Against "Against Subjectivism" (Part 4)

Photo by 9am herday on Unsplash

This is part of an ongoing series. Here is the full series:

Part 1

Part 2

Part 3

7.0 Analytic Subjectivism

Recall that Christian stated that:

Wanting to do something counts in favor of doing that thing! Duh! This is the fundamental intuition behind the Subjectivist tradition dating back most famously to Hume.

I am not on board with this and consider such a move indefensible. I do not think that wanting things “counts in favor” of doing those things. I think this notion of “counting in favor” is philosophical gobbledygook that doesn’t mean anything and reintroduces core features of what I find unacceptable about realism into subjectivist accounts as well. This is why I favor quietism, because it is not Humean in the sense this passage alludes to, which retains objectionable reification. Now, the reason I say this is because Christian attributes features to subjectivist views that a subjectivist doesn’t have to endorse. The various forms of subjectivism he deals with ultimately inherit many of their features from the analytic tradition in which Christian operates and in which the whole realist/antirealist dialectic takes place. Many of the problems with this dialectic are symptomatic of all the standard positions, and when antirealist views are saddled with some of these features, they may turn out to have problems that arise from the poor synergy between these features and their other characteristics. One way to rescue subjectivism, then, is to shrug off these features, e.g., notions of desires “giving” reasons, and so on.

While I am happy to defend subjectivism, I will not defend any accounts of subjectivism that stick their necks out and embrace the faulty notions of analytic philosophers. When Christian turns his attention to analytic subjectivism, then, I will largely abandon it to the wolves. But all is not lost! As I will argue, something akin to analytic subjectivism can readily emerge from the analytic ashes and do just fine. What is this form of subjectivism? I will call it quietist subjectivism. Quietist subjectivism rejects the notion that wanting things “counts in favor” of doing them or “gives” us “reasons” to do them. Instead, one simply holds that we are motivated in certain ways and act on those motivations, and…that’s it. Such an account jettisons normativity as anything distinct from descriptivity. Subjectivists who construe normativity in such a way that they think that something gives, provides, or generates “reasons” that “count in favor” of what we “should” do make a serious mistake: they treat normative talk as something conceptually and/or metaphysically separate and distinct from descriptive facts, as something one needs an account of, as a problem to be solved, then present desires as the solution.

But normative talk is just that: a way of talking. It isn’t a distinct property or phenomenon in need of accounting for in some metaphysically distinct or even conceptually rich sense. All we have to do is account for normative language, and any question of deep metaphysical issues or robust conceptual distinctions dissolves. Analytic subjectivism almost does this, but the kinds of subjectivism Christian addresses don’t adequately finesse the distinction between normative discourse in an abstract, theoretical, or philosophical context from its use in richer, practical, real-world contexts. With a few modifications, it would do just fine. So the position I defend is in some ways similar to analytic subjectivism, which is the very first form of subjectivism Christian addresses and arguably the one many would consider the least defensible.

So let’s address why that is, and how with a few modifications it will work just fine, before addressing the cluster of latter-day forms of subjectivism that follow in its wake. Christian says that analytic subjectivists:

like to draw a straight line between should, reasons, and desires.

Christian draws on Parfit’s characterization of desires:

any state of being motivated, or of wanting something to happen and being to some degree disposed to make it happen, if we can. (As quoted in Parfit, 2011, p. 43)

This account of desires seems fine to me. But Christian charges proponents of this view with saying something vacuous. Here’s the problem. According to Christian, analytic subjectivists hold that we should interpret:

we have most reason to act in some way

as

this act would best fulfill our present fully informed terminal desires

In addition, analytic subjectivists hold that:

we have most reason to do what would best fulfil our present fully informed terminal desires

The problem is that once we combine these claims, the position amounts to holding that:

what would best fulfill our present fulfill [sic] informed terminal desires is what would best fulfil our present fully informed desires.

