With significant recent advances in artificial intelligence and robotics, it is increasingly pressing that we consider the legal and ethical standing of autonomous machines.
With significant recent advances in artificial intelligence and robotics, it is increasingly pressing that we consider the legal and ethical standing of autonomous machines.
Last year, Kevin C. Elliott published three new books on ‘values in science’:
Given that empirical research is often used by moral, social, and political philosophers in scholarship on questions of justice, we thought it would be interesting to chat to Kevin about his recent work and its implications for moral, social, and political philosophy.
Erin Nash: Hi Kevin, thanks for agreeing to be interviewed by Justice Everywhere. To start us off, perhaps you could tell us about how philosophers of science understand the term ‘values’, and how values influence science.
Kevin Elliott: Thanks, Erin. Briefly, I think values are qualities that are desirable or worthy of pursuit, but it’s important to keep in mind that many different things fall under this description. For example, when theories have qualities like explanatory power, wide scope, or predictive success, we typically consider these things to be values because they tend to indicate that a theory is true or reliable. Some label these ‘epistemic values’. Other qualities of theories, such as their tendency to promote environmental protection, public health, economic growth, or gender equality, are also typically regarded as values (often called ‘non-epistemic values’) insofar as they help us to achieve ethical or social goals. The question that permeates all my recent books is what role, if any, ethical and social values like the promotion of environmental protection or gender equality should play in scientific research.
For example, values can subtly influence the questions that get asked in socially relevant fields like agricultural research. There are many ways of trying to improve agricultural production so as to benefit poor farmers, produce more food, and lessen our impacts on the environment; they involve efforts to develop higher-yielding seeds, more ecologically sensitive farming strategies, or more efficient markets for agricultural goods. Even if researchers are not consciously being influenced by values, decisions to focus on some of these questions or approaches rather than others are value-laden insofar as they serve the interests of some individuals and institutions rather than others.
Values can also affect the assumptions and choices that scientists make when they are analysing or interpreting their results. For example, chapter four of Tapestry discusses how economic predictions about the costs of climate change are influenced by value-laden decisions about how much to “discount” costs that occur in the future relative to costs that are borne at present. Values can also influence how much evidence scientists or policy makers demand before they are willing to draw conclusions – this is the basis for what is often called the ‘argument from inductive risk’. In the seventh chapter of Exploring Inductive Risk, Robin Andreasen and Heather Doty discuss studies designed to identify inequalities in the retention or promotion of women university faculty, and especially women faculty of colour. They note that even when the available data suggest that disparities could potentially be present, one still needs to decide how much evidence is sufficient to conclude that problematic forms of discrimination are genuinely occurring. Choosing what rule to use for inferring that discrimination is occurring depends on value-laden decisions about what mistakes we are most concerned to avoid.
Erin: What is unclear, though, is whether philosophers of science take non-epistemic values to play a role in all aspects of science or only in some. For instance, in Tapestry, on one hand you say things like “…scientific reasoning is thoroughly imbued with value influences” (pp. 166–167). But on the other hand, you use caveats such as “…it is often unrealistic to find a perfectly value-neutral way of communicating scientific information” (p. 133). I’ve found this sort of hedging to be common in the literature. But I think this can be quite confusing! If it is the case that value judgements can be avoided in some circumstances, we are left with at least two further questions: (1) How do we identify those circumstances? (2) Where non-epistemic values are currently playing a role in science, should they be, or should they to be in the way or to the extent that they currently do?
Kevin: You make a very perceptive point. I don’t think I have this issue totally sorted out in my own mind. At present, my inclination is to say that non-epistemic values are always at least somewhat relevant to scientific reasoning, but I acknowledge that the extent of their relevance and the best ways of addressing them vary a great deal from case to case. One reason for insisting that values are always relevant is that all scientific research has at least some potential to influence society over the long term. So, given that scientists always run the risk of being incorrect when they draw their conclusions, social values are always somewhat relevant for deciding how much evidence they should be demanding. However, most of the work done in some fields, such as theoretical physics, does not have the sorts of immediate and obvious social impacts that we saw in the research I described earlier about gender discrimination, so in these cases it makes sense to consider social values more indirectly.
