On Agnosticism and Faith

I’d like to think a bit about agnosticism, and in so doing pick up on some themes that have been brushed up against in comment threads in earlier posts. Specifically, I want to explore what it means to be an agnostic about God. The term “agnostic” was coined by Thomas Huxley in the 19th Century, but I’m not mainly interested in pinning down what he meant by it. Rather, I want to think about how the term is used today—and I expect it is used in some different ways.


This issue is interesting to me for a number of reasons, not least of them being that my wife is a self-described “genetically Catholic, Episcopalian agnostic” (who happens to attend a Lutheran church).

Etymologically, “agnostic” literally means “without knowledge.” But on that definition, many self-professing theists would qualify as agnostics. As so many of my students put it, “Belief in God is a matter of faith, not knowledge.” If agnosticism is to qualify as an authentic alternative to theism and atheism, what kind of alternative is it, given that no small number of theists confess to lacking knowledge about God?

Here, it may be helpful to zero in on what they do claim to have in relation to God, namely faith. Perhaps the agnostic is one who has neither faith nor knowledge when it comes to God. The difficulty here, of course, is that “faith” is a difficult concept, used in different ways—and often without any clear meaning or usage in mind at all.

(More often than not, when my students profess to believe something “on faith” and I ask them what that means, they stare at me blankly. “Do you mean,” I ask, “that you don’t have any good reasons to believe it, but you willfully believe it anyway, just because your parents and community will approve of you if you do?” They shift uncomfortably. Eventually I can get an interesting conversation going, but it takes awhile—because so many use the term “faith” unreflectively, without any clear sense of what they really mean by it.)

Hence, it seems to me that defining agnosticism in relation to faith merely shifts the problem to a different place. Furthermore, it may be that there are different kinds of agnosticism, and that not every kind of agnosticism involves an absence of faith (or an absence of faith in every sense, since there may be different kinds of faith). As such, it may be better to explore the concepts of agnosticism and faith in tandem.

Since faith and agnosticism both seem in some way related to beliefs about God, it may be helpful to begin with the notion of belief. What does it mean to believe something? Here, I tend to be rather influenced by the pragmatic philosopher William James, who held that that a belief’s meaning is given by the difference it makes for what we do and how we live. If this is correct, then to believe X is, most essentially, to behave in certain ways—to operate in the world as if X is true. To believe in the existence of the force of gravity is to operate as if stepping off a precipice (when standing on Earth) will lead to a nasty fall. Those who believe in gravity therefore step off precipices (without parachutes) only when they are suicidal. Those who want to experience floating but believe in gravity will, perhaps, strap on a parachute.

But notice a few things here. First, someone who believes in gravity but wants to die may engage in the same outward behavior (stepping off a precipice) as someone who (inexplicably) does not believe in gravity. The former steps off expecting to die, while the latter steps off expecting to float. The difference lies not in the outward behavior but in aims and expectations.

One might therefore argue that the essence of believing in the existence of the force of gravity lies not in what one does, but in what outcomes one expects—based, perhaps, on what one does, but also based on what one sees happening, etc. This way of understanding belief enables us to distinguish between those who just “go through the motions” of belief—doing the sorts of things that people who believe in gravity (and want to live) do, but who honestly expect to float when they step off cliffs—and those who “really believe it,” that is, who really expect to suffer a nasty fall if they fail to follow the appropriate behavioral protocols.

But defining belief in terms of expectations is problematic, too. Suppose you believe that Marge is secretly in love with you and would appreciate text messages expressing your ardor for her. Suppose, however, that Marge is married and that you are convinced she takes that bond so seriously that she would not respond in any way to amorous texts from people other than her husband, even if they are welcome. And so you don’t expect any discernible response from Marge when you text your little love notes. If you expect anything, it is that your messages will generate a wholly indiscernible internal response, one that will look the same (from your vantage point) as what you would expect to see if some very different things were true about her (for example, that your messages are not welcome but that she thinks the easiest solution is to ignore them).

In other words, if belief is construed in terms of expectations, it needn’t be in terms of what you would expect to observe. And if a belief makes no difference for what you would expect to observe, we are led to Anthony Flew’s pragmatic question framed in terms of his invisible gardener fable: If the person who “believes” in the invisible gardener expects the world to look exactly the same as it would were there no gardener, how is believing in the gardener any different from not believing? Here, the most plausible answer (suggested by R.M. Hare in his response to Flew) is that it makes a difference in terms of how the believer operates in the world. This seems right to me: There is a difference between the believer and the skeptic, but the difference lies not in expectations as much as it does in behavioral dispositions. In other words, we’re pushed back to the behavioral understanding of belief.

Now the implication of all of this might be that beliefs come in different kinds—some of which have their primary meaning given by what we expect to experience under a range of conditions, some of which have their primary meaning given by what we are inclined to do, at least given certain motives. But the line here cannot, I think, be neatly drawn. In most cases, to believe something amounts to having both certain expectations and certain behavioral dispositions.

But now consider a somewhat different example—the case of a trust fall. The idea in a trust fall is that someone falls backwards—presumably into the waiting arms of someone else. Suppose you are preparing to do a trust fall. You are standing at the edge of a table, your back to the floor, and Joe is standing behind you on the floor ready to catch you. If you fall backwards into his arms, assuming you have no desire to be hurt, can we say that you believed Joe would catch you? (Let us suppose that Joe is big enough and strong enough to catch you easily if he chooses to.)

We’d surely say you were putting your trust in Joe. But this trust might be automatic (you know Joe well and confidently expect that he will catch you), or a fairly easy decision (you don’t know much of anything about Joe, but you have no special reason to think he’ll let you fall and so you decide to operate as if he will catch you), or a rather anguished one. Perhaps you and Joe have a deeply conflicted relationship, Joe has betrayed you in the past, but he has apologized and seems to want to establish a new and better relationship.

In the first case, it seems natural to say that you believe Joe will catch you. You behave as if he is going to catch you, and you fully expect him to do so—you might even say that you know he will (although you may be wrong about this, in which case you’ll say “I thought I knew” once you pick your bruised self off the floor).

In the second case, it still seems appropriate to say you believe it (but in a weaker sense). You have no reason not to expect him to catch you, and the context of the game and your general sense of human nature leads you to adopt a defeasible presumption that people in this situation will catch you barring any reason not to.

But it also seems appropriate to say that you have faith that Joe will catch you. You can hardly count on Joe’s reliability in this case, since you know nothing about Joe. You are simply extending to Joe a general confidence you have towards people (a confidence that others, who’ve been burned too many times, don’t have). You are optimistic about Joe because you are optimistic about people in general—but you recognize, perhaps, that your optimism isn’t universally shared, that your reasons for optimism are largely anecdotal in character, and that others who lack your optimism also have anecdotal reasons for their reluctance. In other words, you expect Joe to catch you, but you recognize that your reasons for having this expectation are contestable. Nevertheless, you act as if he will.

In the third case, when you fall backwards into Joe’s arms, “belief” that Joe will catch you seems too strong for what is going on. Furthermore the act of falling seems better described as a gesture of faith than a case of having faith. Faith in this case isn’t something you have so much as it is something you are doing. It is something you are doing in the face of uncertainty, when the stakes are substantial. You don’t expect Joe to catch you, but neither do you expect him to let you fall. Your expectations on the matter are conflicted or ultimately silent. But you decide to operate as if he will catch you.

To decide to fall backward is to give Joe a chance to prove his trustworthiness. You might think gestures of this kind are the only way you can begin to test the sincerity of his overtures of friendship. When you let yourself fall, you operate as if he will catch you, but at best what you have is the hope that he will do so.

But under those same conditions of uncertainty, possessing the same absence of any clear expectations about what Joe will do, you might deciding not to make that venture of faith. You clearly don’t know what Joe would have done. You don’t claim to know that he wouldn’t have caught you Your inner expectations are the same as in the third scenario—conflicted or silent—but your behavior is different.

Now clearly, the first and second cases are not cases of agnosticism. But what about the third and the fourth? Here, the most helpful thing might be to distinguish between two dimensions or parameters of belief—one having to do with expectations, the other having to do with behavior. We might then say that in terms of expectations, cases three and four are both instances of agnosticism; but in terms of behavior, case three exemplifies a kind of pragmatic faith and case four exemplifies the absence of such faith. (Case four, however, needs to be distinguished from case five, in which you fully expect that were you to fall, Joe wouldn't catch you, and so you refuse to fall so as to save yourself from the expected pain).

Now I don’t want to go too far with the analogy between believing that Joe will catch you and believing that there is a loving God who transcends the material world and is operating to redeem its evils. And I’ve barely touched in this post on the issue of evidence and the justifiability of belief. My interest here is in providing some talking points for defining notions of “agnosticism” and “faith,” and perhaps distinguishing some different species of each. And so, with that in mind, I’ll stop here and ask what others think.