post

An uncomfortable truth

Statisticians have the wholly unreasonable reputation of being the reprobates of the mathematical world: the drummers of the band, the viola players of the orchestra, the Chalkdusts of the scientific magazines. Much of this contempt seems to stem from statistics not being seen as a ‘real’ discipline within mathematics; statistics is so often the butt of mathematical jokes that it has spawned an entire joke genre of the ‘lies, damned lies and statistics’ type (for example, ‘A statistician is a person who draws a mathematically precise line from an unwarranted assumption to a foregone conclusion’, or ‘Did you know that 87.166253% of all statistics claim a precision of results that is not justified by the method employed?’). A power struggle as to whether statistics is a field of mathematics, an entirely separate field, or a science (or something else) has been ongoing for decades and shows no sign of abating. It may come as little surprise, dear readers, that many of of us with a love of mathematics seem to also have a love for forensic debate about boundaries, classification, and domains.

But there is another, far more urgent power struggle that bridges the worlds of statistics and mathematics, too: the question of how we respond to the emergence of unpleasant historical details surrounding the lives of those we have previously considered heroes of our discipline. This is not a hypothetical question. Books and research papers from the last few decades have increasingly recognised that the founding fathers of statistics built the field on a dark and sinister foundation. The fact is: statistics as we know it is so intertwined with eugenics as to be inseparable. It could be said, in fact, that every time we speak the names of the men who gave their names to the statistical tests we might see as so innocuous, we are breathing life back into the memory of racists, obsessed with measuring ‘difference from the norm’, who built a ‘scientific’ reputation by metaphorically climbing on top of the broken bodies and reputations of entire peoples that they considered less than human.

Don’t believe us?

In Germany, an investigation into the workings of the organisation now called the Max Planck Society, before 1948 known as the Kaiser Wilhelm society, suggested scientists and academics there played a key role in legitimising Hitler’s ideology of racial purity. Not only did they contribute to the theoretical basis for Nazi politics, but in some cases they helped to set up concentration camps, gas chambers and conducted unimaginably cruel experiments on incarcerated humans. But this is Nazi Germany, you might say; nothing so evil could happen on British soil. (But you would be wrong.)

Angela Saini writes in their bestseller Superior: “The truth: that it is perfectly possible for prominent scientists to be racists, to murder, to abuse both people and knowledge—doesn’t sit easily with the way we like to think about scientific research. We imagine that it’s above politics, that it is a noble, rational, and objective endeavour, untainted by feelings or prejudice…[but] science sits at the mercy of the personal political beliefs of those carrying it out.”

Some of the scientists most influential to the thinking inherent in the Holocaust worked at UCL in Gower Street, London, and give their names to statistical measures we continue to teach in classrooms and lecture halls up and down the UK.

The genesis of eugenics

Francis Galton was the cousin of Charles Darwin. He coined the term ‘eugenics’ (‘good breeding’) to describe the process of “improving stock, which is by no means confined to questions of judicious mating, but which, especially in the case of man, takes cognisance of all influences that tend in however remote a degree to give the more suitable races or strains of blood a better chance of prevailing speedily over the less suitable than they otherwise would have had.” As you can see, the word ‘improving’ is used here in a way that depends on a hierarchical framework of ‘suitability’ of ‘races’—in other words, white supremacy.

The US National Human Genome Research Institute defines eugenics as “the scientifically inaccurate theory that humans can be improved through selective breeding of populations. Eugenicists believed in a prejudiced and incorrect understanding of Mendelian genetics that claimed abstract human qualities (eg intelligence and social behaviours) were inherited in a simple fashion.” While the scientific rebuttal alone should be sufficient to discredit eugenics as a discipline, there is a further ethical issue here: the underlying model of white supremacy. White supremacy is the idea that whiteness constitutes superiority and confers rightful power; but this concept of ‘whiteness’ is not just about skin colour. “Racism is based on the concept of whiteness—a powerful fiction enforced by power and violence. Whiteness is a constantly shifting boundary separating those who are entitled to have certain privileges from those whose exploitation and vulnerability to violence is justified by their not being white.” (Kivel, Uprooting Racism, 2011.)

