Close
Unknown Speaker 0:07
Hi, you're listening to a podcast from the Department of Statistical Science here at University College London. My name is Dr. Nathan Green. I'm a statistician here. I'm joined by Dr. Samuel Livingston.
Hello.
Hi, Sam. And we're very pleased today to be talking to Professor Tom Fearn. Also from the department, we're going to be talking about eugenics here at UCL, and in particularly connection with the statistics department. But first of all, Tom, could you please introduce yourself?
Unknown Speaker 0:42
Yes. I'm Professor of Applied Statistics here at UCL. I mean, this statistics you've gotten for what seems like forever, but I looked it up and it's 1989. Yeah, it is forever. Spells five year spells as head of department. About halfway. I'm free of that now. And my main connection with eugenics is that in 2018, I was invited to join the eugenics inquiry that UCL sets up at that point. I guess the other thing I want to say is a disclaimer. I'm not a historian. I'm, if I know anything about eugenics, it's mainly through having taken part in this inquiry and having done a bit of reading around it, but I'm certainly not a historian of the history of science or anything else.
Unknown Speaker 1:30
Well, thanks, Tom. I was going to say, you were the head of department when I joined a long time ago. But I think you've been head of department for a bit longer than that. Even so, we're talking about eugenics today. So I think it's probably natural to start with defining it. So can you just tell us what eugenics is, broadly speaking,
Unknown Speaker 1:51
It will have to be broadly. So being on a thing called the eugenics inquiry was very interesting. We interviewed a lot of expert witnesses. And every one of 40 odd expert witnesses got asked to define eugenics. And needless to say, we got about 40 definitions. But broadly, it's about improving the quality of the human stock by selective breeding, ie by controlling or influencing, influencing in some way who gets to breed and you can have what Galton called positive eugenics, which is encouraging the right, in inverted commas, people to have more children. And you can have negative eugenics, which is encouraging the inverted commas wrong people opt to have too many children. The two enormous problems here, of course, one is you have to define who who are the right people, and who are the wrong people. And clearly what happens is, you define right as people who look like you, it's not entirely about race, it can be about disability, either physical or mental. It can be simply about being poor, and it being your fault for being poor, and therefore you shouldn't have too many children. But there is a very strong racial element in it. The other big problem is that it's on the negative eugenics side, it's about the discouraging, what does discouraging mean? It can range from sort of gentle discouragement, or it can lead to the Nazi death camps, which is a pretty strong discouragement from breeding. So it has had pretty awful consequences when people try to implement it. So that's all a bit of a crude outline, and I don't really want to go into them, but there are subtleties. So what counts as eugenics? So for example, if you give counselling to a couple who are at genetic risk of producing offspring with inherited diseases, is that eugenics or is it not? Is it good or is it bad? And you can certainly get into arguments around the margins.
Unknown Speaker 4:07
You mentioned Galton, and I know he's a central figure here but when did this idea of designing eugenics really become popular? And with Galton in modern times, but the idea dates back?
Unknown Speaker 4:16
I think, if you try to look it up the the texts say that the earliest recorded writings about it are Plato 400 BC, suggesting selective breading of I think soldiers would be a good idea. But really, the modern interest was stimulated by Darwin's publication of origin of the species and and I've got this written down 1859 the ideas arising from that which are that you can the way that the pressures of natural selection improve a species. So this idea is on the twin idea one is the way we're running society is maybe taking off some of these pressures of natural selection. And so maybe that's going to lead to the species, either human race, not improving anymore, in fact, deteriorating. And the more active side of it is, well, can we apply some pressures of our own and improve the human race, where of course it doesn't mean human race, it means our particular corner of the human race. So started in the late 1800s that Galton invented, the word eugenics in the again its written down 1883. And the modern interest began in the end of the 19th, beginning of the 20th century. It started in the UK, it started with Galton but it very rapidly spread to other countries, notably North America and most of Europe, Australia as well.
Unknown Speaker 5:55
Is it worth just noting who Galton is for some people that might not be familiar.
Unknown Speaker 6:01
Francis Galton was a archetypal Victorian scientist. So he never held an academic job, although he had close associations with UCL and I didn't know where his money came from. Once upon a time, you certainly didn't need a job. I think it came from arms manufacturing, gun manufacturing in Wolverhampton. And he was a really very interesting person he was a polymath. He was responsible for weather maps, early use of weather maps, he was responsible for putting on a scientific base, the use of fingerprints in crime investigations, and lots and lots of other things. He was really ingenious, he was very capable of thinking right outside the box, and coming up with lots of new things. And one of the new things he came up with was eugenics.
Unknown Speaker 6:54
And I suppose the natural question to ask next would be when did eugenics ideas fall out of favour?
