Lectures On Classical Mechanics-John Baez PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 76

Lectures on

Classical Mechanics

by

John C. Baez

notes by Derek K. Wise

Department of Mathematics
University of California, Riverside

LaTeXed by Blair Smith

Department of Physics and Astronomy


Louisiana State University

2005
i

c
2005 John C. Baez & Derek K. Wise
ii
iii

Preface
These are notes for a mathematics graduate course on classical mechanics at U.C.
Riverside. I’ve taught this course three times recently. Twice I focused on the Hamiltonian
approach. In 2005 I started with the Lagrangian approach, with a heavy emphasis on
action principles, and derived the Hamiltonian approach from that. This approach seems
more coherent.
Derek Wise took beautiful handwritten notes on the 2005 course, which can be found
on my website:

http://math.ucr.edu/home/baez/classical/

Later, Blair Smith from Louisiana State University miraculously appeared and volun-
teered to turn the notes into LATEXẆhile not yet the book I’d eventually like to write,
the result may already be helpful for people interested in the mathematics of classical
mechanics.
The chapters in this LATEX version are in the same order as the weekly lectures, but
I’ve merged weeks together, and sometimes split them over chapter, to obtain a more
textbook feel to these notes. For reference, the weekly lectures are outlined here.
Week 1: (Mar. 28, 30, Apr. 1)—The Lagrangian approach to classical mechanics:
deriving F = ma from the requirement that the particle’s path be a critical point of
the action. The prehistory of the Lagrangian approach: D’Alembert’s “principle of least
energy” in statics, Fermat’s “principle of least time” in optics, and how D’Alembert
generalized his principle from statics to dynamics using the concept of “inertia force”.
Week 2: (Apr. 4, 6, 8)—Deriving the Euler-Lagrange equations for a particle on an
arbitrary manifold. Generalized momentum and force. Noether’s theorem on conserved
quantities coming from symmetries. Examples of conserved quantities: energy, momen-
tum and angular momentum.
Week 3 (Apr. 11, 13, 15)—Example problems: (1) The Atwood machine. (2) A
frictionless mass on a table attached to a string threaded through a hole in the table, with
a mass hanging on the string. (3) A special-relativistic free particle: two Lagrangians, one
with reparametrization invariance as a gauge symmetry. (4) A special-relativistic charged
particle in an electromagnetic field.
Week 4 (Apr. 18, 20, 22)—More example problems: (4) A special-relativistic charged
particle in an electromagnetic field in special relativity, continued. (5) A general-relativistic
free particle.
Week 5 (Apr. 25, 27, 29)—How Jacobi unified Fermat’s principle of least time and
Lagrange’s principle of least action by seeing the classical mechanics of a particle in a
potential as a special case of optics with a position-dependent index of refraction. The
ubiquity of geodesic motion. Kaluza-Klein theory. From Lagrangians to Hamiltonians.
iv

Week 6 (May 2, 4, 6)—From Lagrangians to Hamiltonians, continued. Regular and


strongly regular Lagrangians. The cotangent bundle as phase space. Hamilton’s equa-
tions. Getting Hamilton’s equations directly from a least action principle.
Week 7 (May 9, 11, 13)—Waves versus particles: the Hamilton-Jacobi equation.
Hamilton’s principal function and extended phase space. How the Hamilton-Jacobi equa-
tion foreshadows quantum mechanics.
Week 8 (May 16, 18, 20)—Towards symplectic geometry. The canonical 1-form and
the symplectic 2-form on the cotangent bundle. Hamilton’s equations on a symplectic
manifold. Darboux’s theorem.
Week 9 (May 23, 25, 27)—Poisson brackets. The Schrödinger picture versus the
Heisenberg picture in classical mechanics. The Hamiltonian version of Noether’s theorem.
Poisson algebras and Poisson manifolds. A Poisson manifold that is not symplectic.
Liouville’s theorem. Weil’s formula.
Week 10 (June 1, 3, 5)—A taste of geometric quantization. Kähler manifolds.
If you find errors in these notes, please email me! I thank Curtis Vinson for catching
lots of errors.
Contents

1 From Newton’s Laws to Langrange’s Equations 2


1.1 Lagrangian and Newtonian Approaches . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Lagrangian versus Hamiltonian Approaches . . . . . . . . . . . . . 6
1.2 Prehistory of the Lagrangian Approach . . . . . . . . . . . . . . . . . . . . 6
1.2.1 The Principle of Minimum Energy . . . . . . . . . . . . . . . . . . 7
1.2.2 D’Alembert’s Principle and Lagrange’s Equations . . . . . . . . . . 8
1.2.3 The Principle of Least Time . . . . . . . . . . . . . . . . . . . . . . 10
1.2.4 How D’Alembert and Others Got to the Truth . . . . . . . . . . . . 12

2 Equations of Motion 14
2.1 The Euler-Lagrange Equations . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.1.1 Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1.2 Lagrangian Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Interpretation of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3 Lagrangians and Noether’s Theorem 20


3.1 Time Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1.1 Canonical and Generalized Coordinates . . . . . . . . . . . . . . . . 21
3.2 Symmetry and Noether’s Theorem . . . . . . . . . . . . . . . . . . . . . . 22
3.2.1 Noether’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3 Conserved Quantities from Symmetries . . . . . . . . . . . . . . . . . . . . 24
3.3.1 Time Translation Symmetry . . . . . . . . . . . . . . . . . . . . . . 25
3.3.2 Space Translation Symmetry . . . . . . . . . . . . . . . . . . . . . . 25
3.3.3 Rotational Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.4 Example Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4.1 The Atwood Machine . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4.2 Disk Pulled by Falling Mass . . . . . . . . . . . . . . . . . . . . . . 29
3.4.3 Free Particle in Special Relativity . . . . . . . . . . . . . . . . . . . 31
3.5 Electrodynamics and Relativistic Lagrangians . . . . . . . . . . . . . . . . 35
3.5.1 Gauge Symmetry and Relativistic Hamiltonian . . . . . . . . . . . . 35
3.5.2 Relativistic Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . 36

v
CONTENTS 1

3.6 Relativistic Particle in an Electromagnetic Field . . . . . . . . . . . . . . . 37


3.7 Alternative Lagrangians . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.7.1 Lagrangian for a String . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.7.2 Alternate Lagrangian for Relativistic Electrodynamics . . . . . . . . 41
3.8 The General Relativistic Particle . . . . . . . . . . . . . . . . . . . . . . . 43
3.8.1 Free Particle Lagrangian in GR . . . . . . . . . . . . . . . . . . . . 44
3.8.2 Charged particle in EM Field in GR . . . . . . . . . . . . . . . . . 45
3.9 The Principle of Least Action and Geodesics . . . . . . . . . . . . . . . . . 46
3.9.1 Jacobi and Least Time vs Least Action . . . . . . . . . . . . . . . . 46
3.9.2 The Ubiquity of Geodesic Motion . . . . . . . . . . . . . . . . . . . 48

4 From Lagrangians to Hamiltonians 51


4.1 The Hamiltonian Approach . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Regular and Strongly Regular Lagrangians . . . . . . . . . . . . . . . . . . 54
4.2.1 Example: A Particle in a Riemannian Manifold with Potential V (q) 54
4.2.2 Example: General Relativistic Particle in an E-M Potential . . . . . 55
4.2.3 Example: Free General Relativistic Particle with Reparameteriza-
tion Invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.2.4 Example: A Regular but not Strongly Regular Lagrangian . . . . . 55
4.3 Hamilton’s Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3.1 Hamilton and Euler-Lagrange . . . . . . . . . . . . . . . . . . . . . 57
4.3.2 Hamilton’s Equations from the Principle of Least Action . . . . . . 59
4.4 Waves versus Particles—The Hamilton-Jacobi Equations . . . . . . . . . . 61
4.4.1 Wave Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4.2 The Hamilton-Jacobi Equations . . . . . . . . . . . . . . . . . . . . 63
Chapter 1

From Newton’s Laws to Langrange’s


Equations

(Week 1, March 28, 30, April 1.)

Classical mechanics is a very peculiar branch of physics. It used to be considered the


sum total of our theoretical knowledge of the physical universe (Laplace’s daemon, the
Newtonian clockwork), but now it is known as an idealization, a toy model if you will.
The astounding thing is that probably all professional applied physicists still use classical
mechanics. So it is still an indispensable part of any physicist’s or engineer’s education.
It is so useful because the more accurate theories that we know of (general relativity
and quantum mechanics) make corrections to classical mechanics generally only in extreme
situations (black holes, neutron stars, atomic structure, superconductivity, and so forth).
Given that GR and QM are much harder theories to use and apply it is no wonder that
scientists will revert to classical mechanics whenever possible.
So, what is classical mechanics?

1.1 Lagrangian and Newtonian Approaches


We begin by comparing the Newtonian approach to mechanics to the subtler approach of
Lagrangian mechanics. Recall Newton’s law:
F = ma (1.1)
wherein we consider a particle moving in n . Its position, say q, depends on time t ∈ ,
so it defines a function,
q : −→ n .
From this function we can define velocity,
n
v = q̇ : −→

2
1.1 Lagrangian and Newtonian Approaches 3

dq
where q̇ = dt
, and also acceleration,
n
a = q̈ : −→ .
n
Now let m > 0 be the mass of the particle, and let F be a vector field on called the
force. Newton claimed that the particle satisfies F = ma. That is:

m a(t) = F (q(t)) . (1.2)

This is a 2nd-order differential equation for q : → n which will have a unique solution
given some q(t0 ) and q̇(t0 ), provided the vector field F is ‘nice’ — by which we technically
mean smooth and bounded (i.e., |F (x)| < B for some B > 0, for all x ∈ n ).
We can then define a quantity called kinetic energy:
1
K(t) := m v(t) · v(t) (1.3)
2
This quantity is interesting because
d
K(t) = m v(t) · a(t)
dt
= F (q(t)) · v(t)
So, kinetic energy goes up when you push an object in the direction of its velocity, and
goes down when you push it in the opposite direction. Moreover,
Z t1
K(t1 ) − K(t0 ) = F (q(t)) · v(t) dt
t0
Z t1
= F (q(t)) · q̇(t) dt
t0

So, the change of kinetic energy is equal to the work done by the force, that is, the integral
of F along the curve q : [t0 , t1 ] → n . This implies (by Stokes’s theorem relating line
integrals to surface integrals of the curl) that the change in kinetic energy K(t1 ) − K(t0 )
is independent of the curve going from q(t0 ) = a to q(t0 ) = b iff

∇×F = 0.

This in turn is true iff


F = −∇V (1.4)
for some function V : n → . This function is then unique up to an additive constant; we
call it the potential. A force with this property is called conservative. Why? Because
in this case we can define the total energy of the particle by

E(t) := K(t) + V (q(t)) (1.5)


4 From Newton’s Laws to Langrange’s Equations

where V (t) := V (q(t)) is called the potential energy of the particle, and then we can
show that E is conserved: that is, constant as a function of time. To see this, note that
F = ma implies
d
[K(t) + V (q(t))] = F (q(t)) · v(t) + ∇V (q(t)) · v(t)
dt
= 0, (because F = −∇V ).

Conservative forces let us apply a whole bunch of cool techniques. In the Lagrangian
approach we define a quantity

L := K(t) − V (q(t)) (1.6)


n
called the Lagrangian, and for any curve q : [t0 , t1 ] → with q(t0 ) = a, q(t1 ) = b, we
define the action to be Z t1
S(q) := L(t) dt (1.7)
t0

From here one can go in two directions. One is to claim that nature causes particles
to follow paths of least action, and derive Newton’s equations from that principle. The
other is to start with Newton’s principles and find out what conditions, if any, on S(q)
follow from this. We will use the shortcut of hindsight, bypass the philosophy, and simply
use the mathematics of variational calculus to show that particles follow paths that are
‘critical points’ of the action S(q) if and only if Newton’s law F = ma holds. To do this,

qs + sδq
n

PSfrag replacements

t0 t1

Figure 1.1: A particle can sniff out the path of least action.

let us look for curves (like the solid line in Fig. 1.1) that are critical points of S, namely:
d
S(qs )|s=0 = 0 (1.8)
ds
1.1 Lagrangian and Newtonian Approaches 5

where
qs = q + sδq
n
for all δq : [t0 , t1 ] → with,
δq(t0 ) = δq(t1 ) = 0.
To show that
d
F = ma ⇔ S(qs )|t=0 = 0 for all δq : [t0 , t1 ] → n with δq(t0 ) = δq(t1 ) = 0 (1.9)
ds
we start by using integration by parts on the definition of the action, and first noting that
dqs /ds = δq(t) is the variation in the path:
Z t1
d d 1

S(qs ) = mq̇s (t) · q̇s (t) − V (qs (t)) dt
ds s=0 ds t0 2
Z t1   s=0
d 1
= mq̇s (t) · q̇s (t) − V (qs (t)) dt
t0 ds 2 s=0
Z t1  
d d
= mq̇s · q̇s (t) − ∇V (qs (t)) · qs (t) dt
t0 ds ds s=0
Note that
d d d d d
q̇s (t) = qs (t) = qs (t)
ds ds dt dt ds
and when we set s = 0 this quantity becomes just:
d
δq(t)
dt
So,
Z t1  
d d

S(qs ) = mq̇ · δq(t) − ∇V (q(t)) · δq(t) dt
ds s=0 dt
t0
Then we can integrate by parts, noting the boundary terms vanish because δq = 0 at t1
and t0 :
Z t1
d
S(qs )|s=0 = [−mq̈(t) − ∇V (q(t))] · δq(t)dt
ds t0
It follows that variation in the action is zero for all variations δq iff the term in brackets
is identically zero, that is,
−mq̈(t) − ∇V (q(t)) = 0
So, the path q is a critical point of the action S iff
F = ma.
The above result applies only for conservative forces, i.e., forces that can be written
as the minus the gradient of some potential. However, this seems to be true of the most
fundamental forces that we know of in our universe. It is a simplifying assumption that
has withstood the test of time and experiment.
6 From Newton’s Laws to Langrange’s Equations

1.1.1 Lagrangian versus Hamiltonian Approaches


I am not sure where to mention this, but before launching into the history of the La-
grangian approach may be as good a time as any. In later chapters we will describe
another approach to classical mechanics: the Hamiltonian approach. Why do we need
two approaches, Lagrangian and Hamiltonian?
They both have their own advantages. In the simplest terms, the Hamiltonian ap-
proach focuses on position and momentum, while the Lagrangian approach focuses on
position and velocity. The Hamiltonian approach focuses on energy, which is a function
of position and momentum — indeed, ‘Hamiltonian’ is just a fancy word for energy. The
Lagrangian approach focuses on the Lagrangian, which is a function of position and veloc-
ity. Our first task in understanding Lagrangian mechanics is to get a gut feeling for what
the Lagrangian means. The key is to understand the integral of the Lagrangian over time
– the ‘action’, S. We shall see that this describes the ‘total amount that happened’ from
one moment to another as a particle traces out a path. And, peeking ahead to quantum
mechanics, the quantity exp(iS/~), where ~ is Planck’s constant, will describe the ‘change
in phase’ of a quantum system as it traces out this path.
In short, while the Lagrangian approach takes a while to get used to, it provides
invaluable insights into classical mechanics and its relation to quantum mechanics. We
shall see this in more detail soon.

