Ground up monkey brains

Short version: It looks like most mammals, at least most large mammals, have the brains they need, while primates, especially large primates, have the brains they can afford.

Longer version: One reason for being interested in monkeys is that they’re brainy mammals. Here’s the conventional graph illustrating that:

Larger mammals tend to have larger brains, but the relationship is non-linear. Multiplying body mass by x doesn’t multiply brain mass by x. Instead it multiplies brain mass by about x^.75. In other words, Brain Mass is proportional to (Body Mass)^.75. Equivalently (taking the logarithm of both sides) Log[Brain Mass] is equal to .75 times Log[Body Mass], plus a constant. So Log[Brain Mass] plotted against Log[Body Mass] gives a straight line with a slope of .75. That means that if one mammal has 16 times the body mass of another, it’s expected to have 8 times the brain mass, 10,000 times the body mass means 1000 times the brain mass, and so on. The thing to note is that primates defy expectations. They have larger brains than would be expected based on their body sizes.

But we’ve recently learned that primates – especially big ones – are even more special than this graph suggests. Susan Herculano-Houzel has pioneered a technique that involves chopping up brains (or parts of brains), dissolving their cells to make a kind of brain soup, and counting cell nuclei. This allows her to estimate how many neurons there are in different brains.

Major findings: Among most mammals, the number of neurons increases more slowly than brain size. Increase brain size by x, and you increase number of neurons by about x^.67. (H-H shows this flipped around. Increase number of neurons by x and you increase brain mass by x^1.5.) But primates are exceptional; the relationship is nearly linear. An x-fold increase in primate brain size corresponds to about an x-fold increase in number of neurons. Humans follow the primate rule here. We have about the same density of neurons as other primates. When you combine the exceptionally large brain sizes of humans (exceptional even relative to our brainy primate relations) with a standard high primate neuron density, you get an animal with an enormous number of neurons. By contrast, a rodent with a human sized brain, if it followed rodent rules for how neuron numbers increase with brain size, would have only 1/7 as many neurons.

Neurons are expensive. Most large animals economize by cutting back on neuron density. A cubic centimeter of cow brain has fewer neurons, and consumes energy at a lower rate, than a cubic centimeter of mouse brain (although of course the cow’s brain is bigger). By contrast, large primates are extravagant, devoting exceptionally large energy budgets to running their brains. And human brains are exceptionally costly. An important question for the study of human evolution is how we paid the bill for such costly brains. That’s a story for later. But another part of the story starts back in the early Cenozoic, when monkeys committed to a different set of rules for building brains.

And here is a chart giving absolute numbers of  cortical neurons (cneurons) for a bunch of species. Scott Alexander has some thoughts about the moral implications. Short version: skip the pork for dinner (and skip the elephant, chimp, and manflesh. But you knew that). Beef might be okay. Better is lobster.

And Werner Herzog is probably okay with you eating chicken.

Share this:

Where is everybody?

5.04 – 4.77 billion years ago

Tomorrow is a big day on Logarithmic History, the origin of our solar system, of the Sun and planet Earth. But is this really such a big deal in a cosmic perspective? After all, stars and planetary systems have been forming a fast clip in the Milky Way and other spiral galaxies since long before this date. So let’s suppose … Suppose there was a star like the Sun, but older by a billion years. And suppose this star had a planet like Earth orbiting around it – call it Urth. And suppose life originated on Urth more or less as on Earth and followed more or less the same evolutionary path. With this head start, intelligent life could have evolved a billion years ago, and today there could be intelligent Urthians (or their self-replicating robot descendants) a billion years ahead of us.

There’s an urban legend that says that Einstein called compound interest the strongest force in the universe. Einstein didn’t actually quite say this, but it’s not a crazy thing to say. For example, consider how compound interest works, backward, on our Logarithmic History calendar. December 30 covers a period 5.46% longer than December 31, December 29 is 11.2% longer (because 1.0546 * 1.0546 = 1.112), and so on. At this rate of compounding we wind up with January 1 covering 751 million years. The same math implies that if we invested 1 dollar at 5.46% interest, compounded annually, then after 365 years we’d have 751 million dollars.

With even the slightest compound rate of increase, a billion year old Elder Race would have plenty of time to fill up a galaxy, and undertake huge projects like dismantling planets to capture more of their suns’ energy. Indeed, they could arguably colonize the whole reachable universe.

Traveling between galaxies – indeed launching a colonization project for the entire reachable universe – is a relatively simple task for a star-spanning civilization, requiring modest amounts of energy and resources. … There are millions of galaxies that could have reached us by now.

Which raises the question, posed by Enrico Fermi in 1950: “Where is everybody?” There are more than 100 billion stars in our galaxies, more than 100 billion galaxies in the visible universe (actually, according to recent estimates, the number is more than  1 trillion). If there are huge numbers of billion year old Elder Races around, why hasn’t at least one of them taken the exponential road and made themselves conspicuous? Yet a recent survey of more than 100,000 galaxies found no evidence of any really advanced civilizations harnessing the power of stars on a large scale.

One possible resolution of the paradox has been suggested recently. The argument goes like this: the easiest route to estimating the number of advanced civilizations in the universe is to multiply point estimates of the probabilities of events like the formation of an Earthlike planet around an appropriate star, the origin of life on such a planet, the origin of human-level intelligence, and so on. But there are enormous uncertainties in these estimates. Properly speaking, instead of just multiplying mean estimates, we should be doing a convolution of the range of estimates. The general point is that if you’re multiplying a whole lot of probabilities, and you’re not certain what those probabilities are, there a strong likelihood that at least one of those probabilities is close to 0, so their product is also close to 0. The authors therefore suggest that there is a strong possibility that there are few or no advanced civilizations, or even other human-level civilizations in the Milky Way or even the observable universe.

And another recent article reinforces this conclusion. Human-level intelligence evolved fairly late in the life of planet Earth. Earth has been around for 4.5 billion years, but it will probably be inhabitable by complex organisms for only about another .8 to 1.3 billion years, before the sun expands and the planet overheats (unless of course we’re still around, and decide to move it further away).

Carter (1983) noticed this coincidence and proposed a resolution to the puzzle based on observation selection effects. Letting t1 denote the lifetime associated with our star, and t2 be the timescale it takes for evolution to produce intelligent life, one can analyze three possibilities: t1 << t2, t1 =~ t2, or t1 >> t2, denoting the situations in which the lifetime of the star either greatly exceeds the timescale associated with intelligent life, approximately equals it, or is greatly exceeded by it. Carter argues that a priori, the possibility that t1 =~ t2 is exceptionally unlikely, leaving t1 << t2 and t1 >>t2 as realistic alternatives. We can also rule out t1 << t2 with high probability, given that intelligent life did not emerge exceptionally early when compared with the Sun’s lifetime. This brings us to the possibility that t1 >> t2. This would mean that most stars will never support intelligent life, as the star will burn out before such life emerges. However, in the rare locations in which intelligent life does emerge, it will find itself emerging within the lifetime of the star, and moreover is most likely to observe t1 =~ t2, consistent with our own observations. Observation selection effects therefore explain why we see these time- scales tightly coupled, even if such an outcome is a priori unlikely.

The authors take Carter’s argument a step further. They look not just at the appearance of intelligent life, but the timing of several other one-time evolutionary transitions (the origins of life, photosynthesis, eukayotes, and sexual reproduction). They conclude that the timing of these events suggests that it typically takes far longer than the expected lifetime of a sunlike star to get all the way to intelligent life. We’re just incredibly lucky.

In short, if this is true, tomorrow on Logarithmic History may mark a truly momentous occasion, not just in the history of the Earth, but in the history of the observable universe.

Blood and brains

Humans are brainy animals. One way to show that is by looking at brain size: our species has the biggest brains, in relation to body size, of any animal. But there’s more to it than that. An earlier post covered the work of Susan Herculano-Houzel. She developed a technique for counting the number of neurons in a brain, or part of a brain. Among most mammals, big animals have a lower density of brain neurons, so they aren’t actually as brainy (measured by neuron number) as you’d think just based on their brain size. Primates however break the usual mammalian rule. Big primates have the same neuron density as little guys, so they really are quite brainy. And humans, with really big brains and (following primate rules) a high density of neurons, stand out even among primates as exceptionally brainy.

This work isn’t much help if we are looking at extinct hominins, when all we’ve got is their fossil skulls. But now there’s some interesting recent research with a new take on the subject. Brains need to be supplied with blood. The more energy they use, the more blood flow is needed. We can now figure out fairly accurately how much blood flow a brain is getting by looking at the size of the hole that lets the carotid artery in through the base of the skull. And then we can apply this technique to look at humans, and at extinct hominins. It turns out that humans are even more exceptional when we look at blood flow to the brain: we’re getting double the flow that you’d expect based on brain size alone.

Early hominins however, Australopithecus and early Homo, aren’t very impressive upstairs, many with less blood flow to the brain than modern apes. Looking at the graph it looks like there are really two grades of brain evolution. In the lower grade, which includes early hominins and modern apes, there is a gradual increase over millions of years. (I’m just guessing here that the ancestors of chimps and gorillas millions of years ago were about as brainy as contemporary hominins, but we’d still like to find more fossils.) And then there is a big leap up to a higher grade with early Homo erectus, and a rapid increase after that. It looks like something major changed with the appearance of Homo erectus, either on the supply side – improvements in food supply making brains more affordable – or on the demand side – a greater fitness payoff to a high energy brain – or both.

Ground up monkey brains

Short version: It looks like most mammals, at least most large mammals, have the brains they need, while primates, especially large primates, have the brains they can afford.

Longer version: One reason for being interested in monkeys is that they’re brainy mammals. Here’s the conventional graph illustrating that:

Larger mammals tend to have larger brains, but the relationship is non-linear. Multiplying body mass by x doesn’t multiply brain mass by x. Instead it multiplies brain mass by about x^.75. In other words, Brain Mass is proportional to (Body Mass)^.75. Equivalently (taking the logarithm of both sides) Log[Brain Mass] is equal to .75 times Log[Body Mass], plus a constant. So Log[Brain Mass] plotted against Log[Body Mass] gives a straight line with a slope of .75. That means that if one mammal has 16 times the body mass of another, it’s expected to have 8 times the brain mass, 10,000 times the body mass means 1000 times the brain mass, and so on. The thing to note is that primates defy expectations. They have larger brains than would be expected based on their body sizes.

But we’ve recently learned that primates – especially big ones – are even more special than this graph suggests. Susan Herculano-Houzel has pioneered a technique that involves chopping up brains (or parts of brains), dissolving their cells to make a kind of brain soup, and counting cell nuclei. This allows her to estimate how many neurons there are in different brains.

Major findings: Among most mammals, the number of neurons increases more slowly than brain size. Increase brain size by x, and you increase number of neurons by about x^.67. (H-H shows this flipped around. Increase number of neurons by x and you increase brain mass by x^1.5.) But primates are exceptional; the relationship is nearly linear. An x-fold increase in primate brain size corresponds to about an x-fold increase in number of neurons. Humans follow the primate rule here. We have about the same density of neurons as other primates. When you combine the exceptionally large brain sizes of humans (exceptional even relative to our brainy primate relations) with a standard high primate neuron density, you get an animal with an enormous number of neurons. By contrast, a rodent with a human sized brain, if it followed rodent rules for how neuron numbers increase with brain size, would have only 1/7 as many neurons.

Neurons are expensive. Most large animals economize by cutting back on neuron density. A cubic centimeter of cow brain has fewer neurons, and consumes energy at a lower rate, than a cubic centimeter of mouse brain (although of course the cow’s brain is bigger). By contrast, large primates are extravagant, devoting exceptionally large energy budgets to running their brains. And human brains are exceptionally costly. An important question for the study of human evolution is how we paid the bill for such costly brains. That’s a story for later. But another part of the story starts back in the early Cenozoic, when monkeys committed to a different set of rules for building brains.

And here is a chart giving absolute numbers of  cortical neurons (cneurons) for a bunch of species. Scott Alexander has some thoughts about the moral implications. Short version: skip the pork for dinner (and skip the elephant, chimp, and manflesh. But you knew that). Beef might be okay. Better is lobster.

And Werner Herzog is probably okay with you eating chicken.

Where is everybody?

Today is a big day on Logarithmic History, the origin of our solar system, of the Sun, Moon, and planet Earth. But is this really such a big deal in a cosmic perspective? After all, stars and planetary systems have been forming a fast clip in the Milky Way and other spiral galaxies since long before this date. So let’s suppose … Suppose there was a star like the Sun, but older by a billion years. And suppose this star had a planet like Earth orbiting around it – call it Urth. And suppose life originated on Urth more or less as on Earth and followed more or less the same evolutionary path. With this head start, intelligent life could have evolved a billion years ago, and today there could be intelligent Urthians (or their self-replicating robot descendants) a billion years ahead of us.

There’s an urban legend that says that Einstein called compound interest the strongest force in the universe. Einstein didn’t actually quite say this, but it’s not a crazy thing to say. For example, consider how compound interest works, backward, on our Logarithmic History calendar. December 30 covers a period 5.46% longer than December 31, December 29 is 11.2% longer (because 1.0546 * 1.0546 = 1.112), and so on. At this rate of compounding we wind up with January 1 covering 751 million years. The same math implies that if we invested 1 dollar at 5.46% interest, compounded annually, then after 365 years we’d have 751 million dollars.

With even the slightest compound rate of increase, a billion year old Elder Race would have plenty of time to fill up a galaxy, and undertake huge projects like dismantling planets to capture more of their suns’ energy. Indeed, they could arguably colonize the whole reachable universe.

Traveling between galaxies – indeed launching a colonization project for the entire reachable universe – is a relatively simple task for a star-spanning civilization, requiring modest amounts of energy and resources. … There are millions of galaxies that could have reached us by now.

Which raises the question, posed by Enrico Fermi in 1950: “Where is everybody?” There are more than 100 billion stars in our galaxies, more than 100 billion galaxies in the visible universe (actually, according to recent estimates, the number is more than  1 trillion). If there are huge numbers of billion year old Elder Races around, why hasn’t at least one of them taken the exponential road and made themselves conspicuous? Yet a recent survey of more than 100,000 galaxies found no evidence of any really advanced civilizations harnessing the power of stars on a large scale.

There is a methodological issue here, involving multiplying probabilities, that might help resolve the paradox. The argument goes like this: the easiest route to estimating the number of advanced civilizations in the universe is to multiply point estimates of the probabilities of events like the formation of an Earthlike planet around an appropriate star, the origin of life on such a planet, the origin of human-level intelligence, and so on. But there are enormous uncertainties in these estimates. Properly speaking, instead of just multiplying mean estimates, we should be doing a convolution of the range of estimates. The general point is that if you’re multiplying a whole lot of probabilities, and you’re not certain what those probabilities are, there a strong likelihood that at least one of those probabilities is close to 0, so their product is also close to 0. The authors therefore suggest that there is a strong possibility that there are few or no advanced civilizations, or even other human-level civilizations in the Milky Way or even the observable universe.

An article published recently reinforces this conclusion. Human-level intelligence evolved fairly late in the life of planet Earth. Earth has been around for 4.5 billion years, but it will probably be inhabitable by complex organisms for only about another .8 to 1.3 billion years, before the sun expands and the planet overheats (unless of course we’re still around, and decide to move it further away).

Carter (1983) noticed this coincidence and proposed a resolution to the puzzle based on observation selection effects. Letting t1 denote the lifetime associated with our star, and t2 be the timescale it takes for evolution to produce intelligent life, one can analyze three possibilities: t1 << t‘, t1 approx= t2, or t1 >>􏰁 t2, denoting the situations in which the lifetime of the star either greatly exceeds the timescale associated with intelligent life, approximately equals it, or is greatly exceeded by it. Carter argues that a priori, the possibility that t1 approx= t2 is exceptionally unlikely, leaving t1 << t2 and t1 >>􏰁 t2 as realistic alternatives. We can also rule out t1 << t2 with high probability, given that intelligent life did not emerge exceptionally early when compared with the Sun’s lifetime. This brings us to the possibility that t1 >>􏰁 t2. This would mean that most stars will never support intelligent life, as the star will burn out before such life emerges. However, in the rare locations in which intelligent life does emerge, it will find itself emerging within the lifetime of the star, and moreover is most likely to observe t1 approx= t‘, consistent with our own observations. Observation selection effects therefore explain why we see these time- scales tightly coupled, even if such an outcome is a priori unlikely.

The authors take Carter’s argument a step further. They look not just at the appearance of intelligent life, but the timing of several other one-time evolutionary transitions (the origins of life, photosynthesis, eukayotes, and sexual reproduction). They conclude that the timing of these events suggests that it typically takes far longer than the expected lifetime of a sunlike star to get all the way to intelligent life. We’re just incredibly lucky.

A more recent article carries the analysis even further. If catastrophes sometimes set back the evolution of life, and if complex organisms are more vulnerable to catastrophe than simple organisms, then we expect that more of the crucial steps in the evolution of intelligent life will happen late in the history of a planet (assuming it generates any intelligent life at all.)

In short, if this is true, tomorrow on Logarithmic History may mark a truly momentous occasion, not just in the history of the Earth, but in the history of the observable universe.

Blood and brains

Humans are brainy animals. One way to show that is by looking at brain size: our species has the biggest brains, in relation to body size, of any animal. But there’s more to it than that. An earlier post covered the work of Susan Herculano-Houzel. She developed a technique for counting the number of neurons in a brain, or part of a brain. Among most mammals, big animals have a lower density of brain neurons, so they aren’t actually as brainy (measured by neuron number) as you’d think just based on their brain size. Primates however break the usual mammalian rule. Big primates have the same neuron density as little guys, so they really are quite brainy. And humans, with really big brains and (following primate rules) a high density of neurons, stand out even among primates as exceptionally brainy.

This work isn’t much help if we are looking at extinct hominins, when all we’ve got is their fossil skulls. But now there’s some interesting recent research with a new take on the subject. Brains need to be supplied with blood. The more energy they use, the more blood flow is needed. We can now figure out fairly accurately how much blood flow a brain is getting by looking at the size of the hole that lets the carotid artery in through the base of the skull. And then we can apply this technique to look at humans, and at extinct hominins. It turns out that humans are even more exceptional when we look at blood flow to the brain: we’re getting double the flow that you’d expect based on brain size alone.

Early hominins however, Australopithecus and early Homo, aren’t very impressive upstairs, many with less blood flow to the brain than modern apes. Looking at the graph it looks like there are really two grades of brain evolution. In the lower grade, which includes early hominins and modern apes, there is a gradual increase over millions of years. (I’m just guessing here that the ancestors of chimps and gorillas millions of years ago were about as brainy as contemporary hominins, but we’d still like to find more fossils.) And then there is a big leap up to a higher grade with early Homo erectus, and a rapid increase after that. It looks like something major changed with the appearance of Homo erectus, either on the supply side – improvements in food supply making brains more affordable – or on the demand side – a greater fitness payoff to a high energy brain – or both.

High fidelity

Arms races have been a big engine of evolutionary progress, both in biological evolution and in the evolution of human societies. Another big driver has been improvements in the fidelity of inheritance. We see this in the evolution of genetic systems, including the evolution of life itself, and of the eukaryotic chromosome. And we’ll see it in human social evolution, including the evolution of language, of writing, of the alphabet, and printing.

Both arms races and improved information transmission may have been factors in the evolution of braininess.

The figure above is from the classic work of Harry Jerison, one of the pioneers in studying the evolution of brain size. It’s several steps away from the raw data, but what it shows is how mammalian Encephalization Quotients (EQs), a measure of brain size relative to body size, evolved over the Cenozoic. The figure might be read as the record of a brainy arms race between prey and predators, leading to increased variance in the EQ bell curve for both.

Primates of course are particularly brainy mammals. One popular explanation for this is a series of arms races within species, with bright monkeys and apes outwitting dimmer ones. This has been called the Machiavellian Intelligence hypothesis (or, in the case of macaques, macachiavellian intelligence).

This hypothesis may not hold up too well, however. One complication is that, contrary to what a lot of modularist evolutionary psychology might suggest, social intelligence in primates is not separate from other sorts of intelligence. The same primate species that are good at solving social problem (e.g. tricking other group members) are also clever about things like tool use and other complex foraging skills. Variation in intelligence across primate species mostly boils down to a single general factor, rather than a bunch of domain-specific aptitudes.

Also, the latest research suggests that variation in diet and ecology, such as the distinction between fruit eaters (brainy) and leaf eaters (not-so-much), accounts a lot of variation in brain size, while differences in social complexity (measured by group size) don’t seem to matter.

An alternative to the Machiavellian Intelligence hypothesis is the cultural intelligence hypothesis, with brainier animals more likely to innovate and more likely to learn others’ innovations. The first part pf this equation holds up: across various groups of organisms, including birds and primates, brainy animals are more flexible in their behavior, more likely to discover new adaptive behaviors, and more successful in colonizing novel environments. The second part is trickier. In recent years we’ve learned that learning useful information by observing others (go ahead, call it culture, if you want to annoy cultural anthropologists) is extremely widespread, and found in organisms like guppies and honeybees that no one thinks are terribly bright. So learning from others doesn’t take special smarts.

Where bigger brained animals may excel is not in how much social learning they do, but in how accurately they do it – in copying fidelity. Theoretical models of the evolution of copying suggest that accurate copying makes a big difference. Small changes in copying fidelity can lead to large changes in the persistence of cultural traits. Of course this will crucially important for human evolution: more on this in days to come.

For a wide-ranging introduction to this rapidly advancing area of research, written by a leader in the field, try Darwin’s Unfinished Symphony: How Culture Made the Human Mind.

Ground up monkey brains

Short version: It looks like most mammals, at least most large mammals, have the brains they need, while primates, especially large primates, have the brains they can afford.

Longer version: One reason for being interested in monkeys is that they’re brainy mammals. Here’s the conventional graph illustrating that:

Larger mammals tend to have larger brains, but the relationship is non-linear. Multiplying body mass by x doesn’t multiply brain mass by x. Instead it multiplies brain mass by about x^.75. In other words, Brain Mass is proportional to (Body Mass)^.75. Equivalently (taking the logarithm of both sides) Log[Brain Mass] is equal to .75 times Log[Body Mass], plus a constant. So Log[Brain Mass] plotted against Log[Body Mass] gives a straight line with a slope of .75. That means that if one mammal has 16 times the body mass of another, it’s expected to have 8 times the brain mass. 10,000 times the body mass means 1000 times the brain mass. The thing to note is that primates defy expectations. They have larger brains than would be expected based on their body sizes.

But we’ve recently learned that primates – especially big ones – are even more special than this graph suggests. Susan Herculano-Houzel has pioneered a technique that involves chopping up brains (or parts of brains), dissolving their cells to make a kind of brain soup, and counting cell nuclei. This allows her to estimate how many neurons there are in different brains.

Major findings: Among most mammals, the number of neurons increases more slowly than brain size. Increase brain size by x, and you increase number of neurons by about x^.67. (H-H shows this flipped around. Increase number of neurons by x and you increase brain mass by x^1.5.) But primates are exceptional; the relationship is nearly linear. An x-fold increase in primate brain size corresponds to about an x-fold increase in number of neurons. Humans follow the primate rule here. We have about the same density of neurons as other primates. When you combine the exceptionally large brain sizes of humans (exceptional even relative to our brainy primate relations) with a standard high primate neuron density, you get an animal with an enormous number of neurons. By contrast, a rodent with a human sized brain, if it followed rodent rules for how neuron numbers increase with brain size, would have only 1/7 as many neurons.

Neurons are expensive. Most large animals economize by cutting back on neuron density. A cubic centimeter of cow brain has fewer neurons, and consumes energy at a lower rate, than a cubic centimeter of mouse brain. By contrast, large primates are extravagant, devoting exceptionally large energy budgets to running their brains. And human brains are exceptionally costly. An important question for the study of human evolution is how we paid the bill for such costly brains. That’s a story for later. But another part of the story starts back in the early Cenozoic, when monkeys committed to a different set of rules for building brains.

And here is a chart giving absolute numbers of  cortical neurons (cneurons) for a bunch of species. Scott Alexander has some thoughts about the moral implications. Short version: skip the pork for dinner (and skip the elephant, chimp, and manflesh. But you knew that). Beef might be okay. Better is lobster.

And Werner Herzog is probably okay with you eating chicken.

Blood and brains

Humans are brainy animals. One way to show that is by looking at brain size: our species has the biggest brains, in relation to body size, of any animal. But there’s more to it than that. An earlier post covered the work of Susan Herculano-Houzel. She developed a technique for counting the number of neurons in a brain, or part of a brain. Among most mammals, big animals have a lower density of brain neurons, so they aren’t actually as brainy (measured by neuron number) as you’d think just based on their brain size. Primates however break the usual mammalian rule. Big primates have the same neuron density as little guys, so they really are quite brainy. And humans, with really big brains and (following primate rules) a high density of neurons, stand out even among primates as exceptionally brainy.

This work isn’t much help if we are looking at extinct hominins, when all we’ve got is their fossil skulls. But now there’s some interesting recent research with a new take on the subject. Brains need to be supplied with blood. The more energy they use, the more blood flow is needed. We can now figure out fairly accurately how much blood flow a brain is getting by looking at the size of the hole that lets the carotid artery in through the base of the skull. And then we can apply this technique to look at humans, and at extinct hominins. It turns out that humans are even more exceptional when we look at blood flow to the brain: we’re getting double the flow that you’d expect based on brain size alone.

Early hominins however, Australopithecus and early Homo, aren’t very impressive upstairs, many with less blood flow to the brain than modern apes. Looking at the graph it looks like there are really two grades of brain evolution. In the lower grade, which includes early hominins and modern apes, there is a gradual increase over millions of years. (I’m just guessing here that the ancestors of chimps and gorillas millions of years ago were about as brainy as contemporary hominins, but we’d still like to find more fossils.) And then there is a big leap up to a higher grade with early Homo erectus, and a rapid increase after that. It looks like something major changed with the appearance of Homo erectus, either on the supply side – improvements in food supply making brains more affordable – or on the demand side – a greater fitness payoff to a high energy brain – or both.

High fidelity

Arms races have been a big engine of evolutionary progress, both in biological evolution and in the evolution of human societies. Another big driver has been improvements in the fidelity of inheritance. We see this in the evolution of genetic systems, including the evolution of life itself, and of the eukaryotic chromosome. And we’ll see it in human social evolution, including the evolution of language, of writing, of the alphabet, and printing.

Both arms races and improved information transmission may have been factors in the evolution of braininess.

The figure above is from the classic work of Harry Jerison, one of the pioneers in studying the evolution of brain size. It’s several steps away from the raw data, but what it shows is how mammalian Encephalization Quotients (EQs), a measure of brain size relative to body size, evolved over the Cenozoic. The figure might be read as the record of a brainy arms race between prey and predators, leading to increased variance in the EQ bell curve for both.

Primates of course are particularly brainy mammals. One popular explanation for this is a series of arms races within species, with bright monkeys and apes outwitting dimmer ones. This has been called the Machiavellian Intelligence hypothesis (or, in the case of macaques, macachiavellian intelligence).

This hypothesis may not hold up too well, however. One complication is that, contrary to what a lot of modularist evolutionary psychology might suggest, social intelligence in primates is not separate from other sorts of intelligence. The same primate species that are good at solving social problem (e.g. tricking other group members) are also clever about things like tool use and other complex foraging skills. Variation in intelligence across primate species mostly boils down to a single general factor, rather than a bunch of domain-specific aptitudes.

Also, the latest research suggests that variation in diet and ecology, such as the distinction between fruit eaters (brainy) and leaf eaters (not-so-much), accounts a lot of variation in brain size, while differences in social complexity (measured by group size) don’t seem to matter.

An alternative to the Machiavellian Intelligence hypothesis is the cultural intelligence hypothesis, with brainier animals more likely to innovate and more likely to learn others’ innovations. The first part pf this equation holds up: across various groups of organisms, including birds and primates, brainy animals are more flexible in their behavior, more likely to discover new adaptive behaviors, and more successful in colonizing novel environments. The second part is trickier. In recent years we’ve learned that learning useful information by observing others (go ahead, call it culture, if you want to annoy cultural anthropologists) is extremely widespread, and found in organisms like guppies and honeybees that no one thinks are terribly bright. So learning from others doesn’t take special smarts.

Where bigger brained animals may excel is not in how much social learning they do, but in how accurately they do it – in copying fidelity. Theoretical models of the evolution of copying suggest that accurate copying makes a big difference. Small changes in copying fidelity can lead to large changes in the persistence of cultural traits. Of course this will crucially important for human evolution: more on this in days to come.

For a wide-ranging introduction to this rapidly advancing area of research, written by a leader in the field, try Darwin’s Unfinished Symphony: How Culture Made the Human Mind.