"Most learning is an illusion"

Jack Cohen occupies a peculiar position among scientists: In addition to his scholarly work, he has published several books in collaboration with the fantasy author Terry Pratchett. Cohen sat down with Martin Eiermann to discuss the link between science and science fiction, the nature of knowledge, and the decline of university education.

The European: Your most recent book – with Ian Stewart and Terry Pratchett – probes a very fundamental epistemological question: How do we know what we know?
Jack Cohen: The whole idea of what constitutes an explanation is very difficult. You might be a science journalist, so if I wanted to explain the origins of a storm to you in a scientific manner, I would assume that you know certain basic terms and concepts like electrical charge or the saturated solution of water in air. But most people don’t know such concepts. I always found it very rewarding to talk with my graduate students because their knowledge was very much like mine, so we could assume that we had much in common. It’s very difficult when you’re faced with an audience that you have less in common with; it’s very difficult to choose the words that will allow them to walk away with the illusion that they have learned something.

The illusion of having learned something?
Most learning is an illusion, yes. Take my partner, Persephone. She has been learning a lot about science since we started dating, but she doesn’t have the background, all the “ifs” and “buts” that I would have in my mind when I am thinking about something like consciousness. I am actually very pleased and amazed by the amount of understanding that we can achieve between people. But whenever I actually sit down to write something, I think, “bloody hell, there’s no way to communicate this.”

Because many things that enable communication are left unsaid: rules of grammar, but also implicit and basic understandings about how the world works. We don’t have to explain gravity to each other to be able to say, "I’ve dropped something.”
Let me give you an example: The story of word processing began with people chipping marks into stones, continued with ink on paper, and has now resulted in computer programs. It’s easy to explain that history because you can point to the evolutionary series and people will generally say, “yes, that looks plausible.” But developments often aren’t quite as neat, especially on evolutionary scales. Stephen Jay Gould introduced the term “exaptation” to describe shifts in the function of a trait over the course of evolution. An adaptational line might start with a reptile that eventually turned into a bird. But there’s a complication: feathers existed before birds could fly, they were almost certainly for warmth or sexual display. Something happened that changed the adaptational line into a different direction. Or take a microwave oven: It started out as a microwave transmitter until an engineer noticed that it would melt the chocolate bar in his pocket and thought, “we can use this machine to heat food.” That’s why the intelligent design people are wrong. They say that once you take away a certain trait, an organism stops working, so it must be a carefully designed thing. The answer is of course that it was a different organism before it acquired that particular trait.

One of your ongoing projects has been to complicate common understandings of evolution…
It’s wrong to tell the story of evolution by saying, “once there was a nerve cell and eventually it turned into Einstein.” We want to say: Once upon a time, creatures gave yolk to their offspring to give them a head-start in life, and that process has continued until today. We are very privileged apes. Lots of people thought of evolution before Darwin. His big contribution was to describe mechanisms by which evolution occurs – and those mechanisms are often quite complex. But it’s such a beautiful simple narrative to say that once there was a nerve cell and it became cleverer and cleverer and cleverer.

You once wrote that we should think of ourselves not as homo sapiens but as pan narrans, the storytelling ape.
If it’s a good story, we’ll accept it as an explanation. We accept the statement and forget to look at the fine print. But of course the universe isn’t like that, it doesn’t have a narrative that goes from A to B to C to D. The universe is multiplex, everything is caused by all the things that came before.

This goes back to the illusion of understanding: The power of narrative can outweigh the power of evidence. But the solution seems not to be to get rid of stories, but to tell better, more accurate, more contingent stories.
I like that. You cannot avoid narratives, but you can think about the perspective of a narrative. In Figments of Reality, we begin by telling the history of the universe in a few pages. And the final bit of the book is an alternative history of the universe as seen through the eyes of an alien civilization. Ian Stewart and I had all the chapters written but they didn’t fit together. We showed the book to Terry Pratchett at a Mongolian restaurant and he said, “it’s easy. What you need are two different ways of thinking about the origins of the universe.” That’s what started our collaboration. Terry said, “I have played in your garden, why don’t you play in mine?”

What is the appeal of combining science and science fiction?
It’s not science fiction! Science fiction would be difficult to combine with science. Scientific explanations of Star Trek or Middle Earth never quite work: If you want to explain how a spaceship can go faster than light, you could come up with some new equations, but I’m not interested in that. It’s not real science. But Terry writes spoof fantasy. It’s not even real fantasy with real swords; it’s about a world in which swords magically disappear. The world that Terry describes doesn’t function according to the laws of nature, it functions according to narrativium. That gives us the chance to say: Let’s imagine that the Wizards in Discworld create a separate world wholly without magic – the Roundworld – that functions according to what we would call the laws of nature. There we can talk about science.

You have said that because physicists study matter and energy, and because everything is made of matter and energy, they think that they can answer any question. What questions cannot be answered by physics?
Questions about complexity. Richard Feynman once wrote that the “goup” in a rock is the same as the “goup” in a frog, only arranged differently. That’s how physicists think: Because it’s all atoms, it must all be the same. They are quite amazed when you point out that when you arrange matter in new ways, it acquires new properties. When you arrange matter into a beetle, for example, it can fly. When different organisms evolve together, you get complicity, the rise of new properties that aren’t predictable when you look at the organisms in isolation.

Is genetic engineering adding further complications?
We cannot genetically engineer genes yet, we can genetically modify them. But the whole story of genes has changed so much even in the last five years. We used to think that the genes were what mattered and the other 95 percent of the genome was mostly junk. Now we know that it’s producing RNA, which determines how genes are turned on and off. Genetics has largely become epigenetics. The picture is much more complex than we thought.

In business or journalism, digital technologies have proved incredibly disruptive and have really accelerated the pace of change. Can the same be said for the impact of modern genetics on biology?
Epigenetics has changed the whole story. A particular piece of RNA in two cells might be doing opposite things in each. That’s very hard to understand, but it’s part of the complexity of life. By looking for simple stories, we often under-estimate how complex life actually is and we convince ourselves that we’ve understood it when we haven’t. Because most things are multicausal, it’s often impossible to talk about it in ordinary language. Sometimes you can talk about it in math, but even math is a very poor tool. Unfortunately the universe doesn’t restrict itself to things that can be described by mathematical equations.

Is there a better tool?
Math is alright and it might be the best tool we have at the moment. Let’s not forget that we are only halfway from homo erectus to human. Why do we expect to be able to do all the things that humans will eventually be able to do? We’ve had modern science only for a terribly short period of time, four hundred years at most. Can you imagine what science will be like when it has been around for a million years?

Do you think that some questions will remain unanswerable?
I like to think that there’s always another door that must be unlocked. But who knows what the world might be like in a thousand years or a million years. Look at something as simple as the mobile telephone, which has changed life for everyone on the planet and particularly in developing countries. Let’s say that we have three major inventions per century – very soon you’ll end up in a very different world. And that trend is probably going to accelerate.

Because of Moore’s Law?
Yes. It seems to be true. We see accelerating change in many different areas. Moore’s Law is a bit like Zipf’s Law in that regard. The linguist George Zipf discovered a very strange relationship in prose: The second-most common word in a text occurs about half as often as the most common word. The next word is about one third as common, and so on. It makes a very nice mathematical curve. The same relationship can be observed in the size of cities and lakes, for example.

When we look at nature, it often seems chaotic and random – but we also observe these hidden structures and an underlying sense of order that has, quite literally, mathematical precision. In other words: Complexity may merely be an outward appearance.
I think there must be some basic principles. But I don’t know why these laws or rules should exist. When the universe created itself, did God reach into a cupboard to pick out a set of laws? I don’t know.

One of the oldest explanations for the structure of the universe is the anthropic principle: The universe is the way it is because that’s what allowed us to evolve and observe it. But that seems like a deeply unsatisfying explanation.
It is, because it’s not true that intelligent life would be impossible if the physical constants were different. You could move several of the constants by quite a bit and the resulting universes would be different, but they would work. Why this particular universe? I have to throw my hands up and say, “bollocks, I don’t know.”

“I don’t know” – that must be one of the most underused phrases in science.
Scientists should be much more willing to say that they don’t know, especially physicists. Stephen Hawking once said that he could look into the mind of God, and all the theologians jumped up and got very excited but he obviously meant it in a different way.

Do you believe in a higher power?
I was brought up to be a rabbi, and the only God I can think about is a very impersonal God, a Spinozan kind of God who is not anthropomorphic at all but more like the rules of nature. But what do I know? I’m a scientist, I know nothing about most things. Not about the important things, anyway. Ask me about sperm or birds or amoebae, and I’ll give you an answer. But ask me about racing cars or a large number of other things and I won’t know what to say. And I’m happy with that. I’ve had eighty years to get used to it.

For a long time, scientists used to be generalists and not specialists. In premodern Germany, the ideal was the Universalgelehrter, the person who commanded knowledge in many subjects.
Specialization is a modern phenomenon and it’s a great pity. We have too many specialists today, even though almost everyone says that we don’t know enough about their particular branch.

I spoke to Stephen Wolfram a while ago, and one of his arguments was that we can gain great insights by applying ideas from one area of science to another. For example, he tries to tackle biological or even philosophical questions by thinking from the perspective of computation.
The danger is you describe everything as the result of simple sets of rules. It gives you a simple story, but you’re losing the complexity – which is what makes the world interesting. The “theory of everything” is only wonderful if you’re looking for something that can be printed on a t-shirt.

But don’t you think that there’s value in the breaking-up of established discourses? Metaphors might not have great explanatory power, but they might open your mind to new interpretations.
Yes, that can’t be bad.

You’ve said that we should reject most explanations we’re presented with because they are most likely wrong. Why?
Doubt is the strongest principle in science. Every statement should be regarded critically. Popper says that things can only be disproved, not proved. But most things are actually very hard to disprove, because you get all kinds of “but if you do this…” arguments. Until you have several convergent lines of evidence pointing to the same conclusion, you should be very critical. Michael Shermer has written a book called “The Believing Brain,” in which he argues that our brains are actually very good at forming beliefs. The brain believes almost anything! It takes sensory data and imbues it with meaning: As an infant, you get fed by your mother, so your brain believes that mommy is herself a source of food. Only when you look at alternative evidence can you begin to discard some of those beliefs.

The history of modern science is often told as the history of the triumph of reason over superstition. Doubt doesn’t feature too strongly in that story.
The average person sees science as the cumulation of facts, as a pile of statements that are true or nearly true. But take theoretical physics, for example: There’s a satisfactory way to explain the puzzles we see when we twist matter about, but I’m not convinced that W-particles and Higgs Bosons are the best explanation. If you want to get notes out of a piano, you can throw it off a tall building and it will shatter and make a sound. But that’s a rather silly way to make music. One problem is that scientists spend fifteen years of their lives trying to find a particular particle or solve a particular problem, so when they see something that fits, it’s tempting to say: “Bingo, we’ve got it.”

Isn’t it the zeal and obsessive pursuit of a hypothesis by scientists that drives science forward?
I believe it is. I spent twenty years looking at sperm, and I proved to my own satisfaction that there are only a few good ones among two hundred million. It was published in the best journal, but no-one has taken it on board. I don’t know why – I believe it’s a question of technology and magic.

What do you mean?
People accept the technology but are very reluctant to accept the magical story that goes with it. They accept the facts of sperm reproduction but not the interpretation, because the underlying biology is so incredibly complex. The more complex the science is, and the more advanced the technology gets, the less distinguishable it is from magic to most people. So the stories that often get told about science aren’t the stories I would want to tell about a particular piece of work. Scientists shouldn’t shy away from complication. Dan Dennett is someone who does this very well. He says that most things we cannot know, and that beliefs should always be left out there to be criticized. I would like to think that I doubt everything, but of course I don’t.

Because it would be paralyzing!

Do you think that scientists have a responsibility to take their knowledge beyond the ivory tower and communicate it?
I certainly think that we have far too many scientists with half the population going to university. I would be happy if two or three percent went to university, because then you could have words like “excellence.” So many universities are simply bums on seats in the lecture hall and it’s expected that the professors don’t let anyone fail. I let six students fail because they were better off doing other things – in one case, committing crimes – but that was considered unacceptable. The politicians think that by putting more people into universities, they can produce more Einsteins. I’m mostly annoyed by politics. Politicians seem to be mainly concerned about getting elected; you can hardly find competence in any of them. The competent ones are the odd ones out.

Isn’t the quest for broad exposure to university education really about increasing the chances of young people to have a good life, to have social mobility?
When I was teaching at Birmingham, we had a tutorial system. I had a handful of students, I saw them for an hour a week, they read their essays to me and I heckled them. That was very useful. But that’s not the situation at most universities today. One of my PhD students back then was a guy named Arthur Jones, who is a noted creationist, and quite a lively spark among them. My colleagues were furious that we gave a PhD to a creationist and said, “he’s doing all the wrong things.” But I was delighted. The first essay assignment I gave him was on the Omphalos hypothesis, the idea that God created an old earth about six thousand years ago. The apparent age of the fossils and the radioactive decay would probably be right if we assume that God is pretty good at this kind of stuff, so all the science still makes sense – what’s different is the origin story.

It’s almost like the Discworld/Roundworld scenario. Creationists can say that God created a world that has its own laws of nature. You don’t need to invoke divine intervention to explain it. But that’s a bit of an absurd idea, isn’t it?
Of course they are wrongly creationists. But I suppose I don’t think like that. One of my students became president of the Royal Society, another invented DNA fingerprinting. I think I’ve done alright. And I still get a lot of joy out of teaching: Out of the moment when students see a live amoeba for the first time, when they understand something that they haven’t understood before. That’s a great pleasure. I’m reading a book by A.C. Grayling at the moment, and I get a lot of pleasure from that, too. Or look at the stream at the edge of the garden here: It’s one of two streams I know in the country that has a particular type of freshwater shrimp. I was delighted to see that when I moved here! There are many pleasures in life, from sex outwards. I try to find little things that give me great pleasure.


comments powered by Disqus

More Conversations

“It is impossible to return to despotism”


With two upcoming elections, Tunisia’s young democracy will be put to the test. Florian Guckelsberger and Lars Mensel sat down with the country’s former Minister weiter...

with Rafik Abdessalem

“Architecture is an expression of values”


Sir Norman Foster is the mastermind behind some of the world’s most iconic buildings. With Max Tholl, he discussed how architecture helps us communicate, where our fascination with bigness stems from, and why we need to do more with less.

with Norman Foster

Related Content: Religion, Evolution, Education


The Origins of Mankind


Decoding Human Evolution

New technologies allow for an unprecedented look at the origins of homo sapiens. They tell us where we came from, and with whom we mingled. read more

by Christopher Stringer


Humanity's Genetic History


The Species Is A-Changing

Genome analysis allows an unprecedented look at the history of humanity. It illustrates that we're a species in flux – and that the best is yet to come. read more

by Joshua Akey
Most Read