ABSOLUTE BRAIN SIZE MATTERS
The questions that I’m thinking about involve how humans are different from other animals. I’m also interested in how that happened, how we became so different in terms of our psychology. The area in psychology that is fertile for a lot of growth is thinking about how our psychology evolved. How did we go from having psychology more like other apes to being like we are now, and what was the process by which that happened? How did either natural selection or random forces of evolution produce what we are today? That’s a hard problem that I'm excited to think about.
What are we doing to try to look at that? Well, we compare different animals. Historically, we’ve been lucky if we could do a comparison of two animals. Say I compare dogs and wolves to each other and try to understand how they’re different from one another, if I can understand how they’re different, then maybe I can make some guesses about how those differences evolved.
My hope for the future is that comparative psychology can move past just comparing pairs of species and look at lots of different species, using the tree of life to make predictions and test ideas about how psychology might evolve in different species so that we can come up with ideas about how our own species may have happened. The idea that we have culture, that we have language, that we can think about the thoughts of others, that we have the ability to deceive or care about others and have empathy—one of the big hypotheses for how humans ended up with these unusual abilities is that evolution favored more complex social skills.
To test that, though, you have to look at lots of different species, and you have to have data on how those different species solve social problems to be able to trace how those social skills may have evolved. That’s been difficult, but it's also exciting to think about how to get around that problem. One of the things we’ve tried to do is pioneer these large-scale collaborations, like the genomicists have done. We published a paper two or three years ago that involved fifty-six co-authors. We got people from all over the world to contribute data on a variety of primate species, and even non-primates, like birds and elephants. We had almost forty species. Everybody had done some cognitive tests with their species they had available to them and, remarkably, it was the first time that people who study animal psychology had ever worked together in this way. We led the charge to do that because we know that if we want to understand the evolution of human social psychology, if we want to test why we are the way we are and the hypotheses that we think are necessary to make us human, we’re going to have to look at a large range of species and understand how they have been shaped by evolution.
We measured inhibitory control, which is basically your ability to not do something that might be counterproductive. We had two measures of this on these forty species. We thought that by testing the big hypothesis—that there’s been selection on social psychology in animals—we might be able to learn about the human case. When we looked at these forty species, that’s not what we saw. We thought it would be that animals with more complex social systems need the ability to control their behavior to not do something that might be counterproductive. You can imagine that if you’re competing with one another, you don’t want to get in a fight with the wrong guy, right? Self-control would seem to be incredibly important in social endeavors. That’s not the pattern we saw when we did the measurements.
The thing that stuck out was that self-control is simply a product of absolute brain size. It had more to do with your feeding ecology: How complex was your diet? How many things do you rely on to survive? That was a big surprise, because the idea that diet is shaping cognition has faded in many circles as the leading hypothesis for thinking about how psychology evolves. So, how do we move forward on testing ideas about the evolution of psychology?
It's interesting to think about how this all came about. It all started in a bar. There's a professor with me at Duke named Charlie Nunn who does what’s called phylogenetics; he uses the tree of life to make predictions about evolution. He invited me to go get a beer. He had realized that no one was using the new phylogenetic techniques within the cognitive sciences, looking at animals and animal psychology. He took me out and asked why I wasn't using phylogenetics, and I told him this sad story about how it’s hard to get enough animals together to do this. He just didn’t want to hear any excuses. He said, “Why don’t we do a series of workshops at the National Evolutionary Synthesis Center at Duke?”
Phylogenetics is using all the new genomic work to look at how different animals are related to each other. Of course, evolution is descent with modification. You can see how many genes different animals share with each other and how they’ve changed or been modified by evolution. It helps us understand how closely related different species are to each other. That’s how we make the tree of life, and then you can use that tree of life to understand how fast change or evolution has occurred. You can also choose which species should be studied to test a different idea about evolution. We haven’t taken advantage of that, historically, while much of biology has been doing this for twenty years.
