Which answers to the now what question collapse into abolitionism (if any)?

ABSTRACT

Moral error theorists face the now what question. How, if at all, ought they to adjust their moral practice after having discovered the error? Various answers have emerged in the literature, including, but not limited to, revisionary fictionalism, revisionary expressivism, and revisionary naturalism. Recently, François Jaquet has argued that there are only two available answers to the now what question, since every extant answer except revisionary fictionalism collapses into abolitionism. This paper provides a response. First, it argues that revisionary naturalism does not collapse into abolitionism. The argument is that abolitionists can neither utter schmoralist naturalist replacement judgments that contain moral terms nor eliminativist naturalist replacement judgments that do not contain moral terms. This means that abolitionists cannot utter replacement judgments at all. Since revisionary naturalists can utter replacement judgments, their view does not collapse into abolitionism. Second, the paper gives reason to believe that this result extends beyond revisionary naturalism to the other revisionary metaethical theories, and that this is good news for moral error theory.

KEYWORDS:

• Moral error theory
• now what question
• revisionary naturalism
• abolitionism
• revisionary expressivism

1. Introduction

Moral error theorists believe that moral judgments such as ‘stealing is morally wrong’ express truth-apt beliefs that ascribe moral properties to objects and actions. They also believe that moral properties are not instantiated. Since moral error theorists believe that moral judgments can only be true if they correctly describe moral properties, they conclude that no moral judgment is true (Mackie 1977; Joyce 2001; Olson 2014; Streumer 2017; Kalf 2018).¹

According to John Burgess, ‘the major philosophical puzzle raised by … moral error theory’ can be expressed by the following question:

if an error theory provides the correct descriptive account of moral discourse, how, if at all, ought we to adjust our practices after having discovered the error? (Burgess 1998, 534).²

This paper is about this so-called ‘now what question’ for moral error theorists (Lutz 2014, 351).

There are different versions of the now what question. The version that I am concerned with here is the following:

What should error theorists do, as a matter of personal policy and in a society of success theorists, with their practice of thinking moral thoughts and uttering moral judgments?

Success theorists have moral beliefs, they think that some of these beliefs are true, and they communicate these beliefs as moral judgments (Sayre-McCord 1986, 6). The version of the now what question that I engage with is addressed to individual error theorists, asking them to consider what they should do with their moral judgments in a society of success theorists.³ The question is not which prescriptive metaethical theory—which view on what we ought to do with our moral judgments—should be implemented by society as a whole (Svoboda 2017).

I engage with this version of the now what question because it is unlikely that we will soon, or even ever, have an entire society comprised of error theorists who also accept the same answer to the now what question. The question whether we can identify a single answer to the now what question that improves our collective practice simply isn’t practically relevant today. Contrastingly, the question what individual error theorists should do in a society of success theorists is highly practically relevant, both for committed error theorists and for those of us who are considering different views in metaethics and who want to know if the error theorists can answer the now what question satisfactorily.

As of today, there are five answers to the now what question that error theorists can embrace. These are conservationism, abolitionism, revisionary expressivism, revisionary naturalism, and revisionary fictionalism.⁴ The five answers can be grouped into three camps. In the first or ‘keep morality’ camp, we find the conservationists, according to whom error theorists should conserve their moral thought and talk (Olson 2014). In the second or ‘abandon morality’ camp, we find the abolitionists, who claim that error theorists should stop thinking moral thoughts and uttering moral judgments (Garner 2007, 2018; Hinckfuss 1987; Marks 2013, 2018).

In the third or ‘revise morality’ camp, we find answers to the now what question that advise error theorists to adjust their moral practice without abandoning it completely. First, we have revisionary expressivists who argue that error theorists should start thinking replacement thoughts and uttering replacement judgments that express emotions or other conative states (Kohler and Ridge 2013; Svoboda 2017). Second, we have revisionary fictionalists who claim that error theorists should start treating morality as a fiction. A revisionary fictionalist can do this by believing propositions about morality as a fiction and uttering replacement judgments that express these beliefs (content fictionalism). She can also do this by make-believing propositions about morality and uttering replacement judgments that express these mental states (force fictionalism) (Joyce 2001). Finally, we have the revisionary naturalists who claim that error theorists should start thinking replacement thoughts and uttering replacement judgments that express true beliefs about natural facts (Lutz 2014).

François Jaquet has recently argued that every answer to the now what question that proposes a revision in our moral practices collapses into abolitionism (Jaquet 2020).⁵ His argument is that the error theorist’s replacement judgments must be genuine moral judgments if they are to avoid collapsing into abolitionism. But replacement judgments are not genuine moral judgments. So we get the collapse claim. In this paper, I argue that naturalism does not collapse into abolitionism. I believe that my arguments apply, mutatis mutandis, to the other prescriptive metaethical theories, and I say a bit about why I think that expressivism fails to collapse towards the end of the paper, but I will not provide a full defence of this universal claim here.

The paper is structured as follows. I first explain in more detail the naturalist’s answer to the now what question (§2). After this, I explain Jaquet’s argument for the collapse claim and I argue that this argument fails because it employs a faulty understanding of abolitionism (§3). Jaquet thinks that abolitionists will be happy to utter replacement judgments because these judgments are not actual moral judgments. I argue that abolitionists do not want to be understood as committing themselves to morality because they believe that morality is all-things-considered prudentially bad. This means that abolitionists do not utter moral or replacement judgments. But if so, then naturalists, expressivists and all the others can utter their replacement judgments. I provide a detailed defence of the claim that the abolitionist’s interlocutors will understand her as uttering genuine moral judgments if she utters moral or replacement judgments (§4-§7). In the penultimate section, I explain why error theorists should be happy with the result that neither naturalism, nor probably the other answers to the now what question, collapse into abolitionism (§8). I end with a conclusion (§9).

