Transcript: 85. Liam Kofi Bright on Du Bois' Philosophy of Science

Note: this transcription was produced by automatic voice recognition software. It has been corrected by hand, but may still contain errors. We are very grateful to Tim Wittenborg for his production of the automated transcripts and for the efforts of a team of volunteer listeners who corrected the texts.
 

PA: Can you just start by reminding listeners of the kind of scientific work that Du Bois himself did? Because we're going to be talking about his philosophy of science, so let's talk about his science first.

LKB: Sure, absolutely. Du Bois was one of those people – very annoying people – who was good at everything and did a lot of things. And so his scientific work often nowadays is primarily thought of as in sociology – he's sort of one of the pioneers of US sociology. But he can also reasonably be said to have done criminology, history... Du Bois certainly counted his historical work as part of his scientific work. Other people may disagree, but Du Bois conceptualised it that way. And economics. I mean, maybe it's already been covered, but Du Bois has this interesting feature where he de facto had two PhDs, where one of his PhDs was from Berlin and was basically in economics or economic history. And then for bureaucratic reasons, he wasn't allowed to defend. And so he took a consolation prize PhD from Harvard in history. And then for bureaucratic reasons, he wasn't allowed to defend. And so he took a consolation prize PhD from Harvard in history, and that was kind of on the suppression of the Atlantic slave trades – it was on the legal and historical sociology of removing the slave trade in America. And so Du Bois covers this very broad range of topics in social life. Du Bois was very well known in his day for this work The Philadelphia Negro, which is a work of urban sociology. In this work, he famously combines both quantitative and qualitative methods to paint a broad picture of the African American life in the seventh ward of Philadelphia. And so this is kind of well representative of Du Bois' scientific work at its strongest, where he adopts a vast majority of research methods and does like really in-depth observational work on this population to try and tell a story about as many facets of their social life as possible. And so it sort of represents Du Bois as this kind of social theorist in the truest sense of where he's really trying to tell you about their economic life, their religious life, crime, the family, just everything he can tell you about this group. So Du Bois as a social scientist is like... really very broad reaching, but maybe it's most natural to call him a sociologist.

PA: Maybe we can say something also about why he thought this was interesting. There's an essay that he wrote called ‘The Study of the Negro Problems’. And in that he wrote that one reason to do what he was doing – to study the condition of what he would have called Negroes, what we would call African Americans in America – was for the sake of, as he put it, ‘the great end of advancing the cause of science in general’, which is a very bold claim. So what is the broader scientific interest of studying the social condition of African Americans?

LKB: Du Bois saw the African American people as a nation of a nation. And that sort of can almost be quasi-literal because there's this famous Paris exhibition in 1900 where Du Bois goes out and he presents a lot of sociological data on the African American people. And famously, he sort of... He represents Black people as a sort of mini-America within America visually. That's how he portrays African Americans. And so that really... the idea is that what we're seeing in America, because of the peculiar conditions of slavery and how that's led to a people being relatively isolated and suddenly introduced to a whole new mode of life, very dissimilar from what people had grown accustomed to in Africa. And when now after the end of the Civil War and the end of slavery, having to sort of build up new institutions for themselves, new modes of political life, their religion was adapting… And so he thought that what this offered social scientists was this like unique opportunity to see a nation develop from relatively simple beginnings right in front of their very eyes as African Americans had to sort of rapidly adapt to often circa the 1900s industrial civilization in America. And so I think Du Bois thought it was a really unusual to the extent that you could see a nation being born right there and then, and African Americans were that nation.

PA: So it's almost like an astronomer who's getting to watch a planet be formed or something.

LKB: That's a great analogy. Yeah.

PA: Okay. We're going to talk about a couple of specific philosophical issues that arise in his science now, both of which you've published on. And one of them is something that you refer to as the value free ideal. And to introduce this, I thought this is a good quote from Du Bois himself. A quote from Du Bois, not from you.

LKB: [Laughter]

PA: So he said, 'The aim of science as such is simple truth. Any attempt to give it a double aim to make social reform the immediate instead of the mediate object of a search for truth will inevitably tend to defeat both objects.' So what he's saying there basically is even if you've got some kind of social agenda for your social science, you should, when you're doing the social science, just pursue the truth and kind of forget about your agenda. So can you explain more why this would be something that now philosophers of science call a value free ideal? Like what does that mean? And why did he think scientists should adopt this value free ideal?

