It is a terrible argument. Myself and a couple other people reviewed Harris on this topic some time ago and offered our own objections, which you can find here:
There are still more issues. Even if the results of a study can be reproduced given the raw data, and even if the findings can be replicated in subsequent studies, that does not ensure that results have identified the effect researchers claim to have found.
This is because studies can rely on invalid measures. If a study claims to measure P, but fails to do so, it may nevertheless pick up on some real pattern other than a successful measurement of P. In these cases, results can replicate and appear legitimate even if they don't show what they purport to show.
There’s been some survey data on this, e.g.:
Fanelli, D. (2009). How many scientists fabricate and falsify research? A systematic review and meta-analysis of survey data. PloS one, 4(5), e5738.
This study reports that ~2% admitted to data fabrication. However, it is of course difficult to get a good estimate by asking people, since this is a case where people have strong incentives to lie. Asking people about suspicions of colleagues may give an overestimate. So I think in general it’s very hard to estimate actual fabrication rates.
One obvious issue is that many instances of fraud that are caught are likely to be cases where the data looked suspicious to others. This requires eyes to be on the data, for someone to notice, for that person to follow through, and most importantly, for the fraud to be sloppy enough that someone noticed it. This means identified cases of fraud are probably from people who are less careful. So we’re seeing the most blatant, obvious, and sloppy fraud. People who are very good at committing fraud are much more likely to go undetected. And that’s scary.
I don't think the claims are stance independent, either, so I don't think there's any loss. In other words, I don't think typical moral claims imply stance-independence or stance-dependence. They don't imply or hint at any particular metaethical position at all. Why suppose that they do?
We don't take causal claims like "It's going to rain tomorrow" to imply a position on how to interpret quantum mechanics. Likewise, it may be that everyday moral claims are simply indeterminate with respect to metaethical presuppositions.
My claim is that normative claims have subtypes. "subjectively wrong" doesn't mean what "objectively wrong" means.
I agree. But I don't think these categories and distinctions regularly figure into everyday normative and evaluative claims. They're philosophical inventions, and have little to do with what ordinary moral and normative discourse is about. At any rate, to the extent that some form of these notions does manifest, I don't think we can readily read it off of the superficial appearance of seemingly fact-stating claims just by examining the structure of toy moral sentences in the abstract. If we want to know what people are doing when they make moral claims, we should be doing empirical work that involves examining actual instances of usage, not hypothetical ones.
In a way that's your position , too, since you think subjective wrongness exists and objective wrongness doesn't.
Depending on precisely what is meant by subjective wrongness, I don't even believe some forms of that exist, either.
You're not making an noncommital statement because you think there is nothing to choose between objectivity and subjectivity.
Sorry, not sure what you mean. Can you clarify or restate? The way I use moral and normative language is idiosyncratic and certainly doesn't reflect ordinary usage. I'm discussing how other people use these terms, not how I use them. If someone wants to know how I use normative language I can just tell them. No need to speculate.
Likewise, realistic metaethics has implications for normative ethics ,not so much in terms of what is wrong , but in terms of how wrong it is.
What do you mean when you say that metaethics has implications for how wrong something is?
Which gets us back to the issue of slightly misrepresenting relativist views...leaving out the stance dependence makes the problem slightly harder to spot.
I'm not sure if what you're referring to is my criticism of Carroll's remark, but my criticism is that he characterizes relativism in terms of agent rather than appraiser relativism.
Moral realists are going to differ with respect to what they think the metaphysical status of the moral facts are. Moral naturalists may see them roughly as a kind of natural fact, so moral facts might be facts about e.g., increases in wellbeing, while non-naturalists would maintain that moral facts aren't reducible to natural facts.
A moral realist would think that there are facts about what is morally right or wrong that are true regardless of what anyone thinks about them. One way to put this is that they aren't made true by our desires, goals, standards, values, beliefs, and so on. Rather, they are true in a way more like how claims about e.g., the mass of an object are true. Facts about the mass of an object aren't made true by our believing them or preferring them to be the case.
Why do you say they're lossy or inaccurate renditions of it? My position on this is that statements like "murder is wrong" are simply normative claims, and they are in no way more indicative of realism or antirealism. I'm still not understanding why you think they'd indicate realism. Why presuppose that such statements have anything to do with expressing metanormative standards at all? They're normative claims, not metaethical ones, and it's not clear to me why we'd imagine a normative claim (i.e., a claim about something being right, wrong, permissible, impermissible, and so on) suggests any particular metanormative stance, unless such a stance were :
(a) explicitly accompanying the remark, e.g., "murder is objectively wrong"
(b) we had background knowledge about the speaker in question that would suggest they're using that way, e.g., a moral realist says "murder is wrong"
or
(c) we had background knowledge about the degree to which such language was typically used to convey claims with particular metanormative presuppositions, e.g., we ran a bunch of surveys and discovered most people from the population the person is from are committed to moral realism
Without such information, I see no particular reason to presume such remarks hint at realism merely by examining the structure of the sentence.
That would be true of casual conversati on, but not philosophical debate.
I disagree. Such norms apply to philosophical conversations as well. For what it's worth, I'm a moral antirealist and I use normative language all the time. I don't think moral realists have any kind of monopoly on, or priority over, straightforward normative claims in any domain, because I don't think normative claims hint at any particular metanormative standards.
Okay. Thanks. With respect, I disagree. I do not think that claims like "murder is wrong" or "pizza is tasty" in any way imply or even hint at normative realism about the claims in question. Both claims are completely consistent with antirealism, and it's not at all clear to me how either would indicate some form of normative realism.
I am not sure the reason you gave, that it's phrased as a one place predicate, is any kind of substantive indication of realism. I can grant that:
"Murder is wrong," is more consistent with, and more likely to be an expression of a realist stance than "I disapprove of murder," but whether it is more consistent with what someone would say if they were a realist about the issue in question relative to some other remark that is less likely to express realism doesn't indicate in absolute terms that it meaningfully hints at realism. However, nothing bars a normative realist from expressing subjective attitudes, and nothing bars an antirealist from employing conventional assertoric language to express subjective (or more generally nonrealist) evaluative standards or normative judgments.
For one thing, expressions of our preferences often exclude any explicit qualification that they are our preferences because in many contexts it would violate Gricean maxims to explicitly indicate that something is a preference, or an expression of our subjective attitudes. To the extent that most people aren't gastronomic realists, a statement like "chocolate ice cream is delicious" doesn't need "...in my opinion" at the end, or "I consider" at the beginning because this is implicit. People may include such qualifications explicitly, but typically only in contexts in which e.g., some contextual goal is relevant, such as not offending someone with a contrary opinion, or to emphasize that you are stating a contrary opinion.
It doesn't seem entirely unreasonable to suggest that these things are viewed on LW in something like the way the analytic/synthetic distinction and reductionism were viewed among empiricist philosophers when Quine wrote "Two Dogmas"
That's fair. I can grant that. Like you, I'm less sure about the general attitude towards moral realism here. I'd have thought inclinations were more towards dissolve-the-dispute than a decidedly antirealist stance. I'd be interested in finding out more about people's metaethical views on LW.
At least one concern I'd have with worries about moral realism and AI alignment is that I think the case for moral realism is much weaker than may be popularly believed in some circles. While a majority of analytic philosophers responding to the 2020 PhilPapers survey favor moral realism (62% favor realism, 26% favor antirealism), I take this more as evidence of problems in the field than an especially good reason to think the "experts" are onto something.
More importantly, I don't think moral realists have any good arguments for their position. Yet the view that moral realism isn't just not true, but that it has very little going for it, doesn't strike me as very popular. I am hoping that can change. If moral realism is as implausible as I think it is (or if it isn't even intelligible), this should reduce our worries about it creating complications with AGI.