Storing Information In Other People's Heads
To function effectively in the world, you need to acquire a whole lot of information. You need to know exactly which medicine is appropriate for each ailment. You need to know how to fix your car and your router and your irrigation system. You need to know the date of every major holiday and how it is observed.
Right? Of course not. That would be crazy.
We don't keep all the information we could possibly need in our own heads, just as we don't make all our own clothes and manufacture our own doorknobs. We rely on a division of labor. And, according to many cognitive scientists, we also rely on "a division of cognitive labor." Beyond consulting Google and smartphones and books, we store some of the information we care about in other people's heads.
Thanks to this clever outsourcing, you don't need to know which medicine is appropriate for each ailment, as long as you can consult a good general physician. And that physician doesn't need to know the ins and outs of tropical foot fungi, as long as she can consult a good specialist. But storing information in other people's heads won't do us much good if we don't know whom to consult for what. It's no good to ask your doctor about the car and your mechanic about your foot fungus.
One way we carve things up is by domains of expertise, and there's evidence that even young children can do this in a way that more-or-less mimics academic disciplines. We know to ask the physicists some kinds of questions and the biologists others, just as we know to consult our physician (and not our mechanic) when it comes to fevers (but not fenders).
More remarkably, still, we seem to do this without having people sorted by professional expertise — as "doctors" or "mechanics." We can parcel up the world into much smaller and more everyday kinds of expertise. You know which friend to ask about why your chocolate soufflé collapsed and which to ask about why your business proposal was dismissed.
How do we keep track of who knows what — of what I'll call our "epistemic community"? And what is it, exactly, that we're tracking when we do so?
One clue about how to answer these questions comes from the way we attribute epistemic terms, such as knowledge and understanding. We might say that Carmen knows why climate change is occurring, or that Sanjay really understands why soufflés can be finicky. Statements like these seem to communicate something about how Carmen and Sanjay should be listed in our epistemic Who's Who -- our guide to who knows what. But what exactly do they tell us?
In a paper forthcoming in the journal Philosophical Studies, Daniel Wilkenfeld, Dillon Plunkett and I set out to answer this question. Specifically, we were interested in when and why people attribute understanding to other people. Is understanding why something occurred (say, understanding why the dinosaurs became extinct) equivalent to knowing why it occurred (say, knowing why the dinosaurs became extinct)? Or might understanding and knowledge track subtly different features of individuals in our epistemic community?
Our hypothesis was that attributions of understanding are pretty special. They're a bit like a gold star for someone's entry in your epistemic Who's Who — a mark that the person doesn't just know why the dinosaurs became extinct in a superficial sense (an asteroid!), but that she can tell you more. She has the kind of expertise that makes her a good person to consult not only if you're asking, "Was it an asteroid that led to the extinction of the dinosaurs, yes or no?" but also when you want to know: "Why would an asteroid have that effect? Which other species were affected? And what else could have led to extinction?" She can offer a certain explanatory depth.
To test these ideas, we conducted a series of experiments probing people's intuitions about when it is and isn't appropriate to say that someone knows or understands why something occurred. For instance, some participants read the following vignette:
"Approximately 65 million years ago, a large asteroid struck the Earth, spewing a large amount of ash and debris into the atmosphere. This debris prevented sunlight from reaching the surface of the Earth, which prevented plants from growing normally. This eventually led to the extinction of the dinosaurs.
Edward is a student who has been taught about dinosaurs and about asteroids. But, as of Monday, he has never been taught about the connection between an asteroid impact and the extinction of the dinosaurs. If asked why the dinosaurs became extinct, Edward would not have been able to give a confident answer, and he would not mention an asteroid impact."
We then asked them how much they agreed with the claim that "On Monday, Edward knew why the dinosaurs became extinct," or with the claim that "On Monday, Edward understood why the dinosaurs became extinct."
Not surprisingly, participants weren't too impressed with Edward's knowledge or understanding on Monday — they didn't agree much with either of these statements.
But, in the experiment, something interesting happened on Tuesday:
"At a lecture on Tuesday, Edward learned that there is a connection between an asteroid impact and the extinction of the dinosaurs. If asked why the dinosaurs became extinct, Edward would say that an asteroid impact caused the dinosaurs to go extinct. If pressed for more detail, however, Edward could not elaborate and would not mention debris, blocked sunlight, or plants."
When we asked participants to rate Edward's knowledge and understanding on Tuesday, we found that ratings were not only higher overall, but also significantly different from each other: Participants were pretty happy with the statement that Edward knew why the dinosaurs became extinct, but significantly less inclined to agree that Edward understood why the dinosaurs became extinct.
In the experiment, Edward went on to learn more and more on Wednesday and Thursday, allowing us to establish the following pattern. At the extremes — knowing very little or knowing quite a lot — attributions of knowledge and understanding went hand in hand. But in the middle, participants were more willing to attribute knowledge than understanding. That is, they demanded greater explanatory depth when it came to crediting Edward with understanding than when it came to crediting Edward with knowledge.
These findings are consistent with our gold star hypothesis: People are stingy in attributing understanding because they want to differentiate the people they should consult — the ones who can answer follow-up questions and help on related problems — from those whose utility is more narrow. But to firm up the connection between understanding and consultation — that is, the retrieval of information from other people's heads — we collected some additional data.
In another study, participants read about individuals with varying levels of explanatory depth, ranging from complete ignorance about why an event occurred (in this case, an omega appearing on a computer screen after a sequence of key presses) all the way to tremendous expertise (in this case, being the world's expert on the relevant software, hardware and underlying physics involved in the omega's appearance). We also asked participants how inclined they would be to consult each individual if they had a related question — not about how to make omegas appear, per se, but about something else related to computers. We found that participants' responses to this question — about whom to consult — were better predicted by their attributions of understanding than by their attributions of knowledge. (Interestingly, knowledge and understanding didn't diverge when it came to consultations about something wholly unrelated — in this case, economics.)
Here's one way to put these findings. When we say that Carmen knows why X occurred, we seem to be saying something relatively narrow about what she knows. It's about why X occurred. But when we say that Carmen understands why X occurred, we're saying something a little stronger. We also seem to be saying that Carmen is a good person to consult when it comes to X-related matters (though not necessarily about Y-related matters). Carmen deserves a gold star in our epistemic Who's Who for the domain relevant to X.
Before you nix me from your epistemic Who's Who, let me tell you why I think this is interesting — and what it tells us about the kinds of epistemic creatures we are.
For one thing, these findings reveal how subtle our use of language is. It usually takes training in philosophy to explicitly recognize nuanced epistemic distinctions, but people seem to track different features of their epistemic community intuitively and systematically. These features are encoded in language in ways that we can communicate with each other, and in ways we can remember for ourselves.
These findings also tell us something about how we manage the division of cognitive labor. If we can't store everything in our own heads, we need good ways to track what we can expect to learn from others. It's not enough to outsource information storage to other minds if we can't retrieve it when we need it back. (That would be like building up a library without bothering to update the card catalog along the way. You'd never find the right book when you needed it.)
So, now you know a little more about why we're such capable epistemic creatures. And, perhaps, you even understand why, too.
Tania Lombrozo is a psychology professor at the University of California, Berkeley. She writes about psychology, cognitive science and philosophy, with occasional forays into parenting and veganism. You can keep up with more of what she is thinking on Twitter: @TaniaLombrozo
Copyright 2021 NPR. To see more, visit https://www.npr.org.