Unknown Speaker 7:01
Well, they haven't entirely but they they took a big hit in the 1930s when Nazi Germany started getting really enthusiastic about these ideas for began to see where they were leading. And a lot of people who have been supporting it have been really interested in it changed their mind at that point, but it didn't disappear. So one of the things that didn't didn't disappear for a long time, was forcible sterilisation, particularly of people who were considered mentally unfit to breed that went on, certainly in the US, Canada and Sweden until the 1970s. And probably later in some other places, but they're the ones that are fairly well documented. And the ideas of eugenics stuck, scientific racism, haven't gone away. There's a really nice book by Angela Saini called Superior, which describes some of the modern ideas of scientific racism, people are still trying to prove there’re superior races.
Unknown Speaker 8:03
What's the kind of bit closer to home and you mentioned Galton, these connections with eugenics here in the UK.
Unknown Speaker 8:12
Well, he was sort of the high priest of eugenics. And Karl Pearson, which is a name I guess, most people listening to this will know, was his very ardent disciple. And if I'm just using a couple of religious analogies, that's because Galton actually described his science of eugenics in religious terms. It was the new religion used the word Jihad holy war at some point, and it was the word fanatics brings to mind. And there was supporters in the end of the 19th, beginning of the 20th century, from all shades of the political spectrum, so that scientific racism these days is sort of associated with the political right. But then it was from the right on the left, and all sorts of people and all sorts of big names. So to pull a few out of that Winston Churchill, John Maynard Keynes, Marie Stopes, quite a surprise to me that when I discovered that I'd always thought of her as an archetypal feminist and the birth control was about liberating women to have control of their own bodies. Birth control was about stopping the poor breeding. Basically, she was a eugenicist. She may have done good as well.
Unknown Speaker 9:26
But it's always a bit weird to learn about how blended some of these ideas are back then, but I guess we don't. The next thing to think about is really, as a discipline of statistics, right? We're a department of statistical science. And everything you've said is very interesting. It's not quite obvious how it connects to statistics as a discipline. So maybe you could explain a bit more about that.
Unknown Speaker 9:52
Yes, again, the names are Galton and Pearson's. Galton really wanted to put this new science on a firm scientific basis, that firm scientific basis was going to move on modern statistics. Basically, he and Pearson were interested in things like heritability. So what proportion of people's characteristics traits are inherited in what proportion and develop nature, the nature versus nurture argument. And between them, they grossly overestimated the influence of nature, either inherited part of people's traits, people's intelligence, for example, people's ability in other ways. And the other thing that we're really interested in was proving that some groups were different, and what different sorts of means better, you're sort of trying to prove they're a superior race. And to do that, you need to prove racial differences. And to do that you need things, tests. And so Galton in person between them, invented a lot of the common tools of modern statistics, things like regression, things like p-values, in order to put a scientific foundation on the eugenics, they're wonderful tools. They're very useful. They're still used, but they were invented for the wrong reasons. Some of them anyway, so the early history of modern statistics is just inextricably linked with eugenics.
Unknown Speaker 11:28
And is this connection why the statistics department at UCL is connected with eugenics?
Unknown Speaker 11:36
Yes, so as I said, Galton never had an academic post. But he had a eugenics laboratory, where they did research into eugenics, including measuring lots of people. So they would know we'd measure people's eyesight and head shape and all sorts of other things and collect lots of what we now call biometric data. And that was attached to UCL in a fairly loose sort of way. More importantly, when for the department. When Galton died in 1911, he left his not inconsiderable fortune, part of it, we UCL or possibly to the University of London, people seem a bit vague about that, to fund the Galton professorship of eugenics. And that was funded at UCL and Karl Pearson was the first holder of that, because Galton although he couldn't specify who should get it, he left a rather strong hint. And so the expected candidate applied and got the job, things things were a bit looser in those days about appointing people to post. And that led directly to the foundation of Department of Applied Statistics in 1911. If it hadn't been forgotten, if it hadn't been for eugenics, there wouldn't have been a Department of Statistics in 1911. And it was the first Department of Statistics in the world. And we're proud of that. But the leading it was a Galton professor of eugenics.
Unknown Speaker 13:04
It's quite funny, because it sounds like they weren't doing very good statistics at this time, sort of following a lot of the principles that we would like to follow as a discipline today getting a lot of things wrong.
Unknown Speaker 13:19
I think, yes. I mean, they're, well, it's hard to forgive both for quite a lot of things. But one of the things it's hard to forgive them for is doing some very bad science. They were guilty of selecting data. They were guilty of defining their variables. I mean, this is social statistics. It's trying to compare different groups of people. So you have to decide what variables you're going to use. And then we're guilty of selecting the variables to prove what they wanted. And they were guilty of Galton, in particular, of making amazing inferences. So Galton was really keen on things running in families. And he was partly keen on that, because he was part of a very famous family. One of my, the people who I met on the eugenics inquiry was fond of referring to Galton as Darwin's racist cousin. He was a second cousin, but he was very proud of being part of that family. And he studied what run in families, and he noticed things like being a High Court Judge tends to run in families conclusion from that was the one inherits the genes for being a high court judge didn't occur to him that there might be other possible explanations for things like that.
