Visual walkthrough — Heisenberg uncertainty principle — Δx Δp ≥ ℏ - 2, ΔE Δt ≥ ℏ - 2

We build the argument in this shape:

---

Step 1 — What a single wave is, and why it hides the particle's position

WHAT. The simplest quantum wave is a plain ripple that repeats forever. We write it $\psi (x)=\cos (kx)$ (the real part is enough to picture). Here:

• $x$ ::: position along a line (metres).
• $\psi$ (the Greek letter "psi") ::: the wave's height at each $x$; think of it as "how much wave is here."
• $k$ ::: the wavenumber — how many radians of wobble fit into one metre. Big $k$ = tightly packed ripples; small $k$ = long lazy ripples.

The wavelength $\lambda$ (distance for one full up-down cycle) is tied to $k$ by $\lambda =\frac{2\pi }{k},$ where $2\pi$ is "one full turn" in radians — it's there because $\cos$ repeats every $2\pi$ of its input, and its input is $kx$, so $x$ must grow by $2\pi /k$ to repeat.

WHY this step. A quantum particle with a perfectly known momentum is exactly this endless single-$k$ wave. To understand the tradeoff we must first see the price of that perfect knowledge.

PICTURE. The endless wave has the same crest-height everywhere. The particle's chance of being found is $|\psi {|}^2$ — and that is equally big everywhere. So a perfectly known momentum means the position is completely unknown.

---

Step 2 — Two waves already start to localize

WHAT. Add two waves of slightly different wavenumbers, $k_1$ and $k_2$. Where their crests line up they reinforce (a tall bump); where a crest meets a trough they cancel (flat). This is interference, and it produces a slow "beat" pattern.

$\psi (x)=\cos (k_{1}x)+\cos (k_{2}x).$

Each term is one of the Step-1 waves; the plus sign means we let them overlap and add heights point by point.

WHY this step. It is the first hint of the whole mechanism: combining different $k$'s carves the flat wave into localized lumps. We are trading momentum-purity for a bit of position information.

PICTURE. The result has fat regions (constructive) separated by dead zones (destructive). The particle is now more likely to be in the fat regions — a first crude "here-ish."

---

Step 3 — Many waves make one clean bump (a wave packet)

WHAT. Keep adding waves whose wavenumbers cover a band of width $\Delta k$ centred on some $k_0$. If we weight them by a smooth (Gaussian) envelope, all the unwanted ripples outside one lump cancel, leaving a single localized hump: a wave packet of width $\Delta x$.

• $\Delta k$ ::: the spread of wavenumbers we used — how many different $k$'s went into the recipe.
• $\Delta x$ ::: the width of the resulting hump in space — the region where the particle actually is.
• $k_0$ ::: the centre of the band, which fixes the packet's average wavelength (and later, average momentum).

WHY this step. This is the object a real particle is (see Wave packets and Fourier analysis): not an endless ripple, not a point, but a hump. Now both "width in space" ($\Delta x$) and "width in wavenumber" ($\Delta k$) are meaningful numbers we can measure.

PICTURE. Look at the two panels. A packet using a wide band of $k$'s is narrow in space; a packet using a narrow band of $k$'s is wide in space. The tradeoff is already visible with the eyes.

---

Step 4 — The exact width law (Fourier)

WHAT. The "narrow-in-space ⇔ broad-in-$k$" trade is not vague — mathematics pins it to an exact floor: $\Delta x\Delta k\ge \frac{1}{2}.$

Reading it term by term:

• $\Delta x$ ::: RMS width of the hump in space (how uncertain the position is).
• $\Delta k$ ::: RMS width of the band of wavenumbers used.
• their product can be shrunk in one factor only by growing the other — you cannot make both tiny.
• $\frac{1}{2}$ ::: the hard floor. You cannot beat it. The smallest product ($=\frac{1}{2}$) happens only for a Gaussian-shaped packet.

WHY this step — and why this tool. We need a statement that is true for every shape of hump, not just a picture. The Fourier transform is the machine that reports "which $k$'s are inside a given shape," and its central theorem is exactly this width inequality. No other tool answers the question "how narrow can a bump be for a given band of ingredients?"

PICTURE. Plot the product $\Delta x\Delta k$ as you slide from a wide packet to a narrow one. It dips down, kisses the line $\frac{1}{2}$ at the Gaussian, and can never cross below.

---

Step 5 — de Broglie turns wavelength into momentum

WHAT. Physics enters through one bridge: a wave of wavenumber $k$ carries momentum $p=\hslash k,$ where $\hslash =h/2\pi \approx 1.055\times 10^{-34}\text{J⋅s}$ is Planck's constant per radian. So the band of wavenumbers $\Delta k$ becomes a band of momenta: $\Delta p=\hslash \Delta k.$

Term by term:

• $p$ ::: the particle's momentum (mass × velocity, roughly).
• $\hslash$ ::: the fixed conversion factor between "ripples per metre" and "momentum." Tiny — this is why the whole effect is invisible for big objects.
• $\Delta p$ ::: the spread of momenta, obtained by scaling the $k$-spread by $\hslash$.

WHY this step. Step 4 was about $k$, an abstract wave count. To make a physics statement we must convert $k$ into a measurable quantity. de Broglie wavelength gives the only bridge that does this: $p=\hslash k$.

PICTURE. The horizontal axis relabels itself. Every tick that read "$k$" now also reads "$p=\hslash k$." The band's width $\Delta k$ becomes $\Delta p=\hslash \Delta k$ — same picture, new units.

---

Step 6 — Substitute and read off Heisenberg

WHAT. Take the Fourier law and multiply both sides by $\hslash$: $\Delta x(\hslash \Delta k)\ge \hslash \cdot \frac{1}{2}.$ The left side is $\Delta x\Delta p$ (from Step 5), giving the result: $\boxed{\Delta x\Delta p\ge \frac{\hslash }{2}}$

Term by term:

• $\Delta x\Delta p$ ::: the product of position-spread and momentum-spread — the joint "fuzziness."
• $\ge$ ::: at least. You can equal $\hslash /2$ (Gaussian) but never go below.
• $\hslash /2$ ::: the universal floor of quantum fuzziness.

WHY this step. It's the payoff: a pure-math width law (Step 4) times a physics bridge (Step 5) is the uncertainty principle. Nothing about measuring disturbance ever appeared — the bound is a property of the wave packet itself.

PICTURE. Same product-curve as Step 4, but the vertical axis now reads $\Delta x\Delta p$ and the floor line sits at $\hslash /2$. The Gaussian point still just touches it.

---

Step 7 — Edge case A: the point particle (Δx → 0)

WHAT. Suppose we try to make the position perfectly sharp: $\Delta x\to 0$. The inequality forces $\Delta p\ge \frac{\hslash }{2\Delta x}\to \infty .$

WHY. A perfect spike needs a band of every wavenumber, so $\Delta k$ (and hence $\Delta p$) blows up. This is Step 1 in reverse: pin the position and the momentum becomes totally unknown.

PICTURE. As the hump is squeezed thinner, the momentum band fans out wider — like squeezing a balloon at one end and watching the other end swell.

---

Step 8 — Edge case B: the pure tone (Δp → 0) and the confined particle

WHAT. The opposite extreme, $\Delta p\to 0$, forces $\Delta x\to \infty$: this is the endless single wave of Step 1. And the real-world consequence — a particle trapped in a box of size $a$ (see Particle in a box) — cannot sit still: $\Delta x\sim a\Rightarrow \Delta p\gtrsim \frac{\hslash }{2a}\Rightarrow E_{\min }\sim \frac{(\Delta p{)}^2}{2m}\sim \frac{{\hslash }^2}{8m{a}^2}\ne 0.$

Term by term:

• $a$ ::: the box size, which caps $\Delta x$ (the particle can't spread wider than the box).
• $\Delta p\gtrsim \hslash /2a$ ::: because $\Delta x$ is capped, $\Delta p$ has a nonzero floor.
• $E_{\min }=(\Delta p{)}^2/2m$ ::: since the average momentum is zero, the spread itself becomes real kinetic energy — the zero-point energy.

WHY. Zero energy would require $\Delta p=0$, hence infinite $\Delta x$ — impossible inside a finite box. So confinement guarantees motion. This is why atoms don't collapse.

PICTURE. Two walls at distance $a$; the ground-state hump fits snugly between them, and its unavoidable curvature (bending of the wave) is the price of nonzero energy.

---

Step 9 — The time twin: ΔE Δt ≥ ℏ/2

WHAT. Replay every step with time $t$ instead of space $x$, and angular frequency $\omega$ instead of wavenumber $k$: $\Delta t\Delta \omega \ge \frac{1}{2},E=\hslash \omega \Rightarrow \Delta E\Delta t\ge \frac{\hslash }{2}.$

Term by term:

• $\Delta t$ ::: how long the state lasts before it changes appreciably (its lifetime) — not a clock error.
• $\Delta \omega$ ::: the spread of frequencies inside the time-signal.
• $\Delta E=\hslash \Delta \omega$ ::: converting frequency-spread to energy-spread with $E=\hslash \omega$.

WHY. A wave that is short in time must contain many frequencies, exactly as a wave short in space contains many wavenumbers. Same Fourier law, same bridge, different pair of variables. A short-lived state therefore has a fuzzy energy — its natural linewidth.

PICTURE. A long-lived ring (nearly one pure frequency, sharp energy) beside a quickly-damped burst (many frequencies, blurred energy).

---

The one-picture summary

Everything collapses into a single trade: space-width × wavenumber-width ≥ ½, upgraded by $p=\hslash k$ into position-width × momentum-width ≥ ℏ/2. Squeeze either, the other springs open; the Gaussian is the one shape that sits exactly on the floor.

Recall Feynman retelling of the whole walkthrough

Picture a wave that goes on forever like ocean swell — every part looks the same, so you have no idea where the particle is, but you know its wavelength (its speed) perfectly. Now try to make a single lump so you know where it is: you can only do it by mixing together many waves of different wavelengths. The tighter you want the lump, the more different wavelengths you must throw into the mix. Wavelength is just a code for momentum, so "many wavelengths" means "many possible momenta." That's the whole secret: sharp where forces blurry how-fast, and blurry where is the price of sharp how-fast. The math (Fourier) even tells you the exact floor — you can never beat ℏ/2, and only a perfectly bell-shaped lump reaches it. Do the same story in time instead of space and you learn that a thing which lives only briefly can't have a sharp energy — live fast, die fuzzy.

---

Active recall