In this paper, I develop a methodological challenge for ethical nonnaturalism. The challenge is methodological because it concerns the way many nonnaturalists argue for their views. I suggest that there is an overlooked problem for a central and prevalent positive argument for nonnaturalism, the argument from ethical phenomenology. This problem, I intend to show, ultimately renders nonnaturalism indefensible—at least in so far as the view is solely based on this argument.

Let us start by clarifying the goals of metaethical theorizing. Here is a useful characterization:

That is, metaethics concerns the nature of moral thought, moral language, moral facts, moral properties, and moral knowledge.¹

Nonnaturalists believe that ethical thought and talk involves nonnatural entities.² What does that mean? Nonnatural entities are thought to be categorically distinct from, or “something over and above,” the natural.³ Nonnaturalists typically do not claim that all ethical entities are nonnatural. Some ethical entities are “mixed”; they consist of a combination of natural and nonnatural entities. (For example, the fact that Anna’s hitting Ben is wrong consists of a natural part—the hitting—and a non-natural part—the hitting’s wrongness.) But, crucially, nonnaturalists claim that the most fundamental ethical entities are “purely” nonnatural (cf. Scanlon 2014, 36–37). In this sense, they are categorically distinct from, or something over and above, natural entities.⁴

Why believe that ethical entities are nonnatural? One prevalent nonnaturalist argument—the argument from ethical phenomenology—takes the form of an inference to the best explanation and consists of two steps: First, describe the phenomenology of ethical deliberation. Second, show that the best explanation for it—the best explanation for why this is what ethical deliberation is like—involves the existence of nonnatural entities.

The typical naturalist response to the argument from ethical phenomenology is that there are better explanations for the phenomenology of ethical deliberation than the existence of nonnatural entities. However, we will pursue a different path here. Our methodological challenge is logically prior to responses of this kind. We will try to show, not that there are better explanations, but that, quite generally, the outlined way of arguing for the existence of nonnatural entities is methodologically problematic. In short, our charge will be that it is methodologically unreasonable to explain or interpret ethical phenomenology by making metaphysical claims without taking into account another, more “external” perspective on ethical thought and talk.

Here is our plan. Section 1 introduces two distinct perspectives on mental processes and argues that both perspectives are important when it comes to understanding how these processes fit into reality. Ethical deliberation is a mental process, and so it will be worth reflecting on how, in general, philosophers should approach these processes. Based on the insights gathered here, section 2 introduces the challenge from lost perspective in the context of David Enoch’s work (Enoch 2011). This section is the heart of the paper. Section 3 discusses two nonnaturalist attempts to meet the challenge Parfit (2011). Both attempts involve the so-called “just too different intuition.” I show why they cannot succeed. At this point, it will hopefully have become clear that the argument from ethical phenomenology runs into a serious methodological problem. It can only get off the ground by presupposing something opponents of nonnaturalism (whether reductionists, expressivists, or error-theorists) deny, namely, that the external perspective is irrelevant for metaethical theorizing. The argument, in other words, begs the question on a methodological level. The final section sums up our main points and recommends a strategy to future nonnaturalists.

1 Reconciling Two Perspectives

As Mark Timmons (1999) and Terence Cuneo (2007b) have helpfully emphasized, the metaethical project can be described as a twofold endeavor. The first part of it is the “internal accommodation project”: developing a theory of ethical thought and talk that fits well with “deeply embedded assumptions” of our ordinary ethical thought and practice (Cuneo 2007b, 854). In other words, the internal accommodation project aims for the theory that best accounts for our internal perspective on ethics, our ethical phenomenology. For example, it is (presumably) a deeply embedded assumption of ethical thought and talk that if an agent has a moral belief, she is pro tanto motivated to act accordingly. So, a plausible metaethical view should account for this feature.

The second part of the metaethical project is the “external accommodation project.” Its goal is to come up with a metaethical theory that fits well with the “scientific world view.” For example, a metaethical view should, at least, not directly contradict scientific insights into human nature as presented by, say, evolutionary biology or empirical psychology. Ideally, a metaethical view would get further evidential support from scientific research such that we, ultimately, get a unified “phenomenological-cum-scientific” theory of ethical thought and talk. However, it might also turn out that the ethical domain is “autonomous,” and that scientific insights are simply irrelevant when it comes to the fundamental ethical entities. If so, the external accommodation project would (maybe trivially) be completed, but more about that later.

These two explanatory projects form the basis of our challenge to nonnaturalism.⁵ In the following, we will distinguish the internal perspective from the external perspective. The internal perspective delivers the stuff relevant for the project of internal accommodation; it grants access to some process or practice “from within.” The external perspective delivers what is necessary for the project of external accommodation; it provides insights into some process or practice “from without,” by means of investigations that are not phenomenological.⁶

Importantly, I take the external accommodation project to cover more than just the methods of the natural sciences. What I mean is the a posteriori investigation of a process or practice that goes beyond phenomenological observations. For example, an anthropological investigation of the practice of monetary transactions counts as external. Such an investigation looks at the practice “from without,” for instance, by focusing on the societal advantages of trade. It is based on insights gathered from the external perspective (and not based on the “phenomenology of money experiences”).

Back to nonnaturalism. Is the idea that there are nonnatural entities the result of external or internal accommodation? As we are about to see in the following section, the claim typically results from an internal accommodation. Nonnaturalists usually start with ethical phenomenology and then proceed to explain it via metaphysical hypotheses that involve nonnatural entities.⁷ But, importantly, these hypotheses are not directly “revealed” by internal, phenomenological analyses. Instead, they are interpretations of our phenomenology. And these interpretations are part of the nonnaturalists’ internal accommodation project because they are solely based on phenomenological appearances.

Now, let us illustrate how both perspectives on mental processes can be brought together. Take the example of human disgust. We could either start investigating disgust “from within,” that is, with its what-it-is-like. This would involve, say, analyzing the stream of thoughts and feelings present in disgust episodes. Or we could assume the external perspective and explain, “from without,” what anchors disgust reactions in the empirical world. This would involve, for instance, analyzing (neuro)physiological processes and disgust’s evolutionary function.

Start with the internal perspective. What is it like to encounter rotten food? You feel a strong inclination or desire not to get too close to the food. Touching it with your bare skin strikes you as repulsive. You might experience nausea. You want to get rid of the rotten food as quickly as possible. And if you imagine having accidentally put it into your mouth, your reactions further escalate. Yuck, away with it!

Now, trying to come up with a theory of disgust, you might discover that there are many other disgusting things. There are greasy, sticky, or malodorous objects, blood, mutilation, waste, hygiene violations, and even some animals (e.g., rats, cockroaches, worms, or flies). This can seem quite puzzling: Why is it that we react to all these different things in the same way?⁸ Do they have something in common that might explain our reaction to them? Is there more to find out and understand about disgust than we can observe from the internal perspective?

Of course there is. But in order to find out more, we need to assume the external perspective. According to a widely accepted scientific theory, disgust is a behavioral extension of the immune system (Rozin, Haidt and McCauley 2008). It helps us to avoid pathogens. Very roughly: disgust is triggered when we encounter something potentially infectious, which helps us to avoid it. So, assuming the external perspective on disgust is quite illuminating. Undoubtedly, our understanding has been enriched by it. On top of the detailed phenomenological descriptions of what it is like to experience disgust, we now also understand what anchors disgust in the world as conceived by the natural sciences. We have a better grasp of its “point”—of why beings like us are disgusted in the first place. We also better understand why there are a whole range of different things that evoke the same disgust reactions. Blood, greasy objects, and rats are all “signs” for the presence of pathogens—and thus to be avoided. In a first and preliminary attempt, we might (partly) characterize disgustingness as something along the lines of being an indicator of the above-some-threshold likelihood of the presence of pathogens.⁹

I take it that disgustingness is a good example because of its evaluative or normative dimension.¹⁰ What renders a property evaluative? (McDowell 1985, 143–146) distinguishes non-evaluative properties that “merely” causally influence our responses from evaluative properties that merit certain responses. His criterion for assigning a property to the evaluative camp is “the possibility of criticism” (1985, 144). Now, I think it is fair to say that a dead rat in one’s fridge merits disgust. If Fred discovered a dead rat in his fridge and showed no signs of disgust while happily starting to eat the open bowl of yoghurt that has been standing right next to the cadaver, we would ask ourselves what is wrong with him. Thus, I side with McDowell and state that disgustingness has an evaluative dimension. So, even in the case of properties with an evaluative or normative dimension, external insights can be quite resourceful.¹¹

The above considerations set the stage for the main claim of the current section:

Let me elaborate. Suppose Danielle wants to investigate the nature of disgust. She only cares for a phenomenological investigation, and so she never even considers taking into account what the sciences have to say. Scrutinizing disgust phenomenology for a few days, she ultimately concludes that disgustingness is a nonnatural property that human beings can apprehend via a special, intuition-like faculty. Some otherwise seemingly unrelated objects (blood and cockroaches, say) instantiate this property, and somehow the human mind can recognize it. Note that nothing in the phenomenology of disgust speaks against Danielle’s disgust nonnaturalism; her view accounts (we may assume) for all the relevant phenomenological data quite well. But now suppose that Danielle’s friend Fatima decides to tell her all the scientific insights about human disgust reactions. She tells her that disgust tracks possible sources of infection and that scientists consider this tracking function as its evolutionary point. Now, here is a crucial question: Coming to learn all the external facts about human disgust reactions, should Danielle’s confidence in disgust nonnaturalism change?

I believe that, upon learning the external facts, it would be rational for Danielle to change her confidence in disgust nonnaturalism. These newly learned facts suggest—and this is a crucial step in my argument—that disgustingness is closely metaphysically linked to something quite natural: the likely presence of pathogens. It is due to this suggested metaphysical link that Danielle should take her disgust nonnaturalism to be less plausible than before.¹² Coming to know the external evidence, it is rational for Danielle to decrease her confidence in the idea that disgustingness is something categorically distinct, something “over and above,” the natural. It must now seem more likely to her that disgustingness fits into reality by being a natural property. (Note that Danielle now understands why blood and cockroaches instantiate disgustingness.) Consequently, she should decrease her confidence in the idea that disgustingness is a nonnatural property.¹³

Based on these considerations, we may formulate a (not entirely catchy) slogan: External evidence can shift the plausibility of metaphysical explanations of the phenomenology of mental processes. As we just saw, the external perspective on human disgust reactions influences the plausibility of Danielle’s disgust nonnaturalism. In virtue of plausibility shifts of this kind, it is methodologically unreasonable to draw metaphysical conclusions about the nature of disgustingness on solely phenomenological grounds. If we want to find out how any mental process fits into the reality that the empirical sciences have taught us so much about, it would be a bad idea to disregard possibly relevant empirical evidence.

We may put two points on record. Firstly, the internal and the external perspectives on disgust complement each other. Reconciling them helps us “anchor” disgust in the natural world. Moreover, adding the external perspective to Danielle’s investigation changes the plausibility of her solely phenomenology-based metaphysical account of disgustingness. So, if you want to write a book titled “Disgust: What It Is and How It Fits into Reality” you should take the external perspective into account. Not doing so would be methodologically unreasonable.

Secondly, our two perspectives deliver characterizations of disgust that look very different but are intimately linked. For example, part of a phenomenological description of disgust is the “yuck”-reaction, a strong inclination to get rid of the disgusting object. There seems to be a large gap between this description and the external story, which includes, besides a list of facts about neurophysiology and muscle twitches, that disgust is an evolutionary tool for tracking and avoiding possibly infectious objects. Despite this gap, there is an intimate connection. Plausibly, the disgustingness of the dead rat in your fridge (partly) consists in the likelihood of its being a source of infection. A close metaphysical link between the dead rat’s disgustingness and some set of scientifically accessible properties can, at least, not be ruled out.^14,15

These two methodological conclusions, I think, apply to mental processes more generally. The case of disgust suggests that, whenever we investigate a mental process, we should take into account both perspectives on it—unless there is reason to believe that one perspective is utterly irrelevant for investigating the respective mental process.¹⁶ As long as we don’t know about such a reason, we should be open to all the internal and external evidence we might get hold on—which lets us formulate two methodological guide lines:

1. When you interpret or explain the phenomenology of mental processes (and the involved entities), take into account both the internal and the external perspective on the respective processes.

2. While the internal and the external perspective might describe mental processes (and the involved entities) in very different ways, do not take this to rule out that the entities mentioned in both descriptions are closely metaphysically linked.

In this section, we have argued that an investigation of the nature of any mental process should take into account and try to reconcile both the internal and the external perspective. This will serve as a fruitful ground for our objection to the argument from ethical phenomenology. As we are going to claim in the upcoming section, the argument violates our first methodological guideline; it constructs a moral metaphysics on phenomenological grounds without taking into account the external perspective.

2 The Challenge from Lost Perspective

Ethical nonnaturalists have a rich history of constructing ethical ontologies out of phenomenological analyses of ethical deliberation. They answer the question of how ethical entities fit into reality by stating that reality comprises more than the sciences would have us believe. There are, they claim, nonnatural ethical entities. Depending on what particular view we are dealing with, these entities are truths, facts, properties, or relations. But whatever they are, the crucial idea is that they are something categorically distinct from, something over and above, the natural.¹⁷ Now, let us take a closer look at one version of the argument from ethical phenomenology.

