Decision Theory Anti-realism

There Is No Fact Of The Matter About Correct Decision Theory

With the recent flurry of posts in the rationalist community about which decision theory ( e.g. CDT, EDT, UDT etc..) it’s time to revisit the theme of this blog: rejecting rationality realism. In this case that means pointing out that there isn’t actually a well-defined fact of the matter about which decision theory is better. Of course, nothing stops us from arguing with each other about the best decision theory but those disagreements are more like debates about what’s the best programming language than disagreements about the chemical structure of Benzene.

Any attempt to compare decision theories must first address the question: what does it mean for one decision theory to be better than another? Unlike many pseudo-problems1 there is a seemingly meaningful answer to this question: one decision theory is better than another to the extent that the choices it recommends lead to better outcomes for the agent. Other than some ambiguity about which theory is better if neither dominates the other it seems like this gives a straightforward criteria for superiority: we just look at actual outcomes and see which decision theory offers the best results for an agent. However, this only appears to give a well-defined criteria because in every day life the subtle differences between the various ways to understand a choice and how to conceptualize making a choice don’t matter.

In particular, the kind of scenarios which distinguish between the various decision theories yield different answers depending on whether you want to know who you should be (i.e. total source code) to do best, how you should program an agent if you want them to do best, which decision rule should you adopt for you to do best, and what choice gives you the best outcome. Furthermore, these scenarios call into question how the supposed ‘choices’ made by the decision theory relate to our intuitive notion in a way that makes them relevant to some notion of good decision making or if they are simply demanding the laws of physics/logic give way to offer them a better outcome in a way that has nothing to do with actual decisions.

Intuitions and Motivation

I’m sure some readers are shaking their heads at this point and saying something like

I don’t need to worry about technical issues about how to understand a choice. I can easily walk through Newcomb style problems and the rules straightforwardly tell me who gets what which is enough to satisfy my intuitive notion that theory X is better. Demanding one specify all these details is nitpicking.

To convince you that’s not enough let me provide an extreme dramatization of how purported payouts can be misleading and the question turns on a precise specification of the question. Consider the following Newtonian, rather than Newcombian, problem. You fall off the top of the empire State building what you do as you fall past the fifth floor? What would one say about the virtues of Floating Decision Theory which tells us that in such a situation we should make the choice to float gently to the ground. Now obviously, one would prefer to float rather than fly but posing the problem as a decision between these two choices doesn’t render it a real choice. Obviously, there is something dubious about evaluating your decision theory based on it’s performance on the float/fall question. At least on one conception a decision theory is no worse for failing to indicate the agent do something impossible for them so we can’t merely blindly assume that anytime we are handed a set of ‘choices’ and told what their payoffs are we can simply take those at face value.

Yet, this is precisely the situation we encounter in the original Newcomb problem as the very assumption of predictability which allows the demon2 to favor the 1 boxers ensures the physical impossibility of choosing any number of boxes other than what you did choose. Of course, the same is (up to quantum mechanical randomness) true of any actual `choice’ by a real person but under certain circumstances we find it useful to idealize it as free choice. What’s different about the Newcomb problem is that, understood naively, it simultaneously asks us to idealize selecting 1 or 2 boxes as a free choice while assuming it isn’t actually. Thus, it’s reasonable to worry that our intuitions about choices can’t just be applied uncritically in Newcomb type problems and now I’ll hope to motivate the concern that there might be multiple ways to understand the question being asked.

Let’s now modify this situation, by imagining that we actually live in the Marvel Universe so there are a number of people (floaters) who respond to large falls by, moments before impact, suddenly decelerating and floating gently to the ground. Now suppose we pose the question of whether, as you fall past the 5th floor, you should choose to have been born a floater or not. Obviously, this question suffers from the same infirmities as the above example in that intuitively there is no ‘choice’ involved in being a floater or not but being a floater. However, we can mask this flaw by instead of phrasing the choice as between being a floater and not instead phrasing it as being between yelling, “Holy shit I’m a floater” and concentrating totally on desperately trying to orient yourself so your feet strike first. Now presuming there is a strong (even exceptionless) psychological regularity that only floaters take the first option it follows that EDT recommends making such a yell while CDT doesn’t.

However, taking a look at the situation it seems clear that the two theories are in some sense answering different questions. If I wanted to know whether or not it is preferable to be the kind of person who yells “Holy shit I’m a floater” then I should consult EDT for an answer. Instead, if I’m interested in what I should do in that situation that doesn’t seem particularly relevant. I believe this should move us to consider the possibility we haven’t asked a clear question when we ask what the right decision theory is and in the next section I will consider a variety of ways the problem we’re trying to solve can be precisified and not they give rise to different decision theories.

Possible Precisifications

Ultimately, there is something a bit weird about asking what decision a real physical agent should take in a given situation. After all, the agent will act just as it’s software dictates and/or the laws of physics require. Thus, as Yudkowsky recognizes, any comparison of decision theories is asking some kind of counterfactual. However, which counterfactual we ask makes a huge difference in what decision theory is preferable. For instance, all of the following are potential ways to precifisify the question of what it means for it to be better for XDT to be a better deciscion theory than YDT.

  1. If there was a miracle that overrode the agent’s programming/physical laws at the moment of a choice then doing so in the manner prescribed by XDT yields better outcomes than doing so in a manner prescribed by YDT.
  2. In fact those actual agents who more often choose the outcome favored by XDT do better than those who choose the outcome favored by YDT.
  3. Those actual agents which adopt/apply XDT do better than those who adopt/apply YDT.
  4. Suppose there is a miracle that overrode physical laws at the moment the agent’s programming/internal makeup is specified then if the miracle results in outcomes more consistent with XDT than YDT the agent does better.
  5. As above except with applying XDT/YDT instead of just favoring outcomes which tend to agree with it.
  6. Moving one level up we could ask about which performs better, agents whose programming inclines them to adopt XDT or YDT when considered.
  7. Finally, if what we are interested in is actually coding agents, i.e., writing AI software, we might ask whether programmers who code their agents to reason in a manner that prefers choice A produce agents that do better than programmers who code agents to reason in a manner that prefers choice B.
  8. Pushing that one level up we could ask about whether programmers who are inclined to adopt/apply XDT/YDT as true produce agents which do better.

One could continue and list far more possibilities but these six are enough to illustrate the point.

For instance, note that if we are asking question 1 CDT outperforms EDT. For the purposes of question 1 the right answer to the Newcomb problem is to be a 2 boxer. After all, if we idealize the choice as a miracle that allows deviation from physical law then the demon’s prediction of whether we would be a two-boxer or one-boxer no longer must be accurate so two-boxes always outperforms one boxing. It doesn’t matter that your software says you will choose only one box if we are asking about outcomes where a miracle occurs and overrides that software.

On the other hand it’s clearly true that EDT does better than CDT with respect to question 2. That’s essentially the definition of EDT.

To distinguish the remaining options we need to consider a range of different scenarios such as demons who punish agents who actually apply/adopt XDT/YDT in reaching their conclusions. Or consider Newcombian demons who punish agents who adopt (or whose programmers adopted one of XDT/YDT).

Ultimately, which criteria we should use to compare decision theories depends on what we want to achieve. Different idealizations/criteria will be appropriate depending on whether we are asking which rule we ourselves should adopt, how we should program agents to act, how we should program agents who program agents etc.. etc… Moreover, I’d suggest that once we’ve fully preciscified the kind of question we want to ask the whole debate about which decision theory is best becomes irrelevant. Given a fully specified question we can just sit down and compute (or do empirical analysis) and when we can’t it indicates that we’ve failed to fully specify what we are asking.

The Use of Decision Theory By Agents

As a postscript I’d note that it’s also misguided to assume that the right way to program some kind of AI agent is to have that agent adopt some kind of decision theory like framework. Many discussions of decision theories seem to presume this by phrasing questions in terms of what decision theory should an AI apply/adopt. However, there is no reason to suppose that the way to produce the behavior favored by XDT is for the agent to actually believe/apply XDT. For instance, if a demon punishes agents who have adopted XDT then the outcomes XDT prefers might be best achieved by agents which explicitly eschew XDT. More pragmatically, it’s not at all clear that the most effective way for agents to reach XDT compatible outcomes is to perform the considerations demanded by XDT. That’s a good way to implement some algorithms but not all.

The reason that decision theory is useful in normal situations (i.e. lacking Omega/Newcombian demons) is that it’s a decent heuristic to assume that the way we internally consider outcomes/make choices doesn’t affect the payout we receive. Under this assumption pretty much all ways of preciscifying the question give the same answer and it offers some good advice for programming agents. However, the usefulness of the framework once we abandon this isn’t clear and can’t simply be assumed.
Thus, not only would I argue that the debate over which decision theory is best is misguided, but that we need to be more careful about the assumptions we make about applicability as well.

Thus, not only would I argue that the debate over which decision theory is best is misguided, but that we need to be more careful about the assumptions we make about applicability as well.