LKB: Right. So what was very important for Du Bois, the way he cashed out this idea that the immediate aim of working scientists should be to discover the truth – to pursue the truth – was that he thought it was very important that when deciding whether or not one had gathered sufficient evidence or sufficiently good evidence to make a conclusion, that you don't let your thoughts about what the social or moral consequences of that conclusion be enter into your decision as to whether you have enough now to publish. So, it can't be the case that something would be convenient to believe. And so therefore we accept it on the basis of quite little evidence because it would be so good were that to be true. Or likewise, if something was an inconvenient truth, you can't keep raising the standards of how much evidence would be required in order to, sort of, forever defer having to make the unfortunate conclusion. And so Du Bois' idea of sort of, what we would now call value free science, is science in which the scientist has kind of strict standards of when their evidence is sufficient to draw a conclusion and those strict standards are independent of the moral or social consequences of drawing that conclusion. Now, that actually matches very well what contemporary philosophers of science typically reject as the value free ideal. So that's kind of almost exactly what you'll find many, many philosophers of science nowadays arguing is just what can't be done.

PA: That's actually really interesting because I think most people who aren't philosophers of science would say, ‘Well, of course Du Bois is right. Like... why would a scientist not pursue the truth? Of course you don't want to have biased scientists trying to, you know, conjure up some kind of evidence for things they hope are true.’

LKB: Yes. So there is this interesting phenomenon wherein Du Bois is defending what I take to be sort of the common sense view of many people, and even many working scientists were against what is now, I'm not sure what about then, but what is now actually the philosophical consensus that that can't be done. I'll say a bit more about why Du Bois thought it would be desirable and then I'll say about why contemporary philosophers of science don't tend to agree. So as was mentioned in the quote that you began with, Du Bois separates the mediate from the immediate aim of science. And so the mediate aim of science for Du Bois, quite generally, is actually explicitly political. Like this is the strange thing about Du Bois as well. He's defending a version of the value free ideal, but because he sees science as ultimately having a political mission. The idea is the mediate aim of science is something like that which guides our construction of the institutions of science, why it is that we fund this thing just in general, why the collective activity of science should be organized in the way it is, like, what is that guides our sense there. And that's for Du Bois is assisting and guiding democratic decision making. So the point of keeping science around is exactly in order to make wiser political decisions. And you can think how that comes up, right? Like you could have agricultural science, which helps us increase crop yields. It could be the epidemiology, which is a kind of social mix. Biological science will help us prevent the spread of disease and maybe even inform public health campaigns, which prevent the diseases arising in the first place. And so, in general, like the reason we have science is because it will help us do that thing. Now, he also clearly thought the pursuit of truth for its own sake was a sort of noble and admirable goal. So it's not that he meant to deny that. It's just he thought the social purpose of science is ultimately… its function is to assist democracies in decision making. But then he got to think about what it takes for science to play that social role in a democracy: and what it takes is trust. People need to believe scientists when they tell them things. Ultimately, the means by which science has influence is by persuading people, like they can't force people to behave a certain way. So if you come up with very good information about how to increase crop yields, but farmers don't believe you, then it's kind of neither here nor there that you did that. And based on some of his own experience, trying to influence policy and public action with regard to race relations in America, he concluded that the public won't trust scientists if they're not seen to be fair-minded. And so in some ways, it's the very fact that it is the common sense view of science that a good scientist is a neutral truth seeker, which makes Du Bois argue that they should therefore behave like neutral truth seekers, because that's what it takes to be fulfilling that role in the public mind. And, you know, just to mention, there's a relatively contemporary incident which might really well illustrate Du Bois' point. So I don't know if listeners know about the Climategate incident wherein some wicked people got their hands on emails that climate scientists were sending to each other. But it was apparent that climate scientists were thinking about how to present their results in such a way as to spur action on ameliorating or preventing climate change. And this was spun as, 'Look at these scientists, they're conspiring against us!', etc. Now, a Du Boisian informed of contemporary climate science might try and argue that, 'Look, in fact, we should be making the kind of actions these scientists are promoting, but exactly the fact that they allowed themselves to be overtly influenced by their political preferences in making these kind of decisions about what evidence to put forward ended up undermining them.' Du Bois would say, 'that's exactly what I'm worried about, and that's the thing I'm trying to prevent!'

PA: That all sounds very reasonable, and maybe even obviously true in some sense. So why is it that philosophers of science are so resistant to advising real scientists – practising scientists – to pursue things in this value-free way?

