The Three Definitions of "Science"

And why most arguments about what "science" is get them muddled

Photo by Rohan Makhecha on Unsplash

Several years ago now, I wrote a book about the universality of science, Eureka: Discovering Your Inner Scientist, which I’m still really proud of, but which was a commercial bomb. I spent a good deal of time around its release talking about the nature of science, and how scientific thinking turns up in all manner of activities we don’t think of as scientific, but sort of burned out on the topic, so haven’t flogged it as relentlessly in the years since.

A couple of Twitter threads have conspired this weekend to drag me back in, though. The first was from Venkatesh Rao:

Venkatesh Rao @vgr

No.

Jessica Hernandez, Ph.D. @doctora_nature

Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science Indigenous knowledge is science

As I noted in response to that one, this is a topic where I have strong feelings, but there’s not a lot of upside to writing about it. Within a couple of days, though, Noah Smith weighed in with this thread:

Noah Smith 🐇🇺🇦 @Noahpinion

There's nothing Eurocentric about modern science. There IS something Eurocentric about the way we teach the history of modern science, which leads a lot of people to mistakenly think that the idea of science itself is Eurocentric.

and also a bunch of follow-ups where he argues with people about their responses. And again, I have strong feelings about this, but they don’t align perfectly with any of the major camps, which means there’s not a huge upside to writing about this.

On the other hand, though, the combination of the two is intensely annoying to me because of those feelings, to the point where the downside of not writing about it is worse than the lack-of-upside to the topic. So I guess I’m in…

The fundamental problem here is one of definitions, which is a problem that plagues all arguments about the nature of science. There are at least three fairly broad things that people mean when they refer to “Science” as a thing, and most of the annoying arguments about the topic happen because people are using different definitions without ever clearly specifying which one they mean. (And sometimes switching between them in the middle of the argument, which doesn’t help anything…)

The three broad meanings of “Science” (as I see it) are:

I) A process for generating reliable models of the world. This is the focus of my book, where I shorthand the process as “Look, Think, Test, Tell”: You look at the world and identify phenomena in need of explantion, you think about why those phenomena happen in that way, you test your theory of how things work with further observations and experiment, and then you tell everybody you know the results of your experiment. This process, iterated many times by many people, generates models that have exceptional predictive and explanatory power.

II) A body of verified knowledge about the world. One of the outcomes of the iterative process of Definition I is a collection of facts: The Earth orbits the Sun, light has both particle and wave character, DNA is the molecule that carries genetic information, etc. These are components of models about the world that have been very thoroughly subjected to the scientific process, and can in most cases be relied on without constantly needing to re-test them (though it’s often useful to poke at even the most fundamental elements of our body of knowledge). Much of this knowledge is embodied in technology: we use our knowledge of orbital mechanics to launch satellites, our knowledge of light to make optical telecom networks, our knowledge of genetics to predict and treat diseases, etc.

III) A collection of institutions and formal practices associated with the people who carry out the process of science. This is the network of universities and laboratories and journals and conferences and all the complicated and expensive apparatus of “doing science” in the modern sense of the term. It also includes the community of researchers making use of the Definition I process at scale, and also the various experts and commentators who offer advice about how to make use of the Definition II knowledge to achieve particular goals.

I list these in order of decreasing universality— just in case you missed the bit where the later ones refer back to the earlier. Absolutely every human capable of functioning as an autonomous individual makes use of the Definition I process of science, but only a tiny minority of people can claim to be “scientists” in the sense of being part of the Definition III institutions of science.

These are closely related, but distinct categories, and nearly all of the arguments about what “Science” is and isn’t involves blurring one or more of these together. Which is what makes these arguments exceptionally annoying, because both sides of the argument can be “right” about their actual claims, but they can fight endlessly because they’re using the same word to mean different things.

You see this in Smith’s thread about how “science is not Eurocentric,” where his claim is mostly that the Definition I process of science is universal and has been practiced by every human civilization that we know anything at all about. (With a dash of reference to the specific people currently participating in the Definition III instutitions of science, who come from all over.)

