In a post that I mostly agreed with, but am also mostly not that interested in, Scott Sumner concludes with the following note:

But I also understand that the part of my brain that tells me that the conventional narrative is stupid, is itself unreliable.

Indeed it’s more than unreliable, it’s a logical contradiction.  The conventional narrative can never, ever be stupid, as ‘stupidity’ is defined as reasoning that falls short of the conventional wisdom.

This is a noble and humble thing to say, but to what extent is it true? My instinctual assertion would be that this statement is almost entirely false, but I am not sure and it could be important. The corollaries of the assertion that the conventional wisdom can be stupid frequently get me into trouble. People very much do not like being told that they, their opinions or their actions are stupid, especially when they are plausibly average or better in context. One could also think of this as ‘having high standards,’ and in at least some contexts my standards are ludicrously high because I find keeping them that high to be useful.

The standard of ‘relative to the conventional wisdom’ is itself high if you take the conventional wisdom to be the wisdom of a relatively small convention. The majority of Americans don’t believe in evolution, or in a market price for water, and legislation regularly goes squarely against both, but it would be reasonable to claim that the conventional wisdom favors both. Scott’s statement need not be a contradiction with the majority of humans being stupid and the majority of them having mostly stupid opinions. In fact, given which opinions count as opinions, Scott’s standard all but guarantees it! Conventional wisdom is necessarily much less stupid than a random person’s opinion, on average, due to the wisdom of crowds.

Certainly, relative stupidity is a thing. One can say “X is stupider than Y, but less stupid than Z,” where X, Y and Z can be proposed courses of action, scientific theories or even people, and have that statement be a not-obviously-stupid way of delivering the desired informational payload. It also does not mean the same thing as “X is not as smart as Y, but is smarter than Z” because reversed stupidity is not intelligence, so the phrasing is necessary. It is not even a contradiction to say “X is smarter than Y, and X is stupider than Y.” They are highly correlated, but not the same scale.

Is absolute stupidity a thing? Could one say “X is stupid” outside the context of an implicit Y that is not as stupid? One can say “that is a stupid answer” without knowing a better one, but that seems to presume that a better answer exists. To say “there are no non-stupid answers to that question” seems to imply that the correct response is not to answer at all, but that is still an answer. There cannot be zero non-stupid moves on a chess board, unless there are also zero moves. On writing it out, it seems right to say that “X is stupid” requires an implicit baseline level of stupidity Y the same way that saying “X is large” requires a baseline size Y.

That leaves the question: What is the correct baseline? Is ‘the conventional wisdom’ a good general choice, leaving anything below it ‘stupid’ and everything at or above it ‘not stupid’? It is not an obviously stupid choice, if one must choose a universal standard. But it lacks context. When a measurement is relative, the goal should be to maximize its informational content and the utility that can be gained from it. My one year old son (happy birthday, Alexander!) is stupid in the sense that all babies everywhere are deeply stupid compared to almost all adults. In some contexts, that is a highly useful thing to keep in mind. You certainly would not want to forget it! In others, a better comparison is to the reference class of other one year olds, in which he is very much not stupid.

I believe that the correct baseline is to keep a full growth mindset, keep your standards high, and set the standard for not being stupid to be the highest standard that is achievable in context given an effort appropriate to the situation’s importance. By default for important things, this should be to make an extraordinary effort but not to shut up and do the impossible. Noting stupidity is noting opportunity for improvement. If you tell me that I am being stupid, that implies that this is by a standard in which I could act non-stupidly, but have chosen not to, and it is likely I should go fix that. It is the exact opposite of calling myself stupid!

By this standard, if Scott Sumner were to believe the conventional wisdom were stupid in the context of economics, that could be seen as him saying ‘I know enough that if I were to believe the conventional wisdom about this, that action would be stupid,’ and I would certainly hold him to higher standards than I would hold conventional wisdom. But he could also be saying ‘there is enough information out there, and the question is important enough, and people are smart enough, that the conventional wisdom should be held to a standard it is failing’ and I think that in much of economics this is also something I would agree with.

There must a reason for this stupidity – to take an easy example, stupidity can be to the advantage of key people who can influence the conventional wisdom, and that’s an easy and plausible story to tell in this context and many others. Another common explanation is that we haven’t had time or experience yet to come up with something less stupid, but we are also making the claim that a less stupid alternative is within reach. However, it can also be seen as a call for explanation. We say ‘given the facts at hand, we really should be doing something less stupid, so given this is so stupid, there must be missing information that explains it. What is going on?’ Alternatively, it could simply be a call for improvement. This is stupid, and that means we can do better.

Telling someone that they are stupid (or not smart enough, or not talented enough…), rather than that their actions or beliefs or ideas are stupid, is then something very different. It is saying that the person in question is incapable of taking what would, for a non-stupid person, be a non-stupid action (or understanding, opinion, plan, etc). On occasion that is useful information. I once asked the question “Am I smart enough to usefully work on this problem?” and was told, point blank by someone who would know, “No.” In context, I was a moron, although of course such words were not used. I am grateful for that answer, because the question is damn important, and if I thought I could meaningfully help I would like to work on it. Instead, I get to work on other problems, which are hopefully problems I can solve. Not saying this when it is true, or telling people they can never say this, is harmful because the informational content is valuable, but such messages have also proven that they can do great psychological harm. How to balance these two needs is a major unsolved problem.

Most people, of course, use these two sets of terms, that (at least in this theory) mean almost exactly opposite things, pretty much interchangeably, and often to mean both things at once, often as a status attack. That makes it really tough to adopt the language and mindset I find most useful, especially since the inner-monologue version tends to use them interchangeably as well because of course I know that I am not stupid in the bad way, so I don’t need to use all these careful words to make that clear. 

There is a set of people that I could call “the set of people who I can freely use the word stupid around without worrying that they are going to take it the wrong way, and who know they can say it back to me as well” and I love spending as much time with such people as possible. It is a joy.  In the meantime, in the quest to become less stupid, it might make sense to consider oneself as stupid as possible, but no stupider.