Visual walkthrough — Cyclic coordinates — corresponding conservation law

We never use a symbol before we draw it. Let us start with the most basic object: a coordinate.

---

Step 1 — What is a "generalized coordinate" $q$? (a number that locates the system)

WHAT. Picture a bead threaded on a wire. Its position along the wire is a single number $q$. If the bead slides, $q$ changes, and $\dot{q}$ (read "$q$-dot") measures how quickly.

WHY. Analytical mechanics refuses to talk about $x,y,z$ arrows. It talks about how many independent numbers you need. One number $q$ per degree of freedom. Everything downstream is built from $q$ and $\dot{q}$, so we anchor them to a picture first.

PICTURE. The red bead sits at value $q$; the red arrow is $\dot{q}$, pointing the way it is moving.

---

Step 2 — What is the Lagrangian $L$? (a single number scoring every position-and-motion)

WHAT. Think of $L$ as a landscape whose height depends on two knobs, $q$ and $\dot{q}$.

WHY. We need $L$ because the whole theory says nature "prefers" certain paths, and $L$ is the score being optimized. The two slopes of this landscape are what the physics reads off:

• $\frac{\partial L}{\partial q}$ — how much $L$ changes if I nudge position and keep speed fixed. This is a generalized force.
• $\frac{\partial L}{\partial \dot{q}}$ — how much $L$ changes if I nudge speed and keep position fixed. This is the generalized momentum.

PICTURE. Two red slope-arrows on the $L$-surface: one along the $q$-axis (the force slope), one along the $\dot{q}$-axis (the momentum slope).

---

Step 3 — The Euler–Lagrange equation (the law every coordinate obeys)

WHAT. For any coordinate $q$, these two slopes are locked together: the time-rate-of-change of the momentum-slope equals the force-slope.

WHY this tool and not another? We could try Newton's $\vec{F}=m\vec{a}$, but that needs arrows and forces we may not know (the wire's constraint force, for example). The Euler–Lagrange equation is the right tool because it needs only the number $L$ — no constraint forces, works in any coordinates. It is the engine of everything below.

PICTURE. A balance beam: on the left pan sits "$\frac{d}{dt}$ of the momentum slope," on the right pan sits "the force slope." The equation says the beam is level — they are always equal.

---

Step 4 — Name the momentum slope: define $p$ (this is Generalized Momentum)

WHAT. We give the momentum slope its own short name, $p$.

WHY the name "momentum"? Try the plainest Lagrangian, a mass on a line: $L=\frac{1}{2}m{\dot{x}}^2-V(x)$. Then $p=\frac{\partial }{\partial \dot{x}}\left(\frac{1}{2}m{\dot{x}}^2\right)=m\dot{x},$ the ordinary momentum from school. Term by term: differentiate $\frac{1}{2}m{\dot{x}}^2$ with respect to $\dot{x}$, the power drops down, giving $m\dot{x}$. So $p$ generalizes familiar momentum to any coordinate — angle, stretch, anything.

PICTURE. Zoom on the $\dot{q}$-axis slope from Step 2 and relabel it with a single red tag $p$.

---

Step 5 — Rewrite the master equation using $p$ (a rate = a force)

WHAT. In Step 3 the first term was "$\frac{d}{dt}$ of the momentum slope." That slope is $p$. So swap it in:

$\frac{dp}{dt}=\frac{\partial L}{\partial q}.$

WHY. This is pure renaming — no new physics, just cleaner reading. Now the sentence is short and physical:

PICTURE. The same balance beam, but the left pan now carries the single tag $\dot{p}$ and the right pan the tag "force $=\partial L/\partial q$."

---

Step 6 — The cyclic condition: kill the right-hand side

WHAT. We now assume the $L$-landscape is perfectly flat in the $q$-direction: sliding $q$ left or right does not change the score at all.

WHY. A flat direction is a symmetry: the system "does not care" where it sits along $q$. Flat slope means zero force slope, so the right side of Step 5 becomes $0$.

PICTURE. The $L$-surface from Step 2, now a level horizontal ridge along $q$ (flat), still sloped along $\dot{q}$. The red arrow along $q$ has collapsed to zero length.

---

Step 7 — The conservation law falls out

WHAT. Put $\partial L/\partial q=0$ into Step 5:

$\frac{dp}{dt}=0⟹\boxed{p=\text{constant}}$

WHY. A quantity whose rate of change is zero never moves — it is frozen for all time. No force slope $\Rightarrow$ nothing to change $p$ $\Rightarrow$ $p$ is conserved. This is the parent note's central result, now seen:

PICTURE. A perfectly flat horizontal line: the value $p$ plotted against time $t$ stays dead level. Contrast (grey) shows a non-cyclic $p$ sloping down.

This is the same story as Noether's Theorem in miniature: a continuous symmetry (the flat direction) hands you a conserved quantity ($p$).

---

Step 8 — Every case: what if the direction is NOT flat? (the projectile)

WHAT. We must not leave a scenario unshown. Take the projectile, $L=\frac{1}{2}m({\dot{x}}^2+{\dot{y}}^2)-mgy$.

• Horizontal $x$: $x$ is absent $\Rightarrow \partial L/\partial x=0\Rightarrow$ flat $\Rightarrow p_x=m\dot{x}$ is constant. The path may curve, but the horizontal oomph is frozen.
• Vertical $y$: $y$ appears (the $-mgy$ term) $\Rightarrow \partial L/\partial y=-mg\ne 0\Rightarrow$ not flat $\Rightarrow {\dot{p}}_y=-mg$. Here $p_y=m\dot{y}$ steadily drains as gravity pulls.

WHY show both. The theorem is an if: only the flat directions freeze. The tilted directions obey the full Step-5 balance with a real force on the right. Seeing them side by side stops the classic error "cyclic means $q$ is constant" — the coordinates $x,y$ both keep changing; it is $p_x$ (and only $p_x$) that holds still. This is exactly Central Force Motion's cousin: flat in $\varphi$ freezes angular momentum while $\varphi$ itself keeps winding.

PICTURE. The parabolic trajectory in red. Along the flat $x$-direction a fixed-length arrow $p_x$ (unchanging). Along the tilted $y$-direction a shrinking arrow $p_y$ under a downward force tag $-mg$.

---

The one-picture summary

Read the arrows top to bottom: a flat direction in $L$ (symmetry) $\to$ zero force slope $\to$ Euler–Lagrange collapses to $\dot{p}=0$ $\to$ $p$ frozen forever. A tilted direction (right branch) keeps its force and lets $p$ drift. This ties straight into Hamiltonian Mechanics, where a frozen $p$ lets you delete a whole coordinate, and into Conservation of Energy, the time-translation twin of this same idea.

Recall Feynman retelling — the whole walkthrough in plain words

Imagine the Lagrangian as a hilly landscape with two knobs: where you are and how fast you go. The physics only listens to the slopes of that landscape. One slope (along the speed knob) we named momentum, $p$. The master law says: the momentum changes at exactly the rate given by the other slope — the one along the position knob, which is the force. Now suppose the landscape is perfectly flat as you slide the position knob: the world looks the same no matter where you are along it. Flat means zero slope, means zero force, means the momentum has nothing pushing it — so it sits still forever. That flat direction is a cyclic coordinate, and its frozen momentum is the conservation law. And beware: it is the momentum that freezes, not your position — a thrown ball's sideways speed stays fixed while the ball itself sails clear across the sky.

Recall Quick self-test

Why does $\dot{q}$ appearing in $L$ NOT ruin cyclicity? ::: Cyclic depends only on the bare coordinate $q$ being absent; the velocity slope is where $p$ lives, so $\dot{q}$ must appear. In Step 5, what do the two sides physically mean? ::: Left $\dot{p}$ = rate of momentum change; right $\partial L/\partial q$ = generalized force. A ball is thrown; which is constant, $x$ or $p_x$? ::: $p_x=m\dot{x}$; the coordinate $x$ keeps increasing.

---

Connections

• Euler–Lagrange Equations — the balance beam of Steps 3 and 5.
• Generalized Momentum — the momentum slope $p$ we named in Step 4.
• Noether's Theorem — flat direction (symmetry) ⇒ conserved $p$, generalized.
• Central Force Motion — flat in $\varphi$ freezes angular momentum.
• Hamiltonian Mechanics — a frozen $p$ lets you drop a coordinate.
• Conservation of Energy — the time-translation analogue.