On the other side, the people who argue that science is fundamentally Eurocentric are talking almost exclusively about the particular set of Definition III institutions and practices that currently exist. And I think it’s equally true that those particular institutions and practices are Eurocentric— yes, there was a lot of Definition II knowledge generated in China and the Islamic world before the “Scientific Revolution,” but I think it would be foolish in the extreme to claim that there’s nothing different about how things were done in Europe after, say, 1700 or so than how they were done elsewhere or prior to that. The Definition III institutions we have now are the result of particular choices made mostly in Europe and spread to the rest of the world from there (largely through a period of imperial expansion), so calling those “Eurocentric” is not an unreasonable distortion.

So, both Smith and his opponents are right, but because they’re largely talking about different things.

In the case of Rao’s argument, the claim he’s responding to is (on its face) that “indigenous knowledge” should be included in the Definition II body of verified knowledge. His response is essentially that it can’t claim that status because it hasn’t been verified through the formal practices of Definition III, which is true.

At the same time, though, I have a lot of sympathy for the claim that at least some categories of “indigenous knowledge” really should be thought of as at least provisionally Definition II scientific knowledge, because while they haven’t gone through the formal institutional verification needed for modern Definition III science, they have been generated through the iterative trial-and-error process of Definition I. That process is truly universal, practiced by basically every human culture at some level. At least for knowledge like “If you have these symptoms, you should eat these plants,” there’s almost certainly something there that’s been tested and passed down through the years.

Of course, what you do about that status is another question. My feeling is that “indigenous knowledge” about herbal remedies and local agricultural practices and the like should be treated as within the Definition II body of knowledge in the sense of not being cavalierly dismissed as worthless by Definition III institutions. But it should absolutely be investigated seriously by those institutions using modern scientific tools and practices to see what’s there and what can be made of it. Folk practices that actually work can almost certainly be learned from and refined and improved, and that’s a worthwhile activity for Definition III institutions to undertake.

Again, both sides have merit, because they’re talking about slightly different things.

A major problem with this particular category of definitional cross-talk, though, is that many claims about the status of “indigenous knowledge” aren’t really made in the service of a drive to better understand the world and add to the Definition II body of scientific knowledge (including useful technology). Instead, they’re made as part of a political project to undermine the status of Definition III institutions and practices— not raising traditional medicine up to the status of medical reserarch, but dragging institutional medicine down to parity with folk remedies. That happens often enough that many folks in and around the Definition III institutions reflexively treat any claim about knowledge generated outside that system as a bad-faith attack on the system, leading to protracted and nasty fights.

(I should note that I do not know anything about Jessica Hernandez, who Rao is responding to, or any history they have together. I don’t know if she’s engaged in this kind of political attack on Definition III institutions (and thus am not accusing her of bad-faith arguments), but his response reads like he thinks she is.)

The same is often true about arguments over Eurocentrism— these aren’t necessarily about removing cultural obstacles to allow the casting of a wider net, but attempts to lower the (perceived) status of Definition III institutions for political reasons. Again, this drives a lot of the responses to such arguments into toxic territory when that’s not necessarily justified.

And, of course, I am not at all without political motives in my own preferences regarding how to talk about this stuff. I’m very much in favor of talking about the Definition I process of science as a human universal because I think that serves the cause of getting more people to realize that science isn’t some arcane activity that only super weird nerds can take part in, but a basic process accessible to everyone. Which, I hope, can boost people’s confidence in the Definition II body of knowledge, and maybe even raise the status of and participation in the Definition III institutions and practices (which are not without problems, but are also an unprecedentedly powerful set of tools for doing the Definition I process at scale). That’s why my book takes the approach it does, and tries to be positive and encouraging throughout. (Which is probably why it sank like a rock after publication, but I tried…).

So there’s my middle-ground position on recent arguments about what “science” is and isn’t, and why these arguments are ultimately kind of frustrating.

---

If you like this, well, first and foremost you should buy my book, but after that, you can click this button to get more of this in your email inbox:

And if you’d like to argue with my definitions or characterizations, the comments will be open: