WEBVTT 00:00:00.000 --> 00:00:02.859 You know, there's this this persistent idea that 00:00:02.859 --> 00:00:05.099 really groundbreaking science is always done 00:00:05.099 --> 00:00:08.119 in a pristine whitewalled laboratory. Right. 00:00:08.259 --> 00:00:10.919 Yeah. You picture the stainless steel, the bubbling 00:00:10.919 --> 00:00:13.160 beakers, maybe a supercomputer humming in the 00:00:13.160 --> 00:00:15.640 corner, hazmat suits, the whole thing. Exactly. 00:00:15.759 --> 00:00:18.339 But history has a funny way of contradicting 00:00:18.339 --> 00:00:21.260 that image, because sometimes the biggest shifts 00:00:21.260 --> 00:00:24.719 in how we see our own species happen in the most. 00:00:25.579 --> 00:00:28.000 The most mundane place is imaginable. Like a 00:00:28.000 --> 00:00:30.600 greyhound bus rattling through rural Illinois 00:00:30.600 --> 00:00:34.119 in the early 1970s. Yes, a bus ride. That is 00:00:34.119 --> 00:00:36.700 exactly where we're starting today. For this 00:00:36.700 --> 00:00:38.560 deep dive, we're looking at a stack of sources 00:00:38.560 --> 00:00:41.460 centered around one specific landmark paper from 00:00:41.460 --> 00:00:44.820 1972. Titled The Apportionment of Human Diversity 00:00:44.820 --> 00:00:47.920 by the evolutionary biologist Richard Lewontin. 00:00:48.320 --> 00:00:49.939 And the reason we're doing a deep dive into this 00:00:49.939 --> 00:00:52.020 today isn't just because it's a classic piece 00:00:52.020 --> 00:00:53.899 of biology though. I mean, it absolutely is. 00:00:54.020 --> 00:00:56.579 It's because this paper essentially took a sledgehammer 00:00:56.579 --> 00:00:59.020 to one of the most ingrained concepts in human 00:00:59.020 --> 00:01:01.679 history, which is the biological concept of race. 00:01:01.939 --> 00:01:05.170 And it did it using math. I think that's the 00:01:05.170 --> 00:01:07.709 key here. Yeah. Before this paper dropped, the 00:01:07.709 --> 00:01:10.650 conversation about race was largely visual or 00:01:10.650 --> 00:01:12.689 social or philosophical. It was what you could 00:01:12.689 --> 00:01:15.590 see. Right. But Lewontin turned it into a raw 00:01:15.590 --> 00:01:18.090 numbers game. He asked a question that seems 00:01:18.090 --> 00:01:21.209 simple, but is actually incredibly complex. Yeah. 00:01:21.269 --> 00:01:23.769 When we look at the biological data, do these 00:01:23.769 --> 00:01:26.629 racial categories we use every day actually exist? 00:01:26.909 --> 00:01:29.209 But it wasn't the final word, was it? Because 00:01:29.209 --> 00:01:31.750 as we went through the sources for this, and 00:01:31.750 --> 00:01:34.329 we have articles from the Royal Society, critiques 00:01:34.329 --> 00:01:36.750 from Cambridge geneticists, even some pretty 00:01:36.750 --> 00:01:38.930 heated social media threads from just recently. 00:01:39.109 --> 00:01:40.930 Oh, yeah. It's still a huge debate. Right. It 00:01:40.930 --> 00:01:43.109 became really clear that this 50 -year -old calculation 00:01:43.109 --> 00:01:46.269 is still sparking massive fights online. It's 00:01:46.269 --> 00:01:48.030 basically the definition of a zombie debate. 00:01:48.230 --> 00:01:50.750 It just won't die. You have the original finding, 00:01:50.870 --> 00:01:52.930 which was a total bombshell. Then you have the 00:01:52.930 --> 00:01:55.689 mathematical counterargument, often called Lewontin's 00:01:55.689 --> 00:01:58.150 fallacy, which we will definitely get into. We 00:01:58.150 --> 00:02:02.010 have to. And now you have political groups weaponizing 00:02:02.010 --> 00:02:04.709 that counter argument on the Internet. It's a 00:02:04.709 --> 00:02:07.230 mess, but it's a fascinating mess because it 00:02:07.230 --> 00:02:09.849 sits right at the intersection of hard math and 00:02:09.849 --> 00:02:12.810 human identity. So the mission for you, the listener, 00:02:12.969 --> 00:02:16.530 and for us today is to parse that mess. We need 00:02:16.530 --> 00:02:19.509 to understand what Lewontin actually calculated 00:02:19.509 --> 00:02:22.729 on that bus ride, why a mathematician named Edwards 00:02:22.729 --> 00:02:25.469 said he missed the point, and where the scientific 00:02:25.469 --> 00:02:28.610 consensus actually lands today. We should probably 00:02:28.610 --> 00:02:31.909 set the scene first, though. Back up to the 1960s. 00:02:31.909 --> 00:02:35.169 Let's do it. What was the scientific vibe regarding 00:02:35.169 --> 00:02:37.750 race at the time? It was a bit of a battleground. 