Meghan Murphy's Speedrun Guide to Logical Fallacies

[Note: The version of this post that was sent by email unfortunately had footnotes missing.]

I can’t help feeling a tinge of awkward self-consciousness whenever I describe myself as part of the modern-day “Rationality” movement. As I’ve said before, it’s perfectly reasonable to be suspicious of a group with such a self-serving and masturbatory name, because isn’t everyone in favor of rational thinking? The core lessons of rationality are — or at least should be — dreadfully pedestrian: you should test your theories with evidence, you shouldn’t believe things that aren’t true, you should make logically coherent arguments, threatening to kill someone is not a valid rebuttal to their arguments, et cetera forever.

Like I said, boring stuff that you would expect from any public intellectual or anyone with even a passing familiarity with logic and critical thinking. It’s at least one reason why I didn’t really understand or appreciate why Eliezer Yudkowsky spent so much time writing The Sequences, a long series of blog posts that aims to break down rationality into delicious bite-sized chunks. But I live in a bubble full of rat nerds, and sometimes I get shocked awake with some cold water thrown at me. In this post, I want to highlight one recent episode that particularly stood out to me as emblematic of the evergreen utility of rationalist concepts,¹ and it’s the debate between Aella and Meghan Murphy on the ethics of the sex industry (prostitution and pornography) on the Calmversations podcast hosted by Benjamin A. Boyce.

To set things up, Meghan Murphy is a feminist thinker who has written extensively about the harms prostitution and pornography have imposed (primarily) on women, including arguing in favor of a total ban on pornography. Aella has worked as a prostitute and produced porn, and her positions on the sex industry are significantly more positive. I should say that I’m friends with Aella, but that hasn’t stopped me from pointedly criticizing her ideas before. I reached out to Meghan Murphy by email, and although this post will be rather critical of her reasoning, she should absolutely be commended for responding to my questions despite the severe snark I previously tweeted her way.

I should also make it clear that any criticism should be construed narrowly, and is not meant as a broad denunciation of either Meghan’s work in general nor — crucially! — her specific criticisms about the sex industry. I want to make sure this point is heavily emphasized, as one of the most useful (and, unfortunately, least adopted) tools in rationality is the ability to cut discussions into discrete pizza slices, such as abstracting from the object level to the meta level, or decoupling distinct ideas from each other. This is an invaluable practice in any discourse because humans are prone to irrational distractions, and very often we get reflexively defensive and assume that an attack on an argument or premise is an attack on a conclusion. To use the parlance, someone could be “accidentally right” and reach a true conclusion despite using faulty logic. So just because I criticize someone’s argument does not mean that their conclusion is false.

Make Your Belief Falsifiable (i.e., Make That Motherfucker Burnable)

If I had a rationality genie grant me one wish, it would be to force everyone to make their theories falsifiable. Put simply, all it means is that if you present a theory, you should be able to articulate at least the theoretical circumstances by which your theory would be “falsified” — proven wrong. Carl Sagan elucidated an excellent illustration of falsifiability when he described the dragon in his garage.

Consider if we were in the same house and I claimed the roof was on fire. Given what you know about the potential risks of fire, you may then want to evacuate the house, or (if you’re feeling frisky) perhaps decline any water and just let the motherfucker burn. But if you look around you and point out that you don’t see any flames, don’t smell any smoke, nor do you hear any fire alarms, you’ve reasonably falsified my claim. If my response is to claim the fire I’m talking about is invisible, doesn’t produce any smoke, and doesn’t trigger any fire alarms, you’d be right to dismiss what I’m saying as complete fucking nonsense because I’m not conveying any useful information to you. An unfalsifiable theory is a broken compass, useless precisely because it’s “north” no matter what direction it’s facing.

One of the very first topics Aella and Meghan broached in their debate was on the subject of the effects of pornography on its users. Meghan’s argument is fairly straightforward (~2:23):² she believes that a consistent exposure by men to violent and/or extreme pornography (and reinforced by orgasmic dopamine feedback loop) would encourage men to replicate the acts depicted on screen in real life. Meghan repeats a version of this argument several times, and ties it explicitly to child abuse (~11:51):

I think that the theory of men seeking out barely legal porn and the amount of men who look for child porn, even online, proves that there’s a lot of men who are seeking out girls and underage women online to jack off to. And I think that we know that molestation and child abuse and sexual abuse and men predating on young girls, or like young women, to young women is a pretty big problem in our society. So I think that adding porn to that mix is for sure not helping and probably making it worse.

This is a logically sound and straightforward argument and, most importantly, testable and falsifiable. If Meghan argues one reason that porn is bad is because it causes bad things, then a falsification would be the absence of said bad things.³ But when Aella asks Meghan “What kind of data would make you update your mind?” Meghan responds “No data” and blithely asserts “I think it’s bad for men to jack to barely legal porn.”

