Woulda, Coulda, Shoulda Followed The Science.
On (some of) the philosophy of science's concrete political implications.
In thinking about the role that science plays in policymaking it helps to (somewhat roughly) distinguish between two sorts of claims that a scientist employed in government can make, and to think about how they are related –
Claims of Efficacy – Descriptive claims about the world and what a government can do to do it given the capital (financial, human, technological, and cultural) at its disposition. For example, when a volcanologist tells government what it can do to redirect lava flows or reinforce buildings against the build-up of volcanic ash.
Claims of Legitimacy – Normative claims about a government should do given a particular state of affairs. For example, when economists on the monetary policy committee declare that interest rates should be raised/lowered in response to inflationary/recessionary pressures.
Governments seek to intervene in the world or, more precisely, particular worldly systems such as ‘the population’, ‘the economy’, or ‘the climate’. Scientists assist governments in this by providing them with one or more account of the system-in-question’s vital characteristics and information on possible interventions into it – that is, they always make claims of efficacy about what is ‘out there’ and what a government can do about it.1 They do not, however, always make claims of legitimacy as this usually falls to elected ministers or other political decision-makers of some description.
Now, I take the distinction between claims of efficacy and legitimacy to be self-evident, almost trivially so. A scientist observing that something can be done (such as culling every cat in the UK to quash coronavirus transmission) is obviously making a very different claim to one declaring that it should be done (our cats are still around…).2 Less self-evident, however, is how these sorts of claim relate to one another. Are they, for example, completely independent of one another? Can any claim of legitimacy be paired with any claim of efficacy? Or is the relation between them more substantive? Are a government’s claims of legitimacy in some way affected by the claims of efficacy that its scientists make? Or – more tantalisingly perhaps! – are the latter meaningfully shaped by the former?
Trying to answer these questions is not just some philosophical distraction. As I will argue in what follows, how we conceive of the relationship between claims of efficacy and legitimacy has significant implications for the role that we give to scientists in our (nominally) liberal, democratic policymaking and how we organise our scientific advisory bodies in government. I will use three such conceptions to illustrate this point – the David Hume-inspired ‘No Ought From Is’ position, the rationalist response to Hume, and Paul Feyerabend’s case for making ethics the measure of scientific truth.
§ NOFI beats to think about science and policymaking to.
The “No Ought From Is” position, acronymised as NOFI, holds that value-judgements about what ought to be done (of which claims of legitimacy form a subset) cannot be derived from descriptive claims about what is (of which claims of efficacy form a subset). Put otherwise, according to NOFI, scientific claims about a system’s vital characteristics and how a politician could intervene tell us nothing about how he ought to intervene in it. This follows from an argument made by David Hume in his Treatise on Human Nature.
There, Hume argued that value-judgements cannot be derived from descriptive claims because each of them is produced by a different, logically independent mental act. Descriptive claims are produced by an act of ‘reason’, the faculty that deals with relations of ideas (such as in mathematics or logic) and with matters of fact.3 Through an act – or a series of them – of reason, we make descriptive claims about the world. Reason alone, however, cannot tell us how to feel and act in response to its descriptive claims – that is, reason alone cannot tell us what ought to happen in light of its accounts of the world. This sort of evaluation is an act of what Human calls the ‘passions’ and is independent of our reasoning about what is real. Hume writes –
It is not contrary to reason to prefer the destruction of the whole world to the scratching of my finger. It is not contrary to reason for me to choose my total ruin, to prevent the least uneasiness of an Indian or person wholly unknown to me. It is as little contrary to reason to prefer even my own acknowledged lesser good to my greater, and have a more ardent affection for the former than the latter.4
This point can be clearly made by noting that the very same descriptive claim can trigger wildly divergent evaluations about what ought to be done in response. For example, where an ordinary person might greet a toxicologist’s report on substance X’s extreme neurotoxicity with apprehension and by concluding that X is best avoided, a poisoner like Locusta might respond with excitement and by looking for ways to secure a baggy or two. The report, produced by an act of the toxicologist’s reason, cannot in isolation tell you how he or others will act in response to it – to understand that you need information about their ‘passions’. For NOFI, scientific description and normative evaluation are two, wholly different activities.
Applied to our two categories, this suggests that claims of legitimacy are logically independent of claims of efficacy. According to NOFI, a scientist can, in principle, make claims about what government can do without influencing internal debates about what they should be doing. They can, NOFI insists, present the facts-of-the-matter without telling decision-makers how to feel (e.g., alarmed, reassured, galvanised, unruffled…) about them.
Now, something like NOFI arguably underlies our current institutional arrangement in the UK. When scientists are convoked by government to assist in policymaking, they are asked to make their claims of efficacy whilst totally refraining from making any claims of legitimacy. Friend-of-the-blog and regular member of government scientific advisory bodies, Neil Ferguson, explained the rationale for this stricture in his witness statement to the COVID-19 inquiry –
I believe that scientists have a key role to play in advising policymakers on the potential impacts of different policy choices in a crisis, but that they should not use the public platform offered to them by that role to campaign or advocate for specific policies. This belief is informed by two considerations:
(a) legitimacy —scientists may have more expertise in a topic but have no more democratic right to determine policy than any other citizen;
(b) trust — it can be difficult to generate the necessary levels of trust between government and independent scientific advisors if those advisors are both advising privately on the policy implications of developing science and advocating (publicly or privately) for specific policy options.5
Taking Ferguson at his word, the UK’s scientific advisory bodies are shaped by a particular, democratically-inflected version of NOFI. He draws a sharp line between claims of efficacy and of legitimacy, identifying them as fundamentally different activities requiring fundamentally different qualifications. As credentialed experts, scientists are better qualified (they “may have more expertise”) than the average citizen to exercise their reason to make sophisticated descriptive claims about the state of the world. However, as per Hume, these claims do not provide any information about the sorts of response that they should motivate. As such, scientists enjoy no special qualification to comment on their policy implications.
Instead, Ferguson insists, invoking our shared democratic ideal, this right is equally shared amongst all citizens, each of us being equally qualified to opine on the appropriate response to a given state of affairs. Of course, in practice, this right is exercised by our elected representatives, the ministers, but the institutional implications remain the same. While, for reasons of professional necessity, government scientists enjoy greater proximity than their fellow citizens to policymaking, they do not enjoy any rightful privilege over its shape and outcomes and so should leave ministers alone to puzzle and debate over the appropriate policy response to a given state-of-affairs.
However, if the link between a scientist’s claims of efficacy and a government’s claims of legitimacy is in fact more substantive than NOFI allows, then institutional arrangements predicated on their logical independence like the UK’s could prove to be fundamentally unstable. With this in mind, let’s turn to two competing conceptions of the link.
§ Priced out and Riled up.
Not everyone agrees with NOFI’s categorical distinction between descriptive and normative claims, and in this section and the next I will consider two conceptions that challenge this distinction and will spell out some of their political implications.
The first of these proceeds by arguing that description and evaluation aren’t fundamentally different activities and that the latter, properly speaking, is but a species of the former. Consider Alasdair Macintyre’s account of one of Hume’s successors and critics, Richard Price –
Price, a Unitarian minister, was perhaps the most important of Hume’s immediate successors. [He] argues in his Review of the Principal Questions and Difficulties of Morals (1757) that moral questions are intellectually grounded just as the rationalists said that they were, and that the apprehension of first principles in morals is not a matter of argument but of grasping their self-evidence.6
Contra Hume, Price argued that normative claims are products of the same faculty of reason as descriptive claims. For Price, we understand what ought to be done by looking upon our worlds with eyes unclouded by bias or feeling – evaluation is effectively a form of description. Interestingly, echoes of Price’s conception were more recently heard during Steven Riley’s (another friend-of-the-blog and a colleague of Neil Ferguson’s) appearance before the COVID inquiry. There, the inquiry’s lead counsel pressed Riley on his understanding of the link between a scientist’s claims of efficacy and of legitimacy –
Inquiry: Professor, this is a point that you expand on in your witness statement, this issue about scientific advocacy [i.e., making claims of legitimacy] or scientific evidence [i.e., making claims of efficacy]. What was the difference of opinion here and what was your take on it?
