The Dawkins Proof

Can We Prove That God Doesn’t Exist?

Let’s forget that story. The whole point of

science is to avoid such stories.

Leonard Susskind

I happen to believe that my ontological disproof is as good as any, but let’s set that aside for the time being. It is much to be commended that after, shall we say, centuries of remaining on the defensive (an attitude motivated partly by the genuine civility of most atheists and the potential for violence on the part of Xtians, Mahometans and their ilk) and simply pointing out the faults of the so-called proofs of the existence of god, atheists have finally whirled about like a well trained Roman legion and pressed the attack with powerful and novel proofs that god simply doesn’t exist. Dimicandum est! Let’s mix it up! The disproofs have become so numerous, it’s quite hard to keep up. The whole situation is pretty damned gratifying.

Consider Richard Dawkins’ The God Delusion, now on the must-read list of every rebellious teenager, which advances the atheological argument by proposing a disproof of the existence of God (or equivalently a proof of the non-existence of God) using concepts of evolutionary biology or, more properly, concepts derived from discussions about evolutionary biology. His conceptual framework is potent, I suppose, because goddists grow particularly ulcerous over the theory of natural selection and have been spending a lot of time recently scouring nature for design-worthy examples of living organisms.

Dawkins’ proof is on the face of it a simple one. It begins from the premise, derived in this instance from the writings of goddists who wish to prove the existence of god, that certain parts of the universe (and accordingly the universe as a whole) are complex. Examples of such complexity are the uniformity of property that characterizes all of the very large number of subatomic particles (proposed by Richard Swinburne) and the structure of the flagellar motor in prokaryotic cells like E. coli and eukaryotic cells like human sperm (proposed by Michael Behe. As of this writing, most of the scientific community has rejected the “irreducible complexity” of the flagellar motor as not true and downright idiotic. But let us assume that at some point an example were produced that wasn’t rejected. For the sake of argument; the structure of the argument itself needs to be examined). For goddists this complexity is such that the only adequate explanation for its occurrence is some sort of creator responsible at the very least for the complex parts of the universe. It is improbable that the observed level of complexity should obtain unless it were intentionally put into place. It is god, so runs the account, that put it into place. Dawkins argues that, if you add the assumption of the existence of a creator to the complex parts of the universe, then the properties of the creator god themselves are complex, so complex in fact as to attain at least the same level of complexity as the entities it supposedly created. This vitiates the explanatory value of a creator god theory because it does not eliminate unexplained complexity from the universe (a universe that now includes a very complex creator god). This is Dawkins’ viewpoint and it is valid as long as the step in the argument that asserts the equivalent complexity of the creator god is defensible. The concept of theic complexity is clearly the core of Dawkins’ disproof, but the theory he proposes is richer than the specifics of the disproof in that it proposes a reasonable alternative to the conceptually challenged designer theory. Traditionally defenders of the designer theory assume that the orderliness of the universe permits of only two explanations. Either the universe was designed or it is the product of “mere chance.” (Mere chance, as I argue elsewhere, is, properly understood, a name for the understanding that probability rigorously defined does not apply to the universe as a whole. You need a domain of possible outcomes to form a proper ratio between actual and possible outcomes. When you deal with the universe as a whole, the domain of possible outcomes is purely imaginary and infinitely large.)  But there is a third alternative, namely natural selection as played out in a series of events such that, however improbable the leap from the beginning to the end of the series, the transitions between intermediate terms in the series are less improbable and, once the series is fully filled in, each step is an instance of a natural law and so has close to 100% probability (or a very high probability) assuming the validity of the natural law (and assuming that a rigorous concept of probability even applies in these nearly totalizing, semi-metaphysical realms; plausibility might be a more suitable alternative.).

Dawkins aside, the goddist account could also lead to logical problems in that an infinite regress of creator gods would be an empty and meaningless recursion like a propositional function that by definition could only take itself and not meaningful names or propositions in its variable position (Consider a language that includes the rule that the variable position (“…”) in a function like “The thing that hits…” can be filled only by “the thing that hits…”.  Obviously such a bit of language is effectively meaningless. The only reason that such a phrase or a proposition including such a phrase does not violate the Law of Non-Contradiction is that, somewhat like a mere grunt, it does not have enough content to violate or to conform to any logical laws.) The same point can be made about utterances. A speaker contradicts himself if he says, “I am not now speaking.” A speaker embroils himself in an infinite regress if he says, “The meaning of every sentence I utter will be specified by the next sentence I utter.”  Contradictions and infinite regress of this kind are equally incomprehensible. (If I think I understand the infinitely regressive assertion, then I haven’t understood it because I haven’t yet heard the speaker’s next utterance.)

If the existence and activity of a creator god is assumed only as a premise of the goddist proof, then Dawkins’ argument is a refutation of that proof and not a positive disproof. However, if god is at least partially defined as the creator of at least those parts of the universe whose complexity is such that they must have been created, then this is a disproof of the existence of god that is at the very least valid from the standpoint of physical theory (empirical science). It may also be a logically rigorous disproof, for supposed propositions about the creator god could turn out to be functions that can only be satisfied by other functions ad infinitum.

Dawkins leaves at least two terms undefined, one of which, “complex” in the sense of “complex organism,” is of course very important (At one point he cites Julian Huxley’s definition). For on our understanding of “complex” rests the validity of that step in the argument that concludes that the creator god must be complex.

The other term, “absolute proof,” is not so critical, but since, when it comes to god, proofs are flying around everywhere, there is some merit in a better understanding of what we mean when we talk about a proof. Dawkins covers himself from trivial objections by asserting that “of course” you cannot absolutely prove the non-existence of anything. “That you cannot prove God’s non-existence is accepted and trivial, in the sense that we can never absolutely prove the non-existence of anything. What matters is not whether (the existence of  - WD) God is disprovable (he isn’t (it isn’t - WD)), but whether his (its – WD) existence is probable.” (p. 54) I suppose Dawkins wishes to acknowledge the difficulty of proving a negative.

Difficulty, not impossibility. Given even the most mundane understanding of “absolutely prove” (and setting aside for the moment philosophical concerns about the meaning of “existence” or “being” or whatever; I will return to that subject below), you can absolutely prove the non-existence of lots of things or supposed things. For example, unicorns that have horns and do not have horns (For simplicity’s sake, all examples are treated as tenseless; the translation to a time and place relative version is straightforward if tedious). Or the man who is and is not Julius Caesar. Let us call the concepts of horned-hornless-unicorns and identical-to-and-not-identical-to-Julius-Caesar-humans inconsistent concepts. A concept is inconsistent in this sense if its meaning is such that it must be transformable into a simple existential proposition such as “Something is a unicorn and that thing is both horned and unhorned,” or “Some person is both Julius Caesar and not Julius Caesar.”  Since either of these propositions, if asserted, implies a contradiction, they cannot be understood. Or, to put it another way, if either of these propositions is true, then everything is true. In my opinion, if a concept is inconsistent in this sense, then the individual it (purportedly) refers to does not exist. And this is an absolute proof of the non-existence of that individual. To cite a rather more practical example, one way we could prove the non-existence of a mean value for a given non-continuous function would be to assume the existence of the mean value for the function and then show that that assumption violates one of our premises (which obviously include the theorems of elementary mathematics). I suppose you could call this a logical disproof and it is by my lights the most absolute sort of proof we have in our arsenal of proofs. Most of the highly laudable disproofs of the existence of god start with a concept of god that has a certain richness of content (e.g. the argument from evil works with a concept of a morally good god). The designer god has a great deal of content. It is defined as a designing and creating entity, i.e. as an entity whose properties are defined by certain supposed properties of the physical universe. Note, however, that if we simplify the concept of god to the barest possible minimum and disprove the existence of that god, then the disproof is proportionally the strongest.