Unknown Speaker 14:40
Yeah, so regression is strange I mean, this is kind of related to regression as the name right for the process of fitting a linear model to data.
Unknown Speaker 14:48
Yes, but I mean, originally what Galton was looking at was the relationship between the heights of fathers and sons. So it was actually regression to the mean. What if you got a very tall father, the son will still be tall but a bit less tall. And so the word regression did have some meaning. But when you go in, but obviously it also involves fitting straight lines to things. And that's the bit that's persistent.
Unknown Speaker 15:15
Well, you mentioned the UCL eugenics inquiry. So I think it's natural to sort of move the discussion on to that at this point. So can you tell us a little bit about what that is? And what it did?
Unknown Speaker 15:27
Yes, it was, it was set up in 2008. And it’s brief was to investigate UCL’s role in the history of eugenics and to make some recommendations about what UCL might do about that. But particularly about things like renaming buildings, and there's a person where there was a person building, the department had to go to the lecture theatre, under other things named after them around the place. It employed a couple of researchers who looked into the history and actually carried out a survey which were with the aim of investigating what opinions about it work about what one should do, were both inside and outside UCL. And he took evidence from 40 odd expert witnesses, underway to town hall meetings as well. And then it reported in 2019, and made a number of recommendations, which may be too many recommendations, but broadly, it recommended that UCL should issue a formal apology. And that's actually happened, only recently happened in January this year, January 2021. It recommended stopping trying to sweep things under the carpet, telling everybody about this, and in particular, telling all our staff and students. And so the recommendation that students really ought to be aware of this bit of our history, it was pretty clear on the naming, we should get rid of names of buildings, prizes, lecture rooms, as named after prominent in eugenesists. And there were also some other sort of obvious recommendations were one was that UCL should support further research into the history of eugenics. And also that it should, UCL in my view is pretty good on equality and diversity. But again, the inquiry recommended, we should be even better, because eugenics has traditionally discriminated against minorities, that we should be very careful that we're not doing that anymore. It recorded into the 19. The Provost who was Michael Arthur accepted all the recommendations. And just as everybody was getting around to implementing them COVID hit and all the people who would have been in charge of taking this forward had lots of other things to do like trying to work out how on earth we were going to run a university when people couldn't come to it, some things have happened, there was a formal apology has been made. The Pearson buildings been de-named, it hasn't yet been renamed, I imagine it's quite quick to de-name something you just take the name down. Renaming it involves committees arguing about what the new name should be, I imagine that might take a long time, they used to have a Galton lecture theatre, that's been de named. And in fact, I saw a photograph this morning of the notice that's gone up outside the Galton lecture theatre, which you know, room 115 explaining why it's been being named. So that's also happened. Some things have happened on the educational front, in that all the stats department intake. Last session in September 2020 got a short talk on the history of the department with a focus on eugenics. The UCL centrally is working on some new material to include in the introductory programme that all new students get an option on during the summer before they come. They're the things I'm aware of, I'm sure there are other things going on in the background. But if it hadn't been for COVID, I might have been prompting a few people to ask what's going on. But it does seem a bit unfair to poke people at the moment because I know how busy they are.
Unknown Speaker 19:18
Lets bring this back to this statistics. You mentioned like some of these, the thing that was the first statistics department in the world and some of these methods that people are introduced to kind of statistics 101. So what do you think about that and what's the general feeling about kind of separating the theory from these historical figures, like the people that have come up with these ideas and the ideas themselves.
Unknown Speaker 19:47
Yes, again, that was another another area where the different expert witnesses gave very varied views on one extreme, or we shouldn't worry about this. It was only a little tiny bit of what they did. And so let's forget it to the other end, we shouldn't be using regression because it was invented by Galton. Now, not neither of those extremes appeals to me, I don't personally have a lot of problem in separating people's good ideas from their bad ideas or the good ideas from the people. I'm extremely happy to go on teaching regression, despite the fact Galton invented it. I do not want to do it in in a lecture theatre that's named after Galton. That seems to be the right separation, it gets harder in some areas, like it's going off topic, but you know, what's your attitude to Wagner's music, and that's harder to separate them, the person and the music. But I think in the case of this, this statistics, it is easy to say there were tools that were invented in order to do something nasty, but they're still valuable tools. tools as a neutral. It's what you do with them. That transforms me.
Yes. What about in the machine gun, for example?
Unknown Speaker 21:13
Yeah. Okay. I mean, I guess I was thinking about mathematical tools.
Unknown Speaker 21:19
Okay. Right. Yeah. I mean, I want to get back to that a little bit. Because I think there's more to be said, but I think before we do, we should, it's kind of wrapped it with the question of what we do next, as a discipline, you know, kind of we've known this, this has been something that everyone could have learned about for a long time, it feels like it's only recently that we've been sort of insisting more that people should learn about it. What happens next?
Unknown Speaker 21:48
Yes, again, one of the things that surprised me. So when I joined this inquiry, I tried circulating numbers of stuff in the statistics department to say you did have things to input. And I was surprised by how many people actually didn't know anything about it. And I think most people knew a bit, there were some people who really didn't know anything. And we were quite surprised and shocked. So I think the statistics community as a whole is very effectively swept this under a carpet for a while. It's been something we don't talk about. And I think, what we are changing now, certainly, at UCL, we're changing, and it's happening in other places. And it's happening to other people. So it is changing, but I think we need to make sure that it, it carries on being open, and that we don't go back to trying to hide it. The other thing we can do is learn some lessons. I mean, as we've already said, we're as I've already said, Pearson And Galton did some very bad statistics, I think we need to be careful, we don't do bad statistics. And the one thing one of the things we can try not to copy is going into a data analysis with an open mind. So it's really easy to see in the data, what you want to see that to look for the things that will support the theory you wanted to prove when you started and maybe ignore the things that seem to conflict with it. And it's quite hard to do that. But I think the lesson from what they did from the way they misled themselves, they grossly overestimated the effect of heredity. And that was they wanted it to be big. Because if it isn't, eugenics doesn't work, but then doesn't work. So they had to sort of prove heredity was dominant, managed to and that's a we need to avoid doing things like that. And I think the other thing, the other lesson we can take from it. But one of the other lessons is we need to think about where what the research that we're doing might lead what it might get used for. Even if we're doing mathematics, you can't pretend that the mathematics is completely neutral. It doesn't matter what people use this for. It's just interesting in its own right. So I've got no particular axe to grind on my example. But it's facial recognition that there's a computer applicant, an artificial intelligence application machine learning application, which has got enormous potential to be useful. And it's got enormous potential to be used for some very dubious purposes for purposes that I wouldn't want to have contributed to. And I'm certainly not saying nobody should work on facial recognition. But I am saying that if you're going to work on that, and on lots of other things you should think about whether you want to work on it or not just assume that this is just research. This is just mathematics. I don't have to think about the consequences? Because do I think whatever you do, and maybe in a lot of cases, it doesn't matter. Maybe not a case it doesn't foreseeably lead anywhere nasty. But there are a lot of cases where it does.
Unknown Speaker 25:11
Yeah, I think that's really interesting. I mean, that's another podcast, I think these applications of algorithms and data, and how it's affecting our lives in the black box algorithm making our decision at the bank or an exam mark. You know, it's kind of insidious, but I mean, it can be terribly biassed. And discriminating.
Unknown Speaker 25:34
Yes. And there's the whole question of algorithmic decision algorithms that are based on the input of data, which tend to reinforce what's been going on and return a well known example is policing algorithms that tell the police where they should concentrate on. And of course, it will tell them to concentrate on wherever arrested loads of people before, which is where they concentrated on before. And that's maybe because they were racist. And so it embeds racism into an objective algorithm. And that's happening to lots of these algorithms. There's nothing wrong with the algorithm. It's what you train it on.
Unknown Speaker 26:13
Everything is not all bad. I think there's lessons people are learning. It's such a new field.
Unknown Speaker 26:20
Yeah. That's it. That's a well known problem. And clearly, there are lots of people working on so. For example Ricardo Silva, you mean Yeah, Ricardo is a another member of our department. I wanted to just come back to the the discussion we were having about prominent contributions to statistics, because for example, the main person, you know, you can't do an introductory statistics course without learning about Pearson's chi squared, Pearson's correlation. We're de naming buildings named after these people. How do you feel about mathematical tools named after these people? Do you think we shouldn't be calling them after who develops them? I mean, it's a difficult question, right?
Unknown Speaker 27:03
Yes, that's, that's a more difficult one. And then things like buildings, I don't see why they shouldn't keep their names to be honest. The names just telling you who invented this and because of the correlation, it's telling you which correlation it is, lots of them. You could call it product moment correlation, I suppose. And maybe I don't think one should try to ban that individual. People want to call it product moment correlation. Just to avoid using Pearson’s name then I think that's fine, because we personally have to be careful because there are two Pearson’s there's Karl who incidentally was was christened Carl person with a C and change the C to a K and Egon Pearson, his son, so the headship of the statistics, dynasty. Karl Pearson was had for 20 odd years and then his son was had for I don't know how many years but certainly over 10 quite a long time. But don't think Egon Pearson didn't have any involvement with eugenics at all. So you can still use Egon Pearson if you're worried about naming. Yeah, at Pearson chi square, totally up to you.
Unknown Speaker 28:09
Okay, so well, in the interest of time, I think we'll probably wrap it up. Just to say thank you very much for speaking to us, Tom. It has been massively informative and very interesting. Thanks for your time.
My pleasure.
Unknown Speaker 28:24
UCL Minds brings together the knowledge, insights and ideas of our community through a wide range of events and activities that are open to everyone.