Stephen Riley: So, I think we should be very careful in describing a view as advocacy and another view as evidence-informed scientific opinion, and I think – I don’t think I say so explicitly here or in other evidence but I think I probably show, I felt that I had an evidence-based opinion that covered recommendations on interventions. As I’ve mentioned before, our scientific discipline includes the study of interventions and I had an evidence-informed opinion for one intervention over another.7
Though he is clearly irked at being described as an advocate, Riley nonetheless insists that his scientific expertise in making descriptive claims of efficacy also qualifies him to make normative claims of legitimacy. For him, there is no clear demarcation between the two – both are ‘evidence-based’ activities and, as a scientist that studies policy interventions (he is an epidemiologist of some form), he is eminently qualified to mobilise and interpret that evidence. In this, Riley effectively rejects Ferguson’s democratic NOFI, much as Price once rejected Hume’s arguments. For if describing possible policy-options and opining on their appropriate shape are the same sort of ‘evidence-based’ activity (much as Price thought that description and evaluation were the same sort of descriptive activity), then it is absurd to insist that a scientist is only more qualified than the everyman to do the former. If claims of efficacy and legitimacy belong to the same species of ‘evidence-based’ statement, then scientists are best placed to make both.
The political implications of this conception are clear and well-precedented in the real world. In many countries – including the UK – central banks have been freed from political control and made ‘independent’. This means that monetary policy has been wholly entrusted to a cadre of unelected technical experts (that is, central bankers), whose expertise in making descriptive claims about the economy and the options available to government is taken to entitle them to make decisions about the shape of government policy without needing to consult elected ministers.8 Or, in my jargon, their competence in making claims of efficacy is taken to entitle them to make claims of legitimacy and to vest them with the authority to make decisions about the government’s course of action. Taking Riley’s reasoning to its logical end would mean granting all government scientists (and not just central bank economists) a similar level of control over the shape of policy.
Independent ‘Public Health Policy Committees’, anyone? 🥶
§ On a Feyerabender.
So, having survived our momentary brush with the dystopian, let’s now turn to the third and final conception that I will consider in this post – Paul Feyerabend’s analysis of scientific descriptions in Ethics as a Measure of Scientific Truth.9 If Riley and Price reject NOFI by – in some sense – collapsing evaluation into a form of description, Feyerabend rejects it by swinging in the opposite direction and arguing that seemingly neutral scientific descriptions smuggle in value-judgements about what matters and how. To understand how he does this, we need to return to the claim that I made about what scientists do. I stated that, at minimum, scientists make descriptive claims about what is real and ‘out there’ – and, when relevant, what a government can do to affect it. In his article, Feyerabend adds to this that descriptive claims about what is real always reflect scientists’ particular value-judgements. Let’s briefly consider how he does this.
Per Feyerabend, scientists working in the ‘Western’ tradition tend to be realists, which is to say they tend to assume that their descriptive claims correspond to features of an observer- and history-independent world.10 And this realism, Feyerabend observes, often contains a normative (which is to say, evaluative) component. When scientists designate certain things as ‘real’ according to their descriptive claims, they effectively dismiss other as ‘unreal’ – scientific materialists insistent upon man’s essential fleshiness dismiss the immaterial soul; evolutionary biologists dismiss the divine creator; and some physicists and chemists even dismiss as illusory any ‘macro’ object (e.g., trains, planes, and automobiles) that falls beyond the reach of their own entities of study (e.g., molecules, atoms, and subatomic paraphernalia).11 More often than not, these dismissals entail the judgement that human behaviours and efforts are best arranged around what is ‘real’, and that what is ‘unreal’ is to be dispensed with as much as possible – e.g., science dismisses the soul and divine creator, so we must stop walking and talking as though they are real. Lying dormant within their descriptive claims about the world are value-judgements about what matters most to human life and practice. To borrow a phrase of Bruno Latour’s, scientists deal not only in ‘matters-of-fact’ but also in ‘matters-of-concern’.12
If so, this suggests that scientists’ claim-making, and the subsequent acceptance of particular claims as ‘true’, is (at least partially) influenced by the regnant conventional understanding of what matters. For example, Feyerabend suggests that Western science’s use of ‘real’ reflects our preference for “forms of coherence that can be managed without too much effort”.13 I take this to mean that we tend to prefer accounts of the world that boil it down to a few basic entities and laws/relations between them, framing it as a closed-off mechanism that one can intervene into or ‘hack’, and tend to reject ones that emphasise other things like complexity or human freedom.