…So the position ends up as a tautology. Insofar as this is an accurate representation of an actual view, I don’t endorse it. Why? Because this account already bakes in some of the features Christian introduces later on, features which presuppose certain concerns analytic philosophers have that presuppose reason-talk. In this case, the main problem is the inclusion of “fully informed.” The purpose of this feature of the account is to distinguish between cases in which we have false information about what would serve our interests from cases in which we don’t. Suppose you wanted to drink a glass of clear liquid you mistakenly believed was water, when in fact it was poison. If you were fully informed, you wouldn’t want to drink it. This feature is supposed to account for the fact that you don’t “have a re ason” to drink the poison water. But this is only included as a result of confused reactions to thought experiments that fail to distinguish terminal and instrumental desires. Our “intuition” that we don’t “have a reason” to drink poison merely because we think it is water (and we want to drink water to quench our thirst) is better understood as recognizing that a particular desire (to quench our thirst) wouldn’t be satisfied in a scenario in which we stipulate a course of action wouldn’t satisfy the desire. We’ll encounter this below when we get to the relevant scenario. But the bigger problem with Christian’s characterization of the analytic subjectivist account is that the analytic subjectivist purportedly offers us an account of what it means to say we “have most reason” to do something, then once this is answered, repeats the question with the analysis excluded from the phrasing “most reason.” If the analytic subjectivist goes along with this, they have been complicit in walking into a linguistic trap they never needed to walk into. Note how Christian says that the analytic subjectivist analyzes:

we have most reason to act in some way

as

this act would best fulfill our present fully informed terminal desires

then they allegedly turn around and say that the answer to what we have most reason to do is to:

best fulfill our present fully informed terminal desires

This is bizarre. Why would someone think that the answer to the question “What do you have most reason to do?” is to say that you have most reason to best fulfill our present fully informed terminal desires, and to interpret the question as asking them what would best fulfill their present fully informed terminal desires, but somehow be so confused as to fail to distinguish the question itself as the culprit behind the terminology, and instead willingly allow one’s position to become some kind of tortured verbal abomination like this:

What would best fulfill our present fulfill informed [sic] terminal desires is what would best fulfil our present fully informed desires

As if one were separately answering the question in such a way that the notion of what one had most reason to do were distinct from the answer, while simultaneously analyzing the question itself in terms of the analysis predicated on one’s answer? This is as silly as doing the following:

Alex: What is your favorite fruit?

Sam: As an analytic bananaist, I interpret the phrase “favorite fruit” as “banana.” Therefore, I interpret your question to be asking me “What is your banana?” So my answer is that my banana is my banana.

Alex: …what?

Alex is right. What? What a bizarre response. The problem with the analytic bananaist’s response is that the question is asking them which account they favor. It is not asking them to offer an analysis of the question of which account they favor in terms of which account they favor, such that to ask them what their “favorite fruit” is is to ask them to offer an analysis of “favorite fruit” in such a way that this is taken to be identified with their answer, then reanalyzed in terms of their answer, such that the question becomes a weird and pointless inquiry that already presupposes their answer. Yet this is how the analytic subjectivist is depicted by Christian. The analytic subjectivist, for some reason, interprets the phrase “most reason to do” in terms of their account, then simultaneously treats “most reason to do” as distinct from the account, ratcheting in the account after this, such that:

Most reason to do = What would best fulfill our present fully informed terminal desires

…and, at the same time, still treating what one has most reason to do as something one can offer an answer to distinct from the phrasing itself, as though there’s still a verbal slot open for the answer to the floating questions “What do you have most reason to do?” distinct from the analysis of the meaning of the question itself in terms of the answer. Let’s see how that works in dialog form.

Alex: What do you have most reason to do?

Sam: As an analytic subjectivist, I interpret the phrase “most reason to do” as “what would best fulfill our present fully informed terminal desires”. Therefore, I interpret your question to be asking me “What would best fulfill your present fully informed terminal desires?” So my answer is that “What would best fulfill my present fully informed terminal desires is what would best fulfill my present fully informed terminal desires.”

Alex: …what?

This is really bizarre. But it’s an understandable error that occurs when one attempts to offer an analysis in certain abstract, theoretical discussions among philosophers rather than simply talking about how you’d speak and what you’d mean by what you say in everyday contexts. In other words, Christian’s characterization of analytic subjectivism saddles them with offering an “account” that ends up an empty analysis of how to interpret a set of sentences outside any relevant contexts of usage. Whose discourse is it even an analysis of? I don’t endorse the view that this is what nonphilosophers mean when they make moral claims. And I don’t endorse the view that sentences have meanings independent of usage. And this isn’t what I mean. I see no reason why I should endorse this account. But I also don’t see why someone couldn’t reject this view, reject the reasons-talk common among philosophers, and the weird abstract contexts in which these analyses are being offered, and still be some sort of subjectivist. Yet Christian proposes that:

[…] to avoid circularity, Subjectivists should use the word “should” in the reason-implying sense, essentially, to mean that we should do what we have most reason to do.

Subjectivists don’t have to do this. They can simply recognize that what one should do in abstract terms is to act on their desires while also appreciating that in any concrete instance this would be realized in concrete terms by the specific desire in question. Suppose the gastronomic analytic subjectivist goes to dinner. Technically they believe that what one ought to choose for dinner is whatever would most satisfy their desires. After they look over the menu, and the waiter asks them what they’d like, do you suppose they’d respond:

“I’d like that which I most desire”?

I don’t think they would. I think they’d pick something in particular. Nothing about being committed to choosing the dish you most desire saddles you with an actual vacuous or problematic tautology in practice. What you would choose at the restaurant in the abstract is whatever you most desire. What you’d choose in concrete instances would be the specific dish. The appearance of a tautology only remains as long as you insist on framing the position in abstract terms.

Christian’s suggestion, on the other hand, may avoid meaningless and nonproblematic circularity (which needn’t be avoided in the first place), but at the cost of reintroducing a reified notion of “reasons” into our account that saddles the subjectivist account with one of the features I reject in realist accounts: an invocation of ineffable (and, I suspect, meaningless) notions of “reason.” Any form of subjectivism that imports such notions inherits precisely one of the central features of moral realism that make it so objectionable in the first place. In other words, if we go back to the original set of claims the analytic subjectivist allegedly endorses, I just wouldn’t endorse one-to-one correspondence between the phrases Christian provides and various “analyses” of those claims. I’d instead just say that there are facts about our desires and when I talk in certain, relevant contexts, what I’m doing is reporting what those desires are. This creates absolutely no problems whatsoever. I’ll say more about this in a bit. For now, Christian continues:

So if we are to make substantive claims, we need to use “most reason”, “should”, and “ought” in this reason-implying sense.

This is not the case. First, as already noted, we don’t need to do this since there was never a problem it’s needed to solve in the first place. Second, I deny reasons-implying claims are substantive, since I don’t think the notion of reasons at play in these accounts (some kind of normative reason) is substantive. As Christian noted, this concept of a reason is ineffable. I don’t think there are substantive but ineffable concepts, because I think that the reason why the terms can’t be understood is that there isn’t anything to understand. Reason-talk among philosophers is just a reification of some quirks in the English language, mistakenly perceived as some kind of independent property, when no such thing exists. It makes about as much sense to think there are reasons as that there are althoughs or howevers.

Third, rejecting analytic subjectivism doesn’t require you to endorse a specific alternative like this, for much the same reason that rejecting Christianity doesn’t require that you become an atheist. Our only options aren’t analytic subjectivism or a belief in allegedly substantive normative reasons. We can reject all of these positions and still make substantive claims. To be fair, I don’t think Christian is likely to have intended a genuine dichotomy (after all, someone could abandon all of these accounts and go be a noncognitivist or something); I just don’t think subjectivists, if they want to avoid tautology, have to embrace ineffable concepts.

The problem with analytic subjectivism as described here is that it tacitly concedes that one must offer some account of reasons, identifies these reasons with desires, then holds that one has reason to do what one desires. In effect, it reifies reasons, treats them as something distinct from desires, then offers an analysis that re-identifies them with desires, thereby treating them simultaneously both as distinct and not distinct, which is why once you unpack it you get this weird tautology. To summarize very quickly why this is silly, the view amounts to something like this:

