Discover more from Good Thoughts
Text, Subtext, and Miscommunication
The Risks of High Decoupling, from Beckstead to Bostrom
People vary a lot in how they think and communicate, which can easily lead to miscommunication and imputing attitudes to others that they do not actually hold. One especially striking dimension of variation concerns what we might call literal-mindedness, or decoupling, whereby high decouplers focus more on the narrowly literal truth of a claim (the “text”), and less on what it pragmatically implicates or socially communicates (the “subtext”). Low decouplers, by contrast, are much more sensitive to subtext or the social implications of an assertion. To roughly characterize the two extremes, you can imagine a high decoupler as exclusively concerning themselves with the question, “Is this proposition literally true?” while a low decoupler evaluates a claim by instead asking something like, “If a politician were to publicly assert this, would they thereby be affiliating with the good guys or the bad guys?”
I’ve previously wondered whether utilitarian sympathies correlate with a high-decoupling cognitive style. Ironically, unsympathetic critics misread me as advancing an insulting subtext (misinterpreting “low decoupling” to mean something like “stupid”). So it’s worth noting that there are serious downsides to excessively high decoupling, such as the risk of neglecting important social implications. High decoupling may be a little bit like being on the autism spectrum in this respect, and may call for a similarly forgiving (rather than suspicious) interpretive stance.
For a classic example of why subtext matters, consider the slogan, “All lives matter.” This text literally expresses a true (and utterly innocuous) proposition. But, as we all know, it has come to implicate, and communicate, something rather different, namely a rejection of what’s communicated by the slogan “Black lives matter.” Presumably, most people who assert the former slogan know full well what they are doing, so it’s reasonable to impute to them hostility towards the Black Lives Matter movement. But it would surely be possible for someone to assert this claim innocuously, unaware of the subtext. And this kind of social cluelessness might be especially likely for an extremely high decoupler, whose style of thinking naturally tends to neglect subtext and (contingent, non-logical) associations. So while many are understandably quick to impute social attitudes on the basis of this sort of talk, it can be worth bearing in mind this alternative explanation, especially when dealing with analytic philosophers (who may be disproportionately likely to be neuro-atypical in these ways).
But also note that it’d be silly to object to a moral theory like utilitarianism for implying that, in fact, all lives do matter. The literal proposition can be evaluated independently of what it conveys to assert it in our current social context, and proper philosophical evaluation often does require such decoupling.
Moral truths can sound bad
Part of the reason I suspect that utilitarian sympathies correlate with high-decoupling cognitive styles is that it’s easy to find utilitarian verdicts that sound awfully discriminatory (see Theses on Mattering, 3 - 5). So if you’re extremely averse to affirming any claim that sounds like something a bad person might say (which is pretty much the definition of “low decoupling”), you’re obviously going to think utilitarianism is utterly appalling. But you’ll also be demonstrably mistaken about this, as the linked post explains. It turns out that “sounds like something a bad person might say” is not a great way to test the theoretical implications of a moral theory.
This is especially the case when it comes to instrumental valuation. Nick Beckstead has come under immense fire for writing in his academic dissertation that, due to the instrumental “ripple effects” on others, “it now seems more plausible to me that saving a life in a rich country is substantially more important than saving a life in a poor country, other things being equal.” You can’t coherently deny this without implicitly denying equal moral weight to the downstream beneficiaries of the positive ripple effects. So he was making a good philosophical point. But it sounds bad, because if a politician (e.g. Trump) were to assert such a claim, they would not be making this subtle philosophical point about the moral importance of indirect beneficiaries, they would be communicating disrespect for the global poor, and a desire not to aid them even when it’s orders of magnitude more cost-effective to do so. (In the real world, we know full well that other things are not equal when it comes to trying to save lives in rich vs poor countries.)
So, context matters. Philosophers are not politicians, and a claim in a philosophy dissertation does not carry the subtext that it would in a political rally. Critics and journalists are not doing anyone any favours by pulling philosophical claims out of context and evaluating them by political lights. They’re just misleading their readers, encouraging them to misattribute vicious attitudes where none in fact exist. It’s simple miscommunication, and the fault in this case lies with the critics (not Beckstead). It’s crucial for moral philosophers to be able to explore their ideas against a presumed background of good faith, rather than having to run their work past PR agents prior to publication.
Don’t be an Edgelord
While truths can sound bad, and some (like those in Beckstead’s dissertation) are objectively innocuous and assertible for good reasons, literal truth is not an excuse for communicating gratuitously hostile or insulting/abusive subtexts (which I take it is what people usually have in mind when calling a claim “offensive”).
There can be more or less offensive ways to convey the same literal information, by radically changing what subtext is communicated along with it. Compare:
(1) Amy says: “Rich kids on average score better on standardized tests.” Depending on contextual details, this could naturally convey the subtext that she thinks rich kids have an unfair advantage, perhaps due to private tutoring and test prep opportunities that other kids lack. By explicitly mentioning “on average”, Amy’s listeners are reminded that statistical generalities don’t justify stereotyping individuals: it would not be remotely surprising for an arbitrarily chosen poor kid to score better than an arbitrarily chosen rich kid (the odds could be anything up to 49.99%, for all Amy’s claim implies). This is an important corollary to bear in mind.
(2) Ben instead says: “Rich kids have higher IQs than poor kids.” When others object, Ben defends himself by pointing out that IQ tests are a kind of standardized test, and IQ is operationally defined by psychologists as “what IQ tests measure”. So his claim is almost certainly true if Amy’s is. But it’s an awfully misleading thing to say because it (contingently, in our social context) implicates the further claims that (i) rich people are inherently smarter, and (ii) every rich kid is smarter than every poor kid (or perhaps that there’s a strong presumption that any given rich kid will be smarter than any given poor kid). As a result, it also conventionally communicates disrespect and hostility towards the poor (regardless of whether Ben himself possesses such negative attitudes), and implicitly invites his audience to share those awful attitudes. That’s a jerk move, so it’s reasonable to criticize Ben for (even inadvertently) communicating such things, especially when these harms are so easily avoided (cf. Amy). (Of course, we might also want to push back against the conventions that lead people to think of intelligence as a marker of innate value or status.)
That said, it makes no sense to demand that Ben retract the explicit claim he made, since (again) the claim is literally true (if Amy’s is). Instead, I’d think he should apologize for (even inadvertently) communicating hostility, and disavow the implicatures (assuming that he does not really believe those).
Which brings us to Nick Bostrom, and his blatantly racist old email from 26 years ago that recently came to light. The “provocative” examples in that email are horrible to read, so I won’t repeat them here. Though I don’t believe that Bostrom intended them this way, such sentences constitute racist abuse: they communicate disrespect and hostility on the basis of race, which is harmful and unacceptable. And while Bostrom apologized for the racial slur that appears in his old email, I didn’t get the clear sense from his apology that he fully grasps this larger point (which is part of what makes it a lame apology, IMO).
But to add a bit of nuance to all the criticism being sent his way, I think it’s worth noting that there’s no evidence here of racist attitudes (either motives or beliefs) on Bostrom’s part. That’s not in any way to defend what he wrote, of course. Racist words can have racist effects, and that’s the larger issue here. But insofar as many are jumping to the conclusion that Bostrom is a racist person, I suspect that isn’t really warranted, and stems from reading in a subtext that he likely never endorsed. (I could be wrong about this; I’ve never met Bostrom, I’m just judging this based on my general sense of his philosophical thinking, and my sense of the kinds of errors that high decouplers could easily fall into.)
Some people have wondered how Bostrom could write such racist things if he’s not himself a racist. But the first sentence of his old email makes this fairly clear, I think:
I have always liked the uncompromisingly objective [sic] way of thinking and speaking: the more counterintuitive and repugnant a formulation, the more it appeals to me given that it is logically correct.
It sounds to me like he’s talking about the classic edgelord thing of enjoying the transgressive tension between intuitive repugnance and (what he took to be) “logical” correctness, i.e. on a strictly literal reading, divorced of all subtext (and thus missing what a reasonable person would ordinarily take the claims in question to communicate).
Further, when the young Bostrom follows up his racist example by saying, “I like that sentence and think it is true,” I don’t think he means anything racist by this. It’s not that he likes the fact that some groups globally have fewer educational opportunities. He just likes the cognitive tension that his edgy sentence generates. Or, so seems the obvious interpretation to me, at least.
That’s not to excuse what he said. Lots of people like to be a bit provocative, at least in defensible ways. But if your “edginess” is coming from implicitly communicating hostility towards marginalized groups, then that’s objectively messed up, even if this constitutes miscommunication and doesn’t accurately reflect your actual attitudes.
To anyone else tempted by the young Bostrom’s “uncompromising” (i.e., actively misleading and offensive) “way of thinking and speaking”, I’d urge you to reconsider. Please don’t say horribly racist things. Don’t be “edgy” at the expense of marginalized people. The literal truth of what you say is not the only thing that matters. What’s more, you can more clearly communicate true claims by avoiding gratuitously edging into repugnance, since the latter obviously invites misreadings (again, compare Amy vs Ben).
Can’t we all just get along?
Despite my great love of philosophical argument and debate, I’m incredibly averse to interpersonal conflict (where the aggro is directed at people, and not just ideas). I really can’t stand it. So (like many) I prefer to avoid heated debates about race or gender, and I’m a little worried that the messy middle ground I’m trying to stake out on this one will just end up making everyone mad at me. (If so, um, please just forget I said anything, and go do something else you’ll enjoy more?)
Despite this, I wanted to write this post because I think it’s worth making clear that there’s no essential conflict between the two admirable values I see playing out in this debate:
(1) Opposition to racism, including unintended racism.
(2) Epistemic integrity and fidelity to the truth (including intolerance of “bullshit”, in the philosophical sense of truth-indifferent assertions).
I’ve seen some people express uncertainty about “what’s so racist about the old email”. I hope my above explanations help to make this clearer (esp. to other high decouplers, who are probably the only people to have this question). Others have demanded that Bostrom affirm things that seem inconsistent with the claim that globally marginalized groups are prevented from attaining their full potential. That seems an unreasonable demand, so I hope I’ve also made it clear why opposition to racism doesn’t generally require affirming falsehoods.1 (Though I agree Bostrom's apology wasn't great; Habiba's thread is illuminating on this point.)
Most of all, I guess I’m hopeful that the above can help turn down the heat on these debates a bit, at least on the margins, by helping to promote mutual understanding (including of how some conflicts here may stem from something as innocuous as differing cognitive styles). In particular, I think it’s worth bearing in mind the possibility of a high decoupler making blatantly racist statements without holding any racist attitudes whatsoever. (People are weird! And stupid, sometimes.) Such inadvertently racist statements are still harmful, so—I fully agree—it remains worthwhile to discourage them. For many practical purposes, the effects are more important than the intent. But insofar as you’re interested in evaluating the person, their actual attitudes surely do matter.
And whereas Brian Leiter confidently asserts that “People who aren't creepy racists (or mentally ill) don't express such views in their 20s,” I guess I’d like to urge a bit more understanding of cognitive diversity (which, again, is not to suggest that the racist statements should be tolerated). High decoupling, especially when taken to extremes, is plausibly a form of neuro-atypicality. But it has advantages as well as disadvantages. So I’d also urge people to steer clear of the stigmatizing language of “mental illness” in relation to this kind of cognitive difference.
A Closing Lament
It’s obviously unfortunate that the young Bostrom wrote that dumb email back in the 1990s. Perhaps less obviously, it’s also unfortunate that it has now come to light (due to, it sounds like, someone digging for dirt). I don’t think it’s informative or useful in any way. Many have found it (understandably) anger-inducing or otherwise upsetting. As a result, the news may have caused more concrete harm than the original email did. Dirt-digging can be a harmful activity.
It could be worth the cost if there was some compensating gain, protection, or moral enlightenment to be had, but I’m not seeing much. In particular, if my interpretation of what motivated his younger self is correct, then the episode doesn’t actually reveal all that much about Bostrom’s moral character. Contrary to those who assume he must be bigoted (and, say, risks evaluating students unfairly), I doubt he poses any such threat. Nor would it seem any kind of cosmic justice for him to suffer severe career or reputational repercussions for his past mistake. The whole episode is just… sad and disappointing.
One broad upshot is that it could be helpful to raise awareness of high decoupling cognitive styles, and so help reduce unnecessary upset that stems from mistaking a kind of cluelessness for a kind of malice or other ongoing threat. (I assume these latter possibilities would be significantly more upsetting.)
But also, more fundamentally: how about everyone tries not to be that clueless. Even if you’re not bigoted—and I’m generally willing to believe the best of people—it is in fact a further matter to avoid doing racist things (like saying things that obviously encourage racism, even if that isn’t one’s intent). So: avoid it. That doesn’t seem too much to ask.
Comments closed: hosting further discussion on these sensitive topics sounds too stressful, sorry.
Nor does it require affirming propositions about which one might reasonably be uncertain or have no view at all. Just as opposition to homophobia shouldn’t be contingent on the (rhetorically useful but morally irrelevant) empirical claim that sexual orientation is innate, so our opposition to racial discrimination shouldn’t be contingent on empirical assumptions about genetics, IQ, or anything else.