00:02:37.969 --> 00:02:40.210 You have to remember, for centuries, race was 00:02:40.210 --> 00:02:42.210 just treated as a given. You look different, 00:02:42.289 --> 00:02:44.370 therefore you're biologically different. But 00:02:44.370 --> 00:02:47.189 by the 60s, you had two major schools of thought 00:02:47.189 --> 00:02:49.490 clashing in anthropology. Right. Our source material 00:02:49.490 --> 00:02:51.870 calls this the Kleins versus Buckets debate. 00:02:52.439 --> 00:02:54.439 Which I love as a visual. It's a perfect way 00:02:54.439 --> 00:02:56.199 to visualize it. Yeah. On one side, you had the 00:02:56.199 --> 00:02:58.460 buckets camp. These were scientists like Siedosius 00:02:58.460 --> 00:03:01.500 Dubzanski. Who was a giant in the field. A total 00:03:01.500 --> 00:03:04.139 giant. And actually, Lewontin's mentor, which 00:03:04.139 --> 00:03:06.360 adds a bit of personal drama to this whole thing. 00:03:06.500 --> 00:03:09.659 But Dubzanski and his camp believed that human 00:03:09.659 --> 00:03:12.599 races were discrete populations. Meaning you 00:03:12.599 --> 00:03:15.060 could draw a hard circle around a group of people, 00:03:15.120 --> 00:03:18.259 say, Europeans or Africans, and say, this is 00:03:18.259 --> 00:03:21.599 a distinct biological bucket. Precisely. They 00:03:21.599 --> 00:03:23.240 believed that if you looked at gene frequencies, 00:03:23.520 --> 00:03:26.099 distinct racial groups were valid biological 00:03:26.099 --> 00:03:28.379 realities. They thought the buckets were real. 00:03:28.740 --> 00:03:30.979 But on the other side, you had the Klein's camp, 00:03:31.139 --> 00:03:33.840 led by anthropologist Frank B. Livingston. And 00:03:33.840 --> 00:03:36.330 he famously said, There are no races. There are 00:03:36.330 --> 00:03:38.990 only clines. Yeah. So for anyone who hasn't taken 00:03:38.990 --> 00:03:41.090 an anthropology class recently, let's unpack 00:03:41.090 --> 00:03:43.710 cline. It's essentially a gradient, right? Yes. 00:03:43.810 --> 00:03:46.110 Think of a hill or an incline. Imagine walking 00:03:46.110 --> 00:03:49.110 from Oslo, Norway, all the way down to Lagos, 00:03:49.110 --> 00:03:51.689 Nigeria. You're walking from a very high latitude 00:03:51.689 --> 00:03:54.610 down to the equator. That's a long walk. A very 00:03:54.610 --> 00:03:57.129 long walk. But the point is, you aren't going 00:03:57.129 --> 00:03:59.569 to step over a magic line on the ground where 00:03:59.569 --> 00:04:01.669 people suddenly stop being one race and start 00:04:01.669 --> 00:04:04.000 being another. It's a smooth transition. Every 00:04:04.000 --> 00:04:05.900 village looks pretty much like the village five 00:04:05.900 --> 00:04:09.759 miles back. Exactly. Skin color, height, hair 00:04:09.759 --> 00:04:12.759 texture. They change gradually based on geography 00:04:12.759 --> 00:04:16.139 and exposure to the sun. Livingstone argued that 00:04:16.139 --> 00:04:18.300 because these changes are continuous, drawing 00:04:18.300 --> 00:04:20.699 a line and saying here's where this race ends 00:04:20.699 --> 00:04:23.839 is scientifically arbitrary. You're trying to 00:04:23.839 --> 00:04:26.899 force a fluid spectrum into a rigid box. Exactly. 00:04:26.899 --> 00:04:29.720 So up until the late 60s, this was mostly a semantic 00:04:29.720 --> 00:04:33.199 fight. One side looks at skin and says, I see 00:04:33.199 --> 00:04:35.639 a gradient. The other side says, I see enough 00:04:35.639 --> 00:04:37.639 difference to make a category. It feels like 00:04:37.639 --> 00:04:39.699 they were just talking past each other. They 00:04:39.699 --> 00:04:41.720 were. And that's largely because they were looking 00:04:41.720 --> 00:04:44.759 at phenotypes, the outside stuff you can see. 00:04:45.000 --> 00:04:47.300 But then technology shifted. And this is the 00:04:47.300 --> 00:04:49.600 real pivot point. We moved from looking at phenotypes 00:04:49.600 --> 00:04:51.740 to looking at genotypes, or at least the molecular 00:04:51.740 --> 00:04:53.920 products of genes. Like blood group proteins. 