A “Letters” discussion between Adam Gopnik and me on ways of knowing

February 16, 2021 • 10:30 am

For a few years I’ve had email conversations about “ways of knowing” with Adam Gopnik, who writes for the New Yorker and has also published several books. Our conversation has centered on whether science is the only way of knowing, or whether there are other ways of knowing as well. Adam defended the arts in this respect, and we’ve had some vigorous back and forths about whether music, painting, and literature in particular can be ways of knowing. I wanted to formalize our thoughts in a systematic way, and so I asked Adam to join me in a series of exchanges on the “Conversation” site of Letter. I just put up the first letter, which you can read by clicking on the screenshot below.

Our discussion critically depends, of course, on what we mean by “science” and “ways of knowing”. I’ve tried to define those carefully in my first letter, construing science rather broadly so that the topic becomes “the methods used by science” as ways of knowing. My position, which you’ll recognize if you’ve read this site for a while, is that yes, the methods of science are the only way of knowing. Religion is not, art is not, and ethics is not. I see philosophy and math as more circumscribed ways of knowing, since they convey knowledge about what holds within a system of axioms, but not about the universe. (I’m not a “mathematical realist”.) Of course philosophy and ethics are often informed by facts about the universe, and mathematics is an indispensable tool for understanding the universe.

Anyway, you can read my first thoughts (and definitions) in the first letter, posted on the site today (click on screenshot below). Adam will respond within a week, and we’ll each produce three or four letters in total, depending on how the conversation goes. I’m looking forward to this because scientists don’t usually get to engage in such a discussion with people who know a lot about the arts. I’m honored that Adam chose to join me in the exchange.


35 thoughts on “A “Letters” discussion between Adam Gopnik and me on ways of knowing

  1. Great letter. I look forward to seeing the exchange and if Mr. Gopnik has something more than an obfuscating cascade of ten dollar words to support his position (i.e. the typical post-modern defense of “other ways of knowing”).

    1. I’d appreciate it if people wouldn’t diss my fellow discussant here. I like Adam and consider him a friend and a thoughtful person. I will take his arguments on board and discuss them appropriately, but please don’t be rude towards either of us. Thank you.

  2. I’m very glad to see you doing this. I once engaged in a public discussion with a friend of mine in the art department, about the similarities and differences between science and art. I argued that science progressed incrementally with time whereas art was somewhat circuitous in its changes, with advancements, if any, due more to technological developments (from science) than from the actual art itself. And in agreement with you, art is not a way of knowing, but rather, a way of emoting, the natural world.

    I have sent your letter to my Honors class to read. I look forward to the rest of this exchange.

  3. Gopnik is one of the few New Yorker writers left who I still read with pleasure. He’s an extremely erudite man who stands up for liberalism and has a lot of common sense. These letters will be good!

  4. I think you’ve made an excellent start Jerry.

    To me, once you’ve defined knowing / knowledge and the methods of science as you have, appropriately IMO, there is no sound argument against the claim that the methods of science are the only way of knowing.

    The issue simplifies down to, if it hasn’t been tested against reality then you’ve got no reason to put any confidence in it.

    As you, I don’t think poorly of most “other ways of knowing” (religion is an exception) but any “knowledge” derived by them is a different category of thing from the knowledge we are talking about. Not necessarily less valuable either. The emotional experience of a moving piece of art could very well be more important to the condition of an individual than any knowledge derived from the methods of science.

  5. “There is only one world, the natural world; the world evolves according to unbroken patterns, the laws of nature; and the only reliable way of learning about the world is by observing it.”[Physicist Sean Carroll]

    Since I first came across this from Sean I’ve always thought it could be edited to give those with an opposing view less trivial things to argue about.

    For example, I’m not sure “world” is the best the term. Perhaps “reality” would be better. Or maybe not.

    I really think “the natural world” should simply be left out. It will readily be misinterpreted as an unwarranted premise.

