Richard Haier: IQ Tests, Human Intelligence, and Group Differences

00:00:00.000 | Let me ask you to this question,

00:00:02.200 | whether it's bell curve or any research on race differences,

00:00:06.340 | can that be used to increase the amount of racism

00:00:12.440 | in the world, can that be used to increase

00:00:14.880 | the amount of hate in the world?

00:00:16.920 | - My sense is there is such enormous reservoirs

00:00:21.920 | of hate and racism that have nothing to do

00:00:27.920 | with scientific knowledge of the data

00:00:31.480 | that speak against that.

00:00:34.080 | That, no, I don't wanna give racist groups a veto power

00:00:39.080 | over what scientists study.

00:00:42.020 | - The following is a conversation with Richard Heyer

00:00:46.960 | on the science of human intelligence.

00:00:49.320 | This is a highly controversial topic,

00:00:51.720 | but a critically important one

00:00:52.960 | for understanding the human mind.

00:00:55.000 | I hope you will join me in not shying away

00:00:57.680 | from difficult topics like this,

00:00:59.980 | and instead, let us try to navigate it

00:01:03.700 | with empathy, rigor, and grace.

00:01:06.460 | If you're watching this on video now,

00:01:08.800 | I should mention that I'm recording this introduction

00:01:11.500 | in an undisclosed location somewhere in the world.

00:01:14.580 | I'm safe and happy, and life is beautiful.

00:01:17.320 | This is the Alex Friedman Podcast.

00:01:20.400 | To support it, please check out our sponsors

00:01:22.360 | in the description.

00:01:23.620 | And now, dear friends, here's Richard Heyer.

00:01:27.960 | What are the measures of human intelligence,

00:01:29.920 | and how do we measure it?

00:01:31.600 | - Everybody has an idea of what they mean by intelligence.

00:01:35.880 | In the vernacular, what I mean by intelligence

00:01:40.080 | is just being smart, how well you reason,

00:01:42.600 | how well you figure things out,

00:01:45.200 | what you do when you don't know what to do.

00:01:48.740 | Those are just kinda everyday common sense definitions

00:01:53.740 | of how people use the word intelligence.

00:01:56.680 | If you wanna do research on intelligence,

00:01:59.560 | measuring something that you can study scientifically

00:02:03.400 | is a little trickier.

00:02:05.060 | And what almost all researchers who study intelligence use

00:02:10.960 | is the concept called the G factor, general intelligence.

00:02:16.740 | And that is what is common.

00:02:19.600 | That is a mental ability that is common

00:02:22.640 | to virtually all tests of mental abilities.

00:02:26.560 | - What's the origin of the term G factor, by the way?

00:02:28.840 | It's such a funny word for such a fundamental human thing.

00:02:32.120 | - The general factor really started with Charles Spearman.

00:02:35.620 | And he noticed, this is like, boy, more than 100 years ago.

00:02:41.600 | He noticed that when you tested people with different tests,

00:02:46.600 | all the tests were correlated positively.

00:02:53.120 | And so he was looking at student exams and things.

00:02:57.360 | And he invented the correlation coefficient, essentially.

00:03:01.600 | And when he used it to look at student performance

00:03:06.360 | on various topics, he found all the scores

00:03:09.920 | were correlated with each other

00:03:11.920 | and they were all positive correlations.

00:03:14.480 | So he inferred from this that there must be

00:03:17.540 | some common factor that was irrespective

00:03:21.180 | of the content of the test.

00:03:23.320 | - And positive correlation means if you do well

00:03:27.000 | on the first test, you're likely to do well

00:03:29.440 | on the second test.

00:03:30.960 | And presumably that holds for tests across even disciplines.

00:03:35.800 | So not within subject, but across subjects.

00:03:39.320 | So that's where the general comes in.

00:03:42.360 | Something about general intelligence.

00:03:45.000 | So when you were talking about measuring intelligence

00:03:46.920 | and trying to figure out something difficult

00:03:50.080 | about this world and how to solve the puzzles of this world,

00:03:53.240 | that means generally speaking, not some specific test,

00:03:56.600 | but across all tests.

00:03:58.120 | - Absolutely right.

00:03:59.300 | And people get hung up on this because they say,

00:04:03.320 | well, what about the ability to do X?

00:04:06.520 | Isn't that independent?

00:04:08.920 | And they said, I know somebody who's very good at this,

00:04:11.960 | but not so good at this, this other thing.

00:04:15.320 | And so there are a lot of examples like that,

00:04:17.440 | but it's a general tendency.

00:04:19.940 | So exceptions really don't disprove,

00:04:23.520 | you know, your everyday experience is not the same

00:04:27.640 | as what the data actually show.

00:04:30.740 | And your everyday experience, when you say,

00:04:32.720 | oh, I know someone who's good at X, but not so good at Y,

00:04:36.560 | that doesn't contradict the statement of about,

00:04:39.120 | he's not so good, but he's not the opposite.

00:04:43.360 | It's not a negative correlation.

00:04:46.520 | - Okay, so we're not, our anecdotal data,

00:04:49.880 | I know a guy who's really good at solving

00:04:53.720 | some kind of visual thing.

00:04:55.840 | That's not sufficient for us to understand actually

00:04:59.400 | the depths of that person's intelligence.

00:05:01.400 | So how this idea of G factor,

00:05:05.060 | how much evidence is there?

00:05:09.360 | How strong, you know, given across the decades

00:05:13.040 | that this idea has been around,

00:05:14.800 | how much has it been held up that there's a universal

00:05:18.080 | sort of horsepower of intelligence

00:05:22.500 | that's underneath all of it?

00:05:24.000 | All the different tests we do to try to get to this thing

00:05:27.320 | in the depths of the human mind,

00:05:30.400 | that's a universal stable measure

00:05:32.960 | of a person's intelligence.

00:05:34.840 | - You used a couple of words in there, stable and--

00:05:38.440 | - Are we gonna have to be precise with words?

00:05:40.600 | I was hoping we can get away with being poetic.

00:05:42.680 | - We can.

00:05:43.520 | There's a lot about research in general,

00:05:46.240 | not just intelligence research, that is poetic.

00:05:49.560 | Science has a poetic aspect to it,

00:05:52.560 | and good scientists are very intuitive.

00:05:55.800 | They're not just, hey, these are the numbers.

00:05:59.200 | You have to kind of step back and see the big picture.

00:06:02.080 | When it comes to intelligence research,

00:06:05.920 | you asked how well has this general concept held up?

00:06:09.920 | And I think I can say,

00:06:12.280 | without fear of being empirically contradicted,

00:06:16.280 | that it is the most replicated finding in all of psychology.

00:06:21.120 | Now, some cynics may say, well, big deal, psychology.

00:06:23.560 | We all know there's a replication crisis in psychology,

00:06:26.360 | and a lot of this stuff doesn't replicate.

00:06:28.480 | That's all true.

00:06:29.720 | There is no replication crisis when it comes to studying

00:06:33.600 | the existence of this general factor.

00:06:36.820 | Let me tell you some things about it.

00:06:38.960 | It looks like it's universal,

00:06:41.520 | that you find it in all cultures.

00:06:44.680 | The way you find it, step back one step,

00:06:47.760 | the way you find it is to give a battery of mental tests.

00:06:51.760 | What battery?

00:06:52.760 | You choose.

00:06:53.800 | Take a battery of any mental tests you want,

00:06:57.080 | give it to a large number of diverse people,

00:07:00.580 | and you will be able to extract statistically

00:07:05.720 | the commonality among all those tests.

00:07:09.020 | It's done by a technique called factor analysis.

00:07:12.460 | People think that this may be a statistical artifact

00:07:17.460 | of some kind.

00:07:18.500 | It is not a statistical artifact.

00:07:21.340 | - What is factor analysis?

00:07:22.700 | - Factor analysis is a way of looking at a big set of data

00:07:26.140 | and look at the correlation among the different test scores,

00:07:29.880 | and then find empirically the clusters of scores

00:07:33.840 | that go together.

00:07:35.620 | And there are different factors.

00:07:37.340 | So if you have a bunch of mental tests,

00:07:39.380 | there may be a verbal factor,

00:07:41.140 | there may be a numerical factor,

00:07:43.340 | there may be a visual spatial factor,

00:07:45.900 | but those factors have variants in common with each other.

00:07:50.220 | And that is the common,

00:07:51.980 | that's what's common among all the tests,

00:07:55.360 | and that's what gets labeled the G factor.

00:07:58.100 | So if you give a diverse battery of mental tests

00:08:01.460 | and you extract a G factor from it,

00:08:04.740 | that factor usually accounts for around half

00:08:07.340 | of the variances, the single biggest factor,

00:08:10.400 | but it's not the only factor,

00:08:12.860 | but it is the most reliable, it is the most stable,

00:08:17.200 | and it seems to be very much influenced by genetics.

00:08:22.200 | It's very hard to change the G factor

00:08:26.660 | with training or drugs or anything else.

00:08:32.900 | We don't know how to increase the G factor.

00:08:34.980 | - Okay, you said a lot of really interesting things there.

00:08:36.920 | So first, just to get people used to it

00:08:40.940 | in case they're not familiar with this idea,

00:08:43.260 | G factor is what we mean.

00:08:45.820 | So often there's this term used, IQ,

00:08:49.100 | which is the way IQ is used,

00:08:53.980 | they really mean G factor in regular conversation.

00:08:57.540 | 'Cause what we mean by IQ, we mean intelligence,

00:09:02.640 | and what we mean by intelligence,

00:09:04.480 | we mean general intelligence,

00:09:05.820 | and general intelligence in the human mind

00:09:08.840 | from a psychology, from a serious,

00:09:10.840 | rigorous scientific perspective actually means G factor.

00:09:13.900 | So G factor equals intelligence,

00:09:15.740 | just in this conversation to define terms.

00:09:18.440 | Okay, so there's this stable thing called G factor.

00:09:22.180 | You said it, now, factor, you said factor many times,

00:09:28.220 | means a measure that potentially could be reduced

00:09:33.220 | to a single number across the different factors

00:09:35.840 | you mentioned, and what you said,

00:09:40.360 | it accounts for half, half-ish.

00:09:43.600 | Accounts for half-ish of what?

00:09:46.640 | Of variance across the different set of tests.

00:09:53.760 | So if you do for some reason well on some set of tests,

00:09:58.560 | what does that mean?

00:10:01.080 | So that means there's some unique capabilities

00:10:03.340 | outside of the G factor that might account for that,

00:10:05.940 | and what are those?

00:10:07.500 | What else is there besides the raw horsepower,

00:10:10.420 | the engine inside your mind that generates intelligence?

00:10:13.340 | - There are test-taking skills,

00:10:16.380 | there are specific abilities.

00:10:20.880 | Someone might be particularly good at mathematical things,

00:10:25.880 | mathematical concepts, even simple arithmetic.

00:10:32.140 | Some people are much better than others.

00:10:34.320 | You might know people who can,

00:10:36.460 | and short-term memory is another component of this.

00:10:41.460 | Short-term memory is one of the cognitive processes

00:10:46.900 | that's most highly correlated with the G factor.

00:10:51.240 | So--

00:10:54.020 | - So all those things like memory,

00:10:56.260 | test-taking skills account for variability

00:10:59.920 | across the test performances.

00:11:02.220 | But you say you can run, but you can't hide

00:11:05.520 | from the thing that God gave you, the genetics.

00:11:09.920 | So that G factor, science says that G factor's there.

00:11:15.120 | Each one of us have--

00:11:16.900 | - Each one of us has a G factor.

00:11:19.380 | - Oh boy.

00:11:20.220 | Some have more than others.

00:11:21.420 | - I'm getting uncomfortable already.

00:11:22.820 | - Well, IQ is a score.

00:11:25.080 | An IQ score is a very good estimate of the G factor.

00:11:32.260 | You can't measure G directly, there's no direct measure.

00:11:36.100 | You estimate it from these statistical techniques.

00:11:39.880 | But an IQ score is a good estimate, why?

00:11:43.080 | Because a standard IQ test is a battery

00:11:46.420 | of different mental abilities.

00:11:48.660 | You combine it into one score,

00:11:51.380 | and that score is highly correlated with the G factor,

00:11:55.700 | even if you get better scores on some subtests than others.

00:12:00.140 | Because again, it's what's common

00:12:02.300 | to all these mental abilities.

00:12:04.300 | - So a good IQ test, and I'll ask you about that,

00:12:08.180 | but a good IQ test tries to compress down

00:12:12.620 | that battery of tests, like tries to get a nice battery,

00:12:16.180 | a nice selection of variable tests into one test.

00:12:21.180 | And so in that way, it sneaks up to the G factor.

00:12:24.180 | And that's another interesting thing about G factor.

00:12:27.460 | Now you give, first of all, you have a great book

00:12:32.380 | on the neuroscience of intelligence.

00:12:34.180 | You have a great course, which is one I first learned.

00:12:37.500 | You're a great teacher, let me just say.

00:12:39.820 | - Thank you.

00:12:40.660 | - Your course at the teaching company,

00:12:44.180 | I hope I'm saying that correctly.

00:12:45.820 | - The Intelligent Brain.

00:12:47.180 | - The Intelligent Brain is when I first heard

00:12:50.500 | about this G factor, this mysterious thing

00:12:53.820 | that lurks in the darkness that we cannot quite shine

00:12:56.300 | a light on, we're trying to sneak up on.

00:12:58.980 | So the fact that there's this measure,

00:13:00.540 | stable measure of intelligence, we can't measure directly.

00:13:03.800 | But we can come up with a battery test

00:13:07.780 | or one test that includes a battery

00:13:10.300 | of variable type of questions that can,

00:13:16.060 | reliably or attempt to estimate in a stable way

00:13:20.900 | that G factor, that's a fascinating idea.

00:13:23.300 | So for me as an AI person, it's fascinating.

00:13:25.820 | It's fascinating there's something stable like that

00:13:27.860 | about the human mind, especially if it's grounded

00:13:30.620 | in genetics, it's both fascinating

00:13:33.660 | that as a researcher of the human mind

00:13:38.220 | and all the human psychological, sociological,

00:13:44.260 | ethical questions that start arising,

00:13:46.580 | it makes me uncomfortable.

00:13:48.260 | But truth can be uncomfortable.

00:13:50.180 | - I get that a lot about being uncomfortable

00:13:54.180 | talking about this.

00:13:56.540 | Let me go back and just say one more empirical thing.

00:13:59.760 | It doesn't matter which battery of tests you use.

00:14:07.060 | So there are countless tests.

00:14:10.700 | You can take any 12 of them at random,

00:14:13.500 | extract a G factor and another 12 at random

00:14:17.100 | and extract a G factor.

00:14:18.780 | And those G factors will be highly correlated

00:14:21.300 | like over 0.9 with each other.

00:14:23.300 | So it is a ubiquitous, it doesn't depend

00:14:27.020 | on the content of the test is what I'm trying to say.

00:14:30.260 | It is general among all those tests of mental ability.

00:14:34.020 | And tests of mental abilities include things like,

00:14:37.900 | geez, playing poker.

00:14:41.260 | Your skill at poker is not unrelated to G.

00:14:46.260 | Your skill at anything that requires reasoning

00:14:49.420 | and thinking, anything, spelling, arithmetic,

00:14:54.260 | more complex things, this concept is ubiquitous.

00:14:59.260 | And when you do batteries of tests in different cultures,

00:15:03.880 | you get the same thing.

00:15:05.820 | - So this says something interesting about the human mind

00:15:08.860 | that as a computer is designed to be general.

00:15:11.840 | So that means you can, so it's not easily made specialized.

00:15:17.780 | Meaning if you're going to be good at one thing,

00:15:23.660 | Miyamoto Musashi has this quote, he's an ancient warrior

00:15:29.780 | famous for the Book of Five Rings

00:15:31.900 | in the martial arts world.

00:15:33.420 | And the quote goes, "If you know the way broadly,

00:15:36.620 | "you will see it in everything."

00:15:38.940 | Meaning if you do one thing,

00:15:42.780 | it's going to generalize to everything.

00:15:47.060 | And that's an interesting thing about the human mind.

00:15:49.700 | So that's what the G factor reveals.

00:15:54.420 | Okay, so what's the difference,

00:15:57.100 | if you can elaborate a little bit further,

00:15:58.940 | between IQ and G factor, just because it's a source

00:16:02.400 | of confusion for people?

00:16:03.580 | - An IQ is a score.

00:16:05.700 | People use the word IQ to mean intelligence,

00:16:08.300 | but IQ has a more technical meaning

00:16:11.080 | for people who work in the field.

00:16:12.860 | And it's an IQ score, a score on a test

00:16:16.580 | that estimates the G factor.

00:16:19.000 | And the G factor is what's common

00:16:22.060 | among all these tests of mental ability.

00:16:24.240 | So if you think about, it's not a Venn diagram,

00:16:27.080 | but I guess you could make a Venn diagram out of it,

00:16:30.620 | but the G factor would be really at the core.

00:16:33.940 | What's common to everything.

00:16:37.580 | And what IQ scores do, is they allow a rank order

00:16:42.580 | of people on the score.

00:16:44.540 | And this is what makes people uncomfortable.

00:16:46.940 | This is where there's a lot of controversy

00:16:48.940 | about whether IQ tests are biased

00:16:51.580 | toward any one group or another.

00:16:54.420 | And a lot of the answers to these questions are very clear,

00:16:59.100 | but they also have a technical aspect of it

00:17:02.100 | that's not so easy to explain.

00:17:04.180 | - Well, we'll talk about the fascinating

00:17:06.180 | and the difficult things about all of this.

00:17:08.380 | So by the way, when you say rank order,

00:17:12.580 | that means you get a number,

00:17:13.820 | and that means one person, you can now compare.

00:17:17.540 | Like you could say that this other person

00:17:20.900 | is more intelligent than me.

00:17:23.020 | - Well, what you can say is IQ scores

00:17:25.820 | are interpreted really as percentiles.

00:17:29.200 | So that if you have an IQ of 140 and somebody else has 70,

00:17:34.200 | the metric is such that you cannot say the person

00:17:38.680 | with an IQ of 140 is twice as smart

00:17:41.800 | as a person with an IQ of 70.

00:17:46.000 | That would require a ratio scale with an absolute zero.

00:17:49.920 | Now you may think you know people with zero intelligence,

00:17:53.160 | but in fact, there is no absolute zero on an IQ scale.

00:17:58.000 | It's relative to other people.

00:18:00.020 | So relative to other people,

00:18:03.240 | somebody with an IQ score of 140 is in the upper

00:18:07.400 | less than 1%, whereas somebody with an IQ of 70

00:18:12.400 | is two standard deviations below the mean.

00:18:15.480 | That's a different percentile.

00:18:18.640 | - So it's similar to like in chess,

00:18:20.920 | you have an Elo rating that's designed to rank order people.

00:18:27.760 | So you can't say it's twice one person.

00:18:30.520 | If your Elo rating's twice another person,

00:18:33.520 | I don't think you're twice as good at chess.

00:18:36.160 | It's not stable in that way,

00:18:37.920 | because it's very difficult to do these kinds of comparisons.

00:18:41.120 | So what can we say about the number itself?

00:18:45.960 | Is that stable across tests and so on or no?

00:18:50.560 | - There are a number of statistical properties of any test.

00:18:54.040 | They're called psychometric properties.

00:18:56.440 | You have validity, you have reliability.

00:18:59.320 | Reliability, there are many different kinds of reliability.

00:19:02.760 | They all essentially measure stability.

00:19:05.960 | And IQ tests are stable within an individual.

00:19:09.760 | There are some longitudinal studies

00:19:11.920 | where children were measured at age 11.

00:19:15.800 | And again, when they were 70 years old

00:19:18.120 | and the two IQ scores are highly correlated with each other.

00:19:22.120 | This comes from a fascinating study from Scotland

00:19:26.240 | in the 1930s, some researchers decided to get an IQ test

00:19:31.240 | on every single child age 11 in the whole country.

00:19:36.120 | And they did.

00:19:37.600 | And those records were discovered in an old storeroom

00:19:42.600 | at the University of Edinburgh by a friend of mine,

00:19:47.280 | Ian Deary, who found the records, digitized them,

00:19:52.280 | and has done a lot of research

00:19:54.200 | on the people who are still alive today

00:19:57.520 | from that original study,

00:19:58.840 | including brain imaging research, by the way.

00:20:01.160 | Really, it's a fascinating group of people who are studied.

00:20:06.160 | Not to get ahead of the story,

00:20:09.300 | but one of the most interesting things they found

00:20:12.560 | is a very strong relationship

00:20:14.680 | between IQ measured at age 11 and mortality.

