This blogpost grows out of a number of recent conversations about “science”– what it is, how to do it, and why. Whenever my research involves truly boring, tedious things (like scoring hours of video footage), my mind starts to wander off to all kinds of such philosophical things.
Y’know, big picture stuff.
(Not that big picture. Not like, why are we spending our short time on the planet earth running from failure and chasing after success when the main thing we treasure in life is to love and be loved and to make the world a better place? Not that. I mean “big picture” compared to my typical, extremely specific topic: why do unrelated female common vampire bats regurgitate into each others’ mouths?. Hope that’s clear. Now onwards.)
Last year I came across the book, Ignorance: How it Drives Science, by neuroscientist Stuart Firestein. According to Firestein, science is about ignorance, not knowledge.
At first glance, this idea seems like pseudo-profundity, like when bad philosophers simply play around with words and their meanings. But I read into it more and I experienced that lovely sensation when, finally, someone clearly expresses some truth that’s been imprisoned vaguely in the back of your mind for so long. To get the central thesis in 20 min, see his TED talk:
Ignorance, when considered as “what we don’t know” is not such a bad thing. Indeed, for most scientists, it’s really a great pleasure to sit around and wonder about all the things that currently make no sense whatsoever. But this is not just a game for working scientists. In today’s world, where the internet and wikipedia is only an arm’s reach away, one can pretty quickly get close to the limits of collective knowledge (at least compared to when I was in middle school and had to pull out the ole Encyclopedia Brittanica and search for that single paragraph). Even if the answer to your question is not easy to find in the peer-reviewed literature (say because it’s behind paywalls), you can get the email address of the world expert, or anyone with the right technical knowledge, and just ask them (and pray for an answer like this). By the way, you should click on that link.
Obviously, a lot of stuff has been figured out over the last few hundred years of scientific investigation. But a curious person also quickly finds that, for some of the questions, nobody knows the answer. For newer fields like biology or the social sciences, some of the basic questions that kids will ask (Why do cats meow? Why do dogs bark?) don’t really have complete, satisfying answers. And a surprising number of the known “facts” (the starting points of many questions) are themselves not true or are based on very bad evidence.
Older fields, like physics and chemistry, are less open to questions and participation by the naive masses (ironically, much to the envy of the rest of academia) because what is known in physics is so deep, rich, and vast, and the active research questions are so technical. As a college freshman, I went rock climbing with a graduate student in physics. I asked him, “Is a flame just gas that’s glowing?” and which led to “What the heck is plasma?” which led to “Ok, but what is matter, really?” which sparked a 20 min question and answer session. Interestingly, he started answering most of my questions by saying “Well, nobody really knows for sure what that is, but here’s what we do know…”. I remember that well because it changed something fundamental about how I think about science. The guy didn’t seem to care or be shocked that I knew absolutely nothing, but he kept saying things like “that’s a really good question.” That’s not what my high school physics teacher ever told me. And that’s maybe how I decided that I liked science afterall.
So with a little effort, almost anyone can find real, legitimate scientific mysteries and enjoy wondering about stuff nobody understands yet. Here are some questions people (mostly young people) asked me at the Great Lakes Bat Festival: “Do vampire bats really like cow blood more than human blood? Do vampire bats make a special sound to say they are hungry? Do they feel bad if they are rejected? Do they get in fights about being rejected?” It’s fun to think about how one would answer these questions. But you have to start by admitting you don’t really know the answer.
I’m no expert on educational psychology, but it’s long been my opinion that one problem with science education is that we don’t teach science students to say “I don’t know.” Science classes often emphasize facts, and this gives the false impression that science is about memorizing lots of these facts. Scientists are people who have collected so many facts that they have eventually run out and need to do experiments to gather more facts.
But that’s not quite how it works. Instead, as Firestein explains, the more you know, the more questions and less certainty you have. The Dunning-Kruger effect tells us that real competence can weaken self-confidence, and that incompetence leads to overestimation of one’s own competence (so when I don’t know something, I probably don’t even know that I don’t know it). But I believe this cognitive bias is further exacerbated by a tendency for teachers to implicitly punish the “I don’t know” answer, in favor of the “here’s a bunch of related information” answer.
