When talking about solar geoengineering, it’s tempting to offer recommendations: “We should do x,” for example, where x might mean deploying some specific version of solar geoengineering, or banning it, or whatever. This note explains why I strive to avoid such we should’s.

Policy analysis informed by science strives to predict the consequences of choices. We might predict, for example, that if emissions of greenhouse gases stop rising in this decade and are constant thereafter, then the odds are one in three that temperatures will warm more than three degrees by the end of this century.

But science alone can’t tell us what to do—how fast to cut emissions or how to deploy solar geoengineering, for example. Those answers involve value judgments and tradeoffs, such as the relative importance of today’s climate versus the climate of the distant future, or the welfare of humans versus coral reefs. Applied psychology can tell us how various groups of people view such trade-offs, but it is impossible to construct an “objective policy-analysis box” into which one feeds only facts and out of which comes answers to open-ended questions such as ‘what is the correct pace of emissions cuts?’

A scientist’s expertise in a particular area comes with no special powers for determining what society should value. If I speak as a scientific expert and make a “we should” statement, I am encouraging listeners to give undue weight to my values. People have good reason to pay attention when scientists make statements about their area of expertise, but they have no reason to give special weight to a scientist’s judgments about values or politics.

Over the course of my career, senior scientists have frequently urged me to avoid work on solar geoengineering. While they have often agreed that solar geoengineering would probably work, they don’t want it discussed publicly because of fear it will be misused—most obviously as an excuse to delay emissions cuts. For example, Ray Pierrehumbert—a brilliant and kind atmospheric scientist and a frequent critic of work on solar geoengineering—once told me that Ken Caldeira and I where roughly correct in contending that solar geoengineering could be useful if used wisely. But he said we were naïve in thinking it would be used wisely or that we could influence how it would be used.

I don’t expect decisions about geoengineering to be guided by benevolent wisdom. Chaos and conflict seem more likely. Yet the suggestion that scientists should downplay or conceal facts about solar geoengineering deeply troubles me.

The Einstein memorial outside the National Academy of Sciences says: “The right to search for truth implies also a duty; one must not conceal any part of what one has recognized to be true.”

If a scientist who has privilege to shape their own research chooses to withhold information about how well solar geoengineering could work because they fear people will make the wrong decisions, then in my view they have violate that duty. I do not want a world where decisions are made by an unelected scientific priesthood. For all its faults, politics is the appropriate arena to thrash out the hard decisions about solar geoengineering.

As I see it, concealing “any part of what one has recognized to be true”—along implying that your political value judgments should be accorded extra weight because you’re considered an expert in some technical domain—are related sins of academic arrogance.

I therefore strive—and sometimes fail—to avoid saying that “we” should or should not do something about solar geoengineering. Instead, I limit myself to contingent, “if this, then that” statements. Or I express my judgments in ways that downplay my special status as an expert, perhaps by saying I would vote “yes” in a referendum on implementing some specific kind of solar geoengineering—while emphasizing that other reasonable people might vote “no”.