This is the first in our new series of featured philosophers. Many more to come. For the schedule look here.
Ambitious Moral Perceptualism and Moral Knowledge from the Armchair
by Preston Werner
For a few years, I’ve been defending the view that all moral justification (realistically construed) bottoms out in the perceptual experience of moral properties.
There are many discussed objections to what I call the Ambitious Perceptualist view. Here, I want to think through some half-baked ideas about the relationship between Perceptualism and the role of thought experiments in moral deliberation and normative theorizing.
We (justifiably) use thought experiments in normative theorizing. But, the thought goes, this is not so for other domains whose epistemologies bottom out in perception. As Michael Milona (2018) puts it:
“[W]ith [empirical inquiry], we rely on actual experiments, evaluative inquiry only seems to require thought experiments…A theory which denies the possibility of evaluative knowledge by mere reflection is going to be highly revisionary; and many would rightly count such a commitment as a serious strike against the theory.”
Consider:
Cat Burning. You want to know whether lighting your cat on fire is morally wrong. You vividly imagine dousing her in gasoline and lighting a match. You form the belief Lighting Mifletset on fire is wrong.
Let’s call this claim, that thought experiments can provide justification for moral beliefs, Datum. As it stands, Datum does not provide support for the claim that there is a deep asymmetry between moral and empirical justification. We can gain empirical knowledge by doing thought experiments:
Couch. You are trying to fit a large and heavy couch through a narrow doorway. You imagine various possible ways of maneuvering the couch until you discover a pattern of movement that will allow the couch to fit using gravity to make it so that you won’t need to lift both sides at once (which would make it too heavy). You form the justified belief The couch will fit through the door by moving it in such and such a way.
The empirical knowledge in a case like Couch is proximately justified by imagination, but the imagination only has the justificatory force that it does because of previous experiences of gravity, large objects, doorways, and so on. The Ambitious Perceptualist could say the same thing about Cat Burning.
Nonetheless, it still feels like there is some asymmetry between Cat Burning and Couch. Perhaps the difference has to do with what we might call, following Sarah McGrath (2011), the “dream test”. Imagine a child counting rocks in order to determine that 2+3=5. In such a case, her experience is playing a causal role in the formation of her justified belief. Nonetheless, the justificatory force of her belief is a priori, as evidenced by the fact that she would not lose her justification for 2+3=5 even if it turned out that she had dreamt of counting the pebbles. The same, we may think, for Cat Burning. Even if your previous experiences of fire, animals, pain, and so on were all dreamt, you’d still be justified in believing that Lighting Mifletset on fire is wrong. Not so with Couch. This illustrates that, at best, experience can enable justified moral beliefs, but not actually justify them.
I’m not so sure that this line of reasoning works. Without holding fixed empirical knowledge about fire and cats, I don’t think we are in a position to be justified in the belief that Lighting Mifletset on fire is wrong. It seems that again, just as in Couch, we need to feed in empirical knowledge for the thought experiment to justify.
Perhaps selecting Cat Burning as my example looks like stacking the deck in favor of the Ambitious Perceptualist. Perhaps instead the asymmetry will re-appear once we look at fundamental normative truths, such as
Pain. Causing pain is pro tanto bad.
Let’s grant that Pain is a necessary truth, and that some people are justified in believing it. If the Ambitious Perceptualist is right, Pain’s justification too must bottom out in moral perceptual experiences. And you might think this can’t be. Now, notice that the objection can’t just be that Pain is a general principle, and principles can’t be perceived. There are plenty of general principles that we know on the basis of experience. It must be that our justification in Pain is somehow not of the right sort to be justified by perceptual experiences. The opponent of Ambitious Perceptualism must flesh out this ‘not of the right sort’ claim. Here are two ways:
- We know that Pain is a necessary truth. But induction over cases won’t be of the right kind to justify knowledge of necessity.
- Our justification for Pain is what explains our justification for particular cases (such as Cat Burning). The Ambitious Perceptualist gets the order of justification wrong. Moral justification is top-down, not bottom-up.