Unknown Speaker 0:07
Hi, you're listening to a podcast from the Department of Statistical Science here at University College London. My name is Dr. Nation Green. I'm a statistician here. I'm joined by Dr. Samuel Livingston.
Hello.
Hi, Sam. And we're very pleased today to be talking to Professor Tom Fearn. Also from the department, we're going to be talking about eugenics here at UCL, and in particularly connection with the statistics department. But first of all, Tom, could you please introduce yourself?
Unknown Speaker 0:42
Yes. I'm Professor of Applied Statistics here at UCL. I mean, this statistics you've gotten for what seems like forever, but I looked it up and it's 1989. Yeah, it is forever. Spells five year spells as head of department. About halfway. I'm free of that now. And my main connection with eugenics is that in 2018, I was invited to join the eugenics inquiry that UCL sets up at that point. I guess the other thing I want to say is a disclaimer. I'm not a historian. I'm, if I know anything about eugenics, it's mainly through having taken part in this inquiry and having done a bit of reading around it, but I'm certainly not a historian of the history of science or anything else.
Unknown Speaker 1:30
Well, thanks, Tom. I was going to say, you were the head of department when I joined a long time ago. But I think you've been head of department for a bit longer than that. Even so, we're talking about eugenics today. So I think it's probably natural to start with defining it. So can you just tell us what eugenics is, broadly speaking,
Unknown Speaker 1:51
It will have to be broadly. So being on a thing called the eugenics inquiry was very interesting. We interviewed a lot of expert witnesses. And every one of 40 odd expert witnesses got asked to define eugenics. And needless to say, we got about 40 definitions. But broadly, it's about improving the quality of the human stock by selective breeding, ie by controlling or influencing, influencing in some way who gets to breed and you can have what Galton called positive eugenics, which is encouraging the right, in inverted commas, people to have more children. And you can have negative eugenics, which is encouraging the inverted commas wrong people opt to have too many children. The two enormous problems here, of course, one is you have to define who who are the right people, and who are the wrong people. And clearly what happens is, you define right as people who look like you, it's not entirely about race, it can be about disability, either physical or mental. It can be simply about being poor, and it being your fault for being poor, and therefore you shouldn't have too many children. But there is a very strong racial element in it. The other big problem is that it's on the negative eugenics side, it's about the discouraging, what does discouraging mean? It can range from sort of gentle discouragement, or it can lead to the Nazi death camps, which is a pretty strong discouragement from breeding. So it has had pretty awful consequences when people try to implement it. So that's all a bit of a crude outline, and I don't really want to go into them, but there are subtleties. So what counts as eugenics? So for example, if you give counselling to a couple who are at genetic risk of producing offspring with inherited diseases, is that eugenics or is it not? Is it good or is it bad? And you can certainly get into arguments around the margins.
Unknown Speaker 4:07
You mentioned Galton, and I know he's a central figure here but when did this idea of designing eugenics really become popular? And with whom gotten in modern times, but the idea dates back?
Unknown Speaker 4:16
I think, if you try to look it up the the texts say that the earliest recorded writings about it are Plato 400 BC, suggesting selective breading of I think soldiers would be a good idea. But really, the modern interest was stimulated by Darwin's publication of origin of the species and and I've got this written down 1859 the ideas arising from that which are that you can the way that the pressures of natural selection improve a species. So this idea is on the twin idea one is the way we're running society is maybe taking off some of these pressures of natural selection. And so maybe that's going to lead to the species, either human race, not improving anymore, in fact, deteriorating. And the more active side of it is, well, can we apply some pressures of our own and improve the human race, where of course it doesn't mean human race, it means our particular corner of the human race. So started in the late 1800s that Galton invented, the word eugenics in the again its written down 1883. And the modern interest began in the end of the 19th, beginning of the 20th century. It started in the UK, it started with Galton but it very rapidly spread to other countries, notably North America and most of Europe, Australia as well.
Unknown Speaker 5:55
Is it worth just noting who Galton is for some people that might not be familiar,
Unknown Speaker 6:01
Francis Galton was a archetypal Victorian scientist. So he never held an academic job, although he had close associations with UCL and I didn't know where his money came from. Once upon a time, you certainly didn't need a job. I think it came from arms manufacturing, gun manufacturing in Wolverhampton. And he was a really very interesting person he was a polymath. He was responsible for weather maps, early use of weather maps, he was responsible for putting on a scientific base, the use of fingerprints in crime investigations, and lots and lots of other things. He was really ingenious, he was very capable of thinking right outside the box, and coming up with lots of new things. And one of the new things he came up with was eugenics.
Unknown Speaker 6:54
And I suppose the natural question to ask next would be when did eugenics ideas fall out of favour?
