Everybody Is Wrong About the Structure of Science (Part 1)
Including me, probably
The new year has gotten off to a spectacular start in the critical area of “people in my social-media feeds being Wrong On The Internet,” at least when it comes to thinkpieces about How Science Works. Some of these are merely new-to-me rather than actually new, but they actually tend to reflect pretty common trends in online wrong-ness giving this a bit of an evergreen quality. Anyway, it’s worth a bit of scattershot commentary spread over a couple of posts on various modes of Internet Wrong that have cropped up.
We’ll start, as we so often do, over on ex-Twitter, with a screenshot of a quote-tweet that is a bit of a two-for-one:
As a physicist, I find this entertaining in both directions— an English professor sneering that economics isn’t a real discipline is amusingly on-brand, and an economist claiming it’s a science on the level of physics likewise.
(A bit of an aside: There’s a long-running semi-joke about economics claiming to be “the physics of the social sciences,” in that it’s the most mathematically formal and rigorous of the lot. My return joke isthat economics actually the astronomy of the social sciences, in that it seems to involve a lot of ad hoc efforts to quantify incredibly complex systems that result in everyone arguing about something that turns out to be proportional to the negative logarithm of the thing you actually would like to measure.)
Page’s quote-tweet is one of many (attempted) dunks on the original claim, which in turn has been met by a lot of “Ah, but all of human knowledge is contingent on a chain of political and historical circumstances and assumptions, so you’re the one who’s Wrong…” Things rapidly descend into “nuh-unh!” “uh-huh!” territory from there.
Both of these positions contain elements of truth, and both are Internet Wrong to some degree. It’s certainly true that the historical development of economic thought has been shaped by underlying sets of assumptions shaped by historical and political circumstances, and you can absolutely imagine different approaches. If you’re in the business of trying to radically re-think things, it can be useful to go back to the start and examine those assumptions, and how things might be different if you made different assumptions. So in that limited sense, Moskowitz is gesturing in the direction of a valid point (though there’s an ironically wrong element of that, too).
At the same time, though, it’s also true that it is not essential to explore those assumptions to have a successful (and lucrative) career as an economist. In fact, it’s almost certainly a better use of time and resources to not go back to the original sources when training students— there have been enormous advances in mathematical techniques and notations since the days of Adam Smith, and it would be foolish to not take advantage of those in streamlining the presentation of the current state of the field. (See the reflections at the end of this old Cosma Shalizi post about Marx, which popped up in replies to the quoted tweet above…)
To make an analogy to a place where I’m on firmer ground regarding the content and history, it’s perfectly possible to learn General Relativity by going back to the original papers by Einstein and others. It can even be illuminating to work through those papers, because they sometimes spell out things that physicists nowadays skip past a little too lightly.
At the same time, though, that would be an inefficient way to learn the subject, because you’d need to slog through the bits where guys trained on 19th-century physics laboriously re-invent branches of mathematics that were not familiar to them. This can be interesting— I saw a fascinating talk once about the many (many, many…) mis-steps Einstein made in the path to GR as he (and Marcel Grossmann) tried to sort out how to correctly capture the physics of relativity in the language of Riemannian geometry. But while those blind alleys are useful to dispel the impression of Einstein as an ineffable genius, at the end of the day, they’re still blind. There’s no useful purpose served by re-making those mistakes for the vast majority of physicists who will use General Relativity to explore open questions in physics.
This is an argument I had with a colleague in history back in the early days of a former Gen Ed curriculum, which put a bunch of emphasis on the teaching difference between “primary” and “secondary” sources. I don’t think I ever fully succeeded in explaining that this was not a significant matter in much of STEM, but the closest I got was “Anything I can check out of the library is a secondary source.” The true “primary sources” in physical science are direct observations, experiments, or calculations; everything else is on an equal footing, a step removed. Unless your primary interest is in exploring the historical development of physics, there’s no added benefit to getting, say, Maxwell’s equations from Maxwell’s original papers versus the latest edition of a modern intro textbook. If anything, you’d be at a disadvantage, because Maxwell’s original work uses a bunch of mechanical analogies that aren’t especially illuminating to modern students and is expressed in a kind of cumbersome notation that has long been superseded.
If, on the other hand, you want an illustration of exploring context being actually useful, the go-to is probably quantum mechanics, specifically the version developed by Bohm in which quantum particles have definite properties that are determined by a non-local potential. That was in large part inspired by picking up an idea originally put forth by de Broglie, but abandoned in the face of opposition from Bohr and the Copenhagen crew. It was assumed for years that a de Broglie style model was impossible, thanks in part to the enormous reputation of John von Neumann, who asserted its impossibility in the 30’s (Grete Hermann spotted the loophole in von Neumann’s argument right away, but nobody paid attention until much later). Bohm revived it in the 1950’s after digging into quantum foundations in the process of writing a textbook, and it’s been a moderately popular alternative approach for decades.
You can argue about the utility (or lack thereof) of the Bohm approach to calculations, specifically, but it was the discovery of Bohm’s non-local model that got John Stewart Bell thinking about quantum foundations in the mid-1960’s. And that led to Bell’s theorem and the modern understanding of quantum entanglement, which is one of the factors that underlies quantum information as a branch of physics and a speculative industry. Which I’m pretty sure we can agree is significant, and would not have happened if Bohm and Bell hadn’t had reasons to go back and look at the context of older ideas.
The core issue, on both sides of this, is an incorrect assumption that the mode of operation characteristic to one discipline is, in fact, universal. The actual truth is somewhere in the middle— there can be value to going back to explore the historical context of original ideas in economics or physics, but it’s not by any means an essential prerequisite to doing modern work in those fields. Asserting either extreme— that there’s no value in understanding the context of original ideas or that there’s no value without it— is stupid and arrogant.
I was originally going to throw in another Internet Wrong topic here, but this took up all the blogging time I had allocated for today, so I’ll pick that up later. If you want to get that as soon as it drops, here’s a button:
And if you feel moved to respond to any of this, the comments will be open:
I took Intro Economics 101 and about a month in, realized all this talk, talk, talk, even speaking Latin ("ceteris paribus") could be avoided with a half page of partial differential equations.
Excellent post!