00:04:54.240 --> 00:04:56.639 Yes, blood proteins and enzymes. So we could 00:04:56.639 --> 00:04:59.740 finally ask, OK, we look different, but how different? 00:04:59.790 --> 00:05:02.420 are we chemically on the inside? Which brings 00:05:02.420 --> 00:05:04.839 us to the setup for our guy, Richard Lewontin. 00:05:05.040 --> 00:05:08.759 So Dobzhansky, his mentor and the big bucket 00:05:08.759 --> 00:05:12.660 guy, asks his student, Lewontin, to write a chapter 00:05:12.660 --> 00:05:15.519 for a book on evolutionary biology. Right. He 00:05:15.519 --> 00:05:17.560 basically hands him this brand new molecular 00:05:17.560 --> 00:05:20.779 data and expects him to finally prove that race 00:05:20.779 --> 00:05:24.079 is real on a firm quantitative basis. Talk about 00:05:24.079 --> 00:05:26.259 a backfire. I mean, Dobzhansky probably thought, 00:05:26.379 --> 00:05:28.420 finally the math is going to prove me right. 00:05:28.879 --> 00:05:31.319 So Lewontin takes the assignment and he decides. 00:05:31.370 --> 00:05:34.529 to run the numbers himself. But here's the kicker. 00:05:34.610 --> 00:05:37.370 He doesn't have access to a university mainframe 00:05:37.370 --> 00:05:39.990 to run these complex statistics. He has a lecture 00:05:39.990 --> 00:05:42.149 to give down in Carbondale, Illinois. Right. 00:05:42.189 --> 00:05:44.329 So he hops on a bus south of Chicago. A three 00:05:44.329 --> 00:05:47.990 or four hour ride. And his toolkit is just laughably 00:05:47.990 --> 00:05:50.129 analog by modern standards. I mean, today we 00:05:50.129 --> 00:05:53.129 have AI that sequences genomes in seconds. What 00:05:53.129 --> 00:05:54.910 was he actually working with on that bus? It 00:05:54.910 --> 00:05:57.209 really is a time capsule. He had a stack of reference 00:05:57.209 --> 00:05:59.329 books full of blood type data from around the 00:05:59.329 --> 00:06:01.769 world. He had a notepad, a little hand calculator. 00:06:02.350 --> 00:06:04.750 And my absolute favorite part, he had a table 00:06:04.750 --> 00:06:07.490 of logarithms to do the multiplication. A table 00:06:07.490 --> 00:06:10.129 of logarithms? I tried to explain log tables 00:06:10.129 --> 00:06:12.029 to a younger colleague the other day, and they 00:06:12.029 --> 00:06:13.990 looked at me like I was churning my own butter. 00:06:14.649 --> 00:06:17.250 It's a lost art. It really is. It's incredibly 00:06:17.250 --> 00:06:19.810 tedious work. He sat there looking up numbers 00:06:19.810 --> 00:06:21.910 in the books, using the log tables to multiply 00:06:21.910 --> 00:06:24.189 them, writing down the results. He spent that 00:06:24.189 --> 00:06:26.649 entire bus ride manually calculating something 00:06:26.649 --> 00:06:29.370 called the fixation index or FST. Let's break 00:06:29.370 --> 00:06:32.009 that term down because fixation index sounds 00:06:32.009 --> 00:06:34.329 a bit intimidating. What is he actually measuring 00:06:34.329 --> 00:06:37.149 there with his notepad? At its simplest, it's 00:06:37.149 --> 00:06:39.949 a way of partitioning diversity. It's a statistical 00:06:39.949 --> 00:06:42.410 method to figure out where the variation in a 00:06:42.410 --> 00:06:45.139 system is actually coming from. Okay. Imagine 00:06:45.139 --> 00:06:48.339 you have a few bags of mixed marbles. The FST 00:06:48.339 --> 00:06:50.899 tells you how much of the color variation is 00:06:50.899 --> 00:06:53.120 due to the difference between the bags versus 00:06:53.120 --> 00:06:56.120 the variation within a single bag. Okay. So Lewontin 00:06:56.120 --> 00:06:57.959 is asking, if I look at all the genetic variation 00:06:57.959 --> 00:07:01.740 in humans, that tiny 0 .1 % of DNA that makes 00:07:01.740 --> 00:07:04.519 us different, where is it located? Is it found 00:07:04.519 --> 00:07:06.699 between the races or is it found between individuals? 00:07:07.220 --> 00:07:10.879 Exactly. He categorized the data into seven huge 00:07:10.879 --> 00:07:13.680 classically defined racial groups. to test the 00:07:13.680 --> 00:07:16.360 bucket theory. And he used the standard classifications 00:07:16.360 --> 00:07:19.879 of the 1970s. Right, Caucasians, Africans, Mongoloids, 00:07:20.000 --> 00:07:23.300 South Asian Aborigines, Arans, Oceanians, and 00:07:23.300 --> 00:07:26.660 Australian Aborigines. He used 17 indirect markers 00:07:26.660 --> 00:07:28.899 since full DNA sequencing wasn't available yet. 00:07:29.040 --> 00:07:31.220 And he wanted to know, is the difference between 00:07:31.220 --> 00:07:34.079 these huge groups the main source of human diversity? 00:07:34.569 --> 00:07:38.050 And this leads us to the 85 .4 % rule. This is 00:07:38.050 --> 00:07:40.170 the number that seemingly ended the debate and 00:07:40.170 --> 00:07:42.589 cemented Lewontin's legacy. It was a statistical 00:07:42.589 --> 00:07:45.230 bombshell. Lewontin found that if you take all 00:07:45.230 --> 00:07:48.649 human genetic variation, a massive 85 .4 % of 00:07:48.649 --> 00:07:50.889 it exists within any given local population. 00:07:51.110 --> 00:07:53.050 Let's contextualize that for a second because... 00:07:53.240 --> 00:07:56.100 85 % is a huge number. If I go to a single town 00:07:56.100 --> 00:07:58.879 in, say, Poland or a village in Thailand, 85 00:07:58.879 --> 00:08:01.500 % of the total genetic diversity of the entire 00:08:01.500 --> 00:08:04.040 human species can be found just between the random 00:08:04.040 --> 00:08:07.480 individuals in that one town. Yes, that is exactly 00:08:07.480 --> 00:08:10.160 what it means. The genetic gap between you and 00:08:10.160 --> 00:08:12.800 your next -door neighbor is, statistically speaking, 00:08:13.040 --> 00:08:15.480 the vast majority of the difference that exists 00:08:15.480 --> 00:08:18.579 in humans anywhere. The variation within the 00:08:18.579 --> 00:08:21.699 group is massive. That is just so counterintuitive. 00:08:21.720 --> 00:08:24.279 We feel like our neighbors like us and someone 00:08:24.279 --> 00:08:26.500 from across the ocean is different. That's the 00:08:26.500 --> 00:08:28.439 visual bias talking. The math says otherwise. 00:08:28.959 --> 00:08:31.300 He then looked at the differences between populations 00:08:31.300 --> 00:08:33.519 within the same race, like the difference between 00:08:33.519 --> 00:08:35.600 Swedes and Italians. And what was that number? 00:08:35.740 --> 00:08:39.139 That accounted for about 8 .3%. Okay, so 85 % 00:08:39.139 --> 00:08:42.019 is individual, 8 % is local nationality. What 00:08:42.019 --> 00:08:44.779 was left for the big races? The buckets. Only 00:08:44.779 --> 00:08:47.779 6 .3%. That's it. That's it. All the history 00:08:47.779 --> 00:08:50.279 of us versus them, the segregation, the global 00:08:50.279 --> 00:08:53.720 identity politics, it's all based on 6 .3 % of 00:08:53.720 --> 00:08:57.299 our genetic variation. That is the finding. Lewontin 00:08:57.299 --> 00:08:59.220 concluded that because the variation between 00:08:59.220 --> 00:09:01.919 races is such a tiny fraction of the total, the 00:09:01.919 --> 00:09:04.659 concept of race has, in his words, virtually 00:09:04.659 --> 00:09:07.940 no genetic or taxonomic significance. He basically 00:09:07.940 --> 00:09:10.159 said the buckets aren't real. Or if they are, 00:09:10.320 --> 00:09:12.789 they are so shallow they don't matter. Now, I 00:09:12.789 --> 00:09:15.110 can hear the skeptics. And frankly, I'd ask this 00:09:15.110 --> 00:09:18.470 too. Was Lewontin just biased? Maybe he was a 00:09:18.470 --> 00:09:20.690 progressive guy who wanted to prove race didn't 00:09:20.690 --> 00:09:23.370 exist, so he found the numbers to match. That 00:09:23.370 --> 00:09:25.529 is a very fair challenge, and it's one that has 00:09:25.529 --> 00:09:28.429 been thrown at him for decades. Right. But we 00:09:28.429 --> 00:09:31.710 actually have a control group of sorts. In 1960, 00:09:32.029 --> 00:09:35.509 Lewontin used these exact same statistical methods 00:09:35.509 --> 00:09:39.120 on fruit flies. Drosophila persimilis. And what 00:09:39.120 --> 00:09:40.919 happened with the flies? Did they have races? 00:09:41.200 --> 00:09:43.799 They did. He found high variance between the 00:09:43.799 --> 00:09:46.799 populations. The math showed that fruit fly populations 00:09:46.799 --> 00:09:50.179 were distinct races. So the method works. It 00:09:50.179 --> 00:09:52.559 is entirely capable of finding race if it exists. 00:09:52.840 --> 00:09:55.620 But the human data just didn't support it. Right. 00:09:55.720 --> 00:09:57.759 The fact that he didn't find it in humans suggests 00:09:57.759 --> 00:10:00.320 the biological reality of humans is just fundamentally 00:10:00.320 --> 00:10:02.240 different from fruit flies. We just haven't been 00:10:02.240 --> 00:10:04.240 isolated long enough, have we? We move around 00:10:04.240 --> 00:10:06.840 too much. Exactly. Humans are a young species 00:10:06.840 --> 00:10:09.559 and we are incredibly mobile. We haven't had 00:10:09.559 --> 00:10:11.580 the millions of years of separation needed to 00:10:11.580 --> 00:10:14.019 develop into true subspecies. We are a messy 00:10:14.019 --> 00:10:18.279 mixing pot. So 1972, the paper drops. It effectively 00:10:18.279 --> 00:10:21.419 ends the old measure the skull era of anthropology. 00:10:21.820 --> 00:10:25.529 It stands as the consensus for decades. But science 00:10:25.529 --> 00:10:27.450 is a pendulum. You push it one way, eventually 00:10:27.450 --> 00:10:30.750 someone pushes back. Always. And that push came 00:10:30.750 --> 00:10:33.970 in 2003 from a Cambridge geneticist named AWF 00:10:33.970 --> 00:10:36.669 Edwards. And he didn't pull any punches. He titled 00:10:36.669 --> 00:10:39.889 his paper, Human Genetic Diversity, Lewontin's 00:10:39.889 --> 00:10:43.029 Fallacy. Ouch. Calling a landmark study a fallacy 00:10:43.029 --> 00:10:45.830 is academic fighting words. What was Edwards' 00:10:45.990 --> 00:10:48.720 problem with the bus ride math? Well... Importantly, 00:10:48.759 --> 00:10:50.940 he didn't dispute the arithmetic. He agreed that 00:10:50.940 --> 00:10:54.100 85 % of variation is within groups. He didn't 00:10:54.100 --> 00:10:56.580 say Lewontin calculated it wrong. His problem 00:10:56.580 --> 00:10:58.379 was with how Lewontin looked at the data points 00:10:58.379 --> 00:11:00.580 relative to each other. He argued Lewontin was 00:11:00.580 --> 00:11:02.559 missing the correlation, right? Yes. This is 00:11:02.559 --> 00:11:04.440 the technical crux of the deeper dive, so stick 00:11:04.440 --> 00:11:06.360 with me here. Yeah. Lewontin looked at genes 00:11:06.360 --> 00:11:09.779 one by one, locus by locus. He asked, does blood 00:11:09.779 --> 00:11:12.179 type separate these groups? No. Does this enzyme 00:11:12.179 --> 00:11:15.259 separate them? No. And then he averaged the results. 00:11:15.460 --> 00:11:17.500 Okay. So he looked at them in isolation. Right. 00:11:17.679 --> 00:11:19.879 Edward said, sure, if you look at them individually, 00:11:20.000 --> 00:11:22.879 the overlap is huge. But if you look at the correlation 00:11:22.879 --> 00:11:26.960 between markers, the pattern, you get a completely 00:11:26.960 --> 00:11:30.240 different answer. I need an analogy here. How 00:11:30.240 --> 00:11:32.500 does looking at the pattern change the result? 00:11:33.120 --> 00:11:35.419 Think of it like a medical diagnosis. If you 00:11:35.419 --> 00:11:37.860 just look at the symptom fever, that doesn't 00:11:37.860 --> 00:11:40.700 tell you much. The flu causes fever. An infection 00:11:40.700 --> 00:11:42.759 causes fever. Heat stroke causes fever. Lots 00:11:42.759 --> 00:11:44.960 of overlap. Right. If you look at cough, same 00:11:44.960 --> 00:11:47.059 thing. You can't categorize the patient based 00:11:47.059 --> 00:11:49.759 on one symptom. But if you look at fever plus 00:11:49.759 --> 00:11:53.279 a specific rash plus a swollen lymph node. Exactly. 