During the COVID-19 pandemic, the UK government has stressed that its policy response is led by the science. Lots of people object that these claims are just window dressing. A tweet from the Liberal Democrat politician Layla Moran sums up these concerns nicely: “the Prime Minister has said from the start that his Government would be led by the science. It seems that was only the case when the science agreed with him”. To confuse matters, though, other governments which also claim to be led by scientific advice have adopted very different measures, most famously Sweden. Is the government led by the science or not? I don’t know. I’m fascinated, though, by the background assumption that policy should be led by the science. Should it be?
There are two familiar arguments as to why science should not lead policy.
The first is that it can’t. At least on a standard understanding, science is in the business of describing the way the world is, whereas policy-making concerns what we ought to do. A long-standing philosophical tradition holds that you can’t derive an “ought” from an “is”; you can’t get “humans ought to eat meat” from “humans evolved to eat meat”. Saying that science should lead policy is a bit like saying that mathematics should conduct orchestras; it sounds nice, but doesn’t make sense.
There’s an obvious response to that argument: what is meant is that scientists ought to lead policy. This claim has the advantage that it makes sense: we could simply give Patrick Vallance or Chris Whitty the power to make policy. At least sometimes, government ministers have presented matters along these lines, smiling apologetically for introducing measures the science has “demanded”.
Unfortunately, this version of science-led policy faces a second familiar challenge, that it’s deeply undesirable. On a standard model, policymaking involves two kinds of inputs: value claims, concerning which ends we ought to pursue, and factual claims, which identify the means to those ends. Typically, in democracies, we think that there can be factual experts, but no value experts; no-one has a better grasp of the “correct” values than anyone else. The value inputs to policy should, ultimately, be decided by the people. Scientists should inform policy by providing factual inputs but having them lead policy is undemocratic.
What can be said in response to these familiar arguments? First, there is some wriggle room to square “scientist-led policy” and democracy. If you think that actual democratic procedures are badly broken such that scientists are more sensitive to public values than are policymakers, then you might argue democracy is best served by having unelected scientists rather than elected representatives decide policy. This is a fascinating argument, but I doubt it’s what Boris Johnson or Matt Hancock have in mind.
A second possible response is to challenge the standard account of policy-making. The division-of-labour between scientists and policy-makers rests on the distinction between fact and value. In recent years, however, many philosophers of science have challenged the once-common assumption that scientific research is or can be entirely uninfluenced by ethical or political values. Consider, for example, epidemiological models for predicting the effects of pandemic mitigation strategies. Building these models involves making assumptions about various parameters, such as the likely degree of adherence to policy guidelines. Unfortunately, we don’t have a lot of evidence about how people in the UK respond to pandemics. Therefore, in setting these parameters, scientists must make educated guesses. These guesses are often, perhaps implicitly, guided by some sense of whether it would be “worse” (in an ethical sense) to under- or over-estimate the likely effects of mitigation strategies. In turn, as John Dupre has argued, it seems that much of the scientific advice we receive about COVID-19, not just our models, is unavoidably value-laden (see nuffieldbioethics.org/blog/following-the-science-in-the-covid-19-pandemic).
However, even if we can’t neatly separate fact and value, it doesn’t follow that policy can or should be led by the science or the scientists. We need to separate out two claims. The first is that scientific practice is or must be shaped by evaluative concerns. The second is that there is no distinction between fact and value. The first claim is plausible, but the second is not. Indeed, we need the fact/value distinction to make sense of the idea that scientific justification is value-laden, much as we need the jam/dough distinction to make sense of the idea of jam doughnuts.
A defender of the traditional division-of-labour can, then, accept that science is value-laden and retrench. The value inputs into policy should be settled by the people; the fact inputs by the scientists; and, if the fact inputs must be partially value-laden, then the relevant values should be democratically legitimate. A more nuanced understanding of the fact/value distinction doesn’t support scientists leading policy. If anything, it weakens it, because it suggests that, even when they seem to be simply presenting the facts, scientists might smuggle their own values into debate.
