General, Science and Democracy

Let me briefly express and then partially exorcise a general worry about current environmental discourse: that it’s depoliticizing. And let me do it in staccatos.

  • Climate Science and Environmental Science are often asked to make pronouncements on enormous and highly variable objects of study.
  • Their conclusions, then, are highly abstracted, highly technical, and non-contextual.
  • This situation is in some ways analogous to specialist talk of ‘the global economy’ – insofar as it is highly abstracted and decontextualized.
  • This technical expertise in economics  tends to blur into technocratic political leanings; the question of ‘what we should do’ can be answered, for these people, by sophisticated Cost-Benefit Analysis.
  • This technical exercise in aggregation is profoundly elitist and anti-democratic. There is no role for discussion, argument, or deliberation.
  • A parallel worry applies to Climate Science – that a highly technical and abstract level of analysis would lead to a technocratic kind of politics. This attitude is not uncommon. James Lovelock (of ‘Gaia’ fame) expressed outright contempt of democracy not so long ago.
  • So – moving quickly – there is an overall worry that technical epistemic expertise in problems which are highly abstracted and decontextualized can slide into an ideal of technocratic political expertise which – it goes without saying –  threatens the political life of the rest of us.

Indeed, this political tilt of many scientists is only surprising with regards to its tedious repetition. Francis Bacon, for one, had the utopian vision of scientists as both epistemic experts and technocrats in The New Atlantis (1627).

Now, caricatured as it is, I think this is a legitimate worry. But it’s not an inevitable one. In fact, I think this is all more or less a consequence of various self-deceptions we have about how science works. Chiefest of these deceptions is the myth that science is a discipline which is free of values – that it represents facts about the world and that’s that. This is the mother of all errors and maybe I’ll summarize some arguments against it in later posts but there are many other dubious assumptions which could be challenged like, say, that science is fully empiricist and free of metaphysics (false – it mostly operates on unquestioned foundations about reality, like that it’s mechanistic) or that our relationship with the truth is an individualistic matter of assessing evidence for our hypotheses (highly contestable, for too many reasons to mention).

There has been a lot of important work in the Philosophy of Science to try and correct these errors, and it leads in one general direction: that science can and should be socialized and democratized. Our epistemic goals and our practical or moral goals are all tied up together; our scientific projects are inextricably bound with the rest of our politics.

If our image of science can change like this then not only does the worry about possible technocratic domination dissolve, but we also see that it was the assumptions about science which failed us. The political anxiety starts to look contingent and remediable once we realize that it was premised on a very poor image of how science works. We get a new way forward when we see that our science doesn’t have to be awkwardly compromised with technocratic ambitions. As a departing shot of optimism, I would add that this new kind of vision also holds promise for resolving the environmental and political stalemates in which we already find ourselves – but that may seem too optimistic to be credible. I hope to expand on these themes in coming posts.

Science and Democracy

I’d like to tack an addendum to that last post.

There’s a nice little term I recall from first year psychology, ‘The Fundamental Attribution Error’,  which marks the tendency for people to attribute causation and responsibility in very suspicious ways.

Example: say both you and some stranger in your class fail an exam. Why did he fail? Why did you fail? When we explain the failure of someone else we are much, much more likely to make an internal attribution. That is, we are likely to say that the person failed because he didn’t study enough or just weren’t smart enough. It’s his fault. But when we explain our own failures, the burden of responsibility magically shifts onto the environment: the room was too hot, there was that late-night call from a friend which made us lose sleep, the exam questions were poorly posed or unrepresentative of course material, etc. This asymmetry seems like it could be a result of a few things but the most obvious explanations involve (1) informational access and (2) motivated cognition.  In the case of (1) we could plausibly hold that this bias is the simple consequence of what we happen to know at a given time. We have access to all the details of our own lives and so a distributed, situational attribution is possible; we were there when that late night phone call came in. Conversely, we know little about the lives of others so, as a short hand, we attribute their failure internally. Of course, the implication here is that a true attribution in either case would involve a wide range of situational variables and that our current bias is because of our ignorance about the lives of others.

Interestingly, this attributional pattern somewhat reverses in the case of success rather than failure. Why did you succeed? Well, I’m pretty clever (internal). Why did they succeed? Well, the exam was pretty easy (situation). Could we explain this reversal by appealing only to (1), the lack of information? No, clearly not. We still have access to more information about our own situation but, instead of using it to explain our success, we make an internal attribution. The same reversal happens for our attributions about the other person. He succeeds and we now explain that success by the situation (the easy exam) rather than his (internal) ability. So instead of being the result of informational access (1), the obvious explanation for this pattern is (2), that we attribute causation – and responsibility – because we are motivated to do so; we pick the story we feel good about.

I want to be agnostic for the moment about whether we think these explanations are sound and instead give a couple of examples to show how this bias generalizes beyond  social psychology.

Firstly, and as concerned the previous post, genetic-behavioural explanations probably follow this pattern:

“Why is she so loving an affectionate?

It’s in her [internally attributable] genes; females are just more caring.

But why are you  so caring?

My parents raised me that way.”

The person whose behaviour we are explaining is more likely to get it attributed internally, and the likelihood of this happening probably increases along with our social distance from that person; men explaining women, whites explaining aboriginals. This could be because of both (1) a lack of information about their lives and (2) some less perceptible motivation to explain things in a way that justifies the beliefs we already hold about these groups.

This suggests a wider political analogy. Liberalism and neo-liberalism (that shibboleth of angry Marxists) are built around the sacraments of ‘individual freedom’ and ‘personal responsibility.’ I take these very seriously, but it’s worth pointing out that from a social-scientific point of view liberals are likely to explain things as the aggregation of individual choices – ‘the market’ being the very ideal.  That is, attributions are internal. This can often be informative but it’s not the only option. For their part, those angry Marxists would reject this kind of attribution, and instead talk of larger social categories like ‘class’, ’empire’, and ‘power relations.’ For them, individual and social behaviour is much more situational or contextual.

This point about leftist situational attributions is well shown by that angry Marxist Brecht in Mother Courage and Her Children. In the (palpitating) play, Brecht apparently casts aspersions on this liberal individualist vision which celebrates solitary choice-and-virtue by having each of Mother Courage’s children embody  a certain virtue – bravery, honesty, and some third one that’s eminently forgettable. As Mother Courage craftily adjusts her allegiances to whosoever is winning the Thirty Years War at the time, her children remain constant and virtuous. At the beginning of the play, each one of them is rewarded for this in some way or another but as the battle-lines and social context shift around them, each child ends up getting killed for some final expression of that virtue. Evidently, Brecht doesn’t have much patience for the internal attribution of choice or virtue; social change is explained by much larger forces in a much wider context.

So it looks like a wide range of fields have substantial discretion – and disagreement – about where causation gets attributed. The Fundamental Attribution Error is far from just social-psychological.

Now, I don’t know the history of the term itself, but it seems that by calling this bias ‘an error’ we are assuming that there is one true story about what causes what and that most of our attributions are just myopic and wrong. This much seems implied by saying that the attribution is a result of (1) limited information.

I want to raise suspicions about this conclusion. We could start doing this by pointing out that ‘attribution’ is a close linguistic cousin to ‘description’ and that the latter is ambivalent about truth. I could describe the faces on the subway as ‘petals on a wet, black bough’ or as ‘bottomless sinkholes for absorbing cultural refuse’ and neither would really be wrong. How we classify things is largely up to us; it depends what we’re trying to show (this generalizes: is a cluster of trees which shares a root system one organism or many organisms? The evolutionary biologist says ‘one’, the ecologist says ‘many’. They are each trying to show different things.) This minimalist relativistic point should be enough to show that our purposes matter an awful lot for our attributions and descriptions, and that these are not only meant to be expressions of some mute realm of facts.

This has fallout for our understanding of the purported error in all its many forms. If we think it really is an error – which I think we should – and we are also suspicious that it’s an error because of its failure to capture some fact of the matter about the way things are (soppingly pejorative) then why is it an error? How could it still be called ‘an error’ if it’s not because of a factual failure? Re-invoking the claim that our descriptions are somewhat purpose-relative, we can say it’s an error because our purposes are bad ones; it’s a moral error. Selectively attributing causation and responsibility as a way of confirming what we already want to believe is simply selfish and bad.

This doesn’t tell us what kind of attributions to make. It doesn’t tell us when people are acting as responsible agents or when they are victims of their situations; it doesn’t resolve the debate about the causal role of genes; it certainly doesn’t recommend that we be liberals or Marxists. What it does do is moralize a large part of what we probably thought were mere factual descriptions and thereby sensitizes us to the possibility that the attributions we are making may be informed in ways we think are bad. Maybe the direction this points us, like any other time we have to select and weigh values, is toward a focus on matters of inclusion and due process. That is, toward democratized causal attributions.

 

 


Says Mother Courage about a commander who is searching for brave soldiers: “he must be a very bad commander….because he needs brave soldiers, that’s why. If his plan of campaign was any good, why would he need brave soldiers, wouldn’t plain, ordinary soldiers do? Whenever there are great virtues, it’s a sure sign something’s wrong.”

 

Science and Democracy

I would like to register a quick complaint about this National Post article concerning suicide in aboriginal communities, posing the titular question:  “Does it run in families like a disease?”

To be fair, the sources they quote don’t give an unequivocal ‘yes’ but I want to clarify a little bit why we think the question is itself obnoxious.

To start, everyone probably has a knee-jerk reaction to the title. Aboriginal communities don’t have a lot going for them; there is poverty, drug abuse, and a history of violence – colonial or otherwise. These are obviously environmental determinants which don’t have to do with genes. Nurture, not nature – right?

Yes and no. It’s becoming more commonplace now to realize that ‘nature-vs-nurture’ is a poor way of posing the question since we think that genes and environments are deeply entangled. Some of the more popular examples of this thought are epigenetic. In rats, the absence of a mother’s affection (licking) can methylate parts of the DNA in the offspring, thereby blocking the expression of certain molecular genes and so making it nervous and anxious. The kicker is that this methylation is heritable in the subsequent generation as well, causing the same outcome. So, to pose the defunct question about this grandchild: is its nervous behavior explained by nature or nurture? Dunno, really. There’s no real answer. To be sure, there’s some kind of genetic explanation – the grandrat’s behaviour is a consequence of its genome – but the structure of the genome is itself caused by an historical environmental change.

A similar kind of study, albeit more contentious, considers the Dutch famine of 1944-45. Pregnant mothers at that time had underweight babies and although those children were then fed normal diets, the next generation was also born underweight. Nature or nurture?

However attractive, epigenetics is but one swell in a rising tide of biological work which is tracing out all the incredibly complex interactions among chunks of DNA, their cellular environment, and the organism’s environment, which make the reductionist story that behaviour-is-really-just-DNA increasingly implausible. But this is all old news.

Now for the bombastic part, which I’m ambivalent on and will only sketch: there is no fact of the matter about whether it’s genes or environment.  Nature won’t reveal the truth to us about whether behaviour follows a straight line from DNA or whether it’s environmentally conditioned. More bombastic still, I think a large part of the choice in which story we tell is a political and ideological, not ‘scientific’. Be honest: don’t you feel attracted to the results of epigenetics because it promises to free you from the sense of fatalism that your destiny is already written?

An intuition pump: Helen Longino, a feminist philosopher of science, gave a very strong example that there are – by and large – two kinds of conceptual models with which scientists approach the question of gender differences. One is called ‘linear-hormonal’ and the other is, let’s say, ‘distributed’. In the former, gender differences are explained in a straight line from genes to neuroendocrinology to behaviour. The other is – as the name suggests – much more distributed; genes interact in complicated ways with each other and with their environments which thereby results in behaviours that are quite a bit more plastic and contingent. Now, the important part is that these models don’t behave as hypotheses about the fact of the matter. Rather, they serve as a way of organizing our evidence. To put it loosely, they are two different ‘ways of seeing’ and we could, in large part, construct our body of scientific knowledge around one of them or the other; we can fit our evidence into either one.

It is not, then, entirely an empirical matter which model we choose. This is where the politics enters. If this gender-behaviour research were undertaken only by men who happened to believe in a fixed, essential image of what women were like then it would be totally unsurprising if their research adopted a linear-hormonal model of explanation and so sought to reduce explanations of female behaviour down to the genes. It’s easy to simplify ‘the other’ if they are a mere object of study and look for some simple essential explanation. If we believe this, then reductionist genetic explanations for things can start to look like they fall out of a set of background beliefs – maybe political ones – and so there’s a concomitant risk that our ‘explanations’ are sometimes nothing more than an expression of the beliefs we started out with.

So that’s the first bombastic thing: sometimes it can be the case that the shape of our science can be political fallout rather than a mere expression of ‘the facts’. It’s a heavy point which would require a lot more argument if we were serious, but just roll with it for now.

Second bombastic thing: which scientific approach we take – reductionist or otherwise – determines what interventions we make in our world. This should be unsurprising, really — it doesn’t take much imagination to see that a world in which we think everything is genetically determined would look very different for our own — but we’ll take an example: ADHD. The last couple of decades have seen a continued explosion of diagnoses for this malfunction and kids have been increasingly hopped up on ritalin. I myself was diagnosed as a child. But why so many; why now? One story we could tell would implicate reductionism itself. If we look for the causes of behaviour inside the individual and, specifically, in the genome then any intervention to mitigate that behaviour will have to take place inside that very individual. But it could have been otherwise; we could have explained the behaviour by the environment: school is boring. Implication? Make school less boring rather than reprogram the kids who, after all, are alright.

So to recap this potted story: politics, or at least our very many background beliefs, can influence the kinds of scientific programs we undertake, but so too can those scientific programs affect the interventions we make in our surrounding world and one another – our politics. This reveals a kind of reciprocity between science and politics which I don’t take to be contentious, but which is very uncommon nonetheless.

 

Now, look back at the National Post byline we started off with. “Does [suicide] run in families like a disease?” starts looking like a very loaded question. What does it presume? Well, firstly, that legitimate go-to explanations are ‘internal’. That is, we should look for genetic and hereditary explanations for behaviour. Even if we think – deep down – that things are more complicated, we can now ask whether there is a risk in using  this explanatory idiom. Is there? Yes, probably. If we tell a story of genetic causation then we risk a fatalistic social attitude towards these people. This comes at the cost of focusing on environmental changes which – after all  – are obvious crises facing aboriginal communities. More crucially, however, is seeing that this reductionist explanatory mode exists quite apart from the cardboard-cutout image of science that pursues mere facts of the matter. This explanatory reductionism precedes evidence; it is a framework, a way of seeing. Now, are we right to think there’s a risk that our adoption of this explanatory mode has been informed by our politics? Do we think there’s a chance that our scientific pursuit for hereditary explanation could actually be a convoluted way of affirming our prior attitudes about the behaviour of certain kinds of people?  More concretely, do we think that a cadre of misogynistic endocrinologists could end up with a scientific image which confirms their prior and essentialized views of what women are like?

If you answered ‘yes’ to any of these questions then you should also think that even asking the question about hereditary suicide is one we should be more careful about. Of course, this is not to say that these are the kinds of questions we should never ask – genetic explanations are incredibly powerful and socially helpful. Rather, it’s a cautionary tale that what we take as mere conjectures about scientific fact can often be bound up with our social and political biases, and can thereby bias the kind of actions we take, which often end up confirming  the views we started out with. The upshot is not that we should stop doing science and do politics instead, but rather that we should give ourselves tools to do both at the same time. Especially since that’s already what we’re doing, though not necessarily doing well.