David Enoch advocates the argument from the moral implications of objectivity (Enoch 2011, 16–49). It runs as follows: In cases of preference conflicts—say, about where to have dinner tonight—it intuitively seems that we should solve the conflict impartially. It would not be okay to declare that Mark’s preference for Italian is more important than Anna’s preference for Indian. Intuitively, they should agree that their preferences count the same, and then find a solution from here on out. Clearly, none of their preferences is mistaken. On the other hand, in a moral conflict, it intuitively seems that the appropriate response is not impartial. For example, if I disagree with someone claiming that not a single refugee from Ukraine should be allowed to cross the German border, she strikes me as mistaken. It seems to me that my opinion has some objective backing—and that an impartial treatment of our “moral preferences” would be deeply misguided. So, there is an internal, phenomenological difference between moral disagreements and conflicts of preference. The former ones have (or seem to have) an objectively right answer. The latter ones don’t. And this, according to Enoch, is “best explained” by a robust nonnaturalist realism ((Enoch 2018, 40); (Enoch 2011, 16–49)).

This argument fits the general pattern of the argument from ethical phenomenology. Starting with phenomenological observations about the differences between moral disagreements and conflicts of preference, it draws a metaphysical conclusion to explain this difference. So, the argument is a suitable target for our methodological worries.¹⁸

There are, of course, many other versions of the argument from ethical phenomenology.¹⁹ However, in the following, I will mostly rely on considerations from Enoch (2011) because they strike me as particularly straightforward. I hope it will become clear that my methodological worries can be extrapolated to different versions of the argument from ethical phenomenology proposed by other nonnaturalist authors. Let us turn to these worries now.

Metaethics, we said, is the project of explaining how ethical thought and talk, and what it is about, fits into reality. Now, trivially, reality does not exhaust itself in phenomenology. As the case of disgust served to show, the phenomenology of a mental process might only be one side of the coin. Sometimes, there is another side—a side that is only revealed if we look at the process from the external perspective. Therefore—and in the absence of reasons to the contrary—we should take into account both perspectives when trying to understand how a mental process and the involved truths, facts, properties, or relations fit into reality. If you want to write a book titled “Ethical Thought and Talk: What It Is and How It Fits into Reality” and you are not planning to even look at the subject matter from an external perspective, chances are you are missing something relevant. This would be methodologically unreasonable. We already saw how external evidence can shift the plausibility of metaphysical claims that solely rest on phenomenological observations. Due to the possibility of such shifts, you should at least give the external evidence a shot at informing your metaphysics. And so we may raise the following challenge:

There is a slight chance that nonnaturalists remain unimpressed by this challenge. They might ask: What could the external perspective possibly contribute to our understanding of ethics? I have a quick and a not-so-quick reply. Here’s the quick one: The question of how ethical thought and talk fit into reality is a descriptive question about the reality we live in. We already know that there are many truths about this reality that cannot be discovered by phenomenological investigations. Therefore, it strikes me as quite commonsensical to at least entertain the possibility that the external perspective—which has proven quite resourceful in teaching us about the nature of reality—has something to contribute here. But since this answer might be considered too superficial, let me try again and present my not-so-quick reply.

Suppose we have two different explanations of the phenomenology of ethical deliberation on the table. One of them is nonnaturalism, according to which the “currencies” of ethical deliberation—values and reasons—essentially involve nonnatural entities. The other one is a broadly “Humean” explanation, according to which values and reasons are grounded in our conative, desire-like attitudes. They are, as (Finlay 2014, 249–250) nicely puts it, “shadow[s] cast by our desires […].” How could the external perspective contribute anything to this debate between the nonnaturalist and the Humean?

Here is one possibility: It might turn out that, from an external perspective, ethical deliberation is an evolutionarily acquired tool for “conative mind-management,” that is, for dealing with conflicts between and hierarchizing our conative attitudes.²¹ As human beings with a capacity for imagination, a limitless time horizon, deeply entrenched social needs, and thus a multitude of conflicting attitudes, we face an enormous evolutionary challenge: managing our minds in order to be coherent agents, and then coordinating our actions with our fellow community members. Investigating the human mind from the external perspective of evolutionary anthropology, we might encounter the hypothesis that ethical deliberation is an evolutionary, cultural tool for solving this challenge.²² Let me be clear: I do not want to argue for this hypothesis. My main point is conditional, but it suffices to answer the question of what the external perspective could possibly contribute. If the external perspective revealed something along these lines, this would (much like in the case of disgust) shift the plausibility of the nonnaturalist and the Humean explanations. How? Well, the nonnaturalist explanation would lose some plausibility points, whereas the Humean explanation would gain some. Why? Because metaethics is concerned with explaining how ethical thought and talk fit into reality and because, as argued above, we should take into account and try to reconcile both perspectives in this process. If the “external point” of ethical deliberation turned out to be conative mind-management, this would fit better with a broadly Humean view, according to which there is a close metaphysical link between values and reasons on the one hand, and conative attitudes on the other hand. Since nonnaturalists reject such a link, their explanation would lose some plausibility points. Additionally, combining a Humean view with our stipulated external story would promise a more parsimonious account of how ethical thought and talk fit into reality.²³ This is how the external perspective could contribute to the metaethical debate between the nonnaturalist and the Humean.

The outlined external story about the evolutionary point of ethical deliberation is, of course, hypothetical. But our general methodological consideration is not. We argued that external investigations into mental processes can (and often do) shift the plausibility of (metaphysical) interpretations of the respective phenomenologies. Thus, we should take into account the external perspective when developing and assessing these interpretations. Importantly, this holds even if external evidence ultimately turns out to be irrelevant for metaethical theorizing. Even in that case, it would still be true that disregarding the external perspective would have been methodologically unreasonable; when we started the investigation, we simply didn’t know.

This means that proponents of the argument from ethical phenomenology face a problem. They proceed in a methodologically unreasonable way. They construct a controversial moral metaphysics on phenomenological grounds without taking into account the external evidence.

Let us put a concrete example on the table. Enoch’s second main argument for nonnaturalism is the argument from deliberative indispensability. Like his first argument, it is a version of the argument from ethical phenomenology. When introducing it, Enoch explicitly disregards the external perspective as irrelevant.

Had we been here in the explanatory business – trying to explain action, or perhaps even deliberation, from a third-person point of view – perhaps desires would have been enough (though I doubt it). But the whole point of the argument of this chapter is the focus on the first-person, deliberative perspective. And from this perspective, desires are not often relevant, and whether they are or are not, the normative commitment is – though perhaps implicit – inescapable. […] [W]e need normative truths even if, viewed from an external perspective, our desires suffice in order to cause our actions and then explain them, because, when deliberating, we know our desires are merely our desires. (Enoch 2011, 76, footnotes left out)

Interestingly, Enoch seems to agree that there is an external perspective from which deliberation could be investigated. But then he dismisses the relevance of possible external insights—desires could help to explain the nature of deliberation—for the purposes of his chapter because desires play no important internal role on the conscious mental stage of deliberation.²⁴ The whole point of his chapter, he suggests, is to better understand the nature of normative truths from a first-person point of view. And, by the end of the chapter, he concludes that we should best think of these truths as nonnatural. So, according to what we have said, Enoch’s approach is methodologically unreasonable; his two main arguments for ethical nonnaturalism construct a moral metaphysics on phenomenological grounds without taking into account the external perspective.

To be fair, however, we should mention that Enoch does consider the external perspective on ethical deliberation later in his book.²⁵ There, he discusses Sharon Street’s Darwinian Dilemma for Moral Realism (2006) as an epistemological challenge to his view. We won’t dive into the details. For our purposes, it suffices to focus on the way Enoch replies to Street’s dilemma. First, he reminds us that metaethics is about scoring plausibility points. Ultimately, he says, metaethicists offer package deals, and the one with the most plausibility points wins. In this spirit, Enoch preliminarily remarks that his view does not need to do “better than competing metanormative theories in every respect, with regard to every problem” (Enoch 2011, 167). And so he sets out to show that his two positive arguments for nonnaturalism scored him more points than he was about to lose due to the epistemological challenge. Ultimately, after having presented his solution to the challenge, he states: “Let me not give the impression that this suggested way of coping with the epistemological challenge is ideal. […] [P]erhaps Robust Realism does lose some plausibility points here. But not, it seems to me, too many, and certainly not as many as you may have thought” (2011, 175). So, Enoch believes that his two main arguments for the existence of nonnatural ethical facts—two different versions of the argument from ethical phenomenology—generate such a significant number of plausibility points that later objections to his view, formulated from an external perspective, can be met via an inferior solution—because he doesn’t lose as many points as he previously scored.

I find this rather unconvincing. It will take the rest of this section to explain why.²⁶ We argued earlier that, when interpreting or explaining mental processes, it is methodologically unreasonable to draw metaphysical conclusions on solely phenomenological grounds. Now, start by noting that this is precisely what Enoch does when he develops his positive arguments for nonnaturalism—even if it is true that he later confronts his metaphysical conclusions with an objection formulated from the external perspective. For all we said above, the external evidence regarding the nature of ethical deliberation may have significantly decreased the plausibility of Enoch’s metaphysical conclusions—in which case we should never have drawn them in the first place.

But nonnaturalists might want to object: Does it really matter when we take into account the external perspective? Enoch clearly does take it into account, so where is the problem? As long as we do take it into account at some point, we should be fine, shouldn’t we? I don’t think so. It actually does matter when we take into account the external perspective because, as long as we don’t, we cannot assign plausibility points to our metaphysics. Without taking into account the external evidence, we simply cannot know how plausible our solely phenomenology-based metaphysical explanation is. But this is a complicated thought, so let me elaborate a little.

As we just saw, Enoch is quite confident that, despite his less than ideal solution to the epistemological challenge, he “certainly” does not lose as many points as he previously scored. Let us reconsider his approach in light of our methodological worries. Enoch first explicitly disregards a perspective it is, we argued, methodologically unreasonable to disregard. This allows him to draw his metaphysical conclusions precisely in the way the way we claimed to be methodologically unreasonable. Later, Enoch confronts his metaphysical picture with objections from the perspective that he previously disregarded. Doing so, he finds that his metaphysical picture, which was drawn, again, in a methodologically unreasonable way, gained such a high (!) number of plausibility points that they “certainly” cannot be outweighed by objections generated by the perspective whose taking into account would have stopped his conclusions from being methodologically unreasonable in the first place.

This strikes me as fishy. When we construct a metaphysics on solely phenomenological grounds, we should expect that, once we add the external perspective to our investigation, the plausibility of our metaphysics might change. (Recall Danielle’s disgust nonnaturalism.) But this means that we cannot—and, importantly, Enoch cannot—confidently distribute plausibility points to his metaphysics before weighing in the external evidence. This, I think, is a crucial implication of our earlier methodological considerations. If these considerations are correct, if drawing metaphysical conclusions on solely phenomenological grounds is methodologically unreasonable, then the plausibility of these conclusions should be considered uncertain as long as we haven’t weighed in the external evidence. In other words, our methodological considerations suggest that the number of plausibility points Enoch’s moral metaphysics scores itself depends on how well it fits with the external evidence. Therefore, Enoch’s allocation of any particular number of plausibility points to his metaphysics—let alone a high number of points—is unwarranted. Enoch simply cannot know how plausible his metaphysics is until he has taken the external evidence into account.²⁷

Consider an analogous case. Tim wants to investigate the nature of taste. At the beginning of his investigation, he explicitly disregards the external perspective. His solely phenomenological investigation leads him to the conclusion that tastiness is a complex, nonnatural property. Later, however, a colleague shows Tim all the tastiness insights that science has to offer (e.g., the evolutionary insight that chocolate is tasty because it is a great source of energy). After considering the scientific evidence, Tim replies: “Ok, I may lose some plausibility points here, but my original, nonnatural hypothesis has gained me so many plausibility points that this loss poses no threat to my overall theory.”

This would clearly be an unsatisfying reply. Why? Well, for the same reason as before. Due to the importance of taking into account both perspectives when investigating how some mental processes (and the involved entities) fit into reality, the plausibility of Tim’s “metaphysics of taste” should be considered uncertain until we weigh in the external evidence. The plausibility of Tim’s view surely depends, among other things, on how well it fits with the best scientific understanding of tastiness. And, thus, Tim cannot reasonably assign a high number of plausibility points to his metaphysics and then compare this number with the number of points he loses in virtue of the scientific facts. Instead, the scientific facts help to determine the plausibility of his metaphysics in the first place. Therefore, Tim cannot reach his preferred final score. The same holds for Enoch, and for the same reasons.

One last comment before we recapitulate and move on. Enoch’s readiness to distribute a high number of plausibility points to his metaphysical picture before having taken into account the external perspective is a perfect example of what I take to be methodologically problematic about many nonnaturalist views. This readiness, I suspect, results from a mindset that already devalues the external perspective’s bearing on metaethical theorizing. For, without such a devaluation, how could we confidently assign a high number of plausibility points to our nonnaturalist metaphysical picture before having even looked at the external evidence? We could only do so, it seems, if we already presupposed that, whatever the external perspective may have to offer, it would be relatively unimportant. I suspect that this presupposition underlies many nonnaturalist approaches. It is a bias that manifests on the methodological level; it manifests in how (some) nonnaturalists approach metaethical theorizing.²⁸

Let us recapitulate. Our methodological considerations, if correct, establish the following: When trying to explain how ethical deliberation and what it is distinctively about fits into reality, we should take into account and try to reconcile the external and the internal data. The argument from ethical phenomenology violates this methodological guideline by drawing metaphysical conclusions on solely phenomenological grounds. Therefore, the argument fails.

What options are nonnaturalists left with? Well, they could give up the argument from ethical phenomenology. But let us not go there (yet). Alternatively, they could feel inclined to dig in their heels and respond: “The external perspective is simply irrelevant for the context of ethics because the fundamental ethical entities are nonnatural.” If true, this response might exculpate the argument from ethical phenomenology. Unfortunately, however, responding in this way is not a real option because it obviously begs the question against naturalism. Metaethical arguments should establish the metaphysical status of ethical entities, not presuppose it.

So, only one option remains for nonnaturalists who want to hold on to the argument from ethical phenomenology. They need an independent argument for the irrelevance of the external perspective. If they were to establish, somehow, that the external perspective couldn’t contribute anything useful regarding the nature of ethical deliberation (and the nature of the involved entities), construing a moral metaphysics on solely phenomenological grounds might turn out legitimate after all. With such an independent argument, nonnaturalists could meet the Challenge from Lost Perspective.

3 The Intuitive Otherness of Ethics

Our previous discussion has shown that if nonnaturalists want to hold on to the argument from ethical phenomenology, they have to independently establish the irrelevance of the external perspective in metaethical theorizing. Their task is, in other words, to establish the “otherness” of ethics. How to do that?

One particularly influential consideration in favor of the otherness of ethics is the so-called just too different intuition.

JTD is wide-spread across the nonnaturalist literature.²⁹ Due to this prevalence, it is worth taking a closer look at two exemplary “applications.”

Start with Enoch. When he develops his argument from deliberative indispensability, he claims—in what I take to be the quintessential paragraph of his book—that the normative truths we are committed to qua deliberators must be nonnatural.

Because only normative truths can answer the normative questions I ask myself in deliberation, nothing less than a normative truth suffices for deliberation. And because the kind of normative facts that are indispensable for deliberation are just so different from naturalist, not-obviously-normative facts and truths, the chances of a naturalist reduction seem rather grim. […] The gap between the normative and the natural, considered from the point of view of a deliberating agent, seems unbridgeable. [Enoch (2011), 80, my emphasis]³⁰

Enoch’s point is straightforward: From the first-person perspective of deliberating agents, the normative truths we are looking for seem so different from natural truths that they couldn’t possibly be natural. Thus, we get the otherness of ethics.

The second exemplary application of JTD is Derek Parfit’s normativity objection against normative naturalism.³¹ To get his objection started, Parfit compares the following two statements:

Parfit observes that appeals to normative facts like (B) strike us to be very different from appeals to natural facts like (C). In his own words: “Given the difference between the meanings of claims like (B) and (C), such claims could not, I believe, state the same fact.”³²

Again, the argument is straightforward: Since appeals to normative facts seem so different from appeals to natural facts, normative facts couldn’t be natural. Thus, we get the otherness of ethics.³³

Now, does this work? Could JTD-based arguments be used as independent arguments for the irrelevance of the external perspective in metaethical theorizing? I don’t think so for the following two reasons: Firstly, Enoch’s and Parfit’s considerations are themselves instances of the argument from ethical phenomenology. According to both authors, phenomenology reveals that ethical facts are very different from natural ones; JTD is a phenomenological datum, after all. Thus, using the intuition to establish the (metaphysical) otherness of ethical entities is just another instance of the argument from ethical phenomenology. Appeals to JTD are not independent. They merely move the bump in the rug.

Secondly, relying on JTD in order to establish the otherness of ethics violates our second methodological guideline (above). Recall: When investigating any mental process, we should expect that the internal data will look very different from the external data. I am inclined to speculate that this is due to the nature of human consciousness (whatever it is). We inhabit a subjective perspective from which experiences come with a “something it is like.” They come with a, well, phenomenology. So, it is not surprising at all that these experiences, as had “from within,” are described very differently from the “external story” about what is going on when we’re having them. This suggests the following: For any property \(P\) that presents itself as part of your phenomenology, the differences between, on the one hand, your phenomenological impression of the nature of \(P\) and, on the other hand, the best external story about the nature of \(P\), provide no reason whatsoever to think that \(P\) is a nonnatural property. We find the same “unbridgeable gap” in the case of water and H\(_{2}\)O (see above). For these two reasons, JTD cannot help nonnaturalists to meet the Challenge from Lost Perspective.³⁴

We are back at square one. We haven’t met the Challenge from Lost Perspective yet; we haven’t established the otherness of ethics. And without the otherness of ethics, the argument from ethical phenomenology does not even get off the ground. Now, there are probably more ways to try to meet the Challenge from Lost Perspective. Nonnaturalists will have more to offer than appeals to JTD. But we won’t turn to these alternative attempts here. Instead, let me point out an interesting big-picture conclusion that follows from our discussion.

It has become clear that there are two general strategies for nonnaturalists. Either they (1) solely rely on the phenomenological perspective, or (2) they take into account and try to reconcile both perspectives. The first strategy falls prey to the Challenge from Lost Perspective. Disregarding the external perspective in one’s (metaphysical) interpretations of ethical deliberation is methodologically unreasonable. Moreover, any purely phenomenology-based attempt to warrant the exclusion of external evidence just moves the bump in the rug. So, here is the big-picture conclusion: If nonnaturalists want to go with the first strategy, they first have to justify the legitimacy of this strategy—but this can only be done by taking the second strategy. Thus, nonnaturalists must move beyond a purely phenomenology-based strategy in any case. They must, on pain of methodological unreasonableness, embrace the external perspective.

However, embracing the external perspective constitutes something close to a paradigm shift for nonnaturalists. As far as I am aware, the most prominent positive arguments for nonnaturalism are versions of the argument from ethical phenomenology. They all maintain, in one way or another, that some part of ethical phenomenology is best explained by the existence of nonnatural ethical entities. This raises what I take to be the million-dollar question for nonnaturalists: Is there a way to legitimize the argument from ethical phenomenology that takes into account both perspectives?

Let me say this much here: I believe there is good reason why nonnaturalists traditionally fend off the relevance of the external perspective in metaethics. If this dam broke, an entire ocean of external, empirical evidence concerning, say, the evolutionary function of deliberation or the origins of ethical intuitions would suddenly have to be weighed in. All of this poses an obvious threat to the nonnaturalist project: It may seem rather unlikely that the existence of nonnatural entities will turn out to remain a better explanation of ethical phenomenology than some externally and internally informed account devoid of such entities.³⁵ This partly explains, I think, the typical nonnaturalist reluctance to acknowledge the external perspective as relevant for metaethical theorizing. But if our considerations are correct, nonnaturalists do not have much choice; they must overcome this reluctance.

4 Conclusion

Nonnaturalists believe that ethical thought and talk involve (robust or not-so-robust) nonnatural ethical entities. In this paper, we have focused on the most prevalent positive argument for this view, the argument from ethical phenomenology. According to it, the claim that some ethical entities are nonnatural is part of the best explanation of why ethical phenomenology is the way it is. Our main conclusion is that the argument is methodologically unreasonable.

We started by stating the goals of metaethical investigations. These investigations try to explain how ethical deliberation—and what, if anything, it is distinctively about—fits into reality. We then argued, quite generally, that investigations of mental processes should take into account and try to reconcile both the internal (phenomenological) and the external (broadly: scientific) perspectives. This, we claimed, is where the argument from ethical phenomenology fails: It draws metaphysical conclusions that are solely based on internal, phenomenological observations. The argument, in other words, blinds out the external perspective. Hence our main challenge:

In order to meet this challenge, we said, nonnaturalists must provide an independent argument for the irrelevance of the external perspective. We discussed one strategy to this effect that involves the just too different intuition. We rejected this strategy for two reasons. The (maybe) more important one was that the just too different intuition cannot provide us with an independent argument for the irrelevance of the external perspective because any argument based on it would just be another instance of the argument from ethical phenomenology.

Our big-picture conclusion was that nonnaturalists must move away from a purely phenomenology-based strategy. Such strategies are methodologically unreasonable because they do not take into account the external perspective; they are unreasonable, that is, unless we already knew that the external perspective is irrelevant for metaethical theorizing. However, to establish that, nonnaturalists would have to, well, move beyond a purely phenomenology-based strategy. Otherwise, they would be arguing in circles, begging the question against those who believe that the external perspective is relevant for metaethical theorizing.³⁶

The big-picture conclusion is especially interesting once we acknowledge that most of nonnaturalism’s supportive considerations are entirely phenomenology-based.³⁷ What exactly this means for the prospects of nonnaturalism is a topic for another occasion. I do think, however, that the loss of the argument from ethical phenomenology leads to a significant decrease in plausibility points—at least as long as nonnaturalists do not defend their approach in a way that isn’t question-begging on the methodological level.

One final question: Could nonnaturalists reject the Challenge from Lost Perspective as illegitimate? I don’t think so. The challenge represents a hard-to-doubt methodological idea: When starting to investigate how any mental process—and what this mental process is distinctively about—fits into reality, we should be open to all kinds of evidence, external and internal. We should not prematurely, that is, without further argument,³⁸ blind out or devaluate a whole perspective on the mental process we are interested in—especially so if this perspective has proven highly resourceful in the context of other mental processes. Ultimately, the best account of the nature of ethical deliberation will be one that hasn’t lost perspective.

References