2. Naturalism

In this section I explain the details of the naturalist answer to the now what question that we need in this paper. The most generic formulation of naturalism is that we should stop uttering moral judgments that express false beliefs about moral facts and we that should start uttering replacement judgments that express true beliefs about natural facts. Following Jaquet, I assume throughout this paper that the relevant natural facts are prudential facts, such as the fact that performing a certain action promotes the satisfaction of my (considered) desires. Following Fletcher, I assume that morality is about behaviour that affects other people, the animals and the environment, and that prudence concerns behaviour that affects my own well-being (Fletcher 2021, 1).

There are two important distinctions within the naturalist camp. First, we have objective and subjective naturalism (Jaquet 2020, 241). Objective naturalists claim that the replacement judgments are beliefs in mind-independent natural facts. For example, rather than believing that donating to charity is right, we should believe that donating to charity maximizes utility. Subjective naturalists say that the replacement judgments are beliefs about our attitudes. Rather than believing that donating to charity is right, we should believe that we approve of donating to charity.

We must also distinguish between schmoralist and eliminativist naturalism (Jaquet 2020, 241). Schmoralists recommend error theorists to continue using moral language albeit with a different meaning. Eliminativists recommend error theorists to abolish moral language altogether and to replace moral talk with different talk, in this case naturalist talk. We now have four options for naturalists. First, we have schmoralist objective naturalists, who state that we should continue to say that donating to charity is right but insist that the new meaning of this phrase is that donating to charity maximizes utility. Second, we have schmoralist subjective naturalists, who state that we should continue to say that donating to charity is right but insist that the new meaning of this phrase is that we approve of donating to charity. Third, we have eliminativist objective naturalists, who state that we should no longer say that donating to charity is right and that we should start saying that donating to charity maximizes utility. Finally, we have eliminativist subjective naturalists, who state that we should no longer say that donating to charity is right and that we should start saying that we approve of donating to charity.

These four naturalist replacement judgments are surrogates of the ordinary moral judgments, only the latter of which are systematically false. This is because replacement judgments are not moral judgments. A schmoralist replacement judgment sounds like a moral judgment because the judgment in question contains moral terms like ‘morally wrong’ and ‘morally right’. But because these terms are used with a new meaning, a schmoralist surrogate judgment does not require moral facts as truth-makers and is therefore not false. Similarly, an eliminativist judgment that doesn’t even contain moral terms—e.g. because it is a judgment about what the agent approves of—also does not require moral facts as truth-makers and it is therefore not false either.

Jaquet has argued that naturalism collapses into abolitionism. I will now turn to his argument for this claim—and my objection to it.

3. The collapse argument and abolitionism

In this section I introduce Jaquet’s argument for the collapse claim. I also argue that this argument is unsound, though it will take me four additional sections to discuss all the details of my objection (§4-§7).

I label Jaquet’s argument for the collapse claim as ‘The Collapse Argument’. It goes as follows.

The Collapse Argument

P1 Unless its replacement attitudes are genuine moral judgments, naturalism is a mere variant of abolitionism.

P2 Naturalism’s replacement attitudes are not genuine moral judgments.

C Therefore, naturalism is a mere variant of abolitionism (Jaquet 2020, 253)

Let me explain this argument in detail. Jacquet thinks that:

abolitionists advocate the abolition of moral beliefs (Jaquet 2020, 240)

This is true. No one denies this. But according to Jaquet, the claim that abolitionists abolish moral beliefs is only the negative component of their view, which explains what error theorists should not have (viz., moral beliefs). As such there are, says Jaquet, various versions of abolitionism, depending on how they think about the positive component of their view; viz., the part of the view that tells error theorists what they are permitted to do once they have jettisoned their systematically false moral beliefs. Thus, Jaquet writes:

Abolitionists are … at liberty to argue that we should replace our moral beliefs with conative attitudes or nonmoral beliefs, and our moral sentences with nonmoral sentences that would express those attitudes (Jaquet 2020, 246).

And so, for example, we can:

consider a variant of abolitionism according to which we should replace our moral judgments with beliefs in natural facts (Jaquet 2020, 250).

The same holds for expressivism, and indeed the other answers to the now what question.

My objection to this argument is that it contains a false presupposition in virtue of its first premise. This renders the argument unsound because now its first premise cannot be true. The presupposition is that abolitionists can replace their moral judgments with replacement judgements, such as the judgment that donating to charity maximizes utility. This presupposition is false. That this is so follows from the reason why abolitionists want to abolish moral thought and talk—a topic that Jaquet does not discuss. Abolitionists cannot accept replacement judgments because they think that morality is all-things-considered prudentially bad.⁶ Morality has some pro tanto prudential good-making features, but it also has pro tanto prudential bad-making features, and the latter outweigh the former.⁷

The first reason why morality is pro tanto prudentially bad for us is that the types of things that morality asks us to do are more often than not, or even always, prudentially bad for us. Morality, says the abolitionist, wants us to be good patriots and go to war if our nation calls on us. But going to war will be (at least psychologically) damaging for me and I will need to leave my thriving business behind (Mackie 1977, 123, 1980, 154). Moreover, morality condemns land reformers as traitors and criminals even though it may be prudentially better for me if we divided the land more equally, for I am bereft of my own parcel (Mackie 1980, 154; Garner 2007, 502). Morality tends to preserve the status quo, blocking reform where it is needed. Note that these examples are not uncontested. But my aim here is not to defend abolitionism—just to explain what its proponents think.

