Tensor Calculus and Differential Geometry: Course Notes

Course Notes
Tensor Calculus and Differential Geometry

2WAH0
Luc Florack
March 21, 2019

Cover illustration: papyrus fragment from Euclid’s Elements of Geometry, Book II [8].
Contents
Preface iii
Notation 1
1 Prerequisites from Linear Algebra 3
2 Tensor Calculus 7
2.1 Vector Spaces and Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Dual Vector Spaces and Dual Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 The Kronecker Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 Inner Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5 Reciprocal Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Bases, Dual Bases, Reciprocal Bases: Mutual Relations . . . . . . . . . . . . . . . . . . . . . . 16
2.7 Examples of Vectors and Covectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.8 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.8.1 Tensors in all Generality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.8.2 Tensors Subject to Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.8.3 Symmetry and Antisymmetry Preserving Product Operators . . . . . . . . . . . . . . . 24
2.8.4 Vector Spaces with an Oriented Volume . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.8.5 Tensors on an Inner Product Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.8.6 Tensor Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.8.6.1 “Absolute Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

CONTENTS i
2.8.6.2 “Relative Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.8.6.3 “Pseudo Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.8.7 Contractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.9 The Hodge Star Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3 Differential Geometry 45
3.1 Euclidean Space: Cartesian and Curvilinear Coordinates . . . . . . . . . . . . . . . . . . . . . 45
3.2 Differentiable Manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3 Tangent Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.4 Tangent and Cotangent Bundle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.5 Exterior Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.6 Affine Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.7 Lie Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.8 Torsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.9 Levi-Civita Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.10 Geodesics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.11 Curvature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.12 Push-Forward and Pull-Back . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.13 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.13.1 Polar Coordinates in the Euclidean Plane . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.13.2 A Helicoidal Extension of the Euclidean Plane . . . . . . . . . . . . . . . . . . . . . . 60

ii CONTENTS
Preface
These course notes are intended for students of all TU/e departments that wish to learn the basics of tensor
calculus and differential geometry. Prerequisites are linear algebra and vector calculus at an introductory level.
The treatment is condensed, and serves as a complementary source next to more comprehensive accounts that
can be found in the (abundant) literature. As a companion for classroom adoption it does provide a reasonably
self-contained introduction to the subject that should prepare the student for further self-study.
These course notes are based on course notes written in Dutch by Jan de Graaf. Large parts are straightforward
translations. I am therefore indebted to Jan de Graaf for many of the good things. I have, however, taken the liberty
to skip, rephrase, and add material, and will continue to update these course notes (the date on the cover reflects
the version). I am of course solely to blame for the errors that might have been introduced in this undertaking.
Rigor is difficult to reconcile with simplicity of terminology and notation. I have adopted Jan de Graaf’s style
of presentation in this respect, by letting the merit of simplicity prevail. The necessary sacrifice of rigor is
compensated by a great number of interspersed “caveats”, notational and terminological remarks, all meant to
train the reader in coming to grips with the parlance of tensor calculus and differential geometry.
Luc Florack
Eindhoven, March 21, 2019.

iv
Notation
Instead of rigorous notational declarations, a non-exhaustive list of examples is provided illustrating the notation
for the most important object types used in these course notes:
• Linear mappings: A , B, etc.

• Matrices: A, B, etc.
• Tensors: A, B, etc.
• Vectors: a, b, etc.
• Covectors: â, b̂, etc.
• Basis vectors: ei , fi , etc. (for holonomic basis also ∂/∂xi , ∂/∂y i , etc.)
• Basis covectors: êi , f̂ i , etc. (for holonomic basis also dxi , dy i , etc.)
• Vector components: ai , bi , etc.
• Covector components: ai , bi , etc.

i ...i i ...i
• Tensor components: Aj11 ...jpq , Bj11 ...jpq , etc.
2
1. Prerequisites from Linear Algebra
Linear algebra forms the skeleton of tensor calculus and differential geometry. We recall a few basic definitions
from linear algebra, which will play a pivotal role throughout this course.
Reminder
A vector space V over the field K (R or C) is a set of objects that can be added and multiplied by scalars, such
that the sum of two elements of V as well as the product of an element in V with a scalar are again elements of
V (closure property). Moreover, the following properties are satisfied for any u, v, w ∈ V and λ, µ ∈ K:
• (u + v) + w = u + (v + w),
• there exists an element o ∈ V such that o + u = u,
• there exists an element −u ∈ V such that u + (−u) = o,
• u + v = v + u,
• λ · (u + v) = λ · u + λ · v,
• (λ + µ) · u = λ · u + µ · u,
• (λµ) · u = λ · (µ · u),
• 1 · u = u.
The infix operator + denotes vector addition, the infix operator · denotes scalar multiplication.
Terminology
Another term for vector space is linear space.
Remark
• We will almost exclusively consider real vector spaces, i.e. with scalar field K = R.
• We will invariably consider finite-dimensional vector spaces.
Reminder
A basis of an n-dimensional vector spacePV is a set {ei }ni=1 , such that for each x ∈ V there exists a unique
n
n-tuple (x1 , . . . , xn ) ∈ Rn such that x = i=1 xi ei .
4 Prerequisites from Linear Algebra
Reminder
• A linear operator A : V → W , in which V and W are vector spaces over K, is a mapping that satisfies
the following property: A (λ v + µ w) = λ A (v) + µ A (w) for all λ, µ ∈ K and v, w ∈ V .
• The set L (V, W ) of all linear operators of this type is itself a vector space, with the following definitions of
vector addition and scalar multiplication: (λA + µB) (v) = λA (v) + µB(v) for all λ, µ ∈ K, A , B ∈
L (V, W ) and v ∈ V .
Notation
The set of square n × n matrices will be denoted by Mn .
Reminder
The determinant of a square matrix A ∈ Mn with entries Aij (i, j = 1, . . . , n) is given by
n n
X 1 X
det A = [j1 , . . . , jn ] A1j1 . . . Anjn = [i1 , . . . , in ] [j1 , . . . , jn ] Ai1 j1 . . . Ain jn .
j1 ,...,jn =1
n!
i1 , . . . , in = 1
j1 , . . . , jn = 1
The completely antisymmetric symbol in n dimensions is defined as follows:


 +1 if (i1 , . . . , in ) is an even permutation of (1, . . . , n) ,
[i1 , . . . , in ] = −1 if (i1 , . . . , in ) is an odd permutation of (1, . . . , n) ,
0 otherwise .

Observation
• Exactly n! of the nn index realizations in [i1 , . . . , in ] yield nontrivial results.

• To determine det A, compose all n! products of n matrix entries such that the row labels as well as the
column labels of the factors in each product are all distinct. Rearrange factors such that the row labels are
incrementally ordered. Assign a sign ±1 to each product corresponding to the permutation (even/odd) of
the column labels observed in this ordered form, and add up everything. Cf. the following example.
Example
Let A ∈ M3 be a 3 × 3 matrix:  
A11 A12 A13
A =  A21 A22 A23  .
A31 A32 A33
Carrying out the recipe yields Table 1.1:
product with row labels ordered sign of column label permutation

A11 A22 A33 [123] = +1
A12 A23 A31 [231] = +1
A13 A21 A32 [312] = +1
A13 A22 A31 [321] = −1
A11 A23 A32 [132] = −1
A12 A21 A33 [213] = −1
Table 1.1: det A = A11 A22 A33 +A12 A23 A31 +A13 A21 A32 −A13 A22 A31 −A11 A23 A32 −A12 A21 A33 .
Prerequisites from Linear Algebra 5
Observation
For a general square matrix A ∈ Mn the following identities hold:
n
X
[i1 , . . . , in ] Ai1 j1 . . . Ain jn = [j1 , . . . , jn ] det A ,
i1 ,...,in =1
X n
[j1 , . . . , jn ] Ai1 j1 . . . Ain jn = [i1 , . . . , in ] det A ,
j1 ,...,jn =1
n
X n
X
[i1 , . . . , in ] [j1 , . . . , jn ] Ai1 j1 . . . Ain jn = n! det A .
i1 ,...,in =1 j1 ,...,jn =1
Products of completely antisymmetric symbols can be expressed in terms of determinants:

 
δi1 j1 . . . δi1 jn
[i1 , . . . , in ] [j1 , . . . , jn ] = det  ... ..  def i1 ...in
.  = δj1 ...jn ,

δin j1 ... δ in j n
and, more generally,

 
n δik+1 jk+1 ... δik+1 jn
X
[i1 , . . . , ik , ik+1 , . . . , in ] [i1 , . . . , ik , jk+1 , . . . , jn ] = k! det  .. ..
.
 
. .
i1 ,...,ik =1 δin jk+1 . . . δin jn
Example
• [i, j, k] [`, m, n] = δi` δjm δkn + δim δjn δk` + δin δj` δkm − δim δj` δkn − δi` δjn δkm − δin δjm δk` .
P3
• i=1 [i, j, k] [i, m, n] = δjm δkn − δjn δkm .
P3
• i,j=1 [i, j, k] [i, j, n] = 2δkn .
P3
• i,j,k=1 [i, j, k] [i, j, k] = 6.
Caveat
One often writes det Aij instead of det A. The labels i and j are not free indices in this case!
Reminder
The cofactor matrix of A ∈ Mn is the matrix Ã ∈ Mn with entries Ãij given by
∂ det A
Ãij = .
∂Aij
The adjugate matrix of A ∈ Mn is a synonymous term for ÃT ∈ Mn .
Notation
Cofactor and adjugate matrix of A ∈ Mn are sometimes indicated by cofA and adjA, respectively.
6 Prerequisites from Linear Algebra
Notation
In any expression of the form Xii , with a matching pair of sub- and superscript i = 1, . . . , n, in which n = dim V
is the dimension of an underlying vector space V , the Einstein summation convention applies, meaning that it is
to be understood as an i-index-free sum:
n
def
X
Xii = Xii .
i=1
This convention applies to each pair of matching sub- and superscript within the same index range. Indices to
which the convention applies are referred to as spatial (or, depending on context, spatiotemporal) indices. Note
that matching upper and lower indices are dummies that can be arbitrarily relabelled provided this does not lead
to conflicting notation.
Observation
• Each term in the polynomial det A contains at most one factor Aij for each fixed i, j = 1, . . . , n.
∂ Ãij ∂ 2 det A
• As a result, Ãij does not depend on Aij for fixed i, j = 1, . . . , n: = = 0.
∂Aij ∂A2ij
d det A ∂ det A dAij dAij

• If A : R → Mn : t 7→ A(t) is a matrix-valued function, = = Ãij (chain rule).
dt ∂Aij dt dt
• AÃT = ÃT A = det A I, or A adjA = adjA A = det A I, i.e. Ajk Ãik = Ãki Akj = det A δji .
• If A is invertible, then ÃT = adjA = det A A−1 .

• Each entry Ãij of Ã = cofA is a homogeneous polynomial of degree n−1 in terms of the entries Aij of A.
• Unlike A−1 , Ã = cofA and ÃT = adjA are always well-defined, even if A is singular.
2. Tensor Calculus
2.1. Vector Spaces and Bases
Ansatz
An n-dimensional vector space V over R furnished with a basis {ei }.
Notation
Unless stated otherwise the Einstein summation convention applies to identical pairs of upper and lower indices.
def Pn
Thus if Xji is any collection of numbers, then Xii = i=1 Xii .
Terminology
The act of constraining one upper and one lower index to be equal—and thus tacitly assuming summation, v.s.—is
called an index contraction, or a trace.
Remark
Contracted indices are dummies that can be arbitrarily relabeled, as opposed to free indices. Thus for example
Xii = Xjj .
Remark

For each basis {ei } of V and each x ∈ V there exists a unique collection of real numbers xi such that
x = xi ei , recall chapter 1.
Terminology
The real numbers xi are referred to as the contravariant components of the vector x relative to the vector basis
{ei }.
Definition 2.1 The Kronecker symbol, or Kronecker-δ, is defined as

1 if i = j
δji =
0 if i 6= j .
Theorem 2.1 Let x = xi ei ∈ V and fj = Aij ei define a change of basis. If xi ei = y i fi , then

xi = Aij y j or, equivalently, y i = Bji xj ,
in which, by definition, Aik Bjk = δji .
Caveat
Pay attention to the occurrence of the components Aij in the transformation of basis vectors ei and contravariant
components xi , respectively.
8 Tensor Calculus
Observation

Given a smooth manifold—to
be defined later—with a coordinate system xi , the collection of linear partial
derivative operators ei ≡ ∂i ≡ ∂/∂xi at any (implicit) fiducial point defines a vector space basis. Upon a
coordinate transformation xi = xi (xj ) the chain rule relates the corresponding bases by
∂ j = Aij ∂i ,
in which fj ≡ ∂ j ≡ ∂/∂xj , and

∂xi
Aij = .
∂xj
This is formally identical to the basis transformation in Theorem 2.1 on page 7.
Notation
On a smooth manifold one often encounters vector decomposition relative to a coordinate basis, i.e. v = v i ∂i .
(The fiducial point is implicit here.)
Terminology
The set {∂i } is called a coordinate (vector) basis, or a holonomic (vector) basis. By definition it spans the tangent
space of the manifold at the fiducial point.
Reminder
• Basis vectors are labeled with lower indices.

• Contravariant vector components are labeled with upper indices.
Remark
The distinction into upper and lower indices is relevant only if the indices are intrinsically constrained to take
values in the range {1, . . . , n}, where n = dim V .
2.2. Dual Vector Spaces and Dual Bases
Ansatz
An n-dimensional vector space V over R.
Definition 2.2 A linear functional on V is a linear mapping from V into R. In other words, f̂ : V → R is a linear
functional if for all x, y ∈ V , λ, µ ∈ R,
f̂ (λx + µy) = λf̂ (x) + µf̂ (y) .
Remark
The set V ∗ consisting of all linear functionals on V can be turned into a vector space in the usual way, viz. let
f̂ , ĝ ∈ V ∗ and λ, µ ∈ R, define the linear combination λf̂ + µĝ ∈ V ∗ by virtue of the identification
(λf̂ + µĝ)(x) = λf̂ (x) + µĝ(x)
for all x ∈ V , recall chapter 1.
Definition 2.3 The dual space is the vector space of linear functionals on V : V ∗ = L (V, R).
2.2 Dual Vector Spaces and Dual Bases 9
Ansatz
An n-dimensional vector space V over R furnished with a basis {ei }, and its dual space V ∗ .
Remark
Each basis {ei } of V induces a basis êi of V ∗ as follows. If x = xi ei , then êi (x) = xi .

Result 2.1 The dual basis êi of V ∗ is uniquely determined through its defining property

êi (ej ) = δji .
Proof of Result 2.1. Let f̂ ∈ V ∗ .
• For any x = xi ei ∈ V we have f̂ (x) = f̂ (xi ei ) = f̂ (ei )xi = f̂ (ei )êi (x) = (f̂ (ei )êi )(x) = (fi êi )(x)
with fi = f̂ (ei ). This shows that the set {êi } spans V ∗ .
• Suppose αi êi = 0 ∈ V ∗ , then αi = αj δij = αj êj (ei ) = (αj êj )(ei ) = 0 ∈ R for all i = 1, . . . , n. This
shows that {êi } is a linearly independent set.
Assumption
vector space V (with basis {ei }) implies
In view of this result, it will henceforth be tacitly understood that agiven
the existence of its dual space V ∗ (with corresponding dual basis êi ).
Observation
The dual basis V ∗ has the same dimension as V .
Terminology
A linear functional f̂ ∈ V ∗ is alternatively referred to as a covector or 1-form. The dual space V ∗ of a vector
space V is also known as the covector space or the space of 1-forms.
Remark
For each covector basis êi of V ∗ and each x̂ ∈ V ∗ there exists a unique collection of real numbers {xi } such

that x̂ = xi êi .
Terminology

The real numbers xi are referred to as the covariant components of the covector x̂ relative to covector basis êi .
We end this section with the dual analogue of Theorem 2.1 on page 7:
Theorem 2.2 The change of basis fj = Aij ei induces a change of dual vector basis
f̂ j = Bij êi ,
in which Akj Bki = δji . Consequently, if x̂ = xi êi = yi f̂ i ∈ V ∗ , then
xi = Bij yj or, equivalently, yi = Aji xj .
Proof of Theorem 2.2.
• Given fj = A`j e` and f̂ i = Bki êk we obtain δji = f̂ i (fj ) = Bki êk (A`j e` ) = Bki A`j êk (e` ) = Bki A`j δ`k =
Bki Akj , from which it follows that B = A−1 .
10 Tensor Calculus
• Thus if we define xi êi = yi f̂ i , then yj f̂ j = yj Bij êi , whence xi = Bij yj .

• Inversion yields the last identity.
Terminology
• Theorem 2.1, page 7, defines the so-called vector transformation law.

• Theorem 2.2, page 9, defines the so-called covector transformation law.
Observation

Recall the observation on page 8. Given a smooth manifold with a coordinate system xi , and a coordinate
transformation xi = xi (xj ), the chain rule (at any fiducial point) states that
∂xj
dxj = Bij dxi in which Bij = .
∂xi
Identifying êi ≡ dxi and f̂ j ≡ dxj , this is formally identical to the dual vector basis transformation in
Theorem 2.2. In particular we have, recall Result 2.1,
dxi (∂j ) = δji .
Notation
On a smooth manifold one often encounters covector decomposition relative to a coordinate basis, i.e. v̂ = vi dxi .
(The fiducial point is implicit here.)
Terminology
The set {dxi } is called a coordinate (dual vector) basis, or holonomic (dual vector) basis. By definition it spans
the cotangent space of the manifold at the fiducial point.
Reminder
• Basis covectors are labeled with upper indices.

• Covariant components are labeled with lower indices.
2.3. The Kronecker Tensor
Ansatz
Notation
Instead of a standard function notation for f̂ ∈ V ∗ one often employs the bracket formalism hf̂ , · i ∈ V ∗ . Function
evaluation is achieved by dropping a vector argument x ∈ V into the empty slot: hf̂ , xi = f̂ (x) ∈ R. In particular
we have
hêi , ej i = êi (ej ) = δji
for dual bases, recall Result 2.1 on page 9.
Definition 2.4 The bilinear map h · , · i : V ∗ × V → R : (f̂ , x) 7→ hf̂ , xi is called the Kronecker tensor.
2.4 Inner Products 11
Remark
The general concept of a tensor will be explained in section 2.8.
Caveat
Do not confuse the Kronecker tensor with an inner product. Unlike the Kronecker tensor an inner product requires
two vector arguments from the same vector space.
Observation
The Kronecker tensor is bilinear:
• ∀ f̂ , ĝ ∈ V ∗ , x ∈ V, λ, µ ∈ R : hλf̂ + µĝ, xi = λhf̂ , xi + µhĝ, xi,
• ∀ f̂ ∈ V ∗ , x, y ∈ V, λ, µ ∈ R : hf̂ , λx + µyi = λhf̂ , xi + µhf̂ , yi.

• Result 2.1 on page 9 shows that the components of the Kronecker tensor are given by the Kronecker symbol.
• If f̂ = fi êi ∈ V ∗ and x = xi ei ∈ V , then bilinearity of the Kronecker tensor implies hf̂ , xi =
fi xj hêi , ej i = fi xj δji = fi xi . Thus Kronecker tensor and index contraction are close-knit.
Remark
• The components of the Kronecker tensor are independent of the choice of basis. (Verify!)
• Given V , the assumption on page 9 implies that the Kronecker tensor exists.
• Since V ∗ is itself a vector space one may consider its dual (V ∗ )∗ = V ∗∗ . Without proof we state that if V
(and thus V ∗ ) is finite-dimensional, then V ∗∗ is isomorphic to, i.e. may be identified with V . This does not
generalize to infinite-dimensional vector spaces.
Notation
The use of the formally equivalent notations f̂ and hf̂ , · i for an element of V ∗ reflects our inclination to regard
it either as an atomic entity, or as a linear mapping with unspecified argument. The previous remark justifies the
use of similar alternative notations for an element of V , viz. either as x, say, or as h · , xi ∈ V ∗∗ ∼ V .
2.4. Inner Products
Ansatz
Definition 2.5 An inner product on a real vector space V is a nondegenerate positive symmetric bilinear mapping
of the form ( · | · ) : V × V → R which satisfies
• ∀ x, y ∈ V : (x|y) = (y|x),
• ∀ x, y, z ∈ V, λ, µ ∈ R : (x|λy + µz) = λ (x|z) + µ (y|z),
• ∀x ∈ V : (x|x) ≥ 0 and (x|x) = 0 iff x = 0.
Definition 2.5 will be our default. In the context of smooth manifolds it is referred to as a Riemannian inner
product or, especially in physics, as a ‘Riemannian metric’. Occasionally we may encounter an inner product on
a complex vector space. Below, z = a − bi ∈ C denotes the complex conjugate of z = a + bi ∈ C.
12 Tensor Calculus
Definition 2.6 An inner product on a complex vector space V is a nondegenerate positive hermitean sesquilinear
mapping of the form ( · | · ) : V × V → R which satisfies
• ∀ x, y ∈ V : (x|y) = (y|x),
• ∀ x, y, z ∈ V, λ, µ ∈ R : (x|λy + µz) = λ (x|y) + µ (x|z),
• ∀x ∈ V : (x|x) ≥ 0 and (x|x) = 0 iff x = 0.
Caveat
• Note that for the complex case (λx + µy|z) = λ (x|z) + µ (y|z), whence the attribute ‘sesquilinear’.
• The term “inner product” is used in different contexts to denote bilinear or sesquilinear forms satisfying
different sets of axioms. Examples include the Lorentzian inner product in the context of special relativity
(v.i.) and the symplectic inner product in the Hamiltonian formalism of classical mechanics.
• Unlike with the Kronecker tensor, the existence of an inner product is never implied, but must always be
stipulated explicitly.
Definition 2.7 A pseudo-Riemannian inner product on a real vector space V is a nondegenerate symmetric bi-
linear mapping of the form ( · | · ) : V × V → R which satisfies
• ∀ x, y ∈ V : (x|y) = (y|x),
• ∀ x, y, z ∈ V, λ, µ ∈ R : (x|λy + µz) = λ (x|y) + µ (x|z),
• ∀ x ∈ V with x 6= 0 ∃ y ∈ V such that (x|y) 6= 0.
A Lorentzian inner product is a pseudo-Riemannian inner product with signature (+, −, . . . , −) or (−, +, . . . , +).
Note that Definition 2.5 is a special case of Definition 2.7, viz. with signature (+, . . . , +). The signature of a
Lorentzian inner product contains exactly one + or one − sign. The notion of signature is explained below in
Definition 2.8.
Terminology
A vector space V endowed with an inner product ( · | · ) is also called an inner product space.
Ansatz
An n-dimensional inner product space V over R furnished with a basis {ei }.
Definition 2.8 The Gram matrix G is the matrix with components gij = (ei |ej ). Its signature (s1 , . . . , sn ) is
defined as the (basis independent) n-tuple of signs si = ±, i = 1, . . . , n, of the (real) eigenvalues of the Gram
matrix for a given ordering of its eigenvectors.
Remark
Clearly, G is symmetric. Another consequence of the definition is that G is invertible in the pseudo-Riemannian
case. For suppose G would be singular, then there exist x = xi ei ∈ V , x 6= 0, and a 0-eigenvector y = y i ei ∈ V ,
y 6= 0, such that gij y j = 0, whence 0 = xi gij y j = xi (ei |ej ) y j = xi ei |y j ej = (x|y). This contradicts the
non-degeneracy axiom of Definition 2.7.
2.4 Inner Products 13
Terminology
The components of the inverse Gram matrix G−1 are denoted by g ij , i.e.
g ik gkj = δji .
An inner product, if it exists, can be related to the Kronecker tensor, as we will see below.
Ansatz
An n-dimensional inner product space V over R.
Theorem 2.3 There exists a bijective linear map G : V → V ∗ , with inverse G−1 : V ∗ → V , such that
• ∀ x, y ∈ V : (x|y) = hG(x), yi,
• ∀ x̂ ∈ V ∗ , y ∈ V : G−1 (x̂)|y = hx̂, yi.

The matrix representations of G and G−1 are given by the Gram matrix G and its inverse G−1 , respectively,
recall Definition 2.8 on page 12.
• For fixed x ∈ V define x̂ ∈ V ∗ as the linear map x̂ : V → R : y 7→ (x|y). Subsequently define the linear
map G : V → V ∗ : x 7→ G(x) ≡ x̂.
• For the second part we need to prove invertibility of G. Suppose x ∈ V, x 6= 0, is such that G(x) = 0.
Then 0 = hG(x), yi = (x|y) for all y ∈ V , contradicting the nondegeneracy property of the inner product.
Hence G is one-to-one1 . Since dim V = n < ∞ it is also bijective, hence invertible.
• To compute the matrix representation G of G, decompose x = xi ei , y = y i ei , and G(x) = G(xi ei ) =

xi Gij êj , with matrix components defined via G(ei ) = Gij êj .
– By construction (first item above) and by virtue of bilinearity of the inner product and the definition
of the Gram matrix, we have hG(x), yi = (x|y) = xi y j (ei |ej ) = gij xi y j .
– By definition of duality, we have hG(x), yi = Gij xi y k hêj , ek i = Gij xi y k δkj = Gij xi y j .
Since xi , y j ∈ R are arbitrary it follows that Gij = gij , recall Definition 2.8.
• Consequently, the inverse G−1 : V ∗ → V has a matrix representation G−1 corresponding to the inverse
Gram matrix: (G−1 )ij = g ij .
Remark
• Note the synonymous formulations: G(x) ≡ x̂ ≡ hG(x), · i ≡ (x| · ). Recall the comment on notation,
page 11. Definition 2.9 below introduces yet another synonymous form.
• The Riesz representation theorem, which states that for each x̂ ∈ V ∗ there exists one and only one x ∈ V
such that hx̂, yi = (x|y) for all y ∈ V , is an immediate consequence of Theorem 2.3, viz. x = G−1 (x̂).
1 The terms “one-to-one” and “injective” are synonymous. A function f : D → C is injective if it associates distinct arguments with
distinct values. If the codomain C of the function is replaced by its actual range f (D), then it is also surjective, and hence bijective.
14 Tensor Calculus
Notation
• We reserve boldface symbols for intrinsic quantities such as vectors, covectors, and (other) linear mappings,
in order to distinguish them from their respective components relative to a fiducial basis.
Definition 2.9 The conversion operators ] : V → V ∗ and [ : V ∗ → V are introduced as shorthand for G,
respectively G−1 , viz. if v ∈ V , v̂ ∈ V ∗ , then
def def
]v = G(v) ∈ V ∗ respectively [v̂ = G−1 (v̂) ∈ V .
Terminology
G is commonly known as the metric, and G−1 as the dual metric. The operators ] : V → V ∗ and [ : V ∗ → V
are pronounced as sharp and flat, respectively.
Caveat
In the literature, the roles of sharp and flat are often reversed! Make sure to adhere to an unambiguous convention.
Result 2.2 Let v = v i ei ∈ V and v̂ = vi êi ∈ V ∗ , then
• ]v = gij v i êj = vj êj = v̂ ∈ V ∗ , with vj = gij v i ;

• [v̂ = g ij vi ej = v j ej = v ∈ V , with v j = g ij vi .
Proof of Result 2.2. This follows directly from the definition of ] and [ using
• ]ei = G(ei ) = gij êj ∈ V ∗ , respectively

• [êi = G−1 (êi ) = g ij ej ∈ V .
Notation
def
Unless stated otherwise we will adhere to the following convention. If v i ∈ R, then vj = gij v i . reversely, if
def
vi ∈ R, then v j = g ij vi .
Caveat
In the foregoing G ∈ L (V, V ∗ ) and G−1 ∈ L (V ∗ , V ), with prototypes G : V → V ∗ , resp. G−1 : V ∗ → V .
However, one often interprets the metric and its dual as bilinear mappings of type G : V × V → R, resp.
G−1 : V ∗ × V ∗ → R. It should be clear from the context which prototype is implied. We return to this
ambiguity in section 2.8.
Reminder
In an inner product space vectors can be seen as “covectors in disguise”, vice versa.
2.5. Reciprocal Bases
Ansatz
2.5 Reciprocal Bases 15

Definition 2.10 The basis {ei } of V induces a new basis [êi of V given by [êi = g ij ej , the so-called recip-
rocal basis.
Definition 2.11 The basis êi of V ∗ induces a new basis {]ei } of V ∗ given by ]ei = gij êj , the so-called

reciprocal dual basis.
Caveat
• Instead of [êi one sometimes writes ei for the reciprocal of ei .

• Likewise, instead of ]ei one sometimes writes êi for the reciprocal of êi .
• Since we conventionally label elements of V by lower indices, and elements of V ∗ by upper indices, we
will adhere to the notation employing the [ and ] operator.
Remark
• The reciprocal basis is uniquely determined by the inner product.

• The Gram matrix of the reciprocal basis is G−1 , with components [êi |[êj

= g ik g j` (ek |e` ) =
g ik g j` gk` = g ij .

• We have [êi |ej = g ik (ek |ej ) = g ik gkj = δji .
Terminology
The vectors [êi and ej are said to be perpendicular for i 6= j.
Remark

For each reciprocal basis [êi of V and each x ∈ V there exists a unique collection of real numbers {xi } such
that x = xi [êi .
Terminology
The real numbers xi are referred to as the covariant components of the vector x relative to the basis {ei }.
Reminder
• Reciprocal basis vectors are labeled with upper indices.

• Covariant components are labeled with lower indices.
• No confusion is likely to arise if we adhere to the notation using the [ and ] converters.
16 Tensor Calculus
2.6. Bases, Dual Bases, Reciprocal Bases: Mutual Relations
Heuristics
Dual and reciprocal bases are close-knit. The exact relationship is clarified in this section.
Sofar, a vector space V and its associated dual space V ∗ have been introduced as a priori independent entities. V
and V ∗ are mutually connected by virtue of the “interaction” between their respective elements, recall Result 2.1
on page 9. One nevertheless retains a clear distinction between a vector (element of V ) and a dual vector (element
of V ∗ ) throughout. That is, there is no mechanism to convert a vector into a dual vector, vice versa.
An inner product provides us with an explicit mechanism to construct a one-to-one linear mapping (i.e. dual
vector) associated with each vector by virtue of the metric G, cf. Theorem 2.3 on page 13. An inner product also
enables us to construct a reciprocal basis that allows us to mimic duality within V proper. “Interchangeability”
of vectors and reciprocal vectors as well as that of vectors and dual vectors is governed by the Gram matrix G
and its inverse.
Ansatz
Observation
Result 2.3 on page 13 states that each vector x ∈ V induces a dual vector x̂ = G(x) ∈ V ∗ . Let x = xi ei
relative to basis {ei } of V . Then2 x̂ = x̂i êi relative to the corresponding dual basis {êi } of V ∗ . According to
Definition 2.10 on page 15 we also have x = xi [êi relative to the reciprocal basis {[êi } of V . Given xi ∈ R we
have xi = x̂i = gij xj .
Caveat
Notice the distinct prototypes of metric G : V → V ∗ and linear map G : V → V represented by the Gram
matrix G, i.e. ei = G ([êi ).
Notation
No confusion is likely to arise if we use identical symbols, x̂i ≡ xi , to denote both dual as well as reciprocal
vector components. The underlying basis ({êi }, respectively {[êi }), and thus the correct geometric interpretation,
should be inferred from the context. By default we shall interpret xi as dual vector components, henceforth
omitting the ˆ symbol, i.e. x̂ = xi êi ∈ V ∗ .
2 For the sake of argument we temporarily write x̂i for dual, and xi for reciprocal vector components.
2.7 Examples of Vectors and Covectors 17
Reminder
• A dual basis does not require an inner product, a reciprocal basis does.
• A dual basis spans V ∗ , a reciprocal basis spans V .

• In an inner product space—with metric G—the components of a vector x ∈ V relative to the reciprocal
basis are identical to the components of the dual vector x̂ = G(x) relative to the dual basis.
• Thus, in an inner product space with fixed metric G, one may loosely identify a vector x ∈ V and its dual
vector x̂ ∈ V ∗ to the extent that x̂ = G(x) and ( · |x) = hx̂, · i.
• The metric and its dual, G, respectively G−1 , provide the natural mechanism to “toggle” between elements
of V and V ∗ . Abbreviations: G(x) = ]x, G−1 (x̂) = [x̂.
• The Gram matrix and its inverse, G, respectively G−1 , or rather the associated linear mappings G : V → V
and G −1 : V → V , provide natural mechanisms to “toggle” between elements of V and their reciprocals
in V .
2.7. Examples of Vectors and Covectors
It is important to learn to recognize covectors and vectors in practice, and to maintain a clear distinction through-
out. Some important examples from mechanics may illustrate this.
Example
• Particle velocity v provides an example of a vector.

• Particle momentum p̂ provides an example of a covector.
• The metric tensor allows us to relate these two: p̂ = G(v) = ]v, or equivalently, v = G−1 (p̂) = [p̂.
Example
• Displacement x relative to a fiducial point in a continuous medium (or in vacuo) provides an example of a
vector.
• Spatial frequency k̂ provides an example of a covector.
• The Kronecker tensor allows us to construct a scalar from these two: hk̂, xi ∈ R (“phase”).
Caveat
• In practice one often refrains from a notational distinction between covectors and vectors, writing e.g. p
and k instead of p̂ and k̂, as we do here.
• By the same token it is a common malpractice to speak of “the momentum vector p” and the “wave vector
k” even if covectors are intended. (Of course, technically speaking covectors are vectors.)
18 Tensor Calculus
2.8. Tensors
2.8.1. Tensors in all Generality
Ansatz
• An n-dimensional vector space V over R.

• A basis {ei } of V .
Definition 2.12 A tensor is a multilinear mapping of type
T : V ∗ × ... × V ∗ ×V × ... × V → R
| {z } | {z }
p q
for some p, q ∈ N0 .
Notation
Recall Definition 2.12. The following notation will be used to indicate tensor prototypes:
def
T ∈ V ⊗ . . . ⊗ V ⊗ V ∗ ⊗ . . . ⊗ V ∗ = Tpq (V ) .
| {z } | {z }
p q
Caveat
• The ordering of arguments matters!

• Thus the prototype may be specified with a different arrangement of linear spaces V and V ∗ .
Terminology

p
• We refer to T ∈ Tpq (V ) sometimes as a q
-tensor.
• T ∈ Tpq (V ) is referred to as a mixed tensor of contravariant rank p and covariant rank q.
• If q 6= 0, T ∈ T0q (V ) is called a covariant tensor of rank q.

• If p 6= 0, T ∈ Tp0 (V ) is called a contravariant tensor of rank p.
• T ∈ T00 (V ) is called a scalar. We identify T with its value in this case, i.e. T ∈ R.
• The words cotensor/contratensor are synonymous to covariant/contravariant tensor.
Example

1
Let T be a 2
-tensor, and x, x1 , x2 , y, y1 , y2 ∈ V , ẑ, ẑ1 , ẑ2 ∈ V ∗ , and λ1 , λ2 ∈ R arbitrary. Then we have
• T(λ1 ẑ1 + λ2 ẑ2 , x, y) = λ1 T(ẑ1 , x, y) + λ2 T(ẑ2 , x, y);

• T(ẑ, λ1 x1 + λ2 x2 , y) = λ1 T(ẑ, x1 , y) + λ2 T(ẑ, x2 , y);
• T(ẑ, x, λ1 y1 + λ2 y2 ) = λ1 T(ẑ, x, y1 ) + λ2 T(ẑ, x, y2 ).
2.8 Tensors 19
Result 2.3 Special cases of low rank tensors:

0
• A 0
-tensor is a scalar, i.e. T00 (V ) = R.

0
• A 1
-tensor is a covector, i.e. T01 (V ) = V ∗ .

1
• A 0
-tensor is a vector, i.e. T10 (V ) = V .
Proof of Result 2.3. The first case is a convention previously explained. The second case is true by Definition 2.2,
page 8, recall the terminology introduced on page 9. The third case is true by virtue of the isomorphism V ∗∗ = V ,
recall the remark on page 11.
Definition 2.13 The outer product f ⊗ g : X × Y → R of two R-valued functions f : X → R and g : Y → R

is defined as follows:
(f ⊗ g)(x, y) = f (x)g(y) for all x ∈ X, y ∈ Y .
Since both vectors as well as covectors can be interpreted as R-valued mappings, the following definition is a
natural one.

p
Definition 2.14 For each (p + q)-tuple of indices (i1 , . . . , ip , j1 , . . . , jq ) the mixed q
-tensor ei1 ⊗ . . . ⊗ eip ⊗
j1 jq
ê ⊗ . . . ⊗ ê ∈ Tpq (V ) is defined by

ei1 ⊗. . .⊗ eip ⊗ êj1 ⊗. . .⊗ êjq (x̂1 ,. . ., x̂p , y1 ,. . ., yq ) = hx̂1 , ei1 i. . .hx̂p , eip ihêj1 , y1 i. . .hêjq , yq i ,
for all x̂1 ,. . ., x̂p ∈ V ∗ and y1 ,. . ., yq ∈ V .
Caveat
• The symbol ⊗ used in Definition 2.14 is related to, but should not be confused with the one used in
Definition 2.12. Mnemonic: ei1 ⊗. . .⊗ eip ⊗ êj1 ⊗. . .⊗ êjq ∈ V ⊗ . . . ⊗ V ⊗ V ∗ ⊗ . . . ⊗ V ∗ .
| {z } | {z }
p q
i ...i
Result 2.4 If T ∈ Tpq (V ) there exists a set of numbers tj11 ...jpq ∈ R such that
i ...i
T = tj11 ...jpq ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq ,
i ...i
viz. tj11 ...jpq = T(êi1 , . . . , êip , ej1 , . . . , ejq ).
Proof of Result 2.4. We have
T êi1 , . . . , êip , ej1 , . . . , ejq

k ...k
= t`11...`qp ek1 ⊗ . . . ⊗ ekp ⊗ ê`1 ⊗ . . . ⊗ ê`q êi1 , . . . , êip , ej1 , . . . , ejq

k ...k
= t`11...`qp hêi1 , ek1 i. . .hêip , ekp ihê`1 , ej1 i. . .hê`q , ejq i
k ...k i ` i ...i
= t`11...`qp δki11 . . . δkpp δj`11 . . . δjqq = tj11 ...jpq .
20 Tensor Calculus
So, for general covector and vector arguments ω̂ 1 , . . . , ω̂ p ∈ V ∗ , v1 , . . . , vq ∈ V ,
i ...i
T (ω̂ 1 , . . . , ω̂ p , v1 , . . . , vq ) = tj11 ...jpq ω1,i1 . . . ωp,ip v1j1 . . . vqjq .
Terminology
• If T ∈ T01 (V ) = V ∗ there exist numbers ti ∈ R such that T = ti êi , viz. ti = T(ei ). These are the
covariant components of T relative to basis {ei }.
• If T ∈ T10 (V ) = V there exist numbers ti ∈ R such that T = ti ei , viz. ti = T(êi ). These are the
contravariant components of T relative to basis {ei }.
i ...i
• If T ∈ Tpq (V ), the numbers tj11 ...jpq ∈ R are referred to as the mixed components of T relative to basis
{ei }.
Caveat
i ...i
Laziness has led to the habit of saying that T = tj11 ...jpq ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq is expressed “relative to
basis {ei }”, where it would have been more accurate to refer to the actual basis {ei1 ⊗. . .⊗eip ⊗ êj1 ⊗. . .⊗ êjq }.
Example
Recall Definition 2.4 on page 10. We have yet another notation for the Kronecker tensor h · , · i ∈ V ⊗ V ∗ , viz.
h · , · i = ei ⊗ êi = δji ei ⊗ êj ,
in which we recognize the Kronecker symbol as the mixed components of the Kronecker tensor relative to basis
{ei }. To see this, insert an arbitrary covector ŷ = yi êi ∈ V ∗ and vector x = xi ei ∈ V on the r.h.s. The result is
ei ⊗ êi (ŷ, x) = hŷ, ei ihêi , xi = yj xk hêj , ei ihêi , ek i = yj xk δij δki = yi xj δji = yi xj hêi , ej i = hŷ, xi .

Note that the Kronecker tensor is parameter free. For this reason it is sometimes called a constant tensor. Since
its coefficients δji do not depend on the chosen basis (an untypical property for tensor coefficients), they are said
to constitute an invariant holor.
Observation
• Tpq (V ) constitutes a linear space subject to the usual rules for linear superposition of mappings, i.e. for all
S, T ∈ Tpq (V ) and all λ, µ ∈ R we declare
(λS + µT) ((co)vectors) = λS((co)vectors) + µT((co)vectors) ,
in which “(co)vectors” generically denotes an appropriate sequence of arbitrary vector and/or covector
arguments.
• We have dim Tpq (V ) = np+q .
i ...i
• Lexicographical ordering of the components tj11 ...jpq of tensors T ∈ Tpq (V ) establishes an isomorphism
p+q
with the “column vector space” Rn .
2.8 Tensors 21
Terminology
i ...i
• The indexed collection of tensor coefficients tj11 ...jpq is also referred to as a holor.
• Physicists and engineers are wont to refer to holors as “tensors”. In a mathematician’s vocabulary the
(basis dependent) holor is strictly to be distinguished from the associated (basis independent) tensor,
i ...i
T = tj11 ...jpq ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq .
• If {ei } is a basis of V , a basis of Tpq (V ) is given by {ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq }. But do recall the
caveat on page 20 concerning terminology.
Caveat
Parlance and notation in tensor calculus tend to be somewhat sloppy. This is the price one must pay for simplicity.
It is up to the reader to resolve ambiguities, if any, and provide correct mathematical interpretations.
Example
It is customary to think of the metric G ∈ L (V, V ∗ ) as a covariant tensor of rank 2. In order to understand this,
let us clarify a previous remark on page 14. For given metric G, define g ∈ V ∗ ⊗ V ∗ = T02 (V ) as follows:
g : V × V → R : (v, w) 7→ g(v, w) ≡ (G(v)) (w) ≡ (v|w) ≡ h]v, wi ,
recall the sharp operator, introduced on page 14.
Notation
• Except when this might lead to confusion, no notational distinction will henceforth be made between such
distinct prototypes as g and G. We will invariably write G, whether it is intended as a covariant 2-tensor (g
in the example, implied when we write G(v, w)), or as a vector-to-covector converter (]-operator, implied
when we write G(v)).
• The ]- and [-operators provide convenient shorthands for the (dual) metric if one wishes to stress its action
as a vector-to-covector (respectively covector-to-vector) converter.
Terminology
Irrespective of the intended interpretation, G will be referred to as the metric tensor.
Theorem 2.4 If S ∈ Tpq (V ), T ∈ Trs (V ), with

i ...i
S = sj11 ...jpq ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq and T = tji11...i j1 js
...js ei1 ⊗ . . . ⊗ eir ⊗ ê ⊗ . . . ⊗ ê ,
r
i ...i i ...i
then S ⊗ T ∈ Tp+r 1 p p+1 p+r j1
q+s (V ), with S ⊗ T = sj1 ...jq tjq+1 ...jq+s ei1 ⊗ . . . ⊗ eip+r ⊗ ê ⊗ . . . ⊗ ê
jq+s
.
Terminology
p+r
For S ∈ Tpq (V ), T ∈ Trs (V ), the tensor S ⊗ T ∈ Tq+s (V ) is also referred to as the outer product or tensor
product of S and T.
Caveat
Notice that we have chosen to adhere to a default ordering of basis and dual basis vectors that span the tensor
product space, recall Definition 2.12, page 18. This boils down to a rearrangement such that the first p + r
arguments of the product tensor S ⊗ T are covectors, and the remaining q + s arguments are vectors.
Proof of Theorem 2.4. Apply Definition 2.13, page 19, accounting for the occurrence of multiplicative coeffi-
cients: ((λf ) ⊗ (µg))(x, y) = ((λf )(x))((µg)(y)) = λµf (x)g(y). Further details are left to the reader (taking
22 Tensor Calculus
into account the previous caveat).
Caveat
The tensor product is associative, distributive with respect to addition and scalar multiplication, but not commu-
tative. (Work out the exact meaning of these statements in formulae.)
Reminder
• A mixed tensor T ∈ Tpq (V ) has contravariant rank p and covariant rank q.

i ...i
• The holor tj11 ...jpq ∈ R of T ∈ Tpq (V ) is labeled with p upper and q lower indices.
• Elements of the underlying basis {ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq } of Tpq (V ), on the other hand, are
labeled with p lower and q upper indices.
• A holor depends on the chosen basis, the tensor itself (by definition!) does not. (We return to this aspect in
Section 2.8.6.)
• The outer product of two tensors yields another tensor. Ranks add up.
2.8.2. Tensors Subject to Symmetries
Ansatz
• Tensor spaces of type Tpq (V ).
Certain subspaces of Tpq (V ) are of special interest. For instance, it may be the case that the ordering of (co)vector
arguments in a tensor evaluation does not affect the outcome.
Remark
Recall a previous remark on page 8 concerning the convention for index positioning. In particular, in Defini-
tions 2.15 and 2.16 below, indices are used to label an a priori unconstrained number of arguments, whence the
upper/lower index convention is not applicable.
Notation
• π = {π(α1 ), . . . , π(αp )} denotes a permutation of the p ordered symbols {α1 , . . . , αp }.

• π(αi ), i = 1, . . . , p, is shorthand notation for the i-th entry of π.
Definition 2.15 Let π = {π(1), . . . , π(p)} be any permutation of labels {1, . . . , p}, then T ∈ T0p (V ) is called a
symmetric covariant tensor if for all v1 , . . . , vp ∈ V
T(vπ(1) , . . . , vπ(p) ) = T(v1 , . . . , vp ) .
Likewise, T ∈ Tp0 (V ) is called a symmetric contravariant tensor if for all v̂1 , . . . , v̂p ∈ V ∗
T(v̂π(1) , . . . , v̂π(p) ) = T(v̂1 , . . . , v̂p ) .

2.8 Tensors 23
Notation
) ⊂ T0p (V ).
W
• The vector space of symmetric covariant p-tensors is denoted p (V
Wp
• The vector space of symmetric contravariant p-tensors is denoted (V ) ⊂ Tp0 (V ).
Wp
• Alternatively one sometimes writes p (V ) = V ∗ ⊗S . . . ⊗S V ∗ and
W
(V ) = V ⊗S . . . ⊗S V .
| {z } | {z }
p p
Example
An example of a symmetric covariant p-tensor on Rn is given by the set of p-th order partial derivatives of a
sufficiently smooth function f ∈ C p (Rn , R) at a fiducial point:
∂ p f (0)
Dp f (0) = êi1 ⊗ . . . ⊗ êip .
∂xi1 . . . ∂xip
Notation
• For simplicity we write fi1 ...ip (0) for the partial derivatives in the above expression.
• The fiducial point (origin in this example) is often suppressed in the notation.
Caveat
The derivative tensor field Dp f : Rn → T0p (V ) : x 7→ Dp f (x)—as opposed to the isolated tensor Dp f (0) ∈
T0p (V )—fails to be a tensor field unless one replaces partial derivatives by covariant derivatives. (We return to
this later.)
It may also happen that an interchange of two arguments causes the outcome to change sign.
Definition 2.16 Let sgn (π) denote the sign of the permutation π, i.e. sgn (π) = 1 if it is even, sgn (π) = −1 if
it is odd. T ∈ T0p (V ) is called an antisymmetric covariant tensor if for all v1 , . . . , vp ∈ V and any permutation
π = {π(1), . . . , π(p)} of labels {1, . . . , p},
T(vπ(1) , . . . , vπ(p) ) = sgn (π)T(v1 , . . . , vp ) .
Likewise, T ∈ Tp0 (V ) is called an antisymmetric contravariant tensor if for all v̂1 , . . . , v̂p ∈ V ∗
T(v̂π(1) , . . . , v̂π(p) ) = sgn (π)T(v̂1 , . . . , v̂p ) .
Notation
) ⊂ T0p (V ).
V
• The vector space of antisymmetric covariant p-tensors is denoted p (V
Vp
• The vector space of antisymmetric contravariant p-tensors is denoted (V ) ⊂ Tp0 (V ).
Vp
• Alternatively one sometimes writes p (V ) = V ∗ ⊗A . . . ⊗A V ∗ and
V
(V ) = V ⊗A . . . ⊗A V .
| {z } | {z }
q p
It is convenient to admit tensor ranks p < 2 in this notation by a natural extension.

24 Tensor Calculus
Assumption
W W0 V V0
• 0 (V
)= (V ) = 0 (V ) = (V ) = R. (This is consistent with Definitions 2.15 and 2.16 if we set
π = id, sgn π = 1 on the empty index set ∅).
• 1 (V ) = 1 (V ) = V ∗ . (This follows from Definitions 2.15 and 2.16, with π = id, sgn π = 1 on {1}).
W V
W1 V1
• (V ) = (V ) = V . (This follows from Definitions 2.15 and 2.16, with π = id, sgn π = 1 on {1}).
2.8.3. Symmetry and Antisymmetry Preserving Product Operators
Ansatz

W V Wp Vp
• Tensor spaces of type p (V ), p (V ), (V ), (V ).
The outer product, recall Definition 2.13 on page 19 and in particular Theorem 2.4 on page 21, does not preserve
W
(anti)symmetry.
V For this reason alternative product operators are introduced specifically designed for p (V ),
p (V ), and their contravariant counterparts.
Definition 2.17 For every v̂1 , . . . , v̂k ∈ V ∗ we define the antisymmetric covariant k-tensor v̂1 ∧ . . . ∧ v̂k ∈
V
k (V ) as follows:
v̂1 ∧ . . . ∧ v̂k (x1 , . . . , xk ) = dethv̂i , xj i for all xj ∈ V .

Example
Let v̂, ŵ ∈ V ∗ , x, y ∈ V , then

hv̂, xi hv̂, yi
(v̂ ∧ ŵ) (x, y) = det = hv̂, xihŵ, yi − hŵ, xihv̂, yi .
hŵ, xi hŵ, yi
Alternatively,
v̂ ∧ ŵ = v̂ ⊗ ŵ − ŵ ⊗ v̂ .
A more general result will be given on page 28.
Terminology
The infix operator ∧ is referred to as the wedge product or the antisymmetric product.
Observation
• ê1 ∧ . . . ∧ êk (e1 , . . . , ek ) = dethêi , ej i = 1.

• êi1 ∧ . . . ∧ êik (ej1 , . . . , ejk ) = δji11 ...i

k
...jk , recall the completely antisymmetric symbols on page 5, cf.
• êi1 ⊗ . . . ⊗ êik (ej1 , . . . , ejk ) = δji11 . . . δjikk .

2.8 Tensors 25
Observation
Furthermore, the wedge product is associative, bilinear and antisymmetric (whence nilpotent, vice versa), i.e. if
û, v̂, ŵ ∈ V ∗ , λ, µ ∈ R, then
• (û ∧ v̂) ∧ ŵ = û ∧ (v̂ ∧ ŵ).

• (λû + µv̂) ∧ ŵ = λû ∧ ŵ + µv̂ ∧ ŵ.
• û ∧ (λv̂ + µŵ) = λû ∧ v̂ + µû ∧ ŵ.
• û ∧ û = 0, or, equivalently, û ∧ v̂ = −v̂ ∧ û.
A similar symmetry preserving product exists for symmetric tensors. To this end we need to introduce the
symmetric counterpart of a determinant.
Definition 2.18 The operator perm assigns a number to a square matrix in a way similar to det, but with a +
sign for each term, recall page 4:
n n
X 1 X
perm A = |[j1 , . . . , jn ]| A1j1 . . . Anjn = |[i1 , . . . , in ] [j1 , . . . , jn ]| Ai1 j1 . . . Ain jn .
j1 ,...,jn =1
n!
i1 , . . . , in = 1
j1 , . . . , jn = 1
Remark
The formal role of the completely antisymmetric symbols in the above expressions is to restrict summation to
relevant terms. The absolute value discards minus signs for odd parity terms, recall the last line in the observation
made on page 4.
Terminology
The operator perm is called the permanent.
Example
Analogous to Table 1.1 on page 4 we have for a 3 × 3 matrix A,
perm A = A11 A22 A33 +A12 A23 A31 +A13 A21 A32 +A13 A22 A31 +A11 A23 A32 +A12 A21 A33 .
Definition
W 2.19 For every k-tuple v̂1 , . . . , v̂k ∈ V ∗ we define the symmetric covariant k-tensor v̂1 ∨ . . . ∨ v̂k ∈
k (V ) as follows:
v̂1 ∨ . . . ∨ v̂k (x1 , . . . , xk ) = perm hv̂i , xj i .

Example
Let v̂, ŵ ∈ V ∗ , x, y ∈ V , then

hv̂, xi hv̂, yi
(v̂ ∨ ŵ) (x, y) = perm = hv̂, xihŵ, yi + hŵ, xihv̂, yi .
hŵ, xi hŵ, yi
Alternatively,
v̂ ∨ ŵ = v̂ ⊗ ŵ + ŵ ⊗ v̂ .
Terminology
The infix operator ∨ is referred to as the vee product or the symmetric product.
26 Tensor Calculus
Observation
The vee product is associative, bilinear and symmetric, i.e. if û, v̂, ŵ ∈ V ∗ , λ, µ ∈ R, then
• (û ∨ v̂) ∨ ŵ = û ∨ (v̂ ∨ ŵ).
• (λû + µv̂) ∨ ŵ = λû ∨ ŵ + µv̂ ∨ ŵ.

• û ∨ (λv̂ + µŵ) = λû ∨ v̂ + µû ∨ ŵ.
• û ∨ v̂ = v̂ ∨ û.
Observation
• dim T0p (V ) = np .
• {êi1 ∧ . . . ∧ êip | 1 ≤ i1 < i2 < . . . < ip−1 < ip ≤ n} is a basis of

V
p (V ).
• {êi1 ∨ . . . ∨ êip | 1 ≤ i1 ≤ i2 ≤ . . . ≤ ip−1 ≤ ip ≤ n} is a basis of p (V ).

W
• For p ≥ 2, dim p (V ) + dim p (V ) ≤ dim T0p (V ), with strict equality iff p = 2.

V W
Caveat
It is customary to suppress the infix operator ∨ in the notation.
Theorem 2.5 Cf. the observations above. We have

V n W n+p−1
dim p (V ) = and dim p (V )= .
p p
• As for the first identity, a strictly increasing ordering of labels in a product of the form êi1 ∧ . . . ∧ êip can
be realized in exactly as many distinct ways as one can select p items out of n different objects.
V V
• In particular, if p > n, then dim p (V ) = 0, i.e. p (V ) = {0}.
• As for the second identity, the number of ways in which a non-decreasing ordering of labels in a product
of the form êi1 ∨ . . . ∨ êip can be realized can be inferred from the following argument3 . Any admissible
ordering of labels can be represented symbolically by a sequence of p dots • , split into subsequences of
equal labels separated by n − 1 crosses × , each of which marks a unit label transition. Examples, taking
n = 5, p = 7:
••×•××•••×• ≡ ê1 ∨ ê1 ∨ ê2 ∨ ê4 ∨ ê4 ∨ ê4 ∨ ê5
••×••••×•×× ≡ ê1 ∨ ê1 ∨ ê2 ∨ ê2 ∨ ê2 ∨ ê2 ∨ ê3
There are exactly (n + p − 1)! permutations of the n + p − 1 symbols in such sequences, with exactly p!,
respectively (n − 1)! indistinguishable permutations among the identical symbols • , respectively × , in any
given realization.
Remark
V W
V between p (V ) and p (V ) with respect to dimensionality scaling with p. In
Notice the qualitative differences
the former case we have dim p (V ) = 0 as soon as p > n = dim V , i.e. no nontrivial antisymmetric tensors
with ranks p exceeding vector space dimension n exist. In contrast, nontrivial symmetric tensors do exist for any
rank.
3 Courtesy of Freek Paans.
2.8 Tensors 27
Example
Vp
Let B p and Bp denote bases for ), in Euclidean 3-space V = R3 .
V
(V ), respectively p (V
p Bp dim p Bp dim
0 {1} 1 0 {1}
1 2 3 1
1 {e1 , e2 , e3 } 3 1 1 ê , ê , ê 3
2 {e1 ∧ e2 , e1 ∧ e3 , e2 ∧ e3 } 3 2 ê ∧ê2 , ê1 ∧ ê3 , ê2 ∧ ê3 3
3 {e1 ∧ e2 ∧ e3 } 1 3 ê1 ∧ ê2 ∧ ê3 1
Observation
Definitions 2.17, page 24, and 2.19, page 25, are straightforwardly complemented by similar ones applicable to
(anti)symmetric contravariant tensors.
Definition 2.20 The linear operator S : T0p (V ) →

W
p (V ), given by
1 X
(S (T)) (v1 , . . . , vp ) = T(vπ(1) , . . . , vπ(p) ) ,
p! π
in which summation runs over all permutations π of the index set {1, . . . , p}, is called the symmetrisation map.
Observation
• If T = ti1 ...ip êi1 ⊗ . . . ⊗ êip ∈ T0p (V ), then S (T) = t(i1 ...ip ) êi1 ⊗ . . . ⊗ êip ∈
W
p (V ), in which
def 1 X
t(i1 ...ip ) = tπ(i1 )...π(ip ) .
p! π
) is symmetric, then T = S (T).

W
• If T = p (V
• The symmetrisation map is a projection, i.e. it is idempotent: S ◦ S = S .
1 1 X
These observations explain the normalization factor , since 1 = 1.
p! p! π
Observation
• The ∨-product is particularly convenient in the context of the symmetrisation operator, for we have
1 i1
S (êi1 ⊗ . . . ⊗ êip ) = ê ∨ . . . ∨ êip .
p!
• T ∨ U = U ∨ T.
W
• The case p = 0, q > 0 is covered by defining λ ∨ U = λU for all λ ∈ R, U ∈ q (V ).
• Rule of thumb: The action of p!S on a ⊗-product amounts to a formal replacement ⊗ → ⊗S ≡ ∨.
The following result generalizes the example given on page 25:
(p + q)!
S (T ⊗ U).
W W
Result 2.5 Let T ∈ p (V ) and U ∈ q (V ), then T ∨ U =
p!q!
28 Tensor Calculus
Definition 2.21 The linear operator A : T0p (V ) →

V
p (V ), given by
1 X
(A (T)) (v1 , . . . , vp ) = sgn (π)T(vπ(1) , . . . , vπ(p) ) ,
p! π
is referred to as the antisymmetrisation map.
Observation
• If T = ti1 ...ip êi1 ⊗ . . . ⊗ êip ∈ T0p (V ), then A (T) = t[i1 ...ip ] êi1 ⊗ . . . ⊗ êip ∈
V
p (V ), in which
def 1 X
t[i1 ...ip ] = sgn (π)tπ(i1 )...π(ip ) .
p! π
) is antisymmetric, then T = A (T).

V
• If T = p (V
• The antisymmetrisation map is a projection, i.e. it is idempotent: A ◦ A = A .
Recall a previous remark on the normalization factor (page 27).
Observation
• The ∧-product is particularly convenient in the context of antisymmetrisation, for we have

1 i1
A (êi1 ⊗ . . . ⊗ êip ) = ê ∧ . . . ∧ êip .
p!
V
• The case p = 0, q > 0 is covered by defining λ ∧ U = λU for all λ ∈ R, U ∈ q (V ).
• Rule of thumb: The action of p!A on a ⊗-product amounts to a formal replacement ⊗ → ⊗A ≡ ∧.
The following result generalizes the example given on page 24.
(p + q)!
A (T ⊗ U).
V V
Result 2.6 Let T ∈ p (V ) and U ∈ q (V ), then T ∧ U =
p!q!
Notation
• The Einstein summation convention,

V recall page 7, is usually adapted in the context of antisymmetric
tensors. In general, if T ∈ p (V ), then we may use either of the following equivalent decompositions:
X
– T = ti1 ...ip êi1 ⊗ . . . ⊗ êip = ti1 ...ip êi1 ⊗ . . . ⊗ êip ,
i1 ,...,ip
X
i1 ip
– T = t|i1 ...ip | ê ∧ . . . ∧ ê = ti1 ...ip êi1 ∧ . . . ∧ êip .
i1 <...<ip
Observation
1
Note that t|i1 ...ip | êi1 ∧ . . . ∧ êip = p! ti1 ...ip ê
i1
∧ . . . ∧ êip , recall the previous observation.
Proof of Result 2.6. Let T = t|i1 ...ip | êi1 ∧ . . . ∧ êip and U = u|ip+1 ...ip+q | êip+1 ∧ . . . ∧ êip+q , with antisymmetric
2.8 Tensors 29
holors ti1 ...ip and uip+1 ...ip+q . Then
1 ∗
T∧U = t|i1 ...ip | u|ip+1 ...ip+q | êi1 ∧ . . . ∧ êip+q =ti ...i ui ...i êi1 ∧ . . . ∧ êip+q
p!q! 1 p p+1 p+q
? (p + q)! (p + q)!
= ti1 ...ip uip+1 ...ip+q A êi1 ⊗ . . . ⊗ êip+q = A (T ⊗ U) .
p!q! p!q!
Notice that all p + q indices can be assumed to be distinct. In step ∗ we have used the fact that there are precisely
p! (respectively q!) orderings of p (respectively q) distinct indices. In step ? we have used the previous observation
on the relationship between wedge product and antisymmetrised outer product. The last step exploits linearity of
A.
), then T ∧ U = (−1)pq U ∧ T.
V V
Result 2.7 Let T ∈ p (V ) and U ∈ q (V
Proof of Result 2.7. Cf. the proof of Result 2.6. To change the positions of the p basis covectors êi1 , . . . , êip in
the product êi1 ∧ . . . ∧ êip+q so as to obtain the reordered product êip+1 ∧ . . . ∧ êip+q ∧ êi1 ∧ . . . ∧ êip requires
p (one for each êik , k = 1, . . . , p) times q (one for each êi` , ` = p + 1, . . . , p + q) commutations, each of which
brings in a minus sign.
Remark
V V
If p + q > n = dim V , then T ∧ U = 0 ∈ p+q (V ). Indeed, k (V ) = {0} for all k > n.
Remark
• Both S as well as A are idempotent: S ◦S = S , respectively A ◦A = A . Thus they are, by definition,

projections from T0p (V ) onto p (V ), respectively onto p (V ).
W V
• Note that we have not defined the property of (anti)symmetry for a genuine mixed tensor.
• Most definitions also apply to the covariant and contravariant parts of a mixed tensor separately, i.e. treating
the mixed tensor as a covariant tensor by freezing its covector arguments, respectively as a contravariant
tensor by freezing its vector arguments.
• A (say covariant) 2-tensor can always be decomposed into a sum of a symmetric and antisymmetric 2-
tensor, viz. T = S + A, with S = S (T) and A = A (T), i.e. S(v, w) = 21 (T(v, w) + T(w, v)) and
A(v, w) = 21 (T(v, w) − T(w, v)). This is consistent with the dimensionality equality for p = 2 in the
observation on page 26.
• A general tensor is typically neither symmetric nor antisymmetric, nor can it be split into a sum of a
symmetric and antisymmetric tensor, recall the dimensionality inequality in the observation on page 26.
• The metric tensor G is symmetric, and so is its inverse, the dual metric tensor G−1 .
• Both the ∧-product as well as the ∨-product obey the “usual” rules for associativity, and distributivity with
respect to tensor addition and scalar multiplication.
30 Tensor Calculus
Example
Let V = R3 , furnished with standard basis {e1 , e2 , e3 }. One conventionally denotes
∂ ∂ ∂
e1 = e2 = e3 = .
∂x ∂y ∂z
It is then customary to denote the corresponding dual basis {ê1 , ê2 , ê3 } of V ∗ by
ê1 = dx ê2 = dy ê3 = dz .
Let u = ui ei , v = v i ei , w = wi ei ∈ V be arbitrary vectors. Then we have
dx(v) = hdx, vi = v1 ,
dy(v) = hdy, vi = v2 ,
dz(v) = hdz, vi = v3 .
Example
Notation as in the previous example.
v1 w1

hdx, vi hdx, wi
(dx ∧ dy) (v, w) = det = det = v 1 w2 − v 2 w1 ,
hdy, vi hdy, wi v2 w2
v1 w1

hdx, vi hdx, wi
(dx ∧ dz) (v, w) = det = det = v 1 w3 − v 3 w1 ,
hdz, vi hdz, wi v3 w3
v2 w2

hdy, vi hdy, wi
(dy ∧ dz) (v, w) = det = det = v 2 w3 − v 3 w2 .
hdz, vi hdz, wi v3 w3
Example
Notation as in the previous example.
 1
v1 w1
  
hdx, ui hdx, vi hdx, wi u
(dx ∧ dy ∧ dz) (u, v, w) = det  hdy, ui hdy, vi hdy, wi  = det  u2 v2 w2 
hdz, ui hdz, vi hdz, wi u3 v3 w3
= u1 v 2 w3 + u3 v 1 w2 + u2 v 3 w1 − u3 v 2 w1 − u1 v 3 w2 − u2 v 1 w3 .
Notice that sofar no inner product has been invoked in any of the definitions and results of this section. With the
help of an inner product certain “straightforward” modifications can be made that generalize foregoing definitions
and results. The underlying mechanism is that of the conversion tools provided by Definition 2.9 on page 14.
Ansatz
• An n-dimensional inner product space V over R.

W V Wp Vp
• Tensor spaces of type p (V ), p (V ), (V ), (V ).
2.8 Tensors 31
Observation
Recall Theorem 2.3, page 13, 2.9 page 14, and Result 2.2, page 14. An inner product on V induces an
VpDefinitionW p
inner product on the spaces (V ) and (V ), as follows. For any v1 , . . . , vk , x1 , . . . , xk ∈ V , set
• (v1 ∧ . . . ∧ vk |x1 ∧ . . . ∧ xk ) = deth]vi , xj i, respectively

• (v1 ∨ . . . ∨ vk |x1 ∨ . . . ∨ xk ) = perm h]vi , xj i.
V W
Likewise for the spaces p (V ) and p (V ):

• v̂1 ∧ . . . ∧ v̂k |x̂1 ∧ . . . ∧ x̂k = dethv̂i , [x̂j i, respectively

• v̂1 ∨ . . . ∨ v̂k |x̂1 ∨ . . . ∨ x̂k = perm hv̂i , [x̂j i,
for any v̂1 , . . . , v̂k , x̂1 , . . . , x̂k ∈ V ∗ .
Section 2.8.5 covers tensors on an inner product space in more generality.
2.8.4. Vector Spaces with an Oriented Volume
Ansatz
• A fixed basis {ei } of V .

V
• A nonzero n-form µ ∈ n (V ).
Terminology
If µ(e1 , . . . , en ) = 1 then µ is called the unit volume form.
Observation
V
• We have dim n (V ) = 1.
V
• Consequently, if µ1 , µ2 ∈ n (V ), then µ1 and µ2 differ by a constant factor λ ∈ R.
• For the unit volume form we have, according to the observation on page 28,
1
µ = ê1 ∧ . . . ∧ ên = [i1 , . . . , in ] êi1 ∧ . . . ∧ êin = [i1 , . . . , in ] A êi1 ⊗ . . . ⊗ êin .

n!
The A -operator on the right hand side may be dropped, since its effect is void when
V combined with
antisymmetric coefficients. We conclude that the covariant components of µ ∈ n (V ) are given by
µi1 ...in = [i1 , . . . , in ].
Caveat
The identification µi1 ...in = [i1 , . . . , in ] holds only relative to the fixed basis used above to construct µ. We refer
to section 2.8.6 for a basis independent evaluation.
Observation
Thus µ = µ|i1 ...in | êi1 ∧ . . . ∧ êin = µi1 ...in êi1 ⊗ . . . ⊗ êin . This shows the convenience of the two variants of
Einstein summation.
32 Tensor Calculus
V
Definition 2.22 Let (V, µ) denote the vector space V endowed with the n-form µ ∈ n (V ).
• (V, µ) is called a vector space with an oriented volume.
• (V, µ) is said to have a positive orientation if µ(e1 , . . . , en ) > 0.
• (V, µ) is said to have a negative orientation if µ(e1 , . . . , en ) < 0.
Terminology
The use of attributes “positive” and “negative” in this context is usually self-explanatory, e.g.
V
• given a fixed basis {ei } of V one refers to µ ∈ n (V ) as a positive volume form if µ(e1 , . . . , en ) > 0,
and
V
• given a basic volume form µ ∈ n (V ) one refers to {ei } as a positively oriented basis if µ(e1 , . . . , en ) >
0, etc.
Notation
Instead of (V, µ) we shall simply write V . The existence and nature of an oriented volume should be clear from
the context.
Observation
Using Definition 2.17 on page 24 we may write, for x1 , . . . , xn ∈ V ,
 1
hê1 , xn i

hê , x1 i ...
j
µ (x1 , . . . , xn ) = dethê , xi i = det  .. ..
,
 
. .
hên , x1 i . . . hên , xn i
which clearly reveals the role of the Kronecker tensor (and thus of the dual vector space V ∗ ) in the definition
of µ. Consistent with a foregoing observation it follows that the covariant components of µ = µi1 ...in êi1 ⊗
. . . ⊗ êin are given by µi1 ...in = µ(ei1 , . . . , ein ) = [i1 , . . . , in ] µ(e1 , . . . , en ) = [i1 , . . . , in ] dethêj , ei i =
[i1 , . . . , in ] det δij = [i1 , . . . , in ].
Definition
V 2.23 Let a1 , . . . , ak ∈ V for some k ∈ {0, . . . , n}, then we define the (n − k)-form µya1 y . . .yak ∈
n−k (V ) as follows:
(µya1 y . . .yak ) (xk+1 , . . . , xn ) = µ (a1 , . . . , ak , xk+1 , . . . , xn ) .

2.8 Tensors 33
Observation
• A trivial rewriting of the previous observation allows us to rewrite Definition 2.23 on page 32 as follows:
 1
hê , a1 i . . . hê1 , ak i hê1 , xk+1 i . . . hê1 , xn i

(µya1 y . . .yak ) (xk+1 , . . . , xn ) = det  .. .. .. ..

.
 
. . . .
hên , a1 i . . . hên , ak i hên , xk+1 i . . . hên , xn i
V
• The holor of µya1 y . . .yak ∈ n−k (V ) for given a1 , . . . , ak ∈ V is established as follows:
ai11 . . . aikk (µyei1 y . . .yeik ) eik+1 , . . . , ein

(µya1 y . . .yak ) eik+1 , . . . , ein =
= ai11 . . . aikk µ (ei1 , . . . , ein ) = µi1 ...in ai11 . . . aikk ,
i
in which we have used the decompositions ap = app eip for p = 1, . . . , k together with multilinearity.
i
• By the same token, expanding also xq = xqq eiq for q = k + 1, . . . , n, we obtain
i
(µya1 y . . .yak ) (xk+1 , . . . , xn ) = µi1 ...in ai11 . . . aikk xk+1
k+1
. . . xinn .
Remark
V
• An oriented volume µ ∈ n (V ) on a vector space V does not require an inner product.
) exists by virtue of the existence of the dual space V ∗ .

V
• An oriented volume µ ∈ n (V
• If µ(e1 , . . . , en ) = 1, then µ = ê1 ∧ . . . ∧ ên .

• More generally, if µ(e1 , . . . , en ) = 1, then µye1 y . . .yek = êk+1 ∧ . . . ∧ ên .
The following example explains the terminology of a “volume form”.
Example
V
Let the notation be as in the examples on page 30 and following. The space 3 (V ) is spanned by
{µ = dx ∧ dy ∧ dz}, which, according to the previous example, is given by
µ(u, v, w) = u1 v 2 w3 + u3 v 1 w2 + u2 v 3 w1 − u3 v 2 w1 − u1 v 3 w2 − u2 v 1 w3 .
This is the (signed) volume of a parallelepiped spanned by the (column vectors corresponding to the) vectors
u, v, w. The result is positive (negative) if the ordered triple u, v, w has the same (respectively opposite) orien-
tation as the basis {e1 , e2 , e3 }.
The example suggests that similar interpretations hold for forms of lower degrees. This is indeed the case, as the
next example shows.
34 Tensor Calculus
Example
3 3
V V
The spaces 1 (R ) and 2 (R ) are spanned by {dx, dy, dz} and {dx ∧ dy, dx ∧ dz, dy ∧ dz}, respectively.
• According to the example on page 30 we have
hdx, vi = v 1 hdy, vi = v 2 hdz, vi = v 3 .
One recognizes the (signed) 1-dimensional volumes (lengths) of the parallel projections of the vector v
onto the x, y, respectively z axis. Recall the examples on page 17 for a physical interpretation in terms of
particle kinematics or wave phenomena.
• According to the example on page 30 we have
1
w1

hdx, vi hdx, wi v
(dx ∧ dy) (v, w) = det = det = v 1 w2 − v 2 w1 .
hdy, vi hdy, wi v2 w2
This is clearly the (signed) 2-dimensional volume (surface area) of the parallelogram spanned by v and w
as it exposes itself from a perspective along the z direction, i.e. the area of the parallelogram after projection
along the z axis onto the x-y plane, cf. Fig. 2.1.
Figure 2.1: Illustration of (dx ∧ dy) (v, w) as the area of the cast shadow of a solar panel onto the ground (x–
y-plane) when lit from above (along z-direction). Note that (v ∧ w) (dx, dy) = (dx ∧ dy) (v, w) provides an
equivalent interpretation of the same configuration.
2.8.5. Tensors on an Inner Product Space
Ansatz
An n-dimensional inner product space V over R.
Heuristics
The same mechanism that allows us to toggle between vectors and dual vectors in an inner product space, viz. via
the ]- and [-operator associated with the metric G, also allows us to establish one-to-one relationships between
Tpq (V ) and Trs (V ), provided total rank is the same: p + q = r + s. This implies that—in an inner product
space—tensor types can be ordered hierarchically according to their total ranks, and that, given fixed total rank, a
tensor can be represented in as many equivalent ways as there are ways to choose its covariant and contravariant
rank consistent with its total rank. The basic principle is most easily understood by an example.
2.8 Tensors 35
Example
Let us interpret the metric G as an element of T02 (V ), as previously discussed. We may consider the following
four equivalent alternatives:
• G : V × V → R : (v, w) 7→ G(v, w);
def
• H : V ∗ × V ∗ → R : (v̂, ŵ) 7→ H(v̂, ŵ) = G([v̂, [ŵ);
def
• I : V ∗ × V → R : (v̂, w) 7→ I(v̂, w) = G([v̂, w);
def
• J : V × V ∗ → R : (v, ŵ) 7→ J(v, ŵ) = G(v, [ŵ).
Note that G ∈ T02 (V ), H ∈ T20 (V ), I ∈ T11 (V ), J ∈ T11 (V ). If {ei } is a basis of V , then
• G = gij êi ⊗ êj = ]ei ⊗ êi = êi ⊗ ]ei ;
• H = g ij ei ⊗ ej = [êi ⊗ ei = ei ⊗ [êi ;
• I = δji ei ⊗ êj = ei ⊗ êi = [êi ⊗ ]ei ;
• J = δij êi ⊗ ej = êi ⊗ ei = ]ei ⊗ [êi .
Caveat
The above example declines from the strict ordering of the spaces V and V ∗ in the domain of definition of a
tensor, as introduced

in Definition 2.12 on page 18. This explains the existence of two distinct tensors, I 6= J, of
equal rank 11 . Note that I(v̂, w) = J(w, v̂).
Definition 2.24 Recall Result 2.2 on page 14 and Definition 2.12 on page 18. For any p, q ∈ N0 we define the
covariant representation ]T ∈ T0p+q (V ) of T ∈ Tpq (V ) as follows:
]T(v1 , . . . , vp+q ) = T(]v1 , . . . , ]vp , vp+1 , . . . , vp+q ) for all v1 , . . . , vp+q ∈ V .
Likewise, the contravariant representation [T ∈ Tp+q
0 (V ) of T ∈ Tpq (V ) is defined as follows:
[T(v̂1 , . . . , v̂p+q ) = T(v̂1 , . . . , v̂p , [v̂p+1 , . . . , [v̂p+q ) for all v̂1 , . . . , v̂p+q ∈ V ∗ .
In general, a mixed representation of T ∈ Tpq (V ) is obtained by toggling an arbitrary subset of (dual) vector
spaces V and V ∗ in the domain of definition of the tensor by means of the metric induced {], [}-mechanism in a
similar fashion.
Example
Suppose T ∈ T12 (V ) is given relative to a basis {ei } of V by T = ti jk ei ⊗ êj ⊗ êk . Let û = ui êi , v̂ = vj êj ,
ŵ = wk êk , u = ui ei , v = v j ej , w = wk ek . Then ]T = tijk êi ⊗ êj ⊗ êk and [T = tijk ei ⊗ ej ⊗ ek . In terms
of components:
T(û, v, w) = ti jk ui v j wk ,
]T(u, v, w) = T(]u, v, w) = tijk ui v j wk , (2.1)
ijk
[T(û, v̂, ŵ) = T(û, [v̂, [ŵ) = t ui vj wk .
The remaining five mixed representations are
T1 (û, v̂, w) = T(û, [v̂, w) = tij k ui vj wk ,

T2 (û, v, ŵ) = T(û, v, [ŵ) = ti j k ui v j wk ,
T3 (u, v̂, ŵ) = T(]u, [v̂, [ŵ) = ti jk ui v j wk , (2.2)
T4 (u, v̂, w) = T(]u, [v̂, w) = ti j k ui v j w k ,
T5 (u, v, ŵ) = T(]u, v, [ŵ) = tij k ui v j wk ,
making up the total of 23 = 8 different representations of the rank-3 tensor T, corresponding to the number of
possibilities to toggle the arguments of the tensor’s 3 slots.
36 Tensor Calculus
Caveat
The example shows that one quickly runs out of symbols if one insists on a unique naming convention for all
distinct tensors that can be constructed in this way, especially for high ranks. One could opt for a systematic
labeling, e.g. using a binary subscript to denote which argument requires a toggle (1) and which does not (0). In
that case we would have, in the above example, T = T000 , ]T = T100 , [T = T011 , T1 = T010 , T2 = T001 ,
T3 = T111 , T4 = T110 , T5 = T101 .
Other labeling policies are possible, e.g. using a similar 3-digit binary code to denote argument type, 0 for a
vector and 1 for a covector, say. This would lead to the following relabeling: T = T100 , ]T = T000 , [T = T111 ,
T1 = T110 , T2 = T101 , T3 = T011 , T4 = T010 , T5 = T001 .
Conventions like these are, despite the mathematical clarity they provide, rarely used. One often simply writes T
for all cases, relying on the particular tensor prototype for disambiguation (“function overloading”).
Notation
• By abuse of notation one invariably uses the same symbol, T say, to denote any of the (covariant, con-
travariant, or mixed) prototypes discussed above.
• By the same token the space Tk (V ) = p+q=k Tpq (V ) collectively refers to all tensors of total rank k.
S
• If a particular prototype T ∈ Tpq (V ) is intended it should be clear from the context.
• Instead of the potentially confusing notation tji11...i

...js for a mixed holor one often writes t
r i1 ...ir
j1 ...js , i.e. one
with a manifest ordering of indices.
Observation
• The invocation of the ]-operator to a vector argument of the tensor is reflected in the holor as an index
lowering.
• The invocation of the [-operator to a dual vector argument of the tensor is reflected in the holor as an index
raising.
Caveat
• Recall that ]ei = G(ei ) = gij êj , respectively [êi = G−1 (êi ) = g ij ej . Thus, at the level of (dual) basis
vectors, ] raises, and [ lowers indices!
• Notice the way indices are raised or lowered in a mixed holor, viz. such as to respect the index ordering, cf.
a previous remark on mixed holor notation.
Reminder
• In an inner product space, tensor (holor) indices can be raised or lowered with the help of the (dual) metric
tensor.
• As a rule of thumb, consistency of free index positioning determines whether and how to use metric and
dual metric components, gij or g ij , in an expression involving ] or [.
2.8.6. Tensor Transformations

2.8 Tensors 37
2.8.6.1. “Absolute Tensors”
Ansatz

• A linear space of mixed tensors, Tpq (V ).
Recall that a holor depends on the chosen basis, but the corresponding tensor itself does not. This implies that
holors transform in a particular way under a change of basis, which is characteristic for tensors. In fact, in the
literature tensors are often defined by virtue of the tensor transformation law governing their holors. Here we
present it essentially as a result.
Result 2.8 Let fj = Aij ei define a change of basis, with transformation matrix A. Define B = A−1 , i.e. AB = I,
or in terms of components, Aik Bjk = δji . Then
i ...i def i ...i

T = tj11 ...jpq ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq = tj11 ...jpq fi1 ⊗ . . . ⊗ fip ⊗ f̂ j1 ⊗ . . . ⊗ f̂ jq ,
iff the holor adheres to the tensor transformation law,
i ...i ` i k ...k
tj11 ...jpq = A`j11 . . . Ajqq Bki11 . . . Bkpp t`11...`qp .
Proof of Result 2.8. Suppose f̂ i = Cji êj . Then, by definition of duality,
def ∗ def
δji = hf̂ i , fj i = hCki êk , A`j e` i = Cki A`j hêk , e` i = Cki A`j δ`k = Cki Akj ,
whence C = A−1 = B. Bilinearity of the Kronecker tensor has been used in ∗. Consequently,
ei = Bi` f` and êj = Ajk f̂ k .
Substitution into the basis expansion of T yields the stated result for the respective holors.
Remark
i ...i
The homogeneous nature of the tensor transformation law implies that a holor equation of the form tj11 ...jpq = 0
holds relative to any basis if it holds relative to a particular one.
Example
With the same notation as used in Result 2.8 we have for a vector v = v i ei = v i fi :
v i = Bji v j ,
also referred to as vector transformation law. Recall Theorem 2.1 on page 7.

38 Tensor Calculus
Example
Let df = ∂i f êi ∈ V ∗ denote the gradient of a scalar function f ∈ C 1 (Rn ), interpreted as a dual (!) vector and
evaluated at a fixed point. The interpretation of this is as follows. If v = v i ei ∈ V , then
∗
hdf , vi = ∂i f v i = (∇f |v) ,
the directional derivative of f along v. The equality marked by ∗ holds only if V is an inner product space, in
which case we identify ∇f = [df ∈ V .
According to the tensor transformation law we have df = ∂ i f f̂ i , with f̂ i = Bji êj , and
∂ i f = Aì ∂` f .
Notice the analogy with “infinitesimal calculus”. Indeed, a change of coordinates, x 7→ x = φ(x), induces a
reparametrization df = ∂i f dxi = ∂ i f dxi , in which
∂f ∂f ∂x` def ∂x`

∂if = i
= = Aì ∂` f with Aì = ,
∂x ∂x` ∂xi ∂xi
by the chain rule.
Caveat
One should distinguish between the new variable x and the coordinate transformation function φ that relates it to
the original variable x by x = φ(x). In practice one often treats x either as a variable or as a function (i.e. x ≡ φ),
depending on context. Likewise for x and φ−1 .
Example
Recall the observation on page 10, which confirms the tensor transformation law for the dual basis vectors dxi :
∂xi j
dxi = dx = Bji dxj ,
∂xj
with B = A−1 , cf. the previous example.
Caveat
Compare the latter example of the transformation law for a basis dual vector, dxi = Bji dxj , to the one on page 37
for the components (holor) of a vector, v i = Bji v j . The apparent formal correspondence has led practitioners
of tensor calculus (mostly physicists and engineers) to interpret the dual basis vector dxi as the “components
of an infinitesimal displacement vector”. This misinterpretation is, although potentially confusing, harmless and
indeed justified to the extent that it is consistent with the stipulated tensor transformation laws for basis dual
vectors, respectively regular vector components.
2.8.6.2. “Relative Tensors”
Recall the unit volume form introduced in section 2.8.4. On page 31 it was anticipated with a modest amount
of foresight that the holor µi1 ...in = [i1 , . . . , in ] depends on the chosen basis. Heuristically this is clear from
the geometric interpretation provided by the example on page 33 in terms of the volume of the parallelepiped
spanned by the basis vectors. Let us scrutinize the situation more carefully.
2.8 Tensors 39
Ansatz

• The unit volume form µ = µ|i1 ...in | êi1 ∧ . . . ∧ êin = µi1 ...in êi1 ⊗ . . . ⊗ êin ∈
V
n (V ), with µi1 ...in =
[i1 , . . . , in ].
The following lemma allows us to compute the holor of µ relative to an arbitrary basis, given its canonical form
[i1 , . . . , in ] relative to a fiducial basis, via Result 2.8, page 37.
Lemma 2.1 Let µ = µi1 ...in êi1 ⊗ . . . ⊗ êin ∈ n (V ), with µi1 ...in = [i1 , . . . , in ] relative to basis {ei } of V ,
V
and fj = Aij ei , so that êi = Aij f̂ j . Setting µ = µi1 ...in f̂ i1 ⊗ . . . ⊗ f̂ in , we have
µj1 ...jn = Aij11 . . . Aijnn µi1 ...in = det A µj1 ...jn .
Proof of Lemma 2.1. In accordance with Result 2.8, page 37, we have
µ = µi1 ...in êi1 ⊗ . . . ⊗ êin = µi1 ...in Aij11 . . . Aijnn f̂ j1 ⊗ . . . ⊗ f̂ jn .
This proves the first identity. Using µi1 ...in = [i1 , . . . , in ], together with an observation made on page 5 in the
context of the determinant, this may be rewritten as
µ = det A [j1 , . . . , jn ] f̂ j1 ⊗ . . . ⊗ f̂ jn .
This establishes the proof of the last identity.
Observation
• Note that det A is the volume scaling factor of the transformation with matrix A.
• If µ (e1 , . . . , en ) = 1, then µ (f1 , . . . , fn ) = det A.
• The holor µi1 ...in = [i1 , . . . , in ] is not invariant under the absolute tensor transformation law, Result 2.8,
in the sense that µi1 ...in = det A µi1 ...in 6= [i1 , . . . , in ] = µi1 ...in .
Heuristics
Clearly, the constant antisymmetric symbol fails to define an invariant holor of a tensor. The definition of the
holor in terms of the antisymmetric symbol requires us to single out a preferred basis relative to which it holds,
i.e. the fiducial basis {ei } has a preferred status among all possible bases. The question arises whether it is
possible to give an invariant expression for the holor, i.e. one that holds in any basis, so that all bases are treated
on equal foot and no a priori knowledge of a preferred one is needed. This is indeed possible. There are two ways
to remedy the problem:
1. without the help of an inner product (Result 2.9), respectively

2. with the help of an inner product (Definition 2.25 and Result 2.10).
Result 2.9 Recall Result 2.8, page 37, for notation. If

µ = µ|i1 ...in | êi1 ∧ . . . ∧ êin = µ|i1 ...in | f̂ i1 ∧ . . . ∧ f̂ in ,
and, by definition,
def −1
µi1 ...in = (det A) Aji11 . . . Ajinn µj1 ...jn ,
then µi1 ...in ≡ µi1 ...in ≡ [i1 , . . . , in ] is an invariant holor.
40 Tensor Calculus
Terminology
• This adapted transformation law is also known as the relative tensor transformation law. The attribute
“relative” refers to the occurrence of (a power of) the Jacobian determinant.
• By virtue of the relative tensor transformation law governing its holor, µ is referred to as a relative tensor.
Strictly speaking this terminology refers to its holor definition!
Proof of Result 2.9. Using an observation on page 5 we find
def −1 −1
µi1 ...in = (det A) Aji11 . . . Ajinn µj1 ...jn = (det A) Aji11 . . . Ajinn [j1 , . . . , jn ]
−1
= (det A) [i1 , . . . , in ] det A = [i1 , . . . , in ] = µi1 ...in .
Remark
• In Lemma 2.1 on page 39 the holor µi1 ...in = [i1 , . . . , in ] is transformed according to the absolute tensor
transformation law, recall Result 2.8 on page 37, which implies that µi1 ...in = det A µi1 ...in 6= µi1 ...in .
• In Result 2.9 the holor µi1 ...in = [i1 , . . . , in ] is transformed according to the relative tensor transformation
law, rendering the holor µi1 ...in = [i1 , . . . , in ] invariant in the sense that µi1 ...in = µi1 ...in = [i1 , . . . , in ]
regardless of the basis.
• Adopting the relative tensor transformation law allows us to interpret µi1 ...in as a ‘covariant alias’ of the
permutation symbol [i1 , . . . , in ], which by definition does not depend on a basis.
• There is no inconsistency other than abuse of notation (to be resolved below). In both interpretations we
have
µ = µi1 ...in êi1 ⊗ . . . ⊗ êin = µi1 ...in f̂ i1 ⊗ . . . ⊗ f̂ in ,
or, equivalently,
µ = µ|i1 ...in | êi1 ∧ . . . ∧ êin = µ|i1 ...in | f̂ i1 ∧ . . . ∧ f̂ in .
Notation
The following convention is used henceforth to support consistent use of the Einstein summation convention and
to resolve ambiguity w.r.t. the meaning of µi1 ...in :
• Henceforth we will adopt µi1 ...in and µi1 ...in as covariant and contravariant aliases for the permutation
symbol, i.e. µi1 ...in ≡ µi1 ...in ≡ [i1 , . . . , in ].
• Consequently, the holor µi1 ...in is subject to the relative tensor transformation law, Result 2.9.
2.8.6.3. “Pseudo Tensors”
Ansatz
• An n-dimensional inner product space V over R, based on our default Definition 2.5.
• The unit volume form µ = µ|i1 ...in | êi1 ∧ . . . ∧ êin = µi1 ...in êi1 ⊗ . . . ⊗ êin ∈
V
n (V ), with invariant holor
definition µi1 ...in = [i1 , . . . , in ].
2.8 Tensors 41
Notation
For brevity one denotes the determinant of the Gram matrix, gij = (ei |ej ), by g = det gij , recall page 5
concerning a remark on the use of indices, and Definition 2.8 on page 12.
Definition 2.25 The Levi-Civita tensor is the unique unit n-form of positive orientation:
√ √
= g µ = g ê1 ∧ . . . ∧ ên = |i1 ...in | êi1 ∧ . . . ∧ êin .
Observation
√ √
• i1 ...in = g [i1 , . . . , in ] = g µi1 ...in .
• = i1 ...in êi1 ⊗ . . . ⊗ êin = |i1 ...in | êi1 ∧ . . . ∧ êin in terms of the appropriate summation conventions.
Vn
Theorem 2.6 The contravariant representation of the Levi-Civita tensor, [ ∈ (V ), recall Definition 2.24 on
page 35, is given by [ = |i1 ...in | ei1 ∧ . . . ∧ ein , with holor representation
1
i1 ...in = √ [i1 , . . . , in ] .
g
Proof of Theorem 2.6. It suffices to prove the validity of the holor representation by index raising:
√ 1
i1 ...in = g i1 j1 . . . g in jn j1 ...jn = g i1 j1 . . . g in jn g [j1 , . . . , jn ] = √ [i1 , . . . , in ] .
g
In the last step we have made use of the fact that g i1 j1 . . . g in jn [j1 , . . . , jn ] = det g ij [i1 , . . . , in ], an observation
made on page 5, together with the fact that det g ij = 1/ det gij .
Observation
1 1
• i1 ...in = √ [i1 , . . . , in ] = √ µi1 ...in relative to the fiducial basis {ei }.
g g
• [ = i1 ...in ei1 ⊗ . . . ⊗ ein = |i1 ...in | ei1 ∧ . . . ∧ ein .
Result 2.10 Recall Result 2.8, page 37, for notation. Let
= |i1 ...in | êi1 ∧ . . . ∧ êin

√
relative to basis {ei }, with i1 ...in = g [i1 , . . . , in ]. Moreover, let
= |i1 ...in | f̂ i1 ∧ . . . ∧ f̂ in
relative to {fi } with fi = Aji ej . Then the pseudotensor transformation law,

def
i1 ...in = sgn (det A)Aji11 . . . Ajinn j1 ...jn ,
√
renders the holor form invariant, in the sense that i1 ...in = g [i1 , . . . , in ].
Proof of Theorem 2.10. By virtue of the transformation properties of gij , recall Result 2.8 on page 37, one obtains
g ij = Aki A`j gk` ,

42 Tensor Calculus
whence p √
g = |det A| g .
By definition of the pseudotensor transformation law we have, using an observation on page 5,
def √
i1 ...in = sgn (det A)Aji11 . . . Ajinn j1 ...jn = sgn (det A) Aji11 . . . Ajinn g µj1 ...jn
√ √ p
= sgn (det A) det A g µi1 ...in = |det A| g µi1 ...in = g µi1 ...in .
Recall that µi1 ...in and µi1 ...in have been used here as aliases for the same (basis independent) permutation symbol
[i1 , . . . , in ].
Remark
• The Levi-Civita tensor is referred to as a pseudotensor, because its form invariant holor (!) transforms
according to the pseudotensor transformation law of Result 2.10, not the absolute tensor transformation
law of Result 2.8.
√
• Form invariance,
√ achieved by virtue of the factor g, should not be confused with numerical invariance.
√
Clearly g 6= g, whence i1 ...in 6= i1 ...in , unlike the formal identity µi1 ...in = µi1 ...in ≡ [i1 , . . . , in ].
• In the case of a non-positive inner product, recall
pDefinition 2.7, slight adaptations may be needed, a typical
√
one being the replacement of the g factor by |g| if g is negative.
The following definition is based on an earlier observation, recall page 5.
Definition 2.26 The generalised Kronecker, or permutation symbol, is defined as
δji11 δji1k
  
...  +1 if (i1 , . . . , ik ) even permutation of (j1 , . . . , jk ) ,
 .. ..  =
δji11 ...ik
...jk = det  . .   −1 if (i1 , . . . , ik ) odd permutation of (j1 , . . . , jk ) ,
δjik1 ... δjikk 0 otherwise .
Observation
• j1 ...jn i1 ...in = µj1 ...jn µi1 ...in = δji11 ...in
...jn .
i ...i
• `1 ...`k j1 ...jn−k `1 ...`k i1 ...in−k = µ`1 ...`k j1 ...jn−k µ`1 ...`k i1 ...in−k = k!δj11 ...jn−k
n−k
.
1 i1 ...in j1
• For the matrix A of a mixed tensor A ∈ T11 (V ) we have det Aij = δ A . . . Ajinn .
n! j1 ...jn i1
2.8.7. Contractions
Ansatz
• A linear space of mixed tensors, Tpq (V ).
2.9 The Hodge Star Operator 43
Definition 2.27 Let T ∈ T11 (V ). The trace, tr T ∈ R, of T is defined as follows:
tr T = T(êi , ei ) .
Remark
• The tr -operator is applicable to any mixed tensor T ∈ Tpq (V ) provided p, q ≥ 1, viz. by arbitrarily fixing
p − 1 covector and q − 1 vector arguments.
• Without “frozen” arguments one may interpret tr T as a mixed tensor tr T ∈ Tp−1
q−1 (V ).
• The tr -operator lowers both covariant and contravariant ranks by one.
Caveat
• In general, the notation tr T is ambiguous for T ∈ Tpq (V ), unless T is symmetric with respect to both
covector as well as vector arguments. (This is manifest if p = q = 1.)
• One could decorate the tr -operator with disambiguating subscripts and superscripts so as to indicate the
covector/vector argument pair it applies to. This is hardly ever done in the literature. Instead, ambiguities
should be resolved from the context.
2.9. The Hodge Star Operator
Ansatz
• An n-dimensional inner product space V over R furnished with a basis {ei }.

√
• The Levi-Civita tensor = g ê1 ∧ . . . ∧ ên = |i1 ...in | êi1 ∧ . . . ∧ êin ∈ n (V ).
V
V V
Definition 2.28 Let a1 , . . . , ak ∈ V . The Hodge star operator ∗ : k (V )→ n−k (V ) is defined as follows:

∗1 = for k = 0,
∗ (]a1 ∧ . . . ∧ ]ak ) = ya1 y . . .yak for k = 1, . . . , n,
followed by linear extension.
This definition could be rephrased by replacing ]a1 , . . . , ]ak ∈ V ∗ by â1 , . . . , âk ∈ V ∗ on the l.h.s. and
a1 , . . . , ak ∈ V by [â1 , . . . , [âk ∈ V on the r.h.s. A similar definition could be given for a contravariant
Vk Vn−k
counterpart ∗ : (V ) → (V ), in which case ]-conversion of the vector arguments a1 , . . . , ak ∈ V is
obsolete.
Remark
V V
The Hodge star operator establishes an isomorphism between k (V ) and n−k (V ), recall the dimensionality
argument on page 26:
V n n V
dim k (V ) = = = dim n−k (V ) .
k n−k
44 Tensor Calculus
Observation
i
• The holor of the tensor ∗ (]a1 ∧ . . . ∧ ]ak ) for fixed ap = app eip , with p = 1, . . . , k, is obtained as follows:
ai11 . . . aikk (∗ (]ei1 ∧ . . . ∧ ]eik )) eik+1 , . . . , ein

(∗ (]a1 ∧ . . . ∧ ]ak )) eik+1 , . . . , ein =
def
ai11 . . . aikk (yei1 y . . .yeik ) eik+1 , . . . , ein

=
= ai11 . . . aikk (ei1 , . . . , ein ) = i1 ...in ai11 . . . aikk .
i
• By the same token, expanding also xq = xqq eiq for q = k + 1, . . . , n, we obtain
i
(∗ (]a1 ∧ . . . ∧ ]ak )) (xk+1 , . . . , xn ) = i1 ...in ai11 . . . aikk xk+1
k+1
. . . xinn .
Result 2.11 Alternatively, we may rewrite Definition 2.28 as follows:

∗1 = 1 for k = 0,
∗ â1 ∧ . . . ∧ âk = y[â y . . .y[âk for k = 1, . . . , n.
In this case we find, setting âp = apjp êjp for p = 1, . . . , k, and using êj = g ij ]ei ,
∗ â1 ∧ . . . ∧ âk = a1j1 . . . akjk ∗ êj1 ∧ . . . ∧ êjk

eik+1 , . . . , ein eik+1 , . . . , ein
= g i1 j1 . . . g ik jk a1j1 . . . akjk (∗ (]ei1 ∧ . . . ∧ ]eik )) eik+1 , . . . , ein

def i1 j1
. . . g ik jk a1j1 . . . akjk (yei1 y . . .yeik ) eik+1 , . . . , ein

= g
= g i1 j1 . . . g ik jk a1j1 . . . akjk (ei1 , . . . , ein )
= i1 ...in g i1 j1 . . . g ik jk a1j1 . . . akjk .
Observation
As a rule of thumb, in order to compute the holor of the Hodge star of a general antisymmetric cotensor of
rank k, the holor of which has free
V indices i1 , . . . , ik , say, raiseVall indices and contract with the antisymmetric
i1 ...in -symbol. That is, if A ∈ k (V ), then the holor of ∗A ∈ n−k (V ) is given by
def 1
∗Aik+1 ...in = (∗A) eik+1 , . . . , ein = i1 ...in g i1 j1 . . . g ik jk Aj1 ...jk .

k!
In other words,
∗A = ∗Aik+1 ...in êik+1 ⊗ . . . ⊗ êin = ∗A|ik+1 ...in | êik+1 ∧ . . . ∧ êin .

In fact it suffices to remember that ∗ (]ei1 ∧ . . . ∧ ]eik ) eik+1 , . . . , ein = i1 ...in , by exploiting multilinearity.
Remark
• Taking k = 0 in the above observation is consistent with ∗1 = .

• Taking k = n shows that ∗ = ∗∗ 1 = 1. This result relies essentially on Definition 2.5. In the general
case of Definition 2.7Vyou may verify that ∗ = ∗∗ 1 = sgn g, and that ∗∗ A = (−1)k(n−k) sgn g A for a
general k-form A ∈ k (V ).
3. Differential Geometry
3.1. Euclidean Space: Cartesian and Curvilinear Coordinates
In this chapter we consider tensor fields, i.e. tensors, in the widest sense as introduced in Chapter 2, attached to all
points of an open subset Ξ ⊂ Rn and considered as “smoothly varying” functions of those points. The space Rn
represents n-dimensional Euclidean space, i.e. “classical” space endowed with a flat, homogeneous and isotropic
structure as we usually take for granted [8, 9, 10].
Terminology
Recall Definition 2.5, page 11. The terminology “n-dimensional Euclidean vector space” is synonymous to
“n-dimensional (real, positive definite) inner product space”.
Definition 3.1 An n-dimensional Euclidean space E is a metric space, i.e. an affine space equipped with a dis-
tance function d : E × E → R, furnished with a mapping + : E × V → E, in which V is an n-dimensional
Euclidean vector space, such that
p
• for all x, y ∈ E there exists a unique v ∈ V such that y = x + v, and d(x, y) = (v|v);
• for all x, y ∈ E and all v ∈ V d(x + v, y + v) = d(x, y);
• for all x ∈ E and all u, v ∈ V (x + u) + v = x + (u + v).
Heuristics
Roughly speaking, a Euclidean space is a Euclidean vector space V that has “forgotten” its origin, or for which
any point may be taken as origin.
Observation
p
The distance function naturally arises from the inner product: d(x, y) = (y − x|y − x), in which y − x denotes
the vector v ∈ V such that y = x + v.
Caveat
Both E and V are usually written as Rn . It should be clear from the context whether a Euclidean space or a
Euclidean vector space is meant.
A point in x ∈ E ∼ Rn can be represented by its Cartesian coordinates xi ∈ R relative to a (globally defined)

Cartesian coordinate system or standard basis {ei }, i = 1, . . . , n, in the usual way: x = xi ei . Here ei ∈ Rn is
the coordinate vector with an entry 1 in i-th position and with remaining n − 1 entries equal to 0.
It is sometimes convenient to consider (curvilinear) coordinates on (an open subset of) Rn . The formal definition
of coordinates, generally applicable to both Euclidean as well as curved spaces, is given in the next section.
46 Differential Geometry
3.2. Differentiable Manifolds
Differentiable manifolds, aka differential manifolds, are spaces that locally look Euclidean. The points of such
manifolds can be labelled with the help of an atlas, i.e. a collection of charts that provide us with coordinates on
open sets that cover the entire manifold.
Definition 3.2 An n-dimensional smooth differentiable manifold is a Hausdorff topological space M furnished
with a family of smooth diffeomorphisms φα : Ωα ⊂ M → φα (Ωα ) ⊂ Rn , with the following properties:
• {Ωα } is an open covering of M, i.e. Ωα is open and ∪α Ωα = M.

• If Ωα ∩ Ωβ 6= ∅, then φα (Ωα ∩ Ωβ ), φβ (Ωα ∩ Ωβ ) ⊂ Rn are open sets, and φα ◦ φ−1 and

β
φβ (Ωα ∩Ωβ )
(thus) φβ ◦ φ−1
α φα (Ωα ∩Ω ) are diffeomorphisms.

β
• The atlas {(Ωα , φα )} is maximal w.r.t. the previous two axioms.
The last axiom ensures that any chart we may come up with is contained in the atlas.
. .
To keep notation simple one often writes x = φα (p) and y = φβ (p) for the coordinates of a fiducial point
−1 −1
p = φα (x) = φβ (y) relative to their corresponding charts. The maps
φβ ◦ φ−1 n n −1
α : φα (Ωα ∩ Ωβ ) ⊂ R → φβ (Ωα ∩ Ωβ ) ⊂ R : x 7→ y = (φβ ◦ φα )(x) (3.1)
φα ◦ φ−1
β
n n
: φβ (Ωα ∩ Ωβ ) ⊂ R → φα (Ωα ∩ Ωβ ) ⊂ R : y 7→ x = (φα ◦ φ−1
β )(y) (3.2)
are called coordinate transformations.
Notation
By abuse of notation Eqs. (3.1–3.2) are often simplified as y = y(x), respectively x = x(y), with proper domains
tacitly suppressed.
Caveat
The thing to keep in mind is that coordinates x, y ∈ Rn are subjective in the sense that they can be rather arbitrarily
chosen. Eqs. (3.1–3.2) ensure that, given their associated charts (φα , φβ ), both refer to the same physical point
−1
p = φ−1
α (x) = φβ (y) ∈ M. The consequence of this observation is that, in all that follows, we must ensure that
numerical statements expressed in terms of coordinates are invariant under coordinate transformations. Here lies
the connection with Chapter 2.
−1
Definition 3.3 The Jacobian of the coordinate transformation φβ ◦ φ−1
α , respectively φα ◦ φβ , is the n×n matrix
of first order partial derivatives
∂(φβ ◦ φ−1 i
α ) (x) . ∂y
i ∂(φα ◦ φ−1 i
β ) (y) . ∂x
i
Sji (x) = j
= j
resp. Tji (y) = j
= j
.
∂x ∂x ∂y ∂y
Observation
The matrices S and T are regular, with S = T inv by virtue of the chain rule. In simplified notation we have
∂y i ∂xk
Ski Tjk = = δji .
∂xk ∂y j
3.3 Tangent Vectors 47
3.3. Tangent Vectors
In differential geometry, the tangent space takes over the role of the archetypical vector space V that lies at the
core of all tensor calculus concepts introduced in Chapter 2. You may see this as a local copy of V attached to
each point x ∈ M of the base manifold, with dim V = dim M = n. For reason of clarity we write TMx for a
local copy of V attached to x, or, by abuse of notation, TMx if x = φα (x) ∈ Rn in some given coordinate chart.
The tangent space is the vector space consisting of all tangent vectors at a fiducial point.
Definition 3.4 Let γ : (−ε, ε) → M : t 7→ γ(t) be a smooth parametrized curve, with γ(0) = p ∈ M, and
f ∈ C ∞ (M) an arbitrary smooth scalar field. Then the tangent vector of γ at p is the linear map given by

∞ d
γ̇(0) : C (M) → R : f 7→ (f ◦ γ) .
dt t=0
Smoothness of γ and f means that the numerically valued functions γ = φα ◦ γ and f = f ◦ φ−1 α are smooth on
the open interval (−ε, ε) ⊂ R, respectively on the open neighbourhood φα (Uα ) ⊂ Rn with p ∈ Uα , in the sense
of standard calculus, i.e. γ i ∈ C ∞ ((−ε, ε)) for each component i = 1, . . . , n, and f ∈ C ∞ (φα (Uα )).
Now recall the observation on page 8 concerning the abstract notation of basis vectors as a partial derivative
operators, initially justified by the formal consistency of their ‘coordinate transformation properties’ (chain rule)
with basis transformations. On a differentiable manifold we can make this more concrete, because we now
actually have something to take derivatives of.
Evaluating the expression in Definition 3.4 on a dummy scalar field with the help of a coordinate chart such that
x = φα (x) ∈ φα (Uα ) ⊂ Rn , we observe that

d d −1
d i ∂ ∗ i ∂
γ̇(0)f = (f ◦ γ) =
f ◦ φα ◦ φα ◦ γ = f ◦ γ = ẋ (0)
f = ẋ (0) f,
dt t=0 dt t=0 dt t=0 ∂xi x(0) ∂xi γ(0)
(3.3)
. i .
in which γ i (t) = xi (t) and γ̇ (t) = ẋi (t). Here we have made use of the aforementioned coordinate prototypes
def
γ : (−ε, ε) ⊂ R → Rn : t 7→ x(t) = γ(t) , (3.4)
def
n
f : φα (Uα ) ⊂ R → R : x 7→ f (x) = (f ◦ φ−1
α )(x) , (3.5)
and of the convenient definition (∗)

∂ ∗ ∂
f = f. (3.6)
∂xi x(0) ∂xi γ(0)
The dummy nature of f on left and right hand sides of Eq. (3.3) justifies the notation introduced without derivative
connotation on page 8.
Notation
Recall Definition 3.4 and Eqs. (3.3–3.6). Instead of γ̇(0) we write

i ∂

v|p = v i
(3.7)
∂x p
in which v|p = γ̇(0) and v i = ẋi (0). The subscripts is reminiscent of the fact that the tangent vector is defined
at p = γ(0). This construction can be repeated for any point x ∈ M and any smooth curve γ ∈ C ∞ ((−ε, ε), M).
Observation
Omitting explicit reference to the fiducial point p for the sake of simplicity, we observe from Eq. (3.3) that
vf = df (v) = hdf , vi .
In particular, if we take f i : M → R : x 7→ xi and vj = ∂j ∈ TMx in an arbitrary coordinate chart x = φα (x),

then vj f i = ∂j xi = δji = dxi (∂j ) = df i (vj ), consistent with the formal notation for dual basis vectors dxi
introduced on page 8.
Result 3.1 Recall the Kronecker tensor definition in Section 2.3. Relative to a coordinate basis we have, at any
implicit point p ∈ M,
hdxi , ∂j i = δji .
3.4. Tangent and Cotangent Bundle
Ansatz
An n-dimensional manifold M.
Definition 3.5 The tangent space at x ∈ M, TMx , is the vector space of tangent vectors vx at point x ∈ M.
Remark
Recall from the previous section that the definition can be stated in intrinsic terms by defining TMx as the span
of n tangent vectors of a family of mutually transversal curves intersecting at x ∈ M. A special case is obtained
by considering the n local coordinate curves induced by the coordinate system at x ∈ M.
Terminology
In the latter case the basis is referred to as a coordinate basis or holonomic basis, in which case the i-th coordinate
basis vector ei is also often written as ∂i , recall the observation on page 8.
Definition 3.6 The tangent bundle, TM = ∪x∈M TMx , is the union of tangent spaces over all x ∈ M.
Observation
• In order to identify base points explicitly, one often incorporates the projection map, π : TM → M :
(x, v) 7→ x, into the definition of the tangent bundle. In particular π(TMx ) = x and π(TM) = M.
Notation
• Recall Definition 3.5. The cotangent space at x ∈ M is denoted by T∗ Mx .

• Recall Definition 3.6. The cotangent bundle is written as T∗ M = ∪x∈M T∗ Mx .
3.5. Exterior Derivative
Ansatz
The tangent bundle TM over an n-dimensional manifold M.
3.6 Affine Connection 49
Definition 3.7 Let ω k = ωi1 ...ik dxi1 ∧ . . . ∧ dxik ∈ k (TM) be an antisymmetric covariant tensor field of rank
V
k, or k-form for brevity, with x ∈ M. The exterior derivative of ω is the (k + 1)-form given by
dω k = dωi1 ...ik ∧ dxi1 ∧ . . . ∧ dxik ∈ k+1 (TM) ,

V
def
with df = ∂i f dxi for any sufficiently smooth scalar field f : M → R.
Caveat
We have loosely identified M ∼ Rn via an implicit choice of coordinates. Strictly speaking this can be done for
sufficiently small neighbourhoods of any fiducial point x ∈ M and a corresponding open set of Rn .
Observation
One readily verifies the following properties:
• dω n = 0 for any ω n ∈ n (TM).
V
def
• d2 ω k = ddω k = 0 for any ω k ∈ and k = 0, . . . , n, in other words: d2 = 0.
V
k (TM)
• One sometimes writes d∧ instead of d.
Exterior derivatives are often used in combination with the Hodge star to extract derivatives from a k-form field
in some desired format. Note that d increases the rank of a k-form by 1, whereas ∗ converts a k-form into an
(n−k)-form. The following example illustrates this up to first order derivatives.
Example
In the following examples, α = α1 dx + α2 dy + α3 dz ∈ T∗ M, with αi ∈ C ∞ (M) and dim M = 3. For
simplicity the manifold is assumed to be Euclidean, and identified with R3 in Cartesian coordinates (in which
the inner product defining Gram matrix equals the identity matrix). Using ∗1 = dx ∧ dy ∧ dz, ∗dx = dy ∧ dz,
∗dy = −dx ∧ dz, ∗dz = dx ∧ dy, ∗(dx ∧ dy) = dz, ∗(dx ∧ dz) = −dy, ∗(dy ∧ dz) = dx, ∗(dx ∧ dy ∧ dz) = 1,
we obtain the following results:
• ∗α = α3 dx ∧ dy − α2 dx ∧ dz + α1 dy ∧ dz;
• dα = (∂x α2 − ∂y α1 ) dx ∧ dy + (∂x α3 − ∂z α1 ) dx ∧ dz + (∂y α3 − ∂z α2 ) dy ∧ dz;
• ∗dα = (∂x α2 − ∂y α1 ) dz − (∂x α3 − ∂z α1 ) dy + (∂y α3 − ∂z α2 ) dx;
• d ∗ α = (∂x α1 + ∂y α2 + ∂z α3 ) dx ∧ dy ∧ dz;
• ∗d ∗ α = ∂x α1 + ∂y α2 + ∂z α3 ;
• et cetera.
Special cases arise if α = df for some f ∈ C ∞ (M), a so-called “exact 1-form”.
3.6. Affine Connection
Ansatz
Recall the observation on page 8 concerning the formal notation for vector decomposition relative to a coordinate
basis, v = v i ∂i . Also recall from infinitesimal calculus that for a scalar function f ∈ C ∞ (M) we may define the
differential df = ∂i f dxi ∈ T∗ M. This suggests that we interpret a vector as a directional derivative operator, as
follows.
Definition 3.8 Let f ∈ C ∞ (M) be a scalar function, and v = v i ∂i ∈ TM. Then

def def
vf = hdf , vi = v i ∂i f .
So, when acting on a scalar field it is tacitly understood that we make the formal identication v = hd · , vi.
Definition 3.9 An affine connection on a manifold M is a mapping ∇ : TM × TM → TM : (v, w) 7→ ∇v w, also

referred to as the covariant derivative of w with respect to v, with the following properties. Let f, f1 , f2 ∈ C ∞ (M),
v, v1 , v2 , w, w1 , w2 ∈ TM, and λ, λ1 , λ2 ∈ R, then
1. ∇f1 v1 +f2 v2 w = f1 ∇v1 w + f2 ∇v2 w for all f1 , f2 ∈ C ∞ (M),

2. ∇v (λ1 w1 + λ2 w2 ) = λ1 ∇v w1 + λ2 ∇v w2 , for all λ1 , λ2 ∈ R,
3. ∇v (f w) = f ∇v w + ∇v f w for all f ∈ C ∞ (M). Here we have defined ∇v f ≡ vf .
Remark
If we leave out the directional vector argument we call ∇w the covariant differential of w, and identify ∇f = df .
Thus the covariant derivative of a local frame vector (i.e. basis vector of the local tangent space) is a vector within
the same tangent space. Consequently it can be written as a linear combination of frame vectors.
Definition 3.10 The Christoffel symbols Γkij are defined by ∇ej ei = Γkij ek .
Caveat
1. Conventions for the ordering of the lower indices of the Christoffel symbols differ among literature sources.
As a rule of thumb I put the differentiation index in second position.
2. The Christoffel symbols are not the components of a tensor, v.i.
To see how the Christoffel symbols transform, consider a vector and dual covector basis transformation:
ek = S`k e` resp. ej = Tjm em .
The matrices T and S are the mutually inverse Jacobian matrices:
∂xi ∂xi
Tji = resp. Sji = .
∂xj ∂xj
Set
k
Γij = ek , ∇ej ei

Substitute the transformations, apply first and third rule of Definition 3.9, ∇ej = Tj` ∇e` , respectively ∇e` (Ti` ek ) =
∂` Tik ek + Ti` ∇e` ek , use the definition of the Christoffel symbols, ∇e` ek = Γm k` em , and finally the chain rule,
∂ j = Tj` ∂` , to obtain the following result:
Result 3.2 The components of the Christoffel symbols in coordinate system x are given relative to those in coor-
dinate system x by
k
Γij = S`k Tjm Tin Γ`nm + ∂ j Ti` .

3.6 Affine Connection 51
Thus the tensor transformation law, recall Result 2.8 on page 37, fails to hold due to the inhomogeneous term,
which arises after any non-affine coordinate transformation.
Remark
k
For any fixed point on the manifold you can solve the system of equations Γij = 0 by slick choice of Jacobian
matrix, i.e. by transforming to a suitable local coordinate system. However, in general it is not possible to achieve
this globally, unless the manifold happens to be flat (zero Riemann curvature, v.i.).
Using the above definitions one can easily prove the following.
Result 3.3 Let v = v i ei , w = wi ei ∈ TM be two vector fields. Then
∇v w = v i Di wj ej .
Here we have defined the components of the covariant derivative Di wj ≡ ∂i wj + Γjki wk .
Observation
• The meaning of Di applied to a holor is rank-dependent; the above definition applies to a holor with only
one, contravariant index.
• Moreover, Di wj does not only depend on the j th component of w, but on all components wk , k = 1, . . . , n.
Definition 3.11 See Definition 3.9. The affine connection can be extended to arbitrary tensor fields by general-
izing the product rule, viz. for any pair of tensor fields, T ∈ Tpq (TM) and S ∈ Trs (TM), say, we require
∇v (S ⊗ T ) = ∇v S ⊗ T + S ⊗ ∇v T .
Observation
Recall the linear trace operator from Definition 2.27. For a covector field ẑ ∈ T∗ M and a vector field w ∈ TM
we have hẑ, wi = tr (ẑ ⊗ v). Observe that
∇v hẑ, wi = ∇v (tr (ẑ ⊗ v)) = tr (∇v (ẑ ⊗ v)) = tr (∇v ẑ ⊗ v) + tr (ẑ ⊗ ∇v v) = h∇v ẑ, wi + hẑ, ∇v wi ,
implying the product rule Di (zj wj ) = Di zj wj + zj Di wj .
i ...i
Result 3.4 See Result 3.3. Let T = Tj11...jqp ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq and v = v i ei be a tensor, respectively
vector field on M, then
i ...i
∇v T = v i Di Tj11...jqp ei1 ⊗ . . . ⊗ eip ⊗ êj1 ⊗ . . . ⊗ êjq .
Here we have defined the components of the covariant derivative
i ...i i ...i ki ...i i i ...i k i ...i i ...i

Di Tj11...jqp ≡ ∂i Tj11...jqp + Γiki1 Tj1 ...j
2
q
p
+ . . . + Γkip Tj11...jqp−1 − Γkj1 i Tkj12 ...jpq − . . . − Γkjq i Tj11...jq−1
p
k.
With the latter definition the covariant derivative operator Di satisfies the product rule for general holor products.
Note that Di is holor-type dependent, involving a ‘Γ-correction term’ for each index.
Terminology
• With affine geometry one indicates the geometrical structure of a manifold endowed with an affine connec-
tion. Note that this does not require the existence of a metric. Without a Riemannian metric, M is referred
to as an affine space or affine manifold.
• If a metric (induced by an inner product on each TMx ) does exist we are in the realm of Riemannian
geometry. Endowed with a Riemannian metric, M is called a Riemannian space or Riemannian manifold.
We refer to section 3.9 for further details.
Assumption
• Recall Definition 2.8, page 12. It is assumed that, if TMx is an inner product space for every x ∈ M, local
bases are chosen in such a way that the coefficients of the Gram matrix, gij (x) = (ei |ej )x , vary smoothly
across the manifold: gij ∈ C ∞ (M). (This is guaranteed if the basis is holonomic.)
• Similar assumptions will henceforth be implicit in the case of tensor fields in general.
Notation
The x-dependency of tensor fields is tacitly understood and suppressed in the notation.
3.7. Lie Derivative
Ansatz
Definition 3.12 The Lie derivative of w ∈ V with respect to v ∈ V is defined as
Lv w = [v, w] ,
in which the so-called Lie bracket is defined by the commutator
[v, w] f = v(wf ) − w(vf ) = v i ∂i wj − wi ∂i v j ∂j f .

for any smooth function f .
Observation
• Lv w is itself a vector, with components [v, w]i = v j ∂j wi − wj ∂j v i .

• No connection is involved, since only derivations of R-valued functions are considered.
• Each term separately in the subtraction does not constitute a vector (it fails to satisfy the covariance re-
quirement), but the difference does.
• The Lie-bracket is antisymmetric.
• The Lie-bracket satisfies the Jacobi identity: [u, [v, w]] + [w, [u, v]] + [v, [w, u]] = 0.
3.8 Torsion 53
3.8. Torsion
Ansatz
• The tangent bundle TM over an n-dimensional manifold M.
• An affine connection.
Definition 3.13 Notation as before. The torsion tensor is given by
T(v, w) = ∇v w − ∇w v − [v, w] .
Remark
The torsion tensor thus reveals the difference between a commutator of ordinary derivatives, as in the Lie deriva-
tive, and of covariant derivatives induced by the affine connection.
Observation
Recall that ∇v w = v i (∂i wj + wk Γjki )∂j , from which you derive the components of the torsion tensor relative to
a coordinate basis:
j k i j
T(v, w) = Tki v w ∂j with Tki = Γjik − Γjki .
Observation
The torsion tensor is antisymmetric: T(v, w) = −T(w, v).
3.9. Levi-Civita Connection
Ansatz
• The tangent bundle TM over an n-dimensional manifold M.

• A Riemannian metric g ∈ T∗ M ⊗ T∗ M.
An important result is that a Riemannian geometry induces a unique torsion-free connection.
Definition 3.14 Recall Definition 3.12, page 52 and Definition 3.13, page 53. A metric connection or Levi-Civita
connection on a Riemannian manifold (M, g) satisfies the following properties:
• z (g(v, w)) = g(∇z v, w) + g(v, ∇z w) for all v, w, z ∈ TM.
• T(v, w) = 0 for all v, w ∈ TM.
Observation
The first requirement is tantamount to the product rule: ∇z (v · w) = ∇z v · w + v · ∇z w for all v, w, z ∈ TM.
Alternatively, in terms of the ]-operator, ∇ ◦ ] = ] ◦ ∇, recall Definition 2.9 on page 14.
Terminology
• The product rule is variously phrased as “covariant constancy of metric”, “preservation of metric”, “metric
compatibility”, et cetera, and can be condensely written as ∇z g = 0 for all z ∈ TM.
• The second requirement is referred to as the requirement of “zero torsion”.
Theorem 3.1 Recall Definition 3.14. The Levi-Civita connection is uniquely given by the Christoffel symbols
1
Γkij = g k` (∂i g`j + ∂j gì − ∂` gij ) .
2
Proof of Theorem 3.1. The proof is based on (i) the general properties of an affine connection, recall Definition 3.9
on page 50, and (ii) the defining properties of the Levi-Civita connection, Definition 3.14, page 53.
Here are the details. Show that relative to a basis the first condition of Definition 3.14 is equivalent to Dm gij = 0,
then cyclicly permute the three free indices; subtract the two resulting expressions from the given one. The result
is
∂m gij − ∂i gjm − ∂j gmi = gi` Γ`jm − Γ`mj + g`j Γìm − Γ`mi − g`m Γìj + Γ`ji .

Subsequently, according to Definition 3.13 on page 53 and the subsequent observation on page 53, the second
condition of Definition 3.14 is equivalent to
Γìj = Γ`ji .
Apply this to the previous equality. Contraction with − 12 g km on both sides, using the fact that g ik gkj = δji ,
finally yields the stated result.
Observation
A metric connection is symmetric:
Γkij = Γkji .
Recall that this is not necessarily the case for an affine connection in general.
3.10. Geodesics
Definition 3.15 The covariant differential ∇v ∈ T11 (TM) of a vector field v = v i ∂i ∈ TM is defined as
∇v = ∇∂k vdxk .
Observation
If we define ∇v = Dv j ∂j , then Dv j = dv j + Γjki v k dxi = Di v j dxi , with Di v j = ∂i v j + Γjki v k .
Remark
The definition can be extended to arbitrary tensor fields T ∈ Tpq (TM):
i ...i
∇T = DTj11...jqp dxj1 ⊗ . . . ⊗ dxjq ⊗ ∂i1 ⊗ . . . ⊗ ∂ip ,
i ...i i ...i
with DTj11...jqp = Di Tj11...jqp dxi , recall Result 3.4.
Heuristics
The conceptual advantage of a (covariant) differential over a (covariant) derivative is that the former is applicable
to vectors known only along a curve, whereas the latter requires us to specify a vector field on a full neighbour-
hood, as the following definition illustrates.
3.11 Curvature 55
Definition 3.16 A geodesic is a curve with “covariantly constant tangent”, or a curve obtained by “parallel
transport” of its tangent vector at any fiducial point. In other words, if v(t) = ẋi (t)∂i for t ∈ I = (t0 , t1 ), then
∇v(t) = 0 for all t ∈ I.
Remark
• It is tacitly understood that ∂i ∈ TMx(t) indicates the coordinate basis vector at point x(t).
• One often identifies v(x(t)) with v(t). It should be clear from the context which of these prototypes is
applicable.
Observation
In terms of components we have Dv k (t)/dt = 0, or ẍk + Γkij ẋi ẋj = 0.
Terminology
The latter second order ordinary differential equation is known as the geodesic equation.
Heuristics
• A geodesic generalizes the concept of a straight line, and can in general be seen as the “straightest line”
between any two sufficiently close points on the geodesic.
• This should not be confused with the “shortest line”, for this would require the additional notion of a
“length” or “distance” on the manifold (i.e. a Riemannian metric).
3.11. Curvature
Definition 3.17 For v, w ∈ TM the (antisymmetric) curvature operator, R(v, w) : TM → TM : u 7→ R(v, w)u, is
defined as
R(v, w) = [∇v , ∇w ] − ∇[v,w] .
Observation
Alternatively, in order to obtain a scalar valued mapping, we would need a covector field bz ∈ T∗ M besides the
three vector field arguments u, v, w ∈ TM involved in Definition 3.17. This naturally leads to the definition of a
fourth order mixed tensor field.
z ∈ T∗ M, the Riemann curvature tensor Riemann ∈ T∗ M⊗TM⊗TM⊗TM

Definition 3.18 For u, v, w ∈ TM, b
is defined as
z , R(v, w)ui .
z, u, v, w) = hb
Riemann(b
Observation
• One should first verify that Definition 3.17 does not require any derivatives of u!
z = zρ dxρ , u = uσ ∂σ , v = v µ ∂µ ,
• In order to find the components of the Riemann curvature tensor, write b
ν
z, u, v, w) = Rσµν zρ uσ v µ wν , in which
w = w ∂ν . Then, using previous result, you find R(b ρ
ρ
Rσµν = hdxρ , R(∂µ , ∂ν )∂σ i = ∂µ Γρσν − ∂ν Γρσµ + Γρλµ Γλσν − Γρλν Γλσµ .
ρ
• Instead of Rσµν the holor is often written as Rρ σµν in order to stress that the first slot of the tensor is
intended for the covector argument.
• Note that R(∂µ , ∂ν ) = [∇∂µ , ∇∂ν ] − ∇[∂µ ,∂ν ] = [∇∂µ , ∇∂ν ], i.e. the curvature operator and, a fortiori, the
Riemann curvature tensor apparently capture the intrinsic non-commutativity of covariant derivatives.
• Thus the ‘second order’ nature of the curvature operator pertains to the geometry of the manifold, not the
differential structure of the vector field it operates on.
• In the case of a Riemannian manifold (with the induced Levi-Civita connection) this is clear from the
ρ
appearance of (up to) second order partial derivatives of the Gram matrix gαβ in Rσµν .
Definition 3.19 The Ricci curvature tensor and the Ricci curvature scalar are defined as
Ric(v, w) = Rµν v µ wν ,
R = g µν Rµν ,
ρ
with v, w ∈ TM, in which the Ricci holor is defined as Rµν = Rµρν .
3.12. Push-Forward and Pull-Back
Ansatz
• A smooth mapping φ : M → N : x 7→ y = φ(x) between manifolds.
• A smooth scalar field f : N → R : y 7→ z = f (y).

• A smooth vector field v ∈ TM.
• A smooth covector field ν ∈ T∗ N.
Definition 3.20
• The pull-back φ∗ f ∈ C ∞ (M) of f ∈ C ∞ (N) under φ is defined such that φ∗ f = f ◦ φ.

• The push-forward φ∗ v ∈ TN of v ∈ TM under φ is defined such that φ∗ v(f ) = v(φ∗ f ).
• The pull-back φ∗ ν ∈ T∗ M of ν ∈ T∗ N under φ is defined such that hφ∗ ν, viM = hν, φ∗ viN .
Terminology
The terminology reflects the “direction” of the mappings involved between (tangent or cotangent bundles over)
the respective manifolds, cf. the diagram (3.8).
3.13 Examples 57
Remark
• Note that the dimensions of the manifolds M and N need not be the same.
Observation
• By decomposing φ∗ v = v µ Sµa ∂y∂a ∈ TN it follows that the components of the push-forward are given by
∂y a
the Jacobian matrix of the mapping: Sµa = . (Note: µ = 1, . . . , m = dim M, a = 1, . . . , n = dim N.)
∂xµ
• By the same token, decomposing φ∗ ν = νa Sµa dxµ ∈ T∗ M, it follows that the components of the pull-back
are given by the same Jacobian.
φ∗
TM −−−→ TN
 

πy
π
y
M −−−→ N (3.8)
φ
x x

π
π

φ∗
T∗ M ←−−− T∗ N
3.13. Examples
3.13.1. Polar Coordinates in the Euclidean Plane
Example
Starting out from a Cartesian coordinate basis at an arbitrary point of the Euclidean plane E (i.e. Euclidean
2-space),
1 0
ex = , ey = ,
0 1
we obtain the local polar coordinate basis by the following transformation:
∂xj ∂xj
ei = ∂ i = ∂j = ej .
∂xi ∂xi
Here we identify x = (x, y), x = (r, φ), with

x(r, φ) = r cos φ
y(r, φ) = r sin φ .
Recall that the Jacobian matrix and its inverse are given by
∂x` ∂xk
Ti` = resp. S`k = ,
∂xi ∂x`
or, in explicit matrix notation (upper/lower index = row/column index):
 x 
p −y
cos φ −r sin φ x + y2
2
Ti` = = ,
 
sin φ r cos φ y
p x
x2 + y 2
!
 x y 
cos φ sin φ p p
S`k = sin φ cos φ =
 x + y2
2 x2 + y 2 
.
− y x
r r − 2
x + y2 2
x +y 2
Consequently, writing “r” and “φ” for index values 1 and 2, respectively,

ex cos φ
er = cos φ sin φ = cos φ ex + sin φ ey = ,
ey sin φ

ex −r sin φ
eφ = −r sin φ r cos φ = −r sin φ ex + r cos φ ey = .
ey r cos φ
Note that the polar basis depends on the base point of the tangent plane, whereas the Cartesian basis is identical
in form at all base points.
3.13 Examples 59
Example
Cartesian coordinates x = (x, y) in the Euclidean plane E induce a local basis for each TEx relative to which
all Γ`nm vanish identically (`, m, n = 1, . . . , 2). In terms of polar coordinates, x = (r, φ), we find, according to
Result 3.2 on page 50:
k
Γij = S`k ∂ j Ti` .
From the previous example it follows that for any (row) index k = 1, 2 and (column) index i = 1, 2,
! !
k
cos φ sin φ ∂

cos φ −r sin φ
0 0
Γir = sin φ cos φ = 1 ,
− ∂r sin φ r cos φ 0
r r r
0 −r
! !
k
cos φ sin φ ∂

cos φ −r sin φ

Γiφ = sin φ cos φ = 1 .
− ∂φ sin φ r cos φ 0
r r r
In other words,
φ 1 φ 1 r k
Γφr = , Γrφ = , Γφφ = −r , all other Γij = 0.
r r
Consequently,
1 1
∇r eφ = eφ , ∇φ er = eφ , ∇φ eφ = −rer , all other ∇ej ei = 0.
r r
This can be immediately confirmed by inspection of the (partial derivatives of the) coordinates of the polar basis
vectors relative to the standard (Cartesian) basis, recall the previous example.
3.13.2. A Helicoidal Extension of the Euclidean Plane
Example
Consider two coordinatizations of 3-space, x = (x, y, φ), and x = (ξ, η, θ), related as follows1 :
 
 x(ξ, η, θ) = ξ cos θ + η sin θ  ξ(x, y, φ) = x cos φ − y sin φ
y(ξ, η, θ) = −ξ sin θ + η cos θ or η(x, y, φ) = x sin φ + y cos φ
φ(ξ, η, θ) = θ , θ(x, y, φ) = φ .
 
The Jacobian matrices are given by

 
` cos θ sin θ −ξ sin θ + η cos θ
∂x
Ti` = =  − sin θ cos θ −ξ cos θ − η sin θ  ,
∂xi 0 0 1
respectively  
k cos φ − sin φ
−x sin φ − y cos φ
∂x
S`k = =  sin φ x cos φ − y sin φ  .
cos φ
∂x`
0 0 1
k
Again, recalling Result 3.2 on page 50: Γij = S`k Tjm Tin Γ`nm + ∂ j Ti` . There are different ways to proceed,

depending on our assumptions about the Γ`nm -symbols in (x, y, φ)-space. Here we pursue the simplest option,
ignoring the remark in footnote 1, and assume Γ`nm = 0, i.e. (x, y, φ) is identified with a Cartesian coordinate
triple in Euclidean 3-space. In this case we may evaluate the Christoffel symbols as in the previous example:
   
cos φ − sin φ −x sin φ − y cos φ cos θ sin θ −ξ sin θ + η cos θ
k ∂ 
Γiξ =  sin φ cos φ x cos φ − y sin φ  − sin θ cos θ −ξ cos θ − η sin θ  ,
∂ξ
0 0 1 0 0 1
   
k ∂
Γiη =  sin φ cos φ x cos φ − y sin φ   − sin θ cos θ −ξ cos θ − η sin θ  ,
∂η
0 0 1 0 0 1
   
k ∂
Γiθ =  sin φ cos φ x cos φ − y sin φ   − sin θ cos θ −ξ cos θ − η sin θ  .
∂θ
0 0 1 0 0 1
A tedious but straightforward computation then reveals that

     
0 0 0 0 0 1 0 1 −ξ
k k k
Γiξ =  0 0 −1  Γiη =  0 0 0  Γiθ =  −1 0 −η  ,
0 0 0 0 0 0 0 0 0
k η ξ η ξ ξ η k k
i.e. all Γij = 0, except Γθξ = −1 , Γθη = 1 , Γξθ = −1 , Γηθ = 1 , Γθθ = −ξ , Γθθ = −η. Note that Γij −Γji = 0,
consistent with the fact that torsion vanishes identically by construction.
1 We interpret this space as an extension of Euclidean 2-space E by attaching to each point a copy of the periodic circle S = {φ ∈
[0, 2π) mod 2π} as an independent dimension. The parametrization can then be seen as a “shift-twist” transformation, whereby the Cartesian
basisvectors ex and ey of the underlying Euclidean 2-space E (at any fiducial point) are rotated in accordance to the periodic translation in
eφ -direction (generating S).
3.13 Examples 61
Example
In the spirit of the previous example we consider a 3-space M, with coordinates (x, y, φ). However, we now
introduce the following Riemannian metric g ∈ T∗ M ⊗ T∗ M:
g(x, y, φ) = dx ⊗ dx + dy ⊗ dy + (x2 + y 2 )dφ ⊗ dφ .
Note that the metric is form invariant under the reparametrization of the previous example:
g(ξ, η, θ) = dξ ⊗ dξ + dη ⊗ dη + (ξ 2 + η 2 )dθ ⊗ dθ .
The induced Levi-Civita connection in (x, y, φ)-coordinates (and therefore also in (ξ, η, θ)-coordinates) follows
from Theorem 3.1, page 54. In matrix notation we have for the metric and its dual, respectively,
   
1 0 0 1 0 0
G= 0 1 0  and G−1 =  0 1 0 .
2 2 2 2
0 0 x +y 0 0 1/(x + y )
The fact that metric and its dual are diagonal greatly simplifies the computation, for we have that
1 kk
Γkij = g (∂i gkj + ∂j gki − ∂k gij ) (no summation over k).
2
Moreover, by virtue of the fact that the coefficients gab are constants for a, b = 1, 2, the above matrix (with row
index i and column index j, and k fixed) will have a null 2 × 2 block in the upper left part:
 
0 0 ∗
Γkij =  0 0 ∗  (row index i, column index j, k fixed.)
∗ ∗ ∗
Finally, the (dual) metric coefficients do not depend on φ. By direct computation, taking these simplifying
remarks into account, we obtain
     
0 0 0 0 0 0 0 0 x
1
Γxij =  0 0 0  Γyij =  0 0 0  Γφij = 2  0 0 y .
x + y2
0 0 −x 0 0 −y x y 0
Note that Γkij − Γkji = 0, consistent with the fact that torsion always vanishes in the case of the Levi-Civita
connection.
Example
Rotational invariance in the x–y plane of the previous example suggests the use of cylinder coordinates x =
(r, ψ, θ) instead of x = (x, y, φ): 
 x(r, ψ, θ) = r cos ψ
y(r, ψ, θ) = r sin ψ
φ(r, ψ, θ) = θ.

To this end we once again invoke Result 3.2, page 50:

k
Γij = S`k Tjm Tin Γ`nm + ∂ j Ti` .

The Jacobian matrices are given by

 
` cos ψ −r sin ψ 0
∂x
Ti` = =  sin ψ r cos ψ 0 ,
∂xi 0 0 1
respectively
x/(x2 + y 2 ) y/(x2 + y 2 )
 
k 0
k ∂x
S` = =  −y/(x2 + y 2 ) x/(x2 + y 2 ) 0 .
∂x`
0 0 1
Bibliography
[1] R. Abraham, J. E. Marsden, and T. Ratiu. Manifolds, Tensor Analysis, and Applications, volume 75 of
Applied Mathematical Sciences. Springer-Verlag, New York, second edition, 1988.
[2] D. Bao, S.-S. Chern, and Z. Shen. An Introduction to Riemann-Finsler Geometry, volume 2000 of Graduate
Texts in Mathematics. Springer-Verlag, New York, 2000.
[3] M. Berger. A Panoramic View of Riemannian Geometry. Springer-Verlag, Berlin, 2003.
[4] R. L. Bishop and S. I. Goldberg. Tensor Analysis on Manifolds. Dover Publications, Inc., New York, 1980.
Originally published by The Macmillan Company in 1968.
[5] M. P. do Carmo. Differential Geometry of Curves and Surfaces. Mathematics: Theory & Applications.
Prentice-Hall, Englewood Cliffs, New Jersey, 1976.
[6] M. P. do Carmo. Riemannian Geometry. Mathematics: Theory & Applications. Birkhäuser, Boston, second
edition, 1993.
[7] E. Cartan. Leçons sur la Géométrie des Espaces de Riemann. Gauthiers-Villars, Paris, second edition, 1963.
[8] Sir Thomas L. Heath. The Thirteen Books of Euclid’s Elements, volume 1 (Books I and II). Dover Publica-
tions, Inc., New York, second edition, 1956.
[9] Sir Thomas L. Heath. The Thirteen Books of Euclid’s Elements, volume 2 (Books III–IX). Dover Publica-
[10] Sir Thomas L. Heath. The Thirteen Books of Euclid’s Elements, volume 3 (Books X–XIII). Dover Publica-
[11] J. Jost. Riemannian Geometry and Geometric Analysis. Springer-Verlag, Berlin, fourth edition, 2005.
[12] J. J. Koenderink. Solid Shape. MIT Press, Cambridge, 1990.
[13] A. P. Lightman, W. H. Press, R. H. Price, and S. A. Teukolsky. Problem Book in Relativity and Gravitation.
Princeton University Press, Princeton, 1975.
[14] D. Lovelock and H. Rund, editors. Tensors, Differential Forms, and Variational Principles. Dover Publica-
tions, Inc., Mineola, N.Y., 1988.
[15] C. W. Misner, K. S. Thorne, and J. A. Wheeler. Gravitation. Freeman, San Francisco, 1973.
[16] H. Rund. The Hamilton-Jacobi Theory in the Calculus of Variations. Robert E. Krieger Publishing Com-
pany, Huntington, N.Y., 1973.
[17] M. Spivak. Calculus on Manifolds. W. A. Benjamin, New York, 1965.
[18] M. Spivak. Differential Geometry, volume 1–5. Publish or Perish, Berkeley, 1975.

Tensor Calculus and Differential Geometry: Course Notes

Uploaded by

Copyright:

Available Formats

Tensor Calculus and Differential Geometry: Course Notes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tensor Calculus and Differential Geometry: Course Notes

Uploaded by

Copyright:

Available Formats

Course Notes

Tensor Calculus and Differential Geometry

March 21, 2019

1 Prerequisites from Linear Algebra 3

2.1 Vector Spaces and Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Dual Vector Spaces and Dual Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.3 The Kronecker Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.4 Inner Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.5 Reciprocal Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.6 Bases, Dual Bases, Reciprocal Bases: Mutual Relations . . . . . . . . . . . . . . . . . . . . . . 16

2.7 Examples of Vectors and Covectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.8.1 Tensors in all Generality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.8.2 Tensors Subject to Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.8.3 Symmetry and Antisymmetry Preserving Product Operators . . . . . . . . . . . . . . . 24

2.8.4 Vector Spaces with an Oriented Volume . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.8.5 Tensors on an Inner Product Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.8.6 Tensor Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

2.8.6.1 “Absolute Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

2.8.6.2 “Relative Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.8.6.3 “Pseudo Tensors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2.9 The Hodge Star Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.1 Euclidean Space: Cartesian and Curvilinear Coordinates . . . . . . . . . . . . . . . . . . . . . 45

3.2 Differentiable Manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

3.3 Tangent Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.4 Tangent and Cotangent Bundle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.5 Exterior Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.6 Affine Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.7 Lie Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.9 Levi-Civita Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.12 Push-Forward and Pull-Back . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.13.1 Polar Coordinates in the Euclidean Plane . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.13.2 A Helicoidal Extension of the Euclidean Plane . . . . . . . . . . . . . . . . . . . . . . 60

Eindhoven, March 21, 2019.

• Linear mappings: A , B, etc.

• Covector components: ai , bi , etc.

The completely antisymmetric symbol in n dimensions is defined as follows:

• Exactly n! of the nn index realizations in [i1 , . . . , in ] yield nontrivial results.

product with row labels ordered sign of column label permutation

Products of completely antisymmetric symbols can be expressed in terms of determinants:

and, more generally,

The adjugate matrix of A ∈ Mn is a synonymous term for ÃT ∈ Mn .

d det A ∂ det A dAij dAij

• If A is invertible, then ÃT = adjA = det A A−1 .

2.1. Vector Spaces and Bases

Definition 2.1 The Kronecker symbol, or Kronecker-δ, is defined as

Theorem 2.1 Let x = xi ei ∈ V and fj = Aij ei define a change of basis. If xi ei = y i fi , then

in which fj ≡ ∂ j ≡ ∂/∂xj , and

• Basis vectors are labeled with lower indices.

2.2. Dual Vector Spaces and Dual Bases

f̂ (λx + µy) = λf̂ (x) + µf̂ (y) .

(λf̂ + µĝ)(x) = λf̂ (x) + µĝ(x)

for all x ∈ V , recall chapter 1.

êi (ej ) = δji .

Proof of Result 2.1. Let f̂ ∈ V ∗ .

in which Akj Bki = δji . Consequently, if x̂ = xi êi = yi f̂ i ∈ V ∗ , then

xi = Bij yj or, equivalently, yi = Aji xj .

Proof of Theorem 2.2.

• Thus if we define xi êi = yi f̂ i , then yj f̂ j = yj Bij êi , whence xi = Bij yj .