Law without a lawgiver

A response to Connor Jennings

1.0 Introduction

It’s great to see more metaethics coming to Substack. I’m not familiar with Connor Jennings or the blog Mind Meandering, but the blog seems new, and has a recent post about metaethics, “Our Moral Attitudes Aren't Mere Preferences” so go check it out and consider subscribing. I’ll be reviewing that post, but hopefully this will spark a conversation.

Connor’s post centers on two main theses:

1. Attempts to construe moral attitudes as preferences are inadequate

2. Mushrooms are bad

I won’t focus much on the latter point, so let’s get that out of the way: mushrooms are great, and Connor’s failure to recognize this is a clear sign of a compromised capacity for detecting gastronomic truths. 

I also partially agree with the former claim: that moral attitudes aren’t mere preferences. I say partially because I think they can be, and often are. How do I know that? Because my moral attitudes are preferences. 

This brings me to a central theme of my own approach to philosophy, and one of my most recurring issues with the way philosophers do philosophy: the presumption of homogeneity in thought. Philosophers speak of how “we” think as though everyone, everywhere (at least if they were rational or were thinking properly or whatever) thinks the same way, experiences the world in the same way, and shares in the same conception of what the world is like. It often seems like they think there are concepts “out there,” and “we” (humans and any other rational agents) “detect” or “discover” these concepts, and thus come to mutually share in a perfectly or near-perfectly overlapping conception of the world. 

I’m not suggesting Connor necessarily thinks this way, but the use of “our” brings this concern to mind. Quite the contrary: the recurrent inclusion of references to mushrooms being bad indicates an awareness that people can and do see the world differently. I really do like mushrooms, and my moral attitudes really do seem like preferences to me. There is already considerable cognitive variation within similar populations (assuming Connor and I are from similar cultures).

There may likewise be considerable cross-cultural variation in normative cognition as well. This is why I am so puzzled when people seem to generalize from how things seem to them to how they “seem” in some unqualified way, as though there is only one way they seem, or one way they seem “to us.” Even the assumption that there is a “typical” way things “seem” can be mistaken: typical among what population? 

That brings me to a few misgivings about the title of the post. Consider the title:

“Our Moral Attitudes Aren't Mere Preferences”

I take issue with two words in the title: “our” and “mere.” 

1.3 Unqualified use of “our”

Who is “our”? All members of their culture? All people who have thought about the topic? All humans? All humans with specific psychological characteristics? Are there normative or other constraints on who is included in “our” or not? If so, what are they?

It’s possible that thinking in distinctively moral terms is a culturally acquired capacity, not a product of biological evolution. If so, then moral attitudes may be something we learn to have, rather than something we and others are disposed to have in virtue of our shared evolutionary history, much as chess intuitions are something we learn to have, rather than something we’re biologically predisposed to exhibit. If moral thinking turns out to be a historical invention, it may even be that nonhuman agents simply wouldn’t think in distinctively moral terms, any more than they’d be disposed to think in terms of European codes of chivalry or the honor promulgated by Kahless the Unforgettable. 

Note that such examples are morality-adjacent, and some would argue that they just are systems of morality. But philosophers are not entitled to help themselves to so sweeping and expansive a conception of “morality,” that the term could be used, without contestation, to encompass any and all normative systems governing the interaction of agents. The broader one’s conception of morality, the shallower it is, until one has diluted the notion so much that it becomes something like “Whatever people care about” or “whatever normative considerations regulate people’s behavior in such-and-such a general way that it applies to basically anything.” You can define “morality” this loosely, but it approaches terminological homeopathy: trading technical inclusion for a notion so devoid of content as to mean very little at all.

If, on the other hand, one does want to try to say something more substantive about morality, I simply haven’t found such efforts to be successful. Attempts to define morality or delineate the moral domain have never achieved anything even remotely approaching unanimity, so while philosophers will make claims like “morality concerns that which one ought to do, all things considered,” none of the rest of us are obliged to accept these definitions as correct. I’ve never seen one that struck me as analytically true, nor have I seen any compelling empirical basis for very broad notions of morality. The notion of “morality” strikes me, at least, as a cultural invention. I am not alone in thinking this. Philosophers have argued variously that:

1. Morality is a historical invention (Machery, 2018)

2. There are no compelling, principled ways of distinguishing moral from nonmoral norms (Stich, 2018)

3. There is no “moral domain”: efforts by both philosophers and psychologists to identify a distinctive nomological cluster of properties that reliably allow us to distinguish moral from nonmoral considerations have consistently failed (Sinnott-Armstrong & Wheatley, 2012; 2014)

In other words, there may be no distinctive set of metanormative properties that distinguish nonmoral from nonmoral considerations in a consistent and principled way. It may be that a disposition to think of certain considerations in moral terms is a learned behavior, and is no more principled or systematic than the spelling of many English words: it’s simply the unprincipled accretion of historical accident within a particular sociolinguistic historical trajectory. If so, there may be entire societies, actual or counterfactual, that don’t have moral attitudes per se. This does not mean they don’t have normative attitudes: I take it that every human society, at least, has normative standards that govern social behavior, promote cooperation within groups, and so on. But we need not necessarily think of these normative standards as distinctively moral. Elaborating on this notion of a distinctive domain would be a titanic digression, so I leave it for future posts. 

For now, I defer to the articles mentioned above, in part to emphasize that my perspective on the matter isn’t unique to me but has already established a beachhead in the literature. My point here is that it’s not even clear that all populations have moral attitudes, where moral attitudes refer to a distinct type of normative attitude. Questions about the origins of distinctively moral thinking may account for how and why some people experience morality the way they do, which I’ll say a bit more about later in this post.

1.4 Use of the term “mere”

I object to the use of the term “mere” to describe preferences as “mere preferences.” Terms carry connotations, and by describing them as “mere” preferences this gives the impression that there’s something inferior, less-than, or undesirable about moral attitudes being preferences. I don’t grant that this is the case. Preferences do all of the relevant motivational and normative work; I could just as readily dismiss the realist’s moral facts as mere “objective facts” about what I should or shouldn’t do. Even if there were such things, I wouldn’t care about them. I only act in accord with my values, my “mere” preferences. If the moral facts don’t accord with my preferences, so much the worse for the moral facts.

Connor notes that some critics hold that claims like “X is wrong” just mean something like “I prefer that not-X.” That is, moral statements about what’s right or wrong express our preferences about what we want to occur or not occur. Connor does something excellent here: explicitly acknowledging that these preferences can be very strong:

The word “preference” sounds a bit soft, but of course, you can have very strong preferences. They don’t believe we just have a trivial distaste for murder, but that we really despise it.

This is a great remark, because when antirealists compare murder to ice cream flavors, people often misunderstand the comparison, mistakenly thinking that you’re comparing the two in terms of their importance and not in terms of both being stance-dependent rather than stance-independent. As an aside: this is a remarkably common mistake, yet I’ve never seen any formal name for it. The mistake is that when someone compares two things, A and B, in terms of quality X, people observe some other similarities (Y) or difference (Z), focus on that, and then criticize the person making the comparison for the allegedly objectionable claim of similarity or difference. Connor goes on to say:

However, strong though it may be, it’s still just a preference - and moral language is just us describing our tastes. Murder has no intrinsic property of disvalue, and were we to run into someone who really loves it, they’d not be making any kind of error. Much like how people who like mushrooms aren’t making an error (however much it pains me to say it).

I take “just” in these cases to mean “is a preference, and not anything else.” That is, I take the intent to characterize it as “just” a preference in a neutral descriptive way. I still worry about the potential implication that there’s something worse. That’s a problem with our language: a term like “just” can be used in an evaluatively neutral and evaluatively loaded way. It is often used, in describing antirealists, in an openly hostile and disparaging way. As a result, I think extra caution is warranted if one wants to avoid the connotation. It doesn’t look to me like it was intended in this case, but it may read that way anyway.