The second reason why morality is pro tanto prudentially bad is that the very act of moralizing has prudential disvalue. Regardless of the content of your moral judgments, making moral judgments per se is pro tanto prudentially bad. As the abolitionist Richard Garner writes:

Not only does the moral overlay inflame disputes and make compromise difficult, the lack of an actual truth of the matter opens the game to everyone. Every possible moral value and argument can be met by an equal and opposing value or argument. The moral overlay adds an entire level of controversy to any dispute, and it introduces unanswerable questions that usurp the original question, which is always some practical question about what to do or support. This ‘moral turn’ guarantees that the participants will be distracted from the real issue, and that the disagreement will flounder in rhetoric, confusion, or metaethics (Garner 2007, 502).

Or consider Ian Hinckfuss, another abolitionist, who writes:

the more that people are motivated by moral concerns, the more likely it is that their society will be elitist, authoritarian and dishonest, that they will have scant respect for most of its members, that they will be relatively inefficient in engendering human happiness, self-esteem or satisfaction, that they will be relatively inefficient in the resolution of conflicts, and that their moralising will exacerbate conflicts, often with physical violence or even war as a result (Hinckfuss 1987, 3)

Again, my aim is not to defend the abolitionists. I just want to explain what they think.

Abolitionists may be extremists and claim that there is nothing pro tanto prudentially good about morality. This would be a rather extreme position indeed. And abolitionists do not need this extreme position to justify abandoning our moral thoughts and judgements. All they need is that morality is more often than not prudentially bad. On the whole, morality blocks reform that is for the better, prudentially speaking. And more often than not, morality spurs violence and war. Abolitionists may be moderates and claim that morality is pro tanto prudentially bad, that it is also pro tanto prudentially good, but that the badness outweighs the goodness.

In light of this, abolitionists can say that our moral thoughts and judgments are pro tanto prudentially good. Indeed, Richard Garner writes that moral thoughts have one pro tanto prudential good-making feature; viz., their ability to combat weakness of will. Garner writes:

The moral abolitionist … recommends that we abandon the practice, or better, replace it with some motivational aids that allow us to acknowledge and deal with things as they are (Garner 2007, 504).

Garner also writes that we should

replace the moral overlay with more effective and less duplicitous devices (Garner 2007, 504, 505).⁸

The reasoning here goes like this. Error theorists, like all human agents, sometimes lack sufficient motivation to do what they ought to do (or what they have sufficient reason to believe they ought to do). Luckily, her private moral thoughts motivate her to do what she ought to do (or what she has sufficient reason to believe she ought to). This is what makes moral judgments pro tanto prudentially good for her.

I am worried about consistency here. If moderate abolitionists claim that moral thought’s ability to motivate is pro tanto prudentially good, then the actions that moral thoughts motivate them to perform must be prudentially good. And this must be so always, or at least sufficiently often. For even if moral thoughts sometimes motivate abolitionists to perform actions that are not prudentially good, it may still be that, on average, keeping moral thoughts leads to more prudential goodness than eliminating these thoughts altogether. Here is my worry. Abolitionists are adamant that morality is usually, if not always, prudentially bad! But then how can it be pro tanto prudentially good for abolitionists to think moral thoughts that motivate them to perform actions that are usually, if not always, prudentially bad? The same is true for our publicly available moral judgments. It may be that publicly available moral judgments have pro tanto prudential good-making features, but abolitionists are adamant that their pro tanto prudential bad-making features outweigh their pro tanto prudential good-making features.

Garner himself seems to think that there is nothing pro tanto prudentially good about our publicly available moral judgments but that our private moral thoughts’ have a pro tanto prudential good-making feature; viz., their ability to combat akrasia. I say this because it would explain why, in the last two quotes, Garner talks about motivational aids when he talks about what abolitionists should replace the moral overlay with. He writes that the new device should not duplicate the moral device. For a duplicate of the moral device will include publically available (replacement) moral judgments, and these would get us all the detrimental effects of moralizing that he and other abolitionists like Hinckfuss discuss. Nevertheless, says Garner, we may want to embrace a wholly internal motivational device to combat weakness of will.

This seems to be an unstable position. If what morality recommends is more often than not prudentially bad (preserving the status quo at the expense of reform that gets me my own parcel), then you should neither get others to do what morality wants you to do nor should you get yourself to do what morality wants you to do. In that case, abolitionists should abolish both publicly available moral judgments and private moral thoughts, not just the former.

On this reading of abolitionism, and this is the reading that I accept in this paper, whatever pro tanto prudential good-making features of moral judgments there may be, all-things-considered, they are prudentially bad because they get other people to perform actions, such as preserving the status quo where land reform is needed, that are prudentially bad. And to this we must add the other pro tanto prudential bad-making feature of publicly available moral judgments; viz., their propensity to spur violence and war. As such, the all-things-considered prudential badness of publicly available moral judgments is overdetermined. The all-things-considered prudential badness of private moral thoughts is not overdetermined in this way. Nevertheless, for reasons just given, private moral thoughts are all-things-considered prudentially bad.

I will return to the question whether abolitionists can think private moral thoughts in §6. For now we just need the result that abolitionists think that publicly available moral judgments are all-things-considered prudentially bad. As such, abolitionists will not want to reintroduce them in whatever form, neither as genuine moral judgments, as conservationists suggest, nor as replacement judgments, as naturalists suggest. This claim directly contradicts Jaquet’s claim that abolitionists can utter replacement judgments.

It will take a bit of effort to see exactly how the various specific formulations of the naturalist answer to the now what question suggest that a discussion merits moral considerations. So, in what follows, I first argue that schmoralist naturalism, which is the view that the error theorist should utter replacement moral judgments that sound like the original moral judgments but that have a different meaning, does not collapse into abolitionism (§4). I then argue that eliminativist naturalism, which is the view that the error theorist should utter replacement moral judgments that do not sound like moral judgments, does not collapse into abolitionism either (§5-§6). Finally, I argue that my argument extends beyond naturalism to expressivism (§7).

4. Why schmoralist naturalism does not collapse into abolitionism

In this section, I first explain why abolitionists can neither utter genuine moral judgments nor schmoralist replacement judgments. I then explain that my argument only has bite on a specific interpretation of the now what question. In the two sections that follow, I use the shape of my argument here to show that abolitionists cannot utter schmoralist naturalist replacement judgments either (§5-§6).

It is clear that abolitionists cannot be conservationists and utter genuine moral judgments. This is because conservationists continue to utter real moral judgments, and their interlocutors (i.e. every ordinary user of moral discourse who is not an error theorist) will naturally and correctly understand the abolitionist to be uttering real moral judgments. But as we know, according to the abolitionist, the problem with conserving moral language is that it blocks reform by preserving the status quo and that it will inflame the conversation, making sensible compromises difficult and possibly igniting violence and war. Therefore, to avoid adding a moral overlay to conversations, abolitionists refrain from uttering genuine moral judgments.

For similar reasons, abolitionists cannot become schmoralist naturalists either. After all, if abolitionists utter schmoralist replacement judgments, such as the judgment that donating is morally right, then their interlocutors are also going to understand these abolitionists as uttering genuine moral judgments. For they hear the abolitionist utter the judgment ‘donating to charity is morally right’ and this judgment sounds like a moral judgment. Of course, it is not really a moral judgment—it only sounds like one but it has a different meaning—but the folk are not metaethicists. The folk cannot entertain the thought that the person they are talking to is using a term that sounds like a moral term but isn’t really a moral term because they are using it with a new meaning. And so, the folk will take the judgment that sounds like a moral judgment for a real moral judgment, they will add the moral overlay to the conversation, and as such they get us the barrier to sensible compromise and possibly even violence and war.

Note that the fact that ordinary users of moral language will normally hear the schmoralist replacement judgments as real moral judgments is the main reason why naturalists continue to utter schmoralist replacement judgments. Naturalists think that overall, morality is prudentially good for us. Take Jaquet’s example of donating to charity. Naturalists think that donating to charity is prudentially good for them, for instance because it satisfies their considered desires (desire-satisfaction theory of prudential value). And the naturalists think that other people will be motivated to donate to charity if they hear the naturalists utter the judgment ‘donating to charity is morally right’. After all, the ordinary users of moral language will normally hear the schmoralist replacement judgments as genuine moral judgments. And so at least some of these people will, upon hearing other people uttering the words ‘donating to charity is morally right’, believe that they are morally required to donate. And some of them will muster the motivation to donate. All of this enables naturalists to use their revised moral language as a tool to get other people to do what is prudentially good for them, while at the same time avoiding judgements that express systematically false beliefs.

Before moving on, note that my objection only works on a practical version of the now what question, which urges error theorists to think about what they should do with their moral judgments in a society of people who have moral beliefs. For in the alternative society in which everyone is an error theorist, people will not tend to hear an error theorist’s replacement judgments as committing her to the idea that topics being discussed merit moral considerations.

Elsewhere Jaquet accepts this alternative formulation of the now what question:

There are, in fact, two versions of the now-what problem. One question is what we should do with our moral beliefs, as individual error theorists in the present society, broadly made of success theorists. Another is what we should do with our moral beliefs, as individual error theorists in a possible future society mainly made of error theorists. Although less urgent than the former, the latter question has more of an ‘existential’ flavour. For it is at the heart of the common worry that wide acceptance of the error theory might lead to the end of civilization. This is the question I will focus on (Jaquet 2021, 40).⁹

In the paper I respond to here, Jaquet does not distinguish these two versions of the now what question (Jaquet 2020). I believe that both versions of the now what question are crucial to our understanding of the merits and demerits of moral error theory. Jaquet’s existential reading of the now what question is important because a convincing answer to this question will do much to remove the fear that a society full of error theorists means the end of civilization. My objection to the Collapse Argument loses its bite in a society of error theorists.

However, as I argued in §1, my interpretation of the now what question is practically (although perhaps not existentially) speaking, much more pressing. It is very unlikely that we will soon, or even ever, have a society fully comprised of error theorists, whereas it is already the case that error theorists need to think about how they are going to accost others that either are not error theorists or do not even know error theory and metaethics. This is my reason for focussing on the practical, non-existential version of the now what question. I believe that my objection to the Collapse Argument has a lot of bite in a society of success theorists.

I have just argued that abolitionists cannot utter real or schmoralist naturalist replacement moral judgments. But can abolitionists utter eliminativist naturalist replacement judgments? I will argue in the next two sections that they cannot.

5. Why eliminativist naturalism does not collapse into abolitionism

Can abolitionists utter eliminativist naturalist replacement judgments? This is an important question, for an objector might say that abolitionists do not raise the suggestion that they want to play the moral game with their interlocutors when they say, for instance, that donating to charity maximizes utility. After all, eliminativist naturalist replacement judgments do not contain moral terms, and so ordinary participants in moral discourse will not, or so the objection goes, hear these judgments as genuine moral judgments. Consequently, abolitionists can utter eliminativist naturalist replacement judgments, such as the judgment that donating to charity maximizes utility. And so, the presupposition in the Collapse Argument, which was that abolitionists can utter replacement judgments, is true after all. Therefore, my objection to the Collapse Argument fails.

I admit that eliminativist naturalist moral judgments do not semantically entail that the utterer of these judgments wants to play the moral game. But I think that these judgments do conversationally imply that the utterer of these judgments accepts that the conversation merits moral considerations. To see how this works, we must first understand the distinction between semantic entailment and conversational implicature.

Semantic entailment is a relation between two judgments where, if the first judgment is true, the second will also be true in virtue of meaning. For example, judgment (1) semantically entails judgment (2):

1. Johnny is a bachelor

2. Johnny is an unmarried man

Judgment (1) semantically entails (2) in virtue of the meaning of the term bachelor in (1), ensuring that (2) is true if (1) is true.¹⁰

Conversational implicature is also a relation between two judgments. It arises if a party to a conversation is justified to accept a particular judgment on the basis of an earlier judgment uttered by someone else, even though the relation is not truth-preserving and even though the relation does not hold in virtue of what is literally (semantically) said (Grice 1989). For example, judgment (3) may conversationally imply judgment (4):

1. Pete’s car broke down

2. Pete will not be present at the meeting

Imagine a staff meeting, suppose that Mary is checking the attendance, and suppose that all the colleagues know that Pete is lazy and doesn’t like exercise, including walking and cycling (describing the context of the conversation to be studied like this is necessary for any implicature to arise and is called priming the context). If Mary utters (3) during this meeting, then Mark, who is present at the meeting, is justified to infer (4) from (3). He is justified to infer (4) because he justifiably believes that everyone in the staff meeting, including Mary, accepts what Paul Grice calls the cooperative principle:

Cooperative Principle

Make your conversational contribution such as is required at the stage at which it occurs, by the accepted purposes or direction of the talk exchange in which you are engaged (Grice 1989, 6)

How do you meet the Cooperative Principle? According to Grice, you can do this by following the four maxims associated with it:

The maxim of quantity: make your contribution neither less nor more informative than is required by your talk exchange.

The maxim of relation: make your contribution relevant.

The maxim of manner: be orderly in how you present your information.

The maxim of quality: be truthful and present information that you are unsure of as such.

We can now explain why Mark is justified to infer (4) from (3).¹¹ He is justified to suppose that Mary accepts the Cooperative Principle and the maxims. Thus, in the context of the meeting in which people are talking about who is and isn’t present, he naturally understands Mary’s remark that Pete’s car broke down as implying that he will not be present. We are talking about who is present in the meeting, we know that Pete is lazy and that his car broke down, and we expect the speaker to be truthful and to say things that are relevant and informative. So, Mark concludes, on the basis of all of this, that what Mary says must be her way of saying that Pete will not attend the meeting. Note that this conversational implicature is cancellable. If Mary realizes that by uttering (3) she implies (4), she may add to (3) something like ‘but mind you, Pete will take the bus’. This cancels the implicature to (4).

Our question is whether abolitionists can utter eliminativist naturalist replacement judgments. In the previous section, I argued that abolitionists can neither utter genuine moral nor schmoralist naturalist replacement judgments. In effect, and although I did not put it in these terms, I argued that moral and schmoralist naturalist replacement judgments semantically entail the judgment that the conversation at hand merits moral considerations (i.e., suggests that talking about morality is to be part of the conversation). After all, the moral judgment ‘giving to charity is morally right’ semantically entails ‘giving to charity is morally relevant (or matters morally)’. This is because it is impossible to talk about an action being morally right without talking about morality per se.

Similarly, the schmoralist naturalist replacement judgment ‘giving to charity is morally right’ also semantically entails—at least for the folk—‘giving to charity is morally relevant (or matters morally)’. Of course, the schmoralist judgment ‘giving to charity is morally right’ doesn’t really carry this entailment, but the folk think that it does because they cannot entertain the thought that their interlocutor uses moral terms with a new, non-moral meaning. Consequently, they will hear the judgment as a genuine moral judgment, and they will accept the entailment as well.

My claim in this section is that although eliminativist naturalist replacement judgments do not semantically entail the judgment that the conversation in question merits moral considerations, they do conversationally imply this. To see this, let’s prime a common conversational context.

Imagine that Dennis is talking to Anne. And imagine that Dennis aims to convince Anne that she should give to charity. Dennis is an eliminativist naturalist and utters (5). Is Anne, who has never studied metaethics, justified to infer (6)?

(5) Giving to charity maximizes utility

(6) Giving to charity is morally right

Anne is justified to infer (6) because she will suppose that Dennis accepts the Cooperative Principle and the maxims. Thus, in the context of an everyday conversation about whether we should give to charity she will, based on the maxim of relation, expect (5) to be relevant. And because everyone always talks about morality if the topic of giving to charity is brought up—especially if we are talking about it in the context of maximizing utility—Anne justifiably believes that Dennis wants to discuss what we are to do from the moral point of view. If Dennis wants to make his contribution to the conversation with Anne relevant for a non-moral discussion of giving to charity, then he should indicate this explicitly and make his contribution relevant for that type of conversation. But we are supposing for the moment that he does not do this. In this context, primed as we have it above, Anne—and everyone else—justifiably infers (6) from (5). When there is talk about maximizing utility and giving to charity, then the conversation is about what we are to do morally speaking.

Note that exactly the same type of reasoning will show that the folk will think that eliminativist subjective replacement judgments like ‘I approve of giving to charity’ will carry the implicature to (6) in the conversational context I just primed. The question who approves of giving to charity will, in everyday conversational contexts, invite the thought that deciding whether we should give to charity merits using moral considerations. And so, abolitionists can neither utter eliminativist objective naturalist nor eliminativist subjective naturalist replacement judgments, as all of these judgments suggest to their interlocutors that the conversation at hand merits moral considerations.

6. Cancelling the implicature

But perhaps abolitionists can rescue their theory by cancelling the implicature. Thus, Mary can cancel the implicature of ‘Pete’s car broke down’ (which was that Pete will not be attending the meeting) by quickly uttering something like ‘but mind you, he will take the bus’. Similarly, Dennis can cancel the implicature of ‘Giving to charity maximizes utility’ (which was that giving to charity is morally right) by adding to (5) something like ‘but mind you, I don’t want to be talking morality here’. This radically changes the context of the conversation, altering which judgments (if any) are to be taken for granted in the remainder of the conversation. If this response works, then abolitionists can safely utter eliminativist naturalist replacement judgments. And in that case, Jaquet’s First Argument is saved after all.

Clearly, this response works to an extent. Everyone who can entertain the possibility that there are no moral values and that moral error theory is true—or at least that the person cancelling the implicature believes that there are no moral values and that moral error theory is true —will understand that the implicature has been cancelled. These people will adjust the context of the conversation accordingly. But here is where the trouble starts. Only a very small number of people can switch from talking about giving to charity from the moral point of view to talking about this form a purely prudential point of view. These are the professional philosophers and their students (and not even all of them, since it is not true that they all know metaethics). But if only very few people can adjust the context of the conversation in this way, then the abolitionist still faces the dire consequences of morality. Most of the time, by uttering eliminativist moral judgments, she risks spurring war and violence, contributing to preserving the status quo and blocking reform where it is needed.

But then what can the abolitionist say? There are two points to make here. First, abolitionists can say whatever they want when they are talking to like-minded abolitionists. As she is talking to her abolitionist peers, who will not interpret her as adding a moral overlay to the conversation at hand, she can talk about maximizing utility and satisfying desires as much as she wants. This commits her to a version of propagandism or Government House abolitionism (Cuneo and Christy 2011). My second point is that the abolitionist can utter replacement judgments when she is talking to the folk if, before that, she managed to teach the folk enough metaethics for them to understand, and take seriously, the abolitionist option (Eriksson and Olson 2019). If that works, then the abolitionist can have the sort of conversation with some of the folk that she already has with her abolitionist peers.

Can abolitionists think (replacement) moral thoughts to themselves, as long as she does not communicate these thoughts to the outside world? The argument I have been pursuing does not entail that they cannot. If you keep your thoughts to yourself, then you will not risk that other people perceive you as adding the moral overlay to a discussion. But earlier I argued that abolitionists will not want to think private (replacement) moral thoughts (§3). For the abolitionist thinks that moral and prudential judgments come apart too often, and so for her, thinking the thought that x-ing is morally required will only motivate her to x, which is too often, if not always, prudentially bad for her.

All of this may mean that abolitionism is not, at the end of the day, your preferred answer to the now what question. You may loath its propagandist elitism, its secrecy, or the fact that it requires so much effort in terms of teaching people metaethics before the view can be put into practice. Or you may think that the abolitionist has things exactly right, agreeing that we are better off without a morality that hampers humanity’s progress by obstructing sensible compromises and by igniting violent wars. Be that as it may, the important point for my purposes is that abolitionists cannot accept naturalist replacement judgments at all, neither schmoralist (§4) nor eliminativist (§5-§6) versions of these judgments. This means that Jaquet’s First Argument for the collapse claim fails. This argument falsely presupposes, by virtue of its first premise, that abolitionists can accept replacement judgements.

7. Extending the argument to expressivism

In this section I argue that if my argument saves naturalism from collapsing into abolitionism, then it also saves expressivism from collapsing into abolitionism. This gives us some reason to think that, mutatis mutandis, none of our current answers to the now what question collapses into abolitionism.

Schmoralist expressivists will continue to use moral terms in their replacement judgments, and the folk will take these judgments to entail that the conversation merits moral considerations. Eliminativist expressivists will eliminate moral terms from their replacement judgments altogether. However, their replacement judgments will, like the eliminativist naturalist’s judgment ‘giving to charity maximize utility’, conversationally imply that its utterer wants to add the moral overlay to the conversation. To see this, we must distinguish between eliminativist expressivists without the quasi-realist agenda, such as A.J. Ayer, and eliminativist expressivists with the quasi-realist agenda, such as Simon Blackburn (Ayer 1936; Blackburn 1984, 1993, 1998, 1999).

Blackburn describes the quasi-realist agenda, or the enterprise of quasi-realism, as follows:

I call the enterprise of showing that there is [no mistake in ordinary moral thinking] – that even on antirealist grounds there is nothing improper, nothing ‘diseased’ in projected predicates – the enterprise of quasi-realism. The point is that it tries to earn, on the slender [metaphysical] basis, the features of moral language … which tempt people to realism. (Blackburn 1984, 171).

Expressivists with the quasi-realist agenda, or quasi-realists for short, employ a number of arguments to earn the right to the features of moral language that tempt people to realism. For example, Blackburn appeals to a minimalist theory of truth to get moral truth, and therefore true moral judgments, without correspondence to reality (Blackburn 1999, 214).

Contrastingly, expressivists without the quasi-realist agenda, or early non-cognitivists for short, do not aim to earn the features of moral language that tempt people to realism. For example, Ayer denies the truth-aptness or primary cognitivist meaning of moral judgments. He does so by employing this two-fold theory of meaning according to which moral judgments are meaningful just in case they are either verifiable in principle or tautologies (Ayer 1936, 48–50). And of course, for Ayer, moral judgments are neither in principle verifiable nor tautologies. Instead, moral judgments are meaningful only in a secondary sense; viz., if they manage to express or arouse emotions (Ayer 1936, 137). Thus the true meaning of the judgment ‘stealing is morally wrong’ is something like ‘Boo! Stealing’. And the true meaning of ‘giving to charity is morally good’ is something like ‘Hurray! Giving to charity’.