The same happened when I reached out by email and asked about her position vis-à-vis the harm on women within the industry. Meghan has demonstrably changed her mind on several issues over time, so I cannot accuse her of being completely immune to new evidence or arguments. Her initial approach to porn involved “feminist porn” projects in school and reading articles by “empowered” sex workers and academics about female agency and how we need to flip the script on the “victim” narrative. But over time, she couldn’t shake off the persistent and gnawing discomfort she had with pornography and eventually connected with and interviewed women who spent many years in the sex industry, who did frontline work, and academics who extensively studied the sex trade (along with reading a lot more second-wave analysis of the topic). In other words, Meghan discovered new evidence and arguments that made her change her mind.

I asked Meghan to explain the apparent contradiction between claiming that no data would change her mind, while simultaneously lucidly explaining how data did indeed change her mind. Meghan’s answer was not responsive to the issue; she reiterated that no data would change her mind because her position is based on ethics.⁴

As shown in numerous instances throughout the debate, Meghan readily provides evidence and support to explain why she is against the sex industry (~31:23):

I mean, I talked to lots and lots of women in prostitution. Women who’ve worked as high-end escorts, women who’ve worked in the street, women who worked in the Downtown Eastside, women who worked in brothels. And they all said they didn’t want to be there. They all have suffered immense physical and emotional trauma. Most of them came from physical and emotional trauma. Most of them came from homes where there was sexual abuse. And it was really a scary, horrific experience for them that is very difficult to get out of. So I think most of the women who are in prostitution and pornography don’t actually want to be there.

But does this mean you can Believe All Women™ regarding what they say about the conditions of the sex industry? Of course not, because Meghan preemptively applies a strict credibility filter, her version of “No True Prostitute” (~37:35):

Typically, I will say that I don’t think that we can ever hear the real truth about your experiences in this industry. Like you can’t say that publicly because this is your job and your income. And most of the women that I’ve talked to who talk about their horrible, traumatic experiences are women who’ve exited prostitution and have had a lot of time outside of prostitution or pornography to look back. Some of them are still in it and struggling to get out or having conflicting feelings about the situation. But most of the women that I’ve talked to who speak really, really honestly about what happened to them and how they felt about the johns, are women who’ve been out of the industry and so are free to speak publicly, but are also better equipped to look back on what was going on truthfully.

According to Meghan, if you speak ill of the industry, you’re telling the truth, but if you speak positively, you’re either still too traumatized or too deluded to speak honestly. Similar to the dragon in my garage, this is a perfect example of Belief in Belief. Crucially, the rationalizations Meghan constructs to ignore inconvenient evidence demonstrates the keen awareness she has about her argument’s vulnerabilities. She’s aware that Aella reporting a positive experience with the sex industry would plainly contradict her stated premise and so Meghan preemptively shores up this vulnerability by declaring Aella’s opinion implacably dishonest and therefore void: (~42:39) “You’re never going to say anything negative about this industry because it’s where you make your money.” When Aella does offer to say something negative, Meghan realizes she needs to hurry up and come up with a different excuse for why Aella’s opinion is still void. Meghan greases the wheels on her goalpost and launches it across the horizon:

Once you’re out of the industry and there’s zero incentive for you to say positive things about the industry, then I would trust your opinion. And I would believe that you were being completely honest. So I feel like if we talked to you in 10 or 15 years, maybe you would say something.

I have no charitable explanation for this move. Meghan creates a new standard on the fly that is conveniently unreachable (unless you have the patience to wait 10–15 years). If you’re not seeing any fire, it’s because the fire is invisible. Heads I win, tails you lose.

Even if this rampart is somehow breached, anytime Meghan’s evidence is challenged, she can still retreat into the safety of the ethical motte by claiming her positions are above evidence anyways (~47:00):

I think that the problem in this conversation is that my issues with prostitution and pornography are primarily ethical. So when you say “data won't change your mind,” I mean, data won’t change my mind because fundamentally, I think that it’s unethical to pay another person for sex.

So which is it? Meghan’s logic is extremely confusing to follow.

My theory is that she realizes conclusory declarations such as “It’s bad when men jerk off to porn” are going to have zero persuasion, so she necessarily has to establish some nexus to real-world harm. But if your argument is consequentialist, stick to consequentialism. If your argument is ethical, stick to ethics. There’s nothing wrong with either approach done separately, but Meghan appears to want to have it both ways to maintain an acrobatic ability to backflip away from ever having to defend her positions. Amazingly, she even managed to do this within the confines of a 280-character limit:

Meghan Murphy @MeghanEMurphy

@SJLanderos @lexfridman @BenjaminABoyce There is no data that will convince me that it is ethical to pay for sex. Also, all 'data' shows abuse and exploitation is rampant in the sex trade. You can say 'data' all you want, but nothing exists that can change the truth or ethics of the trade.

