“There are no facts, only interpretations”—also me,
because my interpretation is that I said this, not Freddy Nietzsche.
Being a scientist, as a matter of identity, is not about what you know. It is about the questions you ask, and the methodology you use to answer them. Talk to any scientist about what they do—this is probably what they’ll bring up first. Why? Because what you know depends in large part not only on what you’re interested in figuring out but also on what constitutes a legitimate attempt at doing so. With any methodology, there are things it can tell you and things it can’t. This also applies to the scientific method as a whole. Scientists, for example, are usually comfortable making the claim that science cannot give you answers to moral questions. Science can’t tell you about meaning, either. Meaning cannot be discerned through empirical experiments. Meaning requires interpretation.
This makes meaning almost fully intractable from a scientific standpoint, because this approach is very different from empiricism—which is how scientists are used to thinking. Empiricism, in the way that I’m using it here, means that whatever you want to know is best figured out through running experiments. Experiments can’t tell you anything about meaning. When it comes to meaning, the only thing experiments can detect are empirical trends in how people tend to report on meaningfulness. But that’s not studying meaning itself; that’s just outsourcing the labor of interpretation to someone else. If, unlike scientists, you are in the business of trying to understand what things means, then interpretation is the tool you must use.
Many people will be familiar with the famous distinction made by C.P. Snow about the two cultures—the sciences and humanities—which is often invoked in lamentation of how nice it’d be if we all just got along (or at least read one another’s papers once in a while). But asking scientists and humanists to share a common culture is like wishing that Argentina shared a common culture with Uzbekistan. They are separate cultures because they cover separate territories. They are different cultures by virtue of their interest in fundamentally different parts of the world—with different concerns, different ways of looking at things, and different stories how they came to do whatever it is they’re currently doing. It is in the nature of Argentina and Uzbekistan to be separate jurisdictions. The separation of the two cultures, likewise, is not at heart a distinction between two “cultures.” It is a distinction between two methodologies.
I think it would be okay if the two cultures/methodologies were simply distinct entities. (There’s no rule that says Argentinians can’t just travel to Uzbekistan if they want to see what the Uzbeks are up to these days.) But that’s not the case. Scientists get really, really offended by even the mere presence of interpretation. It’s as if they feel that once they let anyone with a different way of doing things into their territory, they’re suddenly going to be overrun with intellectual immigrants—and things will just never be like they used to! The scientists just can’t stand the idea of dealing with problems of meaning. It’s a real shame.
Interpretation is really cool, and I don’t know why you don’t want to do it.
More specifically: the people who use interpretative methods to get at questions of meaning are mostly individuals who scientists A) can’t stand, B) have never heard of, or C) think maybe are okay, but have absolutely no bearing on their professional interests.
The standard-bearer of Category A is Freud. There is no one who is more associated with psychology yet absolutely despised by those who practice it (the experimental version of it, anyway) than Vienna’s most famous fondler of the mind. But I think that’s a loss for the field. The entire premise of Sigmund Freud’s enterprise—hairbrained though it was (and which, to enumerate its hairbrainedness in full, would require whipping out a Borges-size map with the same dimensions as the territory itself)—was to use interpretation to address questions of meaning. What was the meaning of a dream? A slip of the tongue? A particular cultural institution? These are all questions Freud felt, deep in his tingly, Oedipal loins, that he could answer with the aid of interpretation.
My favorite candidate for Category B (the Ignored) is the anthropologist Clifford Geertz. Anyone familiar with the behavioral sciences will know the work of Danny Kahneman and Amos Tversky. They are among most influential researchers in the last half century of empirically-informed cognitive, social, and behavioral sciences. Their most influential paper (on Prospect Theory) has, at the time of this writing, around 75k citations. That is an F load of citations. But what about the Interpretation of Cultures, published at almost exactly the same time, by Clifford Geertz? More than 140,000 citations—almost twice as many as K&T! (And in a field where people publish papers at a MUCH slower rate.) Geertz is unequivocally foundational in the qualitative social sciences (e.g., anthropology, sociology, any field that does interviews, etc; if you’re not doing what he said you should, you at least need to explain why). He literally wrote the book on interpretation, how it might be used in a scientific context, and doing so might allow you to discover. Yet most cognitive scientists have never heard of him.
Category C (the Irrelevant) probably has some overlap with Category B, depending on whose work you have come across or happen to admire. But pretty much any literary critic or narrativist historian—one interprets books, the other events on the historical timeline—of good repute could be included here. For example, the critic Harold Bloom is pretty well-known as a scholar, and I know more than a few scientists who have read his work. Then there’s the Metaphysical Club by Louis Menand, a narrative-based history of ideas in American and perhaps my single favorite book I’ve ever read. That’s also a book that’s gotten good reviews from every psych-adjacent person I’ve convinced to read it. And maybe you’ve come across Jill Lapore in the pages of the New Yorker or on her Pushkin podcast. (Okay, now that I think about it basically anyone who writes for the NYer is in the biz of interpretation.) These kinds of people tend to be respected by science-y types. But you know what? So is Dr Suess! (Who is not actually a PhD.) All of the above are great at what they do; it’s just not science.
I would also submit that this is why another instance of Category A (the Reviled) are popularizers of science. Whenever I tell people something to the effect that I’m more interested in telling stories about science than doing science myself, many will respond, “Okay, so you want to do science communication.” Actually, no. Not at all! You know who is a great science communicator? Ed Yong. He is potentially the greatest science communicator in the game at the moment, and his goal is to communicate scientific findings as they are understood by the people who found them in a clearer way than the scientists themselves ever could. In other words, he’s not trying to go beyond the science—for example, to extrapolate the results of a study into a point about how you should apply it in your own life. He’s all about the facts. What most pop science writers do—the ones who many scientists really seem to despise, like Malcolm Gladwell and Steven Pinker—is not science communication, but science interpretation. They are not just here to tell you the facts. They’re here to tell you what they think the facts mean. And that to me is way more interesting!
But because meaning is not part of the scientific purview, no matter how these popular authors construe what they think of the facts—scientists will take issue with them. This doesn’t mean the authors are wrong. What it means is that they’re doing something that’s not allowed within the rules of the scientific game: interpretation. And so when what they’re talking about is represented as “science” the scientists are like, “Woah! Hold on! That’s not what we do at all!” Exactly. You’re interested in the facts. And that’s cool, we appreciate that. But the rest of us are also interested in something else: we want to know what the facts mean. For that, you need someone who is willing to wade into the murky, empirically-unverifiable waters of interpretation.
All this rests on the question “What is meaning?” — which I really don’t want to answer.
Okay, so I’ve made the claim that experiments can’t tell you anything about meaning—only interpretation can. Why? First of all, let me say this is not an easy question to answer. The humanities (and the qualitative social sciences) have been grappling with it for, like, a century. And honestly, I don’t think it’s gone uniformly awesome. Scientists have stayed away from problems of meaning. Their project has had some issues as well (many of which, by the way, are problems of underappreciating the importance of solid interpretations of their data). But with science, it’s easier to point to a trend that looks like progress!
The basic answer rests on my previous argument about how meaning is contextual. In short: there’s no such thing as objective, de-perspectivized meaning. And that’s precisely what science can’t handle that. It can’t study things in fun-sized, individual packaging. This one-and-done approach is labeled as “idiographic.” But given that that’s a term that even I, in my love of obscure and academic-sounding terminology, think feels ham-fisted and overwrought, we can call it N=1. This, of course, is scientific shorthand for something which doesn’t count as science. A study with only one participant isn’t a study. It’s an anecdote!
Let me try to express it this way: because meaning is always based on an individual perspective, all you have to do is change the person looking at it and the meaning changes as well. That’s, like, the main thing that you can’t have happen in a solid empirical experiment. Let’s say I construct an experiment designed to “detect” meaning (whatever that would entail). In science, you would want it so that I can show the results of that experiment to you, a different person, and you, too, can see the same results for yourself. Even if you had a meaning detector, it would buzz for different things depending on who was holding it.
By contrast, science—as typified by the grand, hold-the-universe-in-your-palm-sized-equation school of physics—is “nomothetic.” In the best case scenario it is N = all of the above. That’s what the scientific method strives for. You’d have your measurement, and even if everyone in the world had a look at it they’d all see the same thing. I know I’m not giving a very nuanced picture of the scientific enterprise here, but people write entire books on the basic concept of objectivity. I’m here to make extravagant claims now, ask questions later. So my point is this: the theory of everything would be a theory of everything but meaning. (Fine. Morality can come, too.)
Another important thing here is another thing I already said: meaning is post-hoc. There is no way to predict ahead of time what the meaning of something will be. It may seem like there is. But there’s not. We can detect empirical trends between what someone does and what they report as meaningful, but that’s not because it’s reflecting how meaning works. That trend, as I argued, reflects how we all have a propensity to reach for a fairly limited number of culturally-endorsed stories about what constitutes meaningful activity.
One of the things I tried to convey in that post was that the creation of meaning resists statistical regularities in a way that’s different from other activities. Because meaning is an interpretation—it depends on perspective; it is not objective; etc—the role that an event or experience has can change instantaneously. So don’t get me wrong: there are definitely things that we’re told via our culture that are important and that we should construct meaning from them—like a degree, or a kid, or a book that really changed the way we see things. But whether or not we actually do that is up to us. As a consequence, there are statistical facts about what people endorse as meaningful or not, just like there are statistical facts about anything that occurs in the world. It’s just that these statistical facts don’t necessarily have any bearing on what’s happening for any specific individual considering their own specific experiences. (Actually, I dunno. That kinda sounds like how statistical facts work in general? Maybe that isn’t so different. Still trying to figure that out, I guess.)
