This piece was co-authored by Nason Maani, and his version is cross-posted here
We are largely rationalists. After all this essay—part of our ongoing Ideas about Ideas series—is being written from within universities where we spend our careers trying to understand data to identify fact. While we are both often working on social science contexts, our disciplinary training began in medicine and biomedical science, making us even more wedded to empirically grounded, expert-led, evidence-based analysis. There is no question that this approach has done enormous good and remains the foundation of science and what we do in universities. But are there limits to the insights that can emerge from such rationalism? What would a more honest public health look like if we acknowledged the limits of this approach?
If we look closely enough there are plenty of examples where public health hit a wall, not because the evidence was wrong, but because the question that dominated the questions of the moment was not one that any kind of study can settle. Think for example about the recent pandemic, where much animus and political division was cloaked in “what do the data say?” or “follow the science” when in fact what we should have been asking was “what should we do, balancing data and other things that we value as a society?”. Think of the opioid epidemic that was fueled in part by OxyContin marketing. It was in large part built on technocratic claims about addiction risk that turned out to be an ideology about pain medications masked as data. Or think about any number of academic debates where we preserve the peace by deferring to a methodological difference in perspective when in fact the foundational question was about what we were trying to understand in the first place. Technocratic thinking emerges when we treat the answer to societal questions as though they were merely technical questions. But so very few questions that matter to societies, certainly in health, are technical in nature.
What makes technocratic thinking attractive is worth taking seriously, even as we set out to understand the bounds of that thinking. Technocratic thinking offers the semblance of neutrality at a time when public argument has been swallowed by partisanship. It offers a safe haven for those upset at anger against elites or institutions to think of something that will at some point be returned to. It also offers a way to bridge ideological difference by appealing to shared standards of evidence, and defers to expertise and to data, which seems honest. Even more appealing, technocratic thinking has much success to show for itself. Technocratic thinking emerged in the progressive era and postwar managerial states were genuinely reformist, bringing about gains in safety, fairness, and efficiency that we now take for granted. This approach has particular importance, and appeal for, public health. The field prizes itself on promoting evidence-based approaches that can be applied to improve the health of the public, and to do so in a way that stands apart from ideological considerations. We have previously argued for a small-l-liberal conception of public health which rests, in no small part, on a data-informed approach that transcends faith-based thinking; in other words, what could be argued to be a technocratic public health.
What then might be the challenge? The fundamental issue is that the fact/value distinction is much thinner than it looks. Ultimately, every action and policy choice involves a tradeoff, and a tradeoff is a values question. Michael Sandel has spent a career making this point, most directly in The tyranny of merit, arguing that liberal societies has long tried to settle questions of justice and the common good through procedural and technical reasons, and have ended up with shallower and more brittle public conversations as a result. When seen through this lens, there is very little in public health that is purely a matter of evidence. Ultimately, “to govern, is to choose”. This means that there is little in what we do that is divorced from values which makes the idea that we can get by on a data-based approach alone at face value indefensible and exposes us to legitimate criticism from skeptics of the field.
This philosophical challenge to a technocratic approach translates, perhaps not surprisingly, to practical challenges. This case is neatly presented by James C Scott in Seeing like a state. In an effort to make populations more governable, states sometimes simplify approaches in ways that lose the local, particular, or practical knowledge that people on the ground actually have and that nudge “what we know from the data” in consequential ways. This results in failures of implementation of efforts that should, as suggested by the best available data, work. It is then not surprising that this results in loss of trust in science and evidence-based approaches. Friedrich Hayek made a closely related argument in his 1945 essay The use of knowledge in society, that the knowledge required to coordinate complex systems is dispersed and tacit, and cannot be assembled in any single planner’s head, guided by any single set of data from other areas. There are plenty of real-world examples of where this thinking has resulted in failure, including, for example, the limits of development projects where the best-intended data-informed efforts have collided with realities on the ground and the human capacity to improvise around them. Even in the context of data itself, its quality, coverage, nature, and built in assumptions varies enormously by geographic region, and much more a result of what is measurable in a given context, rather than what is most important in that context. The pattern in all of this can often result in similar challenges: when the technocratic overestimates what is or can be known, implementation disappoints, systems fracture, and we collectively lose a bit of faith in the evidence that underlay the whole project, even if that evidence could not reasonably have been expected to deliver implementation success.
A favorite philosopher, Michael Oakeshott, in a 1947 essay, argued that the rationalist, in his account, is a figure for whom every situation is a problem to be solved by the application of general principles, who values codified knowledge, and distrusts the tacit knowledge that comes only from practice and nuanced values-based thinking. When applied to politics, this approach aims to remake institutions according to a blueprint, resulting in institutions stripped of the very practical wisdom that makes them work. While Oakeshott’s writing emerges from his own conservative priors, the argument extends well to the concern with technocracy we are discussing here. Chirstopher Lasch, similarly, writes more recently about how professional and managerial classes have lost the trust of the public they speak for. The heart of these critiques is similar: it is not that the approaches taken are getting the data wrong, but that a data-informed solution set is being mistaken for an approach that requires judgement and value-informed decision-making.
If we accept then that public health questions are values-laden, we then need to situate technocratic approaches within a more nuanced place in the public health armamentarium. Recommendations do not simply arise from data, but rather they carry values, and we do better when we name them than when we try to hide the values under technical language, and when we seek to articulate a moral as well as empiric case, in anticipation that it will be part of a wider consultation among competing values . With this perspective, the process coming to public judgement, becomes a core part of what public health does not simply a “task” to be added to our list of to dos. This makes us participants in democratic conversations, contributing to what is known in accessible ways that allow for value-based decisions, without us necessarily claiming to settle what should be done. A corollary here is that it behooves us to make sure that the limits of what is known are made plain, minimizing the chance that we lose further authority by over-claiming what we do know.
None of this is an argument against expertise, against evidence, or against the steady work of producing good science. Nor of course does it excuse individual scientists from their other responsibilities as citizens in wider society. At the heart of this argument is the observation that science and expertise are necessary and insufficient. The technocratic temptation to treat science and expertise as though they were sufficient is simply wrong. Ultimately what we care most about is not the kind of question that any single method can settle. Perhaps if we had one goal, it should be to recover an older tradition of public reasoning, where expertise is one voice among several and the public is treated as a serious interlocutor rather than the target of communication. We are not so naïve as not to realize the challenges with this formulation in a world where there is much cynical manipulation of public engagement for political gain, or indeed for personal advancement at a time when public engagement can be so readily monetized. And yet, surely part of our goal should be to sketch aspirations, even if getting to them may prove difficult in the short term.
A public health that knows what it knows and is honest about what it does not know how to know is a public health that we should be aiming towards. That means a public health that argues for its values out loud without hiding them behind the language of evidence. But it is also a public health that listens to other disciplines and, more importantly, to dissent, and is willing to be changed by what it hears. Perhaps we are trying to ensure that public health does not fall prey to the question posted TS Eliot, “where is the wisdom we have lost in knowledge?”. Our collective job is to make sure that does not happen.
__ __ __
The Ideas about Ideas project
The Ideas about Ideas project, co-written with Dr Nason Maani, and cross-posted on the Money, Power, Health substack is a series of reflections of how thought lives in the world. It follows ideas as they emerge in institutions and everyday life, as they move across borders, as they harden into orthodoxy or fade into silence. It argues that ideas are not possessions but shared conditions of living, and that to see them clearly is to recognize both their power and the obligations they impose. Fundamentally, this project is built on the notion that without ideas we have nothing, and that the business of creating a healthier, better world rests on ideas.
__ __ __
Also this week
I enjoyed joining Jessica Steier, of Unbiased Science, for a Substack live conversation about my new book, Why Health? What We Need to Think About When We Think About Health. ICYMI, a recording of our conversation is available here.
The OxyContin example is such a sharp one. I've been writing about opioid potency lately, and so much of what people "know" about these drugs arrived as marketing dressed up as pharmacology. Naming the values out loud seems much more honest than pretending the data decided it for us.
Thank you for this thoughtful and provocative post. It reminds me of our efforts way back in the day to implement higher taxes for alcoholic beverages. It was a top shelf WHO "best buy" and we pushed for it hard. I fully endorsed this policy. And it ended up backfiring socially and politically — because the evidence did not account for the societal backlash that led thousands to start buying their alcoholic beverages across the border.
Your post made me look back on this experience, having spent the past six years trying to change a much smaller ecosystem — a 5000 head hospital. From the get go, I've chosen a very different approach. Primarily, I have been hesitant to apply any policy in our hospital, no matter how well supported, as a top-down imperative. Not because I don't trust the evidence, but because I do not sufficiently understand the implementation context. Every policy needs to account for its effects on a myriad of processes, many of which are tacit and hence only fully understood by the workforce. So trying to change policy without consulting the people affected is at best a lost battle and at worst a recipe for disaster as people scramble to adjust their work around a growing body of ill-fitting requirements. The work around Patient Safety 2.0 and complex adaptive systems has been useful in guiding our thinking. I wonder if we'd have done anything differently if we had considered our understanding of how the world works not as fact but as a sort of imaginary reality (akin to "work as imagined") to be contrasted with the actual behavioural patterns of our population (the equivalent of "work as done"). The complexity inherent in a society is of course orders of magnitude greater than it is in a large organization.
Values absolutely come into the picture. From the perspective of a "technocrat", I think we sometimes take them for granted. We do X because of Y, and Y needs no explanation. Or if it does, it should be explained by somebody else. But nobody's going to want to implement something they don't value. I've seen this happen for example with rules around sponsorship. We assume the independence of our workforce from the influence of pharmaceutical companies is important, but I've found few healthcare workers to really share in this belief, and usually only in the most extreme cases. A study that shows a 20 dollar lunch is enough to influence prescribing behaviour therefore falls on deaf ears, if it is communicated at all.
Partly our reluctance to get into the messy world of values is a direct result of the expectations we set for technocrats. As a government official I was pretty heavily discouraged from speaking directly and candidly to the population. That was the job of ministers. As a hospital official I would find it impossible to do my work if I couldn't and didn't spend significant sums of time communicating directly with the workforce about why patient safety or some other "broadly accepted" end goal is actually personally important to them.
Thanks again for the post and looking forward to a continued discussion!