Robert Plowman is a professor of behavioral genetics at the Institute of Psychiatry, Psychology, and Neuroscience at King's College London. He previously held positions at the University of Colorado Boulder and at Pennsylvania State University, and was elected a Fellow of the Academy of Medical Sciences and of the British Academy for his groundbreaking work in behavioral genetics. And he s the author of the fascinating book, Blueprint: How DNA Makes Us Who We Are. In this episode, we talk about the birth of Behavioral Genetics, the taboos around studying the influences of genes, in particular in human psychology, controversy surrounding the topic of group differences, the first law of behavior genetics, the concept of heritability, nature and nurture, the significance of non-shared environment which is genuinely perplexing, and the prospect of this will land us in some Gattaca-like dystopia: heritability and equality of opportunity, the implications of genetics for parenting and education and other social policies, and other topics. This is important science, and this is a fascinating conversation. And I bring you Robert's interview with me. Robert is one of the most influential people in the field, and one of my good friends, Dr. Robert's book is a must-listen-to-book, so you won t want to miss this one. Thanks to Robert for being on the podcast, and for being kind enough to let me bring you his interview. . And as always, I never want money to be the reason why someone can t listen to the podcast. If you can t afford a subscription, there's an option at Samharris.org to request a free account, and we grant 100% of those requests, and you can't afford it, no questions asked asked? No questions asked, no question asked, 100% asks asked, you get it all the time by Sam Harris and as always I never wants money, I do not want money, you can do it, I just say so, I'm not going to be a millionaire, right? - Sam Harris, I love you, I know you do, I really do, right I do I do that, I'll say so. - Thank you, thank you, right, I said so, right?? - I'm making sense, I mean really, really? - Thanks, Sarah, Sarah? - Sarah? -- Sarah? -- -- Thank you? -- Sarah, Amy?
00:37:23.780Similarly, children growing up in the same family correlate about 0.3 or so.
00:37:28.660In body mass index could be, it's reasonable to think it's nurture, but if those children
00:37:34.300are genetically unrelated, their correlation is zero.
00:37:37.800The other side of the adoption design is to take genetically related people adopted apart.
00:37:43.420These adopted children who correlate zero with the body mass index of their adoptive parents correlate 0.3 with the body mass index of their birth parents, whom they never saw after the first week of life, who had no influence over their environment.
00:37:59.300So that's the sort of evidence that for decades was used to say genetics is important.
00:38:05.360But then people realized, you know, in the 70s and 80s, that it's telling us something very important about the environment.
00:38:13.820Whatever the environment is, it's not making kids in the same family similar to one another.
00:38:20.080It's not making kids similar to their parents.
00:38:23.500And that's what I called in 1987, non-shared environment.
00:39:35.740And so the parents who say are more, if you look at the relationship between parental harsh discipline and children's antisocial behavior, in a family, the child who is more antisocial, the parents are more harsh in their discipline.
00:39:54.680Well, as always, these correlations in psychology have always been assumed to be environmental.
00:40:01.680But I think all your listeners know the adage that correlation does not imply causation.
00:40:07.200Is it necessarily the case that the parents' discipline of the child caused the antisocial behavior?
00:40:13.480Or is it possible that the parents' behavior is reflecting the children's behavior?
00:40:19.160And you can put this in a behavioral genetic design.
00:40:21.480And what you find is that about half of those correlations are due to genetic differences.
00:40:27.040So this is where the nature of nurture comes in.
00:40:30.780And it kind of took me off the track, though, of non-shared environment.
00:40:35.720And the punchline there is after 30 years of trying to find these systematic sources, we haven't been successful.
00:40:41.960I know in one of your conversations with Paul Bloom, you mentioned, you know, Judith Harris's book in the 90s, which really popularized a lot of these concepts, but really results.
00:40:52.560But then also proposed maybe peers are important.
00:40:55.860And that's another reasonable hypothesis.
00:40:57.600You know, your daughters probably won't end up having the same friends.
00:41:01.360Maybe one of them has more academically oriented friends and the other has more athletically oriented friends.
00:41:30.660And it turns out that, you know, kids select friends.
00:41:34.560If they're antisocial, they select friends who are like them in that score.
00:41:39.020And about half of that is due to genetic differences.
