One of the questions that theism must answer is how it is that there can be any discourse about God, a being who transcends the realm of human experience and concepts. There are really two questions here. First, can we say anything meaningful about God? Secondly, if we can say anything meaningful about God, can we say anything literally true about God? The "verificationists" of the first half of the 20^th century denied that former, and many religious people deny the latter. Many religious people, though they believe that human language can speak meaningfully about God, deny that anything in theological language is literally true about God.

Below I provide a sketch of these two negative views of religious language and provide responses to each.

Theism, religious and philosophical, makes assertions like God exists, God is all-powerful, God is an immaterial substance, God knows, God creates, and so on. Now look at what I have just written. What you see on the page are sentences. By a "sentence" I will understand a series of words that are connected together according to rules of grammar and syntax. A meaningful sentence is one that expresses something. Meaningful "declarative" sentences typically make claims about how things are (as opposed to commands, questions, and exclamations that have a different function). A meaningful declarative sentence expresses what philosophers call a proposition or statement. Different sentences can of course express the same proposition. Rex est martuus and the King is dead are two different sentences but they have the same meaning, that is they express the same proposition. Unlike commands and questions, propositions are either true or false.

We also recognize that some propositions are coherent and others are not. A coherent proposition is one that we can conceive or make sense of being true, even if in fact it is not true. By "conceive" I don’t mean forming a mental picture of what is described, but simply being able to think of the proposition as being true or understanding the conditions under which it would be true. In other words, a coherent statement is one that we can understand what it would be like for it to be true. The moon is made of green cheese is false, but it is coherent. Although most of us know that this proposition is false, we know what it would be like for it to be true and we understand that. The lunar body that orbits the earth would be composed of pressed curds of milk, perhaps along with some dye or mould to give it the color green. An incoherent proposition, then, is one that we cannot understand what it would be like for it to be true. Justice smells like dog hair, squares have five sides, and Elvis is dead and Elvis is not dead are all incoherent propositions. The last example here is a self-contradictory statement (and the second to the last a little less obviously so). When one proposition says something is the case and another says that it is not the case, the two propositions are contradictory. All self-contradictory statements are incoherent, for such statements cannot possibly be true. But the first example isn’t as clearly self-contradictory. The contradiction is buried within it (in that justice has no smell). There are plenty of incoherent propositions that are not obviously self-contradictory, but are nonetheless incoherent.

In the early 1920s several philosophers in Vienna, influenced by the work of Ludwig Wittgenstein (1889-1951), began to work on a criterion for meaningful sentences. What is required for a declarative sentence to be meaningful? The Vienna Circle, as this group of philosophers came to be known, developed a theory of meaning known as verificationism. This principle became the central doctrine of the logical positivists – the name of the philosophical movement that originated with the Vienna Circle. The verificationist theory of meaning had a tremendous impact on theism in the first half of this century, the effects of which linger to this day.

Verificationism grouped meaningful declarative sentences into two categories. First, there are sentences like 2 + 3 = 5, all bachelors are unmarried men, and not both A and not-A (where "A" is some term). The first sentence expresses a mathematical proposition, the second a proposition true by definition (a so-called tautology), and the third a proposition of logic. All of these are, to use language from above, (logically) necessary propositions. In contrast, there are factual propositions, or what above I referred to as logically possible propositions. Verificationism is mainly a theory about the meaningfulness of factual sentences. It held that a sentence is factual and meaningful only if it can be verified by empirical observation (i.e., can be seen, smelt, touched, or heard). In other words, meaning is connected what can be verified by the senses (such as done in the natural and physical sciences).

Central to verificationism is the idea of an observation-statement, a statement that reports a sensory observation which could in principle be made. It is snowing outside or there are nine planets in our solar system are both observation statements, since they each report observations which, at least in principle, could be made, even if for practical reasons they cannot (at some given time and place). A more precise version of the verificationist principle can be stated as: a declarative sentence is factual and meaningful if and only if it can in principle be verified by an observation-statement. The kind of verification intended here is "conclusive." So if the appropriate observation could be made, then the observation would conclusively verify the sentence. This is the strong verificationist principle.