What may well be true is that you can never “absolutely” prove the existence of something without some sort of observation, since the consistency of a concept does not entail that there is anything that falls under the concept. Much goddist theology could be reduced to the failure to heed this simple caution.

One could of course argue that the laws of logic might change tomorrow and for that reason a logical proof is not absolutely absolute. This is not a completely foolish argument since non-Euclidean geometries and alternative logical systems have been devised without prior suspicion that they were over the horizon. However, if no proof is absolute unless it somehow anticipates a so far merely projected rejection of current laws of logic, then no proof is absolute. And any “proof” that the current laws of logic do not lead to “absolute” conclusions would itself have to rely on something (the laws of logic perhaps?) or else it would have no means of moving from one step of any proof to another.

If there is a proof that a given concept of god (There are so many) is inconsistent in this sense, then that proof is an absolute proof. Now sciences that are heavily observational (I shall leave the distinction between heavily and lightly observational vague since all sciences including mathematics (and even logic in my opinion) are observational to some degree. Suffice it to say that we usually regard biology as heavily observational and the first order propositional calculus as lightly observational if it is observational at all) do not make much use of logical or mathematical proofs of non-existence as outlined above. The concepts of ether and steady state creation are not, as far as I know, inconsistent. Their existence is disproved (or rendered improbable) because they have not been observed and their existence is not required by current models. Let’s throw in another term. Logical proofs and properly constructed non-logical proofs in the heavily observational sciences are rigorous proofs. Now I think that Dawkins’ proof, if successful, is a largely rigorous proof relying on concepts from biology of the non-existence of god. It is a perfectly valid scientific proof. Entities are proved to exist or not exist by science all the time. One proof is observation and whatever can be directly observed normally does not require extended proof. The usual situation is whether an entity is required by a model of some set of observable facts. If the model has sufficient predictive power and if its predictions are successful and if the entity is a necessary component of the model, then it has been proved that the entity exists. I suggest that an entity required by a successful model of a set of observed facts has some sort of direct or indirect causal relation with some of those observed facts. However, if the model’s predictions are unsuccessful or largely unsuccessful or of it is replaced by a superior model or if the model can be formulated without referring to the entity, then, as far as the scientific model is concerned, the entity does not exist. By Ockham’s Razor, therefore, it has been proved that the entity does not exist. The existence or non-existence of unobserved twin suns and large planets in distant galaxies is proved in this way. Dark matter and dark energy will likewise be subject to this procedure.

However, I think there are logical proofs of the inconsistency of many concepts of god (The fact is that most concepts of god such as the creator, the necessary entity, the completely good thing etc. are logically independent of each other, just as the concepts of horned-unhorned unicorns and tail-bearing-tail-free unicorns are logically independent of each other. The former are just jumbled in the heated imagination of your average goddist. In other words, even if the creator doesn’t exist, there still might exist a necessary entity etc. Of course losing any of the concepts just mentioned would most likely be unacceptable to Xtians and Maxjus in general; but we should still pay attention to their logical independence. The concepts of god buzz around along different trajectories, so we need to keep a fly swatter handy.) A logical version of Dawkins’ proof (viz. that the concept of a creator of complex entities is inconsistent) would a logically rigorous proof, again if it is valid.

The concept of complexity is a different story since it plays a role in Dawkins’ disproof. It needs to be defined or, failing that, better understood. Ultimately, while of some use in biological descriptions, the concept of complexity is so utterly vague in the goddist debate that one wonders whether it is worth rescuing. By adopting the goddist use of “complexity,” Dawkins embroils himself in this vagueness. “Complexity” in this context is used interchangeably with “irreducible complexity.”  I suppose, as gleaned from the way the antagonists in the debate talk, and for lack of an actual definition, that the addition of the notion of irreducibility means that somehow an explanation such as natural selection or some equivalent explanation from the physical sciences would be inadequate. This inadequacy can either mean that natural selection et al. have not adequately explained the complex phenomena in question. In which case the proper phrase would be “unreduced complexity.” Or it could mean that no explanation other than the creator god could ever explain the phenomena. Goddists would seem to need the latter since, if they mean the former, then all they really assert before rushing to the conclusion of a creator god is that the current physical and biological models are not complete, an observation no one, not even presumably the creator god, would deny. The latter, however, is so wildly far-reaching that its defense almost certainly falls into the realm of rubbish speculation. Trying to prove that we will never find an explanation for the little fellow’s possession of a flagellar motor seems not unlike the behavior of the gap toothed mystic staring at his navel. Now Dawkins’ disproof does not rely on the assumption that any complexity in the universe has to be irreducible. It can be irreducible, unreduced or even reduced. None of these alternatives invalidates Dawkins’ inference that the creator god must be characterized as having at least an equivalent level of complexity, at least at this stage of the argument. For this reason the validity bar is legitimately not quite as high for Dawkins as it is for goddists who would seem to need irreducibility or some other form of the inherent explanatory superiority of the creator god concept.

The concept of complexity by itself, however, irreducible or not, is a shoal on which many ships can be wrecked. In the first place, the predicate “complex” is not a simple predicate; it is relation. You cannot say meaningfully that something is complex and leave it at that. A thing is more or less complex than another thing. A particular is not complex by itself; it is only complex by way of comparison with another particular (one of its parts, for example). Another problem is that “complex” is an incomplete predicate. In English you have to give an adequate description of how something is complex. In what respect is it complex? A complex problem, a complex drawing and a complex personality are all different. So, just to say that a particular is complex tout court is close to saying nothing at all about it. Accordingly the phrase “a is so complex that…” as in “a is so complex that it must have been intentionally created,” where a is the name of a particular, doesn’t really say anything unless we specify what a is more complex than and in what respect a is more complex than whatever it is more complex than. (Hume’s part-whole argument against design was an insightful skewed anticipation of this point just as his concept of a vegetative universe anticipated natural selection.) Any part of Dawkins’ disproof that relies on this largely empty notion of complexity will suffer the same problems as goddist proofs that are based on the notion of complexity. (Genetically one sense of “complex” may have preceded the others and so the other senses at some point may have been analogical, as Mill (p. 28) understands analogical predicates. However, it would remain unclear whether some original sense of complex is really the sense used by creationists, i.e. whether their sense may or may not be derivative. Notwithstanding, the many senses of “complex” are now on a morphological equal footing. It remains incumbent on creationists to specify the respect in which they mean something is complex.)

But if you take away the notion of complexity from goddist proofs as largely meaningless then you take away the observational basis which they use as a premise (Swinburne’s proof does not rely on complexity; however, as I argue elsewhere, the idea of improbability is just as big a trap). The goddists begin by looking at certain things such as swimming bacteria and dancing electrons and they call those things complex. Without these initial observation statements no proof. And, even if Dawkins’ notion of complexity does not simply piggy back on the goddist notion, it is equally deficient as long as he also uses a concept that is relational and/or requires further specification. Dawkins in particular assumes the premise that some entity cannot have been created by another entity unless the creator entity is at least as complex as the created entity. In my opinion, the misuse of the concept of complexity is enough to invalidate the goddist blather and eo ipso Dawkins’ disproof. Case closed. This essay could end here and we could all go home and smoke a joint. However, if we keep in mind that complexity as used is a meaningless concept, we can talk around Dawkins’ proof. For one thing, there are many interesting things left to say relating to scientific methodology. We can relieve our mental discomfort at using nonsense terms by replacing “is so complex” in our minds with “possesses some property.”  Thus “Flagellar motors are so complex that they must have been created” becomes “Flagellar motors possess some property such that they must have been created.” Observationally unpromising but at least rid of one level of conceptual confusion.

Dawkins’ assumption that a creator of a universe that possesses a property that requires creation must also possess that property appears to be intuitively obvious, but it hasn’t been proved to be true. The intuitive appeal of the assumption that a creator must be at least as complex as what it creates does waver a bit under inspection. Of course, it is highly counterintuitive to think that a cat could create a king or a Xtian read a book. But the assertion that a simpler entity can create a more complex entity (assuming we know what these terms mean) does not violate any logical laws. (There is a weird mediaeval theory, which is immediately contradicted by everything else the mediaevals say about god, that god is absolutely simple in the sense that it is a simple substance. For us normal folk, the concept of an absolutely simple substance is simply incoherent. Every assertion made about the substance god adds to its complexity if only by negation (If god is not F, then it is G where "G" stands for the complement of F. Cf. Dawkins p. 155). This brings up the related strategy of negative theologians whose trick seems to consist in denying any assertion made about god. If god is not-everything, then it is of course nothing (despite some rather pretty poetry by John of the Cross). The simple substance theory has absolutely no bearing on this discussion. In fact my advice is, don’t even go there. You need to ingest at least half an ounce of shrooms before you can begin to understand the simple substance theory.) And what is the significance of the following examples? (1) A passing storm uproots trees and branches and significantly alters a local ecosystem in favor of “more complex” life forms than had existed there previously. (2) Einstein’s mother. (3) The famous monkeys typing the complete works of Shakespeare in however many years. (4) We humans create a computer program that can replicate and not only perform calculations at a far faster rate than we can but is also much more adept at model creation and even exhibits affects and volitions we cannot understand. Admittedly (1) and (3) border on (or are in fact) chance occurrences, so as examples they would have little appeal to goddists for whom the origination by chance of complex organisms is the unacceptable alternative to creation. Nevertheless, it is not clear that they are not examples of a simpler thing creating something more complex. (2) and (4) appear even more problematic, with apologies to Einstein’s mother. (4) in fact hints at a kind of super evolution where one type of entity intentionally creates something that turns out “more complex” (presumably because much of what it does is incomprehensible to its creators). Because of these examples, I think it is not at all clear that a creating entity must be at least as improbable (because it must be at least as complex) as what it creates. But the source of the vagueness is the vagueness of the term “complex,” not just because the concept that term stands for is relational and incomplete as far as its meaning is concerned, but also, as the examples illustrate, we can project cases where a seemingly simpler thing appears to create a more complex thing.

There are two other terms in this network of arguments whose meaning, or lack thereof, causes trouble. The first is “improbable;” the second is “creates.”  “Improbable” is used in so many ways by goddists as to arouse suspicions of deliberate obfuscation. The flagellar motor example is supposed to imply that natural selection cannot account for its origin and therefore it is improbable in a biological model based solely on natural selection. Some goddists get all highfalutin’ and start babbling about probability theory and Bayesian probability calculations. Mostly, I suspect, “improbable” is just the word that runs through goddists’ minds as they stare at the night sky with their mouths open. As I understand Dawkins, he uses “improbable” to mean, “Whatever sense of ‘improbable’ you choose, it is only reasonable to regard the creator god as equally improbable.”  If you think the flagellar motor could not have just popped on the scene, its creator too could not just have shown up at the door one day. If one needs a creator, so does the other. If you think the chance of the universal constants being what they are is one in a bazillion, so the chance of the creator of the universe being the way it is is at least one in a bazillion plus one. If the starry night makes you feel all queer inside, then think how any reasonable person would react to the thought of something creating the starry night. Dawkins aside, each of these three senses of “improbable’ is severely wanting. Even if natural selection had not yet developed an explanatory model for the appearance of the flagellar motor (and further empirical observation indicates that it has), that is no reason why in time it could not. And any proof that natural selection could never devise an appropriate model to explain the origin of the flagellar motor would have to be a logical (or some other sort of a priori) proof. Good luck. The argument from probability theory is based on such a flagrant mathematical error (for division by infinity either contradicts the initially assumed binomial or is disallowed) that it should be included in elementary texts on probability theory as an example of how not to do mathematics. As for the last, well just watch out for bird shit.

Given the tangle involving the concept of complexity, it is probably not helpful that Dawkins doubles down on this infelicitous term. He can and should say that if something has a property that renders its existence improbable unless it had been created, then whatever created that thing itself has a property (which may or may not be the same property) that also renders its existence improbable unless it too had been created. This won’t be greatly helpful as long as the original property is complexity, but it does clarify that the corresponding property of the creator doesn’t have to be the same property as that possessed by the created entity. Thus restated, Dawkins’ proof asserts that, if certain phenomena are F in such a way that they must have been created, then whatever created them must be G (where “F” and “G” stand for properties). G could be the same property as F, but it doesn’t have to be. Anything that is G must have been created, therefore etc. In this formulation “in such a way that” may include the sense of improbability, which is what induces us to accept the first premise.

The goddists and Dawkins have different tasks. The former must find some property F that allows them to accept as true the protasis. Dawkins needs to show that any such property or a related property G must also characterize the creator. And he must show that a related property must be such that the creator must also have been created. I don’t have any good suggestions for F. In the first place I don’t think there are any such properties. And I’m not in the business of helping goddists. As regards Dawkins the proof would be that, for whatever F you can find, I can find a G such that anything that is G must have been created (or alternatively that G is an inconsistent concept). For any reason you may have that something that is F must have been created is also a reason that some thing that is G, (e.g. something that can create an F) must have been created. An obvious candidate for G would be “can create particulars that are F” or even “creates particulars that are F” assuming that there can be F’s. 

At first blush there is nothing inconsistent in the notion that something can create another thing (Assuming we know what “create” means) and it looks doubtful whether we could find a property G without knowing more about F or in some other way saying something more about G itself. So we may not find a disproof of the existence of a creator solely on the assumption that some things need to be created. In addition, an inconsistent concept would almost certainly have to be a composite of two other concepts both of which cannot apply to the same particular (horned and hornless). And the concept of a creator that created things that are such that they were created is not, on the basis of the little information we have so far, inconsistent. In fact it appears to have no more content than is dictated by the assumption that there are F’s in the world.

But we may be able to argue that the concept of a (complex or simple) creator god is by itself infinitely regressive. Take the following sample argument that the concept of a creator god involves an infinite regress even if the creator god is not as “complex” as the universe it created. Call the physical universe, the universe of natural science and biology, U1. Assume that U1 manifests a sufficient degree of “complexity” (or alternatively possesses some property) such that it must have been created by a creator god. Now call U1 plus the creator god the larger universe U2. U2 consists of the union of U1 and the creator god. But U2 is no less complex than U1, even if the creator god by itself is no more complex (or improbably even less complex) than U1 or elements of U1. The addition of the creator god does not diminish the complexity of U1, so U2 shows a sufficient degree of complexity to require a creator god₂. The union of U2 and creator god₂ is U3 and so on. Just as the repeated utterance that the meaning of my sentence will be specified by the next sentence I utter can never be understood because its understanding would require hearing and understanding an infinite number of utterances, so the universe cannot have been created in a comprehensible way by a creator god_n because its creation would require the creation of an infinite number of universes by an infinite number of creator gods. (Even if the infinite series of creator gods is generated by elementary recursion, this does not add to its comprehensibility; the utterance regress specified above is also generated by a recursive function.) Not to mention the consequent polytheism. Dawkins would observe quite correctly that a theory that involves infinite regress would be rejected on empirical scientific grounds because it lacks explanatory value. But the problem with infinite regress is much deeper. A contradiction is incomprehensible; it entails everything. A theory that involves an infinite regress can be shown upon analysis to involve a propositional function that can only take itself as an argument; since only fulfilled propositions have truth values, an infinite regress entails nothing because it has no meaning (no truth value). In this case the empty propositional function would be something on the order of, "...was created by the thing that was created by..." where it is stipulated that the variable position can be fulfilled only by the same propositional function. Now this infinite regress might be resolved by the stipulation that for any god_n-1 and god_n, god_n-1 is identical to god_n. But the unhappy consequence is that the creator god, so understood, created itself. This is indeed a contradiction because the creator God would have to have existed before it existed. Infinite regress or inconsistency: This is the first logical disproof of the concept of a creator god. In honor of the Dumb Ox, we might call it the First Way. However, it is somewhat weak because it depends on the assumption that a given universe U_n is such that it must have been created. Also it might be avoided by changing one’s position from saying that god created the whole universe to saying that god created only those portions of the universe that it created – basically a version of the space alien theory of creation (Not to mention the fact that Dawkins’ disproof disallows this change of position).

Furthermore, we cannot say one more thing about the particular that is G without inviting even greater inconsistency on the part of the richer concept. Consider the concept of a G that created the whole universe (and not just the F’s in the universe). The concept of a god who created everything is indeed inconsistent simply because the concept of everything is a Typhoid Mary of paradoxes. Xtians cling to the idea that god created everything with the tenacity of wild animals. The reason I suppose is that a god that did not create everything is just Jupiter. In fact they often say that apparent paradox in the god concept (Another example is the problem of evil) is just the result of our limited understanding. Speak for yourself.

To reiterate, if the universe is everything that exists and not just the physical universe, then, if god created the universe, god must have created itself. Therefore it must have existed before it existed. Or, if the universe is everything except god, then god must have created the act of creation. Consequently the act creation must have existed before the act of creation existed. So the concept of a god who created everything is inconsistent. For it is a contradiction to assert that god existed before it was created, i.e. that it existed before it existed, and also that the act of creation existed before it existed. More accurately put, the propositions that are the meaning of those assertions violate the law of Non-Contradiction. (Some might step in at this point and say that god created time when it created the rest of the universe. So arguments based on before and after are not valid. They might do this, but what they say makes about as much sense as saying that dingles blank the spam. It is certainly not the same as the legitimate hypothesis that the current configuration of space-time originated with the Big Bang. For that space-time is the rather precisely defined notion of relativistic and quantum mechanics that permits but obviously does not specify logically consistent alternatives. As for Augustine’s opinion, well, he was a hippo. In fact we can leave time out of it and just use the term “without”. The concept of creating creation is still inconsistent.) Xtian theology is rather muddled about whether god created everything or not. Though that is the common assumption, I presume that they do not mean that god created itself. However, not much is said about things like the act of creation and it seems quite clear that the act of creation also could not have been created by the creator god. But at this point Pandora steps in and the supposedly empty universe turns out to be full of shrieking, jabbering entities that could not possibly have been created by the creator god. Consider the ability to create. Consider planning, the ability to plan, foresight, intent, potential, imagining, understanding (Fill in the blanks). Or take number (presumably only one universe was created), simplicity, complexity, time (mediaeval mumbo jumbo notwithstanding (Cf. Kenny in Martin & Monnier p.211), if a thing is created, then there is by definition a time when it is not and then a time when it is). Without all of these it would be hard to say that some god actually created the universe. Without intentions and capacities god is like the passing storm that blocked the stream, a concomitant circumstance and not a creator. Indeed even in the latter case the creator god could not have created chance or circumstances.

Suppose a Xtian (who is also of needs a thoroughgoing materialist) objects that these uncreated entities are abstract concepts and not real things at all? But in order to maintain such an objection, our Xtian would have to explain what he means by a “real” thing (a rabbit hole if there ever was one). Indeed it is hard to imagine how a creator could create anything without the intention to create that thing. Without that intention, we are left with either a chance concurrence of circumstances or a sort of evolution of one thing out of another (and we certainly don’t want to talk about evolution!). And if intentions could not be in some sense real, we could not distinguish between real intentions and supposed intentions (We thought she intended to run down her boyfriend but he slipped in front of her car). Jurisprudence would come to a standstill. And, if we wished to adhere to a strictly monist viewpoint and to clothe our monism in the language of physical material objects, then intentions (and acts, capacities, imaginings etc.), could be viewed as alternative descriptions of brain states just as a computer program is no more than a structured silica network. Both are very real and neither could have been created by a creator god.

Or else assume that the creator god created our intentions, but it didn’t create its own intentions? Even then we can only understand the concept of a creator god’s intentions by assuming they have something in common with our own intentions. The two types of intention must have something in common in fact or they would not both be intentions. So while the creator god may have created Mahomet’s particular intentions, it still could not have created intentions in general. (I know god is outside time and beyond our limited understanding and all that razzmatazz. That’s the problem with theology. It’s so negative.)

We cannot assert that the creator god created everything on pain of violating the Law of Non-Contradiction. Indeed there are not insignificant parts of the universe that it could not have created. (I can hear your garden variety Xtian preacher now. “De Awmighty don’ need no Law of Non Connadikshun. He can send de Law of Non Connadikshun to boin in Hell like the rest o’ you Commanists.” Whatever. The Awmighty can do what it wants. The fact remains that any assertion that violates the Law of Non-Contradiction is meaningless. Another way of saying this is that it implies everything. Therefore, it is true that the Awmighty eats its own turds.) Inconsistency is avoided if we redefine the creator god as having created some things but not everything. At first blush and from a logical point of view that is a legitimate gambit (although the result leaves the creator god as something more of a constitutional monarch than the all-powerful poobah religionists prefer). Even if the concept of a creator of what needs to have been (and has been) created were not inconsistent, the richer concept of a creator of everything is inconsistent, and it turns out that there are very specific things that we cannot on pain of inconsistency assert that this god created. But indeed the first blush turns out to be a malarial fever. If the creator god did not create its own act of creation and its own capacity for creation, it did not create generic creation and capacity that subsume specific types of creation and capacity. But if it didn’t create the generic creation and capacity, it didn’t create what is critical to the specific creation of the flagellar motor, not to mention the rest of the visible universe. Like a good apprentice, it simply applied a pre-existing mechanism. For if there is no creation or capacity at all, then there are no specific types of creation or capacity. If the creator created it did not create the capacity for creation. Therefore the creator did not create. This concept is inconsistent and so the creator that created everything does not exist. This thing that does not exist is what we call god. QED and all that. It seems that the creator god could not have created the visible universe, because it could not have created creation. This is another logical disproof of the notion of a creator god. It is the Second Way. This proof is stronger than the First Way because it does not depend on the assumption that there is something about the universe that requires its creation.

By way of summary, the creator cannot have created creation or the capacities we associate with creation on pain of violating the Law of Non-Contradiction. A path may be walked that tries to avoid this violation (e.g. the creator created our act of creation but not its own) but the result most likely would end up completely unlike anything we can recognize as creation. It could look like chance occurrence. It would also introduce its own smorgasbord of problematic uncreated phenomena and capacities including generic concepts. (A particularly clever version of this argument was devised by Gilbert Fulmer (Martin & Monnier, pp. 326 ff.) to the effect that a transcendent creator would be subject to a natural law ordaining that its act would regular lawlike consequences so that, stated by way of permissible simplification, that creator would not transcend the natural universe at all.)

Now creation, the capacity to create and everything involved in those concepts is indeed a good candidate for the property G of a Dawkinsian creator. Under these circumstances the Dawkins argument could be framed a) by relying on the original creationist “proof,” viz. the creator god is much too complex not to have been designed and created by something else, or b) as an independent intuition – the creator god must have either been designed or created (if indeed these concepts make any sense) or its existence would have to be explained by a sophisticated mechanism like natural selection.  

It is notable that this gloss on the notion of a creator god throws an unflattering light on the deistic (and apparently Neo-Platonist) god, the god of Voltaire (more or less), the abstract creative force. The notion of creation is anthropomorphic or it doesn’t make any sense. For the idea of a creator without any of the qualities sketched above runs counter to our intuitions of what a creator must be. This should be equally and painfully clear with respect to Voltaire’s view that the abstract creative force is also benevolent and just. What are these but human qualities? Spinoza’s god, namely the universe that creates itself, may be truly non-anthropomorphic if we understand the idea of a universe that creates itself as code for a universe that was not created, a nature that is just there and whose parts behave according to regular patterns that show no evidence of having been imposed from without. But, understood that way, Spinoza’s universe did not literally create itself. Philosophical analysis it can be applied to the concepts of any number of ways that god is understood. Indeed beautiful existential disproofs have been devised using the understanding of god as omnibenevolent, all-powerful and all-knowing (assuming the brick wall postulate that nothing exists whose concept is inconsistent). There is a like contradiction in the Xtian myth of hell where some of the punished are also punishers. If devils consider it good for them to inflict pain on humans, then, if they are successful, they realize an end that is good for them and so are not themselves completely punished. The successful punishment of humans is a reward for the devils, not a punishment. This contradiction was noted by Voltaire in his Dîner de Boulainvilliers (Mélanges p. 1304)

Now we also have non-purely-logical scientific grounds why the concept of a creator god can be rejected without relying on the vague and mathematically disallowed concept of improbability or borrowing the junk concept of complexity from the goddists. I proposed that the concept of creating itself properly understood would be a good candidate for the property G, which would be a more rigorous version of Dawkins’ concept of complexity. A god that did not create the mechanism of creation did not create the mechanism whereby E. coli and its clever little flagellum was created by evolving from a pre-existing species. Or better, if that god did create the evolutionary mechanism specific to that event, it still did not create the generic mechanism of creation through evolution.

Now upon further reflection we can understand that creating or being able to create as the property G is sufficiently rich that it vastly exceeds its explanatory value for phenomena such as the flagellar motor or the universal constants. If we are to accept the concept of a creator or of something that can create (even assuming that concept is consistent, which, as we saw in the case of the creator god, it isn’t), we require more than its hypostasis in an explanatory theory. We need direct observation of this entity. Show us the goods. This replaces the concepts of improbability or complexity with the scientifically more precise concept of direct observation. The creator god is such that if we don’t see it, it doesn’t exist.

The discovery of Neptune after its hypostasis on the basis of irregularities in the orbit of Uranus illustrates this point. When Leverrier proposed that an as yet unobserved planet was responsible for the discrepancies in the observed and predicted positions of Uranus, he hypostatized much more than an entity characterized solely by its gravitational effect on Uranus. This entity was a planet with everything that concept entails. In order to exert gravitational attraction, it must have mass, occupy a position in space-time and have measurable spatial dimensions.  It must consequently be composed of elements and compounds in either solid, liquid or gaseous state or some combination thereof. It would most likely have to be ovoid or nearly spherical. It would most likely need to reflect light and so be visible by way of a telescope. This is the sense on which the concept of an entity exerting gravitational influence on Uranus is too rich to be justified solely on the basis of its explaining observational discrepancies. Neptune does not exist simply because Uranus wanders a bit off course. If Galle had not provided visual verification, it would have been back to the drawing board for Leverrier.

The concept of the creator god is also too rich to be accepted simply because it explains certain natural phenomena. For starters it would have to be distinct from what it created (otherwise the occurrence would not be creation but some wacky form of evolution). The creation would almost certainly have to be intentional. Inadvertent or accidental creation would have no explanatory value since we might just as well settle for the inadvertent occurrence of the phenomena to be explained. Likewise some form of viviparity or cell separation mechanism would have to be different from observed examples of this type of biological mechanism and its particular characteristics would have to be demonstrated or else it would have no explanatory value whatsoever. So the intention to create would almost certainly be required. As well as, as I noted above, the ability to create, the disposition to create and so on. If this entity has intentions then it must have the capacity to choose. It doesn’t make much sense to say something has the intention to create without allowing for the possibility that it may choose not to create. Similarly it must understand what it is doing. It must have some pre-existing idea of what it wants to create. Which means it must want to create. So it must have understanding and will – or something like what we recognize as human or animal understanding and will. Since its creation was successful (We can observe flagellar motors) the creator god must have the ability to realize its designs. Accordingly it must have abilities, capacities and powers. Not all of these have to be exactly the same as their human counterparts, but they must be recognizably similar and they must fulfill their purposes. (There are so many questions. If the creator god makes choices, does it have preferences? Does it like jelly doughnuts or does it prefer Kibbles & Bits? Or does it watch its weight? Does it have an opinion about fake boobs? Or pussy shaving? Of course, it may prefer little boys. Is it a Republican or a Democrat? Maybe (Gasp!) it’s not even American. I suggest these pressing questions require verifiable answers or we must reject the notion of the creator god out of hand.)

Of course your garden variety Xtian would exclaim, “That’s him! Intentions, will, power, even a preference for Krispy Kremes. That’s our turd boy, god of gods, king of kings and his own little bird dropping in that manger long, long ago.” But not so fast. Obviously. This concept is way too rich to be justified solely because it explains (if it does explain) why sperm have propeller screws. Dawkins says the existence of something so definite, so, shall we say anthropomorphic, is not bloody likely. A more rigorous formulation would be that an entity this definite requires direct observation. If we can’t see it, it ain’t there.

Consider another requirement. The creator must either have used something pre-existing to create the universe, like the famous potter and his clay, or else it would have had to create the universe out of nothing. If it used something, what is this something? How can it be described? What does it look like? Wouldn’t the existence of the primordial ooze also need some sort of verification beyond the existence of propeller screws? The idea of creation of something from nothing, on the other hand, involves a bit of a metaphysical tangle. For one thing, Locke’s causal proof of the existence of god involves the assumption that nothing comes from nothing (Obviously unwarranted as Hume pointed out). But, aside from that, one wonders what exactly does creation from nothing (like Merlin saying “Abracadabra!”) explain? Ultimately this aspect of creation itself requires a good deal of splainin’ just like god’s mental state when it did all that creating.

Of course, concepts like “rich” and “definite” are every bit as relational and incomplete as “complex,” the concept I have spent no little time dumping on. I hazard to assert, however, that I am using them in a slightly different way, and that makes a significant difference. To say that a concept is rich, for one thing, doesn’t add much of anything (or anything at all) to the reasons we give for saying that the concept is rich. When we say something is complex we say that it has qualities F, G, H etc. and it is the possession of these qualities that makes the thing complex. But to say that it is complex is to say something more than that it is just F and G and H. Complexity adds further content to the simple enumeration of the qualities a thing possesses. I feel I am using “rich” however as a simple stand-in for the sum of qualities a thing is supposed to possess. There is nothing about the richness of its concept that requires direct verification of whether something we speculate exists really exists or not. Richness is, rather, simply those qualities for which “rich” is a kind of shorthand. So, while “rich” may be relational and incomplete, “intending to create” is not in the same way (Nor are the creator’s hopes and fears and secret desires to produce a nubile little Eve so it can fuck with her mind). And “rich,” in the way I am using it, means nothing more than “intending to create” and the like qualities enumerated above. In this way my definition of “rich” is operational and not informative. What counts are the qualities it stands for. Now Dawkins and his creationist nemeses could argue that they use “complex” in the same operational way. If so it would have been helpful had they explained the real concepts “complex” is supposed to stand for.

Another formulation of an objection to the concept of a creator god from the standpoint of empirical science would be that the concept lacks explanatory value. We have already found good philosophical reasons why this concept cannot be further specified without inviting inconsistency or infinite regress. Now if we broaden our focus on the creationist hypothesis from the “problems” it is supposed to solve and consider the explanation itself we find the idea of a bearded wonder running around with stars and moons on its cloak and waving its magic wand actually raises more problems than it solves. The facts implied by the explanatory theory are in as much if not greater need of explanation as the observed facts it is supposed to explain. A definite bummer. And an entity that cannot be observed and that does not figure in a model with explanatory value does not, for all intents and purposes, exist.

Now the premise of the creationist argument, and eo ipso Dawkins’ (ultimately rejected) premise is that it is highly improbable that certain natural phenomena exist. But of course those natural phenomena do exist. We know that through observation. Dawkins argues that a designer god is not required by the facts assembled to support the creationist theory and that it suffers from the same faults as the “irreducibly complex” universe itself. Nevertheless, just as some of those facts do indeed obtain despite their improbability, so a designer god may exist as a brute fact. It may just be there hidden under a rock somewhere designing away, even though its contribution is not required by the facts and its irreducible complexity is in dire need of further explanation. Who knows? The improbable designer god may have even evolved by minute probable steps from some Ur-designer god that in the beginning designed only relatively (Relatively?!) simple things like hydrogen atoms. Its evolution may have even been spurred by a type of natural selection; it may have displaced other designer gods with less whatever it takes to be good designer gods. However, Dawkins’ argument provides a powerful disinclination to take seriously this bit of speculation barring any real observational evidence. For along the same speculative lines, David Hume may have really been the designer god in mufti, enjoying the sort of cosmic joke that only designer gods can really appreciate. This may be just a brute fact whatever its improbability as long as its obtaining does not violate the Law of Non-Contradiction. But barring actual observational confirmation or the need for the Hume god or any designer god in models that explain actual observed facts, we are left not so much with a disproof as a vast sea of indifference about the status of the Hume god or any other designer gods. Who really cares whether a designer god exists? Who gives a shit about god? I don’t.

Theologians tend to step in at this point and assert that the real god is not the turd boy of the popular imagination, but a complex metaphysical entity that blah, blah, blah. I suppose the implication is that we should never require something so vulgar as actual observation when it comes to turd boy. However, the existence of Allah or Baal or Yahweh or Jesus X or Jupiter is subject to empirical, i.e. observation-based test because of the factual claims made about their interaction with the observable universe. We can empirically disprove the existence of individual gods who either show themselves or purportedly leave traces of their activity in the physical universe. Xtianity and Mahometanism do sometimes shade the way they use the term “god” such that it doesn’t function in any simple way as a referring term whose referent is an individual (a first-order referring term). The way this term is shaded is vague and largely involves a misuse of logic and metaphysics. It is used somewhat like a class term, but it is not a class term (i.e. the set including Allah, Baal, etc.).  It is often understood somewhat like an abstract or predicative term (i.e. like “red” or “sexy” or “good”) but, in most cases where theologians use the term, their intention is not uniquely predicative. Except and not always in cases of extreme pantheism (e.g. Spinoza’s “The universe is goddish” or “The universe gods”), uses of “god” in a predicative way as meaning an abstraction are thoroughly vague. One suspects that at bottom most goddists including Xtian and Mahometan theologians (Some, like D.F. Strauss, don’t mean anything and are quite candid about it) down deep really do mean an individual when they use the term “god.” That is, they use it as a proper name like “Julius Caesar” or “Kayden Kross.”  However, when the individual god begins to look too much like some guy on the side of a mountain somewhere, they start to stress an abstract sense based on all the wonderful qualities the individual is supposed to have. This vagueness explains Dawkins’ perplexities in debating Oxbridge theologians on the one hand, whose god is largely a deistic abstraction, and creationist propagandists whose god is a Wizard of Oz living in a forest cave far far away. But these are not the same thing.

Dawkins’ proof (or our version of it) is valid to the extent that the meaning of “god” is limited to its use as a first-order referring term whose referent is an individual. It is invalid to the extent that “god” is not used in that way but rather in some sense that assertions like “God exists” are not subject to the kind of direct empirical verification in the same way that “There must be a raccoon tipping the garbage cans” or “There must be a real actress starring in Kayden’s First Time” are. But, as we saw, the reason that “God exists” or “God is tipping our garbage cans” are not subject to direct verification and so are not properly the target of Dawkins’ proof is that such assertions are largely meaningless.

Nevertheless, the assertion that god and its existence are not subject to scientific verification are what theologians call NOMA or Non Overlapping Magisteria. This is academic faggot talk for “You got your department and we got ours.” It comes down to saying that concepts and practices such as verifiable observation, consistency of argument and peer review - practices that inform the warp and woof of science (To which I may add that such concepts and practices are not the exclusive domain of science but are equally constitutive of jurisprudence, criminology, sports refereeing and teaching little Timmy not to lie; what distinguishes the institution of science from these phenomena is the practice of model creation and, wherever possible, the use of mathematical techniques) - that these concepts or practices, which in this context are central to establishing the superiority of the model of natural selection over the creationist model, are out of play when it comes to god. Perhaps. But the issue addressed by NOMA is not god so much as theology, i.e. a discipline whose practitioners act very much like scholars in other disciplines. And the magisterium of this discipline overlaps in any number of ways with those of humanistic scholarship and even of the hard sciences. Otherwise, even though a Bible “scholar” by way of example may, according to the rules of this mysterious other magisterium, not be able to say whatever he wants about some text or event, nevertheless his conclusions would presumably not be subject to anything so mundane as establishing the text, philological verification or any sort of peer review, since those are practices belonging to the wrong magisterium. Moreover, the creationist argument for the existence of god based on evidence of design in the physical universe does belong by its very nature to the magisterium of physical science. It has pretensions to gentility. You can’t propose a scientific “proof” for the existence of a creator god and then when scientists proceed to flatten your “proof” claim you operate in some heavenly mystical sphere not accessible to science. Please no shell games.

Something tells me that NOMA is nothing more than a New Age version of the hoary imprecation that faith begins where reason leaves off. And we remember that the question of “Which faith?” causes no end of problems for that worm eaten jingle. As La Mettrie’s Philalète observed, “Je trouve que chaque secte se sert avec plaisir de la raison, autant qu’elle en croit pouvoir tirer quelque secours: cependant dès que la raison vient à manquer, on s’écrie que c’est un article de foi qui est au-dessus de la raison.”

Something may be said for NOMA along the following lines. Let’s take Frazer’s idea of three stages of the explanation of natural events as a guide. The first is the animistic stage where it is believed that objects of the natural world possess human-like will and motivation. The second is the religious stage where it is believed that natural objects do not have these qualities, but that non-human agencies, the gods, who are also endowed with human-like will and motivation, interact with natural objects and cause them to behave the way they do. The third is the scientific stage where it is found that events involving inanimate objects exhibit unvarying regularities that can be expressed in mathematical theorems. A NOMA-like argument would put forth that the religious and the scientific (and why not animistic as well?) views can both be true. However, the criteria for a true assertion within one “magisterium” cannot be applied to the other. Mutatis mutandis scientific proofs of the existence or non existence of entities cannot be applied to entities required by a religious model of the universe. In fact pre-Laplacean physical theories were comfortable with this distinction; much so-called natural philosophy tried to unite in one grand theory explanations by unvarying regularity and explanations by humanoid agency. According to some views the humanoid agent personally intervened in any and every natural event. In fact its interventions could be arbitrary. The deistic view placed the humanoid agency at the beginning of the series and dropped it from succeeding events.

According to NOMA this should be a satisfactory compromise. But why then was it consigned to the scientific attic towards the end of the eighteenth century and why should Dawkins’ proof be considered as more than just scientifically valid? The reason probably lies in supplementary arguments, the amici curiae, so to speak, of the court of intellectual opinion. The historical claims of the Xtian myth were shown to be demonstrably false. The moral pretensions of religion were cast into doubt. The political role of religious sects was viewed with a less than favorable eye. The continuing validity of these supplementary arguments is shown by the second half of Dawkins’ book. The presumption in favor of two non-interacting domains changes into a presumption against the utility or desirability of the religious domain. Since specific religious models are not required to explain any individual observed facts (I hate to break this to you, Boopsie, but there are no miracles) and since the religious magisterium is by definition redundant upon the scientific magisterium, the religious magisterium may be ignored without loss of explanatory power.

But there is more. Individual religious models are not simply redundant alternatives to scientific models. In many instances they entail conclusions about the facts of the observable world that conflict either with direct observations or with scientific models derived from direct observations. To the degree, for example, that the Xtian myth deduces certain conclusions about the natural world from assertions in the Bible that contradict a scientific model, the Xtian myth and the scientific model do not reside in separate magisteria. If you believe in the truth of the Xtian myth, you cannot believe that the earth has a certain geological age, that the earth revolves around the sun, etc. So, while NOMA sounds fine in theory, actual religious models do overlap in practice with scientific models. One suspects that if they did not, the resulting religious model would be so content-free as to be undistinguishable from atheism, a charge often leveled at deism. Moreover, when the scientific model is not several mathematical steps away from simple observations, but rather intuitively linked by common  sense (as is the case with the geological age of the earth or simple evolutionary observations), then it is easy to see that the religious model contradicts observation as well. The real life consequences are unacceptable. Locke and Descartes were warned that they should not propose any theories whose conclusions contradicted either the Bible or papist doctrine or both; and Galileo was punished for ignoring his warning. Today we are told that certain scientific models are false because they contradict certain religious models. This is not NOMA. It is nonsense.

There is a broader conceptual problem. When theologians assert or defend NOMA they assert one or more propositions.  Furthermore their intent is to provide a reasonable defense of religious talk. They want other people to be convinced that theology is constituted as a separate magisterium. To what magisterium do the propositions asserted about or in defense of NOMA belong? Do the propositions that assert NOMA belong to the theological or the scientific magisterium? If they belong to the former, then their justification falls within the theological magisterium whose separateness has not been independently demonstrated. The question is begged. If they belong to the latter, then, to the extent that they say anything about theological issues, what they say is also subject to scientific verification. In that case some or all of theology also falls into the scientific magisterium. To the extent that theologians intend to provide a convincing defense of religious talk one is hard pressed not to conclude that at least part of their defense of NOMA overlaps with the purported scientific magisterium. To the extent that any evidence put forward for NOMA is publicly accessible (e.g. alternative experiences), it falls eo ipso in the domain of science by virtue of being subject to confirmatory examination. These are logical problems with the assertion of NOMA. If they are not answered satisfactorily, NOMA itself is simply incomprehensible. Rather ironic in that NOMA was devised to explain away the incomprehensibility of theological concepts.

But what about the very idea of existence? What do we even mean when we say that something exists or doubt whether a particular or some class of things may exist? Well, if you think the waters have been deep so far, you ain’t seen nothin’ yet. Wading is no longer an option. You gotta learn to swim. Most non-philosophers including some scientists manage quite well talking about things that exist and other things that don’t exist without a great deal (some, but not a great deal) of concern as to what “exists” means. In most situations this is harmless, since some shared background assumptions on the meaning of “exists,” in the context of the utterance, mental or otherwise, are in play, such as when we discover that a tenth planet exists. In most cases – likewise in most cases it is harmless to make Newtonian assumptions about the fundamental laws of the physical universe. All participants in the creationist debate have assumed a minimal working criterion of “exists” to be, “an entity exists if its existence is required by the true observation that the universe is designed.” The questions of what that criterion for existence entails or whatever properties it may involve do not necessarily have to be answered by that minimal working criterion. The issue is whether any particular entity distinct from the physical universe is indeed required by the observed characteristics of the physical universe. Creationists say “Yes.” Anybody with any brains says, “No.”  A different criterion for existence enters when the possibility is raised that the creator god does indeed exist even though its existence is not required by the true observed state of the physical universe. Lacking evidence to the contrary, it is logically possible the butler murdered Sir James even if the evidence pointing to his guilt turns out to be flawed. Likewise it is logically possible that god exists even if the evidence pointing to its existence is flawed. In response to that logical possibility a new minimum criterion is introduced, that of observability. Failing the design justification, we should at least be able to see or otherwise perceive the creator god. It is not my intent here to deal with the validity in the abstract of this new minimum criterion, and there may be legitimate arguments that it is not a necessary condition. I merely observe that it is introduced in a context where some criterion for the existence of the creator god is felt to be necessary or else the design argument would not have had to be introduced in the first place. My interest at this point is to step beyond these working criteria and clarify a few general points about the concept of existence.

A great deal of philosophical blood has been shed over the meaning of “exists” and, in some cases at least, we trust that blood has not been shed in vain. Among those few cases where the meaning of “exists” is important we range the greater number of theological assertions, usually about god or gods and its/their minions. So what do we mean when we say, in a theological context for example, that something exists and what would be the implications of saying that it did not exist? This is not a topic, as one might expect, that fits easily inside a fortune cookie. But, since its resolution or at least discussion is central to rigorous philosophical discussions of god’s existence, it is fruitful to give some flavor here of the issues involved.

Another minimal criterion for existence is that something exists if we can talk about it. More precisely, this minimal criterion states that, for some variable placeholder, x, and for any constant, F, that stands for a property, the proposition “Some x is F” and the proposition “There is an x that is F” are equivalent. In other words, there is no difference in meaning between saying, “Among all the marbles in the box, one of them is green” and saying, “There is definitely a green marble in the box.” At first blush this appears to be wildly counterintuitive. The box may be empty. More tellingly, it posits the equivalence of the following two propositions, “Hamlet is a procrastinator” and “There is someone in the world who is Hamlet and that person is a procrastinator.” Nevertheless in some contexts, this criterion is justified, or at least harmless. For example, an astronomer probably means the same thing if he says “Something is pinging us from Alpha Centauri” and if he says, “There definitely exists something in Alpha Centauri and that thing is pinging us.” This minimal criterion of existence was proposed by Frege who in addition made clear, in a revealing dialogue he had with the German theologian Pünjer, that he felt it was also an exhaustive definition of the concept of existence. In technical terms, the fact that you can quantify over something or a class of things is a sufficient condition for the existence of that thing or class of things. For Frege this assumption was important in his rejection of verifiability as a separate criterion of existence and in his defense of the “existence” of whole numbers. Many have followed Frege in this belief. Gödel felt that his proof that a sufficiently rich syntactically defined logistic system did not generate all the tautologies of a semantic interpretation of that system was a proof of the real existence of some numbers. Meinong, I understand, let it all hang out. As far as he was concerned, everything we can talk about exists and that’s that. Many phenomenologists do not endorse Frege so much as they ignore issues of existence. Whether a thing exists or not is not as important as what kind of being it has, what is it such that without it a thing would not be what it is. One criticism of Frege is that he exposes the vulnerability of his criterion by insisting that any additional criterion of existence presupposes the existence of things that do not satisfy that additional criterion. His argument is not so much a triumph over the principle of perceivable verification as a sign of the inadequacy of his definition. In partial defense of Frege it may be observed that an adequate semantics of a logistic system seems to require it. If there did not exist some entity quantified over in an existential proposition, then that proposition would not have a truth value in a semantics with two values. One criticism of the phenomenologists is that they conflate existence (in the sense of “Does Neptune exist?”) with essential property. This is particularly glaring in Heidegger who does not resort to a theory of mind as a framework for understanding how objects are “constituted.” G.E. Moore once proposed a largely forgotten mixed reaction. In typical fashion he first presented the arguments why something we talk about or imagine must exist (His reason is that there is a difference between imagining a centaur and imagining nothing) and then countered that this led to the absurdity of assuming that things that don’t exist do exist. His solution was to distinguish between “has being” and “is real,” that is between “being” and existence. Everything we can talk about has being, but only some things are real or have existence. Some such distinction must be made and it is to Moore’s credit that he saw the need. But he doesn’t go a whole lot further. For one thing his revised sense of being seems to be exactly equivalent to “being talked about.” Yet the argument that we can’t talk about something that isn’t seems to imply something more. Similarly he doesn’t address what it means to exist in his sense. Famously he talks about sense data like there was no tomorrow. More critically he admits that properties and relations exist, but his assumption that those entities exist seems to be based on nothing more than the fact that we can talk about them - which is the same as the criterion for something having “being.” Moore’s favorite choice of words for a thing’s existing is that it is “in the universe.” I can see nothing more in that than a lexical equivalent. Moore made an important distinction, but he didn’t explore the consequences.

The Fregean analysis of existence seems to imply that real world issues of existence simply don’t, well, exist. Does the hot broad I see depicted on my TV screen really exist or not? Does the Loch Ness monster exist? Does the tenth planet from the sun exist? Does ether exist? Does the unconscious exist?  What do we mean when we say Hamlet doesn’t really exist? We deal with issues like these every day and the answers are often of significance for our actions and the way we lead our lives. And, of course, theory confirmation in the sciences often hinges on questions of the existence of particulars. This appropriate or discriminatory sense of existence is not limited to single individuals. The general might want to know whether the hostile cavalry that his spies warn are hiding in the woods really do exist. But aside from groups that are members of a class whose existence is admitted (The enemy cavalry belongs, after all, to the class of all cavalries, some of which really do exist), there are entire classes about which we can wonder whether they are populated or not. Does anything of the sort exist? For example, do leprechauns exist? Or Martians. Or feelings, and if so in what sense? Once again science deals constantly with these sorts of questions when a certain class of entity is proposed for its role in an explanatory theory. Do gluons exist? Physics struggles with the verifiability of entities that cannot be directly observed. Indeed evolutionary biology must deal with the search for the evidence of phenotype links demanded by a theory.

Another reaction has been that the Frege criterion is wrong or at least inadequate. The verification principle and its variants as proposed by the logical positivists is nothing more, when the dust settles, than a flat rejection of Frege. One criticism of the verification principle has been that it is too restrictive. Another is that the principle itself cannot be verified (more accurately its own verification is inductive and begs the question). Most importantly it raises significant issues about the nature of verification, perception, translatability and meaning – issues that it supposedly laid to rest (If a Martian scientist in translation informs us he has verified the existence of Venusians, how do we know he means the same thing by his words as we do?). In my opinion, for whatever that’s worth, Frege’s criterion is valid in one sense. But what it really establishes is that ontological issues (The meaning of the term “existence”) cannot be decided by a logistic system or even by a formal semantics of a logistic system.  Moreover, one important sense of “existence” – the sense we use when we wonder if a thing exists or not - cannot be defined by a logistic system or by a formal semantics of a logistic system. The logical sense is non-discriminatory, to coin a phrase. Obviously any non-discriminatory sense of existence is unhelpful in deciding whether or not a creator god exists.  

Clearly the logical sense of existence doesn’t say much. God exists but so do talking toads, unrepentant jim jams, and sadly enough a Jenna Jameson who never had her breasts reduced. Presumably Xtians, Mahometans and the rest of the foul crew would be less than thoroughly satisfied with a god that exists because everything exists, that had to share ontological digs with a Jenna that never was or thoughts I never had. Without a bit of fancy footwork, the logical sense of existence could lead to the conclusion that nothing (or Nothing) exists. So what is the sense favored by religionists when they talk about an existing god? To date there hasn’t been a really clear answer to that question; in fact outside the cloisters of Dark Ages philosophy I can’t think of much of an attempt to address it. Your common or garden Xtian works mostly with common or garden notions and so behaves consistently in the case of “exists,” instances of whose everyday use are not that frequent and very possibly mutually unrelated. The physical sciences, and by extension some of the daughter sciences, operate with a rather more rigorous and not entirely unPunjerian concept of “exists.” This concept is discriminatory in that it allows us to distinguish between entities that do and do not exist. Something exists if it is observed (a term that is admittedly almost as troublesome as “exists”) or if its existence is required by a true model. Subatomic strings would exist only in the sense that we could observe them or their effects or if their presence is required by string theory and some version of string theory is true. If something doesn’t fit these criteria, then by Ockham’s Razor it doesn’t exist. Obviously by these criteria god doesn’t have a chance since it has not been observed by any non lunatic and its existence is not required by any physical theory. Dawkins presumably works with the latter concept of “exists” although he doesn’t really get much chance to put it into play since biology doesn’t deal with the arcane near unobservable entities beloved to physicists. (Evolutionary biology does deal with hypostatized events, however, the criteria for whose “existence” if you will don’t pose too many problems at least of a philosophical sort.) Dawkins’ discovery is that, to the extent that God’s existence fits in any of these appropriate or discriminatory senses of existence, it is considerably less likely that God exists than, say, the Loch Ness monster.

Dawkins does wander off the reservation of the discriminatory concept of existence only, and then perforce, in his direct confrontations with some theologians whose notions are neither everyday nor scientific, nor in the end very, shall we say, clear or distinct. Sometimes when you play the home team you have to play by their rules and Dawkins does a masterful job of tying them into knots with their own rules. A more rigorous treatment, however, would need at the very least the added demonstration that the home team’s rules are the purest nonsense.

“All these fancy intellectuals with their logical distinctions really just miss what is front of their own very own noses. I know that I love the Baby Jesus and that I talk to him.” Fuck the Baby Jesus.