00:20:21.680 | So that, you know, 70 years later,

00:20:26.680 | they looked at the survival rates

00:20:30.960 | and they could get death records from everybody.

00:20:33.280 | And Scotland has universal healthcare for everybody.

00:20:37.120 | And it turned out if you divide people

00:20:40.040 | by their age 11 IQ score into quartiles,

00:20:44.040 | and then look at how many people are alive 70 years later,

00:20:51.200 | I know this is in the book, I have the graph in the book,

00:20:54.640 | but there are essentially twice as many people alive

00:20:57.840 | in the highest IQ quartile than in the lowest IQ quartile.

00:21:01.680 | - Interesting. - It's true in men and women.

00:21:03.880 | - Interesting.

00:21:06.920 | - So it makes a big difference.

00:21:08.100 | Now, why this is the case is not so clear

00:21:12.920 | since everyone had access to healthcare.

00:21:15.600 | - Well, there's a lot, and we'll talk about it,

00:21:18.360 | just the sentences you used now

00:21:22.120 | could be explained by nature or nurture.

00:21:25.800 | We don't know.

00:21:27.000 | Now, there's a lot of science that starts to then dig in

00:21:29.720 | and investigate that question.

00:21:31.720 | But let me linger on the IQ test.

00:21:33.720 | How are the test design, IQ test design, how do they work?

00:21:37.400 | Maybe some examples for people who are not aware.

00:21:40.160 | What makes a good IQ test question

00:21:44.040 | that sneaks up on this G factor?

00:21:47.320 | - Well, your question is interesting

00:21:49.760 | because you want me to give examples of items

00:21:53.360 | that make good items.

00:21:55.240 | And what makes a good item is not so much its content,

00:21:59.440 | but its empirical relationship to the total score

00:22:03.200 | that turns out to be valid by other means.

00:22:07.760 | So for example, let me give you an odd example

00:22:12.500 | from personality testing.

00:22:14.320 | - Nice.

00:22:15.560 | - So there's a personality test

00:22:18.040 | called the Minnesota Multiphasic Personality Inventory, MMPI.

00:22:22.720 | Been around for decades.

00:22:24.120 | - I've heard about this test recently

00:22:26.000 | because of the Johnny Depp and Amber Heard trial.

00:22:29.240 | I don't know if you've been paying attention to that.

00:22:31.200 | But they had psychologists--

00:22:32.040 | - I have not been paying attention to it.

00:22:33.560 | - They had psychologists on the stand,

00:22:35.800 | and they were talking, apparently those psychologists did,

00:22:38.900 | again, I'm learning so much from this trial.

00:22:42.240 | They did different, a battery of tests

00:22:45.840 | to diagnose personality disorders.

00:22:50.420 | Apparently there's that systematic way of doing so,

00:22:53.280 | and the Minnesota one is one of the ones

00:22:55.720 | that there's the most science on.

00:22:59.040 | There's a lot of great papers,

00:23:00.440 | which were all continuously cited on the stand,

00:23:03.840 | which is fascinating to watch.

00:23:05.040 | Sorry, a little bit of attention.

00:23:06.480 | - It's okay, I mean, this is interesting

00:23:07.560 | because you're right, it's been around for decades.

00:23:09.560 | There's a lot of scientific research

00:23:11.240 | on the psychometric properties of the test,

00:23:14.840 | including what it predicts with respect

00:23:18.000 | to different categories of personality disorder.

00:23:22.320 | But what I wanna mention is the content

00:23:24.840 | of the items on that test.

00:23:26.880 | All of the items are essentially true/false items.

00:23:31.880 | True or false, I prefer a shower to a bath.

00:23:35.900 | True or false, I think Lincoln was a better president

00:23:41.200 | than Washington.

00:23:42.280 | What have all these, what does that have to do?

00:23:46.580 | And the point is the content in these items,

00:23:49.720 | nobody knows why these items in aggregate predict anything,

00:23:54.720 | but empirically they do.

00:23:57.900 | It's a technique of choosing items for a test

00:24:01.880 | that is called dust bowl empiricism,

00:24:04.820 | that the content doesn't matter,

00:24:07.440 | but for some reason, when you get a criterion group

00:24:10.640 | of people with this disorder and you compare them

00:24:13.600 | to people without that disorder,

00:24:16.120 | these are the items that distinguish.

00:24:18.360 | Irrespective of content, it's a hard concept to grasp.

00:24:22.680 | - Well, first of all, it's fascinating.

00:24:25.040 | 'Cause I consider myself part psychologist

00:24:32.120 | 'cause I love human-robot interaction,

00:24:35.360 | and that's a problem, half of that problem

00:24:38.180 | is a psychology problem 'cause there's a human.

00:24:40.720 | So designing these tests to get at the questions

00:24:45.180 | is the fascinating part.

00:24:46.540 | What does dust bowl empiricism refer to?

00:24:52.660 | Does it refer to the final result?

00:24:57.420 | Yeah, so it's the test is dust bowl empiricism,

00:25:01.500 | but how do you arrive at the battery of questions?

00:25:04.940 | I presume one of the things, now again,

00:25:08.060 | I'm going to the excellent testimony in that trial,

00:25:11.160 | 'cause they also explain the tests,

00:25:14.760 | that a bunch of the questions are kind of,

00:25:19.360 | make you forget that you're taking a test.

00:25:24.060 | Like, it makes it very difficult for you

00:25:26.700 | to somehow figure out what you're supposed to answer.

00:25:31.700 | - Yes, it's called social desirability.

00:25:34.100 | But we're getting a little far afield

00:25:35.580 | 'cause I only wanted to give that example

00:25:37.400 | of dust bowl empiricism.

00:25:39.200 | When we talk about the items on an IQ test,

00:25:43.540 | many of those items in the dust bowl empiricism method

00:25:49.380 | have no face validity.

00:25:52.940 | In other words, they don't look like they measure anything.

00:25:56.460 | - Yes.

00:25:57.540 | - Whereas most intelligence tests,

00:25:59.900 | the items actually look like

00:26:01.500 | they're measuring some mental ability.

00:26:03.860 | So here's one of the--

00:26:04.700 | - Oh, so you were bringing that up as an example

00:26:06.980 | as what it is not.

00:26:08.200 | - Yes.

00:26:09.040 | - Got it.

00:26:09.860 | - Okay, so I don't want to go too far afield on it.

00:26:12.880 | - Too far afield is actually

00:26:14.520 | one of the names of this podcast,

00:26:16.100 | so I should mention that.

00:26:19.080 | - Far afield, yeah.

00:26:19.920 | - Far afield.

00:26:21.120 | Yeah, so anyway, sorry.

00:26:22.240 | So they feel the questions look like

00:26:25.020 | they passed the face validity test.

00:26:28.040 | - And some more than others.

00:26:29.700 | So for example, let me give you a couple of things here.

00:26:33.840 | One of the subtests on a standard IQ test

00:26:37.300 | is general information.

00:26:39.540 | Let me just think a little bit

00:26:41.880 | 'cause I don't want to give you the actual item.

00:26:44.340 | But if I said, how far is it

00:26:47.120 | between Washington DC and Miami, Florida

00:26:51.140 | within 500 miles, plus or minus?

00:26:54.720 | Well, it's not a fact most people memorize,

00:27:00.060 | but you know something about geography.

00:27:02.680 | You say, well, I flew there once.

00:27:04.440 | I know planes fly 500 miles.

00:27:06.980 | You can kind of make an estimate.

00:27:09.160 | But it also seems like it would be very cultural.

00:27:13.880 | So there's that kind of general information.

00:27:20.200 | Then there's vocabulary test.

00:27:22.760 | What does regatta mean?

00:27:27.760 | And I choose that word

00:27:30.100 | because that word was removed from the IQ test

00:27:33.180 | because people complained that disadvantaged people

00:27:36.260 | would not know that word just from their everyday life.

00:27:41.260 | Here's another example from a different kind of subtest.

00:27:47.220 | - What's regatta, by the way?

00:27:48.940 | - Regatta is a--

00:27:51.020 | - I think I'm disadvantaged.

00:27:52.220 | - A sailing competition, a competition with boats.

00:27:55.420 | Not necessarily sailing, but a competition with boats.

00:27:58.780 | - Yep, yep, I'm probably disadvantaged in that way.

00:28:02.180 | Okay, excellent, so that was removed.

00:28:03.860 | Anyway, what you were saying.

00:28:04.980 | - Okay, so here's another subtest.

00:28:07.840 | I'm gonna repeat a string of numbers,

00:28:09.840 | and when I'm done, I want you to repeat them back to me.

00:28:12.680 | Ready?

00:28:13.520 | Seven, four, two, eight, one, six.

00:28:19.360 | - That's way too many.

00:28:22.580 | Seven, four, two, eight, one, six.

00:28:25.480 | - You get the idea.

00:28:26.320 | Now, the actual test starts with a smaller number,

00:28:30.480 | you know, like two numbers, and then as people get it right,

00:28:33.520 | you keep going, adding to the string of numbers

00:28:36.560 | until they can't do it anymore.

00:28:38.640 | Okay, but now try this.

00:28:40.760 | I'm gonna say some numbers, and when I'm done,

00:28:43.800 | I want you to repeat them to me backwards.

00:28:46.680 | - I quit.

00:28:47.800 | - Okay, now, so I gave you some examples

00:28:51.560 | of the kind of items on an IQ test.

00:28:53.600 | - Yes. - General information.

00:28:55.280 | I can't even remember all of it.

00:28:58.640 | General information, vocabulary,

00:29:01.680 | digit span forward and digit span backward.

00:29:06.680 | - Well, you said I can't even remember them.

00:29:08.960 | That's a good question for me.

00:29:10.500 | What does memory have to do with your function?

00:29:14.560 | - Let's hold on.

00:29:15.400 | - Okay, all right.

00:29:16.240 | - Let's just talk about these examples.

00:29:19.720 | Now, some of those items seem very cultural,

00:29:24.720 | and others seem less cultural.

00:29:29.960 | Which ones do you think, scores on which subtests

00:29:35.360 | are most highly correlated with the G factor?

00:29:38.440 | - Well, the two advances less cultural.

00:29:42.560 | - Well, it turns out vocabulary is highly correlated,

00:29:49.580 | and it turns out that digit span backwards

00:29:54.220 | is highly correlated.

00:29:55.720 | - How do you figure?

00:29:58.620 | - Now you have decades of research

00:30:02.240 | to answer the question, how do you figure?

00:30:04.580 | - Right, so now there's good research

00:30:08.460 | that gives you intuition about what kind of questions

00:30:11.320 | get added, just like there's something I've done,

00:30:18.100 | I've actually used for research,

00:30:20.300 | just send me an autonomous vehicle,

00:30:22.020 | like whether humans are paying attention,

00:30:24.380 | there's a body of literature that does

00:30:26.820 | like end-back tests, for example,

00:30:29.060 | where you have to put workload on the brain

00:30:34.060 | to do recall, memory recall,

00:30:37.820 | and that helps you kind of put some work onto the brain

00:30:42.100 | while the person is doing some other task,

00:30:44.260 | and there's some interesting research with that.

00:30:47.700 | But that's loading the memory,

00:30:48.940 | and so there's like research around stably

00:30:52.260 | what that means about the human mind,

00:30:54.100 | and here you're saying recall backwards

00:30:57.140 | is a good predictor.

00:31:00.020 | - It's a transformation.

00:31:01.740 | - Yeah, so you have to do some,

00:31:04.120 | like you have to load that into your brain,

00:31:07.820 | and not just remember it, but do something with it.

00:31:11.220 | - Right, now here's another example

00:31:12.620 | of a different kind of test,

00:31:14.460 | called the Hick paradigm, and it's not verbal at all.

00:31:18.340 | It's a little box, and there are a series of lights

00:31:23.140 | arranged in a semi-circle at the top of the box,

00:31:27.420 | and then there's a home button that you press,

00:31:31.300 | and when one of the lights goes on,

00:31:33.880 | there's a button next to each of those lights,

00:31:37.780 | you take your finger off the home button,

00:31:39.940 | and you just press the button next to the light

00:31:42.780 | that goes on, and so it's a very simple reaction time.

00:31:46.500 | Light goes on, as quick as you can, you press the button,

00:31:49.220 | and you get a reaction time.

00:31:50.540 | From the moment you lift your finger off the button,

00:31:53.820 | when you press the button where the light is,

00:31:57.780 | that reaction time doesn't really correlate

00:32:02.160 | with IQ very much, but if you change the instructions,

00:32:07.160 | and you say three lights are gonna come on simultaneously,

00:32:13.100 | I want you to press the button next to the light

00:32:15.620 | that's furthest from the other two.

00:32:17.860 | So maybe lights one and two go on,

00:32:21.260 | and light six goes on simultaneously.

00:32:24.260 | You take your finger off, and you would press

00:32:26.120 | the button by light six.

00:32:27.900 | That's, that reaction time to a more complex task,

00:32:33.860 | it's not really hard, almost everybody gets it all right,

00:32:38.360 | but your reaction time to that is highly correlated

00:32:42.520 | with the G factor.

00:32:43.740 | - This is fascinating, so reaction time,

00:32:46.020 | so there's a temporal aspect to this.

00:32:48.580 | So what role does time-- - Speed of processing.

00:32:50.860 | It's the speed of processing.

00:32:53.020 | - Is this also true for ones that take longer,

00:32:55.740 | like five, 10, 30 seconds?

00:32:57.700 | Is time part of the measure with some of these ideas?

00:33:01.180 | - Yes, and that is why some of the best IQ tests

00:33:05.620 | have a time limit, because if you have no time limit,

00:33:10.620 | people can do better, but it doesn't distinguish

00:33:15.040 | among people that well.

00:33:17.760 | So that adding the time element is important.

00:33:21.440 | So speed of information processing,

00:33:24.100 | and reaction time is a measure of speed

00:33:27.220 | of information processing, turns out to be related

00:33:30.600 | to the G factor.

00:33:31.920 | - But the G factor only accounts for maybe half

00:33:35.040 | or some amount on the test performance.

00:33:37.560 | For example, I get pretty bad test anxiety.

00:33:42.060 | Like I was never, I mean, I just don't enjoy tests.

00:33:47.060 | I enjoy going back into my cave and working.

00:33:51.300 | Like I've always enjoyed homework way more than tests,

00:33:55.200 | no matter how hard the homework is,

00:33:57.780 | 'cause I can go back to the cave and hide away

00:33:59.980 | and think deeply.

00:34:00.820 | There's something about being watched

00:34:02.700 | and having a time limit that really makes me anxious,

00:34:06.060 | and I could just see the mind not operating optimally

00:34:09.520 | at all, but you're saying underneath there,

00:34:11.640 | there's still a G factor, there's still--

00:34:13.720 | - No question, no question.

00:34:16.360 | - Boy.

00:34:17.280 | - And if you get anxious taking the test,

00:34:19.240 | many people say, oh, I didn't do well 'cause I'm anxious.

00:34:22.140 | I hear that a lot.

00:34:24.960 | Well, fine, if you're really anxious during the test,

00:34:28.320 | the score will be a bad estimate of your G factor.

00:34:32.000 | It doesn't mean the G factor isn't there.

00:34:34.840 | And by the way, standardized tests like the SAT,

00:34:39.080 | they're essentially intelligence tests.

00:34:43.020 | They are highly G loaded.

00:34:45.200 | Now, the people who make the SAT don't wanna mention that.

00:34:49.200 | They have enough trouble justifying standardized testing,

00:34:54.000 | but to call it an intelligence test

00:34:56.000 | is really beyond the pale.

00:34:58.400 | But in fact, it's so highly correlated

00:35:00.700 | because it's a reasoning test.

00:35:03.200 | The SAT is a reasoning test, a verbal reasoning,

00:35:06.160 | mathematical reasoning.

00:35:08.240 | And if it's a reasoning test, it has to be related to G.

00:35:12.240 | But if people go in and take a standardized test,

00:35:17.560 | whether it's an IQ test or the SAT,

00:35:20.120 | and they happen to be sick that day with 102 fever,

00:35:24.600 | the score is not going to be a good estimate of their G.

00:35:29.600 | If they retake the test when they're not anxious

00:35:33.120 | or less anxious or don't have a fever,

00:35:35.740 | the score will go up and that will be a better estimate.

00:35:39.960 | But you can't say their G factor increased

00:35:43.120 | between the two tests.

00:35:45.160 | - Well, it's interesting.

00:35:46.480 | So the question is how wide of a battery of tests

00:35:50.000 | is required to estimate the G factor well?

00:35:53.380 | Because I'll give you as my personal example,

00:35:55.160 | I took the SAT in, I think it was called the ACT

00:35:58.760 | where I was too, also, I took SAT many times.

00:36:02.880 | Every single time I got a perfect on math.

00:36:05.520 | And verbal, the time limit on the verbal

00:36:08.800 | made me very anxious.

00:36:11.000 | I did not, I mean, part of it,

00:36:12.520 | I didn't speak English very well,

00:36:14.240 | but honestly, it was like,

00:36:15.920 | you're supposed to remember stuff.

00:36:17.480 | And like, I was so anxious.

00:36:18.800 | And like, as I'm reading, I'm sweating.

00:36:20.940 | I can't, you know that like,

00:36:22.680 | that feeling you have when you're reading a book

00:36:27.040 | and you just read a page and you know nothing

00:36:30.120 | about what you've read because you zoned out?

00:36:32.640 | That's the same feeling of like,

00:36:35.000 | I can't, I have to, you're like, nope.

00:36:38.040 | Read and understand and that anxiety is like,

00:36:41.120 | and you start seeing like the typography

00:36:44.880 | versus the content of the words.

00:36:47.120 | Like that was, I don't, it's interesting because

00:36:50.800 | I know that what they're measuring,

00:36:55.760 | I could see being correlated with something.

00:36:58.760 | But that anxiety or some aspect of the performance

00:37:02.840 | sure plays a factor.

00:37:07.020 | And I wonder how you sneak up in a stable way.

00:37:10.400 | I mean, this is a broader discussion about,

00:37:12.840 | that's like standardized testing, how you sneak up,

00:37:16.600 | how you get at the fact that I'm super anxious

00:37:19.880 | and still nevertheless measure

00:37:21.440 | some aspect of my intelligence.

00:37:23.080 | I wonder, I don't know if you can say to that,

00:37:26.640 | that time limit sure is a pain.

00:37:28.600 | - Well, let me say this.

00:37:30.560 | There are two ways to approach the very real problem

00:37:34.200 | that you say that some people just get anxious

00:37:36.920 | or not good test takers.

00:37:39.000 | By the way, part of testing is,

00:37:44.000 | you know the answer, you can figure out the answer

00:37:47.520 | or you can't.

00:37:48.440 | If you don't know the answer,

00:37:51.640 | there are many reasons you don't know the answer

00:37:54.000 | at that particular moment.

00:37:55.280 | You may have learned it once and forgotten it.

00:37:58.480 | You may, it may be on the tip of your tongue

00:38:00.600 | and you just can't get it

00:38:01.920 | because you're anxious about the time limit.

00:38:03.880 | You may never have learned it.

00:38:05.920 | You may never, you may have been exposed to it,

00:38:08.720 | but it was too complicated and you couldn't learn it.

00:38:11.440 | I mean, there are all kinds of reasons here,

00:38:14.000 | but for an individual to interpret your scores

00:38:18.840 | as an individual, whoever is interpreting the score

00:38:23.360 | has to take into account various things

00:38:26.280 | that would affect your individual score.

00:38:29.200 | And that's why decisions about college admission

00:38:32.720 | or anything else where tests are used

00:38:35.720 | are hardly ever the only criterion to make a decision.

00:38:40.720 | - And I think people are,

00:38:44.120 | college admissions letting go of that very much.

00:38:46.880 | - Oh yes, yeah.

00:38:48.120 | - But what does that even mean?

00:38:51.040 | Because is it possible to design standardized tests

00:38:55.560 | that do get, that are useful to college admissions?

00:38:58.400 | - Well, they already exist.

00:38:59.720 | The SAT is highly correlated

00:39:02.320 | with many aspects of success at college.

00:39:05.240 | - Here's the problem.

00:39:06.360 | So maybe you could speak to this.

00:39:09.200 | The correlation across the population versus individuals.

00:39:13.360 | So,

00:39:14.200 | our criminal justice system is designed to make sure,

00:39:23.440 | well, it's still, there's tragic cases

00:39:27.280 | where innocent people go to jail,

00:39:29.640 | but you try to avoid that.

00:39:31.320 | In the same way with testing,

00:39:34.400 | it just, it would suck for an SAT to miss genius.

00:39:38.740 | - Yes, and it's possible, but it's statistically unlikely.

00:39:43.240 | So it really comes down to,

00:39:45.880 | do which piece of information

00:39:51.920 | maximizes your decision-making ability?