A challenge to theism follows from each of these versions of verificationism. In a way the challenge is quite obvious. Sentences like God exists and God created the universe are not mathematical propositions, nor are they quite like the principles of logic. Some theistic propositions, like if God exists, then God is eternal or if God exists, then there exists at least one immaterial substance are logically necessary because they are true by definition. In that way they are like the proposition if John is a bachelor, then John is an unmarried male. Nothing follows as to whether John exists. Some philosophers, Descartes and Leibniz for instance, have thought that God exists is logically necessary (because the existence of God is contained in some way in the concept of God). These moves would not impress the verificationist, for on his view no factual proposition can be derived from what is logically necesssary. You simply cannot define things into existence. From a tautology, only a tautology follows. Factual propositions are distinct from tautologies.

Talk about God, though, does seem to involve factual claims. But verificationism appears to render talk about a non-observational, immaterial being, quite problematic, to say the least. Such a being could not, even in principle, be verified. What observation-statement provides conclusive verification that God exists or God created the universe? It is not as though one day we will build a super telescope that will able us to see God, even if at present our resources are a bit limited due to cut backs in the space program. Suppose we could travel back to the beginning of time, the moment of the Big Bang. Would we see God? Probably not. What observation could there be that would even make it probable that God exists? Would the observation-statement describing the expansion of the Universe in its first second of existence raise the probability that God exists or that the created the world?

It follows that God exists is devoid of factual meaning. Notice that the verificationist is not saying that God exists is false. He is saying that it is neither true nor false. Of course an interesting implication of this is that the denial of God’s existence is also meaningless. If a sentence like God exists is meaningless, so is its negation (only what is meaningful can be meaningfully denied). Also, the agnostic who claims not to know whether theism is true recognizes that either there is a God or there is no God (we just can’t which is true). But here again, sentences affirming and denying God’s existence are being used, but no such sentences are factually meaningful.

In his Language, Truth, and Logic, A.J. Ayer (who introduced logical positivism to Great Britain) wrote:

Anthony Flew, another verificationist, develops the anti-theistic case in a slightly different way. His concern is that theism cannot be falsified, in the sense that no empirical observation can count against theism. But this is a requirement for meaningful factual sentences, even on the weak verificationist principle. In "Theology and Falsification" Flew provides a parable to reveal the factual empitness of theism. Two explorers come upon a clearing in the jungle where there is a patch of flowers growing mixed with many weeds. One of the explorers says that there is an invisible gardener who tends to this garden. The other explorer denies this. Various experiments are set up to try to establish the existence of the invisible gardener (e.g., electric fences and blood hounds). None of the tests produce a shred of evidence that the invisible gardener exists. The believer continues to hold his belief, affirming that the invisible gardener makes no sound, has no scent, and he can pass through electrical fences with ease. The unbeliever responds that there really is no difference between this invisible, immaterial gardener and no gardener at all.

Flew’s point in adducing this parable is to show that the refusal of the theist to entertain the possibility of evidence against theism (like the gardener believer in the parable) shows us that the theist’s belief in God is devoid of factual meaning. He says that "sophisticated religious people . . . tend to refuse to allow, not merely anything actually does occur, but that anything conceivably could occur, which would count against their theological assertions" (p.21). But this shows their assertions to not be genuine assertions at all. A meaningful sentence or assertion must deny something if it is to assert anything. But in that case, there will always be some conceivable event or series of events that are inconsistent with any given assertion. If theistic assertions have any factual meaning, then there are (conceivable) events that would count against them. The theist’s refusal to give up theistic belief shows that he does not recognize any such possibility. He thereby shows that his assertions are not genuine assertions at all. They are without factual meaning.

Verificationism, to put it quite pointedly, is a snare and a delusion.

First, theism has nothing to worry about regarding the strong formulations of verificationism, as these principles are intrinsically implausible. A very large number of scientific and common sense statements would be meaningless if either of these principles were true, including many statements that logical positivists themselves accept as meaningful. This highlights a general problem with all versions of verificationism. They have not been able consistently to include many of the statements logical positivists would like to be meaningful and also exclude the sentences of metaphysics and theology.