00:41:42.540So that's what we mean by the nature of nurture.
00:41:44.500But it's also why we haven't found systematic sources of non-shared environment.
00:41:49.460Whenever we find something that looks like it's causing differences between kids and a family, it ends up being a genetic difference in disguise.
00:41:58.560So after 30 years, I came to this what we call gloomy conclusion that non-shared environment is essentially idiosyncratic, stochastic, not systematic.
00:42:09.380So that half of the variants for psychological traits are due to these environmental factors, but they're essentially random chance stuff happens.
00:42:21.720Okay, so this is all, I think, more important than may be obvious to people at first pass here.
00:42:29.540So I just want to linger on this topic.
00:42:31.200So first, what you're saying here is that virtually half of everything we care about in human nature, you know, in our psychology, you know, whether it's susceptibility to various psychopathology, and we'll talk about how we think about disorders and whether the disorder framework is the right framework here.
00:42:52.120But for virtually everything in psychology and in, you know, human difference one could care about, you know, from intelligence to, you know, big five personality traits to susceptibility to things like depression and schizophrenia, the punchline here is something like 50% of human difference.
00:43:15.220It's often on either side, it's often on either side of that halfway mark, I mean, sometimes it's 60%, sometimes it's as high as 80%, you know, later in life for things, is accounted for by genes, and the other half is environment, but it is not the environment that parents or anyone else can systematically control.
00:43:37.260And for the environmental component of things, very often half of what is ascribed to the environment is actually genes in disguise because people, based on their own genetic proclivities, wind up shaping their environment.
00:43:56.320So, I think this is an example you use in your book, you know, you could ask someone, you know, how often does it rain where you live?
00:44:02.640If ever there were an environmental variable that has nothing to do with DNA, well, you know, the weather is certainly that, but then you ask yourself, well, you know, people are free to move, right?
00:44:14.540People can pick the climates in which they live, and maybe some of that is being driven by genetic proclivity, right?
00:44:21.980There's some people who just hate living where it rains, right?
00:44:24.740You know, I count myself as one of those people.
00:44:26.940It's not an accident that I don't live in Seattle.
00:44:29.120And so it is with everything else, you know, how much TV do you watch as a kid?
00:44:35.200How often do your parents read to you?
00:44:37.520This all seems like it's a pure statement of an environmental influence, you know, i.e. nurture, and yet when you strip out the influence of genes, you find that genes are accounting for half of those so-called environmental differences among people.
00:44:54.480I should just pause there, Robert, to ask, did I summarize that point correctly?
00:45:01.460And the point for people to take home is correlation does not imply causation.
00:45:05.560So parents who read a lot to their kids have kids who do better at reading at school.
00:45:10.160And if you don't think about these issues, you might say, sure, it's environmental.
00:45:13.640But I hope after this discussion, people at least pause a minute and say, well, wait a minute now, you know, who are these parents who read a lot to their kids?
00:45:22.380And who, you know, who are these kids who do better at reading?
00:45:25.180It could be due to genetics, or increasingly, I think it's due to parents responding to genetic differences in their kids.
00:45:33.540I have six grandchildren, and I thought, with the first two, I thought, you know, what they're supposed to do is sit there and let you read to them.
00:45:40.340I remember you talking about reading Harry Potter to your older daughter.
00:45:44.160Well, that's what I thought grandchildren were supposed to do.
00:45:46.080And with one grandchild, exactly right, I could read to her all day long, and she'd say, oh, please read some more to me.
00:45:52.220But I've got another grandson who it would almost be abusive for me to make him sit there and let me read to him.
00:46:37.840Well, there's also one enormous environmental difference, too, which is the second child is growing up in the presence of the first, whereas the first had, in our case, five long years of being an only child.
00:46:51.700So, it's hard to figure out how to factor that in, but that's a non-negligible influence there.
00:46:58.920But yeah, I am noticing they're impressively similar in some ways, but they are clearly different people.
00:47:10.200In case people haven't realized this, you know, the first-degree relatives, like parents and offspring or siblings, are 50% similar genetically.
00:47:18.300But that means they're 50% different genetically.
00:47:21.380So, genetics predicts that kids in a family will be different.
00:47:26.100These socialization environmental theories, they have a lot of trouble explaining.
