Heisenberg's Uncertainty Principle — Why Nature Has a Fundamental Fuzziness

The Heisenberg uncertainty principle says that a quantum particle cannot have a perfectly sharp position and a perfectly sharp momentum at the same time. The more tightly one is pinned down, the more the other spreads out, and the product of the two spreads never drops below ħ/2, about 5.3 × 10⁻³⁵ joule-seconds. Werner Heisenberg introduced the idea in 1927. It is not a story about clumsy instruments. It comes from the wave nature of matter, and it explains why atoms have a size at all.

This article covers the math, the wave logic behind it, what Heisenberg got right and wrong in his first explanation, the experiments that settled the old arguments with Einstein, and the concrete things the principle accounts for.

What the principle actually says

The modern, exact statement is an inequality:

Δx · Δp ≥ ħ/2

Here Δx is the spread (standard deviation) in position, Δp is the spread in momentum, and ħ is the reduced Planck constant. The NIST value of ħ is 1.054 571 817... × 10⁻³⁴ J·s, and it is exact by definition. Because ħ is so small, the effect is invisible for a thrown ball and dominant for an electron.

Two details are easy to miss. First, Δx and Δp are not errors in a single measurement. They describe how widely the results are distributed when you prepare many identical systems in the same quantum state and measure position in some runs and momentum in others. Second, the inequality is a lower bound. A state can have a bigger product, but never a smaller one.

The principle generalizes. In 1929 Howard Robertson showed that any two quantities A and B obey ΔA·ΔB ≥ ½|⟨[A,B]⟩|, where [A,B] is the commutator, a measure of how much the order of measurements matters. Position and momentum have a commutator equal to iħ, which gives the inequality above. Other pairs, such as spin components along different axes, have their own versions.

A short history: 1927 and the Copenhagen circle

Heisenberg published his paper in Zeitschrift für Physik in 1927, two years after he developed matrix mechanics, the first full formulation of quantum theory. The Nobel Foundation credits him with "the creation of quantum mechanics," and he received the physics prize for 1932. His 1927 paper argued, using a thought experiment about observing an electron with a gamma-ray microscope, that the imprecision in position and momentum multiplies to roughly Planck's constant.

Heisenberg's own formula was a rough estimate, with δp·δq of order h. The exact inequality with ħ/2 came from Earle Kennard, also in 1927, as the Stanford Encyclopedia of Philosophy records. Robertson's generalization followed in 1929. Niels Bohr, who was working with Heisenberg in Copenhagen at the time, objected to part of the original reasoning. His criticism appeared in an "addition in proof" to Heisenberg's paper, and it set up a disagreement about interpretation that lasted for decades. You can read more about the man himself in our biography of Werner Heisenberg and about his collaborator Niels Bohr.

Why the uncertainty exists: particles are waves

The uncertainty principle is not a statement about bad measurement. It follows from a deeper fact: quantum objects have wave properties.

In quantum mechanics, each particle is described by a wave function, a mathematical object whose squared magnitude gives the probability of finding the particle at each point. The wave function evolves in time according to the Schrödinger equation. (If the wave-particle idea is new, our explainers on the double-slit experiment and the Davisson-Germer experiment show how electrons produce interference patterns.)

Position and momentum are linked through the wave's shape. Momentum is tied to wavelength by p = h/λ, so a particle with a well-defined momentum has a well-defined wavelength. The difficulty is that a wave with one clean wavelength is a long, smooth sinusoid that stretches across all of space. That is a particle with no definite location. To make the wave function narrow, concentrated in one place, you have to add together many waves with different wavelengths. A narrow spike in space therefore contains a broad mixture of momenta.

The sound analogy

Music gives a familiar version of this trade-off. A pure, steady tone has an exact pitch, but it lasts a long time. A sharp click lasts a fraction of a millisecond, and it contains a smear of frequencies, which is why a click sounds like noise rather than a note. You cannot have a sound that is both instantaneous and a single pitch. Engineers call this the time-bandwidth trade-off, and it is a theorem about waves in general, known as the Fourier uncertainty relation.

Quantum mechanics applies exactly this theorem, with position playing the role of time and momentum playing the role of pitch. For any wave, the spreads in position and in wave number k obey Δx·Δk ≥ ½. Multiply by ħ, using p = ħk, and you get Δx·Δp ≥ ħ/2. Position and momentum are what mathematicians call Fourier conjugates, and the uncertainty relation is the same fact that makes a short radio pulse occupy a wide band of frequencies. A state that reaches the minimum, with Δx·Δp exactly ħ/2, is a Gaussian (bell-shaped) wave packet.

A second image: the shadow and the ripple

Imagine trying to say exactly where a ripple on a pond is. If you describe it as a tiny splash at one point, you have lost its wavelength, because a single splash contains ripples of every size. If you describe it as a perfect repeating wave, you have lost its location, because it fills the whole pond. Any description must trade one for the other. The principle says nature does the same, and not just our descriptions of it.

The microscope story, and what Heisenberg got partly wrong

In his 1927 paper, Heisenberg offered an intuitive picture now called the gamma-ray microscope. To locate an electron precisely, you must light it with short-wavelength radiation, since resolving power improves as wavelength shrinks. But a short-wavelength photon carries a lot of momentum, and when it scatters off the electron it gives it an uncontrollable kick. Sharper position knowledge costs a bigger disturbance of momentum.

This "measuring disturbs the system" account is vivid and partly right, and it is how the principle is usually taught. But it blurs two different claims, which the Stanford Encyclopedia of Philosophy separates carefully:

• Preparation uncertainty (Kennard and Robertson): For any quantum state, the spread of position results and the spread of momentum results cannot both be small. This is a property of the state itself. It holds even if no measurement ever takes place.
• Error-disturbance relations (Heisenberg's original idea): The more accurately you measure one quantity, the more you disturb the other. This is about the act of measuring.

The first claim is a rigorous theorem. The second turns out to be much trickier. In 2003, Masanao Ozawa showed that, with certain definitions of error and disturbance, their product need not be bounded below by ħ/2 in the way Heisenberg's formula suggested, and derived a revised relation. In 2012, a group led by Lee Rozema and Aephraim Steinberg reported an experiment using weak measurements on photons that violated the naive Heisenberg-style error-disturbance relationship and supported Ozawa's revised version. Another team using neutron spins reported similar results in the same year. Other physicists have argued about whether those experiments measured error and disturbance in the right way, and that debate continues.

The takeaway is subtle. Heisenberg's inequality for the spread of a prepared state stands unchallenged. What turned out to be shakier was the idea that the principle is simply about the disturbance caused by measuring. A particle really does lack a simultaneously sharp position and momentum before anyone looks.

Energy and time: a relative with different rules

You will often see a second relation, ΔE·Δt ≥ ħ/2. It is useful but trickier than the position-momentum one. Time is not an observable quantum property in the usual sense: there is no standard operator for it, so the derivation of the formula is different, and the meaning of Δt depends on the situation. The Stanford Encyclopedia notes that Bohr treated time as a classical variable and that the two relations are not on equal footing.

The practical meaning is still clear. A state that lasts only a short time cannot have a precisely defined energy. Excited atoms that decay quickly emit light with a broader range of frequencies than those that live longer, which sets the natural linewidth of a spectral line. Ultrashort laser pulses make the same point from the other side. A transform-limited Gaussian pulse has a time-bandwidth product of about 0.44, so a 10-femtosecond pulse (10 × 10⁻¹⁵ s) must contain a frequency spread of at least about 44 trillion hertz. At a wavelength near 800 nanometers, that is a bandwidth of nearly 100 nanometers, which is why femtosecond lasers need exotic broadband materials.

Worked numbers: when does the principle matter?

The inequality is exact for all objects. What changes is whether the consequence is large enough to notice. The table below confines each system to a region of the stated size and calculates the minimum spread in velocity from Δv ≥ ħ/(2mΔx).

System	Mass	Confined to	Minimum velocity spread
Electron in an atom	9.11 × 10⁻³¹ kg	10⁻¹⁰ m (0.1 nm)	about 5.8 × 10⁵ m/s
Proton in a nucleus	1.67 × 10⁻²⁷ kg	10⁻¹⁵ m (1 fm)	about 3 × 10⁷ m/s
One-gram pebble	10⁻³ kg	10⁻¹⁰ m	about 5 × 10⁻²² m/s
0.6 kg ball	0.6 kg	10⁻⁶ m (1 µm)	about 9 × 10⁻²⁹ m/s

Source: calculated from ħ and the CODATA electron mass as listed by NIST. The proton mass is the standard CODATA value.

For the electron, the numbers are enormous: a velocity spread of hundreds of kilometers per second just from being squeezed into an atom. For a pebble, the same confinement produces a velocity spread of about 10⁻²² meters per second, far below anything we could detect. This is why the principle never shows up in daily life. The rule applies to the pebble in exactly the same way. Its consequences are just too small to see.

What the principle explains

The uncertainty principle is not a nuisance. A surprising amount of the visible world is a direct consequence of it.

Why atoms do not collapse

Classically, an electron orbiting a nucleus is an accelerating charge, which should radiate away its energy and spiral into the nucleus in a tiny fraction of a second. Atoms do not collapse. The uncertainty principle supplies the reason. Squeezing an electron into a smaller space forces a larger spread in momentum, and so a larger kinetic energy. The attraction of the nucleus pulls the electron inward, while this quantum "pressure" pushes outward, and the balance fixes the atom's size.

You can see the effect with a back-of-envelope calculation. Take the electron's kinetic energy to be about ħ²/(2mr²) (using p ≈ ħ/r) and add the electrical attraction energy, −e²/(4πε₀r). Minimizing the sum over r gives a radius of about 5.3 × 10⁻¹¹ meters and an energy of −13.6 electron-volts. Those are the measured Bohr radius and the hydrogen ground-state energy, according to NIST's CODATA values. The estimate is rough and lands on the right answer partly by design, but it shows the essential idea: atomic size is set by the contest between electrical attraction and quantum confinement.

The same reasoning at the scale of a nucleus gives a minimum kinetic energy of several million electron-volts for a proton confined to a femtometer, which is the right scale for nuclear energies. The principle sets the energy scale of each level of matter. For the wider framework, see Quantum Mechanics Foundations Explained.

Zero-point energy and liquid helium

A quantum system cannot come to rest, because a particle with zero momentum spread would have to be completely delocalized. Even at absolute zero, every confined system keeps some motion, called zero-point energy. The effect is large for light atoms. Helium is the standout case: it is the only substance known to stay liquid at atmospheric pressure all the way to absolute zero. Its atoms attract each other so weakly, and their zero-point motion is so large, that they never settle into a crystal unless the pressure rises to roughly 25 atmospheres.

Quantum tunneling

If a particle is confined, its momentum is uncertain, and there is a small but real chance of finding it where classical physics says it lacks the energy to be. Particles can therefore pass through barriers they could not climb over. Tunneling lets protons in the Sun's core fuse despite their electrical repulsion, and it is the working principle of the scanning tunneling microscope. Our article on quantum tunneling goes through the experiments.

Vacuum fluctuations and the Casimir effect

The energy-time relation, together with the field description of quantum physics, implies that even empty space is not perfectly quiet. The vacuum has fluctuating fields, and the most direct visible consequence is the Casimir force. Hendrik Casimir predicted in 1948 that two uncharged metal plates placed very close together in a vacuum would attract each other, and precision experiments in the late 1990s, notably by Steve Lamoreaux, measured the force and found agreement with theory to within a few percent. The Lamb shift, a small difference in hydrogen's energy levels, is another consequence of vacuum fluctuations.

A word of caution. Popular accounts often describe "virtual particles" as tiny objects popping in and out of existence by borrowing energy. That is a loose picture of a calculation method, and it should not be taken as a literal description of the vacuum.

The uncertainty principle in technology

The principle is a design constraint in real instruments.

Gravitational-wave detectors. Detectors such as LIGO measure mirror positions to an extraordinary precision. The mirrors weigh 40 kilograms, and they are hit by laser light. A measurement of the mirror's position with light creates random momentum kicks from the photons, which in turn move the mirror. In a 2020 paper, the LIGO team described quantum correlations between the light and the mirrors that let the instrument reach a joint quantum uncertainty about 3 dB below what is called the standard quantum limit, using 200 kW laser beams. In that sense, the instrument works at the edge of the uncertainty principle. See our piece on the LIGO gravitational-wave detection.

Ultrafast lasers. As shown above, shorter pulses require broader bandwidth, which shapes how femtosecond lasers are built.

Electron microscopes. Faster electrons have shorter de Broglie wavelengths and can resolve finer detail, but they also carry more energy, which can damage delicate samples. The wave nature of the electron sets the resolution, and the trade-offs between resolution and sample damage are one of the main design problems.

Quantum computing. Qubits are quantum states, and reading them out changes them. Not every limitation in a quantum computer comes from the uncertainty principle, but the broader principle that quantum states cannot be copied or fully read without disturbance is central. See how quantum computers actually work.

Einstein, hidden variables, and the Bell experiments

Albert Einstein accepted that quantum mechanics made correct predictions. He doubted that it was the final word. His objection is often compressed into the phrase "God does not play dice," and his worry was that the randomness reflected missing information: definite properties that a deeper theory would reveal.

In 1935 Einstein, Boris Podolsky, and Nathan Rosen (EPR) sharpened the challenge. They considered two particles that interact and then separate, with correlated positions and momenta. Measuring the position of one lets you predict the position of the other without touching it. Measuring the momentum of the first lets you predict the momentum of the second. EPR argued that if the second particle is not disturbed by what happens to the first, both its position and momentum must have had definite values all along, which quantum mechanics does not describe. So, they concluded, quantum mechanics is incomplete.

For almost three decades, this looked like a matter of taste. In 1964, John Bell showed that it was not. He derived an inequality that any theory with local hidden variables, meaning definite pre-existing properties and no faster-than-light influences, must obey. Quantum mechanics predicts violations of it. Experiments by John Clauser and Alain Aspect in the 1970s and 1980s found violations, but each left loopholes. In 2015, a team in Delft led by Ronald Hanson reported a loophole-free test using entangled electron spins separated by 1.3 kilometers, with a measured value of S = 2.42 ± 0.20 against a local-realist limit of 2. Aspect, Clauser, and Anton Zeilinger shared the 2022 Nobel Prize in Physics for experiments with entangled photons establishing the violation of Bell inequalities. Our article on the Aspect Bell test tells that story in detail, and quantum entanglement explains what entanglement does and does not mean.

Be careful about what this proves. Bell's theorem rules out local hidden-variable theories. It does not rule out all hidden-variable theories. Pilot-wave theory is non-local and survives, and it reproduces the same uncertainty relations for any observer.

Interpretations: what does the principle mean?

The mathematical content of the principle, and its experimental consequences, are agreed on by everyone. What it says about reality is still argued over. Three well-known readings are:

Interpretation	What it says about uncertainty
Copenhagen (Bohr, Heisenberg)	Position and momentum are not both definable in the same context. Properties depend on the experimental arrangement.
Many-worlds (Everett)	The universal wave function always evolves smoothly. Apparent randomness reflects an observer's branch of that wave function.
Pilot-wave (de Broglie-Bohm)	Particles have definite positions guided by a wave. The uncertainty reflects unavoidable ignorance of initial conditions.

All three make the same predictions for experiments, which is why choosing among them is a philosophical matter for now. Bohr himself, as the Stanford Encyclopedia notes, treated the uncertainty relations less as a statement about disturbance than as a symbolic expression of complementarity: the idea that some descriptions of a system exclude each other.

Common misconceptions

Misconception	What is actually true
"Better instruments could beat the uncertainty."	The preparation uncertainty is a property of quantum states, not of instruments.
"It just means observing a particle disturbs it."	Measurement disturbance is real, but the spread in a prepared state exists with no measurement at all.
"It applies to everyday objects, so a car can be in two places."	The inequality applies, but the spreads are far too small to notice.
"The particle has an exact position and momentum we cannot know."	Local hidden-variable versions of this idea are ruled out by Bell tests.
"It only concerns position and momentum."	Any two non-commuting quantities obey a version of it, such as spin along different axes.
"It means quantum physics is imprecise."	Quantum theory makes some of the most precise predictions in science.

On that last point, the numbers are striking. In a measurement published in 2023, Xing Fan, Gerald Gabrielse, and colleagues reported the electron's magnetic moment as g/2 = 1.001 159 652 180 59, with an uncertainty of 0.13 parts per trillion, which tests the Standard Model's prediction to about one part in 10¹². A theory with built-in uncertainty can still be extraordinarily exact, because the uncertainty applies to individual outcomes and not to the statistical predictions.

The takeaway

The uncertainty principle is best understood as a statement about waves. If a particle is a wave, then a sharp location requires a mixture of many wavelengths, and a sharp wavelength requires a spread-out wave. Heisenberg's first explanation, that measuring disturbs the system, captured part of the picture. The firmer, more surprising claim is that a quantum state has no simultaneously sharp position and momentum to begin with. That one fact fixes the size of atoms, keeps helium liquid at the lowest temperatures, allows the Sun to burn through tunneling, and shapes the best measurements humans have built. It also reminds us why the everyday world looks so solid: the quantum blur is real, but at human scale it is smaller than anything we could ever detect.

FAQ

Q: What does Δx mean in the uncertainty principle? A: Δx is the standard deviation of position results when many identically prepared systems are measured. It is not the error in a single measurement. Any one measurement can be very precise, but across many runs, the spread in position results and the spread in momentum results satisfy Δx·Δp ≥ ħ/2.

Q: Is the uncertainty principle caused by the observer disturbing the particle? A: Only partly. Heisenberg's gamma-ray microscope picture describes measurement disturbance, but the main result, due to Kennard and Robertson, concerns the spread of prepared quantum states, and it holds with no measurement. Experiments in 2012 also showed that the simple error-disturbance version of the principle can be violated, while the preparation version stands.

Q: Does the uncertainty principle mean physics is just guesswork? A: No. Quantum theory predicts probabilities exactly, and the electron magnetic moment has been measured and calculated to roughly 12 digits of agreement. What the principle limits is the sharpness of certain pairs of properties in a single state, not the precision of the theory's statistical predictions.

Q: Why did Einstein reject it? A: Einstein accepted the experimental success of quantum mechanics but believed it was incomplete, and that a deeper theory would restore definite properties. Bell's theorem and the experiments that followed, including the loophole-free test in 2015, ruled out local hidden-variable theories of that kind. Non-local hidden-variable theories, such as pilot-wave theory, remain logically possible.

Q: How small must a system be for the principle to matter? A: It matters when ħ/(2Δx) is comparable to the momentum scales of the system. For an electron confined to an atom of 10⁻¹⁰ m, the minimum velocity spread is about 5.8 × 10⁵ m/s. For a one-gram pebble confined to the same region, it is about 5 × 10⁻²² m/s, far below anything measurable.

Q: What is the energy-time uncertainty relation? A: It says that a state lasting only a short time Δt cannot have a precisely defined energy, with ΔE of at least roughly ħ/(2Δt). It is less rigorous than the position-momentum relation, because time is not an observable in the usual sense. It explains natural spectral linewidths and why ultrashort laser pulses must be broadband.

Q: Does the uncertainty principle apply to anything other than position and momentum? A: Yes. Robertson's 1929 generalization says any two observables that do not commute obey an uncertainty relation. Spin components along different axes are an example, as are the amplitude and phase of light, which is what limits precision in gravitational-wave detectors.

Sources

• Stanford Encyclopedia of Philosophy: The Uncertainty Principle
• NIST CODATA: reduced Planck constant, Bohr radius, and electron mass
• Rozema et al., Violation of Heisenberg's measurement-disturbance relationship by weak measurements, Phys. Rev. Lett. 109, 100404 (2012)
• Hensen et al., Experimental loophole-free violation of a Bell inequality using entangled electron spins separated by 1.3 km, Nature 526, 682 (2015)
• Yu et al., Quantum correlations between the light and kilogram-mass mirrors of LIGO (2020)
• Heisenberg, W. (1927). Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik. Zeitschrift für Physik, 43, 172–198.
• Kennard, E. H. (1927). Zur Quantenmechanik einfacher Bewegungstypen. Zeitschrift für Physik, 44, 326–352.
• Robertson, H. P. (1929). The uncertainty principle. Physical Review, 34, 163–164.
• Bell, J. S. (1964). On the Einstein-Podolsky-Rosen paradox. Physics, 1(3), 195–200.
• Fan, X., Myers, T. G., Sukra, B. A. D., and Gabrielse, G. (2023). Measurement of the electron magnetic moment. Physical Review Letters, 130, 071801.

---

Read next:

• Quantum Mechanics Foundations Explained
• Quantum Tunneling: How Particles Pass Through Walls
• Werner Heisenberg: Uncertainty, Helgoland, and the Nazi Bomb