Ask most people how to think about uncertainty, and they’ll reach for probability without a second thought. Percentages, odds, likelihoods. It’s the language of weather forecasts, poker hands, and insurance premiums, so familiar that it can feel like the only honest way to talk about not knowing something for sure.
But it isn’t the only way. There’s another framework, quieter and less well known, called possibility theory, and once you see the distinction it draws, it’s hard to unsee.
Randomness Isn’t the Only Kind of Uncertainty
Probability theory does a great job with questions like: if I flip this coin a thousand times, roughly how many will land heads? What fraction of people in this city own a car? These are questions about patterns across repeated events or across a population. Probability’s whole structure, likelihoods adding up to a neat total of one, is built for exactly that kind of counting.
But think about a sentence like “the meeting will probably run late.” There’s no population of meetings here. No thousand coin flips. Just one specific meeting, described in fuzzy, everyday language. “Probably” is doing a lot of work in that sentence, and it isn’t really a statistical claim at all.
That’s the gap possibility theory was built to fill. It was introduced by Lotfi Zadeh in 1978, growing out of his earlier work on fuzzy sets, and developed much further since by Didier Dubois and Henri Prade. Instead of one measure, it uses two: possibility and necessity. Possibility asks, how plausible is this, how far is it from being ruled out? Necessity asks, how certain is this, how unavoidable does it look given what we know? Two questions instead of one, better suited to reasoning about a single, specific, vaguely described situation.
A Strange Little Wrinkle
Here’s the part that stuck with me. In probability, a high number is meaningful. If something’s 95 percent likely, you’d genuinely be surprised if it didn’t happen.
In possibility theory, a high number means something much weaker. Something can be highly possible, basically not ruled out, without being expected at all. High possibility just means “nothing here says it can’t happen,” not “this is coming.”
That’s a subtle but real trap. Treating “not impossible” the same as “likely” is an easy mental slip, and it’s the kind of slip that can quietly steer a decision in the wrong direction, whether you’re a doctor weighing an unlikely diagnosis, an analyst sizing up a risk, or just a person trying to figure out whether to worry about something.
Why This Is Worth Sitting With
Nobody needs to run possibilistic calculations at the dinner table. But there’s a bigger idea buried in here worth keeping: the tools we use to think about uncertainty aren’t neutral, and they aren’t the only option. Probability is one lens. It’s an extremely good one, but it was built for a certain kind of question, repeated trials, populations, frequencies. A lot of the uncertainty we actually live with day to day doesn’t look like that at all. It looks like a single, fuzzy, one-off situation described in words like “probably,” “likely,” or “there’s a chance.”
Maybe that’s the real takeaway. Not that everyone needs a new formula, but that it’s worth noticing which kind of uncertainty you’re actually facing before reaching for the nearest number to describe it.
References
Zadeh, L. A. (1978). Fuzzy sets as a basis for a theory of possibility. Fuzzy Sets and Systems, 1(1), 3-28.
Dubois, D., & Prade, H. (2012). Possibility theory: An approach to computerized processing of uncertainty. Springer Science & Business Media.