(Cross-posted from my website. Podcast version here, or search for "Joe Carlsmith Audio" on your podcast app.)

1. Introduction

Sometimes, people try to use meta-ethics to argue that pleasure and pain are the only things that matter. In particular, they say, we have a certain type of direct epistemic access to the goodness of pleasure and the badness of pain – access that we don’t have to other candidate values. So the value/disvalue of pleasure/pain are on solid meta-ethical footing in a way that other values are not.[1] Call this argument “meta-ethical hedonism.”[2] 

I think this is a bad argument. This essay explains why. In brief:

  • Suppose you’re a non-naturalist about meta-ethics (that is, you think that the value of pleasure is something over and above pleasure itself, and irreducible to anything in the natural world). In that case, the argument fails to answer the basic epistemic objection to non-naturalism – namely, that it leaves us without the right type of epistemic access to the non-natural properties in question. In particular, positing a mysterious faculty of “direct acquaintance” with your phenomenal properties doesn’t help (even if such a faculty were otherwise legit), because phenomenal properties are natural properties.
  • Now suppose instead that you are naturalist about meta-ethics (that is, you think that normative properties are identical with some set of natural properties). In that case, your direct acquaintance (if you had it) with phenomenal properties – and thus, with a certain type of natural property – would still do no work in telling you which natural properties matter. Rather, to figure out that out, you need to do some combination of conceptual work, to understand how a concept like “goodness” picks out its natural referent, and empirical work, to check what that referent is. And direct acquaintance with the pleasantness of pleasure does neither.

Maybe this summary makes the issue sound complicated, but it really isn’t. Ultimately, pleasure is just another natural thing, and direct acquaintance, if it exists, just another way of detecting and comprehending a certain type of natural thing. Nothing unique has been offered that would bridge (or explain away) the divide between is and ought.[3] The basic game of normative epistemology remains unaltered.

Indeed, I suspect that the appeal of meta-ethical hedonism derives at least in part from mixing normative epistemology together with the epistemology of consciousness in a manner that allows confusions about the latter to disguise muddiness about both. This isn’t to say that confusions about the epistemology of consciousness aren’t relevant to normative epistemology – I think they might well be. But I think the best meta-ethical view in this vicinity (what I call the “mystery view”) wears its confusion and mysterianism more forthrightly on its sleeve – and I don’t think it gives much support to hedonism in particular.

2. What is meta-ethical hedonism?

In my experience, various strains of meta-ethical hedonism are quite common amongst utilitarian-ish moral realist-ish folks. Developed public arguments for the view, however, seem to me relatively thin on the ground – and many of these focus on a naturalist version that seems to me less interesting (more below).[4] I’m going to try to hone in on what I see as the core claims, and the core mistakes, but I won’t be able to cover all the possible variants.

I see meta-ethical hedonism as trying to do two types of work:

  • First, it’s trying to defend normative realism – and specifically, to defend it against objections to do with our epistemic access to the objective normative truths.[5]
  • Second, it’s trying to defend some form of hedonism about normativity – that is, some view to the effect that pleasure and pain are only things that matter intrinsically.[6]

The argument is generally something like:

  1. We have a distinctive sort of epistemic access to the value of pleasure – one that we lack with respect to other candidate values, norms, and so on.
  2. This sort of epistemic access is immune to the sorts of epistemic worries that otherwise plague normative realism.  
  3. Thus, we are justified in being normative realists, but only justified in believing in the value of pleasure.[7]

Why go in for this sort of argument? The core move is to draw a connection between (a) your epistemic access to your subjective experience (i.e., your phenomenology, “what it’s like to be you”) and (b) your epistemic access to the value of pleasure. Pleasure, after all, seems tied up with subjective experience somehow (though exactly how is a further question). And many people think that our epistemic access to our subjective states is both superior to, and more “direct” than, our epistemic access to the external world. So maybe we have that unique sort of epistemic access to the value of pleasure, too? And maybe this means we can still be normative realists after all?

Thus, consider your experience of that famously unproblematic object/property/way of experiencing/whatever: phenomenal red – the red of a ripe tomato, say, but as it appears in your mind (read: on the screen of your cartesian theater), rather than in the objective world. Aren’t you just so “up close and personal” with phenomenal red? Aren’t you just “confronting” phenomenal red? Isn’t the nature of phenomenal red just “laid bare” to your inner gaze? If someone were to be like: “nah, dude, phenomenal red isn’t the way it seems,” wouldn’t you be like “get out of here with such talk; phenomenal red just is the way phenomenal red seems; my epistemic encounter with phenomenal red is not subject to standard sorts of fallibility, like the sort created by distinctions between appearance and reality.”  

And if we say this, can we say the same about the value of pleasure? The meta-ethicist worries: “sure, pleasure seems good; but maybe pleasure is not as it seems.” Can you reply as follows? “Get out of here with such talk. The goodness of pleasure is like the redness of phenomenal red; I experience it directly; it is ‘laid bare’ before my inner gaze; it’s not subject to the normal distinctions between appearance and reality.”

This, at least, is the sort of thing that meta-ethical hedonists want to say. And they hope to make, from this, an argument like 1-3 above.

2.1 This isn’t about how much you like pleasure

Note the distinction between this argument and a different argument, namely: “Man, I just have such a strong intuition that pleasure is good. No way I’m doubting that. And I have a much stronger intuition that pleasure is good than [blah other normative thing].”

The point here is not about the strength of your conviction that pleasure is good (I’m happy to grant that our convictions about the value of pleasure – and even more so, the disvalue of pain – are unusually strong). The point here is about the source of this conviction – and in particular, whether its source is uniquely immune to normative realism’s epistemic problems, in virtue of some qualitative distinction between it and other sources of normative conviction.

Indeed, despite my disagreements with meta-ethical hedonism, I want to be clear that I, too, like pleasure a lot. In fact, I have written previously in opposition to those who denigrate the sort of optimized but not-obviously-appetizing pleasure (“hedonium”) that hedonistic utilitarians cherish:

pleasure which, for all the disdain with which some throw around the word, would appear to you, if you could experience it, not as a twitching lab rat on a heroin drip, but as something of sublimity and energy and boundlessness; something roaring with life and laughter and victory and love; something oceanic, titanic, breathtaking; something, indeed, beyond all this, far beyond. I am not a hedonist — but I think that casual scorn towards the notion of pleasure, especially actually optimal pleasure (not just ha-ha optimal, cold optimal, sterile optimal), is foolish indeed.

I say this partly because my sense is that some meta-ethical hedonists take themselves to be unusually impressed by the significance of pleasure and pain – to have seen more clearly than others that pleasure is just so good, and pain so bad. And perhaps they have. But I don’t think that’s an important crux, at least here.

Indeed, the crux isn’t, even, the plausibility of hedonism per se. I think there are plenty of interesting arguments for hedonism out there.[8] But we should be careful not to conflate our sympathy (or lack of sympathy) towards those arguments with an assessment of 1-3 above.

2.2 An example of a non-hedonistic intuition, for use in contrast

It will be useful, in what follows, to have an example of a non-hedonistic normative intuition on the table, so that we can compare its epistemic status with our intuitions about the value of pleasure. Consider:

Phoebe the physicist. By day, Phoebe is hard at work on what’s on track to be a Nobel Prize-winning theory of super-duper symmetry. By night, she spends time with her loving family. Phoebe has very strong views about experience machines. In particular, she strongly prefers living in the real world, with her real family, and to be understanding the real physics, over a somewhat-more-pleasurable life in an experience machine, with a fake, unconscious family, developing flimsy and confused theories that are false even about the physics of the machine world. In fact, Phoebe is an official signatory of a strange website called “DON’T YOU DARE PUT ME IN AN EXPERIENCE MACHINE WITHOUT MY CONSENT.com” – a choice she made after being declared sound-of-mind by ten leading psychologists. She also printed out a form with this same statement, signed it in the presence of the local notary, and hung it on her front door.  

But Allan the altruist knows better. Allan has been reading philosophy. He’s become quite convinced that actually, pleasure and pain are the only things that matter intrinsically. Plus, for some reason, he has ninja stealth and advanced tech. He sees that the experiences of his next-door neighbor, Phoebe, are on track to contain less value than they could with his help. So, one day, he drugs and kidnaps Phoebe while she’s asleep, puts her in an experience machine, and replaces her with an unconscious Phoebe-android-doll that her family will never tell from the original. Now, the real Phoebe floats all day in a glass vat – trapped, stuffed with tubes, wrong and foolish about physics (even the simulated physics), alone, surrounded only by simulated cardboard dolls she mistakes for persons, but with more pleasure, and less pain. If she could see her condition, she would be writhing in protest; clawing the tubes from her body; trying to break the glass with her fists. But Allan keeps her nice and quiet.