1.2 Prehistory of the Lagrangian Approach


We’ve seen that a particle going from point a at time t0 to a point b at time t1 follows a
path that is a critical point of the action,
Z t1
S= K − V dt
t0

so that slight changes in its path do not change the action (to first order). Often, though
not always, the action is minimized, so this is called the Principle of Least Action.
Suppose we did not have the hindsight afforded by the Newtonian picture. R Then we
might ask, “Why does nature like to minimize the action? And why this action K−V dt?
Why not some other action?”
‘Why’ questions are always tough. Indeed, some people say that scientists should
never ask ‘why’. This seems too extreme: a more reasonable attitude is that we should
only ask a ‘why’ question if we expect to learn something scientifically interesting in our
attempt to answer it.
There are certainly some interseting things to learn from the question “why is action
minimized?” First, note that total energy is conserved, so energy can slosh back and
forth between kinetic and potential forms. The Lagrangian L = K − V is big when most
of the energy is in kinetic form, and small when most of the energy is in potential form.
1.2 Prehistory of the Lagrangian Approach 7

Kinetic energy measures how much is ‘happening’ — how much our system is moving
around. Potential energy measures how much could happen, but isn’t yet — that’s what
the word ‘potential’ means. (Imagine a big rock sitting on top of a cliff, with the potential
to fall down.) So, the Lagrangian measures something we could vaguely refer to as the
‘activity’ or ‘liveliness’ of a system: the higher the kinetic energy the more lively the
system, the higher the potential energy the less lively. So, we’re being told that nature
likes to minimize the total of ‘liveliness’ over time: that is, the total action.
In other words, nature is as lazy as possible!
For example, consider the path of a thrown rock in the Earth’s gravitational field, as
in Fig. 1.2. The rock traces out a parabola, and we can think of it as doing this in order

K − V small here. . . good!


Spend as much time as possible here

PSfrag replacements

K − V big. . . bad!
Get this over with quick!

Figure 1.2: A particle’s “lazy” motion, minimizes the action.

to minimize its action. On the one hand, it wants to spend a lot much time near the top
of its trajectory, since this is where the kinetic energy is least and the potential energy
is greatest. On the other hand, if it spends too much time near the top of its trajectory,
it will need to really rush to get up there and get back down, and this will take a lot of
action. The perfect compromise is a parabolic path!
Here we are anthropomorphizing the rock by saying that it ‘wants’ to minimize its
action. This is okay if we don’t take it too seriously. Indeed, one of the virtues of the
Principle of Least Action is that it lets us put ourselves in the position of some physical
system and imagine what we would do to minimize the action.
There is another way to make progress on understanding ‘why’ action is minimized:
history. Historically there were two principles that were fairly easy to deduce from ob-
servations of nature: (i) the principle of minimum energy used in statics, and (ii) the
principle of least time, used in optics. By putting these together, we can guess the prin-
ciple of least action. So, let us recall these earlier minimum principles.

1.2.1 The Principle of Minimum Energy


Before physicists really got going in studying dynamical systems they used to study statics.
Statics is the study of objects at rest, or in equilibrium. Archimedes studied the laws of
8 From Newton’s Laws to Langrange’s Equations

PSfrag replacements L1 L2
m1 m2

Figure 1.3: A principle of energy minimization determines a lever’s balance.

a see-saw or lever (Fig. 1.3), and he found that this would be in equilibrium if

m1 L 1 = m 2 L 2 .

Later D’Alembert understood this using his “principle of virtual work”. He considered
moving the lever slightly, i.e., infinitesimally, He claimed that in equilibrium the infinites-
PSfrag replacements dθ
dq1
dq2

Figure 1.4: A principle of energy minimization determines a lever’s balance.

imal work done by this motion is zero! He also claimed that the work done on the ith
body is,
dWi = Fi dqi
and gravity pulls down with a force mi g so,
dWi = (0, 0, −mg) · (0, 0, −L1 dθ)
= m1 gL1 dθ
and similarly,
dW2 = −m2 gL2 dθ
Now D’Alembert’s principle says that equilibrium occurs when the “virtual work” dW =
dW1 +dW2 vanishes for all dθ (that is, for all possible infinitesimal motions). This happens
when
m1 L 1 − m 2 L 2 = 0
which is just as Archimedes wrote.

1.2.2 D’Alembert’s Principle and Lagrange’s Equations


Let’s go over the above analysis in more detail. I’ll try to make it clear what we mean by
virtual work.
1.2 Prehistory of the Lagrangian Approach 9

The forces and constraints on a system may be time dependent. So equal small
infinitesimal displacements of the system might result in the forces Fi acting on the
system doing different amounts of work at different times. To displace a system by δr i for
each position coordinate, and yet remain consistent with all the constraints and forces at a
given instant of time t, without any time interval passing is called a virtual displacement.
It’s called ‘virtual’ because it cannot be realized: any actual displacement would occur
over a finite time interval and possibly during which the forces and constraints might
change. Now call the work done by this hypothetical virtual displacement, F i · δri , the
virtual
P work. Consider a system in the special state of being in equilibrium, i.e.,. when
Fi = 0. Then because by definition the virtual displacements do not change the forces,
we must deduce that the virtual work vanishes for a system in equilibrium,
X
Fi · δri = 0, (when in equilibrium) (1.10)
i

Note that in the above example we have two particles in 3 subject to a constraint
(they are pinned to the lever arm). However, a number n of particles in 3 can be
treated as a single quasi-particle in 3n , and if there are constraints it can move in some
submanifold of 3n . So ultimately we need to study a particle on an arbitrary manifold.
But, we’ll postpone such sophistication for a while.
For a particle in n , D’Alembert’s principle simply says,

q(t) = q0 satisfies F = ma, (it’s in equilibrium)


m
n
dW = F · dq vanishes for all dq ∈ , (virtual work is zero for δq → 0)
m
F = 0, (no force on it!)
If the force is conservative (F = −∇V ) then this is also equivalent to,
∇V (q0 ) = 0
that is, we have equilibrium at a critical point of the potential. The equilibrium will be
stable if q0 is a local minimum of the potential V .
We can summarize all the above by proclaiming that we have a “principle of least en-
ergy” governing stable equilibria. We also have an analogy between statics and dynamics,
Statics Dynamics
equilibrium, a = 0 F = ma
Z t1
potential, V action, S = K − V dt
t0
critical points of V critical points of S
10 From Newton’s Laws to Langrange’s Equations

unstable equilibrium

unstable equilibrium
V

PSfrag replacements stable equilibrium

Figure 1.5: A principle of energy minimization determines a lever’s balance.

 
d ∂T ∂T
− = Qi . (1.11)
dt ∂ q̇i ∂qi

 
d ∂L ∂L
− = 0. (1.12)
dt ∂ q̇i ∂qi

D’Alembert’s principle is an expression for Newton’s second law under conditions


where the virtual work done by the forces of constraint is zero.

1.2.3 The Principle of Least Time


Time now to look at the second piece of history surrounding the principles of Lagrangian
mechanics. As well as hints from statics, there were also hints from the behavior of light,
hints again that nature likes to minimize effort. In a vacuum light moves in straight lines,
which in Euclidean space is the minimum distance. But more interesting than straight
lines are piecewise straight paths and curves. Consider reflection of light from a mirror,
What path does the light take? The empirical answer was known since antiquity, it
chooses B such that θ1 = θ2 , so the angle of incidence equals the angle of reflection. But
this is precisely the path that minimizes the distance of the trajectory subject to the
condition that it must hit the mirror (at least at one point). In fact light traveling from
A to B takes both the straight paths ABC and AC. Why is ABC the shortest path hitting
1.2 Prehistory of the Lagrangian Approach 11

A C

θ1 θ2
PSfrag replacements
B θ2

C0

the mirror? This follows from some basic Euclidean geometry:


B minimizes AB + BC ⇔ B minimizes AB + BC 0
⇔ A, B, C 0 lie on a line
⇔ θ1 = θ2
Note the introduction of the fictitious image C0 “behind” the mirror, this is a trick often
used in solving electrostatic problems (a conducting surface can be replaced by fictitious
mirror image charges to satisfy the boundary conditions), it is also used in geophysics
when one has a geological fault, and in hydrodynamics when there is a boundary between
two media (using mirror image sources and sinks).
The big clue leading to D’Alembert’s principle however came from refraction of light.
Snell (and predecessors) noted that each medium has some number n associated with it,

θ1

medium 1

PSfrag replacements medium 2

θ2

called the index of refraction, such that,


n1 sin θ1 = n2 sin θ2
12 From Newton’s Laws to Langrange’s Equations

(having normalized n so that for a vacuum n = 1). Someone guessed the explanation,
realizing that if the speed of light in a medium is proportional to 1/n, then light will
satisfy Snell’s law if the light minimizes the time it takes to get from A to C. In the case
of refraction it is the time that is important, not just the path distance. But the same
is true for the law of reflection, since in that case the path of minimum length gives the
same results as the path of minimum time.
So, not only is light the fastest thing around, it’s also always taking the quickest path
from here to there!

1.2.4 How D’Alembert and Others Got to the Truth


Sometimes laws of physics are just guessed using a bit of intuition and a gut feeling that
nature must be beautiful or elegantly simple (though occasionally awesomely complex in
beauty). One way to make good guesses is to generalize.
D’Alembert’s principle of virtual work for statics says that equilibrium occurs when
n
F (q0 ) · δq = 0, ∀δq ∈
D’Alembert generalized this to dynamics by inventing what he called the “inertia force”=−m a,
and he postulated that in dynamics equilibrium occurs when the total force = F +inertia
force, vanishes. Or symbolically when,

F (q(t)) − ma(t) · δq(t) = 0 (1.13)
We then take a variational path parameterized by s,
qs (t) = q(t) + s δq(t)
where
δq(t0 ) = δq(t1 ) = 0
and with these paths, for any function f on the space of paths we can define the varia-
tional derivative,
d
δf := f (qs ) (1.14)
ds s=0
Then D’Alembert’s principle of virtual work implies
Z t1  
F (q) − mq̈ · δq dt = 0
t0

for all δq, so if F = −∇V , we get


Z t1  
0= −∇V (q) − mq̈ · δq dt
t0
Z t1  
= −∇V (q) · δq + mq̇δ q̇ dt
t0
1.2 Prehistory of the Lagrangian Approach 13

using

d dV dqs (t)

V (qs (t)) =
ds s=0 dq ds s=0

and

dq̇s2 (t)
δ(q̇ 2 ) =
ds
then we have
Z t1 
dV dqs m dq̇s (t) 
0= − + dt
t0 dq dt 2 ds
Z
d t1  d  m 
= − V qs (t) + q̇s (t) dt
ds t0 dt 2 s=0
 Z t1   
d  m 2
=δ − V qs (t) + q̇s (t) dt
t0 dt 2

therefore Z t1 

δ −V (q) + K dt = 0
t0

so the path taken by the particle is a critical point of the action,


Z
S(q) = (K − V ) dt (1.15)

We’ve described how D’Alembert might have arrived at the principle of least action
by generalizing previously known energy minimization and least time principles. Still,
there’s something unsatisfying about the treatment so far. We do not really understand
why one must introduce the ‘inertia force’. We only see that it’s necessary to obtain
agreement with Newtonian mechanics (which is manifest in Eq.(1.13)).
We conclude with a few more words about this mystery. Recall from undergraduate
physics that in an accelerating coordinate system there is a fictional force = ma, which is
called the centrifugal force. We use it, for example, to analyze simple physics in a rotating
reference frame. If you are inside the rotating system and you throw a ball straight ahead
it will appear to curve away from your target, and if you did not know that you were
rotating relative to the rest of the universe then you’d think there was a force on the ball
equal to the centrifugal force. If you are inside a big rapidly rotating drum then you’ll
also feel pinned to the walls. This is an example of an inertia force which comes from
using a funny coordinate system. In general relativity, one sees that — in a certain sense
— gravity is an inertia force!
Chapter 2

Equations of Motion

(Week 2, April 4, 6, 8.)

In this chapter we’ll start to look at the Lagrangian equations of motion in more
depth. We’ll look at some specific examples of problem solving using the Euler-Lagrange
equations. First we’ll show how the equations are derived.

2.1 The Euler-Lagrange Equations


We are going to start thinking of a general classical system as a set of points in an abstract
configuration space or phase space1 . So consider an arbitrary classical system as living
in a space of points in some manifold Q. For example, the space for a spherical double
pendulum would look like Fig. 2.1, where Q = S 2 × S 2 . So our system is “a particle in

Q = S2 × S2

PSfrag replacements

Figure 2.1: Double pendulum configuration space.

Q”, which means you have to disabuse yourself of the notion that we’re dealing with real
1
The tangent bundle T Q will be referred to as configuration space, later on when we get to the chapter
on Hamiltonian mechanics we’ll find a use for the cotangent bundle T ∗ Q, and normally we call this the
phase space.

14
2.1 The Euler-Lagrange Equations 15

particles, we are not, we are dealing with a single quasi-particle in an abstract higher
dimensional space. The single quasi-particle represents two real particles if we are talking
about the classical system in Fig. 2.1. Sometimes to make this clear we’ll talk about “the
system taking a path”, instead of “the particle taking a path”. It is then clear that when
we say, “the system follows a path q(t)” that we’re referring to the point q in configuration
space Q that represents all of the particles in the real system.
So as time passes “the system” traces out a path

q : [t0 , t1 ] −→ Q

and we define it’s velocity,


q̇(t) ∈ Tq(t) Q
to be the tangent vector at q(t) given by the equivalence class [σ] of curves through q(t)
with derivatives σ̇(t) = dq(s)/ds|s=t. We’ll just write is as q̇(t).
Let Γ be the space of smooth paths from a ∈ Q to b ∈ Q,

Γ = {q : [t0 , t1 ] → Q|q(t0 ) = a, q(t1 ) = b}

(Γ is an infinite dimensional manifold, but we won’t go into that for now.) Let the La-
grangian=L for the system be any smooth function of position and velocity (not explicitly
of time, for simplicity),
L : T Q −→
and define the action, S:
S : Γ −→
by
Z t1
S(q) := L(q, q̇) dt (2.1)
t0

The path that the quasi particle will actually take is a critical point of S, in accord with
D’Alembert’s principle of least action. In other words, a path q ∈ Γ such that for any
smooth 1-parameter family of paths qs ∈ Γ with q0 = q1 , we have

d

S(qs ) =0 (2.2)
ds s=0

We write,
d
as “δ 00
ds s=0
so Eq.(2.2) can be rewritten
δS = 0 (2.3)
16 Equations of Motion

2.1.1 Comments
What is a “1-parameter family of paths”? Well, a path is a curve, or a 1D manifold. So
the 1-parameter family is nothing more nor less than a set of well-defined paths {qs }, each
one labeled by a parameter s. So a smooth 1-parameter family of paths will have q(s)
everywhere infinitesimally close to q(s + ) for an infinitesimal hyperreal . So in Fig. 2.2

qs
PSfrag replacements q0

Figure 2.2: Schematic of a 1-parameter family of curves.

we can go from q0 to qs by smoothly varying s from s = 0 to s = s


What does the condition δS = 0 imply? Patience, we are just getting to that. We will
now start to explore what δS = 0 means for our Lagrangian.

2.1.2 Lagrangian Dynamics


We were given that Q is a manifold, so it admits a covering of coordinate charts. For
now, let’s pick coordinates in a neighborhood U of some point q(t) ∈ Q. Next, consider
only variations qs such that qs = q outside U . A cartoon of this looks like Fig. 2.3 Then

a
U
PSfrag replacements
b

Figure 2.3: Local path variation.


2.1 The Euler-Lagrange Equations 17

we restrict attention to a subinterval [t00 , t01 ] ⊆ [t0 , t1 ] such that qs (t) ∈ U for t00 ≤ t ≤ t01 .
Let’s just go ahead and rename t00 and t01 as “ t0 and t1 ” to drop the primes. We can
use the coordinate charts on U ,

ϕ : U −→ n
x 7−→ ϕ(x) = (x1 , x2 , . . . , xn )

and we also have coordinates for the 1-forms,

dϕ : T U −→ T n ∼= n× n
(x, y) 7−→ dϕ(x, y) = (x1 , . . . , xn , y 1 , . . . , y n )

where y ∈ Tx Q. We restrict L : T M → to T U ⊆ T M, in our case the manifold is


M = Q, and then we can describe it, L, using the coordinates xi , y i on T U . The i are
generalized position coordinates, the y i are the associated generalized velocity coordinates.
(Velocity and position are in the abstract configuration space Q). Using these coordinates
we get
Z t1
δS = δ L(q(t), q̇(t)) dt
t0
Z t1
= δL(q, q̇) dt
t0
Z t1 
∂L i ∂L i 
= δq + i δ q̇ dt
t0 ∂xi ∂y

where we’ve used the given smoothness of L and the Einstein summation convention for
repeated indices i. Note that we can write δL as above using a local coordinate patch
because the path variations δq are entirely trivial outside the patch for U . Continuing,
using the Leibniz rule
d  ∂L  d ∂L ∂L
δq = δq + δ q̇
dt ∂y dt ∂y ∂y
we have,
Z t1 
∂L d ∂L  i
δS = − δq (t) dt
t0 ∂xi dt ∂y i
= 0.

If this integral is to vanish as demanded by δS = 0, then it must vanish for all path
variations δq, further, the boundary terms vanish because we deliberately chose δq that
18 Equations of Motion

vanish at the endpoints t0 and t1 inside U . That means the term in brackets must be
identically zero, or
d ∂L ∂L
i
− i =0 (2.4)
dt ∂y ∂x
This is necessary to get δS = 0, for all δq, but in fact it’s also sufficient. Physicists always
give the coordinates xi , y i on T U the symbols “q i ” and “q̇ i ”, despite the fact that these
also have another meaning, namely the xi and y i coordinates of the quantity,

q(t), q̇(t) ∈ T U.

So in any case, physicists write,

d ∂L ∂L
=
dt ∂ q̇ i ∂q i

and they call these the Euler-Lagrange equations.

2.2 Interpretation of Terms


The derivation of the Euler-Lagrange equations above was fairly abstract, the terms “po-
sition” and “velocity” were used but were not assumed to be the usual kinematic notions
that we are used to in physics, indeed the only reason we used those terms was for their
analogical appeal. Now we’ll try to illuminate the E-L equations a bit by casting them
into the usual position and velocity terms.
So consider,

Q=
1
L(q, q̇) = mq̇ · q̇ − V (q)
2
1
= mq̇ i q̇i − V (q)
2

TERMS MEANING MEANING


(in this example) (in general)
∂L
mq̇ the momentum pi
∂ q̇ i
∂L
−(∇V )i the force Fi
∂q i

When we write ∂V /∂q i = ∇V we’re assuming the q i are Cartesian coordinates on Q.


2.2 Interpretation of Terms 19

So translating our example into general terms, if we conjure up some abstract La-
grangian then we can think of the independent variables as generalized positions and
velocities, and then the Euler-Lagrange equations can be interpreted as equations relat-
ing generalized concepts of momentum and force, and they say that

ṗ = F (2.5)

So there’s no surprise that in the mundane case of a single particle moving in 3 under
time t this just recovers Newton II. Of course we can do all of our classical mechanics
with Newton’s laws, it’s just a pain in the neck to deal with the redundancies in F = ma
when we could use symmetry principles to vastly simplify many examples. It turns out
that the Euler-Lagrange equations are one of the reformulations of Newtonian physics
that make it highly convenient for introducing symmetries and consequent simplifications.
Simplifications generally mean quicker, shorter solutions and more transparent analysis
or at least more chance at insight into the characteristics of the system. The main thing
is that when we use symmetry to simplify the equations we are reducing the number
of independent variables, so it gets closer to the fundamental degrees of freedom of the
system and so we cut out a lot of the wheat and chaff (so to speak) with the full redundant
Newton equations.
One can of course introduce simplifications when solving Newton’s equations, it’s just
that it’s easier to do this when working with the Euler-Lagrange equations. Another good
reason to learn Lagrangian (or Hamiltonian) mechanics is that it translates better into
quantum mechanics.
Chapter 3

Lagrangians and Noether’s Theorem

If the form of a system of dynamical equations does not change under spatial translations
then the momentum is a conserved quantity. When the form of the equations is similarly
invariant under time translations then the total energy is a conserved quantity (a constant
of the equations of motion). Time and space translations are examples of 1-parameter
groups of transformations. Invariance under a group of transformations is precisely what
we mean by a symmetry in group theory. So symmetries of a dynamical system give
conserved quantities or conservation laws. The rigorous statement of all this is the content
of Noether’s theorem.

3.1 Time Translation


To handle time translations we need to replace our paths q : [t0 , t1 ] → Q by paths
q : → Q, and then define a new space of paths,

Γ = {q : → Q}.

The bad news is that the action


Z ∞  
S(q) = L q(t), q̇(t) dt
−∞

typically will not converge, so S is then no longer a function of the space of paths.
Nevertheless, if δq = 0 outside of some finite interval, then the functional variation,
Z ∞
d  
δS := L qs (t), q̇s (t) dt
−∞ ds s=0

will converge, since the integral is smooth and vanishes outside this interval. Moreover,
demanding that this δS vanishes for all such variations δq is enough to imply the Euler-

20
3.1 Time Translation 21

Lagrange equations:
Z
d 
∞ 
δS = L qs (t), q̇s (t) dt
−∞ ds s=0
Z ∞ 
∂L ∂L
= δqi + δ q̇i dt
−∞ ∂qi ∂ q̇i
Z ∞ 
∂L d ∂L
= − δqi dt
−∞ ∂qi dt ∂ q̇i

where again the boundary terms have vanished since δq = 0 near t = ±∞. To be explicit,
the first term in    
∂L i d ∂L i d ∂L
δ q̇ = δq − δq
∂ q̇ i dt ∂ q̇ i dt ∂ q̇ i
vanishes when we integrate. Then the whole thing vanishes for all compactly supported
smooth δq iff
d ∂L ∂L
= .
dt ∂ q̇i ∂qi
Recall that,

∂L
= pi , is the generalized momentum, by defn.
∂ q̇i
∂L
= ṗi , is the force, by the E-L eqns.
∂qi

Note the similarity to Hamilton’s equations—if you change L to H you need to stick
in a minus sign, and change variables from q̇ to pi and eliminate ṗi .

3.1.1 Canonical and Generalized Coordinates


In light of this noted similarity with the Hamilton equations of motion, let’s spend a few
moments clearing up some terminology (I hate using jargon, but sometimes it’s unavoid-
able, and sometimes it can be efficient—provided everyone is clued in).

Generalized Coordinates
For Lagrangian mechanics we have been using generalized coordinates, these are the
{qi , q̇i }. The qi are generalized positions, and the q̇i are generalized velocities. The full
set of independent generalized coordinates represent the degrees of freedom of a particle,
or system of particles. So if we have N particles then we’d typically have 6N general-
ized coordinates (the “6” is for 3 space dimensions, and at each point a position and a
momentum). These can be in any reference frame or system of axes, so for example,
22 Lagrangians and Noether’s Theorem

in a Cartesian frame, with two particles, in 3D space we’d have the 2 × 3 = 6 position
coordinates, and 2 × 3 = 6 velocities,

{x1 , y1 , z1 , x2 , y2 , z2 }, {u1 , v1 , w1 , u2 , v2 , w2 }

where say u = vx , v = vy , w = vz are the Cartesian velocity components. This makes


12 = 6 × 2 = 6N coordinates, matching the total degrees of freedom as claimed. If we
constrain the particles to move in a plane (say place them on a table in a gravitational
field) then we get 2N fewer degrees of freedom, and so 4N d.o.f. overall. By judicious
choice of coordinate frame we can eliminate one velocity component and one position
component for each particle.
It is also handy to respect other symmetries of a system, maybe the particles move on
a sphere for example, one can then define new positions and momenta with a consequent
reduction in the number of these generalized coordinates needed to describe the system.

Canonical Coordinates
In Hamiltonian mechanics (which we have not yet fully introduced) we will find it more
useful to transform from generalized coordinates to canonical coordinates. The canonical
coordinates are a special set of coordinates on the cotangent bundle of the configuration
space manifold Q. They are usually written as a set of (q i , pj ) or (xi , pj ) with the x’s or q’s
denoting the coordinates on the underlying manifold and the p’s denoting the conjugate
momentum, which are 1-forms in the cotangent bundle at the point q in the manifold.
It turns out that the q i together with the pj , form a coordinate system on the cotangent
bundle T ∗ Q of the configuration space Q, hence these coordinates are called the canonical
coordinates.
We will not discuss this here, but if you care to know, later on we’ll see that the
relation between the generalized coordinates and the canonical coordinates is given by
the Hamilton-Jacobi equations for a system.

3.2 Symmetry and Noether’s Theorem


First, let’s give a useful definition that will make it easy to refer to a type of dynamical
system symmetry. We want to refer to symmetry transformations (of the Lagrangian)
governed by a single parameter.

Definition 3.1 (one-parameter family of symmetries). A 1-parameter family of symme-


tries of a Lagrangian system L : T Q → is a smooth map,

F : × Γ −→ Γ
(s, q) 7−→ qs , with q0 = q
3.2 Symmetry and Noether’s Theorem 23

such that there exists a function `(q, q̇) for which


d`
δL =
dt
for some ` : T Q → , that is,
d   d  
L qs (t), q̇s (t) = ` qs (t), q̇s (t)
ds s=0 dt
for all paths q.

