Historical Origin of Uncertainty

The uncertainty principle is often compressed into a slogan: one cannot know position and momentum at the same time. That slogan is too vague. Historically, uncertainty emerged from several linked but distinct developments: wave-packet localization, matrix noncommutativity, Heisenberg’s 1927 analysis of measurement and kinematics, and the later mathematical inequalities of Kennard and Robertson.

This page separates those threads. The formal derivations belong in General Uncertainty Relations and Position-Momentum Uncertainty. The historical point is how uncertainty became a central physical idea during the transition from old quantum theory to modern quantum mechanics.

Wave Packets and Localization

Wave mechanics made the localization problem visible. A plane wave has a sharp wave number and therefore a sharp momentum in the idealized sense:

$$p = \hbar k.$$

But a plane wave is spread through space. To describe a localized particle-like state, one must superpose many wave numbers:

$$\psi(x) = \frac{1}{\sqrt{2\pi}} \int_{-\infty}^{\infty} a(k)e^{ikx}\,dk.$$

A narrow spatial packet requires a broad distribution of $k$ values. Conversely, a sharply concentrated $a(k)$ produces a delocalized wave. This is already a Fourier-analysis tradeoff, before any detailed discussion of measurement apparatus.

For a suitably normalized Gaussian packet, the exact quantum result is

$$\Delta x\,\Delta p = \frac{\hbar}{2}.$$

More general normalized states satisfy

$$\Delta x\,\Delta p \ge \frac{\hbar}{2},$$

provided the variances exist and the position and momentum operators are applied on appropriate domains.

Historically, this wave-packet picture helped dissolve a naive particle-or-wave choice. A localized electron state is not a tiny classical corpuscle with a wave glued to it, and it is not a single infinite sinusoid. It is a state built from amplitudes whose widths in position and momentum representations are linked.

Heisenberg Microscope Thought Experiment

Heisenberg’s 1927 paper used physical thought experiments to argue that classical kinematic concepts could not be carried into quantum mechanics without qualification. The best-known example is the gamma-ray microscope.

To localize an electron optically, one must use short-wavelength radiation. A rough optical resolution scale is

$$\Delta x \sim \frac{\lambda}{\sin\theta},$$

where $\lambda$ is the wavelength and $\theta$ represents the angular aperture scale of the microscope. Shorter wavelength improves localization.

But the photon used to locate the electron carries momentum of order

$$p_\gamma \sim \frac{h}{\lambda}.$$

Scattering that photon from the electron transfers an uncertain recoil momentum. The angular acceptance of the microscope gives an order-of-magnitude momentum disturbance of the form

$$\Delta p \sim \frac{h\sin\theta}{\lambda}.$$

Multiplying the estimates gives a product of order

$$\Delta x\,\Delta p \sim h.$$

This is not the modern proof of the sharp inequality with constant $\hbar/2$. It is a historically important disturbance argument showing why attempts to assign a classical trajectory to an electron run into quantum limits. Heisenberg’s analysis mixed measurement disturbance, wave-optical resolution, and the new noncommutative kinematics. Later work separated these ideas more cleanly.

Noncommutativity

Matrix mechanics supplied the algebraic core. Position-like and momentum-like quantities do not commute:

$$[Q,P] = i\hbar I.$$

This relation is stronger than a statement about a microscope. It says that position and momentum are not represented by ordinary commuting variables. They cannot generally be diagonalized together, and the same quantum state cannot make both associated distributions arbitrarily sharp.

The commutator also links uncertainty to dynamics and representation. In wave mechanics, $Q$ acts by multiplication and $P$ acts by differentiation:

$$(Q\psi)(x) = x\psi(x), \qquad (P\psi)(x) = -i\hbar \frac{d\psi}{dx}.$$

The derivative in $P$ is another way to see why localizing $\psi(x)$ has momentum consequences. Rapid spatial variation requires a spread of Fourier components, and those components are momentum components.

The historical route is therefore two-sided:

• wave packets show a width tradeoff between position and momentum representations;
• matrix mechanics shows that the corresponding observables have a noncommuting algebra.

The modern theory joins these into one statement about states, operators, and probability distributions.

Kennard and Robertson-Style Inequalities

The familiar inequality

$$\Delta x\,\Delta p \ge \frac{\hbar}{2}$$

is more precise than Heisenberg’s microscope estimate. Kennard gave an early wave-mechanical proof for position and momentum in 1927. Robertson then gave a general operator inequality:

$$\Delta A\,\Delta B \ge \frac12 \left\lvert \langle[A,B]\rangle \right\rvert.$$

Here $\Delta A$ and $\Delta B$ are standard deviations of ideal outcome distributions in a specified state. They are not simply errors made by a measuring device. The expectation value is taken in that same state.

For position and momentum,

$$\langle[Q,P]\rangle = i\hbar,$$

so Robertson’s inequality gives the standard canonical bound.

This later mathematical form is the version most often used in modern quantum mechanics. It also clarifies the scope of the claim. The inequality assumes the relevant variances and operator products are well defined. It does not say that every measurement of $A$ mechanically disturbs $B$ by exactly the lower bound. It constrains the preparation statistics assigned by a state.

Measurement Disturbance Versus Intrinsic Uncertainty

The historical language of uncertainty often blends two questions:

• How much does a measurement of one quantity disturb a later measurement of another?
• How narrow can the probability distributions of two observables be in one preparation?

These are related but not identical.

Heisenberg’s microscope is a measurement-disturbance argument. It says that a procedure designed to locate an electron can disturb its momentum. The Robertson inequality is a preparation-uncertainty theorem. It says that a state cannot make two noncommuting observables simultaneously arbitrarily sharp when the commutator expectation enforces a nonzero lower bound.

Modern measurement theory has more refined error-disturbance and joint-measurement statements. Those are important, but they should not be retroactively collapsed into the 1927 argument. The historically responsible summary is:

• Heisenberg identified a deep limitation of classical kinematic description.
• Wave mechanics and Fourier analysis made the width tradeoff mathematically visible.
• Matrix mechanics made the noncommuting algebra explicit.
• Kennard and Robertson supplied the sharp standard-deviation inequalities.

What Uncertainty Changed Conceptually

Uncertainty was not merely a technical correction to measurement practice. It changed what a state could mean.

In classical mechanics, a point in phase space assigns exact position and momentum values at once. A probability distribution may express ignorance about that point, but the underlying kinematic picture still permits simultaneous sharp values.

Quantum mechanics does not use phase-space points as pure states. A quantum state assigns probability distributions for possible measurements, and noncommuting observables cannot generally be treated as if they were ordinary random variables on one shared sample space.

That is why uncertainty sits near the heart of the transition to modern quantum theory. It marks the failure of the old picture in which an electron in an atom always has a classical trajectory, with position and momentum both well defined at every instant.

What This Page Does Not Claim

This page does not claim that uncertainty is only about human knowledge. The standard inequalities are statements about the probability distributions associated with quantum states, not just about what an observer happens to know.

It also does not claim that uncertainty is only about clumsy instruments. Measurement disturbance exists, but the preparation inequality remains even for idealized measurements.

Finally, it does not justify the popular energy-time slogan that energy conservation can be violated for a short time. Energy-time uncertainty has a different status because time is not a universal position-like operator in ordinary nonrelativistic quantum mechanics. Use the dedicated Energy-Time Uncertainty page for that topic.

Common Mistakes

• Saying “uncertainty means measurement errors are large.” The modern inequality concerns standard deviations of outcome distributions in a state.
• Saying “a particle really has exact position and momentum, but quantum mechanics prevents us from knowing both.” That imports a classical phase-space picture the formalism does not assume.
• Treating Heisenberg’s microscope as the rigorous proof of $\Delta x\,\Delta p\ge\hbar/2$.
• Forgetting that wave-packet localization already implies a Fourier width tradeoff.
• Thinking the lower bound is always saturated. Most states have $\Delta x\,\Delta p>\hbar/2$.
• Confusing preparation uncertainty with measurement disturbance.
• Applying energy-time slogans as if they were identical to position-momentum uncertainty.

Cross-Links

• Matrix Mechanics and Operator Ideas
• Commutation Relations in Historical Context
• Equivalence of Matrix and Wave Mechanics
• Heisenberg’s Matrix Mechanics
• Wave Packets
• Interpreting the Wavefunction
• Commutators
• Canonical Commutation Relations
• Noncommuting Observables
• General Uncertainty Relations
• Position-Momentum Uncertainty
• Energy-Time Uncertainty
• Minimum-Uncertainty Wave Packets

References

• W. Heisenberg, “Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik,” Zeitschrift für Physik 43, 172-198, 1927, DOI: 10.1007/BF01397280.
• E. H. Kennard, “Zur Quantenmechanik einfacher Bewegungstypen,” Zeitschrift für Physik 44, 326-352, 1927.
• H. P. Robertson, “The Uncertainty Principle,” Physical Review 34, 163-164, 1929, DOI: 10.1103/PhysRev.34.163.
• W. Heisenberg, The Physical Principles of the Quantum Theory, University of Chicago Press, 1930.
• B. L. van der Waerden, ed., Sources of Quantum Mechanics, Dover, 1968.
• M. Jammer, The Conceptual Development of Quantum Mechanics, 2nd ed., American Institute of Physics, 1989.
• J. Mehra and H. Rechenberg, The Historical Development of Quantum Theory, Springer, 1982-2001.
• L. E. Ballentine, Quantum Mechanics: A Modern Development, 2nd ed., World Scientific, 2014.

Exercises

1. Explain why a sharply localized wave packet requires a spread of momenta.

Solution

A localized packet is a superposition of many plane-wave components. If the wave-number distribution $a(k)$ were concentrated at one value of $k$, the state would be close to a plane wave and spread over space. Narrowing $\psi(x)$ requires combining many $k$ values with phases that interfere constructively near one region and destructively away from it. Since $p=\hbar k$, a spread in $k$ is a spread in momentum.

2. Why is Heisenberg’s microscope not the same thing as the Robertson inequality?

Solution

The microscope is a physical disturbance argument: using short-wavelength light to localize an electron transfers uncertain recoil momentum. The Robertson inequality is a mathematical statement about standard deviations of two observables in one state:

$$\Delta A\,\Delta B \ge \frac12 \left\lvert \langle[A,B]\rangle \right\rvert.$$

The two are historically connected, but they answer different questions.

3. Use $[Q,P]=i\hbar I$ in the Robertson inequality to obtain the position-momentum bound.

Solution

Robertson gives

$$\Delta Q\,\Delta P \ge \frac12 \left\lvert \langle[Q,P]\rangle \right\rvert.$$

For a normalized state,

$$\langle[Q,P]\rangle = \langle i\hbar I\rangle = i\hbar.$$

Therefore

$$\Delta Q\,\Delta P \ge \frac12 \lvert i\hbar\rvert = \frac{\hbar}{2}.$$

4. What is wrong with saying “uncertainty is just a statement about imperfect measuring devices”?

Solution

Real devices do have finite resolution and can disturb systems, but the standard uncertainty inequality is deeper. It constrains the ideal probability distributions assigned by a quantum state. Even before a measurement is performed, the state cannot assign arbitrarily narrow position and momentum distributions when the canonical commutator fixes a nonzero lower bound.