Is belief in God unjustified?
As a part of my recent philosophical wanderings, I'm reading Kai Nielsen's 1985 book Philosophy & Atheism. He wants to show that belief in God is unjustified.
This is the second post in the series -- I encourage you to read Part I if you haven't yet.
In defense of atheism
In Part I, I introduced Nielsen's primary argument for atheism, namely, that sentences used to talk about God are at their core incoherent. The word "God" fails to have an intelligible referent. We are not justified in believing in incoherent things.
In the chapter "In Defense of Atheism," he elaborates on this argument.
First, Nielsen notes that religious discourse tends to reflect particular human commitments and attitudes -- e.g., the feeling of gratitude for one's existence, regardless of the quality of that existence. Two example utterances he puts forth are "God is my Creator to whom everything is owed," and "God is the God of mercy of Whose forgiveness I stand in need."
Fact-stating utterances of this sort always presume some background knowledge. Every utterance exists in a larger context.
"Take the classic example 'The King of France is bald.' We need a context, an application of the Principles of Relevance and the Presumption of Knowledge, to know how to take it. If our context is the present, and the relevant questions are 'What is the King of France like?' or 'Is he bald?' then neither 'The King of France is bald' nor 'The King of France is not bald' would be a correct answer, for the above questions in the above context are not to be answered, but are to be replied to by being rejected. The proper reply--a reply which rejects such questions--is (De Gaulle notwithstanding) 'There is no King of France.' But if our topic is historical and, with some specific period in mind, we are asking 'What bald notables are there?'; 'The King of France is bald' is in such a changed context is an appropriate answer. And here it is a true or false statement." (p. 79)
In statements about the King of France or about God, "the King of France" and "God" are referring expressions -- i.e., there is the assumption that these expressions refer to something that exists. Additionally, there is the presumption that the speaker understands and believes in the reality of the thing being spoken about. In asserting the example utterances I mentioned above, "the religious man presupposes that there is a God and that this God has a certain character. The atheist, on the other hand, does not believe [the utterances] are true because he does not accept the presupposition on which they are made" (p. 79).
The point Nielsen is making here is that we cannot evaluate religious utterances in isolation from the complex activity we refer to as "religion."
Next, Nielsen wants us to recognize that when using expressions in language that refer to existents of some kind, one needs to know how the referring is to be done -- can one point at the entity? Can one identify it indirectly? His claim is that "the concept of God is so incoherent that there could not possibly be a referent for the word 'God'" (p.82). He refers specifically to the non-anthropomorphic Judeo-Christian-Islamic God. Because the concept is so incoherent, he argues, it cannot possibly be true -- the rational thing to do is to reject belief in that God.
So how does he support this claim?
First, he notes that even if incoherent, "God" is not utterly meaningless. There are pieces of what he terms 'God-talk' that are deviant and pieces that are non-deviant -- e.g., commonly accepted uses are such statements as "God so loved mankind that he gave to the world his only son" and "God protect me in my need," while deviant statements include such statements as "God lost weight last week" and "God brews good coffee."
Second, Nielsen clarifies what he means by incoherency.
"[I]n saying that the concept of God is incoherent, I am saying that where 'God' is used nonanthropomorphically, as it is in at least officially developed Jewish and Christian God-talk, there occur sentences such as [the utterances mentioned earlier] which purportedly have a statement-making function, yet no identifiable state of affairs can be characterized which would make such putative religious statements true and no intelligible directions have been given for identifying the supposed referent for the word 'God'" (p. 83).
God cannot be physically pointed to in the world the way we can point to chairs and instances of green things. If God can be identified, it must intra-linguistically. But as mentioned in Part I, what does it mean for a thing to transcend the world, be an ultimate reality, or be an infinite individual? "If in trying to identify God we speak of 'that being upon whom the world can be felt to be utterly dependent' nothing has been accomplished, for what does it mean to speak of 'the world (the universe) as being utterly dependent' or even dependent at all?" (p. 83). If we are puzzled by "God", says Nielsen, we will be equally puzzled by these kinds of descriptive phrases. We know what it means to say that children or nations or lakes are dependent on other things, but we have no sense of what it would mean for the universe to be dependent on something.