Proposal (1) relies on a conflation first pointed out by Al Casullo (1988). It is one thing to have justification for some proposition P, and another thing to have justification for the proposition that if P, then necessarily P. It is wholly compatible with an Ambitious Perceptualist view of the justification of Pain even if moral perceptions can’t provide us with evidence that Pain is necessarily true. We would just need some other story about our justification for the claim that pure normative claims are necessary if true at all.
Proposal (2) relies on the highly contentious and unpopular view that moral justification proceeds by first being justified in general principles and only then reasoning downward to particular cases. This is in conflict with the reflective equilibrium methodology of normative theorizing that many know and love, and anyway, it entails a pretty unrealistic conception of how non-philosophers reason and form (justified) moral beliefs.
In short, while I haven’t given anything like a conclusive argument that there is no deep problem here for Ambitious Perceptualism, it is not obvious how the ‘armchair knowledge’ argument is supposed to work.
(In fact, I think the Ambitious Perceptualist is surprisingly better placed to explain armchair knowledge than her opponent. But such an argument will have to wait for another time.)
I was confused by the “Dream Test”. Perhaps you were not endorsing it but just mentioning it as something you needed to overcome. You write:
“Imagine a child counting rocks in order to determine that 2+3=5. In such a case, her experience is playing a causal role in the formation of her justified belief. Nonetheless, the justificatory force of her belief is a priori, as evidenced by the fact that she would not lose her justification for 2+3=5 even if it turned out that she had dreamt of counting the pebbles.”
I’m not sure I see that. I assume we agree that just because 2+3=5 is analytic we do not think any belief in it automatically justified. So the dream counting has to add some justification in the same way that real world counting can. I don’t see why we should accept that. In a dream adding 2 things to 3 things might produce 6 things in a way that does not happen in reality.
Thanks for the post, Preston. Two quick questions/comments:
(1) I’d like to pick up on your claims about top-down reasoning in ethics (the view that “moral justification proceeds by first being justified in general principles and only then reasoning downward to particular cases”). Two points:
(a) You claim that the top-down method is “unrealistic conception of how non-philosophers reason and form (justified) moral beliefs.” I wonder what the basis for this claim is. It seems to me that we have lots of empirical evidence that people infer (albeit unconsciously) particular moral judgments from more general principles, such as the act/omission distinction, the means/byproduct distinction (a la Double Effect), and a utilitarian-ish principle about how more harmful outcomes are worse. (Or even something like a Pareto optimality principle on which (roughly) harmful outcomes may be justified if they don’t make anyone worse off than other alternatives.)
(b) You suggest that top-down moral reasoning is incompatible with our beloved method of reflective equilibrium. But aren’t some conceptions of reflective equilibrium incompatible with your perceptual view? Might one argue, for example, that the basic principles we start with (and then balance against intuitions) are prima facie justified and not arrived at via moral perception? This brings me to my second question, which is about how you conceive the terms of the debate.
(2) Do you see your perceptualism as saying that moral justification is ALWAYS perceptual or only SOMETIMES? Similarly, do you think of your opponents as saying that perceptual justification in ethics is impossible, or is the debate more about which is the primary mode of justification of ordinary moral beliefs?
Hi David,
Thanks!
I am a bit confused about what to say about McGrath’s ‘dream test’ in light of your example. But perhaps the best way to think about it is to say that a dream *could* enable one to grasp the relevant concepts necessary for having mathematical knowledge, whereas having a dream that you left your wallet on the table isn’t going to even in principle justify your belief that your wallet is on the table.
At the end of the day, I don’t really have a strong view about whether the dream test is a good one. But whether it is or not, I think Ambitious Perceptualism is still in good shape.
*Also I should note here since I didn’t in my original post — thanks to Ben Henke and Noga Gratvol for reading drafts of the original post!
Hi Josh,
Thanks!
On (a): So I think we can separate out two things here:
(i) Do people’s moral intuitions about cases track some kind of (non-wildly-disjunctive) moral principle(s), such that it is plausible to posit some kind of implicit represented principle(s) that their judgments causally depend on?
(ii) Do people’s moral intuitions about cases epistemically depend on any (perhaps implicitly represented) principle(s)?
I take it that the kind of evidence you are gesturing toward is evidence in favor of a positive answer to the first question. So let me — because I’m not fool enough to dispute an empirical claim made by Josh May(!) — grant that the answer to (i) is ‘Yes’.
