‘SERIES ON KNOTS AND EVERYTHING Ecdtoin-charge: Lous H. Kautman Pablsed acts and Physics LH Kaufman How Surfaces nenetia Space Qumeum Topology ited by LH. Rania & RA. Baio £Gpseries on Knots and Everything — Vol. 4 GAUGE FIELDS, KNOTS AND GRAVITY John Baez ogee teats Ue of Caton Reise World Scientific ‘Singapore "New Jersey “London Hong Konga) ee ska ee 159 Mn Rive Ee, I Uke 19 yn Meat Tne, anor NEO iba Congres Calgon Datt ui Srtnbm atria 8 SBN SHL0317295 ~ ISBN 910220340 9b) (ug fl a) 3, Quam pay Ka ten « Genrdtcay (ya) Beoeapatan” 1 Mane cup a | eae seo Crerao ume earn seormiberiareecreertrse meres, ec poping of man i vote, le ray hue he ori (Silane eens 2 Can Se em MA ODT OS Prien Sgapore by UP To Our ParentsPreface ‘Two of the most exciting developments of 20th century physics were general relativity and quantum theory, the Iatter culminating in the ‘standard model’ of particle interactions. General relativity treats gra ity, while the standard model treats the rest of the forces of nature, Unfortunately, the two theories have not yet been assembled into a single coherent picture of the world. In particular, we do not have 1 working theory of gravity that takes quantum theory into account Attempting to “quantize gravity” has led to many fascinating develop ‘ments in mathemati and physis, but it remains a challenge for the ist century. The early 1980s were a time of tremendous optimism concerning string thooey. ‘This theory was very ambitions, taking as its guiding philosophy the idea that gravity could be quantized only by unifying it with all the other forces, As the theory became immersed in ever more ‘complicated technica issues without any sign ofan immediate payoff in testable experimental predictions, some of this enthusiasm diminished among physicists, Ironically, at the same time, mathematicians found string theory an ever mote fertile source of new ideas. A particuacly appealing development to mathematicians was the discovery by Bd- ward Witten in the late 1980s that Chern-Simons theory ~~ a quantum field theory in 3 dimensions that arose as a spin-of of string theory — was intimately related to the invariants of knots and links that had recently been discovered by Vaughan Jones and others. Quantum feld theory and &-dimensional topology’ have became firmly bound together fever since, although there is much that remains mysterious about the relationship, ‘While less popular than string theory, a seemingly very diferent ap-wi Peetu proach to quantum gravity also made dratiatic progress in the Abhay Ashtekar, Carlo Revel, Lee Smolin and others dissuvrtest hens to rewrite general relativity in terms of ‘new variables’ s that it wor closely resembled the other forces of nature allowing shen tw apply ' new set of techniques to the problem of quantizing gravity. “The philosophy of these researchers was far more conservative than that ‘of the string theorists. Instead of attempting a ‘theory of everything describing all forces and all particles, they sttempted to wnverstanl ‘quantum gravity on ils oun, following as closely a8 possible the tr ‘onal guiding principles ofboth general relativity and quantum theory Interestingly, they too were led to the study of knots and Finks. Indes, their approach is often known as the ‘loop representation” of quanti gravity, Furthermore, quantum gravity in 4 dimensions tured out to be closely related to’ Chern-Simons theory in 3 dimensions. Again, there is much that remains mysterious about this, For example, one ‘wonders why Chern-Simons theory shows up so prominently both in string theory and the loop representation of quantum gravity. Perhaps these two approaches are not a8 differen as they sem! Lis the goal of this txt to provide an elementery introduction to some of these developments. We hope that both physicists who wish to lear more differential geometry and topology, and mathematicians who wish to learn more gauge theory and geneal relativity will find this ‘book a useful place to begin. The main prerequisites are some familiar ity with electtomagnetism, special relativity, linear algebra, and vector calculus, together with some ofthat undefinable commodity known as ‘mathematical sophistication’ ‘The book is divided into three parts that treat electromagnetism, sgxoge theory, and general relativity, respectively. Part [of this book introdaces the language of modem differential geometry, and shows how Maxwells equations can be drastically simplified using this lan guage. We stress the coordinate-free approach and the relevance of tlobal topological considerations in understanding such things as the Bohm-Aharonow effect, wormholes, and magnetic monopoles. Part IT introduces the mathematics of gauge theory — fiber bundles, connec- tions and curvature — and then introduces the Yang Mills equation, Cher clases, and Chern-Simons classes, I alo includes a bret intro. ‘duction to knot theory and its relation to Chern-Simons theory. Part Preface x TIL introiuces the basic concepts of Riemannian and semi-Rtiemannian igometry and then concentrates on topies in general relativity of spe cial importance to quantum gravity: the Binstei-Hilbert and Pelatig formulations of the action principle for gravity, the ADM formalism, and canonical quantization. Here we emphasize tentor computations ‘written in the notation used in general relativity. We conclude this part with ashetch of Ashtekat's ‘new variables’ and the way Chern-Simons theory provides a rolution to the Wheeler-DeWitt equation (the basic ‘equation of canonical quantum gravity), ‘While we attempt to explain everything ‘rom scratch’ in a sell: contained manner, we really hope to lure the eader into further study of differential geometry, topology, gauge theory, general relativity and ‘quantum gravity. For this reason, we provide copious notes at the end cof each pert, isting our favorite reading material on all these subjects Indeed, the reader who wishes to understand any of these subjects in epth may find i useful to read some of these references in parallel with our book. his is especially true because we have left out many’ relevant topies in order to keep the book coherent, elementary, and reasonable in size. For example, we have not discussed fermions (or ‘mathematically speaking, spinors) in any detail. Nor have we treated principal bundles. Also, we have not done justice to the experimental aspects of particle physics and general relativity focusing instead upon their common conceptual foundation in gauge theory. The reader will thus have to turn to ther texts to lesen about stich matters, ‘One really cannot learn physics or mathematics except by doing it, Por this reason, this text contains over 300 exercises. Of course, far more exercises are assigned in texts than are actually done by the readers. At the very least, we urge the eader to read and ponder the exercises, the results of which are often used later on. ‘The text also Includes 130 ilustratons, since we wish to emphasize the geometrical and topological aspects of modern physics. Teems appear in boldface when they are defined, and all such definitions are referred to in the index ‘This book is based on the notes of a seminar on knot theory and: quantum gravity taught by J.B. at U. C. Riverside during the school year 1992-1098, ‘The seminar concluded with a conference on the sub- ject, the proceedings of which will appear in a volume entitled Knotsand Quantum Gravity ‘We would like to thank Louis Kauffman for inviting us to write tix book, and also Chris Lee and Ms. H. M. Ho of World Scientie for helping us at every stage of the writing and publication process, We also wish to express our thanks to Edward Heflin and Dardo D. Pix for reading parts of the manascrit and to Carl Yao for helping ws with some IbTeXcomplications. Scott Singer of the Academic Computing Graphics and Visual Imaging Lab ofthe UC. Riverside deserves special ‘hanks end cecognition for helping us to create the book cover. Some ofthe graphics used forthe design of the cover were generated with Mathematics, by Wolfram Research, Inc; these were kindly given to us by Joe Grohens of WR. {I.B.is indebted to many mathematicians and physicists for useful discussions and correspondence, but he would particulary ike to thank Abhay Ashtckar, whose work has done xo much to unify the study of ‘gauge fields, knots and gravity. He woold als like to thank the readers of the USENET physics and mathematics newsgroups, who helped in ‘many ways with the prepartion ofthis book. He dedicates this book to his parents, Peter and Phyllis Bsez, with profound thanks for thei love. He alo gives thanks and love to his mathematical muse, Lisa Raplals {J.M. dedicates this book to his parents, Luis and Crescencin Pérez ‘de Munitin y Mobedano, for their many years of continued love and support. He is grateful to Bleanor Anderson for being a patient and inspiring companion during the long and hard hours taken to complete this book. He also acknowledges José Wudla for many discussions on ‘quantum field theory. Contents Preface I Electromagnetism 1 Maxwells Equations 2 Manifolds 3 Vector Fields 4. Differential Forms 5. Rewriting Maxwell's Equat © DeRham Theory in Electromagnetism Notes to Part I IL Gauge Fields 1 Symmetry 2 Bundles and Connections 3 Curvature and the Yang-Mills Equation 4 Chem-si ons Theory 1s 23 39 69 103 153 159 161 199 a3 267xi G 5. Link Invariants from Gauge Theory 21 Notes to Part It 383 ie sad ss Gauge Fields, Knots, and Gravity 1. Seal-Riemannian Geometey 305 2 Binstein’s Equation set 2 Lagrangians for General Relativity sor 4. The ADM Formalism as 5. The New Variables ar Notes to Part 1 asi Index 457Part I ElectromagnetismChapter 1 Maxwell’s Equations (Our whole progress upto thi point may be described as @ gradual develop. ‘ment of the doctrine of relatorty of all physical phenomena. Postion we ‘mrt evidently acknowisdye to be relative, for we cannot describe the pos ton of body sm any terme which dono express relation. The erdinary language about motion and rest doesnot wo completly exclude the notion of their being measured bsoltly, bet the reason of this i, that in our ordi rary language we lacilyasrue thatthe earth is at rest. There are no landmaris im space; one portion of space is enacly like every other portion, 49 that we can tll whee we are. We ae, at it were, on an wirafed sea, without stars, compas, sowing, wind or tide, and we cannot tll in ‘hat dietion we are going We have nolo which we can ease out to take 1 dead reckoning by se may compute our rate of motion with respect fo the ‘heightoring bodies, but se donot know how these bodies may be moving im ‘space. ~ James Clerk Mazuell, 1870 Starting with Maxwell's beautiful theory of electzomagnetism, and inspired by it, physicists hare made tremendous progress in under standing the basic foroes and particles constituting the physical world. “Maxell showed that two acemingly very diferent forces, the electric land magnetic forces, were simply two aspects of the ‘electromagnetic field”. In 4o doing, be was alo able to explain light as phenomenon in which ripples inthe electri feld create ripples in the magnetic fil ‘which in turn create new ripples in the electric field, and 20 on. Shock” ingly, however, Masel’ theory also predicted that light emitted by @ roving body would travel no faster than ight fom a stationary body. 34 Ld Maxwell's Equations Eventually this led Lorenta, Poincaré and especially Kinstein to realize that our ideas about space and time had to be radically revised. That the motion of » body can only he measured eelative to another body had been understood to some extent since Galileo, Taken in canjune tion with Maxonelle theory, however, this principle foreed the recogni- tion that in addition to the rotational syeametsies of space there must bbe symmetries that mingle the space and time coordinates, These new symmetries also mix the eletsic and magnet felds, charge and cur- rent, energy and momentum, and s0 on, revealing the world to be much ‘more coberent and tightly-kait than had previously been suspected "There are, of course, forces in nature besides electromagnetism, the most obvious of which is gravity. Indeed, it was the simplicity of gravity that gave rise the ist conguerts of modern physics: Kepler’ laws of planetary motion, and thea Newton's laws unifying celestial mechanics wrth the mechanics of filing bodies, However, reconciling the sim plicity of gravity with relativity theory was no easy task! In seeking ‘equations for gravity consistent with his theory of special relativity, Binstein naturally sought to copy the model of Maxwell's equations. However, he result was not merely a theory in which ripples of some feld propagate theoxgh spacetime, but a theory in which the geometry of spacetime itself ripples and bends. Einstin’s equations say, roughly, that energy and momentum affect the metric of spacetime (whereby ‘we meatute time and distance) much ae charges and currents affect the electromagnetic field. This served to heighten hopes that much ot perhaps even all of physics i fundamentally