This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

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

Week 1: (Mar. This approach seems more coherent.ucr. momentum and angular momentum. From Lagrangians to Hamiltonians. I’ve taught this course three times recently. (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. 22)—More example problems: (4) A special-relativistic charged particle in an electromagnetic ﬁeld in special relativity. 11. Fermat’s “principle of least time” in optics. (5) A general-relativistic free particle. In 2005 I started with the Lagrangian approach. 25. (3) A special-relativistic free particle: two Lagrangians. For reference. Examples of conserved quantities: energy. 15)—Example problems: (1) The Atwood machine.C. . Derek Wise took beautiful handwritten notes on the 2005 course. the result may already be helpful for people interested in the mathematics of classical mechanics. 28. with a heavy emphasis on action principles. The prehistory of the Lagrangian approach: D’Alembert’s “principle of least energy” in statics. to obtain a more textbook feel to these notes. Week 3 (Apr. (4) A special-relativistic charged particle in an electromagnetic ﬁeld. 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. and how D’Alembert generalized his principle from statics to dynamics using the concept of “inertia force”. Noether’s theorem on conserved quantities coming from symmetries. and derived the Hamiltonian approach from that. Week 2: (Apr. one with reparametrization invariance as a gauge symmetry. 4. the weekly lectures are outlined here. The ubiquity of geodesic motion. Blair Smith from Louisiana State University miraculously appeared and volunA teered to turn the notes into L TEX .edu/home/baez/classical/ Later. Week 5 (Apr. A The chapters in this L TEX version are in the same order as the weekly lectures. 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. 18. 20. While not yet the book I’d eventually like to write. 13. Kaluza-Klein theory. Riverside. Apr. 6. and sometimes split them over chapter. Twice I focused on the Hamiltonian approach.iii Preface These are notes for a mathematics graduate course on classical mechanics at U. which can be found on my website: http://math. 8)—Deriving the Euler-Lagrange equations for a particle on an arbitrary manifold. Week 4 (Apr. but I’ve merged weeks together. 27. 30. Generalized momentum and force. continued.

4. . A Poisson manifold that is not symplectic. a If you ﬁnd errors in these notes. Hamilton’s equations on a symplectic manifold. 3. Hamilton’s principal function and extended phase space. 5)—A taste of geometric quantization. Regular and strongly regular Lagrangians. The Hamiltonian version of Noether’s theorem. The cotangent bundle as phase space. Week 10 (June 1. Week 8 (May 16. 27)—Poisson brackets. Weil’s formula. Getting Hamilton’s equations directly from a least action principle. Liouville’s theorem. 13)—Waves versus particles: the Hamilton-Jacobi equation. 11. K¨hler manifolds. 25. 18. Poisson algebras and Poisson manifolds. The Schr¨dinger picture versus the o Heisenberg picture in classical mechanics. 20)—Towards symplectic geometry. Darboux’s theorem. 6)—From Lagrangians to Hamiltonians. Hamilton’s equations. Week 7 (May 9. Week 9 (May 23. The canonical 1-form and the symplectic 2-form on the cotangent bundle. please email me! I thank Sheeyun Park and Curtis Vinson for catching lots of errors. continued.iv Week 6 (May 2. How the Hamilton-Jacobi equation foreshadows quantum mechanics.

. . . .3. . . . . .2 Disk Pulled by Falling Mass . . . . . .3 Free Particle in Special Relativity . . . . .4. . 3. .1 Lagrangian and Newtonian Approaches . . . . . . . . . .1 The Atwood Machine . . . . . . . . . . . .1 Gauge Symmetry and Relativistic Hamiltonian . . . 3. . . . . . . . . . . . . . . . . . . . . .1 Canonical and Generalized Coordinates . . . . . . 3 Lagrangians and Noether’s Theorem 3. . . . . . . . . .3. . .Contents 1 From Newton’s Laws to Langrange’s Equations 1. . . . . . . . . . . . . . . . . . . . . . 3. . . . . . . . . . .4 Example Problems . . . . . . 3. . . 1. . . . . . . . . 3.4 How D’Alembert and Others Got to the Truth . . . . . . . . . . . .4. . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. . . .1 Comments . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . 3. .4. . . . . . . . . .1. . . . . .1. . . . . .1 Time Translation . . . 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v . . . . . . . . . . . . . . . . . 1.2. . . .1 The Euler-Lagrange Equations 2. . . . . . . .1 Time Translation Symmetry . . . . . .1 Lagrangian versus Hamiltonian Approaches . . . . . . 2 Equations of Motion 2. . . 3. . . . . . . . . . . .1. . . . . . . .5. . . . . . . . 3. . . . . . . . . . . . . . . . .2. . .2 Relativistic Hamiltonian . . .1. . . . .2 Prehistory of the Lagrangian Approach . . . . . . . . . . . . . . . . . . . . . . . .2. . . . .2 Lagrangian Dynamics . . . . . . . . . . . . . . . . . . . . . . .1 Noether’s Theorem . . . 1. . . 2 2 6 6 7 8 10 12 14 14 16 16 18 20 20 21 22 23 24 25 25 26 28 28 29 31 35 35 36 . . . . . . . . 3. . . . . . . . . . .3 Conserved Quantities from Symmetries . . 3. . . . . . . . . . .5 Electrodynamics and Relativistic Lagrangians .2 Space Translation Symmetry . . .3 The Principle of Least Time . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . . . . .2 Interpretation of Terms . . . . . . . 1.1 The Principle of Minimum Energy . . . . .5. . . . . . . 1. . . . . . 3. . . . . . . . . . . . . . . . . . . . . . . .2 D’Alembert’s Principle and Lagrange’s Equations 1. . . .2 Symmetry and Noether’s Theorem . . . . 2. . . . . .3 Rotational Symmetry . . . . 3. . . . . . .

. . . . . . . . . .1 Example: A Particle in a Riemannian Manifold with Potential V (q) 4. .3 Example: Free General Relativistic Particle with Reparameterization Invariance .8. . . . . . . .2. . . . . 4. . . . . . . . . . . . . . . . . . 3.3. . . . . . . . .6 Relativistic Particle in an Electromagnetic Field .2 Charged particle in EM Field in GR . . .1 Jacobi and Least Time vs Least Action . . . .1 The Hamiltonian Approach . . . . . . . .4. . . . . . 4.8 The General Relativistic Particle . . . . . . . . . . . . . . . . . .4 Example: A Regular but not Strongly Regular Lagrangian . . . . . . . . . . . . . . . .7. . . . . 4.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Hamilton’s Equations . . . . .4 Waves versus Particles—The Hamilton-Jacobi Equations . . . 3. . . . 4. . . . . . . . . . . . . . . . . . . . . . 3. . . . . . . . . 3. 3. . . . .8. . . . . .7 Alternative Lagrangians . .2 The Hamilton-Jacobi Equations . . . . . . . . .7. . . .9. . 4.CONTENTS 1 3. . . . 3. . . . . . . 4. . . . . 4. . . . . . . .2. . . . . . . . . . . . . . . . . . . . .2 Hamilton’s Equations from the Principle of Least Action . . . . . . 3. . . . 4. . . . . . . . . . . . . . . . . . 3.2 Regular and Strongly Regular Lagrangians . . . .9. . . . . 4. . . . .1 Lagrangian for a String . . .1 Free Particle Lagrangian in GR . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . .1 Wave Equations .2 Example: General Relativistic Particle in an E-M Potential . .9 The Principle of Least Action and Geodesics . . . . . .2 Alternate Lagrangian for Relativistic Electrodynamics 3. . . . . . . 37 39 39 41 43 44 45 46 46 48 51 51 54 54 55 55 55 56 57 59 61 61 63 4 From Lagrangians to Hamiltonians 4. . .2 The Ubiquity of Geodesic Motion . . . . 4.2. . . . . . . . . .1 Hamilton and Euler-Lagrange . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The astounding thing is that probably all professional applied physicists still use classical mechanics. atomic structure. superconductivity. a toy model if you will. so it deﬁnes a function. 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. Recall Newton’s law: F = ma (1. From this function we can deﬁne velocity. Its position.Chapter 1 From Newton’s Laws to Langrange’s Equations (Week 1. So.1) wherein we consider a particle moving in n . 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. and so forth). April 1. It used to be considered the sum total of our theoretical knowledge of the physical universe (Laplace’s daemon. what is classical mechanics? 1. the Newtonian clockwork). March 28.1 Lagrangian and Newtonian Approaches We begin by comparing the Newtonian approach to mechanics to the subtler approach of Lagrangian mechanics. depends on time t ∈ . So it is still an indispensable part of any physicist’s or engineer’s education. 30. v=q: ˙ −→ 2 n .) Classical mechanics is a very peculiar branch of physics. but now it is known as an idealization. say q. neutron stars. q : −→ n .

3) .e. Newton claimed that the particle satisﬁes F = ma.4) n . This is a 2nd-order diﬀerential equation for q : → n which will have a unique solution given some q(t0 ) and q(t0 ). t1 ] → n . This in turn is true iﬀ for some function V : n → . that is. dt and also acceleration. and goes down when you push it in the opposite direction.1 Lagrangian and Newtonian Approaches 3 where q = ˙ dq . We can then deﬁne a quantity called kinetic energy: K(t) := This quantity is interesting because d K(t) = m v(t) · a(t) dt = F (q(t)) · v(t) So. a=q: ¨ −→ Now let m > 0 be the mass of the particle.1. for all x ∈ n ).5) F =− V (1. kinetic energy goes up when you push an object in the direction of its velocity. |F (x)| < B for some B > 0. the integral of F along the curve q : [t0 . That is: m a(t) = F (q(t)) . 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(t1 ) = b iﬀ ×F = 0. Why? Because in this case we can deﬁne the total energy of the particle by E(t) := K(t) + V (q(t)) (1. and let F be a vector ﬁeld on force. A force with this property is called conservative. the change of kinetic energy is equal to the work done by the force. Moreover. provided the vector ﬁeld F is ‘nice’ — by which we technically ˙ mean smooth and bounded (i. n called the (1. t1 1 m v(t) · v(t) 2 K(t1 ) − K(t0 ) = = t0 t1 t0 F (q(t)) · v(t) dt F (q(t)) · q(t) dt ˙ So.. This function is then unique up to an additive constant. we call it the potential.2) (1.

let us look for curves (like the solid line in Fig. The other is to start with Newton’s principles and ﬁnd out what conditions. note that F = ma implies d [K(t) + V (q(t))] = F (q(t)) · v(t) + V (q(t)) · v(t) dt = 0. Conservative forces let us apply a whole bunch of cool techniques. constant as a function of time. if any. 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. q(t1 ) = b. To do this.1: A particle can sniﬀ out the path of least action.8) n with q(t0 ) = a. t1 ] → deﬁne the action to be t 1 (1. and derive Newton’s equations from that principle. One is to claim that nature causes particles to follow paths of least action. In the Lagrangian approach we deﬁne a quantity L := K(t) − V (q(t)) called the Lagrangian. we (1.4 From Newton’s Laws to Langrange’s Equations where V (t) := V (q(t)) is called the potential energy of the particle.6) S(q) := t0 L(t) dt From here one can go in two directions. bypass the philosophy. n qs + sδq PSfrag replacements q t0 t1 Figure 1. and then we can show that E is conserved: that is. 1. namely: d S(qs )|s=0 = 0 ds (1.7) . To see this.1) that are critical points of S. and for any curve q : [t0 . (because F = − V ). on S(q) follow from this. We will use the shortcut of hindsight.

The above result applies only for conservative forces. noting the boundary terms vanish because δq = 0 at t1 and t0 : d S(qs )|s=0 = ds t1 t0 [−m¨(t) − q V (q(t))] · δq(t)dt It follows that variation in the action is zero for all variations δq iﬀ the term in brackets is identically zero. this seems to be true of the most fundamental forces that we know of in our universe.1 Lagrangian and Newtonian Approaches 5 where qs = q + sδq for all δq : [t0 . n with. It is a simplifying assumption that has withstood the test of time and experiment. However. and ﬁrst noting that dqs /ds = δq(t) is the variation in the path: F = ma ⇔ d S(qs ) ds = s=0 d ds t1 t0 t1 t1 t0 1 mqs (t) · qs (t) − V (qs (t)) dt ˙ ˙ 2 = = t0 d 1 mqs (t) · qs (t) − V (qs (t)) dt ˙ ˙ ds 2 d mqs · qs (t) − ˙ ˙ ds d V (qs (t)) · qs (t) dt ds d d d d d qs (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. the path q is a critical point of the action S iﬀ F = ma. t1 ] → To show that d S(qs )|t=0 = 0 for all δq : [t0 . d S(qs ) ds t1 Note that = s=0 t0 mq · ˙ d δq(t) − dt V (q(t)) · δq(t) dt Then we can integrate by parts. −m¨(t) − V (q(t)) = 0 q So. forces that can be written as the minus the gradient of some potential. s=0 s=0 s=0 .1.e. i. δq(t0 ) = δq(t1 ) = 0. t1 ] → n with δq(t0 ) = δq(t1 ) = 0 (1..9) ds we start by using integration by parts on the deﬁnition of the action. that is.

And. so this is called the Principle of Least Action. We shall see this in more detail soon. The Hamiltonian approach focuses on energy.1. The key is to understand the integral of the Lagrangian over time – the ‘action’. ‘Hamiltonian’ is just a fancy word for energy. Why do we need two approaches. . Then we might ask. the action is minimized. In the simplest terms. Often. the Hamiltonian approach focuses on position and momentum.1 Lagrangian versus Hamiltonian Approaches I am not sure where to mention this. There are certainly some interseting things to learn from the question “why is action minimized?” First. but before launching into the history of the Lagrangian approach may be as good a time as any. the quantity exp(iS/ ). so energy can slosh back and forth between kinetic and potential forms. “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. will describe the ‘change in phase’ of a quantum system as it traces out this path. Suppose we did not have the hindsight aﬀorded by the Newtonian picture. note that total energy is conserved. Our ﬁrst task in understanding Lagrangian mechanics is to get a gut feeling for what the Lagrangian means. some people say that scientists should never ask ‘why’.6 From Newton’s Laws to Langrange’s Equations 1. The Lagrangian L = K − V is big when most of the energy is in kinetic form. Lagrangian and Hamiltonian? They both have their own advantages. which is a function of position and velocity. t1 S= t0 K − V dt so that slight changes in its path do not change the action (to ﬁrst order). The Lagrangian approach focuses on the Lagrangian. In short. while the Lagrangian approach takes a while to get used to. 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. it provides invaluable insights into classical mechanics and its relation to quantum mechanics. peeking ahead to quantum mechanics. though not always. where is Planck’s constant. 1. In later chapters we will describe another approach to classical mechanics: the Hamiltonian approach. which is a function of position and momentum — indeed. and small when most of the energy is in potential form.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. S. Indeed. while the Lagrangian approach focuses on position and velocity. We shall see that this describes the ‘total amount that happened’ from one moment to another as a particle traces out a path.

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. we’re being told that nature likes to minimize the total of ‘liveliness’ over time: that is. let us recall these earlier minimum principles. 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. used in optics. we can guess the principle of least action. Potential energy measures how much could happen. but isn’t yet — that’s what the word ‘potential’ means. and (ii) the principle of least time. and this will take a lot of action. the higher the potential energy the less lively. So. So. bad! Get this over with quick! Figure 1. it wants to spend a lot much time near the top of its trajectory. In other words. good! Spend as much time as possible here PSfrag replacements K − V big. The rock traces out a parabola. (Imagine a big rock sitting on top of a cliﬀ. Indeed. Statics is the study of objects at rest. if it spends too much time near the top of its trajectory. 1.1 The Principle of Minimum Energy Before physicists really got going in studying dynamical systems they used to study statics.2. On the other hand. By putting these together. to minimize its action.2: A particle’s “lazy” motion. consider the path of a thrown rock in the Earth’s gravitational ﬁeld. minimizes the action. There is another way to make progress on understanding ‘why’ action is minimized: history. Archimedes studied the laws of . Historically there were two principles that were fairly easy to deduce from observations of nature: (i) the principle of minimum energy used in statics.) So.2 Prehistory of the Lagrangian Approach 7 Kinetic energy measures how much is ‘happening’ — how much our system is moving around. . . On the one hand. 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. as in Fig. the total action. 1. with the potential to fall down. or in equilibrium.1. . nature is as lazy as possible! For example. since this is where the kinetic energy is least and the potential energy is greatest. and we can think of it as doing this in order K − V small here. it will need to really rush to get up there and get back down.2. .

0. Later D’Alembert understood this using his “principle of virtual work”. .8 From Newton’s Laws to Langrange’s Equations PSfrag replacements m1 L1 L2 m2 Figure 1. imal work done by this motion is zero! He also claimed that the work done on the ith body is. 1.. i.2. Now D’Alembert’s principle says that equilibrium occurs when the “virtual work” dW = dW1 +dW2 vanishes for all dθ (that is.2 D’Alembert’s Principle and Lagrange’s Equations Let’s go over the above analysis in more detail.e. This happens when m1 L 1 − m 2 L 2 = 0 dW2 = −m2 gL2 dθ which is just as Archimedes wrote. dWi = Fi dqi and gravity pulls down with a force mi g so. 0.4: A principle of energy minimization determines a lever’s balance. and he found that this would be in equilibrium if m1 L 1 = m 2 L 2 . He considered moving the lever slightly. 1. a see-saw or lever (Fig. I’ll try to make it clear what we mean by virtual work. −L1 dθ) = m1 gL1 dθ and similarly.3: A principle of energy minimization determines a lever’s balance. dWi = (0.3). −mg) · (0. inﬁnitesimally. for all possible inﬁnitesimal motions). He claimed that in equilibrium the inﬁnitesPSfrag replacements dq1 dθ dq2 Figure 1.

. For a particle in n . But. 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. Then because by deﬁnition the virtual displacements do not change the forces. The equilibrium will be stable if q0 is a local minimum of the potential V . S = t0 K − V dt critical points of S . Now call the work done by this hypothetical virtual displacement. we have equilibrium at a critical point of the potential. Fi · δri = 0. we must deduce that the virtual work vanishes for a system in equilibrium. i. 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. 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 . (when in equilibrium) (1. V (q0 ) = 0 that is. and yet remain consistent with all the constraints and forces at a given instant of time t. F i · δri .1.2 Prehistory of the Lagrangian Approach 9 The forces and constraints on a system may be time dependent. q(t) = q0 satisﬁes F = ma. without any time interval passing is called a virtual displacement. Statics equilibrium.e. the virtual work. (virtual work is zero for δq → 0) (no force on it!) If the force is conservative (F = − V ) then this is also equivalent to. D’Alembert’s principle simply says. we’ll postpone such sophistication for a while. We also have an analogy between statics and dynamics. n . when Fi = 0. Consider a system in the special state of being in equilibrium. To displace a system by δr i for each position coordinate.. V critical points of V Dynamics F = ma t1 action.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). So ultimately we need to study a particle on an arbitrary manifold. a = 0 potential. However. vanishes for all dq ∈ (it’s in equilibrium) dW = F · dq F = 0. We can summarize all the above by proclaiming that we have a “principle of least energy” governing stable equilibria.

But more interesting than straight lines are piecewise straight paths and curves. 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).3 The Principle of Least Time Time now to look at the second piece of history surrounding the principles of Lagrangian mechanics.11) − (1.5: A principle of energy minimization determines a lever’s balance.2.12) . Why is ABC the shortest path hitting n ∂T = Qi . ∂qi ∂L = 0. ∂qi (1. so the angle of incidence equals the angle of reﬂection. hints again that nature likes to minimize eﬀort. 1. In a vacuum light moves in straight lines. Consider reﬂection of light from a mirror. What path does the light take? The empirical answer was known since antiquity. which in Euclidean space is the minimum distance. In fact light traveling from A to B takes both the straight paths ABC and AC. d dt d dt ∂T ∂ qi ˙ ∂L ∂ 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. it chooses B such that θ1 = θ2 .10 From Newton’s Laws to Langrange’s Equations unstable equilibrium unstable equilibrium V PSfrag replacements stable equilibrium Figure 1. As well as hints from statics. there were also hints from the behavior of light.

such that. Snell (and predecessors) noted that each medium has some number n associated with it.1. it is also used in geophysics when one has a geological fault. The big clue leading to D’Alembert’s principle however came from refraction of light. B. C lie on a line ⇔ θ1 = θ2 Note the introduction of the ﬁctitious image C “behind” the mirror. θ1 medium 1 PSfrag replacements medium 2 θ2 called the index of refraction. n1 sin θ1 = n2 sin θ2 . 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).2 Prehistory of the Lagrangian Approach 11 A C PSfrag replacements θ1 B θ2 θ2 C the mirror? This follows from some basic Euclidean geometry: B minimizes AB + BC ⇔ B minimizes AB + BC ⇔ A. and in hydrodynamics when there is a boundary between two media (using mirror image sources and sinks).

Someone guessed the explanation. D’Alembert’s principle of virtual work for statics says that equilibrium occurs when D’Alembert generalized this to dynamics by inventing what he called the “inertia force”=−m a. then light will satisfy Snell’s law if the light minimizes the time it takes to get from A to C.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).2. not only is light the fastest thing around. not just the path distance. it’s also always taking the quickest path from here to there! 1.13) . for any function f on the space of paths we can deﬁne the variational derivative. realizing that if the speed of light in a medium is proportional to 1/n. One way to make good guesses is to generalize. So. But the same is true for the law of reﬂection. vanishes. We then take a variational path parameterized by s.14) ds s=0 Then D’Alembert’s principle of virtual work implies t1 t0 F (q0 ) · δq = 0. we get 0= t0 t1 − V (q) − m¨ · δq dt q − V (q) · δq + mqδ q dt ˙ ˙ = t0 n (1. Or symbolically when. In the case of refraction it is the time that is important. ∀δq ∈ F (q(t)) − ma(t) · δq(t) = 0 qs (t) = q(t) + s δq(t) F (q) − m¨ · δq dt = 0 q t1 for all δq. since in that case the path of minimum length gives the same results as the path of minimum time.12 From Newton’s Laws to Langrange’s Equations (having normalized n so that for a vacuum n = 1). and he postulated that in dynamics equilibrium occurs when the total force = F +inertia force. so if F = − V . d δf := f (qs ) (1. where δq(t0 ) = δq(t1 ) = 0 and with these paths.

which is called the centrifugal force. there’s something unsatisfying about the treatment so far.1. We use it.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. In general relativity. We only see that it’s necessary to obtain agreement with Newtonian mechanics (which is manifest in Eq. Still. If you are inside the rotating system and you throw a ball straight ahead it will appear to curve away from your target. one sees that — in a certain sense — gravity is an inertia force! .(1. 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. Recall from undergraduate physics that in an accelerating coordinate system there is a ﬁctional force = ma. to analyze simple physics in a rotating reference frame. S(q) = (K − V ) dt (1.13)). We do not really understand why one must introduce the ‘inertia force’. for example.2 Prehistory of the Lagrangian Approach 13 using d V (qs (t)) ds and δ(q 2 ) = ˙ then we have 0= dV dqs m dqs (t) ˙ dt + dq dt 2 ds t0 m d t1 d ˙ dt = − V qs (t) + qs (t) ds t0 dt 2 s=0 t1 d m =δ − V qs (t) + qs (t)2 dt ˙ dt 2 t0 − t1 t1 = s=0 dV dqs (t) dq ds s=0 dqs (t) ˙2 ds therefore δ t0 −V (q) + K dt = 0 so the path taken by the particle is a critical point of the action. This is an example of an inertia force which comes from using a funny coordinate system. We conclude with a few more words about this mystery.