Does this entail a positive response to the second question? I don’t think so, at least not unless there is evidence of a kind I’m not aware of (and would love to be corrected about). Compare our judgments that a particular sentence in English E is grammatical. There is all sorts of semantic processing which goes into the forming of this judgment. Some of it probably follows certain rules which psycholinguists could uncover given sufficient research. Does this entail that our belief that E is grammatical *epistemically* depends on the earlier semantic processing and the rules it engages in? I think intuitively not — the judgment that E is grammatical is epistemically basic. (If you don’t like the linguistic case, think of a case of perceiving some simple property, like a square, and a similar story will hold.)
Ok, so that shows that a positive answer to (ii) doesn’t follow from a positive answer to (i), but I can say a bit more to suggest a negative answer to (ii): Agents generally are unable to articulate such principles, which suggests that these principles are subdoxastic, and thus not up for epistemic assessment.
As for (b): I do agree that Ambitious Perceptualism is probably incompatible with some versions of reflective equilibrium. It is very difficult to see how the Ambitious Perceptualist is going to be able to explain how we could be prima facie justified in general principles, assuming we don’t perceive general principles. One way to try to thread the needle here would be to endorse a kind of default epistemic entitlement for general principles, according to which they get justified for free until and unless they face some challenge from the evidence of moral perception. Strictly speaking, what matters for me is that perception is the only source of *moral evidence*. So if there can be justification without evidence, then perhaps reflective equilibrium can be saved.
That being said, I’m inclined to go for the stronger view and just reject the structure of reflective equilibrium. I think the Ambitious Perceptualist view can capture something quite close to reflective equilibrium. It is just that the general principles that feed into the equilibrium are themselves linked to prior moral perceptual experiences, whether the agent herself can list these off, in the same way that say, ‘physical’ intuitions about gravity and so on are.
But I must admit, this is something I haven’t thought about in much detail, so please do tell me if I’m missing something important.
On (2): I think that *all* (positive) moral justification bottoms out in moral perceptual experiences. As I understand the literature, the standard view, even amongst those who allow for moral perception, is to say that *no* (positive) moral justification bottoms out in moral perceptual experiences. (This is a bit tricky to state precisely because you can be justified in moral beliefs about the actual world based on moral perceptual experiences, but the ultimate justification will have to go back to something a priori. Hopefully that’s clear enough.)
So why is everyone in the debate such an extremist, to one side or the other? I guess the reason is an old and boring one: Providing a story about the justification for something as weird as morality, especially realistically construed, is quite hard. If we go for a hybrid view, we’ll need to posit two distinct, puzzling, and open-to-objection mechanisms to provide us access to the moral facts. So a hybrid view risks having the vices of both pure a priori views and pure a posteriori views.
This doesn’t apply to a priori views, obviously, but for me, my original motivation for seeking out an Ambitious Perceptualism is because I’m generally pretty skeptical of substantive a priori knowledge, and I don’t think morality is made up of a bunch of conceptual truths either. So if someone finds themselves wanting a positive and realist-friendly moral epistemology without having to appeal to the synthetic a priori, they should hope I’m onto something.
Very helpful! I appreciate the distinction between a person’s moral judgment causally depending on a principle and epistemically depending on it. But I worry about the assumption that people “generally are unable to articulate such [moral] principles, which suggests that these principles are subdoxastic, and thus not up for epistemic assessment.” I don’t think inability to articulate is strong evidence that something is subdoxastic. The science seems to suggest that people are generally quite bad at articulating why they have the thoughts and feelings they have (even if they are pretty good articulating the immediate thoughts and feelings). Moreover, I think the science suggests that much of the mind is unconscious and difficult to articulate first-personally. If your assumption were true, then most of our mental life would be subdoxastic!
But maybe that doesn’t sound implausible to you? Perhaps that’s precisely the source of your leanings toward perceptualism—i.e. the thought that much of our introspectively accessible mental states are arrived at non-inferentially. But then at least we’ve identified a more fundamental disagreement to adjudicate. Also, it looks then like the motivation for perceptualism wouldn’t have much to do with morality but with your views about the mind generally. And then wouldn’t your moral epistemology rest largely on an empirical claim about the mind?
