In recent years there has been a lot of debate regarding the proper definition of “atheist,” even on this blog. Traditionally, atheism has been defined as the claim that God does not exist. In the mid-20th century, however, atheist philosopher Antony Flew attempted to redefine atheism. Noting that the Greek prefix “a” is a term of negation, Flew said the proper definition of a-theism is simply “not a theist.” Another popular way of cashing this out has been to define atheism as “one who lacks belief in God.”
What’s the difference between these definitions? The traditional definition is an ontological claim (God is not included among the entities that exist) while the new definition is a psychological description (“I have no belief regarding the existence or non-existence of God”). We might label these two ways of defining atheism as “ontological atheism” and “psychological atheism.”
Why does it matter how we define atheism? It matters because of the burden of proof. A principle of rational discourse is that he who makes a claim bears the burden to defend it. If someone claims that God does not exist (ontological atheism), he bears a burden to demonstrate how he knows this to be true. On the other hand, one who lacks any beliefs with respect to God’s existence (psychological atheism) bears no burden of proof because he is not making a claim to knowledge. He is merely describing the content of his beliefs – that his stock of beliefs does not include a belief regarding the existence or non-existence of God. Flew understood this. He purposely redefined atheism to make it a psychological description so as to absolve atheists from their burden to defend the claim that God does not exist.
Who is Right?
The $64,000 question, then, is which definition of atheism is the right definition? Is atheism “belief in no God” (ontological atheism) or “no belief in God?” (psychological atheism) Historically, atheism has been defined as the belief that no god(s) exists. For example, The Stanford Encyclopedia of Philosophy defines atheism as “the negation of theism, the denial of the existence of God.” And there is good reason for this definition. There are only three possible positions one can take regarding any given truth claim: affirm it, deny it, or reserve judgment. In this context, the truth-claim in question is “God exists.” Theism is the affirmation that the proposition “God exists” is true, while atheism is the denial that the proposition “God exists” is true. If one does not know if the proposition is true or false, then they are an agnostic. It’s that simple!
While many atheists want to define their view as “one who lacks belief in God,” when you probe the content of their beliefs it becomes readily apparent that they do not lack a belief concerning the existence of God at all. They are not neutral with respect to the proposition “God exists.” They believe God does not, or probably does not exist, and thus regardless of how they wish to define atheism, they bear a burden to explain why God’s existence is more improbable than not.
Assessing Flew’s Redefinition
It seems to me that even on Flew’s redefinition, the atheist cannot avoid a burden of proof given the definition of “theist.” A theist is someone who believes God exists. Someone who is “not a theist,” then, is someone who does not believe God exists. That’s the traditional definition of atheism! It is also a claim to knowledge, and thus comes with a burden of proof. Even if we excise the psychological notion of “belief” from the definition of theist, it fares no better. If theism is defined as the proposition that “God exists,” and the a prefix negates that proposition, then atheism is the proposition that “God does not exist.” Once again, this is in line with the traditional definition of atheism. Both definitions entail an ontological affirmation, and thus require the atheist to explain why he believes (or how he knows) God does not exist.
In addition to etymology, there are other reasons to reject this new definition of atheism. First, it invites unnecessary confusion. Words are labels, and labels are shorthand ways of referring to specific ideas or things without having to describe them each time. If I want to talk about a dog, I don’t have to say “one of those little furry creatures with four legs and a tail that has a proclivity for barking.” I simply say “dog” and everyone knows what I am referring to. But what if someone redefined “dog” to mean “any common household pet,” such that both cats and dogs would now both qualify as “dogs.” Surely it would invite confusion. Whenever someone said they own a dog, we would always have to ask, “What kind of dog do you own: a canine dog or a feline dog?” Whenever a single label is used to refer to more than one idea/thing, clarity becomes more difficult to achieve. If we expand the definition of atheism from “belief in no God” to include “no belief in God,” then it must always be asked of an atheist, “What kind of an atheist are you?” The purpose of labels is to clarify and make distinctions, not to obscure. Expanding the definition of atheism will only serve to diminish the label’s descriptive utility.
Secondly, the concept now being labeled as “atheism” (lack of belief in God; not a theist) has long been described as agnosticism. An agnostic is one who either believes that humans cannot possess warranted beliefs concerning the existence or non-existence of God, or one who simply lacks a belief concerning the existence or non-existence of God. In what way does this differ from the new definition of atheism? It appears to be a distinction without a difference. In the end, psychological atheism collapses into agnosticism. If the “agnostic” label captures the essence of this new form of so-called atheism, then it should continue to be described with this label.
