If It's Already On Paper It's a Secondary Source

On "Great Books" for science

Over on ex-Twitter, Zena Hitz of St. John’s¹ tweeted a sweeping call for a Great Books approach, which I replied to defending the existence of textbooks in STEM (screencap because petulant billionaires):

Hitz followed up a bit later with a number of responses arguing that “Great Books” is a superior approach to the sciences, as well, and I fairly quickly bowed out because I was very much Not In The Mood². Then this somehow rose to the level of being a trending topic promoted in my ex-Twitter sidebar, and I got a lot more of it pushed into my feed, culminating in this (again, sorry for the screenshot):

This annoyed me enough that I now feel a need to spell out my position at greater length on the blog.

I will note that I have a fair amount of respect for the general idea of a “Great Books” approach to literature³ and philosophy and history, at least with a sufficiently broad definition of “Great.” But I would say that that is largely because the objects of study in those fields are, to a large extent, objects that exist primarily as text. They’re works of literature, records of events, transcriptions of arguments, and the exact words used to express them are absolutely critical to understanding their meaning. I don’t think you can really claim to understand what Plato was on about if you’ve only read an encyclopedia discussion of his work; I’m not sure I’d fully trust even an interpretation based only on reading Plato in translation, not the original Greek.

The study of STEM is fundamentally different, in that the objects of study are natural phenomena with an independent and continuing existence in the world. When we study physics, the goal is to create and refine the best possible model for predicting the outcomes of experiments and observations. Those outcomes are the real object of study, and the particular words used to describe them are at most a translation of something more fundamental. Even the mathematical formulae used are secondary, and subject to change with future refinements of the models.

Part of my collection of physics texts

This is why a “Great Books” approach, in my opinion, is largely inappropriate for STEM— as another physicist put it, “we don’t study Maxwell, we study EM.” Maxwell’s original presentation of electromagnetism is pioneering work, and a spectacular success in terms of explaining and unifying disparate phenomena. It’s also really cumbersome compared to modern treatments, needing 20 equations to express what we render as 4, and devoting a lot of space to a complicated analogy to fluid flow that isn’t all that illuminating to modern readers and can lead to the thoroughly discredited idea of a “luminiferous aether.”

Maxwell’s work was an essential step on the road to a modern understanding of electromagnetism, but working through Maxwell’s argument in detail is not in any way a prerequisite for working with the modern understanding of electromagnetism. Spending significant time on the details of Maxwell’s reasoning would be a time-consuming distraction from learning either how to work with the modern formulation to design and analyze useful systems, or how the classical understanding has been superseded by quantum electrodynamics. We have a more complete picture of the theory now, and much better tools for applying it. There’s no need to spend time on Maxwell’s relatively crude and cumbersome approach, any more than there’s a need to teach students how to find the precise values of trig functions or logarithms in printed tables now that we have ubiquitous scientific calculators.

I’ve had versions of this sort of argument a lot of times over the years, and I think the closest I’ve ever gotten to getting someone trained in “the humanities” to grasp the point was in the context of an argument about our former Gen Ed system emphasizing the difference between primary and secondary sources. In the course of explaining this to a colleague, I said something like “A primary source is doing the experiment yourself. If I can find it in the library, it’s a secondary source, no matter who wrote it.” (Thus the title of this post…) This is why science courses have labs, after all— the core phenomena being studied are directly accessible, and the best way to learn is to engage with the phenomena directly⁴. “Learning” science only by reading published work— whether “Great Books” or modern textbooks— is inadequate in the same way that “learning” Plato only by reading the SEP would be.

(The actual analogue to a “Great Books” curriculum for STEM would be a lab-heavy program structured around replicating all the key phenomena in whatever subject, and exploring the technical issues involved in making the necessary observations and analyzing the resulting data. That’d be super expensive and in terms of both equipment and faculty labor, but if any classically-inclined billionaires want to explore the idea, my consulting rates are quite reasonable…)

Responses to this structural argument fall into two broad classes, the first being attempts to cite specific Great Books that really do work as the best introduction to something or another. The most successful of these by far is Euclid’s Elements, but that’s very much in the “exception that proves the rule” territory. Plane geometry turns out to be a really special case, and genuinely was worked out pretty completely several thousand years ago. Second place is probably Darwin’s On the Origin of Species which both holds up pretty well as an argument and is highly readable in a florid Victorian sort of way.

Everything else has problems. Galileo straight-up lies about stuff, and Newton’s Principia is famously obscure, possibly deliberately so⁵. I saw a bunch of people pointing to the Einstein, Podolsky, and Rosen paper as a Great Work worth studying, but that’s famously because it’s not that good. People still go back to it because the argument is a little obscure even to the authors—Einstein himself felt that it didn’t capture what he was really after, though exactly what that was remains somewhat hard to figure out.

A subset of these responses are anecdotal experiences of the form “Physics never made sense to me until I read the Principia,” but if you look at who’s making those, they’re generally not in STEM— lawyers, writers, academics in literary fields. Which leads nicely into the second broad category of responses, which amount to changing the terms of the argument. That is, claiming that reading original classic works is superior to more modern treatments by changing the definition of “learning” in a way that makes it true by fiat.

This is the core of the “STEM is pre-professional; we practice science as a liberal art” tweet that got up my nose. Though, to be fair, it does touch on something that’s actually true: when I say that you’re better off learning physics from a modern textbook than reading Newton and Maxwell, “better off” assumes that the goal is to be prepared to understand physics as it exists in the 21st century, and potentially contribute to moving it forward. If you push modern applications aside as grubby “pre-professional” activity, and define “science as a liberal art” in a way that’s unconcerned with the possibility of postsecondary study in the subject⁶, then, sure, some other approach might be better.

But at that point, we’re no longer talking about the same thing— that’s not a Physics degree, it’s more of a Philosophy of Physics degree. Which can be a fine thing, depending on your goals and aspirations. But it does kind of foreclose the possibility of productive curricular discussion, because we’re talking about the relative merits of apples and baseball.

So, I will repeat what I said in one of the various offshoots of the original tweets, and leave it at this: if you want to read Galileo and Newton and Maxwell for historical context and personal enrichment, I’m all for that. It’s an interesting experience— I’ve at least dipped into a lot of this for my more historical books (one, two, three of them). But if your goal is to learn what you need to know to become a working scientist in the 21st century, in whatever field, you will be much better served by taking a class based around a good modern textbook. Anything that comes from a publisher is a secondary source, so you might as well use one that employs the most up-to-date tools and techniques.

---

This is perilously close to violating my rule against responding to things where my initial response is “Oh, go fuck yourself…”, but I think it stays just inside that line. If you want more of this, here’s a button:

And if you want to disagree, the comments will be open:

1. Not the one that’s good at basketball.

2. It would be inappropriate to discuss in detail, but I’m thoroughly fed up with some colleagues in “humanities” departments at work, and it would’ve been hard to keep that frustration from leaking in.

3. I am, in fact, about eight books into a re-read of The Odyssey in a translation picked specifically because Hitz recommended it.

4. This seemed like it worked for the remainder of that conversation, but of course a year or so later we were back to zero. See footnote 2.

5. He also miscalculates the orbit of the moon, though to be fair it’s a horrible calculation that wasn’t sorted for the better part of a century.

6. I’m not sure I know any physicists who went to St. John’s, but I’ve at least talked to people who had a St. John’s alumnus as a graduate student, and they apparently needed to do a ton of catching up in the first year of their Ph.D. program to get through.