Anyway, what I should be really excited to argue for, but for some reason am not, is that experiments can’t detect meaning, fundamentally, because of what meaning is. Or rather, where it lives. Meaning lives in symbols.
This is the explanation given my favorite account, the one from Geertz. So what exactly is a symbol? Well, err... it’s a thing that contains MEANING.
These words—the ones I’m using here!—are symbols, within the symbol system “English.” A crucifix is a symbol within the symbol system “Christianity.” A Birkin Bag is a symbol within the symbol system “socioeconomic status” for some, while for others it resides in the system of “Huh?” A Balinese funeral rite is a symbol within the symbol system “Balinese culture.” One’s middle finger is a symbol within the symbol system “American diplomacy.”
Each of these symbols is meant to be interpreted in terms of its role in the larger system it’s a part of. When you leave the system—for example, when you don’t know anything about designer handbags or Balinese funerals—then the symbol’s meaning evaporates. (Oh, hey look! Meaning is still contextual!) In a way, this is the inverse of experimentalist’s ideal. Interpretations are only N=1. There are no N=all interpretations. You have to look at the specific symbol within the specific symbol system from a specific point of view within that system. So, yeah. The reason that I should be excited to make this argument is that I love symbols. The reason I’m apprehensive about it is that I’m not sure that I’m not sure I’ve explained anything or made things any clearer. I’ve been studying symbols—first as a cognitive scientist, then as a Geertz disciple; and the symbol “symbol” has slightly different significance depending on which system you’re in—for like a decade now. I’m STILL not sure what they are.
This is, perhaps, part of a larger MO. The thing is, I really hate definitions. I know it sounds ludicrous. But I really don’t think they’re helpful. I feel like when people starting arguing about definitions, then everything becomes about definitions and it’s like, for fuck’s actual sake. Who cares? And “meaning” is the worst offender. “What is the meaning of meaning?” is the only question more grating to me than “What is the meaning of it all?” It’s just not a game I’m interested in playing. So here. You want a definition of meaning? Fine. I’ll leave you with this.
Meaning... IS... interpretation...!!!
Everyone loves a third culture but no one actually wants to be part of one.
Louis Menand once wrote—in an essay I really spent a while trying to find but couldn’t and now have to move on because I have other things to do with my life… but here’s one of his NYer pieces on Freud instead!—something to the effect that there are really three cultures: the scientists/experimentalists, who are interest in what there is; the humanists/interpretators, who are interested in what things mean; and the social/behavioral scientists, who are interested in a bit of both.
But I actually think he’s wrong! I don’t think this “third culture” based on the behavioral sciences is actually really a thing in any substantive capacity! But even more than that, I think it should be! I think the scientists shouldn’t get angry when people try, in good faith, to overlay interpretation on top of empirical findings. I get why scientists feel it’s not their purview to do so themselves, but I think it should be part of the scientific pipeline: eventually, we combine our knowledge about what there is with an attempt at grappling with what it means.
I think maybe—ultimately, eventually, in the fullness of time—the overlap here has to do at the intersection of everyone’s mutual interest in symbols. Cognitive scientists, as I alluded to above, love symbols. It’s like their thing—what distinguishes the way they think about the mind from the way other thinkers, with similar interests, think about it. Gary Marcus has an entire Substack about how collections of neurons can create symbols but collections of neural networks can’t! (Okay, that’s not the Substack exactly, but it’s kinda the thrust of his book, The Algebraic Mind, from way back.) And computer science also has a definition, perhaps the most precise one of all, of a symbol. Computers can deal with meaning. Not, like, the full scope of it, in terms of crucifixes and Balinese funerals. But in terms of taking fundamentally meaningless, arbitrary components and combining them in such a way to create a system where symbols can be manipulated. It may not be the most flexible sense of the word, but commands in a computer programming language still mean something—at least within the context of the larger system they’re a part of. Even more exciting: You know whose job it is the do this interpreting? To put together meaningful arrangements from meaningless components? The homunculus eating popcorn in the Cartesian theatre of your MacBook? It’s called an “interpreter” :D
(Strictly speaking it’s not that simple... But it’s also not not that simple, either!)
Anyway, I do think this suggests a way forward for the third/culture. Meaning is interpretation. But what is interpretation? In a word, it’s an algorithm. Now THAT’s the kind of thing scientists can wrap their heads around.