Why I Am an Agnostic
I don't know whether God exists or not. ... Some forms of atheism are arrogant and ignorant and should be rejected, but agnosticism—to admit that we don't know and to search—is all right. ... When I look at what I call the gift of life, I feel a gratitude which is in tune with some religious ideas of God. However, the moment I even speak of it, I am embarrassed that I may do something wrong to God in talking about God.
-- Karl Popper
... in point of epistemological footing, the physical objects and the gods differ only in degree and not in kind.
-- W.V.O. Quine
Introduction:
For as long as I can remember thinking about them, I've felt an irresistible pull toward philosophical problems, as if their gravity held me captive, preventing me from moving on with life. Around the age of 18, I found myself drawn to what I perceived as the most profound philosophical and ethical crossroads I had ever faced: the question of whether or not there is a god.
It seemed obvious to me that if there was a god, one's whole life might have to be led completely differently than if there was not. What the world was ultimately made of, how it behaved, what things were "right" and "wrong" and why. All of these things seemed to pivot entirely around the notion of some powerful deity that set things some way or watched over them. I had powerful intuitions that ran in all directions, many of which I still have. Surely there isn't plain evidence for an interventionist god, I thought. A God theory makes easy sense of concepts like "right" and "wrong" and how there could be anything, but where are the miracles and voices from the sky? If god is all powerful and benevolent, why do bad things happen? Isn't the entire notion of some agential thing running the universal very anthropocentric? If there was a god, how could we even know about them?
Paralyzed by these questions, I wondered by what means I should even attempt to decide. At the time, I failed to answer the important meta-questions and simply dove into the books. I read lots of the popular atheist literature of the 2000-2010s and my skepticism took over. I regret that I seemed almost completely preoccupied with the existence of some Judeo-Christian god. I remember being convinced mostly by the fact that there was no empirical evidence for a god. The budding verificationist/evidentialist that I was at the time closed the books on the god question and eventually found other philosophical problems.
It was not until certain thinkers completely overturned my philosophical world, that I came to reconsider the problem, this time from a few more angles than just what empirical evidence there was. As a growing thinker, I came to believe in and realize the importance of things that were not only either self-evident or empirically true.
In this note I hope to share my journey from what I now consider to have been a naive atheism to agnosticism. I hope to do this by way of switching the focus to two separate meta-questions that I think are prior to the question of whether there is a god:
Why should we believe or not believe in god(s)?
and
What sorts of god(s) is there?
Why Should We Believe?
Allow me to be annoyingly philosophical for a second: Why do we believe anything is true? Because it can be derived from a self-evident base of other facts? Because we have reasons to believe it? Because its the best idea we've got so far? Because we kinda like it?
I take a justificationist view with a strong pragmatist spin. That is, I think "Because we have reasons to believe it" is close to the right answer. I hold that we believe things because there are good reasons to believe them, but also that we are creatures with limited time and limited computational power faced with a limited set of ideas. (Note that a good reason to believe something might simply be that given our limited time, nothing better cuts the cake.) This view allows for things to be true for many different reasons, empirical verification, non-contradiction, indispensability, mathematical necessity, and so on. Over the years, however, I have approached the problem from several different camps and I hope to show that from several other popular epistemological frameworks a similar conclusion should be reached.
In this note, I will to confront the god question from three different angles:
The Justificationist view (with and without any pragmatic spin)
- Do we have reasons to believe in god?
The Critical Rationalist view
- People have proposed the theory that there is a god, can we falsify it or show it to be unfalsifiable?
The Foundationalist view:
- Does the existence of a god follow from self-evident axioms?
and show why I think agnosticism is the best answer.
What Sorts of God(s) Could There Be?
Likely Not a Judeo-Christian One
18 year old me was laser-focused on the Judeo-Christian god. I just couldn't see the evidence for that god. But why that one? Isn't another sort of god just as well a god too? Why wasn't I thinking about the "Necessarily exists if anything is orange" god or "Only does stuff on Tuesday" god? Cleary, the massive historical and cultural impact of Judeo-Christian religions on the world is to blame for some of my close-mindedness. I have noticed the same narrow focus in many others as well, but if we really want to ask if there is a god and we want to stand that theory up to all kinds of measures of its truth, we need to be more open minded about what kind of god there is. The nature of the god we are inquiring about fundamentally changes what will and will not count as justification, refutation, or evidence for the theory.
My younger self was, I still think, correct about the purely evidential case for the Judeo-Christian god being paper thin. I am even more convinced of that now than I was then. I will not beat the dead horse here with all the arguments about why I believe this to be the case. There are many.
From the empirical evidence perspective, this conclusion easily extends to the three epistemic frameworks above. The justificationist ought to say that she doesn't see any empirical signs that there is a god and that the hypothesis has little support. The critical rationalist ought to take the bold claim that there is a Judeo-Christian god and find it quickly false or unfalsifiable. The foundationalist should agree and should instead maybe look for some first-principles proofs that there is such a Judeo-Christian god.
Other Sorts of God(s)?
Many atheists, including my younger self, falsely think that the arguments that show that there is little evidence for the Judeo-Christian god extend to all versions of god. This blindspot was pointed out to me by one of my early philosophical mentors and shocked me considerably as I was quite pleased with myself at having closed the case on the god question.
Let's think about some weird versions of god for a second. What about "Necessarily exists if anything is orange" god? Don't we have evidence for it? The knee-jerk reaction is "No!", but why not!? Something orange surely exists and that's all it takes. But this is equivalent to just saying "At least one thing is orange". That is, until we try to sneak in the baggage of all of the rest of this god's attributes. Typically, when we question if there is a god, we care if there is an impactful god. That god is not merely a sparse bundle of attributes like the property that "at least one thing is orange", but is rather a very large set of attributes like "Created the heavens and earth, watches over us, forgives our sins etc".
We can carelessly say that "Necessarily exists if anything is orange" god exists, if that's really the only attribute they've got, without fear of any fishy metaphysical commitments, but not "Necessarily exists if anything is orange, and shouts from the sky and makes miracles" god. We don't hear anyone shouting our names from the sky and we don't see any strong evidence of miracles, even though some things are indeed orange.
It is easy to overlook the fact that "god" can just be any number of bundles of attributes for which the truth-conditions are all different. As it happens, the most noticeable or impactful bundles of attributes are the most prone to falsification, not having good evidential support, and not being connected in any way to the results of ontological or "necessity" arguments ala Leibniz or Descartes, as we will see later in this note.
How Would We Know What Sort?
Having seen that "god" can come in all shapes and sizes and therefore be open to all kinds of different truth conditions, let's now ask a weirder epistemological question. If we don't know a god's nature how can we assess the truth conditions of their existence? Or, in more philosopher-talk: Without knowing the secondary intension of "god" how can we say they do or don't exist?
Borrowing from philosopher David Chalmers, the primary intension of a word is roughly its meaning to us a priori, prior to other experience. The secondary intension is its meaning to us a posteriori, after learning a bunch more stuff about the world. For example, the primary intension of "Water" could be "clear odorless stuff we drink" whereas the secondary intension is "".
We don't know the secondary intension of "god", nor is it clear how we might figure that out if god is anything less than an interventionist god who pokes stuff or otherwise makes themselves known. What line of communication do we have to the metaphysical status of some god's existence whose nature is completely unknown to us? It is on that grounds, that I will argue that from the above epistemic frameworks, we should be agnostic towards many types of "god". More specifically, we should be agnostic toward any type of "god" for whom we cannot ascertain the truth conditions, falsification grounds, or required proof for.
Belief in God(s), About Whom We Know Nothing?
If we have any epistemic humility, we will admit that we have no access to what sort of "god" god might be if they did in fact exist. This leaves us in a position where the best strategy seems to be to generalize over the broadest "types" of god(s) there could be. Let's say there could be a "visible" type of god and a "furtive" type of god. The visible god is defined as leaving some discoverable evidence of their presence and the furtive god will be defined as leaving none whatsoever. Having some epistemic humility, let's keep these categories very broad and thus open-ended and let us see how different epistemic frameworks should, I think, arrive at the conclusion of agnosticism.
Justificationist
A responsible justificationist at this point should agree that for a plainly visible god, there is little to no evidence that cannot be better accounted for by other theories. We have more reasons to believe those theories, in light of some evidence, than the god theory. That is a point to which many theists would probably agree. If there is some god whose actions or presence we might have noticed, then why have we not noticed it obviously? Perhaps the visible god is "visible but camouflaged" in ordinary events? But then why are the best interpretations of those events better off without this "visible but camouflaged" god?
There routes the justificationist (with a sprinkling of pragmatism) could take for the furtive god. Maybe the hypothesis alone does not have good reasons to compel belief in it, but maybe it have conciliatory reasons to believe it? Maybe our body of hypotheses about the world coheres better when we believe in the furtive god? If anything, I believe this path just leads us to more questions than we started with. I have yet to encounter something for which a furtive or visible god is the best explanation and does not leave more questions open after the fact.
Yet another justificationist route we might take to defend the furtive god could be the indispensability of such an idea. Many otherwise all-out empiricists hold that mathematical objects are platonically real. They make this exception to their otherwise strict empiricism because they believe we simply cannot go on claiming that they are no such objects and doing math. It is difficult to see what indispensable purpose the notion of a furtive god could be needed for. The most frequent one I hear is morality but once we are unconvinced of the idea of a Judeo-Christian god, we see that a more abstract god is capable of no such underpinning of morality. How can a god about whom we know nothing serve as the basis for morality?
There is yet another type of reason why the justificationist might believe something, that I somewhat hesitate to call "evidence". It might be simply impossible to doubt. However, it might also not be a tautology. Clearly "All bachelors are unmarried men." is impossible to doubt, but so is "This is my hand." and that is not a tautology. We believe it simply because it is not really possible to doubt it. One can maybe say they doubt it, but merely saying the words "I don't believe X" do not constitute doubting it or not believing it (especially for me as a dispositionalist about belief). There is no way of acting which can be interpreted as an otherwise rational person doubting that they have hands while looking at them. Thus, sentences like "I have hands." are sometimes called hinge propositions. Is "There is some god." a hinge proposition? It is difficult to see how it could be.
The justificationist can no more say "The furtive god does not exist" than they can say "the furtive god exists". If there is no evidence for a proposition there is no evidence for its negation, whether that evidence is empirical, conciliatory, or holistically necessary for the rest of ones beliefs.
Critical Rationalist
For the critical rationalist the task is evaluating whether the claims that there is some furtive god or that there is some visible can can be falsified. For the visible god hypothesis, the critical rationalist should be on the same page as the justificationist in rejecting it. If there was some visible god, we ought to have seen some evidence within a reasonable amount of time. We have not and therefore the hypothesis is falsified. What about for the furtive god hypothesis?
Critical rationalists will find a major problem with the furtive god hypothesis qua hypothesis. They will say it is not a scientific hypothesis at all because it is not falsifiable. Note that the visible god hypothesis, as construed above, actually is a scientific hypothesis which observation or further theorizing can falsify. So maybe it isn't a scientific hypothesis, but neither is and that is certainly true. Under this epistemic framework, even axioms or "hinge propositions" ought to be allowed to be scrutinized or else they are just some kind of stipulation that is just not a part of the process of inquiry at all.
This leaves the critical rationalist in the same place as the justificationist. They cannot affirm or deny the proposition. If "There is a furtive god" is unfalsifiable, then so is its negation "There is not a furtive god".
Foundationalist
So far, any reader convinced that the existence of some god is a self-evident truth will probably be somewhat frustrated I have left these for last. Most self-evident arguments for god use the very nature of the god itself to demonstrate its proof. I have Leibniz, Anselm, Gödel and Descartes ontological arguments in mind. All involve some property of god's necessitating its own existence. Descartes is something like:
Premise 1: God is defined as a supremely perfect being.
Premise 2: Existence is a necessary attribute of perfection.
Conclusion: Therefore, God necessarily exists.
Regarding what properties we might know a god to have, we can restate the above: none. Maybe premise 1 allow Descartes to work around that? If we allow Descartes the definitional perfection of god AND we allow him the most notoriously dubious second premise, then what have we really proved the existence of? Isn't this an almost empty bundle of attributes again like "Necessarily exists if anything is orange" god? It is unless we pack in all sorts of other assumptions about what a supremely perfect being is. Again, assuming we know the properties of a supremely perfect being is not only objectionable, but the very opposite of epistemically humble.
The problem with all of these arguments is twofold. First, they bake in the secondary intension of "god" into the placeholder, which only has one or few properties to start (being a "supremely perfect being"). Second that property, in this case being a "supremely perfect being", is usually some property that is argued to entail the existence of the thing. This entailment is always difficult to make clear sense of. Is premise 2 clearly true? How do we know what a supremely perfect being is like? What would it look like to be wrong about that?
The foundationalist might also look to indispensability, or hinge propositions as these things are as close to self-evident as epistemological foundations can get, but they will find the same issues the justificationist found when trying those reasons.
The foundationalist might be able to say "If there were some such being with some such attributes which necessitated its own existence, then it would exist.". That claim is true by definition and is perhaps the least interesting way to say even less than "we don't know" that I can think of on the subject.
Conclusion
In this note I hope to have shared some of my journey with this problem and done some persuading that agnosticism is the correct solution, if we want to call it a 'solution'. Regarding the standard Judeo-Christian version of deities, we can confidently say they do not exist. As for the more nebulous or unknown ones, nothing compels us to answer. We have no secondary intension of "god" and that is not to mention the lack of agreement on the primary intension. From every angle one can poke the problem from, the lack of any epistemically valid notion of what god might even be will block us. To claim the negative to is make the same mistake as to claim the affirmative. It is for this reason, that we ought to suspend our belief and simply say we don't know.
It is interesting to walk the same argumentative paths as my younger self, now with a steadier stride and an eye for the bigger picture. It is that bigger picture that I would like to conclude with here. There are factive claims (claims trying to be true or false) like "Today is Tuesday!" and claims which are not factive like "I will do it tomorrow" or "This ice cream is heaven!". Non-factive language does lots of work in human life. I believe that often religious language is doing a sort of pseudo-factive work. I have come to see this work as mostly a positive thing for people. Whether in building community, providing emotional support, or stopping spiraling uncertainty so one can simply go on with life. When many religious people make claims about god(s), they do so in a way that requires none of the typical stage-setting that factive language requires. Consider how different someone convincing you of an impending alien invasion might sound than someone convincing you to believe in god. For many theists, I doubt if their actions are best interpreted as those of one who believes in the visible god hypothesis we defined above, but rather in something more like "I choose to live my life this way and perhaps you should too."
From a belief-dispositionalist perspective, these actions mean that the religious language of some theists may not be factive at all. Something I’d like to reflect on further is how and why that factive-seeming language crosses over into the factive realm and begins to cause problems.