Remark: The simplest case is δL = 0, in which case we really have a way of moving
paths around (q 7→ qs ) that doesn’t change the Lagrangian—i.e., a symmetry of L in the
most obvious way. But δL = dtd ` is a sneaky generalization whose usefulness will become
clear.

3.2.1 Noether’s Theorem


Here’s a statement of the theorem. Note that ` in this theorem is the function associated
with F in definition 3.1.
Theorem 3.1 (Noether’s Theorem). Suppose F is a one-parameter family of symmetries
of the Lagrangian system, L : T Q → . Then,

pi δqi − `

is conserved, that is, it’s time derivative is zero for any path q ∈ Γ satisfying the Euler-
Lagrange equations. In other words, in boring detail,
 
d ∂L  d i 
q(s)q̇(s) q (t) − ` q(t), q̇(t) = 0
dt ∂y i ds s s=0
Proof.
d  d
pi δq i − ` = ṗi δq i + pi δ q̇ i − `
dt dt
∂L ∂L
= i δq i + i δ q̇ i − δL
∂q ∂ q̇
= δL − δL = 0.

“Okay, big deal” you might say. Before this can be of any use we’d need to find a
symmetry F . Then we’d need to find out what this pi δqi − ` business is that is conserved.
So let’s look at some examples.
24 Lagrangians and Noether’s Theorem

Example
1. Conservation of Energy. (The most important example!)
All of our Lagrangian systems will have time translation invariance (because the
laws of physics do not change with time, at least not to any extent that we can tell).
So we have a one-parameter family of symmetries

qs (t) = q(t + s)

This indeed gives,

δL = L̇

for

d d
L(qs ) = L = L̇
ds s=0 dt
so here we take ` = L simply! We then get the conserved quantity

pi δq i − ` = pi q̇ i − L
n
which we normally call the energy. For example, if Q = , and if
1
L = mq̇ 2 − V (q)
2
then this quantity is
1  1
mq̇ · q̇ − mq̇ · q̇ − V = mq̇ 2 + V (q)
2 2
The term in parentheses is K − V , and the left-hand side is K + V .

Let’s repeat this example, this time with a specific Lagrangian. It doesn’t matter what
the Lagrangian is, if it has 1-parameter families of symmetries then it’ll have conserved
quantities, guaranteed. The trick in physics is to write down a correct Lagrangian in the
first place! (Something that will accurately describe the system of interest.)

3.3 Conserved Quantities from Symmetries


We’ve seen that any 1-parameter family

Fs : Γ −→ Γ
q 7−→ qs
3.3 Conserved Quantities from Symmetries 25

which satisfies
δL = `˙
for some function ` = `(q, q̇) gives a conserved quantity

pi δq i − `

As usual we’ve defined 


d 
δL := L qs (t), q̇s (t)
ds s=0

Let’s see how we arrive at a conserved quantity from a symmetry.

3.3.1 Time Translation Symmetry


For any Lagrangian system, L : T Q → , we have a 1-parameter family of symmetries

qs (t) = q(t + s)

because
δL = L̇
so we get a conserved quantity called the total energy or Hamiltonian,

H = pi q̇ i − L (3.1)

(You might prefer “Hamiltonian” to “total energy” because in general we are not in the
same configuration space as Newtonian mechanics, if you are doing Newtonian mechanics
then “total energy” is appropriate.)
For example: a particle on n in a potential V has Q = n , L(q, q̇) = 21 mq̇ 2 − V (q).
This system has
∂L
pi q̇ i = i q̇ i = mq̇ 2 = 2K
∂ q̇
so
H = pi q̇ i − L = 2K − (K − V ) = K + V
as you’d have hoped.

3.3.2 Space Translation Symmetry


For a free particle in n , we have Q = n and L = K = 12 mq̇ 2 . This has spatial translation
symmetries, so that for any v ∈ n we have the symmetry

qs (t) = q(t) + s v
26 Lagrangians and Noether’s Theorem

with
δL = 0

because δ q̇ = 0 and L depends only on q̇ not on q in this particular case. (Since L does not
depend upon q i we’ll call q i an ignorable coordinate; as above, these ignorables always give
symmetries, hence conserved quantities. It is often useful therefore, to change coordinates
so as to make some of them ignorable if possible!)
In this example we get a conserved quantity called momentum in the v direction:

pi δq i = mq̇i v i = mq̇ · v

Aside: Note the subtle difference between two uses of the term “momentum”; here it
is a conserved quantity derived from space translation invariance, but earlier it was a
different thing, namely the momentum ∂L/∂ q̇ i = pi conjugate to q i . These two different
“momentum’s” happen to be the same in this example!
Since this is conserved for all v we say that mq̇ ∈ n is conserved. (In fact that
whole Lie group G = n is acting as a translation symmetry group, and we’re getting a
q(= n )-valued conserved quantity!)

3.3.3 Rotational Symmetry


The free particle in n also has rotation symmetry. Consider any X ∈ so(n) (that is a
skew-symmetric n × n matrix), then for all s ∈ the matrix esX is in SO(n), that is, it
describes a rotation. This gives a 1-parameter family of symmetries

qs (t) = esX q(t)

which has
∂L i ∂L i
δL = δq + i δ q̇ = mq̇i δ q̇ i
∂q i ∂ q̇
now qi is ignorable and so ∂L/∂q i = 0, and ∂L/∂ q̇ i = pi , and

i d i
δ q̇ = q̇
ds s s=0

d d  sX 
= e q
ds dt s=0
d
= Xq
dt
= X q̇
3.3 Conserved Quantities from Symmetries 27

So,

δL = mq̇i Xji q̇ j
= mq̇ · (X q̇)
=0

since X is skew symmetric as stated previously (X ∈ so(n)). So we get a conserved


quantity, the angular momentum in the X direction.
(Note: this whole bunch of math above for δL just says that the kinetic energy doesn’t
change when the velocity is rotated, without changing its magnitude.)
We write,
pi δq i = mq̇i · (X q)i
(δq i = Xq just as δ q̇ i = X q̇ in our previous calculation), or if X has zero entries except
in ij and ji positions, where it’s ±1, then we get

m(q̇i q j − q̇j q i )

the “ij component of angular momentum”. If n = 3 we write these as,

mq̇×q

Note that above we have assumed one can construct a basis for so(n) using matrices
of the form assumed for X, i.e., skew symmetric with ±1 in the respectively ij and ji
elements, otherwise zero.
I mentioned earlier that we can do mechanics with any Lagrangian, but if we want to
be useful we’d better pick a Lagrangian that actually describes a real system. But how
do we do that? All this theory is fine but is useless unless you know how to apply it.
The above examples were for a particularly simple system, a free particle, for which the
Lagrangian is just the kinetic energy, since there is no potential energy variation for a
free particle. We’d like to know how to solve more complicated dynamics.
The general idea is to guess the kinetic energy and potential energy of the particle (as
functions of your generalized positions and velocities) and then let,

L=K−V

So we are not using Lagrangians directly to tell us what the fundamental physical laws
should be, instead we plug in some assumed physics and use the Lagrangian approch to
solve the system of equations. If we like, we can then compare our answers with exper-
iments, which indirectly tells us something about the physical laws—but only provided
the Lagrangian formulation of mechanics is itself a valid procedure in the first place.
28 Lagrangians and Noether’s Theorem

3.4 Example Problems


(Week 3, Apr. 11, 13, 15.)

To see how the formalisms in this chapter function in practise, let’s do some problems.
It’s vastly superior to the simplistic F = ma formulation of mechanics. The Lagrangian
formulation allows the configuration space to be any manifold, and allows us to easily use
any coordinates we wish.

3.4.1 The Atwood Machine


A frictionless pulley with two masses, m1 and m2 , hanging from it. We have

x
PSfrag replacements
`−x
m1

m2

1 d 1
K = (m1 + m2 )( (` − x))2 = (m1 + m2 )ẋ2
2 dt 2
V = −m1 gx − m2 g(` − x)

so

1
L = K − V = (m1 + m2 )ẋ2 + m1 gx + m2 g(` − x)
2
The configuration space is Q = (0, `), and x ∈ (0, `) (we could use the “owns” symbol 3
here and write Q = (0, `) 3 x ). Moreover T Q = (0, `) × 3 (x, ẋ). As usual L : T Q → .
Note that solutions of the Euler-Lagrange equations will only be defined for some time
t ∈ , as eventually the solutions reaches the “edge” of Q.
The momentum is:
∂L
p= = (m1 + m2 )ẋ
∂ ẋ
and the force is:
∂L
F = = (m1 − m2 )g
∂x
3.4 Example Problems 29

The Euler-Lagrange equations say


ṗ = F
(m1 + m2 )ẍ = (m1 − m2 )g
m1 − m 2
ẍ = g
m1 + m 2
 
1 −m2
So this is like a falling object in a downwards gravitational acceleration a = m m1 +m2
g.
It is trivial to integrate the expression for ẍ twice (feeding in some appropriate initial
conditions) to obtain the complete solution to the motion x(t) and ẋ(t). Note that ẍ = 0
when m1 = m2 , and ẍ = g if m2 = 0.

3.4.2 Disk Pulled by Falling Mass


Consider next a disk pulled across a table by a falling mass. The disk is free to move on
a frictionless surface, and it can thus whirl around the hole to which it is tethered to the
mass below.

r m1
PSfrag replacements

`−r no swinging allowed!

m2

Here Q = open disk of radius `, minus it’s center


= (0, `) × S 1 3 (r, θ)
T Q = (0, `) × S 1 × × 3 (r, θ, ṙ, θ̇)
1 1 d
K = m1 (ṙ 2 + r 2 θ̇ 2 ) + m2 ( (` − r))2
2 2 dt
V = gm2 (r − `)
1 1
L = m1 (ṙ 2 + r 2 θ̇ 2 ) + m2 ṙ 2 + gm2 (` − r)
2 2
having noted that ` is constant so d/dt(` − r) = −ṙ. For the momenta we get,
∂L
pr = = (m1 + m2 )ṙ
∂ ṙ
∂L
pθ = = m1 r 2 θ̇.
∂ θ̇
30 Lagrangians and Noether’s Theorem

Note that θ is an “ignorable coordinate”—it doesn’t appear in L—so there’s a symmetry,


rotational symmetry, and pθ , the conjugate momentum, is conserved.
The forces are,

∂L
Fr = = m1 r θ̇ 2 − gm2
∂r
∂L
Fθ = = 0, (θ is ignorable)
∂θ

Note: in Fr the term m1 r θ̇ 2 is recognizable as a centrifugal force, pushing m1 radially


out, while the term −gm2 is gravity pulling m2 down and thus pulling m1 radially in.
So, the Euler-Lagrange equations give,

ṗr = Fr , (m1 + m2 )r̈ = m1 r θ̇ 2 − m2 g


ṗθ = 0, pθ = m1 r 2 θ̇ = J = a constant.

Let’s use our conservation law here to eliminate θ̇ from the first equation:

J
θ̇ =
m1 r 2

so

J2
(m1 + m2 )r̈ = − m2 g
m1 r 3

Thus effectively we have a particle on (0, `) of mass m = m1 + m2 feeling a force

J2
Fr = − m2 g
m1 r 3

which could come from an “effective potential” V (r) such that dV /dr = −Fr . So integrate
−Fr to find V (r):
J2
V (r) = + m2 gr
2m1 r 2
this is a sum of two terms that look like Fig. 3.1
If θ̇(t = 0) = 0 then there is no centrifugal force and the disk will be pulled into the
hole until it gets stuck. At that time the disk reaches the hole, which is topologically the
center of the disk that has been removed from Q, so then we’ve hit the boundary of Q
and our solution is broken.
At r = r0 , the minimum of V (r), our disc mass m1 will be in a stable circular orbit of
radius r0 (which depends upon J). Otherwise we get orbits like Fig. 3.2.
3.4 Example Problems 31

V (r)
PSfrag replacements
attractive gravitational potential

repulsive centrifugal potential

r0 r
no swinging allowed!

Figure 3.1: Potential function for disk pulled by gravitating mass.

Figure 3.2: Orbits for the disc and gravitating mass system.