More importantly, what is the status of the sentences that express the verificationist/falsificationist principles? That is to say, what is their status given what they each claim about meaningful sentences? Given the criteria for meaningfulness that they affirm, are the sentences that express such criteria meaningful themselves? The principles are rather obviously not mathematical or logical assertions, nor is either true by definition. What plausible analysis of the content of "factually meaningful sentences" commits us to "conclusively verifiable or falsifiable by empirical experience"? And if the principles are being asserted as definitions, why should the theist be bound to accept that definition? It seems that the principles in question must be factual, but then it is necessary that they be verifiable or falsifiable. And what argument is there for that? There isn’t must to go on there. We can’t even appeal to examples of what people regard as factual, for this could not conclusively verify verificationism, for – among other things - too many of those examples of allegedly factual statements are not verifiable or falsifiable. Of course, this point might provide an argument for the conclusive falsification of the strong verificationist principle. If any examples of non-conclusively verifiable sentences are factually meaningful, then it is not the case that all factually meaningful sentences are conclusively verifiable. Verificationism would then be meaningful, but simply false. On its own terms, though, it seems that the fate of verificationism would be the same as theistic sentences, for it does not appear that verificationism is conclusively verifiable by any observation-statement.

So we are really only left with the weak verificationist/falsificationist position. And here the theist can respond several ways. First, it seems that theism can respond to Flew’s argument head-on. Is it really that case that the theist admits of no "conceivable" events or states of affairs that would count against theism? Perhaps some do, but this isn’t necessary. Flew, though, seems to jump to this conclusion from the fact that theists exemplify tenacity of belief. Theists hold their beliefs come hell or high water. The conclusion clearly does not follow though. Flew’s parable presents us not merely with someone who holds his Gardener belief come what may, but someone who denies that anything can count against his belief. The theist can happily affirm that there are many things which theism denies, and if these turned out true, there would be evidence against the truth of theism. If this were not so, it is quite difficult to explain the thousands upon thousands of pages that have written by philosophers and theologians since the time of St. Augustine regarding the problem of evil (to be discussed in a later chapter). The fact that theists have typically taken the problem of evil seriously shows that they do most certainly recognize some things that could conceivably occur that would count against theism. Presumably, if God is good, and goodness has moral implications, then there are limits to the degree or kind of evil God could allow. In fact, a believer might be troubled by the amount apparently pointless evil that he sees in the world. He might even recognize this as counting against theism. So what? There are plenty of beliefs that we continue to hold even when we recognize that there is evidence that counts against those beliefs. Why should this render those beliefs meaningless? After all, one can recognize that something counts against one’s beliefs without agreeing that it counts decisively against those beliefs. (See Basil Mitchell (1917- ).

Secondly, why suppose that theism is not in principle verifiable or susceptible to confirmation by any observation-statement? Part of the problem here is that it isn’t clear on where to draw the line separating events that are in principle observable and those that are not. The set of all observation-statements that could be made by some observer is tricky indeed, as people have claimed to have observed all sorts of things (e.g., angels, demons, UFO’s, aliens, the future, great sea monsters, alternative dimensions). Although some have denied that these are genuine observations, what could be advanced against such claims? Before the microscope and telescope were invented, people could not observe very small and very distant things, though such observations were logically possible. There was nothing self-contradictory about a human making such observations. With the aid of the Hubble Telescope and planetary probes more and more regions and details of the Universe are becoming observable, and with this more and more kinds of statements can be confirmed. Setting limits then on the range of observation-statements and what they can confirm seems hopelessly arbitrary.

The only non-arbitrary limit on what is in principle observable would seem to be dictated by what it is coherent to suppose can be observed, what does not involve contradiction. But then for most alleged observations we would show them to be "observable in principle" by showing that they are simply logically possible. But then it is quite difficult to rule out observation-statements that would confirm or verify God’s existence. It is logically possible, after all, that Paul observed things which cannot be uttered when he was taken up to the seventh heaven, or that Moses spoke with God at the Burning Bush. It is possible that one day we will discover life on other planets, and eventually unlock a universal DNA code that delivers a message from God (as an episode of Star Trek: the Next Generation highlighted). In fact, since what is observable in principle cannot be restricted to what can be observed by humans at the present time, it is possible that an afterlife will provide humans with observation statements that confirm many theistic statements. Theistic sentences, because they describe what could be confirmed by yet future observation-statements, are not meaningless. It is simply not possible to show that theistic sentences cannot be confirmed or verified by any observation-statement, since we do not have a complete or even adequate knowledge of all observation-statements.

The best case in favor of the weak verificatonist/falsificationist principle is that the examination of ordinary examples of what people take to be factual sentences can be confirmed or falsified by some observation-statement. The problem here, though, is that there are plenty of people who regard certain sentences as meaningful that are not confirmable or disconfirmable. At best, then we can say that if a sentence is verifiable or can be confirmed/disconfirmed by an observation statement, then it is factually meaningful. But then meaningfulness would not require empirical evidence, so even if there is none for theism, it would not on that account be meaningless. There is no successful verificationist objection to talk about God.

The verificationist objection to talking meaningfully about God is not typically advocated by religious people. Most religious people think that theological discourse is meaningful. The central question about God-talk for religious people is whether talk about God should be construed as literally true. Some think not. They are what we might call theological pessimists. They are pessimistic about the adequacy of human language to refer to or describe God. There is also a tight connection between this negative assessment of theological language and religious epistemology. The idea that human language is inadequate to talk about God is intimately connected to the epistemological view that human reason cannot know anything about God, that is, anything that is literally true about God.

Some religious people argue that that there can be no knowledge about God acquired by the human mind because the human mind operates on the basis of distinctly human concepts, and such concepts could not apply to God. They have been developed or designed to apply to things in the world of physical or finite things, but God is infinite, transcendent, and ultimate, most certainly not a thing in the world. Hence, if there is a God, we could not know by reason that he exists, or even refer to him or talk about him. In fact, on this view, it is not clear that we could know anything literally true about God even on the basis of a divine revelation from God.

What can be said about this line of argument?

Well, it is certainly true that God is not a thing in the world. Perhaps it is also true that human concepts have been designed to apply to things in the word. Would it follow, though, that human concepts could not apply to God? Consider this argument. Teflon was designed to apply to problems in the space program, but how to fry a good omlet or cook a decent batch of bacon are not problems in the space program. Hence, Telfon cannot apply to the problem of cooking eggs or bacon. Not a particularly impressive argument, is it? But its form is the same as the theological pessmist's argument above. Neither argument is valid. Even if it were true that <human concepts are designed to apply to things in the world> and <God is not a thing in the world> it would not logically follow that <human concepts cannot apply to God>. The argument could be made valid if we were to add the following premise: if one thing, X, was not developed or designed to apply to something else, Y, X cannot be applied to Y. If this is true, then the conclusion would follow. But this added premise is obviously not true, as the omlet and bacon example shows.

Clearly, a crucial question for the theological pessimist is whether language and knowledge, rooted in experience, can apply to things beyond experience. The work of the British philosopher David Hume (1711-1776) and the German philosopher Immanuel Kant (1724-1804) is central in this regard. Each in his own way excluded God from the realm of reason or human knowledge by first restricting reason and knowledge to experience, to what can be observed (i.e., seen, smelt, touched, or heard), and then excluding any path from experience to God. Given these philosophical assumptions, no unobservable entity can legitimately be used as an explanation for observable events. Language and human concepts would be limited to experience, and you could never infer what is beyond experience from experience itself. But as noted in the last chapter this is precisely what natural theology claims to do. But the Humean-Kantian restriction knowledge to experience, and of causes of observable effects to causes that are themselves observable, seems implausible. The big bang, black holes, the inside of a star, quarks, leptons, etc are not observable events or entities, but they are rather central to contemporary physics and cosmology. The idea that language and human concepts are developed only for the purpose of encoding previous experience of the physical world seems profoundly mistaken.

The theological pessimist emphasizes the fact that God is unlike anything in the world. But what does this mean? Is God not completely like everything in the world of human experience? Or he is completely unlike everything in the world of human experience? If God is not completely like everything in the world of human experience, then God does share all properties or characteristics with anything else. But if God is completely unlike everything in the world of human experience, then God does not share any properties or characteristics with anything else. It is certainly true that God does not share all properties with anything else, but it's hard to see how God could not share any properties with anything else. If God had nothing in common with anything else, God would have at least one thing in common with everything else, namely the property of "having nothing in common" (as this would be a property of both God and everything else).

This last point actually introduces the most obvious weakness of theological pessimism. It is a philosophical horse that pulls up lame right out of the gate. The idea that no human concepts are applicable to God is self-referentially incoherent. For if it is true that no human concepts apply to God, then at least one human concept must apply to God, namely the concept of "a being who is such that human concepts don't apply to it." The same kind of incoherence lurks with the use of human concepts such as infinity and transcendence as grounds for affirming that no human concept could apply to God. Hence, the theological pessimist's argument really undermines itself, as it presupposes that some concepts are applicable to God, namely infinity and transcendence. In short, if reason cannot uncover any truth about God, it could not discover the truth that "human reason cannot uncover any truth about God." Since any proposition or statement that logically entails a contradiction must be false, extreme theological pessimism must be false. There is no good objection to natural theology to be found in an argument that is self-contradictory.

The theological pessimist position considered above is clearly self-referentially incoherent. The natural way to avoid this incoherence problem is to maintain that some human concepts are applicable to God. A theological pessimist could then limit the scope of the applicable human concepts to some bare minimum, perhaps just to no human concept, other than the inapplicability of human concepts, applies to God. Well, actually, he could do more than this. The theological pessimist could deny the applicability of so-called substantive concepts such as goodness, knowledge, and power in favor of what we might call positive but mere formal concepts or properties. One way of understanding formal properties is in terms of logical necessity. These would be properties that everything must necessarily have. These would include properties like having some properties, being self-identical, sharing some properties with other things, and being the possible referent of human reference. We might call this watered down pessimism, moderate theological pessimism, as opposed to the radical form discussed above.

But wouldn't the moderate theological pessimist already have to have some substantive or non-formal knowledge of God, or way of referring to God, in order to deny that non-formal human concepts are applicable to God? And wouldn't he thereby be caught in a similar self-defeating position as the radical pessimist? Well, a theological pessimist could willingly concede that he must presuppose some non-formal truth about God. In fact, the case for theological pessimism is often based on the premise that God is infinite, transcendent, and ultimate. Since not all things have these properties, much less necessarily, they are not formal properties. But the pessimist will deny that these non-formal properties are positive. They are negative properties. Being infinite simply means not being finite, not being limited. Being transcendent simply means being not being a thing in the physical world. Being ultimate simply means being not relative, or being not dependent on anything else. Philosophical theologians have argued similar points with respect to properties such as timelessness (as meaning not being in time), immutability (as meaning not being subject to change), and immateriality or spirit (as meaning not being material). The theological pessimist can easily advocate this so-called via negativa. The moderate position, then, can be that the only positive properties that apply to God are formal ones, and the only non-formal properties that apply to God are negative ones.

What are we to make of this?

Although one might respond to the moderate pessimist by arguing that natural theology could be developed solely on the basis of predicting negative qualities to God, I think the more important point to see is an interesting difficulty with any purely negative theology (with or without the positive, formal property addition). First, it is difficult to see how one could conclude that there is a God (a seemingly positive statement) if limited to negative properties. If "existence" is a property, then it surely is a positive one. More importantly, though, some negative properties logically imply positive ones. An ultimate being, for instance, is not dependent on anything else for its existence. But ultimately also entails self-sufficiency, and this certainly appears to be a positive property. An infinite being is a being that is not limited. But this negative property seems to imply other properties that are not merely negative. For instance, wouldn't an unlimited being be present everywhere? Omnipresence is a positive property though. And wouldn't an unlimited being be unlimited in power and in knowledge? But this entails being able to do whatever one wills to do and having all knowledge. Omnipotence and omniscience are positive properties.

Now certainly the moderate theological pessimist can avoid talking about God as an ultimate or unlimited being, but the fact is that most don't. John Hick, for instance, sees this way of referring to God as foundational to religion. More crucially, though, it is precisely because theological pessimists conceive of God as wholly other than humans, as ultimate or infinite, that they maintain the inadequacy of human (positive non-formal) concepts to apply to such a being. So the argument here is not that some negative properties imply positive ones, but that (i) there are some negative properties that are essential to the pessimist's case against knowing non-formal positive truths about God and (ii) these negative properties imply positive ones. Hence, it seems that a purely negative theology is not possible, at least not if negative properties such as infinity and absolute independence are essential to that project.

We began with the general pessimist objection that human reason could not know any truth about God. The reason for this objection was that human concepts are limited in their applicability to God, either they don't apply to God at all or their application will be the predication of either negative properties or merely formal positive properties to God. I have argued that the radical pessimist view is necessarily false and hence does not provide adequate support for the claim that human reason could not know any truth about God. Secondly, the moderate position seems unable consistently to avoid the predication of non-formal positive properties to God. In this case also, there is no adequate support for the contention that human reason could not know any truth about God. Negative truth about God is truth about God nonetheless, and if some essential negative divine properties entail positive ones, a purely negative theology is probably impossible.

© Michael Sudduth 1999