Uncertainty

• 1. Introduction

  ›

  Uncertainty touches every domain of human thought, from quantum physics to the price of insurance, from the indecision of Shakespeare's Hamlet to the fine print on a weather forecast. It sounds like a simple word for not knowing things. But specialists in decision theory, statistics, philosophy, and economics have spent centuries arguing over what it actually means and how it differs from risk. The gap between those two ideas turns out to matter enormously. A surgeon deciding whether to operate, an investor weighing a rare catastrophic event, a city planning an outdoor festival while watching storm clouds build on the horizon: each faces something distinct. Understanding what kind of not-knowing they face changes what they can do about it. This documentary follows uncertainty into the places where that distinction has the highest stakes, and asks what it means to act wisely when the information you need simply does not exist.

• 2. Risk Versus the Unmeasurable

  ›

  In 1921, the economist Frank Knight drew a line that would influence how thinkers across many fields approach decision-making. Knight defined uncertainty as a lack of knowledge that is immeasurable and impossible to calculate. Risk, by contrast, involves outcomes whose probabilities can be estimated. To Knight, risk is hypothetically insurable; uncertainty is not. His own words were direct: "Uncertainty must be taken in a sense radically distinct from the familiar notion of risk, from which it has never been properly separated."

  Knight went further, arguing that a known risk carries no special reward. Because its probability distribution is clear, the cost of bearing that risk can simply be priced in. It is only when someone bears a genuine uncertainty, with no known expected probability distribution, that extreme outcomes become possible and entrepreneurial reward becomes conceivable. This is why Knight connected uncertainty directly to entrepreneurship: those who absorb truly unmeasurable risk are, by his definition, entrepreneurs.

  The weather example makes the difference concrete. If it is unknown whether rain will fall tomorrow, that is a state of uncertainty. Once a forecast assigns a 90% chance of sunshine, the situation becomes one of quantified risk. If a business event is planned outdoors and a rain loss would cost $100,000, the expected opportunity loss of the 10% rain probability is $10,000. An insurance company would use that figure as a floor before adding operating costs and profit. Most people, being risk-averse rather than risk-neutral, are willing to pay above that floor, which is precisely why insurance markets exist.

• 3. Varieties of Not-Knowing

  ›

  Vagueness, ambiguity, and second-order uncertainty are three distinct forms that uncertainty takes, and confusing them leads to poor decisions. Vagueness arises when an analyst cannot clearly separate two categories, such as distinguishing a person of average height from a tall person. Lotfi Zadeh's fuzzy logic, and a related framework called subjective logic, offer tools for modeling this kind of graded, boundary-blurring uncertainty.

  Ambiguity cuts deeper. The sentence "He returns from the bank" is the classic illustration: its meaning depends entirely on whether bank means a riverbank or a financial institution. Ambiguity can also describe a situation where the range of possible outcomes is known but their probabilities are not. Daniel Ellsberg became famous for urn experiments that demonstrated how people respond to this kind of ambiguity differently from how they respond to quantifiable risk, and how they tend to avoid ambiguous situations even when the expected values are identical.

  Second-order uncertainty sits above both. Expressed as confidence about outcome probability estimates, it appears as a probability density function layered over first-order probabilities. In statistics and economics, this structure captures not just the odds of an event but the degree of confidence in those odds. Opinions in subjective logic carry this type of uncertainty built in. The term radical uncertainty, popularised by John Kay and Mervyn King in their book published in March 2020, marks the far end of the spectrum: uncertainty that is not resolvable even in principle, because there is no means available to acquire the knowledge that would resolve it. That distinguishes it from Knightian uncertainty, which is immeasurable but may still have a knowable structure.

• 4. Quantum Limits

  ›

  At the subatomic level, uncertainty may not be a matter of ignorance at all. The Heisenberg uncertainty principle places hard limits on what an observer can ever know about the position and velocity of a particle simultaneously. This is not merely a practical limitation of measurement technology. There is genuine controversy in physics over whether such uncertainty is an irreducible property of the universe itself or whether hidden variables exist that would describe a particle's state more precisely than Heisenberg's principle allows.

  If the first view is correct, the universe is not just difficult to measure; it is, at its smallest scales, genuinely indeterminate. That would mean uncertainty is not a gap in human knowledge but a feature of reality. The debate between those positions has not been fully settled, and it shapes how physicists interpret the foundations of modern quantum mechanics, for which the Heisenberg uncertainty principle forms a central pillar.

• 5. Measuring What We Can

  ›

  Where uncertainty cannot be eliminated, it can often be quantified and communicated precisely. The "Guide to the Expression of Uncertainty in Measurement," published by ISO and widely known as the GUM, describes the most commonly used procedure for calculating measurement uncertainty. Related works include the National Institute of Standards and Technology Technical Note 1297 and the Eurachem/Citac publication on quantifying uncertainty in analytical measurement.

  In metrology, physics, and engineering, measurement uncertainty is typically expressed as a range of values likely to enclose the true value. The standard uncertainty assumes an approximately Gaussian distribution and represents one standard deviation. On a graph, this appears as error bars; in notation, as a measured value followed by a plus-or-minus figure. A more concise form encloses the uncertainty in parentheses next to the least significant digits, a convention used by bodies such as IUPAC when stating the atomic mass of elements.

  The statistical implications are precise. When uncertainty represents the standard error of a measurement, the true value falls within the stated range about 68.3% of the time. Doubling the interval width brings that figure to roughly 95.4%, and tripling it to roughly 99.7%. These figures follow directly from the properties of the normal distribution. About 31.7% of atomic mass values on the standard list of elements by atomic mass are expected to have true values that fall outside their stated uncertainty ranges, a built-in acknowledgment that no measurement is perfect.

  Measurement uncertainty also depends on both accuracy and precision. An instrument can be precise, meaning its repeated readings cluster tightly together, without being accurate. When an instrument is inaccurate, the true uncertainty is larger than the standard deviation of repeated measures alone would suggest, because systematic errors are not captured by repeated observation.

• 6. Uncertainty in Public Life

  ›

  Science produces uncertainty as a matter of method. The public, however, often receives a distorted version of it. Journalists may inflate uncertainty by placing scientists with minority views on equal footing with those holding majority views, without clearly describing the state of scientific consensus. They may also downplay it by stripping away the careful tentative language scientists use, leaving findings that sound more definitive than they are. A single-source story with no context from prior research presents its subject as more settled than it is. Framing science as a triumphant quest tends to cast uncertainty as a temporary problem already on its way to resolution.

  The climate debate offers a documented example of deliberate management of this perception. Climate change deniers took the advice of political consultant Frank Luntz to frame global warming as an issue of scientific uncertainty. That framing became a precursor to what journalists call the conflict frame, a way of presenting a scientific question as an ongoing dispute rather than a settled matter. Meanwhile, coverage of fields like plant biotechnology and nanotechnology in the United States tended to lean on economic and social progress frames, which naturally downplay uncertainty in favor of promise and productivity.

  Underlying this gap is a distinction that scientists find difficult to convey. Indeterminacy, where not all parameters of a system and their interactions are known, is different from ignorance, where it is not known what is not known. Both are routinely transformed into the word uncertainty when communicating with the public, because the more precise terms undermine credibility. The irony is that this translation often backfires. Where scientists intend to signal calibrated caution, audiences frequently hear a confession of ignorance. Organizational pressures compound the problem: owners, advertisers, and stockholders may push media organizations to emphasize economic promise and downplay uncertainty claims that could threaten business interests.

• 7. Philosophy and Art

  ›

  Pyrrho was the first philosopher in the Western tradition to make uncertainty a foundational commitment. His position gave rise to Pyrrhonism and Academic Skepticism, the earliest schools of philosophical skepticism in the Hellenistic world. Two concepts from ancient Greek philosophy mark the terrain: aporia, a state of genuine puzzlement or impasse, and acatalepsy, the idea that certain knowledge is unattainable.

  The tradition continues. William MacAskill, a philosopher at Oxford University, has developed the concept of moral uncertainty, defining it as uncertainty about how to act given the absence of certainty in any single moral theory, as well as the study of how people ought to act under those conditions. The question is not just what to do when the facts are unclear; it is what to do when the ethical framework itself is in dispute.

  Artists have inhabited this territory as directly as any philosopher. The indecision of Hamlet stands as a long-running example of uncertainty as dramatic engine. The contemporary artist Martin Creed has spoken openly about his difficulty deciding what artworks to make, turning the artist's own uncertainty into the subject of the work. Uncertainty, in both cases, is not a problem to be solved before the real work begins; it is the real work.