In The Art of Uncertainty: How to Navigate Chance, Ignorance, Risk and Luck (Norton, Mar.), the statistician explores how to think about what one doesn’t know.

You devote a chapter to exploring uncertainty’s role in science. Can you explain that a bit?

Science should be taught not as a set of established facts but as a way of looking at things. The important thing is the scientific method, which produces incremental gains in knowledge by critiquing what other people have said while at the same time respecting a body of evidence. Just because somebody conducts a new study, that doesn’t overturn everything that’s been there before, but you also have to look at the quality of the evidence. That’s part of the scientific process, which generally develops in a reasonable way, though not at all perfectly. The idea of those incremental gains is to balance uncertainty against accumulated evidence.

Is there an aspect of uncertainty that you most want to understand better?

Yes, the measurement of probabilities and what that actually means. Fifty years ago, I learned this idea that probability doesn’t really exist, except possibly at a subatomic level. It’s all a human construct, a measure of our ignorance, and I’ve never really shifted from that. This book is a chance for me to explain that point of view in more detail. In that sense, I’ve only become more convinced that when we use probabilities, they’re all based on human assumptions, and most of the time they’re not even estimating something real in the world—they’re just an expression of our judgment.

Could you expand on your assertion that uncertainty is always subjective?

Statisticians are called upon to analyze data, make statements about the outside world, and, in particular, produce intervals and measures of uncertainty. This requires realizing that every expression of uncertainty is a measure of personal ignorance, of just not knowing. That’s true whether it’s about the future, the present, or what’s happened. We all have our own perspectives, judgments, and assumptions. When we produce key values, those are all based on judgments that we’ve made, using our expertise perhaps, but they are still personal.

What are your thoughts on how politicians should address uncertainty?

The research group I’ve been working with has contributed to an increasing body of evidence finding that, actually, an acknowledgment by experts of humility increases trust, particularly in people who are rather skeptical in the first place. The very people that you’re trying to reach are those that are a bit untrusting of experts, and the claims that are being made become more trusted when they’re accompanied by an acknowledgment that there’s not just one side to the argument.