00:39:56.920 | So, if you just use high school grades, it's okay,

00:40:03.600 | but you will miss some people

00:40:07.160 | who just don't do well in high school,

00:40:09.120 | but who are actually pretty smart,

00:40:11.320 | smart enough to be bored silly in high school,

00:40:13.960 | and they don't care,

00:40:14.920 | and their high school GPA isn't that good.

00:40:17.760 | So you will miss them,

00:40:19.560 | in the same sense that somebody who could be very able

00:40:24.560 | and ready for a college just doesn't do well on their SAT.

00:40:28.320 | This is why you make decisions

00:40:31.640 | with taking in a variety of information.

00:40:36.100 | The other thing I wanted to say,

00:40:38.040 | I talked about when you make a decision for an individual.

00:40:43.800 | Statistically, for groups,

00:40:46.680 | there are many people who have a disparity

00:40:50.020 | between their math score and their verbal score.

00:40:53.040 | That disparity, or the other way around,

00:40:55.600 | that disparity is called tilt.

00:40:58.640 | The score is tilted one way or the other,

00:41:01.720 | and that tilt has been studied empirically

00:41:05.040 | to see what that predicts.

00:41:07.280 | And in fact, you can't make predictions

00:41:09.400 | about college success based on tilt.

00:41:14.840 | And mathematics is a good example.

00:41:16.760 | There are many people,

00:41:18.320 | especially non-native speakers of English,

00:41:21.080 | come to this country, take the SATs,

00:41:23.480 | do very well on the math and not so well on the verbal.

00:41:26.840 | Well, if they're applying to a math program,

00:41:30.380 | the professors there who are making the decision

00:41:33.840 | or the admissions officers,

00:41:35.500 | don't wait so much to score on verbal,

00:41:39.620 | especially if it's a non-native speaker.

00:41:42.120 | - Well, so yeah, you have to try to,

00:41:44.520 | in the admission process, bring in the context.

00:41:47.800 | But non-native isn't really the problem.

00:41:50.760 | I mean, that was part of the problem for me.

00:41:53.720 | But it's the anxiety was, which it's interesting.

00:41:57.960 | It's interesting.

00:41:58.980 | Oh boy, reducing yourself down to numbers.

00:42:06.520 | But it's still true.

00:42:07.840 | It's still the truth.

00:42:09.200 | - Well-- - It's a painful truth.

00:42:10.720 | That same anxiety that led me to be

00:42:13.880 | to struggle with the SAT verbal tests

00:42:18.880 | is still within me in all ways of life.

00:42:24.640 | So maybe that's not anxiety.

00:42:26.860 | Maybe that's something,

00:42:28.760 | you know, like personality is also pretty stable.

00:42:32.400 | - Personality is stable.

00:42:34.480 | Personality does impact the way you navigate life.

00:42:39.480 | - Yeah.

00:42:41.020 | - There's no question.

00:42:42.480 | - Yeah, and we should say that the G factor in intelligence

00:42:45.680 | is not just about some kind of number on a paper.

00:42:50.400 | It also has to do with how you navigate life,

00:42:54.760 | how easy life is for you in this very complicated world.

00:42:59.760 | So personality's all tied into that

00:43:02.880 | in some deep fundamental way.

00:43:05.880 | - But now you've hit the key point

00:43:07.720 | about why we even want to study intelligence.

00:43:11.320 | And personality, I think, to a lesser extent.

00:43:13.360 | But that's my interest is more on intelligence.

00:43:17.480 | I went to graduate school and wanted to study personality,

00:43:20.140 | but that's kind of another story

00:43:22.620 | how I got kind of shifted from personality research

00:43:25.180 | over to intelligence research.

00:43:27.420 | Because it's not just a number.

00:43:30.000 | Intelligence is not just an IQ score.

00:43:32.520 | It's not just an SAT score.

00:43:34.700 | It's what those numbers reflect

00:43:37.640 | about your ability to navigate everyday life.

00:43:42.140 | It has been said that life is one long intelligence test.

00:43:48.000 | (laughing)

00:43:50.800 | And who can't relate to that?

00:43:55.440 | And if you doubt, see, another problem here

00:43:58.480 | is a lot of critics of intelligence research,

00:44:00.820 | intelligence testing, tend to be academics

00:44:04.040 | who, by and large, are pretty smart people.

00:44:07.320 | And pretty smart people, by and large,

00:44:10.080 | have enormous difficulty understanding

00:44:12.920 | what the world is like for people with IQs of 80 or 75.

00:44:17.920 | It is a completely different everyday experience.

00:44:23.080 | Even IQ scores of 85, 90,

00:44:28.080 | there's a popular television program, Judge Judy,

00:44:31.840 | where Judge Judy deals with everyday people

00:44:35.360 | with everyday problems.

00:44:36.840 | And you can see the full range

00:44:39.480 | of problem-solving ability demonstrated there.

00:44:43.280 | And sometimes she does it for laughs,

00:44:45.360 | but it really isn't funny because people who are,

00:44:50.360 | there are people who are very limited

00:44:54.540 | in their life navigation, let alone success,

00:44:59.240 | by having, by not having good reasoning skills,

00:45:04.720 | which cannot be taught.

00:45:06.960 | We know this, by the way, because there are many efforts.

00:45:09.600 | You know, the United States military,

00:45:11.320 | which excels at training people.

00:45:14.160 | I mean, I don't know that there's a better organization

00:45:16.520 | in the world for training diverse people.

00:45:20.320 | And they won't take people with IQs under,

00:45:22.760 | I think, 83 is the cutoff.

00:45:25.400 | Because they have found, they are unable to train

00:45:29.440 | people with lower IQs to do jobs in the military.

00:45:34.400 | - So one of the things that G-Factor

00:45:36.200 | has to do with is learning.

00:45:37.640 | - Absolutely.

00:45:38.580 | Some people learn faster than others.

00:45:42.680 | Some people learn more than others.

00:45:45.480 | Now, faster, by the way, is not necessarily better,

00:45:48.080 | as long as you get to the same place eventually.

00:45:51.980 | But, you know, there are professional schools

00:45:56.360 | that want students who can learn the fastest

00:45:59.720 | because they can learn more, or learn deeper,

00:46:03.120 | or all kinds of ideas about why you select people

00:46:08.120 | with the highest scores.

00:46:09.560 | And there's nothing funnier, by the way,

00:46:12.640 | to listen to a bunch of academics complain

00:46:15.680 | about the concept of intelligence and intelligence testing.

00:46:19.400 | And then you go to a faculty meeting

00:46:21.320 | where they're discussing who to hire among the applicants.

00:46:24.760 | And all they talk about is how smart the person is.

00:46:27.320 | - We'll get to that.

00:46:29.360 | We'll sneak up to that in different ways.

00:46:31.200 | But there's something about reducing a person

00:46:33.020 | to a number that in part is grounded

00:46:35.280 | to the person's genetics

00:46:36.920 | that makes people very uncomfortable.

00:46:38.800 | - But nobody does that.

00:46:40.480 | Nobody in the field actually does that.

00:46:43.800 | That is a worry that is a worry like,

00:46:48.800 | well, I don't wanna call it a conspiracy theory.

00:46:55.880 | I mean, it's a legitimate worry.

00:46:58.360 | But it just doesn't happen.

00:47:01.400 | Now, I had a professor in graduate school

00:47:03.580 | who was the only person I ever knew

00:47:05.860 | who considered the students only by their test scores.

00:47:10.860 | - Yes. - And later in his life,

00:47:14.080 | he kind of backed off that.

00:47:16.400 | But--

00:47:18.160 | - Let me ask you this.

00:47:20.320 | So we'll jump around.

00:47:21.540 | I'll come back to it.

00:47:22.860 | I tend to, I've had political discussions with people.

00:47:29.500 | And actually, my friend Michael Malice, he's an anarchist.

00:47:34.500 | I disagree with him on basically everything

00:47:39.220 | except the fact that love is a beautiful thing in this world.

00:47:44.220 | And he says this test about left versus right,

00:47:50.580 | whatever, it doesn't matter what the test is.

00:47:52.260 | But he believes, the question is,

00:47:54.920 | do you believe that some people are better than others?

00:47:58.900 | The question is ambiguous.

00:48:03.900 | Do you believe some people are better than others?

00:48:06.140 | And to me, sort of the immediate answer is no.

00:48:10.720 | It's a poetic question.

00:48:12.860 | It's an ambiguous question, right?

00:48:15.660 | Like, people wanna maybe,

00:48:18.860 | the temptation to ask better at what?

00:48:20.700 | Better at like sports, so on.

00:48:23.460 | No, to me, I stand with the sort of

00:48:27.420 | the founding documents of this country,

00:48:29.980 | which is all men are created equal.

00:48:32.300 | There's a basic humanity.

00:48:34.380 | And there's something about tests of intelligence.

00:48:39.380 | Just knowing that some people are different,

00:48:43.420 | like the science of intelligence that shows

00:48:45.420 | that some people are genetically

00:48:47.420 | in some stable way across a lifetime,

00:48:52.420 | have a greater intelligence than others,

00:48:56.140 | makes people feel like some people are better than others.

00:49:01.140 | And that makes them very uncomfortable.

00:49:03.580 | And maybe you can speak to that.

00:49:06.940 | The fact that some people are more intelligent than others

00:49:09.540 | in a way that's,

00:49:10.800 | cannot be compensated through education,

00:49:17.900 | through anything you do in life.

00:49:21.820 | What do we do with that?

00:49:24.460 | - Okay, there's a lot there.

00:49:25.900 | We haven't really talked about the genetics of it yet,

00:49:29.940 | but you are correct in that it is my interpretation

00:49:34.940 | of the data that genetics has a very important influence

00:49:39.700 | on the G factor.

00:49:41.380 | And this is controversial.

00:49:42.980 | We can talk about it.

00:49:44.420 | But if you think that genetics,

00:49:47.020 | that genes are deterministic, are always deterministic,

00:49:50.780 | that leads to kind of the worry that you expressed.

00:49:54.040 | But we know now in the 21st century

00:49:58.560 | that many genes are not deterministic,

00:50:00.860 | that are probabilistic,

00:50:02.780 | meaning their gene expression can be influenced.

00:50:07.780 | Now, whether they're influenced

00:50:11.060 | only by other biological variables

00:50:14.020 | or other genetic variables

00:50:16.400 | or environmental or cultural variables,

00:50:19.100 | that's where the controversy comes in.

00:50:21.740 | And we can discuss that in more detail if you like.

00:50:27.140 | But to go to the question about better, are people better,

00:50:31.660 | there's zero evidence that smart people are better

00:50:36.660 | with respect to important aspects of life,

00:50:43.300 | like honesty, even likability.

00:50:47.980 | I'm sure you know many very intelligent people

00:50:50.340 | who are not terribly likable or terribly kind

00:50:53.420 | or terribly honest.

00:50:55.480 | - Is there something to be said?

00:50:56.760 | So one of the things I've recently reread

00:50:59.820 | for the second time,

00:51:01.080 | I guess that's what the word reread means,

00:51:04.500 | the rise and fall of the Third Reich,

00:51:08.980 | which is, I think, the best telling

00:51:12.060 | of the rise and fall of Hitler.

00:51:14.680 | And one of the interesting things about the people

00:51:17.860 | that, how should I say it,

00:51:22.180 | justified or maybe propped up the ideas

00:51:32.260 | that Hitler put forward is the fact

00:51:35.900 | that they were extremely intelligent.

00:51:38.420 | They were the intellectual class.

00:51:40.560 | They were, it was obvious that they thought

00:51:46.220 | very deeply and rationally about the world.

00:51:49.500 | So what I would like to say is,

00:51:51.580 | one of the things that shows to me

00:51:53.860 | is some of the worst atrocities in the history of humanity

00:51:57.940 | have been committed by very intelligent people.

00:52:00.740 | So that means that intelligence

00:52:04.620 | doesn't make you a good person.

00:52:06.420 | I wonder if, you know, there's a G factor for intelligence.

00:52:11.420 | I wonder if there's a G factor for goodness.

00:52:15.840 | You know, they need you in good and evil.

00:52:19.200 | Of course, that's probably harder to measure

00:52:21.720 | 'cause that's such a subjective thing,

00:52:23.240 | what it means to be good.

00:52:25.320 | And even the idea of evil is a deeply uncomfortable thing

00:52:29.400 | 'cause how do we know?

00:52:31.360 | But it's independent, whatever it is,

00:52:33.500 | it's independent of intelligence.

00:52:36.040 | So I agree with you about that.

00:52:37.960 | But let me say this.

00:52:39.360 | I have also asserted my belief

00:52:44.120 | that more intelligence is better than less.

00:52:47.240 | That doesn't mean more intelligent people are better people,

00:52:54.240 | but all things being equal,

00:52:55.840 | would you like to be smarter or less smart?

00:52:58.740 | So if I had a pill, I have two pills,

00:53:01.280 | I said, this one will make you smarter,

00:53:02.880 | this one will make you dumber.

00:53:04.480 | Which one would you like?

00:53:06.480 | Are there any circumstances

00:53:07.960 | under which you would choose to be dumber?

00:53:09.840 | - Well, let me ask you this.

00:53:11.560 | That's a very nuanced and interesting question.

00:53:13.960 | There's been books written about this, right?

00:53:18.360 | Now we'll return to the hard questions,

00:53:21.680 | the interesting questions,

00:53:22.760 | but let me ask about human happiness.

00:53:24.940 | Does intelligence lead to happiness?

00:53:29.200 | - No.

00:53:30.040 | - So, okay, so back to the pill then.

00:53:34.960 | So why, when would you take the pill?

00:53:38.980 | So you said IQ 80.

00:53:41.520 | 90, 100, 110, you start going through the quartiles

00:53:46.520 | and is it obvious, isn't there a diminishing returns

00:53:51.600 | and then it starts becoming negative?

00:53:56.520 | - This is an empirical question.

00:53:59.300 | And so that I have advocated in many forums

00:54:06.160 | more research on enhancing the G factor.

00:54:11.160 | Right now there have been many claims

00:54:14.440 | about enhancing intelligence

00:54:17.060 | with you mentioned the NBAC training,

00:54:19.080 | that was a big deal a few years ago, it doesn't work.

00:54:22.640 | Data's very clear, it does not work.

00:54:24.900 | - Or doing like memory tests, like training and so on.

00:54:28.560 | - Yeah, it may give you a better memory in the short run,

00:54:32.720 | but it doesn't impact your G factor.

00:54:36.000 | It was very popular a couple of decades ago

00:54:40.920 | that the idea that listening to Mozart

00:54:44.560 | could make you more intelligent.

00:54:46.560 | There was a paper published on this

00:54:48.200 | with somebody I knew published this paper.

00:54:50.320 | Intelligence researchers never believed it for a second.

00:54:54.600 | Been hundreds of studies, all the meta-analyses,

00:54:57.800 | all the summaries and so on.

00:54:59.360 | So there's nothing to it, nothing to it at all.

00:55:04.680 | (Luke laughs)

00:55:05.520 | But wouldn't it be something,

00:55:08.720 | wouldn't it be world shaking

00:55:11.800 | if you could take the normal distribution of intelligence,

00:55:15.920 | which we haven't really talked about yet,

00:55:17.460 | but IQ scores and the G factor

00:55:20.360 | is thought to be a normal distribution,

00:55:22.720 | and shift it to the right so that everybody is smarter.

00:55:30.080 | Even a half a standard deviation would be world shaking

00:55:34.280 | because there are many social problems,

00:55:38.660 | many, many social problems that are exacerbated

00:55:43.200 | by people with lower ability to reason stuff out

00:55:48.200 | and navigate everyday life.

00:55:50.580 | - So I wonder if there's a threshold.

00:55:53.740 | So maybe I would push back and say universal shifting

00:55:59.380 | of the normal distribution

00:56:02.200 | may not be the optimal way of shifting.

00:56:05.000 | Maybe it's better to,

00:56:07.300 | whatever the asymmetric kind of distribution is,

00:56:10.580 | is like really pushing the lower up

00:56:13.340 | versus trying to make the people

00:56:17.520 | at the average more intelligent.

00:56:19.600 | - So you're saying that if in fact

00:56:21.020 | there was some way to increase G,

00:56:23.500 | let's just call it metaphorically a pill, an IQ pill,

00:56:27.760 | we should only give it to people at the lower end?

00:56:30.560 | - No, it's just intuitively I can see

00:56:34.880 | that life becomes easier at the lower end if it's increased.

00:56:39.760 | It becomes less and less,

00:56:41.620 | it is an empirical scientific question,

00:56:43.520 | but it becomes less and less obvious to me

00:56:46.140 | that more intelligence is better.

00:56:50.480 | - At the high end, not because it would make life easier,

00:56:56.460 | but it would make whatever problems you're working on

00:57:00.960 | more solvable.

00:57:02.660 | And if you are working on artificial intelligence,

00:57:06.700 | there's a tremendous potential

00:57:09.220 | for that to improve society.

00:57:13.300 | - I understand.

00:57:14.580 | So at the whatever problems you're working on, yes,

00:57:18.980 | but there's also the problem of the human condition.

00:57:21.720 | There's love, there's fear,

00:57:24.320 | and all of those beautiful things

00:57:26.720 | that sometimes if you're good at solving problems,

00:57:29.840 | you're going to create more problems for yourself.

00:57:32.340 | I'm not exactly sure.

00:57:34.440 | So ignorance is bliss, is a thing.

00:57:37.200 | So there might be a place,

00:57:38.400 | there might be a sweet spot of intelligence

00:57:40.960 | given your environment, given your personality,

00:57:43.680 | all of those kinds of things,

00:57:45.040 | and that becomes less beautifully complicated

00:57:48.160 | the more and more intelligent you become.

00:57:50.480 | But that's a question for literature,

00:57:53.160 | not for science, perhaps.

00:57:54.680 | - Well, imagine this.

00:57:56.200 | Imagine there was an IQ pill,

00:57:58.400 | and it was developed by a private company,

00:58:01.580 | and they are willing to sell it to you.

00:58:05.040 | And whatever price they put on it,

00:58:07.880 | you are willing to pay it

00:58:09.480 | because you would like to be smarter.

00:58:11.720 | But just before they give you a pill,

00:58:14.320 | they give you a disclaimer form to sign.

00:58:20.000 | Don't hold us,

00:58:21.000 | you understand that this pill has no guarantee

00:58:25.160 | that your life is going to be better,

00:58:26.960 | and in fact, it could be worse.

00:58:28.880 | - Well, yes, that's how lawyers work,

00:58:32.200 | but I would love for science to answer the question,

00:58:35.160 | to try to predict if your life

00:58:36.640 | is going to be better or worse

00:58:38.680 | when you become more or less intelligent.

00:58:41.200 | It's a fascinating question

00:58:43.240 | about what is the sweet spot for the human condition.

00:58:47.600 | Some of the things we see as bugs

00:58:49.840 | might be actually features,

00:58:51.920 | may be crucial to our overall happiness,

00:58:55.360 | is our limitations might lead to more happiness than less.

00:58:59.080 | But again, more intelligence is better at the lower end.

00:59:02.720 | That's something that's less arguable

00:59:06.440 | and fascinating, if possible, to increase.

00:59:10.080 | - But you know, there's virtually no research

00:59:12.920 | that's based on a neuroscience approach

00:59:15.280 | to solving that problem.

00:59:17.480 | All the solutions that have been proposed

00:59:20.760 | to solve that problem or to ameliorate that problem

00:59:25.360 | are essentially based on the blank slate assumption

00:59:29.640 | that enriching the environment, removing barriers,

00:59:34.640 | all good things, by the way,

00:59:35.960 | I'm not against any of those things,

00:59:38.120 | but there's no empirical evidence

00:59:39.680 | that they're going to improve the general reasoning ability

00:59:45.200 | or make people more employable.

00:59:47.880 | - Have you read "Flowers of Algernon"?

00:59:49.960 | - Yes.

00:59:50.800 | - That's to the question of intelligence and happiness.

00:59:54.640 | - There are many profound aspects of that story.

00:59:59.400 | It was a film that was very good.

01:00:01.800 | The film was called "Charlie,"

01:00:04.760 | for the younger people who are listening to this.

01:00:07.460 | You might be able to stream it on Netflix or something.

01:00:11.720 | But it was a story about a person with very low IQ

01:00:16.720 | who underwent a surgical procedure in the brain

01:00:21.760 | and he slowly became a genius.

01:00:23.800 | And the tragedy of the story is the effect was temporary.

01:00:29.500 | It's a fascinating story, really.

01:00:33.120 | - That goes in contrast to the basic human experience

01:00:36.600 | that each of us individually have,

01:00:38.440 | but it raises the question of the full range of people

01:00:43.440 | you might be able to be,

01:00:45.640 | given different levels of intelligence.

01:00:48.760 | You've mentioned the normal distribution.

01:00:51.940 | So let's talk about it.

01:00:54.520 | There's a book called "The Bell Curve," written in 1994,

01:00:58.360 | written by psychologist Richard Herrnstein

01:01:01.340 | and political scientist Charles Murray.

01:01:04.520 | Why was this book so controversial?

01:01:08.140 | - This is a fascinating book.

01:01:10.740 | I know Charles Murray.

01:01:12.620 | I've had many conversations with him.

01:01:15.340 | - Yeah, what is the book about?

01:01:16.780 | - The book is about the importance of intelligence

01:01:21.620 | in everyday life.

01:01:24.140 | That's what the book is about.

01:01:27.520 | It's an empirical book.

01:01:29.180 | It has statistical analyses of very large databases

01:01:34.180 | that show that essentially IQ scores or their equivalent

01:01:39.180 | are correlated to all kinds of social problems

01:01:44.640 | and social benefits.

01:01:46.800 | And that in itself is not where the controversy

01:01:51.740 | about that book came.

01:01:53.640 | The controversy was about one chapter in that book.

01:01:57.580 | And that is a chapter about the average difference

01:02:02.240 | in mean scores between black Americans and white Americans.

01:02:06.740 | And these are the terms that were used

01:02:08.460 | in the book at the time

01:02:09.740 | and are still used to some extent.

01:02:12.240 | And historically, or really for decades,

01:02:19.040 | it has been observed that disadvantaged groups

01:02:27.980 | score on average lower than Caucasians

01:02:32.920 | on academic tests, tests of mental ability,

01:02:38.280 | and especially on IQ tests.

01:02:40.440 | And the difference is about a standard deviation,

01:02:43.120 | which is about 15 points, which is a substantial difference.

01:02:46.600 | In the book, Hernstein and Murray in this one chapter

01:02:54.160 | assert clearly and unambiguously

01:02:58.260 | that whether this average difference

01:03:01.960 | is due to genetics or not, they are agnostic.

01:03:06.960 | They don't know.

01:03:08.420 | Moreover, they assert they don't care

01:03:11.280 | because you wouldn't treat anybody differently

01:03:14.020 | knowing if there was a genetic component or not

01:03:17.780 | because that's a group average finding.

01:03:20.700 | Every individual has to be treated as an individual.

01:03:24.100 | You can't make any assumption

01:03:26.140 | about what that person's intellectual ability might be

01:03:30.740 | from the fact of a average group difference.

01:03:33.260 | They're very clear about this.

01:03:34.960 | Nonetheless, people took away,

01:03:41.100 | I'm gonna choose my words carefully

01:03:43.180 | 'cause I have a feeling that many critics

01:03:44.760 | didn't actually read the book.

01:03:49.140 | They took away that Hernstein and Murray were saying

01:03:51.860 | that blacks are genetically inferior.

01:03:54.900 | That was the take-home message.

01:03:56.500 | And if they weren't saying it, they were implying it

01:03:59.660 | because they had a chapter that discussed

01:04:02.700 | this empirical observation of a difference.

01:04:05.900 | And isn't this horrible?

01:04:10.500 | And so the reaction to that book was incendiary.

01:04:18.380 | - What do we know about, from that book

01:04:22.500 | and the research beyond,

01:04:24.140 | about race differences and intelligence?

01:04:30.580 | - It's still the most incendiary topic in psychology.

01:04:33.740 | Nothing has changed that.

01:04:35.820 | Anybody who even discusses it is easily called a racist

01:04:40.820 | just for discussing it.

01:04:42.980 | It's become fashionable to find racism

01:04:45.520 | in any discussion like this.

01:04:49.900 | It's unfortunate.

01:04:51.300 | The short answer to your question is

01:04:56.020 | there's been very little actual research on this topic

01:05:01.900 | since 19--

01:05:03.820 | - Since the Bell Curve.

01:05:05.180 | - Since the Bell Curve, even before.

01:05:07.860 | This really became incendiary in 1969

01:05:12.420 | with an article published by an educational psychologist

01:05:15.720 | named Arthur Jensen.

01:05:17.720 | Let's just take a minute and go back to that

01:05:20.480 | to see the Bell Curve

01:05:21.620 | in a little bit more historical perspective.

01:05:25.000 | Arthur Jensen was a educational psychologist at UC Berkeley.

01:05:29.300 | I knew him as well.

01:05:31.440 | And in 1969 or '68, the Harvard Educational Review

01:05:37.780 | asked him to do a review article

01:05:42.620 | on the early childhood education programs

01:05:47.540 | that were designed to raise the IQs of minority students.

01:05:52.540 | This was before the federally funded Head Start program.

01:05:58.540 | Head Start had not really gotten underway

01:06:01.280 | at the time Jensen undertook his review

01:06:04.420 | of what were a number of demonstration programs.

01:06:07.780 | And these demonstration programs were for young children

01:06:13.660 | who were around kindergarten age.

01:06:15.740 | And they were specially designed

01:06:17.720 | to be cognitively stimulating, to provide lunches,

01:06:22.720 | do all the things that people thought would minimize

01:06:28.020 | this average gap of intelligence tests.

01:06:31.780 | There was a strong belief among virtually all psychologists

01:06:36.780 | that the cause of the gap was unequal opportunity

01:06:40.780 | due to racism, due to all negative things in the society.

01:06:45.620 | And if you could compensate for this,

01:06:49.500 | the gap would go away.

01:06:51.220 | So early childhood education back then

01:06:53.940 | was called literally compensatory education.

01:06:57.280 | Jensen looked at these programs.

01:07:00.900 | He was an empirical guy.

01:07:02.180 | He understood psychometrics.

01:07:04.580 | And he wrote a, it was over a hundred page article

01:07:08.780 | detailing these programs

01:07:12.120 | and the flaws in their research design.

01:07:15.380 | Some of the programs reported IQ gains

01:07:17.660 | of on average five points,

01:07:20.380 | but a few reported 10, 20, and even 30 point gains.

01:07:24.720 | One was called the miracle in Milwaukee.

01:07:27.780 | That investigator went to jail ultimately

01:07:30.460 | for fabricating data.

01:07:32.080 | But the point is that Jensen wrote an article that said,

01:07:36.020 | look, the opening sentence of his article is classic.

01:07:40.020 | The opening sentence is, I may not quote it exactly right,

01:07:43.500 | but it's, we have tried compensatory education

01:07:47.100 | and it has failed.

01:07:48.420 | And he showed that these gains were essentially nothing.

01:07:54.540 | You couldn't really document empirically any gains at all

01:07:59.380 | from these really earnest efforts to increase IQ.

01:08:04.100 | But he went a step further, a fateful step further.

01:08:08.780 | He said, not only have these efforts failed,

01:08:12.460 | but because they have had essentially no impact,

01:08:15.940 | we have to re-examine our assumption

01:08:18.460 | that these differences are caused by environmental things

01:08:22.100 | that we can address with education.

01:08:24.420 | We need to consider a genetic influence,

01:08:28.500 | whether there's a genetic influence

01:08:30.580 | on this group difference.

01:08:32.420 | - So you said that this is one of the more controversial

01:08:35.300 | works ever in science. - I think it's the most

01:08:36.940 | infamous paper in all of psychology, I would go on to say.

01:08:41.660 | Because in 1969, the genetic data was very skimpy

01:08:46.660 | on this question, skimpy and controversial.

01:08:49.660 | It's always been controversial,

01:08:50.820 | but it was even skimpy and controversial.

01:08:53.820 | It's kind of a long story that I go into a little bit

01:08:56.500 | in more detail in the book, "Neuroscience of Intelligence."

01:09:00.980 | But to say he was vilified is an understatement.

01:09:06.320 | I mean, he couldn't talk at the American

01:09:08.900 | Psychological Association without bomb threats

01:09:13.140 | clearing the lecture hall.

01:09:15.260 | Campus security watched him all the time.

01:09:18.100 | They opened his mail.

01:09:20.300 | He had to retreat to a different address.

01:09:23.740 | This was one of the earliest kinds,

01:09:28.740 | this is before the internet,

01:09:30.500 | and kind of internet social media mobs,

01:09:35.100 | but it was that intense.

01:09:38.100 | And I have written that overnight,

01:09:42.740 | after the publication of this article,

01:09:45.940 | all intelligence research became radioactive.

01:09:49.740 | Nobody wanted to talk about it.

01:09:51.780 | And then it didn't,

01:09:55.020 | nobody was doing more research.

01:09:58.900 | And then the bell curve came along,

01:10:02.040 | and the Jensen controversy was dying down.

01:10:05.700 | I have stories that Jensen told me about

01:10:08.100 | his interaction with the Nixon White House on this issue.

01:10:10.860 | I mean, this was like a really big deal.

01:10:14.340 | It was some unbelievable stories,

01:10:16.180 | but he told me this,

01:10:17.860 | so I kind of believe these stories.

01:10:20.100 | Nonetheless--

01:10:21.540 | - 25 years later.

01:10:22.940 | - 25 years later.

01:10:23.900 | - So all this silence basically saying,

01:10:27.040 | nobody wants to do this kind of research.

01:10:32.300 | There's so much pressure,

01:10:33.780 | so much attack against this kind of research.

01:10:36.340 | And here's sort of a bold, stupid, crazy people

01:10:41.340 | that decide to dive right back in.

01:10:44.940 | I wonder how much discussion there was.

01:10:46.540 | Do we include this chapter or not?

01:10:48.740 | - Murray has said they discussed it,

01:10:51.140 | and they felt they should include it,

01:10:54.700 | and they were very careful in the way they wrote it,

01:10:59.620 | which did them no good.

01:11:01.080 | So as a matter of fact, when the bell curve came out,

01:11:06.540 | it was so controversial.

01:11:08.480 | I got a call from a television show called Nightline.

01:11:13.300 | It was with a broadcaster called Ted Koppel,

01:11:16.820 | who had this evening show, I think it was on late at night,

01:11:20.620 | talked about news.

01:11:21.600 | It was a straight up news thing.

01:11:24.260 | And a producer called and asked if I would be on it

01:11:28.180 | to talk about the bell curve.

01:11:31.700 | And I said, you know,

01:11:32.900 | she asked me what I thought about the bell curve as a book.

01:11:36.660 | And I said, look, it's a very good book.

01:11:38.580 | It talks about the role of intelligence in society.

01:11:43.100 | And she said, no, no,

01:11:44.380 | what do you think about the chapter on race?

01:11:47.180 | That's what we want you to talk about.

01:11:49.100 | I remember this conversation.

01:11:52.460 | I said, well,

01:11:53.460 | she said, what would you say if you were on TV?

01:11:58.840 | And I said, well, what I would say is that

01:12:02.300 | it's not at all clear

01:12:04.620 | if there's any genetic component to intelligence,

01:12:12.100 | any differences,

01:12:13.620 | but if there were a strong genetic component,

01:12:17.460 | that would be a good thing.

01:12:19.040 | And, you know, complete silence

01:12:23.180 | on the other end of the phone.

01:12:25.300 | And she said, well, what do you mean?

01:12:28.180 | And I said, well, if it's the more genetic any difference is

01:12:32.580 | the more it's biological.

01:12:35.100 | And if it's biological, we can figure out how to fix it.

01:12:39.940 | - I see, that's interesting.

01:12:41.500 | She said, would you say that on television?

01:12:43.740 | - Yes. - I said, no.

01:12:44.980 | (laughing)

01:12:45.820 | And so that was the end of that.

01:12:47.420 | - So that's for more like biology is

01:12:51.940 | within the reach of science

01:12:58.260 | and the environment is a public policy,

01:13:02.020 | social and all those kinds of things.

01:13:04.080 | From your perspective,

01:13:06.900 | whichever one you think is more amenable to solutions

01:13:10.740 | in the short term is the one that excites you.

01:13:13.540 | But you saying that is good,

01:13:16.680 | the truth of genetic differences, no matter what,

01:13:24.060 | between groups is a painful,

01:13:29.060 | harmful, potentially dangerous thing.

01:13:35.660 | Let me ask you to this question,

01:13:38.060 | whether it's bell curve or any research on race differences,

01:13:42.200 | can that be used to increase the amount of racism

01:13:48.300 | in the world?

01:13:49.420 | Can that be used to increase the amount of hate

01:13:51.700 | in the world?

01:13:52.820 | Do you think about this kind of stuff?

01:13:54.420 | - I've thought about this a lot, not as a scientist,

01:13:57.400 | but as a person.

01:13:58.540 | And my sense is there is such enormous risk

01:14:05.940 | enormous reservoirs of hate and racism

01:14:10.940 | that have nothing to do with scientific knowledge

01:14:14.940 | of the data that speak against that.

01:14:19.180 | That no, I don't wanna give racist groups of veto power

01:14:24.180 | over what scientists study.

01:14:27.700 | If you think that the differences,

01:14:30.580 | and by the way, virtually no one disagrees

01:14:33.260 | that there are differences in scores.

01:14:35.460 | It's all about what causes them and how to fix it.

01:14:38.540 | So if you think this is a cultural problem,

01:14:42.920 | then you must ask the problem,

01:14:44.900 | do you want to change anything about the culture?

01:14:49.380 | Or are you okay with the culture?

01:14:51.800 | 'Cause you don't feel it's appropriate

01:14:53.380 | to change a person's culture.

01:14:55.600 | So are you okay with that?

01:14:57.140 | And the fact that that may lead to disadvantages

01:14:59.700 | in school achievement?

01:15:01.580 | It's a question.

01:15:03.540 | If you think it's environmental,

01:15:05.940 | what are the environmental parameters that can be fixed?

01:15:10.380 | I'll tell you one, lead from gasoline in the atmosphere,

01:15:15.380 | lead in paint, lead in water.

01:15:18.780 | That's an environmental toxin

01:15:20.820 | that society has the means to eliminate, and they should.

01:15:25.820 | - Yeah, just to sort of trying to find some insight

01:15:30.260 | and conclusion to this very difficult topic.

01:15:33.580 | Is there been research on environment versus genetics,

01:15:38.580 | nature versus nurture on this question of race differences?

01:15:41.740 | - There is not, no one wants to do this research.

01:15:46.380 | First of all, it's hard research to do.

01:15:48.060 | Second of all, it's a minefield.

01:15:50.620 | No one wants to spend their career on it.

01:15:52.940 | Tenured people don't want to do it, let alone students.

01:15:56.320 | The way I talk about it,

01:15:59.580 | well, before I tell you the way I talk about it,

01:16:02.440 | I want to say one more thing about Jensen.

01:16:04.540 | He was once asked by a journalist straight out,

01:16:08.020 | "Are you a racist?"

01:16:09.480 | His answer was very interesting.

01:16:12.500 | His answer was, "I've thought about that a lot,

01:16:15.800 | "and I've concluded it doesn't matter."

01:16:19.680 | Now, I know what he meant by this.

01:16:23.660 | - The guts to say that, wow.

01:16:25.820 | - He was a very unusual person.

01:16:27.560 | I think he had a touch of Asperger's syndrome,

01:16:29.920 | to tell you the truth,

01:16:30.940 | because I saw him in many circumstances.

01:16:34.140 | - He would be canceled on Twitter

01:16:35.560 | immediately with that sentence.

01:16:37.020 | - Yeah, but what he meant was he had a hypothesis.

01:16:41.060 | And with respect to group differences,

01:16:44.920 | he called it the default hypothesis.

01:16:47.420 | He said whatever factors affect individual intelligence

01:16:51.220 | are likely the same factors that affect group differences.

01:16:54.340 | It was the default, but it was a hypothesis.

01:16:58.260 | It should be tested.

01:16:59.540 | And if it turned out empirical tests

01:17:02.060 | didn't support the hypothesis,

01:17:03.700 | he was happy to move on to something else.

01:17:06.500 | He was absolutely committed to that scientific ideal.

01:17:11.500 | It's an empirical question.

01:17:16.060 | We should look at it, and let's see what happens.

01:17:18.820 | - The scientific method cannot be racist,

01:17:22.100 | from his perspective.

01:17:23.500 | It doesn't matter what the scientists,

01:17:27.340 | if they follow the scientific method,

01:17:30.420 | it doesn't matter what they believe.

01:17:32.060 | - And if they are biased,

01:17:33.780 | and they consciously or unconsciously bias the data,

01:17:38.780 | other people will come along to replicate it.

01:17:42.700 | They will fail, and the process over time will work.

01:17:47.700 | - So let me push back on this idea,

01:17:50.980 | because psychology to me is full of gray areas.

01:17:57.360 | And what I've observed about psychology,

01:18:00.600 | even replication crisis aside,

01:18:04.040 | is that something about the media,

01:18:06.200 | something about journalism,

01:18:08.000 | something about the virality of ideas in the public sphere,

01:18:13.000 | they misinterpret.

01:18:16.040 | They take up things from studies willfully,

01:18:20.400 | or from ignorance, misinterpret findings,

01:18:23.680 | and tell narratives around that.

01:18:26.760 | I personally believe, for me,

01:18:29.840 | I'm not saying that broadly about science,

01:18:31.520 | but for me, it's my responsibility to anticipate

01:18:35.760 | the ways in which findings will be misinterpreted.

01:18:40.440 | So I've thought about this a lot,

01:18:42.840 | 'cause I publish papers on semi-autonomous vehicles,

01:18:47.020 | and those, you know, cars, people dying cars.

01:18:52.960 | There's people that have written me letters saying,

01:18:56.360 | emails, nobody writes letters, I wish they did,

01:18:59.120 | that have blood on my hands,

01:19:01.800 | because of things that I would say,

01:19:03.800 | positive or negative, there's consequences.

01:19:06.200 | In the same way, when you're a researcher of intelligence,

01:19:09.280 | I'm sure you might get emails,

01:19:12.040 | or at least people might believe that a finding

01:19:15.440 | of your study is going to be used

01:19:17.880 | by a large number of people

01:19:19.080 | to increase the amount of hate in the world.

01:19:22.640 | I think there's some responsibility on scientists,

01:19:26.080 | but for me, I think there's a great responsibility

01:19:29.200 | to anticipate the ways things will be misinterpreted,

01:19:35.440 | and there, you have to, first of all,

01:19:37.880 | decide whether you want to say a thing at all,

01:19:41.000 | do the study at all, publish the study at all,

01:19:43.480 | and two, the words with which you explain it.

01:19:47.220 | I find this on Twitter a lot, actually,

01:19:50.680 | which is, when I write a tweet,

01:19:53.240 | and I'm usually just doing so innocently,

01:19:55.340 | I'll write it, it takes me five seconds to write it,

01:20:01.360 | or whatever, 30 seconds to write it,

01:20:04.360 | and then I'll think, all right,

01:20:05.880 | I close my eyes open and try to see

01:20:08.960 | how will the world interpret this,

01:20:11.720 | what are the ways in which this will be misinterpreted?

01:20:14.400 | And I'll sometimes adjust that tweet to see,

01:20:18.480 | yeah, so in my mind, it's clear,

01:20:20.920 | but that's because it's my mind from which this tweet came,

01:20:24.040 | but you have to think, in a fresh mind that sees this,

01:20:26.900 | and it's spread across a large number of other minds,

01:20:32.840 | how will the interpretation morph?

01:20:36.040 | I mean, for a tweet, it's a silly thing, it doesn't matter,

01:20:38.240 | but for a scientific paper and study and finding,

01:20:43.240 | I think it matters, so I don't know.

01:20:47.880 | I don't know what your thoughts about that,

01:20:49.720 | 'cause maybe for Jensen, the data's there,

01:20:53.740 | what do you want me to do?

01:20:55.600 | This is a scientific process that's been carried out,

01:20:59.160 | if you think the data was polluted by bias,

01:21:02.200 | do other studies that reveal the bias,

01:21:04.640 | but the data's there.

01:21:07.240 | I'm not a poet, I'm not a literary writer,

01:21:14.680 | what do you want me to do?

01:21:15.640 | I'm just presenting you the data.

01:21:17.520 | What do you think on that spectrum?

01:21:19.320 | What's the role of a scientist?

01:21:21.080 | - The reason I do podcasts,

01:21:23.500 | the reason I write books for the public

01:21:27.240 | is to explain what I think the data mean

01:21:30.400 | and what I think the data don't mean.

01:21:32.640 | I don't do very much on Twitter

01:21:36.100 | other than to retweet references to papers.

01:21:39.920 | I don't think it's my role to explain these,

01:21:42.480 | 'cause they're complicated, they're nuanced,

01:21:46.440 | but when you decide not to do a scientific study

01:21:51.440 | or not to publish a result

01:21:54.400 | because you're afraid the result

01:21:56.160 | could be harmful or insensitive,

01:22:01.880 | that's not an unreasonable thought,

01:22:04.100 | and people will make different conclusions

01:22:09.240 | and decisions about that.

01:22:11.600 | I wrote about this,

01:22:13.880 | I'm the editor of a journal called Intelligence,

01:22:17.040 | which publishes scientific papers.

01:22:20.600 | Sometimes we publish papers on group differences.

01:22:24.240 | Those papers sometimes are controversial.

01:22:27.080 | These papers are written for a scientific audience,

01:22:29.640 | they're not written for the Twitter audience,

01:22:32.200 | so I don't promote them very much on Twitter,

01:22:35.620 | but in a scientific paper,

01:22:41.280 | you have to now choose your words carefully

01:22:43.800 | also because those papers are picked up by non-scientists,

01:22:48.800 | by writers of various kinds,

01:22:52.040 | and you have to be available to discuss what you're saying

01:22:56.160 | and what you're not saying.

01:22:57.800 | Sometimes you are successful at having a good conversation,

01:23:03.720 | like we are today, that doesn't start out pejorative.

01:23:09.200 | Other times I have been asked to participate in debates

01:23:12.520 | where my role would be to justify race science.

01:23:16.480 | Well, you can see, you start out,

01:23:19.520 | and that was a BBC request that I received.

01:23:25.640 | - I have so much, it's a love-hate relationship,

01:23:28.240 | mostly hate, with these shallow journalism organizations,

01:23:32.960 | so they would want to use you

01:23:35.120 | as a kind of, in a debate setting,

01:23:38.080 | to communicate as to, like,

01:23:39.800 | there is race differences between groups,

01:23:42.280 | and make that into debate,

01:23:44.160 | and put you in a role of--

01:23:46.440 | - Justifying racism.

01:23:49.280 | - Justifying racism. - That's what

01:23:50.120 | they're asking me to do.

01:23:51.120 | - Versus, like, educating about this field

01:23:54.320 | of the science of intelligence, yeah.

01:23:55.840 | - I wanna say one more thing

01:23:57.960 | before we get off the normal distribution.

01:24:01.280 | You also asked me,

01:24:02.480 | what is the science after the bell curve?

01:24:06.160 | And the short answer is there's not much new work,

01:24:09.600 | but whatever work there is supports the idea

01:24:13.360 | that there still are group differences.

01:24:16.240 | It's arguable whether those differences

01:24:18.520 | have diminished at all or not,

01:24:20.640 | and there is still a major problem

01:24:24.920 | in underperformance for school achievement

01:24:28.240 | for many disadvantaged and minority students,

01:24:33.400 | and there so far is no way to fix it.

01:24:37.320 | What do we do with this information?

01:24:39.480 | Is this now a task?

01:24:42.120 | Now, we'll talk about the future

01:24:44.000 | on the neuroscience and the biology side,

01:24:47.760 | but in terms of this information as a society

01:24:51.200 | in the public policy, in the political space,

01:24:53.560 | in the social space, what do we do with this information?

01:24:56.320 | - I've thought a lot about this.

01:24:57.960 | The first step is to have people interested in policy

01:25:02.960 | understand what the data actually show

01:25:07.000 | to pay attention to intelligence data.

01:25:09.880 | You can read policy papers about education

01:25:13.520 | and using your word processor,

01:25:15.920 | you can search for the word intelligence.

01:25:17.960 | You can search a 20,000 word document in a second

01:25:22.200 | and find out the word intelligence does not appear anywhere.

01:25:26.880 | In most discussions about what to do about achievement gaps,

01:25:31.880 | I'm not talking about test gaps,

01:25:33.720 | I'm talking about actual achievement gaps in schools,

01:25:37.400 | which everyone agrees is a problem.

01:25:40.200 | The word intelligence doesn't appear among educators.

01:25:44.000 | - That's fascinating.

01:25:44.840 | - As a matter of fact, in California,

01:25:47.680 | there has been tremendous controversy

01:25:50.080 | about recent attempts to revise the curriculum for math

01:25:54.440 | in high schools, and we had a Stanford professor

01:25:58.320 | of education who was running this review assert

01:26:02.880 | there's no such thing as talent, mathematical talent.

01:26:06.680 | And she wanted to get rid of the advanced classes in math

01:26:12.360 | because not everyone could do that.

01:26:15.400 | Now, of course, this has been very controversial,

01:26:17.320 | they've retreated somewhat,

01:26:19.520 | but the idea that a university professor

01:26:21.720 | was in charge of this who believes

01:26:23.600 | that there's no talent, that it doesn't exist,

01:26:31.920 | this is rather shocking,

01:26:33.600 | let alone the complete absence of intelligence data.

01:26:37.520 | By the way, let me tell you something

01:26:39.040 | about what the intelligence data show.

01:26:42.040 | Let's take race out of it.

01:26:43.640 | Even though the origins of these studies

01:26:48.320 | were a long time ago,

01:26:50.960 | I'm blocking on the name of the report,

01:26:53.960 | the Coleman Report was a famous report about education,

01:26:57.760 | and they measured all kinds of variables about schools,

01:27:01.960 | about teachers, and they looked at academic achievement

01:27:06.320 | as an outcome.

01:27:08.200 | And they found the most predictive variables

01:27:12.720 | of education outcome were the variables the student brought

01:27:17.680 | with him or her into the school,

01:27:20.440 | essentially their ability.

01:27:22.000 | And that when you combine the school

01:27:26.480 | and the teacher variables together,

01:27:29.520 | the quality of the school, the funding of the school,

01:27:31.960 | the quality of the teachers, their education,

01:27:34.720 | you put all the teacher and school variables together,

01:27:37.600 | it barely accounted for 10% of the variance.

01:27:40.300 | And this has been replicated now.

01:27:43.600 | So the best research we have shows that school variables

01:27:50.400 | and teacher variables together account

01:27:54.800 | for about 10% of student academic achievement.

01:27:58.200 | Now, you wanna have some policy

01:28:02.120 | on improving academic achievement,

01:28:04.840 | how much money do you wanna put into teacher education?

01:28:08.400 | How much money do you wanna put into the quality

01:28:11.720 | of the school administration?

01:28:14.280 | You know who you can ask?

01:28:15.420 | You can ask the Gates Foundation,

01:28:17.280 | because they spent a tremendous amount of money doing that.

01:28:21.600 | And at the end of it, because they're measurement people,

01:28:25.160 | they wanna know the data,

01:28:27.600 | they found it had no impact at all.

01:28:29.680 | And they've kind of pulled out of that kind of program.

01:28:33.480 | - So, oh boy.

01:28:35.020 | Let me ask you, this is me talking, but there's--

01:28:41.240 | - Just the two of us.

01:28:42.680 | - Just the two of us, but I'm gonna say

01:28:44.920 | some funny and ridiculous things,

01:28:46.680 | so you surely are not approving of it.

01:28:51.280 | But there's a movie called Clerks.

01:28:53.280 | - I've seen it, I've seen it, yeah.

01:28:56.400 | - There's a funny scene in there

01:28:58.080 | where a lovely couple are talking about

01:29:01.040 | the number of previous sexual partners they had.

01:29:03.960 | And the woman says that, I believe she just had a handful,

01:29:08.960 | like two or three or something like that, sexual partners,

01:29:12.640 | but then she also mentioned that she,

01:29:16.760 | what's that called, fallatio, what's the scientific,

01:29:20.400 | but she went, you know, gave a blowjob,

01:29:23.120 | to 37 guys, I believe it is.

01:29:26.240 | And so that has to do with the truth.

01:29:30.000 | So sometimes, knowing the truth

01:29:34.440 | can get in the way of a successful relationship of love

01:29:40.320 | of some of the human flourishing.

01:29:42.080 | And that seems to me that's at the core here,

01:29:46.380 | that facing some kind of truth that's not able to be changed

01:29:51.380 | makes it difficult to sort of,

01:29:56.100 | is limiting as opposed to empowering.

01:30:00.940 | That's the concern.

01:30:02.300 | If you sort of test for intelligence and lay the data out,

01:30:07.220 | it feels like you will give up on certain people.

01:30:10.220 | You will sort of start binning people,

01:30:13.540 | it's like, well, this person is like,

01:30:16.580 | let's focus on the average people,

01:30:19.060 | or let's focus on the very intelligent people.

01:30:20.980 | That's the concern.

01:30:22.540 | And there's a kind of intuition

01:30:26.700 | that if we just don't measure,

01:30:29.580 | and we don't use that data,

01:30:31.660 | that we would treat everybody equal

01:30:33.620 | and give everybody equal opportunity.

01:30:37.860 | If we have the data in front of us,

01:30:39.680 | we're likely to misdistribute the amount

01:30:44.680 | of sort of attention we allocate,

01:30:46.360 | resources we allocate to people.

01:30:49.400 | That's probably the concern.

01:30:52.160 | - It's a realistic concern,

01:30:55.080 | but I think it's a misplaced concern

01:30:57.660 | if you wanna fix the problem.

01:31:00.560 | If you wanna fix the problem,

01:31:02.040 | you have to know what the problem is.

01:31:03.960 | - Yep.

01:31:05.160 | - Now, let me tell you this.

01:31:06.800 | Let's go back to the bell curve,

01:31:08.880 | not the bell curve, but the normal distribution.

01:31:11.320 | - Yes.

01:31:12.400 | - 16% of the population on average has an IQ under 85,

01:31:17.400 | which means they're very hard.

01:31:22.080 | If you have an IQ under 85,

01:31:24.200 | it's very hard to find gainful employment

01:31:26.760 | at a salary that sustains you,

01:31:31.240 | at least minimally, in modern life.

01:31:34.800 | Okay?

01:31:35.640 | Not impossible, but it's very difficult.

01:31:38.520 | 16% of the population of the United States

01:31:42.400 | is about 51 or 52 million people with IQs under 85.

01:31:47.400 | This is not a small issue.

01:31:52.520 | 14 million children have IQs under 85.

01:31:55.860 | Is this something we wanna ignore?

01:32:00.720 | Does this have any, what is the Venn diagram between,

01:32:04.440 | you know, when you have people with IQs under 85

01:32:08.200 | and you have achievement in school or achievement in life?

01:32:13.140 | There's a lot of overlap there.

01:32:16.840 | This is why, to go back to the IQ pill,

01:32:19.680 | if there were a way to shift that curve

01:32:24.200 | toward the higher end, that would have a big impact.

01:32:31.520 | - If I could maybe, before we talk about

01:32:34.600 | the impact on life and so on,

01:32:37.360 | some of the criticisms of the bell curve.

01:32:39.160 | So Stephen Jay Gould wrote that the bell curve

01:32:42.520 | rests on four incorrect assumptions.

01:32:45.400 | It would be just interesting to get your thoughts

01:32:47.480 | on the four assumptions, which are

01:32:49.520 | intelligence must be reducible to a single number,

01:32:52.520 | intelligence must be capable of rank ordering people

01:32:55.280 | in a linear order,

01:32:56.640 | intelligence must be primarily genetically based,

01:33:00.400 | and intelligence must be essentially immutable.

01:33:04.320 | Maybe not as criticisms, but as thoughts about intelligence.

01:33:09.320 | - Yeah, we could spend a lot of time on him.

01:33:13.600 | - On Stephen Jay Gould?

01:33:14.640 | - Yes. - Yeah.

01:33:15.640 | - He wrote that in what, about 1985, 1984?

01:33:20.640 | His views were overtly political, not scientific.

01:33:26.000 | He was a scientist, but his views on this

01:33:28.600 | were overtly political, and I would encourage

01:33:32.520 | people listening to this, if they really wanna understand

01:33:36.400 | his criticisms, they should just Google

01:33:41.400 | what he had to say, and Google the scientific reviews

01:33:47.400 | of his book, "The Mismeasure of Man,"

01:33:51.440 | and they will take these statements apart.

01:33:54.680 | They were wrong, not only were they wrong,

01:33:57.880 | but when he asserted in his first book

01:34:00.320 | that there was no biological basis, essentially, to IQ,

01:34:05.320 | by the time the second edition came around,

01:34:08.200 | there were studies of MRIs showing that brain size,

01:34:13.200 | brain volume, were correlated to IQ scores,

01:34:16.760 | which he declined to put in his book. (laughs)

01:34:19.880 | - So, okay, I'm learning a lot today.

01:34:21.800 | I didn't know, actually, the extent of his work.

01:34:25.720 | I was just using a few little snippets of criticism.

01:34:28.860 | - That's interesting, so there's a battle here.

01:34:30.680 | He wrote a book, "Mismeasure of Man,"

01:34:32.640 | that's missing a lot of the scientific grounding.

01:34:36.480 | - His book is highly popular in colleges today.

01:34:39.640 | You can find it in any college bookstore

01:34:41.760 | under assigned reading.

01:34:43.400 | It's highly popular. - "The Mismeasure of Man"?

01:34:45.120 | - Yes, highly influential.

01:34:46.720 | - Can you speak to "The Mismeasure of Man"?

01:34:48.600 | I'm undereducated about this, so what,

01:34:50.720 | is this the book basically criticizing

01:34:53.240 | the ideas in the book? - Yeah, yeah,

01:34:55.040 | where those four things came from.

01:34:57.240 | And it is really a book that was really taken apart

01:35:02.240 | point by point by a number of people

01:35:05.800 | who actually understood the data.

01:35:07.960 | And he didn't care.

01:35:09.760 | He didn't care, he didn't modify anything.

01:35:12.480 | - Listen, because this is such a sensitive topic,

01:35:16.560 | like I said, I believe

01:35:18.320 | the impact of the work, as it is misinterpreted,

01:35:26.000 | has to be considered,

01:35:28.160 | because it's not just going to be scientific discourse,

01:35:31.320 | it's going to be political discourse,

01:35:32.840 | there's going to be debates,

01:35:34.560 | there's going to be politically motivated people

01:35:39.440 | that will use messages in each direction,

01:35:41.640 | make something like the bell curve the enemy

01:35:45.240 | or the support for one's racist beliefs.

01:35:50.240 | And so I think you have to consider that,

01:35:55.120 | but it's difficult because

01:35:56.640 | Nietzsche was used by Hitler to justify

01:36:01.440 | a lot of his beliefs,

01:36:02.320 | and it's not exactly on Nietzsche to anticipate Hitler,

01:36:07.320 | or how his ideas will be misinterpreted and used for evil.

01:36:12.520 | But there's a balance there.

01:36:14.560 | So I understand, this is really interesting,

01:36:16.280 | I didn't know, is there any criticism

01:36:18.960 | of the book you find compelling or interesting

01:36:21.400 | or challenging to use from a scientific perspective?

01:36:23.640 | - There were factual criticisms

01:36:25.880 | about the nature of the statistics that were used,

01:36:30.880 | the statistical analyses,

01:36:32.600 | these are more technical criticisms,

01:36:34.640 | and they were addressed by Murray in a couple of articles

01:36:38.160 | where he took all the criticisms and spoke to them.

01:36:41.760 | And people listening to this podcast

01:36:44.840 | can certainly find all those online.

01:36:46.920 | And it's very interesting.

01:36:48.880 | But Murray went on to write some additional books,

01:36:52.800 | two in the last couple of years.

01:36:54.800 | One about human diversity,

01:36:57.880 | where he goes through the data,

01:37:00.240 | refuting the idea that race is only a social construct

01:37:05.240 | with no biological meaning.

01:37:07.960 | He discusses the data, it's a very good discussion,

01:37:11.080 | you don't have to agree with it,

01:37:12.560 | but he presents data in a cogent way,

01:37:16.440 | and he talks about the critics of that,

01:37:19.120 | and he talks about their data

01:37:20.680 | in a cogent, non-personal way.

01:37:23.480 | It's a very informative discussion.

01:37:26.920 | The book is called "Human Diversity."

01:37:28.920 | He talks about race and he talks about gender, same thing,

01:37:32.320 | about sex differences.

01:37:33.720 | And more recently, he's written

01:37:36.960 | what might be his final say on this,

01:37:39.600 | a book called "Facing Reality,"

01:37:41.640 | where he talks about this again.

01:37:44.920 | So, you know, he can certainly defend himself.

01:37:49.640 | He doesn't need me to do that.

01:37:52.120 | But I would urge people who have heard

01:37:55.040 | about him and the bell curve,

01:37:58.160 | and who think they know what's in it,

01:38:00.280 | you are likely incorrect,

01:38:03.480 | and you need to read it for yourself.

01:38:06.320 | - But it is, scientifically,

01:38:09.160 | it's a serious subject, it's a difficult subject.

01:38:13.560 | Ethically, it's a difficult subject.

01:38:16.680 | Everything you said here calmly and thoughtfully

01:38:19.800 | is difficult.

01:38:20.720 | It's difficult for me to even consider

01:38:23.200 | that G factor exists.

01:38:24.760 | I don't mean from like,

01:38:27.880 | that somehow G factor is inherently racist

01:38:30.440 | or sexist or whatever.

01:38:32.280 | It's just, it's difficult in the way that

01:38:35.520 | considering the fact that we die one day is difficult.

01:38:38.240 | That we are limited by our biology is difficult.

01:38:42.880 | And it's,

01:38:46.320 | at least from an American perspective,

01:38:47.800 | you like to believe that everything is possible

01:38:49.960 | in this world.

01:38:51.440 | - Well, that leads us to what I think

01:38:55.440 | we should do with this information.

01:38:58.240 | And what I think we should do with this information

01:39:03.320 | is unusual.

01:39:06.360 | - Uh-oh.

01:39:07.720 | - Because I think what we need to do

01:39:09.560 | is fund more neuroscience research

01:39:12.520 | on the molecular biology of learning and memory.

01:39:15.760 | Because one definition of intelligence

01:39:20.600 | is based on how much you can learn

01:39:24.640 | and how much you can remember.

01:39:26.600 | - Yes.

01:39:27.440 | - And if you accept that definition of intelligence,

01:39:30.920 | then there are molecular studies going on now,

01:39:35.600 | and Nobel Prizes being won on molecular biology

01:39:40.840 | or molecular neurobiology of learning and memory.

01:39:44.240 | Now, the step those researchers,

01:39:48.640 | those scientists need to take when it comes to intelligence

01:39:53.200 | is to focus on the concept of individual differences.

01:39:58.200 | Intelligence research has individual differences

01:40:03.240 | as its heart,

01:40:04.840 | because it assumes that people differ on this variable

01:40:10.600 | and those differences are meaningful

01:40:13.040 | and need understanding.

01:40:15.960 | Cognitive psychologists who have morphed

01:40:19.400 | into molecular biologists studying learning and memory

01:40:23.400 | hate the concept of individual differences historically.

01:40:27.440 | Some now are coming around to it.

01:40:29.240 | I once sat next to a Nobel Prize winner

01:40:34.760 | for his work on memory.

01:40:37.840 | And I asked him about individual differences.

01:40:41.320 | And he said, "Don't go there.

01:40:42.760 | "It'll set us back 50 years."

01:40:44.700 | But I said, "Don't you think they're the key, though,

01:40:49.500 | "to understand, you know,

01:40:50.400 | "why can some people remember more than others?"

01:40:53.960 | He said, "You don't wanna go there."

01:40:55.960 | - I think the 21st century will be remembered

01:40:58.760 | by the technology and the science

01:41:00.440 | that goes to individual differences.

01:41:04.120 | Because we have now data,

01:41:05.560 | we have now the tools to much, much better

01:41:07.400 | to start to measure, start to estimate,

01:41:10.120 | not just on the sort of through tests

01:41:12.080 | and IQ test type of things,

01:41:14.600 | sort of outside the body kind of things,

01:41:18.240 | but measuring all kinds of stuff about the body.

01:41:20.320 | So yeah, truly go into the molecular biology,

01:41:23.160 | to the neurobiology, to the neuroscience.

01:41:27.000 | Let me ask you about life.

01:41:30.360 | (both laugh)

01:41:31.960 | How does intelligence correlate with or lead to

01:41:36.760 | or has anything to do with career success?

01:41:39.040 | You've mentioned these kinds of things.

01:41:41.040 | Is there any data, you've had an excellent conversation

01:41:44.800 | with Jordan Peterson, for example.

01:41:46.720 | Is there any data on what intelligent means

01:41:50.320 | for success in life?

01:41:53.440 | - Success in life, there is a tremendous amount

01:41:57.520 | of validity data that looked at intelligence test scores

01:42:05.760 | and various measures of life success.

01:42:10.760 | Now, of course, life success is a pretty broad topic

01:42:16.040 | and not everybody agrees on what success means,

01:42:22.160 | but there's general agreement on certain aspects of success

01:42:27.160 | that can be measured.

