Someone recently presented me with the following syllogism, which is intended to serve as an argument for moral realism:
P1. If moral realism is false then slavery is not wrong.
P2. Slavery is wrong.
C. Moral realism is true.
I opted to share the response I will share with them as a blog post, which will allow me to revisit the problem of normative entanglement. This notion is worth revisiting, because I believe it is one of the central pillars in popping up the appeal of moral realism through illegitimate distortions of ordinary language. At any rate, here is my response:
What worries me about arguments like these is that they involve normative entanglement. That is, they can give the misleading and typically false impression that if the antirealist denies any of the premises, that the antirealist somehow isn't opposed to the action in question, has bad moral character, is a bad person, and so on. However, none of this is entailed by rejecting moral realism. A moral antirealist can and does maintain the ability to consistently oppose terrible actions and to hold substantive normative moral stances in opposition to those actions, without presuming moral realism. A moral realist who disagrees isn't entitled to simply deny this as a given, since that is one of the very matters of contention in dispute between a realist and an antirealist. As such, there is no way to deny this that I wouldn't reject as begging the question.
Here's how that works. A moral realist can ask a moral antirealist:
“Do you think torturing babies for fun is morally wrong?”
The antirealist is expected to give a “yes” or a “no” response.
The problem with this question is that it does not specify what is meant by “wrong.” Suppose “wrong” is used in a purely normative way, and does presuppose any particular metaethical framework (including realism). If so, then the antirealist can simply respond, without error or inconsistency, “Yes, I think that it’s wrong.” Expressivists, relativists, fictionalists, and all manner of moral antirealists would not be speaking in a way inconsistent with their metaethical views to respond in this way.
However, if the question presupposes a realist conception of wrongness, then to be consistent, the antirealist would need to say “no.”
Unfortunately, when the question doesn’t specify whether the sense of “wrong” entailed by the question presupposes realism, the question is ambiguous between a reading consistent with antirealism and a reading that isn’t. This makes the question effectively impossible to answer in a clear way.
Suppose the realist clarifies that they are presupposing a realist reading of “wrong.” In that case, the question is a trivial one. Moral antirealism just is the view that there are no stance-independent moral facts, so asking an antirealist if something is stance-independently wrong is a waste of time. It serves no genuine dialectical or philosophical purpose.
Instead, I believe the de facto function of such questions is purely rhetorical. That is, it serves no other purpose than to make moral antirealists look bad by giving the misleading impression that the antirealist holds bad normative moral views, rather than revealing a problem with their metaethical views. To be clear, I am not suggesting that you are intentionally engaged in rhetoric or doing anything malicious. Rather, I am suggesting that the form of syllogism you’ve presented has the structure of a kind of rhetorical move that can never achieve any genuine philosophical role, but can only, at best, serve a rhetorical role.
Going back to the question above, the realist could modify the question to ask this:
“Do you think torturing babies for fun is stance-independently morally wrong?”
First, note how pointless this question is. Antirealism just is the view that nothing is stance-independently wrong, so why ask the question? The answer is obviously “no.” The problem is that the question bundles two claims together (a) a metaethical claim about whether anything is stance-independently wrong and (b) a normative claim about whether baby torture is “wrong” (in a normative sense). By bundling the questions together, the question effectively functions as a kind of complex question that forces one to answer two questions simultaneously. Yet the questions could be disambiguated. A person could in fact endorse any of the four following views:
There are stance-independent moral facts, and it is morally wrong to torture babies.
There are stance-independent moral facts, but it is not morally wrong to torture babies.
There are no stance-independent moral facts, but it is morally wrong to torture babies.
There are no stance-independent moral facts, and it is not morally wrong to torture babies.
When someone asks whether it’s stance-independently wrong to torture babies, “yes” entails response (1), while “no” is actually consistent with (2), (3), and (4).
An antirealist could (and probably would) endorse (3): that there are no stance-independent moral facts, but it is morally wrong to torture babies. This is my view. I think it is wrong to torture babies. I’m no less disgusted and outraged and opposed to it than moral realists. I just don’t believe anything is stance-independently wrong, because I think the notion of stance-independent wrongness is nonsensical.
The problem is that if a person is compelled to say “no,” you can’t tell whether they mean (2), (3), or (4) without further specification, which bundles them in at least two possible categories which actually don’t hold it to be wrong to torture babies. For comparison, you could ask someone the following question:
“Which category do you fall into? (a) A perfect person (b) an imperfect person or baby torturer.”
Nobody could plausibly endorse (a), so they’d have to pick (b). But this would be an absurd question to ask, because it could be misleadingly leveraged to imply they might just be a baby torturer. A question like this is unfair, because it funnels answers towards a misleading way of implying the other person might be a bad person.
Questions that bundle metaethical and normative questions together function the same way. They create a wedge between a view that, on philosophical grounds, the antirealist cannot consistently endorse, and every other view, which bundles the antirealist’s normative and metaethical views together, but lumps that view along with every other view where those metaethical and normative views come apart in different combinations, including combinations that involve repugnant and awful normative moral perspectives, e.g., that it’s not wrong to torture babies in *any* respect, not just stance-independently.
More generally, even aside from the misleading nature of the ambiguity built into the very structure of such remarks, there’s an inattention to the pragmatics of such a question. If one person asked another, on a stage, “Do you think it’s stance-independently wrong to torture babies?” and the other person responded with “No,” you can picture the audience gasping in shock and horror. Why? Because the audience may take the person to not only be rejecting the metaethical claim, but the normative claim as well. That is, this could simply be interpreted as the antirealist denying that torturing babies is wrong at all…even though this isn’t something antirealists are committed to.
I think similar concerns apply to your syllogism. The syllogism depends, for its persuasive force, on giving the misleading impression that the antirealist is an evil monster if they answer one way, or is forced to concede to the realist if the answer the other way. I see such syllogisms as a kind of linguistic trap that relies on ambiguity and leveraging misleading features of the pragmatic implications that figure into ordinary moral discourse to put moral antirealists in an illegitimate lose-lose situation where they have no way to answer consistently without looking bad as a result of false and misleading pragmatic implications about what their position entails.
You already know how the dialog tree of the realist works. You know some of the branches are rhetorically effective. So you should prepare counter measures and intercept any realist attempts to score cheap points before they happen.
And I know you're already doing that effectively, but your normative-entanglement blogs usually don't explore these options a lot.
So just for the audience, in the case of the first syllogism you would promptly respond that:
P1. If moral realism is false then slavery is not wrong. -> FALSE
P2. Slavery is wrong. -> TRUE
C. Moral realism is true. -> FALSE
Then the realist will complain and say wrongness is synonymous with stance-independent wrongness and claim that the antirealist can't say P2 is true.
And this is already the most crucial branch point.
The realist willl want to assume realism as self-evident and not needing explanation and then try to score points using this entanglement as you described.
But you must never allow the realist to claim wrongness for themselves and instead press them to explain what stance-independent wrongness is supposed to be. Which of course they can't. Otherwise they would be the first person to have solved meta ethics and they'd leave the discussion to go pick up their nobel prize.
At least a non-naturalist realist has literally zero positive arguments for their position so they MUST attack antirealism and define realism as its negative.
While you must stay on the offense here and not let up, in between formal arguments you can and should make it known that, other than the realist, the antirealist actually has lots of resources to explain their view of morality and what stance-dependent wrongness is. Many of these examples can also be used to score points with an audience. For example observations of moral foundations like fairness in other animals is something that resonates and leaves an impression with people.
And possibly the most rhetorically effective way to turn it around on the realist is to say that exactly the fact that stance-independent wrongness is non-personal makes it appear hollow and suspect. And the only way for morality to be "real" is if it's stance-dependent. And I know you're using this attack as well. You can challenge the realist and dare them to deny that they personally, subjectively think slavery is wrong. Usually, they will agree that they indeed have this stance-dependent position and with that you have established that you both already agree on the antirealist view point. And it will resonate with an audience that you want members of a society to stance-dependently agree with the moral norms and be personally invested instead of treating it as an inexplicable, non-natural abstract thing that may or may not be motivating to the individual. When you have that, any further arguments about the metaphysics of it are just academic and superfluous.
And if they say they don't have a personal stance-dependent sense of right and wrong and would just follow whatever the non-natural moral facts turn out to be, even if they wildly disagree with the norms of the society, then they just look like a psychopath.
Either way the antirealist wins.