Sharpen a particle's position and its momentum blurs — the two spreads have a floor no state can cross. It is not a limit of instruments; it is Fourier geometry: the position and momentum wavefunctions are transforms of one another, and a narrow bump in one forces a wide bump in the other. The engine measures both widths and their product live: a Gaussian packet saturates the floor at exactly ℏ/2, and nothing dips below it. The blue team builds and defends it; the red team tries to break it.
source W. Heisenberg, Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik, Z. Phys. 43, 172 (1927) — doi.org/10.1007/BF01397280. Rendered, not quoted.
For any state, define the spreads Δx and Δp as standard deviations of position and momentum. The law:
Δx · Δp ≥ ℏ/2
Two facts drive it. (1) The momentum wavefunction is the Fourier transform of the position one, so squeezing Δx by a factor k stretches Δp by k — the product is what stays fixed. (2) A Gaussian packet is the unique minimiser: it hits the floor with equality. The current state's live widths:
| quantity | value (ℏ=1) |
|---|
The floor is inherited, not invented. Because x and p are conjugate — the-wavefunction's two faces related by a transform — a bounded product is unavoidable; the Gaussian saturates it.
Push the same floor down to a bound state and it will not let the energy reach zero: the leftover ½ℏω of the-quantum-harmonic-oscillator is this inequality wearing an energy mask. Each sphere is the next one's premise.
The blue team's live check: re-derive Δp from an actual Fourier transform, confirm the Gaussian saturates ℏ/2, and sweep a set of states to confirm none dips below the floor. If red tampers, this badge is where it shows.
Feed the engine one quantum state ψ(x). Two shapes are offered:
| state | ψ(x) | Δx·Δp |
|---|---|---|
| Gaussian | (2πa²)−¼ e−x²/4a² | = ℏ/2 |
| box ground | √(2/L) sin(πx/L) | > ℏ/2 |
The squeeze slider sets Δx. Narrowing the packet in x is exactly what the engine turns into a wider Δp — you feed it a width, it returns the trade.
Gaussian: the minimum-uncertainty packet — it sits on the floor.
Left |ψ(x)|² · right |φ(p)|². Drag the slider: narrow the position bump and the momentum bump widens by the same factor. Every number is computed on the spot, never looked up.
What the machine proves: the momentum spread is Δp = ℏ/2Δx from Fourier conjugacy, so the product Δx·Δp = ℏ/2 for the Gaussian — independent of how hard you squeeze — and > ℏ/2 for every other state. The floor is never crossed.
The blue team's witness (left) re-checks this by real transform; the red team (right) tries to force a state below the floor.
And the floor is specific: it exists only between non-commuting observables (Robertson, 1929). Commuting pairs have no such bound — the principle is Fourier structure, not mysticism, and it does not forbid knowing either quantity alone to any precision.
"Uncertainty is the measuring device disturbing the particle." Cut. The floor is a property of the state's Fourier structure — present with a flawless apparatus. Disturbance is a separate relation.
"Heisenberg proved Δx·Δp ≥ ℏ/2." Corrected. His 1927 paper gave Δx·Δp ∼ ℏ, an order-of-magnitude via the γ-ray microscope. The exact ≥ ℏ/2 is Kennard (1927) and Weyl (1928) — what the engine computes.
"You cannot know position and momentum at all." Kept, corrected. Either alone can be made arbitrarily sharp; only their product is floored.
The red team's move: let a rogue state report Δx·Δp = ℏ/4 — half the floor — and try to slip it past as legal. The blue team's witness (window 7) is watching.
Halve the floor and the reported product falls below ℏ/2 — the witness recomputes, finds a state under the bound, and turns red. Nothing is faked; the violation is real and it is caught.