You will recognize what I’m talking about if you have ever been a graduate teaching assistant and spent hours sitting and reading essays on college science exams, where some answers are written by students having no idea of the answer to the question. Sometimes it takes awhile to realize this. It is painful. From the perspective of the test-grader, the easiest thing to grade would be a blank answer. But nobody encourages that. From the perspective of a test-taker, the rational thing to do when you don’t know the answer is to give it your best shot. That is, you make up an answer. Perhaps throw in some words that you remember hearing from the lecture.
For example, take the following science question: What exactly is happening inside a leaf when it becomes red or yellow in the Autumn?
Leaf changing color. Photo by Nickel Eisen.
Now, here are 2 imaginary student answers.
Student 1: Leaves are green because they contain chlorophyll, which is a pigment. Chlorophyll is necessary for photosynthesis, which is how a plant generates energy. Chlorophyll is green because it absorbs other wavelengths of light such as red and blue. Red light has a wavelength of 620-750 nm. Chlorophyll is contained within chloroplasts within plants cells. When leaves change color, they are losing their chlorophyll. This is because the plant is losing the leaf during the fall. Plants lose their leaves during the fall to conserve resources. This is important for the plant’s survival and reproduction (its biological fitness) and for the evolution of leaves more generally. Both angiosperms and gymnosperms have leaves. Leaves have multiple types of cells, such as guard cells, which function in photosynthesis. During photosynthesis, plants produce sugars using CO2, water, and sunlight. Photosynthesis consists of both light and dark reactions. [This could go on further but you get the point.]
Student 2: I don’t know. Chlorophyll must somehow be lost from the leaf. Maybe yellow and red pigments have been there already and were hidden before but then are revealed by the loss of the green, or perhaps the reds and yellows are created within the leaf. Maybe the red and yellow pigments are actually brought in from elsewhere when the leaf is ready to fall? I don’t know if the yellow and red pigments even serve a function. Maybe they are just byproducts.
So… neither student actually knows the answer to the question. But in my opinion, student 2 is thinking more like a scientist, because she knows that she doesn’t know the answer, while it’s not clear if student 1 is just desperately listing facts that seem relevant or if she thinks she has actually explained the answer. The sad thing is that student 1 would probably get more points on an exam, if the grader is skimming for technical and salient words and phrases such as “photosynthesis”, “chlorophyll”, “chloroplasts”, “pigment” or “absorbs wavelengths of light”. Often, the main thing that matters in a science class is what you can recall from the lectures and reading, not whether you are thinking critically or scientifically.
Most of the time, scientists do not ask questions and get answers. They ask questions and get more and better questions. Look up the literature on red and yellow pigments in leaves and that’s what you’ll find: some answers, but mostly, better and more precise questions. Yet we are often training students to regurgitate facts, without ever admitting the central importance of ignorance in science or “better ignorance” or even asking questions of their own. And we rarely teach students to say that they don’t know.
Is this really a problem? I don’t know.
A related problem is the myth and allure of the single driving explanation when describing scientific research, especially in biology. If you want to make a good story out of research, it helps to have a clear question and a clear answer. In biology, some questions have clear answers, but the most interesting ones actually have a staggering number of correct explanations, which are all confusingly entangled. Take my question: Why do common vampire bats regurgitate food to non-relatives?
Is it because they usually regurgitate food to relatives and offspring? Is it because feeding non-relatives increases their chances of receiving a food donation later? Or is it because their blood diet and metabolism make them so susceptible to starvation? Is it to honestly signal their ability and intention to help, or perhaps to manipulate other bats? Or is it because they are vampire bats and all 3 species of vampire bat share food with non-relatives? Or is it because vampire bat stomachs can hold so much blood, and they have more extra to give the more they have? In other words, is it simply because of diminishing returns? Is it because vampire bats have special stomachs that facilitate regurgitation, or because blood can’t be easily carried back to young like a prey item? Is it because each vampire bat is more likely to help another bat when any individual helps them regardless of whom?
My answers would be yes, probably, probably, probably, probably, yes, yes, yes, probably yes, and maybe.
It sometimes seems like no question in biology has a single, simple explanation. In a nutshell, biology is complex. (To a biologist, even a nutshell is very complex. I just yesterday listened to an hour-long talk by visiting researcher Dr. Amy Litt on the evolution of the molecular mechanisms underlying the development of dry fruit walls, like nutshells).
So how do we deal with the confusing chaos that every question in biology has so many multiple correct and entangled answers? How can we know when two questions are alternatives rather than complements?
