Quassim Cassam wants you to know that conspiracy theorists have bad character. In other words, bad thinking is not just bad thinking; it’s also a vice. Maybe Cassam is right. Intellectual character or the lack thereof is often overlooked, at least in general conversation. It’s not that we have an overabundance of trust and tolerance in our public discourse, which is obvious to anyone, it’s just that we tend to see people who hold a multitude of unjustified beliefs as ill-informed, dim-witted, or maybe even insane. And while we easily find the nerve to accuse our ideological opponents of bad faith or insincerity, that’s mostly a purely moral accusation. Cassam’s idea is more interesting than that. The real problem with people who believe fervently in climate change denialism or 9-11 “trutherism,” is not that they’re dumb, crazy, or motivated by greed or power. Rather, it’s that they lack the intellectual character needed to form sound beliefs.
Think of an accountant who embezzles from their clients, a journalist who can’t find the courage to report the truth, or the employee who coasts on the job. These people are dishonest, cowardly, lazy. It may even be that these morally malformed creatures just can’t help themselves, such is the nature of their bad character. It’s the same with intellectual character. The belief formation process doesn’t go wrong on account of low intelligence or a merely cognitive failure of rationality, rather, what’s to blame is a bad intellectual quality in the person. It’s not clear whether this approach takes the terms psychologists and philosophers already use to name rational shortcomings and simply recasts them as character terms, or, whether it proposes alternative (morally-laden) terms and concepts with allegedly more fundamental causal power, but never mind. I actually don’t doubt that a non-trivially large portion of conspiracy theorists have at least temperamental characteristics that lead them astray – characteristics that can’t be reduced to or equated with bad faith, mental illness, or purely cognitive failings.
My worry is instead over how poisonous this approach could be to our discourse: How about this for a possibility: suppose, every once in a while, our views turn out to be wrong (heaven forbid), but instead of merely expressing our disagreement we accuse our interlocutors of possessing deep character flaws for disagreeing with us? Let’s take one example that’s particularly frustrating to me, the so-called “Hot Hand Fallacy” basketball players, coaches, and fans are said to employ (that this is a topic I would cite as frustrating as opposed to something important like, say, climate change, may reveal something about my character, but that’s a topic for another day).
In basketball, a player who hits many shots in a row is often said to have “the hot hand.” But statisticians have painstakingly pointed out that there’s no such thing. People just have a tendency to see patterns in randomness. Basketball coaches who, in the face of this evidence, still cling to their belief in the Hot Hand are accused by Cassam of having bad intellectual character; it’s not only that they hold a mistaken view about their sport. As for the Hot Hand Fallacy itself, there really is something compelling in the research of psychologists like Gilovich, Vallone, and Tversky, (hereafter GVT) who first identified the belief as a fallacy in the 1985 paper “The Hot Hand in Basketball – On the Misperception of Random Sequences.”
GVT cleverly got a bunch of survey participants to agree that after a player hits a shot or two, the player’s next shot has a higher percentage than normal of going in, and this belief in rising probability was debunked by their research. Further, it’s demonstrable that random sequences of numbers are indistinguishable from what appear to us to be hot streaks. In other words, random numbers sometimes streak as well – 10101111111111 might look like a pattern, but random sequences also behave this way at times. You could even call each term H or M (for hit or miss) rather than 1 or 0 for sake of maximizing the sense of similarity to shooting. Basketball fans have a hard time accepting all this. It’s not an exaggeration to say that conversations on the Hot Hand are often contentious. Tversky says “I’ve been in a thousand arguments over this topic… and I’ve convinced no one.”
The topic even found its way into our pop culture when former Clinton Treasury Secretary, Harvard President, and Winklevii naysayer Larry Summers paid a visit to the Harvard Crimson basketball team for a pep talk. Summers asked the players if they believed in the hot hand. The players all nodded. Summers then informed them that they were wrong, firmly explaining the fallacy, the content of which must imply that years of practice are incapable of building the muscle memory capable of getting a player in the kind of groove that can lead to a genuine hot streak – how inspiring.
Lo and behold, it turns out there’s new data. It purportedly shows that there really is such a thing as a groove, (and therefore the Hot Hand) and that the phenomena can be teased out empirically. The data are apparently persuasive enough that even Summers has given a nod, “Better data plus better statistical techniques means we’re going to understand the world much better.” Openness to new data is certainly the kind of virtue we should all aspire to, but before we break our arms patting ourselves on the back, new data doesn’t resolve the problem of ambiguity that’s lingered all along.
Ludwig Wittgenstein taught us to pay close attention to the way words are used. GVT got people to agree that a player with the Hot Hand who hits a shot would display a rising probability of success on subsequent shots, but again it turned out this phenomena couldn’t be justified statistically, so, fair enough. But to this rabid basketball fan of over 30 years, that definition sounds, while not altogether incorrect, still a bit tortured when presented as exhaustive, perhaps to say something the torturers needed to get on record in order to proceed. But again, never mind, I don’t want to quibble with what some portion of people out there think the hot hand is.
What no fan worth his vintage jersey would believe is that GVT’s definition represents the only common way the concept is used, and, indeed, if the researchers had been as curious about the subtleties in the way people use words as they were about statistical implications, we probably wouldn’t be having this conversation, as thoroughly clarified theses that narrow their headline claims tend to get less attention than sweeping ones. To be fair, GVT did allow that all the things people mean by “hot streak” might not be captured by their analysis. But over the years, the general notion of a hot hand has been said to be dis-proven by clever statistical techniques, and in any case GVT then pivoted back to their exhaustive definition, asserting without authority that however the terms “Hot Hand,” and “hot streak” are employed, the common ways all imply that “the probability of a hit should be greater following a hit than following a miss,” and that “the number of streaks of successive hits or misses should exceed the number produced by a chance process.” In other words, the common notion as defined by GVT happened to line up perfectly with what their research claimed to disprove. But in the end, the researchers are psychologists, not ethnographers, linguists, or Wittgensteinian philosophers, so maybe we can give them a pass.
It’s just that the way the issue has been presented over the years has produced a messy conversation. So per Cassam’s accusation of bad intellectual character, we have little idea what the coaches have in mind when they dismiss the assertion that a fallacy is at play, because too little time has been spent clarifying the issue, too little attention paid to sifting through what might be meant by the key terms allegedly being debunked. Let’s be clear that it won’t do to simply retreat back to GVT’s exclusive definition of the hot hand, because the issue gains its popular traction from the incredulous reactions from basketball fans (and apparently now, coaches) who are all entitled to their untutored uses of the concept, but whose views on the sport they love are dismissed by the statistical cognoscenti. Sure, the data tell us that admonitions from fans such as “Give it to Toney, he’s got the Hot Hand!” turn out to be bits of advice that aren’t any more reliable than following random chance, even if Toney appeared to be on a hot streak.
Everyone should admit that this is an interesting result, and certainly the average fan isn’t entitled to just any belief. But the claim is sweeping in that, in order to demonstrate that the Hot Hand doesn’t exist, we would need to show not only that we do a poor job of spotting hot streaks taking place in real time, or that our odds of predicting whether the next shot will go in aren’t better than random chance. No, we’d also have to show that our retrospective looks at, say, Jamal Crawford’s 16 consecutive made shots in a 2007 NBA game are not memories of a hot-streak in the relevant sense, and in fact each shot was either random or was due to skill only on a case by case basis – that Crawford was in no groove. That, no study has done, and it’s indeed hard to imagine how any could. That might sound like an unreasonable standard to have to meet, but after all, that would be the implication of showing that the Hot Hand doesn’t exist.
If I show you the results of a player’s shots over time (H’s and M’s), say, a string of shots like Crawford’s cited above, and you show me a randomly generated sequence (also call each term in the sequence H or M, for effect) you do not win the argument by simply showing me that random sequences also sometimes cluster in streaks, not without begging the question. The most you’ve done is to raise a case of underdetermination. In other words, one activity is admittedly random (the number generator) and the other, basketball, is at least allegedly influenced by human practice and skill, in a way that allegedly sometimes leads to grooves or “in the zone” experiences. That randomly generated sequences can look just like streaks doesn’t mean “in the zone” experiences aren’t occurring in one of the data sets. Just from the sequences the researchers laid out before us, (the shots and the random sequence) we simply can’t tell if both sets are random or if one contains genuine hot streaks – the data can bear both interpretations.
But before we get too enamored with data, maybe we can just think about it like normal people, which is an avenue that’s been open all along. Have you ever swung a hammer all day? If so, did you get better as the day progressed? Were there parts of the day when you somehow managed to hit the nail right on the head, rather than hitting your thumb, for a long period of time? The question is whether each swing was independent or if the swings at least some of the time bore a relationship to one another. That is, if the swings were related by virtue of being in the pattern of movement you somehow achieved via all those swings. To be fair, the researchers were focused on basketball, not hammering, but both are physical activities, and the Hot Hand skeptics would have us believe that physical grooves cannot exist in shooting a basketball – that the influence of an “in the zone” experience could not be felt throughout successive shooting attempts.
After all, grooves are what Hot Hands are often made of. It’s not necessarily that basketball fans believe that some mystical quality has befallen an athlete or that somehow this power is such that we must throw the athlete the ball and that the next shot will go in. That may be what some fans mean some of the time, but that’s not an exhaustive list of the concept’s uses. A retrospective look at a practiced athlete with muscle memory operating in a groove will do, so far as fair use of the concept goes (notice that this look-back doesn’t necessitate the general belief that each time a player hits a shot or two, that the probability of subsequent shots going in rise). No one needs to consult statisticians or philosophers to confirm how obvious this is, just talk to some folks at a sports bar or ballgame and see if that use passes (“wow, Crawford really had the Hot Hand that night”). Our own reports of our naive experience shouldn’t be a trump card, but to doubt the most obvious forms of our experience in the face of ill-fitting data is to be constitutionally fickle.
So what should I say about Cassam’s mention of the hot hand fallacy when employing his suggested method of assessing intellectual character rather than just sticking to evaluating reasons? Given the irony that the so called fallacy was dragged out by Cassam to demonstrate the bad character of coaches who believe in the Hot Hand, not to mention that the data are now divided on the topic? If there are any character assessments that apply, I’ll leave them to others, and we can imagine the back and forth that could ensue from there. In the wake of that thought, I wonder if our public discourse has the capacity for many negative assessments of one another’s intellectual character – how long we can go before our conversations become fit for cable news rather than reasoned discussion.