Wendell the witness lives across the street, and he can see through walls. Wendell watches the whole affair, and he has the strong intuition that what Allan is doing is wrong, even if it increases the net pleasure in the world – and in particular, that Allan has violated Phoebe’s autonomy is an absolutely unacceptable way. Wendell forms a strong desire to take an axe to the vat, and to set Phoebe free.

Now, obviously, hedonists will have a lot to say about why Allan shouldn’t have done what he did – and a lot, as well, about why Wendell’s intuition tracks valid instrumental considerations and heuristics grounded, ultimately, in hedonistic considerations.[9]Anyone who has ever argued with utilitarians about counter-examples will be familiar with this sort of (not-especially-satisfying) dialectic – and I’m not, here, interested in whether hedonistic attempts to deflate/debunk/rescue intuitions like Wendell’s succeed.  

Rather, I simply want to note what I think even most hedonists will concede: namely, that the intuition presents itself as granting intrinsic significance to something other than pleasure and pain. Indeed, for me at least, the intuition skips over questions about whether Allan made Phoebe’s life worse, or the world worse, and cuts directly to the way Allan has contravened Phoebe’s will. Phoebe’s life is her own. She gets to choose whether or not she goes into the machine. She chose “no” extremely hard – and Allan has done horrible violence to this “no.” The pleasure at stake is not the point.

Or at least, let’s say that this is how Wendell describes his intuition. And let’s say that the meta-ethical hedonist grants this description. Still, the meta-ethical hedonist will argue, this intuition is importantly and qualitatively different from Wendell’s intuition that the pleasure he gets from sinking into a hot bath is good. In particular (and regardless of other differences between the cases[10]), Wendell’s epistemic relationship to the pleasure involves the sort of directness and unmediated-ness and “laid-bare-ness” at stake with phenomenal red. Whereas, according to the meta-ethical hedonist, his epistemic relationship to Allan’s violation of Phoebe’s autonomy is somehow more “distant” and indirect – and thus, more subject to objection.

2.3 Setting aside some issues in philosophy of mind

Let’s call the special form of epistemic access being appealed to here “direct acquaintance.” For some people, including myself, this sort of appeal is already a red flag. In particular, it’s not clear what direct acquaintance consists in (though see here for some discussion),[11] how we’d tell if it exists, and what sort of backdrop conception of mind it assumes – so one worries about smuggling in various confusions and mistakes right off the bat. Indeed, my impression is that advocates of direct acquaintance often take it as a fairly brute notion.[12] Perhaps you can point at it via reference to its properties or its role in the rest of epistemology,[13] but it gets most of its credibility from some kind of pre-theoretic intuition that it’s a thing (indeed, apparently one defense of direct acquaintance is that we’re directly acquainted with it).

More generally, vibes to the effect that introspection gives us some qualitatively distinct/nigh-infallible/unable-to-be-doubted/“self-intimating” epistemic access to our mental states (including, even, our phenomenal states) have been subject to a variety of empirical and theoretical objections over the years.[14] I’m not going to wade into this debate here, but if I was into meta-ethical hedonism, I’d expect to have to grapple with it.

And as a more general red flag re: meta-ethical hedonism (depending on your orientation towards the philosophy of mind – c.f. illusionism), you might worry, as well, about the Cartesian vibes, the “sense data” vibes, the general centrality of qualia to the view in question, and the absence of physical/computational analyses of the main moving parts. Advocates of meta-ethical hedonism will emphasize its compatibility with a wide variety of views in philosophy of mind – and perhaps, technically, yes. But at least in some discussions of the topic, one senses a taste for more metaphysically inflationary (or at least, non-reductive) approaches to consciousness. After all, if consciousness is going to be the only thing that matters, and our epistemic access to consciousness is going to be super different from our epistemic access to other things, then it’s natural to want consciousness to be, well, really quite special and substantive stuff – not just, as it were, another way that matter can move around. (How would you get “direct acquaintance” with movements of matter, anyway? Via what mechanisms could such movements be “laid bare”?).

Indeed, I don’t think it’s an accident that two prominent advocates of (something like) meta-ethical hedonism (Rawlette and Pearce) have quite non-standard views about consciousness: Rawlette has a book arguing for the survival of consciousness after death (see also her work on synchronicity), and Pearce is some sort of idealist-physicalist,[15] who believes that classical computers can’t be conscious because they cannot solve the “binding problem.” Of course, regardless of whether we agree with Rawlette and Pearce on these fronts, we should assess the arguments for meta-ethical hedonism on their own terms (and the fact that some people treat “Cartesian” and “sense-data” and “qualia” like insults isn’t, actually, an argument either). But I think the connection between the appeal of meta-ethical hedonism and the appeal of more inflationary/non-standard views of consciousness is worth tracking, and I discuss it more in the final section.

In the meantime, though, I want to set these issues in the philosophy of mind aside, and basically just grant the meta-ethical hedonist some notion of direct acquaintance with phenomenal states – including, if necessary, phenomenal states construed in some non-reductive sense. This is partly because I do think it intuitive that our epistemic access to e.g. the nature of phenomenal red is in some sense uniquely direct (and I think non-reductionism about consciousness has intuitions in its favor as well). But more importantly, I think that meta-ethical hedonism fails even if we grant its basic assumptions about the philosophy of mind – and I think this failure should be our focus.

In this spirit, I also want to grant a claim that some meta-ethical hedonists seem especially excited to argue about: namely, the claim that there is some distinctive (and maybe: non-representational? not sure if this matters) phenomenal quality associated with pleasure. The thought is, roughly, that there is a particular way that pleasure feels – namely, “good” – and that it’s in virtue of something’s feeling this way that it’s a pleasure at all. This is the way of feeling that the hedonist treats as the only thing worth promoting, protecting, and so forth for its own sake. I’ll sometimes call this feeling “positive valence” or “positive hedonic tone.”

There are lots of other conceptions of pleasure in the literature as well (see here for discussion), and naively, it seems like meta-ethical hedonism will be on shakier ground if you adopt one of those (since if pleasure has no distinctive phenomenal quality, the analogy with our epistemic access to e.g. phenomenal red is weaker). But here, again, I’m happy to grant the meta-ethical hedonist the phenomenology she wants. The core problems lie elsewhere. Let’s turn to those (finally) now.

3. Meta-ethical hedonism for non-naturalists?

As readers of my other essays on meta-ethics might be aware, I tend to group meta-ethical views into two categories.

  • On the one hand, there are the robust, non-naturalist realists, who think that the normative domain is something genuinely and irreducibly extra, relative to the natural world (more here).[16]
  • And on the other hand, there’s everyone else – that is, all the views that accept that the natural world is all there is, and which focus on arguing about which part of it (if any) is the normativity part.[17]

Characterizing “natural,” here, is a bit tricky. Often, it means something like: “you know, the science-y stuff.” But because we are here going to be talking about consciousness a lot, and because consciousness’s status as part of the “science-y stuff” isn’t always clear, I think it’s probably best, here, to construe “natural” more broadly, as indicating the realm of the descriptive as opposed to the evaluative, “is” as opposed to “ought.”[18] Thus, on this definition, dualism about consciousness is compatible with naturalism. Yes, dualists posit stuff that is irreducibly over and above the physical; and maybe, even, stuff that the standard tools of science can’t touch. But the stuff in question is still fundamentally descriptive stuff, is-stuff. It’s the kind of stuff that the nihilist, at least naively (more below on possible disputes), can coherently accept into his/her world-picture, without thereby giving up her nihilism.