geometrical in character, There were, however, severe challenges to these hopes. Attempts by Einstein, Weyl, Kaluza and Klein to further unify our description of the forces of nature using idene from geometry wore largely unsuccessful The reason ir that the careful study of atoms, nuclei and subatomic particles revealed a wealth of phenomena that donot fit easily into any Simple scheme. Each time technology peemitted the study of smaller Gistanceecaes (or equivalently, higher energies), new puzzles rose. In part the reason is that physics at small distance scales is completely omninated by the principle of quantum theory. The naive notion that f particle is'& point tracing out a path in spacetime, or that a fld assigns e number or vector to exch point of spacatime, proved to be ‘holly inadequate, for ene cannot measure the position and velocity Maxwell's Equations 5 ‘of particle simultaneously with arbitrary accuracy, nor the valve of & Feld and its time derivative. Indeed, it tured out that the distinction ‘etween a particle and field was somewhat arbitrary, Much of 20th cen- tury physics hne centered around the task of making sense of microworld and developing a framework with which one ea understand subatomic Dirtcles end the forces between them in the ight of quantum theory. ‘Ous current picture, called the standard model, involves tree forces: clectromagaetism anc the weak and strong nuclear forces. These are all gauge fields", meaning that they are described by equations closely ‘modelled after Maxwel’s equations. These equations describe quantum fields, so the forces can be rogarded as carried by particles: the elec: ‘romagnetic force is arred by the photon, the weak force is carried by the WY and Z particles, and the strong force is earred by gluons. There ‘ce also charged particles that interact with these foree-arrying par ticles, By ‘charge’ here we mean not only the electric charge but also its analogs for the other forces. There ae two main kinds of charged particle, quarks (which feel the strong force) and leptons (which do not). Allof these charged particles have corresponding antiparticles of the tame mast and opposite charge. Somewhat mysteriously, the charged particles come in three fam lies of ‘generations’. The fist generation consists of two leptons, the electron e and the electron neutrino ve, and two quarks, the up and down, or w and d. Mott of the matter we see everyday is made ont ‘of these fire-generation pactcles, For example, according to the stan dard model the proton is @ composite of two up quarks and one down, while the nestron is two downs and an up. ‘There is a second genera. tm of quarks and leptons the muon and mavon newtrino v, and the ‘charmed and rtrange quarks c, s. For the most part these are heavier than the corresponding particles in the fist generation, although all the neutrinos appear to be masslest or nearly xo. For example, the ‘muon is about 207 times as massive as the electron, but almost idea tical in every other respect. Then there isa third, still more massive teneration, containing the fa r and tan neutrino v, and the top and Fottom quarks ¢ and 8, Por many yeaes the top quark was merely con: jectured to exis, but just as thie book went to press, experimentalist Announced that it may finally have been found. Finally, there isa very odd charged particle in the standard model,6 1L1 Maxwell's Bquations the Higge particle, which ie neither a quark nor a lepton. This has not lheen observed either, and is hypothesized to exist primarily to explain the relation between the elecromagnetic and weak forces. ‘Even more pusalng than al the complexities ofthe standard model, however, is the question of where gravity Sts into the pictur! Einstein's ‘equations describing gravity do not take quantum theory into account, ‘and it has proved very dificlt to ‘quantize’ them. We thus have not fone picture of the word, but two: the standard model, in which all forces except gravity are described in accordance with quantum the- ‘ory, and general relativity, in which gravity alone is described, not in fecordance with quantum theory. Unfortunately it seems dificult to ‘obtain guidance from experiment; simple considerations of dimensional fnalysis raggest that quantum gravity effete may become significant at distance scales comparable to the Planck length, f= (ein, wher fis Planck's constant, x ie Newton's gravitational constant, and Cis the speed of light. The Planck length is about 1.616: 10-2 meters, far below the length reales we can probe with particle accelerators, Recent developments, however, hint that gravity may be closer to the gauge theories of the standard model than had been thought. Fascinatngly, the relationship also involves the study of nots in 3 dimensional space. While thie work isin its early stages, and may hot succeed as a theory of physis, the new mathematics involved is fo beatiful that itis eiffcult to reset becoming excited. Unfortu nately understanding these new ideas depends on a thorough mastery of quantusn field theory, general relativity, geometey, topology, and al igebra. Indeed, itis almost certain that nobody is suficiently prepared to understand these ides fully! The reader should therefore not expect to understand them when done with this book. Our goal in this book i simply to start fanly near the beginsing ofthe story and bring the reader far enough along to ase the frontiers of exrrent research in dim outline We mast begin by reviewing some geometry. ‘These days, when mathesnatciane speak of geometry they are usvally refering not to Bucidean geometry but to the many modern generalizations that fll Maxoell’s Equations 7 under the heading of ‘differential geometry’. The frst theory of physics to explicitly use diferential geometry was Einstein's general relativity, in which gravity is explained at the curvature of spacetime. The gauge ‘theories of the standard model are of avery similar geometrical charac- ter (although quantized). But there is also lot of diferential geometry lurking in Maxwell's equations, which after all were the inspieation for both general relativity and gauge theory. So, just aa a good way to ‘master auto repair is to take apart an old car and put in a new engine so that it rons better, we will begin by taking apart Maxwell's equations ‘and patting them back together using modern differential geometry In their classic form, Maxwell's equations describe the behavior of ‘90 vector fields, the electric field E and the magnetic field 3. ‘These felds are defined throughout space, whichis taken to be I However, they are also functions of time, x rel-valued parameter # ‘The electric and magnetic elds depend on the electric charge density _ which isa time-dependent function on space, and also on the electric ‘current density j, which ie time-dependent vector felon space. (For the mathematicians, let us note that unless otherwise epecifed, fune- tone are assumed to be real-valued, and functions and veetor fells on IR® are assumed to be smooth, shat is, infinitely diflerentiable.) In snits where the speed of light is equal to 1, Maxwell's equa- vee ‘There are» numberof interesting thinge shout these equations that sce worth understanding. Fist, there i the tte fact that we can only determine the direction of the magnetic feld experimentally if we know the diference between right and left, This is easieat to se from the Lorents force lav, which says that the fore on 2 charged particle with charge q and velocity 3s Fag(B40%d)8 LI Maxwell's Rquntions "To measure B, ste need only measure the force F on a static particle and divide by q. To figure out B, we can measure the force on charged prtiles with a variety of velocities, However, recall that the definition ‘Of the cross pradvct involves @ completely arbitrary right-hand rule! We typically define x B= (0yBs~ 04By,02Be — Bas 04Py ~ vB) However, this is just a convention; we could ave set x B = (0B, vyBosveBe — Ws BaryBa — veBy)s and all the mathematics of crs produets would work just ag well Tf we used thie left-handed cross product” when figuring out B from ‘measurements of F for various velocities 3, we would get an answer for B with the opposite of the usual siga! ‘It may soem odd that 5 depends on an arbitrary convention this way. In fact, this turns out te bean important clue ae to the mathematical structure of Maxwell's equations ‘Secondly, Maxwells equations aaturally come in two pairs, The pair that doce not involve the electric charge and curtent densities 5 3, 08 viba0 vxb+ Fano, look very much lke the pair that does . 508 vedap vx8- Note the funny minus sgn in the second pair. "The symmetry i clear ct in the vacuum Maxwell equations, where the charge and curreat Adensities vans = =, OB vino vx bs Ao, ved=0 vxb-2=0 ca Maxwells Equations ° "Then the transformation BoB BoB takes the first pair of equations to the second and vice versal This symmetey is called duality and isa clue tat the electric and magnetic fields are pat of a unified whole, the electromagnetic field, Indeed, if se introdice & complex-valied vector field f= 8 +58, duality amounts to the transformation oid, snd the coum Mall equation boil dons toto equations for € . off veaa xen ‘This trick bas very practical applications. For example, one ean use it to find solutions that correspond to plane waves moving along at the speed of light, whieh in the units we are using equal 2 J. fat bea sector in ond ets =f. Fin Be with SB tab. Show that Gua Berea sigs the vactum Manuel uations ‘The symmetry between E and HF does nt, however, extend to the rnon-vaciim Maxwell equttions. We can consider making p and Fcom- plex, and writing dove: viens vxboiE ea, However this amounts to introducing magnetic charge and current den- sity since if we split p and jinto real and imaginary parts, we see that10 1L1 Maxwell's Equations the imaginary parts play the role of magnetic charge and current den tier B= pet ion 7 = iti We get Vv B= om v-E=n "These equations are quite charming, but unfortunately no magnetic charges — so called magnetic monopoles — have been observed! (We vil have a bit more to sty about this in Chapter 6.) We could simply keep these equations and say that p and j are reak-valued on the basis of experimental evidence, But ite a mathematical ax well as a physical challenge to find a better way of understanding this phe- ‘bomenca, It turns out that the formalism of gauge theory makes it seem quite natural Finally, there is the connection between Maxwells equations and special relativity. The main idea of special eelatvity is that in addition to the rymmetriesof space (translations and rotations) and time (trans- lation) there are equally important symmetries mixing space and time, the Loventa transformations. The idea i thet if you and Tare both un sceelerated, 20 that my velocity with respect to you is constant, the Coordinates Twill naturally se, in which Tam at reat will differ from yours, in which you are at rest. If yout coordinate system is (t,2,942) land Tam moving with velocity » in the 2 direction with reapect to you, for example, the coordinates in which [am at rest are given by = (cosh 9}t~ (inde a = ~(sinh §)¢-+ (coahg)e vey where ¢ is & convenient quantity called the rapidity, defined so that tanh = v. Note the close resemblance to the formula for rotations in Maxwell's Bqvations a space. The idea is that just asthe 2, y, and z components of position are all just aspects of something more important, the position itself, ‘pace and time are just aspects of w unitary mbole, spocetime, ‘Maxwell's equations are invariant under these Lorentz transforma- tions — indeed, this was the main fact that Jed Einstein to special relativity! He realized that Maxwell’s equations predict shat any un tccelerated observer will meaure light moving in any direction in the ‘vacuum to have the seme speed. Mathematically speaking, the point is that if we bave a solution of Maxcwel’s equations and we do a Lorents transformation on the coordinates together with a certain transforms: tion of Bp and j; we again have a solution. For example, suppose that we do a Lorentz transformation of veloc- ity v in the = direction, a above. ‘The precise recipe for transforming the charge and current deities is pi = (coshd)o~ (sinh Pic Gh = —[sinh 6p + (cosh pie i = iv i= he Note that p and get mixed up together. In fac, we shall se that they are relly juss two aspects of a single thing called the ‘eursen: whieh has p as its component in the time direction and je, jysJe a8 ite components in the space directions ‘The formula for transforming the electric and magnetic felds wader the same Lorents transformation is somewhat more complicated Bae By = (cosh §)B, ~ (sinh 6)B, EL = (sink #)B, + (cosh #)B, By = Be By = (cosh 6)By + (sinh 9) BL = ~(sinh g)y + (cosh ¢)B. "The mort important mestage here is that the electric and magnetic fields are two aspects of a unified ‘electromagnetic field, Alko, we toe2 1 Maxsell's Equations that the electromagnetic field is mote complicated in character than the current, since it hae sx independent components that teansform in ‘more subile manner. It tums out to be a 2-form’. ‘When we have rewritten Maxwell's equations using the language of diferential geometry, all the things we have just discussed will be ‘much clearer — at leas if we succeed in explaining things well, The key step, which is tomewbat shocking to the uninitiated, isto work as touch as possible in a manner that does not require a choice of Coordinates, Afterall, at far as we can tll, the world was not drawn ‘on graph paper, Coordinates are merely something we introduce for ‘our omn convenience, and the lame of physics should aot care which Coordinates we happen to use. Tf we postpone introducing coordinates Uti itis actually necestay, me will nt have to do anything to show that Maxwells equations are invariant under Lorents transformations; it wil be manifest, Just for fun, let us write down the new version of Maxwell's equa: tions right away. We wil explain what they mean quite abit ater, #0 do not worty if they ae faniyeryptic. They are “=o ade = J. Here F's te ‘electromagnetic feld’ and Jit the ‘current, while the and + operators are slick ways of summarising all the curs, divergences and time derivatives that appear in the oldfashioned version, The {equation dP ~ 0 is equivalent to the fist pair of Maxwells equations, Awhile the equation ad F = J is equivalent to the second pair. The “funny minus sign’ in the second pair will turn out to be a natural consequence of how the + operator works. Tf the reader is too pragmatic to get excited by the terse beauty ‘of this new-fangled version of Maxwell's equations, let us emphasize that thie way of writing them ie a warm-up for understanding, gauge ‘theory, and allows us tostudy Maxwell's equations and gauge theory on ‘curved spacetimes, as one needs to in general relativity. Indeed, we will Start by developing enough diderential goometry to doa fair amount of physics on general spacetimes, ‘Then we will come back to Maxell’ futons, We wara the reader that the next few sections are not really Maxwells Bquations ry solid course in diferential geometry. Whenever tomething is at all tricky to prove we will skip i! The easygoing reader can take some facts on faith; the earful reader may want to get ahold of n good book on differential geometry to help fill in these detail. Some suggestions fon books appear in the notes atthe end of Past 1Chapter 2 Manifolds We therefore reach this rez In the general theory of relents, space and time cannot be defined in such ¢ tay thal differences ofthe sptial co. ordinates can be directly manzued by the anil meceuring-od, or diferent fn the time coordinate by a atendard elo. The mcthod hither employed Jor laying co-ordinates into the space time continuum ino define mannér ths breaks down, and there Seema to be no other way which would allow ut to adapt spree of co-ordinates fo the four-dimensional wiverie 80 that we might expect from these application 4 poticulrly simple formulation of the laws of nature So there ss nothing {for it but to reand all imaginable systems of ex-ondintes, om principle, as ‘equally stele for the description of nature, This comes to requiring tht ‘The general laws of nature are tobe expresied by equations wich hold ‘good fal systems of co-ordinates, that i ace covariant ith rerpect 0 Any subtitutions whatever (generally covasint), — Albert Binstesn In order to do modern physics we need to be able to handle spaces sand spacetimes that are more general than good old IR™. The kinds of spaces we will be concerned with are thoue that look locally like IR, but perkape not globally, Such a space is called an n-dimensional “manifold. For example, the ephere Payee looks locally like the plane I?, which is why some people thought the Encth was ist, These days we call vie sphere S?— the 2-sphere — to indicate that itis a 2-dimensional manifold, Similarly, while the space 1516 1.2 Manifolds 1 live in looks locally like IR, we have no way yet of ruling out the possibility that itis really S°, the S-sphere: wastes te ‘and indeed, in many models of cosmology spaces a 3-sphere, In such & “universe one could, fone had time, aall around the cosmos in a space ship jst as Magellan circumnavigated the globe. More generally, it is ‘even possible that spacetime has more than 4 dimensions, asi assumed in socalled ‘Kalusa-Kleintheoret. Fora while, string theorist sered (Quite sure that the universe mut either be 10 or 26-dimensiaal! More Dragenatcally, there salt of interest in low-dimensional physics, such 2s the behavior of electrons on tia fms and wires. Also, classical Iechanies uses ‘phase spaces that may have very many dimensions "These are some ofthe physical reasons why it i good to generalize vector ealeulus to that it works nicely on any manifold, On the other hand, mathematicians have many reasons of thei own for dealing with manifolds, For example, the st of solutions of an equation is often & ‘manifold (see the equation for the sphere abore). ‘We now bead towards a precite definition of a manifold. First of ail, we remind the reader that topological space isa set X together ‘witha family of subsets of X, called the open sets, required to satisfy the conditions: 1) The empty set and X itself are open, 2) U,V. X are open, ois UV, 8) Ifthe sets Uy CX are open, soit the union UUs ‘The collection of sts taken to be open is called the topology of X. [An open set containing a poiat 2 ¢ X is called « neighborhood of 2 ‘The complement of an open set is ealed closed, ‘A basic exemple is IR, where a ect U is taken to be open if for ‘every 2 € U, all points aufiiently clote to 2 are also in U Manifolds W Pig. 1. An open set in IR? ‘The use of a topology is that it allows us to define continuous func: tions, Roughly spenking, a finction is continuous if it sends nearby points to nearby points. The trick ie making the notion of ‘nearby pre Cine uring open sets. A function f:X —+ ¥ from one topological space to another is defined to be continuous if given any open set UC Y, the inverse image J-'U' C X is open. Fig. 2. A continuous function from X to ¥ Tone has not yet, one should do the fllowing exercise Exercise 2. Show that a function f:IR" — RM se continuous according to ‘the above definition sf ond only sf 2 ecconding (othe epsilon-delta defn ton: forall = € HU" ond lle» 0, there exits > O such that y ~2|< 6 implies f(y) ~ f(2) <« ‘The ides of a manifold i that, like the globe, we can cover it with patches that look just lke R". More precisely, we say that a collection18 12 Manifolds Uy of open sets covers a topological space X if their union ie all of X. Given a topological space X and an open set U C X, we define a chart to bea continuous fanetion gr U—» IR" witha continuous inverse (the inverse being defined on the set e(U)). GG x Fig. 3. A chart on X ‘As long at we work ‘in the chart we can pretend we are working in RY, just as the Baropeans could pretend they lived on IR? as long ts they did not go too far from home. For example, if we have a function f-U > BR, we can turn it into a function on RY by using fog Re sR. Fig. 4, Tarming function on U into funetion an IR al manifold, or n-manifold, Finally, we say that an n-dimer isa topological epace M equipped with charts yg: Ug —» R®, where Up are open sets covering M, such that the transition function va." ie smooth where itis defined, Such 2 collection of charts is called an Manifolds 9 atlas Fig. 5. Two charts and the transition function What docs this definition mean? Firs, every point of M lives in some open subset Uy that loks like R®, ori other words, me can “patch together’ the whole manifold out of bits that look like IR". Second, it means that we can tell using charts if a function on M is smooth, without any ambiguity, because the transition functions between chactt ace smooth, To be precite, we say a function f:M —> IR is smooth if for all a, fo gg':IRY — Bis smooth. Suppore you are uring the chart ga:Ua + IR and [am wing the chart gg: Up —» IR, and let V = U,10Ug be the overlap of our two charts. Suppose that you think the function f is smooth en V, that is, suppose fs yz! ir smooth on a, a8 below: fo SS Fig. 6. Your picture ‘Then I will agree that fis smooth on V, that is, fo yg! will be0 1.2 Manifolds smooth on ya¥ too Fig. 7. My picture Why? Because me can exprets my function in terms of your function and the transition finetion: Soug' = (Fo¢a!)o(eu0 v5") Strictly speaking, the sort of manifold we have defined ere is called ‘smooth manifold. There are alto, for eaample, topological mani folds, where the transition fonctions areonly required tobe continuous. For us, ‘manifold’ will always mean ‘smooth manifold’. Also, we wil always assume our maaiflde are “Hausdoe and ‘paracompact’. These fare topological properties that we prefer toavoid explaining here, which fate satisfied by all but the most bigarre and useless examples, Inthe following exercises we describe some examples of manifolds, leaving the reader to check that they eeally are manifolds. Exercise 8. Given a tpoopical space X and a oubset $ X, define the Induced topology e1 Sto be the topology in which the open sel are ofthe form UNS, here U is open in X- Let 5, the nsphere, be the it sphere om ) gern er Shaw that S* CR wath te induced topolgy is @ manifold Exercise 4. Show that if M 8 « manifold and U 38 an open subset of M, thon with it indad topology 1 « mensfold Manifolds a Exercise 5, Given topological spaces X and, we give XY the product topology in which set i open if and only if it 8 union of acts ofthe form U'% V, where Us open in X and V is open in Y. Show that f Mi fn redmensionel mania and N és an n-diensional marfol, A 8 ie tn (m+ )-dimenstonal manifold Exercise 6. Given topological spaces X and, we give XUY the disjoint tunion topology in which a set is open f and only sf it isthe union of fan open subset of X and on open subset of Y. Show that if MF end N fre mdimensionel manifolds the djint union MUN is an n-dimensional mania ‘There are many diferent questions one can ask abost a manifold, Dut one of the most basic is whether it extends indefinitely in all di zections like IR or i ‘compact’ like $. There it a way to make this precise which proves to be very important in mathematics. Namely, a topological space X is sad to be compact if for every cover of X by open sets Uy there ita finite collection Usy,-.., Uy that covers X. For smaniflds, there is an equivalent definition: @ manifold M is compact if and only if every sequence in M has a convergent subsequence. A basic theorem say thal a subset of IR" ie compact if and only if tie closed and Sts inside a ball of sufficient lange radius, "The study of manifold isa fascinating business in its own right. However, since out goal is to do physics on manifolds, let ua turn to the basic types of fields that liveon manifolds: vector fields and diferential formsChapter 3 Vector Fields ‘And itis «noteworthy fact that ignorant men have long bee in advance ofthe leared about vectors. Ignorant people, ike Faraday, neturally think ‘in vectors. They may know nothing of thei format manipulation, but if they Chink alout vectors, they think of ther as vector, that ie, dietted magn tudes No ignorant man could or would think abot the thre components of a wector separately, end disconnected from one another. That i @ devi of leaned mathematicians, lo enable them to ede vectors. The device i often asf, especially for ealalating purposes, but for general purposes of reatoning the manipulation ofthe selar component instead of the vector ‘self is entirely wrong. — Olver Heniside leavin was one ofthe frst advocates of modern vector analysis, as well ata very sarcati fellow. In the quote above, he was making ‘the point that the great physicist Faraday didnot need to worry about coordinates, because Faraday had a direct physical understanding of vectors. Pietorally, a vector feld on a manifold can be virualized a8 afield of arrows. Por example, a vector field on 5? ie basicaly just a field of arrows tangent to the aphere a2% 113 Vector Fields SS Fig. 1. Vector field on $? “Todocaleulations with vector felds, owever, it is nice to define them in an algebraic sort of way. The key to defining vector felds on manifolds is to note that given Seld of arrows, one can differentiate a function in the diection ofthe arrows, In particular, given a function f and 2 vector field v on IR, we ean form the diectional derivative off in the Airetion v, which we will write simply as vf. let us write a formula for uf in this case. ‘The formula for « di- rectional derivative sbould not be news to the readers of this book, ‘but we wll rewrite i uring some sick physics notation. We will write 21,...,2" for the coordinates on IRM, and rite just 0, forthe partial derivative 9/02". (When we are dealing with three ot fewer dimensions ‘we ill sometimes write z,y,2 instead of 21,2?,2%, and write By, By de for 8,,3,35) Also, ve will use the Kinstein summation convention and aivays sum over repested indices that appear once asa subscript tnd once ae a superscript, Then ifv has components (v!,...,v"), we have the formula vf = ays. If this toems enigmatic, remember that it is just shor! for vol pF eo Infact, since the formula vf = v0, holds for all f, we can be ‘even more slick and write vy Vector Fields 2% What docs this mean, though? ‘The sight of the partial derivatives 8, sitting there with nothing to differentiate is only slightly unnerving ‘we can always put a function f to the right of them whenever we vant. Much odder is that we are saying the vector field isthe linear combination of these partial derivatives. What we are doing might bbe regarded as rather sloppy, since we ace identifying tro diferent, although related, things: the vector eld v, and the operator that takes a directional derivaivein the direction of v. Ia fact, this‘sloppy” attitude rns oat to be extremely convenient, and next we will go eve farther and nse it to define vector feds on manifold. It is important to realize that in mathematic tis often crucial to think about familiar ‘objects in a new way in order to generalize them to a new situation Now lt us define vector Fields on # manifold M. Following the phi losophy outlined above, thes will be entities whose sole ambition in life is to diferentiate functions. First a bit of jargon. The set of smooth (esl-valved) fanetions on a manifold IY is written C™(M), where the (Gis shot for ‘having infinitely many continuous derivatives’. Note that C™() is an algebra over the zeal numbers, meaning that itis closed under (pointwise) adétion and multiplication, as well ar mult plication by eal numbers, and the following batch of rules holds: fre = otf S+(a+h) = (fa) +h Sak) = (Fa) fig+h) = forth (Fgh = fhsoh yas Bf) = (a8)f of +9) = af +09 (oa = af tof, here f,g,k € C™(AM) and a, € IR. Of course itis a commutative algebra, that is, fo = af: Now a veetor fleld » on fis defined tobe. funetion from C= (M) 10 C™(M) satisfying the fllowing properties:6 1.3 Veetor Fields wf +9) = v(N)+H19) was) = an(s) wlfa) = w(Pa+ ola), for all f,g € C™(M) and a € IR. Here we have isolated all the basic rules diectional derivative operator should satisly. The fest two simply amount to linearity, and it is the thied one, the product rule or Leibniz law, that realy captures the essence of differentiation "This definition may sccm painfully abstract. We wil see in bit that it really it just © way of talking about a Held of arrows on Mf For now, note the main good feature ofthis definition: st does not rely ‘on any choice of coordinates on MA basic philosophy of modern physics is that the univere does not eame equipped with 2 coord hate system, While coordinate eyatems are necessary for doing specific concrete calculations, the choice ofthe coordinate aystem to use is a ‘matter of convenience, and ther is often no “best” coordinate system. One shoald strive to write the laws of physica in a manifestly coordinate independent manner, o one ean see what they are rally snying and rot get disteacted by things that might depend on the coordinates, Let Vect(M) denote the ret of all vector felds on IM. We leave it to the reader to check that one can add vector fields and multiply them by fonctions on M as follows. Given va © Vect(M), we define » +22 by (WKN =o) +w), and given v © Veet(M) and 9 € C™(M), we define gu by (oo}(F) = a0) Exercise 7. Show tht + and owe Veet( i). Exercise 8, Show thatthe following ruler for all vw € Voet(Mf) and fge cma) Mote) = forte Uta = forse (ay = 9) ‘Tangent Vectors 2 (dere “1 denotes the constant fnction equa 1 on all of M.) Mather tealy, se summarize theve rules by saying tha Veel(M) 8 a module over oman) 1 turns out that the vector Belds {@,) oa IR” span Veci(IR") as ‘module over C™(A), In other words, every vector field on EI it @ linear combination of the form Dy = VIB HUD, for some functions w* © O=(IR"). It is alo true that the vector fields {0,} 00 B® are linearly independent Exercise 9. Show that f 8, = 0, that ss, 20,f = 0 forall f C™(HY), we must have v= 0 for all. "This impliesthat every vector feld von RY has a unique representation ‘se linear combination vy, we eay that the vector fields {0,} form ‘basis of Veci(IR"). The functions ware called the components of the vector field. ‘Tangent Vectors Often isis nice to think of a vector field on M as really assigning an ‘acton toeach point of M. This kindof arrow i called a tangent vector. For example, we may think of a tangent vetor at point p © S? as a vector in the plane tangent to Fig. 2. Tangent vector2% 1.3 Vector Fields ‘To get a precise definition of a tangent vector at p € M, note that a tangent vector should let us take directional derivatives atthe point p Forexample given a vector field won M, we can take the derivative o(f) of any fenetion f € C™(M), and then evaluate the nection wf) at p ‘We can think ofthe result, »(f){p), ax being the result of diferntiating {fin the direction x," athe point p. In other words, we can define yp 0™(M) > R by elf) = NI), and think of vp as a tangent vector at p. We cal vp the value of vat p. [Note that, has three basic properties, whieh follow from the def nition ofa vector field wAlS +9) = Lf) + 08) sila) = ag) self) = e9(F)alo) + Fedo). Henceforth, me will simply define a tangent vector at p € M to be ' fanesion form C™(Af) to IR satisfying those three properties. Let ‘TyAf, the tangent space at p, denote the set ofall tangent vectors at peM. It now follows rigorously fromm our definitions that foreach p< M, ‘vector feld » € Vect(M) determines a tangent vector vp © TyM. One an aio shom, though it takes a bit of work, that every tangent vector ap is of the form ty for some vector field or other. A related fact, ‘which t much easier to show, isthe following Exercise 10, Let vw € Vee{M), Show that w= w if and only ifm = wy forall pe M. Why de tangent vectors as we have defined them ‘look lke atoms”? Fit of all, we can adel two tangent vector vu © Ty by (+ WA = oN) + Cf) and sultiply tangent vectors by eeal nuraber: (avis) = anf) ‘Tangent Vectors 2 (Now we are using the letters v,w to denote tangent vectors, ot vector fields!) With addition and rultiplieation defined this way, the tangent space is really a vector apace. For example, in Figure 2 we have drawn ‘tangent space to look like a litle plane. The tangent vectors can be thought of ax arows living inthis vector space Exercise 11. Show that TyMf ia vector space over the ral numbers. Another reason thy tangeat vectors realy look like arrows is that curves ave tangent vectors: M Fig. 3. The tangent vector toa curve ia [By a.curve we will always mean a function from IR tore interval to M that is smooth, i,, mich that for aay f © C™(M), ftr(8)) depends smoothly on ¢, Given a curve 7:R + M and any £€ I, the {tangent vecor (2) should bea vector inthe taagent space Tyo M_ We define (2) in the only sensible way posible: itis the function from (C™(AF} to IR that sends any function f € O=(M) to the derivative Srey, In other words, the tangent vector (diferentes fnetions in the direction that in moving ina ime rise 12, Check tht 7) € Ty M sing the definitions TF the curve 7 describes the motion of» partie though apace, the tangent vector 7(¢) represents its velocity: For this reason, we will sometimes write 4 a0 1.3 Veetor Fields for 7(t) especially when we ae aot pastcalaly concerned with which value of ¢ ve ace talking abot Note that for manifolds it generally makes no sense to say that a tangent vector © € TyM is ‘the same’ as another one, w € TyMf, unless the points pand gare the same. For example, there iso “best” way Lo ‘compare tangent vectors at the north pole of 5? to tangent vectors at ‘the equator, It also makes no sense to add tangent vectors at diferent points rn Ss Fig. 4 Tangent vectors at diferent points of S* ‘We mention this becaise the reader may be used to IRS, where one often says the following two vectors are “the same’, even though they fare at different points in IC _ a Fig. 5, Tangent vector at diferent points of IRY "The reason why one can got away with this that for any point p in Covariaat Versus Contravariant 3 IR, the langent veetors (Only € TR, form a basis. ‘This allows one to relate tangent vectors at diferent points of IR” — one can sloppy say that the vector Oe © TRO and the veetor wade € Te are ‘the same’ if o" = wo, even though v and w are nt literally equal Later we vill get a deeper understanding of thie ise, which requires a theory of “parallel transport’ the proces of dragging a vertor at one Doint of a manifold over to another point. Ths turns out to be a crucial idea ia pisies, and infact the oot of gauge theory! Covariant Versus Contravariant [A lot of modesn mathematics and physics eaquires keeping track of ‘whieh things in life are covariant, and which things are contrarian. Let us begia to explain these ides by comparing functions and tangeat vectors, Say we have a function #:M + NV from one manifold to Another. If we have a reslvalved function oa N, say f2N +R, we ‘can get a real-valued function on Mf by composing it with f \e oe Fig. 6, Pulling beck f fom N to MF2 1.3 Vector Fields ‘We call this process pulling back f fom N to M by 4. We define of = fos, and call °F the pullback of f by 4. The point is that while ¢ goes orwardt irom M to NY, the pullback operation #* goes ‘backwards! taking functions on to functions on AY. We say that real-valued functions on manifold are contravariant because of this perverse backwards beh Exercise 19, Let 6:8 — be given by lt) = eh Let = be the url coordinate faction on R. Show that 2 = Exercise 14. Let: BR? be rotation ounterclockwise by an angle 8 Let a,y be the usual coordinate factions on 2. Show that Ge = (comme ~ (6nd oy = (sind}e + (omayy By the way, we say that M+ N is smooth if f © O(N) imple that $7 € 0°(8), Henceforth we wil eve fonction fren ‘mantels fo manifolds are smooth oles othervive tated ad we wil chen eal sah functions nape Exercise 18, Show that thi definition of smoothness is consistent with the previons definitions of oath functions fr +R. and smooth career Row, Using our new jargon, we have: given any map aN M, palling back by ¢ is an operation $:07(M) + 0™(N). ‘Tangent vectors,on the other hand, are covariant: a tangent vector 1» € TyMf and a smooth function @:M — IV gives a tangent vector (00 € TyyN, called the pushforward of» by 8. This is defined by (oN = 0") CCovariant Vereus Contravariant 38 ‘We say we are pushing forward v by 4. Note that we use a subscript asterisk for pushforwards and a superscript for pullbacks! One way to think of the pushforward is that iy sa curve M with tangent vector (0) € Ty(A), the curve $07 isa curve with tangent vector (Boalt) = d6l0'(8) € Ta: — Fig. 7, Pushing forward the tangent veetor of a curve from M to N Exercise 16. Prove that (809}(#) (0). Exercise 17, Show that the pushjowend operation GTM Tay is nar Exercise 18, Show that if 6M +. se can pus forward a vector ld ‘rom AF to obtain o vector lad on satisfying (onde = ol) sahenever dle Exercise 19. Let 6: R? — IR be rotation ounteclctuise by on angle 8 et nd be the coordinate vector fields on EE. Show that at any point of n 4.00 48, (con0)0, — (an 98, (sin 8. + co08),4 1.3 Vector Fields [tis traditional in mathematics, by the way, to write pushforwards and other covariant things with lowered asterisks, and to wite pull backs and other contravarant things with raised asterisks, It might help ar mnemonic to remember that the tangent vectors dare wrt- ten mith the downatars, and are covariant. Tn the next chapter we will discuss things vimilar to tangent vectors, but which are contravari- Gant! Thete things wil have thie indices upstairs. We warn the reader, however, that while the vector field 0, is covariant and has its indices dovenstair, phyicits often think of a vector field» as being its campo- nents o*, These have their indices upstairs, so physicists say that the 1" ace contzavasiant! This is one of thos ite differences that makes Comenunication between the two subjects a bit more dfficlt, Flows and the Lie Bracket (One sor of vector fl that comes up in physic is the vlaty vector fed of ud, such as water, Imagine thatthe velocity vector Bld» ir contot asa feelin a ine that each molecule of water traces tut a curve (¢) a time pane, ith the tangent vector oy equal to the valu of atthe pint ve for all Ifthe curve state at some pont p€ M, that i (0) fall the integral curve through p of the veor eld \ Sak we VAM ty an tly \t Fig. 8. Integral curve through p of the vector feld “40 Flows and the Lie Bracket 35 Calculating the integral curves of a vector feld amounts to solving 1 frs-orderdiferental equation. One at to be careful, because the olution might ‘shoot off to iafinity® in a nite amount of time: Exercise 20. Let v be the vector field 298, + y@y on R¥, Calulate the {integral carves (2) and so which ones ae defined for all We say that the veetor field» is integrable if all the integral curves sre defined for allt Suppose v i an integrable vector fel on M, which we think of as the velocity vector field of some water. If we keep track of how all the molecules of water are moving alg, we have something calle a ‘flow’. et gulp) be the integral curve of v through the point p € M. For each time t, the map cM Mt turns out tobe sath, by a result on the smooth dependence of s- lutions of ferential equations on the iit conditions. Water tht ves at pat time zero will be at f(y) by time ¢, 4 we cll she family ‘maps (4) the flow generated by 2. The desing equation for the ow is (cewtng our equation or 1) 4 Bh) = Yat Exercise 21. Show that dy isthe identity map ideX + X, end that for All 6,4 € R ws have $40 4,= dvs ‘There isan important way to get new vector felds from old ones ‘that ie related to the concept of flows, This is called the Lie bracket or ‘commutator of vector fields, Given v,w € Vect(IM), the Lie bracket [oyu is defined by lo. wit for all f€ C™(M), or, for short, (w() = wh0A A,36 1.3 Veetor Fiede Let us show thatthe Lie bracket defined inthis way actually is vector field onthe manifold M. Itiseasy to prove linearity, so the crucial thing ia the Lebnia rule: we have ufo) = (ow— ww fa) ve Fla + Fela) ~ wi Aa + Foto] ou(f)9 + fea) — wo f)0— fwvta) ule + Fula) Fiore we used the Leibnis lar twice and then used the definition of the Lie brackets, ‘The Lie bracket measures the failure of ‘mixed directional deriva: tives! to commute. OF coarse, ordinary mixed partial derivatives do (m8 We can think of this pictoraly, as folloms: flowing a little bit frst in the 8, direction and then in the 8, direction gets us to the same place tan fe had done it in the other order: Fig. 9. [2,24 However, if we take some other vector fields, this does not usually work: Plows and the Lie Brocket 3 Fig. 10. (oy) #0 ‘We say in this case that the vector fields do not commute, Exercise 22. Consider the normalized vector feds im the r and 8 dive tons on the plane in polar coodinates (not defined at the origi) 28,4 90 Jest 8, — v0 vere Catealte ‘To make the relationship with flows precie, suppose that v gener- ates the lm dy aad w generates the flow yi. Then for any f € C™(M) (ne) = £4400) and siniasly (wie) = L100), 0 one can check that lo, a} /NC@) = grag fale) — FC 60M) If you think about it, tht is related to what we said above, In ‘(de(9,(2)) we are starting at p, Sowing along wo little bit, then along ‘valitiebit, and then evaluating f, while in f(94(4(p)) we are flowing fret along » and then w. The Lie bracket measures (infinitesimally as it were) how these flows fil to commute!8 1.3 Veetor Fields Exercise 23. Check the epation above ‘The Lie bracket of vector fields satisfies yore identities which we will come back to in Part Il. For now, we amply let the reader prove them: Exercise 24, Show that for all vector felis uo, om & manifold, and all real members and fi, we hae: r cafes} + Als identity (+ oa (wfuell=0. Chapter 4 Differential Forms ‘As a herald it's my duty to explain those forms of bnuty. — Goeth, Fant, 1-forms ‘The electric il, the magnetic field, the electromagnetic field on space time, the curent — all these ate examples of diferential forms. The aradient, the cul, andthe divergence can all be thought of ae diferent aspects of single operator d that acts on diflerential forms. The funda ‘mental theorem of calculus, Stokes’ theorem, and Gauss" theorem are all special eases of a ingle theorem about diferental form. So while they are somewhat abstract, diferential forms are a powerfl unifying notion. ‘We begin with forms, Our goal is to generalize the concept of the gradient of function to functions om acbitrary manifold, What ‘we will dois to make up, foreach smooth function fon M, an object called of that is supposed to be like the usual gradient Vf defined on IR". Remember that the ditectional derivative of a function f in the fom IR" in the direction vis just the dot product of Vf with v Viv =f. In other words, the gradient of f it a thing that keeps track of the irectional derivatives off in ll diretions, We want ou a’ to do the same job om any manifold M, 940 14 Differential Forms ‘The gradient ofa function on IR" it a vector field, so one might want to say that of should be a vector field, The problem is the det product in the formula above. On IR ther is well-established way to take the dot product of tangent vector, but manifolds donot come pre-equipped with a way to do this. Geometers call a way of taking dot products of tangent vectors 4 ‘metric’. Infact, we will tee that in general relativity the gravitational field is decribed by the metricon spacetime. Far from there being a single best” metric on a manifold, there are typically lots that satisfy inatein's equation of general relativity. This makes nice to-avoid using a particular metricuness we actually need to. Therefore ‘ne will ot thiak of df ar avector field, but as something ele, a T-form’. The trick isto realize what Vf is doing in the formula Vf-2 = vf. For each vector field» that we choot, this formula spits outa function vf, the directional derivative of f ia'the direction v, In other words, what rally matters isthe operator whe, ‘or, what isthe same thing, wees, Let us itolte the essential properties ofthis map. ‘There is really ‘only one: linearity! The means that UE (etu)= Vf et Vf-w for any vector fields v and w, and VI -(o») =a(VF-0) where g it any smooth funetion on IRM. Since we can pull out any function 9 © C™(IR®) in the above formula, ot just constants, math fematiciane say that wo viy linear over C-(IR") — not just lincar over the real numbers. So, abstracting bit, we define a I-form on any manifold M to be 1a map from Vect{ MM) to C™(M) that is Knear over C™(M), Jn other forme a words, if we feed a vector field sto a I-form a, it spits out a function o(e) in a way satisfying ote +m) fo) +04 $4o0) = 9440. We wie O'(M) to denote the space of all forms on a manifold M. Later on we wil talk about form, forms and 3 on “The bute example ofa Lforn is thin for any mooth function f con M there ta for df defined by aw) = ¥f. (Chink ofthis asa slick way to write VF really a l-farm, we just need to check linearity: vf.) To show that df is Mote) and f(g) We call the I-form df the differenti off ust as we can add vector fields or multiply them by functions, we ‘can do the same for Informs. We ean add two Iforms w andy and get Horm 2 +p by defiing off) = 9af(o) of f,0F the exterior derivative (e+ slo) = (0) +a00), snd we can multiply a Lorm o by a smooth function f and get a form fw by defining (Fe}(o) = foto) Exercise 25. Show that w 4 ond fis anv erally Lforms, i, show Kine ceanty over C=(M). Exercise 26. Show thet 9"(M) is a moule over C™(M) (0 the define Sion in Esereize 8.)2 14 Differential Forme ‘The map 4 C%*(M) + (AM) that sends each function f to its Aiferental df is also called the diferential, or exterior derivative. It is Jnterestng in ite own right, and has the folowing nice properties Exercise 27. Show that aseo= a +e les) = adf Gs an = sah + 9an Aia)= 14994, for ay f.g, hE C(M) and any a € ‘The frst three properties in the execie above ar jut forms of linearity, but the lst ane sa verton ofthe product ule, ot Leibniz tee alfa) = Fda + 94f. Ii isthe Leib law thal males the exterior derivative relly at ike 2 derivative, a if you only want todo part of Exerciae 27 check that the Let law hold! Ts worth mentioning, by the wa, that when Leiba was inventing eal fe Sort guessed that fg) — afd, and cl got right the next da. Tn fac, the vender has seen diferentials befor, in calculus. "They start out as part ofthe expreons for diferentation ay de ad integration [feyee tut soon take on a mysterious if of thrown, ain sin = conde! We bet you remember wondering what the heck these diferentials really dare! In physics one thinks of de as an “ifisitesimal change in position’, fd eo on — but thie ie mystifying in i own right. Parly in the history of ealeulus, the philosopher Berkeley complained about these forms 49 infinitesimale, writing "They are neither finite quantities, aor quanti ties infinitely small, nor yet nothing. May we not call them ghosts of departed quantities?" “More recently, people have worked out an al- ternative appronch to the real numbers, called ‘nonstandard analysis’, that includes a logically satisfactory theory of infnitesimals — puny numbers that are greater than zero but less than any ‘standard real number. Most people these days, however, prefer to think of diferen tials ae forme let us show that dsin2 = coszde i really true as an equation concerning Iforms on the real ine. We need to show that no matter vwhat vector feld we feed these two form, they spit out the same thing. ‘This is uot hard. Any vector feld v on Bis of the form v = §(2]0e, 80 00 oxe hand we have (asin 2)(0) sod on the other hand: (2)Besinz = flz)coss, (cous ds)(v) = (cone) (2) = f(2) on 2 0,2 = f(2) cons. ‘This isin fact just special case of the following Exercise 28. Suppose f(2...,2%) ia function on R Show that = af ae" ‘This means that on IRM the exterior derivative of a function is really jnat a diferent wey of thinking abot it gradient since ia old-fashioned Tanguage we had VF = (Osha Daf) ‘To do the exercise above one needs to se the fact that the vector fee (3) orn a basi vector flds on BU Ln ct, thi implis tat ‘he {fr (2"} form a basin of I-frms on R, The yi that 10) = 0.2" = 8 where the Kronecker delta 6 equals 1 if. = v and O otherwise. Now ssuppore weave a Inform on IR*, Then we ean define some functions = (0“ 1 Differential Forme and we claim that = yd ‘Thin wil imply that the forms {do} span the I-forms on IR". To show that w equals yds", we just need to feed both of them a vector field and show that they spt out the tame function! Feed them v = 8, for example. ‘Then on the ove hand ot) while on the other hand, (oA) (onda )(0) = (onda 0B) using the fact that d2¥(@,) = 8 ‘We leave itso the reader to finish the proof that the I-forms {da} form a basis of 0°(IR") Exercise 29, Show thatthe Mforms (det) are nearly independent i, a ada = 0 the all the factions uy ve 20. Cotangent Vectors Just asa vector field on M gives a tangent vector at each point of MY, 2 ‘form on M gives a kindof vector at each point of M called a 'otan- gent vector, Given a manifold M and a point p € M, a cotangent Vector w at p it defined to be a linear map from the’ tangent spa pM to IR. Let TAF denote the apace of all cotangent vectors at p. or example, if we havea L-form w on M, we can define « cotangent vecior up € Ty by saying that for any vector field v on M, (0M). ere the right-hand side stands for the function w{u) evaluated at the point ale) Cotangent Vectors 4 Exercise 30. For the mathematically inclined show tht the wp i really swell defined by the formula ebove. Thetis, show that w)(p) really depends ‘nly ott not on the values of vat other points Alo, show that form {s dtormned by sts values at points. In other words, ifs,» are tao forme om M with up = vp for every point p © M, then =» y S Fig. 1. A picture ofthe cotangent vector (df)p How can we visualize a cotangent vector? A tangent vector is like 4 little arto; it points somewhere. A cotangent vector dors not. A nice heuristic way to visalie a colangent vector is as a little stack ‘of parallel hyperplanes. For example, if we have a function f on a ‘manifold M, we can visualize df at a point p € IR? by drawing the level curves of fright near p, which look like litte sack of parallel lines. ‘The picture in Figure l is twordimensionl, ao level surfaces are just contour lines, and hyperplanes are just lines, "The binge df is, the more tightly packed the hyperplanes are. When wwe take tangent vector » © TyM, the number df(v) basically just Counts how many litle hyperplanes in the stack df the vector v croses. In Figure 2 we show a situation where d/(v) = 3. By definition, of course, the number df(v) is just the drectanal derivative v(6 14 Difforental Porme Fig. 2. df(v) =9 Actually we must be a bit careful about thinking about df(v) in terms of pictures, because it could be negative! If we think of the litle stack of hyperplanes at ‘contour lines’, we should realy count the ‘numberof them v crosses with a plus sign if is pointing ‘uphill and ‘minus sign ft ie pointing “downhill y Ss 1 Fig af(-) If this way of thinking of I-forms is confusing, foe fee to ignore it — Dut people with a strong taste for Visualization may find it very handy. Now let us explain precisely what we mean by I-forms being dual to veclor field, First ofall, given any vector space V, the dual vector space V" is defined to be the space of all Knear functionals w: VR, Tn pasticular, the cotangent space 77M isthe dual of the tangeat space T,M. More generally, f we have aliuear map from one vector space to Cotangeat Vectors fa another, fVvaw, we automatically get a map from W* to V*, the dual off, written wove and defined by (Fey) = w(F0). ‘Thus the dual of a vector space isa contravariant got of beat: linear maps between vector spaces give rie Lo maps between their duals that so ‘backwards’ Exercise 81. Show thatthe dual ofthe identity map on a vector space V fs the identity map on V". Suppose that we have lincar mape f:¥ -» W and GW +X. Show tha (9f)* =P" ‘This means that cotangent vectors are contravariant. In other vrords, suppose we have aap ¢:M + from one manifold to another ‘rth dp) — a. We saw in the Inst section that there it Knear map 821M > TWN. ‘This gives a dual map, which we write as ¢* going the other way OTN TM, Hw in a cotangent vector at (2), we call gw the pullback of w by ¢. Explicitly, if v € TpM and & © TyN, we have (oy) = 060 We can alto do this ‘pulling back’ globally. That is, given a L-form wo on N, we get a I-form g'w on M defined by (ele = lee) where (7) Exercise 82, Show thatthe plllack of 1forms defined by the formal above relly esits and is unique“3 14 Differential Forms Recall fom the previous ection that we can also pullback functions ow N to functions on M when we have a map 4M —+ N. Taere is ‘marvelous formula saying that the exterior derivative is compatible ‘with pullbacke, Namely, given a function fon Nand a map: Mf —+ N, wre have eh =d9"N. ‘Mathematicians summatie ths by saying that the exterior derivative is natural, For example, if @-IR™ —+ IRV isa difeomorphism repre senting some change of coordinates the above formula implies that we ‘can compute dof a fonction on IR either before or after changing co frdinates, and get she same answer. (We discuss this a it more in the hnert section.) So naturalty ean be rogarded as grand generalization of coordinate independence, "To prove the above equation we just need to show that both sides, smbich ae I-forme on A, give the sarne cotangent vector at every point pin M: CMe = (ED ‘This, in turn, means that ($A) = (AS De) for all v € TpM, To prove thie, work out the left hand side using all the definitions and show it equals the sight hand side: (HUAN) = (AAG = ni) ul8" Ao) (oD) "To make thie more concrete it might be good to work out soane examn- ples: Exercise 89, Let g:IR — R be given by 4{t) = sa. Let de be the wal form on R. Show that duds = costa Exercise $4, Let ¢:R? ~ RP denote rotation counterclockwise by the tangle 9. Let da dy be the noua Basis of forms on RU. Show that Side = contde—snddy Ody = sindde +eonldy Change of Coordinates 49 ‘The formula ee) = aes) isa very ood reason why the differential of function has to be a Lformn instead of a vector fel. Both functions and forms are contravariant, soif ge —+ W and f € C=(N), both sides above are I-forms on N. IF fone tried to make the diferetial of a function be a vector fel, there Srould beno way to write down a sensible formla ike this, since vector fields are covariant. ("Ty it!) Change of Coordinates Indeed, from childhood we have bicome fomilar wih the appearance of phys ‘eal equations in non-Cartesien systoms, suc as polar coordinates, end in hhor-terial systema, such a relating coordinater— Steven Weinberg {The introduction of numbers a condinaes [.] stan ect of violence whose only practical vindication i the special clculstory manegeiity ofthe or ‘binary number contin with sl four baie operations. — Hermann Weyl So far we have been avoiding coordinates as much as possible, The reason, of courte, ie that the world does aot come equipped with co- ordinates! Ae far as we can tel, coordinates are something we impose ‘upon the world when we want to talk about where things are. They are extremely useful, and in many applicatins quite essential. Unfor tunately diferent people might pick diferent coordinates! Sot is good to know how the components of a vector fed or l-form depend oa the coordinates used First let as describe how one can use coordinates locally on any manifold to work with vector fields and differential forms. We described ‘the basic iden back in Chapter 2: given an r-dimensional maniflé M, a chart is adifeomorphism ¢ from an open set U ia M to R. This allows us to do turn calculations on U into calculations on IR™ or example, we can use p to pull back the coordinate functions 2 from IR" to U. Tnstend of calling these functions w"2" as one really Should we usually all them simply 2. This isnot too confusing as long fas we know we are “working in the chart” yi U7 > I. The functions0 14 Differential Forms 2 on U are known at local coordinates on U. Any function on U Can be written ara function f(e!,...2%) of thee local coordinates. ‘Similarly, the coordinate vector Feds d, area basis of vector fields ‘on IR, and we may push these forwards by y-! to a basis of vector felds on U. As with th local coordinates, people usually denote these vector fel simply a 0, These are called the coordinate vector fields asociated to the loal coordinates 2" on U. One thus writes any vector field v on U a8 = 0", In the ame way, the coordinate I-forms dat are a basis of forms on IR, which we may pull back to U by g, obtaining a bass of I-forms on U. "These aze called the coordinate I-forms associated to the local coordinates 2M Those are written simply as do Note that our use of 2" and de? to denote functions and Iforme on U, mbile sloppy, is Exercise 35. Show thatthe coordinate I-forme 2 realy re the dior ental of the locel coordinates 2" on U We can write any Iform w on U a8 ad We should emphasize that it is bad to think of veetor fields v of ‘forma waa being thee components of. Instead, we should think cof them as having components, which depend on the basis used. Por ‘cxample, the usual coordinate functions 2',,..,2" on IR” give a basis {Gy} for Vect(IR"), Given any vector field v oa IRM, F can write it uniquely as van ‘where the ware functions on IR", But ruppose you chose some other ‘coordinates on IR” — that is, rome functions 2,...,2™ on IR™ such that {2,} was another bass for Veci(), Then you would write vav"e, ‘The vector field isthe same in both cases — it i blissfully unaware of which coordinates we mere mortals aze uring, But its components (Change of Coordinates a depend on a coordinate system, and for us to talk to each other, we eed to know how your components, v, are expressed in terms of mine, Fins, since your vector fields form a bass, we can express mine as linear combinations of you a= Tea, vwhere the Ty are a matrix of functions on IR". It is not too bard to figure out these functions. Just apply both sides ofthe equation, which ate vector fields to the coordinate function 2 Oya = Trea" "The partial derivative fais just the Kronecker delta 63, 10 actually swe jst have aa? = 1} We can write this out somewhat more impressively as follows aa ~ oe ‘This implies that _ = Se. ‘Thea, to express the components v'™ in terms of the components wo stant with the fact that v0, = 080, and une the equation above to get = we, Equating coefficients, we get Now we caa talk to each othet! In short, to translate from my com: ponents to yours, T simply multiply by @ matrix of partial derivatives ‘corresponding to the change of coordinates [Horm work the same way, and we leave them as an important ‘exercite fr the reader:2 14 Differential Forme Exercise 36, Inthe situation above, show that =o ‘Show that for any Hformas on B®, writing ae ade" = hae, {your componente, are related to my componente wy by ‘There is am interesting distinction between ‘active’ or ‘passive’ co- ordinate traneformations. A passive coordinate transformation is & ‘change of coordinate functions (on IRS, or on a chart), which is what wre have just been considering. We are not moving poiats of our space around, just changing the funetions we we to describe them, An active Coordinate ransformation it just another name for a diffeomorphism MM it moves she points of AF around, We can push vector felds forwards by a diffeomorphism, and pull funetions and I-forms back. It is nice $o know how these look in the special case of IR* (ora chart), Not surpris ingly, the formulas look similar to the formulas for passive coordinate transformations that me have jut derived! ‘There is, however, something abit tricky about this business. The simplest example of thi rckiness occurs wien people in certain places ‘witch ftom standard time to daglightsoving time in the spring. ‘The rmaemonie formula is spring forward, fll back’. Tis is supposed to feanind you to set your elock forward inthe spring and back inthe fal. "The hard partis remesnbering what setting a clock orwacd’ means! Is one supposed to move the hour hand toa late time, so one has to wake tp earlier than one otherwite would? Oris one supposed to move the hhour hand to an earlier time, 29 oge can slay in bed late? Note that jt takes a clock and a point in time to give a number that we call the ‘hime’ t. More generally, it takes a cootdinate system fogether with point in spacetime to give a number. Changing the coordinate system Change of Coordinates 33 ‘one way has a similar effect to moving points of spacetime around the ‘opposite way Let us now consider theelfect a map g:IR™ —+ IR" has on coordinate vector fields and I-forms. If = m and @ isa difeomorphism, this is sn ‘active coordinate transformation’, but itis actually easier to keep things straight if we work in the general cage. Write 3!,...,2™ for the coordinates on IR, and 2,...,2" for the coordinates on IR" First note that we can pull bck the coordinate functions 2” on IRY 0 functions #*2" on IR” using ¢. The definition is that (#2) = 240) for any point pin IR", In what flloms, we will be sloppy and write oe oa" vehen we really mean : mee ‘The reason we do this is simply that everyone does it, and the 1 will have to get ued to it Now consider the coordinate vector fed Q, on IR™. We can push 0, forward by 4, and we claimn that et ase an ‘Tose this, just apply both sides to any coordinate function 2 on RO and show that we get the same answer. The left hand side gives (@.0,Ne") = a,(6"2) ot oer vrhere inthe last step we aze being sloppy inthe way described above. The right hand side gives, 6.54 Be” yn Be” ete = Ge aest 14 Diffecential Forms whichis the same Finally, consider a coordinate form de. We can pull this Back by We claim that - ode") = Faw Exercise 87. Show this ‘With these basic formulae in hand, you should be able to transform betwen coordinates both actively and passively! "To conclude, we should note that sometimea it i nice to be more ‘general and work with a basse, of wector feds on a chart that are not the coordinate vector elds. ‘These are easy to come by: Exercise $8. Let eo =Ty8e, where Bare the esoninate vector felde associated to local coordinates on an pense end Ts ae nctions am U- Show thatthe vector ide oy re @ Seis of vtor fies on U if and only if Jor each p © U the matris Tap) is nvr I we have such a basis, we automatically get a dual basis of forms fron U such that Pen) = 8, the Kronecker deta [Exercise 39, Use the prsious exercise 0 show tha! the dual basi exists condi ive ‘We can write any vector fd on U as a linear combination paves where v!,...,08 are functions on U, called the components of » in the basis ey.” Similarly, we can write any Mform w on U as a linear combination = ant ‘We will use these more general bases quite a bit ia the next chapter, ‘when we discuss the notion of a ‘metric. This slike an inner product, forms 55 and it will be handy to work with ‘orthonormal bases of vector fields fnd l-forms on a chart. We leave it to the reader to work out howe the components ofa vector Feld or form change when ste perform an arbitrary change of base Exercise 40, Let, be a basis of vector fede on U and let fb the dual esis of forms, Let 4=Te te anther bass of vector fils, end let J be the correnponding deal baie of Porm. Show thet, pear. Show that ifs = wey = whe, then ma The, and that fer = gf" = 0,9 then p-forms By the geometrical product of two vectors, we mean Ike surface content of le paalelogram determined by there vectors we however ix the poviion of le plane im which the perallogram ies. We refer lo fs surface ereas ax scomerialy equal ony when they are equal in content and he in parallel planes. By the feometrial product of three vectra we meam the slid (a parallelepiped) formed from tem. — Hermann Graxoman If you ever seriously wondered how to take cross products in 4 di ‘mensions, you were well on your way to reinventing differential forms. In fact if you ever wondered why the definition of cross products te ‘ques a right-hand rule’, you were gelting close, (This rules etpecially isksome to those who happea to be left-handed ) Differential forms al low one to generalize cross products to any number of dimensions, and it turns out that ifone does things correctly, no right-hand rule is neces: sary! Interestingly, though, it turag out tobe better to define the eros product not for tangent vector (or vector fl) but for cotangent vee- tors (or Horm). If we do this, we get an extra bonus. Namely, we can56 14 Differential Forms show that the gradient, cut and divergence ate all diferent versions of the same thing, and see how Lo define them on arbitrary maciflds. Tet us plunge right in. Let V be a vector pace. We want to beable +o multiply two vectors in V tomehow, and we vant the basic property ‘of the cross product the antiaymmetes, to hold. But we will call thie generalized sort of cross product the “wedge product” (or ‘exterior product’) and write it with aA. We proceed in an abstract, algebraic rort of way. Narmely, we will define a Digger vector space AV, infact an algebra, so that the wedge product of ay numberof vecors in V willie im tie algebra. Fist we wil give the definition ae a mathematician would: the exterior algebra over V, denoted AV, i the algebra generated by V with the relations forall v,w € V. What docs this mean? Roughly, it means that we start with the vectors in V together with an element 1, and then form. an algrbra by taking all Linear combinations of formal products of the form ty A=" Atp, where ty € V; the only relations we impose upon these linear combinations ace those in the definition of an algebra (as Gefined above in Chapter 3) together with the ‘axticommutative’ rule or example, ay V is $dimensional. ‘Then everything in AV is a linear combination of wedge products of elements of V- Suppose V has a basis dy, dy,dz. (We write the bass this way because in abit we will vwant to be a space of cotangent vectors.) Then for starlers we have leav and a,dy,dz © AV, along with al linese combinations of these. But we can also take the wedge product of any two elements v,.e € V and get an element of AV. iE ve unde + ody tude w= undy + wydy twee forms st then we have 2 ww = (weds + dy + ede) 4 (dy + dy + wed) (vat ~ syua)de 1 dy + (oye vv ly A do + (oat — vate)de A dy there all we did is use the definition of an algebra together with the ‘anticommutative’ rule. Notice that this looks a whole lt like the formula for the cross product! If we have a third element of V, say wn ude + uydy + uade, wre can gt another clement of AV, namely wAvAw. This tripe wedge prodsct i lonely related tothe “ple product’ of tree vectors in F°, (9 a), We can also take wedge products of four or more vector, but f V ie Bdimensonal, this alwaye sre: Exercise 41, Show that wrenwesa (i = i) eenannee Compare ths to (8 Exercise 42. Show thet if a,b,ed are four sectors in a $-dimensional space then aN BEA Exercise 48. Descrile AV of Vie Ldimensional, 2-dimensional, or dimensional In general, for any vector space V, we define APY to be the subspace of AV consisting of linear combinations of pfold products of vectors in View. Blementsof AV that lie in APY are aid to have degree p. For example, A'V is jutt V steel, while A°V it by convention defined to be HY, since ‘numbers can be regarded as wedge products of no vectors. Copying the ‘example above, one can show the following:8 14 Dilfereatial Forms Exercise 44, Let V be an nimensional vector space. Show that A°V ie ‘empty forp > m, and tha for 0

dy ‘which would amount toa left-hand rule. In shost, the ‘right-hand rule’ nonsense enters when ve unnaturally try to make the product of two clements of V to come out to an element of V, instead of M'V. This is noted in some physics books, where they say thatthe cross product of two vectors is “prewdovector or ‘axial vector! rather than a true sector. We prefer to say that the wedge product of 2 vectors lies in DV — thie is true in all dimensions Exterior algebra isan interesting subject in itself, ut we do na just vrant to generalize the cross product of vectors; we want to generalize the cross product of vector feds. Actually as aleady mentioned, it i sich better to take products of Iforms! We will do this by copying ‘our construction of AV, with the smooth functions Cm(M) on some manifold M taking the place of the real numbers, and the I-forms (Af) taking the place of the vector space V. Namely, we define the differential forms on M, denoted (A), tobe the algebra generated by 0°(i) with the relations wAp=—Bhe for all, € 9"(M). Tobe precise, we should emphasize that we form (A) as an algebra “over C™(M)’. Thi means, fest of al, that (1) consists of linear combinations of wedge products of Lforms with fune~ tons 28 corfciens, We allow all locally finite linear combinations, that is, shove for which every point pin AF has a neighborhood where only finitely many terms are soazero. Secondly, it means that 9(M) fatisies the rules of an algebra with functions taking the place of num bers. Maybe we should say again what all these rules are. We bave, for alluyy,o € MM) and fg © C=(M), wes ate, wt (uty (otatn wAluan) (wauyan,80 1M Differential Forms wAnteny (wbulAvawrutpny, fae, Rute) = fatto, (sta = fatow. We define the 0-frms, 0°(A), to be the fonctions themselves, and define the wedge product of & function with a differential form to be the ordinary product: / nw = fu. We define the product of a number ‘cand a diferential form w to be the product of the constant function € (P(A) and w. Elements that are linear combinations of products ‘of p Lform are called pforms, and we write the space of forms on Mas P(A), We have (a) = HU. Por example, sxppote M ke IRE. The Oforms on IR" ace just functions, f “The 1-frms all Look like oydat where the coefcients wy are functions, It is easy to check that the Dorms al ook lke Femdet nda where we have put in a factor of 3 because dat det = de" A doh [Also for this reason, we may aa well assume that ww = —Wy- Then on IR for example, we have w= wrads! 4 det Funda! Ade? Funds? Adet Sinilaely the 3-forms lok lke J eyeadet Ade" nds, a sand we may as well assume that djs is totally antisymmetric (that i ‘itches sign when we switch any tio indies). On IR? we get wands! nde? adel Piforms 6 ‘There are no nonzero 4-frms, forms, ete, on RY. In general, there are no nonzero pforms on aa n-dimensional manifold if» > m ‘We leave it forthe reader to show some importaxt facts about dif {erential forms in the following exercises. Exercise 48. Given a vector space V, show tat AV sa graded commu tative or aupercommutative algebra, that fv © APY and w © AYV, then ens (UPaas. ‘Show thet for any manifold I, QE) se graded commutative Exercise 47. Show that differential forms ore coneaveriant. That 5 show thet f$:M— 2 isa map from the manifold M to the manifold 1, there is unique pullback map e290) (4) ring with the usual plllack on O-forme (functions) and f-forms, and satiaing eeu) = agte Swtw = Soren wean) = Songn for allw.u € 9(8) and a ER, Exercise 48. Compare how Lforms and 2forms on BR? tanaform wnder pavity. That is, ft Ps»! be the map Playa) =(-2)-1-2) Imnown asthe ‘parity transformation’. Note tht P maps right handed bases to lefthanded bases and vice verse, Compute (a) when w i the 1form Gud, and when 4 the form heya de® de. In physic, the electri eld Bp called a vector, while the magnetic field Bie called an axial vector, because B changes sign under parity transformation, while B does not. Tn Chapter § we wil se tha itis bes to think of Ue lect td as form on space, andthe magnetic field ay a 2form, In other words, while we may be sed to thinking of«2 14 Differential Forms B = (Be, By, Bx) and B = (By, Byy Be) a8 vector fils, itis better to B= Buda + Eyly + Bale and B= Buty dz + Bye hd + Bye dy By the aboe exercise, this means that they transform diflerently under patity I the reader is frustrated because exterior algebras and difleren- tial forms wer dificult to isualise, we suggest taking © peck ahead to Figures and 4 of Chapter 5. Grassman, the inventor of the ex terice algebra, vitualied a wedge product & \--- A vy as an oriented parallelepiped with sides given by the vectors v,..,t. One must be Careful, however, because the wedge product of I-forme corresponds to 1 parallelepiped in the cotangent space. The Exterior Derivative ‘We know from the first section of thie chapter that the diflereatial is @ nice way to generalize the goed old ‘gradient’ to manifolds. As we sam, the diferential ofa function, or O-form, is « form. Now we will show how to take the diferential af pform and get a (p-+ 1)form: .90(M) + OFAN, ‘This will et us generalize the gradient, the curl and the divergence in one fell swoop, and tee that they are secretly all the same thing. The big clue is that the curl of a gradient i zero: Vx(vpa0 ‘This muggests that we make d satisfy d(df) = 0 for any function f. ‘Another clue is that the various product rales VUls) = (Wf)9+1¥e Ux(fy) = Us xot sxe Vi(fe) = Vive fiw Ve(exw) = (V-vje-vVw ‘The Exterior Derivative 8 should all be special cases of some sort of Leibuis law for differential forms, Since the diferental forms are graded commutative, it turns ‘out that we need a graded version of the Leibniz law. ‘Alerscratching our head fora while, wedefine the exterior deriva tive, or differential, to be the unique set of maps E07(M) + OF*1(0) sch that the following properties hold 1) ds9°UM) -» 0°() agrees with our previous definition. 2) dw +p) = eo + dp and (ov) = ed forall w,y © O(M) and eR. 3) d(whn) = don (Pedy for alle € (A) and w € (81). 4) dds) = 0 forall (4), ‘Toshow that these properties wniquely determine the exterior derive tive, one just needs the fact that any I-form is a locally finite linear combination of thove of the fotm df (with Functions as coeficients). This fact i easy to eee on IR", and can be shown in general axing charts. Then to ealculated of any difleratial form, say fag Nb, swe just ure cules 1) ~ 4): A fdg Nah) = af n(dg ndh) + f ndldg nh) A dg dh + faldg) dh — fag A dfdh) ido nah, "To show that d with these properties is actually well-defined, it euices (by the black magic of algebra) to show that this way of calculating d Js compatible with the eelations in the definition of differential forms. ‘The mos: important one of these is the aaticommutative law wAps naw for forms. Bor d to be well-defined, it had better be true that calew lating dlw A p) gives the same ansner as dlp hw). This i where theot 14 Differential Forms traded Leibne lave is necessary: when w and are I-forms, we have dwn) = dle) Ada nwt nde, wh dp tdo ny ae i) Let us caleulate the exterior derivative of Lforms and 2-forms on IRD, Taking any I-frm meade + ody tad, we gt hy = dag de + day Ady + deh de, hence y the rule for d of function and a ttle extra work hs = (Bj Bay) dy de (Bs Oat) rds +(asy ~Byue Ady In other words, the exterior derivative of « I-form on IR? is esentally just the curl! We need right-hand to define the evr, hoeever, while the ‘exterior derivative involves no right-hand rule. This is because d of a ‘form is «2 form; the right-hand rule only comes in when one tres $o pretend that this 24orm ia Iform, using Uhe star operator as follows: ada = (By And, as noted, this pretense is only possible in 8 dimensions, while we can taked of a -form in any dimention: ~ Dy) da + (Oana ~ Ose) + (Bey ~ Bie Exe given by 49, Show that on RO the exterior derivative of any 1form 3 elute") = Baad 4 de Next, taking a 2-form on yd A dy + pay Nd2 + ads Adz ‘The Exterior Derivative 65 we get ay = dey 8 42 8 dy + dye Ny Ndz + dee A de hd = Biinyhs 32 8 dy 4 Bagels dy f d2 + Oydrdy N de dz (Basey + Bie + Ojra)dE N dy de ‘Thus the exterior derivative of a 2form on Ris just the divergence in disguise, In short, the exterior derivative has as special cases the following familiar operator: © Gradient d:9°(R) + 0°(R°) © Gurl &:a CR) -- PCR) © Divergence &:99(R!) + 9°(R) Tn fact, there isa simple formala for the exterior derivative of any Aifereatial form on IR". Let 1 stand for a multivindex, that is, a tuple (j,..-sip) of distinc integers between I and n. Let do! stand for the pform aah hen fon IRY. Then any p-form on I" can be expressed as wurde! ‘where following the Einetin summation convention we sum over all rmultiindices - We have he = day Nae by the Leibnie law, since (da!) = 0 (as can eaily be checked). More concretely, using the formula for dof a function, we have ty = (Oysr) da hd. Using this formula it it easy to derive an amazing identity alaw)66 14 Differential Forms {for aay difereatial form on IR", Just compute Alda) = dl@yurdo? ode!) = AByurde” A det A de! and note that on the one hand Ada = 3,2 lay the equality of mixed partials, but on the other hand det nde = dah nd” boy the anticommutative law. With a little thought one can see this means that d(as) i equal to the negative of itself, soit is ero. This rule is 9 important that people often write it as fu=0 cor even just e (a ID, d act like the gradient on 0-orms, the curl on Iforms and the divergence on 2-forms, a0 the identity d? = 0 contains within it the identities vxivA= and vv xe) But this identity is better, since it applies to differential forms in any dimension. In fact it applies to any manifold! Here is an easy proof that does not use coordinates. By definition, any pform on a manifold ia linear combination — with constant coeficients — of forms ike w= fads No al So it suffices to prove the identity for pforms of this sort. We have y= do Ndfi A hdl ‘The Exterior Derivative er by the Leibnis law and the fact that d{@f) = Ofor any function, Using the Leibni law and def) =O again, we obtain (du) = 0 11 turns out that the identity d® = 0 and its generalizations have profound consequences for physics slarting with Maxwell's equations tis alo the basis of a very important connection between geometry ad topology, called deRhara theory. We will explore therein Chapter 8. When we do it is important to remember that this identity is just ‘a way of saying that partial derivatives commute! As so often the case, the simplest facts in mathematics lie atthe zoot of some of the most sophisticated developments ‘We will ap up this section by showing that the exterior derivative is natural, We already discussed this for functions in Section 45 it simply meant that d commutes with pullbacks. In fact, this is tru for ‘lifereatial forma of any degree. In other words, for any map ¢:M — NV between manifolds, and any dilferental form w € O°(M), we have # (de) = dg) ‘The proof is easy. By Exercise 47, "is real-linear, 40 itsfces to treat the cate where B= hath ho Aa We then have, using Exercise 47 again together with the naturality of 1 on functions Olde) = Po Nd nerds) Fly hoo Ay BB fan AB hy MO fo NdB fio 46H) Goh $a KoA Gay) = Ag Uo nahi 6--Adh)) = doe) 1s desired.Chapter 5 Rewriting Maxwell’s Equations Hence space of itaelf, and time of tell, will sink into mere shadows, and only union of the fo shall servis. — Hermann Minkowski, ‘The First Pair of Equations ‘We now have developed enough differential gometry to generalize the fiat pair of Maxell equations, vo =o 2.08 vx 848 =o, to any manifold. We claim that they have a very beautiful form as & single equation in terms of diferential forme, fore giving away the answer, let us consider « special cae: the static ene. ‘Then we just have two equations for vector fields on space, cS “ . v-B=0, vxB=0 In the language of diferential forma, the divergence becomes the exte ror derivative on 2-forme on I. Thus, instead of treating the magnetic field as 2 vector field B = (Be, By, B) we wil treat it as the 2-form B= Baby hd + Byde he + Bude dy o0 1.5 Rewriting Maxwells Equations Similacy, the curl becomes the exterior derivative on I-forms on B®, 10 instead of treating the electric field at vector feld E = (Bs, By, Bs) wwe wil teat i as the I-form B= Buda + Eyl + Ba "The frat paie of static Maxwell's equations then become ae , aB Next consider the general time-dependent case. Now we must think ofthe electric and magnetic fields as living on spacetime. We begin by ‘working an Minkowski spacetime, R', using the standard coordinate System, which we wil number a2 (2°,24,24,2°). We will often write ‘instead of 2" for the time coordinate, and 2,9, 2 for the space cor: dinates (z,2?,2"). The electzic and magnetic feds are I-forms and 2-forms on IRS, namely B= Bade + Byly + Bade aad B= Bydy Adz + Bude de + Bud dy We can combine both fede into a unified electromagnetic field Pa 2form on BR, as follom: PoB+ENd I we want to look at all the components, raids” Ade, vee can write them out as 8 matric 0 -E, -E, 0 Be -B, Fe=l5 -B, 0 Be EL By -Be 0 ‘The First Par of Bquations n ‘The beauty of thie way of unifying the electric and magnetic fel ie that the feet pair of Masorll equations became simply aP =o, ‘To see this, fet note that dF =d{B+ Ad) = dB Hab nat ‘Then split up the exterior derivative operator into a spacelike part and a timelike part, Recall shat for any diferential form oe eve is = Ogu det nds! where J ranges overall multi-indioes and = 0,1,2,3. We can thus write du ob a sum of the spacelike part so = Burda’ nde! orhoze ranges over the epaelite’ indices 2,3, and the timelie past aE Oar = Br dod! ‘Phen we have aF = dBsdBnd UpB +d NAB (dob + dt NOE) Act 5B + (GB + dsB) Ad [Note thatthe iret term has no dt init while the second one docs, Also ‘ote that the eecoad one vanishes only ifthe expression in parentheses oes, I follows that dF = 0 is equivalent to the pair of equations 4B = 0, OB+dsB = 0. "Those are just the first pair of Maxwell equations in slighty newfangled2 15 Rewriting Maxwell's Equations Fig. 1. Splitting spacetime into space and time (One advantage of the dferential form language is its generality. We can take our spacetime to be any manifold M, of any dimension, and dlefine the electromagnetic field to be & 2form P on M. The frst pair fof Maxell equations says just that ar=o. Sometimes — but not always — we can split spacetime up into space tnd time, that i, write M ar Lx S for some manifold $ we call ‘space! If ao, we can weve t forthe utudl coordinate on I and split P into an electric and magnetic field Exercise 60. Show that any 2form P on Bx can be uniquely exprested tar BB Adin such a wy that for any local coordinates 2+ on S we have B= Bude and B= PBs de? We can also split the exterior derivative into spacelike and timelike parte a before: Exercise $1. Show tht jor any form won BL S there i «unique way tovrite da) = dt \ dua | dow auth that for an loeal coordinates 2* on S, ‘we hese dso = Byard nds, Aes = Ayan de Adel ‘Whea we split spacetime up into space and time, dF = 0 becomes ‘equivalent to the pai of equations 4B =0, OB 4+dsB=0. The Metre 3 In the static case, when OF coordinate entirely and test static equations 1B = 0, we ean forget about the £ and B as forms on space satisfying the asB = , ds =O. Note that the electric and magnet fields are only defined after we choose a way of splitting spacetime into space and time! If someone hands us @ manifold M, it may be diffeomorphic to Rx $ in many diferent ways, or in no wey at all. In special relativity one learns ‘hat diffrent inertial frames (corresponding to observers moving at constant velocity) will give diferent splittings of spacetimeinto 0, 2s called spacelike, since it points more in the space directions than the time direction. Itz is spacelike, the square foot of 2-2 represents the length ofa straight ruler that stretched from the origin to 2. Ifz has 2-2 <0 we call t timelike, since it points more in the time direction than in the space directions, The velocity Of particle moving slower than the speed of light is timelie. If «is timelike, the square root of x: measures the time a clock would tick off. fait moved from the origin to ima straight ine. If 2-2 = 0, xis called full or lightlike, since t points just as much in the time direction as in the space directions. We should add that sometimes people use she negative of our Minkowski metricand revere the definitions of ypacelike fand timelike. This i just matter of convention — but we wil always {allow the above convention! ‘The notion of metric generalizes these concepts, A semi Riemannian metric (or just"metric)) on a vector space V is & map, nVKV=R, ‘that is bilinear, or lineae in each slot: acu +e) = ealvy) + 90) g(oyewtu') = o(0,10) bealeyw') afer) = (0,2), The Metre 6 and nondegenerate: if 4(2,~) = 0 for all w € V, then v = 0. We say that » € V is spacelike timelike or null depeading on whether g(v,2) is ponitive, negative or ero, If (vw) = 0, we aay that v and w ace orthogonal. Note that null vectors are orthogonal to themselves! Givena metricon V, wecan always find an orthonormal basis for Vs that ia basis {e,) Such that g(eyy¢y) 8 0 ify # v, and it ‘The number of +1's and =1's i independent of the orthonormal basis, tad ifthe number of +1's x p and the number of —1's is 9, we say the metric has signature (p,¢). For example, Minkowski spacetime has signature (3,1), with the Minkowski metric given by n(o,w) =u? pote total + tw So far we have been talking about spacetimes that are vector spaces. [Now let AT be a manifold and consider a situation where the metric depends on where one is, A metric g on M sssigas to each point EM a metric gp on the tangent space Ty, ia @ smoothly varying ‘way. By ‘smoothly varying? we mean that if and w are smooth vector fields on M, the inner product gy(%p,up) is » smooth function on M. By the way, we usually write thie faction rimply as 9, One can show that the rmoothness condition implies that the sig- nature of gp is constant on any connected component of M. We are realy only iaterested in cases where the signature ie constant on all of 1M. Ithe signature of gis (n,0), where dim M = n, we say that is a Riemannian metric, while ifthe signature is (n — 1,1), we say that g is Lorentzian. By a semi-Riemannian manifold te mean ‘manifold equipped with a metric, and similarly for a Riemannian ‘manifold end a Lorentzian manifold. In relativity, spacetime is « Lorentzian manifld, which in the real world appease to be &-dimensional, although other cases are certainly interesting. The easiest way to get ahold of « dimensional Lorentsian ‘manifolds to take Sdimensional manifold S, space’, with a Rieman- nian metric 9p, and let IY, ‘spacetime’, be given by Ix S. Then we ‘ca define @ Lorentzian metre gad 8 fon Mas follows. Let 2 (i = 1,2,8) be local coordinates on an open subset U © S, and let t or 2° denote the coordinate on IR, that is,6 15 Rewriting Maxwell’ Equations ‘time’, Then 2 (j= 0,1,2,3) ate local coordinates on Rex U and we can define the metsic gto be that with componente 00 0 ° weal 8 ‘ o ‘This presets special sort of state spacetime, in mhich space has 1 tate that independent of ime "Phe most basi ue of a Lorentsian matric to measure distances ad times Fr example, path 70,1] > Mis spacelike, that x i ‘tangent vector is everywhere spc, me defies arclength to be [arora ys timelike, we define the proper time along 7 — that is, the time ticked of ty a'clock moving along 7 — to be [eso Orie. We will mainly be interested i some more sophisticated applicx tions ofthe metric, bowever. The most fundamental ofthese i ‘raising tnd lowering indices’, that is, converting between tangent and eotan feat vector. IfV ie a vector space equipped with a metric g, there is ‘natural way to tur an element v ¢ V into an element of V*, namely the linear funeSional g(v,") which eats another clement of V and spits cout a number. Bxercise 52. Use the nondayeneracy of the metre 10 show that the map from to V" given by verge fs an isomorphiom, this, one-to-one and onto. 1 follows that if fie semi-Riemansian manifold the metric de fines an isomorphism between each tangent space TpM and the corze sponding cotangent space Ts. We can picture this as follows: if the tangent vector v isa litle arom, the cotangent vector = g(t) is The Metsie 7 a stack of hyperplanes perpendicular tov, asin Figure 2. The reason for this is that w vanishes on vectors orthogonsl tov. ‘The key point is that one need the metric to know what ‘rthogonal” means! < Fig. 2. Tangent vector v and cotangent veetor (0,-) kOe) Similarly, we can convert between vector feds and I-forms on Mf By using the metric on space, for example, we can think ofthe electric field as 2 vector field instead of x Iform. We need to do this in order to think ofthe electri fel as “pointing in some direction Suppote M is semiRiemaanian manifold. Now that we can visu- lize Mforma on Mf as Selds of little artons, there is a nice way for us to vsualiae pforms for higher pas well. We can draw # wedge product hp of two cotangent vectors at para ttle parallelogram, asin Figure 3. So me can visualize @ 2-4orm on Mas feld of such ‘area elements. Similarly, we can draw a wedge product w fw A of three cotangent vectors ob pas a little parallelepiped, asin Figuce 4, and visualize a form as field of these ‘volume elements! — and 60 on for higher forme > <> Fig. 3. Picture of Ay € APTA* 15 Rewriting Maxwell's Equations We should not tale these pictures tae seriously. For example, if we drew (de +49) 4 (dy +4) and (dy 4 de) a (de az) thie way, we would get differen-looking paralelograms, even though they are equal as elements of APTSIR*. However, there is a lot about ‘these patllelograms that isthe same. Fest, they lie in the same plane VC MP. Second, they have the same area. And third, the pairs (dz + dy,dy-+ ds), (dy +d2,de—de) are bases for V that have the same orientation, in the usual sense of right-handed versus left-handed bases "This is true in general, which x why we think of w Ay € A273M as an “area clement’. Similar statements are true for wedge products of more: colangent vectors, Oa the other hand, itis worth noting that there fre usualy plenty of elements of A'T¢AQ that are aot wedge products Of cotangent vectors — we need to Consider near combinations, too. Sl, witha litle care one can get some good insights abot ciferential forms sing there picture, as we will see. ‘Now let us say some things about how the metric looks in the lan guage of indice. Let ue work in a chart, and let ey be basis of vector fields, Then we ean define the components ofthe metric as fellows: Sn = (0) ‘The Metre 8 If M is redimensional, gu is ann Xm matrix. The aondegeneracy condition implies thie matrix is invertible, 20 let g*” denote the inverse ‘matrix Then we have the following handy formulas, which explain why the process of eoavesting between vector fede and I-forms using the smetric is called raising and lowering indices: Exercise 58, Lets = wey bea vector fald on a chart. Show that the coreg form oy el wre the dt of forms Bxercise 54, Letw = up/ bea Lform on @ chart Show thatthe corre sponding vector field i opal tse, where of = ey [Exercise 55. Let be the Mintowsti meric on BY as defined above. Show (hat it components im the atandard basis are In general, if we have any quantity with some indices, such as 0 ° ates vee can lower or raise any index with the metic and its inverse using the Binsin summation convention. Bug, we can lower @ and get AE Mg = Sandee or mise § and get Aga ae, I we have a fot indices Hosting around it is important to keep track of their onder when we rise and lower them; otherwise things get con fusing. Note that se can even raise and lower indices on the metrie itself