I also worry about the murder/mushroom comparison. While it is true that neither person (the person who loves murder and the person who loves mushrooms), on the antirealist’s view, would necessarily be making an error. However, audiences may be misled into thinking the antirealist isn’t that much more concerned about the murderer than the mushroom lover. Even with Connor’s qualifications about a moral preference involving a really strong preference, strength isn’t the issue here, scope is.

I can have a weak preference for mushrooms :I like them, but I don’t care that much if I have them or not. Or I can have a very strong preference: I can be obsessed with mushrooms, spending all my money on them and insisting I have them with every meal. In both cases, I may not care at all whether anyone else eats mushrooms. Eat what you want.

But I could also have preferences that vary in scope. When it comes to food, I may only care about what I eat. But when it comes to moral attitudes, I have preferences about what other people do. Those preferences could be weak: I can prefer people not take too many ketchup packets from a McDonalds or lick their fingers when turning the pages of a book. Or they can be very strong: I prefer people not launch nuclear weapons or massacre people for fun.

Critics often fail to appreciate this distinction when food/morality comparisons are made. The article doesn’t mix this up, but I again worry that readers may. So this isn’t a criticism of the article, it’s simply a commentary prompted by the article: it’s important to keep all these distinctions clear, because people seem to be deeply invested in misunderstanding antirealist views, and such conflations seem not only ubiquitous but extreme persistent. That’s part of why I find myself, on this blog, making the same points over and over: these sorts of conflations are everywhere. 

2.0 Obviousness and costs

Connor goes on to critique the “moral attitudes as preferences,” perspective, but first shares this remark, as a reason for rejecting this view:

One reason is that murder is just obviously bad, and saying that Jeffrey Dahmer didn’t make any moral error because he had a preference for eating people seems like a massive cost to a theory. 

I think murder is “obviously bad,” I just don’t think it’s obviously stance-independently bad. If Connor thinks it’s obviously stance-independently bad, well, I think it’s obviously not. As an aside: why not include “stance-independently” or “objectively”? Why just say “obviously bad,”? That’s normative language, only, without any explicit reference to the metaethical issues at stake. There’s a hint of normative entanglement in that.

I don’t see any good reason to privilege what seems obvious to realists over what seems obvious to me. I’ve never understood why realists put so much stock in what seems obvious to them. Such appeals are entirely private, and cannot advance the public discourse in any substantive way. That something seems obvious to someone else just isn’t a very good reason for me to believe it. I think moral realism is obviously absurd. As far as I can tell, this has never made any difference to any moral realist.

Connor adds:

They would never accept my appeals to the self evidence of moral properties though, so I’ll try to pick it apart in different ways.

I don’t think anything is self-evident to anyone, but if self-evidence is a thing, it’s just as available to me as it is to realists, and I’d simply find myself saying that it’s self-evident that realism isn’t true. Self-evidence doesn’t privilege realism.

Second, Connor claims that rejecting the view that murder is “just obviously bad” (presumably in a realist) way and that Dahmer didn’t make any moral error (presumably, an error with respect to the stance-independent moral facts) incurs a massive cost.

What are these costs? There are advantages: moral realism isn’t true, and certain conceptions of it probably aren’t even meaningful. But I’ve never been presented with what I took to be a cost to denying moral realism.

3.0 On to preference theory

This brings us to the theory that moral attitudes are preferences. Connor says that if this were true, it’d be “weird” for us to have moral attitudes in the first place. We’re given this example:

There’s something different about thinking an act is wrong and simply having a preference against it. Take my rugby team being kicked out of the URC quarter finals last week. As I watched us get battered into oblivion, I very much had a preference against us losing - but I didn’t think it was morally wrong that we lose.

Even if there is something different about thinking something is wrong and having a preference against it, this example doesn’t illustrate such a difference. All this seems to me to indicate is that you can have preferences you don’t regard as moral. A reasonable preference theory only holds that moral attitudes are a type of preference, not that all preferences are associated with moral attitudes.  