    Trying to nail it down I guess the thing that bothers me is that the methods of science don’t require that there be only one world, the natural world, or unbroken patterns in order to enable us to ascertain knowledge about it. At least not the same patterns with the same consistency that science has actually discovered. Even if there were a “supernatural world” and even if the laws of nature were different, even if they changed, the methods of science would still be effective and would still be the only legitimate way of knowing.

    Sean’s description works fine as a description of what the methods of science have already adduced, which I’m fairly certain was the intention, but it will be taken as a premise by those with opposing views.

  6. Very nice.

    While there’s no one “scientific method”, those tools are familiar: observing reality, checking observations with other people, testing what your ideas predict, resisting your biases, and so on.

    I’d note that in some ways, the arts use this broad methodology, but they use it to do what you’re calling ‘procedural knowledge.’ E.g. a painter may observe the color of various paints and how they look when applied to canvas or mixed. They may check with others as to the quality of their work. They may test ideas for new painting styles, and successfully predict (or fail) whether it works or not. This process does tell them about the real world – not just people’s opinion of their art, but how, for example, paint and canvas combine. It may even discover regularities. But the big difference is that the regularities and the discovery isn’t the point of doing art; the art is. In that way, they may resemble engineers more than scientists; they seek to understand aerodynamics so they can build a better airplane, vs. the scientist who more likely wants to build an airplane to better understand aerodynamics. They seek to understand optics and the properties of their media to paint a better picture, whereas the ‘artist as scientist’ would be painting lots of pictures to better understand optics and the properties of their media.

    Still, the arts can provide insights later confirmed by scientific methods.

    I agree, but this is not saying much, as hypothesis generation is sort of the wild west “anything goes” part of science. Doing art can generate hypotheses because anything can, in principle, generate hypotheses. To cite the three classic scientific folktales, getting hit by an apple, having a dream, or sitting in your bathtub can generate hypotheses, so yes, doing art can too.This not to say these things are scientific methodology, it is to say that science’s methodology doesn’t dictate, restrict, or care how you came up with an idea; the methodology only really kicks in when you decide you want to test it.

    I ask those who claim otherwise to give me one bit of knowledge conveyed by the arts that could not have been derived using the only genuine way we have of knowing—observing the universe.

    Bound up in the concept of ‘ways of knowing’ is the notion of confidence. Drawing characters out of a hat might give you E=mc^2, but that method provides no confidence that you are right. What distinguishes the scientific way of knowing from religious introspection, art appreciation, etc. is not necessarily the conclusions it reaches – for these other practices can reach the same conclusions on occasion – it’s that the method of science provides mechanisms to estimate and increase your confidence in that conclusion.

    I mention that as a “be careful what you ask for” caution. There might indeed be the odd case where some artist does art and in so doing discovers some previously unknown true regularity about the world. However, it’s hard to say humanity knows it, until it’s been reproduced, tested, etc…. i.e. until (to paraphrase The Martian) we science the crap out of it. 🙂

    1. This reply is not necessarily only for the above comment. I’m getting sort of carried away :^)

      Art, music, literature, movies, poetry – these things are really about personal experience, and drawn from personal experience. These products exist in the moment – then they end in silence, in the dark. We can think back about them, but that is not what we mean by them. Their production I’d say is also about personal experience, that of the artist or subject.

      The products of the natural sciences as construed by PCC(E) can exist independently of personal experience. They are impersonal. The core of natural sciences is problem solving – asking questions about observations that still exist when we aren’t looking, and proposing precise solutions with clear predictions.

  7. Great letter – this will be interesting. “Other ways of knowing” is, though I am not a real philosopher, a symptom of essentialism, and related to the consilience of inductions. I expand a bit :

    Music as pure sound expresses something that cannot be expressed any other way. I like to think I “get” music from around the world, but this could really be off. Another listener, from a very different region, are likely to not “get” DA DA DA DAHHHHHH – even if they were in the concert hall, but who knows. And then some music sat around for a long time before it meant something to me. There is no correct personal experience, and that’s why we call it subjective.

    So I think the notion that everything is somehow delivering the same essential thing, because humans are producing it – producing what can be a personal knowledge – is a fantasy. There is every reason for these things that cannot be expressed any other way to mostly keep to themselves, and we are asking too much from them by asserting they are just “other ways” to get the same essential product. That sounds like the imposition of Plato’s essentialism (I learned this from Dawkins’ writing). The notion of consilience (W. Whewell, I learned about that from E. O. Wilson) also seems related to “other ways of knowing” :



    … but it is really captivating to think about how everything is related, or not – amazing here in 2021 to view over the centuries in this way.

  8. I tend towards a moderate “No”. However, it may be a matter of definition.

    The problem arises from conceiving a “science broadly construed” to which I agree, including the plummer, but excluding the artist. Let me chip away your argument by looking at representational art first. An artist is trying to create an accurate representation of some living person. She has somehow soaked up some feature of reality by observation and distilled it into a document that is more-or-less in “agreement with reality”. You’ll find cultural processes to more reliably produce accurate representations, and the creation of accurate and masterful representations make up a large part of what most people think art is (I have a broader view).

    We could use this painting as a “document of knowlege” and ask people to use the painting to identify the living person, much like a photofit to identify a suspect of a crime.

    Let‘s turn to more abstract art, like ancient cave paintings. Are they a way of knowing? The image itself, like a thesis, theory or formula can convey knowledge. It seems not far fetched to imagine a hunter describing an animal he had encountered to those who stayed behind in the cave. And they could literally gain knowledge by looking at it.

    One argument might be that the cave painting is merely a medium or an artifact, not art itself. But we are concerned with what it expresses, in much the same way as we would judge “On the Origins of the Species” not as a book, but what it says about the world.

    Another conundrum is that representations, often artistic, are also a feature of science proper. We may say they are “science” if the purpose is documentation of nature only, and then we have simply defined it away — but art can document, wheras science imagery must document. If Darwin was an artist, and his tree of life scribble made it into art galleries, would it cease to be knowledge? You see here, there is some blurriness or unclarity in the argument, at least to my mind. This seems more about rigour and expressed intentions rather than the form per se.

    I see “knowing” as some content of conciousness that is analogical to some feature of reality. “Knowledge” is shareable knowing through some means of communication. These means of communication are media, over which we have collaborative, creative control (if it was one-sided, we could not communicate, and not share the knowledge). “Truth” is a degree of analogy. Since our mind operates with categories, created out of analogies, truth has to do with fixity of patterns, or persistence of those analogical “links” across the category, making truth abstract. Art can express that abstraction.

    One aspect of art is a form of communication of knowledge where the communication is one sided, and exploratory. For the sake of illustration, consider it knowledge conveyed in a language the recipient does not understand but where the recipient can detect patterns that match with their own experience, such as “pulling out”, magnifying, re-categorising elements that were previously hidden, filling in missing information or through bisociation (merging or blending). I’d say in summary, art cannot be categorically excluded from being a “way of knowing”.

    1. “One argument might be that the cave painting is merely a medium or an artifact, not art itself. But we are concerned with what it expresses, in much the same way as we would judge ‘On the Origins of the Species’ not as a book, but what it says about the world.”

      I think what you’re describing here makes literally everything that exists a “way of knowing.” Every single thing in the universe can be observed and tell us something about the universe; as you’ve said, art can “communicate” things, but a process being a form of communication does not make that process a way of knowing, but merely a way of transmitting or extrapolating knowledge.

      I think the same is true of what follows in your post: it’s not that art is a “way of knowing,” but merely that, like a scientific paper or a mathematical formula, it can communicate something. For art to be a “way of knowing,” it must uniquely shed light to the artist on something concrete via its production. If art — cave art, for example — tells anyone besides the artist anything about the world, it’s merely another document, and the only way to know anything from it is to use scientific methods to document and analyze it, extrapolating from it (hopefully accurately, using scientific methods) new information. Art can express something, but it cannot, through its production, cause the discovery of something. It’s not a “way of knowing” anything concrete unless used and analyzed as a document, and then it’s the analysis that is the “way of knowing,” not the art.

      And, since I know this comes up: with regard to art providing the artist with a “way of knowing” the artist’s own feelings, or changing the artist’s perception, etc., the same could be said about any other process, including those that are very clearly inimical to the very idea of discovering truths about the world around us. Astrology, voodoo, homeopathy, and other processes and fields that deal in obvious falsities can still lead people to discover feelings, change their perceptions of things, etc., but that does not make them “ways of knowing.” If a process does not have the capacity to produce scientific knowledge, then it is not a “way of knowing,” except in only the broadest and least useful sense that would encompass just about all things, from jogging to typing a reply on a blog.

      1. Good reply. It brings one part more into focus, “ways of knowing” as “activity that produces knowledge” — can art do it, when plumbing can as well? I would also answer this with “yes” and here again the issue is perhaps rigour, not the principle itself. The plumber tries to fix something, thus there’s a science process, “broadly construed” at work that is about making a guess what might be wrong, and checking for it.

        But in my art example, the artist also can solve a “problem” for example, say, trying to create an accurate represenation of a patron, and they would use artistic means to solve that problem. Why would that not count?

        It’s not that I don’t see where Jerry is coming from. It seems he relies too much on defining “art” fairly narrowly and idealized, but more commonly science is narrow defined, and art is broadly construed. I would construe both broadly and see that overlap. Put in another way, “science” seems to mean some form of posing a question, and finding an answer through testing. Jerry seems to exclude artistic processes, because he doesn’t recognize them that way, but they can work that way.

        Remember, the issue I pose is that when plumbing is in the club, why is art categorically excluded.

        1. I think it’s because plumbing, by its nature, requires the discovery of concrete realities of the universe. An artist’s portrait of someone isn’t really solving a problem, but giving an interpretation of what is already known, no? The artist’s portrait may end up communicating some information to someone at some point, but that only means that the creation of the art itself communicated information, not that it discovered it, and I think the distinction between mere communication and discovery is critical. So, as I see it, there are two problems with the example of an artist painting a portrait: (1) there is no discovery of new information about the material world through the process of painting it (beyond, perhaps, the artist’s own aesthetic preferences and skills), and (2) even the portrait itself is a mere interpretation of the reality of the subject’s material existence.

          Regarding point (2): this is different from the plumber solving the problem of why a toilet won’t flush in that the plumber discovers something true about the material universe through a broadly construed scientific method. while the artist merely interprets and cannot even provide a completely accurate representation of the subject’s material existence. Figuring out that x = y + z is solving a problem, but an artist drawing flowers around the x in the equation because the artist feels that the letter x has a certain beauty that can be conveyed through such imagery is not solving a problem.

          I am enjoying this discussion! Thank you 🙂

          1. I think it’s because plumbing, by its nature, requires the discovery of concrete realities of the universe. An artist’s portrait of someone isn’t really solving a problem, but giving an interpretation of what is already known, no?

            It seems to me that what is tripping this up is reliance on a “why” question which is easily visible in the plumber example. “Why is the wall wet, even though the pipes seem intact?” whereas the artist seems to go with a “how” question, like “how to reproduce the effect of light that is hitting the hair?”

            However, the plumber also implies a “how”, namely, “how to find the water leak?” which expresses his goal. The artist also has a “why” question, which is buried in the process but still at play: “why doesn’t the representation look like the original?” for example. You need to find out a great deal about the universe to eventually create a faithful reproduction.

            Creative problems are called “wicket problems” and they are a type of problem where neither the issue nor the solution are known at first, but are figured out in tandem. However, not knowing exactly the nature of the problem to figure out (which includes learning something about reality through testing, adjustment and repetition) is not the same as not having anything to solve or to learn.

            I agree with your point made earlier that one gets quickly into a situation where a lot of things can be seen as “science, broadly construed”.

  9. An interesting twist in this discussion is how artistic tools like aesthetic judgement have affected, for good and bad, basic scientific thinking, particularly in fundamental physics. Here is a description of this phenomenon in relation to the book “Lost in Math: How Beauty Leads Physics Astray”, which is very critical of this approach:

    “Whether pondering black holes or predicting discoveries at CERN, physicists believe the best theories are beautiful, natural, and elegant, and this standard separates popular theories from disposable ones.
    The belief in beauty has become so dogmatic that it now conflicts with scientific objectivity: observation has been unable to confirm mindboggling theories, like supersymmetry or grand unification, invented by physicists based on aesthetic criteria”

    1. Of course, this is her opinion. She has some valid points, but sometimes tends to throw the baby out with the bathwater. She was kind enough to let me review her book in a guest post on her blog, even though the review is not wholly positive.

      1. Aesthetic criteria are useful for coming up with models of reality that are easier to reason with, lend themselves to more effective mathematical treatment, can more easily be taught, etc. So that is always a worthy goal to pursue.
        But assuming that the fundamental structure of reality should satisfy criteria of beauty, elegance, etc. is, in my opinion, just an unfounded metaphysical assumption

        1. I slightly disagree. I think modern science has a pretty good track record of discovering that complexity is very often produced by a set of fairly simple underlying principles or elements. Every proton, neutron, and election may be the same as every other, but mash them together and you get the periodic table and crazily complex chemistry.

          So the search for ever-more ‘elegant’ theories has some empirical support in that simple-giving-rise-to-complex appears to be a general trend. That doesn’t mean it’s going to be true in every case, but it’s probably a good way to bet.

          1. On the other hand, we have efforts like string theory (Ed Witten: “the math of string theory is so beautiful that it has to be true”) and supersymmetry which engaged a large number of researchers for many years and so far have nothing to show for their significant use of human resources

            1. Sure. Supersymmetry keeps failing. Still, we pretty much know the standard model can’t be all there is, because it doesn’t explain dark matter.

              AIUI the problem with string theory is the practical impossibility of building the instruments needed to test various versions of it, which is not the fault of the theory. It’s not like the universe came with a guarantee that all it’s underlying principles must be accessible via 21st century human technology. If it turns out we need Earth-orbit-sized detectors to find out the underlying principles of the universe, well that really stinks for us, but it could be the case.

              In any event, my point was that positing the standard theory’s jumble arises from some simpler set of principles or positing an elegant theory of all particles and forces isn’t an “unfounded metaphysical assumption,” because it’s foundation is a long history of past empirical observations discovering that complex behavior often arises from simple laws, forces, and particles.

  10. I enjoyed this very much and am glad. I consider this to be a very important topic of discussion and am glad you’re advocating for what I consider to be the only objective answer to the question.

  11. How fortunate that we can read these letters. I look forward to them and the start you provide is chewy, smart as usual and the result of thinking hard about this.

    I’ve long been an admirer of Gopnik–whip smart, considerate, a range that seems without a border. Carroll’s definition of naturalism is one I like and have used. It made me think about a piece he wrote on Poetic Naturalism (writing them for Preposterous Universe) and I suspect you’ve seen it but I copy a bit of his short discussion on what he means here:

    “I like to talk about a particular approach to naturalism, which can be thought of as Poetic. By that I mean to emphasize that, while there is only one world, there are many ways of talking about the world. “Ways of talking” shouldn’t be underestimated; they can otherwise be labeled “theories” or “models” or “vocabularies” or “stories,” and if a particular way of talking turns out to be sufficiently accurate and useful, the elements in its corresponding vocabulary deserve to be called real.”

  12. Your next debate should be about the accusation that music theorist Heinrich Schenker’s teachings about music theory infer his racist political views. This raises the interesting but contestable issue of whether the symbolic language of music (pitch primarily) conveys meaning just like utterances made using words which convey thought. Secondary issue is whether music conveys emotion, i.e. whether composers felt emotion or intended to express them in their music, as opposed to listeners forming their private emotional states from music…as if these were cause and effect. For those who infer politics or emotions from music, they are in the same company as astrologists and fortune tellers. As such the burden is on them to prove any correlation apart from their own subjective emotions.

  13. Cave art, drawings, give a little insight to homo sapien cognitive ability 40 – 6o thousand plus yrs in the past. As we saw just recently with a piece of cave art it showed what animal resided in the area and now extinct and all that means… environmental conditions, it’s diet say.
    In cosmology, the Hyperfine splitting of the ground state of a hydrogen atom has been sent out into interstellar space (early 70’s) within the 2 Pioneer spacecraft as a diagrammatic representation as a way of basically, showing we are a intelligent being. It’s not art (or is it?) but superfine splitting is considered a universal truth which is why these drawings were on board.
    In this, does it show art is a way other than pure optic pleasure, to understanding, convey facts but not how to achieve that, only science can do that. At best, art is a tool of science?
    I’m not sure if I’m even on track with this comment and certainly if you’re an alien 👽 would you consider the Pioneer diagrams useful knowledge for further investigation, reply here.

    1. His last word will be his concession speech. 🙂

      Jerry is clearly right. You can have all the ideas you like about how the World works, but if you don’t test them against reality, you don’t know anything.

      Take Euclid’s elements as an example. He starts with five geometric axioms and from them deduces all sorts of exciting things like “the angles of a triangle add up to two right angles” and “the square of the hypotenuse is equal to the sum of the squares of the other two sides in a right triangle”, but there’s no information at all about the real world in Euclid. In fact, the whole of Euclidian geometry is technically a tautology. All the theorems and everything are inherent in the five axioms. This is true of all maths.

      You might then have an idea that the Universe obeys Euclidian geometry but, until you test it, you don’t know. One way to test it might be to draw some triangles and measure the angles to see if they add up to 180 degrees and at human scales they do. Now you have learned something. Of course, you also have to accept that your triangle drawing and measuring equipment have physical limitations and also curved things look flat on a small enough scale and therefore your knowledge might later be modified.

      1. Interesting, yes.

        There’s usually a big deal made about how mathematics and natural sciences work so well together – one is said to be a “queen”, or is unreasonably effective – and that’s all well and good.

        But if something works with some other unrelated thing, it doesn’t matter if it is a mathematical proof or Aztec pottery – it works. Mathematics works exceptionally well with natural sciences, period. There is nothing that meant it was supposed to.

        I think that leads to confusion – that somehow, some way, e.g. Bach is supposed to be important for cosmology. While that is a grand romantic fantasy, and cosmologists might derive important personal experience from Bach performances, it is an enormous confusion to claim that the products of cosmological inquiry and the personal experience of sound waves are equal ways of knowing, or that the things learned from either inform the other, or that the two forms of knowledge are different variations of the same essential thing. The two examples I picked (it could be any other) have also had plenty of time to show some connection – some sort of mystical, deep, meaningful relationship between the function of quasars and the counterpoint of Bach (remember the book Godel Escher Bach?), but it hasn’t happened yet, and it won’t. Not that it wouldn’t make a grand fantasy story, though.

        By the way, was the Gougu / Pythagorean theorem published first in Elements?

        1. I think Euclid’s Elements has the first known general proof of the Pythagorus theorem. I suppose technically, we should call it Euclid’s Theorem but we call it Pythagorus because it’s always been called Pythagorus in the European tradition, much like Fermat’s Last Theorem (featured in GEB) is technically Wiles’ Theorem.

  14. During the Renaissance, today’s cultural divide between science and art did not exist. The first academic post of the young mathematician Galileo Galilei—who later on created the subject we now call Physics—was in 1588 as an instructor in the Accademia delle Arti del Disegno in Florence. Galileo taught the subjects of perspective and chiaroscuro. In those days, these subjects were not conventionally assigned to one or the other of two separate, mutually exclusive, realms.

Leave a Reply