00:47:31.380Why are two kids in the same family with the same parents so different, when, presumably, it's these parents that are causing differences in the kids' development?
00:47:41.500Yeah, so, again, there are implications for parenting here and social policy, education.
00:47:50.780We'll get to those in the back half of our conversation.
00:47:53.920But, you know, all of this, again, is, this is a bit of a high-wire act to talk about these things without having people freak out.
00:48:00.760But we are really just talking about the facts of human psychology insofar as we have come to know them.
00:48:08.840And no doubt we will be wrong about certain things.
00:48:13.360Certain assumptions will be proven wrong in the fullness of time.
00:48:16.080But the idea that genetics doesn't account for a lot of what we care about in human nature, the door seems to be closed to that thesis.
00:48:29.220I mean, the blank slate thesis is no longer on the table.
00:48:32.280And, you know, it's empowering in some ways.
00:48:35.820It pushes your intuitions around in others.
00:49:12.920And then, when we find correlations, like between parents reading to kids and kids reading ability at school, you can't assume that's environmental.
00:49:21.160They're often genetic effects in disguise.
00:49:22.920So, I find what helps people put this together is if I tell you that if one of your daughters had been switched at birth in the maternity ward and raised in a different family, she would have grown up to be very similar to who she is, even though she was raised in a different family.
00:49:41.060And that's not hypothetical, because we have studies of identical twins reared apart.
00:49:46.720And this wonderful documentary that won an award last year called Three Identical Strangers, about three identical twins, and just how similar they are despite being raised in quite different family environments.
00:49:59.480So, it's a dramatic illustration of this point, because your daughter would be her identical twin.
00:50:05.380She's still 100% genetically who she is, even though she's raised in a different family.
00:50:10.540So, I think that helps people to understand it, that we'd be very much who we are, even if we had been raised in a different family with different parents.
00:50:45.100There are monozygotic twins and dizygotic twins, identical and fraternal twins.
00:50:50.040And identical twins share the same DNA.
00:50:54.420They're 100% identical, barring some surprising mutation genetically.
00:51:01.140And whereas fraternal twins are like ordinary siblings, they share 50% of their DNA, but they just share the same environment all the way down to the womb.
00:51:13.980So, in these studies where you can compare identical twins to fraternal twins, and you can really strip out the influence of shared environment.
00:51:26.200Because, again, you're looking at one group that has identical DNA and one group that has only 50% similar DNA, and yet shared environment.
00:51:36.000And then you have these other studies where you have identical twins separated at birth and raised in different families.
00:51:43.200And you are a pioneer in doing this work.
00:51:47.000And maybe, I don't know if you want to talk about the Colorado Adoption Project, or I don't know how you want to enter this.
00:51:53.000But let's talk a little bit more about the logic of these experiments and why they have been so compelling.
00:51:58.580Well, that was a great description of the twin method.
00:52:01.980But the punchline there is that if a trait, take like musical ability, which hasn't been studied very much, you know, and it's hard to measure.
00:52:10.500But what we'd be saying is if genetic influences are important, you'd have to predict that identical twins would be more similar in their musical ability than non-identical twins.
00:52:20.280And the extent to which they're more similar than fraternal twins, I say non-identical because in UK they call them non-identical rather than fraternal.
00:52:30.160So if a trait like musical ability is heritable, you'd have to predict that the MZ, monozygotic twins, are more similar than the dizygotic fraternal twin.
00:52:40.060And the extent to which identical twins are more similar estimates the magnitude of genetic influence.
00:52:46.780So as you said, that's a pretty powerful test of genetic influence.
00:52:51.220But the main assumption there is called the equal environments assumption.
00:52:54.780What if identical twins are treated more similarly than non-identical twins?
00:53:13.280But the neat thing is we have this other method that's completely different called the adoption method.
00:53:18.140And that's a wonderful situation to be in because the adoption method also has its possible, well, it has its assumptions and possible flaws, but they're completely different.
00:53:29.120And these two methods, the twin method and the adoption method, converge on this conclusion that everything that we study in psychology is heritable.
00:53:37.580So the adoption method, though, is in some ways more powerful.
00:53:41.340You can really see it with identical twins reared apart.
00:53:46.480So much more typical are biological parents who adopt their child away at birth.
00:53:52.360And then you can study those adopted children and their adoptive parents who give them their family environment, but not their genes.
00:54:00.040They're not genetically similar to them.
00:54:01.740So it's another powerful way of getting at genetic environmental influences.
00:54:07.180And I gave you the example of body mass index and how adopted children don't correlate with their adoptive parents in terms of body mass index, even though they share food and lifestyle.
00:54:18.660Whereas parents who share genes and environment with their children correlate about 0.3 for body mass index.
00:54:25.080And the real killer data is that these adopted children correlate 0.3 for body mass index with their birth parents who they never saw after the first week of life.
00:54:37.980So I think together, that's a very powerful indication, not only of genetic influence, but of the unimportance of what we call shared environment.
00:54:47.780You know, that traditional view of nurture.
00:54:50.760Which is, we should just pause to acknowledge how counterintuitive this is.
00:54:54.500We're talking about parents who have their own eating habits, which they then lavish upon their children from birth onward.
00:55:04.260And it turns out those habits, stripped of their underlying genetic cause, is not what contributes to the body mass index of a child as he or she grows up.
00:55:29.700When I give a public lecture, it's sort of the first question I get is, yeah, but what about epigenetics?
00:55:34.860You know, the environment changes genes.
00:55:36.560Well, you know, as I say in the book, you only inherit differences in DNA sequence.
00:55:44.140You start life as a single cell with half, 3 billion base pairs of DNA from your mother and 3 billion from your father.
00:55:52.240And that DNA is the same DNA in the trillions of cells in your body.
00:55:56.580We do pick up some mutations as we go along.
00:55:59.260But the genes that are expressed of those 3 billion nucleotide bases of DNA in the double helix of DNA, we don't have the same DNA expressed in all of our cells.
00:56:17.520You know, the cells in your liver do different things from the cells in your blood and from the cells in your brain.
00:56:24.620Different bits of DNA are turned on and off in response to the environment.
00:56:28.940But what we inherit are the DNA differences.
00:56:34.760And if a DNA difference correlates with an outcome like schizophrenia or alcoholism or reading disability, then that means that that DNA difference was expressed somewhere.
00:56:50.320But some, you know, people have really used epigenetics, which literally means above genetics, beyond genetics, to try and argue against Mendelian genetics.
00:56:58.940And I think there's, after the initial excitement about epigenetics, I think people are calming down about it and realizing, yeah, gene expression is important.
00:57:09.380Everything between inherited DNA and behavior is important.
00:57:13.300We call that expression transcriptomics and tabulomics and the brain.
00:57:21.160Everything in between DNA and behavior is important to understand.
00:57:24.140But it's important to realize all we inherit are DNA sequence differences.
00:57:31.060And if they're making a difference in terms of traits and if they're correlating with differences, individual differences in traits, well, then they're being expressed on some level.
00:57:52.040But that's not to say all of these other things are not important.
00:57:55.980But I just do, I'd like to argue against this idea that epigenetics somehow invalidates genetics because it doesn't.
00:58:03.140Yeah, there's another detail here which is interesting and has important implications, and it's that we're not tending to talk about single genes having some overwhelming trait effect.
00:58:19.200We're talking about thousands of genes contributing tiny effects to any one of these traits, whether it's a susceptibility to schizophrenia or intelligence or anything else that interests us.
00:58:37.060Well, it has, you tell me what significance you see.
00:58:40.780I mean, the one thing that jumped out for me immediately, which I believe you mentioned in your book, is that it gives a somewhat less than hopeful picture that any single drug target will be a high leverage target for us in improving ourselves in whatever way we might hope to.
00:59:00.940Well, the most important thing we've learned from the DNA revolution in the last 10, really 5 years, is that genetic influence on complex traits and common disorders of the sort we've been talking about are not due to one gene, certainly.
00:59:54.880Everyone learns about genetics from Mendel.
00:59:57.120And Mendel was studying disorders in pea plants, like wrinkled seeds, you know.
01:00:02.000And so they're hardwired and deterministic.
01:00:05.720And he showed through that that that's the way genes work in heredity.
01:00:10.780But what's important to realize is that despite these thousands of single gene disorders, many of which are extremely debilitating, lethal for the people who have them, they're very rare, one in 100,000, one in 500,000.
