Defying most experts

Feb 15, 2024

1.0 Introduction

Consider this this remark about me on Twitter:

Bro Lance Bush is so wacky. If a majority of experts in a field believe something, of course my credence in that belief being true is going to go up.
His "it's all the experts that are wrong not me" shtick is also wild and we wouldn't accept this for nearly any other field.

This comment first begins with the claim that my views are wacky, then states that if a majority of experts in a field believe something, that their credence is going to go up. This pair of remarks implies that I don’t think this. There is also a two minute clip from a conversation I had with Matthew Adelstein over two years ago (from a conversation on 12/29/2021), presumably because that’s supposed to support this claim. The clip begins here (at 31:04) and runs for about two minutes (to 33:05).

Here’s the problem: I explicitly state that I do take what the majority of philosophers believe to be evidence. Here (at 31:51) is the relevant point where I do so. I’ll transcribe the remarks, as awkward as they are:

This is broadly why I’m so skeptical of this high rate of realism among philosophers. As far as I’m concerned I don’t take that as evidence that uh—it’s some degree of evidence for realism—but I take it as stronger evidence that there is something wrong with the way people are being taught philosophy, there’s something wrong with the practices of philosophy, I take it to be embarrassing that that many philosophers endorse realism…

I explicitly say, in the very clip provided, that the fact that most philosophers endorse moral realism is evidence of moral realism. Yet these remarks imply that I don’t agree that “If a majority of experts in a field believe something, of course my credence in that belief being true is going to go up.”

This is simply false. Of course it does. I have said so on this blog, and I have said as much numerous times, some of which is probably written or recorded. Here is an excerpt from this blog post written nearly a year ago:

Suppose we discovered that a majority of those surveyed endorsed a particular philosophical position. Perhaps 70% endorsed physicalism or 90% believed God didn’t exist or 96% thought two-boxing was the best solution to Newcomb’s problem. What should we make of these results? Should we conclude that if a majority of philosophers endorse a particular view, that this is evidence that this view is true? And that the stronger the majority, the stronger the evidence?
Yes.
We should take this to be some evidence that the views in question are true. However, such evidence may be qualified by a variety of considerations, and must be weighed against all the other evidence one has for and against a particular philosophical position.

CharlotteTheLib has not provided an accurate characterization of my views. The tweet gives the false impression that I do not think what a majority of experts think is evidence that would cause me to update my credence in favor of the view.

What I don’t take it to be is decisive evidence, and I doubt many people (including realists themselves) take it to be decisive, in the sense that what the majority believe settles in the matter in their favor. It’s some evidence. That there is some evidence for a view does not mean that the evidence is very strong, nor does it mean that such evidence comes anywhere close to outweighing evidence against the view.

For me, evidence comes very cheaply. The fact that lots of people believe in God is some evidence that God exists. The fact that people have reported seeing Bigfoot is some evidence that Bigfoot exists. And so on. In all of these cases, we weigh such evidence, alongside with corroborating evidence, against evidence to the contrary, and it is the evidence for and against a view, in total, that determines where we fall when it comes our beliefs on an issue.

The question, in this case, is the strength of the evidence provided by the fact that 62% of respondents to the PhilPapers survey endorsed moral realism. I wrote an extensive nine-part series where I argued that while what most philosophers believe is some evidence, but, as I said above, “such evidence may be qualified by a variety of considerations, and must be weighed against all the other evidence one has for and against a particular philosophical position.”

I present a variety of reasons why the PhilPapers survey results do not provide strong evidence in favor of moral realism. You can find these posts here and here is the abstract for that series:

Surveys that indicate most analytic philosophers endorse a particular philosophical position do not necessarily provide strong evidence that the position is true. The strength of such appeals depends on the degree to which analytic philosophy confers appropriate expertise on its proponents, and on a host of background assumptions about the degree to which their judgments are independent of one another and free of parochial and biasing influences that would render their conclusions unrepresentative of the rest of the world’s population under similar reflective conditions. Until and unless legitimate skepticism about analytic philosophy’s methods is overcome, we should put little stock in what most analytic philosophers think.

In this series, I:

Question whether philosophers have the relevant forms of expertise such that we should regard them as experts on whether moral realism is true
Discuss causal factors that should lead us to doubt the strength of the evidence provided by a raw proportion, rather than the causal direction of change in baseline belief caused by studying philosophy
Discuss the role selection effects may play in influencing the proportion of philosopher who endorse a position
Discuss possible variation in how philosophers interpreted the PhilPapers survey questions
Discuss how degree of independence in the factors causing people to arrive at their conclusions is relevant to the degree of evidence provided by the total number and proportion of people who endorse a view
Speculate on the possible role of sociological forces in shaping the current academic landscape, and its potential influence on survey responses
Discuss the representativeness of the survey respondents and the role demographic variables could play in influence their responses

I’ll pick out two examples of the sorts of points I raise to illustrate why these considerations are important for assessing how much stock to put into what most philosophers believe.

2.0 Causal considerations and base rates

(1) Suppose 100% of nonphilosophers are moral realists. If 62% of philosophers are moral realists, and 26% endorse moral antirealism what does this tell us? It may be that studying philosophy causes 38% of people to stop endorsing moral realism and 26% to endorse moral antirealism. Is that good evidence that moral realism is true? I don’t think so. If a particular perspective is so ubiquitous that literally everyone who doesn’t study philosophy believes it to be true, and the only effect that studying philosophy has is to cause over a third of people to stop believing something, this could instead be evidence that people have a very strong bias in favor of moral realism and that studying philosophy can partially but not fully overcome this bias. One might even see this as evidence in favor of moral antirealism.

Of course, the “100%” is totally made up for illustrative purposes (after all, I think hardly any nonphilosophers are moral realists!). The point is simply that looking at the total percentage who hold a view can give us the misleading impression that philosophers all started from some neutral starting point, and whichever direction the overall body of philosophers lean towards, this is evidence in favor of that view. But without knowing if the initial starting point was neutral between the views, it’s hard to know whether the current proportion who endorse a particular position was caused by studying philosophy, or is instead largely caused by antecedent commitments.

For comparison, imagine 95% of a population are theists. Now imagine some of those people began studying the philosophy of religion, and develop an expert knowledge of the topic. We survey these people, and it turns out only 65% of them are theists. The best explanation for this decline in belief might be that expertise in philosophy of religion tends to cause many people to stop believing in God. Would that be good evidence that God exists? I’m inclined to think: no, not really, and perhaps it’s evidence to the contrary.

This point has been completely overlooked by people who appeal to the 2020 PhilPapers survey results: they simply ignore base rates of belief, and treat the proportion who endorse a view as though the entirety of the respondents views have been caused by studying philosophy. And there is simply no good reason to believe this is true.

It’s worth noting that moral realists themselves routinely claim that most people believe moral realism is true, that moral realism is “intuitive,” that it is the “commonsense” view, and that is our “default position.” So it’s entirely plausible that moral realists themselves would be committed to the presumption of an extremely high base rate of belief in moral realism among nonphilosopher populations prior to studying the topic. Ironically, I think this isn’t true, but they do. So in a somewhat amusing twist, since I think almost nobody is a moral realist prior to studying philosophy, I actually think the fact that 62% of philosophers endorse moral realism is probably stronger evidence for moral realism than my critics think that it is, all else being equal. In short, the suggestion that I don’t think it’s evidence isn’t merely false, it’s the opposite of true.

3.0 Potential selection effects

Another potential problem is this: What if people who pursue degrees in philosophy are more likely to be or to come to endorse moral realism than other people? If so, the proportion of people who endorse moral realism may be a selection effect.

Without knowing whether there are systematic differences in the sorts of people who become professional philosophers and people who don’t, we are not in a position to be certain that professional philosophers are psychologically representative of people in general (though there is some evidence they are different, and I’ve argued elsewhere that there are good reasons for hypothesizing that they will differ in relevant ways). As such, their may be idiosyncrasies in the way they are disposed to think. And since people are not randomly selected to become philosophers, it’s possible that professional philosophers are systematically biased towards moral realism.

This is, of course, speculative. But are ther any examples of what philosophers believe where this is a very plausible hypothesis?

Yes, there are.

Take a look at the overall proportion of philosophers who endorse theism and atheism:

Among general respondents, the rate of theism was very low, at just 19%. What happens if we look at philosophers who specialize in philosophy of religion?

…suddenly it flips: philosophers who specialize in philosophy of religion are much more likely to endorse theism than philosophers in general, up to nearly 70%.

What’s the takeaway? Should it be that this is good evidence for theism, because most specialists endorse theism? I don’t think so. Here’s an alternative explanation: people who already endorse theism or are inclined towards theism are more likely than other philosophers to specialize in philosophy of religion.

Of course, specialists in metaethics are slightly more likely to endorse moral realism (65%) than philosophers in general (62%), but that’s little comfort to those who downplay or are unaware of potential selection effects: first, if there are really fantastic arguments for moral realism, why isn’t the number even higher? Second, a bias in favor of moral realism may be a general feature of philosophers relative to nonphilosophers, not just specialists. That’s a bit more speculative, but the lack of a big difference between what specialists and philosophers in general think hardly vindicates the notion that expertise specially in metaethics lends itself towards moral realism.

I’m not claiming to know why exactly we observe the propoprtions we do among specialists and philosophers in general. My point is simply that when experts endorse a position, this isn’t necessarily because their knowledge and competence on the topic has caused them to do so.

4.0 Unacceptable Shtick

Next, CharlotteTheLib says:

His "it's all the experts that are wrong not me" shtick is also wild and we wouldn't accept this for nearly any other field.

First, note the use of the snarl term shtick. This term refers to a kind of performative routine a person engages in, often for entertainment purposes, a kind of gimmick. The implication serves to denigrate my views, giving the impression that my views on the matter are somehow insincere and performative, like a party trick.

Even if my views on these matters were just shtick, that’s hardly relevant to whether I’m correct. This is also an uncharitable and unkind way to characterize my views.

However, this remark is objectionable in more substantive ways, as well.

First, note the line “it’s all the experts that are wrong not me.” All the experts? I don’t disagree with all the experts. I disagree with at least 62% of them on this particular issue. 26% endorse antirealism. Aren’t they experts, too? The rest at least don’t unequivocally endorse moral realism. So at least another 26-38% of the other experts also don’t endorse moral realism. The impression in this remark is that I am somehow defying the entire field on this issue. I’m not.

There are ways in which I do defy most of the rest of the field, but that’s largely because I agree with other “experts” in the field. Even where I differ from most of the people in the field, I don’t just stomp my feet and insist they’re all mistaken. I offer explanations for why I think there are widespread flaws in the methods and presumptions of analytic philosophers. I’m not the first or only person to raise such objections; academic philosophy has hosted dissident metaphilosophical critiques from numerous sources within the ambit of Western philosophy, including pragmatists, positivists, Wittgenstein, ordinary language philosophers, and experimental philosophers. I am in a lot of good company. Consider this fairly recent critique of the entire field from Peter Unger. Should we consider this work unacceptable? If there’s one field that should host critical perspectives of the very foundations of the field itself, I would have hoped it was philosophy.

Third, the majority of respondents to this survey aren’t experts in metaethics. They specialize in other areas. Just what is this remark implying? Of the 1719 respondents, only 228 specialize in metaethics, or 13.3%. The vast majority of respondents, a full 86.7% don’t specialize in metaethics, and are therefore probably not experts on the topic. Incidentally, I do work in and publish in metaethics, including writing my dissertation on metaethics.

Fourth, disagreeing with most, or even all other experts is not always “wild,” even if it were this wouldn’t necessarily be a bad thing. Then we’re given the claim that we wouldn’t “accept” this in other fields. First: sure we would. Progress in science is often driven by someone disagreeing with “all the experts” (I take this to be non-literal; this is fine, because neither I nor the doctor I’m about to describe necessarily disagreed with literally all the experts).

Consider the case of Ignaz Semmelweis, the Hungarian doctor whose pioneering work led to the recognition that washing hands was an important part of human hygiene and could dramatically cut down on mortality rates, a discovery that was made in the context of the high mortality rate associated with childbirth. Was this proposal heralded with parades and celebration? Quite the opposite:

You'd think everyone would be thrilled. Semmelweis had solved the problem! But they weren't thrilled.
For one thing, doctors were upset because Semmelweis' hypothesis made it look like they were the ones giving childbed fever to the women.
And Semmelweis was not very tactful. He publicly berated people who disagreed with him and made some influential enemies.
Eventually the doctors gave up the chlorine hand-washing, and Semmelweis — he lost his job.
Semmelweis kept trying to convince doctors in other parts of Europe to wash with chlorine, but no one would listen to him.

Sometimes “all the experts” are just wrong, and it takes someone to come along and make a case to the contrary. Science has worked this way many, many times, with unpopular views eventually winning people over in spite of what all the experts think, solely on the merits of the quality of the arguments and evidence in favor of the view.

Lastly, CharlotteTheLib appears to endorse Christianity and utilitarianism. According to the 2020 PhilPapers survey, 67% of philosophers are atheists, while only 19% are theists. Not all of that 19% are going to be Christians, so the total number of Christians is something less than 19%.

What about utilitarians? Utilitarianism is a form of consequentialism. About 30.1% of philosophers endorse consequentialism:

Consequentialism is also, by a small margin, the least common choice among the three major choices, and even manages to be less common than “Other.” Note that “Other” includes “Accept a combination of views.” The number in the parentheses above indicates the total proportion of participants who only endorsed consequentialism, which is only 21.4%. The remaining 9.2% endorsed consequentialism along with some other view, which arguably indicates that they are not consequentialists since such views are often characterized as exclusive.

Note that there are forms of consequentialism other than utilitarianism, so the true proportion of utilitarians is going to something less than 21.4%, if we only count someone as a utilitarian if they don’t endorse some hybrid view. That puts the number of utilitarians at quite a bit less than the number of moral antirealists (~26%).1

So CharlotteTheLib appears to endorse two views that are either clearly less common among experts than antirealism, or maybe about as unpopular as moral antirealism is. It’s a bit strange, in light of this, to take issue with me disagreeing with what experts think, given that CharlotteTheLib appears to disagree with what most experts think on these topics.

I’m not happy with the way my views were characterized in this post, but I’d be happy to have a discussion with CharlotteTheLib. Perhaps we could find some common ground, or I could make a case that my views aren’t quite so wacky as they might seem.

Corrections

(1) In an earlier version of this post, I misinterpreted the results for the number of consequentialists. The number who choose “Other” and accept more than one position is included in the total number of results, with those who exclusively favor the position included alongside this in parentheses. I overlooked this and assumed those who chose “consequentialist” chose it exclusively. Thanks to a discord user poggywoggy for the correction. The original text read:

Consequentialism is also, by a small margin, the least common choice among the three major choices, and even manages to be less common than “Other.” Note that “Other” includes “Accept a combination of views” at 15.7%, so it’s possible that the proportion who endorse some combination of consequentialist and non-consequentialist views would push the total number up as high as ~47%, if every single one of those people endorsed consequentialism and something else, which still isn’t enough to be a majority.
Note that there are forms of consequentialism other than utilitarianism, so the true proportion of utilitarians is going to something less than 30%, if we only count someone as a utilitarian if they don’t endorse some hybrid view. That would probably put the number of utilitarians at less than the number of moral antirealists (~26%).

(2) I also accidentally put in the numbers for deontologists during my last correction, rather than consequentialists. This has also been corrected.