While Feyerabend’s argument is (to me) basically compelling, it isn’t perfect. For starters, while both our formally scientific and more colloquial uses of the predicate “real” do seem governed by value-judgements of some sort, it remains unclear whether Feyerabend was aiming at science-in-certain-contexts or science-in-general. It remains unclear, that is, whether he meant that scientists’ descriptive claims are only sometimes (perhaps mostly) evaluative or always so. The fact that scientists who describe our lives and everyday effects as ‘real’ seldom, if ever, recommend that we do away entirely with the idea of trains, planes, and automobiles suggests to me that reality is more complex than Feyerabend’s curt exposition allows.14 Secondly, Feyerabend’s terms are imprecise. ‘Western science’ is, I confess, a notion broad enough to test the outer-bounds of my credulity and patience and, even if we accept his analysis of realism’s normative content, it is flatly implausible that all scientific claim-making is shaped by the lone preference that he identifies. Surely, different disciplines will have different boutique concerns – are we really to believe that quantum physics or astronomy are governed by the exact same norms as obstetrics or behavioural genetics?
That said, his argument certainly applies to the descriptive claims made by government scientists. The scientific accounts of the world presented to ministers necessarily reflect a number of value-judgements made by scientists about what concepts and techniques are most relevant to understanding a given situation and what matters-of-concern a government should be organising its efforts around. For example, the notorious compartmental models of SARS-CoV-2’s spread presented to British ministers by Ferguson, Riley, and friends in 2020 reflect their judgement that what mattered most (both in the sense of understanding and intervening) at the time was case-numbers/disease transmission, deaths from COVID-19, and demand for intensive care unit beds. And these claims are made and evaluated in contexts (i.e., modern governments) that value intervention-facilitating simplifications. Unlike Hume and Ferguson, but like Price and Riley, Feyerabend muddies the clean line that NOFI draws between scientific description and evaluation. Unlike Price and Riley, however, he does this by identifying description as a form of evaluation rather than the reverse.
The political implications of this argument are less clear than before. Nonetheless, I will very briefly draw out a couple as I see them. If the descriptive claims made by government scientists inevitably reflect the preferences and norms of those who make and evaluate them, a good place to begin would be to stop overlooking or obscuring this fact. We should stop treating scientists as ‘apolitical’ or ‘value-free’ agents (as Hume, Ferguson, Price, and Riley each do in their own way), allow them to express their preferences, perhaps even explicitly as policy recommendations, and start thinking of institutional reforms that might guarantee space to possible counter-preferences. We could, assuming that different disciplines reflect different preferences, widen the diversity of the government’s scientific advisory roster and insist that non-natural scientific disciplines like ethics, theology, social anthropology, and economics must always be represented there. Or, following Roger Koppl’s suggestion, we could think about practices like ‘red teaming’ whereby groups of advisers are asked to scrutinise and poke holes in each other’s claims and recommendations.15 Or, finally, we could conclude that that no amount of interdisciplinary origami can do justice to the rich array of preferences woven through our social fabric. In this case, we might turn to a system of inviolable civil liberties that places a portion of people’s everyday lives beyond the reach of policymakers and scientists and guarantees their (people’s, that is) capacity to broadly pursue their own preferences, whatever the situation or crisis. Too-broadly-speaking, Feyerabend’s argument pushes us to find ways of enshrining value- and preference-pluralism at the highest levels of government.
Of course, NOFI, Price-Rileyism, and Feyerabendery are not the only conceptions available to us. They have just served to illustrate my overarching point – that how we conceive of scientific evaluation’s relationship to normative evaluation has important, concrete political implications for how we organise our governmental institutions.
Paul Rabinow and Nikolas Rose, ‘Biopower Today’, BioSocieties 1, no. 2 (1 June 2006): 195–217, https://doi.org/10.1017/S1745855206040014.
Telegraph Reporters, ‘Government Considered Killing All Domestic Cats at Start of Covid Pandemic’, The Telegraph, 1 March 2023, https://www.telegraph.co.uk/politics/2023/03/01/government-considered-killing-domestic-cats-start-covid-pandemic/.
Alasdair Macintyre, A Short History of Ethics (Routledge Classics): (Routledge, 2002), 164.
David Hume, ‘A Treatise of Human Nature’, https://www.gutenberg.org/files/4705/4705-h/4705-h.htm, 2022 2003, bk. 2, https://www.gutenberg.org/ebooks/4705/pg4705-images.html.
Neil Ferguson, ‘INQ000249526 - First Witness Statement of Professor Neil Ferguson, Dated 11/07/2023. UK Covid-19 Inquiry Archives’, UK Covid-19 Inquiry, 2023, https://covid19.public-inquiry.uk/documents/inq000249526-first-witness-statement-of-professor-neil-ferguson-dated-11-07-2023/.
Macintyre, A Short History of Ethics (Routledge Classics):, 170. (Emphases mine).
UK Covid-19 Inquiry, ‘UK Covid-19 Inquiry - Module 2 Hearing - 17 October 2023’, 19 October 2023,
Annelise Riles, Financial Citizenship: Experts, Publics, and the Politics of Central Banking, Illustrated edition (Ithaca, NY London: Mario Einaudi Centre for International Studies, Cornell University, 2018), 9; Mitchel Y. Abolafia, Stewards of the Market: How the Federal Reserve Made Sense of the Financial Crisis (Cambridge, Massachusetts: Harvard University Press, 2020), 153–71.
Paul Feyerabend, Conquest of Abundance: A Tale of Abstraction versus the Richness of Being, ed. Bert Terpstra, 2nd ed. edition (Chicago, Ill.: University of Chicago Press, 2001), chap. 9.
Feyerabend, 170 & 245.
Feyerabend, 247.
Bruno Latour, ‘Why Has Critique Run out of Steam? From Matters of Fact to Matters of Concern’, Critical Inquiry 30, no. 2 (January 2004): 231, https://doi.org/10.1086/421123.
Feyerabend, Conquest of Abundance, 247.
As, to be fair, does the fact that many of these scientists go on penning essays, giving talks, and publishing pop-sci books that smugly describe our everyday lives and cares as ‘illusions’, even as their authors admit that in practice we, mere fleshy mortals, are condemned to go on leading and worrying about them…
Roger Koppl, ‘Public Health and Expert Failure’, Public Choice 195, no. 1 (17 September 2021): 101–24, https://doi.org/10.1007/s11127-021-00928-4.
I continue to maintain that studying, working in or calling oneself a scientist doesn’t make it so - particularly if you only focus on part of the picture because that’s all you’re aware of, and you reject anything outside of it. One can’t know everything, but at least one should be open that much exists outside of the limited realms they've learned about. Most often the attitude of scientific experts is “If I don’t know it, it can’t exist.”
Not that it wasn’t evident before covid, but scientists have certainly revealed their arrogance during this period. As far back as April 2020, I read a news piece about an Oxford-AZ researcher rueing the funds then spent on PPE, saying the money should go to the likes of her because what she was doing was more important. Not that I ever believed in the “science” behind vaccination, nor even back then believed any forthcoming vaccine would be safe, she gave me even more reason not to trust researchers. Since then I’ve repeatedly observed not only the same arrogance but also defensiveness among such science types, plus a strong obsessiveness about their work including overstating its value. That’s what I see as one of the reasons they can be easily influenced through grant money.
I appreciate your writings even though I don’t always find it easy to follow all of them. I’m glad I persisted with this one as I learned a lot. I’m kind of on the same page as Feyerabend, but I like the way you’ve presented both sides. Like him, I’d say that medical scientists dismiss (or dissociates from) physiology and pathophysiology if acknowledging it confuses their practice of prescriptions. That’s why adverse reactions are often played down or decided to be all in the patient’s head. The evident limitations in the minds/attitudes of many scientists and government officials revealed during covid makes broad personal choice even more necessary.