1. My account of what it means to say I have most reason to do something is that it would fulfill my desires.

2. I think we have most reason to fulfill our desires.

3. So, my position is that it would fulfill my desires to fulfill my desires.

The mistake here is in step two. If you already think talk of what you have most reason to do reduces to what you desire, why would you say you have most reason to fulfill your desires? Recall: if you think “favorite fruit” just means “banana,” why would you say “My favorite fruit is banana”? This would be like saying “Banana is banana,” which is ridiculous. In most contexts, you wouldn’t say these sorts of things, but as we’ll see, this concealed tautology is perfectly respectable in certain contexts. Because of this, in a way, the tautologous aspect of analytic subjectivism isn’t the problem. Ironically, Christian thinks the only way for the analytic subjectivist to avoid the problem of one’s account revealing itself to be a tautology is to embrace a nonreductive, substantive notion of reasons distinct from desires. I argue that the opposite is the case: that the subjectivist has already conceded too much to Christian in granting that there is any meaningful conception of ineffable normative reasons apart from desires in the first place, such that they need to offer an “account” of them distinct from the reduction of reasons to desires.

For the analytic subjectivist, if you were asked why you have most reason to fulfill your desires, the answer would be “because it would fulfill my desires,” which is weird and vacuous and makes your position sound silly. But what an analytic subjectivist could do instead is insist that facts about what one has most reason to do just are facts about what would or wouldn’t fulfill one’s desires. And then asking someone: “What do you have most reason to do?” just becomes the question: “What would fulfill your desires?” At that point, it would be a little weird to say “I would fulfill my desires.” Is this vacuous? It sure seems like it. But that’s only because the question is being posed outside any specific context in which someone is asking which desires would be fulfilled by which actions. Instead, the question is posed in a context in which the analytic reduction of “most reason” to desire fulfillment hasn’t yet been implemented, such that the question leaves open what one might have “most reason” to do in a way that isn’t necessarily reducible to facts about desire fulfillment.

For comparison, if someone asks what your “favorite fruit” is, they are not asking you to offer a definition of “favorite fruit” that reduces it to whatever your favorite fruit is, such that to ask what someone’s favorite fruit is is to ask them a vacuous question from the outset: “What is your [whatever their answer is]?” which will yield the response “[their answer.]” This would be an insane question to ask. Rather, the person posing the question is asking something like

Of the possible fruits one might prefer, of which there are more than one, which one do you prefer?

To answer:

Banana.

…already tacitly acknowledges that something other than “banana” is a viable option. But if you’re an analytic subjectivist, do you interpret “most reason” in this open-ended way, such that one could in principle have “most reason” to do something other than what would fulfill their desires, or not?

There is no straightforward way to answer this question. Why? Because there are two separate answers: an unconditional and a conditional.

The unconditional answer should be “yes”: not conditional on my analysis of “most reason,” one could slot some other answer into a response to the question of what one has most reason to do other than what would fulfill one’s desires.

The conditional answer would be “no”: conditional on “most reason” being analyzed in accord with the analytic subjectivist’s response, the question itself asks “What would fulfill your desires?” to which the analytic subjectivist responds “what would fill my desires.”

This distinction reveals something extremely important:

The tautology only emerges conditional on the analytic subjectivist’s account being accepted (at least conditionally) in the first place.

Why is that important? Because the “tautology” emerges only on the conditional victory of the analytic subjectivist. If they’ve already won, if we’ve already granted that “most reason” = “what would fulfill your desires,” they can close shop there and do a victory dance. The problem would be with the person posing the question. There’s nothing wrong with responding to a vacuous question with a vacuous answer. Ask empty questions, get empty answers.

There’s no other analysis of “most reason” on offer if we accept the conditional response, so there are facts about what would fulfill our desires and it wouldn’t make any sense to ask if one has “most reason” to do something else instead, because “most reason” just means what would fulfill one’s desires.

The reason the analytic subjectivist’s response appears to be a problematic tautology is a result of failing to distinguish these unconditional and conditional responses. The problem is that the question is structurally framed in a way that presents itself as an unconditional response. As a result, if one offers the conditional response, such that one reanalyzes the question in terms of the analysis, the question itself participates in setting up the tautology:

Question: What would fulfill your desires?

Answer: What would fulfill my desires.

Structurally, the analytic subjectivist is expected to simultaneously treat the question of what they have “most reason” to do in unconditional terms: as not reduced to “what would fulfill my desires” and at the same time treat it as conditional, such that it does mean this. Now when we go back to Christian’s characterization, we can see how this has occurred.

Here’s the analytic subjectivist’s conditional response:

These subjectivists use words like ‘reason’ and ‘should’ in the internal senses. So they interpret the claim

we have most reason to act in some way

as

this act would best fulfill our present fully informed terminal desires

Then the analytic subjectivist is expected to offer an unconditional response in which the notion of “most reason” is held open to alternative possible accounts or reductions. Only Christian doesn’t pose the question that generates the problem. Christian just says:

we have most reason to do what would best fulfil our present fully informed terminal desires

But if the analytic subjectivist already thinks “most reason” just means “what would best fulfil our present fully informed terminal desires” why would they then say that they have “most reason” to do what would best fulfil our present fully informed terminal desires? Conditional on their analysis, they wouldn’t; this would be redundant and an empty tautology. It’s only when “most reason” is taken to be unconditional and thus ex hypothesi not reduced to the analytic subjectivist’s analysis that they’d make the claim that

we have most reason to do what would best fulfil our present fully informed terminal desires

…This is exactly like the banana example. If you’re asking me what my favorite fruit is in an unconditional sense, then it makes perfect sense to say:

My favorite fruit is banana.

If you’re asking in a conditional sense, where we reinterpret “favorite fruit” in line with my answer, now you are asking me something a bit awkward like

What is your banana?

Now, if I am required to verbally reiterate your distinction between “favorite fruit” and “banana” in my response, such that I am required, again, by the very setup you yourself are imposing on me, to slot “banana” in for “favorite fruit,” now I can’t just say:

Banana.

I have to say:

My favorite fruit banana is my banana.

The lesson we should extract from this is that analytic philosophers completely trip all over themselves creating pseudoproblems that result from a failure to be attentive to language, mostly to disambiguate and distinguish what one means in various contexts of usage. They conflate different contexts, smashing them together into a single question or single phrase or analysis or account, don’t keep the various contexts in which one might say one or another thing distinct, demand responses to questions that look awkward outside any real-world or practical contexts of usage, then declare victory when the result, which is often tortured, vacuous, and redundant verbiage looks weird or vacuous or empty or whatever. But such weird, vacuous language only emerges if one twists English into enough pretzels that one more or less compels a rival response to sound weird in the contrived, abstract contexts one has created. At least one problem is that these contexts often have little or no bearing on what a proponent of a position would say in any practical, concrete context. They don’t raise any actual practical problems with a position. And the theoretical problems are not genuine problems, either. They are pseudoproblems that only emerge when we aren’t careful with language.

To put it bluntly: objections to analytic subjectivism largely involve abusing language to generate the misleading impression of a problem. The analytic subjectivist’s position should be rejected insofar as the analytic subjectivist themselves buys into the practices of analytic philosophy that generate the problem. If analytic subjectivism is (perhaps a bit ironically) defined in terms of its commitment to these practices, then I reject it. If it isn’t, then analytic subjectivism in some form can be rescued.

But I’d prefer to just grant that the position is committed to its own internally self-negating confusions, and suggest quietist subjectivism as an alternative. Whether one wants to call this a form of analytic subjectivism is up to them. Words, after all, are subordinate to us, though people who do analytic philosophy too much seem to have it the other way around. In any case, quietist subjectivism dispenses with the verbal confusions characteristic of so much of analytic philosophy and stops Christian’s critique cold. Subsequent analyses of modifications to subjectivism are unnecessary, since we never give the initial ground that requires is to buy into the notion of construing normative considerations in a “reason-implying” sense, which I take to be a concession that there are ineffable reasons and any successful account would need to explain what we have “reasons” to do.

Have I rescued analytic subjectivism? Not necessarily. It depends on how much the analytic subjectivist wants to participate in the language games analytic philosophers engage in. I’ve maintained that there’s nothing inherently wrong with asserting a tautology. Carrots are carrots. As far as I can tell, that’s true. And I am not going to get struck by lightning for saying it. There’s nothing vacuous or obviously empty about saying that claims about what one has most reason to do will be understood as claims about what would promote one’s desires.

The real problem here isn’t the tautology in such a position, but that it accepts as meaningful to even say that one “most reason to fulfill our desires” as if there were some conception of should that didn’t reduce to fulfilling one’s desires while simultaneously maintaining that there isn’t. The whole way “I have most reason to fulfill my desires” is framed pragmatically implies that something other than desire fulfillment could be ratcheted into the slot for “should,” but, conditional on analytic subjectivism, it couldn’t. So really we should see the statement as an innocuous tautology conditional on analytic subjectivism, and there really shouldn’t be a problem with it consisting in a tautology. But it sounds weird, and if that’s reason enough to reject the view, we can finesse what a subjectivist is up to without paying tacit lip service to the legitimacy of “most reason” as a nonreductive notion while simultaneously claiming to have reduced it in the way we ourselves talk.

7.1 A detour about zapples

There’s some serious linguistic strangeness going on with the preceding critique of analytic subjectivism, and I really want to drive home what the issue is. So consider this section another attempt at approaching the topic. It may be slightly redundant, but this is a blog and part of what I am doing is thinking things through as I write. If you think you’ve understood well enough what I take the issue to be, you can skip ahead to the next section.

Suppose a new product comes along: the zapple. You examine it thoroughly and determine that zapples are just apples. Not a specific kind of apple or anything like that. They’re just plain ol’ apples. Now someone asks you:

What is a zapple?

And you say:

A zapple is an apple.

Would it be much of a critique of your view to say:

Oh, so if you say “This zapple is an apple” you mean to say “This apple is an apple”? That’s a tautology. Why would you go around saying something vacuous like that?

Do you see the problem? It isn’t the zapple-apple reductionist who proposed that zapples were something distinct from apples. And if they were to be in charge of how we spoke about zapples, they might suggest we drop all talk of zapples and just call them apples. So they’d never need to say “zapples are apples.” They’d just talk about apples. But it was other people who insisted on calling something a zapple rather than an apple. All the zapple-apple reductionist has done is close this verbal loop. Yet if they were engaged with zapple nonreductionists, the nonreductionists may insist on verbally distinguishing zapples from apples. In response, the zapple-apple reductionist may, in exasperation, insist that zapples just are apples, only to be met with:

But that’s a tautology! “Zapples are apples” just means “Apples are apples” on your view.

Indeed it does. But this isn’t vacuous or meaningless or a problem in the context of arguing with zapple-apple nonreductionists. In exactly the same way, a person who reduces talk of what one should do to talk of what would promote one’s desires can reduce the former to the latter in a way that would yield tautologies in an entirely innocent way: innocent in exactly the same way that someone who concluded that zapples were apples wouldn’t be saying anything vacuous or confused or mistaken in the context of a dialectic in which others held that zapples aren’t or at least might not be apples.

The “tautology” plays a dialectical role: it closes a verbal loop, and yields a tautology that effectively dissolves an undischarged demand for an account of zapples. Determining that there is no distinct thing as a zapple may simply be the correct position. Just so, judging that there is no distinct content to or conception of should-talk apart from it referring to one’s desires can settle any dispute about whether one should do something other than what one desires. One can even allow that a kind of open question could be meaningfully posed, but note that this would simply involve someone using the same terms to mean something other than what the analytic subjectivist themselves means: it’d be a tautology for the analytic subjectivist to say “I should do what I desire” but not for others. There’s just no problem with this.

But if an analytic subjectivist agrees with everyone else that reasons are something distinct from desires, while simultaneously trying to reduce reasons to desires, then of course they’re going to have a problem. I think one could perhaps still call themselves a subjectivist while denying this or any other problematic features Christian associates with the view, but let’s just grant that subjectivists hold some weird, objectionable view that buys into some of the assumptions of non-subjectivists but can’t successfully reconcile these with their account of reasons as desires. Fine. Then we can reject the view. But a suspiciously similar-looking phoenix can readily rise from the ashes, and simply abandon those dubious features it couldn’t reconcile.

A something-like-a-subjectivist like myself doesn’t have to grant that there are reasons, that reasons are desires, or that reasons “give” us reasons. My account simply rejects any notion of normative reasons altogether, and thus denies that I need to offer an “analysis” of “should” that treats it as distinct from facts about what would fulfill my desires in the first place. What I think is simple:

1. There are facts about our desires.

2. There are facts about what is conducive to achieving those desires.

3. There are no further normative facts at all. It isn’t the case that, e.g., we have “most reason” to act to fulfill our desires or anything of the sort.

This presents an alternative to analytic subjectivism which I’ll call quietist subjectivism or if that isn’t descriptive enough, reasons-eliminativist subjectivism. On this account, reasons-talk isn’t ineffable, reasons-talk and should-talk just is (at least for the quietist subjectivist) talk of desires. Insofar as this allegedly saddles the subjectivist with empty tautologies: it does not. The tautologies only emerge in contexts in which the subjectivist is clarifying their position on the meaning of should/reason talk. There is nothing vacuous about a non-subjectivist asking, “How do you use reasons-talk and should-talk?” and a subjectivist responding “to refer to desires.” If the non-subjectivist then says, “Wait, so if I asked what you should do, you’d say ‘fulfill my desires’? but all this would mean is that it would fulfill your desires to fulfill your desires! Why would you present me with a tautology?”

The subjectivist can point out that in the context of the question the meaning of “should” is being treated as theoretically open to non-subjectivist analyses that don’t reduce should-talk to desire talk. Since the analytic subjectivist does reduce should-talk to desire-talk, when they respond that one “should do what one desires,” all they are doing is collapsing the putative autonomy of the two concepts. This results in a tautology, but it isn’t a vacuous one: it is one that, in a dialectical context in which the subjectivist is distinguishing their position from non-subjectivist positions, illustrates the collapse, or reduction. This does not carry over outside this context into the subjectivist going around asserting tautologies like an idiot in any real-world contexts. In those contexts, people are almost always asked a context-specific practical should question. In other words, we can distinguish two types of “What should you do?” questions:

• Abstract theoretical
  What should you do?

• Contextualized practical
  What should you do?

Note how both look identical. But that is because sentences mean different things in different contexts. Let’s unpack what these questions involve in their proper contexts:

• Abstract theoretical

  What should you do? = What is your account of normativity? That is, when it comes to the question of how you deliberate about what choice to make and proceed to take action, what theory captures the considerations relevant to doing so?

• Contextualized practical

  (For this, I have to provide a hypothetical context. Here it is. You are deciding whether to eat leftovers or go out to have a nice dinner. What should you do?)

Note how in the second case, responding to “What should you do?” with “stay home and eat leftovers” or “go out to dinner” makes perfect sense: if you want to save money and eat healthy, stay home. If you want to splurge and enjoy yourself more, then go out to have a nice dinner. There’s no tautology here. If you responded with “what I desire,” that might seem vacuous, but that’s only because you aren’t clarifying what it is you desire. Think of it this way. If the question “what should you do?” in this situation amounted to “what do you desire?” in this situation, you could respond with “whatever I want.” This might be a glib and unhelpful response, and one might even say it’s vacuous, but so what? You’re not theorizing here. In a practical context, this couldn’t generate any serious philosophical problems for you. It is only in theoretical contexts where the nonspecificity is both mandated, because there is no specific context, and seemingly vacuous. But that seeming vacuity is a byproduct of philosophers failing to distinguish practical and abstract contexts, such that the sense of vacuity bleeds into how they’d interpret the remark in practical contexts. If the response to literally every question: “What should you do?” in every practical context amounted to “What I desire,” which was the equivalent of saying “What I should do,” you really would have an empty theory. But this isn’t what a quietist subjectivist (or an analytic subjectivist, for that matter) is actually committed to. In any practical context, they can just report the desire:

What should you do?

Go out and have a nice dinner.

There is nothing remotely vacuous about this.

With respect to the first, abstract theoretical question, the quietist subjectivist can respond: My account reduces considerations of what I should do to considerations of my desires. When deliberating, I would ideally choose that course of action most consistent with my desires (and I’d try to do my best given the information available to me). Note that on this view, a reduction can result in a tautology without any issue at all. Recall that if you think zapples are apples, such that you can substitute “zapple” for “apple,” then you can translate the statement “Zapples are apples,” which doesn’t look like a tautology, even though “Apples are apples,” clearly is a tautology. But this is an innocent tautology because the reduction of zapples to apples occurs in contexts where the status of zapples is in dispute: others may think they aren’t apples. That you think they are as a matter of definition does result in “Zapples are apples” meaning exactly the same thing as “Apples are apples,” but that you’re stuck with this tautology isn’t your fault; it’s a byproduct of a dialectic in which others hold incorrect or unspecified views.