01:42:30.120 | And those--

01:42:33.360 | - Including life expectancy, like you said.

01:42:35.080 | - Life expectancy, now there's life success.

01:42:38.240 | Life expectancy, I mean, that is such an interesting finding

01:42:47.080 | but IQ scores are also correlated to things like income.

01:42:52.480 | Now, okay, so who thinks income means you're successful?

01:42:59.120 | That's not the point.

01:43:01.400 | The point is that income is one empirical measure

01:43:06.400 | in this culture that says something

01:43:09.360 | about your level of success.

01:43:11.560 | You can define success in ways

01:43:13.520 | that have nothing to do with income.

01:43:16.080 | You can define success based

01:43:18.640 | on your evolutionary natural selection success.

01:43:23.480 | But for variables, and even that by the way

01:43:29.840 | is correlated to IQ in some studies.

01:43:34.000 | So however you wanna define success, IQ is important.

01:43:39.000 | It's not the only determinant.

01:43:44.320 | People get hung up on, well, what about personality?

01:43:46.840 | What about so-called emotional intelligence?

01:43:49.400 | Yes, all those things matter.

01:43:52.080 | The thing that matters empirically,

01:43:54.520 | the single thing that matters the most

01:43:56.440 | is your general ability,

01:43:59.280 | your general mental intellectual ability,

01:44:01.800 | your reasoning ability.

01:44:03.680 | And the more complex your vocation,

01:44:07.200 | the more complex your job, the more G matters.

01:44:11.600 | G doesn't matter in a lot of occupations

01:44:14.680 | don't require complex thinking.

01:44:17.160 | And there are occupations like that and G doesn't matter.

01:44:21.040 | Within an occupation,

01:44:24.160 | the G might not matter so much.

01:44:28.360 | So that if you look at all the professors at MIT,

01:44:33.280 | and had a way to rank order them on,

01:44:39.840 | there's a ceiling effect is what I'm saying.

01:44:42.560 | - Also, when you get past a certain threshold,

01:44:47.200 | then there's impact on wealth, for example,

01:44:49.920 | or career success.

01:44:51.840 | However, that's defined in each individual discipline,

01:44:54.280 | but after a certain point, it doesn't matter.

01:44:56.800 | - Actually, it does matter in certain things.

01:44:59.320 | So for example, there is a very classic study

01:45:03.280 | that was started at Johns Hopkins

01:45:06.920 | when I was a graduate student there.

01:45:08.800 | I actually worked on this study at the very beginning.

01:45:11.360 | It's the study of mathematically

01:45:12.680 | and scientifically precocious youth.

01:45:15.680 | And they gave junior high school students,

01:45:20.120 | age 11 and 12, the standard SAT math exam.

01:45:25.120 | And they found a very large number of students

01:45:31.440 | scored very high on this exam.

01:45:33.840 | Not a large number.

01:45:35.120 | I mean, they found many students when they cast the net,

01:45:39.120 | they're all a Baltimore.

01:45:40.760 | They found a number of students

01:45:42.560 | who scored as high on the SAT math

01:45:45.280 | when they were 12 years old as incoming Hopkins freshmen.

01:45:50.160 | And they said, "Gee, now this is interesting.

01:45:53.600 | "What shall we do now?"

01:45:56.960 | And on a case-by-case basis,

01:45:59.840 | they got some of those kids

01:46:01.680 | into their local community college math programs.

01:46:05.040 | Many of those kids went on to be very successful.

01:46:10.280 | And now there's a 50-year follow-up of those kids.

01:46:13.880 | And it turns out,

01:46:17.680 | these kids were in the top 1%.

01:46:21.160 | Okay, so everybody in this study is in the top 1%.

01:46:24.800 | If you take that group, that rarified group,

01:46:28.120 | and divide them into quartiles,

01:46:30.560 | so that you have the top 25% of the top 1%,

01:46:35.360 | and the bottom 25% of the top 1%,

01:46:39.800 | you can find on measurable variables of success

01:46:45.800 | variables of success,

01:46:48.040 | the top quartile does better than the bottom quartile

01:46:51.720 | in the top 1%.

01:46:53.800 | They have more patents, they have more publications,

01:46:56.360 | they have more tenure at universities.

01:46:59.480 | And this is based on,

01:47:00.840 | you're dividing them based on their score at age 12.

01:47:05.960 | - I wonder how much interesting data

01:47:10.240 | is in the variability, in the differences.

01:47:12.720 | So, but that's really, oh boy.

01:47:16.480 | That's very interesting, but it's also,

01:47:19.760 | I don't know, somehow painful.

01:47:21.040 | I don't know why it's so painful.

01:47:22.800 | That G-factor's so determinant,

01:47:26.640 | of even in the nuanced top percent.

01:47:30.760 | - Well, this is interesting that you find that painful.

01:47:32.800 | Do you find it painful that people with charisma

01:47:36.800 | can be very successful in life,

01:47:40.720 | even though having no other attributes

01:47:42.720 | other than they're famous and people like them?

01:47:45.760 | Do you find that painful?

01:47:47.400 | - Yes, if that charisma is untrainable.

01:47:51.120 | So one of the things, again,

01:47:53.320 | this is like I learned psychology from the Johnny Depp trial.

01:47:56.760 | But one of the things the psychologist,

01:48:00.840 | the personality psychologist,

01:48:02.320 | he can maybe speak to this,

01:48:03.440 | 'cause he had interest in this for a time,

01:48:07.200 | is she was saying that personality,

01:48:10.660 | technically speaking, is the thing that doesn't change

01:48:15.400 | over a lifetime.

01:48:16.960 | It's the thing you're,

01:48:19.280 | I don't know if she was actually implying

01:48:20.600 | that you're born with it.

01:48:21.680 | - Well, it's a trait.

01:48:22.760 | - It's a trait that's--

01:48:23.600 | - It's a trait that's relatively stable over time.

01:48:27.200 | I think that's generally correct.

01:48:28.960 | - So to the degree your personality is stable over time,

01:48:33.120 | yes, that too is painful.

01:48:36.840 | 'Cause what's not painful is the thing,

01:48:39.600 | if I'm fat and out of shape,

01:48:42.360 | I can exercise and become healthier in that way.

01:48:47.360 | If my diet is a giant mess

01:48:50.640 | and that's resulting in some kind of conditions

01:48:53.880 | that my body's experiencing,

01:48:55.320 | I can fix that by having a better diet.

01:48:58.040 | That's sort of my actions,

01:49:00.480 | my willed actions can make a change.

01:49:03.840 | If charisma is part of the personality

01:49:06.540 | that's the part of the charisma

01:49:09.160 | that is part of the personality that is stable,

01:49:11.880 | yeah, yeah, that's painful too.

01:49:15.400 | 'Cause it's like, oh shit, I'm stuck with this.

01:49:18.400 | I'm stuck with this.

01:49:19.880 | - Well, I mean, and this pretty much generalizes

01:49:22.880 | to every aspect of your being.

01:49:24.980 | This is who you are, you gotta deal with it.

01:49:27.520 | And what it undermines, of course,

01:49:29.360 | is a realistic appreciation for this,

01:49:32.560 | undermines the fairly recent idea

01:49:37.560 | prevalent in this country,

01:49:40.760 | that if you work hard, you can be anything you wanna be.

01:49:44.020 | Which has morphed from the original idea

01:49:47.460 | that if you work hard, you can be successful.

01:49:50.900 | Those are two different things.

01:49:52.840 | - Yeah.

01:49:53.900 | - And now we have,

01:49:55.720 | if you work hard, you can be anything you wanna be.

01:50:00.860 | This is completely unrealistic.

01:50:03.440 | I'm sorry, it just is.

01:50:05.220 | Now you can work hard and be successful,

01:50:06.820 | there's no question.

01:50:08.540 | But you know what, I could work very hard

01:50:11.740 | and I am not going to be a successful

01:50:15.060 | theoretical physicist, I'm just not.

01:50:18.740 | - That said, I mean, we should,

01:50:20.700 | 'cause we had this conversation already,

01:50:22.900 | but it's good to repeat.

01:50:24.420 | The fact that you're not going to be

01:50:27.700 | a theoretical physicist,

01:50:30.140 | is not judgment on your basic humanity.

01:50:32.820 | Returning again to the old man,

01:50:35.580 | which means men and women are created equal.

01:50:39.100 | So again, some of the differences we're talking about

01:50:41.580 | in quote unquote success, wealth,

01:50:44.860 | number of, whether you win a Nobel Prize or not,

01:50:50.620 | that doesn't put a measure on your basic humanity

01:50:55.780 | and basic value and even goodness of you as a human being.

01:51:00.780 | 'Cause that, your basic role and value in society

01:51:06.860 | is largely within your control.

01:51:10.900 | It's some of these measures that we're talking about.

01:51:14.580 | It's good to remember this.

01:51:18.300 | One question about the Flynn effect.

01:51:21.180 | What is it?

01:51:23.260 | Are humans getting smarter over the years,

01:51:26.300 | over the decades, over the centuries?

01:51:28.820 | - The Flynn effect is, James Flynn,

01:51:31.540 | passed away about a year ago,

01:51:34.020 | published a set of analyses,

01:51:39.640 | going back a couple of decades,

01:51:43.660 | when he first noticed this,

01:51:46.060 | that IQ scores, when you looked over the years,

01:51:51.500 | seemed to be drifting up.

01:51:53.220 | Now this was not unknown to the people who make the test,

01:51:58.900 | because they re-norm the test periodically,

01:52:02.300 | and they have to re-norm the test periodically,

01:52:05.340 | because what 10 items correct meant

01:52:09.780 | relative to other people 50 years ago

01:52:13.780 | is not the same as what 10 items mean relative today.

01:52:18.660 | People are getting more things correct.

01:52:21.340 | Now, the scores have been drifting up about three points,

01:52:25.060 | IQ scores have been drifting up

01:52:27.340 | about three points per decade.

01:52:30.160 | This is not a personal effect, this is a cohort effect.

01:52:34.060 | Well, it's not for an individual, but--

01:52:37.180 | - The world, how do you, so what's the explanation?

01:52:39.580 | - And this has presented intelligence researchers

01:52:42.460 | with a great mystery.

01:52:44.660 | Two questions.

01:52:46.300 | First, is it effect on the 50% of the variance

01:52:50.900 | that's the G factor, or on the other 50%,

01:52:55.180 | and there's evidence that it is a G factor effect.

01:52:58.640 | And second, what on earth causes this,

01:53:02.740 | and doesn't this mean intelligence and G factor

01:53:05.860 | cannot be genetic, because the scale of natural selection

01:53:10.260 | is much, much longer than a couple of decades ago.

01:53:15.260 | And so it's been used to try to undermine the idea

01:53:19.780 | that there can be a genetic influence on intelligence.

01:53:23.500 | But certainly, it can be, the Flynn effect

01:53:28.660 | can affect the non-genetic aspects of intelligence,

01:53:32.140 | because genes account for maybe 50% of the variance.

01:53:37.140 | Maybe higher, it could be as high as 80% for adults,

01:53:40.980 | but let's just say 50% for discussion.

01:53:46.480 | So the Flynn effect, it's still a mystery.

01:53:50.360 | - It's still a mystery, that's interesting.

01:53:51.200 | - It's still a mystery,

01:53:52.200 | although the evidence is coming out.

01:53:54.080 | I told you before I edited a journal on intelligence,

01:53:56.760 | and we're doing a special issue in honor of James Flynn.

01:54:00.520 | So I'm starting to see papers now

01:54:02.520 | on really the latest research on this.

01:54:04.840 | I think most people who specialize in this area,

01:54:09.720 | of trying to understand the Flynn effect,

01:54:12.200 | are coming to the view, based on data,

01:54:16.520 | that it has to do with advances in nutrition and healthcare.

01:54:20.420 | And there's also evidence that the effect is slowing down,

01:54:28.040 | and possibly reversing.

01:54:30.440 | - Oh boy.

01:54:31.320 | So how would nutrition,

01:54:33.320 | so nutrition would still be connected to the G factor.

01:54:38.320 | So nutrition as it relates to the G factor,

01:54:41.640 | so the biology that leads to the intelligence.

01:54:45.080 | - Yes.

01:54:45.920 | - That would be the claim,

01:54:46.920 | the hypothesis being tested by the research.

01:54:52.160 | - Yes, and there's some evidence from infants

01:54:55.320 | that nutrition has made a difference.

01:55:00.600 | So it's not an unreasonable connection.

01:55:02.860 | But does it negate the idea that there's a genetic influence?

01:55:07.720 | Not logically at all.

01:55:10.040 | But it is very interesting.

01:55:11.880 | So that if you take an IQ test today,

01:55:15.560 | but you take the score and use the tables

01:55:21.040 | that were available in 1940,

01:55:25.020 | you're gonna wind up with a much higher IQ number.

01:55:28.040 | So are we really smarter than a couple of generations ago?

01:55:34.040 | No, but we might be able to solve problems a little better.

01:55:40.700 | And make use of our G because of things like Sesame Street

01:55:45.700 | and other curricula in school.

01:55:47.900 | More people are going to school.

01:55:51.120 | So there are a lot of factors here to disentangle.

01:55:56.220 | - It's fascinating though.

01:55:57.580 | It's fascinating that there's not clear answers yet.

01:56:00.460 | That as a population, we're getting smarter.

01:56:05.140 | When you just zoom out, that's what it looks like.

01:56:06.900 | As a population, we're getting smarter.

01:56:08.100 | And it's interesting to see what the effects of that are.

01:56:10.780 | I mean, this raises the question.

01:56:12.060 | We've mentioned it many times,

01:56:14.820 | but haven't clearly addressed it,

01:56:16.500 | which is nature versus nurture question.

01:56:19.320 | So how much of intelligence is nature?

01:56:22.100 | How much of it is nurture?

01:56:23.900 | How much of it is determined by genetics versus environment?

01:56:27.700 | - All of it.

01:56:28.820 | - All of it is genetics.

01:56:30.260 | - No, all of it is nature and nurture.

01:56:34.300 | - Yeah, so yes.

01:56:35.780 | Yes.

01:56:36.780 | - Okay.

01:56:37.620 | That's not as-

01:56:40.740 | - But how much of the variance can you apportion to either?

01:56:44.500 | Most of the people who work in this field say that

01:56:47.380 | that is the framing of that.

01:56:50.340 | If the question is framed that way, it can't be answered

01:56:54.500 | because nature and nurture

01:56:55.920 | are not two independent influences.

01:56:59.420 | They interact with each other.

01:57:01.300 | And understanding those interactions is so complex

01:57:06.500 | that many behavioral geneticists say it is today impossible

01:57:11.500 | and always will be impossible to disentangle that

01:57:16.980 | no matter what kind of advances there are in DNA technology

01:57:21.140 | and genomic informatics.

01:57:24.660 | - But there's still, to push back on that,

01:57:26.820 | that same intuition from behavioral geneticists

01:57:31.620 | would lead me to believe that there cannot possibly

01:57:34.380 | be a stable G factor.

01:57:37.040 | 'Cause it's super complex.

01:57:38.960 | - Many of them would assert that as a logical outcome.

01:57:43.820 | But because I believe there is a stable G factor

01:57:49.120 | from lots of sources of data, not just one study,

01:57:52.480 | but lots of sources of data over decades,

01:57:54.940 | I am more amenable to the idea

01:58:00.160 | that whatever interactions between genes

01:58:03.880 | and environment exist, they can be explicated,

01:58:08.880 | they can be studied,

01:58:11.680 | and that information can be used as a basis

01:58:16.920 | for molecular biology of intelligence.

01:58:19.400 | - Yeah, so, and we'll do this exact question,

01:58:21.880 | 'cause doesn't the stability of the G factor

01:58:26.880 | give you at least a hint that there is a biological basis

01:58:32.920 | for intelligence?

01:58:34.000 | - Yes, I think it's clear that the fact that

01:58:38.040 | an IQ score is correlated to things like thickness

01:58:42.960 | of your cortex, that it's correlated to

01:58:47.040 | glucose metabolic rate in your brain,

01:58:51.000 | that identical twins reared apart

01:58:58.120 | are highly similar in their IQ scores.

01:59:01.900 | These are all important observations

01:59:06.360 | that certainly more than, that indicate,

01:59:09.320 | not just suggest, but indicate

01:59:11.880 | that there's a biological basis.

01:59:13.480 | And does anyone believe intelligence

01:59:15.240 | has nothing to do with the brain?

01:59:16.880 | I mean, it's so obvious.

01:59:20.920 | - Well, indirectly, definitely has to do with it,

01:59:23.680 | but the question is, environment interacting with the brain,

01:59:27.360 | or is it the actual raw hardware of the brain?

01:59:32.360 | - Well, some would say that the raw hardware of the brain,

01:59:39.600 | as it develops from conception through adulthood,

01:59:46.380 | or at least through the childhood,

01:59:49.980 | that that so-called hardware that you are assuming

01:59:53.780 | is mostly genetic, in fact,

01:59:57.340 | is not as deterministic as you might think,

02:00:00.980 | that it is probabilistic,

02:00:03.380 | and what affects the probabilities

02:00:05.420 | are things like in uterine environment,

02:00:08.220 | and other factors like that, including chance,

02:00:13.980 | that chance affects the way the neurons

02:00:18.860 | are connecting during gestation.

02:00:21.480 | It's not, hey, it's pre-programmed.

02:00:26.720 | So there is pushback on the concept

02:00:30.520 | that genes provide a blueprint,

02:00:32.720 | that it's a lot more fluid.

02:00:36.740 | - Well, but also, yeah, so there's a lot,

02:00:38.940 | a lot happens in the first few months of development.

02:00:43.940 | So in nine months inside the mother's body,

02:00:51.060 | and in the few months afterwards,

02:00:57.040 | there's a lot of fascinating stuff,

02:00:58.580 | like including chance and luck, like you said,

02:01:01.720 | how things connect up.

02:01:03.040 | Man, the question is, afterwards,

02:01:06.160 | in your plasticity of the brain,

02:01:07.480 | how much adjustment there is relative to the environment,

02:01:10.000 | how much that affects the G factor,

02:01:12.340 | but that's where the whole conclusions of the studies

02:01:15.560 | that we've been talking about is,

02:01:18.280 | that seems to have less and less and less of an effect

02:01:20.880 | as pretty quickly.

02:01:23.840 | - Yes, and I do think there is more of a genetic,

02:01:27.540 | by my view, and I'm not an expert on this,

02:01:30.860 | I mean, genetics is a highly technical and complex subject.

02:01:34.280 | I am not a geneticist, not a behavioral geneticist,

02:01:37.940 | but my reading of this, my interpretation of this,

02:01:42.900 | is that there is a genetic blueprint, more or less,

02:01:47.900 | and that has a profound influence

02:01:51.020 | on your subsequent intellectual development,

02:01:54.180 | including the G factor.

02:01:56.340 | And that's not to say things can't happen to,

02:02:01.340 | I mean, if you think of that genes provide a potential,

02:02:05.500 | fine, and then various variables impact that potential.

02:02:10.500 | And every parent of a newborn, implicitly or explicitly,

02:02:15.780 | wants to maximize that potential.

02:02:19.480 | This is why you buy educational toys.

02:02:21.940 | This is why you pay attention to organic baby food.

02:02:25.620 | This is why you do all these things,

02:02:28.840 | because you want your baby to be as healthy

02:02:31.460 | and as smart as possible.

02:02:33.580 | And every parent will say that.

02:02:35.980 | - Is there a case to be made,

02:02:37.380 | can you steel me on the case,

02:02:39.820 | that genetics is a very tiny component of all of this,

02:02:47.000 | and the environment is essential?

02:02:49.460 | - I don't think the data supports

02:02:50.980 | that genetics is a tiny component.

02:02:53.620 | I think the data support the idea

02:02:55.140 | that the genetics is a very important,

02:02:57.820 | and I don't say component, I say influence.

02:03:01.080 | A very important influence.

02:03:03.540 | And the environment is a lot less than people believe.

02:03:07.700 | Most people believe environment plays a big role.

02:03:10.780 | I'm not so sure.

02:03:11.620 | - I guess what I'm asking you is,

02:03:13.200 | can you see where what you just said, it might be wrong?

02:03:18.200 | Can you imagine a world,

02:03:22.120 | and what kind of evidence would you need to see,

02:03:25.580 | to say, you know what, the intuition, the studies so far,

02:03:29.880 | like reversing the directions.

02:03:31.680 | So one of the cool things we have now more and more,

02:03:34.720 | is we're getting more and more data,

02:03:36.040 | and the rate of the data is escalating

02:03:40.080 | because of the digital world.

02:03:41.800 | So when you start to look at a very large scale of data,

02:03:46.180 | both on the biology side and the social side,

02:03:49.600 | we might be discovering some very counterintuitive things

02:03:52.280 | about society.

02:03:53.820 | We might see the edge cases that reveal

02:03:57.540 | that if we actually scale those edge cases,

02:04:00.080 | and they become like the norm,

02:04:02.640 | that we'll have a complete shift in our,

02:04:06.220 | like you'll see G-factor be able

02:04:09.580 | to be modified throughout life,

02:04:11.940 | in the teens and in later life.

02:04:15.200 | So is it any case you can make,

02:04:18.280 | or for where your current intuitions are wrong?

02:04:20.780 | - Yes, and it's a good question,

02:04:22.280 | because I think everyone should always be asked,

02:04:24.540 | what evidence would change your mind?

02:04:26.560 | It's certainly not only a fair question,

02:04:29.800 | it is really the key question for anybody working

02:04:32.940 | on any aspect of science.

02:04:36.580 | I think that if environment was very important,

02:04:41.580 | we would have seen it clearly by now.

02:04:45.860 | It would have been obvious that school interventions,

02:04:49.820 | compensatory education, early childhood education,

02:04:53.500 | all these things that have been earnestly tried,

02:04:56.780 | and well-funded, well-designed studies,

02:04:59.480 | would show some effect, and they don't.

02:05:02.060 | - They don't.

02:05:02.900 | What if the school, the way we've tried school,

02:05:05.680 | compensatory school sucks, and we need to do better?

02:05:08.300 | - That's what everybody said at the beginning,

02:05:09.780 | that's what everybody said to Jensen.

02:05:11.660 | He said, well, maybe we need to start earlier.

02:05:15.460 | Maybe we need not do pre-kindergarten,

02:05:18.500 | but pre-pre-kindergarten.

02:05:19.900 | Yeah, it's always an infinite,

02:05:22.220 | well, maybe we didn't get it right.

02:05:24.460 | But after decades of trying, 50 years,

02:05:28.220 | 50 or 60 years of trying,

02:05:30.700 | surely something would have worked to the point

02:05:34.660 | where you could actually see a result,

02:05:37.140 | and not need a probability level at .05 on some means.

02:05:42.140 | So that's the kind of evidence that would change my mind.

02:05:47.420 | - Population-level interventions like schooling

02:05:52.380 | that you would see, like this actually has an effect.

02:05:57.380 | - Yes, and when you take adopted kids,

02:06:01.020 | and they grow up in another family,

02:06:04.060 | and you find out when those adopted kids are adults,

02:06:06.740 | their IQ scores don't correlate with the IQ scores

02:06:09.740 | of their adoptive parents,

02:06:11.660 | but they do correlate with their IQ scores

02:06:14.300 | of their biological parents, whom they've never met.

02:06:17.840 | I mean, these are important,

02:06:20.500 | these are powerful observations.

02:06:22.740 | - And it would be convincing to you

02:06:24.220 | if the reverse was true.

02:06:26.420 | - Yes, that would be more.

02:06:27.700 | And there is some data on adoption

02:06:30.540 | that indicates the adopted children

02:06:34.340 | are moving a little bit more

02:06:35.860 | toward their adoptive parents.

02:06:39.280 | But it's, to me, the overwhelming,

02:06:44.020 | I have this concept called the weight of evidence,

02:06:47.260 | where I don't interpret any one study too much.

02:06:50.440 | The weight of evidence tells me genes are important.

02:06:53.460 | But what does that mean?

02:06:54.460 | What does it mean that genes are important,

02:06:56.700 | knowing that gene expression,

02:06:59.300 | genes don't express themselves in a vacuum,

02:07:02.620 | they express themselves in an environment.

02:07:05.820 | So the environment has to have something to do with it,

02:07:08.540 | especially if the best genetic estimates

02:07:11.420 | of the amount of variance are around 50,

02:07:14.220 | or even if it's as high as 80%,

02:07:17.420 | it still leaves 20% of non-genetic.

02:07:21.460 | Now, maybe that is all luck, maybe that's all chance.

02:07:25.580 | I could believe that, I could easily believe that.

02:07:28.860 | So, but I do think,

02:07:32.980 | after 50 years of trying various interventions,

02:07:36.500 | and nothing works, including memory training,

02:07:40.060 | including listening to Mozart,

02:07:41.780 | including playing computer games,

02:07:43.940 | none of that has shown any impact

02:07:46.380 | on intelligence test scores.

02:07:48.980 | - Is there data on the intelligence,

02:07:51.980 | the IQ of parents as it relates to the children?

02:07:57.020 | - Yes, and there is some genetic evidence

02:08:00.020 | of kind of an interaction

02:08:04.340 | between the parents' IQ and the environment,

02:08:08.580 | high IQ parents provide an enriched environment,

02:08:13.240 | which then can impact the child in addition to the genes,

02:08:17.860 | it's that environment.

02:08:19.980 | So there are all these interactions that,

02:08:23.860 | but it's not, you know,

02:08:25.060 | think about the number of books in a household.

02:08:28.900 | This was a variable that's correlated with IQ, and--

02:08:33.020 | - It is. - Yeah.

02:08:34.300 | Well, why?

02:08:35.580 | Especially if the kid never reads any of the books.

02:08:39.680 | It's because more intelligent people

02:08:42.940 | have more books in their house.

02:08:45.420 | And if you're more intelligent,

02:08:47.900 | and there's a genetic component to that,

02:08:51.100 | the child will get those genes or some of those genes

02:08:55.140 | as well as the environment,

02:08:57.300 | but it's not the number of books in the house

02:09:00.300 | that actually directly impacts the child.

02:09:04.020 | So the two scenarios on this are,

02:09:06.420 | you find that,

02:09:07.540 | and this was used to get rid of the SAT test,

02:09:12.940 | oh, the SAT score is highly correlated

02:09:15.100 | with the social economic status of the parents.

02:09:18.260 | So all you're really measuring is how rich the parents are.

02:09:21.260 | Okay, well, why are the parents rich?

02:09:24.660 | - Yes. - Okay.

02:09:27.540 | And so you could, the opposite kind of syllogism

02:09:32.540 | is that people who are very bright make more money.

02:09:37.900 | They can afford homes in better neighborhoods

02:09:42.300 | so their kids get better schools.

02:09:44.500 | Now, the kids grow up bright.

02:09:47.620 | Where in that chain of events does that come from?

02:09:50.540 | Well, unless you have a genetically informative

02:09:55.260 | research design, where you look at siblings

02:09:58.940 | that have the same biological parents and so on,

02:10:02.600 | you can't really disentangle all that.

02:10:05.460 | Most studies of social economic status and intelligence

02:10:10.460 | do not have a genetically informed design.

02:10:13.900 | So any conclusions they make about the causality

02:10:17.660 | of the social economic status being the cause of the IQ

02:10:22.660 | is a stretch.

02:10:25.380 | And where you do find genetically informative designs,

02:10:30.000 | you find most of the variance in your outcome measures

02:10:35.140 | are due to the genetic component.

02:10:38.860 | And sometimes the SES adds a little,

02:10:43.300 | but the weight of evidence is it doesn't add

02:10:46.300 | very much variance to predict what's going on

02:10:49.280 | beyond the genetic variance.

02:10:52.620 | So when you actually look at it in some,

02:10:56.020 | and there aren't that many studies

02:10:57.500 | that have genetically informed designs,

02:11:00.140 | but when you do see those,

02:11:02.980 | the genes seem to have an advantage.

02:11:05.300 | - Sorry for the strange questions,

02:11:06.740 | but is there a connection between fertility

02:11:13.100 | or the number of kids that you have and G factor?

02:11:17.260 | So you know, the kind of conventional wisdom

02:11:19.960 | is people of, maybe is it higher economic status

02:11:24.960 | or something like that are having fewer children?

02:11:28.620 | I just loosely hear these kinds of things.

02:11:30.880 | Is there data that you're aware of

02:11:33.100 | in one direction or another on this?

02:11:35.300 | - Well, strange questions always get strange answers.

02:11:38.980 | - Yes, all right.

02:11:42.140 | So do you have a strange answer for that strange question?

02:11:44.700 | - The answer is there were some studies

02:11:48.180 | that indicated the more children in a family,

02:11:52.000 | the firstborn children would be more intelligent

02:11:57.840 | than the fourth or fifth or sixth.

02:12:00.620 | It's not clear that those studies hold up over time.

02:12:05.220 | And of course, what you see also is that families

02:12:10.260 | where there are multiple children, four, five, six, seven,

02:12:14.740 | you know, really big families,

02:12:16.700 | the social economic status of those families

02:12:23.640 | usually in the modern age is not that high.

02:12:26.720 | Maybe it used to be the aristocracy

02:12:30.860 | used to have a lot of kids, I'm not sure exactly.

02:12:33.420 | But there have been reports of correlations

02:12:39.040 | between IQ and fertility.

02:12:41.600 | But I'm not sure that the data are very strong

02:12:46.220 | that the firstborn child is always the smartest.

02:12:50.440 | It seems like there's some data to that,

02:12:52.760 | but I'm not current on that.

02:12:54.860 | - How would that be explained?

02:12:56.040 | That would be a nurture.

02:12:58.600 | - Well, it could be nurture.

02:13:00.220 | It could be in uterine environment.

02:13:03.100 | I mean-- - Boy, biology's complicated.

02:13:06.200 | - It's, and this is why this, you know,

02:13:09.920 | like many areas of science, you said earlier

02:13:13.220 | that there are a lot of gray areas

02:13:15.580 | and no definitive answers.

02:13:19.060 | This is not uncommon in science

02:13:24.020 | that the closer you look at a problem,

02:13:27.820 | the more questions you get, not the fewer questions,

02:13:32.620 | because the universe is complicated.

02:13:35.660 | And the idea that we have people on this planet

02:13:40.020 | who can study the first nanoseconds of the Big Bang,

02:13:43.900 | that's pretty amazing.

02:13:46.860 | And I've always said that if they can study

02:13:50.680 | the first nanoseconds of the Big Bang,

02:13:53.220 | we can certainly figure out something about intelligence

02:13:57.100 | that allows that.

02:13:58.900 | - I'm not sure what's more complicated,

02:14:00.940 | the human mind or the physics of the universe.

02:14:05.560 | It's unclear to me.

02:14:07.180 | I think we overemphasize-- - Well, that's a very

02:14:10.040 | humbling statement. (laughs)

02:14:12.920 | - Maybe it's a very human-centric, egotistical statement

02:14:15.740 | that our mind is somehow super complicated,

02:14:18.000 | but biology is a tricky one to unravel.

02:14:21.640 | Consciousness, what is that?

02:14:25.340 | - Well, I've always believed that consciousness

02:14:29.280 | and intelligence are the two real fundamental problems

02:14:34.800 | of the human brain.

02:14:35.920 | And therefore, I think they must be related.

02:14:40.640 | (both laugh)

02:14:43.400 | - Yeah, heart problems like walk together,

02:14:46.400 | holding hands kind of idea.

02:14:48.960 | - You may not know this, but I did some of the early research

02:14:52.440 | on anesthetic drugs with brain imaging,

02:14:54.800 | trying to answer the question, what part of the brain

02:14:57.720 | is the last to turn off when someone loses consciousness?

02:15:01.440 | And is that the first part of the brain to turn on

02:15:04.440 | when consciousness is regained?

02:15:06.960 | And I was working with an anesthesiologist

02:15:09.080 | named Mike Alkire, who's really brilliant at this.

02:15:11.800 | These were really the first studies of brain imaging

02:15:15.400 | using positron emission tomography long before fMRI.

02:15:20.400 | And you would inject a radioactive sugar

02:15:24.720 | that labeled the brain, and the harder the brain was working,

02:15:28.880 | the more sugar it would take up.

02:15:30.960 | And then you could make a picture

02:15:32.240 | of glucose use in the brain.

02:15:35.060 | And he was amazing.

02:15:38.520 | He managed to do this.

02:15:40.320 | In normal volunteers, he brought in an anesthetized

02:15:44.400 | as if they were going into surgery.

02:15:46.620 | And he managed all the human subjects requirements

02:15:51.480 | on this research.

02:15:53.040 | And he was brilliant at this.

02:15:56.920 | And what we did is we had these normal volunteers

02:16:01.920 | come in on three occasions.

02:16:05.000 | On one occasion, he gave them enough anesthetic drug

02:16:08.640 | so they were a little drowsy.

02:16:13.200 | And on another occasion, they came in

02:16:16.840 | and he fully anesthetized them.

02:16:18.640 | And he would say, "Mike, can you hear me?"

02:16:25.900 | And the person would say, "Uh, yeah."

02:16:29.960 | And then we would scan people

02:16:33.960 | and under no anesthetic condition.

02:16:37.480 | So same person.

02:16:39.480 | And we were looking to see if we could see

02:16:43.480 | the part of the brain turn off.

02:16:46.640 | He subsequently tried to do this with fMRI,

02:16:49.400 | which has a faster time resolution.

02:16:51.680 | And you could do it in real time as the person went under

02:16:55.200 | and then regain consciousness,

02:16:56.700 | where you couldn't do that with PET.

02:16:57.940 | You had to have three different occasions.

02:17:00.340 | And the results were absolutely fascinating.

02:17:03.300 | We did this with different anesthetic drugs

02:17:05.980 | and different drugs impacted different parts of the brain.

02:17:09.680 | So we were naturally looking for the common one.

02:17:13.940 | And it seemed to have something to do with the thalamus

02:17:16.900 | and consciousness.

02:17:19.460 | This was actual data on consciousness.

02:17:23.660 | Real, actual consciousness.

02:17:25.600 | - What part of the brain turns on?

02:17:28.280 | What part of the brain turns off?

02:17:30.740 | - It's not so clear.

02:17:32.160 | - But maybe has something to do with the thalamus.

02:17:35.880 | - The sequence of events seemed to have the thalamus in it.

02:17:39.920 | - Boy.

02:17:41.080 | - Now, here's the question.

02:17:42.600 | Are some people more conscious than others?

02:17:45.720 | Are there individual differences in consciousness?

02:17:49.520 | And I don't mean it in the psychedelic sense.

02:17:53.000 | I don't mean it in the political consciousness sense.

02:17:55.760 | I just mean it in everyday life.

02:17:57.260 | Do some people go through everyday life

02:17:59.460 | more conscious than others?

02:18:01.560 | And are those the people we might

02:18:03.680 | actually label more intelligent?

02:18:06.040 | So now, the other thing I was looking for

02:18:08.980 | is whether the parts of the brain we were seeing

02:18:11.520 | in the anesthesia studies were the same parts

02:18:14.560 | of the brain we were seeing in the intelligence studies.

02:18:17.780 | Now, this was very complicated, expensive research.

02:18:22.520 | We didn't really have funding to do this.

02:18:24.480 | We were trying to do it on the fly.

02:18:26.400 | I'm not sure anybody has pursued this.

02:18:29.680 | I'm retired now.

02:18:31.560 | He's gone on to other things.

02:18:33.960 | But I think it's an area of research

02:18:36.480 | that would be fascinating to see the parts.

02:18:40.200 | There are a lot more imaging studies now of consciousness.

02:18:43.160 | I'm just not up on them.

02:18:45.120 | - But basically, the question is which imaging,

02:18:48.120 | so newer imaging studies to see in high-resolution

02:18:52.360 | spatial and temporal way, which part of the brain

02:18:55.320 | lights up when you're doing intelligence tasks?

02:19:00.200 | And which parts of the brain lights up

02:19:02.160 | when you're doing consciousness tasks

02:19:03.600 | and see the interplay between them?

02:19:05.800 | Try to infer.

02:19:06.960 | I mean, that's the challenge of neuroscience.

02:19:09.120 | Without understanding deeply, looking from the outside,

02:19:14.120 | try to infer something about how the whole thing works.

02:19:19.420 | - Well, imagine this.

02:19:21.060 | Here's a simple question.

02:19:23.040 | Does it take more anesthetic drug

02:19:25.640 | to have a person lose consciousness

02:19:32.480 | if their IQ is 140 than a person with an IQ of 70?

02:19:37.240 | - That's an interesting way to study it, yeah.

02:19:41.320 | I mean, if there is a, if the answer to that is a stable yes,

02:19:46.320 | that's very interesting.

02:19:47.920 | - So I tried to find out.

02:19:50.200 | And I went to some anesthesiology textbooks

02:19:53.160 | about how you dose.

02:19:55.720 | And they dose by weight.

02:19:57.040 | And what I also learned, this is a little bit off subject,

02:20:04.200 | anesthesiologists are never sure how deep you are.

02:20:09.880 | - Yeah.

02:20:10.720 | - And they usually tell by poking you with a needle.

02:20:13.660 | And if you don't jump, they tell the surgeon to go ahead.

02:20:17.120 | I'm not sure that's literally true, but it's--

02:20:20.840 | - Well, it might be very difficult to know precisely

02:20:23.800 | how deep you are.

02:20:26.320 | It has to do with the same kind of measurements

02:20:28.400 | that you were doing with the consciousness.

02:20:30.500 | It's difficult to know.

02:20:34.640 | - So I don't lose my train of thought.

02:20:35.960 | I couldn't find in the textbooks anything

02:20:38.100 | about dosing by intelligence.

02:20:41.000 | I asked my friend, the anesthesiologist,

02:20:43.740 | he said, no, he doesn't know.

02:20:45.720 | I said, can we do a chart review

02:20:47.520 | and look at people using their years of education

02:20:52.060 | as a proxy for IQ?

02:20:54.280 | Because if someone's gone to graduate school,

02:20:56.720 | that tells you something.

02:20:58.280 | You can make some inference as opposed to someone

02:21:00.240 | who didn't graduate high school.

02:21:02.680 | Can we do a chart review?

02:21:04.360 | And he says, no, they never really put down the exact dose.

02:21:08.880 | No, he said, no.

02:21:10.560 | So to this day, the simple question,

02:21:15.560 | does it take more anesthetic drug to put someone under

02:21:19.240 | if they have a high IQ or less, or less?

02:21:23.440 | It could go either way.

02:21:24.640 | Because by the way, our early PET scan studies

02:21:27.520 | of intelligence found the unexpected result

02:21:32.520 | of an inverse correlation between glucose metabolic rate

02:21:37.360 | and intelligence.

02:21:38.520 | It wasn't how much a brain area lit up.

02:21:42.960 | How much it lit up was negatively correlated

02:21:45.920 | to how well they did on the test,

02:21:47.960 | which led to the brain efficiency hypothesis,

02:21:51.120 | which is still being studied today.

02:21:53.680 | And there's more and more evidence that the efficiency

02:21:57.820 | of brain information processing is more related

02:22:01.360 | to intelligence than just more activity.

02:22:06.360 | - Yeah, and it'll be interesting, again,

02:22:10.680 | it's a total hypothesis how much the relationship

02:22:14.240 | between intelligence and consciousness,

02:22:17.000 | it's not obvious that those two, if there's correlation,

02:22:19.800 | they could be inversely correlated.

02:22:24.320 | Wouldn't that be funny?

02:22:26.000 | If you, the consciousness factor,

02:22:31.000 | the C factor plus the G factor equals one.

02:22:35.680 | It's a nice trade-off.

02:22:40.000 | - You get a trade-off, how deeply you experience the world

02:22:43.080 | versus how deeply you're able to reason through the world.

02:22:48.040 | - What a great hypothesis.

02:22:50.360 | Certainly somebody listening to this can do this study.

02:22:53.800 | - Even if it's the aliens analyzing humans

02:22:57.360 | a few centuries from now.

02:22:59.080 | Let me ask you from an AI perspective.

02:23:01.140 | I don't know how much you've thought about machines,

02:23:06.800 | but there's the famous Turing test,

02:23:09.680 | test of intelligence for machines,

02:23:12.600 | which is a beautiful, almost like a cute formulation

02:23:17.320 | of intelligence that Alan Turing proposed.

02:23:22.320 | Basically conversation being if you can fool a human

02:23:26.760 | to think that a machine is a human that passes the test.

02:23:34.960 | I suppose you could do a similar thing for humans.

02:23:39.960 | If I can fool you that I'm intelligent,

02:23:43.100 | then that's a good test of intelligence.

02:23:46.660 | You're talking to two people,

02:23:50.180 | and the test is saying who has a higher IQ.

02:23:57.880 | It's an interesting test,

02:24:03.440 | 'cause maybe charisma can be very useful there.

02:24:06.660 | You're only allowed to use conversation,

02:24:09.200 | which is the formulation of the Turing test.

02:24:11.120 | Anyway, all that to say is what are good tests

02:24:15.000 | of intelligence for machines?

02:24:17.100 | What do you think it takes to achieve

02:24:20.680 | human-level intelligence for machines?

02:24:23.120 | - Well, I have thought a little bit about this,

02:24:25.520 | but every time I think about these things,

02:24:29.920 | I rapidly reach the limits of my knowledge and imagination.

02:24:34.920 | So when Alexa first came out,

02:24:40.380 | and I think there was a competing one.

02:24:46.440 | Well, there was Siri with Apple, and Google had Alexa.

02:24:50.440 | - No, no, Amazon had Alexa.

02:24:52.600 | - Amazon had Alexa, and Google has something.

02:24:56.400 | So I proposed to one of my colleagues

02:24:58.420 | that he buy one of these, one of each,

02:25:02.920 | and then ask it questions from the IQ test.

02:25:06.180 | - Nice.

02:25:08.800 | - But it became apparent that they all search the internet

02:25:13.800 | so they all can find answers to questions

02:25:16.880 | like how far is it between Washington and Miami,

02:25:21.080 | and repeat after me.

02:25:22.640 | Now, I don't know if you said to Alexa,

02:25:26.080 | I'm going to repeat these numbers backwards to me.

02:25:30.420 | I don't know what would happen, I've never done it.

02:25:32.780 | But so one answer to your question is,

02:25:36.200 | try, you're gonna try it right now, let's try it.

02:25:39.200 | Let's try it.

02:25:40.040 | - No, no, no. - Yes, Siri.

02:25:41.360 | - So it would actually probably go to Google search,

02:25:46.080 | and it would be all confusing kind of stuff.

02:25:48.540 | It would fail.

02:25:50.300 | - Well, then I guess there's a test that it would fail.

02:25:54.620 | - Well, but that's not, that has to do more with the,

02:25:58.280 | you know, the language of communication versus the content.

02:26:03.620 | So if you did an IQ test to a person

02:26:06.700 | who doesn't speak English,

02:26:07.700 | and the test was administered in English,

02:26:10.300 | that's not really the test of--

02:26:11.580 | - Well, let's think about the computers

02:26:13.220 | that beat the Jeopardy champions.

02:26:15.420 | - Yeah, so that, because I happen to know

02:26:20.040 | how those are programmed, those are very hard-coded,

02:26:22.380 | and there's definitely a lack of intelligence there.

02:26:25.340 | There's something like IQ tests.

02:26:30.340 | There's a guy, artificial intelligence researcher,

02:26:35.500 | Francois Chollet, he's at Google,

02:26:38.120 | he's one of the seminal people in machine learning.

02:26:40.860 | He also, as a fun aside thing,

02:26:43.820 | developed an IQ test for machines.

02:26:46.020 | - Oh, I haven't heard that.

02:26:47.700 | I would just like to know about that.

02:26:48.780 | - I'll actually email you this,

02:26:51.260 | 'cause it'd be very interesting for you.

02:26:53.260 | It doesn't get much attention,

02:26:54.480 | because people don't know what to do with it.

02:26:57.260 | But it deserves a lot of attention,

02:27:00.820 | which is, it basically does a pattern type of tests,

02:27:04.640 | where you have to do, you know, one standard one is,

02:27:09.560 | you're given three things, and you have to do a fourth one,

02:27:13.500 | that kind of thing.

02:27:14.340 | So you have to understand the pattern here.

02:27:16.780 | And for that, it really simplifies

02:27:20.860 | to,

02:27:21.700 | so the interesting thing is,

02:27:26.360 | he's trying not to achieve high IQ,

02:27:30.740 | he's trying to achieve like pretty low bar for IQ.

02:27:34.980 | Things that are kind of trivial for humans.

02:27:37.640 | And they're actually really tough for machines.

02:27:41.900 | Which is seeing, playing with these concepts

02:27:44.860 | of symmetry, of counting.

02:27:48.700 | Like if I give you one object, two objects, three objects,

02:27:52.140 | you'll know that the last one is four objects,

02:27:54.740 | you can like count them.

02:27:56.420 | You can cluster objects together.

02:27:59.080 | It's both visually and conceptually.

02:28:01.060 | We could do all these things with our mind

02:28:03.140 | that we take for granted.

02:28:04.740 | The objectness of things.

02:28:07.340 | We can like figure out what spatially is an object and isn't.

02:28:12.340 | And we can play with those ideas.

02:28:17.380 | And machines really struggle with that.

02:28:19.780 | So he really cleanly formulated these IQ tests.

02:28:22.720 | I wonder what like that would equate to for humans with IQ.

02:28:27.060 | But it'd be a very low IQ.

02:28:28.740 | But that's exactly the kind of formulation like,

02:28:32.940 | okay, we wanna be able to solve this.

02:28:34.660 | How do we solve this?

02:28:35.900 | And he does this as a challenge,

02:28:37.140 | and nobody's been able to,

02:28:38.840 | it's similar to the Alexa Prize,

02:28:42.060 | which is Amazon is hosting a conversational challenge.

02:28:45.620 | Nobody's been able to do well on his.

02:28:48.780 | But that's an interesting,

02:28:50.900 | those kinds of tests are interesting

02:28:52.620 | 'cause we take for granted all the ability of the human mind

02:28:57.420 | to play with concepts and to formulate concepts

02:29:03.420 | out of novel things.

02:29:06.960 | So like things we've never seen before.

02:29:10.100 | We're able to use that.

02:29:11.820 | I mean, that's, I've talked to a few people

02:29:14.460 | that design IQ tests sort of online.

02:29:17.180 | They write IQ tests.

02:29:19.000 | And I was trying to get some questions from them.

02:29:21.740 | And they spoke to the fact that

02:29:23.900 | we can't really share questions with you

02:29:25.540 | because part of the,

02:29:27.940 | like first of all, it's really hard work

02:29:31.580 | to come up with questions.

02:29:33.460 | It's really, really hard work.

02:29:34.980 | It takes a lot of research, but it also takes a lot,

02:29:38.300 | novelty generating.

02:29:40.140 | You're constantly coming up with really new things.

02:29:44.300 | And part of the point is that

02:29:47.780 | they're not supposed to be public.

02:29:50.020 | They're supposed to be new to you when you look at them.

02:29:53.540 | It's interesting that the novelty is fundamental

02:29:56.060 | to the hardness of the problem.

02:29:57.620 | At least a part of what makes the problem hard

02:30:01.540 | is you've never seen it before.

02:30:03.100 | - Right, that's called fluid intelligence

02:30:05.860 | as opposed to what's called crystallized intelligence,

02:30:08.780 | which is your knowledge of facts.

02:30:12.060 | You know things.

02:30:13.900 | But can you use those things to solve a problem?

02:30:17.540 | Those are two different things.

02:30:19.460 | - Do you think we'll be able to,

02:30:21.540 | 'cause we spoke,

02:30:22.740 | and I don't wanna miss an opportunity to talk about this,

02:30:25.660 | we spoke about the neurobiology,

02:30:27.380 | about the molecular biology of intelligence.

02:30:30.260 | Do you think one day we'll be able to modify the biology

02:30:33.820 | of, or the genetics of a person

02:30:39.500 | to modify their intelligence,

02:30:42.860 | to increase their intelligence?

02:30:43.980 | We started this conversation

02:30:45.100 | by talking about a pill you could take.

02:30:47.060 | Do you think that such a pill would exist?

02:30:49.140 | - Metaphorically, I do.

02:30:51.300 | And I am supremely confident that it's possible

02:30:56.180 | because I am supremely ignorant

02:30:58.660 | of the complexities of neurobiology.

02:31:01.100 | (Lex laughing)

02:31:02.660 | And so I have written--

02:31:03.980 | - Ignorance is bliss.

02:31:04.940 | - Well, I have written that the nightmares

02:31:07.340 | of neurobiologists,

02:31:09.620 | understanding the complexities,

02:31:11.260 | this cascade of events that happens at the synaptic level,

02:31:16.260 | that these nightmares are what fuel some people to solve.

02:31:24.900 | So some people, you have to be undaunted.

02:31:28.420 | I mean, yeah, this is not easy.

02:31:31.260 | Look, we're still trying to figure out cancer.

02:31:33.980 | It was only recently that they figured out

02:31:37.900 | why aspirin works.

02:31:41.020 | These are not easy problems,

02:31:42.980 | but I also have the perspective of the history of science

02:31:47.980 | is the history of solving problems

02:31:52.980 | that are extraordinarily complex.

02:31:57.540 | - And seem impossible at the time.

02:31:58.740 | - And seem impossible at the time.

02:32:01.020 | - And so one of the things you look at

02:32:03.900 | at companies like Neuralink,

02:32:06.060 | you have brain-computer interfaces,

02:32:08.460 | you start to delve into the human mind

02:32:10.180 | and start to talk about machines measuring,

02:32:12.420 | but also sending signals to the human mind,

02:32:14.740 | you start to wonder what impact that has on the G factor,

02:32:19.740 | modifying in small ways or in large ways the functioning,

02:32:26.740 | the mechanical, electrical,

02:32:29.860 | chemical functioning of the brain.

02:32:34.180 | - I look at everything about the brain.

02:32:37.460 | There are different levels of explanation.

02:32:39.780 | On one hand, you have a behavioral level,

02:32:42.420 | but then you have brain circuitry,

02:32:45.940 | and then you have neurons,

02:32:49.940 | and then you have dendrites,

02:32:52.580 | and then you have synapses,

02:32:55.140 | and then you have the neurotransmitters

02:32:59.900 | and the presynaptic and the postsynaptic terminals.

02:33:05.740 | And then you have all the things

02:33:07.580 | that influence neurotransmitters.

02:33:10.620 | And then you have the individual differences among people.

02:33:15.620 | Yeah, it's complicated,

02:33:18.700 | but 51 million people in the United States

02:33:23.700 | have IQs under 85 and struggle with everyday life.

02:33:28.940 | Shouldn't that motivate people to take a look at this?

02:33:36.940 | - Yeah, and to treat it seriously.

02:33:39.540 | Yeah, but I just want to linger one more time

02:33:42.860 | that we have to remember that the science of intelligence,

02:33:47.340 | the measure of intelligence

02:33:52.020 | is only a part of the human condition.

02:33:54.660 | The thing that makes life beautiful

02:33:56.420 | and the creation of beautiful things in this world

02:33:59.260 | is perhaps loosely correlated,

02:34:04.380 | but it's not dependent entirely on intelligence.

02:34:08.740 | - Absolutely, I certainly agree with that.

02:34:12.060 | - So for anyone sort of listening,

02:34:15.580 | I'm still not convinced that

02:34:18.020 | more intelligence is always better

02:34:23.140 | if you want to create beauty in this world.

02:34:26.060 | I don't know.

02:34:27.180 | - Well, I didn't say more intelligence is always better

02:34:29.860 | if you want to create beauty.

02:34:31.340 | I just said all things being equal,

02:34:34.340 | more is better than less.

02:34:36.380 | That's all I mean.

02:34:37.420 | - Yeah, but that's sort of that I just want to sort of say,

02:34:40.820 | 'cause to me, one of the things that makes life great

02:34:45.820 | is the opportunity to create beautiful things,

02:34:48.620 | and so I just want to sort of empower people to do that

02:34:52.860 | no matter what some IQ test says.

02:34:56.300 | At the population level,

02:34:57.580 | we do need to look at IQ tests to help people.

02:35:01.460 | And to also inspire us, yeah,

02:35:03.540 | to take on some of these

02:35:04.380 | extremely difficult scientific questions.

02:35:07.820 | Do you have advice for young people in high school,

02:35:12.460 | in college, whether they're thinking about career

02:35:17.460 | or they're thinking about a life they can be proud of?

02:35:20.660 | Is there advice you can give?

02:35:22.240 | Whether they want to pursue psychology or biology

02:35:28.860 | or engineering, or they want to be artists

02:35:30.980 | and musicians and poets.

02:35:33.700 | - I can't advise anybody on that level

02:35:37.620 | of what their passion is. - Poetry.

02:35:40.140 | - But I can say if you're interested in psychology,

02:35:45.500 | if you're interested in science,

02:35:47.860 | and the science around the big questions

02:35:52.860 | of consciousness and intelligence and psychiatric illness,

02:36:00.180 | we haven't really talked about brain illnesses

02:36:03.820 | and what we might learn from,

02:36:05.900 | if you are trying to develop a drug

02:36:08.980 | to treat Alzheimer's disease,

02:36:11.020 | you are trying to develop a drug

02:36:13.540 | to impact learning and memory,

02:36:17.480 | which are core to intelligence.

02:36:20.300 | So it could well be that the so-called IQ pill

02:36:23.660 | will come from a pharmaceutical company

02:36:26.220 | trying to develop a drug for Alzheimer's disease.

02:36:29.220 | - Because that's exactly what you're trying to do, right?

02:36:31.380 | Yeah, just like you said. - Well, what will that drug do

02:36:34.700 | in a college student that doesn't have Alzheimer's disease?

02:36:38.340 | So I would encourage people who are interested in psychology,

02:36:43.340 | who are interested in science,

02:36:47.660 | to pursue a scientific career

02:36:52.100 | and address the big questions.

02:36:55.140 | And the most important thing I can tell you,

02:36:59.700 | if you're gonna be in kind of a research environment,

02:37:04.340 | is you gotta follow the data where the data take you.

02:37:07.380 | You can't decide in advance where you want the data to go.

02:37:10.340 | And if the data take you to places

02:37:12.420 | that you don't have the technical expertise to follow,

02:37:16.180 | like I would like to understand more

02:37:19.540 | about molecular biology,

02:37:21.480 | but I'm not gonna become a molecular biologist now,

02:37:24.820 | but I know people who are.

02:37:27.020 | And my job is to get them interested

02:37:29.460 | to take their expertise into this direction.

02:37:32.620 | And it's not so easy, but...

02:37:36.420 | - And if the data takes you to a place

02:37:39.100 | that's controversial, that's counterintuitive in this world,

02:37:42.760 | no, I would say it's probably a good idea

02:37:49.140 | to still push forward boldly,

02:37:52.140 | but to communicate the interpretation of the results

02:37:56.380 | with skill, with compassion,

02:37:58.960 | with a greater breadth of understanding of humanity,

02:38:04.780 | not just the science, of the impact of the results.

02:38:08.760 | - One famous psychologist wrote about this issue,

02:38:12.360 | that somehow a balance has to be found

02:38:15.960 | between pursuing the science and communicating it

02:38:19.940 | with respect to people's sensitivities.

02:38:22.000 | The legitimate sensitivities.

02:38:24.580 | Somehow, he didn't say how.

02:38:27.400 | - Somehow. - Somehow.

02:38:28.740 | And this is-- - Every part of that sentence,

02:38:30.380 | somehow, and balance is left up

02:38:34.960 | to the interpretation of the reader.

02:38:37.380 | Let me ask you, you said big questions,

02:38:39.420 | the biggest, or one of the biggest.

02:38:42.440 | We already talked about consciousness and intelligence,

02:38:46.380 | one of the most fascinating, one of the biggest questions,

02:38:48.420 | but let's talk about the why.

02:38:51.260 | Why are we here?

02:38:53.220 | What's the meaning of life?

02:38:55.200 | - Oh, I'm not gonna tell you.

02:38:56.980 | - You know, but you're not gonna tell me?

02:38:58.980 | This is very, I'm gonna have to wait for your next book.

02:39:03.380 | - The meaning of life, you know,

02:39:05.540 | we do the best we can to get through the day.

02:39:09.480 | - And then there's just the finite number of the days.

02:39:15.380 | Are you afraid of the finiteness of it?

02:39:18.060 | You think about your death? - I think about it

02:39:19.140 | more and more as I get older.

02:39:21.180 | - Yeah, I do.

02:39:22.500 | And it's one of these human things,

02:39:26.100 | that it is finite, we all know it.

02:39:28.400 | Most of us deny it and don't wanna think about it.

02:39:35.300 | Sometimes you think about it in terms of estate planning,

02:39:39.340 | you try to do the rational thing.

02:39:41.780 | Sometimes it makes you work harder

02:39:44.580 | 'cause you know your time is more and more limited

02:39:46.900 | and you wanna get things done.

02:39:50.740 | I don't know where I am on that.

02:39:53.500 | It is just one of those things

02:39:57.260 | that's always in the back of my mind.

02:39:59.660 | And I don't think that's uncommon.

02:40:03.120 | - Well, it's just like G-factor intelligence,

02:40:06.100 | it's a hard truth that's there.

02:40:09.300 | And sometimes you kinda walk past it

02:40:12.300 | and you don't wanna look at it, but it's still there.

02:40:15.300 | - Yeah, yes, you can't escape it.

02:40:19.660 | And the thing about the G-factor intelligence

02:40:22.780 | is everybody knows this is true on a personal daily basis.

02:40:27.780 | Even if you think back to when you were in school,

02:40:33.720 | you know who the smart kids were.

02:40:36.760 | When you are on the phone talking

02:40:40.860 | to a customer service representative

02:40:43.660 | that in response to your detailed question

02:40:46.700 | is reading a script back to you

02:40:49.300 | and you get furious at this.

02:40:51.420 | Have you ever called this person a moron

02:40:54.740 | or wanted to call this person a moron?

02:40:56.820 | You're not listening to me.

02:40:58.340 | Everybody has had the experience of dealing with people

02:41:01.380 | who they think are not at their level.

02:41:03.840 | It's just common because that's the way human beings are.

02:41:09.180 | That's the way life is.

02:41:11.020 | - But we also have a poor estimation

02:41:15.360 | of our own intelligence.

02:41:16.620 | We have a poor, we're not always a great,

02:41:19.220 | our judgment of human character of other people

02:41:22.980 | is not as good as a battery of tests.

02:41:26.420 | That's where bias comes in.

02:41:31.260 | That's where our history, our emotions,

02:41:34.820 | all of that comes in.

02:41:35.740 | So people on the internet,

02:41:37.940 | there's such a thing as the internet

02:41:39.900 | and people on the internet will call each other dumb

02:41:42.580 | all the time.

02:41:43.420 | And that's the worry here

02:41:48.420 | is that we give up on people.

02:41:52.980 | We put them in a bin just because of one interaction

02:41:56.700 | or some small number of interactions

02:41:59.380 | as if that's it, they're hopeless.

02:42:01.900 | That's just in their genetics.

02:42:03.440 | But I think no matter what the science here says,

02:42:07.980 | once again, that does not mean we should not have compassion

02:42:12.820 | for our fellow man.

02:42:14.860 | - That's exactly what the science does say.

02:42:17.540 | It's not opposite of what the science says.

02:42:22.060 | Everything I know about psychology,

02:42:25.020 | everything I've learned about intelligence,

02:42:28.660 | everything points to the inexorable conclusion

02:42:32.680 | that you have to treat people as individuals respectfully

02:42:37.680 | and with compassion because through no fault of their own,

02:42:42.180 | some people are not as capable as others.

02:42:45.120 | And you wanna turn a blind eye to it,

02:42:48.620 | you wanna come up with theories

02:42:51.500 | about why that might be true, fine.

02:42:54.300 | I would like to fix some of it as best I can.

02:42:57.900 | - And everybody is deserving of love.

02:43:02.140 | Richard, this is a good way to end it, I think.

02:43:06.020 | - I'm just getting warmed up here, wasn't I?

02:43:07.540 | - I know.

02:43:08.380 | I know you can go for another many hours,

02:43:11.820 | but to respect your extremely valuable time,

02:43:15.180 | this was an amazing conversation.

02:43:16.620 | Thank you for the teaching company,

02:43:19.980 | the lectures you've given

02:43:22.340 | with the New York Science of Intelligence,

02:43:24.460 | just the work you're doing.

02:43:25.900 | It's a difficult topic,

02:43:28.060 | it's a topic that's controversial and sensitive to people,

02:43:31.380 | and to push forward boldly and in that nuanced way,

02:43:35.060 | just thank you for everything you do.

02:43:36.840 | And thank you for asking the big questions

02:43:39.420 | of intelligence, of consciousness.

02:43:42.140 | - Well, thank you for asking me.

02:43:43.460 | I mean, there's nothing like good conversation

02:43:45.540 | on these topics.

02:43:46.620 | - Thanks for listening to this conversation

02:43:48.980 | with Richard Heyer.

02:43:50.380 | To support this podcast,

02:43:51.500 | please check out our sponsors in the description.

02:43:54.380 | And now, let me leave you with some words

02:43:56.220 | from Albert Einstein.

02:43:58.260 | "It is not that I'm so smart,

02:44:00.540 | "but I stay with the questions much longer."

02:44:03.340 | Thank you for listening, and hope to see you next time.

02:44:07.260 | (upbeat music)

02:44:09.840 | (upbeat music)

02:44:12.420 | [BLANK_AUDIO]

Richard Haier: IQ Tests, Human Intelligence, and Group Differences | Lex Fridman Podcast #302

Chapters