3.4.3 Free Particle in Special Relativity


In relativistic dynamics the parameter coordinate that parametrizes the particle’s path
in Minkowski spacetime need not be the “time coordinate”, indeed in special relativity
there are many allowed time coordinates.
Minkowski spacetime is,
n+1
3 (x0 , x1 , . . . , xn )
if space is n-dimensional. We normally take x0 as “time”, and (x1 , . . . , xn ) as “space”,
but of course this is all relative to one’s reference frame. Someone else travelling at some
high velocity relative to us will have to make a Lorentz transformation to translate from
our coordinates to theirs.
This has a Lorentzian metric

g(v, w) = v 0 w 0 − v 1 w 1 − . . . − v n w n
= ηµν v µ w ν
32 Lagrangians and Noether’s Theorem

where  
1 0 0 ... 0
 0 −1 0 . . . 0 
 
 0 0 −1 0 
ηµν = 
 .. .. .. . 
 . . . .. 
0 0 0 . . . −1
In special relativity we take spacetime to be the configuration space of a single point
particle, so we let Q be Minkowski spacetime, i.e., n+1 3 (x0 , . . . , xn ) with the metric
ηµν defined above. Then the path of the particle is,

q : (3 t) −→ Q

where t is a completely arbitrary parameter for the path, not necessarily x0 , and not
necessarily proper time either. We want some Lagrangian L : T Q → , i.e., L(q i , q̇ i ) such
that the Euler-Lagrange equations will dictate how our free particle moves at a constant
velocity. Many Lagrangians do this, but the “best” ones(s) give an action that is inde-
pendent of the parameterization of the path—since the parameterization is “unphysical”
(it can’t be measured). So the action
Z t1  
i i
S(q) = L q (t), q̇ (t) dt
t0

for q : [t0 , t1 ] → Q, should be independent of t. The obvious candidate for S is mass times
arclength, Z t1 q
S=m ηij q̇ i (t)q̇ j (t) dt
t0

or rather the Minkowski analogue of arclength, called proper time, at least when q̇ is a
timelike vector, i.e., ηij q̇ i q̇ j > 0, which says q̇ points into the future (or past) lightcone
and makes S real, in fact it’s then the time ticked off by a clock moving along the path q :
[t0 , t1 ] → Q. By “obvious candidate” we are appealing somewhat to physical intuition and

Timelike
Lightlike
Spacelike

generalization. In Euclidean space, free particles follow straight paths, so the arclength
or pathlength variation is an extremum, and we expect the same behavior in Minkowski
3.4 Example Problems 33

spacetime. Also, the arclength does not depend upon the parameterization, and lastly,
the mass m merely provides the correct units for ‘action’.
So let’s take p
L = ηij q̇ i q̇ j (3.2)
and work out the Euler-Lagrange equations. We have
∂L ∂ p
pi = = ηij q̇ i q̇ j
∂ q̇ i ∂ q̇ i
2ηij q̇ j
=m p
2 ηij q̇ i q̇ j
ηij q̇ j mq̇i
=m p =
i
ηij q̇ q̇ j kq̇k

(Note the numerator is “mass times 4-velocity”, at least when n = 3 for a real single
particle system, but we’re actually in a more general n + 1-dim spacetime, so it’s more
like the “mass times n + 1-velocity”). Now note that this pi doesn’t change when we
change the parameter to accomplish q̇ 7→ αq̇. The Euler-Lagrange equations say,
∂L
ṗi = Fi = =0
∂q i
The meaning of this becomes clearer if we use “proper time” as our parameter (like
parameterizing a curve by it’s arclength) so that
Z t1
kq̇kdt = t1 − t0 , ∀ t0 , t1
t0

which fixes the parametrization up to an additive constant. This implies kq̇k = 1, so that
q̇i
pi = m = mq̇i
kq̇k
and the Euler-Lagrange equations say

ṗi = 0 ⇒ mq̈i = 0

so our (free) particle moves unaccelerated along a straight line, which is as we desired
(expected).

Comments
This Lagrangian from Eq.(3.2) has lots of symmetries coming from reparameterizing the
path, so Noether’s theorem yields lots of conserved quantities for the relativistic free
34 Lagrangians and Noether’s Theorem

particle. This is in fact called “the problem of time” in general relativity, here we see it
starting to show up in special relativity.
These reparameterization symmetries work as follows. Consider any (smooth) 1-
parameter family of reparameterizations, i.e., diffeomorphisms

fs : −→

with f0 =  . These act on the space of paths Γ = {q : → Q} as follows: given any
q ∈ Γ we get 
qs (t) = q fs (t)
where we should note that qs is physically indistinguishable from q. Let’s show that
˙
δL = `, (when E-L eqns. hold)

so that Noether’s theorem gives a conserved quantity

pi δq i − `

Here we go then.
∂L i ∂L i
δL = δq + i δ q̇
∂ q̇ i ∂q
i
= pi δ q̇
mq̇i d i  
= q̇ fs (t)
kq̇k ds s=0
mq̇i d d i  
= q fs (t)
kq̇k dt ds s=0
mq̇i d i   f (t)
= q̇ fs (t)
s
kq̇k dt ds s=0
mq̇i d i 
= q̇ δfs
kq̇k dt
d 
= pi q̇ i δf
dt
where in the last step we used the E-L eqns., i.e. dtd pi = 0, so δL = `˙ with ` = pi q̇ i δf .
So to recap a little: we saw the free relativistic particle has
p
L = mkq̇k = m ηij q̇ i q̇ j

and we’ve considered reparameterization symmetries


 
qs (t) = q fs (t) , fs : →
3.5 Electrodynamics and Relativistic Lagrangians 35

we’ve used the fact that 


d i
i
δq := q fs (t) = q̇ i δf
ds s=0

so (repeating a bit of the above)

∂L i ∂L i
δL = δq + i δ q̇
∂q i ∂ q̇
= pi δ q̇ , (since ∂L/∂q i = 0, and ∂L/∂ q̇ i = p)
i

= pi δ q̇ i
d
= pi δq i
dt
d
= pi q̇ i δf
dt
d
= pi q̇ i δf, and set pi q̇ i δf = `
dt
so Noether’s theorem gives a conserved quantity

pi δq i − ` = pi q̇ i δf − pi q̇ i δf
=0

So these conserved quantities vanish! In short, we’re seeing an example of what


physicists call gauge symmetries. This is a good topic for starting a new section.

3.5 Electrodynamics and Relativistic Lagrangians


We will continue the story of symmetry and Noether’s theorem from the last section with
a few more examples. We use principles of least action to conjure up Lagrangians for
our systems, realizing that a given system may not have a unique Lagrangian but will
often have an obvious natural Lagrangian. Given a Lagrangian we derive equations of
motion from the Euler-Lagrange equations. Symmetries of L guide us in finding conserved
quantities, in particular Hamiltonians from time translation invariance, via Noether’s
theorem.
This section also introduces gauge symmetry, and this is where we begin.

3.5.1 Gauge Symmetry and Relativistic Hamiltonian


What are gauge symmetries?

1. These are symmetries that permute different mathematical descriptions of the same
physical situation—in this case reparameterizations of a path.
36 Lagrangians and Noether’s Theorem

2. These symmetries make it impossible to compute q(t) given q(0) and q̇(0): since if
q(t) is a solution so is q(f (t)) for any reparameterization f : → . We have a high
degree of non-uniqueness of solutions to the Euler-Lagrange equations.

3. These symmetries give conserved quantities that work out to equal zero!

Note that (1) is a subjective criterion, (2) and (3) are objective, and (3) is easy to
test, so we often use (3) to distinguish gauge symmetries from physical symmetries.

3.5.2 Relativistic Hamiltonian


What then is the Hamiltonian for special relativity theory? We’re continuing here with
the example problem of §3.4.3. Well, the Hamiltonian comes from Noether’s theorem
from time translation symmetry,

qs (t) = q(t + s)

and this is an example of a reparametrization (with δf = 1), so we see from the previous
results that the Hamiltonian is zero!

H = 0.

Explicitly, H = pi δ q̇ i − ` where under q(t) → q(t + s) we have δ q̇ i = q̇ i δf , and so


δL = d`/dt, which implies ` = pi δq i . The result H = 0 follows.
Now you know why people talk about “the problem of time” in general relativity
theory, it’s glimmerings are seen in the flat Minkowski spacetime of special relativity.
You may think it’s nice and simple to have H = 0, but in fact it means that there is no
temporal evolution possible! So we can’t establish a dynamical theory on this footing!
That’s bad news. (Because it means you might have to solve the static equations for the
4D universe as a whole, and that’s impossible!)
But there is another conserved quantity deserving the title of “energy” which is not
zero, and it comes from the symmetry,

qs (t) = q(t) + s w

where w ∈ n+1 and w points in some timelike direction.


In fact any vector w gives a conserved quantity,

∂L i ∂L i
δL = δq + i δ q̇
∂q i ∂ q̇
= pi δ q̇ , (since ∂L/∂q i = 0 and ∂L/∂ q̇ i = pi )
i

= pi 0 = 0
3.6 Relativistic Particle in an Electromagnetic Field 37

qs

q
PSfrag replacements w

since δq i = w i , δ q̇ i = ẇ i = 0. This is our `˙ from Noether’s theorem with ` = 0, so


Noether’s theorem says that we get a conserved quantity

pi δq i − ` = piw

namely, the momentum in the w direction. We know ṗ = 0 from the Euler-Lagrange


equations, for our free particle, but here we see it coming from spacetime translation
symmetry;

p =(p0 , p1 , . . . , pn )
p0 is energy, (p1 , . . . , pn ) is spatial momentum.

We’ve just about exhausted all the basic stuff that we can learn from the free particle.
So next we’ll add some external force via an electromagnetic field.

3.6 Relativistic Particle in an Electromagnetic Field


The electromagnetic field is described by a 1-form A on spacetime, A is the vector potential,
such that
dA = F (3.3)
is a 2-form containing the electric and magnetic fields,
∂Ai ∂Aj
F µν = − (3.4)
∂xj ∂xi
n+1
We’d write (for Q having local charts to ),

A = A0 dx0 + A1 dx1 + . . . An dxn


38 Lagrangians and Noether’s Theorem

and then because d2 = 0

dA = dA0 dx0 + dA1 dx1 + . . . dAn dxn


and since the “Ai ” are just functions,
dAi = ∂µ Ai dxµ
using the summation convention and ∂µ := ∂/∂xµ . The student can easily check that the
components for F = F01 dx0 ∧ dx1 + F02 dx0 ∧ dx2 + . . ., agrees with the matrix expression
below (at least for 4D).
So, for example, in 4D spacetime
 
0 E1 E2 E3
 −E1 0 B3 −B2 
F =  −E2 −B3

0 B1 
−E3 B2 −B1 0
where E is the electric field and B is the magnetic field. The action for a particle of
charge e is Z t1 Z
S=m kq̇k dt + e A
t0 q
here
Z t1
kq̇k dt = proper time,
t0
Z
A = integral of A along the path q.
q

Note that since A is a 1-form it can be integrated (it is a linear combination of some basis
1-forms like the {dxi }).

(Week 4, April 18, 20, 22.)

Note that since A is a 1-form we can integrate it over an oriented manifold, but one
can also write the path integral using time t as a parameter, with Ai q̇ i dt the differential,
after dq i = q̇ i dt.
The Lagrangian in the above action, for a charge e with mass m in an electromagnetic
potential A is
L(q, q̇) = mkq̇k + eAi q̇ i (3.5)
so we can work out the Euler-Lagrange equations:
∂L q̇i
pi = i
=m + eAi
∂ q̇ kq̇k
= mvi + e Ai
3.7 Alternative Lagrangians 39

where v ∈ n+1 is the velocity, normalized so that kvk = 1. Note that now momentum is
no longer mass times velocity! That’s because we’re in n + 1-d spacetime, the momentum
is an n + 1-vector. Continuing the analysis, we find the force

∂L ∂  j

Fi = = e A j q̇
∂q i ∂q i
∂Aj
= e i q̇ j
∂q
 
So the Euler-Lagrange equations say (noting that Ai = Aj q(t) :

ṗ = F
d  ∂Aj
mvi + eAi = e i q̇ j
dt ∂q
dvi ∂Aj dAi
m = e i q̇ j − e
dt ∂q dt
dvi ∂Aj ∂Ai
m = e i q̇ j − e j q̇ j
dt ∂q ∂q
 
∂Aj ∂Ai
=e i
− j q̇ j
∂q ∂q

the term in parentheses is F ij = the electromagnetic field, F = dA. So we get the following
equations of motion

dvi
m = eF ij q̇ j , (Lorentz force law) (3.6)
dt

(Usually called the “Lorenz” force law.)

3.7 Alternative Lagrangians


We’ll soon discuss a charged particle Lagrangian that is free of the reparameterization
symmetry. First a paragraph on objects other than point particles!

3.7.1 Lagrangian for a String


So we’ve looked at a point particle and tried
Z
S = m · (arclength) + A
40 Lagrangians and Noether’s Theorem

or with ‘proper time’ instead of ‘arclength’, where the 1-from A can be integrated over
a 1-dimensional path. A generalization (or specialization, depending on how you look at
it) would be to consider a Lagrangian for an extended object.
In string theory we boost the dimension by +1 and consider a string tracing out a 2D
surface as time passes (Fig. 3.3).

becomes

Figure 3.3: Worldtube of a closed string.

Can you infer an appropriate action for this system? Remember, the physical or
physico-philosophical principle we’ve been using is that the path followed by physical
objects minimizes the “activity” or “aliveness” of the system. Given that we presumably
cannot tamper with the length of the closed string, then the worldtube quantity analogous
to arclength or proper time would be the area of the worldtube (or worldsheet for an open
string). If the string is also assumed to be a source of electromagnetic field then we need
a 2-form to integrate over the 2D worldtube analogous to the 1-form integrated over the
pathline of the point particle. In string theory this is usually the “Kalb-Ramond field”,
call it B. To recover electrodynamic interactions it should be antisymmetric like A, but
it’s tensor components will have two indices since it’s a 2-form. The string action can
then be written Z
S = α · (area) + e B (3.7)

We’ve also replaced the point particle mass by the string tension α [mass·length−1 ] to
obtain the correct units for the action (since replacing arclength by area meant we had to
compensate for the extra length dimension in the first term of the above string action).
This may still seem like we’ve pulled a rabbit out of a hat. But we haven’t checked that
this action yields sensible dynamics yet! But supposing it does, then would it justify our
guesswork and intuition in arriving at Eq.(3.7)? Well by now you’ve probably realized
that one can have more than one form of action or Lagrangian that yields the same
dynamics. So provided we supply reasonabe physically realistic heuristics then whatever
Lagrangian or action that we come up with will stand a good chance of describing some
system with a healthy measure of physical verisimilitude.
3.7 Alternative Lagrangians 41

That’s enough about string for now. The point was to illustrate the type of reasoning
that one can use in conjuring up a Lagrangian. It’s particularly useful when Newtonian
theory cannot give us a head start, i.e., in relativistic dynamics and in the physics of
extended particles.

3.7.2 Alternate Lagrangian for Relativistic Electrodynamics


In § 3.6, Eq.(3.5) we saw an example of a Lagrangian for relativistic electrodynamics
that had awkward reparametrization symmetries, meaning that H = 0 and there were
non-unique solutions to the Euler-Lagrange equations arising from applying gauge trans-
formations. This freedom to change the gauge can be avoided.
Recall Eq.(3.5), which was a Lagrangian for a charged particle with reparametrization
symmetry
L = mkq̇k + eAi q̇ i
just as for an uncharged relativistic particle. But there’s another Lagrangian we can use
that doesn’t have this gauge symmetry:
1
L = mq̇ · q̇ + eAi q̇ i (3.8)
2
This one even has some nice features.

• It looks formally like “ 21 mv 2 ”, familiar from nonrelativistic mechanics.

• There’s no ugly square root, so it’s everywhere differentiable, and there’s no trouble
with paths being timelike or spacelike in direction, they are handled the same.

What Euler-Lagrange equations does this Lagrangian yield?

∂L
pi = = mq̇i + eAi
∂ q̇ i
∂L ∂Aj
Fi = i = e i q̇ j
∂q ∂q

Very similar to before! The E-L eqns. then say

d  ∂Aj
mq̇i + eAi = e i q̇ j
dt ∂q
mq̈i = eF ij q̇ j

almost as before. (I’ve taken to using F here for the electromagnetic field tensor to avoid
clashing with F for the generalized force.) The only difference is that we have mq̈i instead
of mv̇i where vi = q̇i /kq̇k. So the old Euler-Lagrange equations of motion reduce to the
42 Lagrangians and Noether’s Theorem

new ones if we pick a parametrization with kq̇k = 1, which would be a parametrization


by proper time for example.
Let’s work out the Hamiltonian for this
1
L = mq̇ · q̇ + eAi q̇ i
2
for the relativistic charged particle in an electromagnetic field. Recall that for our
reparametrization-invariant Lagrangian
p
L = m q̇i q̇ i + eAi q i
we got H = 0, time translation was a gauge symmetry. With the new Lagrangian it’s
not! Indeed
H = pi q̇ i − L
and now
∂L
pi = = mq̇ i + eAi
∂ q̇ i
so
H = (mq̇i + eAi )q̇ i − ( 12 mq̇i q̇ i + eAi q̇ i )
= 12 mq̇i q̇ i

Comments. This is vaguely like how a nonrelativistic particle in a potential V has


H = pi q̇ i − L = 2K − (K − V ) = K + V,
but now the “potential’ V = eAi q̇ i in linear in velocity, so now
H = pi q̇ i − L = (2K − V ) − (K − V ) = K.
As claimed H is not zero, and the fact that it’s conserved says kq̇(t)k is constant as
a function of t, so the particle’s path is parameterized by proper time up to rescaling
of t. That is, we’re getting “conservation of speed” rather than some more familiar
“conservation of energy”. The reason is that this Hamiltonian comes from the symmetry
qs (t) = q(t + s)
instead of spacetime translation symmetry
n+1
qs (t) = q(t) + s w, w∈
the difference is illustrated schematically in Fig. 3.4.
Our Lagrangian
L(q, q̇) = 21 mkq̇k2 + Ai (q)q̇ i
has time translation symmetry iff A is translation invariant (but it’s highly unlikely a
given system of interest will have A(q) = A(q + sw)). In general then there’s no conserved
“energy” for our particle corresponding to translations in time.
3.8 The General Relativistic Particle 43

4 5
3 4 w
2 3
1 2
PSfrag replacements 0 1

qs (t) = q(t + s) qs (t) = q(t) + sw

Figure 3.4: Proper time rescaling vs spacetime translation.

3.8 The General Relativistic Particle


In GR spacetime, Q, is an (n + 1)-dimensional Lorentzian manifold, namely a smooth
(n + 1)-dimensional manifold with a Lorentzian metric g. We define the metric as follows.
1. For each x ∈ Q, we have a bilinear map
g(x) : Tx Q × Tx Q −→
(v, w) 7−→ g(x)(v, w)
or we could write g(v, w) for short.
2. With respect to some basis of Tx Q we have
g(v, w) = gij v i w j
 
1 0 ... 0
 0 −1 0 
 
gij =  .. . . . ... 
 . 
0 0 . . . −1

Of course we can write g(v, w) = gij v i w j in any basis, but for different bases gij will
have a different form.
3. g(x) varies smoothly with x.
44 Lagrangians and Noether’s Theorem

3.8.1 Free Particle Lagrangian in GR


The Lagrangian for a free point particle in the spacetime Q is
p
L(q, q̇) = m g(q)(q̇, q̇)
p
= m gij q̇ i q̇ j
just like in special relativity but with ηij replaced by gij . Alternatively we could just as
well use
L(q, q̇) = 12 mg(q)(q̇, q̇)
= 12 mgij q̇ i q̇ j
The big difference between these two Lagrangians is that now spacetime translation
symmetry (and rotation, and boost symmetry) is gone! So there is no conserved energy-
momentum (nor angular momentum, nor velocity of center of energy) anymore!
Let’s find the equations of motion. Suppose then Q is a Lorentzian manifold with
metric g and L : T Q → is the Lagrangian of a free particle,
L(q, q̇) = 21 mgij q̇ i q̇ j
We find equations of motion from the Euler-Lagrange equations, which in this case start
from
∂L
pi = i = mgij q̇ j
∂ q̇
The velocity q̇ here is a tangent vector, the momentum p is a cotangent vector, and we
need the metric to relate them, via
g : Tq M × Tq M −→
(v, w) 7−→ g(v, w)
which gives
Tq M −→ Tq∗ M
v 7−→ g(v, −).
In coordinates this would say that the tangent vector v i gets mapped to the cotangent
vector gij v j . This is lurking behind the passage from q̇ i to the momentum mgij v j .
Getting back to the E-L equations,
∂L
pi = i
= mgij q̇ j
∂ q̇
∂L ∂  
Fi = i = i 12 mgjk (q)q̇ j q̇ k
∂q ∂q

= 12 m∂i gjk q̇ j q̇ k , (where ∂i = ).
∂q i
3.8 The General Relativistic Particle 45

So the Euler-Lagrange equations say


d
mgij q̇ j = 21 m∂i gjk q̇ j q̇ k .
dt
The mass factors away, so the motion is independent of the mass! Essentially we have a
geodesic equation.
We can rewrite this geodesic equation as follows
d
gij q̇ j = 12 ∂i gjk q̇ j q̇ k
dt
∴ ∂k gij q̇ k q̇ j + gij q̈ j = 12 ∂i gjk q̇ j q̇ k

∴ gij q̈ j = 1
∂g
2 i jk
− ∂k gij q̇ j q̇ k

= 1
2
∂i gjk − ∂k gij − ∂j gki q̇ j q̇ k
where the last line follows by symmetry of the metric, gik = gki . Now let,

Γijk = − ∂i gjk − ∂k gij − ∂j gki
the minus sign being just a convention (so that we agree with everyone else). This defines
what we call the Christoffel symbols Γijk . Then

q̈i = gij q̈ j = −Γijk q̇ j q̇ k


∴ q̈ i = −Γijk q̇ j q̇ k .
So we see that q̈ can be computed in terms of q̇ and the Christoffel symbols Γijk , which
is really a particular type of connection that a Lorentzian manifold has (the Levi-Civita
connection), a connection is just the rule for parallel transporting tangent vectors around
the manifold.
Parallel transport is just the simplest way to compare vectors at different points in
the manifold. This allows us to define, among other things, a covariant derivative.

3.8.2 Charged particle in EM Field in GR


We can now apply what we’ve learned in consideration of a charged particle, of charge e,
in an electromagnetic field with potential A, in our Lorentzian manifold. The Lagrangian
would be
L = 21 mgjk q̇ j q̇ k + eAi q̇ i
which again was conjured up be replacing the flat space metric ηij by the metric for GR
gij . Not surprisingly, the Euler-Lagrange equations then yield the following equations of
motion,
mq̈i = −mΓijk q̇ j q̇ k + eF ij q̇ j .
If you want to know more about Lagrangians for general relativity we recommend the
paper by Peldan [Pel94], and also the “black book” of Misner, Thorne & Wheeler [WTM71].
46 Lagrangians and Noether’s Theorem

3.9 The Principle of Least Action and Geodesics


(Week 4, April 18, 20, 22.)

3.9.1 Jacobi and Least Time vs Least Action


We’ve mentioned that Fermat’s principle of least time in optics is analogous to the prin-
ciple of least action in particle mechanics. This analogy is strange, since in the principle
of least action we fix the time interval q : [0, 1] → Q. Also, if one imagines a force on a
particle resulting from a potential gradient at an interface as analogous to light refraction
then you also get a screw-up in the analogy (Fig. 3.5).

PSfrag replacements
light particle
n high V high
faster slower

n low light particle


V low
slower faster

Figure 3.5: Least time versus least action.

Nevertheless, Jacobi was able to reinterpret the mechanics of a particle as an optics


problem and hence “unify” the two minimization principles. First, let’s consider light in
a medium with a varying index of refraction n (recall 1/n ∝ speed of light). Suppose it’s
in n with its usual Euclidean metric. If the light s trying to minimize the time, its trying
to minimize the arclength of its path in the metric

gij = n2 δij

n
that is, the index of refraction n : → (0, ∞), times the usual Euclidean metric
 
1
 ..
0 
δij = 
 . 
0 1
3.9 The Principle of Least Action and Geodesics 47

This is just like the free particle in general relativity (minimizing it’s proper time)
except that now gij is a Riemannian metric

g(v, w) = gij v i w j
where g(v, v) ≥ 0

So we’ll use the same Lagrangian,


q
L(q, q̇) = gij (q)q̇ i q̇ j

and get the same Euler-Lagrange equations,


d2 q i
+ Γijk q̇ j q̇ k (3.9)
dt2
if q is parameterized by arclength or more generally
q
kq̇k = gij (q)q̇ i q̇ j = constant.

As before the Christoffel symbols Γ are built from the derivatives of the metric g.
Now, what Jacobi did is show how the motion of a particle in a potential could be
viewed as a special case of this. Consider a particle of mass m in Euclidean n with
potential V : n → . It satisfies F = ma, i.e.,
d2 q i
m = −∂i V (3.10)
dt2
How did Jacobi see (3.10) as a special case of (3.9)? He considered a particle of energy
E and he chose the index of refraction to be
r
2 
n(q) = E − V (q)
m
which is just the speed of a particle of energy E when the potential energy is V (q), since
r r
2 21
(E − V ) = mkq̇k2 = kq̇k.
m m2
Note: this is precisely backwards compared to optics, where n(q) is proportional to the
reciprocal of the speed of light!! But let’s see that it works.
q
L = gij (q)q̇ i q̇ j
p
= n2 (q)q̇ i q̇ j
p
= 2/m(E − V (q))q̇ 2
48 Lagrangians and Noether’s Theorem

where q̇ 2 = q̇ · q̇ is just the usual Euclidean dot product, v · w = δij v i w j . We get the
Euler-Lagrange equations,
r
∂L 2 q̇
pi = i = (E − V ) ·
∂ q̇ m kq̇k
r
∂L 2
Fi = i = ∂ i (E − V (q)) · kq̇k
∂q m
1 −2/m∂i V
= p · kq̇k
2 2/m(E − V q)

Then ṗ = F says,
dp q̇i 1 kq̇k
2/m(E − V (q)) · = − ∂i V p
dt kq̇k m 2/m(E − V )

Jacobi noticed that this is just F = ma, or mq̈i = −∂i V , that is, provided we reparame-
terize q so that, p
kq̇k = 2/m(E − V (q)).
Recall that our Lagrangian gives reparameterization invariant Euler-Lagrange equations!
This is the unification between least time (from optics) and least action (from mechanics)
that we sought.

3.9.2 The Ubiquity of Geodesic Motion


We’ve seen that many classical systems trace out paths that are geodesics, i.e., paths
q : [t0 , t1 ] → Q that are critical points of
Z t1 p
S(q) = gij q̇ i q̇ j dt
t0

which is proper time when (Q, g) is a Lorentzian manifold, or arclength when (Q, g) is a
Riemannian manifold. We have

1. The metric at q ∈ Q is,

g(q) : Tq Q × Tq Q →
(v, w) 7→ g(v, w)

and it is bilinear.

2. w.r.t a basis of Tq Q
g(v, w) = δij v i w j
3.9 The Principle of Least Action and Geodesics 49

3. g(q) varies smoothly with q ∈ Q.