Thirdly, to define atheism as “a lack of belief in God” turns atheism into a description of one’s psychological state, rather than a position. Since psychological states are neither true nor false, defined as a psychological description, atheism is neither true nor false. Only propositions have truth value.
Fourthly, to suggest that those who think the proposition “God does not exist” is false and those who lack beliefs with respect to God’s existence should both be called “atheists” makes as much sense as suggesting that those who think the proposition “God exists” is true and those who lack beliefs with respect to God should both be called theists! Indeed, it seems a bit strange to associate the absence of belief with either “atheism” or “theism” since both labels are used to describe beliefs about the truth or falsity of God’s existence.
Finally, the new definition spreads its net too wide. If an atheist is one who lacks belief with respect to the question of God’s existence (or is simply not a theist), then babies and dogs should also be classified at atheists. This is absurd! There is a cognitive element to atheism that the new definition is clearly lacking, and thus not deserving of the atheism label.
Having Said That…
While I think there are good historical, logical, and practical reasons for rejecting the new definition of atheism, I acknowledge the fact that words are defined by usage. If the more expansive definition of atheism becomes generally accepted by English speakers, then one cannot deny that the new definition is a legitimate definition (even if we do not agree with the change in usage). There is no principled reason for thinking the historic definition of atheism could not be changed or expanded to include other meanings.
Indeed, something similar happened with the word “agnostic.” The term was coined by T. H. Huxley, which he defined as a method of knowledge encapsulated roughly by the principle ‘the certainty with which you believe X should be proportional to the evidence for X.’ As applied to the question of God’s existence, he did not think it was possible to know whether God exists. Over time, however, it also came to refer to one who makes no claim to personally know whether God exists, but considers it possible that the answer can be discovered. To distinguish between these different positions and to restore clarity in the public square, it became necessary to augment “agnostic” with other descriptors such as “soft” and “hard.” If the public comes to accept the new definition of atheism as legitimate, those who adopt the “atheist” label will required to provide similar augments to accurately identify their view. Personally, I think the simplest solution is to abandon the attempt to redefine or expand the definition of atheism, opt for a new label to describe their position, stick with “agnostic,” or forego a label altogether. In support of the last option, consider the fact that we do not regularly have labels to describe a lack of belief in something. For example, there is no label for those who lack a belief in aliens. So if truly and only lacks belief in God, perhaps they are not deserving of a label because they have nothing to contribute to the conversation.
The move to redefine atheism has been motivated less by a desire for greater clarity, and more by a desire to avoid a burden of proof. While I think there are good reasons to reject this redefinition as invalid, unnecessary, and unhelpful, even if it comes to be accepted as legitimate, the “atheist” who adopts it still bears a burden of proof if he thinks the proposition “God exists” is false or probably false. In my experience, the vast majority of those who prefer to define atheism as “no belief in God,” “not a theist,” or “one who lacks belief in God” fit this description. They think God does not, or probably does not exist, and as such, they bear a burden of proof no matter what they call themselves or how they define their terms. While definitions are important, let’s not allow them to distract us from the main thing: What does the self-described atheist believe concerning the proposition “God exists”? If they have an opinion on the matter, then they also have a burden of proof.
Not all atheists would argue that these two definitions are mutually exclusive. Many would acknowledge that “atheist” has been historically defined as the belief that no God exists, but argue that this is but one legitimate way of defining the word. They object to those who wish to exclude the new definition as illegitimate, or automatically assume that anyone who adopts the atheist label must believe God does not exist.
An atheist might object, “While I tend to think God does not exist, I am not certain of this and thus bear no burden of proof.” What he fails to understand is that certainty is not a prerequisite for belief, and hence is not a prerequisite for a burden of proof. Most of our beliefs are held with a level of justification that falls sort of certainty. Indeed, most theists would probably admit that they are not certain God exists. How would the atheist respond if the theist said, “Since I am not certain that God exists, I bear no burden of proof”? Surely the atheist would object to this, and rightly so. One need not be certain of the truth of some X to believe X,and if they believe X is true they have the burden to demonstrate how they know X is true.
Certainty is not a prerequisite for labeling oneself an atheist either. See “Dawkins is an agnostic? Why certainty is irrelevant to defining atheism.”
William Lane Craig and J.P. Moreland, Philosophical Foundations for a Christian Worldview (Downers Grove, IL: InterVarsity Press, 2003), 156.
Consider the fact that the early Christians were called “atheists” by their pagan neighbors because Christians did not believe in the Roman gods.
“Agnostic” comes from the Greek word gnosis,meaning “knowledge.” The a prefix negates this meaning, hence “lack of knowledge.”