it pushes us to think more deeply about the problem of existence, and about what might count as a solution to it. It also leads us to new philosophical insights, one of which (I shall argue) is a solution to the ultimate problem of existence. Let’s see.

14.1 How to Stop the Infinite Re-Asker

The question is: Doesn’t NB simply bring up the question of existence all over again, causing us to ask, “But what accounts for the existence of the necessary being?” And doesn’t this just generate an infinite regression of necessary beings, leaving us no better off than when we started—saddled with unexplained existences? Doesn’t NB just make “re-askers” of us all? These are very good questions, indeed profound questions. But I believe there is a very good reply, indeed a profound reply, one which was put concisely by Peter van Inwagen in his textbook Metaphysics:

But [to ask, of a necessary being, “Why does it exist?” or “What explains its existence?” ] is to neglect the fact that a necessary being is one whose non-existence is impossible. Thus, for any necessary being, there is by definition a sufficient reason [i.e., explanation] for its existence: there could hardly be a more satisfying explanation for the existence of a thing than that its non-existence was impossible.[23]

This reply may sound too simple at first, but its surface simplicity belies its true depth. I submit that once one carefully mulls over van Inwagen’s answer, and once one fully understands its meaning and significance, necessary existence comes into focus as the only existentially self-subsuming mode of existence, and necessity grounded explanation is revealed as the only self-subsuming, self-explanatory form of explanation we possess. Let’s see why.

A necessary being, if one existed, would by hypothesis be a being whose essential nature is such that its nonexistence is absolutely impossible. It follows that in the case of a necessary being N, the explanation of it existence (the explanatory ground of its existence), is a necessary logical fact internal to its own being—internal in the sense that it obtains independently of all possible circumstances. And what is that internal fact? The impossibility of N’s nonexistence. Van Inwagen is right: Surely the impossibility of some being N’s nonexistence would be a logically sufficient condition of N’s existence. (This is a matter of pure logic.) In this exact sense, a necessary being contains within itself the sufficient condition of its own existence. This is self-explanation if there ever was such.

The reflective mind reaches out beyond the category of contingency because, as we have seen, if all things are merely contingent, then there is no explanation for why contingent things exist, nothing that might account for the existence of contingent beings as a whole. And what is the only place to land outside of contingency? Explanation in terms of noncontingent being—in other words, explanation grounded in a necessary existence.

Given the reasoning above, it seems clear that the necessary being hypothesis, if true, would leave behind no existential facts “considered brute for the purposes of the explanation”—contrary to Mackie’s claim that all explanations presuppose something brute. I conclude that NB seems to be just the type of explanation we need in order to bring a regress of explanations of contingent existence to a rationally satisfying conclusion. What other explanatory option do we have?

So, the concept of necessary existence presents us with an existentially self-subsuming explanation, one that could form a rock-bottom endpoint for a universal explanatory regress. This is an existential basis that would account at once both for its own existence and for the contingent existence of everything else, leaving no brute existence in its wake. Many philosophers have thought that if contingent existence were to be explained by reference to a metaphysically necessary ground of being (or, à la Parsons, if all contingencies are ultimately grounded in something necessary), then no existence would be left brute. All existence would then be accounted for, since a necessary being, if one were to exist, would be existentially self-explanatory in the sense defined above (and in contrast to a contingent being, whose existence can only be explained in terms of something external to it). Although Parsons does not examine and drill into this explanatory option, this is the explanatory advantage we get if we postulate a being whose essence is such that its nonexistence is simply, fundamentally, absolutely, and metaphysically impossible.

Perhaps it is now clear why many philosophers maintain that if explanations of contingent existence ultimately regress back to a metaphysically necessary ground of being, then all existence, including the rock bottom level, is accounted for, and brute fact is conquered. In Metaphysics, van Inwagen also wrote:

Why should there be anything at all?… If we could show that there was a necessary being … we should have an answer to our question. For if there were a necessary being, then it would be impossible for there to be nothing. And if we could show that it was impossible for there to be nothing, that, surely, would count as an answer to our question.[24]

Does NB resolve the problem of existence so as to leave no facts of existence brute? It certainly looks that way to me. If so, then Mackie and Parsons are wrong when they claim that, as a matter of necessity, every explanation must leave irreducible, pure brute facts in its wake. The category of necessary being provides us with a self-subsuming explanatory ground, one whose explanatory field accounts both for its own existence and the existence of the contingent beings dependent on it for their existence, leaving no brute facts of existence in its wake.

15. A Powerful Objection to All of This: The Dilemma Argument

During review of this paper, an anonymous referee for Internet Infidels raised a very powerful objection to my argument: If the universe is contingent, then its existence cannot possibly be explained in terms of necessity alone, for the following reason. Consider the connection between a necessarily existent God and a contingently existing universe. Is the connection necessary, or is it contingent? It must be one or the other, for there is no middle ground between necessity and contingency. If the connection is necessary, then the choice to create is a necessary consequence of a necessarily existing being, in which case the existence of the universe would be necessary and not contingent (which contradicts the working assumption that the universe exists contingently). This is so because it is a standard principle of logic that anything that follows necessarily from that which is necessary must also be necessary. So the connection between the necessary being and the universe cannot be necessary, it must be contingent. But, if the connection is contingent, then what explains or accounts for it? Certainly not something wholly necessary, for the reason already given.

So, either the connection has no explanation, and theism is stuck with a brute contingency; or else the connection is explained in terms of some other contingent fact, which in turn must be either brute or contingent, and the same problem recycles. Either way, we inevitably end up with some fact that is brutely contingent. Thus, it follows that pure, unresolved, and brute contingency is inescapable if there is any contingency at all. It is not only impossible to account for the contingent in terms of the necessary, then, but impossible to account for the contingent at all. Our explanation of the universe must, in the end, bottom out in pure brute contingent fact.

15.1 The Dilemma for the Theist

What a great counterargument. Of all the objections that might be raised against my argument, I find this to be the most powerful objection of all. (Bear in mind that every philosophical argument, no matter what the subject matter and no matter what side of the issue, is subject to objections—there are no exceptions anywhere in philosophy.) I value this objection because it causes us to probe the issue further than we otherwise would, and because it leads to new discoveries (which I shall present shortly). So, what can a philosophical theist say in response to this very thoughtful objection? Quite a bit, actually.

To begin with, the reviewer is certainly right that an act of choice by a necessary being (to create a contingent universe) cannot itself be necessary, for if the choice is necessary, then the material universe exists necessarily, not contingently, contrary to the standing assumption that the universe is contingent. For the following is indeed a standard principle of any acceptable system of modal logic: If proposition P is necessary, and P logically implies Q, then proposition Q must also be necessary. In other words: anything that follows necessarily from something that is itself necessary must also be necessary. So the divine choice to create cannot follow necessarily from the divine nature, as it cannot be necessitated by preexisting conditions in any way; the choice to create has to be contingent.

But if we suppose that the choice is contingent, then we face the reviewer’s next question: What explains it? Certainly the theist cannot reply that the choice has no explanation, for then the divine choice to create is itself a brute (unexplained) contingency, and the philosophical theist is doing the same thing he criticizes the atheist for doing—accepting a purely brute contingent fact as the ultimate explanatory ground for the material universe.

Clearly this won’t do. If the theist accepts a brute contingent fact as an ultimate explanatory ground for the universe, then theism and atheism both end up in an explanatory tie, with each equally embracing brute contingency in the end. It would follow that neither theism nor atheism takes explanation to a deeper level than the other, in which case philosophical theism offers no explanatory advantage over atheism. Two consequences would follow, both fatal to philosophical theism:

(1) If philosophical theism offers no explanatory advance over atheism, then theism offers no good explanatory reason to believe in God.

(2) If theism does not explain more than its rival, then by one swipe of Ockham’s Razor, atheism becomes the more reasonable philosophical view of the two.

15.1.1 Ockham’s Razor

Ockham’s Razor is a widely accepted principle of explanation stating that if two explanations equally account for the very same phenomenon, but one explanation is simpler than the other, then the simpler explanation is to be preferred, where one explanation is “simpler” than another if it posits (hypothesizes) fewer explanatory entities. In other words, if explanation A explains everything that explanation B explains, but in a simpler way, then A is to be preferred (all else equal).[25]

For example, suppose explanation A accounts for certain ancient structures in Central America in terms of ordinary human knowledge and skills existing in Central America at the time the structures were built. Explanation B, by contrast, accounts for the same structures by postulating the same knowledge and skills, but in addition postulates that space aliens from a distant galaxy landed in ancient times and helped build the structures. In this case, explanation A explains everything B explains, but in a simpler way, since it postulates fewer explanatory entities. Ockham’s Razor requires that we give our explanation a “shave” by cutting out the unnecessary explanatory entities—the space aliens—since explanation A explains everything that we need to explain and without reference to them. Since the space aliens are an unnecessary posit, we cut them out of the explanatory picture; we shave our account down to the essentials.

Ockham’s Razor does not apply when explanation A posits more explanatory entities than explanation B but also explains more than B explains; it only applies if two explanations equally explain the same data. For example, suppose the police find 60 shoeprints at a crime scene, all made by size 12 Bruno Magli shoes, all with the same wear patterns, and so on. Explanation A is that two individuals left the shoeprints, with each one leaving 30 of the prints, with both wearing matching shoes, matching wear patterns, and so on. Explanation B is that one single individual left all 60 shoeprints. Ockham’s Razor favors B on the grounds that it explains everything A explains, but in a simpler way. But suppose that half of the shoeprints were made by size 12 shoes and half were made by size 8 shoes. In this case, the first explanation (two individuals left the prints) is preferred, even though it posits more explanatory entities, because it explains more of the data: the “one individual” explanation certainly can’t explain the two sizes of shoeprints. In this case the Razor does not shave the explanation down from A to B, from two individuals to one individual.

Now, if theism and atheism tie, with neither one accounting for more than the other, then the Razor bids us shave God out of the explanatory picture. For in this case philosophical theism postulates one more explanatory entity (God) without making any explanatory gain in so going, without explaining any more. Atheism is then preferred because it explains everything theism explains, but in a simpler way.

If this is the case, then Parsons is certainly right: theism is a logically superfluous hypothesis, an unneeded explanation. God is just an unnecessary explanatory epicycle, a spinning wheel doing no explanatory work, an unnecessary cog in the cosmic clockwork—indeed, as Parsons argues, “a Creator with nothing to do.”

The upshot is that the theist cannot hold that there is no explanation for the divine choice to create. From a rational standpoint, even the divine choice needs an explanation; it cannot be left brute. But now a new problem arises. As we have seen, in scientific and many other explanatory contexts, when we fully and adequately explain some phenomenon P, we typically cite antecedent causal conditions that are logically sufficient for P to occur. In other words, we explain P by citing the sufficient condition that completely caused P to be the case. It is important to keep in mind that a sufficient condition must logically imply that which it would explain. For example, we explain a pot of water boiling by citing (a) the universal law of physics stating that water always boils at 100° centigrade under standard temperature and pressure, and (b) the initial conditions, including the fact that the pot contains pure water, that the standard conditions obtain, and that the pot has been heated to 100° centigrade. Notice that given (a) and (b), the water must boil: it cannot do anything else; the boiling becomes inevitable. In other words, (a) and (b) together imply that the water is boiling.

Now at first glance it seems that only an explanation citing a sufficient cause could fully account for or explain the divine choice. But suppose that there is such an explanation for the divine choice, and call it the logically sufficient condition S. Since a sufficient condition must logically imply that which it would explain, the fact that S obtains must logically imply that the choice gets made. Now the question arises: Is S necessary, or is S contingent? It has to be one or the other; there is no middle ground.

S can’t be necessary, for that would make the divine choice necessary, and we have already seen that the divine choice must be contingent. (Since only the necessary is implied by the necessary, if the divine choice is contingent, it cannot logically follow from that which is necessary. That is, if S were necessary, then the choice to create would also have to be necessary—but we’ve already seen that a necessary divine choice is untenable.)

So the antecedent causal condition S must be contingent. But if S is contingent, then the reviewer’s question arises again: What explains it? The theist’s answer cannot be “nothing,” for then we are back at pure, brute contingency again. And then Ockham’s Razor swings into action, and atheism becomes the most reasonable view of the two.

However, the answer cannot be another contingent condition either, some underlying condition S* that is the sufficient cause of S, for then we are back where we started, with the reviewer asking, “What caused or accounts for S*? For then we are trapped again in an infinite regression, with the same question looping over and over again—with no end in sight.

It is beginning to look as if the reviewer may be right. For it seems that either way the theist turns, whether he says there is no explanation for S, or there is an explanation, theism ultimately loses this explanatory contest, and atheism emerges the simpler and therefore more reasonable view. What’s a philosophical theist to do? Either way, it seems, atheism emerges the more reasonable explanation. This is called a dilemma.

16. A Solution Proposed: The Logic of Personalistic Explanation

But not so fast. There may be a way out of this dilemma for the theist. My proposal requires taking a closer look at two things: (1) the logic of personalistic explanation and (2) the necessary being hypothesis. Let’s start with the logic of personalistic explanation.

In everyday life, people make things we call “choices.” And in the usual circumstances, we suppose that their choices are “freely” made. Thus we naturally speak of things we call “free choices.” (For example: “Don’t blame me for your stomach ache, you freely chose to eat three extra hamburgers.”) Now, in everyday life, when we call someone’s decision a “free choice,” we would normally agree that his choice was not completely caused, necessitated, or totally predetermined by the sum total of all preexisting circumstances, conditions, and forces at work in the world at the time of his choice such that he had to do what he did under the circumstances. That is, we don’t normally suppose that under the circumstances, he could not possibly have done otherwise. For in everyday contexts, if we are told that someone’s action was completely caused, necessitated, or totally predetermined by antecedent circumstances, conditions, or forces, in such a way that the person could not possibly have done otherwise under the circumstances, then we naturally and commonsensically suppose that the person did not act of his own free will; we say that his choice was not a free choice.

For instance, in a restaurant suppose a man knocks over a tray containing six plates of spaghetti and meatballs, splattering sauce all over your brand new white suit. Once you learn that the cause was a massive and unexpected heart attack, you feel sorry for him, or you visit him in the hospital and bring him a card; you don’t blame him or hold him morally responsible for what happened. Even if the man is legally to blame, the man is not morally to blame, for under the circumstances he couldn’t have done otherwise once the heart attack struck. This is the way the words “free” and “choice” are used in everyday, common-sense situations.

But this can be deepened and made more precise. We typically hold that when someone does something of his own free will, it is the case that under the circumstances he chose to do one thing, but at the same time it was physically possible that, under the very same circumstances, he choose to do a different thing instead. Let C be a choice of A over B, made at time T by a freely choosing agent M, and let S be a large conjunction stating all preexistent causes operating, and conditions and circumstances obtaining, at time T affecting agent M. In short, let S be all the causes and conditions in the universe operating on M at the time. From the common-sense standpoint, to claim that M acted freely, and thus could have done otherwise, is to claim that at time T, given the sum total of all prior circumstances and conditions S, the chooser M had the power to choose A, and at the very same time and under the same circumstances, M also had the power to choose B; thus M had both powers at the same time, and it was up to M which power to exercise, which course of action to initiate. This is why some philosophers characterize the capacity to make a free choice as the possession of “dual power.”

16.1 Freedom and (Lack of a) Sufficient Cause

The common-sense account of free choice adumbrated above explains why it is that in explanatory contexts, in the case of a free choice, we do not suppose a free choice has a sufficient causal condition, a sufficient causal explanation stating a sufficient condition for the choice. For remember that a sufficient causal condition is a condition the obtaining of which implies that which it would explain: If S is a sufficient causal condition for P, then once S obtains or exists, then P must occur next; nothing but P can occur next, P becomes inevitable. Thus if we were to suppose that a free choice C has a sufficient cause S, a sufficient explanation accounting for why the free choice was made, then we would have to admit that once the sufficient condition S obtained, the choice C became inevitable; nothing else but C could happen next, the choice C had to happen. This would mean that under the circumstances the person could not have chosen otherwise. And common sense says that such a choice was not freely made. (How could it have been freely made when under the antecedent sufficient condition S, the person could not have chosen otherwise?)

The moral of the story is that a “choice” caused by a sufficient condition would not be a “free” choice as we use the word “free” in ordinary explanatory circumstances. It would simply be another happening strictly predetermined by antecedent circumstances and conditions, including universal laws of motion. It would be like the fall of a rock off a cliff, or the apparent movement of the Sun across the sky—just one more event causally necessitated and rigidly determined by the sum total of all preexisting conditions in the universe. (To say that B is “causally necessitated” or “rigidly determined” by A is to say that once A occurs, B must occur next; nothing but B can occur next, for B was inevitable once A happened.)

Thus the property of having a sufficient causal condition (and thus a sufficient explanation) is conceptually incompatible with the everyday concept of a free choice. This is why common sense has no room for a free choice that possesses a sufficient causal explanation.

This, I submit, is also why the theist is not required to provide a sufficient explanation for the divine choice to create a world. To repeat: In common-sense explanatory contexts, we do not give sufficient explanations for free choices anymore than we give the weight in pounds for a note on the musical scale, or the color of a particular moment in time. To attempt to do either would be a “category mistake.”

16.2 Explaining Free Choices

Does it follow then that a free choice has no explanation? If so, then once again theism is stuck with a purely brute contingent fact, namely an unexplained, brute divine choice at the root of the universe, and for reasons already given, Ockham’s Razor slices away and the atheist wins this explanatory contest.

But it does not follow that a divine choice, or any free choice for that matter, has no explanation. In everyday explanatory contexts, we actually have a perfectly adequate, rationally satisfying manner of explaining the free choices of others, a working model of explanation that we employ successfully everyday to make sense of others’ choices. In order to see this, we need to examine how we actually make sense of free choice in ordinary contexts.

Suppose on a snowy day a father decides to build a sled for his young children so they can play in the snow, and an explanation for his choice (to build the sled) is demanded. In ordinary circumstances, the man’s choice to build the sled would make sense, i.e., it will be adequately explained, once we learn that (a) the father had the power, skills, and knowledge to build the sled (for example, suppose he is a carpenter), and (b) he built it out of love for his children (i.e., he loved his children, he believed they would benefit from the sled, and he wanted to do something good for them). That would normally be all the explanation we would need, wouldn’t it? With that explanation, his choice now makes sense; it has been made intelligible. And once his choice has been made intelligible, explanation is at an end (with respect to the choice itself), and we move on to something else, don’t we?

Notice that part (a) of the explanation gives the powers and capacities of the father, and part (b) gives his reasons for acting. Now, I submit that examination of our ordinary explanatory practices shows that in the typical case, when the powers and capacities of an agent (someone who makes a choice) are sufficient for the effect, and the reasons cited are very good reasons, then the agent’s choice makes sense—it is intelligible, it is adequately explained, and no further explanation is called for. Thus we reach explanatory finality with respect to the choice. Notice that if we had been given illogical reasons for the choice above—for instance, if we had been told that the man built the sled because he wanted to prove that broccoli is not a healthy food—then the explanation would not be rationally satisfying; good reasons are required. But in the above case, we were not given illogical reasons: the explanation cited provided very good reasons for the action. That a father built a sled out of love for his children strikes us right away as a good reason, because we know that fathers normally love their children, and we know that love is by its nature inherently diffusive: it naturally goes out, it shares, it does not want to be bottled up selfishly inside itself. In addition, we know that love tends to seek to enhance the good of the beloved. Thus a loving action by a loving father makes perfectly good sense in itself, which is why no further explanation is needed. (Indeed, it would be hard to think of an action that is more intrinsically intelligible than a purely loving one.) This is why, I believe, the explanation above for the man making the sled reaches explanatory finality.

Further examples would show that we typically account for the free choices of others in terms of the reasons they have for their choices (along with their powers or capacities to implement those reasons), and that when we cite good reasons for a choice, along with adequate powers or capacities, we typically attain a rationally satisfying explanation for the choice, an explanation that makes sense of the choice and ends the questioning (with respect to the choice). Look and see for yourself: This is how we reach explanatory finality with respect to choices; this is how we make choices intelligible.

Now this is important: Unlike scientific explanations, which do cite sufficient conditions, notice that a personalistic explanation—an explanation of a choice—does not cite a causally sufficient condition for the choice. Common sense says that the fact that the father loves his children, that he knows how to make a sled, that he knows that a sled would be good for them, and so on, that does not in itself constitute a sufficient causal condition for his choice to build the sled; these factors do not completely cause him to build a sled, for (at least from the common-sense standpoint) we normally suppose that the father could have had those very same reasons and yet could have chosen not to act on them. Likewise, he could have had those same powers and could have chosen not to put them into effect. This is what we normally mean when we say that under the circumstances, he could have chosen otherwise. And this is why we normally suppose that the reasons and capacities of a chooser are not in themselves sufficient for the choice; we suppose that by themselves they do not guarantee that the choice actually gets made. Indeed, isn’t this why we give the father moral credit for making the choice—because he didn’t have to, that is, under the circumstances, he could have stayed inside to watch TV instead? Again, a personalistic explanation explains a choice not by citing a sufficient condition for the choice, but by making sense of the choice (by making the choice rationally intelligible), and it does this by making sense of the choice in terms of good reasons.

16.3 Problem Solved?

There is no good reason why the explanatory model of personalistic explanation cannot be applied to the (hypothesized) divine choice to create a world. The capacity part of the explanation will be obvious: According to philosophical theism, God (by hypothesis) possesses unlimited power, knowledge, and love. The reason part of the explanation is only slightly less obvious: Philosophical theists propose that the creation of the world was an act of pure love, undertaken to share being (existence or “be-ing”) with creatures, and an act of love is by its nature a free act. In short: God freely created the world out of love. Now, love is a pretty good reason to do something, a reason that is intrinsically intelligible apart from all else. And with that, it seems that we have a rationally satisfying explanation for the divine choice, one that is conclusive in the sense that no further explanation is called for, and we reach explanatory finality (with respect to the choice to create).

Note that the fact that God has sufficient power and also a good reason for creating a world is not a sufficient condition or a sufficient cause for God’s creating a world. For it does not logically follow that God creates a world just because God has a good reason to create a world. Having a good reason to create does not necessitate that God creates for, on the standard theistic assumption that God has “free will,” God could have had a good reason to create a world and yet have chosen not to act on it. In other words, God could have done otherwise.

Applying what I have argued above, the lack of a sufficient causal condition for the divine choice is not a defect in the theist’s explanation. Nothing is wrong with not citing a sufficient condition for the divine choice, for an account in terms of good reasons is not only all that we can ever logically hope for in an explanation of a free choice, it is also normally a perfectly rationally satisfying and explanatorily adequate explanation for a free choice (whether that be a free choice of a divine being, or a free choice of a created being). For it is conceptually wrong-headed to demand a sufficient causal explanation, a sufficient condition, if what we are explaining is really a free choice. Theism succeeds by providing a personalistic explanation here, one that makes the divine choice rationally intelligible. And with this, we have done all we can in the context; explanation has bottomed out. And no fact is left brute.

17. Back to the Necessary Being Hypothesis: A Closer Look

Now let us take a closer look at the necessary being hypothesis, which I shall state as follows:

NB: A necessary being, with sufficient power, knowledge, and other attributes, freely chose to create a universe.

Notice that NB breaks down into two parts:

  1. There exists a necessary being with sufficient power, knowledge, and other attributes requisite to creating a universe.

  2. This being freely chose to create a universe.

Now, notice that the first part of NB is about being, while the second part is about choice. I submit that being is more fundamental ontologically than choice, for the simple reason that a being can exist without making a choice, but a choice cannot exist without being made by a being, in other words without a being making the choice. The idea of a disembodied choice somehow existing on its own without a chooser making it be is simply incoherent. You might say that a choice is to a being as a smile is to a face: just as you cannot have a smile without a face that carries it, but you can have a face without a smile “in” it, you cannot have a choice without a being that makes that choice, but you can have a being existing without a choice being made.

In other words, being ontologically supports or underlies choice in the way that substance underlies attribute in Aristotle’s metaphysics. Aristotle held that an attribute such as redness cannot exist on its own, as an independent entity; it can only exist “in” some substance that is red, in some thing that in a sense “carries” or supports it in existence. On the other hand, the substance that is red, for instance a ripe apple, can exist on its own without the redness in it, as it did before it became ripe. The redness thus depends on the apple for its existence, but not vice versa. Aristotle says that the substance that is red, for instance the ripe apple, has “primary” existence insofar as it carries or supports in existence the attribute, while the attribute exists in only a “secondary” sense “in” the apple.

NB thus presents us with a two-level explanatory structure, with one part explanatorily more fundamental than the other. I shall express the relationship between the parts by saying that the second part—choice—is explanatorily “grounded” in the first part—being. We could also say that the first part is ontologically primary, while the reference of the second part is ontologically secondary. The first level of NB is the existence of a necessary being itself. The second part is a free choice to create, a free act of creation made by a necessary being. The first part is pure unoriginate necessity. But in the second part, insofar as the choice is free, it is, of course, contingent.

At this point an interesting aspect of NB emerges: NB unites the necessary and the contingent, fusing them into one explanatory structure with the contingent part explanatorily grounded in the necessary part. For part one of NB is about necessity with respect to existence, while part two is about contingency with respect to an act of choice. I shall call this unique explanatory structure a “modal fusion explanation”: a theoretical unification of the necessary and the contingent within a single, unified explanation.

This may sound odd at first, since the necessary and the contingent are logical opposites, yet notice that the explanatory link joining the necessary and the contingent inside NB is natural and intuitive, for we all understand, from a first-person perspective, an act of choice. And we all understand what it means to attribute agency (the capacity to make a free choice) to a being. After all, we do this every day when we explain the actions of other people. Indeed, the concept of agency is partly constitutive of our concept of personhood: Our ordinary understanding of ourselves and our fellow human beings makes no sense without it.

So NB fuses together the necessary and the contingent in one explanation, with the contingent part explanatorily grounded in the necessary. I therefore conclude that, contrary to Parsons, on the theistic hypothesis the material universe is ultimately grounded in that which is necessary—not in a mere brute, purely unexplained contingent fact about a choice.

17.1 The Philosophical Advantage of a Necessary Being Hypothesis

In sum, on the basis of the above arguments, the theist can claim all of the following:

  1. Why does a contingent, material universe exist? Philosophical theism provides us with a complete, logically sufficient cause, that is, an explanation stating a logically sufficient causal condition for the existence of the whole thing, while atheism (as Parsons himself admits) leaves the existence of the material universe a completely unexplained or brute contingent fact.[26]

  2. Upon close examination, the theistic explanation is cleared of the charge that it leaves pure brute facts in its wake.

  3. Philosophical theism traces explanation back and back until it finally grounds its account of the material universe in an explanatorily ultimate basis of existence, an explanatory ground that is itself logically necessary and explanatorily self-subsuming (or self-explanatory) with respect to its own existence. And from the beginning of philosophy, the logically necessary (unlike the contingent) has been recognized as a rationally satisfying stopping point for explanatory regress.

  4. In short, of the two views under consideration in this paper, philosophical theism alone makes sense of the existence of the material universe.

18. The Argument Continues: Theism vs. Grand Unified Theory

Parsons takes us close to the heart of the issue when he directly compares the theistic hypothesis with the prospect of a well-confirmed, physics-based “theory of everything”: a completed, grand unified physics that accounts for all known physical particles, forces, and properties of matter. Parsons asks: Why would anyone find a completed, grand unified physics intellectually unsatisfying? Why, in the face of a well-confirmed scientific theory of everything, would anyone want anything more? Why would someone want to keep asking “But why does it all exist?” until they reach a theistic explanation? Why not just end our inquiries with the most basic features of the physical universe, whatever they may turn out to be, as revealed by a completed physics, and go onto more interesting problems?

As we have seen, Varghese has already given an answer to these questions, and I have defended his answer. Varghese argued that even in the face of a completed, grand unified physics of an eternal universe, the question remains: Why does an eternal universe exist? But here Parsons is interested in another argument Varghese gives for the incompleteness of science, and for the explanatory advantage offered by philosophical theism. Varghese claims that, by its nature, no physical (or material) entity or collection of physical entities could ever be self-explanatorily ultimate, while God or a necessary being, as hypothesized, would be:

The only viable explanation, then, for the existence of any one of the entities or all of the entities that make up the universe would be the existence of an ultimate uncaused being—a being that did not receive existence from anyone or anything else and can completely explain its own existence. This self-explanatory being is commonly called “God” and is the explanatory ultimate demanded by all non-self-explanatory entities from sub-atomic particles to galaxies. Cosmological arguments do not reason from the fact that everything in the universe has a cause in space and time to the conclusion that the universe has a cause in space and time; these arguments point out, rather, that everything in the universe is non-self-explanatory, which means that the explanation of the universe does not lie in itself but must lie in a self-explanatory being. (p. 21)

Parsons asks: But exactly what does Varghese mean by self-explanatory here? Does he merely mean uncaused, where an uncaused being is one that simply has not received existence from anything else? Or would he say that being uncaused is logically sufficient for, but not identical with, being self-explanatory? If either of these accounts is what Varghese means by self-explanatory, then the atheist has every right to reply that the most basic state of matter, or the most fundamental state of the physical universe as presented by our best physics, is uncaused in this sense and therefore is the self-explanatory being. The problem for Varghese is that it seems that he does not rule out either possibility, for he never says why the most basic state of the physical universe itself could not be self-explanatory in this sense.

So even if we insist on pushing the cosmological explanatory regress until we find a self-explanatory being, the atheist can argue that the quantum vacuum, or the ultimate space-time foam, or whatever is considered the most fundamental state of the physical cosmos, is a legitimate stopping point by Varghese’s own “stopping point” criterion. And Parsons’ claim would stand: theism is an unnecessary explanatory epicycle—unless Varghese is claiming something else here.

Is Varghese claiming that there is something more to being self-explanatory than merely being uncaused, some further property or condition that constitutes self-explanation? That this is indeed what Varghese is claiming is evident, I think, from the presence of the word “and” in the definition:

[A]n ultimate uncaused being—a being that did not receive existence from anyone or anything else and can completely explain its own existence (p. 16, my emphasis).

But how does something “explain its own existence”? Varghese does not say; it is merely an explanatory posit in his paper. (My exploration of the concept of necessary being above can be seen as my contribution toward filling in this gap in Varghese’s argument.) However, Varghese later suggests a further condition on explanatory ultimacy (p. 16), which I shall render this way:

X is self-explanatory if X is that on which all else depends and if X is such that X could not depend on anything else.

This is an improvement, for self-explanation now has a modal dimension: a self-explanatory being not only does not have an external explanation, it could not have one and it could not have had one. However, Parsons’ reply overlooks this crucial modal distinction and makes two inferences that I believe are logically invalid. He makes these inferences when he asserts that this—the property of being self-explanatory and thus not in need of further explanation—would be true of anything that is the apex of all explanations, including the most basic states of the material universe if they are explanatorily ultimate (p. 11).

18.1 Analysis of This: The First Inference

Parsons claims that the “most basic states of the material universe” could turn out to be self-explanatory in the sense specified by Varghese—namely, not only do they not have an underlying explanation, but they could not have one. But this seems wrong; it is simply not logically possible that something contingent could be such that (a) it has no underlying explanation and (b) it could not have had one. For it is an a priori truth that, for any contingent entity, there is in principle some logically possible state of affairs that could have been the cause or ground of its existence, even if no such ground or cause has been discovered. In other words, it is necessarily true that for any contingent entity X, there is a possible world W in which X has a cause in W. Therefore, the most basic states of the material universe, being contingent, could not be such that (as Parsons claims) they logically “could not depend on anything else.” Contrary to Parsons’ claim, the most basic physical states of the universe therefore could not be self-explanatory in the sense Varghese has specified, namely that X is self explanatory if (a) X is that on which all else depends and (b) X is such that X could not depend on anything else.

On the other hand, a metaphysically necessary being, if one existed, would be self-explanatory in the sense specified by Varghese, for by hypothesis, such a being could not even possibly have, or have had, an external explanation or external ground of its existence. (If a being even possibly has an external cause or ground or explanation of its existence, it would not be a necessary being, but a contingent one.) This is a pure conceptual truth. In short, the “could not possibly have an external cause” clause cannot properly be attributed to any logically contingent being, although it logically can be said of a necessary being.

Thus, contrary to Parsons’ claim, Varghese has defined self-explanatory in such a way that it cannot be said of the most fundamental state of the material universe as a whole, while it logically can be said of a necessary being posited to exist apart from the physical universe and as its underlying cause.

18.2 The Second Inference

Parsons next takes up Varghese’s suggestion that being self-explanatory entails that it would be logically absurd to ask for an explanation. Says Varghese: “[I]t is as absurd to ask for an explanation for the existence of a self-explanatory being as it is to ask ‘Why is a circle round?'” Thus Varghese gives us another elaboration on the concept of self-explanation:

X is self-explanatory if it would be logically absurd to ask for an explanation of X. Asking why X exists, if X is self-explanatory, would be like asking, “Why is a circle round?”

Parsons replies that (a) the most basic states of the physical universe could also be called self-explanatory in this sense, for (b) if physics were to reach a grand unified theory of everything, it would be logically absurd to seek a deeper explanation of the physical universe.

But this inference too seems to be mistaken, for Parsons has admitted that the universe is a contingent being. However, given that its existence is contingent, there is always at least the logical possibility that it has an external cause, even if one has not been found. And if there is at least the logical possibility of a cause, then it will always make sense to ask, “What is the cause?” And if it makes sense, then it is not logically absurd to seek the cause. Therefore, even if we do reach a well-confirmed grand unified theory of everything in physics, there can be nothing logically absurd, nothing contradictory or inconsistent, in seeking the cause of the most fundamental state of the physical universe (as revealed by a completed physics). I conclude that Parsons is wrong when he supposes that the causal question would be logically absurd when applied to the most fundamental state of matter as presented by a completed physical theory.

On the other hand, the same cannot be said of a metaphysically necessary being, for it would indeed be logically absurd to seek an external or more fundamental explanation for a necessary being, for reasons we have already seen. This is a conceptual truth. I conclude, contrary to Parsons, that “self-explanatory”—as specified by Varghese—could not accurately characterize the most fundamental states of the material universe, while it logically could be said of the existence of a necessary being.

18.3 Is Theism Explanatory at All?

Parsons wonders: How does being self-explanatory differ from simply being inexplicable? He argues that in the absence of clarification—in the absence of any further positive content—it is not clear that there is any difference, and thus it is not clear that there is any content to the concept of self-explanatory being (as employed by Varghese). If so, Parsons asks, isn’t Varghese simply positing another inexplicable brute fact as the cause of the universe, a brute fact as the bottom level of explanation at the base of existence, when he posits God’s necessary existence? And wouldn’t this be a brute fact of just the sort that Varghese claims he wants to avoid—the type of brute fact Varghese accuses the atheist of settling for?

Parsons would have a good point here if he had successfully deflated the concept of self-explanatory being. But he has not. For as I just argued, Varghese has given enough content to the concept of a self-explanatory being to show that no material being or state of a contingent, physical universe could be ultimately self-explanatory. This implies that if there is a self-explanatory being, it is not a material being. In addition, in my examination of necessary being and self-subsumptive explanation above, I provided further content to the idea of a self-explanatory being.

This is why I cannot follow Parsons when he claims that Varghese leaves us with a God that amounts not so much to a self-explanatory being as to an inexplicable being, one that is “precisely the sort of ultimate brute fact that [theists] decry.”

19. A Further Worry

But a worry remains: Is theism, offered as an ultimate explanation for the existence of a contingent universe, a real explanation? Or is it only a faux explanation masquerading as a real one?

Well, the necessary being hypothesis is certainly not a scientific explanation that deduces the explanandum from a statement of universal laws of nature and a statement specifying initial conditions. But not all legitimate explanations are scientific or “covering law” explanations. We employ many different kinds of explanation as we go about making sense of our world. For example, most explanations employed in everyday life are not scientific ones, and most explanations in academic fields such as history, sociology, and economics are not strictly speaking scientific explanations. The explanations that we use to explain the actions of other people are usually personalistic explanations, not scientific ones. In order to show that theism is not a real explanation, Parsons would need to examine all known types of explanation and then show that philosophical theism fits none of these categories. That would go some way toward showing that philosophical theism is not a real explanation, but even that would not be enough: he would also have to show that no other types of legitimate explanation exist.

19.1 Five Models of Explanation

The problem with this approach is that our philosophical understanding of explanation itself is still at an incomplete level, at least when compared to other areas of philosophy. We lack both an adequate descriptive account of what explanations are (the description problem), and an adequate account of how explanations are justified (the justification problem). Showing that philosophical theism fits no known category of explanation, and that no other categories of explanation (known or unknown) exist, would require doing more work on the description problem, certainly more than Parsons does in his paper. At least five models of explanation exist[27] in the philosophical literature:

  1. On the familiarity model, that which needs explanation is something unfamiliar, and it is explained by relating it (via a familiar relationship) to something already familiar (thus making it familiar). For example, a physicist explains the flow of electricity through a copper wire by saying that it is similar to water flowing through a garden hose.

  2. On the deductive-nomological or “covering law” model, we explain something by showing that it can be logically deduced from a universal law of nature and a specific statement of initial conditions. For example, the curved path of a cannonball is explained by deducing it from a statement of the ball’s initial position inside the cannon, the forces acting on it, and an application of the relevant laws of physics.

  3. On the unification model, we explain something by showing how it is really a part of an explanatorily unified whole.

  4. On the necessity model, a phenomenon is explained when we see why it had to occur or had to be—why its occurrence or existence was necessary.

  5. Personalistic explanation (also called “intentionalistic explanation”) occurs when we explain something by saying that a person freely brought it about for a reason. In this type of explanation, the explanans refers to the powers, knowledge, and purposes or reasons of a rational agent, including the power to freely bring something about. Thus we might explain a freshly mown lawn by referring to the homeowner’s power to mow it, his knowledge of how to mow it, his desire that the lawn be mowed before company comes over for dinner, and so on.

In everyday life, a personalistic explanation is generally an intellectually satisfying explanation when (a) the person cited exists, (b) the person possesses the powers, beliefs, and desires cited, and (c) the reasons cited are good reasons. For example, suppose an artist’s studio contains a beautiful painting of a rose. That the artist painted the work because he wanted to give his arthritic fingers some exercise is not a very satisfying explanation. That he painted it as a gift for his wife on their 50th anniversary, as an act of love because the rose is her favorite flower, is a more satisfying explanation; and if the reason fits his character, since it is a very good reason, no further explanation is called for.

The philosophical theist can be seen as offering a necessity explanation in response to the contrastive question, “Why does God exist, rather than not exist?” In addition, the theist offers a personalistic explanation for the divine choice to create, and a sufficient explanation (providing a logically sufficient condition) for the existence of the contingent universe as a whole. The philosophical theist argues that theism makes explanatory progress by taking our explanatory account of the material universe one level deeper compared to atheism, for it takes us beyond the seemingly brute contingent existence of the material universe to a metaphysical ground of existence that accounts both for the existence of the material universe and its own existence in one fell swoop. In every other area of philosophy, explanatory progress is a good reason to accept a hypothesis. Says the theist: Why not here too? As Alexander Pruss observes in his fine book on the principle of sufficient reason, “Is not all philosophical research itself a quest for explanation?” And to this I would add, “ever deeper explanation.”[28]

I am suggesting in this paper that theism is philosophically attractive for the same reason that any philosophical theory in any area is attractive: It makes explanatory progress by pushing our explanations to a deeper level. Parsons disagrees, of course, but I see no argument in his paper that shows that philosophical theism does not provide a legitimate necessity explanation for the first question (“Why does God exist?”), and I am aware of no good argument for the claim that philosophical theism fails to give us a legitimate personalistic explanation for the divine choice to create.

20. One Last Time: In Defense of Theistic Explanation

Parsons and Varghese agree on the crucial point that the material universe exists contingently. But on the daring inquiry principle, it is reasonable to seek the cause or ground of the material universe’s existence since it exists contingently. Since the universe is contingent, if it does have a ground, sufficient explanation, or cause of its existence, that ground must lie in something external to it, in something more fundamental—in something noncontingent. For we have seen that if all things are contingent, then there is no explanation for why there are contingent beings. In other words, the universe does not exist of its own nature; it does not exist “on its own steam” so to speak. (If it did, it would be a necessary being, not a contingent one, for it would then exist no matter what the circumstances, existing in all possible worlds, not just in some.)

Of course, Parsons would agree, because he writes:

Given that we do someday acquire the fabled Theory of Everything (TOE), and are rationally assured that it is indeed the TOE for which we have so long sought, our quest for a scientific understanding of the physical cosmos will have reached a satisfactory terminus. As Mackie observed earlier, in every scientific explanation, the explanans are temporarily regarded as brutely factual, i.e., as simply given. We may proceed to inquire about these, but only by regarding their explanations as in turn unexplained….

There is no reason to think that in thus pursuing ever-deeper explanations, that we might not someday reach explanatory “rock bottom,” i.e., those physical facts that undergird all other explanations, and which themselves admit of no deeper account or explanation. In fact, reaching such “rock bottom” seems to be the goal of fundamental physics…. Of course, the “rock bottom” itself would remain unexplained (my emphasis)

So Parsons and Varghese agree here: If matter is indeed metaphysically ultimate, then the “rock-bottom” level is unexplained and thus brute. This is precisely why a philosophical theist such as Varghese looks for a deeper explanation. And in doing so, the philosophical theist is simply doing what philosophers always do: seeking a deeper explanation when something stands in need of further explanation.

This is where the concept of a metaphysically necessary ground of being as the ultimate cause of the material universe enters the picture. For the reasons I have already given, the theistic hypothesis seems to be just what the doctor ordered; it appears to fit the bill perfectly. What else could possibly explain the existence of contingent beings? And the hypothesis comes with an unexpected bonus: a necessary being would be a self-explanatory ground of being, thus a “rock-bottom” level that is not itself brute.

It is worth observing that this theistic solution to the ultimate existential problem is in harmony with a major current of thought in the Western philosophical tradition, one going back to the first generation of philosophers. The story of these revolutionary thinkers, who first appeared in ancient Greece six centuries before the birth of Christ, is worth at least a summary retelling.

Ancient Greek philosophers were the first thinkers in history to reject myth and unquestioned authority as sources of answers to the fundamental questions of life, and to seek in their place independent rational explanations—explanations based on independent, reasoned argument and observable evidence.

Starting with Thales of Miletus (625-546 B.C.), the ancient Greek philosophers sought the arche of all things, by which they meant the ultimate ground of all being, the foundational principle of all existence. Why a single starting point? The argument (still accepted in physics today) was that (a) the cosmos displays a remarkable overall unity; (b) this unity suggests it probably originated from one ultimate source or arche; thus (c) we can expect a single first principle of explanation, a unified, fundamental explanation that singly accounts for the cosmos overall. This would be the arche of the cosmos. (Physicist Leon Lederman once joked that physicists seek a final theory of everything that is so compact it can be written as a single equation on a t-shirt.)

However, when an arche is posited, it is natural to ask: “But what caused the arche? It is supposed to explain the cosmos, but what explains its existence?” The traditional philosophical response was that the arche, if it is to be a reasonable ultimate explanation, cannot be something that also stands in need of a prior cause or more fundamental explanation. For if the arche were to need an external cause or an underlying explanation, then it would not be explanatorily ultimate, and reference to it wouldn’t solve the problem of existence. But what category of being would thus be explanatorily ultimate? For reasons I have given above, reflection on necessity and contingency long ago suggested the idea of a metaphysically necessary being, understood as one whose intrinsic nature is such that its nonexistence would be absolutely impossible. Thus, L. P. Gerson, a scholar of classical Greek thought, observes that the ancient Greek philosophers generally concluded that God must be a necessary being because:

That which explains the existence of things can’t be something whose existence also needs explaining [in terms of something deeper than itself]. So, all the major Greek philosophers conceived of God as a being in some sense necessary—not just another thing that might not have existed—-hence not in itself in need of further explanation.[29]

Since a necessary being, by hypothesis, would contain the ground or explanation of its own existence within its own essence, its existence would not stand in need of further explanation. And, says the advocate of the “necessary being” hypothesis, only such a being could provide a rationally satisfactory, self-subsuming terminus to the regress of questions of existence.

This may explain why the following dialectic occurs throughout history in most great systems of philosophical thought. One thing is explained in terms of a deeper thing, or things at one level are explained in terms of things at a deeper level. The starting level is notable for its stark contingency. (Indeed, the contingency is what raises the philosophical questions in the first place.) Next, things at this level are explained in terms of an even more fundamental level. The explanatory process only bottoms out when the explanations get down to a ground of necessity: here thought naturally hits “rock-bottom.” For as I have argued above, once we reach necessity, we see that no deeper explanation is possible nor needed.

What does the theist conclude? That philosophical inquiry reaches a natural and intellectually satisfying end only when it reaches the rock-bottom explanatory level of necessary being.

21. Concluding Thoughts

Parsons concludes:

Varghese and his allies … have given us no reason for thinking that the universe must have been caused by God, or that the original and most basic features of the universe cannot be an uncaused reality that explains everything else but is itself unexplained….

It appears, therefore, that he choice between theism and atheism boils down to a choice between ultimate brute facts—God or the most basic features, whatever they might be, of the physical cosmos. I have argued in this paper that the choice of God is not intellectually compelling, that it is entirely rational and reasonable to see the universe as needing no cause and capable of existing with no metaphysical support. The universe, insofar as it is explicable, is explicable in naturalistic terms. No logical or metaphysical principle, at least none mentioned by Varghese, requires that the world have a creator.

In short, says Parsons, we have no rational need to suppose that God exists. But Parsons’ argument contains undefended gaps at a number of crucial points, as I have argued in this paper. On the other hand, Varghese’s argument, when supplemented by the additional considerations suggested in this paper, makes a good case that the choice between theism and atheism does not simply boil down to a choice between one ultimate brute fact or another. Rather, it comes down to a choice between one view that makes rational sense of the existence of the material universe, and an opposing view that, bottom line, does not. This is a choice between theism on the one hand, and Parsons’ atheism and scientific naturalism on the other, respectively.

This all adds up to a perfectly good explanatory reason to embrace theism. After all, what better reason is there to hold a belief, than that it is the best explanation for something we are trying to make sense of (in this case, the existence of the whole of what there is)?