  1. For instance, any attempt to answer what makes one programming language better than another reveals substantial disagreement about which tradeoffs are desirable and no agreed upon framework for resolving them. Indeed, we in some sense all recognize that which programming language tradeoffs are desirable is context dependent. 
  2. Or in Yudkowsky’s formulation, Omega. 

Legally Recognizing Digital Human Memory

Or Using Smartphones and Smart Speakers to Solve All Violent Crime

We are at the point, or very near it, that our technology could virtually eliminate unsolved violent crime. For instance, suppose we all constantly captured audio with all our smart devices and uploaded it to the cloud. Our devices could update the uploaded buffer when we log into our devices later and fail to report an emergency1. Unlike biological memory this could be reliably used in court and captured even by murder victims.

Why don’t we use this kind of tech to take a bite out of crime? Well, for the moment it still might strain our bandwidth and storage resources but 5G and ever cheaper storage make this a temporary issue but even people who can afford those resources aren’t so inclined. Now one might think it’s out of fear of technical loss of privacy. What if amazon, google or some hacker figures out how to access our buffered audio?

But that’s not really a convincing worry since it’s pretty easy to secure the buffered audio more securely than our devices themselves are secured. I mean anyone who can hack our cell phones and smart speakers can just enable listening in directly while we could split the secrets to decrypt the audio between multiple big tech companies. The real problem is that we are creating a record that can be subpoenaed and used against us or our intimates without our permission.

What we need to solve this problem is some digital legal analog of biological memory. That is a class of digital records that, like biological memory, need not be produced if the creator chooses not to. Of course, it’s actually a bit more subtle because we can all be compelled to testify but when we do so we not only maintain our 5th amendment protections we can also simply lie or evade. In conjunction this prevents fishing expeditions that a digital record would allow (e.g. I’m sure my husband did something unsavory during the last 48 hours let’s subpoena his audio record so we can use it against him in the divorce).

One possibility is to only use such records as a supercharger for testimony (unless the individual who created it has died). In other words the only access to such data would be by allowing it’s creator to review the tape and indicate what happened in a deposition that would be checked for perjury/accuracy by a third party (special master?) against the actual tape. Maybe it’s not the best solution but we need something that lets us treat our digital memories like our organic ones with respect to our control over their revelation.


  1. One might naturally worry about the friends and family members who commit a great deal of violent crime scheming to impersonate you to clear the incriminating information but if we always keep a sufficiently long buffer so as to make the failure to report us missing during that time suspicious we can at least minimize the risk. 

  Category: Law
  Comments: None

Social Justice Extremism Comes To Mathematics

Math Isn't Immune To Ideological Censorship

Even after reading the counterarguments it seems clear that part of the mathematical community suppressed a mathematical [article] (https://arxiv.org/pdf/1703.04184.pdf) because it was ideologically inconvenient. As detailed in the quillette piece a paper accepted for publication (and actually published online) at the Mathematical Intelligencer was yanked from publication because some mathematicians felt it’s mathematical modeling of the greater male variability hypothesis might be discouraging to prospective female mathematicians and/or be picked up by conservative media.

Obviously (if the account is close to accurate) this is a horrific violation of academic norms. If there is anything that academic freedom means it’s that ideas can’t be suppressed because they might support politically inconvenient conclusions and it appears that is exactly what happened. What makes the whole situation truly absurd is that the people complaining about the paper were the ones doing the real damage to gender equality. Obviously, one unheralded paper entertaining a greater male variability hypothesis will have a lot less of a harmful effect than even the chance of a controversy over pulling it and the subsequent attention and overreactions1. Not to mention that explaining the gender imbalance via bias and harassment is far more discouraging to potential female mathematicians than invoking a biological difference in male/female variability2.

Some mathematical blogs critisize the argument in the paper (see back and forth in comments) but don’t allege it’s absurd or incoherent. A more substantial counterargument is made in the comments at ycombinator where it is suggested that a rogue editor deliberately pushed the piece through and directed changes to make it an overtly political piece supporting his own views. However, looking at the actual [article] (https://arxiv.org/pdf/1703.04184.pdf) reveals a piece perfectly appropriate to the Intelligencer or indeed many peer reviewed journals (perhaps excepting the relatively unspecialized nature of the content)3:. Moreover, if the editor had truly gone rogue rather than just having made a decision the board disagreed with she would have been summarily fired.

Of course, at this point it’s just one isolated incident (and new facts could always emerge to change the narrative) but it’s critical to express our disapproval now so this doesn’t become acceptable behavior. Ideally, I’d like to see some kind of resolution or statement of principles at an AMA meeting so hopefully we don’t need to move to a boycott or anything. Of course, political and ideological considerations do affect what gets published all the time but it’s critical to reinforce the norm that such considerations shouldn’t matter which makes our response in the few clear cut cases like this all the more important.

What to do about the original paper is more of a puzzle. I’m tempted to say that some other periodical should publish it merely to avoid giving even an apparent victory to the forces of censorship. However, there aren’t many equivalents to the Intelligencier and maybe it really isn’t a great model (though hopefully peer review could tell us). Also, at this point publishing the paper really would send a very different politicized message that other periodicals would be understandably reluctant to send. So other than moving quickly to the end of the journal format I don’t really know what should be done about the paper.


  1. I do not expect many women inclined to fight through group theory and differential geometry (as well as any gender specific barriers) to be deterred because some mathematicians aired an idea about gender differences in a journal. 
  2. Moreover, prospective female mathematicians have no reason to believe that conditional on being the sort of woman even considering going into math they are any less likely to be good than their male colleagues. Indeed, after conditioning they might expect to be better if it is true boys are more encouraged to enter these fields as well. 
  3. Indeed, the inclusion of the appendix citing support for the greater male variability hypothesis seems necessary to insulate against criticisms that it’s perpetuating an unacademic, unsupported talking point. 

A Liberal Case Against Roe v. Wade

Why Roe Skepticism Doesn't Make One A Judicial Extremist

So I think that a right to have an abortion isn’t just a good idea but a vital component of maintaining a modern, egalitarian and free society. It’s important enough to our way of life and makes so much difference in outcomes that it should be enshrined in our constitution with other fundamental principles like free speech or freedom of religion.

Problem is that it wasn’t. Certainly not explicitly. Since I believe that it’s an appropriate function of the US supreme court to expound new legal theories, rights and remedies (as part of the great tradition of common law as the founders intended) in itself this isn’t a problem. The court certainly could have validly developed a theory of personal autonomy which coherently marked the right to receive medical treatment to terminate a pregnancy as protected in much the same way as it has coherently developed the theory of personal autonomy interests in people’s private sexual liaisons (accepting the consequences even when it reaches results like protecting fundamentalist mormon derived polygamy despite the unusual partisan valence).

Unfortunately, the court simply hasn’t done that. If banning medical treatments that would let someone terminate a pregnancy implicates personal autonomy then banning medical treatments that might let save me from dying of cancer must do so to an even greater degree. Yet the court does not recognize a constitutional right to try medication that hasn’t been approved by the FDA even when the approval was denied merely because the FDA judges the harmful side effects too severe or the evidence for efficacy too weak or not yet fully presented. In other words even when the government merely thinks my fully informed (and not mislead into fakery) judgement about what risks to take or harms to endure for a chance at a cure is unwise or ought not to be encouraged this isn’t seen as infringing on my right to personal autonomy. Even though those who want to ban abortive medical services are similarly expressing a value judgement about the relative merits of the consequences of doing so.

Nor is this some narrow exception justified by the compelling governmental need to incentivize efficacy trials or exclude quackery. Even when the government bans use of Marijuana as medicine the court doesn’t require any showing from the government that it’s not denying citizens the personal autonomy to effectively treat their medical condition as it would if it felt that any true constitutional right were implicated. It might not fall on identity based fault lines but surely being forced to lose one’s sight to glaucoma is rather than take an intoxicating medication, if not on par with having carry a child to term1, still an undeniably substantial infringement on personal autonomy. The court isn’t any more sympathetic when individuals seek to assert a constitutional right to drugs like peyote even when, as in Employment Division v. Smith, the individual is seeking access to peyote as part of a deeply felt belief in the religious principles of the Native American Church.

Intuitively, maybe you feel this is all silly. Surely none of this crap is important the way abortion is. How dare I suggest denying some wacky religious use of peyote is comparable to forcing women to carry a child to term. Indeed, I agree with you. These aren’t anywhere near as important and retaining women’s right to choose. But that’s the point. Constitutional protections for abortion don’t flow from any generally applied/developed theory of a right to personal autonomy or anything else. They flow from a value judgement about the importance (and correctness) of protecting a women’s right to choose. Even on a liberal living constitution theory of jurisprudence that’s not kosher. The supreme court is supposed to articulate general theories/conceptions of constitutional rights not weigh in individually on policy judgements simply because they happen to be really important or bad.

Having said all this I still believe that reliance interests favor respecting stare decisis when it comes to Roe. Personally, I actually believe they should respond by actually taking seriously the idea of a right to personal autonomy that is infringed by government restrictions on what medical treatments or drugs one can access. However, there is surely a strong argument that even someone who has a liberal judicial philosophy should react to the strong policy considerations that animated Roe and vote to reverse this outlying opinion.

Now, I don’t expect most people who read this to agree with this last conclusion. However, if you at least found these worries/arguments cogent, serious and requiring a substantive rebuttal you certainly can’t join the howling masses in suggesting that an unwillingness to uphold Roe v. Wade is indicative of a right wing judicial extremists. If you really believe that the Senate should confirm any academically/judicially qualified candidate without disqualifying extremist views then you can’t simply insist that Kavanaugh must be such an extremist because he happens to hold a non-crazy position that you’re convinced is not only really wrong but also really important.

It’s still perfectly reasonable to think that abortion is important enough that the Senate shouldn’t confirm anyone who will vote to overturn Roe. I hold such a position myself. But let’s not confuse that straightforward disagreement with the claim that somehow the conservatives have violated a norm of judicial non-extremism. Just admit that some principles can be important enough to vote against nominees who don’t support them no matter how reasonable or even mainstream (relative to the political landscape) their views are.


  1. Note that even if it turns out no one actually is faced with that choice, e.g., the commercial drugs now manage equal effect, the point is that the court obviously wouldn’t even require the government to demonstrate such a point to uphold the restriction of Marijuana. 

Definitional Creep and Psychological Salient

Recognizing Harms Without Eroding Our Terms

This nytimes article made salient a certain harmful dynamic that often occurs in liberal circles.

There is some class of behavior like domestic abuse or sexual assault which has a relatively clear cut and highly salient definition, e.g., domestic abuse (unmodified) is physical violence by an intimate partner or family member and sexual assault is sexual touching despite a clear lack of consent.

Then people (correctly!!) observe that some related behavior which doesn’t quite fall under the existing definition can be just as bad and harmful and start using the existing word to describe that behavior. For instance, in the article below using digital smart devices to harass, intimidate and stalk ex-partners and calling it domestic abuse. Or calling situations where an older man puts a young impressionable woman in a situation she feels uncomfortable and unsafe and applying substantial pressure (or legal but unsavory social threats) to coerce her into sexual activities sexual assault.

Individuals who then protest that such behavior isn’t domestic abuse or sexual assault are often accused of not taking the suffering of such related behavior seriously and this creates a strong social pressure to expand the category. When those on the right stand up and mock such expansions as not really being domestic abuse or sexual assault it further cements the narrative that this is all about whether or not one takes the suffering of victims seriously.

This is unfortunate because measuring harm is not the only, or even primary, purpose served by such categorizations. Rather, such categorizations serve to clearly delineate certain kinds of blameworthy activity which can be clear adjudicated to provide deterrence (legal or social) as well as creating a clear, psychologically salient bright line which people will be reluctant to cross.

For instance, while having someone grab your ass without permission or tear off an item of clothing is obviously sexual assault it is almost surely less harmful and traumatic than being tricked into a sexual relationship with someone based entirely on lies and finding out they were mockingly journaling the entire course of their deception online. Yet the former is sexual assault, in part, because we can clearly define the offense while it is hard to draw useful lines about how much deception in a relationship or how much sharing unflattering details with friends justifies punishment. It’s deeply unfortunate but the sad truth is that the lack of clear standards means we can’t use punishment to deter all kinds of deeply harmful behavior.

When we use the fact that some kinds of non-violent behavior perpetrated by intimates are just as harmful as physical abuse or that some types of technically consensual relationships have, via lies or pressure, the same kinds of harms as sexual assault to justify expanding the terms we erode the deterrent barriers that make negative behavior less likely.

Yes, there are ways to hurt people we can’t yet deter well but every time social norms and fear of legal consequences cause someone to refrain from hitting their spouse or touching someone without permission that’s a victory. The more we blur those lines that are being crossed the more we imperil those victories. This doesn’t mean one shouldn’t try to introduce new terms that characterize other classes of behavior we also wish to deter but there is no cause to undermine the terminological clarity of the words we already have. Respecting such terminological clarity doesn’t deny the suffering of victims who don’t fall under the description. It’s just a recognition the world is far from perfect and we need to preserve our partial victories.

Stop Calling Subjects Ethically Fraught

It's An Excuse Not An Argument

Listening to the Last Week Tonight on Gene Editing (it’s pretty good) and seeing this debate about paying organ donors I’m compelled to call out the practice of simply asserting that something is ethically fraught or troublesome.

Both with respect to not compensating organ donors (something which could save huge numbers of lives) and with (mostly prospective) limits on eliminating genetic disease or even barring improvement I think we let people who are simply uncomfortable with change off the hook by constantly repeating the supposed truism that the issue is ethically fraught or there are serious ethical concerns. It’s basically a free pass that excuses the fact that they are putting their discomfort ahead of people’s welfare.

Under all the scenarios/conditions seriously being considered No, there aren’t ethical concerns. Fears like letting a bank reposes your kidney are no more relevant to the proposals on the table than the fear that debtors will enslave people is to wages. Similarly, concerns about racially motivated eugenics programs have no plausible relationship to any kind of gene therapy even being prospectively considered.

Of course, we should hear potential concerns about such policies just like we would for any other policy/technology. However, opponents should be on the spot to either shut up or come up with compelling arguments suggesting harms. Based on the fact that the opponent in the WSJ to paying for organ donation is reduced to arguments like “The introduction of money for a precious good comes at the cost of the ability for one to aspire to virtue” makes me doubt they can come up with such arguments.

I’d add that I think philosophers are partially to blame on this point. As a matter of philosophical interest we correctly find clever new arguments seeking to show that paid organ donation is actually somehow problematically coercive or otherwise wrong more interesting than the obvious argument that it saves lives. However, just as physicists need to convey to the public that the very thing which makes theories which deviate from the standard model interesting also makes them less likely I think philosophers need to do this as well.

How to Provide Better Incentives to Organ Donors

Three experts discuss strategies to address the shortage of organs available for people who need transplants.

Repudiation Of Partisan Inference

No I'm Not On Sara Huckabee Sander's Side

I felt it was worth another post just to clarify that my [last post] (https://rejectingrationality.org/blog/gay-cakes-and-conservative-hens/) doesn’t indicate that I’m joining up with Sara Huckabee Sanders and Trump in this recent civility fight. Their behavior in calling out incivility and trying to use it as a political weapon (and lying about it) is worse than anything the Red Hen did to her and these are small potatoes relative to the normal harms they inflict.

I just tend to see them as beyond the moral pale and there is no use to excoriate them once again. And just because they are bad actors isn’t any reason for us to lose sight of the fact that denying people service is a big deal even when it’s merely based on partisan affiliation and we need a good justification to do it. Nowhere near as good a justification as separating families at the border or enacting a Muslim ban and the like but we don’t have to sink.

Gay Cakes And Conservative Hens

No, It's Not Ok Because It's Partisan Identity

So in all the discussion about denying Sara Huckabee Sanders service at the Red Hen I’ve seen several places argue that it wasn’t a big deal to deny her service because they were very polite about the incident. While it’s surely better to be asked to leave politely this fails to grasp what’s upsetting about it and is completely inconsistent with our attitude about the harm in Masterpiece Cakeshop where the customers were denied service because they were gay1. Whether or not you think it’s warranted any honest analysis should admit that it is still likely very upsetting and, should it spread, has the potential to be a very big deal.

Don’t get me wrong. The long history of discrimination against homosexuals and their status as a vulnerable minority in much of the country makes it much worse to deny someone service based on sexual orientation. Also, I firmly believe it should be legal to deny someone service because of their partisan affiliation2. However, just because it’s less bad doesn’t mean it isn’t still plenty unpleasant. Note that I also think Sara Huckabee Sanders is a pretty awful person and spokespeople are surely different than random Trump supporters but I’ve seen many people defend such partisan denials of service generally not just for high ranking officials and that is what I feel deserves comment.

Humans are acutely sensitive to exclusion. It’s bad enough not to be invited to a party or know that you aren’t attractive or cool enough to be waived into a club. It’s much worse to be overtly told you don’t belong and asked to leave. This is especially true when you are asked to leave a place of public accommodation such as a restaurant and doubly true when one is there with family (especially children who might not understand).

Of course there isn’t the same history of discrimination as with race, sexual orientation or religion but surely the immediate emotional response of feeling hurt, confused and angry will share a fair bit in common. If you are skeptical imagine a Bedouin who has lived his entire life out of the reach of modern technology or history on his first visit to the United States who is politely asked to leave a restaurant because of his race or his religion. Do you doubt the very real pain and humiliation that will cause — even though that diner’s lack of contextual knowledge means that our fraught racial and religious history can’t play a causal factor?

At this point I’m sure that many liberals are marshaling reasons why it’s not at all similar. Partisan affiliation is a belief not an immutable characteristic, they aren’t a minority almost half the country is on their side, it doesn’t have the same potential to normalize harmful discrimination.

But religion is every bit as much a belief people choose (in both cases many people simply believe what their peers do) and people do identify very strongly with their partisan affiliation. Whether or not it is objectively different in some sense being excluded for partisan affiliation feels like being excluded for who you are to many people.

It’s certainly true that Trump supporters are hardly a minority in America but they are very much a minority in many areas (including large parts of DC). It’s also true that there are large swaths of the country where there is strong, vocal support for homosexual equality but we don’t feel that merely knowing that this would never happen in San Francisco obviates the harm to the couple denied service in Masterpiece Cakeshop. Nor would this intuition change if 40 of the 50 states were strong and vocal defenders of homosexual equality. Besides, you don’t really think the question turns on the fact that Trump supporters are 40 some percent of the population rather than 15% (the percentage of African-Americans in the US) and you’d surely never accept the response in the racial case.

Moreover, in terms of normalizing harmful discrimination, it is surely far more likely that (as many people openly suggested following the incident) people wearing MAGA hats will be denied service in a wide range of establishments in blue parts of the country sometime in the near future than homosexuals will face similar widespread denial of service (if for no other reason than the stronger norm against such discrimination). Note that if you want to argue this isn’t truly harmful discrimination you need to challenge one of the points above about this being a very painful and unpleasant experience.

Again, it’s not as bad as discrimination based on sexual orientation but we’re being hypocrites if we don’t recognize that it’s still a pretty intense and humiliating experience which we should avoid inflicting on people unless we have a very good reason.


  1. Ok, technically because they planned to use the cake at a gay wedding. 
  2. I’m unsure about legal protections based on gender identity since that turns on tough questions about how fast norms are changing versus the long term harms from such legislation but this uncertainty doesn’t stem from any doubt about the degree of pain imparted by such denials of service but merely from concern about the long term costs of such legislation and the hopeful belief that such denials of service will quickly become the kind of silly extremism (like refusing to sell to people with bare knees in the US) that is so rare as to no longer be seen as a threat. 

Civility Without Concession

We Can Be Civil Without Letting Trump Tell Us What That Means

I’m really disappointed by the responses from the left to the recent kerfufle about Sara Huckabee Sanders being denied service in a red hen restaurant. These responses all seem to have the following structure:

  • If we let those in power define what is civil and uncivil we allow them to silence protest.
  • Historical protests like the civil rights movement we now revere were often seen as uncivil.
  • Trump and his supporters are deeply uncivil and have no respect for these norms.
  • Therefore we should accept or even cheer incivility towards Trump voters.

But this makes two big logical errors.

  1. Just because we shouldn’t let Trump and those in power dictate a self-serving standard of civility doesn’t mean that the concept itself is flawed or that it’s an inappropriate ideal to aspire towards. It just means we should apply the concept correctly not let those in power dictate a self-serving conception of civility.
  2. It conflates the desirability of showing civility and respect with the appropriateness of criticizing people for not doing so. There are plenty of things, e.g., chewing with your mouth open, that it’s both inappropriate to do as well as to call others out for doing.

Now a valid takeaway from this argument is that overly mechanistic rules of civility (e.g. never mention a politician’s personal life) shouldn’t be followed blindly but that’s not the same as the conclusion that it should be seen as appropriate or even desirable to simply abandon civility all together.

Just Don’t Be A Dick

The right answer here is pretty simple. All we need to do is treat people, even those whose political views we see as harmful, as if they are human beings who we, ceterus parabus, don’t want to suffer.

This doesn’t mean we place their feelings ahead of the welfare of the country. If the next democratic president finds the best way to respond to Trump’s personal attacks is in kind then they shouldn’t let their concern about hurting Trump’s feelings stand in the way (though I suspect the voters who aren’t already deep in Trump’s camp are more likely to be swayed by statesman like restraint). If the same congressmen and staffers who are working to undermine a woman’s right to choose are having abortions themselves we shouldn’t let the personal impact stop us from exposing such facts.

What it means is that we shouldn’t be mean to them or refuse them service just because it makes us feel good or because they somehow deserve it. Refusing to serve Ms. Sanders doesn’t score any political points, it risks losing them. Even those on the left skeptical that such behavior will galvanize Trump supporters admit it’s a possible risk while identifying no corresponding benefit.

The right definition of civility doesn’t require rolling over and showing your belly it merely demands we refrain from needless petty antagonism and rudeness.

However, just because that’s what we aspire to doesn’t mean we should excoriate those on the left who fall short. That risks giving the misleading impression that their behavior is somehow worse than that of the administration. It isn’t. Ms. Sanders behaves plenty badly herself and any discussion of the restaurant’s choice not to serve her should keep that firmly in focus. However, just because she behaves even worse doesn’t mean we can’t urge our own side to take the high road and avoid unnecessary rudeness and incivility.

Real World Applications

Now this still leaves open two big questions.

  1. When are we simply being rude and mean to satisfy our emotional desires as opposed to engaging in a necessary form of protest which unfortunately might hurt some Trump supporters.
  2. At what point does the harm from violating norms about civility outweigh the short term tactical gains from some kind of protests that might be seen as uncivil.

While these are hard questions in the abstract in practice I think they are actually pretty damn easy.

Regarding 1, I think deep down we know the answer in 99% of cases if we just take a step back and think about it. Just imagine the person you are about to be rude to or insult is an old friend from High School who fell in with a bad crowd in college. Would you still do it? Or better yet imagine the person you’re attacking is actually a liberal from a parallel universe who just switched places with their double in this universe. Just put yourself in a position where you see the harm you’re doing to the Trump supporter as a negative (like it would be for anyone else) and ask if the benefit outweighs the harm. If your too emotionally involved to step back like this your probably also too compromised to know if you’re actually helping or making the situation worse so you probably shouldn’t do it anyway.

Point 2 is a serious concern when it comes to civil rights movements demanding recognition for groups that society has seen as unworthy. In such cases there is a real tradeoff between the moral potency and enthusiasm offered by making it clear that the equality or moral worth of your group isn’t up for debate and engaging in civil dialog with norms that call that into question. For instance, advocates of LGBT rights had (have) to choose between being seen as being deliberately provocative and uncivil by forcing their sexuality into the public square in ways that were seen as inappropriate and putting it in people’s faces even though the analogous heterosexual conduct wasn’t. One choice risked jeopardizing the support of moderates while the other risked undermining the momentum and moral force of the movement,

However, in the Trump context it’s Trump supporters who are pushing aside existing cultural norms while it’s the left who are desperately trying to defend (what was previously) the mainstream. When Trump’s trying to break up families at the border, eliminate Obamacare or undermine Roe. v. Wade we don’t gain anything by pushing positions which make mainstream Americans feel insulted or uncomfortable. Now isn’t the time to suggest, however true it might be, that normal everyday behaviors make one a racist or transphobic. It’s time to talk about how taking children from their parents is wrong and unamerican and that the rich don’t deserve more tax cuts and in this context there’s just no real payoff to insulting or alienating the moderates we need to persuade to stop Trump.

Wrong About Me Too

Not Ideal But I Was Way Off

In an earlier post I was very critical of the form the me too movement took predicting that simply conveying me too in the absence of any clear agreement on what was being claimed (did someone once ask you out in a way you found creepy or did your boss threaten to have you blacklisted from the industry if you didn’t sleep with him) would do more harm than good. I didn’t believe that merely expressing vague pro-social feelings of unknown intensity on social media would do much to persuade anyone to change their ways nor would it do much to change anyone’s justified belief in the prevalence of any particular kind of behavior. To be clear, I always supported the goal of eliminating the impunity with which many powerful men harassed I just thought it would require sharing emotionally detailed accounts of actual harassment to shift people’s beliefs.

Now I still believe I was right about many of the unimportant issues. MeToo didn’t directly do much to change people’s estimates of the incidence of harassment and I do think that the temptation for those who wouldn’t have otherwise judged their experiences as qualifying to participate (with the best of motives) might have lead some people to see harassment as less serious1 but all these effects were trivially small if real at all and I totally missed the real importance of the movement it created common knowledge that there was widespread condemnation of sexual harassment and that many people were finally willing to take accusations against powerful and admired figures seriously. Note, that conveying common knowledge didn’t require shifting belief in any particular frequency (of incidence or of people willing to stand up) and I totally missed this possibility[^facebook].

Was MeToo the optimal vehicle to accomplish this? I don’t know but it seems to have worked quite well and there is a good chance any little tweak I might have preferred would have undermined it. So I was totally, extremely wrong about all the parts that mattered and I believe in owning up to that. Indeed, I even vaguely remember someone mentioning the theory I endorse here on facebook at the time but I wasn’t convinced.

In retrospect I let my annoyance that people seemed to be just doing things because they felt emotionally satisfying rather than having any particularly good justification biased me so even when someone made the right argument I missed it.


  1. For instance, someone assumed that those posting mostly just to be supportive were the extent of harrassment with rare exceptions.