Is everyone who lacks belief in a god an atheist? Or is there always some middle category—we’ll call it “agnostic”—such that (a) they don’t believe any gods exist but also, at the same time, (b) they can never be called an atheist? Those in the know will know I am of course talking about the endless debates everyone frustratingly has all the time with Steve McRae, whom despite a lot of elaborate effort I couldn’t pin down to any coherent position on the matter. Here is an outcome report on that attempt and what facts and reality actually dictate in the matter.

Backstory

My involvement all began with my article on Misunderstanding the Burden of Proof, in which I agreed with McRae that everyone bears a burden of proof to assert any belief or disbelief. Thus even atheists bear a burden of proof to warrant any belief that god probably doesn’t exist. But I disagreed with McRae by pointing out that anyone who is already calling themselves an atheist has already met that burden. So they don’t need to meet it “again.” The burden is now back on theists to prove atheists are wrong about that.

This does mean an atheist should be able to give an accounting of why they are an atheist. They don’t “have” to in the sense that they are under no obligation to every theist to waste their time doing this. We all have better things to do. But in principle they should be able to, otherwise their atheism is not warranted. Even atheists who are highly uncertain about their atheism—atheists who conclude the probability of there being any god is close to 50% (many of whom thus prefer to call themselves “agnostics”)—must be able to justify their uncertainty. If they cannot give an account of why they think the probability of any god’s existence on presently available evidence is even so high as 50%, then their believing that is without justification, and thus not rationally founded.

In short, every probability you could ever possibly assign to “some god exists” is a positive belief that requires justification. There is no such thing as a truly “negative” belief within any coherent logic of epistemology. Every so-called negative belief is in fact a positive assertion of some probability or range of probabilities of the truth of the thing not believed. Even “I’m not sure whether x is true or false” or “I don’t know whether x is true or false” logically entails asserting that the probability of x being true so far as you know is or is close to 50%. Which is a positive assertion about x. You therefore must still have a justifiable reason to believe that assertion.

That justification can be fairly simple, however, e.g. “I’ve seen no evidence for or against x being likely” or “I have seen equally balanced arguments for and against x being likely,” but that’s still a justification, it’s still a listing of evidential reasons warranting your belief (in this case, that the probability of x is close to 50/50). And that still entails positive truth claims about the world and your experience of it (e.g. have you really seen no such evidence? have you really seen only balanced arguments for and against?).

But second to this epistemological point is a separate semantic point: that what people in common discourse mean by the word “atheist” is, in actual empirical fact, most commonly just anyone who lacks belief in the existence of any god. McRae wants to change the English language so that this would be somehow semantically disallowed, because he has some sort of weird phobia against being called an atheist himself, and, to satisfy that strange emotional need, he’d rather pave the earth with leather than just wear shoes.

Consequently McRae pisses off a lot of people who prefer the ordinary meaning of English words, and who need that meaning to deal with Christian propagandists trying to distort ordinary language to push their false narrative of the world (such as with the “you aren’t really an atheist” rigmarole or the “you have to waste your time arguing with me or else your atheism isn’t justified” game). But, being insensitive to that socio-political reality, McRae simply won’t let ordinary people call themselves atheists for merely lacking belief, because he himself doesn’t want to be called an atheist for merely lacking belief. Which evinces a completely incorrect understanding of how language works. And McRae really doesn’t like being told that.

That’s the backstory. Which led to the following.

Linguistic Imperialism as Fool’s Errand

Starting back on the 25th of February in 2019, Steve McRae and I agreed to discuss the issue of the epistemic and semantic status of atheism and nontheism in a special Facebook thread I hosted on my account.

Steve asked whether I “think atheists would have a stronger position if they adopted the more philosophical understanding of it in a positive case?” But no answer to this question can be intelligible without working through two subordinate questions, which relate to what it even means to “have a stronger position” or why definitions of words have anything to do with that. So we first had to work through those two things to see if we got anywhere. Then, and only then, could we repeat this process for the original question. Here, I shall be getting only to the bottom of those preliminary questions. (As to the concluding question, I already answered it in my Burden article.)

Steve of course meant the most common definition of “atheism” specifically in use in philosophy (which is not the definition most widely encoded in the brains of the wider public), that being of a positive belief that there is no god; as opposed to merely lacking belief in any gods, which he wishes we coud refer to with a term like nontheist instead (even though these are etymologically identical, a– simply being the Greek counterpart to the Latin non-). But is it more useful to only call “atheism” the positive assertion of the belief that there is no god?

The difference is, again, just a matter of what probability one asserts for the existence of a god: “positive atheism” (also known as “hard” or “strong” atheism) simply asserts that that probability is low enough to be reasonably certain there are no gods; while “negative atheism” (also known as “soft” or “weak” atheism) simply asserts that that probability is not that low, but is also not high enough to suspect there is a god either. Which, again, entails negative atheists are asserting the probability of there being any god is too close to 50% to be confident either way.

That latter position of course can be most easily communicated in regular discourse by calling oneself an agnostic. And generally we’ll all know what you thus mean by that. But that again means in the colloquial, now commonly-familiar sense of agnostic (as someone unsure, undecided, less confident about the status of god); not the rare philosophical sense, having to do with a supposed impossibility of having any knowledge of god, which is now an obscure and antiquated epistemological position almost no one you meet on the street will ever have heard of. But by the most commonly used definitions of all these words, all agnostics in either sense are also atheists. And though they can still dictate how they prefer people refer to them (after all, everyone gets to name themselves), they cannot thereby command how language works generally. If they are atheists by common definition, they are atheists by common definition. Denying this is denying reality. It’s to confuse personal preferences with linguistic facts.

McRae wants instead for common English usage today to “change” to align with hyper-technical usage only widely known in academic philosophy. But, first, that’s not how linguistic change happens. It’s a vain dream to think your wishing or raging at a vast linguistic community will affect it. It won’t. Language will go where language goes; and until then language is what it is. All the railing against it you might conjure will not change either fact. But also, second, even McRae’s claim about academic philosophy is too presumptive; in actual fact philosophers have long employed multiple different definitions of the word “atheist” and there really isn’t an official definition even in that field.

As the Stanford Encyclopedia of Philosophy puts it:

While identifying atheism with the metaphysical claim that there is no God (or that there are no gods) is particularly useful for doing philosophy, it is important to recognize that the term “atheism” is polysemous—i.e., it has more than one related meaning—even within philosophy.

It then gives examples of even more restrictive definitions (such as atheist meaning specifically a philosophical naturalist) and broader definitions (such as atheist meaning anyone lacking belief), which latter the Encyclopedia admits is “a legitimate definition in the sense that it reports how a significant number of people use the term.” Exactly.

Nevertheless, as that same entry points out, a greater utility (for certain limited purposes) of the middle definition tends to be the more common view in the ivory tower. Particularly as philosophers tend to want to talk about propositions (assertions) rather than “psychological states.” Although I believe there remains confusion as to whether these are actually in fact different things. Once you frame all assertions as assertions about a probability—as all assertions of fact actually are—the distinction dissolves. Because every psychological state with respect to the truth of x corresponds to a positive proposition about the probability of x.

But that rare definition only holds sway in the rarefied community of the ivory tower. Not the wider public discourse. And only the latter matters outside the narrow confines of an obscure community. Anyone who grounds their beliefs and attitudes in reality accepts this and makes their peace with it. They don’t get pointlessly outraged by it. Language is what language is. And we should work with how things are, not how we wish they were.

Framing the Discussion in Correct Terms

I then broke this debate down into two propositions:

  • (1) I have said there is rarely any point in arguing over definitions. The only thing that matters is whether you understand how someone else is defining a term, and that they understand how you are defining that term. Anything else results in failed communication, defeating the purpose of words altogether.

Outside that single statement, there is no unique sense in which one definition is “right.” One can talk about which definitions are most commonly understood (so that if you use a word without defining it, what can you expect your audience to have “heard” you say?), but that’s an empirical question, not an analytical or logical one. Likewise which definitions are most “useful” can vary enormously. Because people can disagree as to which outcome measures are the more important (a word can be “useful” in many different ways; which of those ways are we to prefer and why?); and even when acknowledging each other’s chosen utility function, the question of whether a certain definition actually is useful in that sense is again an empirical question, not an analytical or logical one.

And this distinction is important. Empirical questions can only be resolved by reference to facts, not logic. So is there anything really to discuss here, as to how “atheist” is defined? For example, if most people simply understand it to mean one who isn’t convinced theism is true, should we go around acting like that’s not what most people understand it to mean? Or do most people understand it that way? These are questions in communication theory.

And then:

  • (2) I have said that when we treat epistemic assertions coherently, they are always probabilistic; and when you reduce every assertion to its assertion about probability, the question of who is an atheist dissolves as moot.

I will set aside the rare exceptions (e.g. certain propositions about “uninterpreted present experience” can be, epistemically, absolutely true or false; and incoherent statements have no content capable of being true or false), because they will not be relevant to this issue, which is a question about intelligible assertions regarding external reality, which always have a nonzero probability of being false. Because all such statements have a nonzero probability of being false, all such statements reduce to the assertion of a probability (or range of probabilities). Always.

Once we recognize that, the question “Do you believe a god exists?” can only ever really be answered with a declaration (however couched in coded words) of what you believe to be the probability that a god exists. Even claiming you do not know the probability entails the declaration of an epistemic probability: that the affirmation of the fact in question is no more or less likely than the denial. Because if you did not believe that, then it is logically necessarily the case that you believe the affirmation of the fact in question is more or less likely than the denial. This is a necessary consequence of the laws of logic. Either q (affirmation is as likely as denial) or ~q (affirmation is not as likely as denial). There is no third possibility available. So what epistemic probability do you assign to the existence of god? That’s all that matters. What you “call” that position is irrelevant. Except with regard to communication theory. Otherwise, honestly—is there any position you can take that does not entail an epistemic probability for god? These are questions in epistemology, not communications.

I am of course here stating what I’ve argued. Whether or in what ways McRae agrees or disagrees with either is what we wanted to discuss. I shall note again that there are two separate and distinct issues here: one is a question in communications theory (e.g. how do we successfully communicate with the public); the other a question in epistemology (e.g. what positions with respect to the existence of god are even possible and how can we logically characterize those distinctions). I had to call attention to that distinction again and again as McRae kept blurring them. We must stop blurring them.

Outcome Number One

In regard to communications theory, if we call ordinary language the Common Tongue and the peculiar dialect employed by philosophers Elvish, my point is that if you are talking to people who don’t speak Elvish but only Common, you need to speak Common; else they won’t understand you. And if they are speaking Common to you, you can’t criticize them for making errors in Elvish. Because they aren’t speaking Elvish.

And as the Stanford Encyclopedia of Philosophy says: the lack-of-belief (i.e. the broader) definition of atheist “is a legitimate definition in the sense that it reports how a significant number of people use the term.” Indeed, when you ask bunches of people from all walks of life, you will discover this is the definition most assume to be normative—in Common. Not in Elvish. But we weren’t speaking Elvish. And most people don’t. Communication will always fail if you interpret statements made in Common as if they were made in Elvish, or if you keep speaking Elvish to people who only speak Common.

And there are certainly philosophers who have agreed. Antony Flew, for example, specifically argued in his paper “The Presumption of Atheism” that the colloquial meaning should be adopted even by philosophers, in order to align philosophy with ordinary language:

I want the originally Greek prefix ‘a’ to be read in the same way in ‘atheist’ as it customarily is read in such other Greco-English words as ‘amoral’, ‘atypical’, and ‘asymmetrical’. In this interpretation an atheist becomes: not someone who positively asserts the non-existence of God; but someone who is simply not a theist. Let us, for future ready reference, introduce the labels ‘positive atheist’ for the former and ‘negative atheist’ for the latter.

In truth, if you go around in public saying “atheist,” most people will understand the term to be inclusive of both positive and negative atheism. Such that when someone says they are an “atheist,” no one will by that fact alone know whether they mean positively or negatively, requiring further explanation or inquiry. This is simply how the language is. So if you want to communicate with the public, you have to assume this is the case. Because it is. That’s just the way Common works. And you don’t have a magical nanorobotic cloud that can envelop the earth and rewire the neurons in the brains of billions of people to change that. You’re stuck with how they are wired. And to communicate successfully, you have to listen and to speak as their brains are programmed to understand. Period.

It’s not even particularly better in philosophy, as there are enough philosophers using “atheist” in both senses that one will not know which sense any specific philosopher is using unless they specifically tell you. And since many do indeed use the fully inclusive sense shared by the wider public, even if most use the hyper-technical sense that excludes negative atheism as a third and separate category from “atheist” and “theist,” however labeled, you still won’t know which philosopher is which unless they tell you. So in actual practice there is no utility in specializing the definition like that. Nothing is gained in efficiency or effectiveness of communication by that even within philosophy. Even in philosophy, one simply just always has to stipulate which sense you mean. And that’s that.

I couldn’t get McRae to acknowledge any of this or explain why he refused to acknowledge it. He wouldn’t even recognize that most people arguing with him were taking him as attempting to dictate to them how Common works (in which event he was totally wrong), not how Elvish works (which none of them actually care about). So I gave up. As far as I could tell, he seems resolutely disinterested in adopting a sound theory of communication, particularly with the public.

Outcome Number Two

I also think McRae kept confusing ontological with epistemic probability in our discussion. It seemed he kept falsely assuming propositions can be assigned an ontological probability—which, outside purely analytical contexts divorced from the impact of evidence, is never the case.

The ontological dichotomy between “a god exists” and “no god exists” (hereafter G and ~G) is represented in the logic of epistemic probability (or P) by the rule that P(G) + P(~G) must always equal 1. If P(G) + P(~G) < 1 then you have violated the Law of Excluded Middle (you’ve left out an option, whatever fills the remaining probability space). If P(G) + P(~G) > 1 then you have violated the Law of Non Contradiction (you’ve included a possibility that is both G and ~G). But once you correctly ramify P(G) and P(~G), you have a continuum, not a binary.

We can never simply “know” whether “God exists” is true or false; we can only ever know how probable it is. And we can only know how probable it is given what we know at any given time. Thus as our knowledge changes, so might that probability. So when making verbal statements about God, only the logic of epistemic probability applies. And this follows because knowledge is “justified true belief,” yet it is logically impossible to ever have absolute knowledge that something external to us is or is not “true,” because we can only have knowledge that it is probable (or improbable—or neither probable nor improbable, which entails it’s at or near 50/50, still a probability).

This is why knowledge only exists when stated as a probability. We can hide this with all manner of couched language or elided assumptions about how “certain” we are of what we are saying at any given time. But we cannot avoid the fact that there is always a nonzero probability we are wrong in any statement we make, in any belief we have (see How Not to Be a Doofus). Which entails we are only ever really affirming the converse probability to that, which is then the probability that we are right. For instance, if we believe there is a 5% chance we are wrong about something, we are ipso facto declaring a 95% chance we are right about it. Thus, all knowledge, and thus all belief, is probabilistic.

You can “believe God does not [or does] exist or hold no belief either way (Suspend judgment),” but that is simply couched language hiding what you really mean, which is that you can believe the probability that God exists is low enough to be confident He doesn’t, or high enough to be confident He does, or neither high nor low enough to be confident either way. And those three options then exhaust all possibilities. But you cannot pick a third option when the dichotomy is “affirmation is as likely as denial, or affirmation is not as likely as denial.” Because then as defined those two options exhaust all logical possibilities. Thus you can’t escape the conclusion that you must be affirming some range of values, however wide or vague, for P(G). And that entails 1 – P(G) = P(~G). Thus every probability you assign to “God exists” entails an equally positive assertion of the probability “God does not exist.” And vice versa. There is no way to avoid this with semantics about “suspending judgment.” That’s just code for “I don’t know that P(G) is < or > P(~G)” which logically entails “for all I know P(G) = 0.5 or near enough to be uncertain whether or not G exists.”

And note here I am not talking about what we will call each probability assignment, i.e. who counts as an atheist or agnostic; I am here only talking about what is true regardless of what words we assign to these possibilities, which is that all positions regarding G (no matter how you demarcate them) entail claiming some range for P(G), and this is the only useful information to have. “What that position is called” is not useful information, except with respect to communication theory. Which I already discussed above. I feel like McRae had a hard time grasping this.

I think epistemically McRae wants to assert for himself “P(G) is near enough to 0.5 for me to lack confidence in either G or ~G.” Which is fine. As long as it’s rationally founded. That is, as long as “close to 0.5” actually is where the evidence you’ve seen rationally leaves it—one can question that (and to be honest, IMO, McRae’s probability assignment here is not even remotely plausible), but one would then be making a different argument, one to do neither with the epistemic options available nor how to name them. Otherwise, the epistemic options can be as many as you want. You can split the probability space in two, three, five, a million pieces, and name them all if you want. Then the question becomes whether people who affirm one or more of those positions “should” or “should not” be called atheists, or when they should or shouldn’t. Which then is only a matter of effective communication: what the people you are communicating with already understand the relevant words to mean.

McRae wanted to insist on dividing the possibility space into three (high P(G); low P(G); and middling P(G)), and incorrectly assumed doing that requires separate words for each division of that space (“theism”; “atheism”; and then “agnosticism” or “nontheism,” or something). Neither is true. Dividing the space into three rather than two is simply an arbitrary decision. We could divide it into two, three, or ten if we wanted to. It would dictate nothing. How people’s brains have learned to label which demarcations will be what it is regardless of how you or any fringe linguistic community does it. And even if we do divide it into three, in that very fashion, it does not follow that we cannot name supersets of them, e.g. the last two options can together have one label inclusive of both (which is what “atheism” has evolved in common use to mean), while also at the same time having specific separate labels for each (such as “agnosticism” for the third option and some phrase like “strong atheism” for the second). There is no logical necessity that it be otherwise. And in actual linguistic fact, it simply is not otherwise.

McRae had a hard time being clear as to what he even means by “believing” something as opposed to “suspending judgment.” But these are again just code words for confidence levels. “Believing x” simply means “Confident that x is true.” “Disbelieving x” simply means “Confident that x is false.” And “suspending judgment” simply means “Not confident that x is true or false.” Which is itself a belief. Because any statement about your level of “confidence” is simply a disguised assertion of probabilities, an assertion you thus believe to be true.

Everyone’s threshold may differ from everyone else’s, but each individual is only reasoning coherently if their own thresholds are consistent. At some probability of x, you will “be confident that x” and therefore “believe x“; or “be confident that ~x” and therefore “disbelieve x.” And at every other probability, you will “lack confidence or belief that x or x,” a.k.a. “suspend judgment.” In which case you believe the probability is too close to call. And in practice, people usually demarcate more than just those three options, describing instead different and varying degrees of confidence or belief. We also have to distinguish this from the fact that people will also make decisions not just on their degree of belief, but also on their calculated risk of being wrong—such that when it costs more to be wrong, we usually require a higher confidence that something is true (or false) before acting on it.

One must not confuse these things. Risk tolerance is different from degrees of belief; which is different from what you choose to call each degree of belief; which might then even be different from what you need to call each degree of belief in order to be understood when communicating with other people (or to understand them in turn).

Instead, McRae insisted his tripartite division was “not arbitrary as it is a function of logic.” Wrong. He was choosing to divide the probability space into three areas: high-P, low-P, and middling-P. There is no objective reason why anyone has to do that. We could have divided the total area into six, or ten, simply by adding more labels to describe each division (such as “high confidence,” “very high confidence,” and so on). And in practice people often do exactly that. People can also divide the space simply into two if they want: high-P and low-P, being merely whether someone thinks P(G) is above or equal-or-below 50%. Nothing in logic prohibits their doing that. And even if we divide it into three, we still have to arbitrarily choose how much lower than 50% will count as “low-P” and not “middling-P” (similarly for “high-P”). Which is always just arbitrary. There is no demarcation dictated by logic.

So McRae can’t “force” people to change how they use words by resorting to logic. Every system of demarcating words suggested here is perfectly logical. Overlapping terms is perfectly logical. Binary terminology is perfectly logical. Trinary terminology is perfectly logical. It’s all perfectly logical. So the only consideration that remains is what terminology people actually use, and thus will actually be understood. Thus it just goes back to communication theory.

Bad Analogies

McRae tried to argue such things as that Christian misuse of the word “evolution” to argue against evolution theory is like the Common-tongue “misuse” of the Elvish term “atheist.” But there is no analogy here. The word “evolution” is used in all sorts of ways in Common (“language evolves”; “I evolved as a person”; “evolution is any change over time”; and so on). That in Elvish it might specifically refer to speciation by natural selection is irrelevant to this fact, and is never an argument for changing common usage. If Christians specifically mis-define “evolution” when arguing about the scientific theory, then they are the ones violating the principles of language; but people using colloquial definitions of “atheist” in common discourse aren’t doing that—because that’s correct usage in Common. Only if they tried to insist the Common definition was the “only” one used in Elvish would they be breaking the rules of language. But they aren’t wrong even to say that the Common definition is sometimes still used in Elvish. Because it is.

Even McRae’s insisting God’s existence is “either true or false” gets us nowhere. Because all we have access to is the probability of its being true or false. So when people refer to what they believe (what they do or don’t have confidence in), they are only ever talking about that probability. Which spans a continuum. It is not binary. So there is no route from here to what McRae wants either.

Similarly, McRae tried arguing that he would treat differently his conclusion that there was a 70% chance my claim to have bought a cat was true and his conclusion that there was a 70% chance my claim to have been abducted by aliens was true. But that’s incoherent. If he actually believes there is a near 1 in 3 chance I didn’t actually buy a cat, then he would have to admit he’s not that confident that I did, but will only accept it on balance of probability. Because that’s literally what “there is a 70% chance you really did buy a cat” means. Likewise, if (remarkably!) I could present enough evidence to get McRae’s confidence as high as 70% that I was abducted by aliens, then he is literally conceding there really is a near 2 in 3 chance I really was abducted by aliens! He should act accordingly. Which means, he should act as if there is only a near 1 in 3 chance I wasn’t abducted by aliens. He should not be treating the two cases differently with respect to belief—only if he assigned them a different probability of being true would that make sense.

Of course it will be different if we are talking about the cost of trusting me in either case and being wrong, which might be lower for the cat thing than the aliens thing, but that’s not the same issue—that’s a matter for decision theory, not epistemology: what we should do about our degree of belief, not what our degree of belief should be.

Conclusion

McRae of course complains about confusions people fall into who aren’t philosophers and haven’t really thought much about this (or even about semantics or neurolinguistics in general), such as when they start talking about babies or even rocks being atheists. But this is no different than the confusions that result from people arguing over whether a taco is a sandwich. They don’t know why their brain tells the difference between them, so they struggle to formulate an analytical explanation for it. In result, they often come up with a false definition of taco and sandwich—and here I mean false in the sense that their vocalized definition does not correctly describe what their brain is actually doing when distinguishing tacos from sandwiches.

Any more careful thought would resolve this error. Our brains are telling us, quite simply, that a sandwich is any food between bread that is not a taco. Our brain has learned to exclude special conditions in order to recognize the difference between a sandwich and a taco. It is thus simply defining a sandwich as not a taco. There is nothing philosophically systematic about it. It’s just how language works. Similarly, rocks are incapable of having cognitive belief states, and “atheist” is only a label for a cognitive belief state. Rocks can therefore no more be atheists than they can be Republicans.

Similarly, babies cannot comprehend the question, “Does a god exist?” and therefore can have no epistemic state regarding the answer. They are therefore no more capable of being atheists than theists. But that’s only true with a highly abstract definition of “God.” If instead we define theism as any belief in the existence of anything resembling a god, then it’s actually more likely that all babies are theists, contrary to the usual line atheists maintain.

Child psychologists have long suspected that babies begin life essentially solipsistic in their cognitive disposition toward the world—which means, in this broad sense of “theism” just proposed, they are born believing they are god. But they very soon move toward displacing that belief with the belief that their parents are gods—as in, having absolute power over the world, e.g. babies readily think their parents are responsible for the rain, gravity, and everything else that happens in the world, or even in their own bodies, which is why infants get so angry that their parents don’t fix their every physical discomfort. Agency overdetection (a neurologically innate cognitive bias) then leads them to start seeing gods everywhere (in trees, dolls, the weather). Then, usually, they are subsequently, seamlessly encouraged to displace those beliefs, in their parents as gods or in a world full of gods, with the belief in an abstract celestial god—unless for some reason they find such attempts unconvincing and thus abandon their belief in gods altogether. But by that point they are no longer babies and are indeed properly described as atheists.

But the fact that average people don’t know anything about child psychology or don’t notice that “atheist” is a descriptor for a cognitive state, and thus trip up over such things, doesn’t tell us anything about what they really understand the meaning of “atheist” to be in practice, which is actually physically encoded in their brains wholly regardless of what they erroneously think that encoding to be. No matter how wrong they are about whether a taco is a sandwich when asked to reason it out, their brains already know the difference and have no difficulty with it. They spot tacos from sandwiches with 100% consistency without even having to think how. Likewise “atheism” and “atheist.”

In the end, when it comes to epistemology, the only thing that matters is what probability you assign to “God exists.” Labels are irrelevant. And when it comes to communication (understanding what other people are saying, and being understood in turn), the only thing that matters is what someone will understand you to mean when you say you are an atheist or that someone else is an atheist. And the fact is, most people won’t know whether that means a positive or a negative atheist. And that tells us their brains have been enculturated to recognize no difference between them by that word. Therefore, if we want to be understood, and if we want to understand other people, we have to use the word that way. If we want someone to know more specifically what kind of atheist we are or that someone else is, we simply have to tell them. Because neurolinguistically, that’s just how it is. This is why phrases like “positive” or “strong” atheist and “negative” or “weak” atheist had to be formulated. That fact alone is evidence of the conclusion.

Hence that being the case in no way prevents you from using more specific descriptors if you want to, like “agnostic” or “weak atheist” or whatever you want, including simply stating outright the ranges of probabilities you conclude for the existence of a god or speaking in alien languages (like Elvish) among fringe linguistic communities who understand those languages—as long as the people you are talking to have brains suitably programmed to know what you mean by those utterances. And I have to caution here, in the case of “nontheist,” no such programming exists. Not even in the brains of most academic philosophers. So no one will recognize that as meaning anything other than atheist, which means no one will be able to distinguish merely by that word between either type of atheism. It therefore has no special utility in communication.

The linguistic facts here are already settled. And the logical possibilities are arbitrary and limitless. You should speak so as to be understood, and learn how people actually use words to understand them. Everything else is fringe semantic gameplay. Warring over the meaning of “atheist” or “nontheist” is therefore a monumental waste of time. You can surely find better things to do with your limited resources in life.

Discover more from Richard Carrier Blogs

Subscribe now to keep reading and get access to the full archive.

Continue reading