Charlie saw a real opportunity there and encouraged me to get my community together, so we did these three workshops. The group of people we brought together were a wide variety of people who study animals in the wild, people who study animals in captivity, people who are biologists by training and use these phylogenetic techniques. We had Mike Tomasello, Carel van Schaik, Josep Call, Daniel Haun, I was there, obviously, and Laurie Santos was there—a lot of great talent in our area of research. The first workshop was literally like therapy because we all were saying this was impossible. The funny part is, I already revealed what the big finding was, which was that absolute brain size matters.
We all agreed at the workshop that we never wanted to hear anyone talk about how absolute brain size could predict cognitive ability because we felt strongly that there’s domain specificity, that evolution shapes species to have different ability to be flexible and show flexible problem solving, and that flexibility wasn’t always going to be related to brain size. We really said, “We want to move past that. That’s why we need this phylogenetic team."
Each successive meeting, though, we got more and more optimistic. One of the big things that kept people from being depressed and saying this wasn't going to be possible was a study we did with the lemurs at the Duke Lemur Center. We compared six of the species there on the same set of problem-solving tasks, and we were able to show our colleagues that we could do a meaningful comparison of different species that told us about similarities and differences among them. In demonstrating that, we thought, well, why can’t we expand this to more species?
The problem that comparative psychology has always had what’s known as the Beach-Bitterman problem. The Beach-Bitterman problem asks how we are supposed to compare species that have fins, that have hands, that have trunks, and for that to be a fair comparison. And that’s just talking about the morphology. Phylogenetics helps us get out of that problem because it allows us to compare species that are closely related, that have similar morphology, and compare them almost as an individual. So, we can get multiple individuals, multiple clusters of species across the tree of life, and look at how they are different from each other. If we get six or seven sets of these species that differ in the same way, then we’ve got a pattern. If you want to have a species get better at remembering things, for instance, it ends up that in all seven of these different families, they all have a larger brain or they all have a more complex social system.
We’ve never been able to approach this problem before, so by showing that the lemurs could be compared in this way, it helped us get away from the Beach-Bitterman trap for the first time. By using phylogenetic tools, we could compare closely related species but make conclusions about patterns that we see in multiple sets of distantly related species, and essentially look for convergence, which is the same process happening multiple times across the tree of life.
People got excited. We started going to all our friends who we knew had interesting animals and interesting populations of animals that could participate. Was it a random sampling? No. Was it a perfect sampling? No. It was our attempt at a proof of concept. We chose inhibitory control on purpose because we knew that that was something people could do quickly, it wouldn’t require training, and it was a problem that all animals would need to be able to solve. And we were right. We had everything from swamp sparrows to elephants participate in this. We found this interesting pattern that brain size and how many types of food you eat are what matter. But it was really a proof of concept. We published it in Proceedings of the National Academy of Sciences and it’s been very well received.
The question is, what’s next? How do we move this field forward? We know we can do this now. We know, as a community, that we can get people to work together, we can identify a problem, we can solve the Beach-Bitterman trap, and we can test the big evolutionary hypotheses for psychological evolution. We could even put them into competition against each other. We can say, “Okay, was it really social evolution? Was it foraging? Is it brain size? Is it relative brain size?” We can put those things into competition.
A lot of people have ideas about how psychology evolves, whether it’s my PhD advisor talking about the importance of cooking, which emanates from thinking about increases in energy and how that affects brain evolution, or somebody arguing about the importance of self-control in evolution. My own PhD advisor, Richard Wrangham, has the idea about the importance of cooking and how it's fundamental to the evolution of human cognition. The energy that it would take to have a human brain evolve, the only way to get there, in his mind, is through cooking.
Then you have somebody like Robert Sapolsky, thinking about the importance of self-control and how self-control would be crucial to cognitive evolution in primates and beyond. My own idea, together with Richard Wrangham, is thinking about selection for friendliness and how if you become more interested in a larger range of social partners, it has a huge impact on the evolution of psychology. These are all fascinating hypotheses, but a hypothesis is only as good as it is testable. Using phylogeny is the way we’re going to test these big ideas and potentially falsify them.
In thinking about selection for friendliness, I need to find a range of times in evolution where I think that selection pressure may have had a big impact. I’m interested in species that live on islands, but also have a population on a continent. The thinking is that when you’ve escaped predation, there’s selection against a defensive fear response. What’s the point if there are no predators that can eat you? I would suspect that something similar to the cognitive evolution that we see between dog and wolf has happened in these island populations.
That’s where phylogenetics and looking at the tree of life and testing these hypotheses is super powerful. It means moving beyond just comparing two species and saying, “Here’s this one species or population that lives on an island and one that doesn’t live on an island; they’re different in the way that I thought they’d be different. Look, this hypothesis is supported.” That’s not convincing to an evolutionary biologist. For an evolutionary biologist studying bacterial evolution, they’d never be impressed by that. So, we can move towards doing evolutionary science by looking at, say, species that live on islands in a variety of island environments.
Another interest of mine is looking at species that are invading urban environments. Let’s say that there were no coyotes west of the Mississippi fifty years ago and now they’re in every state east of the Mississippi. I would predict that coyotes have been selected for friendliness and that their psychology has been shaped in ways that I would predict based on our hypothesis. Through looking at phylogeny, we can choose dozens of species that I would predict had changed based on my hypothesis. We can get collaborators to work with those animals and come up with a task that would challenge their problem-solving ability, to see if that hypothesis was as predictive as we thought it was and potentially falsify it. That’s the progress we’ve never been able to make. That’s the power of this phylogenic approach.
If people are wondering what cognitive evolution might be, the evolution of morphology, the evolution of diet—it’s just studying how the mental processes have evolved. The way to think about cognitive evolution is that it’s the hidden life inside your mind. The underlying assumption of cognitive evolution is that animals also have a hidden life inside their minds. It’s a rejection of the idea that someone like B.F. Skinner or J.B. Watson, of the traditional school of psychology known as behaviorism, would have advocated. Cognitive evolution rejects the notion of a uni-dimensional measure of intelligence, that there’s one form of intelligence, and differences between species are just quantitative and measured at how fast they can learn things.
Instead, the idea of cognitive evolution is that there are different types of intelligence. The crucial piece is that they vary independently. For instance, you can have a species that has a lot of self-control but no working memory, relative to another species that might have a lot of working memory and no self-control. You could have a species that is very good at understanding what others intend, but isn’t very good at understanding the causal principles of the world, things like gravity or understanding that when things are connected to each other, they move together. We can measure all these types of cognitive abilities, and we can look at which species have the most flexibility across all these different domains or different types of intelligence. That’s the exciting program of cognitive evolution. It is radical in that we’re saying we’re going to be able to see inside the minds of other animals and make inferences about what’s going on. That was rejected by behaviorism as almost a pseudoscience.
One thing that people sometimes misinterpret is that B.F. Skinner didn’t argue that animals didn’t have minds; he just didn’t think they were important. Today, we know based on neuroscience and how development works and also our own studies of animal psychology, that that just doesn’t fit the facts.
It took us seven years to organize and put together this group of people, get everybody to agree to participate, and get enough species together where we could do our phylogenetic tests, using the tree of life to test these evolutionary hypotheses. And then we had to coordinate. We had done our six species of lemurs, but we had to coordinate because there were people who were in South America working with golden lion tamarins; there were people in Thailand working with elephants; there were people in Germany who were working with chimpanzees and bonobos; and then we had people in California who were working with squirrels. It was ridiculous in a way.
We had to have everybody give a reasonable approximation of our cognitive task or our problem so that the comparison would be meaningful. That meant we had to get all these people to send us their videos, we had to watch the methods and make sure that it seemed right, and we had to have everybody agree that this seemed like a reasonable way forward. So, it was a lot of work. It took us seven years. Then we finally ran the analysis. I was putting my money on the idea that being social and being in a richer social group would probably explain the difference between species’ ability to solve this problem. A lot of people were betting on that. This hypothesis has been kicking around since the late ‘70s, and this was the biggest test yet. It ends up that there was absolutely no evidence to support it whatsoever. The main pattern that came out, and I mean screaming out of this data, was that it is absolute brain size. Relative brain size not so much, which was interesting because there’s been a debate about which of those measures is more powerful. I can tell you it’s absolute brain size if you’re talking about self-control.
There’s some evidence from this for the diversity of food that you eat; that was an important pattern as well. It was a wonderful moment in science where you get to find out what kind of scientist you are. I remember my postdoc came in and he told me what we had found and I was like, “Go back and find anything else. I don’t care what you find in the data." He just said, “Look at the pattern. This isn’t going to go away.” In some ways, it made me super proud that this is what we found because it was not what we were looking for. We must have been honest in the science.
Absolute brain size would be MRI. You could do it that way. There are ways to approximate it without having to be scanned. You could go old-fashioned calipers like a paleoanthropologist would measure a fossil. You could do it on your head to approximate. We did that in class as kids in intro anthropology class.
I was talking about between species, but what about within species? You want to talk about a tough topic. We have a study that we’re working up right now, which I feel confident enough to talk about. We have data on dogs, and dogs have become a very powerful tool if you want to study psychology. One of the reasons they’re powerful is because we can study thousands of them and they have incredible diversity in their behavior and their morphology, and that includes their brain size.
There is significant variation across different breeds in brain size. We were able to look at ten different types of problem solving that we think probably account for about five different types of intelligence that we think dogs have. We have demonstrated that these types of intelligence vary independently within dogs. When we looked at absolute brain size across different dog breeds in a sample of 7,000 dogs, we have evidence that brain size matters for two of the five types of intelligence. For three, it doesn’t matter at all, but for two it matters a lot. One is working memory: Larger dogs are better at remembering for longer things they’ve seen in the past. Small dogs struggle more.
What’s fun and interesting is that we have a game where dogs have the opportunity to be obedient or disobedient after you told them not to take food. The manipulation is that in one condition you watch after you’ve told them not to take the food, the other condition you turn your back on them. Small dogs are much more likely to be disobedient, strategically, when you turn your back. That’s related to brain size.
When you get different brain size across a species, you also get different proportions of different parts of the neuroanatomy that are bigger or smaller, but in dogs, it just scales one to one, the entire brain. We have one type of intelligence larger brained dogs do better with, this working memory, but when it comes to being deceptive based on what you can see, it’s small dogs that are more intelligent and flexible. So, it may be good in some cases to have a larger brain, but it may be problematic as we continue to look at different types of intelligence.
The working memory finding does support our cross-species comparison, because working memory and self-control have been argued to be what’s known as executive function. These things are related to almost every problem we solve. People see them as a package that has been called "general intelligence" because it’s involved in almost everything we do. That’s to differentiate from something that is more specialized, like a social problem-solving skill such as the one I was talking about where dogs are being deceptive based on what you can see or not; that’s a more specialized skill that people like to call "domain specific."
I would say the underlying research program of everything I’m doing is all about humans. I’m interested in how humans happened: What is it that makes us human? How did we get this way? And I don’t think you can know what it is to be human if you don’t first know what it is to be not human. How in the world can you know that whatever it is you think that’s special about humans is even specific to our species if you don’t know what other species are like? It’s very likely that whatever you think is special about humans, probably other animals do it.
If somebody has a great idea about what might make us human, especially in my area—I’m interested in psychology—we use animals to try to potentially falsify or challenge that idea. How do we connect with people who are specifically interested in the animals when our research program is all about studying human evolution? That’s a good question. I connect through my own brain because, while everything we do is aimed at understanding humans, along the way of making discoveries, we learn a lot about the animals themselves and are very quick to find ways to try to apply that.
I work with the Department of Defense, and we’re very heavily involved in trying to understand how you get the best dog for the different jobs that they do. Dogs have more jobs than ever, whether it’s finding bombs or assisting people with disabilities. We work towards trying to solve those kinds of problems. Whether it’s figuring out how we have better welfare for animals, or how we take better care of them when they’re under our care, or how we get people to think about conservation and motivating people to care more about animals, we work very hard on all those issues.
The reason I got into all of this and why I got excited to think about how animal minds work is because I had a pet dog as a kid. His name was Oreo. There were many times in interacting with him that I was just curious as to whether he loved me like I loved him. Was it different? What was that like? And like many people, I got fascinated by the problem. I didn’t go to college knowing that this is what I was going to become, but I had Frans de Waal as a professor, who introduced me to the field of animal psychology. I worked very closely starting as an undergraduate with Mike Tomasello. By the time I finished undergraduate, I’d already published seven papers working on chimpanzee cognition. I was convinced that animals had minds, potentially. At the time I started, people like Daniel Povinelli had done his pioneering work arguing that chimpanzees didn't have the ability to understand the mental lives of others. Working with Mike Tomasello, we did a whole bunch of pioneering experiments. As an accident, it became apparent that dogs were doing something that even chimpanzees weren’t doing. And when I went to graduate school with Richard Wrangham, we became very interested in the effect of domestication. We both were exposed to the work of Dmitry Belyayev, a Russian geneticist who had experimentally domesticated foxes over a sixty-year period, and were heavily influenced by his thinking on domestication.
Being at Harvard, around people like Steve Pinker and running into Dan Dennett, I was thinking about how we could explain all of the unique phenomena that humans seem to create. How do we answer Darwin’s ultimate challenge? The ultimate person who influenced me is Charles Darwin. How do we get from a common ancestor with apes to the species that we are? In interacting with all these different great minds, I realized there were so few people working on the hard problem.
One of Darwin’s greatest challenges was not just identifying the thing that’s unique—and there’re lots of people in cognitive psychology who have tried to find out what it is about language and culture that’s different in us than other animals—but what the process is. Once we identify what it is, well, how did it happen? How are we going to have a testable hypothesis? There’s a seven-million-year process that led from our common ancestor with our living great ape cousins that we can actually measure and study and us. How are we going to fill the void? We can’t look at fossils and make very many rich inferences about their psychology. It’s only going to take us so far.
That’s why using phylogenetics, the tree of life, and trying to look for evidence of distantly related species that have been changed in similar ways is going to be our most powerful tool in looking for convergence. Getting people to collaborate in ways they have not collaborated before will be the answer to the biggest question Darwin ever asked, which addresses this gap between us and other animals: How did natural selection or other random processes get us from animal psychology to human psychology?
The big thing that dogs helped us learn is that you can have evolution in how animals solve problems that happened as an accident, basically. We think dogs evolved through selection for friendliness. Basically, wolves were able to take advantage of interacting with humans as humans became more sedentary in their lifestyle. The evidence for the beginning of dog evolution from genomic comparisons is that dogs probably began evolving 25,000 to 40,000 years ago. That means that it’s hunter-gatherers that, in their interactions with wolves, led to the evolution of dogs. Before ten years ago, people thought it would have been agriculturalists, so that’s a big change in the thinking. That’s why we think that it was wolves that chose us and began interacting with us in a different way.
We think the wolves that could interact with humans and be friendly and non-aggressive would have been at a huge advantage, because your choice is try to catch an elk and get kicked in the face or maybe not get a meal, or you can sneak into a human settlement at night, eat garbage, eat human feces and you have a very stable food resource. We think that there was selection, that those wolves that could take advantage of human settlements and garbage bred together, and that led to changes in their morphology, physiology, and psychology. They became friendlier. They could interact with us as if we were their own social partners, and they could read our communicative intentions. That’s why they’re so good at reading our gestures. That’s the hypothesis.
But what about other species? What I just argued is that natural selection can select for friendliness and lead to domestication or the syndrome that we know as domestication. Domestication doesn’t require artificial selection. It doesn’t require human guidance to generate a domesticated animal, at least the first steps. That’s kind of a radical idea.
The reason we got so interested in bonobos was because Richard Wrangham and I realized that bonobos would provide an amazingly strong test case for our idea that selection for friendliness can happen as a result of natural selection, and that the domestication syndrome, this change in morphology, physiology, and psychology that we think comes as a package, through changing how development occurs in a species, the bonobos would be a great way to potentially falsify this idea. The reason we picked bonobos was because there was evidence that bonobos were less aggressive and friendlier than their almost genetically indistinguishable closest relative, chimpanzees, that we’re very familiar with.
So, Richard and I set about using ten years of comparisons between chimp and bonobo morphology, physiology, and cognition, and development. What we found, to encapsulate it very quickly, is bonobos are the dog of our ape family. They have changes in their morphology that are very similar to the changes you see between dogs and wolves. For instance, they have smaller canines and their skulls are more infant-like. Even as adults, they’re juvenile in their morphology. When it comes to their physiology, they deal with stress in a very different way than chimpanzees, in ways that you would predict if there was selection for friendliness. And we can see that in their corticosteroid response, your hormones that are involved in how you deal with stress and fear, and also in their testosterone. For instance, a male chimpanzee that’s challenged has an increase in testosterone, but bonobos don’t; they have an increase in their cortisol. They get upset if there’s a social problem and they want to hug or have sex. That’s a very different response to stress and social anxiety. It’s much more similar in what you see in how domesticated animals deal with social stress.
What we found is that bonobos are the ape that never grows up. They have very juvenile psychology even as adults. So, many of the patterns of development where chimpanzees begin, let’s say with spatial cognition, bonobos and chimpanzees are very similar when they’re juveniles, but chimpanzees move way past where bonobos ever arrive in their development. That’s exactly what we see wolf versus dog.
It’s more interesting than that because it’s not just that bonobos are frozen as juveniles, they also have extended windows of development on both ends. The sexual behavior that bonobos are so famous for, the big surprise for me as I started to work with them is you see their sociosexual behavior as early as six months, eight months of age. These tiny infant bonobos are rubbing their genitals on their mothers or on their peers. And we have a comparison. Vanessa Woods and I did a comparison of bonobo and chimp infants that had no parents around them. The bonobos were essentially playing doctor with each other, but we see none of that behavior in chimps. This is important because what we see in bonobos is the same thing we see in domesticated animals: these extended windows of development. You can see patterns of growth that occur earlier and later.
That’s why then it becomes extremely tempting to think about our own species' evolution because the big pattern when you compare ourselves to development in other great apes is expanded windows of development. Especially when you think about our brain. If you want to get expanded windows of development in humans, how do you do it? There hasn’t been a hypothesis that addresses that problem. How do you deal with changes in human morphology, especially late in human evolution?
We have slightly smaller brains than our recent ancestors of 20,000 to 40,000 years ago—how do you deal with the feminization or juvenilization of our morphology? The fact that we have much less robust facial morphology than the humans of 50,000 to 80,000 years ago? There’s no hypothesis that deals with that. How do you deal with the fact that our self-control, the pruning that occurs in our prefrontal cortex, doesn’t end until we’re twenty-two? How do we deal with the fact that we’re born with a brain that’s underdeveloped, yet we know, comparing ourselves to other great apes, our social skills are way advanced? The answer is that there’s selection for friendliness late in human evolution. Just like dogs, bonobos, the selection for friendliness led to expanded windows of development and led to changes in morphology, physiology, and psychology. It’s the answer for our species, too.