The Halo Defense
Once upon a time, A Relatively Famous Guy On The Internet was accused of having been simultaneously dating multiple women, without those women’s knowledge, those women (according to the accusation) having been convinced that they were his exclusive partners all along. Then, one of his friends, another Relatively Famous Guy On The Internet, defended him by claiming that he is “a great friend, a wonderful human being, and that his podcast was incredibly helpful to a lot of people”.
This is a great example of what I came to call the “halo defense”.
The halo defense involves “defending” someone from allegations by bringing up traits of X that are meant to be interpreted positively, without addressing the allegations on the basis of which X is being “attacked”.
I’ll give two more examples from my personal experience:
A university professor (with a PhD in ethics!) is accused by several of his female students of sexual harassment. The dean demands that the faculty staff take the professor’s side, because he is a “vital part of their community”, not engaging to any extent with the validity of accusations.
A (different) university professor is afflicted by a strange curse: students express little to no interest in his lectures. He kinda threatens his students that if they don’t start regularly attending the lectures, he will not let them pass the exams (where he technically doesn’t have the right to deny them passing the exams). His colleague, during her own class, tries to give the same students reasons to attend the lectures that have nothing to do with their intellectual value.
The halo defense seems to occur not too infrequently. For example, it is one of the main drivers of community disputes. I haven’t seen it named yet, and the name “halo defense” fits perfectly.
It is a special case of the halo effect:
the tendency for positive impressions of a person, company, country, brand, or product in one area to positively influence one’s opinion or feelings of a person, company, country, brand, or product in another area”
of the noncentral fallacy:
X is in a category whose archetypal member gives us a certain emotional reaction. Therefore, we should apply that emotional reaction to X, even though it is not a central category member
and of the affect heuristic:
Finucane et al. found that for nuclear reactors, natural gas, and food preservatives, presenting information about high benefits made people perceive lower risks; presenting information about higher risks made people perceive lower benefits; and so on across the quadrants.5 People conflate their judgments about particular good/bad aspects of something into an overall good or bad feeling about that thing.
The common pattern underlying all of them is a spillover of evaluative judgments made on one axis (e.g., “helpful to people”, “efficient”) to all other axes.
PS: A phenomenon related to the halo defense is that people seem to be insufficiently aware of the fact that many persons’ distribution of behaviors is long- and heavy-tailed, and thus extrapolating the behavior observed in most circumstances to all circumstances may not work. Relatedly, the typical mind fallacy leads people to reasoning in terms of (something like) asking themselves a question “Under what circumstances and modifications of my own mind, could I imagine myself doing such a thing?”. Upon this question returning a null result (“Never. Under no circumstance or modification of my own mind.”), they come to conclude that this allegation cannot be true.
How much is this our brain doing lazy reasoning, and how much is this strategically correct reasoning under cultural constraints.
E.g. if I admit that I believe allegations against A, then I the norms of our culture demand that I must stop associating with A. But if A has a lots of good qualities, then the cost of stopping associating with them is high, so I might want to take that into account.
I.e, the debate that is superficially about [are the allegations about A true] is actually about [should we kick out A], and most people know this on some level and act accordingly.
If this is what is going on, then the only way to stopp this “fallacy” is to change the incentive some how. This would include making it common knowledge that after we find out the truth of the allegations, there is a second step of waying the pros and cons of having this person around. But that can get into very taboo territory.
I’m not actually sure people are miscalibrated here, at least for close friends? If someone accuses my best friend of murder, I really do think that they are much more likely to be lying than my friend is to have committed murder (barring exceptional circumstance like self-defense). This is entirely because of my judgement of their character, so that “when would I commit this crime?” really does tell me a lot about when they would.
People who are generally honest, law-abiding, and moral are in fact less likely to lie, break the law, and violate common morality.
It’s seems likely that people are miscalibrated about weaker links, though, due to the reasons you’ve cited, the bonds of friendship, and the way that ill-doers will act differently (consciously or unconsciously) around people they think they can get away with harming.
I’ll add to this: I think generally someone people kind and moral in one circumstance is Bayesian evidence of them being so in other circumstances, AND there are specific patterns of being sometimes kind and sometimes nasty that most people are under-aware of.
You have a good friend. They are always happy and cheerful. Once they hurt you by accident, but they sent you a cake the next day, and you forgave them before you could ask for an apology. They’ve had a couple people be mean to them in the past, and you feel sorry for them, but there’s been nothing but cheer between you.
The previous paragraph includes three traits of Narcissistic Personality Disorder (constant happy exterior, niceties instead of apologies, believing a story in which they’ve always been the victim), and thus the previous paragraph is in fact Bayesian evidence that they may be horrible to other people.
Relatedly, if someone you know is accused of acting poorly toward another person, the observation that they never act that way toward you may provide nearly zero Bayesian evidence in their favor.
A very common example:
Alice: “My boyfriend is controlling and abusive toward me, and disrespectful of other women.”
Bob: “I don’t believe that. He never acted that way toward me!”
Alice: ”...Yeah. Duh. You’re not a woman.”
A related example from my own life:
Bob: “You should be wary of Charlie.“
Alice: “Why? He’s only ever been nice to me!”
Bob: “Yes, and he’s also only ever been nice to me, but many of our friends report him being bad to them AND being horrible to people who he seems to consider ‘beneath him.’ I trust our friends, and so I’m worried about the possibility that Charlie is specifically kind when he’s around us.”
Alice: “Charlie isn’t the type of person who would do that. I can tell.”
Bob: “…”
I think in general people are vastly uncalibrated (in both directions) of the degree, and in some cases even the direction, of positive correlations/positive manifolds of different kinda-normal behavior with extreme behavior.
As an example of something with a superficially opposite moral of your post, I’ve seen multiple people confuse meta-honesty (of the form where someone will cheerfully tell you they’re lying to you/lying to other people) with object-level honesty.
I think the implicit mental motion is that if somebody’s telling you they’re lying to you about X, they’re less likely to lie to you about things other than X. Or if somebody tells you they’re lying to other people, you’re special and they won’t ever lie to you, would they? Or something like “all politicians/CEOs lie, at least this dude’s honest about it.”
Whereas I much more have the view that any lie you see is the tip of the iceberg, and somebody cheerfully telling you how much they lie is probably positively correlated with the number of lies you don’t know about.
Relatedly, “figuring out whether someone’s a psychopath” is one of the situations where I trust empiricism and ordinary scientific reasoning over vibes, even for social questions.
Great post. I’m curious about generalizations / other things in some class. Feels like there’s a whole toolbox for “aggressively not answering the question”. Another example besides the Halo Defense is agreeing on an abstracted claim (while not addressing the particular claim). E.g.:
A: “Do you agree that Charles generally does not carry his firearm, and that on the morning of the 9th he put his firearm into his backpack?” B: “I totally agree with you that murder is wrong. It’s wrong when Republicans do it, and it’s wrong when Democrats do it. And we can definitely discuss these things, and as I’ve said before, and I’ll say it again, if someone commits murder then he should go to jail.”
Part of how it works is simply distraction / filibustering. But also it kinda bends the Gricean implicature, where you’re kinda compelling the discourse to have been about the abstracted claim. Sometimes it works on A; even if it doesn’t, it might work on some of the audience; and even if it doesn’t, it can give B plausible deniability, where an audience member might be like “well the conversation just didn’t get clarified”, as opposed to if B more directly said “uh no comment” or something.
You should probably disagree with me here, but that’s exactly what I mean when I say “x person / group selects on vibes (or rather, someone else’s testimony of someone’s “vibes”)” in the context of not having a clear and transparent selection policy.
I guess halo effect may be something that affects judgment and not necessary for “vibes based selection ”. But I’d find it helpful to clarify my thoughts.
I mean, to the extent that people are selecting based on “vibes” but are not aware of it and actually think that they’re selecting based on something else, this is totally a halo effect thing.
So it is basically a failure to decouple the values when they diverge?
IDK how broadly you tend and intend to use the term “value”, but for me, applying it here is quite a stretch.
If you’re measuring the performance of a program, and notice that it’s very fast/time-efficient, and consequently conclude that it must therefore also be space-efficient (use as little memory as possible), and elegantly written, then you’re definitely confusing various axes of evaluation — i.e., the thing I pointed at in the part of the post you’re replying — but it seems like a big stretch to call time-efficiency, space-efficieny, and source code elegance three distinct values that are being confused here.
Oh, I didn’t mean value as “measure of worth” but value as in the math definition of “amount denoted by reference”.
The “halo defence” is not, I think, an attempt to actually state that factually the outcome is less likely or has not happened.
It is, rather, that X is not important enough to warrant focus/punishment because of the positive effects of Y.
It is socially costly (basically, negative EV) to be publically seen to claim that X is somehow not a problem, hence the deflection.
To use an absurd example—imagine that my neighbour is a unique person in the world with some sort of superhuman ability to cure cancer or do incredible research that no-one else can do with really high expected value for society. He also may or may not have murdered a few people.
It is less costly socially to argue, even if everyone knows that this isn’t really true, that he probably didn’t do the murders, because he’s such a good guy in other areas, as opposed to arguing the utilitarian “well yeah, he’s a murderer, but he’s saved 5000 people from cancer so like, whatever”.
People could be useful pillars or functionaries in a community, and losing those pillars would come at great cost to the community or group. Isn’t it reasonable for the group to weigh the cost to remove vs the cost to look the other way/ privately censure/ or warn with more leniency than someone less important (Positive and negative costs of looking the other way/ private censure/ lenient warnings all considered)?
My guess is this kind of calculus is happening at a half-conscious level, but it comes with feelings like “we can’t lose <person x>.”
And I think groups really can schism or collapse over stuff like this, either via the argument about what to do, the visibly and costly holding to the same standards (and thus maybe losing a core leader), or the letting them get away with it. There may be no clean heuristic answer here.
As one tangential example, I still like MIT Opencourseware Physics by Dr Lewin, for example, and as of a few years ago still found his classes hard to replace. Yet he was found to have done online sexual harassment. What would we do if say Feynman was found guilty of that or worse? Could we really replace his lectures? Would we be able to recommend them to a niece in good conscience?
As I said, I doubt there are many easy answers down this road.