GEOMETRIC ALGEBRA FOR PHYSICISTS CHRIS DORAN and ANTHONY LASENBY University of Cambridge
published by the press syndicate of the university of cambridge The Pitt Building, Trumpington Street, Cambridge, United Kingdom cambridge university press The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York, NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarc´ on 13, 28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http://www.cambridge.org 0001 C Cambridge University Press, 2003
This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2003 Printed in the United Kingdom at the University Press, Cambridge Typeface CMR 10/13 pt
System LATEX 2ε [TB]
A catalogue record for this book is available from the British Library Library of Congress Cataloguing in Publication data ISBN 0 521 48022 1 hardback
Contents
Preface Notation
ix xiii
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Introduction Vector (linear) spaces The scalar product Complex numbers Quaternions The cross product The outer product Notes Exercises
1 2 4 6 7 10 11 17 18
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9
Geometric algebra in two and three dimensions A new product for vectors An outline of geometric algebra Geometric algebra of the plane The geometric algebra of space Conventions Reflections Rotations Notes Exercises
20 21 23 24 29 38 40 43 51 52
3 3.1 3.2 3.3
Classical mechanics Elementary principles Two-body central force interactions Celestial mechanics and perturbations
54 55 59 64
v
CONTENTS
3.4 3.5 3.6
Rotating systems and rigid-body motion Notes Exercises
4 4.1 4.2 4.3 4.4 4.5 4.6 4.7
Foundations of geometric algebra Axiomatic development Rotations and reflections Bases, frames and components Linear algebra Tensors and components Notes Exercises
84 85 97 100 103 115 122 124
5 5.1 5.2 5.3 5.4 5.5 5.6 5.7
Relativity and spacetime An algebra for spacetime Observers, trajectories and frames Lorentz transformations The Lorentz group Spacetime dynamics Notes Exercises
126 127 131 138 143 150 163 164
6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8
Geometric calculus The vector derivative Curvilinear coordinates Analytic functions Directed integration theory Embedded surfaces and vector manifolds Elasticity Notes Exercises
167 168 173 178 183 202 220 224 225
7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8
Classical electrodynamics Maxwell’s equations Integral and conservation theorems The electromagnetic field of a point charge Electromagnetic waves Scattering and diffraction Scattering Notes Exercises
228 229 235 241 251 258 261 264 265
vi
69 81 82
CONTENTS
8 8.1 8.2 8.3 8.4 8.5 8.6 8.7
Quantum theory and spinors Non-relativistic quantum spin Relativistic quantum states The Dirac equation Central potentials Scattering theory Notes Exercises
267 267 278 281 288 297 305 307
9 9.1 9.2 9.3 9.4 9.5 9.6 9.7
Multiparticle states and quantum entanglement Many-body quantum theory Multiparticle spacetime algebra Systems of two particles Relativistic states and operators Two-spinor calculus Notes Exercises
309 310 315 319 325 332 337 337
10 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9
Geometry Projective geometry Conformal geometry Conformal transformations Geometric primitives in conformal space Intersection and reflection in conformal space Non-Euclidean geometry Spacetime conformal geometry Notes Exercises
340 341 351 355 360 365 370 383 390 391
11 11.1 11.2 11.3 11.4 11.5 11.6 11.7
Further topics in calculus and group theory Multivector calculus Grassmann calculus Lie groups Complex structures and unitary groups The general linear group Notes Exercises
394 394 399 401 408 412 416 417
12 12.1 12.2 12.3
Lagrangian and Hamiltonian techniques The Euler–Lagrange equations Classical models for spin-1/2 particles Hamiltonian techniques
420 421 427 432
vii
CONTENTS
12.4 12.5 12.6
Lagrangian field theory Notes Exercises
439 444 445
13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8
Symmetry and gauge theory Conservation laws in field theory Electromagnetism Dirac theory Gauge principles for gravitation The gravitational field equations The structure of the Riemann tensor Notes Exercises
448 449 453 457 466 474 490 495 495
14 Gravitation 14.1 Solving the field equations 14.2 Spherically-symmetric systems 14.3 Schwarzschild black holes 14.4 Quantum mechanics in a black hole background 14.5 Cosmology 14.6 Cylindrical systems 14.7 Axially-symmetric systems 14.8 Notes 14.9 Exercises Bibliography Index
viii
497 498 500 510 524 535 543 551 564 565 568 575
1
Introduction
The goal of expressing geometrical relationships through algebraic equations has dominated much of the development of mathematics. This line of thinking goes back to the ancient Greeks, who constructed a set of geometric laws to describe the world as they saw it. Their view of geometry was largely unchallenged until the eighteenth century, when mathematicians discovered new geometries with different properties to the Greeks’ Euclidean geometry. Each of these new geometries had distinct algebraic properties, and a major preoccupation of nineteenth century mathematicians was to place these geometries within a unified algebraic framework. One of the key insights in this process was made by W.K. Clifford, and this book is concerned with the implications of his discovery. Before we describe Clifford’s discovery (in chapter 2) we have gathered together some introductory material of use throughout this book. This chapter revises basic notions of vector spaces, emphasising pictorial representations of the underlying algebraic rules — a theme which dominates this book. The material is presented in a way which sets the scene for the introduction of Clifford’s product, in part by reflecting the state of play when Clifford conducted his research. To this end, much of this chapter is devoted to studying the various products that can be defined between vectors. These include the scalar and vector products familiar from three-dimensional geometry, and the complex and quaternion products. We also introduce the outer or exterior product, though this is covered in greater depth in later chapters. The material in this chapter is intended to be fairly basic, and those impatient to uncover Clifford’s insight may want to jump straight to chapter 2. Readers unfamiliar with the outer product are encouraged to read this chapter, however, as it is crucial to understanding Clifford’s discovery. 1
INTRODUCTION
1.1 Vector (linear) spaces At the heart of much of geometric algebra lies the idea of vector, or linear spaces. Some properties of these are summarised here and assumed throughout this book. In this section we talk in terms of vector spaces, as this is the more common term. For all other occurrences, however, we prefer to use the term linear space. This is because the term ‘vector ’ has a very specific meaning within geometric algebra (as the grade-1 elements of the algebra).
1.1.1 Properties Vector spaces are defined in terms of two objects. These are the vectors, which can often be visualised as directions in space, and the scalars, which are usually taken to be the real numbers. The vectors have a simple addition operation rule with the following obvious properties: (i) Addition is commutative: a + b = b + a.
(1.1)
a + (b + c) = (a + b) + c.
(1.2)
(ii) Addition is associative:
This property enables us to write expressions such as a + b + c without ambiguity. (iii) There is an identity element, denoted 0: a + 0 = a.
(1.3)
(iv) Every element a has an inverse −a: a + (−a) = 0.
(1.4)
For the case of directed line segments each of these properties has a clear geometric equivalent. These are illustrated in figure 1.1. Vector spaces also contain a multiplication operation between the scalars and the vectors. This has the property that for any scalar λ and vector a, the product λa is also a member of the vector space. Geometrically, this corresponds to the dilation operation. The following further properties also hold for any scalars λ, µ and vectors a and b: (i) (ii) (iii) (iv)
λ(a + b) = λa + λb; (λ + µ)a = λa + µa; (λµ)a = λ(µa); if 1λ = λ for all scalars λ then 1a = a for all vectors a. 2
1.1 VECTOR (LINEAR) SPACES
b
b c
a+b
a
a
a
b+c a+b a+b+c
b Figure 1.1 A geometric picture of vector addition. The result of a + b is formed by adding the tail of b to the head of a. As is shown, the resultant vector a + b is the same as b + a. This finds an algebraic expression in the statement that addition is commutative. In the right-hand diagram the vector a + b + c is constructed two different ways, as a + (b + c) and as (a + b) + c. The fact that the results are the same is a geometric expression of the associativity of vector addition.
The preceding set of rules serves to define a vector space completely. Note that the + operation connecting scalars is different from the + operation connecting the vectors. There is no ambiguity, however, in using the same symbol for both. The following two definitions will be useful later in this book: (i) Two vector spaces are said to be isomorphic if their elements can be placed in a one-to-one correspondence which preserves sums, and there is a one-to-one correspondence between the scalars which preserves sums and products. (ii) If U and V are two vector spaces (sharing the same scalars) and all the elements of U are contained in V, then U is said to form a subspace of V.
1.1.2 Bases and dimension The concept of dimension is intuitive for simple vector spaces — lines are onedimensional, planes are two-dimensional, and so on. Equipped with the axioms of a vector space we can proceed to a formal definition of the dimension of a vector space. First we need to define some terms. (i) A vector b is said to be a linear combination of the vectors a1 , . . . , an if scalars λ1 , . . . , λn can be found such that b = λ1 a1 + · · · + λn an =
n 0001
λi ai .
(1.5)
i=1
(ii) A set of vectors {a1 , . . . , an } is said to be linearly dependent if scalars 3
INTRODUCTION
λ1 , . . . , λn (not all zero) can be found such that λ1 a1 + · · · + λn an = 0.
(1.6)
If such a set of scalars cannot be found, the vectors are said to be linearly independent. (iii) A set of vectors {a1 , . . . , an } is said to span a vector space V if every element of V can be expressed as a linear combination of the set. (iv) A set of vectors which are both linearly independent and span the space V are said to form a basis for V. These definitions all carry an obvious, intuitive picture if one thinks of vectors in a plane or in three-dimensional space. For example, it is clear that two independent vectors in a plane provide a basis for all vectors in that plane, whereas any three vectors in the plane are linearly dependent. These axioms and definitions are sufficient to prove the basis theorem, which states that all bases of a vector space have the same number of elements. This number is called the dimension of the space. Proofs of this statement can be found in any textbook on linear algebra, and a sample proof is left to work through as an exercise. Note that any two vector spaces of the same dimension and over the same field are isomorphic. The axioms for a vector space define an abstract mathematical entity which is already well equipped for studying problems in geometry. In so doing we are not compelled to interpret the elements of the vector space as displacements. Often different interpretations can be attached to isomorphic spaces, leading to different types of geometry (affine, projective, finite, etc.). For most problems in physics, however, we need to be able to do more than just add the elements of a vector space; we need to multiply them in various ways as well. This is necessary to formalise concepts such as angles and lengths and to construct higher-dimensional surfaces from simple vectors. Constructing suitable products was a major concern of nineteenth century mathematicians, and the concepts they introduced are integral to modern mathematical physics. In the following sections we study some of the basic concepts that were successfully formulated in this period. The culmination of this work, Clifford’s geometric product, is introduced separately in chapter 2. At various points in this book we will see how the products defined in this section can all be viewed as special cases of Clifford’s geometric product.
1.2 The scalar product Euclidean geometry deals with concepts such as lines, circles and perpendicularity. In order to arrive at Euclidean geometry we need to add two new concepts 4
1.2 THE SCALAR PRODUCT
to our vector space. These are distances between points, which allow us to define a circle, and angles between vectors so that we can say that two lines are perpendicular. The introduction of a scalar product achieves both of these goals. Given any two vectors a, b, the scalar product a · b is a rule for obtaining a number with the following properties: (i) (ii) (iii) (iv)
a·b = b·a; a·(λb) = λ(a·b); a·(b + c) = a·b + a·c; a·a > 0, unless a = 0.
(When we study relativity, this final property will be relaxed.) The introduction of a scalar product allows us to define the length of a vector, |a|, by √ |a| = (a·a). (1.7) Here, and throughout this book, the positive square root is always implied by √ the symbol. The fact that we now have a definition of lengths and distances means that we have specified a metric space. Many different types of metric space can be constructed, of which the simplest are the Euclidean spaces we have just defined. The fact that for Euclidean space the inner product is positive-definite means that we have a Schwarz inequality of the form |a·b| ≤ |a| |b|.
(1.8)
The proof is straightforward: (a + λb)·(a + λb) ≥ 0 2
⇒ a·a + 2λa·b + λ b·b ≥ 0 2
⇒ (a·b) ≤ a·a b·b,
∀λ ∀λ (1.9)
where the last step follows by taking the discriminant of the quadratic in λ. Since all of the numbers in this inequality are positive we recover (1.8). We can now define the angle θ between a and b by a·b = |a||b| cos(θ).
(1.10)
Two vectors whose scalar product is zero are said to be orthogonal. It is usually convenient to work with bases in which all of the vectors are mutually orthogonal. If all of the basis vectors are further normalised to have unit length, they are said to form an orthonormal basis. If the set of vectors {e1 , . . . , en } denote such a basis, the statement that the basis is orthonormal can be summarised as ei ·ej = δij . 5
(1.11)
INTRODUCTION
Here the δij is the Kronecker delta function, defined by 0002 1 if i = j, δij = 0 if i = j.
(1.12)
We can expand any vector a in this basis as a=
n 0001
ai ei = ai ei ,
(1.13)
i=1
where we have started to employ the Einstein summation convention that pairs of indices in any expression are summed over. This convention will be assumed throughout this book. The {ai } are the components of the vector a in the {ei } basis. These are found simply by ai = ei ·a.
(1.14)
The scalar product of two vectors a = ai ei and b = bi ei can now written simply as a·b = (ai ei )·(bj ej ) = ai bj ei ·ej = ai bj δij = ai bi .
(1.15)
In spaces where the inner product is not positive-definite, such as Minkowski spacetime, there is no equivalent version of the Schwarz inequality. In such cases it is often only possible to define an ‘angle’ between vectors by replacing the cosine function with a cosh function. In these cases we can still introduce orthonormal frames and use these to compute scalar products. The main modification is that the Kronecker delta is replaced by ηij which again is zero if i = j, but can take values ±1 if i = j. 1.3 Complex numbers The scalar product is the simplest product one can define between vectors, and once such a product is defined one can formulate many of the key concepts of Euclidean geometry. But this is by no means the only product that can be defined between vectors. In two dimensions a new product can be defined via complex arithmetic. A complex number can be viewed as an ordered pair of real numbers which represents a direction in the complex plane, as was realised by Wessel in 1797. Their product enables complex numbers to perform geometric operations, such as rotations and dilations. But suppose that we take the complex number z = x + iy and square it, forming z 2 = (x + iy)2 = x2 − y 2 + 2xyi.
(1.16)
In terms of vector arithmetic, neither the real nor imaginary parts of this expression have any geometric significance. A more geometrically useful product 6
1.4 QUATERNIONS
is defined instead by zz ∗ = (x + iy)(x − iy) = x2 + y 2 ,
(1.17)
which returns the square of the length of the vector. A product of two vectors in a plane, z and w = u + vi, can therefore be constructed as zw∗ = (x + iy)(u − iv) = xu + vy + i(uy − vx).
(1.18)
The real part of the right-hand side recovers the scalar product. To understand the imaginary term consider the polar representation z = |z|eiθ ,
w = |w|eiφ
(1.19)
so that zw∗ = |z||w|ei(θ − φ) .
(1.20)
The imaginary term has magnitude |z||w| sin(θ − φ), where θ − φ is the angle between the two vectors. The magnitude of this term is therefore the area of the parallelogram defined by z and w. The sign of the term conveys information about the handedness of the area element swept out by the two vectors. This will be defined more carefully in section 1.6. We thus have a satisfactory interpretation for both the real and imaginary parts of the product zw ∗ . The surprising feature is that these are still both parts of a complex number. We thus have a second interpretation for complex addition, as a sum between scalar objects and objects representing plane segments. The advantages of adding these together are precisely the advantages of working with complex numbers as opposed to pairs of real numbers. This is a theme to which we shall return regularly in following chapters. 1.4 Quaternions The fact that complex arithmetic can be viewed as representing a product for vectors in a plane carries with it a further advantage — it allows us to divide by a vector. Generalising this to three dimensions was a major preoccupation of the physicist W.R. Hamilton (see figure 1.2). Since a complex number x + iy can be represented by two rectangular axes on a plane it seemed reasonable to represent directions in space by a triplet consisting of one real and two complex numbers. These can be written as x + iy + jz, where the third term jz represents a third axis perpendicular to the other two. The complex numbers i and j have the properties that i2 = j 2 = −1. The norm for such a triplet would then be (x + iy + jz)(x − iy − jz) = (x2 + y 2 + z 2 ) − yz(ij + ji).
(1.21)
The final term is problematic, as one would like to recover the scalar product here. The obvious solution to this problem is to set ij = −ji so that the last term vanishes. 7
INTRODUCTION
Figure 1.2 William Rowan Hamilton 1805–1865. Inventor of quaternions, and one of the key scientific figures of the nineteenth century. He spent many years frustrated at being unable to extend his theory of couples of numbers (complex numbers) to three dimensions. In the autumn of 1843 he returned to this problem, quite possibly prompted by a visit he received from the young German mathematician Eisenberg. Among Eisenberg’s papers was the observation that matrices form the elements of an algebra that was much like ordinary arithmetic except that multiplication was non-commutative. This was the vital step required to find the quaternion algebra. Hamilton arrived at this algebra on 16 October 1843 while out walking with his wife, and carved the equations in stone on Brougham Bridge. His discovery of quaternions is perhaps the best-documented mathematical discovery ever.
The anticommutative law ij = −ji ensures that the norm of a triplet behaves sensibly, and also that multiplication of triplets in a plane behaves in a reasonable manner. The same is not true for the general product of triplets, however. Consider (a + ib + jc)(x + iy + jz) = (ax − by − cz) + i(ay + bx) + j(az + cx) + ij(bz − cy).
(1.22)
Setting ij = −ji is no longer sufficient to remove the ij term, so the algebra does not close. The only thing for Hamilton to do was to set ij = k, where k is some unknown, and see if it could be removed somehow. While walking along the Royal Canal he suddenly realised that if his triplets were instead made up of four terms he would be able to close the algebra in a simple, symmetric way. 8
1.4 QUATERNIONS
To understand his discovery, consider (a + ib + jc + kd)(a − ib − jc − kd) = a2 + b2 + c2 + d2 (−k 2 ) − bd(ik + ki) − cd(jk + kj),
(1.23)
where we have assumed that i2 = j 2 = −1 and ij = −ji. The expected norm of the above product is a2 + b2 + c2 + d2 , which is obtained by setting k 2 = −1 and ik = −ki and jk = −kj. So what values do we use for jk and ik? These follow from the fact that ij = k, which gives ik = i(ij) = (ii)j = −j
(1.24)
kj = (ij)j = −i.
(1.25)
and
Thus the multiplication rules for quaternions are i2 = j 2 = k 2 = −1
(1.26)
and ij = −ji = k,
jk = −kj = i,
ki = −ik = j.
(1.27)
These can be summarised neatly as i2 = j 2 = k 2 = ijk = −1. It is a simple matter to check that these multiplication laws define a closed algebra. Hamilton was so excited by his discovery that the very same day he obtained leave to present a paper on the quaternions to the Royal Irish Academy. The subsequent history of the quaternions is a fascinating story which has been described by many authors. Some suggested material for further reading is given at the end of this chapter. In brief, despite the many advantages of working with quaternions, their development was blighted by two major problems. The first problem was the status of vectors in the algebra. Hamilton identified vectors with pure quaternions, which had a null scalar part. On the surface this seems fine — pure quaternions define a three-dimensional vector space. Indeed, Hamilton invented the word ‘vector ’ precisely for these objects and this is the origin of the now traditional use of i, j and k for a set of orthonormal basis vectors. Furthermore, the full product of two pure quaternions led to the definition of the extremely useful cross product (see section 1.5). The problem is that the product of two pure vectors does not return a new pure vector, so the vector part of the algebra does not close. This means that a number of ideas in complex analysis do not extend easily to three dimensions. Some people felt that this meant that the full quaternion product was of little use, and that the scalar and vector parts of the product should be kept separate. This criticism misses the point that the quaternion product is invertible, which does bring many advantages. The second major difficulty encountered with quaternions was their use in 9
INTRODUCTION
describing rotations. The irony here is that quaternions offer the clearest way of handling rotations in three dimensions, once one realises that they provide a ‘spin-1/2’ representation of the rotation group. That is, if a is a vector (a pure quaternion) and R is a unit quaternion, a new vector is obtained by the double-sided transformation law a0002 = RaR∗ ,
(1.28)
where the * operation reverses the sign of all three ‘imaginary’ components. A consequence of this is that each of the basis quaternions i, j and k generates rotations through π. Hamilton, however, was led astray by the analogy with complex numbers and tried to impose a single-sided transformation of the form a0002 = Ra. This works if the axis of rotation is perpendicular to a, but otherwise does not return a pure quaternion. More damagingly, it forces one to interpret the basis quaternions as generators of rotations through π/2, which is simply wrong! Despite the problems with quaternions, it was clear to many that they were a useful mathematical system worthy of study. Tait claimed that quaternions ‘freed the physicist from the constraints of coordinates and allowed thoughts to run in their most natural channels’ — a theme we shall frequently meet in this book. Quaternions also found favour with the physicist James Clerk Maxwell, who employed them in his development of the theory of electromagnetism. Despite these successes, however, quaternions were weighed down by the increasingly dogmatic arguments over their interpretation and were eventually displaced by the hybrid system of vector algebra promoted by Gibbs.
1.5 The cross product Two of the lasting legacies of the quaternion story are the introduction of the idea of a vector, and the cross product between two vectors. Suppose we form the product of two pure quaternions a and b, where a = a1 i + a2 j + a3 k,
b = b1 i + b2 j + b3 k.
(1.29)
Their product can be written ab = −ai bi + c,
(1.30)
c = (a2 b3 − a3 b2 )i + (a3 b1 − a1 b3 )j + (a1 b2 − a2 b1 )k.
(1.31)
where c is the pure quaternion
Writing c = c1 i + c2 j + c3 k the component relation can be written as ci = 001aijk aj bk , 10
(1.32)
1.6 THE OUTER PRODUCT
where the alternating tensor 001aijk is defined by if ijk is a cylic permutation of 123, 1 001aijk = −1 if ijk is an anticylic permutation of 123, 0 otherwise.
(1.33)
We recognise the preceding as defining the cross product of two vectors, a×b. This has the following properties: (i) a×b is perpendicular to the plane defined by a and b; (ii) a×b has magnitude |a||b| sin(θ); (iii) the vectors a, b and a×b form a right-handed set. These properties can alternatively be viewed as defining the cross product, and from them the algebraic definition can be recovered. This is achieved by starting with a right-handed orthonormal frame {ei }. For these we must have e1 ×e2 = e3
etc.
(1.34)
so that we can write ei ×ej = 001aijk ek .
(1.35)
Expanding out a vector in terms of this basis recovers the formula a×b = (ai ei )×(bj ej ) = ai bj (ei ×ej ) = (001aijk ai bj )ek .
(1.36)
Hence the geometric definition recovers the algebraic one. The cross product quickly proved itself to be invaluable to physicists, dramatically simplifying equations in dynamics and electromagnetism. In the latter part of the nineteenth century many physicists, most notably Gibbs, advocated abandoning quaternions altogether and just working with the individual scalar and cross products. We shall see in later chapters that Gibbs was misguided in some of his objections to the quaternion product, but his considerable reputation carried the day and by the 1900s quaternions had all but disappeared from mainstream physics.
1.6 The outer product The cross product has one major failing — it only exists in three dimensions. In two dimensions there is nowhere else to go, whereas in four dimensions the concept of a vector orthogonal to a pair of vectors is not unique. To see this, consider four orthonormal vectors e1 , . . . , e4 . If we take the pair e1 and e2 and attempt 11
INTRODUCTION
Figure 1.3 Hermann Gunther Grassmann (1809–1877), born in Stettin, Germany (now Szczecin, Poland). A German mathematician and schoolteacher, Grassmann was the third of his parents’ twelve children and was born into a family of scholars. His father studied theology and became a minister, before switching to teaching mathematics and physics at the Stettin Gymnasium. Hermann followed in his father’s footsteps, first studying theology, classical languages and literature at Berlin. After returning to Stettin in 1830 he turned his attention to mathematics and physics. Grassmann passed the qualifying examination to win a teaching certificate in 1839. This exam included a written assignment on the tides, for which he gave a simplified treatment of Laplace’s work based upon a new geometric calculus that he had developed. By 1840 he had decided to concentrate on mathematics research. He published the first edition of his geometric calculus, the 300 page Lineale Ausdehnungslehre in 1844, the same year that Hamilton announced the discovery of the quaternions. His work did not achieve the same impact as the quaternions, however, and it was many years before his ideas were understood and appreciated by other mathematicians. Disappointed by this lack of interest, Grassmann turned his attention to linguistics and comparative philology, with greater immediate impact. He was an expert in Sanskrit and translated the Rig-Veda (1876–1877). He also formulated the linguistic law (named after him) stating that in Indo-European bases, successive syllables may not begin with aspirates. He died before he could see his ideas on geometry being adopted into mainstream mathematics.
to find a vector perpendicular to both of these, we see that any combination of e3 and e4 will do. A suitable generalisation of the idea of the cross product was constructed by 12
1.6 THE OUTER PRODUCT
a b∧a
a∧b b
b
θ
θ a
Figure 1.4 The outer product. The outer or wedge product of a and b returns a directed area element of area |a||b| sin(θ). The orientation of the parallelogram is defined by whether the circuit a, b, −a, −b is right-handed (anticlockwise) or left-handed (clockwise). Interchanging the order of the vectors reverses the orientation and introduces a minus sign in the product.
the remarkable German mathematician H.G. Grassmann (see figure 1.3). His work had its origin in the Barycentrischer Calcul of M¨ obius. There the author introduced expressions like AB for the line connecting the points A and B and ABC for the triangle defined by A, B and C. M¨ obius also introduced the crucial idea that the sign of the quantity should change if any two points are interchanged. (These oriented segments are now referred to as simplices.) It was Grassmann’s leap of genius to realise that expressions like AB could actually be viewed as a product between vectors. He thus introduced the outer or exterior product which, in modern notation, we write as a ∧ b, or ‘a wedge b’. The outer product can be defined on any vector space and, geometrically, we are not forced to picture these vectors as displacements. Indeed, Grassmann was motivated by a projective viewpoint, where the elements of the vector space are interpreted as points, and the outer product of two points defines the line through the points. For our purposes, however, it is simplest to adopt a picture in which vectors represent directed line segments. The outer product then provides a means of encoding a plane, without relying on the notion of a vector perpendicular to it. The result of the outer product is therefore neither a scalar nor a vector. It is a new mathematical entity encoding an oriented plane and is called a bivector. It can be visualised as the parallelogram obtained by sweeping one vector along the other (figure 1.4). Changing the order of the vectors reverses the orientation of the plane. The magnitude of a∧b is |a||b| sin(θ), the same as the area of the plane segment swept out by the vectors. The outer product of two vectors has the following algebraic properties: 13
INTRODUCTION
a a∧c
a∧b
c
b a
a∧(b + c)
b+c Figure 1.5 A geometric picture of bivector addition. In three dimensions any two non-parallel planes share a common line. If this line is denoted a, the two planes can be represented by a ∧ b and a ∧ c. Bivector addition proceeds much like vector addition. The planes are combined at a common boundary and the resulting plane is defined by the initial and final edges, as opposed to the initial and final points for vector addition. The mathematical statement of this addition rule is the distributivity of the outer product over addition.
(i) The product is antisymmetric: a∧b = −b∧a.
(1.37)
This has the geometric interpretation of reversing the orientation of the surface defined by a and b. It follows immediately that a∧a = 0,
for all vectors a.
(1.38)
(ii) Bivectors form a linear space, the same way that vectors do. In two and three dimensions the addition of bivectors is easy to visualise. In higher dimensions this addition is not always so easy to visualise, because two planes need not share a common line. (iii) The outer product is distributive over addition: a∧(b + c) = a∧b + a∧c.
(1.39)
This helps to visualise the addition of bivectors which share a common line (see figure 1.5). While it is convenient to visualise the outer product as a parallelogram, the 14
1.6 THE OUTER PRODUCT
actual shape of the object is not conveyed by the result of the product. This can be seen easily by defining a0002 = a + λb and forming a0002 ∧b = a∧b + λb∧b = a∧b.
(1.40)
The same bivector can therefore be generated by many different pairs of vectors. In many ways it is better to replace the picture of a directed parallelogram with that of a directed circle. The circle defines both the plane and a handedness, and its area is equal to the magnitude of the bivector. This therefore conveys all of the information one has about the bivector, though it does make bivector addition harder to visualise. 1.6.1 Two dimensions The outer product of any two vectors defines a plane, so one has to go to at least two dimensions to form an interesting product. Suppose then that {e1 , e2 } are an orthonormal basis for the plane, and introduce the vectors a = a 1 e 1 + a2 e 2 ,
b = b 1 e1 + b 2 e2 .
(1.41)
The outer product a ∧ b contains a∧b = a1 b1 e1 ∧e1 + a1 b2 e1 ∧e2 + a2 b1 e2 ∧e1 + a2 b2 e2 ∧e2 = (a1 b2 − a2 b1 )e1 ∧e2 ,
(1.42)
which recovers the imaginary part of the product of (1.18). The term therefore immediately has the expected magnitude |a| |b| sin(θ). The coefficient of e 1 ∧ e2 is positive if a and b have the same orientation as e1 and e2 . The orientation is defined by traversing the boundary of the parallelogram defined by the vectors a, b, −a, −b (see figure 1.4). By convention, we usually work with a right-handed set of reference axes (viewed from above). In this case the coefficient a1 b2 − a2 b1 will be positive if a and b also form a right-handed pair. 1.6.2 Three dimensions In three dimensions the space of bivectors is also three-dimensional, because each bivector can be placed in a one-to-one correspondence with the vector perpendicular to it. Suppose that {e1 , e2 , e3 } form a right-handed basis (see comments below), and the two vectors a and b are expanded in this basis as a = ai ei and b = bi ei . The bivector a ∧ b can then be decomposed in terms of an orthonormal frame of bivectors by a∧b = (ai ei )∧(bj ej ) = (a2 b3 − b3 a2 )e2 ∧e3 + (a3 b1 − a1 b3 )e3 ∧e1 + (a1 b2 − a2 b1 )e1 ∧e2 . 15
(1.43)
INTRODUCTION
The components in this frame are therefore the same as those of the cross product. But instead of being the components of a vector perpendicular to a and b, they are the components of the bivector a ∧ b. It is this distinction which enables the outer product to be defined in any dimension.
1.6.3 Handedness We have started to employ the idea of handedness without giving a satisfactory definition of it. The only space in which there is an unambiguous definition of handedness is three dimensions, as this is the space we inhabit and most of us can distinguish our left and right hands. This concept of ‘left’ and ‘right’ is a man-made convention adopted to make our life easier, and it extends to the concept of a frame in a straightforward way. Suppose that we are presented with three orthogonal vectors {e1 , e2 , e3 }. We align the 3 axis with the thumb of our right hand and then close our fist. If the direction in which our fist closes is the same as that formed by rotating from the 1 to the 2 axis, the frame is right-handed. If not, it is left-handed. Swapping any pair of vectors swaps the handedness of a frame. Performing two such swaps returns us to the original handedness. In three dimensions this corresponds to a cyclic reordering, and ensures that the frames {e1 , e2 , e3 }, {e3 , e1 , e2 } and {e2 , e3 , e1 } all have the same orientation. There is no agreed definition of a ‘right-handed’ orientation in spaces of dimensions other than three. All one can do is to make sure that any convention used is adopted consistently. In all dimensions the orientation of a set of vectors is changed if any two vectors are swapped. In two dimensions one does still tend to talk about right-handed axes, though the definition is dependent on the idea of looking down on the plane from above. The idea of above and below is not a feature of the plane itself, but depends on how we embed it in our three-dimensional world. There is no definition of left or right-handed which is intrinsic to the plane.
1.6.4 Extending the outer product The preceding examples demonstrate that in arbitrary dimensions the components of a∧b are given by (a∧b)ij = a[i bj]
(1.44)
where the [ ] denotes antisymmetrisation. Grassmann was able to take this idea further by defining an outer product for any number of vectors. The idea is a simple extension of the preceding formula. Expressed in an orthonormal frame, the components of the outer product on n vectors are the totally antisymmetrised 16
1.7 NOTES
products of the components of each vector. This definition has the useful property that the outer product is associative, a∧(b∧c) = (a∧b)∧c.
(1.45)
For example, in three dimensions we have a∧b∧c = (ai ei )∧(bj ej )∧(ck ek ) = 001aijk ai bj ck e1 ∧e2 ∧e3 ,
(1.46)
which represents a directed volume (see section 2.4). A further feature of the antisymmetry of the product is that the outer product of any set of linearly dependent vectors vanishes. This means that statements like ‘this vector lies on a given plane’, or ‘these two hypersurfaces share a common line’ can be encoded algebraically in a simple manner. Equipped with these ideas, Grassmann was able to construct a system capable of handling geometric concepts in arbitrary dimensions. Despite Grassmann’s considerable achievement, the book describing his ideas, his Lineale Ausdehnungslehre, did not have any immediate impact. This was no doubt due largely to his relative lack of reputation (he was still a German schoolteacher when he wrote this work). It was over twenty years before anyone of note referred to Grassmann’s work, and during this time Grassmann produced a second, extended version of the Ausdehnungslehre. In the latter part of the nineteenth century Grassmann’s work started to influence leading figures like Gibbs and Clifford. Gibbs wrote a number of papers praising Grassmann’s work and contrasting it favourably with the quaternion algebra. Clifford used Grassmann’s work as the starting point for the development of his geometric algebra, the subject of this book. Today, Grassmann’s ideas are recognised as the first presentation of the abstract theory of vector spaces over the field of real numbers. Since his death, his work has given rise to the influential and fashionable areas of differential forms and Grassmann variables. The latter are anticommuting variables and are fundamental to the foundations of much of modern supersymmetry and superstring theory. 1.7 Notes Descriptions of linear algebra and vector spaces can be found in most introductory textbooks of mathematics, as can discussions of the scalar and cross products and complex arithmetic. Quaternions, on the other hand, are much less likely to be mentioned. There is a large specialised literature on the quaternions, and a good starting point are the works of Altmann (1986, 1989). Altmann’s paper on ‘Hamilton, Rodriques and the quaternion scandal’ (1989) is also a good introduction to the history of the subject. The outer product is covered in most modern textbooks on geometry and 17
INTRODUCTION
physics, such as those by Nakahara (1990), Schutz (1980), and Gockeler & Schucker (1987). In most of these works, however, the exterior product is only treated in the context of differential forms. Applications to wider topics in geometry have been discussed by Hestenes (1991) and others. A useful summary in provided in the proceedings of the conference Hermann Gunther Grassmann (1809–1877), edited by Schubring (1996). Grassmann’s Lineale Ausdehnungslehre is also finally available in English translation due to Kannenberg (1995). For those with a deeper interest in the history of mathematics and the development of vector algebra a good starting point is the set of books by Kline (1972). There are also biographies available of many of the key protagonists. Perhaps even more interesting is to return to their original papers and experience first hand the robust and often humorous language employed at the time. The collected works of J.W. Gibbs (1906) are particularly entertaining and enlightening, and contain a good deal of valuable historical information.
1.8 Exercises 1.1
1.2
Suppose that the two sets {a1 , . . . , am } and {b1 , . . . , bn } form bases for the same vector space, and suppose initially that m > n. By establishing a contradiction, prove the basis theorem that all bases of a vector space have the same number of elements. Demonstrate that the following define vector spaces: (a) the set of all polynomials of degree n; (b) all solutions of a given linear ordinary differential equation; (c) the set of all n × m matrices.
1.3 1.4 1.5
1.6
Prove that in Euclidean space |a + b| ≤ |a| + |b|. When does equality hold? Show that the unit quaternions {±1, ±i, ±j ± k} form a discrete group. The unit quaternions i, j, k are generators of rotations about their respective axes. Are rotations through either π or π/2 consistent with the equation ijk = −1? Prove the following: (a) a·(b×c) = b·(c×a) = c·(a×b); (b) a×(b×c) = a·c b − a·b c; (c) |a×b| = |a| |b| sin(θ), where a·b = |a| |b| cos(θ).
1.7
Prove that the dimension of the space formed by the exterior product of m vectors drawn from a space of dimension n is n! n(n − 1) · · · (n − m) = . 1 · 2···m (n − m)!m! 18
1.8 EXERCISES
1.8 1.9
Prove that the n-fold exterior product of a set of n dependent vectors is zero. A convex polygon in a plane is specified by the ordered set of points {x0 , x1 , . . . , xn }. Prove that the directed area of the polygon is given by A = 12 (x0 ∧x1 + x1 ∧x2 + · · · + xn ∧x0 ). What is the significance of the sign? Can you extend the idea to a triangulated surface in three dimensions?
19
published by the press syndicate of the university of cambridge The Pitt Building, Trumpington Street, Cambridge, United Kingdom cambridge university press The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York, NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarc´ on 13, 28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http://www.cambridge.org 0001 C Cambridge University Press, 2003
This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2003 Printed in the United Kingdom at the University Press, Cambridge Typeface CMR 10/13 pt
System LATEX 2ε [TB]
A catalogue record for this book is available from the British Library Library of Congress Cataloguing in Publication data ISBN 0 521 48022 1 hardback
Contents
Preface Notation
ix xiii
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Introduction Vector (linear) spaces The scalar product Complex numbers Quaternions The cross product The outer product Notes Exercises
1 2 4 6 7 10 11 17 18
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9
Geometric algebra in two and three dimensions A new product for vectors An outline of geometric algebra Geometric algebra of the plane The geometric algebra of space Conventions Reflections Rotations Notes Exercises
20 21 23 24 29 38 40 43 51 52
3 3.1 3.2 3.3
Classical mechanics Elementary principles Two-body central force interactions Celestial mechanics and perturbations
54 55 59 64
v
CONTENTS
3.4 3.5 3.6
Rotating systems and rigid-body motion Notes Exercises
4 4.1 4.2 4.3 4.4 4.5 4.6 4.7
Foundations of geometric algebra Axiomatic development Rotations and reflections Bases, frames and components Linear algebra Tensors and components Notes Exercises
84 85 97 100 103 115 122 124
5 5.1 5.2 5.3 5.4 5.5 5.6 5.7
Relativity and spacetime An algebra for spacetime Observers, trajectories and frames Lorentz transformations The Lorentz group Spacetime dynamics Notes Exercises
126 127 131 138 143 150 163 164
6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8
Geometric calculus The vector derivative Curvilinear coordinates Analytic functions Directed integration theory Embedded surfaces and vector manifolds Elasticity Notes Exercises
167 168 173 178 183 202 220 224 225
7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8
Classical electrodynamics Maxwell’s equations Integral and conservation theorems The electromagnetic field of a point charge Electromagnetic waves Scattering and diffraction Scattering Notes Exercises
228 229 235 241 251 258 261 264 265
vi
69 81 82
CONTENTS
8 8.1 8.2 8.3 8.4 8.5 8.6 8.7
Quantum theory and spinors Non-relativistic quantum spin Relativistic quantum states The Dirac equation Central potentials Scattering theory Notes Exercises
267 267 278 281 288 297 305 307
9 9.1 9.2 9.3 9.4 9.5 9.6 9.7
Multiparticle states and quantum entanglement Many-body quantum theory Multiparticle spacetime algebra Systems of two particles Relativistic states and operators Two-spinor calculus Notes Exercises
309 310 315 319 325 332 337 337
10 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9
Geometry Projective geometry Conformal geometry Conformal transformations Geometric primitives in conformal space Intersection and reflection in conformal space Non-Euclidean geometry Spacetime conformal geometry Notes Exercises
340 341 351 355 360 365 370 383 390 391
11 11.1 11.2 11.3 11.4 11.5 11.6 11.7
Further topics in calculus and group theory Multivector calculus Grassmann calculus Lie groups Complex structures and unitary groups The general linear group Notes Exercises
394 394 399 401 408 412 416 417
12 12.1 12.2 12.3
Lagrangian and Hamiltonian techniques The Euler–Lagrange equations Classical models for spin-1/2 particles Hamiltonian techniques
420 421 427 432
vii
CONTENTS
12.4 12.5 12.6
Lagrangian field theory Notes Exercises
439 444 445
13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8
Symmetry and gauge theory Conservation laws in field theory Electromagnetism Dirac theory Gauge principles for gravitation The gravitational field equations The structure of the Riemann tensor Notes Exercises
448 449 453 457 466 474 490 495 495
14 Gravitation 14.1 Solving the field equations 14.2 Spherically-symmetric systems 14.3 Schwarzschild black holes 14.4 Quantum mechanics in a black hole background 14.5 Cosmology 14.6 Cylindrical systems 14.7 Axially-symmetric systems 14.8 Notes 14.9 Exercises Bibliography Index
viii
497 498 500 510 524 535 543 551 564 565 568 575
1
Introduction
The goal of expressing geometrical relationships through algebraic equations has dominated much of the development of mathematics. This line of thinking goes back to the ancient Greeks, who constructed a set of geometric laws to describe the world as they saw it. Their view of geometry was largely unchallenged until the eighteenth century, when mathematicians discovered new geometries with different properties to the Greeks’ Euclidean geometry. Each of these new geometries had distinct algebraic properties, and a major preoccupation of nineteenth century mathematicians was to place these geometries within a unified algebraic framework. One of the key insights in this process was made by W.K. Clifford, and this book is concerned with the implications of his discovery. Before we describe Clifford’s discovery (in chapter 2) we have gathered together some introductory material of use throughout this book. This chapter revises basic notions of vector spaces, emphasising pictorial representations of the underlying algebraic rules — a theme which dominates this book. The material is presented in a way which sets the scene for the introduction of Clifford’s product, in part by reflecting the state of play when Clifford conducted his research. To this end, much of this chapter is devoted to studying the various products that can be defined between vectors. These include the scalar and vector products familiar from three-dimensional geometry, and the complex and quaternion products. We also introduce the outer or exterior product, though this is covered in greater depth in later chapters. The material in this chapter is intended to be fairly basic, and those impatient to uncover Clifford’s insight may want to jump straight to chapter 2. Readers unfamiliar with the outer product are encouraged to read this chapter, however, as it is crucial to understanding Clifford’s discovery. 1
INTRODUCTION
1.1 Vector (linear) spaces At the heart of much of geometric algebra lies the idea of vector, or linear spaces. Some properties of these are summarised here and assumed throughout this book. In this section we talk in terms of vector spaces, as this is the more common term. For all other occurrences, however, we prefer to use the term linear space. This is because the term ‘vector ’ has a very specific meaning within geometric algebra (as the grade-1 elements of the algebra).
1.1.1 Properties Vector spaces are defined in terms of two objects. These are the vectors, which can often be visualised as directions in space, and the scalars, which are usually taken to be the real numbers. The vectors have a simple addition operation rule with the following obvious properties: (i) Addition is commutative: a + b = b + a.
(1.1)
a + (b + c) = (a + b) + c.
(1.2)
(ii) Addition is associative:
This property enables us to write expressions such as a + b + c without ambiguity. (iii) There is an identity element, denoted 0: a + 0 = a.
(1.3)
(iv) Every element a has an inverse −a: a + (−a) = 0.
(1.4)
For the case of directed line segments each of these properties has a clear geometric equivalent. These are illustrated in figure 1.1. Vector spaces also contain a multiplication operation between the scalars and the vectors. This has the property that for any scalar λ and vector a, the product λa is also a member of the vector space. Geometrically, this corresponds to the dilation operation. The following further properties also hold for any scalars λ, µ and vectors a and b: (i) (ii) (iii) (iv)
λ(a + b) = λa + λb; (λ + µ)a = λa + µa; (λµ)a = λ(µa); if 1λ = λ for all scalars λ then 1a = a for all vectors a. 2
1.1 VECTOR (LINEAR) SPACES
b
b c
a+b
a
a
a
b+c a+b a+b+c
b Figure 1.1 A geometric picture of vector addition. The result of a + b is formed by adding the tail of b to the head of a. As is shown, the resultant vector a + b is the same as b + a. This finds an algebraic expression in the statement that addition is commutative. In the right-hand diagram the vector a + b + c is constructed two different ways, as a + (b + c) and as (a + b) + c. The fact that the results are the same is a geometric expression of the associativity of vector addition.
The preceding set of rules serves to define a vector space completely. Note that the + operation connecting scalars is different from the + operation connecting the vectors. There is no ambiguity, however, in using the same symbol for both. The following two definitions will be useful later in this book: (i) Two vector spaces are said to be isomorphic if their elements can be placed in a one-to-one correspondence which preserves sums, and there is a one-to-one correspondence between the scalars which preserves sums and products. (ii) If U and V are two vector spaces (sharing the same scalars) and all the elements of U are contained in V, then U is said to form a subspace of V.
1.1.2 Bases and dimension The concept of dimension is intuitive for simple vector spaces — lines are onedimensional, planes are two-dimensional, and so on. Equipped with the axioms of a vector space we can proceed to a formal definition of the dimension of a vector space. First we need to define some terms. (i) A vector b is said to be a linear combination of the vectors a1 , . . . , an if scalars λ1 , . . . , λn can be found such that b = λ1 a1 + · · · + λn an =
n 0001
λi ai .
(1.5)
i=1
(ii) A set of vectors {a1 , . . . , an } is said to be linearly dependent if scalars 3
INTRODUCTION
λ1 , . . . , λn (not all zero) can be found such that λ1 a1 + · · · + λn an = 0.
(1.6)
If such a set of scalars cannot be found, the vectors are said to be linearly independent. (iii) A set of vectors {a1 , . . . , an } is said to span a vector space V if every element of V can be expressed as a linear combination of the set. (iv) A set of vectors which are both linearly independent and span the space V are said to form a basis for V. These definitions all carry an obvious, intuitive picture if one thinks of vectors in a plane or in three-dimensional space. For example, it is clear that two independent vectors in a plane provide a basis for all vectors in that plane, whereas any three vectors in the plane are linearly dependent. These axioms and definitions are sufficient to prove the basis theorem, which states that all bases of a vector space have the same number of elements. This number is called the dimension of the space. Proofs of this statement can be found in any textbook on linear algebra, and a sample proof is left to work through as an exercise. Note that any two vector spaces of the same dimension and over the same field are isomorphic. The axioms for a vector space define an abstract mathematical entity which is already well equipped for studying problems in geometry. In so doing we are not compelled to interpret the elements of the vector space as displacements. Often different interpretations can be attached to isomorphic spaces, leading to different types of geometry (affine, projective, finite, etc.). For most problems in physics, however, we need to be able to do more than just add the elements of a vector space; we need to multiply them in various ways as well. This is necessary to formalise concepts such as angles and lengths and to construct higher-dimensional surfaces from simple vectors. Constructing suitable products was a major concern of nineteenth century mathematicians, and the concepts they introduced are integral to modern mathematical physics. In the following sections we study some of the basic concepts that were successfully formulated in this period. The culmination of this work, Clifford’s geometric product, is introduced separately in chapter 2. At various points in this book we will see how the products defined in this section can all be viewed as special cases of Clifford’s geometric product.
1.2 The scalar product Euclidean geometry deals with concepts such as lines, circles and perpendicularity. In order to arrive at Euclidean geometry we need to add two new concepts 4
1.2 THE SCALAR PRODUCT
to our vector space. These are distances between points, which allow us to define a circle, and angles between vectors so that we can say that two lines are perpendicular. The introduction of a scalar product achieves both of these goals. Given any two vectors a, b, the scalar product a · b is a rule for obtaining a number with the following properties: (i) (ii) (iii) (iv)
a·b = b·a; a·(λb) = λ(a·b); a·(b + c) = a·b + a·c; a·a > 0, unless a = 0.
(When we study relativity, this final property will be relaxed.) The introduction of a scalar product allows us to define the length of a vector, |a|, by √ |a| = (a·a). (1.7) Here, and throughout this book, the positive square root is always implied by √ the symbol. The fact that we now have a definition of lengths and distances means that we have specified a metric space. Many different types of metric space can be constructed, of which the simplest are the Euclidean spaces we have just defined. The fact that for Euclidean space the inner product is positive-definite means that we have a Schwarz inequality of the form |a·b| ≤ |a| |b|.
(1.8)
The proof is straightforward: (a + λb)·(a + λb) ≥ 0 2
⇒ a·a + 2λa·b + λ b·b ≥ 0 2
⇒ (a·b) ≤ a·a b·b,
∀λ ∀λ (1.9)
where the last step follows by taking the discriminant of the quadratic in λ. Since all of the numbers in this inequality are positive we recover (1.8). We can now define the angle θ between a and b by a·b = |a||b| cos(θ).
(1.10)
Two vectors whose scalar product is zero are said to be orthogonal. It is usually convenient to work with bases in which all of the vectors are mutually orthogonal. If all of the basis vectors are further normalised to have unit length, they are said to form an orthonormal basis. If the set of vectors {e1 , . . . , en } denote such a basis, the statement that the basis is orthonormal can be summarised as ei ·ej = δij . 5
(1.11)
INTRODUCTION
Here the δij is the Kronecker delta function, defined by 0002 1 if i = j, δij = 0 if i = j.
(1.12)
We can expand any vector a in this basis as a=
n 0001
ai ei = ai ei ,
(1.13)
i=1
where we have started to employ the Einstein summation convention that pairs of indices in any expression are summed over. This convention will be assumed throughout this book. The {ai } are the components of the vector a in the {ei } basis. These are found simply by ai = ei ·a.
(1.14)
The scalar product of two vectors a = ai ei and b = bi ei can now written simply as a·b = (ai ei )·(bj ej ) = ai bj ei ·ej = ai bj δij = ai bi .
(1.15)
In spaces where the inner product is not positive-definite, such as Minkowski spacetime, there is no equivalent version of the Schwarz inequality. In such cases it is often only possible to define an ‘angle’ between vectors by replacing the cosine function with a cosh function. In these cases we can still introduce orthonormal frames and use these to compute scalar products. The main modification is that the Kronecker delta is replaced by ηij which again is zero if i = j, but can take values ±1 if i = j. 1.3 Complex numbers The scalar product is the simplest product one can define between vectors, and once such a product is defined one can formulate many of the key concepts of Euclidean geometry. But this is by no means the only product that can be defined between vectors. In two dimensions a new product can be defined via complex arithmetic. A complex number can be viewed as an ordered pair of real numbers which represents a direction in the complex plane, as was realised by Wessel in 1797. Their product enables complex numbers to perform geometric operations, such as rotations and dilations. But suppose that we take the complex number z = x + iy and square it, forming z 2 = (x + iy)2 = x2 − y 2 + 2xyi.
(1.16)
In terms of vector arithmetic, neither the real nor imaginary parts of this expression have any geometric significance. A more geometrically useful product 6
1.4 QUATERNIONS
is defined instead by zz ∗ = (x + iy)(x − iy) = x2 + y 2 ,
(1.17)
which returns the square of the length of the vector. A product of two vectors in a plane, z and w = u + vi, can therefore be constructed as zw∗ = (x + iy)(u − iv) = xu + vy + i(uy − vx).
(1.18)
The real part of the right-hand side recovers the scalar product. To understand the imaginary term consider the polar representation z = |z|eiθ ,
w = |w|eiφ
(1.19)
so that zw∗ = |z||w|ei(θ − φ) .
(1.20)
The imaginary term has magnitude |z||w| sin(θ − φ), where θ − φ is the angle between the two vectors. The magnitude of this term is therefore the area of the parallelogram defined by z and w. The sign of the term conveys information about the handedness of the area element swept out by the two vectors. This will be defined more carefully in section 1.6. We thus have a satisfactory interpretation for both the real and imaginary parts of the product zw ∗ . The surprising feature is that these are still both parts of a complex number. We thus have a second interpretation for complex addition, as a sum between scalar objects and objects representing plane segments. The advantages of adding these together are precisely the advantages of working with complex numbers as opposed to pairs of real numbers. This is a theme to which we shall return regularly in following chapters. 1.4 Quaternions The fact that complex arithmetic can be viewed as representing a product for vectors in a plane carries with it a further advantage — it allows us to divide by a vector. Generalising this to three dimensions was a major preoccupation of the physicist W.R. Hamilton (see figure 1.2). Since a complex number x + iy can be represented by two rectangular axes on a plane it seemed reasonable to represent directions in space by a triplet consisting of one real and two complex numbers. These can be written as x + iy + jz, where the third term jz represents a third axis perpendicular to the other two. The complex numbers i and j have the properties that i2 = j 2 = −1. The norm for such a triplet would then be (x + iy + jz)(x − iy − jz) = (x2 + y 2 + z 2 ) − yz(ij + ji).
(1.21)
The final term is problematic, as one would like to recover the scalar product here. The obvious solution to this problem is to set ij = −ji so that the last term vanishes. 7
INTRODUCTION
Figure 1.2 William Rowan Hamilton 1805–1865. Inventor of quaternions, and one of the key scientific figures of the nineteenth century. He spent many years frustrated at being unable to extend his theory of couples of numbers (complex numbers) to three dimensions. In the autumn of 1843 he returned to this problem, quite possibly prompted by a visit he received from the young German mathematician Eisenberg. Among Eisenberg’s papers was the observation that matrices form the elements of an algebra that was much like ordinary arithmetic except that multiplication was non-commutative. This was the vital step required to find the quaternion algebra. Hamilton arrived at this algebra on 16 October 1843 while out walking with his wife, and carved the equations in stone on Brougham Bridge. His discovery of quaternions is perhaps the best-documented mathematical discovery ever.
The anticommutative law ij = −ji ensures that the norm of a triplet behaves sensibly, and also that multiplication of triplets in a plane behaves in a reasonable manner. The same is not true for the general product of triplets, however. Consider (a + ib + jc)(x + iy + jz) = (ax − by − cz) + i(ay + bx) + j(az + cx) + ij(bz − cy).
(1.22)
Setting ij = −ji is no longer sufficient to remove the ij term, so the algebra does not close. The only thing for Hamilton to do was to set ij = k, where k is some unknown, and see if it could be removed somehow. While walking along the Royal Canal he suddenly realised that if his triplets were instead made up of four terms he would be able to close the algebra in a simple, symmetric way. 8
1.4 QUATERNIONS
To understand his discovery, consider (a + ib + jc + kd)(a − ib − jc − kd) = a2 + b2 + c2 + d2 (−k 2 ) − bd(ik + ki) − cd(jk + kj),
(1.23)
where we have assumed that i2 = j 2 = −1 and ij = −ji. The expected norm of the above product is a2 + b2 + c2 + d2 , which is obtained by setting k 2 = −1 and ik = −ki and jk = −kj. So what values do we use for jk and ik? These follow from the fact that ij = k, which gives ik = i(ij) = (ii)j = −j
(1.24)
kj = (ij)j = −i.
(1.25)
and
Thus the multiplication rules for quaternions are i2 = j 2 = k 2 = −1
(1.26)
and ij = −ji = k,
jk = −kj = i,
ki = −ik = j.
(1.27)
These can be summarised neatly as i2 = j 2 = k 2 = ijk = −1. It is a simple matter to check that these multiplication laws define a closed algebra. Hamilton was so excited by his discovery that the very same day he obtained leave to present a paper on the quaternions to the Royal Irish Academy. The subsequent history of the quaternions is a fascinating story which has been described by many authors. Some suggested material for further reading is given at the end of this chapter. In brief, despite the many advantages of working with quaternions, their development was blighted by two major problems. The first problem was the status of vectors in the algebra. Hamilton identified vectors with pure quaternions, which had a null scalar part. On the surface this seems fine — pure quaternions define a three-dimensional vector space. Indeed, Hamilton invented the word ‘vector ’ precisely for these objects and this is the origin of the now traditional use of i, j and k for a set of orthonormal basis vectors. Furthermore, the full product of two pure quaternions led to the definition of the extremely useful cross product (see section 1.5). The problem is that the product of two pure vectors does not return a new pure vector, so the vector part of the algebra does not close. This means that a number of ideas in complex analysis do not extend easily to three dimensions. Some people felt that this meant that the full quaternion product was of little use, and that the scalar and vector parts of the product should be kept separate. This criticism misses the point that the quaternion product is invertible, which does bring many advantages. The second major difficulty encountered with quaternions was their use in 9
INTRODUCTION
describing rotations. The irony here is that quaternions offer the clearest way of handling rotations in three dimensions, once one realises that they provide a ‘spin-1/2’ representation of the rotation group. That is, if a is a vector (a pure quaternion) and R is a unit quaternion, a new vector is obtained by the double-sided transformation law a0002 = RaR∗ ,
(1.28)
where the * operation reverses the sign of all three ‘imaginary’ components. A consequence of this is that each of the basis quaternions i, j and k generates rotations through π. Hamilton, however, was led astray by the analogy with complex numbers and tried to impose a single-sided transformation of the form a0002 = Ra. This works if the axis of rotation is perpendicular to a, but otherwise does not return a pure quaternion. More damagingly, it forces one to interpret the basis quaternions as generators of rotations through π/2, which is simply wrong! Despite the problems with quaternions, it was clear to many that they were a useful mathematical system worthy of study. Tait claimed that quaternions ‘freed the physicist from the constraints of coordinates and allowed thoughts to run in their most natural channels’ — a theme we shall frequently meet in this book. Quaternions also found favour with the physicist James Clerk Maxwell, who employed them in his development of the theory of electromagnetism. Despite these successes, however, quaternions were weighed down by the increasingly dogmatic arguments over their interpretation and were eventually displaced by the hybrid system of vector algebra promoted by Gibbs.
1.5 The cross product Two of the lasting legacies of the quaternion story are the introduction of the idea of a vector, and the cross product between two vectors. Suppose we form the product of two pure quaternions a and b, where a = a1 i + a2 j + a3 k,
b = b1 i + b2 j + b3 k.
(1.29)
Their product can be written ab = −ai bi + c,
(1.30)
c = (a2 b3 − a3 b2 )i + (a3 b1 − a1 b3 )j + (a1 b2 − a2 b1 )k.
(1.31)
where c is the pure quaternion
Writing c = c1 i + c2 j + c3 k the component relation can be written as ci = 001aijk aj bk , 10
(1.32)
1.6 THE OUTER PRODUCT
where the alternating tensor 001aijk is defined by if ijk is a cylic permutation of 123, 1 001aijk = −1 if ijk is an anticylic permutation of 123, 0 otherwise.
(1.33)
We recognise the preceding as defining the cross product of two vectors, a×b. This has the following properties: (i) a×b is perpendicular to the plane defined by a and b; (ii) a×b has magnitude |a||b| sin(θ); (iii) the vectors a, b and a×b form a right-handed set. These properties can alternatively be viewed as defining the cross product, and from them the algebraic definition can be recovered. This is achieved by starting with a right-handed orthonormal frame {ei }. For these we must have e1 ×e2 = e3
etc.
(1.34)
so that we can write ei ×ej = 001aijk ek .
(1.35)
Expanding out a vector in terms of this basis recovers the formula a×b = (ai ei )×(bj ej ) = ai bj (ei ×ej ) = (001aijk ai bj )ek .
(1.36)
Hence the geometric definition recovers the algebraic one. The cross product quickly proved itself to be invaluable to physicists, dramatically simplifying equations in dynamics and electromagnetism. In the latter part of the nineteenth century many physicists, most notably Gibbs, advocated abandoning quaternions altogether and just working with the individual scalar and cross products. We shall see in later chapters that Gibbs was misguided in some of his objections to the quaternion product, but his considerable reputation carried the day and by the 1900s quaternions had all but disappeared from mainstream physics.
1.6 The outer product The cross product has one major failing — it only exists in three dimensions. In two dimensions there is nowhere else to go, whereas in four dimensions the concept of a vector orthogonal to a pair of vectors is not unique. To see this, consider four orthonormal vectors e1 , . . . , e4 . If we take the pair e1 and e2 and attempt 11
INTRODUCTION
Figure 1.3 Hermann Gunther Grassmann (1809–1877), born in Stettin, Germany (now Szczecin, Poland). A German mathematician and schoolteacher, Grassmann was the third of his parents’ twelve children and was born into a family of scholars. His father studied theology and became a minister, before switching to teaching mathematics and physics at the Stettin Gymnasium. Hermann followed in his father’s footsteps, first studying theology, classical languages and literature at Berlin. After returning to Stettin in 1830 he turned his attention to mathematics and physics. Grassmann passed the qualifying examination to win a teaching certificate in 1839. This exam included a written assignment on the tides, for which he gave a simplified treatment of Laplace’s work based upon a new geometric calculus that he had developed. By 1840 he had decided to concentrate on mathematics research. He published the first edition of his geometric calculus, the 300 page Lineale Ausdehnungslehre in 1844, the same year that Hamilton announced the discovery of the quaternions. His work did not achieve the same impact as the quaternions, however, and it was many years before his ideas were understood and appreciated by other mathematicians. Disappointed by this lack of interest, Grassmann turned his attention to linguistics and comparative philology, with greater immediate impact. He was an expert in Sanskrit and translated the Rig-Veda (1876–1877). He also formulated the linguistic law (named after him) stating that in Indo-European bases, successive syllables may not begin with aspirates. He died before he could see his ideas on geometry being adopted into mainstream mathematics.
to find a vector perpendicular to both of these, we see that any combination of e3 and e4 will do. A suitable generalisation of the idea of the cross product was constructed by 12
1.6 THE OUTER PRODUCT
a b∧a
a∧b b
b
θ
θ a
Figure 1.4 The outer product. The outer or wedge product of a and b returns a directed area element of area |a||b| sin(θ). The orientation of the parallelogram is defined by whether the circuit a, b, −a, −b is right-handed (anticlockwise) or left-handed (clockwise). Interchanging the order of the vectors reverses the orientation and introduces a minus sign in the product.
the remarkable German mathematician H.G. Grassmann (see figure 1.3). His work had its origin in the Barycentrischer Calcul of M¨ obius. There the author introduced expressions like AB for the line connecting the points A and B and ABC for the triangle defined by A, B and C. M¨ obius also introduced the crucial idea that the sign of the quantity should change if any two points are interchanged. (These oriented segments are now referred to as simplices.) It was Grassmann’s leap of genius to realise that expressions like AB could actually be viewed as a product between vectors. He thus introduced the outer or exterior product which, in modern notation, we write as a ∧ b, or ‘a wedge b’. The outer product can be defined on any vector space and, geometrically, we are not forced to picture these vectors as displacements. Indeed, Grassmann was motivated by a projective viewpoint, where the elements of the vector space are interpreted as points, and the outer product of two points defines the line through the points. For our purposes, however, it is simplest to adopt a picture in which vectors represent directed line segments. The outer product then provides a means of encoding a plane, without relying on the notion of a vector perpendicular to it. The result of the outer product is therefore neither a scalar nor a vector. It is a new mathematical entity encoding an oriented plane and is called a bivector. It can be visualised as the parallelogram obtained by sweeping one vector along the other (figure 1.4). Changing the order of the vectors reverses the orientation of the plane. The magnitude of a∧b is |a||b| sin(θ), the same as the area of the plane segment swept out by the vectors. The outer product of two vectors has the following algebraic properties: 13
INTRODUCTION
a a∧c
a∧b
c
b a
a∧(b + c)
b+c Figure 1.5 A geometric picture of bivector addition. In three dimensions any two non-parallel planes share a common line. If this line is denoted a, the two planes can be represented by a ∧ b and a ∧ c. Bivector addition proceeds much like vector addition. The planes are combined at a common boundary and the resulting plane is defined by the initial and final edges, as opposed to the initial and final points for vector addition. The mathematical statement of this addition rule is the distributivity of the outer product over addition.
(i) The product is antisymmetric: a∧b = −b∧a.
(1.37)
This has the geometric interpretation of reversing the orientation of the surface defined by a and b. It follows immediately that a∧a = 0,
for all vectors a.
(1.38)
(ii) Bivectors form a linear space, the same way that vectors do. In two and three dimensions the addition of bivectors is easy to visualise. In higher dimensions this addition is not always so easy to visualise, because two planes need not share a common line. (iii) The outer product is distributive over addition: a∧(b + c) = a∧b + a∧c.
(1.39)
This helps to visualise the addition of bivectors which share a common line (see figure 1.5). While it is convenient to visualise the outer product as a parallelogram, the 14
1.6 THE OUTER PRODUCT
actual shape of the object is not conveyed by the result of the product. This can be seen easily by defining a0002 = a + λb and forming a0002 ∧b = a∧b + λb∧b = a∧b.
(1.40)
The same bivector can therefore be generated by many different pairs of vectors. In many ways it is better to replace the picture of a directed parallelogram with that of a directed circle. The circle defines both the plane and a handedness, and its area is equal to the magnitude of the bivector. This therefore conveys all of the information one has about the bivector, though it does make bivector addition harder to visualise. 1.6.1 Two dimensions The outer product of any two vectors defines a plane, so one has to go to at least two dimensions to form an interesting product. Suppose then that {e1 , e2 } are an orthonormal basis for the plane, and introduce the vectors a = a 1 e 1 + a2 e 2 ,
b = b 1 e1 + b 2 e2 .
(1.41)
The outer product a ∧ b contains a∧b = a1 b1 e1 ∧e1 + a1 b2 e1 ∧e2 + a2 b1 e2 ∧e1 + a2 b2 e2 ∧e2 = (a1 b2 − a2 b1 )e1 ∧e2 ,
(1.42)
which recovers the imaginary part of the product of (1.18). The term therefore immediately has the expected magnitude |a| |b| sin(θ). The coefficient of e 1 ∧ e2 is positive if a and b have the same orientation as e1 and e2 . The orientation is defined by traversing the boundary of the parallelogram defined by the vectors a, b, −a, −b (see figure 1.4). By convention, we usually work with a right-handed set of reference axes (viewed from above). In this case the coefficient a1 b2 − a2 b1 will be positive if a and b also form a right-handed pair. 1.6.2 Three dimensions In three dimensions the space of bivectors is also three-dimensional, because each bivector can be placed in a one-to-one correspondence with the vector perpendicular to it. Suppose that {e1 , e2 , e3 } form a right-handed basis (see comments below), and the two vectors a and b are expanded in this basis as a = ai ei and b = bi ei . The bivector a ∧ b can then be decomposed in terms of an orthonormal frame of bivectors by a∧b = (ai ei )∧(bj ej ) = (a2 b3 − b3 a2 )e2 ∧e3 + (a3 b1 − a1 b3 )e3 ∧e1 + (a1 b2 − a2 b1 )e1 ∧e2 . 15
(1.43)
INTRODUCTION
The components in this frame are therefore the same as those of the cross product. But instead of being the components of a vector perpendicular to a and b, they are the components of the bivector a ∧ b. It is this distinction which enables the outer product to be defined in any dimension.
1.6.3 Handedness We have started to employ the idea of handedness without giving a satisfactory definition of it. The only space in which there is an unambiguous definition of handedness is three dimensions, as this is the space we inhabit and most of us can distinguish our left and right hands. This concept of ‘left’ and ‘right’ is a man-made convention adopted to make our life easier, and it extends to the concept of a frame in a straightforward way. Suppose that we are presented with three orthogonal vectors {e1 , e2 , e3 }. We align the 3 axis with the thumb of our right hand and then close our fist. If the direction in which our fist closes is the same as that formed by rotating from the 1 to the 2 axis, the frame is right-handed. If not, it is left-handed. Swapping any pair of vectors swaps the handedness of a frame. Performing two such swaps returns us to the original handedness. In three dimensions this corresponds to a cyclic reordering, and ensures that the frames {e1 , e2 , e3 }, {e3 , e1 , e2 } and {e2 , e3 , e1 } all have the same orientation. There is no agreed definition of a ‘right-handed’ orientation in spaces of dimensions other than three. All one can do is to make sure that any convention used is adopted consistently. In all dimensions the orientation of a set of vectors is changed if any two vectors are swapped. In two dimensions one does still tend to talk about right-handed axes, though the definition is dependent on the idea of looking down on the plane from above. The idea of above and below is not a feature of the plane itself, but depends on how we embed it in our three-dimensional world. There is no definition of left or right-handed which is intrinsic to the plane.
1.6.4 Extending the outer product The preceding examples demonstrate that in arbitrary dimensions the components of a∧b are given by (a∧b)ij = a[i bj]
(1.44)
where the [ ] denotes antisymmetrisation. Grassmann was able to take this idea further by defining an outer product for any number of vectors. The idea is a simple extension of the preceding formula. Expressed in an orthonormal frame, the components of the outer product on n vectors are the totally antisymmetrised 16
1.7 NOTES
products of the components of each vector. This definition has the useful property that the outer product is associative, a∧(b∧c) = (a∧b)∧c.
(1.45)
For example, in three dimensions we have a∧b∧c = (ai ei )∧(bj ej )∧(ck ek ) = 001aijk ai bj ck e1 ∧e2 ∧e3 ,
(1.46)
which represents a directed volume (see section 2.4). A further feature of the antisymmetry of the product is that the outer product of any set of linearly dependent vectors vanishes. This means that statements like ‘this vector lies on a given plane’, or ‘these two hypersurfaces share a common line’ can be encoded algebraically in a simple manner. Equipped with these ideas, Grassmann was able to construct a system capable of handling geometric concepts in arbitrary dimensions. Despite Grassmann’s considerable achievement, the book describing his ideas, his Lineale Ausdehnungslehre, did not have any immediate impact. This was no doubt due largely to his relative lack of reputation (he was still a German schoolteacher when he wrote this work). It was over twenty years before anyone of note referred to Grassmann’s work, and during this time Grassmann produced a second, extended version of the Ausdehnungslehre. In the latter part of the nineteenth century Grassmann’s work started to influence leading figures like Gibbs and Clifford. Gibbs wrote a number of papers praising Grassmann’s work and contrasting it favourably with the quaternion algebra. Clifford used Grassmann’s work as the starting point for the development of his geometric algebra, the subject of this book. Today, Grassmann’s ideas are recognised as the first presentation of the abstract theory of vector spaces over the field of real numbers. Since his death, his work has given rise to the influential and fashionable areas of differential forms and Grassmann variables. The latter are anticommuting variables and are fundamental to the foundations of much of modern supersymmetry and superstring theory. 1.7 Notes Descriptions of linear algebra and vector spaces can be found in most introductory textbooks of mathematics, as can discussions of the scalar and cross products and complex arithmetic. Quaternions, on the other hand, are much less likely to be mentioned. There is a large specialised literature on the quaternions, and a good starting point are the works of Altmann (1986, 1989). Altmann’s paper on ‘Hamilton, Rodriques and the quaternion scandal’ (1989) is also a good introduction to the history of the subject. The outer product is covered in most modern textbooks on geometry and 17
INTRODUCTION
physics, such as those by Nakahara (1990), Schutz (1980), and Gockeler & Schucker (1987). In most of these works, however, the exterior product is only treated in the context of differential forms. Applications to wider topics in geometry have been discussed by Hestenes (1991) and others. A useful summary in provided in the proceedings of the conference Hermann Gunther Grassmann (1809–1877), edited by Schubring (1996). Grassmann’s Lineale Ausdehnungslehre is also finally available in English translation due to Kannenberg (1995). For those with a deeper interest in the history of mathematics and the development of vector algebra a good starting point is the set of books by Kline (1972). There are also biographies available of many of the key protagonists. Perhaps even more interesting is to return to their original papers and experience first hand the robust and often humorous language employed at the time. The collected works of J.W. Gibbs (1906) are particularly entertaining and enlightening, and contain a good deal of valuable historical information.
1.8 Exercises 1.1
1.2
Suppose that the two sets {a1 , . . . , am } and {b1 , . . . , bn } form bases for the same vector space, and suppose initially that m > n. By establishing a contradiction, prove the basis theorem that all bases of a vector space have the same number of elements. Demonstrate that the following define vector spaces: (a) the set of all polynomials of degree n; (b) all solutions of a given linear ordinary differential equation; (c) the set of all n × m matrices.
1.3 1.4 1.5
1.6
Prove that in Euclidean space |a + b| ≤ |a| + |b|. When does equality hold? Show that the unit quaternions {±1, ±i, ±j ± k} form a discrete group. The unit quaternions i, j, k are generators of rotations about their respective axes. Are rotations through either π or π/2 consistent with the equation ijk = −1? Prove the following: (a) a·(b×c) = b·(c×a) = c·(a×b); (b) a×(b×c) = a·c b − a·b c; (c) |a×b| = |a| |b| sin(θ), where a·b = |a| |b| cos(θ).
1.7
Prove that the dimension of the space formed by the exterior product of m vectors drawn from a space of dimension n is n! n(n − 1) · · · (n − m) = . 1 · 2···m (n − m)!m! 18
1.8 EXERCISES
1.8 1.9
Prove that the n-fold exterior product of a set of n dependent vectors is zero. A convex polygon in a plane is specified by the ordered set of points {x0 , x1 , . . . , xn }. Prove that the directed area of the polygon is given by A = 12 (x0 ∧x1 + x1 ∧x2 + · · · + xn ∧x0 ). What is the significance of the sign? Can you extend the idea to a triangulated surface in three dimensions?
19
Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required. Geometric Algebra for Physicists by Chris Doran Paperback $84.73. Ships from and sold by SpeedyHen. Download geometric algebra for physicists or read online books in PDF, EPUB, Tuebl, and Mobi Format. Click Download or Read Online button to get geometric algebra for physicists book now. This site is like a library, Use search box in the widget to get ebook that you want. Geometric algebra is a powerful mathematical language with applications across a range of subjects in physics and engineering. This book is a complete guide to the current state of the subject with early chapters providing a self-contained introduction to geometric algebra. Download or Read Book Geometric algebra has established itself as a powerful and valuable mathematical tool for solving problems in computer science, engineering, physics, and mathematics. Reading Geometric Algebra For Physicists Pdf Book everyone. Download file Free Book PDF Geometric Algebra For Physicists Pdf at Complete PDF Library. This Book have some digital formats such us: paperbook, ebook, kindle, epub, and another formats. Here is The Complete PDF Book Library. It's free to register here to get Book file PDF Geometric.
Publisher Description
Geometric algebra is a powerful mathematical language with applications across a range of subjects in physics and engineering. This book is a complete guide to the current state of the subject with early chapters providing a self-contained introduction to geometric algebra. Topics covered include new techniques for handling rotations in arbitrary dimensions, and the links between rotations, bivectors and the structure of the Lie groups. Following chapters extend the concept of a complex analytic function theory to arbitrary dimensions, with applications in quantum theory and electromagnetism. Later chapters cover advanced topics such as non-Euclidean geometry, quantum entanglement, and gauge theories. Applications such as black holes and cosmic strings are also explored. It can be used as a graduate text for courses on the physical applications of geometric algebra and is also suitable for researchers working in the fields of relativity and quantum theory.
Geometric Algebra For Physicists Download Free
![Geometric algebra for physicists download 1 Geometric algebra for physicists download 1](https://images-na.ssl-images-amazon.com/images/I/41Zyilj1aqL._SY291_BO1,204,203,200_QL40_.jpg)
![Algebraic geometry vs geometric algebra Algebraic geometry vs geometric algebra](https://cdn.eeweb.com/tools/assets/math_help/geometry-shapes-solids.png)