Next, Connor says:

If we’re going to accept the Preference Theory, we’re going to need to explain exactly why some preferences have a moral character to them, and some of them don’t.

I agree. Such an account will involve engagement with a massive amount of empirical research in anthropology, psychology, linguistics, history, and other fields. It would require many dedicated researchers working on this topic for a long time to begin to get a handle on it. We’re not there yet. I would not even begin to suggest that this is infeasible and that we should fall back on realism without a better understanding of moral psychology.

This is the part I agree with: 

You might think that it’s the intensity of the preference. We think murder is wrong because we strongly dislike it, and our other preferences are relatively minor. However, I think this fails. There are absolutely some things that I think are wrong, that I don’t have an intense preference against. 

See my distinction between strength and scope above. If moral attitudes were preferences, I don’t think what would distinguish them is that they were strong preferences. 

Connor also says:

Of course, there’s another way out which is to simply deny that there’s a difference in the experience [sic] our moral attitudes and our other preferences - but all I can say is that seems obviously false. 

Again, who is “our”? I experience my moral attitudes as preferences. Connor (and probably quite a few other people) don’t. Given this, there are going to be lots of interesting questions to ask about what’s going on when people have “moral attitudes.” One possibility is that there is no distinctive psychological state associated with moral attitudes: they may be preferences most or all of the time for some people, and only sometimes, rarely, or never for others. They may involve a variety of psychological states, and it may be that biological and cultural evolution, as well as individual differences, all contribute in various interconnected ways in shaping interpersonal variation in how we experience (or even if we experience) moral attitudes, and if so, what the nature of those attitudes is.

4.0 The saw scenario

Connor presents this scenario:

Imagine you wake up in a Saw scenario. You’re in a cage, and beside you is another person in a cage. They share all the relevant moral similarities to you - they’re not really old and close to death, they’re not Hitler, and they also subscribe to my blog. There’s no plausible reason for a stranger to value you differently. Across the room is a pair of buttons with a stranger standing beside them. Button one tortures you for 1 hour. Button two tortures the other caged person for 1 hour and 1 second. For some reason, the button pusher is obligated to choose an button (let’s say, for example, if they don’t, the world will implode. Or maybe they’ll be forced to use a Stairmaster for 30 seconds. Haunting). What should they do?

I don’t think there’s a stance-independent fact of the matter about what they should do. If it were me, I’d push the button that resulted in less suffering. This looks like a normative question to me, though, and has little to do with metaethics. Connor says:

Were I in the cage, it seems obvious that they should torture me.

I don’t know what this experience is like, because I apparently don’t have it. I can talk about what I’d do, and what I’d prefer someone else would do. But I have phenomenology that it’s “obvious” that it’s “true” that someone “should” opt for less torture to occur, independent of their goals or values. I don’t recall ever having experiences like this. Realist phenomenology is utterly alien to me. Nothing about my position on the situation in any way involves any substantive sense of some “truth” out there. It feels like someone asking me what I’d prefer to have for lunch. There’s a sense in which if my options were “a sandwich” or “a slightly less appealing sandwich,” it’s obvious I’d opt for the former. However, it’s obvious only in the trivial sense that I have a decent enough sense of my own preferences that if you present me with two options, and one is clearly preferable to the other, I’m going to favor it.

Regarding the torture situation, I suppose the goal is to draw a distinction between what I’d prefer and what I think they “should do,” with the point being that what they should do differs from what I’d prefer that they do. Speaking for myself, I again simply do not think there’s a fact of the matter about what they should do. I can talk about what I think I would do if I were them, and were occupying their point of view, which is torture “me” since it results in less total suffering. But this doesn’t show that my moral standards aren’t preferences. I can both:

1. Prefer not to be tortured

2. Prefer that people to opt to minimize suffering, all else being equal

To be clear, I’m not defending the view that, for anyone else, moral attitudes just are preferences. But they do appear to be preferences for me. 

Connor next says:

One possible out is to say that our moral statements don’t describe our preferences, but our second order preferences. To which I say, “Ah hah! So they don’t just describe our preferences then!”. My men have risen from the trenches and driven you back an inch. A fine victory that I will presumably win a medal for. Perhaps, some kind of hat.

This consideration does start to reveal some of the cracks in what I suppose you could call “First-order preference theory,” the view that moral attitudes exclusively consist of first-order preferences. This isn’t a very plausible view from the outset, and isn’t how I think of my own moral attitudes. I’d think of them more in terms of them being any combination of first-order, second-order, or higher-order preferences. 

The problem here is that the notion of a “preference” is a bit of folk psychology that probably doesn’t accurately reflect the actual structure of human cognition. Our “preferences” may or may not be arranged hierarchically, such that we can speak of first-order and second-order “preferences.” it may be that a more complete account of human cognition would carve up the motivational and evaluative space of human thought in a more complicated and non-obvious way. Preference theory should have already been thought of to encompass second-order preferences from the outset, as that’s the more plausible account and probably more closely mirrors human cognition. But even this inclusion isn’t sufficient, as it may fail to capture qualitative distinctions in human cognition that we think of as “preferences,” but actually fragment into a number of functionally distinct psychological states.

Connor argues that moral attitudes don’t seem to be like second-order preferences. However, an example of an attitude that doesn’t seem to be like this would, at  best, only show that they aren’t all second-order preferences. But why would they need to be? Why couldn’t moral attitudes include both first-order and second-order preferences? 

But Connor reports a moral stance that may not appear to be a first-order or second-order preference. I can’t dispute that: perhaps it simply isn’t. If so, does this mean that moral attitudes are not preferences? Even if we grant for the sake of argument that Connor has at least one moral preference that isn’t an attitude, this would at best only show that, for one person, at least some moral attitudes are not reducible to preferences. I’m not sure we should generalize from how things seem to draw inferences about an entire category, moral attitudes. When I introspect, my moral attitudes do uniformly appear to be preferences. If I were to generalize, we’d run into an immediate conflict. However, I am not disposed to presume everyone thinks the way I do. 

Those commenting on philosophical topics should not presume that people (humans and non-humans alike, e.g., aliens or extraterrestrials) necessarily think the same way, or share the same concepts or modes of cognition. We already know that some people lack mental imagery or an internal narrative. Human cognition can vary dramatically; there is little reason to presume everyone, everywhere, has “moral attitudes,” without good evidence for such a claim. And such a claim would require a robust operationalization of what a moral attitude is, coupled with evidence that such a phenomenon is shared across cultures. Unfortunately, the use of “our” is rarely accompanied by any specification that “our” means “all humans” or “all agents” or whatever, so it’s hard to even say who this is supposed to refer to.

Having said all that, I don’t think it’s generally the case that moral attitudes are simply preferences. I want to describe something roughly like what I think may be going on with people who report that their moral attitudes aren’t preferences.

5.0 Laws without a Lawgiver

Christians often claim that atheism is inconsistent with moral realism. Sometimes, they’ll say that this is because “laws require a lawgiver.” The idea seems to be that it makes little sense for a secular moral realist to propose that there are “rules” that have “authority” over us, but there isn’t some agent who ratified those rules and acts to enforce them. Secular moral realists often scoff at this, insisting that moral realism doesn’t require theism. Perhaps not. But I think there’s something to the intuition that “laws require a lawgiver.”

Humans are cultural specialists. We are highly adapted to acquiring, improving upon, and passing on the cumulative cultural knowledge of our ancestors to future generations. This cultural knowledge is used to flexibly adapt to local environments and efficiently extract resources from those environments. Cumulative cultural knowledge means just that: generation after generation, we don’t simply pass on a static amount of cultural knowledge, but we add to that knowledge before passing it on.

Not all innovations involve the direct manipulation of our environments: while we do pass on knowledge of how to hunt, how to build shelters, and so on, we also pass on the norms and traditions of our cultures themselves: taboos, codes of conduct, and other norms regulating personal conduct, interactions within our societies, and interactions with other societies.

There are a whole host of psychological factors that facilitate and reinforce cultural norms. I don’t think we yet know entirely how biological and cultural evolutionary forces interact to shape how, exactly, these processes function, but they include psychological phenomena most people are probably familiar with: guilt, shame, anger, disgust, a sense of vengeance, compassion, respect, indignation, trust, admiration, and so on. 

With respect to people who find moral realism intuitive, here’s what I think is going on: I think such people have tend to grow up in societies that have undergone thousands of years of of enculturation under monotheistic religious systems that depict the world in terms of the cosmic forces of good and evil. Generation after generation, people were taught from an early age that there is a God and there are good and evil forces at play in the world. This God has issued rules, and there are righteous and sinful actions. People may have all sorts of preferences: lust, greed, malice, envy, and so on. Yet we’re taught to suppress and control these darker urges. As monotheism has receded and secularism has taken hold, vestiges of this external, cosmic, divine sense of purpose, meaning, and value “from the outside” remains, a legacy of a time when there really was a lawgiver issuing these moral laws. While people may have stopped believing in God, an analog to a king on a throne issuing divine decrees that constitute the moral law in a very literal sense: a literal law, enforced by a literal being, people still feel strongly disposed to believe there really is some kind of cosmic moral law, even if there is no lawgiver to construct and enforce that law. 

I don’t think you need religion for this, but I think it reinforces a preexisting predisposition to internalize the norms of one’s culture. If your culture holds that certain actions are wrong, and that one should feel guilt or shame about them, we can find ourselves feeling these negative emotional experiences even when we contemplate actions we’d like to perform. This results in an induced motivation to perform or abstain from actions even when they don’t align, directly, with our preferences. People may have a preference to do something bad, but feel a kind of constraint “from the outside” a sense that you “shouldn’t do that.”

What puzzles me is this: to me, the suggestion that some combination of cultural and biological factors dispose us towards a disinclination to perform certain sorts of actions in such away that, on introspection, doesn’t seem like a preference, seems to me like a perfectly reasonable and straightforward hypothesis about what’s actually going on when people feel a “pull from the outside” to act in certain ways. 

Yet for whatever reason, people seem inclined to think of this almost like a sixth sense, as if (in some cases, literally) they’re sensing moral facts. From my perspective, this appears to be a sort of extrasensory capacity for detecting or perceiving a special kind of fact. I find it very strange that this is taken so seriously, while “it’s probably something mundane about our psychology,” isn’t the more obvious go-to explanation. Why think one possesses what amounts virtually, if not literally, to something akin to a paranormal power, or at the very least, an extremely mysterious power? Some realists seem comfortable just insisting they find nothing mysterious about it. Well, perhaps not. Then again, perhaps people who buy into astrology or astral projection or literal magic don’t find anything especially mysterious about it. How comfortable people are with their own intuitions isn’t a good indicator of the merits of those intuitions.

I don’t think moral attitudes are just preferences. I don’t think “moral attitudes” picks out any distinct category that exhibits a shared set of traits. I think it’s more likely there is no such distinct thing as a “moral attitude,” and that moral judgment will eventually be characterized in terms of a mature psychology that better captures the actual psychological phenomena that characterize human cognition. I predict that absolutely nothing about that resulting picture will, in any way, be best explained by invoking moral realism.

References

Machery, E. (2018). Morality: A historical invention. In K. J. Gray & J. Graham (Eds.), Atlas of moral psychology (pp. 259-265). New York, NY: The Guilford Press.

Sinnott-Armstrong, W., & Wheatley, T. (2012). The disunity of morality and why it matters to philosophy. The Monist, 95(3), 355-377.

Sinnott-Armstrong, W., & Wheatley, T. (2014). Are moral judgments unified?. Philosophical Psychology, 27(4), 451-474.

Stich, S. (2018). The moral domain. In K. Gray & J. Graham (Eds.), Atlas of moral psychology (pp. 547- 555). New York, NY: Guilford Press.