Hi Preston. Interesting stuff.
I’d like to ask three question, the first two following up, to an extent, on your exchange with Josh.
(1) What does “bottoming out” consist in? I mean, you say epistemic dependence is different from non-causal dependence, and that’s good. But do you have a substantive account of epistemic dependence? Is it that currently popular grounding relation? A specific kind of grounding? Sometihng else?
(2) You know I have the temperament of an extremist, but still, I find the fact that all the relevant views here are so extreme problematic. Perhpas we can proceed by stages. First, do the more phenomenological stuff, and see in what ways justified moral beleifs seem to be justified. I suspect the answer won’t be unified. (Why *would* it be unified?). Then, we can try to vindicate each of those ways. Maybe we won’t be successful at all, and will have to go skeptical. Maybe we’ll vindicate all, and everything will be rosy. Maybe we’ll vindicate some but not all, so maybe we’ll have to go locally but not globally skeptical within morality, or perhpas we’re going to have to revise some of our moral-epistemic practices in order to better vindicate our knowledge of morality. And sure, it’s also possible that only perceptoin will be vindicated, and then whatever moral beleifs can’t get justified in a way that “bottoms out” in perception are screwed. But whether this is so remains to be seen. Can’t you at least go non-committal on the other methods that seem to do work?
(Myself, I’m not a huge fan of perception. I mean, I’ll rely on perception like the next guy, but I don’t think perception is that special compared to other seemings. So I don’t share your general empiricist motivation here.)
(3) Consider “Couch” again. I agree that we can have a justified belief and even knowledge just based on the thought experiment. Still, it seems to me that when we then go ahead and move the couch, and everything works according to plans, we are justified in increasing our credence or in being more confident in the judgment to which we already had some access earlier on.
In “Cat Burning”, though, this doesn’t seem the case. There are reasons not to go ahead with the experiment, of course, but if somehow you do, and you “see” it’s wrong, I don’t think you should be even a tiny little bit more confident of the wrongness of this than just based on the thought experiment. No?
If so, this seems like an important difference, doesn’t it?
Hi again Josh,
I do think you are right that this turns on larger issues about what makes a state epistemically dependent. And yes, it is true that I am attracted to a pretty narrow conception of what is properly agential (and thus up for epistemic assessment). You say, for example, “If your assumption were true, then most of our mental life would be subdoxastic!” — I just would have thought that we should all think this is true! But perhaps my bias is showing.
I’m not an accessibilist about justification — in fact, I am much more attracted to externalist conceptions of justification. But I do think that there is something to the ‘what would you have me do?’ line of thinking that internalists often push. When we have no way to know where our beliefs are coming from, I think that tells in favor (though not conclusively) of a lack of *epistemic* dependence. (Though of course defeaters can still hang around to undermine our justification.)
Anyway, I think none of the above is going to convince you, and that is fair enough. This turns on quite deep issues in epistemology proper and I don’t have anything novel to add to those debates. I’m merely stating my allegiances.
As for why I’m a perceptualist, it isn’t, as far as I can tell(!) grounded in any conception of how wide or narrow the states up for epistemic assessment are. Ultimately, I think that there is empirical reason to think that moral information is encoded in perceptual processing, and that this moral information can serve to provide evidence for moral beliefs. I wholly grant that I *am* going out on an empirical limb here, and it is something I’m in the process of writing about. Strictly speaking, this is compatible with even a Susanna Siegel-ian view according to which even our perceptions are up for rational assessment. I happen to disagree with her view (this is in her new book), but I think for reasons orthogonal to my motivations for Ambitious Perceptualism.
Hi David,
Thanks for these thoughts! My numbers correspond to yours:
(1) I don’t have a fully worked out conception of epistemic dependence. The rough idea is just that epistemic dependence is something like causal/etiological dependence + the basing state is itself up for epistemic assessment (be it a perceptual experience, a seeming, or whatever). Of course what the basing relation consists in, and separating out basing relations from enabling conditions is a quite difficult one, but as far as I can tell I think I can plug in whatever plausible theory you like without trouble here.