In 1961, Ernst Mayr clarified our thinking in biology in an essay about cause and effect in biology. He popularized the clear distinction between two very different kinds of questions and their corresponding answers. He labeled these, proximate and ultimate. Proximate explanations are answers to mechanistic “how” questions. Like, how does a bat produce sounds that are ultrasonic? Ultimate explanations are evolutionary “why” questions. Why does a bat make calls that are ultrasonic? Hence, proximate explanations and ultimate explanations can never be alternative hypotheses.
In discussing animal behavior, Niko Tinbergen later split these 2 categories into 4 levels of explanation. All organismal biologists know this framework, and this was the first lesson I learned in Intro to Animal Behavior as an undergraduate. Proximate causes include both developmental answers (how does the behavior emerge during the animal’s lifetime?) and mechanistic answers (how does it actually work in real time?). Ultimate causes can relate to phylogeny (when did the behavior first evolve in evolutionary time? and where in the evolutionary tree?). Or they can relate to function (what is the evolutionary advantage of performing the behavior?).
But it doesn’t stop there. Questions at the single level of evolutionary function, for example, can relate to its origin, why the trait first evolved, or alternatively, function can ask what maintains it now. These are also clearly different questions, not competing hypotheses.
Even when considering the more specific question, “what is currently maintaining an evolved trait X?” — one has to consider that there might be several selective pressures acting across different species or situations, or perhaps in all situations simultaneously, perhaps even interactively, such that the presence of one factor determines the importance of the others. In a statistical model, there can be the interesting 2-way interaction (e.g., the ability of X to predict A depends itself on factor Y), the slightly yucky 3-way interaction (e.g., the ability of X to predict A depends on Y to some variable degree based on Z), and the dreadful 4-way interaction (e.g., the ability of X to predict A depends on Y to some degree based on Z…oh, and forgot to mention– that whole thing I just said is true to some degree based on factor Q. Got that?). There are even 5-way interactions and beyond, but to hell with trying to understand those.
To make matters even more complex, proximate mechanisms are not isolated from ultimate mechanisms. Indeed, they often help determine the dynamics of selection that give traits their function. Proximate explanations (mechanisms and development) also put constraints on the directions in which evolution can possibly go (and hence what kinds of ultimate explanations are possible). This leads to so-called reciprocal causation, where proximate mechanisms help drive ultimate ones.
So attempts to cleave a clear framework in biology are soon muddled by trying to interweave more and more reality, and hence more complexity, and hence more potential confusion. Below is a recent attempt to illustrate cause and effect in biology in an easily grasp-able framework. The authors were arguing that the red arrows were not previously recognized enough by previous generations of biologists (whose blue-only frameworks were too simple):
from Laland, K. N., Sterelny, K., Odling-Smee, J., Hoppitt, W., & Uller, T. (2011). Cause and effect in biology revisited: is Mayr’s proximate-ultimate dichotomy still useful?. Science, 334(6062), 1512-1516.
See, we added some new arrows, and it’s still simple, right?
The practical solution to complexity is reductionism. I know that reductionism is normally used as a dirty word. But reductionism to me is actually quite beautiful and under-appreciated in how it allows us to move beyond utter stupefying awe when looking upon something that is holistically complex. By reductionism, I mean breaking big concepts down into components in a hierarchy of levels, which allows each level to be understood in terms of black boxes, which themselves are reducible to other black boxes, and so on. The complex interactions between parts at one level or even different levels can give rise to the emergent properties that seem to just magically appear at the next higher level.
The whole of science is organized in this pleasant way. A community ecologists can tell you about populations of different bats interacting, a population biologist can tell you what a population of vampire bats is doing, and an organismal biologist can tell you something about what an individual vampire bat is doing, and a neuroscientist can tell you what its brain is doing, and a cell biologist can tell you what the neuron is doing, and a biochemist can tell you what the cell membrane is doing, and a chemist can tell you what the molecules are doing, and a physicist can maybe tell you what the carbon atom is doing. And a particle physicist can tell you what the heck a Higgs boson is, because apparently they found one, whatever it is, and how that gives particles mass…or something [?]. My point is that somebody out there understands each part, even though nobody understands the whole big thing. So science is collectively wiser than any one scientist.
In this way, science is both a form and product of collective intelligence. Matt Ridley gave a pretty insightful talk on this principle as applied to cultural and technological innovation, which he illustrated so well by the simple observation that no single person knows how to make a pencil: