WEBVTT 00:00:00.000 --> 00:00:02.690 Welcome back to the deep dive. Today, we are 00:00:02.690 --> 00:00:04.750 starting with a word that, I mean, it defines 00:00:04.750 --> 00:00:07.049 so much of our world, our education system, our 00:00:07.049 --> 00:00:10.169 technology. And right now, especially, the AI 00:00:10.169 --> 00:00:13.250 revolution, that word is intelligence. It's the 00:00:13.250 --> 00:00:15.490 standard we use for everything, measuring capability, 00:00:15.929 --> 00:00:18.850 progress, sometimes even human worth. But what 00:00:18.850 --> 00:00:21.010 if I told you that the very foundation of how 00:00:21.010 --> 00:00:22.710 we've traditionally defined it, you know, as 00:00:22.710 --> 00:00:25.170 this single, quantifiable thing, what if that 00:00:25.170 --> 00:00:27.429 whole idea is fundamentally flawed? I mean, not 00:00:27.429 --> 00:00:29.829 just flawed, but academically rejected and historically 00:00:29.829 --> 00:00:32.630 just a social disaster. disaster. You've brought 00:00:32.630 --> 00:00:34.710 us some really fascinating sources today, all 00:00:34.710 --> 00:00:36.890 diving deep into this powerful critique from 00:00:36.890 --> 00:00:38.929 the evolutionary biologist Stephen Jay Gould, 00:00:39.450 --> 00:00:41.570 especially the arguments he laid out in his book, 00:00:41.909 --> 00:00:44.210 The Mismeasure of Man. And this isn't some minor 00:00:44.210 --> 00:00:46.090 critique, is it? It's a full -on rejection of 00:00:46.090 --> 00:00:48.469 a concept that has haunted us for over a century. 00:00:48.710 --> 00:00:51.049 Let's unpack this. Our mission for this deep 00:00:51.049 --> 00:00:53.509 dive is to look at six specific characteristics 00:00:53.509 --> 00:00:56.630 of intelligence, the core assumptions that Gould 00:00:56.630 --> 00:00:59.049 says we just have to reject, and to understand 00:00:59.049 --> 00:01:01.390 why these flawed ideas are, well, still causing 00:01:01.390 --> 00:01:03.490 so many problems today. What's fascinating here 00:01:03.490 --> 00:01:06.269 is just how ambitious the original project was. 00:01:06.909 --> 00:01:09.230 If you go back to the early 20th century, psychology 00:01:09.230 --> 00:01:14.930 was trying really hard to be seen as a hard science. 00:01:15.230 --> 00:01:17.590 Like physics. Exactly, like physics or chemistry. 00:01:17.829 --> 00:01:19.469 They weren't just looking for theories. They 00:01:19.469 --> 00:01:23.030 wanted objective, quantifiable laws to describe 00:01:23.030 --> 00:01:25.810 how the brain worked, almost like measuring gravity. 00:01:26.269 --> 00:01:28.450 And this wasn't just some academic exercise in 00:01:28.450 --> 00:01:31.400 an ivory tower. No. These scholars were actively 00:01:31.400 --> 00:01:34.040 pushing for public policies. We're talking educational 00:01:34.040 --> 00:01:36.739 tracking, immigration laws, social engineering, 00:01:37.000 --> 00:01:38.980 all based on these numbers they were trying to 00:01:38.980 --> 00:01:41.540 define. So the stakes were incredibly high from 00:01:41.540 --> 00:01:43.879 the very beginning. Immense. And our sources 00:01:43.879 --> 00:01:46.920 really stress that even though careful modern 00:01:46.920 --> 00:01:48.879 study has pretty much rejected the philosophical 00:01:48.879 --> 00:01:52.459 basis of these tests, the shadows of those old 00:01:52.459 --> 00:01:54.959 ideas are definitely still with us. And Gould 00:01:54.959 --> 00:01:56.760 just starts his demolition right at the root 00:01:56.760 --> 00:01:58.819 of the whole thing. He says the first mistake. 00:01:59.040 --> 00:02:01.140 The philosophical one that underpins everything 00:02:01.140 --> 00:02:04.180 was how we started treating intelligence, not 00:02:04.180 --> 00:02:07.120 as a set of skills or processes, but as a singular, 00:02:07.480 --> 00:02:11.360 measurable object. So his first point to reject 00:02:11.360 --> 00:02:15.340 is this idea of intelligence as a single, unitary, 00:02:15.699 --> 00:02:18.240 abstractable entity, this sort of central stuff. 00:02:18.719 --> 00:02:20.879 Right. And he had a specific philosophical term 00:02:20.879 --> 00:02:23.979 for this mistake. Reification. Reification. Yeah. 00:02:23.979 --> 00:02:25.719 And it's key to understanding this whole trap. 00:02:26.039 --> 00:02:28.800 It's when you take an abstract concept. something 00:02:28.800 --> 00:02:31.939 like cleverness or competence. And you wrongly 00:02:31.939 --> 00:02:34.900 turn it into a concrete thing, a hard entity 00:02:34.900 --> 00:02:37.039 that you think exists physically and can be measured, 00:02:37.379 --> 00:02:39.460 like your height or weight. So it's mistaking 00:02:39.460 --> 00:02:42.520 a process for a physical part. Like a mental 00:02:42.520 --> 00:02:44.280 organ. Precisely. It's like trying to define 00:02:44.280 --> 00:02:46.400 courage and saying, OK, we need to find the courage 00:02:46.400 --> 00:02:48.240 muscle in the brain. Right. And if that muscle 00:02:48.240 --> 00:02:51.020 is bigger, you're just more courageous. Gould 00:02:51.020 --> 00:02:52.860 insisted intelligence isn't like that. He said 00:02:52.860 --> 00:02:55.680 it was richly multiple and independent. There's 00:02:55.680 --> 00:02:57.919 no single well that all your abilities draw from. 00:02:58.199 --> 00:03:00.259 OK. So if you believe it's a single thing, the 00:03:00.259 --> 00:03:03.659 logical next step is to assume it has to be located 00:03:03.659 --> 00:03:06.280 somewhere. Exactly. Which brings us to the second 00:03:06.280 --> 00:03:08.419 characteristic we have to reject. And it flows 00:03:08.419 --> 00:03:11.680 right from reunification. the belief that this 00:03:11.680 --> 00:03:14.740 central stuff must have a physical location in 00:03:14.740 --> 00:03:18.039 the brain and The main scientific tool they used 00:03:18.039 --> 00:03:21.000 to justify this was something called factor analysis. 00:03:21.000 --> 00:03:25.719 Okay, it's a complex statistical method but basically 00:03:25.719 --> 00:03:28.159 psychologists like Charles Spearman used it to 00:03:28.159 --> 00:03:32.620 argue that if different mental tasks like memory 00:03:32.620 --> 00:03:35.509 logic patterns If they all seem to correlate, 00:03:35.569 --> 00:03:37.689 then there must be one single thing causing it. 00:03:37.810 --> 00:03:40.210 Right. A single underlying general ability. Spearman 00:03:40.210 --> 00:03:43.270 called it G for general intelligence. And G was 00:03:43.270 --> 00:03:45.349 held up as the scientific proof that there really 00:03:45.349 --> 00:03:48.110 was this unitary thing in the head. So factor 00:03:48.110 --> 00:03:50.210 analysis didn't actually find the thing. It was 00:03:50.210 --> 00:03:52.710 a statistical method they used to create a score. 00:03:52.849 --> 00:03:54.889 And then they made the philosophical leap to 00:03:54.889 --> 00:03:57.750 say that score represented a real physical thing. 00:03:57.789 --> 00:04:00.229 That's the crucial distinction. Yes. OK. So now 00:04:00.229 --> 00:04:02.889 you believe this central stuff exists. It's somewhere 00:04:02.889 --> 00:04:06.520 in the brain. The next impulse, obviously, is 00:04:06.520 --> 00:04:09.659 to measure it, put a number on it. And this gives 00:04:09.659 --> 00:04:13.620 us the third rejected characteristic, the belief 00:04:13.620 --> 00:04:16.720 that all that complexity can be boiled down to 00:04:16.720 --> 00:04:20.040 one number for each individual. Which we all 00:04:20.040 --> 00:04:22.339 know is the IQ score. Right. And here's where 00:04:22.339 --> 00:04:24.699 the story gets really interesting. Because the 00:04:24.699 --> 00:04:27.560 man who invented the first scale, Alfred Binet, 00:04:27.850 --> 00:04:30.170 was just trying to solve a practical problem. 00:04:30.389 --> 00:04:32.110 This was in France, right, with school children. 00:04:32.329 --> 00:04:34.769 Exactly. He wanted a way to identify kids who 00:04:34.769 --> 00:04:37.189 needed extra help in school, specifically to 00:04:37.189 --> 00:04:39.790 help them catch up. Our source material is very 00:04:39.790 --> 00:04:42.589 clear on this. Benet himself explicitly denied 00:04:42.589 --> 00:04:44.509 that his test could measure the intelligence 00:04:44.509 --> 00:04:46.930 as a single quantity. So he didn't even believe 00:04:46.930 --> 00:04:48.990 in his own test and the way it was later used. 00:04:49.149 --> 00:04:51.709 He saw it as a diagnostic tool. a rough guide 00:04:51.709 --> 00:04:54.250 for intervention. He said it was not a lifetime 00:04:54.250 --> 00:04:57.310 brand of a person's worth. For him, a low score 00:04:57.310 --> 00:04:59.930 was a starting point for help, not a final judgment. 00:05:00.230 --> 00:05:03.170 Wow. So Benet himself warned everyone against 00:05:03.170 --> 00:05:06.730 the exact misuse that became, well, rampant almost 00:05:06.730 --> 00:05:09.370 immediately. Especially in America. The next 00:05:09.370 --> 00:05:11.430 generation of psychologists just stripped all 00:05:11.430 --> 00:05:14.930 that nuance away. And that transition from a 00:05:14.930 --> 00:05:17.790 tool to a label. That leads us straight to the 00:05:17.790 --> 00:05:20.230 fourth characteristic Gould rejects. And this 00:05:20.230 --> 00:05:22.730 one is maybe the most socially devastating. The 00:05:22.730 --> 00:05:26.689 use of that single number to rank people in a 00:05:26.689 --> 00:05:29.550 single series of worthiness. This is where the 00:05:29.550 --> 00:05:32.490 academic error meets real -world cruelty. Once 00:05:32.490 --> 00:05:34.689 you accept that a single number defines a person, 00:05:35.189 --> 00:05:38.569 you are forced into this massive, brutal simplification 00:05:38.569 --> 00:05:41.069 of reality. A dichotomization, as the sources 00:05:41.069 --> 00:05:43.850 call it. Yes. Suddenly, complex individuals, 00:05:43.949 --> 00:05:46.069 each with their own unique profile of strengths 00:05:46.069 --> 00:05:48.370 and weaknesses, are just flattened into smart 00:05:48.370 --> 00:05:51.089 and stupid, or normal versus feeble -minded. 00:05:51.250 --> 00:05:53.589 And the consequences were huge. Immediate and 00:05:53.589 --> 00:05:56.370 massive. This ranking system was seized by policymakers. 00:05:56.509 --> 00:05:58.610 It gave a kind of pseudo -scientific justification 00:05:58.610 --> 00:06:00.870 for their own prejudices. The score became a 00:06:00.870 --> 00:06:03.329 weapon. To justify the status quo. Absolutely. 00:06:03.730 --> 00:06:05.949 The sources we looked at are explicit. The ranking 00:06:05.949 --> 00:06:07.990 was invoked constantly to argue that oppressed 00:06:07.990 --> 00:06:10.709 and disadvantaged groups, races, classes, or 00:06:10.709 --> 00:06:13.300 sexes are innately inferior and deserve their 00:06:13.300 --> 00:06:15.980 status. It was the intellectual engine of the 00:06:15.980 --> 00:06:18.220 eugenics movement. That immigration restrictions 00:06:18.220 --> 00:06:21.879 segregation all of it was propped up by this 00:06:21.879 --> 00:06:24.120 idea of a single ranked number. Which provides 00:06:24.120 --> 00:06:27.279 a chillingly perfect bridge to the next point 00:06:27.279 --> 00:06:30.120 because for those rankings to mean anything for 00:06:30.120 --> 00:06:32.519 social policy they had to be seen as permanent. 00:06:32.680 --> 00:06:35.759 So the fifth rejected characteristic is the belief 00:06:35.759 --> 00:06:39.920 that intelligence is innate, inherited, and almost 00:06:39.920 --> 00:06:42.639 entirely fixed by your genetics. And the conviction 00:06:42.639 --> 00:06:45.699 they had is just, it's astounding to look back 00:06:45.699 --> 00:06:48.079 on. You have figures like H .H. Goddard who brought 00:06:48.079 --> 00:06:51.240 the Binet scale to the U .S. He literally argued 00:06:51.240 --> 00:06:53.720 that intelligence was a unitary mental process, 00:06:54.199 --> 00:06:56.220 conditioned by a nervous mechanism, which is 00:06:56.220 --> 00:06:58.819 inborn. He even suggested it might be a single 00:06:58.819 --> 00:07:01.019 gene. A single gene for all of intelligence. 00:07:01.040 --> 00:07:03.990 A single gene. undoubtedly recessive to normal 00:07:03.990 --> 00:07:06.629 intelligence. That's a direct quote. And that 00:07:06.629 --> 00:07:10.250 thinking was everywhere. Lewis Turman, who developed 00:07:10.250 --> 00:07:13.589 the Stanford -Binet IQ test, that became the 00:07:13.589 --> 00:07:16.550 standard, he argued for the relatively greater 00:07:16.550 --> 00:07:19.170 importance of endowment over training. He saw 00:07:19.170 --> 00:07:21.170 disparities between groups and just concluded 00:07:21.170 --> 00:07:24.129 it must be innate genetic inferiority. It wasn't 00:07:24.129 --> 00:07:26.589 looking at environment or opportunity or history. 00:07:26.850 --> 00:07:29.149 Not at all. The test score was proof enough for 00:07:29.149 --> 00:07:31.290 him. And you can't forget someone like Cyril 00:07:31.290 --> 00:07:33.709 Byrd, who for decades was a giant in the field 00:07:33.709 --> 00:07:36.810 arguing for what he called innate general cognitive 00:07:36.810 --> 00:07:39.209 ability. Even though his data was later discredited. 00:07:39.250 --> 00:07:41.870 Yes, but the damage was done. The core belief 00:07:41.870 --> 00:07:46.689 that this single measurable thing is mostly genetic, 00:07:47.730 --> 00:07:50.269 that became the dominant and I'd say toxic idea 00:07:50.269 --> 00:07:53.069 in the field for decades. OK, so if it's innate 00:07:53.069 --> 00:07:55.410 and it's genetic, then it has to be unchangeable. 00:07:55.519 --> 00:07:57.959 inescapable, which is the sixth and final point 00:07:57.959 --> 00:08:00.620 we have to reject. The belief that intelligence 00:08:00.620 --> 00:08:02.959 is a fixed quantity that can't be increased, 00:08:03.540 --> 00:08:06.120 that it's but little affected by any later influences 00:08:06.120 --> 00:08:09.220 like education or environment. Right. And the 00:08:09.220 --> 00:08:11.779 consequence of that belief is just total fatalism. 00:08:12.240 --> 00:08:14.660 Why bother trying to improve education for certain 00:08:14.660 --> 00:08:17.620 groups? Why invest in social programs if their 00:08:17.620 --> 00:08:20.709 rank is fixed at birth? This is where Gould really 00:08:20.709 --> 00:08:23.230 brings the hammer down, isn't it? Oh, absolutely. 00:08:23.470 --> 00:08:26.990 He calls this belief in fixity a form of brutal 00:08:26.990 --> 00:08:30.689 pessimism that was founded on, well, on nothing. 00:08:31.269 --> 00:08:33.029 The research didn't support it, but the social 00:08:33.029 --> 00:08:36.110 utility of the idea was immense for those in 00:08:36.110 --> 00:08:38.669 power. It let them off the hook. It did. And 00:08:38.669 --> 00:08:40.649 Gould spends a lot of time attacking what he 00:08:40.649 --> 00:08:44.710 calls the false equation of heritable with permanent 00:08:44.710 --> 00:08:47.889 or unchangeable. That's a key distinction. It's 00:08:47.889 --> 00:08:50.009 vital. We know lots of traits are heritable. 00:08:50.450 --> 00:08:52.570 Height, for example. But that doesn't mean it's 00:08:52.570 --> 00:08:54.950 unchangeable. Nutrition can dramatically change 00:08:54.950 --> 00:08:57.529 the outcome. Heritability is a statistical measure 00:08:57.529 --> 00:09:00.250 about variation in a population, not a statement 00:09:00.250 --> 00:09:02.409 about an individual's fixed destiny. But that 00:09:02.409 --> 00:09:04.909 idea of a fixed, unchangeable number was necessary 00:09:04.909 --> 00:09:07.190 for the whole hereditarian theory to hold together. 00:09:07.269 --> 00:09:09.690 Through the linchpin. And Gould just completely 00:09:09.690 --> 00:09:13.429 dismantled it. So, okay. What does this all mean 00:09:13.429 --> 00:09:17.110 for us today? If we accept Gould's critique... 00:09:17.580 --> 00:09:20.740 that this entire historical concept of intelligence 00:09:20.740 --> 00:09:24.860 as a single innate fixed thing, if that's all 00:09:24.860 --> 00:09:27.000 academically unsound. So what does all this mean? 00:09:27.580 --> 00:09:29.480 The author of one of our main sources concludes 00:09:29.480 --> 00:09:31.679 that because that whole framework has been rejected 00:09:31.679 --> 00:09:34.379 by careful study, the word intelligence itself 00:09:34.379 --> 00:09:36.559 is almost becoming meaningless, at least when 00:09:36.559 --> 00:09:39.120 it's used in that old single -number way. Which 00:09:39.120 --> 00:09:41.649 brings us right into this moment. right into 00:09:41.649 --> 00:09:44.090 the age of AI. We have these large language models 00:09:44.090 --> 00:09:46.289 that are, I mean, they're incredibly effective. 00:09:46.370 --> 00:09:48.870 They can pass the bar exam. They can write code. 00:09:49.149 --> 00:09:51.049 They can do things that would score incredibly 00:09:51.049 --> 00:09:53.250 high on any of those old 20th century tests. 00:09:53.970 --> 00:09:56.850 And yet, most people hesitate to call them intelligent 00:09:56.850 --> 00:09:58.850 in the same way a person is. Right, you don't 00:09:58.850 --> 00:10:01.509 hear people describe an AI as conscious or wise 00:10:01.509 --> 00:10:04.419 or having common sense. Exactly. The AI seems 00:10:04.419 --> 00:10:06.759 to perform all these complex intellectual tasks 00:10:06.759 --> 00:10:09.620 without having that mysterious, singular, central 00:10:09.620 --> 00:10:11.759 stuff that we were trying so hard to find in 00:10:11.759 --> 00:10:14.360 the human brain. The AI kind of forces our hand. 00:10:14.700 --> 00:10:17.179 It makes us realize that maybe that concept was 00:10:17.179 --> 00:10:19.460 the problem all along. Maybe it was just an error 00:10:19.460 --> 00:10:22.720 of reification. A human invention. That really 00:10:22.720 --> 00:10:25.179 reframes the whole conversation. The old guard 00:10:25.179 --> 00:10:27.740 wanted a single number for worth. Now we have 00:10:27.740 --> 00:10:30.419 machines that can ace their tests, and we still 00:10:30.419 --> 00:10:32.159 instinctively feel like something is missing. 00:10:32.460 --> 00:10:34.299 And that connects this whole critique right back 00:10:34.299 --> 00:10:35.980 to learning and teaching, which I know is a key 00:10:35.980 --> 00:10:39.440 interest for you, our listener. If we stop chasing 00:10:39.440 --> 00:10:43.700 one single measurable quantity, what should education 00:10:43.700 --> 00:10:46.720 be aiming for? Yeah. puts it beautifully, it 00:10:46.720 --> 00:10:48.419 says that whatever students leave school with 00:10:48.419 --> 00:10:51.559 is far more diverse than anything we can conceive 00:10:51.559 --> 00:10:54.480 as general intelligence. It's not about maximizing 00:10:54.480 --> 00:10:58.200 a G score. It's about cultivating that rich multiplicity 00:10:58.200 --> 00:11:00.779 of skills. Creative thinking, emotional intelligence, 00:11:01.100 --> 00:11:04.039 collaboration, technical competence, all the 00:11:04.039 --> 00:11:06.320 things a single number just catastrophically 00:11:06.320 --> 00:11:08.940 fails to capture. Gold's critique frees us from 00:11:08.940 --> 00:11:11.759 that brutal pessimism. It says potential is fluid, 00:11:11.899 --> 00:11:14.460 not fixed. It's a profound takeaway. The whole 00:11:14.460 --> 00:11:18.360 historical concept of intelligence was a mistake 00:11:18.360 --> 00:11:21.559 from the start, rooted in this desire for a simple, 00:11:21.559 --> 00:11:24.399 scientific -sounding number. And its application 00:11:24.399 --> 00:11:27.419 led to social catastrophe, all based on a number 00:11:27.419 --> 00:11:30.000 that its own inventor warned us not to misuse. 00:11:30.580 --> 00:11:32.860 We have to reckon with that history to find a 00:11:32.860 --> 00:11:35.840 better, more honest language to talk about human 00:11:35.840 --> 00:11:38.809 potential. which leaves us with a final provocative 00:11:38.809 --> 00:11:41.389 thought for you to take away. The flaw of the 00:11:41.389 --> 00:11:44.629 early 20th century was trying to define intelligence 00:11:44.629 --> 00:11:47.370 like a law of physics. Now that we have these 00:11:47.370 --> 00:11:49.529 AI systems that show incredible capabilities 00:11:49.529 --> 00:11:53.210 without fitting that old definition, maybe the 00:11:53.210 --> 00:11:55.250 whole mission was flawed from the start. Should 00:11:55.250 --> 00:11:57.350 we just stop trying to measure intelligence altogether 00:11:57.350 --> 00:12:00.090 and focus only on measuring observable, specific, 00:12:00.210 --> 00:12:02.429 and multiple capabilities? Abandon the noun and 00:12:02.429 --> 00:12:04.730 just focus on the verbs. Exactly. Something to 00:12:04.730 --> 00:12:06.549 think about is you define competence in your 00:12:06.549 --> 00:12:06.950 own life.