An important distinction to keep in mind is that Lorentzian manifolds represent space-
times, whereas Riemannian manifolds represent that we’d normally consider as just space.
We’ve seen at least three important things.
n
(1) In the geometric optics approximation, light in Q = acts like particles tracing
out geodesics in the metric
gij = n(q)2 δij
where n : Q → (0, ∞) is the index of refraction function.
n
(2) Jacobi saw that a particle in Q = in some potential V : Q → traces out
geodesics in the metric
2
(E − V )δij
gij =
m
if the particle has energy E (where1 V < E).
(3) A free particle in general relativity traces out a geodesic on a Lorentzian manifold
(Q, q).
In fact all three of these results can be generalized to cover every problem that we’ve
discussed!
(10 ) Light on any Riemannian manifold (Q, q) with index of refraction n : Q → (0, ∞)
traces out geodesics in the metric h = n2 g.
(20 ) A particle on a Riemannian manifold (Q, q) with potential V : Q → traces out
geodesics w.r.t the metric
2
h = (E − V )g
m
if it has energy E. Lots of physical systems can be described this way, e.g., the
Atwood machine, a rigid rotating body (Q = SO(3)), spinning tops, and others.
All of these systems have a Lagrangian which is a quadratic function of position, so
they all fit into this framework.
(30 ) Kaluza-Klein Theory. A particle with charge e on a Lorentzian manifold (Q, q)
in an electromagnetic vector potential follows a path with
e
q̈i = −Γijk q̇ j q̇ k + F ij q̇ j
m
where
F ij = ∂i Aj − ∂j Ai
but this is actually geodesic motion on the manifold Q × U (1) where U (1) = {eiθ :
θ ∈ } is a circle.
1
The case V > E, if they exist, would be classically forbidden regions.
50 Lagrangians and Noether’s Theorem

Let’s examine this last result a bit further. To get the desired equations for motion on
Q × U (1) we need to given Q × U (1) a cleverly designed metric built from g and A where
the amount of “spiralling”—the velocity in the U (1) direction is e/m. The metric h on

Q × U (1)
PSfrag replacements
a geodesic

Q
the apparent path

Q × U (1) is built from g and A in a very simple way. Let’s pick coordinates xi on Q where
i ∈ {0, . . . , n} since we’re in n + 1-dimensional spacetime, and θ is our local coordinate
on S 1 . The components of h are

hij = gij + Ai Aj
hθi = hθi = −Ai
hθθ = 1

Working out the equations for a geodesic in this metric we get


e
q̈i = −Γijk q̇ j q̇ k + F ij q̇ j
m
q̈θ = 0,
if q̇θ = e/m

since F ij is part of the Christoffel symbols for h.


To summarize this section on least time versus least action we can say that every
problem that we’ve discussed in classical mechanics can be regarded as geodesic motion!
Chapter 4

From Lagrangians to Hamiltonians

In the Lagrangian approach we focus on the position and velocity of a particle, and
compute what the particle does starting from the Lagrangian L(q, q̇), which is a function

L : T Q −→

where the tangent bundle is the space of position-velocity pairs. But we’re led to consider
momentum
∂L
pi = i
∂ q̇
since the equations of motion tell us how it changes
dpi ∂L
= i.
dt ∂q

4.1 The Hamiltonian Approach


In the Hamiltonian approach we focus on position and momentum, and compute what
the particle does starting from the energy

H = pi q̇ i − L(q, q̇)

reinterpreted as a function of position and momentum, called the Hamiltonian

H : T ∗ Q −→

where the cotangent bundle is the space of position-momentum pairs. In this approach,
position and momentum will satisfy Hamilton’s equations:

dq i ∂H dpi ∂H
= , =−
dt ∂pi dt ∂qi

51
52 From Lagrangians to Hamiltonians

where the latter is the Euler-Lagrange equation


dpi ∂L
=−
dt ∂qi
in disguise (it has a minus sign since H = pq̇ − L).
To obtain this Hamiltonian description of mechanics rigorously we need to study this
map
λ : T Q −→ T ∗ Q
(q, q̇) 7−→ (q, p)
where q ∈ Q, and q̇ is any tangent vector in Tq Q (not the time derivative of something),
and p is a cotangent vector in Tq∗ Q := (Tq Q)∗ , given by

λ ∂L
q̇ −→ pi =
∂ q̇ i
So λ is defined using L : T Q → . Despite appearances, λ can be defined in a coordinate-
free way, as follows (referring to Fig. 4.1). We want to define “ ∂∂L
q̇ i
” in a coordinate-free

Tq Q

TQ
T(q,q̇) Tq Q
PSfrag replacements

Tq Q
Q
q

Figure 4.1:

way; it’s the “differential of L in the vertical direction”—i.e., the q̇ i directions. We have
π : T Q −→ Q
(q, q̇) 7−→ q

and

dπ : T (T Q) −→ T Q
4.1 The Hamiltonian Approach 53

has kernel1 consisting of vertical vectors:

V T Q = ker dπ ⊆ T T Q

The differential of L at some point (q, q̇) ∈ T Q is a map from T T Q to , so we have



(dL)(q,q̇) ∈ T(q,q̇) T Q

that is,

dL(q,q̇) : T(q,q̇) T Q −→ .

We can restrict this to V T Q ⊆ T T Q, getting

f : V(q,q̇) T Q −→ .

But note
V(q,q̇) T Q = T (Tq Q)
and since Tq Q is a vector space,
Tq Q
T(q,q̇) Tq Q ∼
= Tq Q
PSfrag replacements V(q,q̇) T Q
in a canonical way2 . So f gives a linear map

p : Tq Q −→ T(q,q̇) Tq Q

that is,
q Q
p∈ Tq∗ Q

this is the momentum!

(Week 6, May 2, 4, 6.)

Given L : T Q → T ∗ Q, we now know a coordinate-free way of describing the map

λ : T Q −→ T ∗ Q
(q, q̇) 7−→ (q, p)

given in local coordinates by


∂L
pi = .
∂ q̇ i
1
The kernel of a map is the set of all elements in the domain that map to the null element of the
range, so ker dπ = {v ∈ T T Q : dπ(v) = 0 ∈ T Q}.
2
The fiber Tv V at v ∈ V of vector manifold V has the same dimension as V .
54 From Lagrangians to Hamiltonians

We say L is regular if λ is a diffeomorphism from T Q to some open subset X ⊆ T ∗ Q.


In this case we can describe what out system is doing equally well by specifying position
and velocity,
(q, q̇) ∈ T Q

or position and momentum


(q, p) = λ(q, q̇) ∈ X.

We call X the phase space of the system. In practice often X = T ∗ Q, then L is said to
be strongly regular.

4.2 Regular and Strongly Regular Lagrangians


This section discusses some examples of the above theory.

4.2.1 Example: A Particle in a Riemannian Manifold with Po-


tential V (q)
For a particle in a Riemannian manifold (Q, q) in a potential V : Q → has Lagrangian

1
L(q, q̇) = mgij q̇ i q j − V (Q)
2

Here
∂L
pi = = mij q̇ j
∂ q̇ i
so

λ(q, q̇) = q, mg(q̇, −)

so3 L is strongly regular in this case because

Tq Q −→ Tq∗ Q
v −→ g(v, −)

is 1-1 and onto, i.e., the metric is nondegenerate. Thus λ is a diffeomorphism, which in
this case extends to all of T ∗ Q.
3
The missing object there “−” is of course any tangent vector, not inserted since λ itself is an operator
on tangent vectors, not the result of the operation.
4.2 Regular and Strongly Regular Lagrangians 55

4.2.2 Example: General Relativistic Particle in an E-M Poten-


tial
For a general relativistic particle with charge e in an electromagnetic vector potential A
the Lagrangian is
1
L(q, q̇) = mgij q̇ i q j − eAi q̇ i
2
and thus
∂L
pi = i = mgij q̇ i q j + eAi .
∂ q̇
This L is still strongly regular, but now each map

λ |Tq Q : Tq Q −→ Tq∗ Q
q̇ 7−→ m g(q̇, −) + eA(q)

is affine rather than linear4 .

4.2.3 Example: Free General Relativistic Particle with Repa-


rameterization Invariance
The free general relativistic particle with reparameterization invariant Lagrangian has,
p
L(q, q̇) = m gij q̇ i q̇ j

This is terrible from the perspective of regularity properties—it’s not differentiable when
gij q̇ i q̇ j vanishes, and undefined when the same is negative. Where it is defined

∂L mgij q̇ j
pi = =
∂ q̇ i kq̇k

(where q̇ is timelike), we can ask about regularity. Alas, the map λ is not 1-1 where defined
since multiplying q̇ by some number has no effect on p! (This is related to the reparameter-
ization invariance—this always happens with reparameterization-invariant Lagrangians.)

4.2.4 Example: A Regular but not Strongly Regular Lagrangian


Here’s a Lagrangian that’s regular but not strongly regular. Let Q = and

L(q, q̇) = f (q̇)


4
All linear transforms are affine, but affine transformations include translations, which are nonlinear.
In affine geometry there is no defined origin. For the example the translation is the “+eA(q)” part.
56 From Lagrangians to Hamiltonians

so that
∂L
p= = f 0 (q̇)
∂ q̇
This will be regular but not strongly so if f 0 : → is a diffeomorphism from to some

proper subset U ⊂ . For example, take f (q̇) = eq̇ so f 0 : → (0, ∞) ⊂ . So

positive slope
L(q, q̇) = eq̇ /

or

slope between
p −1 and 1
L(q, q̇) = 1 + q̇ 2 /

and so forth.

4.3 Hamilton’s Equations


Now let’s assume L is regular, so

λ : T Q −→ X ⊆ T ∗ Q
(q, q̇) 7−→ (q, p)

This lets us have the best of both worlds: we can identify T Q with X using λ. This lets
us treat q i , pi , L, H, etc., all as functions on X (or T Q), thus writing

q̇ i (function on T Q)

for the function

q̇ i ◦ λ−1 (function on X)

In particular
∂L
ṗi := (Euler-Lagrange eqn.)
∂q i
which is really a function on T Q, will be treated as a function on X. Now let’s calculate:
∂L i ∂L i
dL = dq + i dq̇
∂q i ∂ q̇
= ṗi dq + pi dq̇ i
i
4.3 Hamilton’s Equations 57

while

dH = d(pi q̇ i − L)
= q̇ i dpi + pi dq̇ i − L
= q̇ i dpi + pi dq̇ i − (ṗi dq i + pi dq̇ i )
= q̇ i dpi − ṗi dq̇ i

so
dH = q̇ i dpi − ṗi dqi .
Assume the Lagrangian L : T Q → is regular, so

λ : T Q −→ X ⊆ T ∗ Q
(q, q̇) 7−→ (q, p)

is a diffeomorphism. This lets us regard both L and the Hamiltonian H = pi q̇ i − L as


functions on the phase space X, and use (q i , q̇ i ) as local coordinates on X. As we’ve seen,
this gives us

dL = ṗi dq i + pi dq̇ i
dH = q̇ i dpi − ṗi dq i .

But we can also work out dH directly, this time using local coordinates (q i , pi ), to get
∂H ∂H
dH = i
dpi + i dq i .
∂p ∂q
Since dpi , dq i form a basis of 1-forms, we conclude:

∂H ∂H
q̇ i = , ṗi = −
∂pi ∂qi
These are Hamilton’s Equations.

4.3.1 Hamilton and Euler-Lagrange


Though q̇ i and ṗi are just functions of X, when the Euler-Lagrange equations hold for
some path q : [t0 , t1 ] → Q, they will be the time derivatives of q i and pi . So when
the Euler-Lagrange
 equations hold, Hamilton’s equations describe the motion of a point
x(t) = q(t), p(t) ∈ X. In fact, in this context, Hamilton’s equations are just the Euler-
Lagrange equations in disguise. The equation
∂H
q̇ i =
∂pi
58 From Lagrangians to Hamiltonians

really just lets us recover the velocity q̇ as a function of q and p, inverting the formula
∂L
pi =
∂ q̇ i
which gave p as a function of q and q̇. So we get a formula for the map
λ−1 : X −→ T Q
(q, p) 7−→ (q, q̇).
Given this, the other Hamilton equation
∂H
ṗi = −
∂q i
is secretly the Euler-Lagrange equation
d ∂L ∂L ∂L
i
= i, or ṗ =
dt ∂ q̇ ∂q ∂q i
These are the same because
∂H ∂ i
 ∂L
= p i q̇ − L = − .
∂q i ∂q i ∂q i

Example: Particle in a Potential V (q)


n
For a particle in Q = in a potential V : n → the system has Lagrangian
m
L(q, q̇) = kq̇k2 − V (q)
2
which gives
p = mq̇
p g ij pj
q̇ = , (though really that’s q̇ = )
m m
and Hamiltonian
 
i 1 2 kpk2
H(q, p) = pi q̇ − L = kpk − −V
m 2m
1
= kpk2 + V (q).
2m
So Hamilton’s equations say
∂H p
q̇ i = ⇒ q̇ =
∂pi m
∂H
ṗi = − i ⇒ ṗ = −∇V
∂q
The first just recovers q̇ as a function of p; the second is F = ma.
4.3 Hamilton’s Equations 59

Note on Symplectic Structure


Hamilton’s equations push us toward the viewpoint where p and q have equal status as
coordinates on the phase space X. Soon, we’ll drop the requirement that X ⊆ T ∗ Q where
Q is a configuration space. X will just be a manifold equipped with enough structure to
write down Hamilton’s equations starting from any H : X → .
The coordinate-free description of this structure is the major 20th century contribution
to mechanics: a symplectic structure.
This is important. You might have some particles moving on a manifold like S 3 ,
which is not symplectic. So the Hamiltonian mechanics point of view says that the
abstract manifold that you are really interested in is something different: it must be a
symplectic manifold. That’s the phase space X. We’ll introduce symplectic geometry
more completely in later chapters.

4.3.2 Hamilton’s Equations from the Principle of Least Action


Before, we obtained the Euler-Lagrange equations by associating an “action” S with any
q : [t0 , t1 ] → Q and setting δS = 0. Now let’s get Hamilton’s equations directly by
assigning an action S to any path x : [t0 , t1 ] → X and setting δS = 0. Note: we don’t
impose any relation between p and q, q̇! The relation will follow from δS = 0.
Let P be the space of paths in the phase space X and define the action

S : P −→

by
Z t1
S(x) = (pi q̇ i − H)dt
t0

where pi q̇ i − H = L. More precisely, write our path x as x(t) = q(t), p(t) and let
Z t1  
d i 
S(x) = pi (t) q (t) − H q(t), p(t) dt
t0 dt

we write dtd q i instead of q̇ i to emphasize that we mean the time derivative rather than a
coordinate in phase space.
Let’s show δS = 0 ⇔Hamilton’s equations.
Z
δS = δ (pi q̇ i − H)dt
Z

= δpi q̇ i + pi δ q̇ i − δH dt
60 From Lagrangians to Hamiltonians

then integrating by parts,


Z

= δpi q̇ i − ṗi δq i − δH dt
Z 
∂H ∂H 
= δpi q̇ i − ṗi δq i − i δq i − δpi dt
∂q ∂pi
Z   
i ∂H  i
 ∂H 
= δpi q̇ − + δq −ṗi − i dt
∂pi ∂q
This vanishes ∀δx = (δq, δp) if and only if Hamilton’s equations
∂H ∂H
q̇ i = , pi = −
∂pi ∂q i
hold. Just as we hoped.
We’ve seen two principles of “least action”:
1. For paths in configuration space Q, δS = 0 ⇔Euler-Lagrange equations.
2. For paths in phase space X, δS = 0 ⇔Hamilton’s equations.
Additionally, since X ⊆ T ∗ Q, we might consider a third version based on paths in position-
velocity space T Q. But when our Lagrangian is regular we have a diffeomorphism λ :

T Q → X, so this third principle of least action is just a reformulation of principle 2.
However, the really interesting principle of least action involves paths in the extended
phase space where we have an additional coordinate for time: X × .
Recall the action
Z
S(x) = (pi q̇ i − H) dt
Z
dq i
= pi dt − H dt
dt
Z
= pi dq i − H dt

We can interpet the integrand as a 1-form


β = pi dq i − H dt
on X × , which has coordinates {pi , q i , t}. So any path
x : [t0 , t1 ] −→ X
gives a path
σ : [t0 , t1 ] −→ X ×
t 7−→ (x(t), t)
4.4 Waves versus Particles—The Hamilton-Jacobi Equations 61

and the action becomes the integral of a 1-form over a curve:


Z Z
i
S(x) = pi dq − H dt = β
σ

4.4 Waves versus Particles—The Hamilton-Jacobi Equa-


tions

(Week 7, May 9, 11, 13.)

In quantum mechanics we discover that every particle—electrons, photons, neutrinos,


etc.—is a wave, and vice versa. Interestingly Newton already had a particle theory of
light (his “corpuscules”) and various physicists argued against it by pointing out that
diffraction is best explained by a wave theory. We’ve talked about geometrized optics, an
approximation in which light consists of particles moving along geodesics. Here we start
with a Riemannian manifold (Q, g) as space, but we use the new metric

hij = n2 gij

where n : Q → (0, ∞) is the index of refraction throughout space (generally not a con-
stant).

4.4.1 Wave Equations


Huygens considered this same setup (in simpler language) and considered the motion of
a wavefront:

/
62 From Lagrangians to Hamiltonians

and saw that the wavefront is the envelope of a bunch of little wavelets centered at points
along the big wavefront:

 
  
balls of radius 

 
centered at points
of the old wavefront


  
 

In short, the wavefront moves at unit speed in the normal direction with respect to the
“optical metric” h. We can think about the distance function

d : Q × Q −→ [0, ∞)

on the Riemannian manifold (Q, h), where

d(q0 , q1 ) = inf (arclength)


Υ

where Υ = {paths from q0 to q1 }. (Secretly this d(q0 , q1 ) is the least action—the infimum
of action over all paths from q0 to q1 .) Using this we get the wavefronts centered at q0 ∈ Q
as the level sets
{q : d(q0 , q) = c}
or at least for small c > 0, as depicted in Fig. 4.2. For larger c the level sets can cease to

PSfrag replacements q0

Figure 4.2:

be smooth—we say a catastrophe occurs—and then the wavefronts are no longer the level
sets. This sort of situation can happen for topological reasons (as when the waves smash
into each other in the back of Fig. 4.2) and it can also happen for geometrical reasons
4.4 Waves versus Particles—The Hamilton-Jacobi Equations 63

Figure 4.3:

(Fig. 4.3). Assuming no such catastrophes occur, we can approximate the waves of light
by a wavefunction:
ψ(q) = A(q)eik d(q,q0 )
where k is the wavenumber of the light (i.e., its color) and A : Q → describes the am-
plitude of the wave, which drops off far from q0 . This becomes the eikonal approximation
in optics5 once we figure out what A should be.
Hamilton and Jacobi focused on distance d : Q × Q → [0, ∞) as a function of two
variables and called it W =Hamilton’s principal function. They noticed,
• /
q1 p 1
∂,
W (q0 , q1 ) = (p1 )i , •
∂q1i q0
where p1 is a cotangent vector pointing normal to the wavefronts.

4.4.2 The Hamilton-Jacobi Equations


We’ve seen that in optics, particles of light move along geodesics, but wavefronts are level
sets of the distance functions: kgc_[WSOJ
uo D=
z rke_YSLD: 72-
{ _SD 2q+ qq )
  {k q2q% q8q %  %!
! % % 1 qq   
%) + D {  
-2 3:D S_k z 
7= LSY_ekr 
DJ y
NSW[_cgkot
at least while the level sets remain smooth. In the eikonal approximation, light is described
by waves
ψ : Q −→ 
ψ(q1 ) = A(q1 )eik W (q0 ,q1 )
5
Eikonal comes form the Greek word for ‘image’ or ‘likeness’, in optics the eikonal approximation is
the basis for ray tracing methods.
64 From Lagrangians to Hamiltonians

where (Q, h) is a Riemannian manifold, h is the optical metric, q0 ∈ Q is the light source,
k is the frequency and
W : Q × Q −→ [0, ∞)
is the distance function on Q, or Hamilton’s principal function:

W (q0 , q1 ) = inf S(q)


q∈Υ

where Υ is the space of paths from q0 to q and S(q) is the action of the path q, i.e.,
its arclength. This is begging to be generalized to other Lagrangian systems! (At least
retrospectively with the advantage of our historical perspective.) We also saw that

• /
q1 p 1
,∂ W (q0 , q1 ) = (p1 )i , •
∂q1i q0

“points normal to the wavefront”—really the tangent vector

pi1 = hij (p1 )j

points in this direction. In fact kp1i is the momentum of the light passing through q1 .
This foreshadows quantum mechanics! (Note: in QM, the momentum is a derivative
operator—we get p by differentiating the wavefunction!)
Jacobi generalized this to the motion of point particles in a potential V : Q → ,
using the fact that a particle of energy E traces out geodesics in the metric
2(E − V )
hij = gij .
m
We’ve seen this reduces point particle mechanics to optics—but only for particles of fixed
energy E. Hamilton went further, and we now can go further still.
Suppose Q is any manifold and L : T Q → is any function (Lagrangian). Define
Hamilton’s principal function

W :Q× ×Q× −→

by
W (q0 , t0 ; q1 , t1 ) = inf S(q)
q∈Υ

where 
Υ = q : [t0 , t1 ] → Q, q(t0 ) = q0 , & q(t1 ) = q1
and Z t1  
S(q) = L q(t), q̇(t) dt
t0
4.4 Waves versus Particles—The Hamilton-Jacobi Equations 65

Now W is just the least action for a path from (q0 , t0 ) to (q1 , t1 ); it’ll be smooth if (q0 , t0 )
and (q1 , t1 ) are close enough—so let’s assume that is true. In fact, we have

• /
p
(q1 , t1 )1
∂ ,
W (q0 , q1 ) = (p1 )i , •
∂q1i (q0 , t0 )

where p1 is the momentum of the particle going from q0 to q1 , at time t1 , and

∂W
= −(p0 )i , (-momentum at time t0 )
∂q0i
∂W
= −H1 , (-energy at time t1 )
∂t1
∂W
= H0 , (+momentum at time t0 )
∂t0

(H1 = H0 as energy is conserved). These last four equations are the Hamilton-Jacobi
equations. The mysterious minus sign in front of energy was seen before in the 1-from,

β = pi dq i − H dt

on the extended phase space X × . Maybe the best way to get the Hamilton-Jacobi
equations is from this extended phase space formulation. But for now let’s see how
Hamilton’s principal function W and variational principles involving least action also
yield the Hamilton-Jacobi equations.
Given (q0 , t0 ), (q1 , t1 ), let
q : [t0 , t1 ] −→ Q

be the action-minimizing path from q0 to q1 . Then

W (q0 , t0 ; q1 , t1 ) = S(q)

Now consider varying q0 and q1 a bit

t0 t1
66 From Lagrangians to Hamiltonians

and thus vary the action-minimizing path, getting a variation δq which does not vanish
at t0 and t1 . We get

δW = δS
Z t1
=δ L(q, q̇) dt
t0
Z t1  
∂L i ∂L i
= δq + i δ q̇ dt
t0 ∂q i ∂ q̇
Z t1   t1
∂L i i

i
= δq − ṗ i δq dt + p i δq
t0 ∂q i t0
Z t1  
∂L
= i
− ṗi δq i dt
t0 ∂q

the term in parentheses is zero because q minimizes the action and the Euler-Lagrange
equations hold. So we δq i have

δW = p1i δq1i − p0i δq0i

and so
∂W ∂W
, and = −p0i
∂q1i ∂q0i
These are two of the four Hamilton-Jacobi equations! To get the other two, we need to
vary t0 and t1 :
Now change in
W will involve ∆t0

• and ∆t1
• •

t0 t1
t0 + ∆t0 t1 + ∆t1

(you can imagine ∆t0 < 0 in that figure if you like).


We want to derive the Hamilton-Jacobi equations describing the derivatives of Hamil-
ton’s principal function
W (q0 , t0 ; q1 , t1 ) = inf S(q)
q∈Υ

where Υ is the space of paths q : [t0 , t1 ] → Q with q(t0 ) = q, q(t1 ) = q1 and


Z t1
S(q) = L(q, q̇) dt
t0
4.4 Waves versus Particles—The Hamilton-Jacobi Equations 67

where the Lagrangian L : T Q → will now be assumed regular, so that


λT Q −→ X ⊆ T ∗ Q
(q, q̇) 7−→ (q, p)
is a diffeomorphism. We need to ensure that (q0 , t0 ) is close enough to (qi , t1 ) that there
is a unique q ∈ Υ that minimizes the action S, and assume that this q depends smoothly
on U = (q0 , t0 ; q1 , t1 ) ∈ (Q × )2 . We’ll think of q as a function of U :

(Q ×  )2 → Υ
q • (t1 , q1 )
u 7→ q
defined only when (q0 , t0 ) and
(t0 , q0 ) • (q1 , t1 ) are sufficiently close.

Then Hamilton’s principal function is


W (u) := W (q0 , t0 ; q1 , t1 ) = S(q)
Z t1
= L(q, q̇) dt
t0
Z t1  
= pq̇ − H(q, p) dt
t
Z 0t1
= p dq − H dt
t0
Z
= β
C

where β = pdq − H(q, p)dt is a 1-form on the extended phase space X × , and C is a
curve in the extended phase space:
 
C(t) = q(t), p(t), t ∈ X × .

Note that C depends on the curve q ∈ Υ, which in turn depends upon u = (q0 , t0 ; q1 , t1 ) ∈
(Q × )2 . We are after the derivatives of W that appear in the Hamilton-Jacobi relations,
so let’s differentiate Z
W (u) = β
C
with respect to u and get the Hamilton-Jacobi equations from β. Let us be a 1-parameter
family of points in (Q × )2 and work out
Z
d d
W (us ) = β
ds ds Cs
68 From Lagrangians to Hamiltonians

where Cs depends on us as above

C* s •
• . . . . ..
... . . .
.V ....... . . . . . . . . . . . . . ...........W Bs X×
.. . .. . . . .
. . .. .......... .
................. . ... .....
As

..... •
,
C0

Let’s compare Z Z Z Z Z
β and = β+ β+ β
Cs As +Cs +Bs As Cs Bs
Since C0 minimizes the action among paths with the given end-points, and the curve
As + Cs + Bs has the same end-points, we get
Z
d
β=0
ds As +Cs +Bs
(although As + Cs + Bs is not smooth, we can approximate it by a path that is smooth).
So Z Z Z
d d d
β= β− β at s = 0.
ds Cs ds Bs ds As
Note
Z Z
d d
β= β(A0r ) dr
ds As ds
= β(A00 )
where A00 = v is the tangent vector of As at s = 0. Similarly,
Z
d
β = β(w)
ds Bs
where w = B00 . So,
d
W (us) = β(w) − β(v)
ds
where w keeps track of the change of (q1 , p1 , t1 ) as we move Cs and v keeps track of
(q0 , p0 , t0 ). Now since β = pi dqi − Hdt, we get
∂W
= pi1
∂q1i
∂W
= −H
∂t1
4.4 Waves versus Particles—The Hamilton-Jacobi Equations 69

and similarly
∂W
= −pi0
∂q0i
∂W
=H
∂t0
So, if we define a wavefunction:

ψ(q0 , t0 ; q1 , t1 ) = eiW (q0 ,t0 ;q1 ,t1 )/~

then we get
∂ψ i
= − H1 ψ
∂t1 ~
∂ψ i
= p1 ψ
∂q1i ~

which at the time of Hamilton and Jacobi’s research was interesting enough, but nowadays
it is thoroughly familiar from quantum mechanics!
Bibliography

[Pel94] Peter Peldan. Actions for gravity, with generalizations: A review. Classical
and Quantum Gravity, 11:1087, 1994.

[WTM71] J. A. Wheeler, K. S. Thorne, and C. W. Misner. Gravitation. W. H. Freeman,


New York, 1971.

70

You might also like