April 4. 2. We’ll look at some speciﬁc examples of problem solving using the Euler-Lagrange equations. Q”. 2. and normally we call this the phase space.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 . 6. So our system is “a particle in Q = S2 × S2 PSfrag replacements Figure 2. For example.) In this chapter we’ll start to look at the Lagrangian equations of motion in more depth.1. 1 14 . So consider an arbitrary classical system as living in a space of points in some manifold Q. the space for a spherical double pendulum would look like Fig. 8.Chapter 2 Equations of Motion (Week 2. which means you have to disabuse yourself of the notion that we’re dealing with real The tangent bundle T Q will be referred to as conﬁguration space.1: Double pendulum conﬁguration space. later on when we get to the chapter on Hamiltonian mechanics we’ll ﬁnd a use for the cotangent bundle T ∗ Q. First we’ll show how the equations are derived. where Q = S 2 × S 2 .

2) can be rewritten δS = 0 (2. we have d S(qs ) ds We write.1) (2. L : T Q −→ and deﬁne the action. for simplicity). instead of “the particle taking a path”.1. q(t1 ) = b} (Γ is an inﬁnite dimensional manifold. d ds so Eq.(2. but we won’t go into that for now.2. t1 ] −→ Q and we deﬁne it’s velocity. We’ll just write is as q(t). It is then clear that when we say.) Let the Lagrangian=L for the system be any smooth function of position and velocity (not explicitly of time. “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. Sometimes to make this clear we’ll talk about “the system taking a path”. 2. q) dt ˙ The path that the quasi particle will actually take is a critical point of S.3) s=0 s=0 =0 as “δ (2. S: S : Γ −→ by t1 S(q) := t0 L(q.1 The Euler-Lagrange Equations 15 particles. So as time passes “the system” traces out a path q : [t0 . in accord with D’Alembert’s principle of least action. t1 ] → Q|q(t0 ) = a.2) . The single quasi-particle represents two real particles if we are talking about the classical system in Fig. ˙ ˙ Let Γ be the space of smooth paths from a ∈ Q to b ∈ Q. Γ = {q : [t0 . q(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 σ(t) = dq(s)/ds|s=t. In other words. we are not. we are dealing with a single quasi-particle in an abstract higher dimensional space. a path q ∈ Γ such that for any smooth 1-parameter family of paths qs ∈ Γ with q0 = q1 .

2 qs PSfrag replacements q0 Figure 2.1 Comments What is a “1-parameter family of paths”? Well. A cartoon of this looks like Fig. . let’s pick coordinates in a neighborhood U of some point q(t) ∈ Q. So the 1-parameter family is nothing more nor less than a set of well-deﬁned paths {qs }. Next. so it admits a covering of coordinate charts. 2.2: Schematic of a 1-parameter family of curves. 2. So a smooth 1-parameter family of paths will have q(s) everywhere inﬁnitesimally close to q(s + ) for an inﬁnitesimal hyperreal . or a 1D manifold.3 Then Q a PSfrag replacements U b Figure 2. we are just getting to that. 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.1.3: Local path variation. 2. We will now start to explore what δS = 0 means for our Lagrangian. consider only variations qs such that qs = q outside U . For now.2 Lagrangian Dynamics We were given that Q is a manifold. each one labeled by a parameter s.16 Equations of Motion 2.1. a path is a curve. So in Fig.

y) = (x1 . using the coordinates xi . (Velocity and position are in the abstract conﬁguration space Q). using the Leibniz rule d ∂L d ∂L ∂L δq = δq + δq ˙ dt ∂y dt ∂y ∂y we have. . further. xn . q) dt ˙ t0 t1 = = t0 ∂L i ∂L i δq + i δ q dt ˙ ∂xi ∂y where we’ve used the given smoothness of L and the Einstein summation convention for repeated indices i. The i are generalized position coordinates. . Using these coordinates we get t1 δS = δ t0 t1 L(q(t). t1 δS = t0 d ∂L ∂L − δq i (t) dt i i ∂x dt ∂y = 0. Continuing. . y) −→ dϕ(x. x2 . . in our case the manifold is M = Q. Let’s just go ahead and rename t0 and t1 as “ t0 and t1 ” to drop the primes. y n ) where y ∈ Tx Q. q(t)) dt ˙ δL(q. . xn ) and we also have coordinates for the 1-forms.1 The Euler-Lagrange Equations 17 we restrict attention to a subinterval [t0 . then it must vanish for all path variations δq. . . t1 ] ⊆ [t0 . t1 ] such that qs (t) ∈ U for t0 ≤ t ≤ t1 . y i on T U . . . the y i are the associated generalized velocity coordinates. We can use the coordinate charts on U . dϕ : T U −→ T n ∼ n × n = (x. If this integral is to vanish as demanded by δS = 0. and then we can describe it. ϕ : U −→ n x −→ ϕ(x) = (x1 . . 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 . L. the boundary terms vanish because we deliberately chose δq that . We restrict L : T M → to T U ⊆ T M. . . y 1 .2.

namely the x and y coordinates of the quantity. the terms “position” and “velocity” were used but were not assumed to be the usual kinematic notions that we are used to in physics. q(t). ˙ So in any case. but in fact it’s also suﬃcient. or ∂L d ∂L − i =0 (2. q(t) ∈ T U. Physicists always give the coordinates xi . So consider. physicists write. 2. When we write ∂V /∂q i = . d ∂L ∂L = i dt ∂ q i ˙ ∂q and they call these the Euler-Lagrange equations. y i on T U the symbols “q i ” and “q i ”. Now we’ll try to illuminate the E-L equations a bit by casting them into the usual position and velocity terms.2 Interpretation of Terms The derivation of the Euler-Lagrange equations above was fairly abstract.18 Equations of Motion vanish at the endpoints t0 and t1 inside U . That means the term in brackets must be identically zero. q) = mq · q − V (q) ˙ 2 1 = mq i qi − V (q) ˙ ˙ 2 TERMS ∂L ∂qi ˙ ∂L ∂q i MEANING (in this example) mq ˙ −( V )i MEANING (in general) the momentum pi the force Fi V we’re assuming the q i are Cartesian coordinates on Q. despite the fact that these ˙ i i also have another meaning.4) i dt ∂y ∂x This is necessary to get δS = 0. indeed the only reason we used those terms was for their analogical appeal. for all δq. Q= 1 ˙ ˙ L(q.

Simpliﬁcations generally mean quicker. Of course we can do all of our classical mechanics with Newton’s laws. 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.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. . One can of course introduce simpliﬁcations when solving Newton’s equations. Another good reason to learn Lagrangian (or Hamiltonian) mechanics is that it translates better into quantum mechanics. and they say that p=F ˙ (2. it’s just that it’s easier to do this when working with the Euler-Lagrange equations.2 Interpretation of Terms 19 So translating our example into general terms. 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. and then the Euler-Lagrange equations can be interpreted as equations relating generalized concepts of momentum and force. 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. if we conjure up some abstract Lagrangian then we can think of the independent variables as generalized positions and velocities.2. The main thing is that when we use symmetry to simplify the equations we are reducing the number of independent variables. shorter solutions and more transparent analysis or at least more chance at insight into the characteristics of the system.

demanding that this δS vanishes for all such variations δq is enough to imply the Euler20 → Q}. Nevertheless.1 Time Translation To handle time translations we need to replace our paths q : [t0 . dt s=0 . Moreover. t1 ] → Q by paths q : → Q. if δq = 0 outside of some ﬁnite interval. Γ = {q : The bad news is that the action S(q) = ∞ −∞ L q(t). The rigorous statement of all this is the content of Noether’s theorem. So symmetries of a dynamical system give conserved quantities or conservation laws. Invariance under a group of transformations is precisely what we mean by a symmetry in group theory. δS := ∞ −∞ d L qs (t). then the functional variation. 3. qs (t) ˙ ds will converge. since the integral is smooth and vanishes outside this interval. Time and space translations are examples of 1-parameter groups of transformations. 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). so S is then no longer a function of the space of paths.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. q(t) dt ˙ typically will not converge. and then deﬁne a new space of paths.

dt ∂ qi ˙ ∂qi Recall that. is the force. . these are the {qi . and at each point a position and a momentum). let’s spend a few moments clearing up some terminology (I hate using jargon. ˙ ∂qi Note the similarity to Hamilton’s equations—if you change L to H you need to stick in a minus sign.1 Time Translation 21 Lagrange equations: δS = = = ∞ −∞ ∞ −∞ ∞ −∞ d L qs (t). or system of particles. The qi are generalized positions.1 Canonical and Generalized Coordinates In light of this noted similarity with the Hamilton equations of motion. so for example. qi }. Generalized Coordinates For Lagrangian mechanics we have been using generalized coordinates. and the qi are generalized velocities. the ﬁrst term in ∂L i d ∂L d ∂L i δq = ˙ δq − δq i i ∂q ˙ dt ∂ q ˙ dt ∂ q i ˙ vanishes when we integrate. is the generalized momentum. These can be in any reference frame or system of axes. The full ˙ ˙ set of independent generalized coordinates represent the degrees of freedom of a particle. by the E-L eqns. To be explicit. but sometimes it’s unavoidable. by defn.1. and sometimes it can be eﬃcient—provided everyone is clued in).3. ∂L = pi . qs (t) ˙ dt ds s=0 ∂L ∂L δqi + δ qi dt ˙ ∂qi ∂ qi ˙ ∂L d ∂L − δqi dt ∂qi dt ∂ qi ˙ where again the boundary terms have vanished since δq = 0 near t = ±∞. ˙ ˙ 3. Then the whole thing vanishes for all compactly supported smooth δq iﬀ ∂L d ∂L = . So if we have N particles then we’d typically have 6N generalized coordinates (the “6” is for 3 space dimensions. ∂ qi ˙ ∂L = pi . and change variables from q to pi and eliminate pi .

22 Lagrangians and Noether’s Theorem in a Cartesian frame. y2 . {u1 . The canonical coordinates are a special set of coordinates on the cotangent bundle of the conﬁguration space manifold Q. matching the total degrees of freedom as claimed. form a coordinate system on the cotangent bundle T ∗ Q of the conﬁguration space Q. with two particles. We will not discuss this here. By judicious choice of coordinate frame we can eliminate one velocity component and one position component for each particle. They are usually written as a set of (q i . x2 . one can then deﬁne new positions and momenta with a consequent reduction in the number of these generalized coordinates needed to describe the system. w1 . and 2 × 3 = 6 velocities. 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. hence these coordinates are called the canonical coordinates. It turns out that the q i together with the pj . 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. We want to refer to symmetry transformations (of the Lagrangian) governed by a single parameter. v1 . q) −→ qs .f. in 3D space we’d have the 2 × 3 = 6 position coordinates. maybe the particles move on a sphere for example. 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.o. z2 }. overall. with q0 = q .1 (one-parameter family of symmetries). Deﬁnition 3. which are 1-forms in the cotangent bundle at the point q in the manifold. v2 . let’s give a useful deﬁnition that will make it easy to refer to a type of dynamical system symmetry. w2 } where say u = vx . u2 . It is also handy to respect other symmetries of a system. v = vy . z1 . w = vz are the Cartesian velocity components. A 1-parameter family of symmetries of a Lagrangian system L : T Q → is a smooth map. F : × Γ −→ Γ (s.2 Symmetry and Noether’s Theorem First. y1 . {x1 . pj ) or (xi . 3. and so 4N d. but if you care to know. This makes 12 = 6 × 2 = 6N coordinates. pj ) with the x’s or q’s denoting the coordinates on the underlying manifold and the p’s denoting the conjugate momentum.

