AE 738 Tensors

AE 738 : Tensors for Engineers
Contents
1 Review of index notation and mapping 5
1.1 Summation convention . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Kronecker delta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Dot product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Permutation symbol . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 Cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Differentiation notation . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.7 Mappings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.7.1 Injective or one-to-one mapping . . . . . . . . . . . . . . . . 11
1.7.2 Surjective or onto mapping . . . . . . . . . . . . . . . . . . . 11
1.7.3 Bijective or one-to-one correspondence mapping . . . . . . . 12
1.7.4 Composition of two mappings . . . . . . . . . . . . . . . . . 13
1.8 Homomorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.9 Isomorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2 Vector space 17
2.1 Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Vector space or Linear space . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Linear combination . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4 Linearly independent set . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5 Linear span . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.6 Basis, components and dimension . . . . . . . . . . . . . . . . . . . 24
2.6.1 Standard basis . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.7 Vector subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.8 Linear sum of two subspaces . . . . . . . . . . . . . . . . . . . . . . 30
2.9 Direct sum of two subspaces . . . . . . . . . . . . . . . . . . . . . . 32
3 Homomorphism of vector spaces or linear transformation 36

3.0.1 Linear operator . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.0.2 Identity operator . . . . . . . . . . . . . . . . . . . . . . . . 38
3.0.3 Zero transformation . . . . . . . . . . . . . . . . . . . . . . . 38
1
CONTENTS
3.1 Range space R(T ) . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

3.2 Null space N (T ) or ker(T ) . . . . . . . . . . . . . . . . . . . . . . . 39
3.3 Rank and nullity of a homomorphism . . . . . . . . . . . . . . . . . 40
3.4 Singular homomorphism . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5 Nonsingular homomorphism . . . . . . . . . . . . . . . . . . . . . . 44
3.6 Isomorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.7 Invertible T : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.8 Representation of homomorphism or LT by a matrix . . . . . . . . 46
4 Homomorphism as vector space 50

4.1 Product of homomorphism . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Algebra of linear operators . . . . . . . . . . . . . . . . . . . . . . . 54
5 Linear functional or linear form 56

5.1 Dual space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2 Second dual space . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.3 Transposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6 Tensor product 61
6.1 Bilinear functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.2 Tensor product as homomorphism . . . . . . . . . . . . . . . . . . . 65
6.3 Matrix representation of tensor products . . . . . . . . . . . . . . . 68
6.4 Tensor spaces: a general multilinear function . . . . . . . . . . . . 69
6.5 Tensor multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.6 Contraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
7 Inner product space 74

7.1 Inner product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.2 Reciprocal basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
7.2.1 Length of a vector . . . . . . . . . . . . . . . . . . . . . . . 77
7.2.2 Angle between two vectors . . . . . . . . . . . . . . . . . . . 78
7.2.3 Differential arc length . . . . . . . . . . . . . . . . . . . . . 78
7.2.4 Transposition of a tensor in an inner product space . . . . . 78
7.3 Physical interpretation of gij . . . . . . . . . . . . . . . . . . . . . . 79
7.3.1 Physical components of tensors . . . . . . . . . . . . . . . . 85
7.4 Affine space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.5 Some applications to the continuum mechanics . . . . . . . . . . . . 86
7.6 Orthonormal set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.7 Gram-Schmidt orthonormalization . . . . . . . . . . . . . . . . . . . 90
7.8 Orthogonal complement . . . . . . . . . . . . . . . . . . . . . . . . 92
AE 738 2 Dr. Krishnendu Haldar

CONTENTS
8 Symmetric and skew/anti-symmetric tensor 95

8.1 Permutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8.2 Symmetric tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8.3 Skew-symmetric tensor . . . . . . . . . . . . . . . . . . . . . . . . . 98
8.4 Exterior algebra (popular name) or Grassmann algebra . . . . . . . 100
8.4.1 Exterior or wedge product . . . . . . . . . . . . . . . . . . . 101
8.5 Basis of Λk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.6 Hodge duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8.7 Determinant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.8 Trace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.9 Interior product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8.10 Hodge duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
9 Invariants in 3D 113
9.1 Cofactor tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10 Orthogonal tensors 116
11 Skew-symmetric tensors 118
12 Spectral decomposition of a second order tensor 119

12.1 Spectral decomposition of a symmetric second order tensor . . . . . 121
12.1.1 Two repeated eigenvalues . . . . . . . . . . . . . . . . . . . 122
12.1.2 Three repeated eigenvalues . . . . . . . . . . . . . . . . . . . 123
12.2 Symmetric positive definite . . . . . . . . . . . . . . . . . . . . . . . 123
12.3 Cayley-Hamilton theorem [3D] . . . . . . . . . . . . . . . . . . . . . 124
13 Representation of a rigid rotation tensor 125
14 Representation of a skew-symmetric tensor 126
15 Tensor calculus 128

15.1 Normed space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
15.2 Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
15.3 Partial derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
15.4 Product rule for bilinear mapping . . . . . . . . . . . . . . . . . . . 134
15.5 Chain rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
15.6 Gradient and divergence . . . . . . . . . . . . . . . . . . . . . . . . 137
15.6.1 Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
15.6.2 Divergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
15.6.3 Curl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
15.7 Divergence theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

CONTENTS
15.8 Product rule for any mapping . . . . . . . . . . . . . . . . . . . . . 142
16 Covariant derivative in Riemannian space 142

16.0.1 Christoffel symbols of 2nd kind . . . . . . . . . . . . . . . . . 142
16.0.2 Christoffel symbols of 1st kind . . . . . . . . . . . . . . . . . 143
16.0.3 Symmetry properties of Christoffel symbols . . . . . . . . . 143
16.1 Ricci identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
16.2 Gradient and divergence in Riemannian space . . . . . . . . . . . . 144
16.2.1 Gradient of vector-valued scalar function . . . . . . . . . . . 144
16.2.2 Gradient of vector-valued vector mapping . . . . . . . . . . 145
16.2.3 Gradient of vector-valued tensor mapping . . . . . . . . . . 147
16.2.4 Divergence of vector-valued vector mapping . . . . . . . . . 148
16.2.5 Divergence of vector-valued tensor mapping . . . . . . . . . 149

1 Review of index notation and mapping
Consider the following sets of numbers:
a11 , a22 , a12 , a21
or
a11 , a22 , a12 , a21
or
a11 , a22 , a12 , a22
In the index notation, they are represented by: aij , aij and aij respectively, where
each of the indices (i, j) takes values 1 and 2.
1.1 Summation convention

Let n quantities be denoted by x1 , x2 , x3 , ..., xn . Consider the expression:
n
X
ai xi = a1 x1 + a2 x2 + ... + an xn
i
This expression will be written ai xi by introducing the summation convention of

EINSTEIN. When an index appears twice in a term, a summation is implied. The
range of the summation is known from the context. Note carefully that:
1. The convention does not apply to numerical indices. For instant a2 x2 stands
for single term.
2. The repeated index may be replaced by any other index. For instance ai xi =
at xt . For this reason, the repeated index is called a dummy index.
3. If an index is not dummy, it is called free index. Thus in the expression aik xi ,
k is the free index.
Some examples:
1. If n = 3, ak xk = a1 x1 + a2 x2 + a3 x3 .
∂ϕ ∂ϕ ∂ϕ
2. If ϕ is a function of x1 , x2 , x3 , ..., xn , then dϕ = ∂x1
dx1 + ∂x2
dx2 +...+ ∂xn
dxn =
∂ϕ
∂xi
dxi

1.2 Kronecker delta
3. Let yi = αit xt and zi = βit yt . We can rewrite yt = αtk xk and now we can
substitute back as zi = βit (αtk xk ) = [βit αtk ]xk = γik xk . Here γik = βit αtk ,
where t is the dummy index and i, k are the free index.
4. Akmn Bkn
l
= (A1mn B1n
l
) + (A2mn B2n
l
) + ... + (AN l 1 l 1 l
mn BN n )=(Am1 B11 + Am2 B12 +
..) + (A2m1 B21
l
+ A2m2 B22
l
+ ...) + ...
5. Consider the summation (a1j xj )2 + (a2j xj )2 + ... + (anj xj )2 . This expression
is written as (ais xs )(ait xt ) or ais ait xs xt .
1.2 Kronecker delta

This is defied as:
(
1 if i = j
δij = (1)
0 if i ̸= j
Note that,
δ1j xj = δ11 x1 + δ12 x2 + ... + δ1n xn = x1

δ2j xj = δ21 x1 + δ22 x2 + ... + δ2n xn = x2
..
.
δij xj = δi1 x1 + δi2 x2 + ... + δin xn = xi
(2)
Thus if we multiply xk by δik , we simply replace the index k of xk by i. Moreover,
δii = δ11 + δ22 + δ33 = 3 (3)
1.3 Dot product

If the set {ê1 , ê2 , ê3 } contains a set of orthonormal unit vectors, then
êi · êj = δij (4)
For a Cartesian system, you can think as

     
1 0 0
ê1 = 0
  ê1 = 1
 ê3 = 0

0 0 1
Let a = ai êi and b = bj êj
a · b = (ai êi ) · (bj êj ) = ai bj (êi · êj ) = ai bj δij = ai bi (5)

1.4 Permutation symbol
Figure 1: Even(left) and odd(right) permutations.
1.4 Permutation symbol


1
 if i, j, k form an even permutation
ϵijk = −1 if i, j, k form an odd permutation (6)

0 if two or more indices are equal

and we can write
ei · (ej × ek ) = ϵijk
1.5 Cross product
ê1 ê2 ê3

a×b = a1 a2 a3
b1 b 2 b3
= (a2 b3 − a3 b2 )ê1 + (a3 b1 − a1 b3 )ê2 + (a1 b2 − a2 b1 )ê3
= ϵijk aj bk êi (7)
Basically we represents a = aj êj and b = bk êk . So we can also write the compo-
nent form as:
(a × b)i = ϵijk aj bk (8)
If we replace a and b, then
b × a = ϵikj bk aj êi = −ϵijk aj bk êi = −a × b

1.5 Cross product
Let c = ar êr , then

c · (a × b) = ϵijk cr aj bk (êi · êr ) = ϵijk cr aj bk δir = ϵijk ci aj bk
 
c1 c2 c3
= det a1 a2
 a3 
b1 b2 b3
 
c · e1 c · e2 c · e3
= det a · e1
 a · e2 a · e3 
b · e1 b · e2 b · e3
If we put c = ei , b = ek , and a = ej , then
   
ei · e1 ei · e2 ei · e3 δi1 δi2 δi3
ei · (ej × ek ) = ϵijk = det  ej · e1 ej · e2 ej · e3  = det  δj1 δj2 δj3  .
ek · e1 ek · e2 ek · e3 δk1 δk2 δk3
So we can write,
   
δi1 δi2 δi3 δp1 δp2 δp3
ϵijk ϵpqr = det δj1 δj2
 δj3 det δq1
  δq2 δq3 
δk1 δk2 δk3 δr1 δr2 δr3
   
δi1 δi2 δi3 δp1 δq1 δr1
= det δj1 δj2
 δj3 det δp2
  δq2 δr2  ∵ det[A] det[B] = det[A] det[B]t
δk1 δk2 δk3 δp3 δq3 δr3
  
δi1 δi2 δi3 δp1 δq1 δr1
= det δj1 δj2
 δj3  δp2 δq2 δr2  ∵ det[A] det[B] = det([A][B]t )
δk1 δk2 δk3 δp3 δq3 δr3
   
δim δpm δim δqm δim δrm δip δiq δir
= det  δjm δpm δjm δqm δjm δrm  = det  δjp δjq δjr 
δkm δpm δkm δqm δkm δrm δkp δkq δkr
We are now interested in
   
δii δiq δir 3 δiq δir
ϵijk ϵiqr = det  δji δjq δjr  = det  δij δjq δjr 
δki δkq δkr δik δkq δkr
= 3(δjq δkr − δjr δkq ) + δiq (δjr δik − δij δkr ) + δir (δij δkq − δjq δik )
= 3(δjq δkr − δjr δkq ) + (δjr δkq − δjq δkr ) + (δjr δkq − δjq δkr )
= (δjq δkr − δjr δkq )
Therefore, the Epsilon-delta identity reads
ϵijk ϵiqr = δjq δkr − δjr δkq (9)

1.5 Cross product
Example 1 Show that
a × b × c = (a · c)b − (a · b)c
Figure 2: Even(left) and odd(right) permutations with i, j, and k.
Ans: Consider
a = ai ei , b = bj ej , c = ck ek
Now
b × c = ϵmjk bj ck em = Am em .
Then,
a×b×c = ϵrim ai Am er
= ϵrim ai (ϵmjk bj ck )er
= (ϵmri ϵmjk )ai bj ck er
= (δrj δik − δrk δij )ai bj ck er
= δrj δik ai bj ck er − δrk δij ai bj ck er
= (δik ai )(δrj bj )ck er − (δij ai )bj ck (δrk er )
= (ak ck )(br er ) − (aj bj )(cr er )
= (a · c)b − (a · b)c
Example 2 Show that
(a × b) · (a × b) = (a · a)(b · b) − (a · b)2

1.6 Differentiation notation
Consider,
a × b = ϵmij ai bj em , a × b = ϵpqr aq br ep .
Then
(a × b) · (a × b) = (ϵmij ai bj em ) · (ϵpqr aq br ep )
= ϵmij ϵpqr ai bj aq br (em · ep )
= ϵmij ϵpqr ai bj aq br (δmp )
= ϵmij ϵmqr ai bj aq br
= (δiq δjr − δir δjq ) ai bj aq br
= δiq δjr ai bj aq br − δir δjq ai bj aq br
= aq br aq br − ar bq aq br
= (aq aq )(br br ) − (ar br )(aq bq )
= (a · a)(b · b) − (a · b)2
1.6 Differentiation notation

Let
a = a(x1 , x2 , x3 )
ai = ai (x1 , x2 , x3 )
aij = aij (x1 , x2 , x3 )
Then
∂a
a, i = = ∂i a
∂xi
∂ai
ai, j = = ∂ j ai
∂xj
∂aij
aij, k = = ∂k aij (10)
∂xk
Example 3 Show that
∇ × ϕu = ∇ϕ × u + ϕ∇ × u
where ∇ = e1 ∂x∂ 1 + e2 ∂x∂ 2 + e3 ∂x∂ 3 = ei ∂i , ϕ(x1 , x2 , x3 ) is a scalar function, and

u(x1 , x2 , x3 ) is a vector function.

1.7 Mappings
∇ × ϕu = ej ∂j × ϕuk ek = ϵijk (∂j (ϕuk ))ei

= ϵijk ((∂j ϕ)uk )ei + (ϕ(∂j uk ))ei
= ϵijk (∂j ϕ)uk ei + ϕϵijk (∂j uk )ei
= ∇ϕ × u + ϕ∇ × u (11)
1.7 Mappings
1.7.1 Injective or one-to-one mapping
Definition 1.1 Let X and Y be the two sets. Then a mapping
T :X →Y
is called injective or one-to-one mapping1 if
1. x1 ̸= x2 =⇒ T x1 ̸= T x2 or equivalently T x1 = T x2 =⇒ x1 = x2
2. the domain2 of the map D(T ) = X
3. the image of the map I(T ) ⊂ Y
where, x1 , x2 ∈ X , and T x1 , T x2 ∈ Y. We will use the notation T x instead of
T (x), and T x reads as ‘ T operates on x’.
In the other words, an injective mapping means that for every input, there is a
unique output. An injective mapping is shown in Fig. 3.
1.7.2 Surjective or onto mapping

T :X →Y
is called surjective or onto mapping if
1. the domain of the map D(T ) = X
2. the image of the map I(T ) = Y
Every element in Y is the image of at least one element in the domain. All the
elements in the Y, as well as in the domain X , are used. However, the mapping
may not be unique. An injective mapping is shown in Fig. 4.
1
The terms ‘mapping’ and ‘function’ are technically synonymous. We generally use ‘mapping’
when the sets could be anything. The terminology ‘function’ is followed when Y = R or C.
However, this is more like a personal preference rather than a convention.
2
The ‘domain’ of a mapping is the set into which all the inputs of the mapping are considered.
The image of a mapping is the set of all output values it may produce.

1.7 Mappings
a y
b z
c u
d w
Figure 3: An injective mapping T : X → Y.
c x
d y
e z
Figure 4: A surjective mapping T : X → Y.
1.7.3 Bijective or one-to-one correspondence mapping

T :X →Y
is called bijective or one to one mapping if the mapping is injective and surjective,
i.e.,
1. x1 ̸= x2 =⇒ T x1 ̸= T x2 or equivalently T x1 = T x2 =⇒ x1 = x2
2. the domain of the map D(T ) = X
3. the image of the map I(T ) = Y
All the element are used in Y and there is one-one correspondence with the ele-
ments in X . Also, elements in X are equal in number of the elements in Y. An
injective mapping is shown in Fig. 5.

1.7 Mappings
a x
b y
c z
d z
Figure 5: A bijective mapping T : X → Y.
1.7.4 Composition of two mappings
Figure 6: Composition of two mappings. Read it ‘ T2 composed with T1 ’.
Definition 1.4 Let X , Y and Z be the three sets. Consider the following map-
pings
T1 : X → Y, T2 : Y → Z
such that D(T2 ) = I(T1 ), then the composition mapping T2 ◦ T1 is defined as
(T2 ◦ T1 ) : X → Z
such that
(T2 ◦ T1 )x = T2 (T1 x) ∈ Z, ∀x ∈ X
Sometimes, we simply write T2 ◦ T1 ≡ T2 T1 . For example, if T1 x = 4x + 3 = y ∈ Y,

and T2 y = y 3 ∈ Z, then T2 (T1 x) = (4x + 3)3 ∈ Z.

1.8 Homomorphism
1.8 Homomorphism
Figure 7: Homeomorphism between two sets.
Let X and Y be the two sets, equipped with some operations 3 . Then X and Y
are known as algebraic structures.
Definition 1.5 A homomorphism is a map between two algebraic structures that

preserves the operations of the structures (Fig. 7).
Consider the mapping
T : X → Y,
and let us denote ‘∗’ as a list of generic operations in X , and ‘⋆’ as a list of generic
operations in Y. Then,
T (x1 ∗ x2 ) = T x1 ⋆ T x2 ∈ Y, ∀x1 , x2 ∈ X .
For example, If ∗ = ⋆ = + is an addition operation, then
T (x1 + x2 ) = T x1 + T x2 , ∀x1 , x2 ∈ X .
This means, we first apply the operation on some elements in X and then map the
resulting element into Y. It will be the same if we map those elements first into
Y and then perform the operations in Y.
Example 4 Let X and Y are equipped with the binary operations + and ·, i.e.,
∗ = ⋆ = {+, ·}, and consider the map

x 0
T : X → Y, such that T x = ∈ Y, x ∈ X .
0 x
3
An operation takes input value/s to a well-defined output value, e.g., binary operations like
+, ·(multiplication), etc.

1.8 Homomorphism
So Y is a set of 2 × 2 matrices as per the given mapping. If x1 , x2 ∈ X , then

x1 + x2 0 x1 0 x2 0
T (x1 + x2 ) = = + = T x1 + T x2
0 x1 + x2 0 x1 0 x2
and

x1 · x 2 0 x1 0 x2 0
T (x1 · x2 ) = = · = T (x1 ) · T (x2 )
0 x1 · x2 0 x1 0 x2
So the mapping preserves the operations and structures, and so the map is homo-
morphism. We say that X and Y are homomorphic to each other.
Example 5 Consider ∗ = + in X and ⋆ = · in Y. Then the exponential function

exp(x) satisfies exp(x1 + x2 ) = exp(x1 ) · exp(x2 ), i.e, an additive operation in
X and a multiplicative operation on Y. As we are not familiar with the group
structures, we can exclude such scenarios at the present.
Homomorphisms give us ways to relate different algebraic objects. Often one
is more interested in how objects relate to each other than how they actually are.
In the algebraic setting, homomorphisms give a tool for talking about that. Many
algebraic objects may appear different, e.g., the group symmetries of the triangle,
the group of permutations of three letters, but turn out to be the same as they
are homomorphic. If there is a homorphism between two sets, they are actually
identical as far as their algebraic structure is concerned. Consider the symmetry
Figure 8: Rotational symmetry transformation of a square.
operation C4 , generated by a single discrete rotation 2π

4
with respect to the axis
passing through the centroid of a square abcd (Fig. 8). Then the following set with
multiplication operation
X = {E, C4 , (C4 )2 , (C4 )3 },
represents the operations in Fig. 8. The multiplication table is given in Table 1.

Let us consider another set
√
Y = {1, i, −1, −i}, (i = −1)

1.9 Isomorphism
E C4 (C4 )2 (C4 )3
E E C4 (C4 )2 (C4 )3
(C4 )3 (C4 )3 E C4 (C4 )2
(C4 )2 (C4 )2 (C4 )3 E C4
C4 C4 (C4 )2 (C4 )3 E
Table 1: The multiplication table for X
1 i −1 −i
1 1 i −1 −i
−i −i 1 i −1
−1 −1 −i 1 i
i i −1 −i 1
Table 2: The multiplication table for Y
and the multiplication table of this set is given in Table. 2.

It is possible that the map T : X → Y is a homomorphism if we make the
following correspondence
T E = 1, T C4 = i, T (C4 )2 = −1, T (C4 )3 = −i
as they have similar multiplication tables, i.e., having the same rearrangement.
Note that T (C4 · (C4 )2 ) = T (C4 )3 = −i and T (C4 ) · T (C4 )2 = i · (−1) = −i. It
is then satisfying T (C4 · (C4 )2 ) = T (C4 ) · T (C4 )2 . You can also check that this
relation is satisfied for the other elements. So, T is a homomorphism.
Very often, several groups, which arise in different contexts in everyday life
and consequently with different physical meanings attached to the elements, are
isomorphic to one abstract group, whose properties can then be analyzed once and
for all.
1.9 Isomorphism
A bijective homomorphism is known as isomorphism.

2 Vector space
2.1 Field
Figure 9: Essential elements of a field and the closer properties under addition and
multiplication. The boundary of the set in the diagram is just to schematically
represent a set.
Suppose F is equipped with two binary operations called addition and multi-
plication and denoted by + and ·, respectively. This means ∀a, b ∈ F, we have
a + b ∈ F and a · b ∈ F. This algebraic structure (F, +, ·) is called a field if the
following postulates are satisfied :
1. Closed under addition: a + b ∈ F, ∀a, b ∈ F
2. Addition is commutative: a + b = b + a, ∀a, b ∈ F.
3. Addition is associative: (a + b) + c = a + (b + c), ∀a, b, c ∈ F.
4. ∃ an element 0 ∈ F (called zero) such that: a + 0 = a ∀a ∈ F.
5. To each element a ∈ F, ∃ −a ∈ F such that: a + (−a) = 0 ∀a ∈ F, and

a ̸= 0.
6. Closed under multiplication: a · b ∈ F, ∀a, b ∈ F
7. Multiplication is commutative: a · b = b · a ∀a, b ∈ F.
8. Multiplication is associative: (a · b) · c = a · (b · c) ∀a, b, c ∈ F.
9. ∃ an non-zero element 1 ∈ F (called one/unity) such that: a · 1 = a ∀a ∈ F,

and a ̸= 0.
10. To each element a ∈ F, ∃ a−1 ∈ F such that: a · a−1 = 1 ∀a ∈ F, and a ̸= 0.

2.2 Vector space or Linear space
11. Multiplication is distributive with respect to addition: a · (b + c) = a · b + a · c

∀a, b, c ∈ F.
We simply denote (F, +, ·) as F. The elements of F are called scalars.
Example 6 The set of natural numbers N, integers I, and rational numbers Q

are examples of fields as they obey all the eleven rules for a field. However, in the
course, we mostly consider field as the set of real numbers R.

Suppose V is a non-empty set equipped with a field F, is called a linear or vector
space if the following conditions are fulfilled:
(A1) Closed under addition: x + y ∈ V, ∀x, y ∈ V
(A2) Addition is commutative: x + y = y + x, ∀x, y ∈ V.
(A3) Addition is associative: (x + y) + z = x + (y + z), ∀x, y, z ∈ V.
(A4) ∃ an element 0 ∈ V (called zero) such that: x + 0 = x ∀x ∈ V.
(A5) To each element x ∈ V, ∃ −x ∈ V such that: x + (−x) = 0 ∀x ∈ V, and

x ̸= 0.
An operation is called vector addition if all the above operations hold. Now an
operation called scalar multiplication if a scalar (or a number) a ∈ F can be
combined with every element x ∈ V to give an element ax ∈ V such that the
following rules hold:
(B1) Closed under scalar multiplication: ax ∈ V, ∀a ∈ F and x ∈ V
(B2) Multiplication by scalar is associative: a(bx) = (ab)x ∀a, b ∈ F and x ∈ V.
(B3) ∃ an non-zero element 1 ∈ F (called one/unity) such that: 1x = x ∀x ∈ V.
(B4) Multiplication by scalar is distributive with respect to vector addition: a(x+

y) = ax + ay ∀a ∈ F and ∀x, y ∈ V.
(B5) Multiplication by scalar is distributive with respect to scalar addition: (a +

b)x = ax + bx ∀a, b ∈ F and ∀x ∈ V.
So a vector space is constructed with the above 5 + 5 = 10 rules.4

4
There should be no confusion about the use of the word vector. Here by vector we do not
mean the vector quantity which we have defined in vector algebra as a directed line segment.

Example 7 Let V = Rm×n , the set of m × n matrices. Then V is a vector space

with the field F = R. Let, x = [A]mn , y = [B]mn , and z = [C]mn ∈ V. So,
1. Closed under addition: [A]mn + [B]mn = [A + B]mn ∈ V Let

   
a11 a12 · · · a1n b11 b12 · · · b1n
[A]mn =  ... .. ..  , [B]mn =  ... .. ..  .
 
. .  . . 
am1 am2 · · · amn m×n bm1 bm2 · · · bmn m×n
where aij , bij ∈ R. Then

 
a11 + b11 a12 + b12 · · · a1n + a1n
[A + B]mn = 
 .. .. .. 
. . . 
am1 + bm1 am2 + bm2 · · · amn + bmn m×n
 
c11 c12 · · · c1n
=  ... .. ..  ∈ V.

. . 
cm1 cm2 · · · cmn m×n
since cij = aij + bij ∈ F as aij , bij ∈ F.
2. Addition is commutative: [A]mn + [B]mn = [B]mn + [A]mn Note that,

 
a11 + b11 a12 + b12 · · · a1n + a1n
[A]mn + [B]mn = [A + B]mn = 
 .. .. .. 
. . . 
am1 + bm1 am2 + bm2 · · · amn + bmn m×n
 
b11 + a11 b12 + a12 · · · b1n + a1n
= 
 .. .. .. 
. . . 
bm1 + am1 bm2 + am2 · · · bmn + amn m×n
= [B + A]mn = [B]mn + [A]mn
since aij + bij = bij + aij ∈ F as aij , bij ∈ F. Similarly, you can verify the
rest of the following relations.
3. Addition is associative: ([A]mn + [B]mn ) + [C]mn = [A]mn + ([B]mn + [C]mn )
4. Existence of a zero element [0]mn ∈ V such that: [A]mn + [0]mn = [A]mn
5. Additive inverse: [A]mn + (−[A]mn ) = [0]mn
6. Closed under scalar multiplication: a[A]mn ∈ V, ∀a ∈ F

7. Multiplication by scalar is associative: a(b[A]mn ) = (ab)[A]mn ∀a, b ∈ F

8. ∃ an non-zero element 1 ∈ F (called one/unity) such that: 1[A]mn = [A]mn
9. Multiplication by scalar is distributive with respect to vector addition: a([A]mn +
[B]mn ) = a[A]mn + a[B]mn , ∀a ∈ F
10. Multiplication by scalar is distributive with respect to scalar addition: (a +
b)[A]mn = a[A]mn + b[A]mn ∀a, b ∈ F
Observe that we cannot multiply two matrices in the vector space! We certainly
need more tools to do so, and we discover those as we progress.
Example 8 Set of all n-tuples of real numbers Rn constitutes a vector space.
Substitute m = 0 in the previous example and immediately get the result as a
special case of Rm×n . These n-tuples of real numbers Rn are the vectors for the
physicists!
Example 9 All polynomials of degree n, Fn , over a field constitute a vector space.
Let Fn (x) denote the set of all polynomials of x of degree n over a field F. Let us
consider the following polynomials:
f (x) = a0 + a1 x + a2 x2 + · · · + an xn , f (x) ∈ Fn (x), ai ∈ F
g(x) = b0 + b1 x + b2 x2 + · · · + bn xn , g(x) ∈ Fn (x), bi ∈ F
h(x) = c0 + c1 x + c2 x2 + · · · + cn xn , h(x) ∈ Fn (x), ci ∈ F
Moreover consider the zero polynomial as
0̂(x) = 0 + 0x + 0x2 + · · · + 0xn , 0̂(x) ∈ Fn (x), 0 ∈ F
Then,
1. f (x) + g(x) ∈ Fn (x). Note that
f (x) + g(x) = (a0 + b0 ) + (a1 + b1 )x + (a2 + b2 )x2 + · · · + (an + bn )xn
= c0 + c1 x + a2 x2 + · · · + cn xn ∈ Fn (x)
since ai + bi = ci ∈ F as ai , bi ∈ F.
2. f (x) + g(x) = g(x) + f (x) as
f (x) + g(x) = (a0 + b0 ) + (a1 + b1 )x + (a2 + b2 )x2 + · · · + (an + bn )xn
= (b0 + a0 ) + (b1 + a1 )x + (b2 + a2 )x2 + · · · + (bn + an )xn
= (b0 + b1 x + b2 x2 + · · · + bn xn ) + (a0 + a1 x + a2 x2 + · · · + an xn )
= g(x) + f (x)
since ai + bi = bi + ai ∈ F as ai , bi ∈ F. Similarly check the remaining points.

2.3 Linear combination
3. f (x) + (g(x) + h(x)) = (f (x) + g(x)) + h(x).
4. f (x) + 0̂(x) = f (x)
5. f (x) + (−f (x)) = 0̂(x)
6. αf (x) ∈ Fn (x), where α ∈ F
7. k1 (k2 f (x)) = (k1 k2 )f (x), k1 , k2 ∈ F.
8. 1f (x) = f (x)
9. k1 (f (x) + g(x)) = k1 f (x) + k1 g(x), k1 ∈ F.
10. (k1 + k2 )f (x) = k1 f (x) + k2 f (x), k1 , k2 ∈ F.
Note that the polynomial are nonlinear with respect to x but the polynomial space
is linear!
2.3 Linear combination

Let x ∈ V and x1 , x2 , · · · , xn ∈ V. Then x is said to be linear combination of
{xi ∈ V, i = 1, n} such that:
x = a1 x 1 + a2 x 2 + · · · + an x n = ai x i
where {ai ∈ F, i = 1, n}.
2.4 Linearly independent set

Let V be a vector space over the field F. A finite set of vectors {x1 , x2 , · · · , xn } ∈ V
is said to be linearly independent if every relation of the form
a1 x1 + a2 x2 + · · · + an xn = 0, ai ∈ F, 1 ≤ i ≤ n
implies each ai = 0. Otherwise (i.e., if at least one ai ̸= 0) it is called linearly

dependent.
Example 10 Consider the following n vectors in V = Rn , where,

     
1 0 0
0 1 0
e1 =  ..  , e2 =  ..  , · · · , en =  ..  .
     
. . .
0 0 1

If ai ∈ F for i = 1, n are n arbitrary scalars, then

a1 e1 + a2 e2 + · · · + an en = 0.
So,
           
a1 0 0 0 a1 0
 0  a2   0  0  a2  0
 ..  +  ..  , · · ·  ..  =  ..  =⇒  ..  =  .. 
           
. .  .  .  .  .
0 0 an 0 an 0
So a1 = a2 = . . . an = 0 and the vectors are linearly independent.
Figure 10: (a) Two mutually perpendicular vectors and (b) two oblique vectors.
Example 11 Consider Fig. 10. Which one of them is linearly independent?
The first figure is obvious. Let’s examine the second case. We write,

a1 b1 0 a1 b 1 α 0
α +β = , i.e, = . (12)
a2 b2 0 a2 b 2 β 0
If the determinant of the coefficient matrix is not equal to zero, then we always get
α = β = 0. Then a and b are linearly independent. However, if the determinant
is zero, then
a2 b2
a1 b2 − a2 b1 = 0, or = .
a1 b1
This means a and b have the same direction (Fig. 11). We can express any of
them with a scalar multiplication of the other vector, say b = λa, where λ belongs
to the field and represents a scaling.

Figure 11: Two linearly dependent vectors.
Example 12 Let
     
 1 0 −1 
S=  2 , 3 , 0  ⊂ V = R3 .
   
0 1 1
 
Check if they are linearly independent or not

Let a, b, c ∈ F. Then, by definition,
       
1 0 −1 0
a 2 + b 3 + c  0  = 0 .
0 1 1 0
This gives,
a−c = 0
2a + 3b = 0
b + c = 0.
The coefficient matrix

 
1 0 −1
A = 2 3 0  .
0 1 1
and det A = 1 ̸= 0. So the only solutions are a = b = c = 0 and the set S is

linearly independent.

2.5 Linear span
Example 13 Show that the set {1, x, 1 + x + x2 }, where x ∈ F is linearly inde-

pendent in the space of all polynomials.
Let a, b, c ∈ F. Then,
a(1) + b(x) + c(1 + x + x2 ) = 0

(a + c) + (b + c)x + c(x2 ) = 0
=⇒ (a + c) = 0, (b + c) = 0, c = 0 for all x.
So a = b = c = 0, and the set is linearly independent.
2.5 Linear span

Let V be a vector space and S ⊂ V. Then linear span of S or Span(S) is the set
of all linear combinations of the elements of S. Thus we have
Span (S) = a1 x1 + a2 x2 + · · · + ap xp
where S := {x1 , x2 , · · · , xp } and a1 , a2 , · · · , ap are from F.
2.6 Basis, components and dimension

A set of vectors B = {g 1 , g 2 , · · · , g n } (g i ∈ V) is said to be a basis of V if
1. B is linearly independent
2. Span(B) = V
So any x ∈ V can be expressed as x = xi g i , where xi s are called components of V
with respect to the basis {g i }. If x represents position vector, then those scalars
are called coordinates.
A vector space can have infinite many different basis, but all of them will have
the same number of elements. The number of elements in a basis is called the
dimension of the space V. In this case dim V = n.
Theorem 2.1 Let B = {g 1 , g 2 , · · · , g n } be a basis of V. Then every element x

can be uniquely expressed as x = xi g i .
Let us assume the x could be expressed in a different way and x = ai g i . This
means
x = xi g i = ai g i =⇒ (xi − ai )g i = 0i .
By definition, {g i } is linearly independent set. So, ai = xi , i.e., the expression is

unique.

Example 14 Show that the vectors

     
 1 2 1 
2 , 1 , −1
1 0 2
 
form a basis of V = R3 .
Let a, b, c ∈ F. We start with
       
1 2 1 0
a 2 + b 1 + c −1 = 0 .
1 0 2 0
We further write,
    
1 2 1 a 0
2 1 −1  b  = 0
1 0 2 c 0
Now det A = −11 ̸= 0, and the only solutions are a = b = c = 0. So the vectors
are linearly independent. Next we need to show that they also span the space.
Consider an arbitrary vector
 
k1
k2  ∈ V = R3 .
k3
Then,
    
1 2 1 a k1
2 1 −1  b  = k2  .
1 0 2 c k2
As detA ̸= 0, for any arbitrary k1 , k2 , k3 , the system is solvable uniquely. So, the
considered vectors span R3 , and they are a basis.
2.6.1 Standard basis

Consider the set Bs = {e1 , e2 , · · · , en }, where,
     
1 0 0
0 1 0
e1 =  ..  , e2 =  ..  , · · · , en =  ..  . (13)
     
. . .
0 0 1
Note that they are linearly independent. This particular basis is known as standard
basis in Rn .

Example 15 Let V = R2×2 . Show that

1 0 0 1 0 0 0 0
α= ,β = ,γ = ,δ =
0 0 0 0 1 0 0 1
is a basis.
First step is to show that the set is linearly independent. Let a, b, c, d ∈ F. Then,
aα + bβ + cγ + dδ = 0,
i.e.,

1 0 0 1 0 0 0 0 0 0 a b 0 0
a +b +c +d = =⇒ =
0 0 0 0 1 0 0 1 0 0 c d 0 0
So, a = b = c = d = 0 and the set is linearly independent. The second step is to

show that the set also span V. Consider a generic vector

k1 k2
K= ∈ V,
k3 k4
then
K = k1 α + k2 β + k3 γ + k4 δ,
i.e., the vectors α, β, γ, δ span the space. The set is linearly independent and span
the space as well. So the set of vectors is a basis.
Example 16 What is the dimension of the space V = {0}, i.e., space only with
the zero element?
Dimension ≡ Number of basis elements in V,
First check that (by yourself) V = {0} is a vector space. Since {0} is the only
element, the only possible basis is B = {0}. Then,
α0 = 0
always satisfies for any α ̸= 0, i.e., the considered space can not have any basis.
So dim V = 0.

2.7 Vector subspaces

Definition 2.1 Let V be a vector space over the field F and W ⊂ V. Then W
is called a subspace of V if W itself is a vector space over F with respect to the
operations of vector addition and scalar multiplication in V.
Theorem 2.2 The necessary and sufficient condition for a nonempty subset W
of a vector space V over field F to be a subspace is
a, b ∈ F, and x, y ∈ W =⇒ ax + by ∈ W. (14)
Proof:
First note that any subset of V will follow operations of vector additions and scalar
multiplications, i.e., A2, A3, B2, B4, B5. We do not need prove again those.
Necessary condition
Consider W is a subspace of V. Then W must be closed under scalar multiplication
and vector addition. This means
a ∈ F, x ∈ W =⇒ ax ∈ W
b ∈ F, y ∈ W =⇒ by ∈ W
Now as ax ∈ W and by ∈ W, then ax+by ∈ W. Hence the condition is necessary.
Sufficient condition
Now suppose W is a non-empty subset of V satisfying the given condition ax+by ∈
W for all a, b ∈ F, and x, y ∈ W. We need to satisfy that
1. Closed under addition: x + y ∈ W, ∀x, y ∈ W
2. ∃ an element 0 ∈ W
3. To each element x ∈ W, ∃ −x ∈ W such that: x + (−x) = 0
4. Closed under scalar multiplication: ax ∈ W, ∀a ∈ F and x ∈ W
5. ∃ an non-zero element 1 ∈ F (called one/unity) such that: 1x = x, x ∈ W
As ax + by ∈ W,
1. let a = 1, b = 1, then 1.x + 1.y = x + y ∈ W . Thus W is closed under

vector addition.

2. Let a = 0, b = 0, then 0.x + 0.y = 0 ∈ W . Thus the zero vector 0 ∈ W.
3. Let a = −1, b = 0, then −1.x + 0.y = −x ∈ W . Thus the additive inverse

−x ∈ W.
4. Let y = 0, then ax + b0 = ax ∈ W . Thus W is closed under scalar

multiplication.
5. Let a = 1, and y = 0, then 1x + b0 = x ∈ W .
This proves that W is itself a vector space and subspace of V.
Figure 12: Representation of example 17.
Example 17 Let us consider a vector space V = R3 , and W := (w1 , w2 , 0) ⊂ V,

where w1 , w2 ∈ F. Show that W is a subspace of V.
Let x = (x1 , x2 , 0), and y = (y1 , y2 , 0) ∈ W, and x1 , x2 , y1 , y2 ∈ F. Then
αx + βy = α(x1 , x2 , 0) + β(y1 , y2 , 0) (α, β ∈ F)

= (αx1 , αx2 , 0) + (βy1 , βy2 , 0)
= (αx1 + βy1 , αx2 + βy2 , 0)
Now, αx1 + βy1 ∈ F and αx2 + βy2 ∈ F. So, αx + βy ∈ W, and W is a subspace

of V.
Example 18 Let us consider a vector space V = R3 , and W := (w1 +2w2 , w2 , −w1 +

3w2 ) ⊂ V, where w1 , w2 ∈ F. Show that W is a subspace of V.

Let,
x = (x1 + 2x2 , x2 , −x1 + 3x2 ) x1 , x2 ∈ F
y = (y1 + 2y2 , y2 , −y1 + 3y2 ) y1 , y2 ∈ F
αx + βy = α(x1 + 2x2 , x2 , −x1 + 3x2 ) + β(y1 + 2y2 , y2 , −y1 + 3y2 ) α, β ∈ F

= (αx1 + βy1 + 2(αx2 + βy2 ), αx2 + βy2 , 0), −(αx1 + βy1 ) + 3(αx2 + βy2 )
= (p + 2q, q, −p + 3q) ∈ W, p = αx1 + βy1 , q = αx2 + βy2 ∈ F
Figure 13: Intersection of any two subspaces is again a subspace.
Theorem 2.3 Intersection of any two subspaces W1 and W2 of V is also a sub-

space
Proof: Let x, y ∈ W1 ∩ W2 , then
x ∈ W1 ∩ W2 =⇒ x ∈ W1 also x ∈ W2
y ∈ W1 ∩ W2 =⇒ y ∈ W1 also y ∈ W2
As W1 is a subspace and x, y ∈ W1 , then αx + βy ∈ W1 . Similarly, W2 is a
subspace and x, y ∈ W2 , then αx + βy ∈ W2 . So, αx + βy ∈ W1 ∩ W2 , and
W1 ∩ W2 is a subspace of V (Fig. 13).
However, union of two subspaces may not be a subspace. Consider V = R3 ,
and
W1 := {(0, 0, z), z ∈ F} ⊂ V
W2 := {(0, y, 0), y ∈ F} ⊂ V.
Then, α(0, 0, z) + β(0, y, 0) = (0, βy, αx) ∈/ W1 and ∈ / W2 . So, (0, βy, αx) ∈
/
W1 ∪ W2 , i.e., the union is not closed under vector addition (Fig. 14).

2.8 Linear sum of two subspaces
Figure 14: Union of two subspaces may not be a subspace.

Definition 2.2 Let V be a vector space over F, and W1 , W2 be the two subspaces
of V. Then the linear sum of the subspaces W1 , and W2 , denoted by W1 + W2 , is
the set of all sums w1 + w2 such that w1 ∈ W1 and w2 ∈ W2 .
Thus,
W1 + W2 := {w1 + w2 ; w1 ∈ W1 , w2 ∈ W2 }
Lemma 2.4 W1 + W2 is a subspace.

Let w′1 , w′′1 ∈ W1 , and w′2 , w′′2 ∈ W2 . Consider, v ′ = w′1 + w′2 ∈ W1 + W2 , and
v ′′ = w′′1 + w′′2 ∈ W1 + W2 . Then,
αv ′ + βv ′′ = α(w′1 + w′2 ) + β(w′′1 + w′′2 )

= (αw′1 + βw′′1 ) + (αw′2 + βw′′2 )
| {z } | {z }
∈W1 ∈W2
∈ W1 + W2
Theorem 2.5 If W1 and W2 are the two subspaces of V, then
dim(W1 + W2 ) = dim W1 + dim W2 − dim(W1 ∩ W2 )
Let Bγ = {γ 1 , γ 2 , . . . , γ k } be a basis of W1 ∩ W2 and dim(W1 ∩ W2 ) = k. We can

write a basis of W1 as
Bα = {γ 1 , γ 2 , . . . , γ k , α1 , α2 , . . . , αt }
and a basis of W2 as
Bβ = {γ 1 , γ 2 , . . . , γ k , β 1 , β 2 , . . . , β m }

So,
dim W1 + dim W2 − dim(W1 ∩ W2 ) = (k + t) + (k + m) − k = k + t + m.
We need to show that dim(W1 + W2 ) = k + t + m. Our claim is then,
B = {γ 1 , γ 2 , . . . , γ k , α1 , α2 , . . . , αt , β 1 , β 2 , . . . , β m }
is a basis of W1 + W2 . First we will show that the set B is linearly independent.
Consider,
(c1 γ 1 + c2 γ 2 + · · · + ck γ k + a1 α1 + a2 α2 + · · · + at αt ) + (b1 β 1 + b2 β 2 + · · · + bm β m ) = 0(15)
| {z } | {z }
∈W1 ∈W2
(b1 β 1 + b2 β 2 + · · · + bm β m ) = − (c1 γ 1 + c2 γ 2 + · · · + ck γ k + a1 α1 + a2 α2 + · · · + at αt ) ∈ W2
| {z } | {z }
∈W2 ∈W1
=⇒ b1 β 1 + b2 β 2 + · · · + bm β m ∈ W1 ∩ W2
So, it can be written with respect to the basis of W1 ∩ W2 , and we write,
b1 β 1 + b2 β 2 + · · · + bm β m = d1 γ 1 + d2 γ 2 + · · · + dk γ k
=⇒ b1 β 1 + b2 β 2 + · · · + bm β m − d1 γ 1 − d2 γ 2 − · · · − dk γ k = 0
But {γ 1 , γ 2 , . . . , γ k , β 1 , β 2 , . . . , β m } is a basis of W2 . We get
b1 = b2 = · · · = bm = 0 = d1 = · · · = dk
From (15),
(c1 γ 1 + c2 γ 2 + · · · + ck γ k + a1 α1 + a2 α2 + · · · + at αt ) = 0.
Moreover, = {γ 1 , γ 2 , . . . , γ k , α1 , α2 , . . . , αt } is a basis of W1 and so c1 = · · · =
ck = a1 = · · · = at = 0. This implies
c 1 = · · · = c k = a1 = · · · = at = b 1 = b 2 = · · · = b m = 0
and B = {γ 1 , γ 2 , . . . , γ k , α1 , α2 , . . . , αt , β 1 , β 2 , . . . , β m } is linearly independent.
Next step is to show that the set B spans W1 +W2 . Let w1 ∈ W1 , and w2 ∈ W2
such that w = w1 + w2 ∈ W1 + W2 . Now
w = w1 + w2
= (c1 γ 1 + c2 γ 2 + · · · + ck γ k + a1 α1 + a2 α2 + · · · + at αt )
+(c′1 γ 1 + c′2 γ 2 + · · · + c′k γ k + b1 β 1 + b2 β 2 + · · · + bm β m )
= (c1 + c′1 )γ 1 + (c2 + c′2 )γ 2 + · · · + (ck + c′k )γ k
+a1 α1 + a2 α2 + · · · + at αt + b1 β 1 + b2 β 2 + · · · + bm β m
= linear combination of B
We then have:

2.9 Direct sum of two subspaces
1. B is linearly independent.
2. Span B = W1 + W2 .
So B is a basis, and hence the proof.

Definition 2.3 Let V be a vector space over F. Then V is said to be the direct
sum of W1 , W2 if
1. W1 + W2 = V
2. W1 ∩ W2 = {0}
We then write
V = W1 ⊕ W 2 .
Also note that
dim V = dim W1 + dim W2 − dim(W1 ∩ W2 )

= dim W1 + dim W2 − dim{0}
= dim W1 + dim W2 (16)
Example 19 Let V = R2 over the field F. Then W1 := {(a, 0) : a ∈ F} and

W2 := {(0, b) : b ∈ F} are the subspaces of R2 . Now any element (x, y) ∈ R2 can
be uniquely expressed as (x, y) = (x, 0) + (0, y). Then R2 = W1 ⊕ W2 .
Example 20 Consider the following subset of V = R3

     
 1 3 1 
S := 2 ,  6  , 2 .
0 −1 1
 
1. Show that W := Span S is a subspace.
2. Find a basis and the dimension of W.
Ans-1: Note that

     
1 3 1
w1 2 + w2 6 + w2 2 ∈ W
    
0 −1 1

Or,
 
w1 + 3w2 + w3
2w1 + 6w2 + 2w3  ∈ W
−w2 + w3
Consider two elements of W

   
x1 + 3x2 + x3 y1 + 3y2 + y3
x = 2x1 + 6x2 + 2x3  and y = 2y1 + 6y2 + 2y3 
−x2 + x3 −y2 + y3
So,
 
α1 (x1 + 3x2 + x3 ) + α2 (y1 + 3y2 + y3 )
α1 x + α2 y = α1 (2x1 + 6x2 + 2x3 ) + α2 (2y1 + 6y2 + 2y3 )
α(−x2 + x3 ) + α2 (−y2 + y3 )
 
(α1 x1 + α2 y1 ) + 3(α1 x2 + α2 y2 ) + (α1 x3 + α2 y3 )
= 2(α1 x1 + α2 y1 ) + 6(α1 x2 + α2 y2 ) + 2(α1 x3 + α2 y3 )
−(α1 x2 + α2 y2 ) + (α1 x3 + α2 y3 )
 
p1 + 3p2 + p3
= 2p1 + 6p2 + 2p3 
−p2 + p3
where p1 = α1 x1 + α2 y1 , p2 = α1 x2 + α2 y2 , etc. So α1 x + α2 y ∈ W, and W is a

subspace of V.
Ans-2: We first check linear independency. Consider

       
1 3 1 0
a 2 + b 6 + c 2 = 0
      
0 −1 1 0
and we construct the coefficient matrix

    
1 3 1 a 0
A := 2 6 2  b  = 0
0 −1 1 c 0
The space of linear combinations of the column vectors of A, i.e., the span of the
column vectors of A, is known as column space, colA, and W = colA. Similarly,
the span of the row vectors also forms a subspace, known as row space, rowA,
which is definitely different from W. Note that rowAt = colA = W.

If the det A =
̸ 0, then a, b, c are all zero, and the given set is a basis. However,
in our case det A = 0. One convenient way is to find the echelon form of At such
that row-shuffling implies different linear combinations of the column vectors of
A.
     
1 2 0 1 2 0 1 2 0
R2 →R2 −3R1 R →R −R2
3 6 −1 − −−−−−−→ 0 0 −1 −−3−−−3−−→ 0 0 −1
R3 →R3 −R1
1 2 1 0 0 1 0 0 0
So a basis is
    
 1 0 
2 ,  0  ,
0 −1
 
and dim W = 2. A simple representation of the subspace is shown in Fig. 15.
Figure 15: A subspace W in R3 .
Example 21 Consider the following subspace W of V = R4

     

 1 2 3 
     
−2 , , 8  .
3

W := Span   5   1  −3

 
−3 −4 −5
 
Find a basis and the dimension of W.

The coefficient matrix becomes

 
1 2 3
−2 3 8
A :=  
5 1 −3
−3 −4 −5
Then starting from At to obtain the echelon form, we write,

     
1 −2 5 −3 1 −2 5 −3 1 −2 5 −3
R2 →R2 −2R1 R3 →R3 −2R2
2 3 1 −4 −− −−−−−→ 0 7 −9 2  −− −−−−−→ 0 7 −9 2 
R3 →R3 −3R1
3 8 −3 −5 0 14 −18 4 0 0 0 0
So a basis is
   

 1 0 
   
−2 ,  7  ,

  5  −9
 
3 2
 
and dim W = 2.

3 Homomorphism of vector spaces or linear trans-
formation
Definition 3.1 Let V and U be two vector spaces over the field F. Consider the
map
T :V→U
such that T v = u ∈ U, v ∈ V. The map T is called homomorphism or linear

transformation (LT) of V into U if
1. T (x + y) = T x + T y, x, y ∈ V.
2. T (ax) = aT x, a ∈ F, x ∈ V.
The above two conditions could be combined into a single condition as
T (ax + by) = aT x + bT y, a, b ∈ F x, y ∈ V (17)
A linear transformation T : V → U is called isomorphic if the mapping is bijective,

and we write V ∼ = U. T is known as a second order tensor that maps vectors to
vectors.
Example 22 Consider the relation between the instantaneous angular momentum

vector L, and angular velocity vector ω as
L = Iω.
The mass moment of inertia I is a second order that maps the vector ω to L.
These two vectors may not necessarily be in the same direction. A scalar multipli-
cation with a vector also gives a vector. However, like a tensor operation, scalar
multiplication can not change the direction of a vector. The above relation also
follows the linear relationship (17), i.e.,
I(αω1 + βω2 ) = αIω1 + βIω2
for any scalars α, β and for any vectors ω1 , ω2 .
Example 23 The polarization vector P , and the electric field vector E are con-
nected as
P = ε0 χe E.
Here, ε0 , a scalar, is the electric permittivity of free space. χe , a scalar too, is the
electric susceptibility of the concerned medium. Note that P and E have the same

direction as ε0 χe is a scalar. However, some materials, e.g., liquid crystals, have
the capability to change the direction of the polarization from the applied electric
field. In this case electric susceptibility, χe , acts like a second order tensor, i.e.,
P = ε0 χe E.
The above relation also follows the linear relationship (17).
Example 24 The traction vector t, and the unit normal n to a plane (cut) at a
point are connected as
t = σn
where σ is the Cauchy stress and a second order tensor. The above relation also
follows the linear relationship (17).
Example 25 Show that T : V → U, where dim V = 3 and dim U = 2, defined by

 
v1
v
T v2  = 1
v2
v3
is homomorphism of V into U.
Let,
   
x1 y1
x = x2  ,
 y = y2  .
 (18)
x3 y3
Then
     
ax1 by1 ax1 + by1
ax 1 + by 1
T (ax + by) = T ax2  + by2  = T ax2 + by2  =
ax2 + by2
ax3 by3 ax3 + by3
   
x1 y1
ax1 by1 x1 y1
= + =a +b = aT x2 + bT y2 
  
ax2 by2 x2 y2
x3 y3
= aT x + bT y
So, T is a homomorphism or a LT. Note that T behaves as a projection tensor

which projects a vector (v1 , v2 , v3 )t to a plane that contains (v1 , v2 )t components.

3.1 Range space R(T )
3.0.1 Linear operator

T becomes a linear operator if V = U, i.e.,
T : V → V.
3.0.2 Identity operator

I is called an identity operator if
I : V → V, such that Iv = v
3.0.3 Zero transformation

0̂ is called a zero transformation if
0̂ : V → U, such that 0̂v = 0 ∈ U
Lemma 3.1 A LT maps 0 ∈ V to a zero 0 ∈ U, i.e., T 0 = 0.

Proof: Let’s start with T (v + 0) = T v + T 0. Now, v + 0 = v, so T (v + 0) = T v,
and then T v = T v + T 0, or T 0 = 0. This means a zero is always mapped to a
zero.
Lemma 3.2 T (−v) = −T v.

Proof: Let’s start with T (v + (−v)) = T v + T (−v). Now, v + (−v) = 0, so
T (v + (−v)) = T 0 = 0, and then 0 = T v + T (−v), or T (−v) = −T v.
Lemma 3.3 T (αi v i ) = αi T v i .

Pn−1 Pn−1
Proof:
Pn−1Let’s start with T
Pn−2 ( α i v i + αn v n ) = T (
Pn−2 αi v i ) + αn T v n . Now,
T( αi v i ) = T ( αi v i + αn−1 v n−1 ) = T ( αi v i ) + αn−1 T v n−1 , and so
on. Eventually we write,
T (α1 v 1 + · · · + αn v n ) = α1 T v 1 + · · · + αn T v n (19)
3.1 Range space R(T )

Definition 3.2 Let V and U be two vector spaces over F and T be a homomor-
phism from V to U. Then range of T , written as R(T ), is basically the image set
I(T ) of the mapping T , i.e., I(T ) = R(T ) ⊆ U.
R(T ) := {T v = u ∈ U, v ∈ V} .
Lemma 3.4 R(T ) is a subspace of U

3.2 Null space N (T ) or ker(T )
Figure 16: Range space, R(T ), of homomorphism T .
Proof: As V is a vector space, and v 1 , v 2 ∈ V (Fig. 16), then α1 v 1 + α2 v 2 ∈ V,

where α1 , α2 ∈ F. Now,
T (α1 v 1 + α2 v 2 ) = α1 T v 1 + α2 T v 2
= α1 u1 + α2 u2
where, u1 = T v 1 ∈ R(T ), and u2 = T v 2 ∈ R(T ). As, α1 v 1 + α2 v 2 is a vector in

V, after the mapping α1 u1 + α2 u2 belongs to R(T ). So, u1 , u2 ∈ R(T ) implies
α1 u1 + α2 u2 ∈ R(T ). This proves that R(T ) is a subspace of U (see equation
(14)).
3.2 Null space N (T ) or ker(T )

Definition 3.3 Let V and U be two vector spaces over F and T be a homomor-
phism from V to U. Then, the null space is defined as
N (T ) := {v ∈ V : T v = 0 ∈ U} ⊆ V.
It is also known as kernel of T or ker(T ).
Lemma 3.5 N (T ) is a subspace of V

Proof: Let v 1 , v 2 ∈ N (T ) ⊆ V (Fig. 17). Then, by definition, T v 1 = 0 and
T v 2 = 0. Let α1 , α2 ∈ F, then
T (α1 v 1 + α2 v 2 ) = α1 T v 1 + α2 T v 2
= 0+0=0
So, α1 v 1 + α2 v 2 ∈ N (T ), and N (T ) is a subspace of V (see equation (14)).
Lemma 3.6 If Span {v 1 , . . . , v n } = V, then Span {T v 1 , . . . , T v n } = R(T )

3.3 Rank and nullity of a homomorphism
Figure 17: Null space, N (T ) of homomorphism T .
Proof: Let v = α1 v 1 + · · · + αn v n ∈ V. Then T v = α1 T v 1 + · · · + αn T v n ∈ R(T ).

So, Span {T v 1 , . . . , T v n } = R(T ). Note that ’span’ does not say whether the
set is linearly independent or not. It says that the any vector can be expressed
as linear combination of {T v 1 , . . . , T v n }. It may happen that some of them are
equal to each other or zero.

Let T is a homomorphism from V to U. Then
Rank T = dim R(T )

Null T = dim N (T )
Theorem 3.7 Show that:
Rank T + Null T = dim V
Proof: Let N (T ) be the null-space of V, and {γ 1 , γ 2 , . . . , γ k } be a basis of N (T ).

Let {γ k+1 , . . . , γ n } in V such that {γ 1 , . . . , γ k , γ k+1 , . . . , γ n } be a basis of V. Now,
Span {T γ 1 , . . . , T γ k , T γ k+1 , . . . , T γ n } = R(T ). (20)
As, γ 1 , γ 2 , . . . , γ k ∈ N (T ), we immediately write T γ 1 = T γ 2 = · · · = T γ k = 0.

So,
Span {T γ k+1 , . . . , T γ n } = R(T ). (21)
We now want to check if {T γ k+1 , . . . , T γ n } is linearly independent or not. We

equate their linear combination to zero and write
αk+1 T γ k+1 + · · · + αn T γ n = 0,
T (αk+1 γ k+1 + · · · + αn γ n ) = 0.

So, αk+1 γ k+1 + · · · + αn γ n ∈ N (T ), and it could be expressed as a linear combi-

nations with the basis of N (T ). We write,
αk+1 γ k+1 + · · · + αn γ n = β1 γ 1 + β2 γ 2 + · · · + βk γ k ,
β1 γ 1 + β2 γ 2 + · · · + βk γ k − αk+1 γ k+1 − · · · − αn γ n = 0
Now, {γ 1 , . . . , γ k , γ k+1 , . . . , γ n } be a basis of V. So,
β1 = β2 = · · · = βk = 0 = −αk+1 = −αk = · · · = −αn .
As αk+1 = αk = · · · = αn = 0, then {T γ k+1 , . . . , T γ n } is a basis of R(T ), and
dim R(T ) = n − k = dim V − dim N (T ) =⇒ dim R(T ) + dim N (T ) = dim V
or
Rank T + Null T = dim V (22)
Example 26 Let T : V → V, where V = R3 , is a linear operator defined by

   
x x + 2y − z
T y  =  y + z  (23)
z x + y − 2z
Find R(T ), a basis of R(T ), Rank T and N (T ), a basis of N (T ) , Null T .

Ans: We write
       
x + 2y − z 1 2 −1
 y + z  = x 0 + y 1 + z  1 
x + y − 2z 1 1 −2
Then,
      
 1 2 −1 
Span  0 , 1 ,
    1  = R(T ).
1 1 −2
 
We construct the matrix

 
1 2 −1
A := 0 1 1  ,
1 1 −2

where the space of linear combinations of the column vectors of A, i.e., the column
space is equal to R(T ). Conducting the row-operations of At we write
     
1 0 1 1 0 1 1 0 1
R2 →R2 −2R1 R →R −R2
 2 1 1 − −−−−−−→ 0 1 −1  −−3−−−3−−→ 0 1 −1
R3 →R3 +R1
−1 1 −2 0 1 −11 0 0 0
So a basis of R(T ) is
    
 1 0 
0 ,  1  ,
1 −1
 
and Rank T = dim R(T ) = 2. Let,

 
a1
a = a2  ∈ N (T )
 (24)
a3
Then, T a = 0. This gives us,
       
a1 a1 + 2a2 − a3 1 2 −1 a1 0
T a2 =
   a2 + a3  =⇒ 0 1 1
  a2   0 (25)
a3 a1 + a2 − 2a3 1 1 −2 a3 0
Now,
     
1 2 −1 1 2 −1 1 2 −1
R3 →R3 −R1 R →R +R2
0 1 1  −
−−−−−−→ 0 1 1  −−3−−−3−−→ 0 1 1 
1 1 −2 0 −1 −1 0 0 0
So,
    
1 2 −1 a1 0
O
0 1 1   a2  = 0
0 0 0 a3 0
where a3 is the free component. We then express the other two components in
terms of a3 . We write,
a1 + 2a2 − a3 = 0
a2 + a3 = 0 (26)
Or,
a2 = −a3
a1 = 3a3 (27)

3.4 Singular homomorphism
and so,
   
a1 3
a2  = a3 −1 (28)
a3 1
That is,
  
 3 
−1 (29)
1
 
is a basis of N (T ), and Null T = dim N (T ) = 1. Check that Rank T + Null T =

2 + 1 = 3. Note that when T acts on any vector on the line a3 (3, −1, 1) ∈ N (T ),
Figure 18: R(T ) is a plane containing the two basis vectors and N (T ) is a line
containing its basis.
the mapped value would be the origin. Any other vectors in V will stay on the
plane (hatched, see Fig. 18).
3.4 Singular homomorphism

Let
T : V → U.
If any non zero v ∈ V gives T v = 0, i.e., v ∈ N (T ), and Null T = dim N (T ) ̸= 0,
then T is a singular homomorphism.

3.5 Nonsingular homomorphism
3.5 Nonsingular homomorphism

Let
T : V → U.
If T v = 0 =⇒ v = 0, i.e., N (T ) = {0}, and Null T = dim N (T ) = 0, then T is
a nonsingular homomorphism.
Theorem 3.8 Let T : V → U is nonsingular. Then T carries each linearly

independent subset of V to a linearly independent subset in U.
Proof: Let {γ 1 , γ 2 , . . . , γ n } be a linearly independent subset of V. Note that,
{T γ 1 , . . . , T γ n } ∈ R(T ). To show that {T γ 1 , . . . , T γ n } is linearly independent,
consider
α1 T γ 1 + · · · + αn T γ n = 0, (30)
T (α1 γ 1 + · · · + αn γ n ) = 0.
As T is nonsingular (i.e N (T ) = {0}), then
α1 γ 1 + · · · + αn γ n = 0.
However, the considered set {γ 1 , γ 2 , . . . , γ n } is linearly independent, and so α1 =
· · · = αn = 0. Then from (30) {T γ 1 , . . . , T γ n } is also linearly independent.
3.6 Isomorphism
When T : V → U is bijective, we call it isomorphism. Moreover, T is isomorphic
if
1. T is nonsingular
2. dim V = dim U
Proof: To show isomorphism, we need to prove that the mapping is bijective, i.e.,
T is both injective and surjective.
T is injective:
We want to show T v 1 = T v 2 =⇒ v 1 = v 2 , where v 1 , v 2 ∈ V. Let us start with
T v1 = T v2,
T v 1 − T v 2 = 0,
T (v 1 − v 2 ) = 0
As, T is nonsingular, T (v 1 − v 2 ) = 0 means (v 1 − v 2 ) = 0 or v 1 = v 2 . So, T is
injective.

3.7 Invertible T :
T is surjective:
We need to show that R(T ) = U. As T is nonsingular, then N (T ) = {0}, and
Rank T + Null T = dim V
gives
dim R(T ) + 0 = dim V = dim U (given).
As, dim R(T ) = dim U, and R(T ) ⊆ U then R(T ) = U, i.e., T is surjective. So,
T is injective and surjective, i.e., bijective.
Lemma 3.9 If T : V → U is isomorphic and {γ 1 , γ 2 , . . . , γ n } is a basis of V,

then {T γ 1 , . . . , T γ n } is also a basis of U.
Proof: From Theorem 3.8, we say that as T is nonsingular, then the linearly
independent set {γ 1 , γ 2 , . . . , γ n } in V transforms to a linearly independent set
{T γ 1 , . . . , T γ n } in R(T ). A nonsingular T also implies dim V = dim U, and so
R(T ) = U. So, {T γ 1 , . . . , T γ n } is also a linearly independent set in U. Now from
Lemma 3.6, Span {T v 1 , . . . , T v n } = R(T ) = U. So, {T γ 1 , . . . , T γ n } is a basis of
U.
3.7 Invertible T :
When T : V → U is an isomorphism, then T is invertible. We write,
T −1 : U → V
and if T v = u, then T −1 u = v. When T is invertible, it is also nonsingular and
dim V = dim U.
Theorem 3.10 T −1 is a homomorphism.

Proof. Let u1 , u2 ∈ U, such that T v 1 = u1 , and T v 2 = u2 . Then by definition,
v 1 = T −1 u1 , and v 2 = T −1 u2 . Moreover, α1 u1 + α2 u2 ∈ U. We start with
T (α1 v 1 + α2 v 2 ) = α1 T v 1 + α2 T v 2
= α1 u 1 + α2 u 2 ∈ U
=⇒ α1 v 1 + α2 v 2 = T −1 (α1 u1 + α2 u2 ).
Now
T −1 (α1 u1 + α2 u2 ) = α1 v 1 + α2 v 2
= α1 T −1 u1 + α2 T −1 u2 .
So, T −1 is a homomorphisms.

3.8 Representation of homomorphism or LT by a matrix

Let BV = {g 1 , · · · g n } be a basis of V, and BU = {f 1 , · · · f m } be a basis for U.
Given,
T : V → U.
If T acts on vectors, g j , then each n vectors T g j is uniquely expressible as a linear

combination
T g j = Tij f i (31)
of f i . The scalars T1j , · · · , Tmj are the components of T g j with respect to the basis
BU . The n×m matrix [T ]BV −BU or simply [T ] (or [Tij ]) is the matrix representation
of T relative to the pair of basis BV and BU . If v = v1 g 1 + · · · + vn g n ∈ V, then
u = Tv = T (vj g j )
= vj (T g j )
= vj Tij f i
ui f i = (Tij vj )f i
Or we write,
ui = Tij vj
For example, if dim V = 2 and dim U = 3, we write

 
f1
T g1 T T T
= 11 21 31 f 2  (32)
T g2 T12 T22 T32
f3
where the matrix form of [T ]BV −BU is given by

T11 T21 T31
[T ]BV −BU = . (33)
T12 T22 T32
As ui = Tij vj , in the matrix notation, the components are connected as

   
u1 T11 T12
u2  = T21 T22  v1
v2
u3 T31 T32
where the matrix is transpose of (33), i.e.,
[u] = [T ]tBV −BU ][v] (34)

Example 27 Let T be a linear operator on V = R3 such that

   
a 2b + c
T  b  = a − 4b . (35)
c 3a
Find the matrix representation of T , [T ]B−B , with respect to the basis B where
      
 1 1 1 
B = {g 1 , g 2 , g 3 } := 1 , 1 , 0 .
   
1 0 0
 
Ans: Considering T g 1 we can write

   
1 3
T 1 = −3
   (36)
1 3
Now we want to express (3, −3, 3)t as a linear combination of the given basis. One
writes,
       
3 1 1 1
−3 = α1 1 + α2 1 + α3 0
3 1 0 0
and we have
α1 + α2 + α3 = 3
α1 + α2 = −3
α1 = 3
Or
α1 = T11 = 3, α2 = T21 = −6, α3 = T31 = 6
Similarly,
   
1 2
T 1 = −3 (37)
0 3
Now we want to express (2, −3, 3)t as a linear combination of the given basis. One
writes,
       
2 1 1 1
−3 = β1 1 + β2 1 + β3 0
3 1 0 0

and we have
β1 + β2 + β3 = 2
β1 + β2 = −3
β1 = 3
Or
β1 = T12 = 3, β2 = T22 = −6, β3 = T32 = 5
Finally,
   
1 0
T 0 = 1
   (38)
0 3
Now we want to express (0, 1, 3)t as a linear combination of the given basis. One
writes,
       
0 1 1 1
1 = γ1 1 + γ2 1 + γ3 0
3 1 0 0
and we have
γ1 + γ2 + γ3 = 0
γ1 + γ2 = 1
γ1 = 3
Or
γ1 = T13 = 3, γ2 = T23 = −2, γ3 = T33 = 1.
So,
 
3 −6 6
[T ]B = 3 −6 5
3 −2 1
Example 28 Consider a differential operator
D:V→V (39)
where V be the vector space of all polynomials
f (x) = a0 + a1 x + a2 x2 + a3 x3 , ai ∈ F.
Write the transformation matrix [D] relative to the basis BV and BR(D) .

Ans: Note that B = {1, x, x2 , x3 } be a basis of V. Now, Df (x) = a1 + 2a2 x +

3a3 x2 ∈ R(D) ⊆ V. So, BR(D) = {1, x, x2 } be a basis of R(D). Then we write
D1 = 0 = 0 + 0x + 0x2
Dx = 1 = 1 + 0x + 0x2
D x2 = 2x = 0 + 2x + 0x2
D x3 = 3x2 = 0 + 0x + 3x2
So,
 
0 0 0
1 0 0
[D]BV −BR(D) =
0

2 0
0 0 3

4 Homomorphism as vector space
Figure 19: Homomorphism as a vector space.
Consider the example of σn = t. If you keep changing the traction vector t,

say, by changing boundary conditions, the values of σ will keep changing without
changing n. So, we get a bunch of different second order tensors.
We now consider the set of all linear transformations from a vector space V to
U, and denote it Hom (V, U). We then impose addition and scalar multiplication
in Hom (V, U) over field F. Let T 1 and T 2 be two homomorphisms in Hom (V, U).
We define,
1. (T 1 + T 2 )v = T 1 v + T 1 v, where v 1 , v 2 ∈ F
2. (cT )v = c(T v), where c ∈ F
Then Hom (V, U) becomes a vector space over field F (Fig. 19). It satisfies all the
vector space rules A1 to A5 , and B1 to B5 . We only show the closer properties
here.
Closed under addition

We need to show that if T 1 , T 2 ∈ Hom (V, U) implies T 1 + T 2 ∈ Hom (V, U).
Consider, a vector α1 v 1 + α2 v 2 ∈ V. Then
(T 1 + T 2 )(α1 v 1 + α2 v 2 ) = T 1 (α1 v 1 + α2 v 2 ) + T 2 (α1 v 1 + α2 v 2 )
= α1 (T 1 v 1 + T 2 v 1 ) + α2 (T 1 v 2 + T 2 v 2 )
= α1 (T 1 + T 2 )v 1 + α2 (T 1 + T 2 )v 2
So, T 1 + T 2 ∈ Hom (V, U)

4.1 Product of homomorphism
Closed under scalar multiplication

We need to show that if T ∈ Hom (V, U) implies cT ∈ Hom (V, U). Consider, a
vector α1 v 1 + α2 v 2 ∈ V. Then
cT (α1 v 1 + α2 v 2 ) = cT (α1 v 1 ) + cT (α2 v 2 )

= cα1 T v 1 + cα2 T v 2
= α1 (cT v 1 ) + α2 (cT v 2 )
So, cT ∈ Hom (V, U). One can show that the other laws for the vector space are
satisfied.
Lemma 4.1
dim Hom (V, U) = dim V. dim U
We are not proving this. Consult any linear algebra book.
Figure 20: The composition mapping is called the product of homomorphism. We

simply write T 2 ◦ T 1 = T 2 T 1 .
Theorem 4.2 Let T 1 : V → U, and T 2 : U → W be the two homomorphisms.

Then the composition mapping T 2 ◦ T 1 ∈ Hom (V, W), i.e., again a homomor-
phism.

Consider a vector α1 v 1 + α2 v 2 ∈ V. Then
T 2 ◦ T 1 (α1 v 1 + α2 v 2 ) = T 2 (T 1 (α1 v 1 + α2 v 2 ))
= T 2 (α1 T 1 v 1 + α2 T 1 v 2 )
= α1 T 2 T 1 v 1 + α2 T 2 T 1 v 2
= α1 (T 2 ◦ T 1 )v 1 + α2 (T 2 ◦ T 1 )v 2
So, T 2 ◦ T 1 ∈ Hom (V, W), and we simply write the product T 2 T 1 , i.e, (T 2 T 1 )v =
T 2 (T 1 v). Commutative diagrams are very useful to represent composition map-
pings. We can represent the above mentioned composition mapping as
T1 /
V U . (40)
T2
T 2 ◦T 1 =T 2 T 1

W
Consider, T : V → U is a homomorphism and I U is the identity transformation in

U. Then, from (41)(a)
T /U IV /
V V V . (41)
IU T
IUT T IV

U U
we can write I U (T v) = (I U T )v or T v = (I U T )v or I U T = T . Similarly, if I V is

the identity transformation in V, from (41)(b) T (I V v) = (T I V )v or (T I V )v = T v
or T I V = T . We can then write,
IU T = T = T IV (42)
Theorem 4.3 If T : V → U, is isomorphic, then T −1 T = I V and T T −1 = I U .

Then, from (43)(a)
T / T −1 /
V U U V . (43)
T −1 T
T −1 T =I V T T −1 =I U

V U
T −1 u = v means T −1 (T v) = v or (T −1 T )v = I V v. We thus write T −1 T = I V .

Similarly, T v = u means T (T −1 u) = u or (T (T −1 ))u = u or (T T −1 )u = u.
This implies, T T −1 = I U .
Theorem 4.4 If T : V → U, is isomorphic, then T −1 : U → V is also isomorphic.

Proof: T is isomorphic means T is nonsingular and dim V = dim U. Let, u1 , u2 ∈
U. Then v 1 = T −1 u1 , and v 2 = T −1 u2 . To prove injectivity, let us start with
T −1 u1 = T −1 u2
T T −1 u1 = T T −1 u2
I U u1 = I U u2
u1 = u2 .
So, T −1 is injective. This also means T −1 is nonsingular as T −1 (u1 − u2 ) = 0

implies u1 − u2 = 0. Moreover, dim V = dim U implies T −1 is surjective. We thus
conclude that T −1 is bijective and hence isomorphic.
Theorem 4.5 If T : V → U, and S : U → V are isomorphic such that ST = I V ,

and T S = I U , respectively, then S = T −1 and S −1 = T .
V T / U U
S / V (44)
S T
ST =I V T S =I U

V U

4.2 Algebra of linear operators
From the commutative diagram (44), we can directly write S = T −1 . We can also
check it in the following way.
T −1 u = T −1 (I U u) = (T −1 I U )u = (T −1 (T S))u = ((T −1 T )S)u

= (I V S)u = I V (Su) = Su =⇒ T −1 = S
Similarly,
V S −1 / U U T −1 / V (45)
T −1 S −1
T −1 S −1 =I V S −1 T −1 =I U

V U
From the commutative diagram (45), we can directly write S −1 = T .

In a similar way, one can show that S −1 = T (try it!).

Theorem 4.6 Let us consider the vector space Hom (V, V). Such a structure is
called an ‘algebra’ as it is equipped with the multiplication of vectors. If A, B, C
∈ Hom (V, V) (linear operators), then the following additional product rules hold
true for an algebraic structure on the top of a vector space structure:
1. Closed under multiplication: AB ∈ Hom (V, V)
2. Multiplication is associative: A(BC) = (AB)C
3. Multiplication is distributive with respect to addition: A(B+C) = AB+AC

and (A + B)C = AC + BC
4. For each scalar α ∈ F, α(AB) = (αA)B = A(αB).
If there is an element I ∈ Hom (V, V) such that IT = T = T I for each T ∈

Hom (V, V), we call Hom (V, V) a (linear) algebra with identity and call I the
identity of the algebra.
Proof: We have the following three rules to prove these relations:
i (A + B)v = Av + Bv

ii (αA)v = αAv
iii (AB)v = A(Bv)
where v ∈ V. To prove that the multiplication is associative, we start with
A(BC)v = A(B(Cv)) = (AB)Cv

=⇒ A(BC) = (AB)C
Try yourself to prove the other relations! Note that, as V = U, we can write from
(42)
IT = T = T I (46)
Lemma 4.7 From Theorem 4.3, one can write for an algebraic structure Hom (V, V),
T −1 T = I = T T −1 (47)

5 Linear functional or linear form
Figure 21: Linear functional or linear form.
Linear functional or linear form is a special case of homomorphism when
ϕ : V → R, such that ϕ⟨v⟩ ∈ R, , v ∈ V
Then ϕ ∈ Hom (V, R). The notation ϕ⟨v⟩ is same as T v. We are just using ϕ⟨v⟩
instead of ϕv. Many researchers use ϕ(v) as well. Note that
ϕ⟨α1 v 1 + α2 v 2 ⟩ = α1 ϕ⟨v 1 ⟩ + α2 ϕ⟨v 2 ⟩, v 1 , v 2 ∈ V, and α1 , α2 ∈ F.
5.1 Dual space

We call V ∗ := Hom (V, R) as the dual space of V. The elements of V ∗ are often
called covector or 1-forms. Note that dim V ∗ = dim Hom (V, R) = dim V · dim R =
dim V.
Theorem 5.1 Let BV = {g 1 , . . . , g n } be a basis of V. Consider the set BV ∗ =

{ϕ1 , . . . , ϕn } of V ∗ such that
ϕi ⟨g j ⟩ = δij . (48)
Then BV ∗ = {ϕ1 , . . . , ϕn } is a basis of V ∗ (see figure 22).

Proof: Consider the following linear combination
α1 ϕ1 + α2 ϕ2 + · · · + αn ϕn = 0

5.1 Dual space
Figure 22: We construct {ϕ1 , ϕ2 , · · · , ϕn } in the following ways. When ϕ1 acts

on g 1 , we get 1, i.e., ϕ1 ⟨g 1 ⟩ = 1. Any other vectors other than g 1 will give 0.
Similarly, when ϕ2 acts on g 2 , it gives 1. All other vectors give 0, and so on.
Now,
(α1 ϕ1 + α2 ϕ2 + · · · + αn ϕn )⟨g j ⟩ = 0⟨g j ⟩ = 0
α1 ϕ1 ⟨g j ⟩ + α2 ϕ2 ⟨g j ⟩ + · · · + αn ϕn ⟨g j ⟩ = 0
αi ϕi ⟨g j ⟩ = 0
αi δij = 0
=⇒ αj = 0 for j = 1, · · · , n
As the set {ϕ1 , . . . , ϕn } is linearly independent, and dim V ∗ = n, then {ϕ1 , . . . , ϕn }
is a basis of V ∗ . This basis, BV ∗ is called the dual basis of BV . Such a dual basis
is unique (we are not proving its uniqueness). It is a standard practice to denote
ϕ as g ∗ , so we renaming and rewriting the dual basis as
BV ∗ = {g ∗1 , . . . , g ∗n }, and so g ∗i ⟨g j ⟩ = δij (49)
Let v ∗ ∈ V ∗ , and we write
v ∗ = c1 g ∗1 + c2 g ∗2 + · · · + cn g ∗n
v ∗ ⟨g i ⟩ = ci
=⇒ v ∗ = {v ∗ ⟨g i ⟩}g ∗i
Similarly, if v = V, then
v = c1 g 1 + c2 g 2 + · · · + cn g n
g ∗i ⟨v⟩ = ci
=⇒ v = {g ∗i ⟨v⟩}g i

5.2 Second dual space
5.2 Second dual space

It follows that there exists linear isomorphisms from a given linear space V to its
dual V ∗ . Now one can associate a natural isomorphism from V ∗ to its second dual,
i.e. to the dual V ∗∗ of the dual V ∗ of V. Then
V ∗∗ := Hom (V ∗ , R).
dim V ∗∗ = dim Hom (V ∗ , R) = dim V ∗ · dim R = dim V ∗ = dim V.

We consider a basis {g ∗∗ } ∈ V ∗∗ , a dual basis of {g ∗ } ∈ V ∗ . Then,
g ∗∗ ∗
i ⟨g j ⟩ = δij . (50)
Comparing (49)(b), and (50) we write
g ∗∗ ∗ ∗
i ⟨g j ⟩ = g i ⟨g j ⟩. (51)
The above relation naturally connects with g ∗∗i to g i , and for a finite dimensional
space, a natural isomorphism is established between V, and V ∗∗ . Such an isomor-
phism is also known as canonical isomorphism. It could be shown that in such a
scenario the double dual space becomes exactly same as the original space, i.e.,
V ∗∗ = V. We then identify, g ∗∗
i = g i , and then
g ∗i ⟨g j ⟩ = g i ⟨g ∗j ⟩ and v ∗ ⟨v⟩ = v⟨v ∗ ⟩. (52)
Moreover, it is a standard practice to write
g ∗i ≡ g i . (53)
We call v = v i g i ∈ V a contravariant vector, and v ∗ = vi g i ∈ V ∗ , a covariant

vector or covector.
vi : ith contravariant component of v,

vi : ith covariant component of v.
and
{g i } : is called covariant basis,

{g i } : is called contravariant basis.
Example 29 Consider the basis

2 3
B = {g 1 , g 2 } := , ∈ V = R2
1 1
Find its dual basis B = {g 1 , g 2 } ∈ V ∗ .

5.3 Transposition
Ans: Let v ∈ V, and our goal is to find g 1 ⟨v⟩ and g 1 ⟨v⟩, i.e., the explicit form of
the mapping or function. Let us consider the standard basis {e1 , e2 } ∈ V. Then,
g 1 ⟨v⟩ = g 1 ⟨v1 e1 + v2 e2 ⟩ = v1 g 1 ⟨e1 ⟩ + v2 g 1 ⟨e2 ⟩
If we know g 1 ⟨e1 ⟩ and g 2 ⟨e2 ⟩, then the explicit form for g 1 ⟨v⟩ is known. Consid-
ering the properties of (48) one can write
g 1 ⟨g 1 ⟩ = g 1 ⟨2e1 + 1e2 ⟩ = 2g 1 ⟨e1 ⟩ + 1g 1 ⟨e2 ⟩ := 1

g 1 ⟨g 2 ⟩ = g 1 ⟨3e1 + 1e2 ⟩ = 3g 1 ⟨e1 ⟩ + 1g 1 ⟨e2 ⟩ := 0
Solving the two equations we get g 1 ⟨e1 ⟩ = −1, and g 1 ⟨e2 ⟩ = 3. Then
g 1 ⟨v⟩ = −v1 + 3v2
Similarly,
g 2 ⟨v⟩ = g 2 ⟨v1 e1 + v2 e2 ⟩ = v1 g 2 ⟨e1 ⟩ + v2 g 2 ⟨e2 ⟩
Considering the properties of (48) one can write
g 2 ⟨g 1 ⟩ = g 2 ⟨2e1 + 1e2 ⟩ = 2g 2 ⟨e1 ⟩ + 1g 2 ⟨e2 ⟩ := 0

g 2 ⟨g 2 ⟩ = g 2 ⟨3e1 + 1e2 ⟩ = 3g 2 ⟨e1 ⟩ + 1g 2 ⟨e2 ⟩ := 1
Solving the above two equations we get g 2 ⟨e1 ⟩ = 1, and g 2 ⟨e2 ⟩ = −2. Then
g 2 ⟨v⟩ = v1 − 2v2
5.3 Transposition
Let v ∈ V and {g i } be a basis of V. Let v ∗ ∈ V ∗ , and {g i } be the dual basis in
V ∗ (Fig. 23). Then,
v = vigi and v ∗ = vi g i
Now consider u ∈ U and {f i } be a basis of U. Let u∗ ∈ U ∗ , and {f i } be the dual

basis in U ∗ . Then,
u = ui f i and u∗ = ui f i
We define,
v ∗ ⟨v⟩ = u∗ ⟨u⟩. (54)

5.3 Transposition
Figure 23: Different mappings and functionals that induce a transposition.
Let us further consider that T : V → U, and T t : U ∗ → V ∗ . i.e,

T v = u, and T t u∗ = v ∗
then (54) could be rewritten as,
(T t u∗ )⟨v⟩ = u∗ ⟨T v⟩. (55)
We call T t is the transposition or transpose of T . As T : V → U, then g i ∈ V,
means T g i ∈ U. So, T g i ∈ U could be expressed as a linear combination of the
basis of U. We write,
T g j = Tij f i .
Similarly, T t : U ∗ → V ∗ , then f i ∈ U ∗ , means T t f i ∈ V ∗ . So, T t f i ∈ V ∗ could be
expressed as a linear combination of the basis of V ∗ . We write,
T t f j = Tijt g i .
From (55), we can write
(T t f j )⟨g i ⟩ = f j ⟨T g i ⟩
t p
(Tpj g )⟨g i ⟩ = f j ⟨Tmi f m ⟩
t
Tpj [g p ⟨g i ⟩] = Tmi [f j ⟨f m ⟩]
t
Tpj δpi = Tmi δjm
Tijt = Tji

6 Tensor product
6.1 Bilinear functionals
Let V, U are two vector spaces. A map L : V × U → R is called bilinear if for any
v, v 1 , v 2 ∈ V, u, u1 , u2 ∈ U, and α1 , α2 ∈ F, it is linear in each variable; that is,
L⟨α1 v 1 + α2 v 2 , u⟩ = α1 L⟨v 1 , u⟩ + α2 L⟨v 2 , u⟩,

L⟨v, α1 u1 + α2 u2 ⟩ = α1 L⟨v, u1 ⟩ + α2 L⟨v, u2 ⟩.
The extension of this definition to functionals of more than two variables are
simple, and such functions are called multilinear functionals.
Theorem 6.1 Let V and U are two vector space and v ∈ V, u ∈ U. If v ∗ ∈ V ∗

and u ∈ U ∗ . Then the following construction
L⟨v, u⟩ = [v ∗ ⟨v⟩][u∗ ⟨u⟩] (56)
is bilinear.
Proof: Consider,
L⟨α1 v 1 + α2 v 2 , u⟩ = [v ∗ ⟨α1 v 1 + α2 v 2 ⟩][u∗ ⟨u⟩]

= [α1 v ∗ ⟨v 1 ⟩ + α2 v ∗ ⟨v 2 ⟩][u∗ ⟨u⟩]
= α1 [v ∗ ⟨v 1 ⟩][u∗ ⟨u⟩] + α2 [v ∗ ⟨v 2 ⟩][u∗ ⟨u⟩]
= α1 L(v 1 , u) + α2 L(v 2 , u)
Similarly,
L⟨v, α1 u1 + α2 u2 ⟩ = [v ∗ ⟨v⟩][u∗ ⟨α1 u1 + α2 u2 ⟩]

= [v ∗ ⟨v⟩][α1 u∗ ⟨u1 ⟩ + α2 u∗ ⟨u2 ⟩]
= α1 [v ∗ ⟨v⟩][u∗ ⟨u1 ⟩] + α2 [v ∗ ⟨v⟩][u∗ ⟨u2 ⟩]
= α1 L(v, u1 ) + α2 L(v, u2 )
Therefore, L is bilinear. There are also other bilinear functionals. However, this
particular form is our further interest to explore tensors. All such real valued
bilinear functionals (V × U → R) again form a vector space (see Fig. 24) when the
following rules further added:
(L1 + L2 )⟨v, u⟩ = L1 ⟨v, u⟩ + L2 ⟨v, u⟩

cL1 ⟨v, u⟩ = c · L1 ⟨v, u⟩

Figure 24: A tensor product space (vector space) V ∗ ⊗ U ∗ is generated due to

bilinear functionals V × U → R.
This space is generally denoted by the symbol V ∗ ⊗U ∗ , where ⊗ is a tensor product,

and you read V ∗ ⊗ U ∗ as V ∗ tensor U ∗ . We then write
L := v ∗ ⊗ u∗ ∈ V ∗ ⊗ U ∗
It should be noted that the tensor product is defined on V ∗ ⊗ U ∗ , not on V ⊗ U.
If v ∗ ⊗ u∗ ∈ V ∗ ⊗ U ∗ , then by definition,
v ∗ ⊗ u∗ ⟨v, u⟩ = [v ∗ ⟨v⟩][u∗ ⟨u⟩]
One can extend the above relation as
v ∗1 ⊗ · · · ⊗ v ∗n ⟨v 1 , · · · , v n ⟩ = [v ∗1 ⟨v 1 ⟩] · · · [v ∗n ⟨v n ⟩]
Let {g i } (i = 1, ..., n) and {f j } (j = 1, ..., n) form the basis of V and U, and let
{g i } (i = 1, ..., n) and {f j } (j = 1, ..., n) form the basis of V ∗ and U ∗ .
Lemma 6.2 Consider the following bilinear functional

V × U → R.
The collection of all such bilinear functionals generate the vector space V ∗ ⊗ U ∗ .
An element v ∗ ⊗ u∗ ∈ V ∗ ⊗ U ∗ could be then expressed as
v ∗ ⊗ u∗ = vi uj g i ⊗ f j
where g i ⊗ f j is a basis of V ∗ ⊗ U ∗ . Moreover, dim(V ∗ ⊗ U ∗ ) = dim V ∗ · dim U ∗ .
Proof. First we will show that g i ⊗ f j is linearly independent, and then will
show that it span the space.

Linearly independent
Consider
cij g i ⊗ f j = 0 ⊗ 0.
Applying the defined bilinear mapping at left hand side, we write
cij g i ⊗ f j ⟨g m , f n ⟩ = 0
cij [g i ⟨g m ⟩][f j ⟨f n ⟩] = 0
i
cij [δm ][δnj ] = 0 =⇒ cmn = 0
So g i ⊗ f j are linearly independent.
Span the space

Let us start with
v ∗ ⊗ u∗ ⟨v, u⟩ = [v ∗ ⟨v⟩][u∗ ⟨u⟩]
= [vi g i ⟨v j g j ⟩][um f m ⟨un f n ⟩]
= [vi v j g i ⟨g j ⟩][um un f m ⟨f n ⟩]
= [vi v j δij ][um un δmn ]
= [vi v i ][um um ] (57)
Moreover,
g i ⊗ f j ⟨v, u⟩ = [g i ⟨v⟩][f i ⟨u⟩]
= [v i ][uj ]
= v i uj (58)
From (57) and (58), we finally write
v ∗ ⊗ u∗ ⟨v, u⟩ = vi uj g i ⊗ f j ⟨v, u⟩
∴ v ∗ ⊗ u∗ = vi uj g i ⊗ f j (59)
as v and u are arbitrary vectors. So g i ⊗ f j is a basis. Thus the total number of

basis elements are dim V ∗ · dim U ∗ = dim(V ∗ ⊗ U ∗ ).
V ∗ × U ∗ → R.
The collection of all such bilinear functionals generate the vector space V ⊗ U. An
element v ⊗ u ∈ V ⊗ U could be then expressed as
v ⊗ u = v i uj g i ⊗ f j
where g i ⊗ f j is a basis of V ⊗ U. Moreover, dim(V ⊗ U) = dim V · dim U.

Consider
cij g i ⊗ f j = 0 ⊗ 0.
Applying the defined bilinear mapping at left hand side, we write
cij g i ⊗ f j ⟨g m , f n ⟩ = 0
cij [g i ⟨g m ⟩][f j ⟨f n ⟩] = 0
cij [δm
i
][δnj ] = 0 =⇒ cmn = 0
So g i ⊗ f j are linearly independent.
v ⊗ u ⟨v ∗ , u∗ ⟩ = [v⟨v ∗ ⟩][u⟨u∗ ⟩]
= [v i g i ⟨vj g j ⟩][um f m ⟨un f n ⟩]
= [v i vj g i ⟨g j ⟩][um un f m ⟨f n ⟩]
= [v i vj δij ][um un δmn ]
= [vi v i ][um um ] (60)
Moreover,
g i ⊗ f j ⟨v ∗ , u∗ ) = [g i ⟨v ∗ ⟩][f j ⟨u∗ ⟩]
= [vi ][uj ]
= vi uj (61)
From (60) and (61), we finally write
v ⊗ u ⟨v ∗ , u∗ ⟩ = v i uj g i ⊗ f j ⟨v ∗ , u∗ ⟩
∴ v ⊗ u = v i uj g i ⊗ f j . (62)
as v and u are arbitrary vectors. So g i ⊗ f j is a basis. Thus the total number of

basis elements are dim V · dim U = dim(V ⊗ U). Similarly,

V ∗ × U → R.
The collection of all such bilinear functionals generate the vector space V ⊗ U ∗ . An
element v ⊗ u∗ ∈ V ⊗ U ∗ could be then expressed as
v ⊗ u ∗ = v i uj g i ⊗ f j
where g i ⊗ f j is a basis of V ⊗ U ∗ . Moreover, dim(V ⊗ U ∗ ) = dim V · dim U ∗ .

6.2 Tensor product as homomorphism
and
V × U ∗ → R.
The collection of all such bilinear functionals generate the vector space V ∗ ⊗ U. An
element v ∗ ⊗ u ∈ V ∗ ⊗ U could be then expressed as
v ∗ ⊗ u = vi uj g i ⊗ f j
where g i ⊗ f j is a basis of V ∗ ⊗ U. Moreover, dim(V ∗ ⊗ U) = dim V ∗ · dim U.

Definition 6.1 For any vectors v ∈ V and u ∈ U, the tensor product v ⊗ u is
also defined as a map
v ⊗ u : U ∗ → V.
such that
(v ⊗ u)w∗ = (u⟨w∗ ⟩)v
for w∗ ∈ U ∗ . Then the mapping is homomorphism (try it yourself )5 , and g i ⊗ f j

is a basis of Hom (U ∗ , V).
Consider cij (g i ⊗ f j ) = 0 ⊗ 0 ∈ Hom (U ∗ , V). Then
cij (g i ⊗ f j )f m = (0 ⊗ 0)f m
cij [f j ⟨f m ⟩]g i = 0
cij δjm g i = 0
cim g i = 0
Now for each m, cim = 0 as {g i } is a basis. So, g i ⊗ f j is linearly independent.

5
Take an element αw∗1 + βw∗2 ∈ U ∗ and proceed

Span the space

Now,
(v ⊗ u)u∗ = [⟨u⟩u∗ ]v = uj uj v = uj uj v i g i
Moreover,
(g i ⊗ f j )u∗ = [⟨f j ⟩u∗ ]g i = uj g i
Substituting back in the first expression, we write
(v ⊗ u)u∗ = uj uj v i g i = v i uj (uj g i ) = v i uj (g i ⊗ f j )u∗ = (v i uj g i ⊗ f j )u∗
As u∗ is arbitrary in the above expression, we finally write
v ⊗ u = v i uj g i ⊗ f j . (63)
and so gi ⊗ f j spans the space Hom (U ∗ , V). Comparing (62) and (63) we see that
the two spaces V ⊗ U and Hom (U ∗ , V) share the same element, same basis and
also the same dimension. We thus identify these two spaces are equal. i.e,
v ⊗ u ∈ V ⊗ U = Hom (U ∗ , V).
Similarly,
Definition 6.2 For any vectors v ∈ V and u ∈ U, the tensor product v ∗ ⊗ u∗ is

defined as a homomorphism
v ∗ ⊗ u∗ : U → V ∗
such that if w ∈ U, then
(v ∗ ⊗ u∗ )w = (u∗ ⟨w⟩)v ∗
and
v ∗ ⊗ u∗ ∈ U ∗ ⊗ V ∗ = Hom (U, V ∗ ).
Definition 6.3 For any vectors v ∈ V and u∗ ∈ U ∗ , the tensor product v ⊗ u∗ is
v ⊗ u∗ : U → V
such that if w ∈ U, then
(v ⊗ u∗ )w = (u∗ ⟨w⟩)v
and
v ⊗ u∗ ∈ V ⊗ U ∗ = Hom (U, V).

Finally,
Definition 6.4 For any vectors v ∗ ∈ V ∗ and u ∈ U, the tensor product v ∗ ⊗ u is
v∗ ⊗ u : U ∗ → V ∗
such that if w∗ ∈ U ∗ , then
(v ∗ ⊗ u)w∗ = (u⟨w∗ ⟩)v ∗
and
v ∗ ⊗ u ∈ V ∗ ⊗ U = Hom (U ∗ , V ∗ ).
We can extend this to the higher order spaces.
Theorem 6.6 Consider the basis {g i } ∈ V and {f i } ∈ U. Then {g i ⊗ f j } is
a basis of V ⊗ U = Hom (U ∗ , V). Then any T ∈ V ⊗ U could be expressed as
T = T ij g i ⊗ f j .
As g i ⊗ f j is a basis of Hom (U ∗ , V), any T ∈ Hom (U ∗ , V) can be expressed as
T = T ij g i ⊗ f j . In summary,
T = T ij g i ⊗ f j ∈ V ⊗ U ≡ Hom (U ∗ , V)
T̂ = T̂ij g i ⊗ f j ∈ V ∗ ⊗ U ∗ ≡ Hom (U, V ∗ )
T̆ = T̆i·j g i ⊗ f j ∈ V ∗ ⊗ U ≡ Hom (U ∗ , V ∗ )
Ť = Ť·ji g i ⊗ f j ∈ V ⊗ U ∗ ≡ Hom (U, V)
Note that Ti·j means i is the first index. A dot (·) is used to denote the second
index. T ij , T̂ij , T̆i·j , and Ť·ji are the components with respect to the associated
basis. Also note that Tj·i ̸= T·ji .
Example 30 Let u ⊗ v ∈ U ⊗ V, and v ∗ ⊗ w ∈ V ∗ ⊗ W. Then
(u ⊗ v)(v ∗ ⊗ w) = (u ⊗ w)[v⟨v ∗ ⟩] (64)
Ans: Note that the left hand side is the composition mapping or product of two
homomorphisms. Consider the following commutative diagram,
v ∗ ⊗w / V∗
W∗ (65)
u⊗v
(u⊗v )◦(v ∗ ⊗w )=(u⊗v )(v ∗ ⊗w )

U

6.3 Matrix representation of tensor products
If p∗ ∈ W ∗ , then
(u ⊗ v)(v ∗ ⊗ w)p∗ = (u ⊗ v) ◦ (v ∗ ⊗ w)p∗ = (u ⊗ v)((v ∗ ⊗ w)p∗ )

= (u ⊗ v)([w⟨p∗ ⟩]v ∗ )
= u[v ∗ ⟨v⟩][w⟨p∗ ⟩]
= (u[w⟨p∗ ⟩])[v⟨v ∗ ⟩]
= ((u ⊗ w)p∗ )[v⟨v ∗ ⟩]

= (u ⊗ w)[v⟨v ∗ ⟩] p∗
6.3 Matrix representation of tensor products

Consider the tensor
T = Tij f i ⊗ g j ∈ U ∗ ⊗ V ∗ . (66)
We assume dim U = dim V = 2. The basis {f 1 , f 2 } ∈ U ∗ , and the basis {g 1 , g 2 } ∈

V ∗ . Also note that the basis {f 1 , f 2 } ∈ U, and the basis {g 1 , g 2 } ∈ V. Tij are the
components of the tensor in (66), and for this particular case we can write
T = T11 f 1 ⊗ g 1 + T12 f 1 ⊗ g 2 + T21 f 2 ⊗ g 1 + T22 f 2 ⊗ g 2 .
As,
T : V → U ∗,
we write, T g j = Tij f i , i.e.,
T g 1 = T11 f 1 + T21 f 2
T g 2 = T12 f 1 + T22 f 2 .
So in the matrix form,

1
T g1 T11 T21 f
= (67)
T g2 T12 T22 g −f i f 2
i
Now consider T v = u∗ , and we write,
T (v j g j ) = um f m
v j (T g j ) = um f m
v j Tmj f m = um f m
=⇒ v j Tmj = um (68)

6.4 Tensor spaces: a general multilinear function
So,
u1 = T11 v 1 + T12 v 2
u2 = T21 v 1 + T22 v 2
In a matrix notation,

u1 T11 T12 v 1
=
u2 T21 T22 v 2
which is transpose of (67). Similarly, if we consider u∗ ⊗ v ∗ = ui vj f i ⊗ g j , then

the matrix form reads

∗ ∗ u1 v1 u2 v1
(u ⊗ v )ij = .
u1 v2 u2 v2 g −f i
i
Now,

u1 v1 u1 v2 v 1 u1 (v1 v 1 + v2 v 2 ) u
∗ ∗
[(u ⊗ v )v]j = 2 = 1 2 = [v ⟨v⟩] 1 .
∗
u2 v1 u2 v2 v u2 (v1 v + v2 v ) u2
i.e., the component representation of (u∗ ⊗ v ∗ )v = [v ∗ ⟨v⟩]u.
6.4 Tensor spaces: a general multilinear function

The vector spaces in multilinear mapping only with V or V ∗ are called tensor
spaces. In general a multilinear function on V ∗ × ... × V ∗ × V × ... × V forms a
tensor of type (r, s) (r, V ∗ and s, V) denoted by
Tsr = V ⊗ ... ⊗ V ⊗ V ∗ ⊗ ... ⊗ V ∗ .
| {z } | {z }
r times s times
A tensor of type (r, 0) is called contravariant tensor and one of type (0, s) is called
covariant tensor. Suppose {g i } and {g i } are the bases of V and V ∗ respectively,
then
g i1 ⊗ · · · ⊗ g ir ⊗ g j1 ⊗ · · · ⊗ g js
form a basis of Tsr . Therefore an (r, s)-type tensor T can be uniquely expressed as
T = T·ij11...i j1 js
...j s g i1 ⊗ · · · ⊗ g ir ⊗ g ⊗ · · · ⊗ g ∈ Ts
r r
Note that this is one particular format that is often useful. You have to start with
covariant basis and contravariant basis will come after them. You can not have a
suffling of covariant and contravariant basis.
Example 31 T ∈ T12 implies T = T·kij g i ⊗ g j ⊗ g k .

6.5 Tensor multiplication
6.5 Tensor multiplication

The tensor multiplication of tensors from Tsr11 and Tsr22 is defined as a mapping
Tsr11 ⊗ Tsr22 → Tsr11+s

+r2
2
So, if T ∈ Tsr11 , and S ∈ Tsr22 , then
T ⊗ S ∈ Tsr11+s
+r2
2
This means if
i ...i
T = T· j11 ...jr1s1 g i1 ⊗ · · · ⊗ g ir1 ⊗ g j1 ⊗ · · · ⊗ g js1 ∈ Tsr11
and
m ...m
S = S· n11 ...nsr22 g m1 ⊗ · · · ⊗ g mr2 ⊗ g n1 ⊗ · · · ⊗ g ns2 ∈ Tsr22
then
i ...i m ...m
T ⊗ S = T· j11 ...jr1s1 S· n11 ...nsr22 g i1 ⊗ · · · ⊗ g ir1 ⊗ g m1 ⊗ · · · ⊗ g mr2
⊗g j1 ⊗ · · · ⊗ g js1 ⊗ g n1 ⊗ · · · ⊗ g ns2 ∈ Tsr11+s
+r2
2
.
The following relations now hold true.
1. Tensor multiplication is associative: Let A ∈ Tsr11 , B ∈ Tsr22 , and C ∈

Tsr33 . Then
A ⊗ (B ⊗ C) = (A ⊗ B) ⊗ C ∈ Tsr11+s
+r2 +r3
2 +s3
2. Tensor multiplication is distributive with respect to addition: Let

A ∈ Tsr11 , B ∈ Tsr22 , and C ∈ Tsr22 . Then
A ⊗ (B + C) = A ⊗ B + A ⊗ C ∈ Tsr11+s
+r2
2
3. For each α ∈ F:
α(A ⊗ B) = αA ⊗ B = A ⊗ αB ∈ Tsr11+s
+r2
2
when A ∈ Tsr11 and B ∈ Tsr22 .
The vector space equipped with such an addition and multiplication is called tensor
algebra.

6.6 Contraction
6.6 Contraction
Consider tensors of type Tsr or (r, s) such that r, s ≥ 1.
Definition 6.5 The contraction of a tensor is an operation which equalizes a con-

travariant index and a covariant index and sum over the repeated index.
We can also think of contraction as a mapping
[·] : Tsr → Ts−1

r−1
r−1
such that a tensor space Tsr becomes Ts−1 . For example, let us consider the tensor
T = T·kmp g m ⊗ g p ⊗ g k ∈ T12
Contracting p and k we obtain a tensor um such that
·T = T·kmp g m [g p ⟨g k ⟩] = T·kmp g m [δpk ] = (T·pmp )g m = um g m ∈ T01
Note that:
1. Every contraction of a tensor removes one contravariance and one covariance.
2. After q contractions a tensor of type (q, q) is reduced to a tensor (0, 0).

One needs to be specific about the contracting basis elements if multiple options
are there. Otherwise, one should start contracting from the last.
A double contraction is a mapping
[:] : Tsr → Ts−2

r−2
r−2
such that a tensor space Tsr becomes Ts−2 . The symbol ‘:’ is used when two
contractions take place. Consider,
mp
T = T·kl g m ⊗ g p ⊗ g k ⊗ g l ∈ T22
Then
mp mp m
: T = T·kl [g m ⟨g k ⟩][g p ⟨g l ⟩] = T·kl [δk ][δlp ] = T·kl
kl
∈ T00
A n-contraction is a mapping
[•] : Tsr → Ts−n

r−n
r−n
such that a tensor space Tsr becomes Ts−n . Unfortunately, the symbol ‘•’ (or
sometimes :) is used in most of the time beyond double contraction. You need to
understand it from the context.

6.6 Contraction
Single contraction between two tensors are defined in the following way. First
we will consider for simplicity a pure contravariant (r1 , 0) and a pure covariant
tensor (0, s2 ) such that
−1
[·] : T0r1 × Ts02 → Tsr21−1
A single dot ‘·’ is used for a single contraction. For example, if T = T ijk g i ⊗g j ⊗g k
and S = Sqr g q ⊗g r , then, a single contraction between r and k (start from extreme
right) is given by
T · S = (T ijk g i ⊗ g j ⊗ g k ) · (Sqr g q ⊗ g r )
= T ijk Sqr g i ⊗ g j ⊗ g q [g k ⟨g r ⟩]
= T ijk Sqr g i ⊗ g j ⊗ g q [δrk ]
= T ijk Sqk g i ⊗ g j ⊗ g q
If nothing is mentioned, it is a standard practice to start from the last member.
A double contraction between two pure contravariant (r1 , 0) and a pure covariant
tensor (0, s2 ) such that
−2
[:] : T0r1 × Ts02 → Tsr21−2
Continuing the previous example we write
T : S = (T ijk g i ⊗ g j ⊗ g k ) : (Sqr g q ⊗ g r )
= T ijk Sqr g i [g j ⟨g q ⟩][g k ⟨g r ⟩]
= T ijk Sqr g i [δjq ][δrk ]
= T ijk Sjk g i
A generalized full contraction is given by
[•] : T0r1 × Ts02 → T0r1 −s2 if r1 > s2 or Ts02 −r1 if s2 > r1 .
If T = T ijkl g i ⊗ g j ⊗ g k ⊗ g l and S = Sqr g q ⊗ g r , then
T • S = (T ijkl g i ⊗ g j ⊗ g k ⊗ g l ) : (Sqr g q ⊗ g r )
= T ijkl Sqr g i ⊗ g j [g k ⟨g q ⟩][g l ⟨g r ⟩]
= T ijkl Sqr g i ⊗ g j [δkq ][δlr ]
= T ijkl Skl g i ⊗ g j
The following double contraction is equivalent to a bilinear functional of a tensor
product.
u ⊗ v : u∗ ⊗ v ∗ = u ⊗ v ⟨u∗ , v ∗ ⟩ = [u⟨u∗ ⟩][v⟨v ∗ ⟩].

6.6 Contraction
For example, if T = u ⊗ v ⊗ w ⊗ x ⊗ y and S = a∗ ⊗ b∗ ⊗ c∗ , where a∗ , b∗ , c∗ ∈ V ∗

and u, v, w, x, y ∈ V, then a 3-contraction (generalized) could be written as
(u ⊗ v ⊗ w ⊗ x ⊗ y) • (a∗ ⊗ b∗ ⊗ c∗ ) = u ⊗ v[w⟨a∗ ⟩][x⟨b∗ ⟩][y⟨c∗ ⟩]
However, as explained in Example –,
(u ⊗ v ⊗ w ⊗ x ⊗ y)(a∗ ⊗ b∗ ⊗ c∗ ) = u ⊗ v ⊗ w ⊗ x ⊗ b∗ ⊗ c∗ [y⟨a∗ ⟩]

7 Inner product space
7.1 Inner product
We may think of a vector as a geometric object. Then it makes sense to talk about
the length of a vector and angle between two vectors. We will now study a certain
type of scalar valued bilinear functional on pair of vectors, known as inner product.
Definition 7.1 An inner product is a map
g :V ×V →R
with the following properties: For u, v ∈ V, and α ∈ R
1. g⟨u, v⟩ is bilinear,
2. g⟨u, v⟩ = g⟨v, u⟩,
3. g⟨u, u⟩ > 0, if u ̸= 0.
For such a bilinear mapping, g ∈ V ∗ ⊗ V ∗ , and if {g 1 , g 2 , · · · , g n } be a basis of V ∗ ,

then we can write
g = gij g i ⊗ g j
Now,
g⟨g i , g j ⟩ = gmn g m ⊗ g n ⟨g i , g j ⟩
= gmn [g m ⟨g i ⟩][g n ⟨g j ⟩]
= gmn δim δjn
= gij (69)
Moreover,
g⟨g i , g j ⟩ = g⟨g j , g i ⟩
=⇒ gij = gji (70)
We call g⟨g i , g j ⟩ the inner product of g i and g j and the notation is
g⟨g i , g j ⟩ = g i · g j = gij
and
gi · gj = gj · gi.

7.2 Reciprocal basis
A vector space V equipped with the inner product is called an inner product space.
We will denote such space as ·V. By including the inner product, we have an
alternative of the dual space V ∗ to produce scalars. However, dual space carries
functionals where as the inner product space carries the vectors!
g is known as metric tensor or fundamental tensor. gij is called the components
of the metric tensor or fundamental tensor as it is directly linked with the measure
of distance in a generalized way.

We note that the dual space dissolves in the actual vector space when equipped
with the inner product. However, if {g 1 , g 2 , · · · , g n } be a basis of V, we can
construct a set of vectors {g 1 , g 2 , · · · , g n } such that
g i · g j = δji
where δji is called the Kronecker delta. The basis {g 1 , g 2 , · · · , g n } spans the same
vector space V again. Such a basis is known as reciprocal basis.
From this construction, if v = v i g i ∈ V, then by taking the inner product with
i
g we have
v · g i = (v j g j ) · g i = v j (g j · g i ) = v j δji = v i .
This gives the ith component of v in the {g 1 , g 2 , · · · , g n } basis system. Let u =

ui g i and v = v j g j belong to V. Then
u · v = (ui g i ) · (v j g j ) = ui v j (g j · g j ) = ui v j gij = ui (v · g j )gij = (ui gij g j ) · v.
This implies that u can also be expressed as u = uj g j . By comparing with the

above relation we get
uj = gij ui .
Since we can express u as u = ui g i w.r.t the basis {g i } or as u = ui g i w.r.t the

basis {g i }, then
u = ui g i = ui g i .
We call
ui : ith contravariant component of u,

ui : ith covariant component of u.

and
{g i } : is called covariant basis,

{g i } : is called contravariant basis.
Similarly, we define g = g ij g i ⊗ g j such that
g⟨g i , g j ⟩ = g ij = g i · g j .
Then
u · v = (ui g i ) · (vi g j ) = ui vj (g j · g j ) = ui vj g ij = (ui g ij g j ) · v.
So, we can write
uj = g ij ui .
The two operators
gij : uj → ui , g ij : uj → ui
enable us to lower and raise the component index. Moreover,
g i = gij g j , or g i = g ij g j .
Therefore, lowering or raising the index for the reciprocal basis can be made in
the same manner. Note that
g k · g i = gij g k · g j
δik = gij g kj = gij g jk .
This means [g ij ] = [gij ]−1 . In a tensor space with inner product, for a, b, c,
d, x, y ∈ ·V we write
1. (a ⊗ b)c = (b · c)a
2. (a ⊗ b) : (c ⊗ d) = (a · c)(b · d)
3. (a ⊗ b)(c ⊗ d) = (a ⊗ d)(b · c)
4. (a ⊗ b ⊗ c)d = (a ⊗ b)(c · d)
5. (a ⊗ b ⊗ c) : x ⊗ y = a(b · x)(c · y)

Note that, all the dual space operations are now represented by the dot · instead
of ⟨ ⟩.
For T ∈ ·V ⊗ ·V, we can write it in four possible ways:
T = T ij g i ⊗ g j = Tij g i ⊗ g j = Ti·j g i ⊗ g j = T·ji g i ⊗ g j
All the components T ij , Tij , Ti·j , and T·ji are different. T ij , Ti·j , and T·ji are also
called associate tensor components of Tij .If we consider T = T ij g i ⊗ g j , then
T g k = T ij (g i ⊗ g j )g k = T ij (g j · g k )g i = T ij δjk g i = T ik g i
=⇒ g m · T g k = g m · T ik g i = T ik g m · g i = T ik δim = T mk
In summary,
g i · T g j = T ij , g i · T g j = Tij , g i · T g j = T·ji , g i · T g j = Ti·j
Now,
g i · T g j = T ij , (g im g m ) · T (g jn g n ) = T ij , g im g jn (g m · T g n ) = T ij
=⇒ g im g jn Tmn = T ij .
Similarly, knowing the components with respect to a basis, one can directly cal-
culate components with respect to other basis through gij and g ij . We also have
four versions of g s:
g = g ij g i ⊗ g j
= gij g i ⊗ g j
= δi·j g i ⊗ g j = g i ⊗ g i
= δ·ji g i ⊗ g j = g i ⊗ g i
You operate g on any vector v i g i or vi g i , you will get that vector back, i.e., gv = v.
So g is indeed an identity tensor!
7.2.1 Length of a vector

The norm or length of a vector x ∈ V is defined as
p √
|x| := g⟨x, x⟩ = x · x
A vector space equipped with such a norm is called Euclidean vector space.

7.2.2 Angle between two vectors

The angle θ(x, y) between two non-zero vectors x, y ∈ V is defined as
x·y
cos θ(x, y) :=
|x||y|
7.2.3 Differential arc length

The differential arc length along a curve is given by
√
ds = dx · dx
7.2.4 Transposition of a tensor in an inner product space
Figure 25: Transpose of a linear operator in inner product space.
The notion of transpose is similar to what we studied in the dual space. How-
ever, we have only one vector space ·V. Consider,
T : ·V → ·V
such that v and T v belong to ·V. Let a and u be the any other two vectors such
that v · a = T v · u. Then the transpose is the mapping T t u = a, and satisfies the
relation (Fig. 25)
v · T t u = T v · u. (71)
If v = g i , and u = g j , then form the above equation,
gi · T tgj = T gi · gj =⇒ [T t ]ij = [T ]ji

7.3 Physical interpretation of gij

Consider ·V = R3 and position of a point with respect to the standard basis
{e1 , e2 , e3 } is denoted by
r = x1 e1 + x2 e2 + x3 e3
The correspondence between points (x1 , x2 , x3 ) and the three scalar parameters
Figure 26: A local (curvilinear) coordinate system. u1 , u2 , u3 are the coordinate

lines.
(u1 , u2 , u3 ) defines a local coordinate system in ·V. We write
x1 = ϕ1 (u1 , u2 , u3 ), x2 = ϕ2 (u1 , u2 , u3 ), x3 = ϕ3 (u1 , u2 , u3 ).
We assume that the mappings are bijective and have continuous derivatives, so
that the correspondence between (x1 , x2 , x3 ) and (u1 , u2 , u3 ) is unique.
If we only vary one local coordinate, say u1 , while the other two local coordi-
nates, (u2 , u3 ), are fixed, then we call the locus of the points (x1 , x2 , x3 ) a coordinate
line. So
cu1 = {ϕ1 (u1 , c2 , c3 ), ϕ2 (u1 , c2 , c3 ), ϕ3 (u1 , c2 , c3 )} → u1 coordinate line

cu2 = {ϕ1 (c1 , u2 , c3 ), ϕ2 (c1 , u2 , c3 ), ϕ3 (c1 , u2 , c2 )} → u2 coordinate line
cu3 = {ϕ1 (c1 , c2 , u3 ), ϕ2 (c1 , c2 , u3 ), ϕ3 (c1 , c2 , u3 )} → u3 coordinate line
are the three coordinate lines (Fig. 26). If the coordinate lines are straight,
we call it rectilinear coordinate system, else it is a curvilinear coordinate system.
Consider a point P ′ on the coordinate line u1 , obtained by a small change δu1 .

Then the position vector of the new point with respect to the standard basis could
be written as
OP ′ = ϕ1 (u1 + δu1 , c2 , c3 )e1 + ϕ2 (u1 + δu1 , c2 , c3 )e2 + ϕ3 (u1 + δu1 , c2 , c3 )e3 .
Figure 27: P P ′ becomes tangent as δu1 → 0.
So, as per Fig. 27,
P P ′ = OP ′ − OP
= ϕ1 (u1 + δu1 , c2 , c3 )e1 + ϕ2 (u1 + δu1 , c2 , c3 )e2 + ϕ3 (u1 + δu1 , c2 , c3 )e3

− ϕ1 (u1 , c2 , c3 )e1 + ϕ2 (u1 , c2 , c3 )e2 + ϕ3 (u1 , c2 , c3 )e3

∂ϕ1 1 ∂ϕ2 1 ∂ϕ3 1

= δu e 1 + δu e 2 + δu e3 .
∂u1 ∂u1 ∂u1
Then,
PP′ ∂x1 ∂x2 ∂x3

lim = g 1 = e 1 + e 2 + e3
δu1 →0 δu1 ∂u1 ∂u1 ∂u1
∂
= (x1 e1 + x2 e2 + x3 e3 )
∂u1
∂r
=
∂u1
∂r ∂r
Similarly, g 2 = ∂u 2 and g 3 = ∂u3 . These three local tangents vectors
n ∂r ∂r ∂r o
{g 1 , g 2 , g 3 } := 1
, , (72)
∂u ∂u2 ∂u3
are known as covariant basis or natural basis (Fig. 28). As
r(u1 , u2 , u3 ) = x1 (u1 , u2 , u3 )e1 + x2 (u1 , u2 , u3 )e2 + x3 (u1 , u2 , u3 )e3 (73)

Figure 28: Covariant or tangent basis vectors {g 1 , g 2 , g 3 }.
we write
∂r 1 ∂r ∂r
dr(u1 , u2 , u3 ) = 1
du + 2 du2 + 3 du3
∂u ∂u ∂u
= du1 g 1 + du2 g 2 + du3 g 3 . (74)
So,
dr · dr = (dui g i ) · (duj g j ) = (g i · g j )dui duj
or
ds2 = dr · dr = gij dui duj . (75)
Here ds is the differential arc-length distance along dr with respect to the local
coordinates {u1 , u2 , u3 }. Recall that, gij (u1 , u2 , u3 ) is known as metric tensor or
fundamental tensor. It is associated with the measure of distance in a generalized
way.
1. Riemannian space (gij is positive definite i.e. eigenvalues are ≥ 0)

(a) Non-Euclidean space: curved and Gaussian curvature ̸= 0
(b) Euclidean: flat and Gaussian curvature = 0
2. Semi/Pseudo Riemannian space (gij is negative definite i.e. at least one
eigenvalue is negative)

Figure 29: Covariant or tangent basis vectors {g 1 , g 2 } for the polar coordinates.
(a) Lorentz space: Only one eigenvalue of gij is negative
Example 32 Find ds2 for the polar coordinate system (Fig. 29)
We identify u1 = r and u2 = θ. Then,
x1 (r, θ) = r cos θ, x2 (r, θ) = r sin θ
The position vector is written as
r = x1 (r, θ)e1 + x2 (r, θ)e2 = r cos θe1 + r sin θe2
Therefore,
∂r
g1 = = cos θe1 + sin θe2
∂r
∂r
g2 = = −r sin θe1 + r cos θe2
∂θ
We compute
g11 = g 1 · g 1 = 1, g12 = g 1 · g 2 = 0, g21 = g 2 · g 1 = 0, g22 = g 2 · g 2 = r2
The arc-element with respect to the {g 1 , g 2 } system is
ds2 = gij dui duj

= g11 (du1 )2 + g22 (du2 )
= dr2 + (rdθ)2

Similarly, we can find metric to any curvilinear system!

As the correspondence between (x1 , x2 , x3 ) and (u1 , u2 , u3 ) is unique (bijective),
the inverse function could be written as
u1 = ψ 1 (x1 , x2 , x3 ), u2 = ψ 2 (x1 , x2 , x3 ), u3 = ψ 3 (x1 , x2 , x3 ).
Considering u3 = ψ 3 (x1 , x2 , x3 ) = c3 , some constant, then
Figure 30: g 1 is normal to the u3 = ψ 3 (x1 , x2 , x3 ) surface.
∂ψ 3 ∂ψ 3 ∂ψ 3
∇ψ 3 (x1 , x2 , x3 ) = e 1 + e 2 + e3
∂x1 ∂x2 ∂x3
or
∂u3 ∂u3 ∂u3
g3 = e 1 + e2 + e3 .
∂x1 ∂x2 ∂x3
Then g 3 = ∇u3 (x1 , x2 , x3 ) is perpendicular to the tangent plane 6 at P (Fig. 30).
6
Recall that gradient of a constant surface is normal to the tangent plane. Suppose we have a
surface xy 3 = z + 2. To find the normal to the surface at (1, 1, −1), we consider f = xy 3 − z = 2.
Then ∇f = y 3 e1 + 3xy 2 e2 − e3 = e1 + 3e2 − e3 is the normal vector at (1, 1, −1).
(1,1,−1)

Now,
∂u3 ∂u3 ∂u3 ∂x1 ∂x2 ∂x3
g3 · g3 = e +
1 1
e 2 + e3 · e 1 + e2 + e3 .
∂x ∂x2 ∂x3 ∂u3 ∂u3 ∂u3
∂u3 ∂x1 ∂u3 ∂x2 ∂u3 ∂x3
= + +
∂x1 ∂u3 ∂x2 ∂u3 ∂x3 ∂u3
∂u3
=
∂u3
= 1
Similarly,
∂u3 ∂u3 ∂u3 ∂x1 ∂x2 ∂x3
g3 · g2 = e +
1 1
e 2 + e3 · e 1 + e2 + e3 .
∂x ∂x2 ∂x3 ∂u2 ∂u2 ∂u2
∂u3 ∂x1 ∂u3 ∂x2 ∂u3 ∂x3
= + +
∂x1 ∂u2 ∂x2 ∂u2 ∂x2 ∂u2
∂u3
=
∂u2
= 0
and g 3 · g 1 = 0. Essentially we have,
g i · g j = δji (76)
Therefore {g 1 , g 2 , g 3 } is the reciprocal basis of {g 1 , g 2 , g 3 } (Fig. 31).
Example 33 Find the reciprocal basis {g 1 , g 2 } for the polar coordinates, as shown
in Fig. 29.
We first find
p x2
r= (x1 )2 + (x2 )2 , θ = tan−1
x1
So,
∂r ∂r
g 1 = ∇r = e1 + e2
∂x1 ∂x2
x1 x2
= p e1 + p e2
(x1 )2 + (x2 )2 (x1 )2 + (x2 )2
= cos θe1 + sin θe2

7.4 Affine space
Figure 31: Covariant {g 1 , g 2 , g 3 } and contravariant basis {g 1 , g 2 , g 3 } with respect

to a local coordinate system.
and
∂θ ∂θ
g 2 = ∇θ = 1
e1 + 2 e2
∂x ∂x
1
x x2 x2 1
= − p − 1 2 e1 + p − 1 e2
(x1 )2 + (x2 )2 (x ) (x1 )2 + (x2 )2 (x )
2 1
x x
= −p e1 + p e2
(x1 )2 + (x2 )2 (x1 )2 + (x2 )2
sin θ cos θ
= − e1 + e2
r r
1
Moreover, g 1 · g 1 = g 11 = 1, g 1 · g 2 = g 12 = g 21 = 0, g 2 · g 2 = g 22 = r2
.
7.3.1 Physical components of tensors

Discussed in the class
7.4 Affine space

Discussed in the class

7.5 Some applications to the continuum mechanics
Figure 32: Local coordinates in the reference (B0 ) and current configuration (Bt ).

Let B0 be a generic representation of a solid body at t = 0. It is also known as
reference configuration. After deformation, the body takes a new configuration Bt ,
known as current configuration. Let P ∈ B0 be an arbitrary point with position
vector
X = X 1 e1 + X 2 e2 + X 3 e3
with respect to the standard basis {e1 , e2 , e3 }. So the body is embedded in an

Euclidean space. In many cases, it is convenient and necessary to use curvilinear
(local) coordinates. When the problem domain or the geometry of the body might
be of a particular shape, e.g., a spherical shell, a membrane structure, a cylindrical
structure, etc., appropriate curvilinear system is often convenient and necessary.
We use (U 1 , U 2 , U 3 ) as a curvilinear system at P . Then,
X(U 1 , U 2 , U 3 ) = X 1 (U 1 , U 2 , U 3 )e1 + X 2 (U 1 , U 2 , U 3 )e2 + X 3 (U 1 , U 2 , U 3 )e3
We then write,
∂X
1. GI = ∂U I
2. GI = ∇X U I
3. Covariant basis: {G1 , G2 , G3 }
4. Contravariant basis: {G1 , G2 , G3 }

Similarly, we use (u1 , u2 , u3 ) as a curvilinear system at p ∈ Bt . Then,
x(u1 , u2 , u3 ) = x1 (u1 , u2 , u3 )e1 + x2 (u1 , u2 , u3 )e2 + x3 (u1 , u2 , u3 )e3
We then write,
∂x
1. g i = ∂ui
2. g i = ∇x ui
3. Covariant basis: {g 1 , g 2 , g 3 }
4. Contravariant basis: {g 1 , g 2 , g 3 }
We also call U I as material coordinates, and ui as spatial coordinates. During
deformation, consider the following maps between the material and spacial coor-
dinates.
u1 = ϕ1 (U 1 , U 2 , U 3 , t), u2 = ϕ2 (U 1 , U 2 , U 3 , t), u3 = ϕ3 (U 1 , U 2 , U 3 , t).
Or
uk = ϕk (U I , t). (77)
ϕk is also known as deformation map. Consider a neighborhood point P ′ ∈ B0

with position vector X + dX. Then,
dX = dU I GI .
The point P ′ maps to the point p′ with the position vector x + dx, and we write
dx = duk g k .
A differential change in the position vector in Bt could be written in terms of

covariant basis at p as,
∂ϕk
k I
dx = du g k = dU g k
∂U I
∂ϕk
I
= g ⊗ G dU I GI
∂U I k
∂ϕk
I
= g ⊗ G dX
∂U I k
We call
∂ϕk
F = g ⊗ GI (78)
∂U I k
Figure 33: A material line in the reference configuration passing through a point
P . At α = 0, we have the position vector of the point P , X = C(α = 0).
the deformation gradient. It simply maps a differential length dX ∈ B0 to a

differential length dx ∈ Bt , i.e.,
dx = F dX (79)
We can think of dX as a tangent to a curve C(α), passing through P ∈ B0 ,

such that C(α0 ) = X, where α ∈ R is a parameter (Fig. 33). C(α) is known as
material line in the reference configuration. Now, there are infinite-many material
lines passing through P such that dX could be a tangent to each curves, passing
through P . Total set of all material line elements dX, belonging to the point P
forms a vector space, the so-called tangent space (Fig. 34) TX (at P ), and dX ∈ TX
Similarly, c(α) is known as material line in the current configuration, and Tx
is the tangent space at x ∈ Bt , and dx ∈ Tx . Then
F : TX → Tx
i.e., deformation gradient is a homomorphism between the two tangent spaces

TX , and Tx . It maps a differential length dX to dx. Let dX, dY ∈ TX , and
dx, dy ∈ Tx such that (Fig. 35)
dx = F dX, and dy = F dY (80)
Consider,

Figure 34: Set of all material line elements dX, belonging to the point P forms a
vector space, known as tangent space TX .
Figure 35: Change in length and angle between material line elements.
dx · dy − dX · dY = F dx · F dy − dX · dY
= F dX · F dY − dX · dY
= dX · F t F dY − dX · dY
= dX · (F t F − I)dY
The quantity C = F t F is called right Cauchy-Green tensor. So, we write
dx · dy − dX · dY = dX · (C − I)dY
We further define (C − I) = 2E or E = 21 (C − I), the Green strain tensor. The

factor 2 or 1/2 is there to be consistent with the definition of engineering strain.

7.6 Orthonormal set
So, Green strain tensor E is a bilinear mapping
E : TX × TX → R (81)
such that,
E⟨dX, dY ⟩ = dX · 2EdY ∈ R
E represents the change in length and angle between material line elements. From
the bilinear mapping (81), we can say that E ∈ T∗X ⊗ T∗X , or rather in an inner
product space, E ∈ T# #
X ⊗ T#X , where TX is spanned by the reciprocal basis of
TX . We then write
E = EIJ GI ⊗ GJ
7.6 Orthonormal set

Let ·V be an inner product space. S ⊂ ·V be an orthonormal set if any two distinct
vectors in S are orthonormal. If si , sj ∈ S, then
si · sj = δij
and |si | = 1. If the magnitude is not unity, S is called orthogonal.
Theorem 7.1 If S = {s1 , · · · , sn }, is any finite orthonormal set in ·V such that

n = dim ·V, then S forms a basis in ·V, known as orthonormal basis.
Proof: S ⊂ ·V be an orthonormal set of vectors. Consider
α1 s 1 + · · · + αn s n = 0
Taking dot product with sk , we write αk |sk |2 = 0. Note that k is not summed,
rather it is just one term. As |sk |2 = 1, we can further write αk = 0. So, Sm is
linearly independent.
Moreover, If v ∈ ·V, then v = αk sk = (v · sk )sk (k = 1, n). So any arbitrary
vector in ·V could be spanned by S.
7.7 Gram-Schmidt orthonormalization

Let BV = {g 1 , . . . , g n } be a basis of ·V. We will now construct a set of orthonormal
basis (Fig. 36) B̂V = {s1 , . . . , sn } from BV = {g 1 , . . . , g n }.
g
1. Construct s1 = |g 1 |
1

7.7 Gram-Schmidt orthonormalization
Figure 36: Change in length and angle between material line elements.
2. Consider h2 = g 2 − (g 2 · s1 )s1 . Note that h2 · s1 = g 2 · s1 − (g 2 · s1 ) = 0, so

h2 ⊥ s1 . Construct
h2
s2 =
|h2 |
3. Consider h3 = g 3 − (g 3 · s1 )s1 − (g 3 · s2 )s2 . Note that h3 · s1 = g 3 · s1 − (g 3 ·

s1 )(s1 · s1 ) − (g 3 · s2 )(s2 · s1 ) = 0, so h3 ⊥ s1 . Similarly, h3 · s2 = 0, and
h3 ⊥ s2 . So we have h3 ⊥ {s1 , s2 }. Construct
h3
s3 =
|h3 |
4. Finally,
n−1
X hn
hn = g n − (g n · si )si , and sn =
i=1
|hn |
Example 34 Let ·V = R3 and consider a basis

     
1 2 4
g1 = 1 , g2 =
   1 , g 3 = 2 .
 
0 −2 1
Construct an orthonormal basis.

Ans: We start with
 
1
g 1
s1 = 1 = √ 1
|g 1 | 2 0

7.8 Orthogonal complement
Then,
     
2 1 1/2
1 1
h2 = g 2 − (g 2 · s1 )s1 =  1  − √ (2 + 1) √ 1 = −1/2
−2 2 2 0 −2
So,
 
√ 1/6
h2
s2 = = 2 −1/6 .
|h2 |
−2/3
Finally,
 
10/9
h3 = g 3 − (g 3 · s1 )s1 − (g 3 · s2 )s2 = −10/9 .
5/9
and
 
2/3
h3
s3 = = −2/3 .
|h3 |
−1/3
Therefore, {s1 , s2 , s2 } is an orthonormal basis.

Let ·V be an inner product space and W is a subspace of ·V. The orthogonal
complement of W, written as W ⊥ and read as ‘W perpendicular’, is defined as
n o
⊥
W := p ∈ ·V, p · s = 0, ∀s ∈ W .
Thus W ⊥ is the set of all vectors in ·V which are orthogonal to every vector in W.
Theorem 7.2 W ⊥ is a subspace of ·V.

Proof: Let p, q ∈ W ⊥ , and s ∈ W. Then by definition,
p · s = 0, q · s = 0.
Let α1 , α2 ∈ F. Then,
(α1 p + α2 q) · s = α1 p · s + α2 q · s = 0.
So, α1 p + α2 q ∈ W ⊥ ⊂ ·V. This implies that W ⊥ is a subspace.

Theorem 7.3 Projection theorem: Let W be any subspace of ·V. Then
W ⊕ W ⊥ = ·V.
Proof: Let dim ·V = n, and dim W = m. Let BW = {f 1 , · · · , f m } be an orthonor-

mal basis of W. Let v ∈ ·V, and motivated by the Gram-Schmidt algorithm,
consider
m
X
h=v− (v · f i )f i .
i=1
Then, h · f 1 = v · f 1 − (v · f 1 ) = 0. Similarly, h · f 2 = · · · = h · f m = 0. This

implies,
h ⊥ {f 1 , · · · , f m } and h ∈ W ⊥ .
Now,
m
X m
X
⊥
v =h+ (v · f i )f i , where h ∈ W , and (v · f i )f i ∈ W
i=1 i=1
This implies the linear sum
W + W ⊥ = ·V. (82)
Let u ∈ W ∪W ⊥ . This means u ∈ W, also u ∈ W ⊥ . Then by definition, u·u = 0,

implies u = 0. As u is arbitrary,
W ∩ W ⊥ = {0}. (83)
and equations (82), (83) imply
W ⊕ W ⊥ = ·V. (84)
Lemma 7.4 If W ⊕ W ⊥ = ·V, then every vector v ∈ ·V can be uniquely expressed

as
v = w + w⊥ .
where w ∈ W and w⊥ ∈ W ⊥ . w and w⊥ are then called orthogonal projections

of v on the subspaces W and W ⊥ .
Example 35 If T is a self adjoint (i.e. T = T t ) operator on ·V, then R(T ) =

N (T )⊥ .

Ans: Let v ∈ R(T ). Then there exists a vector, say u, such that T u = v. Let
s ∈ N (T ) such that T s = 0. Now,
v · s = T u · s = u · T t s = u · T s = u · 0 = 0.
Then v ⊥ s, and so
v ∈ N (T )⊥ , and v ∈ R(T ) =⇒ R(T ) ⊆ N (T )⊥ .
Now,
dim R(T ) + dim N (T ) = dim ·V (85)
As N (T ) is a subspace of V, we can further write,
N (T ) ⊕ N (T )⊥ = ·V =⇒ dim N (T ) + dim N (T )⊥ = dim ·V (86)
From (85) and (86), we write dim R(T ) = dim N (T )⊥ . Moreover, R(T ) ⊆
N (T )⊥ . So we conclude R(T ) = N (T )⊥ .

8 Symmetric and skew/anti-symmetric tensor
8.1 Permutations
A permutation of a set A = {1, . . . , r}, r ∈ I+ , is a bijection σ : A → A. More
correctly, σ may be thought of as a reordering of the list 1, 2, . . . , r from its nat-
ural increasing order to a new order σ(1), σ(2), . . . , σ(r). So, there are exactly r!
permutations of (1, 2, . . . , r). A simple way to describe a permutation σ : A → A
is by the matrix
1 2 ... r
σ(1) σ(2) . . . σ(r)
The order 1, 2, . . . , r is known as natural order of degree r. A permutation by
exchanging the order of two adjacent elements is called a transposition. So, you
can get any permutation by subsequent transpositions. The number of such trans-
portations are not unique.
If σ is a permutation of degree r, one can pass from the sequence 1, 2, . . . , r
to σ(1), σ(2), . . . , σ(r) by succession of interchanges of pairs (transpositions), and
this can be done in a variety of ways; however, no matter how it is done, the
number of interchanges is either even or odd, respectively. One defines the sign of
a permutation by (
1, if σ is even
sgn σ =
−1, if σ is odd
We denote the collection of all permutations as P(r).
8.2 Symmetric tensor

Consider a tensor T = v 1 ⊗ a ⊗ v 3 ⊗ a, where v 1 , a, v 3 ∈ ·V. If you change
the position of the 2 nd and 4 th vector, the tensor remains the same. We say
that the tensor is symmetric in the second and the fourth positions. We end
up with a different tensor if we exchange any other two positions. Now consider
T = a ⊗ a ⊗ a ⊗ a. If you exchange any two positions, the tensor remains the
same. We then call the tensor symmetric. We say that a tensor T is symmetric
if the tensor remains unchanged when every positions are interchanged. Consider
the tensor v 1 ⊗ v 2 , which is not symmetric as per our definition. We now define
1
Sym (v 1 ⊗ v 2 ) = (v 1 ⊗ v 2 + v 2 ⊗ v 1 )
2
Observe that if you change the position of v 1 and v 2 , Sym (v 1 ⊗ v 2 ) does not
change. The factor 1/2 is introduced to nullify the repeated sums when the tensor

is already symmetric, say a ⊗ a, then Sym (a ⊗ a) = a ⊗ a. Now how about

Sym (v 1 ⊗ v 2 ⊗ v 3 ) ? Let me give you
1
Sym (v 1 ⊗ v 2 ⊗ v 3 ) = (v 1 ⊗ v 2 ⊗ v 3 + v 1 ⊗ v 3 ⊗ v 2 + v 2 ⊗ v 1 ⊗ v 3 + v 2 ⊗ v 3 ⊗ v 1
6
+v 3 ⊗ v 2 ⊗ v 1 + v 3 ⊗ v 1 ⊗ v 2 )
You change any two positions of vectors in Sym (v 1 ⊗ v 2 ⊗ v 3 ), resulting in the

same tensor again and again. But how do you come up with such an expression?
Consider v 1 ⊗ · · · ⊗ v r ∈ T0r and v 1 , . . . , v r ∈ ·V = T01 , where v i = vik g k . We
can then immediately write the symmetrization operation
1 X
Sym (v 1 ⊗ · · · ⊗ v r ) = v σ(1) ⊗ · · · ⊗ v σ(r) . (87)
r!
σ∈P(r)
where the sum is taken over the r! permutations of the integers 1, ..., r. If we write
v 1 ⊗ · · · ⊗ v r = v1i1 · · · vrir g i1 ⊗ · · · ⊗ g ir ,
then
h1 X i i
i
Sym (v 1 ⊗ · · · ⊗ v r ) = v1σ(1) · · · vrσ(r) g i1 ⊗ · · · ⊗ g ir (88)
r!
σ∈P(r)
Similarly, if T ∈ T0r such that
T = T i1 i2 ...ir g i1 ⊗ · · · ⊗ g ir
then
h1 X i
Sym T = T iσ(1) ...iσ(r) g i1 ⊗ · · · ⊗ g ir . (89)
r!
σ∈P(r)
The symmetric components of the tensor T are often written as

1 X iσ(1) ...iσ(r)
T [i1 ...ir ] = T .
r!
σ∈P(r)
We denote the set of all symmetric tensors of order-r as S r . Moreover S r is again

a vector space.
Similarly, we can consider the space Tr0 , and the results will be the same. Only
the basis will be covariant type. How about the space of a mixed tensor Tsr ? Note

that by exchanging one contravariant position and one covariant position, the new
tensor does not belong to Tsr space anymore. Such a situation does not arise for
a pure contravariant and a pure covariant space, i.e., for Tor or Tr0 . Symmetry
operation is not defined for mixed tensors.
Example: Let us find Sym (v 1 ⊗ v 2 ). The number mapping becomes
1 2
σ(1) = 1 σ(2) = 2
σ(1) = 2 σ(2) = 1
where the top row is the natural order. From (87) we write
1 X
Sym (v 1 ⊗ v 2 ) = v σ(1) ⊗ v σ(2)
2!
σ∈P(r)
1
= (v 1 ⊗ v 2 + v 2 ⊗ v 1 )
2
If we express v 1 = v1i1 g i1 , and v 2 = v2i2 g i2 , then from (88) we write
h1 X i i
σ(1) iσ(2)
Sym (v 1 ⊗ v 2 ) = v1 v2 g i1 ⊗ g i2
2!
σ∈P(r)
h1 i
= (v1i1 v2i2 + v1i2 v2i1 ) g i1 ⊗ g i2
2!
Example: Let us find Sym (v 1 ⊗ v 2 ⊗ v 3 ). Then
1 2 3
σ(1) = 1 σ(2) = 2 σ(3) = 3
σ(1) = 1 σ(2) = 3 σ(3) = 2
σ(1) = 2 σ(2) = 1 σ(3) = 3
σ(1) = 2 σ(2) = 3 σ(3) = 1
σ(1) = 3 σ(2) = 2 σ(3) = 1
σ(1) = 3 σ(2) = 1 σ(3) = 2
1 X
Sym (v 1 ⊗ v 2 ⊗ v 3 ) = v σ(1) ⊗ v σ(2) ⊗ v σ(3)
3!
σ∈P(r)
1
= (v 1 ⊗ v 2 ⊗ v 3 + v 1 ⊗ v 3 ⊗ v 2 + v 2 ⊗ v 1 ⊗ v 3 + v 2 ⊗ v 3 ⊗ v 1
6
+v 3 ⊗ v 2 ⊗ v 1 + v 3 ⊗ v 1 ⊗ v 2 )

8.3 Skew-symmetric tensor
If we expressv 1 = v1i1 g i1 , v 2 = v2i2 g i2 , and v 3 = v3i3 g i3 , then from (87) we write

h1 X i i i
i
Sym (v 1 ⊗ v 2 ⊗ v 3 ) = v1σ(1) v2σ(2) v3σ(3) g i1 ⊗ g i2 ⊗ g i3
3!
σ∈P(r)
h1
= (v1i1 v2i2 v3i3 + v1i1 v2i3 v3i2 + v1i2 v2i1 v3i3 + v1i2 v2i3 v3i1 + v1i3 v2i2 v3i1
3! i
+v1i3 v2i1 v3i2 ) g i1 ⊗ g i2 ⊗ g i3
The symmetric product of symmetric tensors Sym A ∈ S p ⊂ T0p and Sym B ∈

S q ⊂ T0q is the symmetric tensor Sym (A ⊗ B) ∈ S p+q . In general, symmetric
multiplication is
1. Commutative: Sym (A ⊗ B) = Sym (B ⊗ A).
2. Associative: Sym ((A ⊗ B) ⊗ C) = Sym (A ⊗ (B ⊗ C)).
3. Distributive: Sym ((A + B) ⊗ C) = Sym (A ⊗ C) + Sym (B ⊗ C).
4. α Sym (A ⊗ B) = Sym (αA ⊗ B) = Sym (A ⊗ αB), α ∈ F.
So, we have an algebra of symmetric tensors. The first point shows that the algebra
is commutative.

In continuation to the previous section, a tensor is skew symmetric if it changes its
sign when any two positions are interchanged. We write the anti-symmetrization
operation
1 X
Skw (v 1 ⊗ · · · ⊗ v r ) = sgn σ v σ(1) ⊗ · · · ⊗ v σ(r) . (90)
r!
σ∈P(r)
where the sum is taken over the r! permutations of the integers 1, ..., r. Here
T ∈ T0r and v 1 , . . . , v r ∈ V. If we write
v 1 ⊗ · · · ⊗ v r = v1i1 · · · vrir g i1 ⊗ · · · ⊗ g ir ,
then
h1 X i i
i
Skw (v 1 ⊗ · · · ⊗ v r ) = sgn σ v1σ(1) · · · vrσ(r) g i1 ⊗ · · · ⊗ g ir (91)
r!
σ∈P(r)

If we consider
T = T i1 i2 ...ir g i1 ⊗ · · · ⊗ g ir ,
we can then immediately write
h1 X i
Skw T = sgn σ T iσ(1) ...iσ(r) g i1 ⊗ · · · ⊗ g ir . (92)
r!
σ∈P(r)
The skew-symmetric components of the tensor T are often written as

1 X
T (i1 ...ir ) = sgn σ T iσ(1) ...iσ(r) .
r!
σ∈P(r)
Example: Let us find Skw (v 1 ⊗ v 2 ). The number mapping becomes

1 2 Permutation
σ(1) = 1 σ(2) = 2 even
σ(1) = 2 σ(2) = 1 odd
1 X
Skw (v 1 ⊗ v 2 ) = sgn σ v σ(1) ⊗ v σ(2)
2!
σ∈P(r)
1
= (v 1 ⊗ v 2 − v 2 ⊗ v 1 )
2
If we express v 1 = v1i1 g i1 , and v 2 = v2i2 g i2 , then from (91) we write
h1 X i i
i
Skw (v 1 ⊗ v 2 ) = sgn σ v1σ(1) v2σ(2) g i1 ⊗ g i2
2!
σ∈P(r)
h1 i
i 1 i2 i2 i1
= (v v − v1 v2 ) g i1 ⊗ g i2
2! 1 2
Example: Let us find Skw (v 1 ⊗ v 2 ⊗ v 3 ). Note that

1 2 3 Permutation
σ(1) = 1 σ(2) = 2 σ(3) = 3 even
σ(1) = 1 σ(2) = 3 σ(3) = 2 odd
σ(1) = 2 σ(2) = 1 σ(3) = 3 odd
σ(1) = 2 σ(2) = 3 σ(3) = 1 even
σ(1) = 3 σ(2) = 2 σ(3) = 1 odd
σ(1) = 3 σ(2) = 1 σ(3) = 2 even

8.4 Exterior algebra (popular name) or Grassmann algebra
1 X
Skw (v 1 ⊗ v 2 ⊗ v 3 ) = sgn σ v σ(1) ⊗ v σ(2) ⊗ v σ(3)
3!
σ∈P(r)
1
= (v 1 ⊗ v 2 ⊗ v 3 − v 1 ⊗ v 3 ⊗ v 2 − v 2 ⊗ v 1 ⊗ v 3 + v 2 ⊗ v 3 ⊗ v 1
6
−v 3 ⊗ v 2 ⊗ v 1 + v 3 ⊗ v 1 ⊗ v 2 )
1. Note that by exchanging any two positions, we get a negative sign out.
2. Repeating any two vectors twice always gives zero! This is true for any order.
For example, Skw (v 1 ⊗ a ⊗ v 3 ⊗ a) = 0 as a is repeated twice.
If we express v 1 = v1i1 g i1 , v 2 = v2i2 g i2 , and v 3 = v1i3 g i3 , then from (91) we write

h1 X i i i
i
Skw (v 1 ⊗ v 2 ⊗ v 3 ) = sgn σ v1σ(1) v2σ(2) v3σ(3) g i1 ⊗ g i2 ⊗ g i3
3!
σ∈P(r)
h1
= (v1i1 v2i2 v3i3 − v1i1 v2i3 v3i2 − v1i2 v2i1 v3i3 + v1i2 v2i3 v3i1 − v1i3 v2i2 v3i1
3! i
+v1i3 v2i1 v3i2 ) g i1 ⊗ g i2 ⊗ g i3
We denote the set of all skew-symmetric tensors of order-r as Ωr . Moreover Ωr

is again a vector space. If Skw A ∈ Ωp ⊂ T0p and Skw B ∈ Ωq ⊂ T0q , then
Skw (A ⊗ B) ∈ Ωp+q has the following properties.
1. Anti-commutative: Skw (A ⊗ B) = −Skw (B ⊗ A).
2. Associative: Skw ((A ⊗ B) ⊗ C) = Skw (A ⊗ (B ⊗ C)).
3. Distributive: Skw ((A + B) ⊗ C) = Skw (A ⊗ C) + Skw (B ⊗ C).
4. α Skw (A ⊗ B) = Skw (αA ⊗ B) = Skw (A ⊗ αB), α ∈ F.
The first point shows that the algebra is anti-commutative.
8.4 Exterior algebra (popular name) or Grassmann alge-

bra
The concept of an exterior algebra was originally introduced by H. Grassman for
the study of vector subspaces. Subsequently Elie Cartan developed the theory of

exterior differentiation and successfully applied to the study of differential geome-

try and differential equations. Recently, exterior algebra are indispensable tool in
the study of differential geometry of manifolds.
Symbol to denote the space of skew-symmetric tensors of type (r, 0) is Ωr . The
skew-symmetric tensor space of type (0, r) is denoted by Λr .
T ∈ Ωr : is called exterior r-vector ,

T ∈ Λr : is called exterior r-form.
Skew-symmetric tensors of type Λr play an especially important role in the dif-

ferential geometry. They provide the tool to generalization of curl, divergence,
stokes law in any arbitrary dimensions, and play essential roles in any branch of
mechanics. Any tensor belongs to Λr is covariant in nature, and are known as
r-forms. However, all the derivations are equally valid to the r-vectors.
Similar to the symmetric tensors, tensor product of two skew-symmetric tensors
are not skew-symmetric. Hence it is required to define such a product by which
tensor product of two skew-symmetric tensors are again skew symmetric. This
product is called the exterior product or wedge product.
8.4.1 Exterior or wedge product

The symbol for this product is a wedge, ∧. We start this topic by defining exterior
product as
(p + q)!
T ∧S = Skw (T ⊗ S)
p!q!
when T ∈ Λp , S ∈ Λq , and T ∧ S ∈ Λp+q .
Case-1: Let T = v 1 ∈ Λ1 and S = v 2 ∈ Λ1 , then
(1 + 1)! 1 X
T ∧ S = v1 ∧ v2 = Skw (v 1 ⊗ v 2 ) = 2! sgn σ v σ(1) ⊗ v σ(2)
1!1! 2!
σ∈P(r)
X
= sgn σ v σ(1) ⊗ v σ(2)
σ∈P(r)
= (v 1 ⊗ v 2 − v 2 ⊗ v 1 )
Note that for two 1-forms,
v 1 ∧ v 2 = −v 2 ∧ v 1 . (93)

Case-2: Let T = v 1 ∧ v 2 ∈ Λ2 and S = v 3 ∈ Λ1 , then

(2 + 1)! 3!
T ∧ S = v1 ∧ v2 ∧ v3 = Skw ((v 1 ∧ v 2 ) ⊗ v 2 ) = Skw ((v 1 ⊗ v 2 − v 2 ⊗ v 1 ) ⊗ v 3 )
2!1! 2!
3! h 1 2 3 2 1 3
i
= Skw (v ⊗ v ⊗ v ) − Skw (v ⊗ v ⊗ v )
2!
3! h i
= Skw (v 1 ⊗ v 2 ⊗ v 3 ) + Skw (v 1 ⊗ v 2 ⊗ v 3 )
2!
3!
= (2) Skw (v 1 ⊗ v 2 ⊗ v 3 )
2!
= 3! Skw (v 1 ⊗ v 2 ⊗ v 3 )
1 X
= 3! sgn σ v σ(1) ⊗ v σ(2) ⊗ v σ(3)
3!
σ∈P(r)
X
= sgn σ v σ(1) ⊗ v σ(2) ⊗ v σ(3)
σ∈P(r)
Extending, we simply write
v 1 ∧ v 2 . . . ∧ v p = p! Skw (v 1 ⊗ v 2 · · · ⊗ v p )
X
= sgn σ v σ(1) ⊗ v σ(2) · · · ⊗ v σ(p) ∈ Λp (94)
σ∈P(p)
where each v i ∈ Λ1 , i.e., 1-form.
Theorem 8.1 If T = v 1 ∧ v 2 . . . ∧ v p ∈ Λp and S = u1 ∧ u2 . . . ∧ uq ∈ Λq then

T ∧ S = (−1)pq S ∧ T
Proof:
T ∧ S = v 1 ∧ v 2 . . . ∧ v p ∧ u1 ∧ u2 . . . ∧ uq (95)
exchanging position of v p and u1
= (−1)v 1 ∧ v 2 . . . ∧ v p−1 ∧ u1 ∧ v p ∧ u2 . . . ∧ uq
exchanging position of v p−1 and u1
= (−1)2 v 1 ∧ v 2 . . . ∧ u1 ∧ v p−1 ∧ v p ∧ u2 . . . ∧ uq
= (−1)p u1 ∧ v 1 ∧ . . . ∧ v p−1 ∧ v p ∧ u2 . . . ∧ uq
= (−1)p (−1)p u1 ∧ u2 ∧ . . . ∧ v p−1 ∧ v p ∧ . . . ∧ uq
= (−1)p+...q times u1 ∧ u2 ∧ . . . ∧ uq ∧ v 1 · · · v p−1 ∧ v p
= (−1)pq S ∧ T (96)
The exterior product has the following general properties

8.5 Basis of Λk
1. Associativity: (A ∧ B) ∧ C = A ∧ (B ∧ C)
2. Anticommutativity: If A is of degree p and B is of degree q, then
A ∧ B = (−1)pq B ∧ A.
In particular, ∀a, b ∈ Λ1 , a ∧ b = −b ∧ a.
3. Distributivity: (A + B) ∧ C = A ∧ C + B ∧ C.
4. If α ∈ F, then α(A ∧ B) = (αA ∧ B) = (A ∧ αB).
8.5 Basis of Λk
Case-1: Suppose T = c ∈ Λ0 , i.e., a scalar, and the basis is {1}
Case-2: Suppose T = v 1 ∈ Λ1 , i.e., a covector or 1-form and v 1 = vi11 g i1 .
Obviously {g i1 } is a basis of Λ1 .
Case-3: Suppose T = v 1 ∧ v 2 ∈ Λ2 , i.e., a covector or 2-form and v 1 = vi11 g i1 and
v 2 = vi22 g i2 . Now
v 1 ∧ v 2 = vi11 g i1 ∧ vi22 g i2
= (v11 g 1 + v21 g 2 + v31 g 3 ) ∧ (v12 g 1 + v22 g 2 + v32 g 3 ) assuming dim ·V=3.
= (v11 g 1 ∧ v22 g 2 + v11 g 1 ∧ v32 g 3 ) + (v21 g 2 ∧ v12 g 1 + v21 g 2 ∧ v32 g 3 )
+(v31 g 3 ∧ v12 g 1 + v31 g 3 ∧ v22 g 2 )
= (v11 v22 − v21 v12 )g 1 ∧ g 2 + (v21 v32 − v31 v22 )g 2 ∧ g 3 + (v11 v32 − v31 v12 )g 1 ∧ g 3
(97)
We used the fact g i ∧ g j = −g j ∧ g i if i ̸= j and g i ∧ g j = 0 if i = j. This shows

that {g 1 ∧ g 2 , g 2 ∧ g 3 , g 1 ∧ g 3 } span any element v 1 ∧ v 2 ∈ Λ2 . Now to show that
they are linearly independent, consider
C1 g 1 ∧ g 2 + C2 g 2 ∧ g 3 + C3 g 1 ∧ g 3 = 0
Operating on g 2 at both sides,
C1 (g 1 ∧ g 2 )g 2 + C2 (g 2 ∧ g 3 )g 2 + C3 (g 1 ∧ g 3 )g 2 = 0 (98)
Note that,
(g 1 ∧ g 2 )g 2 = (g 1 ⊗ g 2 − g 2 ⊗ g 1 )g 2 = g 1 (g 2 · g 2 ) − g 2 (g 1 · g 2 ) = g 1
(g 2 ∧ g 3 )g 2 = (g 2 ⊗ g 3 − g 3 ⊗ g 2 )g 2 = g 2 (g 3 · g 2 ) − g 3 (g 2 · g 2 ) = −g 3
(g 1 ∧ g 3 )g 2 = (g 1 ⊗ g 3 − g 3 ⊗ g 1 )g 2 = g 1 (g 3 · g 2 ) − g 3 (g 1 · g 2 ) = 0

8.5 Basis of Λk
and we write (98)
C1 g 1 − C2 g 3 = 0 =⇒ C1 = C2 = 0.
Again from (98), C3 = 0. So, {g 1 ∧ g 2 , g 2 ∧ g 3 , g 1 ∧ g 3 } are linearly independent,

and a basis of Λ2 . A concise way of writing (97) is
h X i
v1 ∧ v2 = sgn σ vi1σ(1) vi2σ(2) g i1 ∧ g i2 (i1 < i2 ≤ 3)
σ∈P(2)
We can easily extend our calculations to the higher dimension and write
h X i
v1 ∧ v2 = sgn σ vi1σ(1) vi2σ(2) g i1 ∧ g i2 (i1 < i2 ≤ n = dim · V)
σ∈P(2)
n

and we have 2
number of terms in the basis {g i1 ∧ g i2 } (i1 < i2 ≤ n). We then
write
h X i
1 2 k
v ∧ v ∧ ... ∧ v = sgn σ vi1σ(1) vi2σ(2) . . . vikσ(k) g i1 ∧ g i2 . . . ∧ g ik (i1 < i2 < · · · < ik ≤ n)
σ∈P(k)
Note that for k = n, we write

h X i
1 2 n
v ∧ v ∧ ... ∧ v = 1 2
sgn σ viσ(1) viσ(2) . . . viσ(n) g i1 ∧ g i2 . . . ∧ g in
n
(i1 < i2 < · · · < in ≤ n)
σ∈P(n)
h X i
1 2 n
= sgn σ vσ(1) vσ(2) . . . vσ(n) g1 ∧ g2 . . . ∧ gn (99)
σ∈P(n)
Space Dimensions Basis

Λ0 1 {1}
1 n
Λ 1
{g i1 }
n
Λ2 2
{g i1 ∧ g i2 } i1 < i2 ≤ n
3 n i1 i2 i3
Λ 3
{g ∧ g ∧ g } i 1 < i2 < i3 ≤ n
··· · · · ···
k n
Λ k
{g ∧ g ∧ g i3 . . . ∧ g ik } i1 < i2 < i3 < . . . ik ≤ n
i1 i2
··· · · · ···
n n
Λ n
{g ∧ g ∧ g 3 . . . ∧ g n }
1 2

8.6 Hodge duality
All the above results are equally applicable to r-vectors, and we write
Space Dimensions Basis
0
Ω 1 {1}
1 n
Ω 1
{g i1 }
n
Ω2 2
{g i 1 ∧ g i2 } i1 < i2 ≤ n
3 n
Ω 3
{g i1 ∧ g i2 ∧ g i3 } i1 < i2 < i3 ≤ n
··· · · · ···
n
Ωk k
{g i1 ∧ g i2 ∧ g i3 . . . ∧ g ik } i1 < i2 < i3 < . . . ik ≤ n
··· · · · ···
n
Ωn n
{g 1 ∧ g 2 ∧ g3 . . . ∧ gn}
8.6 Hodge duality

The dimension of the skew-symmetric tensor spaces over a vector space ·V of
dimension n have a symmetry, np = n−p n

. The Hodge star operator is an iso-
morphism between the pairs of these spaces of equal dimension.
⋆ := Λp → Λn−p .
We define Hodge star operator by specifying it on the basis of Λp , obtained from
the standard basis (e1 , ..., en ). Then
⋆(ei1 ∧ ... ∧ eip ) = sgn σ(ej1 ∧ ... ∧ ejn−p ). i1 < · · · < ip & j1 < · · · < jn−p
where {j1 , . . . jn−p } ∈ {n}/{i1 , . . . , ip }, and σ ∈ P(n) is the permutation {i1 , . . . , ip , j1 , . . . , jn−p }.
Example 36 Consider a 2D case, and let {e1 , e2 } be the basis of Λ1 . Then

⋆e1 = e2
⋆e2 = −e1
⋆(e1 ∧ e2 ) = 1
Example 37 In 3D, where {e1 , e2 , e2 } be the basis of Λ1 . Then
⋆e1 = e2 ∧ e3 , ⋆(e1 ∧ e2 ) = e3
⋆e2 = −e1 ∧ e3 , ⋆(e2 ∧ e3 ) = e1
⋆e3 = e1 ∧ e2 , ⋆(e1 ∧ e3 ) = −e2
and ⋆(e1 ∧ e2 ∧ e3 ) = 1.
Example 38 Consider
a ∧ b = (a2 b3 − a3 b2 )e2 ∧ e3 − (a3 b1 − a1 b3 )e1 ∧ e3 + (a1 b2 − a2 b1 )e1 ∧ e2
⋆(a ∧ b) = (a2 b3 − a3 b2 )e1 + (a3 b1 − a1 b3 )e2 + (a1 b2 − a2 b1 )e3
= a×b

8.7 Determinant
Example 39 Consider
a ∧ b ∧ c = [(a2 b3 − a3 b2 )c1 + (a3 b1 − a1 b3 )c2 + (a1 b2 − a2 b1 )c3 ]e1 ∧ e2 ∧ e3

⋆(a ∧ b ∧ c) = [(a2 b3 − a3 b2 )c1 + (a3 b1 − a1 b3 )c2 + (a1 b2 − a2 b1 )c3 ]
= c · (a × b)
This is the volume cell.
8.7 Determinant
Definition 8.1 Let T ∈ Hom(·V, ·V) and nonsingular. Then the determinant of
T , det T ∈ R, is defined as the factor
T v 1 ∧ · · · ∧ T v n = det T (v 1 ∧ · · · ∧ v n )
where v 1 , . . . , v n ∈ Λ1 are linearly independent and dim ·V = n.

We can write from (99)
h X i
v1 ∧ v2 ∧ . . . ∧ vn = 1
sgn σ vσ(1) 2
vσ(2) n
. . . vσ(n) g1 ∧ g2 . . . ∧ gn
σ∈P(n)
= [Θ] g 1 ∧ g 2 . . . ∧ g n (100)
So, we write
T v 1 ∧ · · · ∧ T v n = det T [Θ] (g 1 ∧ g 2 . . . ∧ g n ) (101)
Note that T v k = T (vikk g ik ) = vikk T g ik , and recall {T g 1 , T g 2 , . . . T g n } is again a

basis in ·V when {g 1 , g 2 , . . . , g n } is a basis in ·V, and T is nonsingular. We again
write from (99)
h X i
1 2 n
Tv ∧ Tv ∧ ... ∧ Tv = sgn σ vσ(1) vσ(2) . . . vσ(n) T g 1 ∧ T g 2 . . . ∧ T g n
1 2 n
σ∈P(n)
= [Θ] T g 1 ∧ T g 2 . . . ∧ T g n (102)
(101) and (102) implies
[Θ] T g 1 ∧ T g 2 . . . ∧ T g n = det T [Θ] (g 1 ∧ g 2 . . . ∧ g n )

T g 1 ∧ T g 2 . . . ∧ T g n = det T (g 1 ∧ g 2 . . . ∧ g n )
Note that the definition of determinant does not depend on the selection of a basis
or set of vectors. Hence determinant is an invariant. If we consider T = Ti·j g i ⊗g j ,

8.7 Determinant
then
T g 1 = (Ti·j g i ⊗ g j )g 1 = Ti·1 g i
T g 2 = (Ti·j g i ⊗ g j )g 2 = Ti·2 g i
..
.
T g n = (Ti·j g i ⊗ g j )g n = Ti·n g i
This gives,
T g 1 ∧ ... ∧ T g n = Ti1·1 g i1 ∧ · · · ∧ Tin·n g in

X
·1 ·2 ·n
= sgn σ Tσ(1) Tσ(2) . . . Tσ(n) (g 1 ∧ ... ∧ g n )
σ∈P(n)
and so
X
det T = det[Ti·j ] = ·1
sgn σ Tσ(1) ·2
Tσ(2) ·n
. . . Tσ(n) (103)
σ∈P(n)
Now consider the transpose of T , T t = Ti·j g j ⊗ g i = [T t ]j·i g j ⊗ g i . We then write
T t g 1 ∧ ... ∧ T t g n = T1·i1 g i1 ∧ · · · ∧ Tn·in g in

X
·τ (1) ·τ (2)
= sgn τ T1 T2 . . . Tn·τ (n) (g 1 ∧ ... ∧ g n )
τ ∈P(n)
and so
·τ (1) ·τ (2)
X
det T t = sgn τ T1 T2 . . . Tn·τ (n) . (104)
τ ∈P(n)
Let us now compare the right hand sides of (103) and (104). We will justify
that they are equal by some simple arguments. Let us consider one particular
permutation of (103) among n!. Let say σ(2) = 8(< n). Then there is always a
·τ (8)
term T8 in (104) such that τ (8) = 2. So for a specific permutation of σ, there
exist a permutation in τ such that these two terms are exactly equal with the
sign. However, we are skipping the detail here. We can then immediately write
det T = det T t .
Example 40 det u ⊗ v = 0

8.8 Trace
Let T = u ⊗ v. Then,
(u ⊗ v)v 1 ∧ ... ∧ (u ⊗ v)v n = det(u ⊗ v)(v 1 ∧ ... ∧ v n )
(v · v 1 )u ∧ ... ∧ (v · v n )u = det(u ⊗ v)(v 1 ∧ ... ∧ v n
(v · v 1 )(v · v 2 )...(v · v n )[u ∧ ... ∧ u] = det(u ⊗ v)(v 1 ∧ ... ∧ v n )
0 = det(u ⊗ v)(v 1 ∧ ... ∧ v n )
So, det(u ⊗ v) = 0.
Example 41 det(αI) = αn
Let T = αI. Then,
αIv 1 ∧ ... ∧ αIv n = det(αI)(v 1 ∧ ... ∧ v n )
αv 1 ∧ ... ∧ αv n = det(αI)(v 1 ∧ ... ∧ v n )
αn (v 1 ∧ ... ∧ v n ) = det(αI)(v 1 ∧ ... ∧ v n )
So, det(αI) = αn .
Example 42 det(ST ) = (det S)(det T )
(ST )v 1 ∧ ... ∧ (ST )v n = det(ST )(v 1 ∧ ... ∧ v n )

S(T v 1 ) ∧ ... ∧ S(T v n ) = det(ST )(v 1 ∧ ... ∧ v n )
det(S)(T v 1 ∧ ... ∧ T v n ) = det(ST )(v 1 ∧ ... ∧ v n )
det(S) det(T )(v 1 ∧ ... ∧ v n ) = (det S)(det T )(v 1 ∧ ... ∧ v n )
So, det(ST ) = (det S)(det T ). Note that,
m m
det[Tkn ] = det[T ·n gkm ] = det[T ·n ] det[gkm ] = (det T ) det[gkm ].
Similarly,
det[T mk ] = det[T ·n
m kn m
g ] = det[T ·n ] det[g kn ] = (det T ) det[g kn ].
So, det T is not equal to det[Tij ] or det[T ij ], unless det[gij ] = det[g ij ] = 1.
8.8 Trace
Definition 8.2 Let T ∈ Hom(·V, ·V). Then the trace of T , tr T ∈ R, is defined
as
Xn
(v 1 ∧ ... ∧ T v i ... ∧ v n ) = tr T (v 1 ∧ ... ∧ v n )
i=1
where v 1 , ..., v n ∈ ·V and dim ·V = n. In the summation, when i = 1, it is T v 1 ,

when i = 2, it is T v 2 , and so on.

8.8 Trace
We can also select a basis {g 1 , ..., g n ∈ ·V}, then (like the definition of determinant)
n
X
(g 1 ∧ ... ∧ T g i ... ∧ g n ) = tr T (g 1 ∧ ... ∧ g n )
i=1
Example 43 tr (αS + βT ) = αtr S + βtr T
n
X
(g 1 ∧ ... ∧ (αSg i + βT g i )... ∧ g n ) = tr (αS + βT )(g 1 ∧ ... ∧ g n )
i=1
n
X n
X
1 i n
(g ∧ ... ∧ αSg ... ∧ g ) + (g 1 ∧ ... ∧ βT g i ... ∧ g n ) = tr (αS + βT )(g 1 ∧ ... ∧ g n )
i=1 i=1
Xn Xn
α (g 1 ∧ ... ∧ Sg i ... ∧ g n ) + β (g 1 ∧ ... ∧ T g i ... ∧ g n ) = tr (αS + βT )(g 1 ∧ ... ∧ g n )
i=1 i=1
α(tr S)(g ∧ ... ∧ g ) + β(tr T )(g 1 ∧ ... ∧ g n ) = (α tr S + β tr T )(g 1 ∧ ... ∧ g n )
1 n
So, tr (αS + βT ) = αtr S + βtr T .
Example 44 tr I = n
n
X
(g 1 ∧ ... ∧ Ig i ... ∧ g n ) = tr I(g 1 ∧ ... ∧ g n )
i=1
Ig ∧ ... ∧ g + g ∧ Ig 2 ∧ ... ∧ g n + · · · = tr I(g 1 ∧ ... ∧ g n )
1 n 1
g 1 ∧ ... ∧ g n + g 1 ∧ g 2 ∧ ... ∧ g n + · · · = tr I(g 1 ∧ ... ∧ g n )

n(g 1 ∧ ... ∧ g n ) = tr I(g 1 ∧ ... ∧ g n )
So, tr I = n.
Example 45 tr (u ⊗ v) = u · v
n
X
(g 1 ∧ ... ∧ (u ⊗ v)g i ... ∧ g n ) = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
i=1
(u ⊗ v)g ∧ ... ∧ g + g ∧ (u ⊗ v)g 2 ∧ ... ∧ g n + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
1 n 1
(v · g 1 )u ∧ ... ∧ g n + g 1 ∧ (v · g 2 )u ∧ ... ∧ g n + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )

(v j g j · g 1 )u ∧ ... ∧ g n + g 1 ∧ (v j g j · g 2 )u ∧ ... ∧ g n + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )

8.9 Interior product
continuing we get
(v 1 u ∧ ... ∧ g n ) + (g 1 ∧ v 2 u ∧ ... ∧ g n ) + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )

v 1 (u ∧ ... ∧ g n ) + v 2 (g 1 ∧ u ∧ ... ∧ g n ) + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
v 1 (uk g k ∧ ... ∧ g n ) + v 2 (g 1 ∧ uk g k ∧ ... ∧ g n + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
v 1 (u1 g 1 ∧ ... ∧ g n ) + v 2 (g 1 ∧ u2 g 2 ∧ ... ∧ g n ) + · · · = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
(v 1 u1 + v 2 u2 + · · · )(g 1 ∧ ... ∧ g n ) = tr (u ⊗ v)(g 1 ∧ ... ∧ g n )
This gives tr (u ⊗ v) = u · v.

We saw that exterior product increases the dimension of the space. In contrast,
interior product lowers the degree of the forms. interior product is a map
iv : ·V × Λk → Λk−1
Let T = v 1 ∧ v 2 . . . ∧ v k ∈ Λk , the operation is defined as
iv · T = v · T . (105)
So,
iv · T = v · T
= v · (v 1 ∧ v 2 . . . ∧ v k )
= v · k! Skw (v 1 ⊗ v 2 · · · ⊗ v k )
Case-1: Suppose T = c ∈ Λ0 , i.e., a scalar. Then iv : ·V × Λ0 → Λ−1 , and Λ−1 is

not defined. We consider
iv · T = 0.
Case-2: Suppose T = v 1 ∈ Λ1 , i.e., a covector or 1-form. Then iv : ·V × Λ1 → Λ0 ,

and
iv · T = v · 1! Skw (v 1 ) = v · v 1 .
Case-3: Suppose T = v 1 ∧ v 2 ∈ Λ2 , i.e., 2-form. Then iv : ·V × Λ2 → Λ1 , So
iv · T = v · 2! Skw (v 1 ⊗ v 2 ) = v · (v 1 ⊗ v 2 − v 2 ⊗ v 1 )
= (v · v 1 )v 2 − (v · v 2 )v 1
= (iv · v 1 )v 2 − (iv · v 2 )v 1

or
iv · (v 1 ∧ v 2 ) = (iv · v 1 )v 2 − (iv · v 2 )v 1 (106)
Case-4: Suppose T = v 1 ∧ v 2 ∧ v 3 ∈ Λ3 , i.e., 3-form. Then iv : ·V × Λ3 → Λ2 , So
iv · T = v · 3! Skw (v 1 ⊗ v 2 ⊗ v 3 )
= v · (v 1 ⊗ v 2 ⊗ v 3 − v 1 ⊗ v 3 ⊗ v 2 − v 2 ⊗ v 1 ⊗ v 3 + v 2 ⊗ v 3 ⊗ v 1
−v 3 ⊗ v 2 ⊗ v 1 + v 3 ⊗ v 1 ⊗ v 2 )
= (v · v 1 )v 2 ⊗ v 3 − (v · v 1 )v 3 ⊗ v 2 − (v · v 2 )v 1 ⊗ v 3 + (v · v 2 )v 3 ⊗ v 1
−(v · v 3 )v 2 ⊗ v 1 + (v · v 3 )v 1 ⊗ v 2
= (v · v 1 )[v 2 ⊗ v 3 − v 3 ⊗ v 2 ] − (v · v 2 )[v 1 ⊗ v 3 − v 3 ⊗ v 1 ]
−(v · v 3 )[v 2 ⊗ v 1 − v 1 ⊗ v 2 ]
= (iv · v 1 )v 2 ∧ v 3 + (iv · v 2 )v 3 ∧ v 1 + (iv · v 3 )v 1 ∧ v 2
or
iv · (v 1 ∧ v 2 ∧ v 3 ) = (iv · v 1 )v 2 ∧ v 3 + (iv · v 2 )v 3 ∧ v 1 + (iv · v 3 )v 1 ∧ v 2 (107)
Case-5: Suppose A = v 1 ∧ v 2 ∈ Λ2 , i.e., 2-form. Then
iv · (A ∧ v 3 ) = v · 3! Skw (v 1 ⊗ v 2 ⊗ v 3 )
= [(iv · v 1 )v 2 − (iv · v 2 )v 1 ] ∧ v 3 + (iv · v 3 )v 1 ∧ v 2
= iv · (v 1 ∧ v 2 ) ∧ v 3 + (iv · v 3 )v 1 ∧ v 2 (from (106))
= (iv · A) ∧ v 3 + A(iv · v 3 )
Case-6: Suppose B = v 2 ∧ v 3 ∈ Λ2 , i.e., 2-form. Then
iv · (v 1 ∧ B) = v · 3! Skw (v 1 ⊗ v 2 ⊗ v 3 )
= (iv · v 1 )v 2 ∧ v 3 − v 1 ∧ [(iv · v 2 )v 3 − (iv · v 3 )v 2 ]
= (iv · v 1 )B − v 1 ∧ (iv · B) (from (106))
In general, we write
iv · (T ∧ S) = (iv · T ) ∧ S + (−1)ord T T ∧ (iv · S) (108)
We also have
iv · (T + S) = iv · T + iv · S (109)

8.10 Hodge duality
8.10 Hodge duality

The dimension of the skew-symmetric tensor spaces over a vector space ·V of
dimension n have a symmetry, np = n−p n

. The Hodge star operator is an iso-
morphism between the pairs of these spaces of equal dimension.
⋆ := Λp → Λn−p .
We define Hodge star operator by specifying it on the basis of Λp , obtained from

the standard basis (e1 , ..., en ). Then
⋆(ei1 ∧ ... ∧ eip ) = sgn σ(ej1 ∧ ... ∧ ejn−p ). i1 < · · · < ip & j1 < · · · < jn−p
where {j1 , . . . jn−p } ∈ {n}/{i1 , . . . , ip }, and σ ∈ P(n) is the permutation {i1 , . . . , ip , j1 , . . . , jn−p }.
Example 46 Consider a 2D case, and let {e1 , e2 } be the basis of Λ1 . Then
⋆e1 = e2
⋆e2 = −e1
⋆(e1 ∧ e2 ) = 1
Example 47 In 3D, where {e1 , e2 , e2 } be the basis of Λ1 . Then
⋆e1 = e2 ∧ e3 , ⋆(e1 ∧ e2 ) = e3
⋆e2 = −e1 ∧ e3 , ⋆(e2 ∧ e3 ) = e1
⋆e3 = e1 ∧ e2 , ⋆(e1 ∧ e3 ) = −e2
and ⋆(e1 ∧ e2 ∧ e3 ) = 1.
Example 48 Consider
a ∧ b = (a2 b3 − a3 b2 )e2 ∧ e3 − (a3 b1 − a1 b3 )e1 ∧ e3 + (a1 b2 − a2 b1 )e1 ∧ e2

⋆(a ∧ b) = (a2 b3 − a3 b2 )e1 + (a3 b1 − a1 b3 )e2 + (a1 b2 − a2 b1 )e3
= a×b
Example 49 Consider
a ∧ b ∧ c = [(a2 b3 − a3 b2 )c1 + (a3 b1 − a1 b3 )c2 + (a1 b2 − a2 b1 )c3 ]e1 ∧ e2 ∧ e3

⋆(a ∧ b ∧ c) = [(a2 b3 − a3 b2 )c1 + (a3 b1 − a1 b3 )c2 + (a1 b2 − a2 b1 )c3 ]
= c · (a × b)
This is the volume cell.

9 Invariants in 3D
We will now consider T ∈ Hom (·V, ·V) such that dim ·V = 3. We can immediately
write the three invariants as
T u ∧ T v ∧ T w = det T (u ∧ v ∧ w)
T u ∧ v ∧ w + u ∧ T v ∧ w + u ∧ v ∧ T w = tr T (u ∧ v ∧ w)
T u ∧ T v ∧ w + u ∧ T v ∧ T w + T u ∧ v ∧ T w = I2 (u ∧ v ∧ w)
We are familiar with det T , tr T in the previous section, and it is a standard

practice to name I3 = det T and I1 = tr T As we progress, we will identify I2 .
Note that u, v, w ∈ ·V. If we take advantage that I1 , I2 , I3 are scalars, operating
with Hodge star on the both sides, i.e., using the relation
⋆(a ∧ b ∧ c) = a · (b × c) = [a, b, c]
we write
[T u, T v, T w] = I3 [u, v, w] (110)
[T u, v, w] + [u, T v, w] + [u, v, T w] = I1 [u, v, w] (111)
[T u, T v, w] + [u, T v, T w] + [T u, v, T w] = I2 [u, v, w] (112)
9.1 Cofactor tensor

Definition 9.1 Let T ∈ Hom(·V, ·V), and u, v ∈ ·V. Then the cofactor tensor,
denoted by Cof T is defined by7
Cof T (u × v) := T u × T v. (113)
Example 50 Show that (det T )I = (Cof T )t T

7
The components are the same as the cofactor of a matrix. If
 
  A22 A23 A21 A23 A21 A22
A11 A12 A13  A32 A33 − A31 A33 A31 A32 
[Aij ] = A21 A22 A23  then [Cof A]ij =  
 ... ... ... 
A31 A32 A33
... ... ...

9.1 Cofactor tensor
We start from (110) and write
[T u, T v, T w] T u · (T v × T w) T u · Cof T (v × w)
det T = = =
[u, v, w] u · (v × w) u · (v × w)
t
⇒ (det T )(u · (v × w)) = T u · Cof T (v × w) = (Cof T ) T u · (v × w)
h i h i
t
⇒ (det T )I u · (v × w) = (Cof T ) T u · (v × w)
The last line gives
(det T )I = (Cof T )t T . (114)
Example 51 Show that (Cof T )t = I2 I − (tr T )T + T 2

We start from (111)
[T u, v, w] [u, T v, w] [u, v, T w]
tr T = + +
[u, v, w] [u, v, w] [u, v, w]
and replacing u by T u we get
[T 2 u, v, w] [T u, T v, w] [T u, v, T w]
tr T = + +
[T u, v, w] [T u, v, w] [T u, v, w]
2
⇒ tr T [T u, v, w] = [T u, v, w] + [T u, T v, w] + [T u, v, T w]
⇒ tr T [T u, v, w] − [T 2 u, v, w] + [u, T v, T w] = [T u, T v, w] + [T u, v, T w] + [u, T v, T w]
⇒ tr T [T u, v, w] − [T 2 u, v, w] + [u, T v, T w] = I2 [u, v, w]
⇒ tr T (T u · (v × w)) − T 2 u · (v × w) + u · (T v × T w) = I2 (u · (v × w))
⇒ tr T (T u · (v × w)) − T 2 u · (v × w) + u · Cof T (v × w) = I2 (u · (v × w))
h i h i h i h i
⇒ (tr T )T u · (v × w) − T 2 u · (v × w) + (Cof T )t u · (v × w) = I2 I u · (v × w)
This gives
h ih i h i h i
2 t
(tr T )T − T + (Cof T ) = I2 I
and finally we arrive at
(Cof T )t = I2 I − (tr T )T + T 2 . (115)
Example 52 Show that I2 = tr (Cof T )t

9.1 Cofactor tensor
We know from (112) that8

I2 [u, v, w] = [T u, T v, w] + [u, T v, T w] + [T u, v, T w]
= (T u × T v) · w + u · (T v × T w) + (T w × T u) · v
= Cof T (u × v) · w + u · Cof T (v × w) + Cof T (w × u) · v
= (u × v) · (Cof T )t w + (Cof T )t u · (v × w) + (w × u) · (Cof T )t v
= [u, v, (Cof T )t w] + [(Cof T )t u, v, w] + [u, (Cof T )t v, w]
= tr (Cof T )t [u, v, w]
So we write
I2 = tr (Cof T )t . (116)
Example 53 Show that I2 = 21 [(tr T )2 − tr T 2 ]

Taking the ‘trace’ of both sides of (115), and using (116) we get
1
I2 = 3I2 − (tr T )2 + tr T 2 ⇒ I2 = [(tr T )2 − tr T 2 ] (117)
2
Example 54 Show that Cof T t = (Cof T )t

From (115), taking transpose of both sides, we write
Cof T = I2 I − (tr T )T t + (T t )2 .
Replacing T with T t in the above equation we get
Cof T t = I2′ I − (tr T t )T + T 2
= I2′ I − (tr T )T + T 2
where,
1 1
I2′ = [(tr T t )2 − tr (T t )2 ] = [(tr T )2 − tr (T )2 ] = I2 .
2 2
So,
Cof T t = I2 I − (tr T )T + T 2 = (Cof T )t .
Finally we get
Cof T t = (Cof T )t . (118)
8
Recall that a · (b × c) = b · (c × a) = c · (a × b)

Example 55 Show that T −1 = (Cof T )t / det T
From (115)
T (Cof T )t = I2 T − (tr T )T 2 + T 3 = (Cof T )t T = (det T )I. (119)
So we write,
(Cof T )t (Cof T )t
T = T = I.
det T det T
and it means
(Cof T )t
T −1 = (120)
det T
Example 56 Show that det T = 61 [(tr T )3 + 2tr T 3 − 3(tr T )(tr T 2 )]

From (119) we write
(det T )I = I2 T − (tr T )T 2 + T 3 .
Taking ‘trace’ at both sides we get,
3(det T ) = I2 tr T − (tr T )(tr T 2 ) + tr T 3
1
= [(tr T )2 − tr T 2 ] tr T − (tr T )(tr T 2 ) + tr T 3
2
1 1
= (tr T )3 − (tr T 2 )(tr T ) − (tr T )(tr T 2 ) + tr T 3
2 2
1 3
= (tr T )3 − (tr T 2 )(tr T ) + tr T 3
2 2
1
= [(tr T )3 − 3(tr T 2 )(tr T ) + 2tr T 3 ]
2
So we get
1
det T = [(tr T )3 − 3(tr T 2 )(tr T ) + 2tr T 3 ]. (121)
6
10 Orthogonal tensors
Let R ∈ Hom (·V, ·V) and dim ·V = 3 such that Rt = R−1 or alternatively
RRt = Rt R = I. (122)
Then R is said to be orthogonal tensor. It has the following properties

1. It preserves the length two vectors after transformation (Fig. 37):
Ru · Ru = u · Rt Ru = u · Iu = u · u. This means |Ru| = |u|.
2. It preserves the angle between two vectors after transformation (Fig. 37):
Ru · Rv = u · Rt Rv = u · Iv = u · v. So,
Ru · Rv u·v
cos θ = =
|Ru||Rv| |u||v|
3. As RRt = Rt R = I, det(RRt ) = det I implies (det R)2 = 1 or det R = ±1.
Qu
v
θ
Qv
θ u
(a) (b)
Figure 37: θ and the lengths of the vectors remain the same after an orthogonal
transformation.
When det R = 1, we call it proper orthogonal. It represents rotations, and we call

n o
+ t t
R ∈ Orth := R ∈ Hom (·V, ·V), RR = R R = I, det R = 1
When det R = −1, we call it improper orthogonal. It represents reflections, and

we call
n o
R ∈ Orth− := R ∈ Hom (·V, ·V), RRt = Rt R = I, det R = −1
Introducing rotation tensor as R = cos({e}, {e′ }) ∈ Orth+ (Rij is the cosine of

the angle between ei and e′j axes), we have the following transformations:
1. for a vector v: v ′ = Rv

2. for a second order tensor T = v ⊗ u: T ′ = v ′ ⊗ u′ = Rv ⊗ Ru = R(v ⊗
u)Rt = RT Rt .
3. for a n-th order tensor: T ′ = Rv 1 ⊗ Rv 2 ⊗ · · · ⊗ Rv n . It is difficult to write

it in a compact form like second order tensor. However, in the component
form we can write Ti′1 i2 ...in = Ri1 j1 Ri2 j2 . . . Rin jn Tj1 j2 ...jn .
11 Skew-symmetric tensors
u
Wu = w × u
Figure 38: Axial or dual vector.
Let W ∈ Skw and dim ·V = 3 such that W t = −W . Then W has the

following properties
1. u · W v = W t u · v = −v · W u.
2. If v = u, then from the above relation u · W u = 0. This means W u ⊥ u.
3. There exist a vector ω such that ω × u = W u. When ω is given, W is

fixed and vice versa. ω is known as axial or dual vector corresponding to
W (Fig. 38).
If ω = (ω1 , ω2 , ω3 )t , then
 
0 −ω3 ω2
[W ]ij =  ω3 0 −ω1  (123)
−ω2 ω1 0

12 Spectral decomposition of a second order ten-
sor
Let T ∈ Hom (·V, ·V), where dim ·V = 3. If n be the unit eigenvector, by definition
we write
T n = λn or (T − λI)n = 0. (124)
where I is the identity tensor. As we know that there are four possibilities to
express T and I, i.e.,
T = T ij g i ⊗ g j , T = Tij g i ⊗ g j , T = T·ji g i ⊗ g j , T = Ti·j g i ⊗ g j .

I = g ij g i ⊗ g j , I = gij g i ⊗ g j , I = δ·ji g i ⊗ g j = g i ⊗ g i ,
I = δi·j g i ⊗ g j = g i ⊗ g i .
Note that the matrix representations of the last two forms of I are diagonal. Now
going back to the eigenvalue problem, we have to make sure that both T and I
should be expressed with respect to the same basis for the subtraction operation.
Let us consider the form T = Tij g i ⊗ g j (also Tij = Tji ), then we should consider
I = gij g i ⊗ g j . Let n = nk g k (we can also consider n = nk g k ), then
h i
(T − λI)n = (Tij − λgij )g i ⊗ g j nk g k
= (Tij − λgij )nk δkj g i
= (Tij − λgij )nj g i
Thus from (124) we get
(Tij − λgij )nj = 0
The above form is known as generalized eigenvalue problem, i.e., in the form of
An = λBv. However, solving a simple eigenvalue problem is convenient. Note
that nj = g jk nk and we write
(Tij − λgij )g jk nk = 0, (Tij g jk − λgij g jk )nk = 0,

(Ti·k − λδik )nk = 0,
(Ti·j − λδji )ni = 0 (just renaming the indices)
The condition for non-zero eigenvectors is
det(Ti·j − λδji ) = 0

and the characteristics polynomial is written as:
P(λ) = λ3 − I1 λ2 + I2 λ − I3 ,
1
where I1 = tr T = λ1 + λ2 + λ3 , I2 = [tr (T )2 − tr T 2 ] = λ1 λ2 + λ2 λ3 + λ3 λ1 , and
2
det T = λ1 λ2 λ3 .
Theorem 12.1 Let T ∈ Hom (·V, ·V), and dim ·V = 3. If T has three distinct eigenvalues,
and n(i) is the eigenvector corresponding to i-th eigenvalue, then {n(1) , n(2) , n(3) }
is a basis of ·V.
Consider,
c1 n(1) + c2 n(2) + c3 n(3) = 0
Operating both sides with (T − λ1 I) we get
c1 (T − λ1 I)n(1) + c2 (T − λ1 I)n(2) + c3 (T − λ1 I)n(3) = 0
As n(1) is the eigenvalue of λ1 , we have (T − λ1 I)n(1) = 0. We write,
c2 (T n(2) − λ1 n(2) ) + c3 (T n(3) − λ1 n(3) ) = 0

⇒ c2 (λ2 n(2) − λ1 n(2) ) + c3 (λ3 n(3) − λ1 n(3) ) = 0
⇒ c2 (λ2 − λ1 )n(2) + c3 (λ3 − λ1 )n(3) = 0
Operating the last equation with (T − λ2 I) we get
c2 (λ2 − λ1 )(T − λ2 I)n(2) + c3 (λ3 − λ1 )(T − λ2 I)n(3) = 0
As n(2) is the eigenvalue of λ2 , we have (T − λ2 I)n(2) = 0. We write,
c3 (λ3 − λ1 )(T − λ2 I)n(3) = 0

c3 (λ3 − λ1 )(T n(3) − λ2 n(3) ) = 0
c3 (λ3 − λ1 )(λ3 − λ2 )n(3) = 0
Now n(3) is not a null vector, and all the eigenvalues are distinct, i.e., λ1 ̸= λ2 ̸=
λ3 . This means c3 = 0. Similarly, starting with the operators (T − λ2 I), and
(T − λ3 I) give c1 = 0. And starting with the operators (T − λ1 I), and (T − λ3 I)
give c2 = 0. This means {n(1) , n(2) , n(3) } is linearly independent. Now, one can
express any vector v = v 1 n(1) + v 2 n(2) + v 3 n(3) , i.e., it spans the space. This
shows {n(1) , n(2) , n(3) } is a basis. One can extend this algorithm to show that if
dim ·V = n, {n(1) , n(2) , . . . , n(n) } is a basis.

12.1 Spectral decomposition of a symmetric second order tensor
12.1 Spectral decomposition of a symmetric second order

tensor
Theorem 12.2 Let T ∈ Hom (·V, ·V), and also T ∈ Sym , where dim ·V = 3.
Then any two distinct eigenvectors are orthonormal. If T has three distinct eigenvalues,
and n(i) is the eigenvector corresponding to i-th eigenvalue, then {n(1) , n(2) , n(3) }
is an orthonormal basis of ·V.
Let
T n(i) = λi n(i) , and T n(j) = λj n(j) , (i ̸= j)
Now,
n(j) · T n(i) = n(j) · λi n(i) , ⇒ T t n(j) · n(i) = λi n(j) · n(i) , ⇒ T n(j) · n(i) = λi n(j) · n(i) .
We write,
λj n(j) · n(i) = λi n(j) · n(i) , ⇒ (λj − λi )n(j) · n(i) = 0.
As λi ̸= λj , we get n(j) ⊥ n(i) . So, for three distinct eigenvalues, {n(1) , n(2) , n(3) }
is an orthonormal basis.
For three distinct eigenvalues, when {n(1) , n(2) , n(3) } is an orthonormal basis
of ·V, we write the tensor T as
T = λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3) . (125)
This is called spectral decomposition of T with respect to the basis system {n(i) ⊗
n(j) }. Note that the matrix form of T will be diagonal in this case. One of the
advantages, among many, of spectral decomposition is the following:
T 2 = (λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3) )(λ1 n(1) ⊗ n(1)
+λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3) )
= λ21 [n(1) (n(1) · n(1) ) ⊗ n(1) ] + λ22 [n(2) (n(2) · n(2) ) ⊗ n(2) ] + λ23 [n(3) (n(3) · n(3) ) ⊗ n(1) ]
= λ21 n(1) ⊗ n(1) + λ22 n(2) ⊗ n(2) + λ23 n(3) ⊗ n(3)
In general,
T n = λn1 n(1) ⊗ n(1) + λn2 n(2) ⊗ n(2) + λn3 n(3) ⊗ n(3) , (n ∈ R). (126)

12.1 Spectral decomposition of a symmetric second order tensor
12.1.1 Two repeated eigenvalues

Consider λ2 = λ3 = λ ̸= λ1 . Then we have,
T n(1) = λ1 n(1) , T n(2) = λn(2) , T n(3) = λn(3) .
We also have,
n(1) ⊥ n(2) and n(1) ⊥ n(3) .
If we consider n̂ = αn(2) + βn(3) , it also satisfy T n̂ = λn̂. We can show this in
the following ways:
T (αn(2) + βn(3) ) = αT n(2) + βT n(3) = α(λn(2) ) + β(λn(3) ) = λ(αn(2) + βn(3) ).
This means that any linear combinations of the eigenvectors n(2) and n(3) will
n(3)
n(1) 2
1 n(2)
Figure 39: Any vector in the 2 − 3 plane, i.e., linear combination of n(2) and n(3)
is an eigenvector.
be an eigenvector with the eigenvalue λ. Moreover, n(2) and n(3) need not to be
orthonormal to each other. Note that the choice of n(1) is unique and it has to be
perpendicular to the plane containing n(2) and n(3) (Fig. 39). If we select that
n(2) ⊥ n(3) , the spectral decomposition in this case becomes
T = λ1 n(1) ⊗ n(1) + λn(2) ⊗ n(2) + λn(3) ⊗ n(3)
= (λ1 − λ)n(1) ⊗ n(1) + λI,
where the identity tensor is given by
I = n(1) ⊗ n(1) + n(2) ⊗ n(2) + n(3) ⊗ n(3) .

12.2 Symmetric positive definite
12.1.2 Three repeated eigenvalues

In this case λ1 = λ2 = λ3 = λ, means every directions are eigenvectors. If we
select any three orthonormal vectors, then the spectral decomposition becomes
T = λI,
I = n(1) ⊗ n(1) + n(2) ⊗ n(2) + n(3) ⊗ n(3) .
12.2 Symmetric positive definite

A second order tensor T is positive definite if for all v ∈ ·V,
v · T v ≥ 0, v · T v = 0 =⇒ v = 0.
If we decompose a tensor to its symmetric and skew-symmetric part, T = T s +T ss ,

then
v · T v = v · (T s + T ss )v = v · T s v + v · T ss v = v · T s v + 0.
So, even for a general tensor, it is sufficient to study the symmetric part of it.
When a symmetric tensor T is positive definite, we say T ∈ Psym .
Theorem 12.3 Let T ∈ Hom (·V, ·V), and also T ∈ Sym , where dim ·V = 3. If
T has strictly positive eigenvalues, then T ∈ Psym .
We consider the form
T = λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3)
where λ1 , λ2 , λ3 > 0. Let v = v1 n(1) + v2 n(2) + v3 n(3) . Then,
T v = (λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3) )(v1 n(1) + v2 n(2) + v3 n(3) )
= λ1 v1 n(1) + λ2 v2 n(2) + λ3 v3 n(3) .
So,
v · T v = λ1 v12 + λ2 v22 + λ3 v32 > 0,
and T ∈ Psym .
Theorem 12.4 Let T ∈ Hom (·V, ·V), and also T ∈ Psym , where dim ·V = 3.
Then all the eigenvalues of T are strictly positive.

12.3 Cayley-Hamilton theorem [3D]
Let T n = λn. Then,

n · Tn
n · T n = λn · n, ⇒ λ = .
n·n
By definition, as T ∈ Psym , n · T n > 0. Also n · n > 0. These two facts prove
that λ > 0.
Theorem 12.5 Let T ∈ Hom (·V, ·V), and also T ∈ Psym , where dim ·V = 3.
Then there exists a U ∈ Psym such that U 2√= T . The tensor U is called positive
definite square root of T and we write U = T .
Let us start from T n = λn, where λ > 0 as T ∈ Psym . We then write
(T − λI)n = 0
(U 2 − λI)n = 0
√ √
(U + λI)(U − λI)n = 0
√
Let n̂ = (U − λI)n. Then continuing we get
√ √
(U + λI)n̂ = 0 ⇒ U n̂ = − λn̂
√
This mean − λ < 0 is an eigenvalue of U . This is not possible as U ∈ Psym .
We the only option is n̂ = 0, and we write
√ √
(U − λI)n = 0 ⇒ U n = λn.
√
This shows n is also the eigenvectors of U with eigenvalues λ. If the spectral
decomposition of T is
T = λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3)
then
p p p
U= λ1 n(1) ⊗ n(1) + λ2 n(2) ⊗ n(2) + λ3 n(3) ⊗ n(3) .
12.3 Cayley-Hamilton theorem [3D]

From (115) we write
(Cof T )t T = I2 T − I1 T 2 + T 3 ⇒ (det T )I = I2 T − I1 T 2 + T 3 .
Or
T 3 − I1 T 2 + I2 T − I3 I = 0.
Operating on n, one can write (λ3 −I1 λ2 +I2 λ−I3 )n = 0 or λ3 −I1 λ2 +I2 λ−I3 = 0
as n ̸= 0. So, T satisfies its own characteristic equation λ3 − I1 λ2 + I2 λ − I3 = 0.

13 Representation of a rigid rotation tensor
Qq
Qr
p
r
Qp = p, (λ = 1)
Figure 40: p is the eigenvector with λ = 1.
Let R ∈ Orth, and let us study the associated eigenvalue problem Rn = λn.
We write,
Rn · Rn = λ2 n · n ⇒ n · Rt Rn = λ2 n · n ⇒ n · n = λ2 n · n.
This gives that the only real eigenvalues are λ = ±1. Further considering R is a
rotation tensor, i.e., R ∈ Orth+ , we have the following information:
Rt R = I, ⇒ Rt (R − I) = −(R − I)t ⇒ (det Rt ) det(R − I) = − det(R − I)t
As det Rt = 1, we have det(R − I) = 0. So, one real eigenvalue for a rotation
tensor is λ = 1. We have Rp = p, where p is the associated eigenvector. p is
known as the axis of rotation which does not change if operated by the rotation
tensor R. Note that Rt p = p as Rt Rp = Ip means Rt (Rp) = p or Rt p = p.
Let us select two other mutually perpendicular unit vectors q and r, which are
perpendicular to the axis p. We then construct a basis {p, q, r}. We have,
p · Rq = Rt p · q = p · q = 0 ⇒ p ⊥ Rq
p · Rr = Rt p · r = p · r = 0 ⇒ p ⊥ Rr
Rr · Rq = r · (Rt R)q = r · q = 0 ⇒ Rr ⊥ Rq.
This means, q and r, after transformation remains perpendicular to the axis of
rotation p. The initial angle (90◦ ) between q and q remains unchanged after

transformation (Fig. 40). So, Rq and Rr lie in the q-r plane and could be span
by the uni basis {q, r}. We can write
Rq = αq + βr, Rr = γq + δr (127)
Imposing the constraint that det R = 1, we can write

[Rp, Rq, Rr] [p, αq + βr, γq + δr]
det R = = =1
[p, q, r] [p, q, r]
This gives the constraint condition αδ − βγ = 1. Therefore, there exists an angle
θ, known as angle of rotation, that satisfies constraint relation by selecting α =
δ = cos θ, and β = −γ = sin θ. Substituting back in (127), we write
Rq = cos θq + sin θr, Rr = − sin θq + cos θr (128)
Note that |Rq| = |q| = 1, |Rq| = |q| = 1, and θ is the angle of rotation from q.
We then have the following components of R
Rpp = p · Rp = p · p = 1, Rpq = p · Rq = 0, Rpr = p · Rr = 0

Rqp = q · Rp = q · p = 0, Rqq = q · Rq = α, Rqr = q · Rr = γ
Rrp = r · Rp = r · p = 0, Rrq = r · Rq = β, Rrr = r · Rr = δ.
The expression of R with respect to the basis {p, q, r} becomes,
R = p ⊗ p + αq ⊗ q + δr ⊗ r + γq ⊗ r + δr ⊗ q.
Substituting the parametric relation of θ for the constants α, β, γ, δ we get
R = p ⊗ p + (q ⊗ q + r ⊗ r) cos θ + (r ⊗ q − q ⊗ r) sin θ.
14 Representation of a skew-symmetric tensor

Let p be a eigenvector of W ∈ Skw such that W p = λp. Then p · W p = λp · p
and so λp · p = 0 ⇒ λ = 0. This also implies W p = 0 ⇒ ω × p = 0. So, the axial
vector ω is parallel to p and we write ω = ωp.
We then construct an orthonormal basis {p, q, r} such that p = q×r, q = r×p,
and r = p × q. We write
p · W q = W t p · q = −W p · q = 0 ⇒ p ⊥ W q
p · W r = W t p · r = −W p · r = 0 ⇒ p ⊥ W r
p · W p = q · W q = r · W r = 0 = q · W p = r · W p.

q
Wr
Wq
p
r
Wp = 0, (λ = 0)
Figure 41: p is the eigenvector with λ = 0.
We then have the following components of W
Wpp = p · W p = 0, Wpq = p · W q = 0, Wpr = p · W r = 0

Wqp = q · W p = 0, Wqq = q · W q = 0, Wqr = q · W r = −ω
Wrp = r · W p = 0, Wrq = r · W q = ω, Wrr = r · W r = 0.
We used the fact
q · W r = q · (ω × r) = ω · (r × q) = −ω · p = −ωp · p = −ω
r · W q = W t r · q = −W r · q = ω
The final form of W becomes
W = ω(r ⊗ q − q ⊗ r).