First and foremost, philosophy is in the explanation business. Her customers are human beings asking the fundamental questions of existence. Among the most fundamental questions that can be asked are the two contrastive questions treated in this paper: “Why is there something (rather than nothing)?” and “Why this universe (rather than some other)?” Philosophical theism offers a reasonable explanation, arguably the only one that seems to deliver the goods. When he says that there is no answer to the most fundamental question of all, Parsons turns away the last vendor before the business day has ended. That’s not good business practice, and not good practice in philosophy, either. Says the philosophical theist: Why stop short of the only rational explanation that ultimately makes sense of it all?[30]


Endnotes

[1] For an online summary of Peirce’s philosophy, I recommend the Stanford Encyclopedia of Philosophy entry on Charles Sanders Peirce at <http://plato.stanford.edu/entries/peirce/>

For a serious treatment of explanatory reasoning, I recommend Peter Lipton, Inference to the Best Explanation, 2nd edition (New York: Routledge, 2004.)

[2] Although I shall not enter this issue here, Gilbert Harman has argued, correctly in my view, that all forms of inductive reasoning are reducible to (can be understood as special cases of) explanatory reasoning. See his “The Inference to the Best Explanation,” Philosophical Review 74 (1965): 88-95. In passing I shall raise an interesting question here for those who want to dig more deeply. Explanatory reasoning moves to its conclusion on the grounds that the conclusion, if true, would be the best explanation of the facts. But why should the fact that a potential explanation would best explain some set of facts be a reason to believe that the explanation is true, that is, that it corresponds to an independently existing reality? Why suppose that reality tends to give up its secrets to human explanatory reasoning? In short, why suppose that explanation is truth-tropic? Scientists as scientists simply assume that explanatory reasoning tends to take us to real truth, to correspondence with an independently existing reality. But philosophers ask: What is the epistemic justification for believing that explanatory reasoning takes us to real truth—in other words, that it is reliable? Is there a way to justify explanatory reasoning that would be both noncircular and purely empirical? Notice that it would be logically circular or question-begging, in at least two different ways, to reply: The best explanation of our successful explanatory practices in science is that explanatory reasoning is truth-tropic.

The following imaginary dialogue is designed to suggest some of the pitfalls that await anyone who seeks to produce a noncircular defense of explanatory inference:

Person A: I believe that explanatory reasoning is reliable, that is, I believe that by its nature it tends to take us toward truth. And by truth I mean correspondence with reality. Therefore, if something is the best explanation of the facts, that is a good reason to think it really is true, i.e., that it corresponds with reality.

Person B: But how do you know that explanatory reasoning is reliable? On what basis do you make that claim? How do you logically justify it? How do you rule out the possibility that explanatory reasoning has generally given us nothing but illusion and falsehood, not reality?

Person A: Well, here is my argument:

  1. Our explanatory practices have generally taken us to truth in the past.

  2. The best explanation of this success is that explanatory reasoning is reliable, that is, it inherently tends to take us to truth.

  3. Therefore, explanatory reasoning is reliable, by its nature it tends toward truth.

Person B: Wait a minute! Doesn’t your first premise assume what you are supposed to be proving—that explanatory reasoning is reliable? How do you know that our explanatory practices have generally taken us to truth in the past? How do you justify that premise without assuming what you are trying to show, that explanatory reasoning is reliable?

Person A: I see, yes… I can’t use that premise to establish the reliability of explanatory reasoning without begging the question at issue.

Person B: Furthermore, even if we grant you premise 1 of your argument, notice that in the next premise you employ explanatory reasoning to argue for the legitimacy of… explanatory reasoning! Isn’t that also circular? Isn’t that like using P to prove P? It is like saying: “I trust Joe because Ed says that Joe is trustworthy, and I trust Ed because Joe says that Ed is trustworthy.” Your argument seems to be doubly question-begging.

Person A: Hmmm… I see.

Person B: Your argument reminds me of an argument that is said to have been given by the Reverend Billy Graham:

  1. I believe in God because the Bible says that God exists.

  2. But why do you believe that the Bible is true?

  3. Because God inspired it.

Person A: Yes, I see the circularity there. Can I go back to my first premise? You asked me how I know that explanatory reasoning has generally taken us to the truth in the past. Here is how I would justify my claim. What other explanation is there for the fact that explanatory reasoning has allowed us to make so many successful predictions in science as well as in everyday life? And what other explanation is there for the fact that it has given us so many useful theories? Theories that help us build bridges and jet airplanes and computers and skyscrapers? It must be that explanatory reasoning has been successfully taking us to truth, for there is no better explanation!

Person B: There you go again, using explanatory reasoning to justify explanatory reasoning! Do you see the circle that you are in?

[3] See Henry Margenau and Roy A. Varghese, Cosmos, Bios, and Theos (Chicago, IL: Open Court, 1992).

[4] In connection with this, here is a short but interesting argument to ponder: Complete, absolute, uncompromising nothingness would entail the absence of even the possibility that something begins to exist, since a possibility is a something, just as it would entail the absence of even the potential for something coming into existence (since a potential is also something, not nothing). Given this, wouldn’t it follow necessarily that it is a priori logically impossible that something comes into existence out of pure, absolute, uncompromising nothingness? (And wouldn’t this entail that something or other has always existed since something exists now?) Perhaps the medievals were right: ex nihilo, nihilo fits (“from nothing, nothing comes”).

[5] For elaboration we can turn to the subject of cosmic evolution. In The Little Book of the Big Bang, University of Washington astronomer Craig Hogan lists the cosmological eras the material universe has passed through in its evolution to the present state, each explained within scientific cosmology. Hogan’s list includes:

The Planck epoch. The Cosmic Inflation epoch. The Creation of Light epoch. The Creation of Matter epoch. The Electroweak epoch. The Strong epoch. The Weak Decoupling epoch. The Nucleosynthesis epoch. The Spectral decoupling epoch. The Matter/radiation equality epoch. The Matter/radiation decoupling epoch.

Thus, just as geologists trace the history of the earth’s crust back in time, cosmologists trace the history of the whole cosmos step by step back in time. But it is instructive to notice that at each point in the story, cosmologists only explain how already existing matter changes; at no point is a beginning out of absolutely nothing given an explanation. See Craig J. Hogan, The Little Book of the Big Bang (New York, NY: Copernicus, 1998), pp. 22-23.

[6] Stephen Hawking, A Brief History of Time (New York, NY: Bantam Books, 1988), p. 174.

[7] Penrose in quoted in Alan Lightman and Roberta Brawer, eds., Origins: The Lives and Worlds of Modern Cosmologists (Cambridge, MA: Harvard University Press, 1990), p. 433.

[8] Someone in an even more metaphysical mood might also wonder about another question. Abstract mathematics is capable of describing the concrete, physical universe. Models of the universe consist of an immensely complicated mathematical structure expressed in differential equations. What accounts for the (abstract) existence of that mathematical structure? That a material universe exists is philosophically puzzling enough; that there should be an awesome mathematical structure describing it from above, so to speak, should be even more wondrous, at least to the philosophically inclined. And the whole ensemble—abstract and concrete combined—seems even more wondrous if one is aware of the enormous epistemological challenges presented by mathematical knowledge: That the grand structure of the physical universe has been found to conform to independent mathematical principles known purely a priori by abstract human reason (and thus epistemically justified on grounds independent of sense experience of the world) is one of the most intriguing issues yet unsolved within the philosophy of science. This was the famous question raised by the physicist Eugene Wigner in his now classic 1960 article, “The Unreasonable Effectiveness of Mathematics in the Natural Sciences.” See the Wikipedia entry on The Unreasonable Effectiveness of Mathematics in the Natural Sciences at <http://en.wikipedia.org/wiki/ The_Unreasonable_Effectiveness_of_Mathematics_in_the_Natural_Sciences>

The online encyclopedia quotes Wigner:

The miracle of the appropriateness of the language of mathematics for the formulation of the laws of physics is a wonderful gift which we neither understand nor deserve. We should be grateful for it and hope that it will remain valid in future research and that it will extend, for better or for worse, to our pleasure, even though perhaps also to our bafflement, to wide branches of learning.

[9] David Hume, Dialogues Concerning Natural Religion (New York, NY: Library of Liberal Arts, 1947.)

[10] William J. Wainwright, Philosophy of Religion, 2nd ed. (Belmont, CA: Wadsworth Publishing, 1998.)

[11] Laurence BonJour, Epistemology: Classic Problems and Contemporary Responses (New York, NY: Rowman & Littlefield Publishers, 2002), p. 196.

[12] Sometimes it helps to hear an abstract idea explained in more than one way, so here is another way to put the idea. The problem with an infinite regress of explanations, all the members of which are explanations of contingent things, is that: (a) each being in the regress, being logically contingent, receives existence from something prior in the series; and (b) each being in the regress transmits existence to something following it, but nothing in the regress is a genuine initiator of existence (where an initiator transmits existence to something without first needing existence transmitted or given to it). Existence or being would be intrinsic to an initiator of existence, a part of its internal make-up, so to speak, not just a gift from an external source of being. But if everything that exists is only a receiver and transmitter of existence, and nothing is a real initiator, then it would seem that each and every explanation of existence is explanatorily inconclusive in the same way that “The new battery is dead” is an inconclusive explanation of the fact that the car won’t start. So within such a regression, no conclusive explanation of existence will ever be found.

[13] In philosophy, an infinite regression that does not resolve the question at issue is called a “vicious” infinite regression; one that does is called a “benign” infinite regression.

[14] I want to thank an anonymous reviewer from Internet Infidels for a good criticism of an earlier version of this infinite regression argument; the reviewer’s excellent objection caused me to improve my argument considerably.

[15] In one form or another, this principle was employed implicitly in all the great systems of philosophy up until Leibniz, functioning in those systems somewhat like a computer program running in the background within a larger program. It is an interesting historical fact that it was never named and given explicit philosophical expression until Leibniz recognized it in his work in the late seventeenth century.

[16] Mackie’s original argument can be found in J. L. Mackie, The Miracle of Theism: Arguments for and Against the Existence of God (New York, NY: Oxford University Press, 1983).

[17] One might wonder why every scientific explanation must include an explanans considered brute for the purpose of the explanation. Basically, because scientific explanations are always internal to the universe as a whole; that is, science always explains one thing A within the universe in terms of another thing B also within the universe (with B the cause of A). But doesn’t this imply that science alone will never answer the fundamental question, “Why does the material existence as a whole exist?”

[18] The distinction between the necessary and the contingent is absolutely central to logical theory. The most fundamental logical ideas—logical inference, valid argument, implication, and so on—cannot even be stated, let alone precisely defined, without employing the necessary-contingent distinction. (For instance, in the case of a valid argument, we say that the conclusion follows “necessarily” from the premises, that P implies Q if and only if it is not even possible that P is true but Q is false, and so on.)

Incidentally, the first introductory logic textbook to place the elementary modal distinctions front and center, and to emphasize throughout the text their fundamentality to logical theory, was Possible Worlds by Ray Bradley and Norman Swartz (Hackett Publishing, 1979). I used this wonderful text early in my career, and still return to it from time to time to appreciate the theoretical beauty of its presentation. I recommend this text to anyone who wants to explore necessity and contingency in more depth.

Contingency and necessity are the focus of the specialized branch of logical theory known as modal logic, so named because contingency and necessity are sometimes called “modes” of truth and falsity. Modal logic was founded in the fourth century B.C. by the Greek philosopher Aristotle (322-384 B.C.), the founder of logical theory itself. (Aristotle qualifies as the founder of logical theory because he was the first individual in world history to write an actual treatise on the subject; indeed, he wrote five.) For an important contemporary defense of the necessary-contingent distinction and its application to the question of the existence of God, see Alvin Plantinga, The Nature of Necessity (Oxford, UK: The Clarendon Press, 1974.)

[19] If a necessary being is absolutely unconditioned, what is its relation to time? Philosophers have debated two different theories. According to one view, a necessary being would exist outside of time altogether, unaffected by it and thus not subject to the “flow” of time. According to a second theory, necessary existence implies existing at all times, past, present and future, in an unending existence within the flow of time.

[20] See Steven M. Duncan, The Proof of the External World: Cartesian Theism and the Possibility of Knowledge (Eugene, OR: Wipf and Stock Publishers, 2008.)

[21] John-Paul Sartre, Nausea (New York, NY: New Directions Publishing, 1969), p. 188.

[22] Mixing up a sufficient condition with a necessary one has been a widespread source of confusion:

Condition S is a sufficient condition for a phenomenon P if and only if it is the case that if S obtains then P must obtain; S is all that is needed in order that P obtains, so that once S obtains P is guaranteed.

Condition N is a necessary condition for some phenomenon P if and only if it is the case that in the absence of N, P will not obtain; N is a requirement for P, that is, P cannot obtain without N.

For example, the presence of oxygen is a necessary condition for human life (since in the absence of oxygen human life is not possible), but it is not by itself sufficient for human life (it is not enough all by itself to do the job). The presence of oxygen is thus a necessary condition for human life but not a sufficient one. On the other hand, jumping in the Atlantic Ocean is a sufficient condition for becoming wet, since it is all that is needed to get the job done, but it is certainly not necessary for getting wet (since there are other ways to get wet).

Many philosophers hold that an explanation for some phenomenon P is not adequate or complete (and thus is not rationally satisfying) unless it provides a sufficient causal condition or sufficient reason for the existence of P. In order to qualify as sufficient, a reason or causal condition must meet a very high standard: it must logically imply that which it would explain. (In logic, to say that proposition A “logically implies” proposition B is to say that if A is true then B must be true, B cannot possibly be false given that A is true, or it is impossible that A is true and yet B is false.) Another way to put this would be: The presence of a sufficient condition must in itself guarantee the presence of the phenomenon it explains.

For example, suppose we want to understand why a loaf of bread has become moldy, and suppose someone offers this explanation: The bread is one month old. Notice that this would not count as a sufficient reason or sufficient condition for the moldy bread, for not all month-old bread is moldy. (Bread stored in a freezer for a month will not get moldy.) Being a month old may be a necessary condition for being moldy, i.e., it may be one requirement among others, but it is not a sufficient condition, i.e., it is not enough by itself to get the job done. A sufficient reason for the moldy bread would have to cite those conditions a, b, c…n such that the bread must be moldy once they obtain; that is, the bread cannot do anything else but get moldy once they obtain.

[23] See Peter van Inwagen, Metaphysics (Boulder, CO: Westview Press, 1992), p. 98.

[24] Peter van Inwagen, Metaphysics, p. 98.

[25] Named after the great English philosopher and logician William of Ockham (1288-1348), a Franciscan monk who made original contributions to many areas of philosophy, this principle only applies when two explanations tie, that is, when both equally account for the same phenomenon. It should be noted that there is a debate in philosophy over this question: Is the Razor truth-tropic? That is, if we follow it, are we likely to get closer to the truth? What justifies Ockham’s Razor?

[26] Recall that a sufficient condition logically implies that which it would explain, and a sufficient explanation for some phenomenon P is a complete explanation in the sense that it states a logically sufficient condition for P.

[27] I am following Peter Lipton here. See chapter 2 (“Explanation”) of his Inference to the Best Explanation.

[28] See Alexander Pruss, The Principle of Sufficient Reason: A Reassessment (Cambridge, UK: Cambridge University Press, 2006), p. 16.

[29] See L. P. Gerson, God and Greek Philosophy: Studies in the Early History of Natural Theology (London, UK: Routledge, 1990), p. 14.

[30] In addition to the anonymous reviewer for Internet Infidels, I wish to thank Steven Duncan, Arthur DiQuattro, Norm Lund, Richard Kopcinski, and Samuel Sundquist for helpful comments on an earlier draft of this paper. I also want to thank Robert Coburn and Peter van Inwagen for personal correspondence concerning a previous paper of mine published on the Secular Web (“Contra Carrier: Why Theism is Needed to Explain Everything,” June 2006); their comments helped me think through some of the issues contained in this paper.


Copyright ©2009 by Paul Herrick. This electronic version is copyright ©2009 by Internet Infidels, Inc. with the written permission of Paul Herrick. All rights reserved.

all rights reserved