LKB: So, it's nowadays such an unpopular position in philosophy of science that it's almost an activity for grad students to try and think of a new reason to not get scientists to do this. I'm just going to go through two of what I think of as the main and most influential arguments against this, and then maybe I'll say a bit about how Du Bois might respond to them. So the first, maybe most influential argument against this, is historically associated with Richard Rudner and Carl Hempel, maybe nowadays with Heather Douglas. And according to this argument, it's just literally impossible to do this thing. So it's not like there is an option wherein scientists just pursue the truth and use that to guide their decisions about when to conclude things and when not to. And the question is, should scientists do that or not? Rather, there's no such thing as doing that, and there's only the question of 'Are scientists being honest about value judgments?' And then the thought is, well, if you're going to do this thing, you need to be honest and overt about what it is you're doing and make the decisions consciously of that in mind so you can make wiser decisions. I'll say a bit more about why people believe that in a second. And the second broad line of argument is that whether or not it's possible to do that, it's actually... You will do better, more reliable science if you do allow people to make explicit and importantly a plurality of value judgments in pursuing their scientific research. Okay? So the first of those arguments that it's literally impossible. Here the idea is to sort of question one of the presuppositions of Du Bois' claim, where what has been said is that scientists shouldn't let their value judgments about what would be a good thing to believe or not enter into their decisions about when they've gathered enough evidence. But what's that 'enough' used there? (I just made air quotes. You can't see that audience, but that 'enough'... be very suspicious of it!) What does that mean? Well, they'll argue there's no such thing as a sort of purely objective, purely epistemic, just to do with truth seeking idea of ‘enough’ evidence. What 'enough' means is given the values at stake in this particular decision, it is acceptable, morally, to stop gathering evidence here and make a conclusion or to require that we gather more evidence. And that can change depending on what's at stake. And so, in particular, you might think there are some circumstances where it would be very, very bad to reject a claim which turns out later to be true. In this kind of sense, it might be very low risk to take a certain action, even if this is false, but we'd get a huge benefit in, you know, administering this vaccine if it turns out it actually works. So in that case, given the cost, you can afford to gather less evidence. On the other hand, there might be some circumstances where it would be disastrous to accept as true a claim we later learn to be false. For instance, you know, if there is some question as to whether a new baby formula is in fact poisonous, you might think that no, it's really inappropriate to accept anything other than the highest standards of evidence in that circumstance. And in those cases, it's apparent that people pushing this line of argument will say, ‘That's just making apparent what's always going on. You're always having to make some decision about what constitutes sufficient evidence. And there's just no such thing as doing that in a value neutral way. You're always making decisions about what the costs of different kinds of error are.' And so to those people, Du Bois is simply asking for the impossible. You can't ask scientists to just decide based on purely truth seeking criteria when they've got enough evidence. There's no such thing. To the second group of scientists, this argument is mainly associated, I guess, with Helen Longino and Philip Kitcher. What these people will say is, when we are trying to pursue science as a kind of collective enterprise, what we really want to happen is for people to spread out and explore different theories, different ideas, different methodologies... as a means of appropriately exploring logical space, as a means of appropriately making sure that all of the options are being considered and appropriately weighed. And it turns out that a really good way of doing that is to let scientists be overtly motivated by sort of different political goals. And that will lead to different areas of hypothesis space or different hypotheses being plausible to them. And so the value pluralism amongst the scientists actually leads to better science because it doesn't homogenize, it ensures there's diversity amongst the scientific community. As long as you set things up such that scientists can appropriately debate the issues. And so, either, because it's impossible, or, because it's just undesirable. Most philosophers of science tend to disagree with Du Bois.

PA: And do you think Du Bois could respond to these problems and defend this value free ideal that he's espousing?

LKB: Yes, I do in the second case. In the first case, I think it might require resources which 
Du Bois himself wasn't in a position to draw upon. So in the first case, that is the argument that is simply impossible to make value neutral decisions about when there is enough evidence. I'm not aware of Du Bois ever really questioning the presupposition that there is at least in theory, a sort of a neutral property scientific truth seeking level of evidence which is the appropriate level of evidence to require of a hypothesis. And so it seems that, sort of, Du Bois never developed the resources to defend that claim because it almost never occurred to him to question that claim. I say it almost never occurred, perhaps it really didn't ever occur, I don't know. So in that case, if it's going to be defended, one has to draw on resources which Du Bois himself didn't develop. Now, in contemporary philosophy, you might think there are resources because there are a number of philosophers in philosophy of science and in epistemology who are working on things like what's called objective Bayesianism, or sometimes relatedly what's called the Accuracy First Program. And what these people are trying to do is start from fairly basic simple rules, or maybe even in some sense value judgments, but the way those value judgments are of the form I wish to have accurate beliefs or I wish to know how the world really is or something like that. And beginning with just those kind of value judgments, which are the sort which Du Bois would endorse as the immediate aim for a working scientist, develop more complex rules about, for instance, how to treat new data, or how to gather new evidence, or whatnot. And so I think what Du Bois would have to do is try and integrate his ideal from the value free working scientist with those programs and argue that, ‘Actually, I can defend the claim that there is some, for any given hypothesis, there is some level of evidence which is the proper epistemic truth-seeking level of evidence to require for that hypothesis.’ So that idea has recurred as a means of responding to critiques of the value free ideal, but so far it hasn't persuaded philosophers of science largely because they haven't been persuaded that any of these programs can successfully play that role. But whether or not something like that could be developed in the Du Boisian vein is itself a sort of open question. So that would be, I think, how Du Bois would have to try and respond to that. On the second problem that, you know, it's whether or not it's possible to develop a value-free science, it's simply undesirable because we do better to allow for plurality of political perspectives; then, in that case, I think Du Bois really does have resources internal to his own research program, which allows him to put a pretty strong or convincing argument against them, returning to that mediate versus immediate aim of science. So what people in this tradition are arguing is that we'll do better at getting the truth by allowing for a variety of political perspectives. But what Du Bois might say is, 'that's trading off the immediate aim of science for the mediate aim, and that's the wrong way around.' Why? Well, because what one is doing in the case where we allow science to be a place in which people sort of carry out political disputes by other means is, well, ‘You're politicizing the science. You're making it such that now it's a sort of factional matter’: whether or not one adopts a certain line of hypotheses over another. And that's exactly what we're trying not to do with science. The immediate aim of science, the reason we're ultimately designing these institutions is to make it a resource which a democracy can safely draw upon in guiding its policies. And so Du Bois can grant that, of course, maybe we'll be more accurate or more efficiently explore hypotheses space if we allow ourselves this political plurality, but it's not worth being more accurate if it makes science less trusted, and there's a real risk that will make science less trusted.

PA: It's like you'd impress the philosophers of science while not impressing the public. [Laughter] He really limits the public and not the philosophers of science.

LKB: Right, exactly.

PA: Is this actually a position that Du Bois holds on to consistently? Because there's a later essay that was published in 1944 called ‘My Evolving Program for Negro Freedom' where he seems like he might be dropping, or even repudiating, the value free ideal.

LKB: So my sense is that Du Bois pretty clearly does repudiate the value-free ideal, although not in a way which involves questioning some of these presuppositions. So I'll say more on what I mean there. As far as I can tell, later in his life, maybe coinciding with, basically, the end of his scientific career, Du Bois concluded that what was happening with African Americans was such a sort of moral emergency and moral calamity that one simply had to be willing to accept conclusions and act on their basis before an ideal scientific process would say there's enough evidence to act on this: we need to act now; we simply lack the time and resources to properly investigate what the optimal course of action would be. And so in such a moral emergency, Du Bois thinks, ‘Okay, it just has to be the case that we're allowed to roll back and require less evidence before we go forward.’ So Du Bois does seem to conclude that in his later life. Now, note that that's consistent with thinking, and indeed he actually says in the essay, there is actually what you would require ideally in a proper scientific inquiry. So he hasn't dropped the idea that there is such a thing as what pure truth-seeking would tell you that you should do in this case. He simply thinks that there are certain kind of emergency situations where you can suspend that.

PA: I see. So that actually just seems like a kind of pragmatic caveat rather than giving up on the idea, on the value of the ideal of what science ought to be under good circumstances.

LKB: Yes. I think the under good circumstances is the key there's because it is giving up on it as an action-guiding ideal. He does think that in fact scientists working on issues broadly relating to race relations in America should be willing to actually suspend in their behaviour, the value-free ideal and act on the basis of less evidence than would otherwise be required. But it's not giving up as a sort of ideal ideal in the sense that ultimately in the good state, we might revert to requiring our scientists that they pay the value-free ideal.

PA: Right. Okay. Let's move on to talking about one other topic you've discussed in respect of Du Bois' science and this is what you've called triangulation. So this involves how to use more than one method in scientific inquiry at the same time. So can you explain what this is and why Du Bois thought it was a good idea?

LKB: Certainly. And I should say, actually, I think Du Bois called it triangulation. He might even be the person who coined that term. Du Bois, as I said, was kind of an all-round social scientist: he was trained in the use of many different methods and he would deploy them in his own research. And in the preface to The Philadelphia Negro, he gives the rationale for doing this. And his rationale is as follows: that in the case of social scientific inquiry, we have a lot of different potential approaches or methodologies which we can adopt, but no particular reason to believe that any of them are all that good. Like social science circa the 1890s, and maybe circa today still, hasn't exactly got a great track record of establishing unambiguously truths of a very high degree of confidence. And so no one method can boast of being very reliable. And in that circumstance, Du Bois says, we should adopt the following approach: we run as many of the methods as we can afford to run and we adopt those conclusions which they agree upon. Now what he says is they agree upon in practice later in the book, when you see him applying this approach very explicitly, he seems to really say which a majority agree upon. So if you can get that, well, there's at one point in The Philadelphia Negro where he runs four methods at once and then says, 'I'll adopt the conclusions, which at least three of them agree on.' So triangulation is this kind of means of responding to low confidence in any one of your methods by running multiple of them simultaneously and then performing a kind of aggregation on the results of that by sort of using it explicitly, actually in the case of Du Bois, as a vote, basically, on what the best answer to this question is and adopting the majority answer. This really stood out to me because, you know, I've done this work, Du Bois was about on the value free ideal, where it was clear that Du Bois' beliefs about the role of science in a democracy was making a difference to how he thought scientists should behave in terms of adopting the value free ideal. This is kind of an interesting reverse case where it seems that Du Bois is trying to import some of the methods of democracy into scientific practice where voting, or sort of democratic aggregation, is deployed on scientific methods themselves as a means of coming to reliable conclusions. And I thought there's a really interesting parallel between sort of Du Bois' defences of democracy as a means of generating the best decisions and Du Bois' practice as a scientist in trying to help that democracy.

PA: Okay. So one thing I find kind of puzzling about this view is what counts as saying that two things are different methods, because I could imagine a scientist saying, 'Well, I'm going to collect data from different sources. For example, like I'll look at their tax returns, which have been made publicly available, but I'll also look at their average monthly expenditure or something.' But that doesn't sound like two different methods to me. That just sounds like two sort of data streams that could be combined into one method. So what does it mean to triangulate between different methods? Does it mean that we have two different methods that can't really be given weights independently of how reliable they are or what exactly is going on there?

LKB: Right. That's a great question. It's one which Du Bois wasn't fully explicit on in his own work. What seems to be the case is Du Bois had in mind broadly methods which seem sort of independent in their mode of access to the world. And so actually he might count the two things you're looking at there as separate methods, because what's going on is it's well possible to imagine someone who lies on a tax returns. And so that could be influenced by a different kind of thing than their expenditures, which is limited by how much money they actually have. And so for Du Bois, two methods have to be different in just the sense that the way they get at the world, the way they generate data are sort of susceptible to different influences from the world. And so their errors could cancel out. That was his phrase. Now, when I've been thinking about Du Bois, me and co-authors have tried to model Du Bois's practice. We treat it as something like, it's a kind of probabilistic claim that the methods have to be independent conditional on what the truth is. And that just means it has certain kind of properties of their probability of getting the answer right in certain conditions. But Du Bois is just never explicit about this actually. So it's hard to say definitively. I think what we can say most confidently is that they have to get at the world in different ways, but then spelling out what getting at the world in different ways means will be difficult.

PA: I see. So we're not talking about like a really highfalutin, like a totally different theory of science stands behind one method rather than another. The point is just that it's sort of giving you two chances at getting the right answer, something like that.

LKB: Exactly that. Although note that he doesn't mean to exclude that kind of case. So it can happen that one way in which say ethnographic methods of participant observation and quantum economic methods of modelling the world, the reason people adopt those different methods is that they have in the background broadly different theories of the kind of thing science is. For Du Bois, that's fine as long as it turns out that they then end up with independent modes of access to the world for whatever reason.

PA: Right. Okay. And I take it you think this is a good position he's adopting. So it's right to say that if we have these more or less independent ways of arriving at an answer to the same question, then we should in fact use all the ways that we have available and then go through some kind of voting procedure or weighted assignment of truth values. Is that right?

LKB: Yes. My work along with co-authors is a defence of this because while Du Bois advocates this method of triangulation as a mode of ensuring errors cancel out and ultimately we get more accurate answers, he doesn't really say why he expects that to work. But it was exactly noticing the parallel to his work on democratic theory, which made us realize that, after all, there is a long tradition of arguing that the reason democracy can be expected to make the wise decisions is because democracy allows us to take advantage of the sort of the pooled information of the entire population. And so maybe any one of us doesn't know that much about this, sort of, economic situation facing the nation; but if you could somehow aggregate or pool all of our information, you'd have a pretty good picture of the state of affairs. And Du Bois had made that case about for democracy himself – he had argued that this was a good feature of democracy. And so what we noticed is that some of the ways that mathematicians have tried to uphold that claim about democracy could be applied to Du Bois' procedure. And it could be sort of demonstrated that under those conditions, he talks about where, and we don't know if any one method is reliable, but we have a lot of them and we think at least some of them might be minimally reliable. Under exactly those conditions, Du Bois's triangulation approach does turn out to be a reliable, maybe the most reliable method we have of getting at the truth. And so we think it's very defensible because of its similarities to what democratic theorists have discussed.

PA: Okay. So one last question about really just the role of science in Du Bois's work generally. We've been talking about him as a social scientist, also as a theorist of science, but he's much more than that. So as you said, he does everything and he writes books like The Souls of Black Folk, where he does draw on what he would think of as science, like history, as you mentioned, sociology... But there's also all sorts of other things going on, moral considerations, political considerations, aesthetic considerations… And this despite the fact that, as you said earlier, he really thinks the scientists should kind of separate their work as a scientist from their, let's say, political activism. So how did the scientific conclusions that he drew and the work that he did as a scientist fit into the broader framework of a work like The Souls of Black Folk?

LKB: Great. So I think there are sort of two things going on here. One more idiosyncratic to Du Bois and one broader theoretical point. So first, there's this idiosyncratic point which can't be avoided if one is involved in Du Bois's scholarship, which is that Du Bois was not lacking for humility, shall we say. Wait, [laughter] not lacking for confidence rather. So Du Bois may have thought that generally scientists would do less reliable or at least less trusted work if they were seen to be also politically engaged. But Du Bois for himself was like, you know, ‘I can be trusted and...’

PA: ‘I'll get away with this.’

LKB: ’...I'll get away with this.’ So I think that there's just this element of Du Bois wherein he does kind of make an exception for himself. So I think that's part of what's going on, but it's not all of what's going on. So another thing that happens for Du Bois throughout his life is he gradually changes his theory of why he hasn't succeeded in persuading people. So he famously, he says, you know, ‘when I was younger and I was thinking about sort of race relations and racism in America, I thought a lot of the problem was ignorance. I thought that what was going wrong was that a lot of white people were simply unaware of the conditions black people faced and the actual truths about black people's equality and standing. And so I, you know, I just needed to do the work and persuade them of this. And so for instance, some of his early work in criminology was just going through and being more careful with the data and showing that many of the claims that were being made about black disposition to criminality were simply false. They couldn't be supported by the data, often by the data which those very people were trying to use to support it. So it's just undermining it that way. And he found this just didn't persuade as many people as he thought. It didn't have anything like the kind of rhetorical effectiveness. And so I think Souls represents a kind of a middle period in Du Bois where he still ultimately thinks that what we need to do is bring to bear more accurate information about the conditions of black folk and their inner lives and their material conditions. But that extra rhetorical work has to be done to make it clear to people that, like, this is something you should care about, that I am someone who's kind of an engaged figure who can be trusted as a kind of a representative of these people. And indeed as a pitch to black folk themselves, it's like, 'I am honestly representing your interests because I embody values you'll find attractive.' And so in Souls he's trying to meld that kind of rhetorically effective, ostentatious display of artistic and ethical high value with a scientifically informed accurate picture of black people and their conditions and what we'll make for effective strategies for reform. And so I think that's kind of what's happening in Souls. Gradually, over the course of his life he becomes less and less convinced that there is any presentation of facts, which he can do, which will serve to really dissuade American racists. But at this point he still believes it. It's just he thinks, I think, he thinks that it needs to be supplemented of a more engaging mode of writing.

PA: Okay. So thank you very much, Liam, for that fascinating discussion of Du Bois on his philosophy of science. Next time we're going to be going on to look at other topics in 20th century Africana thought. But for now I'll thank Liam very much for coming on the podcast.

LKB: Thank you, Peter.