1 (Noether’s Theorem). But δL = dt is a sneaky generalization whose usefulness will become clear. In other words. L : T Q → .2.3.. q(t) ˙ =0 “Okay. Suppose F is a one-parameter family of symmetries of the Lagrangian system. qs (t) ˙ ds for all paths q.2 Symmetry and Noether’s Theorem 23 such that there exists a function (q.1. that is.e. in boring detail.1 Noether’s Theorem in this theorem is the function associated Here’s a statement of the theorem. Theorem 3. it’s time derivative is zero for any path q ∈ Γ satisfying the EulerLagrange equations. q) for which ˙ δL = for some : T Q → . a symmetry of L in the d most obvious way. d L qs (t). that is. big deal” you might say. = s=0 d dt d dt 3. Then we’d need to ﬁnd out what this pi δqi − business is that is conserved. So let’s look at some examples. Then. Before this can be of any use we’d need to ﬁnd a symmetry F . d ∂L d i q(s)q(s) ˙ q (t) dt ∂y i ds s Proof. . Note that with F in deﬁnition 3. in which case we really have a way of moving paths around (q → qs ) that doesn’t change the Lagrangian—i. is conserved. Remark: The simplest case is δL = 0. d pi δq i − dt = pi δq i + pi δ q i − ˙ ˙ = d dt − q(t). qs (t) ˙ pi δqi − s=0 ∂L i ∂L i δq + i δ q − δL ˙ ∂q i ∂q ˙ = δL − δL = 0. qs (t).

this time with a speciﬁc Lagrangian. For example. if it has 1-parameter families of symmetries then it’ll have conserved quantities.24 Lagrangians and Noether’s Theorem Example 1. and the left-hand side is K + V . So we have a one-parameter family of symmetries qs (t) = q(t + s) This indeed gives. It doesn’t matter what the Lagrangian is. ˙ δL = L for d L(qs ) ds so here we take = s=0 d ˙ L=L dt = L simply! We then get the conserved quantity pi δq i − = pi q i − L ˙ which we normally call the energy. Let’s repeat this example.) 3. at least not to any extent that we can tell). if Q = 1 ˙ L = mq 2 − V (q) 2 then this quantity is mq · q − ˙ ˙ 1 mq · q − V ˙ ˙ 2 1 = mq 2 + V (q) ˙ 2 The term in parentheses is K − V . The trick in physics is to write down a correct Lagrangian in the ﬁrst place! (Something that will accurately describe the system of interest. 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.3 Conserved Quantities from Symmetries Fs : Γ −→ Γ q −→ qs We’ve seen that any 1-parameter family n . and if . guaranteed.

) 1 For example: a particle on n in a potential V has Q = n . L(q.3. if you are doing Newtonian mechanics then “total energy” is appropriate. This has spatial translation ˙ 2 n we have the symmetry symmetries. q) = 2 mq 2 − V (q). ˙ ˙ This system has ∂L pi q i = i q i = mq 2 = 2K ˙ ˙ ˙ ∂q ˙ so H = pi q i − L = 2K − (K − V ) = K + V ˙ as you’d have hoped.3.3. L : T Q → .3 Conserved Quantities from Symmetries 25 which satisﬁes δL = ˙ for some function = (q. q) gives a conserved quantity ˙ pi δq i − As usual we’ve deﬁned δL := d L qs (t). we have a 1-parameter family of symmetries qs (t) = q(t + s) because ˙ δL = L so we get a conserved quantity called the total energy or Hamiltonian. 3. we have Q = n and L = K = 1 mq 2 . H = pi q i − L ˙ (3.1 Time Translation Symmetry For any Lagrangian system.1) (You might prefer “Hamiltonian” to “total energy” because in general we are not in the same conﬁguration space as Newtonian mechanics.2 Space Translation Symmetry For a free particle in n . qs (t) ˙ ds s=0 Let’s see how we arrive at a conserved quantity from a symmetry. so that for any v ∈ qs (t) = q(t) + s v . 3.

here it is a conserved quantity derived from space translation invariance. It is often useful therefore.3. Consider any X ∈ so(n) (that is a skew-symmetric n × n matrix). These two diﬀerent ˙ “momentum’s” happen to be the same in this example! Since this is conserved for all v we say that mq ∈ n is conserved. that is. hence conserved quantities. and we’re getting a q(= n )-valued conserved quantity!) 3.26 Lagrangians and Noether’s Theorem with δL = 0 because δ q = 0 and L depends only on q not on q in this particular case. then for all s ∈ the matrix esX is in SO(n). but earlier it was a diﬀerent thing. (Since L does not ˙ ˙ i i depend upon q we’ll call q an ignorable coordinate. these ignorables always give symmetries. and ∂L/∂ q i = pi . and ˙ δqi = ˙ d i q ˙ ds s s=0 d d sX = e q ds dt d = Xq dt =Xq ˙ s=0 . This gives a 1-parameter family of symmetries qs (t) = esX q(t) which has δL = ∂L i ∂L i δq + i δ q = mqi δ q i ˙ ˙ ˙ ∂q i ∂q ˙ now qi is ignorable and so ∂L/∂q i = 0.3 Rotational Symmetry The free particle in n also has rotation symmetry. as above. namely the momentum ∂L/∂ q i = pi conjugate to q i . it describes a rotation. 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 = mqi v i = mq · v ˙ ˙ Aside: Note the subtle diﬀerence between two uses of the term “momentum”. (In fact that ˙ n whole Lie group G = is acting as a translation symmetry group.

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. without changing its magnitude. We’d like to know how to solve more complicated dynamics. ˙ mq×q Note that above we have assumed one can construct a basis for so(n) using matrices of the form assumed for X. pi δq i = mqi · (X q)i ˙ (δq i = Xq just as δ q i = X q in our previous calculation).3. But how do we do that? All this theory is ﬁne but is useless unless you know how to apply it. If n = 3 we write these as. for which the Lagrangian is just the kinetic energy. a free particle. we can then compare our answers with experiments.3 Conserved Quantities from Symmetries 27 So. 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. . I mentioned earlier that we can do mechanics with any Lagrangian. the angular momentum in the X direction. then we get m(qi q j − qj q i ) ˙ ˙ the “ij component of angular momentum”. instead we plug in some assumed physics and use the Lagrangian approch to solve the system of equations. i. or if X has zero entries except ˙ ˙ in ij and ji positions.e. (Note: this whole bunch of math above for δL just says that the kinetic energy doesn’t change when the velocity is rotated. skew symmetric with ±1 in the respectively ij and ji elements. The above examples were for a particularly simple system. So we get a conserved quantity. since there is no potential energy variation for a free particle. L=K−V So we are not using Lagrangians directly to tell us what the fundamental physical laws should be.) We write.. δL = mqi Xj q j ˙ i˙ = mq · (X q) ˙ ˙ =0 since X is skew symmetric as stated previously (X ∈ so(n)). If we like. where it’s ±1. but if we want to be useful we’d better pick a Lagrangian that actually describes a real system. otherwise zero.

11. ) x ).) To see how the formalisms in this chapter function in practise. and allows us to easily use any coordinates we wish. ) (we could use the “owns” symbol (x.4 Example Problems (Week 3. hanging from it. Moreover T Q = (0. The momentum is: ∂L p= = (m1 + m2 )x ˙ ∂x ˙ and the force is: ∂L = (m1 − m2 )g F = ∂x . 15. 13. It’s vastly superior to the simplistic F = ma formulation of mechanics. Apr.1 The Atwood Machine A frictionless pulley with two masses. x). We have PSfrag replacements x m1 m2 −x 1 d 1 K = (m1 + m2 )( ( − x))2 = (m1 + m2 )x2 ˙ 2 dt 2 V = −m1 gx − m2 g( − x) so 1 L = K − V = (m1 + m2 )x2 + m1 gx + m2 g( − x) ˙ 2 The conﬁguration space is Q = (0. m1 and m2 . ). As usual L : T Q → . The Lagrangian formulation allows the conﬁguration space to be any manifold. ) × Note that solutions of the Euler-Lagrange equations will only be deﬁned for some time t ∈ .4. as eventually the solutions reaches the “edge” of Q. 3. ˙ here and write Q = (0. and x ∈ (0.28 Lagrangians and Noether’s Theorem 3. let’s do some problems.

) × S 1 × × 1 1 d ˙ K = m1 (r 2 + r 2 θ 2 ) + m2 ( ( − r))2 ˙ 2 2 dt V = gm2 (r − ) 1 1 ˙ ˙ ˙ L = m1 (r 2 + r 2 θ 2 ) + m2 r 2 + gm2 ( − r) 2 2 is constant so d/dt( − r) = −r. and x = g if m2 = 0. Note that x = 0 ˙ ¨ when m1 = m2 . minus it’s center = (0. m1 +m2 It is trivial to integrate the expression for x twice (feeding in some appropriate initial ¨ conditions) to obtain the complete solution to the motion x(t) and x(t). r m1 PSfrag replacements −r m2 no swinging allowed! Here Q = open disk of radius .3. θ) T Q = (0. and it can thus whirl around the hole to which it is tethered to the mass below. For the momenta we get.4 Example Problems 29 The Euler-Lagrange equations say p=F ˙ (m1 + m2 )¨ = (m1 − m2 )g x m1 − m 2 x= ¨ g m1 + m 2 So this is like a falling object in a downwards gravitational acceleration a = m1 −m2 g. θ. ) × S 1 (r. r. ˙ ∂θ pr = having noted that (r. ¨ 3. The disk is free to move on a frictionless surface. ˙ ∂L = (m1 + m2 )r ˙ ∂r ˙ ∂L ˙ pθ = = m1 r 2 θ.2 Disk Pulled by Falling Mass Consider next a disk pulled across a table by a falling mass. θ) ˙ ˙ .4.

At r = r0 . 3. ) of mass m = m1 + m2 feeling a force Fr = J2 − m2 g m1 r 3 which could come from an “eﬀective potential” V (r) such that dV /dr = −Fr . pushing m1 radially out. and pθ . At that time the disk reaches the hole. 3.1 ˙ If θ(t = 0) = 0 then there is no centrifugal force and the disk will be pulled into the hole until it gets stuck. . So. the Euler-Lagrange equations give. pr = F r . so then we’ve hit the boundary of Q and our solution is broken. which is topologically the center of the disk that has been removed from Q. is conserved. the minimum of V (r). (θ is ignorable) ∂θ Fr = ˙ Note: in Fr the term m1 r θ 2 is recognizable as a centrifugal force. ∂L ˙ = m1 r θ 2 − gm2 ∂r ∂L Fθ = = 0. ˙ Let’s use our conservation law here to eliminate θ from the ﬁrst equation: ˙ θ= so (m1 + m2 )¨ = r J2 − m2 g m1 r 3 J m1 r 2 Thus eﬀectively we have a particle on (0. The forces are. So integrate −Fr to ﬁnd V (r): J2 + m2 gr V (r) = 2m1 r 2 this is a sum of two terms that look like Fig. ˙ ˙ (m1 + m2 )¨ = m1 r θ 2 − m2 g r ˙ pθ = m1 r 2 θ = J = a constant. the conjugate momentum. our disc mass m1 will be in a stable circular orbit of radius r0 (which depends upon J). rotational symmetry.2. while the term −gm2 is gravity pulling m2 down and thus pulling m1 radially in. ˙ pθ = 0.30 Lagrangians and Noether’s Theorem Note that θ is an “ignorable coordinate”—it doesn’t appear in L—so there’s a symmetry. Otherwise we get orbits like Fig.