If in practical contexts the notion of “should” weren’t in dispute, there is no need to be theoretically open about one’s meaning, so the tautologies need not emerge to threaten the subjectivist’s account. In those contexts, responding to practical questions about what one should do with “what I desire” would instead be used to express either an unwillingness or inability to respond, whether one is being glib or is confused. You may have heard conversations like this:

Parent: Well, if you’re not going to stay here and do your chores, what are you going to do?

Child: Whatever I want!!!

It would be silly to interpret the child as asserting a tautology. Keep the theoretical and practical contexts distinct, and subjectivists just don’t face any serious threat or need to invoke ineffable notions of reasons to explain what they mean and to offer a substantive account.

Incidentally, I think Christian has accurately diagnosed the problem with the analytic subjectivist position he outlines:

The issue with Analytical Subjectivism is not using reason in the internal sense alone. It’s using reason in the internal sense paired with using “most reason” or “should” in the internal or desire-fulfillment sense.

Christian believes the proper solution is to abandon this circularity by using the term “should” in the “reason-implying” sense:

Instead, and to avoid circularity, Subjectivists should use the word “should” in the reason-implying sense, essentially, to mean that we should do what we have most reason to do. This makes reasons alone to be the “key battle ground in ethics.” So if we are to make substantive claims, we need to use “most reason”, “should”, and “ought” in this reason-implying sense.

As I have argued, subjectivists do not need to embrace a reason-implying sense of “should” to make substantive claims. And I have argued that this isn’t the only or best path for the subjectivist to take. Instead, the subjectivist can just reject the demand that, in theoretical contexts, they are required to offer an account of “most reason” or “should” talk in such a way that they are saddled with weird tautologies. They can instead argue that such talk, in these conversations, is the artificial, theoretical construction of analytic philosophers, that it doesn’t map onto practical contexts or ordinary language, and that an account that reduces “should”-talk and “reason”-talk to desires can offer tautological responses in some of these contexts without this rendering their account vacuous or in any other way philosophically suspect. This quietistic move demonstrates that the subjectivist can speak and reason in ways that make perfect sense and don’t saddle them with vacuity.

In short, the problem with the analytic subjectivist as Christian depicts them here is that they appear to be complicit in tacitly conceding:

1. That there is an open question about normative questions about what one ought to do (or what one has “most reason” to do, or “should” do, etc.)

2. That this question is, in fact, closed: what one ought/has most reason to/should do is what would most fulfill their desires

Christian has thus crafted a position with an internal tension, then criticized it for having that tension. He then presents only one solution, and proceeds from there, without considering the way of resolving the tension I’ve proposed here. The critique of the position only lands insofar as the analytic subjectivist themselves fails to recognize that their position exists in a kind of Necker-cubean superposition, failing to collapse the open question into a closed one. But if they do close it, the problem dissolves, and the tautology is only a tautology in a cross-theoretical context, i.e., between proponents of the question being closed by analysis and those who think it remains open. That is, it only appears to be an empty tautology if we are discussing matters as if the question were closed, when in fact it isn’t empty in any context in which the reductive analysis that closes the verbal loop between normative terms (should, ought, most reason) and descriptive terms (what would most fulfill desires) remains open. For comparison, if the question of what a zapple is remains open, it is not at all empty or vacuous to propose that zapples are apples. But once we agree that zapples are apples by definition, now asking “What is a zapple?” becomes a closed question. But it is only a closed question once the analysis is accepted. It is equivalent to asking “What is an apple?” to which “An apple” is an entirely appropriate response in that context. Note, though, that “An apple” may seem glib or dismissive precisely because, if the question were closed, asking someone what an apple was would probably not be taken to ask whether an apple was an apple, but to ask something further, such as to ask about its properties or category memberships (e.g., a red or green fruit known for being crisp, slightly sweet…). Careful attention to language dissolves any problems analytic subjectivism might face on grounds of asserting empty tautologies.

Since the subjectivist doesn’t need to embrace should-talk in the reason-implying sense, I could end my rebuttal here. I have provided a way for analytic subjectivists to embrace a bit of quietism and thereby embrace quietist subjectivism. This allows a form of the position more or less like the original to emerge unscathed. But let’s continue anyway, to address Christian’s other remarks.