00:11:53.379 --> 00:11:55.360 The pattern of symptoms gives you a specific 00:11:55.360 --> 00:11:58.919 diagnosis. Edwards argued that Lewontin was looking 00:11:58.919 --> 00:12:01.120 at the symptoms individually and saying, these 00:12:01.120 --> 00:12:03.419 don't tell us anything. Edwards showed that if 00:12:03.419 --> 00:12:05.019 you look at enough genetic markers simultaneously, 00:12:05.200 --> 00:12:07.779 if you look at the pattern, the probability of 00:12:07.779 --> 00:12:09.759 correctly classifying someone into a racial group 00:12:09.759 --> 00:12:13.179 approaches 100%. So Edwards is saying, I can 00:12:13.179 --> 00:12:15.539 take a DNA sample and tell you if your ancestors 00:12:15.539 --> 00:12:18.200 came from Europe or Africa, therefore Lewontin 00:12:18.200 --> 00:12:20.919 is wrong. He's saying that the existence of these 00:12:20.919 --> 00:12:23.500 statistical clusters means biological race is 00:12:23.500 --> 00:12:26.220 a valid concept and that denying it based on 00:12:26.220 --> 00:12:29.500 the 85 % figure is a fallacy. He's saying the 00:12:29.500 --> 00:12:32.000 buckets do exist if you just use enough data 00:12:32.000 --> 00:12:34.840 to find them. This feels like a paradox. How 00:12:34.840 --> 00:12:38.399 can we be 85 % identical within our groups, which 00:12:38.399 --> 00:12:41.120 Lewontin proved, but also easily sorted into 00:12:41.120 --> 00:12:43.700 clusters, which Edwards proved? Those two things 00:12:43.700 --> 00:12:46.059 sound mutually exclusive. They definitely feel 00:12:46.059 --> 00:12:48.240 that way, but they aren't. And this is where 00:12:48.240 --> 00:12:50.179 the response from other scientists comes in to 00:12:50.179 --> 00:12:52.779 save us. People like Jonathan Marks and Rasmus 00:12:52.779 --> 00:12:55.299 Grenfeldt -Winder. Winder calls this two sides 00:12:55.299 --> 00:12:58.159 of the coin. Explain that. How can both be mathematically 00:12:58.159 --> 00:13:00.750 true? You have to realize they're answering two 00:13:00.750 --> 00:13:02.830 fundamentally different mathematical questions. 00:13:03.250 --> 00:13:05.990 Question A is variance partitioning. That's Lawontin. 00:13:06.210 --> 00:13:08.429 He's asking how much of the total diversity of 00:13:08.429 --> 00:13:11.230 the species is explained by race. And the answer 00:13:11.230 --> 00:13:13.610 is very little. Race is a terrible explanation 00:13:13.610 --> 00:13:15.830 for human diversity. And question B. Question 00:13:15.830 --> 00:13:18.090 B is clustering analysis. That's Edwards. He's 00:13:18.090 --> 00:13:20.029 asking, can I find a pattern that allows me to 00:13:20.029 --> 00:13:22.289 sort these people into groups? And the answer 00:13:22.289 --> 00:13:25.330 is yes. So both are true simultaneously. We are 00:13:25.330 --> 00:13:28.149 mostly the same, but we have just enough distinct 00:13:28.149 --> 00:13:30.450 patterns to be sorted if you look closely enough. 00:13:30.649 --> 00:13:34.250 Exactly. And a 2007 study by David Witherspoon 00:13:34.250 --> 00:13:37.490 actually confirmed this. They used distinct populations 00:13:37.490 --> 00:13:40.470 and hundreds of genetic loci and found that yes, 00:13:40.710 --> 00:13:42.970 Lewontin's observation about the distribution 00:13:42.970 --> 00:13:45.929 of differences holds up, and yes, Edwards' ability 00:13:45.929 --> 00:13:48.879 to classify people is also real. But, and here's 00:13:48.879 --> 00:13:51.460 the really big but, does the ability to sort 00:13:51.460 --> 00:13:54.240 people mean that the race label is actually important? 00:13:54.659 --> 00:13:57.720 That is the crux of the modern debate. Biologist 00:13:57.720 --> 00:13:59.879 Charles Roseman points out that Lewontin wasn't 00:13:59.879 --> 00:14:01.559 claiming you couldn't sort people into piles. 00:14:01.759 --> 00:14:04.659 He was focused on taxonomic significance. Which 00:14:04.659 --> 00:14:07.159 means? It means, does this grouping actually 00:14:07.159 --> 00:14:09.779 explain the biology? Look at it this way. If 00:14:09.779 --> 00:14:11.620 I hand you a book and say, this is a red book, 00:14:11.879 --> 00:14:13.860 does that tell you anything about the story inside? 00:14:14.320 --> 00:14:16.120 Not at all. It just helps me put it on a shelf. 00:14:16.200 --> 00:14:18.740 It tells me about the cover. Exactly. Edwards 00:14:18.740 --> 00:14:21.580 proved you can shelve the books by color. Lewontin 00:14:21.580 --> 00:14:23.480 proved that the color of the cover tells you 00:14:23.480 --> 00:14:26.100 almost nothing about the content. The story, 00:14:26.179 --> 00:14:28.480 the genetic variation that determines your health, 00:14:28.580 --> 00:14:32.059 your personality, your potential? is overwhelmingly 00:14:32.059 --> 00:14:34.360 individual. That's a really powerful distinction. 00:14:34.559 --> 00:14:36.519 Just because you can categorize something doesn't 00:14:36.519 --> 00:14:38.340 mean the category is the most important thing 00:14:38.340 --> 00:14:40.940 about it. Correct. And this is exactly where 00:14:40.940 --> 00:14:44.139 the modern context gets messy, because this nuance, 00:14:44.340 --> 00:14:46.679 this idea that, yes, we can cluster, but no, 00:14:46.720 --> 00:14:49.399 it doesn't explain much, is often totally lost 00:14:49.399 --> 00:14:51.919 in today's culture wars. Right. We saw this in 00:14:51.919 --> 00:14:54.659 the sources about social media. The term Lewontin's 00:14:54.659 --> 00:14:57.500 fallacy has apparently seen a huge resurgence 00:14:57.500 --> 00:15:01.370 online 50 years later. It has. Geneticists Jedediah 00:15:01.370 --> 00:15:03.509 Carlson and Kelly Harris actually did some research 00:15:03.509 --> 00:15:06.309 on this. They noted that far -right and white 00:15:06.309 --> 00:15:09.149 nationalist accounts on Twitter frequently cite 00:15:09.149 --> 00:15:11.529 Lewontin's fallacy as a rhetorical device. And 00:15:11.529 --> 00:15:13.370 we're not taking sides here just reporting what 00:15:13.370 --> 00:15:15.669 the researchers found. These groups use Edwards' 00:15:15.850 --> 00:15:18.330 critique to dismiss scientific arguments against 00:15:18.330 --> 00:15:22.029 biological race. Exactly. They're grabbing Edwards' 00:15:22.269 --> 00:15:24.549 clustering argument and using it to say, see? 00:15:25.550 --> 00:15:28.710 Science says race is real and deep and fundamental. 00:15:28.950 --> 00:15:30.490 They're ignoring the other side of the coin. 00:15:30.669 --> 00:15:33.110 The fact that the cluster only accounts for a 00:15:33.110 --> 00:15:37.490 tiny 6 .3 % of our actual biology. They are mistaking 00:15:37.490 --> 00:15:40.190 a sorting mechanism for a deep biological divide. 00:15:40.840 --> 00:15:42.879 They're pretending the book cover is the story. 00:15:43.039 --> 00:15:45.620 It's fascinating how a highly specific statistical 00:15:45.620 --> 00:15:49.399 critique from 2003 is being cited by people who, 00:15:49.440 --> 00:15:51.299 let's be honest, probably haven't read the original 00:15:51.299 --> 00:15:53.919 1972 paper or Edwards paper, for that matter. 00:15:53.980 --> 00:15:57.000 It is. But despite all the Internet noise, the 00:15:57.000 --> 00:16:00.139 scientific consensus has largely held firm. Countless 00:16:00.139 --> 00:16:02.559 subsequent studies, even with modern full DNA 00:16:02.559 --> 00:16:05.539 sequencing, have confirmed Lewontin's main finding. 00:16:05.700 --> 00:16:07.980 The vast majority of human difference is individual. 00:16:08.279 --> 00:16:10.509 Right. The Royal Society even published a special 00:16:10.509 --> 00:16:13.450 issue in 2022 celebrating the 50th anniversary 00:16:13.450 --> 00:16:16.389 of Lawontin's paper. It cemented its status as 00:16:16.389 --> 00:16:18.730 a landmark study. The experts generally agree 00:16:18.730 --> 00:16:20.789 that Edwards had a valid mathematical point about 00:16:20.789 --> 00:16:23.029 clustering, but Lawontin was right about the 00:16:23.029 --> 00:16:25.710 big picture. We are an overwhelmingly mixed and 00:16:25.710 --> 00:16:28.450 shallowly divided species. It really changes 00:16:28.450 --> 00:16:30.570 how you look at the person standing next to you. 00:16:30.649 --> 00:16:33.330 If the science holds up, the stranger isn't actually 00:16:33.330 --> 00:16:35.750 that strange. It should change how we see everyone. 00:16:37.269 --> 00:16:39.570 Lewontin's math forces you to confront a really 00:16:39.570 --> 00:16:42.250 weird reality. If you needed a kidney match or 00:16:42.250 --> 00:16:44.669 a blood transfusion, your neighbor, who looks 00:16:44.669 --> 00:16:47.549 just like you, might be a worse match than someone 00:16:47.549 --> 00:16:49.269 from the complete opposite side of the planet. 00:16:49.409 --> 00:16:51.350 Because the genes that determine those internal 00:16:51.350 --> 00:16:53.590 things aren't the ones that code for skin color 00:16:53.590 --> 00:16:56.210 or eye shape. Right. We obsess over the genes 00:16:56.210 --> 00:17:00.049 we can see. The paint job, so to speak. But the 00:17:00.049 --> 00:17:02.049 engine, the transmission, the wiring, that stuff 00:17:02.049 --> 00:17:05.009 varies wildly from person to person, completely, 00:17:05.210 --> 00:17:07.170 regardless of where they are from. It suggests 00:17:07.170 --> 00:17:09.490 that our social intuition is basically just lying 00:17:09.490 --> 00:17:12.089 to us. We see different faces and our brain immediately 00:17:12.089 --> 00:17:14.890 says different category, but the DNA says rounding 00:17:14.890 --> 00:17:17.589 error. It's a classic case of our eyes tricking 00:17:17.589 --> 00:17:20.769 our brains. We are highly visual creatures, so 00:17:20.769 --> 00:17:24.210 we over -index on visual differences. Laurentin 00:17:24.210 --> 00:17:26.650 used math to strip away that visual bias and 00:17:26.650 --> 00:17:29.740 show us the raw data. Thinking back to that bus 00:17:29.740 --> 00:17:31.900 ride, it's amazing that he did this with a pencil 00:17:31.900 --> 00:17:34.579 and a log table. But I'm mostly struck by the 00:17:34.579 --> 00:17:36.880 philosophical weight of it. He wasn't just doing 00:17:36.880 --> 00:17:39.619 sums. He was dissolving boundaries. He was. And 00:17:39.619 --> 00:17:42.140 I think that's why this paper still matters so 00:17:42.140 --> 00:17:44.819 much today. We live in a world that is obsessed 00:17:44.819 --> 00:17:47.700 with putting people into boxes. It's efficient. 00:17:47.900 --> 00:17:50.880 It makes data processing easy. It makes politics 00:17:50.880 --> 00:17:53.500 easy. Us versus them is the oldest story in the 00:17:53.500 --> 00:17:56.569 book. It's comforting to have a team. But biology 00:17:56.569 --> 00:17:59.690 is telling us that the U .S. is actually incredibly 00:17:59.690 --> 00:18:03.210 diverse and the them is mostly just us with a 00:18:03.210 --> 00:18:05.730 slightly different tan. So here is a provocative 00:18:05.730 --> 00:18:07.529 thought to leave you with as we wrap up this 00:18:07.529 --> 00:18:10.250 deep dive. We spend so much energy, political 00:18:10.250 --> 00:18:13.009 capital and social angst focusing on the differences 00:18:13.009 --> 00:18:15.329 between these global groups. We build walls, 00:18:15.490 --> 00:18:18.470 both real and metaphorical, based on that 6 .3 00:18:18.470 --> 00:18:20.390 percent of difference. Meanwhile, we completely 00:18:20.390 --> 00:18:24.539 ignore the 85 percent. Exactly. If 85 % of what 00:18:24.539 --> 00:18:26.599 makes you biologically unique is found right 00:18:26.599 --> 00:18:29.099 there in your own local community, within your 00:18:29.099 --> 00:18:33.180 own so -called group, why do we continue to prioritize 00:18:33.180 --> 00:18:35.500 the tiny fraction of differences that separate 00:18:35.500 --> 00:18:39.069 us globally? Why is the label on the box more 00:18:39.069 --> 00:18:41.730 important to us than the contents? Science can 00:18:41.730 --> 00:18:43.869 give us the numbers, but it can't answer that 00:18:43.869 --> 00:18:46.369 question for us. That one is on us to figure 00:18:46.369 --> 00:18:48.890 out. A massive thank you to you, the listener, 00:18:49.009 --> 00:18:51.029 for joining us on this deep dive into the math 00:18:51.029 --> 00:18:53.109 of who we really are. It's a pleasure to unpack 00:18:53.109 --> 00:18:55.430 this. Keep questioning the categories. We'll 00:18:55.430 --> 00:18:56.089 see you next time.