3.4 Example Problems

31

PSfrag replacements

V (r)

attractive gravitational potential repulsive centrifugal potential

r0 no swinging allowed!

r

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 (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

ηµν

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 (x0 , . . . , xn ) with the metric ηµν deﬁned above. Then the path of the particle is, q : ( 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 , q 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 independent of the parameterization of the path—since the parameterization is “unphysical” (it can’t be measured). So the action

t1

=

1 0 0 ... 0 0 −1 0 . . . 0 0 0 −1 0 . . . .. . . . . . . . 0 0 0 . . . −1

S(q) =

t0

˙ L q i (t), q i (t) dt

**for q : [t0 , t1 ] → Q, should be independent of t. The obvious candidate for S is mass times arclength,
**

t1

S=m

t0

ηij q i (t)q j (t) dt ˙ ˙

or rather the Minkowski analogue of arclength, called proper time, at least when q is a ˙ timelike vector, i.e., ηij q i q j > 0, which says q 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 ˙ ˙ (3.2) L = m ηij q i q j and work out the Euler-Lagrange equations. We have pi = ∂ ∂L = m i ηij q i q j ˙ ˙ i ∂q ˙ ∂q ˙ 2ηij q j ˙ =m 2 ηij q i q j ˙ ˙ j ηij q ˙ mqi ˙ =m = iqj q ˙ ηij q ˙ ˙

(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 q → αq. The Euler-Lagrange equations say, ˙ ˙ pi = F i = ˙ ∂L =0 ∂q i

The meaning of this becomes clearer if we use “proper time” as our parameter (like parameterizing a curve by its arclength) so that

t1 t0

q dt = t1 − t0 , ˙

∀ t0 , t1

which ﬁxes the parametrization up to an additive constant. This implies q = 1, so that ˙ pi = m and the Euler-Lagrange equations say pi = 0 ⇒ m¨i = 0 ˙ q 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 qi ˙ = mqi ˙ q ˙

e. Consider any (smooth) 1parameter family of reparameterizations. i. (when E-L eqns. This is in fact called “the problem of time” in general relativity. These reparameterization symmetries work as follows... fs : → → Q} as follows: given any ¢ £¡ . hold) so that Noether’s theorem gives a conserved quantity pi δq i − Here we go then: δL = ∂L i ∂L i δq + i δ q ˙ ∂q i ∂q ˙ = pi δ q i ˙ mqi d i ˙ = q fs (t) ˙ q ds ˙ s=0 mqi d d i ˙ = q fs (t) q dt ds ˙ s=0 fs (t) mqi d i ˙ q fs (t) ˙ = q dt ˙ ds s=0 mqi d i ˙ = q δfs ˙ q dt ˙ d = pi q i δf ˙ dt = pi q i δf . dt pi = 0. i.34 Lagrangians and Noether’s Theorem particle. ˙ d where in the last step we used the E-L eqns.e. Here we see it starting to show up in special relativity. These act on the space of paths Γ = {q : q ∈ Γ we get qs (t) = q fs (t) where we should note that qs is physically indistinguishable from q. diﬀeomorphisms fs : −→ with f0 = . Let’s show that δL = ˙. so δL = ˙ with So to recap a little: we saw the free relativistic particle has L=m q =m ˙ ηij q i q j ˙ ˙ and we’ve considered reparameterization symmetries qs (t) = q fs (t) .

Given a Lagrangian we derive equations of motion from the Euler-Lagrange equations. via Noether’s theorem. . realizing that a given system may not have a unique Lagrangian but will often have an obvious natural Lagrangian. dt and set pi q i δf = ˙ so Noether’s theorem gives a conserved quantity pi δq i − = pi q i δf − pi q i δf ˙ ˙ =0 So these conserved quantities vanish! In short. 3. We use principles of least action to conjure up Lagrangians for our systems. in particular Hamiltonians from time translation invariance. These are symmetries that permute diﬀerent mathematical descriptions of the same physical situation—in this case reparameterizations of a path. Symmetries of L guide us in ﬁnding conserved quantities.5. and ∂L/∂ q i = p) ˙ ˙ = pi δ q i ˙ d = pi δq i dt d = pi q i δf ˙ dt d ˙ = pi q i δf.5 Electrodynamics and Relativistic Lagrangians 35 we’ve used the fact that δq i := so (repeating a bit of the above) δL = d i q fs (t) ds = q i δf ˙ s=0 ∂L i ∂L i δq + i δ q ˙ ∂q i ∂q ˙ = pi δ q i . we’re seeing an example of what physicists call gauge symmetries.3. This section also introduces gauge symmetry. (since ∂L/∂q i = 0.1 Gauge Symmetry and Relativistic Hamiltonian What are gauge symmetries? 1. 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. and this is where we begin.

3.4.5.2 Relativistic Hamiltonian What then is the Hamiltonian for special relativity theory? We’re continuing here with the example problem of §3. so we often use (3) to distinguish gauge symmetries from physical symmetries. and so ˙ ˙ ˙ i δL = d /dt. (Because it means you might have to solve the static equations for the 4D universe as a whole. (since ∂L/∂q i = 0 and ∂L/∂ q i = pi ) ˙ ˙ = pi 0 = 0 . H = pi δ q i − where under q(t) → q(t + s) we have δ q i = q i δf . and that’s impossible!) But there is another conserved quantity deserving the title of “energy” which is not zero.3. Well. and (3) is easy to test. which implies = pi δq . so we see from the previous results that the Hamiltonian is zero! H = 0. These symmetries give conserved quantities that work out to equal zero! Note that (1) is a subjective criterion. Now you know why people talk about “the problem of time” in general relativity theory. 3. the Hamiltonian comes from Noether’s theorem from time translation symmetry. In fact any vector w gives a conserved quantity. it’s glimmerings are seen in the ﬂat Minkowski spacetime of special relativity. qs (t) = q(t) + s w where w ∈ n+1 and w points in some timelike direction. The result H = 0 follows. 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. and it comes from the symmetry. These symmetries make it impossible to compute q(t) given q(0) and q(0): since if ˙ q(t) is a solution so is q(f (t)) for any reparameterization f : → . qs (t) = q(t + s) and this is an example of a reparametrization (with δf = 1). δL = ∂L i ∂L i δq + i δ q ˙ ∂q i ∂q ˙ = pi δ q i . You may think it’s nice and simple to have H = 0. (2) and (3) are objective. Explicitly. We have a high degree of non-uniqueness of solutions to the Euler-Lagrange equations.36 Lagrangians and Noether’s Theorem 2.

This is our ˙ from Noether’s theorem with ˙ ˙ Noether’s theorem says that we get a conserved quantity pi δq i − = pi w = 0. so namely. We’ve just about exhausted all the basic stuﬀ that we can learn from the free particle. An dxn ).3.3) is a 2-form containing the electric and magnetic ﬁelds. . . . We know p = 0 from the Euler-Lagrange ˙ equations. pn ) is spatial momentum. such that dA = F (3. F µν = We’d write (for Q having local charts to ∂Ai ∂Aj − ∂xj ∂xi n+1 (3.6 Relativistic Particle in an Electromagnetic Field The electromagnetic ﬁeld is described by a 1-form A on spacetime. . for our free particle. . A is the vector potential. p1 .4) A = A0 dx0 + A1 dx1 + . . .6 Relativistic Particle in an Electromagnetic Field 37 qs q PSfrag replacements w since δq i = w i . . 3. pn ) p0 is energy. the momentum in the w direction. . (p1 . p =(p0 . . but here we see it coming from spacetime translation symmetry. . δ q i = w i = 0. So next we’ll add some external force via an electromagnetic ﬁeld.

April 18. 22. q) = m q + eAi q i ˙ ˙ ˙ (3. ˙ t0 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 }). ˙ The Lagrangian in the above action. . but one can also write the path integral using time t as a parameter. 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. . dAi = ∂µ Ai dxµ using the summation convention and ∂µ := ∂/∂xµ . agrees with the matrix expression below (at least for 4D). . with Ai q i dt the diﬀerential. for a charge e with mass m in an electromagnetic potential A is L(q. for example.) Note that since A is a 1-form we can integrate it over an oriented manifold..38 Lagrangians and Noether’s Theorem and then because d2 = 0 dA = dA0 dx0 + dA1 dx1 + . So. (Week 4. ˙ i i after dq = q dt. 20. . The action for a particle of charge e is t1 S=m t0 q dt + e ˙ q A here t1 q dt = proper time. The student can easily check that the components for F = F01 dx0 ∧ dx1 + F02 dx0 ∧ dx2 + . dAn dxn and since the “Ai ” are just functions.5) so we can work out the Euler-Lagrange equations: pi = qi ˙ ∂L + eAi =m i ∂q ˙ q ˙ = mvi + e Ai .

F = dA. So we get the following equations of motion m dvi = eF ij q j . we ﬁnd the force Fi = ∂L ∂ = i e Aj q j ˙ i ∂q ∂q ∂Aj = e i qj ˙ ∂q So the Euler-Lagrange equations say (noting that Ai = Aj q(t) : p=F ˙ ∂Aj d ˙ mvi + eAi = e i q j dt ∂q ∂Aj dAi dvi = e i qj − e ˙ m dt ∂q dt ∂Aj dvi ∂Ai = e i qj − e j qj m ˙ ˙ dt ∂q ∂q ∂Ai ∂Aj =e − j qj ˙ i ∂q ∂q the term in parentheses is F ij = the electromagnetic ﬁeld. ˙ dt (Lorentz force law) (3.3.1 Lagrangian for a String So we’ve looked at a point particle and tried S = m · (arclength) + A . normalized so that v = 1. First a paragraph on objects other than point particles! 3. Continuing the analysis. Note that now momentum is no longer mass times velocity! That’s because we’re in n + 1-d spacetime.6) (Usually called the “Lorenz” force law.7 Alternative Lagrangians We’ll soon discuss a charged particle Lagrangian that is free of the reparameterization symmetry.7 Alternative Lagrangians 39 where v ∈ n+1 is the velocity. the momentum is an n + 1-vector.) 3.7.

3). But we haven’t checked that this action yields sensible dynamics yet! But supposing it does. To recover electrodynamic interactions it should be antisymmetric like A. Can you infer an appropriate action for this system? Remember.(3. In string theory we boost the dimension by +1 and consider a string tracing out a 2D surface as time passes (Fig.3: Worldtube of a closed string. depending on how you look at it) would be to consider a Lagrangian for an extended object. but it’s tensor components will have two indices since it’s a 2-form. 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. In string theory this is usually the “Kalb-Ramond ﬁeld”. then would it justify our guesswork and intuition in arriving at Eq. becomes Figure 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). A generalization (or specialization. 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.40 Lagrangians and Noether’s Theorem or with ‘proper time’ instead of ‘arclength’. The string action can then be written S = α · (area) + e B (3. 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). 3. 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. . call it B.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. where the 1-from A can be integrated over a 1-dimensional path. This may still seem like we’ve pulled a rabbit out of a hat.

.3.(3. (I’ve taken to using F here for the electromagnetic ﬁeld tensor to avoid clashing with F for the generalized force.5) we saw an example of a Lagrangian for relativistic electrodynamics that had awkward reparametrization symmetries. It’s particularly useful when Newtonian theory cannot give us a head start. (3. 1 • It looks formally like “ 2 mv 2 ”.7 Alternative Lagrangians 41 That’s enough about string for now. What Euler-Lagrange equations does this Lagrangian yield? ∂L = mqi + eAi ˙ ∂qi ˙ ∂Aj ∂L ˙ Fi = i = e i q j ∂q ∂q pi = Very similar to before! The E-L eqns. So the old Euler-Lagrange equations of motion reduce to the ˙ ˙ ˙ . which was a Lagrangian for a charged particle with reparametrization symmetry L = m q + eAi q i ˙ ˙ just as for an uncharged relativistic particle.8) • There’s no ugly square root.e. so it’s everywhere diﬀerentiable. The point was to illustrate the type of reasoning that one can use in conjuring up a Lagrangian. 3. in relativistic dynamics and in the physics of extended particles. familiar from nonrelativistic mechanics. meaning that H = 0 and there were non-unique solutions to the Euler-Lagrange equations arising from applying gauge transformations. Eq.6. This freedom to change the gauge can be avoided.) The only diﬀerence is that we have m¨i instead q of mvi where vi = qi / q . i.(3. then say d ∂Aj mqi + eAi = e i q j ˙ ˙ dt ∂q m¨i = eF ij q j q ˙ almost as before. Recall Eq. they are handled the same.5). But there’s another Lagrangian we can use that doesn’t have this gauge symmetry: 1 L = mq · q + eAi q i ˙ ˙ ˙ 2 This one even has some nice features.2 Alternate Lagrangian for Relativistic Electrodynamics In § 3. and there’s no trouble with paths being timelike or spacelike in direction.7.

With the new Lagrangian it’s not! Indeed H = pi q i − L ˙ and now pi = so ˙ ˙ ˙ H = (mqi + eAi )q i − ( 1 mqi q i + eAi q i ) ˙ ˙ 2 ˙ ˙ = 1 mqi q i 2 Comments. In general then there’s no conserved “energy” for our particle corresponding to translations in time. q) = 2 m q 2 + Ai (q)q 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)). so the particle’s path is parameterized by proper time up to rescaling of t. n+1 . we’re getting “conservation of speed” rather than some more familiar “conservation of energy”. This is vaguely like how a nonrelativistic particle in a potential V has but now the “potential’ V = eAi q i is linear in velocity.4. That is. and the fact that it’s conserved says q(t) is constant as ˙ a function of t. Our Lagrangian 1 L(q. Recall that for our reparametrization-invariant Lagrangian L=m qi q i + eAi q i ˙ ˙ we got H = 0. The reason is that this Hamiltonian comes from the symmetry qs (t) = q(t + s) instead of spacetime translation symmetry qs (t) = q(t) + s w. Let’s work out the Hamiltonian for this 1 L = mq · q + eAi q i ˙ ˙ ˙ 2 for the relativistic charged particle in an electromagnetic ﬁeld. ˙ As claimed H is not zero. so now ˙ H = pi q i − L = 2K − (K − V ) = K + V. time translation was a gauge symmetry. which would be a parametrization ˙ by proper time for example. w∈ the diﬀerence is illustrated schematically in Fig.42 Lagrangians and Noether’s Theorem new ones if we pick a parametrization with q = 1. ˙ ∂L = mq i + eAi ˙ ∂qi ˙ H = pi q i − L = (2K − V ) − (K − V ) = K. 3.

8 The General Relativistic Particle 43 PSfrag replacements 4 5 3 4 2 3 1 2 0 1 w qs (t) = q(t + s) qs (t) = q(t) + sw Figure 3. 0 0 −1 0 gij = . . is an (n + 1)-dimensional Lorentzian manifold. w) = gij v i w j 1 0 .4: Proper time rescaling vs spacetime translation. We deﬁne the metric as follows. w) for short.. 3. . w) −→ g(x)(v. w) or we could write g(v.3. but for diﬀerent bases gij will have a diﬀerent form. . 3. . . 0 0 . Q. −1 Of course we can write g(v.. we have a bilinear map . 2. For each x ∈ Q. g(x) : Tx Q × Tx Q −→ (v. g(x) varies smoothly with x.. . With respect to some basis of Tx Q we have g(v.8 The General Relativistic Particle In GR spacetime. w) = gij v i w j in any basis. . namely a smooth (n + 1)-dimensional manifold with a Lorentzian metric g. 1. . .

q) ˙ ˙ gij q i q j ˙ ˙ The Lagrangian for a free point particle in the spacetime Q is just like in special relativity but with ηij replaced by gij . nor velocity of center of energy) anymore! Let’s ﬁnd the equations of motion. ∂q i 1 L(q. ∂L = mgij q j ˙ i ∂q ˙ ∂L ∂ Fi = i = i 1 mgjk (q)q j q k ˙ ˙ ∂q ∂q 2 pi = ˙ ˙ = 1 m∂i gjk q j q k . q) = m ˙ =m g(q)(q. 2 (where ∂i = ∂ ). q) = 1 mg(q)(q. and we ˙ need the metric to relate them. −).8. Suppose then Q is a Lorentzian manifold with metric g and L : T Q → is the Lagrangian of a free particle.44 Lagrangians and Noether’s Theorem 3. 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 q i to the momentum mgij v j . and boost symmetry) is gone! So there is no conserved energymomentum (nor angular momentum. which in this case start from ∂L pi = i = mgij q j ˙ ∂q ˙ The velocity q here is a tangent vector. the momentum p is a cotangent vector. w) which gives Tq M −→ Tq∗ M v −→ g(v.1 Free Particle Lagrangian in GR L(q. w) −→ g(v. q) ˙ 2 = 1 mgij q i q j ˙ ˙ 2 The big diﬀerence between these two Lagrangians is that now spacetime translation symmetry (and rotation. We ﬁnd equations of motion from the Euler-Lagrange equations. Alternatively we could just as well use ˙ ˙ L(q. ˙ Getting back to the E-L equations. via g : Tq M × Tq M −→ (v. q) = 2 mgij q i q j ˙ ˙ ˙ .

which ¨ ˙ jk is really a particular type of connection that a Lorentzian manifold has (the Levi-Civita connection). Parallel transport is just the simplest way to compare vectors at diﬀerent points in the manifold. The Lagrangian would be 1 L = 2 mgjk q j q k + eAi q i ˙ ˙ ˙ which again was conjured up be replacing the ﬂat space metric ηij by the metric for GR gij . gik = gki .3. ¨ jk ˙ ˙ So we see that q can be computed in terms of q and the Christoﬀel symbols Γi . among other things. and also the “black book” of Misner. the Euler-Lagrange equations then yield the following equations of motion. 3. This allows us to deﬁne.2 Charged particle in EM Field in GR We can now apply what we’ve learned in consideration of a charged particle. q ˙ ˙ ˙ If you want to know more about Lagrangians for general relativity we recommend the paper by Peldan [Pel94]. This deﬁnes what we call the Christoﬀel symbols Γi . ˙ dt The mass factors away. a covariant derivative. of charge e. in our Lorentzian manifold. Γijk = − ∂i gjk − ∂k gij − ∂j gki the minus sign being just a convention (so that we agree with everyone else). m¨i = −mΓijk q j q k + eF ij q j . We can rewrite this geodesic equation as follows d ˙ ˙ gij q j = 1 ∂i gjk q j q k ˙ 2 dt ˙ ˙ ∂k gij q k q j + gij q j = 1 ∂i gjk q j q k ˙ ˙ ¨ 2 ∴ gij q j = ¨ = 1 2 1 ∂g 2 i jk ∴ ∂i gjk − ∂k gij − ∂j gki q j q k ˙ ˙ − ∂k gij q j q k ˙ ˙ where the last line follows by symmetry of the metric. so the motion is independent of the mass! Essentially we have a geodesic equation. Not surprisingly. in an electromagnetic ﬁeld with potential A.8. Then jk qi = gij q j = −Γijk q j q k ¨ ¨ ˙ ˙ ∴ q i = −Γi q j q k .8 The General Relativistic Particle 45 So the Euler-Lagrange equations say d 1 ˙ ˙ mgij q j = 2 m∂i gjk q j q k . a connection is just the rule for parallel transporting tangent vectors around the manifold. . Thorne & Wheeler [WTM71]. Now let.

. 0 1 . 3. 1] → Q.5: Least time versus least action. Jacobi was able to reinterpret the mechanics of a particle as an optics problem and hence “unify” the two minimization principles. First.) 3. times the usual Euclidean metric 1 . Nevertheless.9 The Principle of Least Action and Geodesics (Week 4. its trying to minimize the arclength of its path in the metric that is.1 Jacobi and Least Time vs Least Action We’ve mentioned that Fermat’s principle of least time in optics is analogous to the principle of least action in particle mechanics. since in the principle of least action we ﬁx the time interval q : [0.46 Lagrangians and Noether’s Theorem 3. let’s consider light in a medium with a varying index of refraction n (recall 1/n ∝ speed of light). Also. 20. the index of refraction n : δij = gij = n2 δij n → (0. Suppose it’s in n with its usual Euclidean metric. 0 . April 18.9. PSfrag replacements n high light faster light slower particle slower particle faster V high n low V low Figure 3. If the light s trying to minimize the time. 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.5). 22. This analogy is strange. ∞).

Now. L= = = gij (q)q i q j ˙ ˙ n2 (q)q i q j ˙ ˙ 2/m(E − V (q))q 2 ˙ (3.3.9)? He considered a particle of energy E and he chose the index of refraction to be n(q) = 2 E − V (q) m which is just the speed of a particle of energy E when the potential energy is V (q).10) . w) = gij v i w j where g(v. Consider a particle of mass m in Euclidean n with potential V : n → . i. what Jacobi did is show how the motion of a particle in a potential could be viewed as a special case of this. ˙ Note: this is precisely backwards compared to optics.9) gij (q)q i q j ˙ ˙ As before the Christoﬀel symbols Γ are built from the derivatives of the metric g. v) ≥ 0 So we’ll use the same Lagrangian: L(q.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. q) = ˙ and get the same Euler-Lagrange equations: d2 q i + Γi q j q k = 0 jk ˙ ˙ dt2 if q is parameterized by arclength or more generally q = ˙ gij (q)q i q j = constant. since 2 (E − V ) = m 21 m q ˙ m2 2 = q . where n(q) is proportional to the reciprocal of the speed of light!! But let’s see that it works. It satisﬁes F = ma.e.. ˙ ˙ (3.10) as a special case of (3. m d2 q i = −∂i V dt2 How did Jacobi see (3.

We get the ˙ ˙ ˙ Euler-Lagrange equations. ˙ 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.48 Lagrangians and Noether’s Theorem where q 2 = q · q is just the usual Euclidean dot product. g) is a Lorentzian manifold. The metric at q ∈ Q is. paths q : [t0 .9. i. We have 1. w) and it is bilinear. v · w = δij v i w j . w) = δij v i w j . 2.2 The Ubiquity of Geodesic Motion We’ve seen that many classical systems trace out paths that are geodesics. or m¨i = −∂i V . ˙ d dt 2/m(E − V (q)) · Jacobi noticed that this is just F = ma.r. 3.e. t1 ] → Q that are critical points of t1 S(q) = t0 gij q i q j dt ˙ ˙ which is proper time when (Q. g(q) : Tq Q × Tq Q → (v.t a basis of Tq Q g(v. w) → g(v. or arclength when (Q. w. that is. g) is a Riemannian manifold. pi = Fi = ∂L = ∂qi ˙ ∂L = ∂i ∂q i 1 = 2 2 q ˙ (E − V ) · m q ˙ 2 (E − V (q)) · q ˙ m −2/m∂i V · q ˙ 2/m(E − V q) 1 qi ˙ = − ∂i V q ˙ m q ˙ 2/m(E − V ) Then p = F says. q = 2/m(E − V (q)).. provided we reparameq terize q so that.

r. gij = (3) A free particle in general relativity traces out a geodesic on a Lorentzian manifold (Q. (2 ) A particle on a Riemannian manifold (Q.3. In fact all three of these results can be generalized to cover every problem that we’ve discussed! (1 ) Light on any Riemannian manifold (Q. (3 ) Kaluza-Klein Theory. 1 The case V > E. the Atwood machine. q). ∞) is the index of refraction function. (1) In the geometric optics approximation. Lots of physical systems can be described this way. g(q) varies smoothly with q ∈ Q.t the metric 2 h = (E − V )g m if it has energy E. We’ve seen at least three important things. q) with potential V : Q → traces out geodesics w. where n : Q → (0. An important distinction to keep in mind is that Lorentzian manifolds represent spacetimes. n acts like particles tracing traces out . if they exist. e. so they all ﬁt into this framework. q) with index of refraction n : Q → (0. spinning tops. whereas Riemannian manifolds represent that we’d normally consider as just space. light in Q = out geodesics in the metric gij = n(q)2 δij (2) Jacobi saw that a particle in Q = geodesics in the metric n in some potential V : Q → 2 (E − V )δij m if the particle has energy E (where1 V < E). would be classically forbidden regions. ∞) traces out geodesics in the metric h = n2 g.. a rigid rotating body (Q = SO(3)). A particle with charge e on a Lorentzian manifold (Q.9 The Principle of Least Action and Geodesics 49 3.g. and others. All of these systems have a Lagrangian which is a quadratic function of position. q) in an electromagnetic vector potential follows a path with e ˙ qi = −Γijk q j q k + F ij q 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.

n} since we’re in n + 1-dimensional spacetime. 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 qi = −Γijk q j q k + ¨ ˙ ˙ qθ = 0. 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. . ¨ if qθ = e/m ˙ since F ij is part of the Christoﬀel symbols for h. Let’s pick coordinates xi on Q where i ∈ {0. 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. . and θ is our local coordinate on S 1 .50 Lagrangians and Noether’s Theorem Let’s examine this last result a bit further. . . 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! e F ij q j ˙ m .

Chapter 4 From Lagrangians to Hamiltonians In the Lagrangian approach we focus on the position and velocity of a particle. But we’re led to consider momentum ∂L pi = i ∂q ˙ since the equations of motion tell us how it changes dpi ∂L = i. In this approach.1 The Hamiltonian Approach In the Hamiltonian approach we focus on position and momentum. and compute what the particle does starting from the Lagrangian L(q. which is a function ˙ L : T Q −→ where the tangent bundle is the space of position-velocity pairs. position and momentum will satisfy Hamilton’s equations: dq i ∂H = . and compute what the particle does starting from the energy H = pi q i − L(q. called the Hamiltonian H : T ∗ Q −→ where the cotangent bundle is the space of position-momentum pairs. q). dt ∂pi 51 dpi ∂H =− dt ∂qi . dt ∂q 4. q) ˙ ˙ reinterpreted as a function of position and momentum.

q) Tq Q ˙ Tq Q q Figure 4.1). 4. λ can be deﬁned in a coordinate∂L free way. p) ˙ where q ∈ Q. the q i directions. We have ˙ π : T Q −→ Q (q. We want to deﬁne “ ∂ qi ” in a coordinate-free ˙ PSfrag replacements way. ˙ and p is a cotangent vector in Tq∗ Q := (Tq Q)∗ . Despite appearances. q) −→ (q. q) −→ q ˙ and dπ : T (T Q) −→ T Q Tq Q TQ T(q. and q is any tangent vector in Tq Q (not the time derivative of something). ˙ To obtain this Hamiltonian description of mechanics rigorously we need to study this map λ : T Q −→ T ∗ Q (q.1: Q .52 From Lagrangians to Hamiltonians where the latter is the Euler-Lagrange equation ∂L dpi = dt ∂qi in disguise (it has a minus sign since H = pq − L).e. given by q −→ pi = ˙ λ ∂L ∂qi ˙ So λ is deﬁned using L : T Q → .. it’s the “diﬀerential of L in the vertical direction”—i. as follows (referring to Fig.

p) ˙ given in local coordinates by pi = 1 PSfrag replacements Tq∗ Q ∂L . 4.) Given L : T Q → T ∗ Q.q) T Q = T (Tq Q) ˙ and since Tq Q is a vector space. so we have ˙ ∗ (dL)(q.q) Tq Q ˙ q Q . so ker dπ = {v ∈ T T Q : dπ(v) = 0 ∈ T Q}. ∂qi ˙ The kernel of a map is the set of all elements in the domain that map to the null element of the range. getting f : V(q. ˙ ˙ We can restrict this to V T Q ⊆ T T Q.q) T Q −→ .q) T Q ˙ T(q. May 2. ˙ But note V(q. q) −→ (q. T(q.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.4.q) : T(q. Tq Q V(q. 6.q) T Q −→ . p∈ this is the momentum! (Week 6. we now know a coordinate-free way of describing the map λ : T Q −→ T ∗ Q (q. So f gives a linear map p : Tq Q −→ that is. q) ∈ T Q is a map from T T Q to .q) ∈ T(q.q) T Q ˙ ˙ that is. dL(q.q) Tq Q ∼ Tq Q = ˙ in a canonical way2 . 2 The ﬁber Tv V at v ∈ V of vector manifold V has the same dimension as V .

q) ∈ T Q ˙ or position and momentum (q. q) = mgij q i q j − V (Q) ˙ 2 Here pi = so λ(q. −) is 1-1 and onto. q) ∈ X..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. Thus λ is a diﬀeomorphism.2 Regular and Strongly Regular Lagrangians This section discusses some examples of the above theory. In practice often X = T ∗ Q. 4. mg(q. 3 .1 Example: A Particle in a Riemannian Manifold with Potential V (q) has Lagrangian For a particle in a Riemannian manifold (Q. not inserted since λ itself is an operator on tangent vectors. (q. not the result of the operation. the metric is nondegenerate. q) = q.2. The missing object there “−” is of course any tangent vector.e. then L is said to be strongly regular. −) ˙ ˙ so3 L is strongly regular in this case because Tq Q −→ Tq∗ Q ∂L = mgij q j ˙ ∂qi ˙ v −→ g(v. q) in a potential V : Q → 1 ˙ L(q. ˙ We call X the phase space of the system. p) = λ(q. 4. i. In this case we can describe what our system is doing equally well by specifying position and velocity. which in this case extends to all of T ∗ Q.

Alas. −) + eA(q) ˙ ˙ 4. 4 . Where it is deﬁned ˙ ˙ pi = mgij q j ˙ ∂L = i ∂q ˙ q ˙ (where q is timelike).2 Example: General Relativistic Particle in an E-M Potential For a general relativistic particle with charge e in an electromagnetic vector potential A the Lagrangian is 1 ˙ ˙ L(q. Let Q = L(q. In aﬃne geometry there is no deﬁned origin. q −→ m g(q.2. which are nonlinear. but aﬃne transformations include translations. and undeﬁned when the same is negative. we can ask about regularity. q) = m ˙ gij q i q j ˙ ˙ This is terrible from the perspective of regularity properties—it’s not diﬀerentiable when gij q i q j vanishes.2. q) = f (q) ˙ ˙ All linear transforms are aﬃne. L(q.4 Example: A Regular but not Strongly Regular Lagrangian and Here’s a Lagrangian that’s regular but not strongly regular.3 Example: Free General Relativistic Particle with Reparameterization Invariance The free general relativistic particle with reparameterization invariant Lagrangian has. q) = mgij q i q j − eAi q i ˙ 2 and thus ∂L ˙ pi = i = mgij q i q j + eAi . ∂q ˙ This L is still strongly regular. For the example the translation is the “+eA(q)” part.2.) 4. but now each map λ |Tq Q : Tq Q −→ Tq∗ Q is aﬃne rather than linear4 . the map λ is not 1-1 where deﬁned ˙ since multiplying q 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 Regular and Strongly Regular Lagrangians 55 4.

3 Hamilton’s Equations λ : T Q −→ X ⊆ T ∗ Q (q.) ∂q i which is really a function on T Q. q) −→ (q. Now let’s calculate: dL = ∂L i ∂L i dq + i dq ˙ ∂q i ∂q ˙ = pi dq i + pi dq i ˙ ˙ ∂L = f (q) ˙ ∂q ˙ This will be regular but not strongly so if f : → is a diﬀeomorphism from ∼ proper subset U ⊂ .. For example. This lets us treat q i . so This lets us have the best of both worlds: we can identify T Q with X using λ. ∞) ⊂ . So ˙ p= so that to some .56 From Lagrangians to Hamiltonians positive slope L(q. will be treated as a function on X. p) ˙ ∼ Now let’s assume L is regular. all as functions on X (or T Q). H. thus writing qi ˙ for the function q i ◦ λ−1 ˙ In particular pi := ˙ (function on X) (function on T Q) ∂L (Euler-Lagrange eqn. q) = eq˙ ˙ G or slope between −1 and 1 G L(q. take f (q) = eq˙ so f : → (0. q) = ˙ and so forth. 1 + q2 ˙ 4. etc. pi . L.

q) −→ (q. so ∼ . we conclude: qi = ˙ These are Hamilton’s Equations. when the Euler-Lagrange equations hold for ˙ ˙ some path q : [t0 . As we’ve seen. pi ). they will be the time derivatives of q i and pi . this time using local coordinates (q i . and use (q .4. In fact. i ∂p ∂q Since dpi . ˙ ˙ λ : T Q −→ X ⊆ T ∗ Q (q.3 Hamilton’s Equations 57 while dH = d(pi q i − L) ˙ = q i dpi + pi dq i − dL ˙ ˙ = q i dpi − pi dq i ˙ ˙ ˙ = q i dpi + pi dq i − (pi dq i + pi dq i ) ˙ ˙ ˙ ˙ so Assume the Lagrangian L : T Q → dH = q i dpi − pi dqi . p) ˙ is a diﬀeomorphism. This lets us regard both L and the Hamiltonian H = pi q i − L as ˙ i i functions on the phase space X. dq i form a basis of 1-forms. ˙ ˙ But we can also work out dH directly. Hamilton’s equations are just the EulerLagrange equations in disguise. Hamilton’s equations describe the motion of a point x(t) = q(t). ˙ this gives us dL = pi dq i + pi dq i ˙ ˙ dH = q i dpi − pi dq i . t1 ] → Q. ∂H . So when the Euler-Lagrange equations hold. q ) as local coordinates on X. to get dH = ∂H ∂H dpi + i dq i . p(t) ∈ X.1 Hamilton and Euler-Lagrange Though q i and pi are just functions of X.3. ∂pi pi = − ˙ ∂H ∂qi 4. in this context. The equation qi = ˙ ∂H ∂pi is regular.

58 From Lagrangians to Hamiltonians really just lets us recover the velocity q as a function of q and p. = 2m p m g ij pj ) m (though really that’s q = ˙ So Hamilton’s equations say ∂H ∂pi ∂H pi = − i ˙ ∂q qi = ˙ ⇒ ⇒ q= ˙ p=− V ˙ The ﬁrst just recovers q as a function of p. ˙ . p) −→ (q. q). ∂q i ∂q ∂q Example: Particle in a Potential V (q) For a particle in Q = n ∂H ∂q i ∂L ∂q i or p= ˙ in a potential V : n → the system has Lagrangian m q 2 − V (q) ˙ L(q. q) = ˙ 2 which gives p = mq ˙ p q= . inverting the formula ˙ pi = ∂L ∂qi ˙ which gave p as a function of q and q. i dt ∂ q ˙ ∂q These are the same because ∂L ∂ ∂H ˙ = i pi q i − L = − i . So we get a formula for the map ˙ λ−1 : X −→ T Q (q. the second is F = ma. ˙ m and Hamiltonian H(q. ˙ Given this. the other Hamilton equation pi = − ˙ is secretly the Euler-Lagrange equation d ∂L ∂L = i. p) = pi q i − L = ˙ 1 p 2 p 2− −V m 2m 1 p 2 + V (q).

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. This is important. t1 ] → Q and setting δS = 0. 4. t1 ] → X and setting δS = 0. ˙ Let P be the space of paths in the phase space X and deﬁne the action S : P −→ by t1 S(x) = t0 (pi q i − H)dt ˙ where pi q i − H = L.3. which is not symplectic. Now let’s get Hamilton’s equations directly by assigning an action S to any path x : [t0 . Note: we don’t impose any relation between p and q. q! The relation will follow from δS = 0. We’ll introduce symplectic geometry more completely in later chapters.4. The coordinate-free description of this structure is the major 20th century contribution to mechanics: a symplectic structure. δS = δ = (pi q i − H)dt ˙ δpi q i + pi δ q i − δH dt ˙ ˙ dt . X will just be a manifold equipped with enough structure to write down Hamilton’s equations starting from any H : X → . we obtained the Euler-Lagrange equations by associating an “action” S with any q : [t0 . That’s the phase space X.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. You might have some particles moving on a manifold like S 3 . p(t) and let ˙ t1 S(x) = t0 pi (t) d i q (t) − H q(t). More precisely.2 Hamilton’s Equations from the Principle of Least Action Before. we’ll drop the requirement that X ⊆ T ∗ Q where Q is a conﬁguration space. write our path x as x(t) = q(t). p(t) dt d we write dt q i instead of q i to emphasize that we mean the time derivative rather than a ˙ coordinate in phase space. Soon. Let’s show δS = 0 ⇔Hamilton’s equations.

60 From Lagrangians to Hamiltonians then integrating by parts. which has coordinates {pi . t) . We’ve seen two principles of “least action”: 1. So any path x : [t0 . we might consider a third version based on paths in positionvelocity space T Q. δS = 0 ⇔Euler-Lagrange equations. But when our Lagrangian is regular we have a diﬀeomorphism λ : ∼ T Q → X. Just as we hoped. δS = 0 ⇔Hamilton’s equations. However. t}. t1 ] −→ X × t −→ (x(t). = = = δpi q i − pi δq i − δH dt ˙ ˙ δpi q i − pi δq i − ˙ ˙ ∂H i ∂H δq − δpi dt ∂q i ∂pi ∂H ∂H dt δpi q i − ˙ + δq i −pi − i ˙ ∂pi ∂q ∂H . t1 ] −→ X gives a path σ : [t0 . δp) if and only if Hamilton’s equations qi = ˙ pi = − hold. since X ⊆ T ∗ Q. the really interesting principle of least action involves paths in the extended phase space where we have an additional coordinate for time: X × . For paths in conﬁguration space Q. Additionally. q i . Recall the action S(x) = = = We can interpet the integrand as a 1-form β = pi dq i − H dt (pi q i − H) dt ˙ pi dq i dt − H dt dt 2. pi dq i − H dt on X × . For paths in phase space X. ∂pi ∂H ∂q i This vanishes ∀δx = (δq. so this third principle of least action is just a reformulation of principle 2.

∞) is the index of refraction throughout space (generally not a constant). g) as space. May 9.1 Wave Equations Huygens considered this same setup (in simpler language) and considered the motion of a wavefront: G . neutrinos. We’ve talked about geometrized optics. 11. etc.) In quantum mechanics we discover that every particle—electrons.—is a wave. photons.4. and vice versa. 13. 4.4 Waves versus Particles—The Hamilton-Jacobi Equations (Week 7. but we use the new metric hij = n2 gij where n : Q → (0. Here we start with a Riemannian manifold (Q. 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. an approximation in which light consists of particles moving along geodesics.4.4 Waves versus Particles—The Hamilton-Jacobi Equations 61 and the action becomes the integral of a 1-form over a curve: S(x) = pi dq i − H dt = β σ 4.

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: 1$0%×ÐÖÑ)&ÕÒÔÓ1$0%('×ÐÖÑ)&ÕÒÔÓ (' 1$0%×Ð1$ÖÑ)&ÕÒ0%ÔÓ ×ÐÖÑ(')&ÕÒÔÓ ('Ô 1$1$0%0%×ÐÖÑ×Ð)&ÕÒÖÑÔÓ)&ÕÒÔÓ ('(' 11$$0%1$0%0%×ÐÖÑ×Ð)&ÕÒÖÑÔÓ×Ð)&ÕÒÖÑÔÓ)&ÕÒÔÓ('('(' 1×1$0%ÐÖÑ)&ÕÒ$ÔÓ0%('×Ð1$ÖÑ)&ÕÒ0%ÔÓ ×ÐÖÑ(')&ÕÒÔÓ (' 1$0%×ÐÖÑ)&ÕÒÔÓ (' balls of radius centered at points of the old wavefront In short. q1 ) is the least action—the inﬁmum of action over all paths from q0 to q1 . ∞) on the Riemannian manifold (Q. 4.2) and it can also happen for geometrical reasons . q1 ) = inf (arclength) Υ where Υ = {paths from q0 to q1 }. This sort of situation can happen for topological reasons (as when the waves smash into each other in the back of Fig.2: be smooth—we say a catastrophe occurs—and then the wavefronts are no longer the level sets.2. h).) Using this we get the wavefronts centered at q0 ∈ Q as the level sets {q : d(q0 . 4. the wavefront moves at unit speed in the normal direction with respect to the “optical metric” h. For larger c the level sets can cease to PSfrag replacements q0 Figure 4. as depicted in Fig. We can think about the distance function d : Q × Q −→ [0. q) = c} or at least for small c > 0. (Secretly this d(q0 . where d(q0 .

They noticed that: ∂ W (q0 .. 5 .4. This becomes the eikonal approximation in optics5 once we ﬁgure out what A should be. in optics the eikonal approximation is the basis for ray tracing methods. particles of light move along geodesics.q0 ) where k is the wavenumber of the light (i.4 Waves versus Particles—The Hamilton-Jacobi Equations 63 Figure 4. Hamilton and Jacobi focused on distance d : Q × Q → [0. which drops oﬀ far from q0 . 4. its color) and A : Q → describes the amplitude of the wave. we can approximate the waves of light by a wavefunction: ψ(q) = A(q)eik d(q.e. Assuming no such catastrophes occur.4. but wavefronts are level sets of the distance functions: gy Eikonal comes form the Greek word for ‘image’ or ‘likeness’.3).3: (Fig. ∞) as a function of two variables and called it W =Hamilton’s principal function.2 The Hamilton-Jacobi Equations k th a uo Ñz rkevhX UPE × Ô{ h Pqqq A7 C ! k P 7 qVq 7 % %1 { qqqq 1 3 )3 7 % 7 I % % %) 7A C QX hk{ Ô ! EP Ua hvekrz Ñ× y ht xgkot We’ve seen that in optics. 4. q1 ) = (p1 )i i ∂q1 q1 q0 • • p1 G where p1 is a cotangent vector pointing normal to the wavefronts.

In the eikonal approximation. the momentum is an operator that acts to diﬀerentiating the wavefunction. or Hamilton’s principal function: W (q0 . ∞) is the distance function on Q.e. q0 ∈ Q is the light source.. In fact kp1i is the momentum of the light passing through q1 . Hamilton went further. h) is a Riemannian manifold. This is begging to be generalized to other Lagrangian systems! (At least it is retrospectively. Jacobi generalized this to the motion of point particles in a potential V : Q → . 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.q1 ) where (Q. i. and we now can go further still. This foreshadows quantum mechanics! After all. light is described by waves ψ(q1 ) = A(q1 )eik W (q0 .) We also saw that ∂ W (q0 . k is the frequency and W : Q × Q −→ [0. in quantum mechanics. its arclength. Deﬁne Hamilton’s principal function W :Q× ×Q× −→ ¤ ψ : Q −→ G . with the advantage of our historical perspective. q1 ) = (p1 )i . m We’ve seen this reduces point particle mechanics to optics—but only for particles of ﬁxed energy E. i ∂q1 q1 p 1 q0 • • “points normal to the wavefront”—really the tangent vector pi = hij (p1 )j 1 points in this direction. Suppose Q is any manifold and L : T Q → is any function (Lagrangian). using the fact that a particle of energy E traces out geodesics in the metric hij = 2(E − V ) gij . h is the optical metric.64 From Lagrangians to Hamiltonians at least while the level sets remain smooth.

we have ∂ W (q0 . (+energy at time t0 ) ∂t0 (H1 = H0 as energy is conserved).4 Waves versus Particles—The Hamilton-Jacobi Equations 65 by W (q0 . it’ll be smooth if (q0 . Maybe the best way to get the Hamilton-Jacobi equations is from this extended phase space formulation. t1 ) = inf S(q) q∈Υ where Υ = q : [t0 . (-energy at time t1 ) ∂t1 ∂W = H0 . t0 . t1 ) = S(q) • • G . β = pi dq i − H dt on the extended phase space X × . Given (q0 . and ∂W = −(p0 )i . But for now let’s see how Hamilton’s principal function W and variational principles involving least action also yield the Hamilton-Jacobi equations. q(t) dt ˙ Now W is just the least action for a path from (q0 . These last four equations are the Hamilton-Jacobi equations. q1 . q(t0 ) = q0 . t0 ) where p1 is the momentum of the particle going from q0 to q1 .4. The mysterious minus sign in front of energy was seen before in the 1-form. i ∂q1 p (q1 . t1 ) 1 (q0 . & q(t1 ) = q1 and S(q) = t0 t1 L q(t). let q : [t0 . (q1 . at time t1 . t1 ). t0 ) to (q1 . t1 ] → Q. q1 ) = (p1 )i . Then W (q0 . t1 ). t1 ) are close enough—so let’s assume that is true. In fact. (-momentum at time t0 ) i ∂q0 ∂W = −H1 . q1 . t0 ). t0 ) and (q1 . t0 . t1 ] −→ Q be the action-minimizing path from q0 to q1 .

66 From Lagrangians to Hamiltonians Now consider varying q0 and q1 a bit q t0 t1 and thus vary the action-minimizing path. We get δW = δS t1 =δ t0 t1 L(q. we need to vary t0 and t1 : Now change in W will involve ∆t0 and ∆t1 • • • • t0 t0 + ∆t0 t1 t1 + ∆t1 (you can imagine ∆t0 < 0 in this ﬁgure if you like). q) dt ˙ ∂L i ∂L i δq + i δ q dt ˙ ∂q i ∂q ˙ ∂L i δq − pi δq i dt + pi δq i ˙ ∂q i ∂L − pi δq i dt ˙ ∂q i t1 t0 = t0 t1 = t0 t1 = t0 the term in parentheses is zero because q minimizes the action and the Euler-Lagrange equations hold. getting a variation δq which does not vanish at t0 and t1 . . So we δq i have i i δW = p1i δq1 − p0i δq0 and so ∂W = p1i . i ∂q1 and ∂W = −p0i i ∂q0 These are two of the four Hamilton-Jacobi equations! To get the other two.

q1 . and assume that this q depends smoothly on U = (q0 . We need to ensure that (q0 . p(t). t ∈ X × . t1 ) are suﬃciently close. p) dt ˙ p dq − H dt β C = t0 t1 = t0 = where β = pdq − H(q. We’ll think of q as a function of U : q (t0 . t1 ) that there is a unique q ∈ Υ that minimizes the action S. p) ˙ (Q × )2 → Υ u→q deﬁned only when (q0 .4 Waves versus Particles—The Hamilton-Jacobi Equations 67 We want to derive the Hamilton-Jacobi equations describing the derivatives of Hamilton’s principal function W (q0 . t1 ) = inf S(q) q∈Υ where Υ is the space of paths q : [t0 . q1 . . q) dt ˙ where the Lagrangian L : T Q → is a diﬀeomorphism. p)dt is a 1-form on the extended phase space X × . q0 ) • • (t1 . q(t1 ) = q1 and t1 S(q) = t0 L(q. t0 ) is close enough to (qi .4. q1 ) Then Hamilton’s principal function is W (u) := W (q0 . q1 . t1 ] → Q with q(t0 ) = q. t0 . t0 . and C is a curve in the extended phase space: C(t) = q(t). ¥ will now be assumed regular. q) −→ (q. t1 ) ∈ (Q × )2 . so that λT Q −→ X ⊆ T ∗ Q (q. q) dt ˙ pq − H(q. t0 ) and (q1 . t0 . t1 ) = S(q) t1 = t0 t1 L(q.

. • . . ... . We are after the derivatives of W that appear in the Hamilton-Jacobi relations.. As . .. we get d ds β=0 As +Cs +Bs (although As + Cs + Bs is not smooth. . t1 ) ∈ (Q × )2 ... . . .. . t0 . . Similarly.. .. .. .... . . So d d d β= β− β at s = 0. .... . . . ..... . . .. . Let us be a 1-parameter family of points in (Q × )2 and work out where Cs depends on us as above X× C0 D Let’s compare β Cs and As +Cs +Bs = As β+ Cs β+ Bs β Since C0 minimizes the action among paths with the given end-points.. . .. .. and the curve As + Cs + Bs has the same end-points. we can approximate it by a path that is smooth). .. ds Cs ds Bs ds As Note d d β(Ar ) dr β= ds As ds = β(A0 ) where A0 = v is the tangent vector of As at s = 0.. .68 From Lagrangians to Hamiltonians Note that C depends on the curve q ∈ Υ. Bs . q1 . so let’s diﬀerentiate W (u) = β with respect to u and get the Hamilton-Jacobi equations from β. .. . . ..... .. . .. .. ... . • .... . ... d ds β = β(w) Bs C d d W (us ) = ds ds β Cs .. . • CB s • . . ... which in turn depends upon u = (q0 ... . ..

but nowadays it is thoroughly familiar from quantum mechanics! . Now since β = pi dqi − Hdt. if we deﬁne a wavefunction: ψ(q0 . So.4 Waves versus Particles—The Hamilton-Jacobi Equations 69 where w = B0 .. t0 ). p1 . p0 . t1 ) as we move Cs and v keeps track of (q0 .t1 )/ then we get ∂ψ i = − H1 ψ ∂t1 i ∂ψ = p1 ψ i ∂q1 At the time of Hamilton and Jacobi’s research this would have been new. q1 . we get ∂W = pi 1 i ∂q1 ∂W = −H ∂t1 and similarly ∂W = −pi 0 i ∂q0 ∂W =H ∂t0 So. t0 .. t1 ) = eiW (q0 . d W (us ) = β(w) − β(v) ds where w keeps track of the change of (q1 .t0 .q1 .4.

A. 70 . H. [WTM71] J. Classical and Quantum Gravity. W. Gravitation. 1994. New York. S. K. 1971. 11:1087. W. Freeman. Misner. Wheeler.Bibliography [Pel94] Peter Peldan. and C. with generalizations: A review. Actions for gravity. Thorne.

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd