On Understanding and Explaining
They're not the same, but not completely distinct, either.
One of the tabs I opened last week but didn’t have time to respond to is Charlie Becker’s “Simplicity, Expertise, and Bullshit”, taking issue with the apocryphal quote about how “If you can't explain it simply, you don't understand it well enough.” This general sentiment is attributed to a lot of people in a lot of different forms— other versions have it as “You don’t really understand a subject until you can explain it to your grandmother”— but always with the link between understanding and communication. As you might well imagine, given that my tenuous nerd fame arises from explaining physics to my dog, this is a subject I’ve thought quite a bit about, and as a result, I have Thoughts about Becker’s take.
Becker is coming at this from a sort of policy journalism perspective, which is reflected in the title of his post, and leads him to a worry that the expectation that explanations will be simple makes people susceptible to bullshit. It also leads to this paragraph, whose final sentence made me do a double-take:
This is where we get to the crux of the issue, why I think the connection between understanding and simplicity is false. Understanding something and communicating that understanding are two entirely different skills. Communication is important, but conventional wisdom places too much relative importance on the value of communicating your understanding.
As someone who moves between the worlds of academic science and public communication, I assure you that at no point in my career have I felt that communication skills were over-weighted in professional evaluations. Not for one single minute.
He has a part of a good point, though, when noting that ability to understand a subject and ability to communicate that understanding are not identical, but I think it’s a step too far to say that they’re entirely different. There’s enough overlap that I don’t think they can be completely distinguished, which you can see by considering the limiting cases: if you don’t understand something, there’s nothing to communicate, and if you can’t explain something, it’s hard to see how you would demonstrate that you understand it.
This is actually demonstrated by his principal example of a communicator with understanding, Richard Feynman, and the disastrous debut of his formulation of QED at the Pocono conference in 1948. Feynman’s presentation was built around demonstrations that he could calculate the correct values for various quantities using his intuitive approach, but it was a complete disaster as a piece of communication, because he hadn’t fully nailed down his approach in a way that allowed him to communicate it. That was actually probably a Good Thing in the end, as his humiliating failure at that workshop spurred him to actually sort the whole thing out and write it up in a way that made it comprehensible to others. He hated the process of publishing, but embarrassment forced him through it, to the great benefit of physics at large.
The other issue I have with Becker’s discussion is that most of the terms in the problematic “quote” are subjective. Becker recognizes one piece of this when he repeatedly notes that Feynman’s target audience was first-year students at Caltech, but leaves out the second subjective element, namely Feynman’s expectation for what they would come away with. When Feynman said he couldn’t explain spin at a level appropriate for Caltech frosh, he meant that he couldn’t explain it to students at that level in a way that would meet his own standards for accuracy. Since we don’t have a good record of whatever explanation Feynman attempted, it’s impossible to say where the breakdown was, but it’s entirely possible that the issue wasn’t with his understanding but with the level he wanted them to reach.
This is a problem that comes up a lot in general-audience explanations of physics: professional physicists have very strict ideas about what counts as “understanding” a subject, and sometimes that leads to unreasonable expectations. I’ve gotten any number of comments and emails over the years where people tell me my explanation of some quantum concept is trash because it fails to capture some incredibly abstruse point of the formalism that only matters in a handful of edge cases. Those edge cases really matter to the research specialty of the complainers, though, which leads them to maintain that my explanation is horribly inadequate. But that’s often not a reasonable expectation for a non-technical audience.
(And keep in mind that my nerd fame is as limited and tenuous as it is in part because I’m less inclined to “dumb down” my explanations than people whose books sell orders of magnitude more copies tan mine… I’ve also gotten a lot of emails and comments praising me for not skipping past some abstruse point or another.)
One of the most important distinctions between internal communications in a scientific community and public outreach to non-scientific communities is that the expectations have to be adjusted to the audience. You can’t expect first-year Caltech students to understand a complex technical subject at the same level as the Nobel laureate who invented the field, but it’s equally unreasonable to set “Caltech frosh” as the bar for understanding when talking to a general audience. The level of understanding you can reasonably expect your audience to achieve depends on the audience.
This is a bit like the folk belief that children learn languages more readily than adults, which an actual linguist once explained to me is not actually true. By objective measures— number of vocabulary words learned, mastery of grammatical concepts, etc.— adults and children learning a second language pick things up at about the same rate. The apparent difference is a matter of expectation— it takes an 8-year-old and a 48-year-old the same amount of time to reach the level of proficiency that we expect of an 8-year-old. Getting to that level doesn’t seem like as much of an accomplishment for the adult, though, because they’re still way short of the level of fluency we expect from someone who can legally drive a car.
If you’re trying to communicate a complicated subject to people with minimal background in it, you need to properly calibrate your expectation for what they will know afterwards. The standard for explanation can’t be “Can your single lecture bring this complete layperson up to the level appropriate for a highly intelligent and motivated first-year student at one of the world’s best technical universities?” The target needs to be much lower.
And it’s important to stress that this is not “dumbing down” the subject, or patronizing the audience. It’s a matter of setting reasonable expectations given your audience and their background. In the same way that you wouldn’t hand a first-year French student a modern literary novel and expect them to get anything out of it, but might give them a copy of Le petit prince, you wouldn’t expect a random person coming out of one lecture to grasp the full subtleties of the spin-statistics theorem, but you might be able to get across some of the essential elements needed to explain a particular phenomenon.
The place where I definitely agree with Becker is that there’s a lot of specialized skill involved in communicating a complex subject like physics, that’s at least partly distinct from the specialized skill involved in doing physics. This involves things like a flair for metaphors and some skill in language, but also a kind of mental modeling of the audience you’re attempting to reach, and the level you want to bring them up to. That requires a reasonably high level of understanding of the subject matter, but also a willingness to set aside some of the abstruse details that are essential for deep understanding in the interest of communicating more cleanly.
On Becker’s core concern regarding “bullshit,” I think this partly reflects the same issue. That is, I don’t think it’s unreasonable to expect explanations to non-technical audiences to be (relatively) simple, but there should also be a reasonable expectation for the level of understanding that these can impart. A lot of the problems associated with people bullshitting after hearing simple explanations aren’t problems with the explanations themselves, but with Caltech frosh thinking they’re Nobel laureates after hearing a single lecture.
If you like this, you should definitely buy my books, and might also consider clicking this button:
If you want to take issue with any of my claims here, the comments will be open:
Did the dog end up getting tenure, or did he go into industry?
To me the spectrum here is understanding, explaining and *doing*.
I think there are things I understand--that I am, in some sense, literate with. But being able to communicate that understanding to another person takes something more. There are things I understand implicitly or kind-of-sort-of but I would get tongue-tied trying to lay it out and would have to practice/review to do it. Having my communication produce understanding in another person is yet another challenge level up. Being able to *do* what I understand--to produce research or enact a skill--is beyond that.
I think college pedagogy is sometimes unreasonably focused on getting students to *doing* in everything that we want them to take seriously, when in fact most faculty rely on a big web of understandings that don't necessarily translate into the ability to *do*.