01:00:27.040So they're very rare, fortunately, and don't really contribute much to the heritability of the traits that we study.
01:00:35.620The heritability of complex traits and common disorders, medical as well as psychological, are due to thousands of tiny DNA differences.
01:00:51.680If you're trying to do a bottom-up approach, as neuroscientists would want to do, where you go from genes to brain to behavior, it's going to be very hard if each of those DNA effects are so tiny.
01:01:03.420You know, you're going to definitely have to get away from a modular approach to neuroscience, where you think, you know, this gene does this, and then that has that effect.
01:01:13.220You know, it's going to take more like a systems network sort of approach to be able to deal with the brain from this perspective, which we talk about as polygenic.
01:01:23.380That is, every trait is influenced by many, many genes.
01:01:26.420And that would include traits in the brain, you know, neurotransmitter levels, whatever.
01:01:31.040But the other word that's important is pleiotropy.
01:01:36.020Every DNA difference has many, many effects.
01:01:40.500So, you know, you name these genes based on a disorder, you know, like this gene caused diabetes, but then you find out that gene affects hundreds of other things.
01:01:52.040So, this polygenic point that you're making is critically important, and it's really hard for people to understand because they're still thinking about genetics from a single gene, hardwired, deterministic perspective.
01:02:09.080So, let's break apart a few of these concepts.
01:02:11.820So, yeah, there's an analog point to make about the brain.
01:02:15.240There are very few parts of the brain that only do one thing, where you can say this is the part of the brain that, you know, recognizes faces, right?
01:02:24.840Well, you know, even fusiform cortex does other things.
01:02:27.800So, the real picture is, you know, of pleiotropy, where any one gene, in this case, contributes to many traits, and also, and this is a point you make in the book, our concept of disorders, like schizophrenia, is itself misleading.
01:02:48.540And it makes sense to talk more in terms of dimensions for traits, as opposed to these kind of terminal disorders.
01:02:56.300And you use an analogy, which really drives home the point, with height, you know, and maybe you want to talk about height and the imaginary problem of giantism to clarify this concept.
01:03:37.840It's a matter of more or less, that is any DNA difference that is more prevalent in the giants, it will be distributed in the distribution.
01:03:47.380So, people who are higher than average are more likely to have that DNA difference.
01:03:51.860So, the DNA research, I think, puts the nail in the coffin of diagnoses.
01:03:57.940Now, you might say, well, that's just a stupid example.
01:04:00.280I mean, why would anyone divide height, which is so normally distributed, into a dichotomy?
01:04:08.000But I think that's what we're doing with most other disorders.
01:04:12.180You know, depression, no one thinks depression, you wake up one day and you're depressed.
01:04:17.100Depressive symptoms are almost, they're quantitatively distributed.
01:04:21.440And you never find genes for a disorder.
01:04:26.280Any gene you find is distributed through the population.
01:04:30.240Like, more concretely, one of the first of these effects that were identified using these new approaches called genome-wide association was a DNA difference that was associated with body mass index.
01:04:42.580So, this gene had an A and a T, you know, the four nucleotide bases of DNA, A, C, T's, and G's.
01:04:51.320And in the old, in thousands of years ago, we were all TT.
01:04:56.060But then someone got a mutation that was an A.
01:04:58.900And that A seemed to have been adaptive.
01:05:01.620The story used to be that it allows you to conserve fat.
01:05:04.760And in the Stone Age, that would be a good thing because you never knew when your next meal was coming.
01:05:09.540But now, that makes you more likely to become obese in a fast food nation.
01:05:16.840So, if you have two A's, you're three pounds heavier than someone who has one A.
01:05:22.260And if you have no A's, well, that one A makes you three pounds heavier than someone with no A's, TT.
01:05:29.160So, there's a six-pound difference between TT and AA.
01:05:32.280That's what we mean by an association.
01:05:34.700So, that was found for obesity initially.
01:05:37.100But then they found that that DNA difference works quantitatively throughout the distribution.
01:05:44.880That is, if you and your sibling, you have an A and they don't, you're likely to be, if we get a lot of siblings like that, three pounds heavier on average.
01:05:56.380But that only accounts for 1% of the variance of body mass index.
01:06:03.220And when that was published in Science in 2007, people, oh, well, 1%, I mean, what's that?
01:06:08.800Turns out, it's one of the biggest effects that we can find for complex disorders, complex traits in common disorders.
01:06:15.480So, it's so important to realize that these polygenic scores, that is, you can put these thousands of DNA differences together, because any one of them just doesn't account for enough variance to predict or to try and understand it mechanistically.
01:06:32.480But you can put them all together, aggregate them in a polygenic score, and make pretty substantial predictions.
01:06:42.640Like, we can predict 25% of the variance in height and about 10% of the variance in weight by putting all of these together.
01:06:51.620So, for weight, this one DNA difference I was talking about accounts for 1%, but then these other DNA differences account for 9%.
01:07:01.540So, altogether, you can predict about 10% of the variance.
01:07:04.960But these polygenic scores are all necessarily perfectly normally distributed, because it's the central limit theorem of statistics.
01:07:16.000You know, you flip a coin, and you flip 100 coins, and you get this normal distribution of heads and tails.
01:07:25.300You know, you either have one allele or the other allele, or two of them.
01:07:29.140So, these polygenic scores are perfectly normally distributed, so that the genetic liability for everything, any disorder, autism, schizophrenia, coronary heart disease, it's perfectly normally distributed.
01:07:42.660So, I think that is really, I think, ought to put the nail in the coffin of diagnoses, because I really believe in psychiatry and psychology.
01:07:51.280These diagnoses have held us back tremendously, and all of the DNA studies, these genome-wide association studies, are case control studies.
01:08:02.200So, the whole game is to find these people who meet these, what I think are arbitrary diagnostic criteria, and you call them cases, like schizophrenics, and everybody else is a control.
01:08:16.060And that's really held us back, because it's just simply not true.
01:08:20.060Yeah, because everybody else who's normal, so-called normal, could be just like the 6'3 person who's not classed as a giant, but still shares all of these increased height probability genes.
01:08:32.560And it means that we all have thousands of genes for schizophrenia.
01:08:40.200And I think, you know, if you have a very high, we call it polygenic score for schizophrenia, it probably takes, we all have stresses that would freak us out.
01:08:49.920And as you've mentioned several times in your podcast, if you did have a genetic propensity towards schizophrenia, you probably ought to be careful about some of the psychedelics, for example.
01:09:08.200You know, if you have the genetic propensity, it doesn't mean you're going to become alcoholic, or you're necessarily going to become schizophrenic.
01:09:14.140You're just more likely to be, and given the stresses and strains of life, you're more likely to be tipped over the edge than someone else.
01:09:24.700And these genes are very likely contributing to who you are in noticeable ways that puts you on this spectrum, which has schizophrenia as its terminus.
01:09:38.240So, like, these genes for height that would render a giant 6'5 or beyond are also operative in you at the height of 5'10.
01:09:50.460It's just you have a different complement of, and we're now talking about many, many genes for any one of these traits.
01:09:58.140The picture is, I think the phrase you used is that the abnormal is normal.
01:10:01.920I mean, we're all on every spectrum that we could posit exists in the population.
01:10:08.680We're all somewhere on it, and whether we have a symptomology that's interesting or not is the only difference, right?
01:10:16.160So it gives us a finer-grained way of thinking about human difference and the boundaries between what is considered, you know, normative or normal and pathological.
01:10:30.680I mean, it's common to—I think you referenced this in your book—ever since Aristotle, the analogy between madness and genius has been drawn, and I think probably too much has been made of that.
01:10:42.940But this is susceptible to, in the end, a genetic analysis.
01:10:47.120I mean, we can look at the genotype of whatever we want to call genius and the genotype of whatever we want to call madness and just see how much genetic real estate they share.
01:11:15.540We're alleviating symptoms rather than curing a disorder.
01:11:20.220It all has to do with psychology-aping medical sciences, where, you know, a lot of this does work if you have a simple cause, a simple environmental cause.
01:11:31.000If you'd like to continue listening to this podcast, you'll need to subscribe at samharris.org.
01:11:38.360You'll get access to all full-length episodes of the Making Sense podcast and to other subscriber-only content, including bonus episodes and AMAs and the conversations I've been having on the Waking Up app.
01:11:50.300The Making Sense podcast is ad-free and relies entirely on listener support.
01:11:54.600And you can subscribe now at samharris.org.