Erin: So how can we determine whether certain value influences are appropriate?
Kevin: In my Tapestry book, I suggest three criteria for determining whether value influences are appropriate: transparency, representativeness, and engagement. First, it is important for scientists to be as open as possible about the details of their work and the ways in which values might have influenced it so that others can recognise those influences. Second, when scientists make value judgements, those judgements should be informed by ethical principles and social priorities. Third, it is important to engage key stakeholders in efforts to identify important value judgements and to reflect on how to address them. However, much more needs to be said about the nature of these criteria and how they work together. For example, I don’t think we can draw the simple conclusion that whenever these three criteria are met, value influences are appropriate, or whenever they are not met, value influences are inappropriate. I see them more as rules of thumb that can help us to incorporate values in science more responsibly.
Erin: I like your criteria, but I have a few concerns. For instance, with regards to your second criterion, it doesn’t seem like we have broad common ground on ethical principles and social priorities within our societies. That isn’t necessarily a bad thing, especially if we endorse value pluralism. A commitment to democratic values and a concern for the proper place of science and scientists in democratic societies also motivated W.E.B. Du Bois’ early defence of the value-free ideal for science, as Liam Kofi Bright has recently explained. Along similar lines, Stephen John has argued that because people hold different values, science that is value-laden may fail to contribute to ‘public knowledge’. Do you find these arguments persuasive?
Kevin: I’m very sympathetic to the sorts of concerns that you are highlighting. This is partly why I suggest with Ted Richards in the concluding chapter of Exploring Inductive Risk that in some cases, scientists should try to avoid making controversial value judgements themselves. For example, when there is a great deal of disagreement about how to interpret the available scientific evidence, it might make the most sense for scientists to report the state of the evidence as clearly as possible, and let others decide what conclusions they want to draw. Similarly, sometimes it is possible to report more than one way of interpreting the available evidence so that decision makers can decide which approach fits better with their values. For example, economists studying climate change can report how their analyses of climate impacts differ based on the decision to use different discount rates. However, Ted and I would respond to the work of Bright and John by emphasising that even in these cases, an array of implicit value judgements have probably still played a role in how the available evidence was collected, analysed, and communicated. Thus, despite the limitations of my criteria, I don’t think we can avoid them. Scientists need to acknowledge value judgements as best they can (transparency), make them as responsibly as possible (representativeness), and invite critical reflection about them from an array of perspectives (engagement).
Erin: What advice can be derived from the values in science literature for political philosophers, social theorists, and policymakers who use empirical research to support their normative arguments?
Kevin: It’s really important for scholars and practitioners who draw on scientific research, and perhaps social-science research in particular, to recognise the potential for this research to be subtly influenced by non-epistemic values. As I noted earlier, the questions scientists ask, the assumptions underlying their interpretation and analyses, the evidence they demand before drawing conclusions, and the ways their results are framed and communicated can all involve value-laden judgements. Thus, when this research is informing important decisions that will have social consequences, it is important to scrutinise potential value influences and recognise how they may have influenced the research and its communication.
Erin: So perhaps this debate is best thought of as being situated at the interface of philosophy of science and moral, social or political philosophy? If this is the case, how do you think moral, social, and political philosophers might be able to contribute to, and help advance, this debate?
Kevin: I totally agree. We actually just had a discussion about the need to bring together moral, social, and political philosophy with the philosophy of science at a conference session devoted to my Tapestry book. An important theme was the fact that my criteria (transparency, representativeness, and engagement) need a good deal more elaboration, and this is the kind of work that moral, social, and political philosophers are well placed to do. For example, moral philosophers can help us think through the ethical principles that are most appropriate for guiding particular value judgements, such as decisions about what discount rates ought to be used when analysing the economic costs of climate change. Moreover, there will almost always be disagreements about these ethical principles, so we also need political philosophers to provide guidance about how to address these disputes. What forms of engagement should we employ for responding to disagreements about important value judgements? Which stakeholders should be involved in the deliberations? One of the most obvious lessons to be gleaned from the recent work on values in science is that philosophers of science desperately need guidance from moral and political philosophers, so I’m really grateful that you provided this opportunity to talk about my books on this blog!
A full review of Kevin Elliott’s A Tapestry of Values: An Introduction to Values in Science by Erin Nash can be found here.
Why do we trust experts to take care of our health and not to take care of our interests in the political realm? This is a very old question of democratic theory. Epistocracy is a neologism frequently used in recent works to refer to a form of government by those who know more or are wiser than the mass.
Two different aspects might differentiate an epistocracy from a democracy: the absence of political equality in the selection of the rulers, or the absence of egalitarian accountability. In addition to these undemocratic aspects, an epistocracy would differ from other non-democratic regimes by some mechanism allowing people who distinguish themselves from the mass by their wisdom or expertise to rule or at least enjoy an important degree of political power. The best example and – to my knowledge – the most interesting challenge to our democratic convictions is Jason Brennan’s idea of an “epistocratic council”. Members of this council would be selected on a meritocratic basis, passing a competency exam. And all citizens would have an equal voice in the choice of the expertise criteria.
Leaving aside the practical challenges such as the choice of the people in charge of preparing the exam, what would be wrong with such an epistocratic council?
Can Iris Young’s analysis of structural injustice, problematic norms, individual guilt and forward-looking responsibility contribute to contemporary feminism as #MeToo broaches the subject of bad dates and male privilege?
This blog post comes with a trigger warning as it contains discussions of sexual harassment and sexual assault and controversial opinions regarding them.
Last week much of my facebook feed was again full of discussion regarding accounts of sexual abuse and comment regarding the #MeToo movement.
One of last week’s stories concerned the behaviour of a male celebrity who publicly endorses calls to end sexual abuse in the entertainment industry and beyond. A woman who dated the celebrity detailed to a reporter how he acted in a pushy, forceful manner. She explained how he ignored non-verbal cues to slow down their encounter and end more overtly sexual activity, and then re-initiated sexual activity after she stated that she was feeling pressured. The woman left their date feeling violated and miserable. The report has been broadly discussed with a spectrum of opinions emerging regarding the case: some sympathetic to the celebrity, others criticising his hypocrisy and abuse and some suggesting we are reluctant to acknowledge this as abuse because we want to protect ourselves from facing the reality of the problems we have encountered in our own sex lives.
Last week also saw criticism of the #Metoo movement gain momentum. A letter was published by concerned women in the French entertainment industry who believe that the movement has gone too far and begun to stigmatise men who make clumsy and persistent advances. The letter also suggested that the movement has begun to undermine women’s power and self esteem, prohibit people’s legitimate sexuality, censor artistic expression and prevent the enjoyment of art made by perpetrators of sexual violence. The letter too has provoked extensive discussion, clarification, criticism and response.
How should we think about these cases and positions? There are two sorts of understandings that opponents in the debate often identify each other as falling into.
From June 14th to June 16th, the Amsterdam Centre for Contemporary European Studies (ACCESS EUROPE) organised an international conference on “Solidarity and European integration”. In his contribution to the panel “European solidarity and justice: normative issues”, Andrea Sangiovanni presented his dispositional analysis of the concept “solidarity”. He defines solidarity as a (complex) disposition to sacrifice one’s own self-interest (narrowly understood) for the good of others. In order to distinguish solidarity from utilitarian altruism, love, enlightened self-interest, and fairness, he further specifies it as being a disposition to sacrifice that is impersonal, narrow, and person-directed. It is a disposition to sacrifice one’s own self-interest for the sake of overcoming an adversity faced by other member states or EU citizens. Such a dispositional analysis is, I believe, much more promising than, for instance, an analysis of solidarity as a mental state. It enables us to reach a better understanding of the conditions that are most conducive to the development of solidarity and the factors that hinder it. In this post, I develop some thoughts on how to address this issue in the European context.
Relational egalitarians hold what matters for justice is that all members of a society “stand in relations of equality to others.” The idea that all human beings are moral equals is widely shared: it underlies the Universal Declaration of Human Rights and many national constitutions. How will this norm be affected by the arrival of “big data,” the collecting and analysing of huge amounts of data about individuals? Internet companies and government services collect data about individuals’ activities, including geographic locations, shopping behaviour and friendships. Many individuals voluntarily share such information on social media, some also track their physical activities in meticulous details. Experts expect that “people analytics” – big data applied to the measurement of work performance – will have a revolutionary impact on labour markets.
This summer my 2-year-old daughter and I were looking at a world map together. I would have liked to tell her something about the different continents and countries (about all the different languages, the food, music, local customs), but wasn’t able to because the sight of the map prompted only thoughts such as “There is war here, there are people starving there, refugees drowning here…” So I remained silent. We are currently overwhelmed by negative news. Almost everywhere things seem to go awfully wrong: more than 65 million refugees worldwide, 470000 deaths in Syria, the terror of ISIS, right-wing populists gaining more votes everywhere, Donald Trump for president, the Brexit, growing child poverty in Europe’s strongest economy (Germany), burning asylum seeker centres… (I could go on and on). Of course, the news we get through the media has always been mainly negative, but now it seems to have reached a new dimension. Whether this impression is accurate or not, it is certainly unsettling, raising perturbing questions: How long will we still be able to live in peace and with our basic human rights protected? Will the fear of terrorist attacks soon be part of our daily lives? Have all attempts after 1945 to create a more peaceful world been in vain? What kind of world will my children find themselves in? To what extent do our governments and we carry responsibility for what is going on? What does justice require from us as individuals? Is there a moral justification for focusing on one’s own comparatively small problems and not trying to help solving the big, global ones? How many resources are we allowed to spend on our own children? These kinds of questions are far from new, but they currently pose themselves with particular urgency.
This post has been published anonymously to protect the identity of its author, who is still receiving messages of hate.
Recently I did a radio interview in which I argued for equal access to certain social services, such as health care, for migrants and refugees. I did not focus on the instrumental value migrants have for countries – I did not focus on the economic and health benefits everyone has if people on the same territory are able to work, and are healthy and sane. I focused on the broader ethical arguments for equal access, even though I mentioned the instrumental arguments too. Perhaps I should have expected that not everyone would agree with my views. But nothing could possibly have prepared me for the hate mail that I received after the interview. In this post, I try to describe the experience and make a plea for greater solidarity in standing against such hate.
Normative theorists are not a species known for an oversupply of consensus. But one of the most heated debate of recent years has led to a kind of consensus: the debate about “situationism”, which was raised as a challenge to virtue ethics. With virtue ethicists referring to the character of virtuous agents for guidance about moral behaviour, situationists drew attention to the problem that human behaviour is greatly influenced by the situations they find themselves in. For example, they are more altruistic when exposed to the good smells of a bakery. They are more likely to cooperate in a game call “Community Game” than in one called “Wall Street Game” even if they payoffs are the same. And if they are told to play the role of “prison guards”, while others play the role of “prisoners”, the situation can easily get out of hand. Reading such accounts, one might think that all talk about individual agency and responsibility had been based on an illusion: on an account of a “Cartesian” or “Kantian” self, or on an “Aristotelian” notion of stable character, that simply do not exist. All that there is, it seems, are situational forces.
I spend a disproportionate amount of my free time following the ins and outs of American politics. And one of the most interesting/baffling things about the nominations for the 2016 presidential election is the sheer capacity of the average Republican voter to stomach policy proposals that seem tailor-made to benefit the tiny minority of the wealthiest at the expense of everybody else. For example, all of the Republican front-runners have come out with some form of tax plan that cuts taxes on the wealthiest 1% by hundreds of thousands of dollars. Yet this doesn’t scare away nearly as many voters as you’d expect — in fact, the race is wide open, and some pundits even suggest the GOP are favourites at this stage. Voting against your own interests is, arguably, a global phenomenon – I’m sure many in the UK will say it happened in the elections in May – but it does seem to be particularly prevalent in the US, perhaps because one of the parties has moved so far to the right on social and economic issues that there is not yet any equivalent in Europe.
Page 1 of 2