I see the question of whether normativity constitutes an irreducibly extra realm, over and above the natural world, as the most interesting one in meta-ethics. Once we answer “no” to that (as I think we likely should), it becomes unclear to me how substantive additional debate in meta-ethics becomes: we all agree on what exists, so it seems plausible that we’re mostly arguing about how to use words[19] – and even granted that it ends up making most sense to label some particular bit of the natural world “normativity,” it’s not clear what grounds the authority of that bit over our practical deliberation.[20]

What’s more, it’s centrally in the context of a non-naturalist meta-ethics that the most pressing epistemic problems for normative realism arise. In particular, a metaphysics that places normative properties beyond the natural world makes it extremely unclear how our natural minds ever get the right sort of epistemic access to them. Exactly how to frame this objection most perspicuously is a somewhat open question (I give my own account here), but as an intuition pump: consider that on non-naturalism, all of your evaluative beliefs, intuitions, and attitudes are the product of a causal history that is not sensitive to, explained by, or dependent on the existence or the distribution of the non-natural normative properties, since those non-natural properties are (at least according to the most standard forms of non-naturalist realism) “causally inert.” That is, if it were the case that the non-natural property of goodness supervened on “eating bricks” instead of on “pleasure,” or if there weren’t any non-natural properties at all, you’d have all the same evaluative attitudes that you do now. In this sense, your attitudes do not co-vary with the normative properties. Such properties live in their own separate, inaccessible realm, floating on top of (but never influencing or interacting with) the one where all your belief-formation takes place. And this makes it very hard to see how some specific set of beliefs about the content of this realm (i.e., “non-natural goodness supervenes on pleasure but not on eating bricks”) could be justified. And the same holds, indeed, for the belief that the realm exists at all.

This is the objection to non-naturalist realism that I take most seriously. And I’m most interested in meta-ethical hedonism insofar as it purports to show that in virtue of something about direct acquaintance, this objection does not apply to the goodness of pleasure. If true, this would grant new hope to non-naturalist realism.

But I don’t think it’s true.

3.1 The basic problem

The basic problem is that pleasure is just another natural thing. This is true even if we construe pleasure as a phenomenal property, and even if we construe phenomenal properties in a non-reductive, non-physicalist sense. Still, direct acquaintance with phenomenal properties will not grant us epistemic access to some further, non-natural realm of normative facts. All it gets us is more “is.”

Of course, we can posit an additional, yet-further epistemic faculty for accessing such a normative realm (e.g. “rational intuition”). But non-hedonists can do this, too – a focus on phenomenal properties, and on the directness of our acquaintance with them, is not required.

Alternatively, we can try to just state, baldly, that the same faculty that puts us into direct acquaintance with phenomenal properties also gives us access to the further, non-natural realm of normativity. But this is a substantive expansion of our usual notion of direct acquaintance – one that does not inherit whatever antecedent plausibility the notion had in the context of our access to our phenomenal states, and which raises the same problems that other posited forms of access to non-natural normativity (e.g., rational intuition) do.

Here I’ll go through these considerations in turn. Before jumping in, though, I want to acknowledge that I do feel sympathy for some vibe in the vicinity of the views I’m objecting to, here. In particular, I do think that the problems involved in positing any sort of epistemic access (whether direct or indirect) to non-reductively-construed phenomenal states mirror the problems involved in positing epistemic access to a non-natural normative realm, in a way that might prompt suspicion that they will stand or fall together. And more generally, consciousness seems sufficiently mysterious, from the perspective of a broadly reductionist worldview, that if such a worldview is wrong in a way that makes room for epistemic access to a non-natural normative realm, we might expect the mistake to be tied up with consciousness somehow. But in my opinion, the view that emerges from taking these sorts of considerations seriously – what I call the “mystery view” – does not look very much like meta-ethical hedonism. I discuss this issue more in the final section of the essay.

3.2 Some diagrams

For now, let’s look at some diagrams. Here’s a diagram of the basic epistemic problem for non-naturalist realism.

Figure 1: The basic epistemic for non-naturalist realism

That is, in brief, the problem is that the green “my epistemology” node is only touching the orange “whatever natural properties normativity supervenes on” node (and the “supervenience” arrow doesn’t transfer the “touching” property in the right way). Thus, you could erase the blue “non-natural normative properties” node, or you could move it (and its supervenience arrow) to hover over some other set of natural properties (e.g., “eating bricks”), and the green “my epistemology” node wouldn’t notice. Hence the red “no epistemic access” arrow between my epistemology and normative realm.

Now let’s see what happens if we add in the machinery of meta-ethical hedonism:

Figure 2: The picture suggested by non-naturalist meta-ethical hedonism

What’s changed? Well, we have a new (and more mysterious) type of epistemic relationship in the mix (“direct acquaintance”); we’ve taken a stand on the specific sort of natural property that the non-natural normative properties supervene on (namely, the phenomenology of pleasure); and we’ve got a new yellow arrow pointing at that property as well, which (if we construe it in terms of supervenience) seems a bit worrying given the problems that the yellow supervenience arrow in the last diagram created (though I’m leaving open the option of construing the phenomenology of pleasure in more reductionist/physicalist terms as well).[21]

But for all this, have we improved our epistemic position with respect to the non-natural normative properties? I don’t see how we have (though we’ve introduced a sufficiently mysterious form of epistemic access, and a sufficiently more complicated ontology, that we might be more liable to confuse ourselves). For example, if we wiped away the “non-natural normative properties” node, or moved it to hover over some other set of natural properties (i.e., “eating bricks”), the “my epistemology” node still wouldn’t notice. Perhaps direct acquaintance gives us some qualitatively different form of epistemic access to a certain type of natural property – namely, pleasure. But the goodness of pleasure is still floating on top of pleasure itself, off in that extra realm that our natural minds can’t touch. Appeals to direct acquaintance don’t change that. So the red “no epistemic access” arrow persists.

3.3 Dialogue with the nihilist

One way to bring this out is to note that it seems totally possible to accept everything in the “natural world” part of the diagram above, and to remain a nihilist. Thus:

Meta-ethical hedonist: “Dude, isn’t it just so true that pleasure involves positive valence?”

Nihilist: “Oh yeah, no doubt. Possible that’s true by definition or something.”

Meta-ethical hedonist: “And like, don’t you just confront pleasure’s positive valence, when you experience pleasure? Isn’t the positive valence just ‘laid bare?’”

Nihilist: “Sure, I guess. Not totally sure what you mean, but let’s go with it.”

Meta-ethical hedonist: “So how could you doubt that supervening on positive valence is an additional, irreducibly non-natural property of goodness?”

Nihilist: “Um… sorry, why would I think that in the first place?”

Meta-ethical hedonist: “Well, like, positive valence feels good, right?”

Nihilist: “Sure.”  

Meta-ethical hedonist: “And your epistemic relationship to positive valence is like your epistemic relationship to phenomenal red: you’re not going to be wrong about how it feels, or what about it’s like, right?”

Nihilist: “Sure.”  

Meta-ethical hedonist: “So, you’re not going to be wrong that positive valence is good!”

Nihilist: “I never said that positive valence is good. All I said was that it feels good. When I said that it feels good, I meant to refer to a descriptive feature of the world, an ‘is’-feature. And I’m happy to grant that I have some sort of especially-direct access to this sort of descriptive feature of the world. But by hypothesis, goodness construed as a non-natural, irreducibly normative property is something else entirely.”

That is, in the context of a non-naturalist meta-ethic, the meta-ethical hedonist risks equivocating between a descriptive understanding of the goodness of pleasure (i.e., the sense in which pleasure “feels good”) and a normative understanding (i.e. the sense in which pleasure is to-be-promoted, an intrinsic source of practical reasons, and so on). No one denies (or at least, no one in this debate need deny) that pleasure feels good in the former sense. But the whole point of non-naturalism is that descriptive claims like “pleasure feels good” are irreducibly different from normative claims like “pleasure is to-be-promoted” – indeed, it’s this irreducible difference that prompts the non-naturalist to posit a whole extra normative realm, on top of the natural world, to serve as the “territory” for our normative “maps.” And nothing has been said, yet, to give us access to this territory.

3.4 Positing access to the non-natural normative realm?

OK, but suppose we start trying to say such things. One option here would be to posit a distinct epistemic faculty that grants us access to the non-natural realm – for example, “rational intuition.”

Figure 3: Appeal to rational intuition?

This move brings its own problems (I discuss a few here), but for our purposes, the more important point is that the non-hedonist can make it just as easily. In particular, notice that the “direct acquaintance” and the “phenomenology of pleasure” bits of the diagram aren’t doing any work – something we can see by replicating the same move on our first diagram:

Figure 4: Non-hedonists can make that move too.

Thus, for example, you could just as easily use your rational intuition to get access to the fact that Allan’s violation of Phoebe’s autonomy (an example natural property that could be in the orange box) was wrong (an example normative property that could be in the blue box). Hedonistic judgments aren’t privileged.

Alternatively – and here I think we get closer to the spirit of meta-ethical hedonism – we could try to extend the faculty of direct acquaintance to include direct acquaintance with non-natural, normative properties as well. Thus, in a diagram:

Figure 5: Direct acquaintance with non-natural properties?

Here, our introspection itself slips the surly bonds of the natural world. When we look inside our minds, what we “confront,” what is “laid bare,” is not simply what it’s like to experience pleasure – or more broadly, to be us. Rather, we confront, directly, that extra, mysterious realm of what should be, the realm of ought, must, value, reasons. Pleasure’s non-natural to-be-promoted-ness is presented to us the way that the brightness of phenomenal yellow is presented to us – not merely in a manner that prompts unusual confidence about it (Wendell, after all, might feel quite confident that Allan’s violation of Phoebe’s autonomy was wrong), but in a manner sufficiently “unmediated” as to leave no room for distinction between its being presented to us and its being the case, between its appearance and its reality.[22]

I think that to the extent meta-ethical hedonists want to be non-naturalists, this is the version of the view they should go for. But now, note, we’ve entered into more controversial and extravagant territory than we did when we posited that introspection gives you direct acquaintance with phenomenal properties. Admittedly, we didn’t really know what we meant, when we posited this – unclarity that one worries is serving as a kind of “realism of gaps,” wherein the non-naturalist realist seizes on some especially obscure part of our epistemology and says “that super-obscure part is where the epistemic access to the non-natural normative realm happens” (more on this in section 5). But even setting this unclarity aside: to the naïve ear, “introspection gives us direct acquaintance with what it’s like to be us” sounds a lot more intuitively plausible than “introspection gives us direct acquaintance with non-natural, irreducibly normative properties, over and above what it’s like to be us.”

Granted the first, has the meta-ethical hedonist given us any reason to accept the second? One argument might be:

  1. Pleasure has the property of being good in a non-natural, irreducibly normative sense.  
  2. You have direct acquaintance with pleasure.
  3. Direct acquaintance with pleasure involves direct acquaintance with all of pleasure’s properties.
  4. Thus, in being directly acquainted with pleasure, you are directly acquainted with its non-natural goodness.

But where did we get premise 3? Maybe the thought is: pleasure is a phenomenal thing, so its properties are exhausted by its phenomenal presentation? But (on non-naturalism) pleasure’s goodness isn’t a phenomenal thing – rather, it’s an irreducibly normative thing, supervening on a phenomenal thing. Maybe direct acquaintance gets us into contact with all the phenomenal properties of a phenomenal thing[23] – but why think it gets us to the normative properties floating, on top, in another realm?

Indeed, my sense is that the argument that actually drives (non-naturalist) meta-ethical hedonists is something more like:

  1. When I look at pleasure in my experience, I’m like “man, that is so good.”
  2. Also, when I look at phenomenal yellow in my experience, I’m like “man, that is so bright.”
  3. When I do this with phenomenal yellow, I’m on uniquely solid ground, because I’m in direct acquaintance with phenomenal yellow.
  4. Also, when I do this with pleasure, I’m in direct acquaintance with pleasure.
  5. Thus, my judgment that pleasure is good is on the same, uniquely solid ground as my judgment that phenomenal yellow is bright.[24]

But even granted 1-4, 5 doesn’t follow. In particular, we’ve been assuming that the reason your judgment about the brightness of phenomenal yellow is on uniquely solid ground is that in being directly acquainted with phenomenal yellow, you’re directly acquainted with its brightness as well. But, as before, the brightness of phenomenal yellow is a phenomenal property, whereas (by hypothesis, on non-naturalism) the goodness of pleasure is not. So the two judgments would only be on comparably solid ground if, in being directly acquainted with pleasure, you were also directly acquainted with the realm of irreducible normativity. And that’s precisely the question at issue – a question that the meta-ethical hedonist has thus far not illuminated.

We can imagine generating other possible stories about direct acquaintance, for use in attempting to justify the claim that it lays bare not just the phenomenal domain but the extra, irreducibly normative domain as well. Indeed, much of Bastian Stern’s dissertation is devoted to generating, examining and shooting down various options in this vein. The need for “generation,” though, derives (at least as far as I can tell) from the comparative dearth of serious efforts, in this respect, on the part of meta-ethical hedonists themselves (the published defenses of something-like meta-ethical hedonism I’ve seen are in a more naturalist meta-ethical vein). That is, many meta-ethical hedonists seem to assume that once they’ve been granted direct acquaintance with phenomenal properties, they can get direct acquaintance with a distinct set of irreducibly normative properties as well, for free, without saying how and why the former extends to the latter, or giving much of a story of how direct acquaintance works more broadly. And I don’t think that’s right.

Plus, even if we got ourselves into the idea of direct acquaintance as a source of access to the irreducibly normative domain, this idea would lead to the same problems we run into when we appeal to “rational intuition” as a source of such access – except, without partners-in-guilt like math, philosophy, and general synthetic a priori reasoning as comforts. Notably, for example (and even setting aside objections to do with the mysteriousness of the faculty and form of access at issue), we would be giving up on the causal (or at least, the explanatory) inertness of the irreducibly normative, and positing a source of explanation of our judgments not countenanced by science – a source which plausibly makes empirical predictions about the moral views of aliens, super-intelligent AI systems, and the like, and which needs to find some place amidst a physical world generally thought to be causally closed. If we’re up for paying these sorts of costs, in the context of an otherwise mysterious epistemic faculty, then plausibly, we should be up for them in the context of rational intuition as well, in which case we can begin to diffuse the epistemic objection to non-naturalism without any special focus on direct acquaintance.

I’ll have more to say about moves in this broad vein at the end of the essay. For now, let’s move on to naturalist variants of meta-ethical hedonism.

4. Meta-ethical hedonism for naturalists?

Thus far I’ve been focused on a version of meta-ethical hedonism that attempts to defend meta-ethical non-naturalism. But perhaps some meta-ethical hedonists out there have been frustrated: “Obviously, I never meant to claim something so metaphysically extravagant as non-naturalism in meta-ethics. No, no, my normative realism is quite scientifically respectable, I assure you.” And indeed, various published views in the vicinity of meta-ethical hedonism – notably, Rawlette (2016) and Pearce (1995) – are in a more naturalist vein.[25]

4.1 Verbal disputes?

I want to briefly discuss this version of meta-ethical hedonism, but I admit that I have less energy for it. This is partly because, as I mentioned above, I tend to see the distinction between non-naturalism and naturalism as more important than the distinctions between different naturalist meta-ethical views (e.g., naturalist realism vs. subjectivism vs. nihilism). In particular, I worry that the latter are often verbal. After all, the whole point of naturalism is that the world can be exhaustively described in purely non-normative terms. So then what are we arguing about, except the best way to use some additional set of terms – namely, the normative ones? What ultimately exists isn’t at issue. We could taboo all the normative terms, without loss of any ability to describe the world accurately (though perhaps, with some loss of convenience).

Indeed, talking with the straw naturalist meta-ethical hedonist in my head, it can feel like the main thing they’re telling me is “no, really, there exists such a thing as positively valenced experience,” and once I grant that, which I’m fairly happy to do, I cease to have a clear grip on what we’re arguing about. Did the nihilist ever need to disagree, for example, that positively valanced experience exists? Yet it can feel like my straw naturalist meta-ethical hedonist takes this fact to have disproven nihilism.[26] Maybe this is just what it feels like to be a naturalist about meta-ethics? I.e., if I stamp my foot hard enough about equating goodness with blah-natural-thing (e.g., positively valenced experience), it ceases to be possible to accept the existence of blah-natural-thing, but remain a nihilist? News to the nihilists, to be sure. Then again, the nihilists were always non-naturalists at heart, and someone has to be confused…

4.2 Is there any distinctive epistemic problem for naturalist realism?

Beyond these worries about verbal disputes, though, there’s another reason I’m less interested in naturalist versions of meta-ethical hedonism: namely, objections to do with how we can get the right sort of epistemic access to normative properties seem much less pressing, conditional on naturalism about meta-ethics.[27] After all, on naturalism, normative properties are natural properties, so our epistemic diagram (regardless of whether we’re hedonists) looks more like:

Figure 6: The epistemic benefits of naturalism about meta-ethics

To flesh this out a bit more: many naturalist forms of moral realism rest heavily on a story about the concept of goodness.[28] Thus, for example, the naturalist might claim that it is an analytic truth, akin to the truth that all bachelors are unmarried, that goodness is blah-natural-thing (for example, pleasure). That is, on this view, to say that something is good just is to say that it’s pleasant. Goodness is pleasure because that’s the definition of goodness.

Alternatively, the naturalist’s conceptual story might pick out the set of natural properties identical with goodness in some more indirect way, by saying that goodness is identical with whatever natural property meets blah-conditions: for example, causally regulating our use of the term “good,” or playing blah-functional role in “mature folk morality,” or being such that blah-people would have blah-attitudes towards it under blah-conditions.

Either way, once the naturalist’s conceptual story is in place, all the rest of the epistemic work, in figuring out which things are good, is of a quite familiar, empirical type. Thus, if goodness is pleasure by definition, then figuring out what’s good just means: figuring out what’s pleasant. Or if goodness is “whatever property causally regulates our use of the term ‘good’”, we can get out there and try to figure out what that property is, and what has it, too. In both cases, we’re in a vastly superior epistemic position than we were in, in the non-naturalist section above, when we thought that goodness was “that irreducibly extra property of to-be-promoted-ness, beyond the natural world, that supervenes on some stuff but not others.” Back then, even after we knew what we were looking for (did we?), we had no clue how to find it. We kept scraping our fingers against the edges of the natural world, trying to detect the location of something beyond it – but we could never break through.

So at least compared to the non-naturalist, normative epistemology has never been the naturalist’s main problem (rather, the naturalist’s main problem has always been something like: normative properties seeming irreducibly different from natural properties). But this leaves meta-ethical hedonism, at least as I’ve characterized it, with a much weaker motivation in a naturalist context. What is the epistemic problem that it solves for pleasure, but to which other candidate intrinsic values still fall prey? And what role does direct acquaintance have in the solution?     

I’m not sure. Indeed, my vague sense is that meta-ethical hedonists often want to be naturalists, but they don’t necessarily want to offer a specific account of how their naturalism works – of what makes it the case that “goodness” picks out one natural property vs. another.[29] But this makes it hard to evaluate exactly what work a given appeal to direct acquaintance with the phenomenology of pleasure is supposed to do, but which our epistemic relationship to Allan’s violation of Phoebe’s autonomy can’t do. Is direct acquaintance with the phenomenology of pleasure supposed to help us figure out the definition of goodness? Is it supposed to help us decide between different indirect ways “goodness” might pick out a natural property as a referent – e.g., causal regulation, functional role in mature folk morality, etc? Is it supposed to help us in the empirical project of figuring out what those natural properties are (e.g., which natural properties causally regulate our use of the term “good”), and which things have them?

Meta-ethical hedonists rarely say. And regardless, none of these options sound to me attractive. I won’t, here, attempt to canvass all possible ways of pursuing them, but I will look at one of them – Rawlette’s – in more depth, in an effort to illustrate what seems to me a broader barrier.

4.3 Rawlette’s view

Of the published, naturalist views in the vicinity of meta-ethical hedonism, Rawlette (2016) goes the furthest towards explaining the sort of naturalism she has in mind. Her view is a version of “goodness = pleasure by definition.” In particular, she thinks that our concept of goodness (and indeed, of normativity more broadly) comes from the phenomenology of pleasure.

I’m not sure exactly how Rawlette’s theory of concept acquisition goes, but here’s one attempt at reconstruction. First, you wake up in the world without the concept of goodness or normativity.[30] Then you have that particular sort of raw phenomenology that experiences of stuff like hot baths, cookies, and sex have in common – phenomenology that, for Rawlette, does not itself involve anything being represented as good or normative (since you don’t yet have a concept of goodness or normativity).[31] Then, you form some concept of this way of feeling, the same way you form a concept of phenomenal red when you see it for the first time. And, um, that’s your concept of goodness. That is, goodness just means something like “that special way that pleasure feels.”[32]

True, maybe you go around foolishly applying that concept (and concepts derived from it, like “wrongness”) to non-experiential stuff – e.g., things that experience machines are missing, Allan’s violation of Phoebe’s autonomy, etc. But that’s because of mistakes involved in mis-associating your internal experiences of pleasure/pain phenomenology with the external world. I.e., in the same way you end up thinking that redness exists in external objects, rather than only in your mind, you end up thinking that “that special way that pleasure/pain feels” exists in the external world (even apart from its existence in other minds), when in fact it does not.

Of course, this view requires biting various bullets. Notably, for example, you have to say (in the face of “open question” arguments to the contrary) that “this instance of pleasure is not even pro tanto good” is literally incoherent – a view Rawlette defends in Chapter 4. And coming to think that “that special way that pleasure feels” exists in external, unconscious objects is a strange mistake. Still, I think it’s an interesting view in various ways, and worth consideration.

For present purposes, though, what matters is that once the full view is spelled out, we can see that its structure and motivation don’t actually fit very well with meta-ethical hedonism as I’ve characterized it.[33] In particular, all the important work, in support of hedonism, is being done by Rawlette’s story about the definition of goodness – a story that our direct acquaintance with the phenomenology of pleasure doesn’t do much to help us assess. That is, Rawlette’s argument is basically:

  1. There is a certain special way that pleasure feels, which you have direct acquaintance with.
  2. Also, “goodness,” by definition, is identical with this special way of feeling.

And it’s true that from this, you can get to hedonism – and indeed, to the claim that you have direct acquaintance with the goodness of pleasure (this, it turns out, is just a restatement of 1 – a way of saying that you’re directly acquainted with the pleasantness of pleasure). But 2 is doing all the heavy lifting – and you don’t have direct acquaintance with the truth of that.

Thus, imagine trying to use Rawlette’s argument to persuade Wendell the witness that he should discard his intuition that Allan’s autonomy violation was wrong, but keep his intuition that the pleasure of his hot bath was good.

Wendell: I have both of these intuitions pretty strongly. Why should I give up one, but not the other?

Hedonist: You’re directly acquainted with the pleasantness of your hot bath experience. But you’re not directly acquainted with the autonomy violation – rather, it’s out there in the external world.

Wendell: OK, sure, but why does that matter?

Hedonist: One more thing: goodness just is pleasantness, by definition.

Wendell: Oh, huh. That doesn’t sound right. Why would I think that?

Hedonist: Moore’s open question argument doesn’t show that the claim I just made is false. Also, if we accept it, we have the makings of a form of moral realism that satisfies various nice criteria. I’ll send you a PDF of Rawlette (2016), Chapter 4.  

Wendell: Wait, I thought you were going to make an argument in favor of keeping my pleasure intuition, but not my respecting-autonomy intuition, because direct acquaintance gives me special access to the truth of the one, but not the other.  

Hedonist: I am! You’re directly acquainted with the goodness of pleasure, and in that sense directly acquainted with the truth of the claim that pleasure is good. But you’re definitely not directly acquainted with the wrongness of the autonomy violation. Indeed, we know from the very definition of goodness that the autonomy violation was goodness-promoting (also, we’re consequentialists, because reasons).  

Wendell: But how do I know that that’s the right definition of goodness?

Hedonist: Let me send you this PDF.

Wendell: But is direct acquaintance going to give me access to the truth of the detailed philosophical claims in that PDF?

Hedonist: Um… maybe it will make them seem plausible? Like, maybe you can look in your experience, check out what pleasure feels like, and then be like: “yeah, my concept of goodness just is something like ‘that special way pleasure feels’; the raw, non-representational phenomenology of pleasure is the very source of my concept.”

Wendell: Hmm, that sounds like a pretty specific thesis about my concept acquisition, which seems hard to evaluate via introspection alone. And regardless, it doesn’t sound like a phenomenal thing, and phenomenal things were the things I was ready to grant that I have direct acquaintance with.

Hedonist: OK, but you do also have direct acquaintance with the goodness of pleasure.

Wendell: If I accept your definition of goodness. In your language, you’re basically just repeating the claim that I have direct acquaintance with the pleasantness of pleasure.  

Hedonist: Yes, that’s right.

That is, if you accept Rawlette’s naturalist hedonism, then you get to the claim that you have direct acquaintance with the goodness of pleasure. But your direct acquaintance with your phenomenology does not, itself, give you much reason to accept Rawlette’s hedonism.

4.4 This problem generalizes

Of course, Rawlette’s is only one form of naturalist hedonism. But even without canvassing all of the other options, I expect this problem to generalize. That is, while it’s possible to describe various forms of naturalistic, hedonistic moral realism on which we have direct acquaintance with the goodness of pleasure, it’s a substantially further step to say that direct acquaintance with pleasure gives us any reason to accept such a view. And indeed, in general, direct acquaintance seems poorly suited to do the type of work that a naturalist realist needs to do, in arguing for one normative-ethical view over another. This work, after all, is either centrally conceptual (i.e., figuring out the true story about how the concept of goodness picks out its referent) or empirical (figuring out which stuff has what natural properties) – and by granting that we have direct acquaintance with our phenomenology, we have not granted that we have direct acquaintance to the relevant truths in these conceptual or empirical domains.

Indeed, in a sense, the naturalist’s problem, here, is the same problem that the non-naturalist had in section 3. In both cases, we granted direct acquaintance with the pleasantness of pleasure. In both cases, though, what was needed was an argument for some specific connection between pleasure and goodness (in the case of non-naturalism: supervenience; and in the case of naturalism: identity). And in neither case did direct acquaintance give us epistemic access to a connection of this form (even if we can argue for such a connection on other grounds).

5. The mystery view

This hasn’t been an exhaustive survey of all the ways you might try to get some form of meta-ethical hedonism going. But I hope that the basic challenge is fairly clear. Ultimately, from the perspective of normative epistemology, positing direct acquaintance with phenomenal states changes little – at least on its own. All it gives you is a (mysterious) form of epistemic access to a certain bit of the natural world – a certain bit of the “is.” But figuring out what’s up with the “ought” remains as hard as it ever was – and in the case of non-naturalism: too hard.

Perhaps we might wonder, though: what explains the prima facie appeal of meta-ethical hedonism, if it’s such a non-game-changer? Well, at least to a certain type of a person, I’m not actually sure the view has much prima facie appeal. That is, a certain type of meta-ethicist, attuned to the “is-ought” gap, will be immediately pessimistic that this gap can be crossed or dissolved via our epistemic access to our phenomenology. The question was never whether pleasure is pleasant.

To the extent meta-ethical hedonism does have prima facie appeal, though, I think that part of it comes from the fact that our intuitions about the goodness of pleasure are, indeed, unusually strong and hard-to-doubt; that those intuitions are formed in response to introspection on the pleasantness of pleasure; and that it is indeed intuitive that our epistemic access to the pleasantness of pleasure is in some sense uniquely direct. So it’s easy to smoosh it all together and to say that you’re getting the same sort of epistemic access to pleasure’s goodness that you’re getting to pleasure’s pleasantness – especially if you’re up for equivocating between the descriptive sense in which pleasure is good (i.e., it feels good) and the normative sense (i.e., it’s to-be-promoted), and/or saying that these two senses are the same due to the truth of a naturalist hedonism that it’s easy to forget you’re supposed to be in the midst of justifying.

But I think there’s another source of appeal as well, which I don’t think is wholly misguided.

Above I mentioned a worry about a “normative realism of the gaps.” On this worry, the (non-naturalist) normative realist, scrambling to find some way of holding on to non-natural, causally-inert, somehow-known-about normative properties in the face of an otherwise naturalistic world picture, seizes on the aspects of that picture that remain most confusing and/or incomplete, and tries to use then as a wedge to make room for realism. Thus, perhaps: “Consciousness! It’s still super confusing and seems pretty different from other physical stuff. And we’ve got some kind of confusing and different form of epistemic access to it? Thus, maybe, that’s where the normative realism thing happens too?”

In a bad form, this sort of reasoning serves to merely multiply confusions. Here I’m reminded of cases in which someone groping for a theory of consciousness decides to appeal to quantum mechanics, despite not really understanding how quantum mechanics works (in certain sorts of conversations, if someone says the word “quantum,” you can tell things are heading downhill).[34] Mixing not-understood-thing A with not-understood-thing B rarely leads to additional clarity. And to the extent one is attached to A (in this case, non-natural normativity), one might then end up unduly motivated to keep B (in this case, consciousness) not-understood, in the same way God-of-the-gaps theists can end up unduly motivated to resist the onward march of science.

Then why is such a mixing-together so tempting? Chalmers (1995) suggests “the Law of Minimization of Mystery: consciousness is mysterious and quantum mechanics is mysterious, so maybe the two mysteries have a common source.” But as the example of consciousness and quantum mechanics makes clear: sometimes two things are mysterious and also still very different from each other. Thus, as Chalmers notes, quantum mechanics does not, itself, offer any additional resources for crossing the “explanatory gap” between the physical/functional and the phenomenal – the gap that makes consciousness mysterious in the first place.[35] And in a sense, the realist who seeks after normativity in the realm of the phenomenal runs into a similar problem. Maybe the normative and the phenomenal are both different from the physical – but they’re different from each other as well. So positing blah type of epistemic access to the latter leaves the former untouched. The relevant gap (this time: is vs. ought, rather than functional vs. phenomenal) has not been crossed.

That said, I don’t think an interest in connections between normative epistemology and the epistemology of (phenomenal) consciousness is wholly misguided. In particular (and in disanalogy with connections between quantum mechanics and consciousness), both non-naturalists about normativity and dualists about consciousness face a structurally similar epistemic objection. Both of them posit something that floats above the physical world, and hence outside the network of physical causes that usually give us knowledge about stuff. And in both cases, this leaves it worryingly unclear how we could ever get epistemic access to the existence or the distribution of the relevant floating stuff. Naively, that is, one expects to end up saying “I am conscious,” with your physical mouth, because you are conscious; and to end up saying “pleasure is good,” with your physical mouth, because pleasure is good. But it doesn’t sound like you do, if consciousness and goodness are outside the normal nexus of physical causes that generate your mouth movements – and thus, it doesn’t sound like the consciousness/normativity utterances produced by your mouth would be reliable indicators of the things they’re supposed to be about. I wrote about this objection in the context of non-natural normativity here. For discussion of it in the context of dualism about consciousness, see e.g. Yudkowsky (2016), and Chalmers (2018) on “the coincidence problem.”

One option, here, is to become a reductionist about both normativity and consciousness. And indeed, I think that’s probably the way to go. But many people, including me, have a harder time being reductionists about consciousness (for example, reductionism about consciousness can lead to illusionism in the same way that reductionism about normativity can lead to nihilism, but most people just aren’t up for illusionism – and I don’t blame them). To the extent they try to hold on to a non-reductionist theory of consciousness, then, while also holding on to the idea that you say that you’re conscious because you’re conscious, they are implicitly assuming some kind of funky set-up where the non-physical consciousness stuff still explains your mouth movements despite its wholly-non-physical character. That is, they’re implicitly assuming some answer to the epistemic problem for dualism about consciousness – even if they can’t say exactly what it is. But if we have an answer like that, then it seems very non-crazy to wonder whether it might apply to the epistemic problem for non-natural normativity as well, given the structure similarities at stake (see here for a description of what that might look like) – and thus, to wonder whether the two non-reductionisms might stand (rather than fall) together.

In my head, I call this view “the mystery view,” due to its appealing to a mysterious answer to the epistemic objections in question, and to a broader connection with vibes about philosophical humility, noticing when you’re still confused, and not-discarding-important-stuff just because it feels clever. At some point, I’m hoping to write more about the strengths and weaknesses of the mystery view as I see it (it’s my main meta-ethical alternative to a more reductionist naturalism/anti-realism/subjectivism).

For now, though, the main thing I want to note is that while the mystery view bears some superficial resemblance to meta-ethical hedonism (both are interested in connections between the epistemology of consciousness and the epistemology of normativity), they’re actually quite different, and I don’t think the mystery view does much to support hedonism.

In particular, even while the mystery view draws support for non-naturalist realism from consciousness as a partner-in-guilt, it also draws support from a variety of additional domains/partners-in-guilt – notably, mathematics and synthetic a priori reasoning more generally – in which somewhat similar epistemic problems arise. In this sense, it makes room for the relevant solution to these problems (to the extent there is a unified, non-reductionist solution) to apply beyond the realm of the phenomenal. Maybe, in some sense, it ends up being the “same faculty” (something something rational intuition?) that accesses the normative facts, the phenomenal facts, the mathematical facts, the synthetic a priori facts, and so on – or maybe not. But regardless, my best-guess is that phenomenal stuff doesn’t end up privileged, and so “pleasure is good” doesn’t end up privileged, either, relative to other intuitions like “Allan’s violation of Phoebe’s autonomy was wrong.”

Put another way, and even regardless of whether further domains like math and a priori reasoning get included: the mystery view notices a structural similarity between the “ways of knowing” required for different non-reductionist views to withstand epistemic critique, and so suggests that maybe this way of knowing is possible even if we don’t have a good story for how it works. But it doesn’t say much about the content of what gets to be known, in those different domains – rather, it just says that knowledge, in all of them, is possible. Whereas meta-ethical hedonism says, specifically, that it’s only with respect to phenomenal stuff that normative knowledge is possible.

Beyond this, though, in my head there’s a broader difference in vibe between the mystery view and meta-ethical hedonism. In my head, the mystery view is trying, in a sense, to do less philosophy – to hold onto some basic, important-seeming stuff, even if we don’t have a super clear story about how, lest we become too-clever-by-half, and lose something precious in the process. Whereas meta-ethical hedonism, in my head, is trying to do more philosophy – to tell a very specific story about a specific domain (the phenomenal) in which we’re allowed to have normative knowledge, and thus to fit our minds and our ethics to a more procrustean bed. That is, meta-ethical hedonism feels like it’s missing some element of the humility thing I mentioned above (even if its otherwise humble about e.g. our understanding of phenomenal consciousness). But I’ll leave explaining and justifying this “humility thing” for another time.

In closing, and regardless of how we feel about the mystery view: I think the hedonists should be humbler than meta-ethical hedonism tries to leave them. Indeed, I wrote this post partly because I have a sense that some hedonist normative realists, upon encounter with the epistemic objections to normative realism, have some inner sense that “those don’t apply to me, because I get my normative knowledge from a special source – namely, direct acquaintance.” Indeed, I worry that they emerge from this encounter bolstered in their confidence in hedonism, by a perception of having survived an assault to which the non-hedonists succumb. I think they’re wrong about this: they’re in the same meta-ethical boat as the rest of us.

  1. ^

    For convenience, I’ll focus on the value of pleasure going forward. All of my comments can be likewise applied to the disvalue of pain.

  2. ^

    The term “hedonism” is sometimes used to refer to the view that well-being is entirely a matter of pleasure and pain, but here I’m using the term more broadly, to refer to the view that pleasure and pain are the only (intrinsic) sources of normative reasons, period (all the rest of the normative domain is some combination of an instrumental and a conceptual matter). This view is often taken to imply hedonistic utilitarianism, but it could in principle lead to a more egoistic hedonism as well – including “person-moment” egoism, on which any given person moment’s only terminal goal should be to maximize its own hedonistic satisfaction (a form of egoism less subject than the traditional version to objections from reductionism about personal-identity).

  3. ^

    Nothing unique, that is, from the appeal to direct acquaintance alone. Of course, we can make other arguments to try to bridge or explain away this divide (indeed, as I suggest in section 4.3, this is how I understand the argument in Rawlette (2016), which rests centrally on a story about the concept of goodness). But it’s important to keep track of which bits of a given argument for hedonism are doing what work. More on this below.

  4. ^

    The most sustained defense of something in the vicinity I’ve seen is Sharon Hewitt Rawlette’s (2016) book The Feeling of Value: Moral Realism Grounded in Phenomenal Consciousness (podcast discussion here) – though as I’ll discuss below, I don’t actually think that Rawlette’s view leans very heavily on the distinctive type of epistemic access we have to our phenomenology. Neil Sinhababu has a few nearby papers (notably, “The Epistemic Argument for Hedonism,” though see also his more recent “Pleasure is Goodness; Morality is Universal”). Adam Lerner has an unpublished paper – “Fine Tuning Evolutionary Debunking Arguments” -- discussing a view like this, which he was kind enough to share with me. David Pearce’s The Hedonistic Imperative discusses similar themes in Chapter 2, Sections 2.13-2.20 (see also his comments in this podcast). Indeed, I’ve wondered how much of the amorphous meta-ethical hedonism pervading various of my social circles comes from Pearce (or perhaps, from some other originator-who-influenced-Pearce; perhaps one popular on the Felicifia forum back in the day – so much of contemporary discourse in effective altruism, rationalism, transhumanism, etc originated on old forums, mailing lists, and so on: do not underestimate). My colleague Bastian Stern also wrote his dissertation objecting to something closely akin to meta-ethical hedonism. He cites a few other possible advocates in a footnote on page 6 – though at a glance, the citations look more like “someone made a comment suggesting they maybe endorse a view like this,” rather than “someone actually defended this view in any depth.”

  5. ^

    Here I’m understanding normative realism as the claim that normative facts are objective – a claim compatible with their being reducible to facts about the natural world (naturalism) or not (non-naturalism).

  6. ^

    Sometimes, the claim here is a bit weaker – namely, that pleasure and pain are the only things that we know matter intrinsically – but this distinction won’t matter much in what follows.

  7. ^

    This isn’t the most precise statement of the argument (e.g., you aren’t necessarily justified in being a normative realist overall, even if you can answer one of the epistemic objections to realism), but hopefully the gist comes through.

  8. ^

    See e.g. Bradley (2011) for one example.

  9. ^

    For an example of this sort of discussion, see Rawlette’s 80,000 hours podcast.

  10. ^

    Thus, for example, the question here is not whether the hedonistic intuition might be more reliable because e.g. the autonomy violation occurs in the context of a crazy-thought-experiment type thing, whereas hot baths are more normal; or because it’s more easily debunked; or because it’s weaker; etc.

  11. ^


    From the SEP: “We said above that what distinguishes the classical, Russellian notion of acquaintance is, minimally, that (i) it is a non-intentional form of awareness: acquaintance with something does not consist in forming any judgment or thought about it, or applying any concepts to it; and (ii) it is real relation requiring the existence of its relata; one cannot be acquainted with some thing, property or fact that does not exist. Let us use the label “acquaintance theory” broadly to stand for any view, in the philosophy of mind or epistemology, in which acquaintance so understood plays a central role.”

    See also Chalmers (2010, p. 286): “Acquaintance can be regarded as a basic sort of epistemic relation between a subject and a property. Most fundamentally, it might be seen as a relation between a subject and an instance of a property: I am most directly acquainted with this instance of phenomenal greenness. This acquaintance with an instance can then be seen to confer a derivative relation to the property itself. Or, in the experience-based framework, one might regard acquaintance as most fundamentally a relation between a subject and an experience that confers a derivative relation between the subject and the phenomenal properties of the experience. I will usually abstract away from these fine details, however. What is central is the shared feature that whenever a subject has a phenomenal property, the subject is acquainted with that phenomenal property.”

  12. ^

    From the SEP: “Acquaintance is typically taken to be simple and thus indefinable. The philosophers for whom acquaintance plays a critical philosophical role are usually unapologetic about their inability to define the concept (e.g., Fumerton 1995: 76–7). They will quite plausibly argue that analysis requires conceptual ‘atoms’—simple ideas out of which other ideas are built. The concept of acquaintance, for them, is arguably the most fundamental concept on which all epistemology, and perhaps all philosophy of mind, is built. Still, one philosopher’s conceptual atom is another’s complete mystery, and unless one can convince oneself that one understands what acquaintance is, one will not be able to understand views that invoke it.”

  13. ^

    Chalmers (2010, p. 286) suggests that we can view acquaintance as “a theoretical notion that is inferred to give a unified account of the distinctive conceptual and epistemic character that we have reason to believe is present in the phenomenal domain.”

  14. ^

    For example: if your beliefs about a mental state, and the mental state itself, are distinct entities, why can’t they come apart? What about someone who, expecting to be burned, momentarily misclassifies a cold sensation as hot (h/t Chalmers)? What about introspective illusions, like the illusion that your visual experience is clear and high resolution throughout most of your visual field, when in fact it rapidly starts to suck as you reach the periphery? What about all those empirical studies (did they replicate?) of people who seem be super wrong about why they’re doing/saying what they’re doing/saying? What about Freud? (Did Freud replicate?) What about cases where you start out cold, then gradually become warmer, such that something something maybe your belief that you’re cold is unsafe or something and so doesn’t always count as knowledge according to a highly specific discourse in philosophy that it’s not clear you care about? (See “luminosity” -- 1, 2, 3.) What about speckled hens?

  15. ^

    Wikipedia: “British philosopher David Pearce defends what he calls physicalistic idealism (‘the non-materialist physicalist claim that reality is fundamentally experiential and that the natural world is exhaustively described by the equations of physics and their solutions’) and has conjectured that unitary conscious minds are physical states of quantum coherence (neuronal superpositions).”

  16. ^

    My paradigm advocate of this view David Enoch. See also Derek Parfit, Thomas Nagel, and many others, though the “robustness” of their realism isn’t always clear.

  17. ^

    I’ll generally be assuming cognitivism – i.e., the view that normative claims are candidates for truth or falsehood – in what follows, for both naturalist and non-naturalist views.

  18. ^

    “These distinctions are problematic for blah reason!” “I have cached the idea that these distinctions are problematic even though I don’t actually have a good account of why!” Maybe, maybe. But I think some distinction in the vicinity is actually animating a huge amount of meta-ethical discourse, so for now I’m going to use it.

  19. ^

    Compare the debate between the second and third tribes here.

  20. ^

    Bastian Stern’s dissertation has a nice articulation of this: “I personally care deeply about whether pleasant and unpleasant experiences are, respectively, good and bad in the way in which they strike us as good and bad; if pleasure and suffering really come with this kind of evaluative property, then I want to promote pleasure and end suffering. In contrast, I find it hard to see why anyone should have significant de dicto concern for things that are good in some more deflationary sense. For instance, I am not particularly concerned to promote things qua thing-which-instantiate-the-property-which-causally-regulates-our-use-of-“good”, or qua thing-considered-good-by-mature-folk-morality, or qua thing-disposed-to-prompt-approval-under-such-and-such-circumstances.” (p. 16-17).

  21. ^

    Here, I’m assuming that supervenience indicates distinctness of identity, and that the “identity” option covers all forms of reductionism.

  22. ^

    The example of the brightness of phenomenal yellow is from Sinhababu here.

  23. ^

    Though some approaches to speckled hens would dispute this.  

  24. ^

    See e.g. Sinhababu (unpublished) for vibes like this.

  25. ^

    Though some of Pearce’s comments seem a bit ambiguous in this respect to me, e.g. “I would say that for reasons we simply don't understand, pleasure-pain axis discloses the world's inbuilt metric of value and disvalue” (see here). My memory is that Sinhababu’s paper doesn’t take a stance of naturalism vs. non-naturalism, though I think his other work is fairly naturalistic.

  26. ^

    See, e.g. Pearce here – comments that strike me as either confused, or resting on a certain type of foot-stamping about “goodness = pleasure.”

  27. ^

    Stern (unpublished) makes this point well: “After all, insofar as a pleasant experience’s instantiation of value is merely a matter of its instantiating a property of some more mundane sort – say, the property of being pleasant itself, or that of causally regulating our use of “good” – then there is no deep mystery to be resolved regarding how we can become justified in our evaluative beliefs in the first place. In fact, precisely this is one central motivation for favouring such less demanding forms of realism about normativity! As a result, the Hedonic-Evaluative Acquaintance Thesis is of far greater theoretical interest if discussed against the background of a robustly non-naturalist conception of value” (p. 19).

  28. ^

    There are lots of varieties of naturalist realism. I haven’t tried to understand them all, and it’s possible that what I say here won’t apply to all of them. 

  29. ^

    I didn’t see such an account in Pearce, for example. And to the extent Sinhababu (unpublished) wants his argument construed in naturalistic terms, I have the same objection there (though I think the argument makes most sense on a non-naturalist construal, on which normative intuitions are understood as a form of “perception of the non-natural realm,” but one that could be reliable or unreliable).

  30. ^

    I’m not sure if the temporal ordering here is supposed to be literal.

  31. ^

    Here Rawlette differs from Stern, who argues that pleasure essentially involves representing something as good.

  32. ^

    See Chalmers (2010), Chapter 8, for a more detailed story about the content of phenomenal concepts – a story that involves a few extra distinctions that could in principle be incorporated into the story above.

  33. ^

    This isn’t necessarily an argument against Rawlette in particular: she need not be trying to do what the meta-ethical hedonist, as I’ve characterized them, is trying to do.

  34. ^

    Here I’m mostly talking about the role of this sort of discourse in a certain type of new age context. It’s totally possible for advocates of quantum theories of consciousness to understand the relevant quantum mechanics quite well (and also, to still be wrong).

  35. ^

    Chalmers: “When it comes to the explanation of experience, quantum processes are in the same boat as any other. The question of why these processes should give rise to experience is entirely unanswered.”

21

New Comment
3 comments, sorted by Click to highlight new comments since: Today at 3:44 PM

This is a very, very long post.

There's a lot that I feel I ought to reply to here (I'm one of those unsatisfying-to-argue-with hedonic utilitarian moral realistishs (kinda)) and I think Pearce has a point or two (though I've talked with him about our many differences of opinion).

But it's a very, very long post.

Imma have to pace myself.

One of the other problems with hedonism is that its difficult to get an altruistic (ot any extent over complete egoism) theory out of it. Only my pain exists for me .. I don't feel other people's suffering directly. I might suppose by analogy that their pains are bad for them, but I don't know it by direct acquaintance...and what is supposed to tell me that I have a duty to ameliorate suffering I don't feel? I could bundle it into some additional axiom:-

  1. Pain is bad.
  2. I have a duty to reduce all pain, including pain that doesn't exist for me phenomenally. That is a thing I should do.

But 2 is obviously normative, and isn't obviously naturalistic.

It might be the case that 2-like statements can be built out of naturalistic elements...but it could be the case that they are then doing all the lifting, and 1 isn't necessary. It could then be the case that I do have a duty to support some kind of preferences or values that I don't have direct access to....but not necessarily hedonistic ones.

This, I think, is a key point, and one that could be stressed more forcefully:

“I suspect that the appeal of meta-ethical hedonism derives at least in part from mixing normative epistemology together with the epistemology of consciousness in a manner that allows confusions about the latter to disguise muddiness about both.”

Many of these arguments seem to appeal to questionable views about consciousness; if we reject those views, then it’s not clear how plausible the rest of the argument is, or indeed, if elements of the argument aren’t even intelligible (because they rely on confusions about consciousness that can’t be made coherent), then we’re not even dealing with an argument, just the appearance of one.

This points towards a deeper worry I have about arguments like these. While you raise what I take to be credible epistemic concerns, it’s unclear whether metaethical hedonism can even get to the stage of being evaluated in this way if we cannot first assess whether it can offer us an account of normative realism that isn’t vacuous, self-contradictory, or unintelligible.

Take the claim that there are stance-independent normative moral facts. A naturalist might end up identifying such facts with certain kinds of descriptive claims. If so, it’s unclear how they can capture the kinds of normativity non-naturalists want to capture. While such accounts can be intelligible, it’s unclear whether they can simultaneously be both intelligible and nontrivial: such accounts would amount to little more than descriptive identifications of moral facts with some set of natural facts. Without bringing the unintelligible elements back in, this takes morality out of the business of having the overriding authority to mandate what we should and shouldn’t do independent of our goals and values. 

Naturalism ends up delivering us a completely toothless notion of moral “norms”: these are norms that I either already cared about because they aligned with my goals, or still don’t care about because they don’t align with my goals. In the former case, I would have acted on those goals anyway, and realism adds nothing to my overall motivation, while in the latter case, I would at worst simply come to recognize I have no interest in doing what’s “morally good.” And what is the naturalist going to say? That I am “incorrect”? Well, so be it. That I am “irrational”? Again, so what? All these amount to are empty labels that have no authority.

But with non-naturalist realist, what would it even mean for there to be a normative fact of the relevant kind? The kinds of facts that purport to have this kind of authority are often described as e.g., irreducibly normative, or as providing us with some kind of decisive, or external reasons that “apply” to us independent of our values. I don’t think proponents of such views can communicate what this would mean in an intelligible way. 

When I go about making decisions, I act in accordance with my goals and interests. I am exclusively motivated by those goals. If there were irreducibly normative facts of this kind, and they “gave me reasons,” what would that mean? That I “should” do something, even if it’s inconsistent with my goals? Not only am I not interested in doing that, I am not sure how I could, in principle, comply with such goals, unless, and only unless, I had the goal of complying with whatever the stance-independent moral facts turned out to be. As far as I can tell, I have no such goal. So I’m not even sure I could comply with those facts.

When it comes to pleasure and pain, these can either be trivially described so as to just be, by definition, states consistent with my goals and motivations, e.g., states I desire to have and to avoid, respectively. If not, it’s unclear what it would mean to say they were “intrinsically” good.

Philosophers routinely employ terms that may superficially appear to be meaningful. But, scratch the surface, and their terms simply can’t thread the conceptual needle.

In short, there is a deeper, and more worrisome problem with many accounts of moral realism: not only do they face seemingly insurmountable epistemic problems, and in the case of non-naturalist realism metaphysical problems but that at the very least non-naturalist realism also faces a more basic problem, which is that it’s so conceptually muddled it’s unclear whether there is an intelligible position to reject in the first place.