He continues to discuss the dependent universe example:
"What are the sufficient conditions for the universe being dependent? What would make it true or false or what would even count for the truth or falsity of the putative statement 'The universe is dependent' or 'The universe is not dependent?' To answer by speaking of 'God,' e.g., the universe is dependent because God is its final cause, is to pull oneself up by one's own bootstraps, for talk of the dependency of the universe was appealed to in the first place in order to enable us to identify the alleged reference of 'God'. And to speak of a logically necessary being upon whom the universe depends is to appeal to a self-contradictory conception, for only propositions or statements, not beings, can either be logically necessary or fail to be logically necessary. Yet to speak of a 'factually necessary being' upon whom the universe depends is again to pull oneself up by one's own bootstraps; for what would count toward establishing the truth or falsity of a statement asserting or denying the existence of such an alleged reality?" (pp. 83-84).
If God exists, he somehow exists necessarily. But if the concept of a logically necessary being is self-contradictory, then it cannot be true that any being must exist simply because its existence is logically necessary. Nor, argues Nielsen, is there sense in the claim that there is anything which categorically must exist.
The problem is in determining what the referent of "God" is. The problem is that the phrases used to describe God (e.g., "a self-existent being," "a self-caused being") have the same problem of purportedly being referring expressions, namely, that there is no way of discovering their referents.
Another point he makes is that perhaps believers feel that they are in the presence of an "ultimate reality" that is taken to be God. But if this is to be taken as a nonanthropomorphic God and transcendent to the world, "it should still be evident that 'a transcendent X' could not be 'an X whose presence was given in experience.' Something given in experience would eo ipso be nontranscendent, for it would automatically be part of the spatio-temporal world" (p. 84).
It's possible that some people will argue that Nielsen is assuming too much -- some experiences, particularly experiences of God, may not be materially grounded in the spatio-temporal world. Philosophers have certainly posited that thought may have immaterial aspects. Could some kind of immaterial thought account for experience of God? Perhaps so.
The crux of Nielsen's argument
Nielsen's argument rests on this fact: If there is religious truth, the statements expressing those religious beliefs must be true. If he can show that the statements are indeterminate and incoherent, then, he says, there is nothing in those statements that could constitute something true or false.
One example Nielsen focuses on is what it means for God to act. In the case of a statement such as "God is the God of mercy of Whose forgiveness I stand in need," further statements are entailed: that God does or can do things and that God acts or can act in certain ways. "[I]t is utterly senseless to speak of being merciful if one could not even in principle act, do or fail to do merciful acts" (p. 86). Nielsen says that anyone, not just reductionists and materialists, can recognize the truth of this. To say that a being acted mercifully implies that the being acted; to act implies that the being acting is an agent that can perform actions. This may seem trivial; however, Nielsen argues that even if one allows for bodiless action, e.g., as in cases of chemical agents or forces producing effects, "there is still a physically specifiable something which reacts in a determinate physically specifiable way" (p. 88). If God is indeed non-anthropomorphic, realized as Pure Spirit, not a reality with a body or a spatio-temporal location, how does God act? We have no idea, says Nielsen, "of what it would be like for something to be done, for something to do something, for an action to occur, without there being a body in motion" (p. 88).
In saying that God can act, God is conceived of as being able to do things. But we can only understand doing things when there is something identifiable doing the doing.
"X is only identifiable as an agent, and thus X an only be intelligibly said to be an agent if X has a body. For agency to be logically possible, we must have a discrete something specifiable in spatio-temporal terms. But the transcendent God of Judaism and Christianity is thought to be a wholly independent reality, wholly other than the world which is utterly dependent on this 'ultimate reality' and is said to be ultimately unintelligible without reference to this nonphysical mysterium tremendum et fascinans" (p.88).
But then it is senseless to speak of God as an agent who acts, and sentences about God are therefore incoherent.
One might argue, here, that utterances like "God loves all His creation" and "God is all merciful" are symbolic or metaphorical, but are not themselves literal true/false statements. Nielsen remarks that some theologians will then recourse to describing God as "Being-itself" or "the source and unity of all beings," but the problem remains the same -- what is the referent of "Being" or "Being-itself"? For a statement such as "God is not a being, but rather Being-itself within which all other beings have their being" to be be intelligible, then "being-itself" must be a genuine referring expression. But it is not, for much the same reasons why "God" in the utterances discussed above is not.
Nielsen notes that some people will now bring up ineffability -- that there are ineffable truths that cannot be put into words and religious truths are of this category. The first part he agrees with; there are certainly "some things which are literally unsayable or inexpressible but are nonetheless given in those experiences of depth where human beings must confront their own existence" (p. 91). The second part he claims is incoherent.
First, if one claims that religious truths are ineffable, then some people with the proper experience can in a sense understand the concept of God but cannot literally express what they know to be true. Statements about the concept of God are not true or false statements; they merely hint at what cannot be literally stated. No sentences about God can literally express facts or assert that certain things are true or false, though they could be sensical, given their metaphorical or symbolic use.
"But if an utterance P is metaphorical, this entails that it is logically possible for there to be some literal statement G which has the same conceptual content. 'Metaphorical,' for that matter 'symbolic' or 'analogical,' gets its meaning by being constrastable with 'literal.' There can be no intelligible metaphorical or symbolic or analogical God-talk if there can be no literal God-talk. Thus the ineffability thesis is internally incoherent." (pp. 91-92)
Furthermore, Nielsen argues that if knows something that is literally inexpressible, then trivially, one cannot communicate it. One cannot be justified in saying that it is, in fact, God you experience, know, or encounter because one cannot significantly say that if one does particular acts or has particular experiences, one will come to know God. If one says that God cannot be described, then the word "God" is meaningless -- "we cannot even say that something is if it is indescribable" (p. 93).
"'What is unsayable is unsayable,' is a significant tautology. Only if one could at least obliquely or metaphorically express one's experience of the Divine could one's God-talk have any significant, but on the present radical ineffability thesis even the possibility of obliquely expressing one's knowledge or belief is ruled out. So, given such a thesis, there could be no confessional community or circle of faith; in fine, the thesis is reduced to the absurd by making it impossible for those who accept such a thesis to acknowledge the manifest truth that the Judeo-Christian religion is a social reality. On this simple consideration alone, we should surely rule out the ineffability thesis." (p. 93)
We don't understand the concept of God
Here, Nielsen reminds us that what he wants is for the believer to show how God-talk is a coherent form of language. "Faith presupposes a minimal understanding of what you take on faith, and if my arguments are correct, we do not have that understanding of a nonathropomorphic concept of God" (p. 94).
He acknowledges that so far, his arguments have relied on verificationist principles, and that it clearly not the case that sentences are only meaningful if verifiable. Indeed, he goes so far as to claim that only sentences can be meaningful or not meaningful, and only statements can be true or false; many meaningful sentences fail to make statements ("Could you pass the butter?"). However, he also argues that some form of verifiability is correct in determining factual significance.
What makes a meaningful utterances fact-stating? Nielsen argues that "a statement has factual significance only if it is at least logically possible to indicate the conditions or set of conditions under which it could be to some degree confirmed or infirmed, i.e., that it is logically possible to state evidence for or against its truth," (p. 95). If you disagree, he says, try to think of a statement that everyone would agree has factual content that is not verifiable in principle.
In Judaism and Christianity, God is conceived of as a nonanthropomorphic, transcendent being upon whom the universe is dependent. Believers must accept certain that allegedly factual statements are true, such as "There is an infinite, eternal Creator of the world." Believers take these kinds of statements to be factual. Yet Nielsen argues that these pieces of God-talk are not directly confirmable or infirmable -- we have no idea how to establish their truth or falsity -- and thus they are, in reality, not factual statements at all. Because the utterances fail to be fact-stating, there is a fundamental incoherency at the heart of these religions. One has no reason to cling to incoherent beliefs.
Nielsen didn’t really win me until he addressed the question of ineffability with the same paring-knife back-alley brutality with which he took apart the earlier problems. All along, I kept thinking, “That’s great, that’s right, but what about type II error? What if God did exist but were of a nature such that we did not have the faculties to understand that nature?” When Nielsen addressed ineffability, I doffed my cap and closed my mouth.
But, on second thought: it could be the case that God, or whatever thing it is that people are getting at when we talk about divine agents, is ineffable but not, as in the case of Nielsen’s argued ineffable God, radically ineffable. In this scenario, God has a nature that we can’t understand and thus can’t successfully describe (God is ineffable) but can experience and can misunderstand, and this misunderstanding can take any one of several similar but opposed forms (that is, different denominations or different religions).
I think this is meaningfully distinct from the “radical ineffability” so brutally eviscerated above, in that people could create a shared account of God as a way of trying to make sense of their contact with the ineffable; that account would become a cultural tradition entrenched in such a way that it informs following generations’ experience of the divine; and so the circle (cycle?) of faith can still be built, and the object of faith still discussed, in a scenario in which the divine is ineffable, precisely because of the human ability that we know so well from Gazzaniga et al., constructing stories in the face of something misunderstood. Those circles of faith might lead to true religious experiences after the initial misconception, or might lead to self-reaffirming and false experiences taken for true religious experiences; there’s no way to tell the difference because there’s no way to assess whether the former type of experience exists.
So, although we would still have no justification for believing in God and especially, as before, we have no reason for selecting one account of God over another, in this scenario it could be the case that humans can discuss an ineffable God through the medium of a somewhat inaccurate signifier. Frankly, we do that sort of thing all the time with other concepts that have been subject to less thorough examination. Just ask Broude. The difference in this case is that, unlike some of the concepts for which someone might drag a precise definition from us, in this case “God” is always going to fall short of its true referent because humans aren’t equipped to understand that divine referent; but Nielsen is simply incorrect that that would prevent us from discussing an ineffable divine. Rather, if the divine is ineffable but can in some way be experienced, we would expect to see a history of divisively divergent accounts of God’s nature: exactly what we do see in the world. This is hardly parsimonious, but I think that it outlines the circumstances under which Nielsen’s other arguments could be valid and God could still exist.
Sorry, I just woke up an hour ago, so my semiotics could be a heap of rubbish. Of course nothing I say there suggests that one should believe in any given religion, or that any one among them is likely to be right, just that there are circumstances under which, even given Nielsen’s arguments, there still might be an ineffable divine being known to humans through certain interpretive masks, and under those circumstances, given that Nielsen’s other arguments still establish the weight of logic as being against the case of divinity, having a true belief about God would ultimately be determined in a manner orthogonal to reasoning and more by variation in the error term. It’s certainly more likely that the ultimate referent is merely the apophenically perceived pattern of agency that humans infer when observing the apparently purposive results of a vast number of chance events. But, while there is no reason for one to cling to incoherent beliefs, it remains imaginable that these beliefs could be unjustified but true; and so we arrive back at home with type II error still sitting in the drawing room, sipping the tea that has grown cold since the beginning of our argument. How uplifting.
Oct. 30, 2011, 1:06 p.m.
Interesting points! Thanks for sharing.
Nielsen definitely sits on the “God is radically ineffable” side of the fence. Part of this, I think, is how he defines what a non-anthropomorphic God is — i.e., that part of being a non-anthromorphic God is being transcendent and ineffable. But you’re right — we could take a slightly less radical approach. Narratives could be constructed out of our misunderstanding that do in varying ways inaccurately refer to the being that is misunderstood. I don’t doubt that we do discuss all sorts of less-closely examined things with inaccurate signifiers on a regular basis (language is imperfect), but I’m curious — did you have any particular concepts in mind as examples?
A point that Nielsen brings up later (see Part III, soon) is that the statement “There is a God” is unfalsifiable and thus shouldn’t rationally be the basis for one’s beliefs. Do you think this is subject to the same problem as his arguments on God-talk, namely, that if one takes a somewhat less radical approach, there might actually be some evidence conceivable that makes the statement falsifiable after all?
Nov. 12, 2011, 11:04 a.m.