Unknown Speaker 7:01
Well, they haven't entirely but they they took a big hit in the 1930s when Nazi Germany started getting really enthusiastic about these ideas for began to see where they were leading. And a lot of people who have been supporting it have been really interested in it changed their mind at that point, but it didn't disappear. So one of the things that didn't didn't disappear for a long time, was forcible sterilisation, particularly of people who were considered mentally unfit to breed that went on, certainly in the US, Canada and Sweden until the 1970s. And probably later in some other places, but they're the ones that are fairly well documented. And the ideas of eugenics stuck, scientific racism, haven't gone away. There's a really nice book by Angela Saini called Superior, which describes some of the modern ideas of scientific racism, people are still trying to prove there’re superior races.
Unknown Speaker 8:03
What's the kind of bit closer to home and you mentioned Galton, these connections with eugenics here in the UK.
Unknown Speaker 8:12
Well, he was sort of the high priest of eugenics. And Karl Pearson, which is a name I guess, most people listening to this will know, was his very ardent disciple. And if I'm just using a couple of religious analogies, that's because Galton actually described his science of eugenics in religious terms. It was the new religion used the word Jihad holy war at some point, and it was the word fanatics brings to mind. And there was supporters in the end of the 19th, beginning of the 20th century, from all shades of the political spectrum, so that scientific racism these days is sort of associated with the political right. But then it was from the right on the left, and all sorts of people and all sorts of big names. So to pull a few out of that Winston Churchill, John Maynard Keynes, Marie Stopes, quite a surprise to me that when I discovered that I'd always thought of her as a archetype or feminist and the birth control was about liberating women to have control of their own bodies. Birth control was about stopping the poor breeding. Basically, she was a eugenicist. She may have done good as well.
Unknown Speaker 9:26
But it's, it's always a bit weird to learn about how blended some of these ideas are back on them, but I guess we don't. The next thing to think about is really, as a discipline of statistics, right? We're a department of statistical science. And everything you've said is very interesting. It's not quite obvious how it connects to statistics as a discipline. So maybe you could explain a bit more about that.
Unknown Speaker 9:52
Yes, again, the names are Galton and Pearson's. Galton really wanted to put these new science of use Next on a firm scientific basis, that firm scientific basis was going to move on modern statistics. Basically, he and Pearson were interested in things like heritability. So what what proportion of people's characteristics traits are inherited in what proportion and develop nature, the nature versus nurture argument. And between them, they grossly overestimated the influence of nature, either inherited part of people's traits, people's intelligence, for example, people's ability in other ways. And the other thing that we're really interested in was proving that some groups were different, and what different sorts of means better, you're sort of trying to prove they're a superior race. And to do that, you need to prove racial differences. And to do that you need things, tests. And so Galton in person between them, invented a lot of the common tools of modern statistics, things like regression, things like p-values, in order to put a scientific foundation on the eugenics, they're wonderful tools. They're very useful. They're still used, but they were invented for the wrong reasons. Some of them anyway, so the early history of modern statistics is just inextricably linked with eugenics.
Unknown Speaker 11:28
And is this connection why the statistics department at UCL is connected with eugenics?
Unknown Speaker 11:36
Yes, so as I said, Galton never had an academic post. But he had a eugenics laboratory, where they did research into eugenics, including measuring lots of people. So they would know we'd measure people's eyesight and head shape and all sorts of other things and collect lots of what we now call biometric data. And that was attached to UCL in a fairly loose sort of way. More importantly, when for the department. When Galton died in 1911, he left his not inconsiderable fortune, part of it, we UCL or possibly to the University of London, people seem a bit vague about that, to fund the Galton professorship of eugenics. And that was funded at UCL and Karl Pearson was the first holder of that, because Galton although he couldn't specify who should get it, he left a rather strong hint. And so the expected candidate applied and got the job, things things were a bit looser in those days about appointing people to post. And that led directly to the foundation of Department of Applied Statistics in 1911. If it hadn't been forgotten, if it hadn't been for eugenics, there wouldn't have been a Department of Statistics in 1911. And it was the first Department of Statistics in the world. And we're proud of that. But the leading it was a Galton professor of eugenics.
Unknown Speaker 13:04
It's quite funny, because it sounds like they weren't doing very good statistics at this time, sort of following a lot of the principles that we would like to follow as a discipline today getting a lot of things wrong,
Unknown Speaker 13:19
I think, yes. I mean, they're, well, it's hard to forgive both for quite a lot of things. But one of the things it's hard to forgive them for is doing some very bad science. They were guilty of selecting data. They were guilty of defining their variables. I mean, this is social statistics. It's trying to compare different groups of people. So you have to decide what variables you're going to use. And then we're guilty of selecting the variables to prove what they wanted. And they were guilty of Galton, in particular, of making amazing inferences. So Galton was really keen on things running in families. And he was partly keen on that, because he was part of a very famous family. One of my, the people who I met on the eugenics inquiry was fond of referring to Galton as Darwin's racist cousin. He was a second cousin, but he was very proud of being part of that family. And he studied what run in families, and he noticed things like being a High Court Judge tends to run in families conclusion from that was the one inherits the genes for being a high court judge didn't occur to him that there might be other possible explanations for things like that.
Unknown Speaker 14:40
Yeah, so regression is strange I mean, this is kind of related to regression as the name right for the process of fitting a linear model to data.
Unknown Speaker 14:48
Yes, but I mean, originally what Galton was looking at was the relationship between the heights of fathers and sons. So it was actually regression to the mean written. What if you got a very tall father, the son will still be tall but a bit less tall. And so the regress the word regression did have some meaning. But when you go in, but obviously it also involves fitting straight lines to things. And that's the bit that's persistent.
Unknown Speaker 15:15
Well, you mentioned the UCL eugenics inquiry. So I think it's natural to sort of move the discussion on to that at this point. So can you tell us a little bit about what that is? And what it did?
Unknown Speaker 15:27
Yes, it was, it was set up in 2008. And it’s brief was to investigate UCL’s role in the history of eugenics and to make some recommendations about what UCL might do about that. But particularly about things like renaming buildings, and there's a person where there was a person building, the department had to go to the lecture theatre, under other things named after them around the place. It employed a couple of researchers who looked into the history and actually carried out a survey which were with the aim of investigating what opinions about it work about what one should do, were both inside and outside UCL. And he took evidence from 40 odd expert witnesses, underway to town hall meetings as well. And then it reported in 2019, and made a number of recommendations, which may be too many recommendations, but broadly, it recommended that UCL should issue a formal apology. And that's actually happened, only recently happened in January this year, January 2021. It recommended stopping trying to sweep things under the carpet, telling everybody about this, and in particular, telling all our staff and students. And so the recommendation that students really ought to be aware of this bit of our history, it was pretty clear on the naming, we should get rid of names of buildings, prizes, lecture rooms, as named after prominent in eugenesists. And there were also some other sort of obvious recommendations were one was that UCL should support further research into the history of eugenics. And also that it should, UCL in my view is pretty good on equality and diversity. But again, the inquiry recommended, we should be even better, because eugenics has traditionally discriminated against minorities, that we should be very careful that we're not doing that anymore. It recorded into the 19. The Provost who was Michael Arthur accepted all the recommendations. And just as everybody was getting around to implementing them COVID hit and all the people who would have been in charge of taking this forward had lots of other things to do like trying to work out how on earth we were going to run a university when people couldn't come to it, some things have happened, there was a formal apology has been made. The Pearson buildings been de-named, it hasn't yet been renamed, I imagine it's quite quick to de-name something you just take the name down. Renaming it involves committees arguing about what the new name should be, I imagine that might take a long time, they used to have a Galton lecture theatre, that's been de named. And in fact, I saw a photograph this morning of the notice that's gone up outside the Galton lecture theatre, which you know, room 115 explaining why it's been being named. So that's also happened. Some things have happened on the educational front, in that all the stats department intake. Last session in September 2020 got a short talk on the history of the department with a focus on eugenics. The UCL centrally is working on some new material to include in the introductory programme that all new students get an option on during the summer before they come. They're the things I'm aware of, I'm sure there are other things going on in the background. But if it hadn't been for COVID, I might have been prompting a few people to ask what's going on. But it does seem a bit unfair to poke people at the moment because I know how busy they are.
Unknown Speaker 19:18
Lets bring this back to this statistics. You mentioned like some of these, the thing that was the first statistics department in the world and some of these methods that people are introduced to kind of statistics 101. So what do you think about that and what's the general feeling about kind of separating the theory from these historical figures, like the people that have come up with these ideas and the ideas themselves.
Unknown Speaker 19:47
Yes, again, that was another another area where the different expert witnesses gave very varied views on one extreme, or we shouldn't worry about this. It was only a little tiny bit of what they did. And so let's forget it to the other end, we shouldn't be using regression because it was invented by Galton. Now, not neither of those extremes appeals to me, I don't personally have a lot of problem in separating people's good ideas from their bad ideas or the good ideas from the people. I'm extremely happy to go on teaching regression, despite the fact Galton invented it. I do not want to do it in in a lecture theatre that's named after Galton. That seems to be the right separation, it gets harder in some areas, like it's going off topic, but you know, what's your attitude to Wagner's music, and that's harder to separate them, the person and the music. But I think in the case of this, this statistics, it is easy to say there were tools that were invented in order to do something nasty, but they're still valuable tools. tools as a neutral. It's what you do with them. That transforms me.
Yes. What about in the machine gun, for example?
Unknown Speaker 21:13
Yeah. Okay. I mean, I guess I was thinking about mathematical tools.
Unknown Speaker 21:19
Okay. Right. Yeah. I mean, I want to get back to that a little bit. Because I think there's more to be said, but I think before we do, we should, it's kind of wrapped it with the question of what we do next, as a discipline, you know, kind of we've known this, this has been something that everyone could have learned about for a long time, it feels like it's only recently that we've been sort of insisting more that people should learn about it. What happens next?
Unknown Speaker 21:48
Yes, again, one of the things that surprised me. So when I joined this inquiry, I tried circulating numbers of stuff in the statistics department to say you did have things to input. And I was surprised by how many people actually didn't know anything about it. And I think most people knew a bit, there were some people who really didn't know anything. And we were quite surprised and shocked. So I think the statistics community as a whole is very effectively swept this under a carpet for a while. It's been something we don't talk about. And I think, what we are changing now, certainly, at UCL, we're changing, and it's happening in other places. And it's happening to other people. So it is changing, but I think we need to make sure that it, it carries on being open, and that we don't go back to trying to hide it. The other thing we can do is learn some lessons. I mean, as we've already said, we're as I've already said, Pearson And Galton did some very bad statistics, I think we need to be careful, we don't do bad statistics. And the one thing one of the things we can try not to copy is going into a data analysis with an open mind. So it's really easy to see in the data, what you want to see that to look for the things that will support the theory you wanted to prove when you started and maybe ignore the things that seem to conflict with it. And it's quite hard to do that. But I think the lesson from what they did from the way they misled themselves, they grossly overestimated the effect of heredity. And that was they wanted it to be big. Because if it isn't, eugenics doesn't work, but then doesn't work. So they had to sort of prove heredity was dominant, managed to and that's a we need to avoid doing things like that. And I think the other thing, the other lesson we can take from it. But one of the other lessons is we need to think about where what the research that we're doing might lead what it might get used for. Even if we're doing mathematics, you can't pretend that the mathematics is completely neutral. It doesn't matter what people use this for. It's just interesting in its own right. So I've got no particular axe to grind on my example. But it's facial recognition that there's a computer applicant, an artificial intelligence application machine learning application, which has got enormous potential to be useful. And it's got enormous potential to be used for some very dubious purposes for purposes that I wouldn't want to have contributed to. And I'm certainly not saying nobody should work on facial recognition. But I am saying that if you're going to work on that, and on lots of other things you should think about whether you want to work on it or not just assume that this is just research. This is just mathematics. I don't have to think about the consequences? Because do I think whatever you do, and maybe in a lot of cases, it doesn't matter. Maybe not a case it doesn't foreseeably lead anywhere nasty. But there are a lot of cases where it does.
Unknown Speaker 25:11
Yeah, I think that's really interesting. I mean, that's another podcast, I think these applications of algorithms and data, and how it's affecting our lives in the black box algorithm making our decision at the bank or an exam mark. You know, it's kind of insidious, but I mean, it can be terribly biassed. And discriminating.
Unknown Speaker 25:34
Yes. And there's the whole question of algorithmic decision algorithms that are based on the input of data, which tend to reinforce what's been going on and return a well known example is policing algorithms that tell the police where they should concentrate on. And of course, it will tell them to concentrate on wherever arrested loads of people before, which is where they concentrated on before. And that's maybe because they were racist. And so it embeds racism into an objective algorithm. And that's happening to lots of these algorithms. There's nothing wrong with the algorithm. It's what you train it on.
Unknown Speaker 26:13
Everything is not all bad. I think there's lessons people are learning. It's such a new field.
Unknown Speaker 26:20
Yeah. That's it. That's a well known problem. And clearly, there are lots of people working on so. For example Ricardo Silva, you mean Yeah, Ricardo is a another member of our department. I wanted to just come back to the the discussion we were having about prominent contributions to statistics, because for example, the main person, you know, you can't do an introductory statistics course without learning about Pearson's chi squared, Pearson's correlation. We're de naming buildings named after these people. How do you feel about mathematical tools named after these people? Do you think we shouldn't be calling them after who develops them? I mean, it's a difficult question, right?
Unknown Speaker 27:03
Yes, that's, that's a more difficult one. And then things like buildings, I don't see why they shouldn't keep their names to be honest. The names just telling you who invented this and because of the correlation, it's telling you which correlation it is, lots of them. You could call it product moment correlation, I suppose. And maybe I don't think one should try to ban that individual. People want to call it product moment correlation. Just to avoid using Pearson’s name then I think that's fine, because we personally have to be careful because there are two Pearson’s there's Karl who incidentally was was christened Carl person with a C and change the C to a K and Egon Pearson, his son, so the headship of the statistics, dynasty. Karl Pearson was had for 20 odd years and then his son was had for I don't know how many years but certainly over 10 quite a long time. But don't think Egon Pearson didn't have any involvement with eugenics at all. So you can still use Egon Pearson if you're worried about naming. Yeah, at Pearson chi square, totally up to you.
Unknown Speaker 28:09
Okay, so well, in the interest of time, I think we'll probably wrap it up. Just to say thank you very much for speaking to us, Tom. It has been massively informative and very interesting. Thanks for your time.
My pleasure.
Unknown Speaker 28:24
UCL Minds brings together the knowledge, insights and ideas of our community through a wide range of events and activities that are open to everyone.