Be Skeptical of Things You Dislike and Credulous of Things You Like

It’s trivial to end up bombarded with an endless avalanche of studies, each purporting to show how coffee or whatever will give you all sorts of cancers or, depending on the day of the week, cure them all. For most of us, it’s just not practically possible to scrutinize every coffee study that gets published, so we probably ignore individual findings and their exact methodology and resort to broad heuristics instead (e.g., “all the cool kids drink coffee”). But let’s say you run a coffee roasting company, and anytime you come across a study purporting to show the benefits of coffee, you can’t hit that retweet button fast enough, but when it’s a negative finding, you use a Ouija board to invoke the ghosts of dead statisticians past to poke holes in the methodology. If your skepticism is deployed in only one direction, there is a serious risk of incurring a rotator cuff injury from patting yourself too much on the back from being (allegedly) correct. To use another navigation analogy: if you credulously accept every right turn, but are severely skeptical of every left turn, you’ll just keep going in circles.

Aella has been known for her Twitter polls for a while, and more recently she shifted toward working with large survey datasets built from internet responses numbering in the hundreds of thousands. Throughout, sample and selection bias have been a persistent criticism of Aella’s surveys. This indeed can be a real problem, as the fake anecdote about liberal insularity (“I can’t believe Nixon won. I don’t know anyone who voted for him”) purported to demonstrate, but it’s also a problem we already learned to live with as it also affects virtually every quote unquote real study out there. If you think selection bias is an indelible sin, what exactly can you believe?

Following a discussion on how porn companies physically exploit women, Aella happens to mention that one of her surveys found that women tend to prefer more violent porn than men (~19:11). Meghan is immediately skeptical that this finding applies universally to all women, and as a rebuttal she cites a study linked in a 2017 Vice article purporting to show that actually more men seek out violent porn in total because there are more men that consume porn period. This is a logically sound and well-reasoned objection on how much extrapolation you can reasonably draw from a potentially skewed sample. However, by Meghan’s own admission, she never even read this study she references (“I think what their research says, and I didn’t actually read the research, I just read this article”) but nevertheless she believes the conclusion is credible enough to relay to her audience.

As I said above, we don’t have infinite time to carefully scrutinize every potential study we come across, so Meghan’s credulity here is not necessarily ill-placed. But elsewhere in the debate, we get an excellent mirror test for how Meghan evaluates the credibility of studies with conclusions she disagrees with. When Aella brought up her own surveys, Meghan was immediately and consistently skeptical of the methodology and conclusions (e.g., ~21:00 “Who are these women? Where did they come from?”, ~39:46 “Well, you did a TikTok survey. I mean, I don’t know who these people are. I don’t think that’s like an official study,” etc.)

Meghan relayed her criticism in further detail when I asked her to outline what was wrong with Aella’s survey methodology and what changes she would make to the surveys. Here is her response in full:

They aren’t real studies. They are not peer reviewed, the questions begin with incredibly biased assumptions, we don’t know who is even answering the surveys beyond the fact of being her anonymous followers, all of whom are porn consumers. They offer nothing revealing about the population at large and her stated conclusions make absolutely zero sense if you look at the questions and numbers of respondents. She has zero training in statistics, data science, or survey methodology. Her survey can’t answer the question implied by the title & conclusion — “Do women prefer violent porn?” — as the question implies that she’s comparing women’s preference for violent porn vs non-violent porn and assumes that all women like/consume porn. Moreover, the vast majority of porn consumers are men which she completely ignores in her stated conclusion. She also very clearly has a definition of ‘violence’ that is very different from many other people’s definition of violence, though she wouldn’t know this because she presumes ‘regular porn’ is not violent, which is complete and total bullshit. Most porn is violent. And women are more likely to view ‘violence’ differently than men. Either way, all the responses are going to be subjective. Such a tiny tiny number of women she surveyed responding they liked ‘violent porn’ (whatever that means) that to conclude ‘women prefer more violent porn’ is frankly insane, if not an intentional misrepresentation of her own ‘findings.’ Her conclusion implies that women’s preference for violent porn is greater than men’s preference for violent porn, but she didn’t construct her “experiment” (in this case, setup, survey, and analysis) in such a way as to be able to answer this question. In fact, the data she did collect indicates that neither women nor men like violent porn. She is not a researcher in any way shape or form and her ‘research’ is frankly embarrassing.

Whether Meghan’s criticisms are valid is beside the point because the problem is the obvious dissonance: she never investigated whether the study she casually referenced (and didn’t even read) from the Vice article met the exacting criteria she outlines above. I directly asked Meghan to explain the discrepancy between the skepticism she displayed against Aella’s methodology versus the credulity she displayed on a study she didn’t even read. Meghan claimed that, contrary to what she said on video, she did in fact read the study she was referencing.⁵ She also claimed that it was “not reasonable for [Aella] to try to bring up her own junk data as proof of anything” but also that it wouldn’t matter if it were true that women preferred violent porn because it “doesn’t make violent porn ok.”

If the results don’t matter to begin with, then why waste time refuting them? Although I can’t read Meghan’s mind, I believe that the inconsistent criticism is pretextual. This is a clear example of the Arguments as Soldiers mentality, where once you know which “side” you’re on, you feel an obligation to support all arguments on that side, no matter how terrible they might be. In fairness, although it’s intellectually dishonest, it also is not necessarily unreasonable behavior from either an activist or a religious proselytizer. After all, their goal is to win the cause, not to seek truth.

Maybe Meghan is indeed acting honestly here. If so, and in line with my devotion to falsifiability, there is a plainly obvious way to prove me wrong, one which I asked Meghan to do directly. If my accusation — that Meghan is making up excuses to discredit a study just because it’s inconvenient for her preferred conclusion — is false, all she would need to do to prove me wrong is outline what evidence would convince her that her disfavored conclusion is true. Meghan has already refused to do so multiple times.

It bears repeating that Meghan has changed her mind about issues many times. She’s capable of lucidly describing support for her arguments in a logically coherent fashion. But at least within the confines of this debate, time and time again, as soon as her conclusions are challenged, she demonstrates remarkable agility in either evading the criticism, or landing on a separate justification. There’s a general coherence and accord to her positions, so it wouldn’t make sense to claim she arrives at her conclusions haphazardly, but it’s almost as if as soon as she reaches her destination, she burns up the epistemological boats behind her. No turning back now.

The Soothing Comfort of a Cocoon

I know what it’s like to be a zealous partisan hack, because it’s what I do for work as a public defender. My job is not to seek truth; my job is to zealously defend my client, so fuck your so-called evidence. How do we know that’s actually blood? It could be anything! And what if the cops planted that evidence? Who knows! But even within the confines of an arena where I am professionally and ethically obligated to be a shill, there’s a limit. If I make a habit of pretending that all my arguments (even the shitty ones) are above reproach, it’s reasonable for judges and prosecutors to stop taking me seriously, because my voice would be hoarse from crying wolf.

Even within the role of an advocate, I would think (hope?) that there is merit to maintaining intellectual honesty. I recognize there is an undeniable seduction to hijacking your audience’s reptile brain by blustering confident assertions, since confident assertion is a social hack that is perennially exploitable. But I would also think (hope?) it’s easy to reach a ceiling on the persuasion scale with these parlor tricks. Maybe you can convince the people already on your side, but for everyone else, how effective is this really? Setting aside the impact this may have on one’s persuasion acumen, there is also a serious risk of suffocating yourself with an impenetrable epistemological cocoon of your own making.

For the purposes of clarity, I intentionally avoided diving into the object level debate so far. But now that we got the meta level out of the way, it’s likely that I agree with Meghan on a host of issues. I personally encourage avoiding pornography because it’s probably harmful on net for the consumer. The production side is admittedly not a facet I thought extensively about (chalking this up to male privilege is not unreasonable), but I find a significant portion of the industry’s dynamics to be extremely distasteful. For example, Meghan posted a documentary about the “barely legal” genre that I found too disturbing to watch. In part it’s because I agree with Meghan on some topics and because I wish to detach from seeing debate as warfare that I believe it’s especially important for me to critique her arguments (which, it bears repeating, are distinct from her conclusions). Because so much criticism gets reflexively interpreted as hostile warfare, and because of our fallibility of credulously accepting faulty reasoning that flatters our priors, I think we have a duty to be especially vigilant about policing bad arguments for conclusions we might agree with.

1. Credit to @SJLanderos for first pointing out some of the issues I cover here.

2. Auto-generated (read: slightly inaccurate) transcript of the debate can be found here.

3. There may be other reasons why the conclusion (porn is bad) remains true, but if so, you should make a habit of preregistering your hypotheses to avoid any accusation or appearance of post hoc justifications.

4. Full answer: “Because there is no data that can make that industry anything but exploitative, abusive, and unethical, just as there is no data that can make slavery, or child labour, or murder, or animal abuse, or any of the other things we have decided as a society are too unethical to accept, ethical.”

5. Meghan never explained this discrepancy.