This means that various peoples throughout history have been marginalised and oppressed as part of white supremacy, even those who might in other contexts be considered ‘white’. This is because race is not a scientific concept—it is a social construct, and remarkably hard to pin down. It is, as Jenée Desmond-Harris says, neither “permanent, scientific, objective, logical, consistent, [nor] able to stand up to scrutiny.” So why does race loom so large in our society and our histories? Desmond-Harris also argues, along with many others, that the concept of race itself was invented by scientists to justify slavery.

Intersections with racism

A Galton board in action: ball bearings are poured through the top and bounce through the pegs into a bell curve beneath.

Galton’s father made money from the slave trade, which he inherited and which allowed him to work as an explorer and a scientist. Galton travelled to Africa, describing the people who lived there as ‘savage races’. In our office, we have a Galton board: a lovely object, a dynamic normal graph, showing as you rotate it hundreds of bouncing, jostling ball bearings which eventually settle into a beautiful bell curve (see opposite). But both the words ‘Galton’ and ‘bell curve’ (the name of a 1994 book by Herrnstein and Murray explicitly connecting race and intelligence) don’t feel right in our mouths. They taste horrible. Would we be better off ignorant, not worrying about these issues and just enjoying the mathematical ideas here? Perhaps, as individuals. But, as Aris Winger, assistant professor of mathematics at Georgia Gwinnett College, says, ‘How white people interrogate their own racial identity is a requirement for  any meaningful racial, professional or personal progress in maths. The future of the discipline is tied to whiteness, how we understand whiteness, and white supremacy.’ Because we, as white people, have the luxury of not having to deal with racism in our professional lives—no one has ever assumed we are not mathematicians because of the colour of our skin. So we also have a painting in our office—the one you see displayed here—which helps remind us of some of these issues.

Karl Pearson is often described as the father of statistics; working as recently as the 1900s, in many ways he founded the discipline. He was the first professor of national eugenics at UCL. He was racist and antisemitic, writing that Jewish immigrants were a ‘parasitic race’ and that those who were not white were ‘inferior’.

His namesake, the Pearson product-moment correlation coefficient (or PMCC) is something we have both taught, and used in our research—it is the first port of call for those investigating linear correlation, or asking ‘might these two numbery things have a relationship?’ This concept has many, many names—but the ‘Pearson’ part seems remarkably sticky. It’s almost as if someone is doing a PR job behind the scenes. It is remarkable how insistent people can be on proper credit and value for scientific work when said work involves white men, and how these principles seem to vanish without trace when minoritised people are involved. It’s almost as if we have a skewed (stats joke) idea of what constitutes authority and status based on kyriarchy (the social system that keeps all intersecting oppressions in place).

Implications and complexities

In a world where toppling statues of slavers has become international news, it is worth considering what a scientist must do in order that we stop hallowing them in curricula, in assessment, in the simple act of speaking their name. Is a mathematical breakthrough ever neutral? Are there figures in our discipline for whom their ideologies are so distasteful that presenting their work to students, or in our research without proper contextualisation is in fact unethical, particularly in a subject like statistics where the two seem so very intertwined?

It is impossible, of course, and indeed undesirable to remove mathematicians’ ideas from the wider picture; but at what point does it become reasonable to strike their names from their most famous methods? Remove their faces from our textbooks? Or at least include historical notes alongside their mathematical legacies telling the dark stories of their distasteful ideologies which, in some cases, allowed their work to continue? If Galton’s work was created on the backs of murdered and dehumanised slaves, should all those using it be aware of this fact? If Pearson’s legacy was built on the idea that humans can be used as data without regard for their humanity, should that history lesson not stand next to his statistical methods, as a warning for all future statistics students not to make the same mistakes?

The empty pedestal of the statue of Edward Colston in Bristol, the day after its toppling in June 2020.

Indeed, these are not just hypothetical questions: in 2020, UCL announced that it was renaming buildings which had been named after both Pearson and Galton because of their connection with eugenics, and they have said this is one of many measures they plan to take in this vein. Of course the venerable institution with which we are associated, Cambridge University, does not escape critique; for example, there have been calls for the removal of the stained glass window in Gonville \& Caius College which commemorates statistician Ronald Fisher, who was similarly enamoured of eugenics. In 2020, Daniela Witten, professor of statistics at the University of Washington, ran a successful campaign for the Committee of Presidents of Statistical Societies to remove Fisher’s name from the Fisher award and lectureship (now called the COPSS distinguished achievement award and lectureship), saying “If Quaker Oats can retire Aunt Jemima, then statistics can retire the Fisher lecture. I hope that this tiny little gesture of a name change can be just the beginning of a much larger conversation in our field about how we can work to create a statistical community that is committed to our shared goals of diversity, equity, and inclusion.”

The felling of one stone monument to the slaver Edward Colston signalled for many the beginning of the important work of questioning which other figures’ more metaphorical pedestalling might similarly be challenged. The Colston statue was toppled by protesters in June 2020 in the context of longstanding campaigns by Bristol residents to remove references to Colston throughout the city. They advocated to remove the statue, or even just to have a small plaque added, pointing out that the 84,000 African lives he helped sell into slavery were worth considering as a counterpoint to his much celebrated ‘philanthropy’. However, the veneration of history’s white men has a certain inertia, and multiple attempts to draft a plaque that satisfied everybody failed spectacularly, with one local councillor claiming that vandalising it would be an appropriate course of action should it ever appear. The plaque was never mounted.

Possible futures

But what does this mean for mathematics, and in particular, statistics? Cases like that of Colston’s statue and others such as Cecil Rhodes’ statue outside Oriel College, Oxford, seem to demonstrate that long dead men can still count on the support of powerful institutions, individuals, and even government ministers whenever people have the temerity to threaten to peel a layer or two of varnish from their legacies and expose the rotten cores beneath. As mathematicians, maths communicators, and maths educators seeking to do the same within our discipline, perhaps we should expect the same forces to swing into action to defend the ‘good’ names of Galton, Pearson, Fisher and their ilk. How then do we develop the collective power of our community to call for change? What should our demands even be? And how do we counter the arguments that it was ever thus; that these people were born in different times; and that we should not seek to ‘vandalise’ our history by tit-for-tat retaliatory silences, overwrites, and heavy-handed edits?

We do not have the solution. But we are mathematicians (and statisticians, and we do not differentiate between the two): to us this is familiar ground. We humbly suggest the same approach we take to our beloved discipline of mathematics when we are stuck: listen, learn, think, spend time with the problem; most of all, seek out greater expertise.

One of the places we often turn to when navigating these issues is the excellent podcast Mathematically Uncensored. In the seemingly complicated and sometimes fraught discussion about inclusive practices in mathematics education, Aris Winger and Pamela E Harris, associate professor of mathematics at Williams College in Massachusetts and featured in Chalkdust issue 10, challenge us with a very simple question: what does it actually cost us to be inclusive? If we are only prepared to take action that is comfortable, that does not materially cost us in some way, then are we really making any meaningful challenge to the status quo? Ultimately, no matter how horrified we might be at reading some of these biographies, if we cannot even ‘give up’ somehow the names of some long dead men, or take action to prioritise the inclusion of those whose contributions have been historically marginalised or completely erased—do we really care at all?

The artwork below is a painting by Lucy Rycroft-Smith. The name of the piece is Galton Under Fire (2021), so named in reference to the frustration the artist feels at the apparently fireproof reputation of some mathematicians, Francis Galton included. It poses the question: how unethical do someone’s actions have to be before we choose to take action against them as a mathematical community, and what should that action be?