A seemingly never-ending controversy around the practice of science is its relationship to politics. A frequently heard view that ‘science should be apolitical’ clashes with the equally often heard ‘everything is political’. That tension sits at the intersection of both more fundamental debates (e.g., the possibility of ‘value neutral’ science) and very concrete legal rules regarding academic free speech. The tension also feels like it has become more acute in recent years.

Both sentiments may persist because they feel somewhat right. The following is an attempt to help resolve that controversy, clarify what actually is right (and what is wrong!) about both intuitions and why the distinction between science and political action matters.

At the heart of this attempt is the simple suggestion that science and political action are associated with different modes of engagement.

Scientific integrity, to me, requires engaging in good faith with all of the relevant evidence which includes a demand to engage with arguments by 'the other side'. Politics or activism do not require me to do that in any shape or form (though some such engagement might be instrumentally useful). That's the actual tension.

This matters because scientists, as individuals, are subject to both of these sets of norms. As individuals, they have a right to political expression and political action. In fact, this is one of the elements of academic freedom. Presumably this means that as politically engaged individuals they have the same rights and (only) the same obligations as other citizens. Acting as scientists, though, they have a commitment to good faith evaluation of the totality of evidence in ways that citizens don’t; and because it is their good faith evaluation of evidence that is owed, it is inappropriate, in particular for state actors, to tell them what that evaluation should be (and that is also part of academic freedom).

The fact that there are two very different standards here creates problems when the two contexts are blurred, confused or conflated. If a position presented in public discourse represents my own political advocacy it is fine if the way that I arrived at it or the way I promote it are highly selective (because there is no norm that requires political advocacy to be even handed). But it’s misleading if that advocacy is erroneously read by others as the outcome or product of the norms of scientific engagement when it is not.

This suggested distinction between science and politics side-steps long-standing, distracting, debate about ‘objective’ versus ‘subjective’: One can readily agree that my evaluation of evidence as a scientist rests, in some way or to some degree, on my entire person and that my evaluation of the evidence will consequently be irreducibly subjective in some way. There remains a difference between me trying to evaluate a total body of evidence in good faith by the standards of my discipline and me merely strategically picking evidence to bolster a claim regardless of what competing evidence might exist or how such an overall evaluation would go.*

Nor does the suggested difference between scientific and political modes of engagement rest on the presence or absence of values, and one does not have to commit to the (unrealistic) demand that science be ‘value free’.

Finally, the distinction makes clear that the relevant norms are about process (action) not about outcomes. This has been obscured by the ambiguity inherent in the phrase ‘to be political’, which fails to distinguish clearly between an (inherent) property of a statement and an activity.

Scientific conclusions may well ‘be political’ in the sense of having implications for politics. In fact, this seems inevitable if we want research to be ‘societally relevant’ and choose to assess its quality in part by ‘impact’. Even if all conclusions are political in this sense, though, there is still a difference between a ‘political’ and a ‘scientific’ mode of engagement. Scientists can still fall afoul of the distinction between doing science and doing politics in ways that are misleading or transgressive. At the same time the distinction makes clear that scientific conclusions aren’t inappropriate simply because they have political implications, even where those implications are strong and contentious.

Scientists cannot owe specific conclusions. This follows from what it is they do owe: a particular mode of engagement toward the collection and evaluation of evidence. What conclusion to draw is to be determined by that process. So asking scientists to stay clear of politics in the sense of ‘avoid politically relevant conclusions’ is a non-starter for multiple reasons and incompatible with academic freedom.

Equally unproblematic seems the idea that scientists, as private citizens, should be free to engage in the political process (in the best case, they may even bring to it useful skills and expertise). In the action sense of political (that is, ‘active in the political process’ or ‘engaging in political action’), scientists clearly also have the right ‘to be political’.

Despite the fact that, individually, these two sets of considerations seem straightforward and clear, their intersection is not. The conflict between the competing demands is real. Most importantly, I think, we have created conditions where they are virtually guaranteed to conflict through the way we have fostered the public engagement of scientists.

Over the last decade, it’s been a consistent development across many countries that scientists are not just encouraged but increasingly expected to engage with the public. In some of these contexts, it may be clear whether scientists’ communications reflect an underlying scientific (as opposed to political) mode of engagement: for example, scientists discussing their work at science fairs. But in others it really is not. Chief among those is online social media. If I make a public pronouncement on a ‘political issue’ (as in ‘topic with political implications’), others have no way of knowing whether I’m commenting in my capacity as a scientist or as a private citizen from my statement alone. This means they also have no way of knowing, just from the conclusion or piece of evidence offered, what epistemic standards and what underlying mode of engagement it reflects.

They may be able to guess something from the nature of the topic (me saying ‘we need more playgrounds in this neighbourhood’ seems a far cry from a research profile that says ‘works on human rationality with a particular interest in communication across social networks’) but in many cases they won’t. In fact, things may be least clear on those topics which most relate to my academic expertise, where my words are likely to have greatest credibility and greatest effect (if they have any).

This is made all the more difficult by the fact that social media is so restrictive in the way it pushes one to discuss only a single thing at once. Human communication itself is intrinsically selective in that I’m typically choosing one thing from many different things I could say. So we routinely interpret both what is said and what is not as indicative of underlying speaker intention. Online, identifying speaker intention is additionally complicated by the fact that one often knows little about the speaker. This is presumably why it’s become routine, online, to interpret every bit of data available, including who people follow and who they retweet. That then further shifts focus away from (just) what has actually been said, in this case toward the sources of content.

These features of online social media are in direct opposition to what the scientific mode of engagement needs: good faith evaluation of all of the evidence requires me to be able to tackle different arguments and individual pieces of evidence bit by bit. It is impossible to scrutinise a substantial body of relevant considerations all it once, let alone to simultaneously discuss them. Furthermore, good faith evaluation requires me to be able to do the evaluating before committing to a final position. If my conclusion is unalterably set in stone, come what may, I’m not really basing it on the evidence in any meaningful way (and whatever it is I’m doing isn’t really science).

In other words, the specific pressures of the medium make it incredibly difficult to talk about things from the perspective of science because analysis that is step by step, part by part, and (still) withholds a strong conclusion or stance is very likely to be met with interpretations of speaker intent and position that far outrun those actually in play. An illustration of this in my social media corner of the world is the way the slightest acknowledgement of reasonable, let alone good, AI performance on a specific task is invariably (mis)interpreted as promotion of AI as such. In short, bit-by-bit evaluation and discussion of evidence is misread as activism about some unstated underlying position.

Conversely, what genuinely is political advocacy seems likely to be misread as representing even-handed scientific evaluation. This is bad because it may falsely suggest the rigour and robustness that the scientific mode of engagement is meant to deliver and may lead listeners to make inferences about the underlying body of evidence that are at odds with what’s actually there.

So the confusion goes in both directions. I also think it’s not just non-scientists getting confused, I think we as scientists are sometimes also confusing ourselves. I don’t know how to devise communication in ways that avoid such confusion. I’ve had a social media account for a few years now that has a profile tied to myself as a scientist. I’ve used that with a self-imposed restriction to posts that are related to (and relevant for) my science. But implementing decisions about what and what not to post in line with that goal is frequently challenging in practice and the limits of what that allows me to talk about have often been quite uncomfortable.

I very much don’t have the solution for public engagement by scientists; and I have more questions than answers about my own engagement (both online and offline) with issues of pressing political concern. Maybe being clearer about the respective demands of the two modes of engagement is part of finding those answers.

———————————————————————————————————————————————

* It seems worth note that the value of the distinction doesn’t require it to be possible to draw a clear line between good faith evaluation and strategic cherry picking in every case any more than the value of the everyday concepts like ‘cup’ and ‘bowl’ require that we can identify some definitive boundary between them.