I spoke of descriptive quasi-realism and descriptive early non-cognitivism in the previous paragraph. I now return to the prescriptive versions of these theories and ask if abolitionists can utter eliminativist quasi-realist or eliminativist early non-cognitivist replacement judgments (inclusive or). First, and quite clearly, eliminativist quasi-realism is out, for this is an internally contradictory theory. You cannot have a theory that aims to allow its users to think and utter moral judgments that feel and sound like real moral judgments—and yet takes much away from that aim by eliminating the moral terms from the replacement moral judgements that contribute to the cognitivist realist feel of these judgments.

Early expressivists do not want to mimic our current moral judgments. As a descriptive theory, this eventually created too much trouble for the view. Indeed, the theory accrued so many objections that even though ‘in the 1930s and 1940s [it] bested the competition and dominated the scene of analytic metaethics,’ hardly anyone writing today embraces it (Darwall, Gibbard, and Railton 1992, 120). As is well-known, one of early non-cognitivism’s main problems is its inability to explain fundamental moral disagreement; viz., disagreement about what the right ethical feelings are, or about which ethical sense we ought to have (Ayer 1936, 147). But as a prescriptive metaethical theory, early non-cognitivism may allow the abolitionist to utter eliminativist judgments such as ‘Boo! Stealing’ and ‘Hurray! Giving to charity’.

The trouble, though, is that these judgments conversationally imply that their utterers add the moral overlay to a conversation. To see this, consider once more Dennis and Anne and their conversational context. Dennis is an eliminativist naturalist and utters (7). Is Anne, who has never studied metaethics, justified to infer (8)?

1. Hurray! Giving to charity.

2. Giving to charity is morally right

For all the same reasons that Anne is justified to infer (6) from (5) above, she is justified to infer (8) from (7). Basically, because everyone always talks about morality if the topic of giving to charity is brought up—especially if Dennis supports it rather vehemently by expressing such a positive emotion about it—Anne justifiably believes that Dennis wants to discuss what we are to do from the point of view of morality.

I conclude that abolitionists cannot utter schmoralist naturalist replacement judgments (§4), that they cannot utter eliminativist naturalist replacement judgements either (§5-§6), and that my argument extends beyond naturalism to expressivism and probably, mutatis mutandis, to the other answers to the now what question (§7).

8. Good news for error theory

There is at least one important reason why error theorists should be happy with the result that the various extant prescriptive metaethical views do not collapse into abolitionism. It is that the availability of non-abolitionist answers to the now what question renders it psychologically possible for people to accept the error theory.

One of the major complaints about error theory is that the view impoverishes our normative deliberation. Matt Lutz writes:

moral thought, discourse, and action play a large role in all of our lives. Error theorists, it is widely supposed, put themselves in a position where they must abandon all of these commitments, thereby leaving a gaping hole in their normative lives (Lutz 2014, 352).

Imagine what it would be like to do your practical deliberation without the aid of moral terms and moral judgments—and what it would be like to lack publically available moral judgments to communicate your considerations in moral terms.

For instance, imagine what it would be like to engage with Peter Singer’s shallow pond thought experiment without access to moral judgments:

if I am walking past a shallow pond and see a child drowning in it, I ought to wade in and pull the child out. This will mean getting my clothes muddy, but this is insignificant, while the death of the child would presumably be a very bad thing (Singer 1972, 231).

Without moral judgments, you would not be able to say to yourself that you morally ought to wade in and pull the child out, nor would you be able to discuss the matter in moral terms. It may still be true that you should save the child as far as prudence is concerned. Even without morality, prudence will command you to save the child—you will lose your friends and ability to sleep at night if you don’t. Stronger still, it may also even be true that, in this case and other cases like it, the prudential considerations are so important (carry so much normative weight) that they silence whatever other considerations there may be that speak against wading in to save the child. So it may still be true that, even without morality, you ought to save the child all-things-considered.

But for too many people, this is not enough. Too many people will not be able to live their lives without any tools to moralize their decisions and discussions. Joshua Greene, a philosopher and neuroscientist, writes that our brains are ‘moral brains’ that adopt immediate emotional responses and conscious judgments in order to foster cooperation and enhance the agent’s chance to remain a member of the group, which brings all sorts of evolutionary advantages (Greene 2013, 60). Roughly this sort of message can also be found in the work of Jonathan Haidt and Richard Joyce (Haidt 2001, 2012; Joyce 2016). William Lycan writes that ‘to produce a genuine freedom from moral intuitions, one would need a steady diet of hard drugs, or some other very powerful alienating force’ (Lycan 1985: n29). Similarly, Nolan et. al write that people are ‘unable to refrain from making positive moral judgments’ (Nolan, Restall, and West 2005, 314). We can find similar remarks in Peter Singer’s The Expanding Circle and in Derek Parfit’s On What Matters (Singer 1981; Parfit 2012).

True, some philosophers, such as Garner, deny that it is too hard to stop moralizing (Garner 2007, 505). Garner argues that, at least compared to getting rid of the concepts of folk psychology, cutting back on moral pronouncements is possible. But people like Garner are very much the minority. The evidence, both from empirical studies and from philosophical (phenomenological) reflection on our own practices, suggests that refraining from moralizing is extremely difficult.

If I am right that their current answers to the now what question do not collapse into abolitionism, then this is good news for error theory. For this would render it psychologically possible for people to accept the error theory. Presumably, and although people cannot live without any type of moral judgments at all, they can live with replacement judgments of various sorts, such as various naturalist and expressivist replacement moral judgments.

This argument makes two important assumptions. First, it assume that error theorists should aim for a view that they and other people can actually accept. As David Lewis puts the point elsewhere, you should.

never put forward a … theory … you cannot yourself believe in your least philosophical and most commonsensical moments (Lewis 1986, 135)

We can, and perhaps should, say much more about Lewis’ maxim.¹² But I think that Lewis is right and that we should not defend a view that we simply cannot believe or accept. A view in practical philosophy—a view that is ultimately about what people should do—that people cannot accept is, for that reason, disqualified. Fortunately, if I am right, error theory is acceptable, for we have access to the various prescriptive answers to the now what question.

Second, the argument I just gave also assumes that we do not yet know which prescriptive metaethical theory we should embrace. Had we already had our preferred answer to the now what question—say conservationism—then naturalism and expressivism are false, and whether a false view collapses into another view is neither good nor bad news.¹³ However, I imagine a dialectical situation in which we do not yet know which prescriptive metaethical theory we ought to embrace. In that situation—and I think that this is our current situation—we need a plausible response to the now what question to render error theory acceptable. Had Jaquet been right that every extant answer to the now what question (except fictionalism) collapses into abolitionism, then, if neither of these two views is plausible, the error theory becomes psychologically difficult to accept. And so in that situation, it is good news for the error theorist that she can choose from many candidate answers to the now what question. Or indeed, and remaining within this dialectical situation, we may imagine a metaethicist who is aware of the various descriptive metaethical theories and the arguments for and against them, but who is unsure about whether she should embrace the error theory rather than, say, robust realism. One thing that may hold her back is the difficulty of embracing error theory. In this case too it will be good news for the error theory if our undecided metaethicist has a good number of prescriptive metaethical theories to choose from.

9. Conclusion

François Jaquet has argued that all the error theorists’ answers to the now what question except fictionalism collapse into abolitionism (Jaquet 2020). This paper has provided a response, arguing that various distinct answers to the now what question fail to collapse and that this is good news for error theorists. Error theory has many problems to overcome, but lack of recourses to answer the now what question does not seem to be one of them.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In what follows I will sometimes, for ease of exposition, omit the qualifier ‘moral’ in moral error theory. I do this because I am only concerned with moral error theorists and not, for instance, with error theorists about colour. I will also sometimes write that error theorists believe that all moral judgments are false (rather than, e.g., ‘untrue’) even though many philosophers—including some error theorists—believe that this is problematic (Kalf 2018; Perl 2019). I do this because whether or not error theorists can say that all moral judgments are false is a moot point in the context of this paper, which is about a question that lies further downstream (viz., the now what question).

2 The ought is not a moral ought but the all-things-considered ought, which combines the various extant pro tanto oughts (prudential, epistemic, etc.) but disregards moral oughts. The error theorist’s main task is to avoid the Scylla of continuing to accept false moral beliefs and the Charybdis of losing moral judgments. Accepting false moral beliefs is bad because you will loosen your epistemic standards and as such you open the floodgates to a whole bunch of other false beliefs. Eventually you run the risk of becoming an ‘epistemic wreck’ (Joyce 2005, 299). But abolishing moral judgments altogether is not a good idea either because (many) people want to continue to use moral judgments’ motivational power to get themselves to do what they (all-things-considered, but not morally) ought to do.

3 From now on I will sometimes just write about moral judgments, and when I do, I will—unless stated otherwise—have both private moral thoughts and publicly available moral judgments in mind.

4 Henceforth, I will often omit the qualifier ‘revisionary’ since the versions of these views that I tend to discuss are the revisionary metaethical theories that can be adopted by error theorists. When I have the hermeneutic or descriptive versions of these metaethical theories in mind, I state this explicitly.

5 The only exception, according to Jaquet, is fictionalism. To aid readability, I will not add this exception clause every time I write about Jaquet’s collapse claim, especially since we agree that fictionalism fails to collapse.

6 Objective theories of prudence, such as the objective list theory, allow us to stop here. We just say that morality is prudentially bad, full stop. Subjective theories of prudence, such as hedonism or the desire-satisfaction theory, require us to say that morality is prudentially bad for an agent, or for some agents. My arguments go through on both objective and subjective theories of prudence. After all, abolitionist moral error theorists can be naturalist realists about prudence, claiming that prudential facts are natural facts in the world, and therefore objective, albeit without the irreducibly normativity that makes objective moral facts too queer to exist. In what follows, and to aid readability, I will sometimes write that morality is prudentially good or bad full stop, and I will sometimes write that it is prudentially good or for us, where the ‘for us’ signifies a subjective theory that I believe abolitionists will want to say holds for many though not necessarily all of us.

7 Why not consider whether morality is all-things-considered bad for us simpliciter, taking into account not just prudential normativity but also epistemic, legal, aesthetic and other sorts of normativity? I believe that this is what abolitionists should do but I also think that abolitionists think that the importance of non-prudential norms pales into insignificance compared to prudential norms. In this paper I follow contemporary abolitionists and assume that the question whether morality is overall prudentially bad for us probably tell us whether morality is overall bad for us simpliciter.

8 I thank an anonymous referee for asking me to engage with these quotes.

9 I thank an anonymous referee for this quote and for asking me to state much more explicitly that my objection to Jaquet’s Collapse Argument only works on one of these two interpretations of the now what question.

10 Strictly speaking, semantic entailments hold between the propositions that these judgments express, independent of the context in which these judgments are uttered. But for ease of exposition, and because the error theorists I engage with all think that moral judgements express truth-apt propositions, I omit this qualification here.

11 There is considerable debate in literature about the exact mechanisms behind the notion of a conversational implicature, with different people offering different formulations of the cooperative principle and the maxims (Levinson 2000). In this paper I adopt Grice’s original account because I think that it explains the phenomena I’m interested in sufficiently well.

12 See Bart Streumer (2017) and Chris Daly and David Liggins (2010).

13 I thank an anonymous referee for asking me to engage with this point.