Even if we blur the fact/value distinction, policy should not be science-led. This implies that politicians who promote science-led policy are, at best, stupid, and, at worst, mendacious – inappropriately buck-passing blame for unpopular decisions. That might be right, but there’s something appealing about the slogan. So, I will now explore a third, more attractive defence of science-led policy.
To set the scene, consider a hypothetical example. You are driving to a hotel with your friend. You come to a crossroads and don’t know which way to go. Your friend, who knows the local area very well, shouts “turn left”. You turn around and object: “this is my car, and I decide what to do – you just tell me the facts”. That would be an extremely odd response. Given that your goals are clear – to get to the hotel – and your friend knows the area, it seems fine for her to tell you what to do.
This example provides a useful model for saving the ideal of science-led policy. “Stopping the pandemic” is a legitimate democratic goal with widespread public support. Therefore, it may seem that handing power to the scientists isn’t anti-democratic, so much as an efficient way of cutting out the middleman. Furthermore, not only could we do this, but we have reasons we should do it, as a way of ensuring that policy is not distorted by short-term political interests. Consider the example which provoked Layla Moran’s outrage, the decision to reopen various sectors of the economy. It’s easy to think of this as a case where policymakers are faced with a problem: scientists tell them that loosening lockdown measures risks a second spike, but business leaders demand looser measures to stay afloat. To say that the pandemic is a case where we ought to “follow the science” is to say that, because “ending the pandemic” ought to be the key value governing policymaking, policymakers must defer to the scientists, even if doing so will anger a politically powerful lobby. It is a way of giving scientists a certain sort of authority over policymakers, where that authority is derived from a concern about policymakers pursuing their short-term political interest rather than the democratic values they ought to pursue.
It is useful here to draw an analogy with climate change, another area where figures such as Greta Thunberg enjoin us to follow the science. It seems odd to say that climate policy can be based solely on the facts, because the facts of climate change don’t tell us to do anything. What does seem plausible is to say that there is widespread agreement on the value judgement that massive loss of life due to extreme climate change would be a very bad thing. As such, if scientists tell us that we are headed towards massive temperature rises and this only be avoided if we change our lifestyles, then we should change our lifestyles.
So, is it fine to have scientists take charge of COVID-19 policy? I don’t think so. To see why, I will explore two dis-analogies between driving to the hotel and tackling COVID-19.
The first dis-analogy concerns the value considerations at play. The appeal of passenger-led driving stems from an assumption that you only have one goal – to get to the hotel – which is known to both you and your passenger. The value issues are simple. Responding to a pandemic is far more complex. It’s not clear that ending the pandemic is, or ought to be, our only goal, regardless of all other value considerations. Lockdown measures had significant costs for many in the population – restricting their access to non-emergency medical care, employment opportunities, intimate relations, and so on. We can’t simply assume that the goal of stopping the pandemic is so valuable as to justify all of these costs. (And this is even before we get to the complex intergenerational justice issues raised by the fact that lockdown tended to protect the elderly at the cost of the young). Note that I am not saying that lockdown measures were ethically or politically unjustifiable. Rather, I am saying that simply handing decision-making to the scientists is inappropriate, because doing so hides complex, difficult, ethical and political choices.
The second dis-analogy is more complex. Particularly in complex and fast-moving cases such as a pandemic, scientific knowledge is often piecemeal and subject to significant uncertainty, in a way in which knowledge about directions to the hotel is not. Consider, for example, the question of whether wearing facemasks in public places has a significant effect on viral transmission. We have lots of different kinds of evidence which suggests it might. On the other hand, that evidence is often incomplete and difficult to interpret. It is methodologically difficult to combine different sorts of evidence, ranging from behavioural observations to statistical models. Indeed, some scientists even worry that facemask wearing may promote the spread of disease, because we will touch our faces more often.
I don’t know how to adjudicate these disputes. What I do know is that any claim along the lines that “facemask wearing will limit the spread of COVID-19” is subject to far greater uncertainty than “the hotel is to the left”, and that most of our scientific knowledge about COVID-19 is similarly uncertain. We may be able to make educated guesses about cause-and-effect or about the possible effects of interventions, but those claims are highly uncertain.
To draw out the implications of this uncertainty, consider a variation on my hotel example. This time, matters are more desperate: night is drawing in, and you can hear wolves howling. You come to the crossroads and don’t know which way to go. Your passenger is also a stranger to the area. She leans forward and says, “turn left.” You check: “are you certain?” “No, but that’s my best guess based on my memory of the map you lost.”
Assuming she isn’t an inveterate liar or lacking a sense of direction, you are probably better-off relying on your friend’s advice than not. Still, you shouldn’t just do what she says unthinkingly. After all, there is a decent chance that she might be wrong, and, if so, you might end up being eaten alive by wolves. Maybe you could gather more evidence instead. Unfortunately, gathering information might also be costly. What you must do is to decide, given everything you know – about the costs of an error, the costs of waiting, the likely success and costs of more investigation – whether her level of certainty is certain enough to take the left turn. This is a complex judgement, which requires ethical judgement – just how bad would it be if she is wrong?
Something similar is true of policymaking. Even if we make the wildly implausible assumption that policy should be guided by the single goal of stopping the pandemic, scientists are highly uncertain about how to do that. For example, they may say it likely that a policy of mandatory facemask wearing will limit the spread of disease, but acknowledge there is a possibility the opposite will occur. This kind of uncertainty does not mean that scientists’ claims are irrelevant to policy. Highly uncertain claims, hedged with all sorts of caveats, are often the best claims we have. However, we need to make judgements about how certain these predictions must be to be “certain enough” to form the basis of policy. Those judgements involve value judgements about which kinds of mistakes we are willing to risk. Those are political judgements. Given the science we have, it is inappropriate to expect scientists to lead, rather than inform, policy.
As the example of driving to the hotel shows, handing decision-making over to experts can be appropriate. However, for that case to be a useful analogy for policy-making we need simple goals and certain experts. In a pandemic, we have neither. What, then, is the appeal of saying that policy is science-led?
Here is a tentative answer to that challenge: we are deeply confused about what science is. At least sometimes, we think of scientists as people who do not commit themselves to claims unless they are highly certain of them. Consider how “this appeared in a peer reviewed scientific paper” functions as shorthand for “uncontroversial”. This model of science provides an obvious justification for why policy-makers should take scientists’ claims seriously (even when doing so is politically uncomfortable): scientists’ claims are highly likely to be true.
There’s lots to be said for this model of science in general, and as a justification for why policymakers should defer to scientists’ factual claims. (In fact, I’ve said lots of it myself elsewhere). Still, it is a model for understanding the role of mature, well-established science in policy. Clearly, much “science” isn’t like this: much of our “knowledge” is closer to educated guesses. That’s particularly true in a pandemic situation.
I suspect, then, that the appeal of “policy-led science” rests on a bait-and-switch: we are drawn to the idea because we are drawn to a notion of science as certain, but our actual science is often far messier and scrappier. This isn’t to say that there is anything wrong with our actual science: the world is a messy and scrappy place. Nor is it to say that messy and scrappy science has no place in policy: it does and should inform our decisions. However, our actual science cannot displace the need for distinctively political judgements. Both the lockdown measures introduced in the UK and the laxer measures introduced in Sweden might be based on the science. The science cannot tell us which of those measures must be adopted: that is a political judgement.
Do I think scientists have played too large a role in our response to COVID-19? No. If anything, I don’t think they have been listened to enough. Still, we can’t think properly about these issues unless we recognise why scientists cannot and should not lead policy.