(2) I am not *in principle* committed to some extremist view of moral epistemology being true. If different routes can be vindicated, then they are all the more welcome to join the fold, as far as I’m concerned. But I think that we shouldn’t be optimistic. One deeper disagreement between us that I don’t think we can resolve here is that, for me, evidence for some fact has to bear some relation to the fact. And for widely known reasons, getting that relation between us and the *moral* facts is extremely difficult. On your view, a seeming without a defeater is enough, so it makes some sense that there may be multiple sources of moral evidence, because, and I hope I’m not speaking out of turn here, on your view, evidence is cheap! Whereas on my view, evidence is…uhh…expensive.
Put another way, my credence that *no* source of moral evidence is available is higher than that there are multiple sources of moral evidence available. But again, I’m open to the possibility that a priori routes to moral justification could be vindicated.
(3) This is a nice point. I will have to think about it more. Here’s a rambling thought that doesn’t fully address the worry, but is related:
Suppose we had a case where we had a judgment resulting from a thought experiment that it is permissible to Phi. Then we found ourselves witnessing the situation imagined in the thought experiment and it seemed to us that it is impermissible to Phi. Which intuition should we give more credence, everything else equal? My view entails the latter. That might seem implausible, but I’m not so sure. After all, thought experiments are quite limited — we don’t literally imagine every detail of a scenario, we only imagine what we take to be the morally relevant subvening facts. Not so when we actually witness some scenario. This means that actually witnessing could provide us, in principle, with new information about which features are morally relevant in the first place.
I take it that this is, in a way, like Couch. This isn’t exactly the case as you stated it, but suppose you try to move the couch and it turns out your plan won’t work — there is a lamp on the ceiling that you didn’t take into account, and it blocks the couch from fully getting in the doorway. Your thought experiment failed, because you didn’t take into account all of the relevant facts when you were imagining the moving of the couch. This looks quite similar to what was going on in the moral case!
So now we can try to reason backwards: When you *do* successfully move the couch, that can rationally up your credence. Why? Because there was always the chance you were overlooking something. But why not say the same in the moral case? That is, because you aren’t literally imagining a complete possible world when you do a thought experiment, there is always the chance you are missing some morally relevant feature. So when you witness the action and it confirms your thought experiment evidence, you can up your credence.
Now this may seem implausible for Cat Burning, but I suspect that’s because we already start off with such a high credence that Cat Burning is bad. So perhaps better to think about a case where a thought experiment gives us an intuition but not a strong one, and then compare that case with actually witnessing the experiment in the real world.
Thanks, Preston.
Re epistemic dependence: I kind of suspect that the details will matter, but I’m not sure at this stage. Perhaps just one thing now – the claim that all justification (of moral beliefs) depends on perception, is it about doxastic or propositional justification? Or both? Why?
Re (3), I agree that this is what you should say. I also agree that it may be right in some cases. I find it hard to believe, though, that it will be right in all. But it’s worth pursuing. I also think that there are going to be times in which the thought experiment in ethics will be more reliable than the perception itself (you’re right, the world is messy; it’s also messy in ways that divert our attention, that trigger our biases, …). And this too will be true of some science cases (you can, I’m sure, describe a scientific thought experiment that will leave us so confident, that if we do the experiment and get a different result we’ll question the accuracy of the measuring devices). But it still seems to me importantly different, somehow. Not sure I can say more…
Hi David,
Re: Doxastic/propositional justification — The short answer is both. I understand doxastic justification in (what I take to be) the classic way: Propositional justification + the right basing relation. Moral perceptions, assuming no defeaters and so forth, provide propositional justification. If our moral beliefs are based on these perceptions, they will be doxastically justified.
As for the thought experiments vs. actual moral perceptual experiences case. I certainly agree that, for the reasons you say, perceptual experiences could, in cases of the sort you flag, be *less* reliable than thought experiments. And of course I think this is also true in analogous cases about other mundane empirical matters. So I think, so far, I’m still on ok grounds.
I will say that I understand the sense that it is importantly different, and I would welcome any further thoughts. The original post was only trying to illustrate how hard it is to flesh out this gut intuition. I am certainly open to other ways of precisifying the asymmetry here. I want to figure out the most powerful way to understand the armchair objection, but I’m having trouble fleshing things out in a better way than the (bad) ways I gave in the post. This has made me suspicious that there is no there there. But I’m open to being convinced.