15 Tensor calculus
15.1 Normed space
Definition 15.1 Let V is a vector space together with a function || · || : V → R,
called a norm, satisfying the following four conditions.
1. ||v|| > 0 for all v ∈ V
2. ||v|| = 0 if and only if v = 0
3. ||av|| = |a| · ||v|| for all v ∈ V and a ∈ R
4. ||v + w|| ≤ ||v|| + ||w|| for all v, w ∈ V. (triangle inequality)
Then the vector space is called normed linear space of simply normedspace
Example 1: If V = R, then ||v|| = |v|, v ∈ R.
p
Example 2: If V = En (Euclidean), then ||v|| = v12 + · · · + vn2 .
Example 3: If V = E2 (Euclidean), then we can define many norms. For example,
||v|| = |v1 | + |v2 |.
If V is an inner product space ·V, then we can say that

√
||v|| = v · v,
the length of the vector v. Note that normed space does not always mean an inner
product space, but an inner product space can always have a norm. Norms always
give rise to a metric, which is the distance between two points. Here the points
means the members of the set. The distance function between any two points are
written as
d(v, w) = ||w − v||.
This particular norm, among many other possible norms, known as metric.
15.2 Differentiation
Let ·V and ·U are the two inner product spaces with a metric norm. Let v ∈ V
and a map f (v) ∈ ·U. Note that f need not to be a linear transformation, i.e.,

f v = u. It is a mapping with the argument v. f could be a scalar valued, vector

valued or a tensor valued functions. For example,
f (v) = v · v ∈ ·U = R is a vector-valued scalar mapping/function

f (v) = av ∈ ·U = ·V is a vector-valued vector mapping
f (v) = v ⊗ v ∈ ·U = Hom (·V, ·V) is a vector-valued tensor mapping
and so on. Let f be defined at the neighborhood of 0 ∈ ·V and have values in ·U.
If f (v) approaches 0 ∈ ·U faster than v, we write
f (v) = o(v), as v → 0.
or simply
f (v) = o(v)
if
||f (v)||
lim = 0.
v →0 ||v||
For example, consider f (t) = ta for a > 1, then f (t) = o(t). Let D be a subset of
Figure 42: When v approaches 0 ∈ ·V, F (v) approaches 0 ∈ ·U faster than v.
·V and consider f : D → ·U (Fig. 42). We say f is differentiable at x ∈ D if the

difference
f (x + v) − f (x)

is equal to a linear transformation of v (say, T v ∈ ·U) plus a term (i.e. o(v) ∈ ·U)
that approaches zero faster than v. We write
f (x + v) − f (x) = D x f (x)[v] + o(v) (129)
where T is rewritten as D x f (x), known as derivative (or Fréchet derivative) of f

with respect to x and it belongs to Hom (·V, ·U). Alternatively we can write
||f (x + v) − f (x) − D x f (x)[v]||
lim = 0. (130)
v →0 ||v||
Note that the vector v could be any vector in ·V. Now, we freeze the direction
of v, and shrinks or expand v by considering the vector αv, where α ∈ R. From
(129) we can write
f (x + αv) − f (x) = D x f (x)[αv] + o(αv). (131)
Note that v is fixed and α is a variable. This means o(αv) approaches zero faster
than α, i.e., o(αv) = o(α) and we write
||f (x + αv) − f (x) − D x f (x)[αv]||
lim = 0
α→0 ||α||
f (x+αv )−f (x)
|| α
− D x f (x)[v]||
lim |α| = 0
α→0 |α|
f (x + αv) − f (x)
|| lim − D x f (x)[v]|| = 0.
α→0 α
The final form becomes
f (x + αv) − f (x)
D x f (x)[v] = lim , (132)
α→0 α
and we call D x f (x) directional derivative or Gateaux derivative. If we further
expand f (x + αv) in a Taylor series, we get
∂f (x) 1 ∂f (x) 2
f (x + αv) = f (x) + (αvi ) + (α vi vj ) . . . .
∂xi 2 ∂xi ∂xj
Now,
d ∂f (x)
f (x + αv) = (vi )
dα α=0 ∂xi
d2 ∂f (x)
f (x + αv) = (vi vj )
dα2 α=0 ∂xi ∂xj
..
. (133)

Then one can write

d 1 d2
f (x + αv) = f (x) + f (x + αv) α+ f (x + αv) α2 . . . .
dα α=0 2 dα2 α=0
So, (132) reads
d
D x f (x)[v] = f (x + αv) (134)
dα α=0
Fréchet derivative pertains if Gateaux derivative exists at each direction. So it

may happen that Gateaux derivative exists at a particular direction but not at all
direction. In such a case Fréchet derivative does not exist.
Example 57 Find the derivative and directional derivative (along v ∈ ·V) of

ϕ(u) = u · u, where u ∈ D ⊂ ·V and ϕ : ·V → R (vector-valued scalar function)
First calculate
ϕ(u + v) = (u + v) · (u + v) = u · u + u · v + v · u + v · v
= ϕ(u) + 2u · v + o(v).
Then,
ϕ(u + v) − ϕ(u) = 2u · v + o(v)
and comparing it with (129) we get D u ϕ(u)[v] = 2u · v.

In this case, the derivative D u ϕ(u) is a linear functional in this case as it maps from
·V to R, i.e., D u ϕ(u) ∈ Hom (·V, R) = ·V ∗ ≡ ·V. So, D u ϕ(u) is a vector and it
operates on a vector v to produce a scalar. So we identify D u ϕ(u)[v] = D u ϕ(u)·v
and the derivative D u ϕ(u) = 2u. Note that the operation on [v] is a general
expression and for this case it becomes a dot product.
For the direction derivative we write,
d
D u ϕ(u)[v] = ϕ(u + αv)
dα α=0
d
= [(u + αv) · (u + αv)]
dα α=0
d
= (u · u + αu · v + αv · u + α2 v · v)
dα α=0
= (u · v + 2v · u + 2αv · v)
α=0
= 2u · v
So the directional derivative D u ϕ(u) = 2u.

Example 58 Find the derivative of ϕ(A) = det A, with respect to A, where

A ∈ Hom (·V, ·V) (tensor-valued scalar function) and dim ·V = 3
Recall that
det(A − λI) = −λ3 + I1 λ2 − I2 λ + I3
For λ = −1, we can write
det(A + I) = 1 + I1 + I2 + I3 .
Now selecting V ∈ Hom (·V, ·V)
ϕ(A + V ) = det(A + V ) = det A(I + A−1 V ) = det A det(I + A−1 V )

= det A(1 + tr (A−1 V ) + o(V ))
= det A + det A tr (A−1 V ) + o(V ).
= det A + det A tr (A−1 V ) + o(V ).
So,
D A ϕ(A)[V ] = det A tr (A−1 V )

= (det A)A−t : V
So,
D A ϕ(A) = (det A)A−t (135)
In this case, the derivative D A ϕ(A) : Hom (·V, ·V) → R and so D A ϕ(A) ∈
Hom (·V, R) = ·V ∗ ≡ ·V. So, D A ϕ(A) ∈ Hom (Hom (·V, ·V), R) =
∗ ∗ ∗
Hom (·V, ·V) = Hom (·V , ·V ) ≡ Hom (·V, ·V). So it is a second order tensor.
We identify D A ϕ(A)[V ] = D A ϕ(V ) : V , as two second order tensors produce a
scalar only by a double contraction.
Then
D A ϕ(A)[V ] = D A ϕ(A) : V = (det A)A−t : V (136)
So,
D A ϕ(A) = (det A)A−t (137)

15.3 Partial derivative
15.3 Partial derivative

Consider f (A, u) which is a vector function with a vector (u) and a tensor (A)
as its arguments. Then, the partial derivative of f with respect to u is defined as
d
∂u f (A, u) = D u f (A, u) = f (A, u + αv) (138)
dα α=0
Similarly, the partial derivative of f with respect to A is defined as

d
∂A f (A, u) = D A f (A, u) = f (A + αV , u) (139)
dα α=0
Example 59 Find the derivative of ϕ(A, u) = u · Au, where A ∈ Hom (·V, ·V),
u ∈ D ⊂ ·V and ϕ : Hom (·V, ·V) × ·V → R (tensor- and vector valued scalar
function)
We denote Dϕ(A, u)[v] = ∂u ϕ(A, u)[v] and Dϕ(A, u)[V ] = ∂A ϕ(A, u)[V ], where
v ∈ ·V and V ∈ Hom (·V, ·V).
ϕ(A, u + v) = (u + v) · A(u + v) = u · Au + u · Av + v · Au + v · Av
= ϕ(A, u) + At u · v + Au · v + o(v).
So,
ϕ(A, u + v) − ϕ(A, u) = At u · v + Au · v + o(v)

= (At u + Au) · v + o(v).
=⇒ ∂u ϕ(A, u) = At u + Au
Now following the definition (138),
d
∂u ϕ(A, u)[v] = ϕ(A, u + αv)
dα α=0
d
= (u + αv) · (A(u + αv))
dα α=0
d
u · Au + αv · Au + αu · Av + α2 v · Av

=
dα α=0

= v · Au + u · Av + 2αv · Av
α=0
= v · Au + u · Av
∂u ϕ(A, u) · v = (Au + At u) · v =⇒ ∂u ϕ(A, u) = Au + At u
Similarly ∂A ϕ(A, u) = u ⊗ u. (work this out!).

15.4 Product rule for bilinear mapping
Figure 43: x ∈ D ⊂ ·V, v ∈ ·V and f : D → ·F, and g : D → ·G. π : ·F ×·G → ·H

is a general bilinear mapping, where h(x) ∈ ·H.
15.4 Product rule for bilinear mapping

It will be frequently necessary to compute the derivative of a product π(f (x), g(x))
of two mappings f (x) and g(x), where x ∈ D ⊂ ·V. Moreover, π is a general
bilinear mapping (Fig. 44)
π : ·F × ·G → ·H
Depending on the vector spaces, the structure of the mapping π is determined.

For example, consider the following bilinear mappings.
π(ϕ, w) = ϕw ϕ ∈ ·F = R, w ∈ ·G, and π ∈ ·H = ·G

π(u, w) = u·w u ∈ ·F, w ∈ ·G, and π ∈ ·H = R
π(u, w) = u⊗w u ∈ ·F, w ∈ ·G, and π ∈ ·H = Hom (·F, ·G)
π(S, w) = Sw S ∈ ·F = Hom (·G, ·G), say, w ∈ ·G, and π ∈ ·H = ·G

15.5 Chain rule
and so on. Consider h(x) = π(f (x), g(x)) and, if v ∈ ·V, then
h(x + v) =
π(f (x + v), g(x + v))
=
π(f (x) + D x f (x)[v] + o(v), g(x) + D x g(x)[v] + o(v))
≈
π(f (x) + D x f (x)[v], g(x) + D x g(x)[v])
=
π(f (x), g(x) + D x g(x)[v]) + π(D x f (x)[v], g(x) + D x g(x)[v])
=
π(f (x), g(x)) + π(f (x), D x g(x)[v])
+π(D x f (x)[v], g(x)) + π(D x f (x)[v], Dg(x)[v])
= h(x) + π(f (x), D x g(x)[v]) + π(D x f (x)[v], g(x)) + o(v)
h(x + v) − h(x) = π(f (x), D x g(x)[v]) + π(D x f (x)[v], g(x)) + o(v)
So,
D x h(x)[v] = π(D x f (x)[v], g(x)) + π(f (x), D x g(x)[v]), (140)
If ·V = R, and t, α ∈ R then
D t h(t)[α] = π(D t f (t)[α], g(t)) + π(f (t), D t g(t)[α])

D t h(t)α = π(D t f (t)α, g(t)) + π(f (t), D t g(t)α)
= απ(D t f (t), g(t)) + απ(f (t), D t g(t)) as π is bilinear
and
d
D t h(t) = h(t) = ḣ(t) = π(ḟ (t), g(t)) + π(f (t), ġ(t)). (141)
dt
We have the following product rules for different π:

d
(ϕv) = ϕ̇v + ϕv̇
dt
d
(u · v) = u̇ · v + u · v̇
dt
d
(u ⊗ v) = u̇ ⊗ v + u ⊗ v̇
dt
d
(T S) = Ṫ S + T Ṡ
dt
d
(T : S) = Ṫ : S + T : Ṡ
dt
d
(T v) = Ṫ v + T v̇
dt

15.5 Chain rule
Figure 44: x ∈ D ⊂ ·V, v ∈ ·V and f : D → ·F, and g : D → ·G.
15.5 Chain rule

Consider the composition function h = g ◦ f . Then h(x) = g(f (x)) as shown in
Fig. 44.
h(x + v) = g(f (x + v)) = g(f (x) + D x f (x)[v] + o(v)))
≈ g(f (x) + D x f (x)[v])) = g(f (x)) + D f (x) g(f (x))[Df (x)[v]]
= h(x) + D f (x) g(f (x))[D x f (x)[v]] + o(Df (x)[v]).
So we write the chain rule as
D x h(x)[v] = D f (x) g(f (x))[D x f (x)[v]]. (142)
If D ⊂ ·V = R, then
D t h(t)[α] = D f (t) g(f (t))[D t f (t)[α]]
D t h(t)α = D f (t) g(f (t))[D t f (t)α]
D t h(t)α = D f (t) g(f (t))[D t f (t)]α
D t h(t) = D f (t) g(f (t))[D t f (t)]
d ˙
h(t) = D f (t) g(f (t))[f (t)]. (143)
dt
Example 60 Find dtd (det A(t))
Note that ·V = R. Let f (t) = A(t), and g = det() such that g(f (t)) = det A(t).
So, following equation (136) we write
d ˙ = (det A)A−t : Ȧ
(det A(t)) = D A(t) (det A(t))[A(t)] (144)
dt
15.6 Gradient and divergence

Consider x ∈ D ⊂ ·V, v ∈ ·V and ·F and ·G are the two inner product spaces
with a metric norm.
15.6.1 Gradient
1. Consider a vector valued scalar function
h(x) : D → R.
then we write
D x h(x)[v] = grad ϕ(x) · v
Note that grad ϕ ∈ ·V is a vector.
2. Similarly consider a vector values vector function
h(x) : D → ·V
then
D x h(x)[v] = [grad h(x)]v
and here grad h(x) ∈ ·V ⊗ ·V is a second order tensor. We can keep on

continue in this notion for the higher order tensors.
3. In general, if
h(x) : D → ⊗n−1 · V, i.e, h(x) ∈ ⊗n · V
then
D x h(x)[v] = [grad h(x)]v
and here grad h(x) ∈ ⊗n+1 · V is a (n + 1)th order tensor.9
Example 61 Show that grad (ϕw) = w⊗grad ϕ+ϕ grad w, where ϕ(x) ∈ ·F =
R, and w(x) ∈ ·G.
9
The notation ⊗n · V = ·V ⊗ ·V ⊗ · · · ⊗ ·V means n copies of ·V.

Consider h = π(ϕ, w) = ϕw. From product rule
D x h(x)[v] = π(D x f (x)[v], g(x)) + π(f (x), D x g(x)[v])
we write
D x h(x)[v] = π(D x ϕ(x)[v], w(x)) + π(ϕ(x), D x w(x)[v])

= π(grad ϕ(x) · v, w(x)) + π(ϕ(x), [grad w(x)]v)
= (grad ϕ · v)w + ϕ([grad w]v)
= (w ⊗ grad ϕ)v + (ϕ grad w)v
h i
grad h v = w ⊗ grad ϕ + ϕ grad w v
So,
grad (ϕw) = w ⊗ grad ϕ + ϕ grad w (145)
Example 62 Show that grad (u · w) = (grad u)t w + (grad w)t u, where where
u(x) ∈ ·F, and w(x) ∈ ·G.
Consider h = π(u, w) = u · w. From product rule
we write
D x h(x)[v] = π(D x u(x)[v], w(x)) + π(u(x), D x w(x)[v])

= π([grad u]v, w) + π(u, [grad w]v)
= [grad u]v · w + u · [grad w]v
= v · [grad u]t w + [grad w]t u · v
(grad h) · v = ([grad u]t w + [grad w]t u) · v
So,
grad (u · w) = [grad u]t w + [grad w]t u (146)
Example 63 Show that grad (S t u) = S t (grad u)+u·grad S, where u(x) ∈ ·F,

and S(x) ∈ Hom (·F, ·F)
Consider h = π(S, u) = S t u = u · S. Note that grad S is a third order tensor
and u · A is ui Aijk , when A is a third order tensor. Then u · (Aw) = ui Aijk wk =
(ui Aijk )wk = (u · A)w. From the product rule

we write
D x h(x)[v] = π(D x u(x)[v], S(x)) + π(u(x), D x S(x)[v])
= π([grad u(x)]v, S(x)) + π(u(x), [grad S(x)]v)
= [grad u]v · S + u · [grad S]v

grad h v = [grad u]v · S + u · [grad S]v
= S t [grad u]v + (u · [grad S])v
h i
t
= S [grad u] + u · [grad S] v
we write
grad (S t u) = S t (grad u) + u · (grad S) (147)
15.6.2 Divergence
Divergence will make sense when T (x) is a vector or higher order tensor. Consider
T (x) ∈ ⊗n · V Then
div T (x) = grad T (x) : I
where I ∈ ·V ⊗·V as long as the argument variable x ∈ ·V. Note that grad T (x) ∈
⊗n+1 · V and so div T (x) ∈ ⊗n+1−2 · V = ⊗n−1 · V. If n = 1, then without any
ambiguity we can write
div T (x) = grad T (x) : I = tr (T (x))
Example 64 Show that div (ϕw) = w · grad ϕ + ϕ div w, where ϕ(x) ∈ ·F = R,
and w(x) ∈ ·G.
We know grad (ϕw) from (145) . Taking trace for both sides
tr (grad (ϕw)) = tr (w ⊗ grad ϕ) + tr (ϕ grad w)
=⇒ div (ϕw) = w · grad ϕ + ϕ div w (148)
Example 65 Show that div (S t u) = S : (grad u) + u · div S, where u(x) ∈ ·F

and S(x) ∈ Hom (·F, ·F).
We know grad (S t u) from (147) . Taking trace for both sides
grad (S t u) : I = S t (grad u) : I + u · (grad S) : I
grad (S t u) : I = S t (grad u) : I + u · ((grad S) : I)
tr (grad (S t u)) = tr (S t (grad u)) + u · div S
=⇒ div (S t u) = S : (grad u) + u · div S (149)

If we consider the vector u = a, a constant, then
a · div S = div (S t a). (150)
Sometimes the above relation is used as the definition of the divergence of a second
order tensor.
Example 66 Show that div (u ⊗ w) = u(div w) + (grad u)w, where u(x) ∈ ·F,
and w(x) ∈ ·G.
Let a be a constant vector. Then
div (u ⊗ w) · a = div ((w ⊗ u)a) (by using(150))
= div ((u · a)w) = w · grad (u · a) + (u · a) div w (by using(148))
h i
= w · (grad u)t a + (u · a) div w (by using(146))
= (grad u)w · a + u (div w) · a
So we get
div (u ⊗ w) = (grad u)w + u (div w) (151)
Example 67 Show that div (ϕS) = ϕ(div S) + S grad ϕ, where ϕ ∈ R, S ∈

Hom (·V, ·V)
div (ϕS) · a = div ((ϕS t )a) (by using(150))
= div (ϕ(S a)) = (S t a) · grad ϕ + ϕ div (S t a)
t
(by using(148))
= a · S grad ϕ + a · ϕdiv S (by using(150))
So we get
div (ϕS) = S grad ϕ + ϕ div S (152)
15.6.3 Curl
Consider the map
ϕ : Rn → R3 .
Then curl is defined as
(grad ϕ(x) − (grad ϕ(x))t )v = curl ϕ(x) × v
Generalizing curl in any dimension and any order requires exterior calculus, which
is out of scope in this course.

15.7 Divergence theorem
15.7 Divergence theorem

Consider a bounded region D ⊂ ·V with a boundary ∂D. x is a generic point on
D. Moreover, we have an outward unit normal n at the boundary points. We
denote a differential area as dA ∈ ∂D and a differential volume element dV ∈ D.
If
ϕ:D→R u : D → ·F S : D → Hom (·F, ·F)
then the divergence theorem gives us

Z Z
ϕn dA = grad ϕ dV (153)
Z ∂D ZD
u · n dA = div u dV (154)
∂D
Z ZD
Sn dA = div S dV (155)
∂D D
You are already familiar with (153) and (154) from your elementary vector calculus
course. So, we will prove only (155).
Z Z Z
a· Sn dA = a · Sn dA = S t a · n dA
∂D Z∂D ∂D
Z
t
= div (S a) dV = a · div (S) dV
DZ D
= a· div (S) dV
Z Z D
=⇒ Sn dA = div S dV
∂D D
R R
Example 68 Show that ∂D
(u ⊗ n) dA = D
grad u dV
Z Z Z
a· (u ⊗ n) dA = a · (u ⊗ n) dA = (a · u)n dA
∂D Z∂D Z∂D
= grad (a · u) dV = (grad u)t a dV

DZ D
= a· grad u dV
Z Z D
=⇒ (u ⊗ n) dA = grad u dV
∂D D

15.8 Product rule for any mapping
15.8 Product rule for any mapping

We now remove the restriction that h(x) = π(f (x), g(x)) is bilinear. If v ∈ ·V,
then
h(x + v) =
π(f (x + v), g(x + v))
=
π(f (x) + D x f (x)[v] + o(v), g(x) + D x g(x)[v] + o(v))
≈
π(f (x) + D x f (x)[v], g(x) + D x g(x)[v])
=
π(f (x), g(x)) + D f (x) π(f (x), g(x))[D x f (x)[v]]
+D g(x) π(f (x), g(x))[D x g(x)[v]] + o(v)
= h(x) + ∂f (x) π(f (x), g(x))[D x f (x)[v]]
+∂g(x) π(f (x), g(x))[D x g(x)[v]] + o(v)
h(x + v) − h(x) = ∂f (x) π(f (x), g(x))[D x f (x)[v]]
+∂g(x) π(f (x), g(x))[D x g(x)[v]] + o(v)
So,
D x h(x)[v] = ∂f (x) π(f (x), g(x))[D x f (x)[v]] + ∂g(x) π(f (x), g(x))[D x g(x)[v]],(156)
16 Covariant derivative in Riemannian space

We have all the tools to consider derivative of any tensor mappings which are true
for any vector spaces and free from any basis. We will now consider a Riemannian
space ·V = R3 . Let x ∈ D ⊂ ·V, where D is a region of ·V. Let u(x) be a
local coordinates and {g 1 (u), g 2 (u), g 3 (u)} and {g 1 (x), g 2 (x), g 3 (x)}10 are the
covariant and contravariant basis, respectively. Then we write the metric as
ds2 = gij (u)dui duj
where gij is the metric tensor and positive definite when the space is Riemannian.
Denoting the local coordinate system at a point x + dx by u + du, the change
in covariant basis becomes {g i + dg i }. Now our task is to find dg i .
16.0.1 Christoffel symbols of 2nd kind

We can write,
∂g i j
dg i = du (157)
∂uj
10
You can always substitute back the coordinate relation between x and u to express
{g 1 (x), g 2 (x), g 3 (x)} as {g 1 (u), g 2 (u), g 3 (u)}

∂gi
Now ∂uj
can be spanned by the basis {g k (u)}
∂g i
= Γkij g k
∂uj
and
∂g i k
j
· g = Γkij (u). (158)
∂u
We call Γij as Christoffel symbols of 2nd kind.
k
16.0.2 Christoffel symbols of 1st kind

∂g
Now ∂uji can be spanned by the basis {g k (u)} too. Then
∂g i
j
= Γijk g k
∂u
and
∂g i
· g = Γijk (u). (159)
∂uj k
We call Γijk as Christoffel symbols of 1nd kind. Note that
∂g i ∂g i ∂g
Γijk = j
· gk = j
· gmk g m = gmk ji · g m = gmk Γm
ij . (160)
∂u ∂u ∂u
Also note that
g im Γijk = g im gih Γhjk = δhm Γhjk = Γm
jk (161)
16.0.3 Symmetry properties of Christoffel symbols
∂g i k ∂ ∂x k ∂ ∂x ∂g j k
Γkij = j
· g = j i
· g = i j
· g k
= i
· g = Γkji
∂u ∂u ∂u ∂u ∂u ∂u
and
∂g i ∂ ∂x ∂ ∂x ∂g j
Γijk = · g k = · g k = · g k = · g k = Γjik
∂uj ∂uj ∂ui ∂ui ∂uj ∂ui
∂g
Example 69 Show that ∂uji = −Γimj g m
We take advantage the relation g i · g m = δm i
by taking derivative with respect to
uj we write
∂g i ∂g
j
· g m + g i · mj = 0
∂u ∂u
∂g i
·g = −Γimj
∂uj m
∂g i
= −Γimj g m (162)
∂uj
16.1 Ricci identities
16.1 Ricci identities

We now discuss the relation between the Christoffel symbols with the metric gij .
We denote We start from gij = g i · g j and write
∂
∂k gij = (∂k g i ) · g j + g i · (∂k g j ) (where ∂k ≡ )
∂uk
= Γikj + Γjki . (163)
Now cyclically changing the position of i, j and k we write
∂i gjk = Γjik + Γkij (164)

∂j gki = Γkji + Γijk . (165)
Now (165)+(163)-(164) gives
1
Γijk = − ∂k gij + ∂i gjk + ∂j gki (166)
2
Similarly,
1 mi
Γm
jk
mi
= g Γijk = g − ∂k gij + ∂i gjk + ∂j gki (167)
2
This means that knowing the metric, we will be able to compute Christoffel sym-
bols.
16.2 Gradient and divergence in Riemannian space

This is continuation of Section 15.6. We have the space ·V = Rn with x ∈ D ⊂ Rn .
We consider this Rn space as Euclidean (E n ) or flat and a curvilinear coordinate
system u such that x = ψ(u). The local coordinate system has {g i } as local
(covariant) basis , where {ek } is a basis of E n . So we have a Riemannian space
Rn , with a metric tensor gij , embedded in E n .
16.2.1 Gradient of vector-valued scalar function

Consider a scalar function ϕ(u), such that
ϕ = D → R.
and we have
Dϕ(u)[v] = grad ϕ(u) · v.

for a scalar function ϕ(u). The directional derivative along g i could be written as
D gi ϕ(u)[g i ] = grad ϕ(u) · g i . (168)
From (134) we further write

d
D gi ϕ(u)[g i ] = ϕ(u + αg i )
dα α=0
d
= ϕ(u , u , . . . , ui + α, . . . , un )
1 2
dα α=0
∂ϕ
=
∂ui
where u = uk g k . So we write from (168)
∂ϕ
grad ϕ(u) · g i =
∂ui
which implies
∂ϕ i
grad ϕ(u) = g. (169)
∂ui
16.2.2 Gradient of vector-valued vector mapping

Consider a scalar function ϕ(u), such that
ϕ = D → Rn .
and we have
Dϕ(u)[v] = [grad ϕ(u)]v
for a vector mapping ϕ(u). Now directional derivative along g i could be written
as
D gi ϕ(u)[g i ] = [grad ϕ(u)]g i . (170)
From (134) we further write

d
D gi ϕ(x)[g i ] = ϕ(u + αg i )
dα α=0
d
= ϕ(u1 , u2 , . . . , ui + α, . . . , un )
dα α=0
∂ϕ
=
∂ui
We write from (170)

∂ϕ
[grad ϕ(u)]g i =
∂ui
which implies
∂ϕ
grad ϕ(u) = ⊗ gi. (171)
∂ui
If ϕ = ϕi g i , then
∂ϕ ∂ k ∂ϕk k ∂g k
= (ϕ g k ) = g k + ϕ
∂ui ∂ui ∂ui ∂ui
k
∂ϕ
= g + ϕk Γm ki g m
∂ui k
So, we write
∂ϕ ∂ϕm
i
grad ϕ(u) = i
⊗g = i
+ ϕ Γki g m ⊗ g i = ϕm g m ⊗ g i .
k m
∂u ∂u i
It is a standard practice to denote the covariant derivative of a contravariant

component as
∂ϕm
ϕm = + ϕk Γm
ki
i ∂ui
If ϕ = ϕi g i , in a similar way we get
∂ϕ ∂ k ∂ϕk k ∂g k
= (ϕk g ) = g + ϕk
∂ui ∂ui ∂ui ∂ui
∂ϕk k
= g − ϕk Γkmi g m , from equation (162)
∂ui
So, we write
∂ϕ i
∂ϕ
m k

m i m i
grad ϕ(u) = i
⊗ g = i
− ϕ k mi g ⊗ g = ϕm g ⊗ g .
Γ
∂u ∂u i
It is a standard practice to denote the covariant derivative of a covariant compo-

nent as
∂ϕm
ϕm = − ϕk Γkmi
i ∂ui

16.2.3 Gradient of vector-valued tensor mapping

Consider a tensor function T (u), such that T ∈ Hom (Rn , Rn )
T = D → Hom (Rn , Rn ).
We then write, in a similar procedure as explained in the previous subsections,
∂T
grad T (u) = ⊗ gk . (172)
∂uk
If T = T ij g i ⊗ g j , we write
∂T ∂T ij
k tj i it j k ij
grad T (u) = ⊗ g = + T Γkt + T Γkt g i ⊗ g j ⊗ g = T gi ⊗ gj ⊗ gk .
∂uk ∂uk k
where
∂T ij
T ij = k
+ T tj Γikt + T it Γjkt
k ∂u
If T = Tij g i ⊗ g j , we write
∂T k
∂T
ij t t

i j k ij
grad T (u) = k
⊗ g = k
− T Γ
it jk − Tjt ik g ⊗ g ⊗ g = T
Γ gi ⊗ gj ⊗ gk .
∂u ∂u k
where
∂Tij
Tij = − Tit Γtjk − Tjt Γtik (173)
k ∂uk
Similarly we have T·ij . (work this out!)

k
Example 70 Show that the covariant derivative of the metric coefficient gij is
always zero.
Replacing gij in place of Tij in equation (173)
∂gij
gij = − git Γtjk − gjt Γtik
k ∂uk
∂gij
= − Γjki − Γikj
∂uk
∂gij
= − (Γjki + Γikj )
∂uk
∂gij ∂gij
= − from equation (163)
∂uk ∂uk
= 0 (174)

∂g
Example 71 Show that Γiik = 12 g ij ∂uijk .
As gij = 0, the first line of the previous example reads

k
∂gij
= git Γtjk + gjt Γtik
∂uk
∂gij
g ij k = g ij git Γtjk + g ij gjt Γtik
∂u
= δtj Γtjk + δti Γtik
1 ∂gij
= Γjjk + Γiik = 2Γiik =⇒ Γiik = g ij k (175)
2 ∂u
√
∂ g
Example 72 Show that Γiik = √1 .
g ∂uk
Recall from (144)

d
(det A(t)) = (det A)A−t : Ȧ
dt
Now replace A = g = gij g i ⊗ g j and dtd with ∂u∂ k and write
∂g ∂g
k
= gg −t : , where g = det g
∂u ∂uk
∂gij
= gg ij k
∂u
Now from (175), we write
√
1 ij ∂gij 1 ∂g 1 ∂ g
Γiik = g = =√ (176)
2 ∂uk 2g ∂uk g ∂uk
16.2.4 Divergence of vector-valued vector mapping

We have derived that
grad ϕ(u) = ϕi g i ⊗ g j .
j
Then
div ϕ(u) = grad ϕ(u) : I = tr (grad ϕ(u)) = ϕi
i
i
∂ϕ
= + ϕk Γiki
∂ui √
∂ϕi k 1 ∂ g
= +ϕ √ from (176)
∂ui g ∂uk
√
1 ∂ gϕk
= √ (177)
g ∂uk

16.2.5 Divergence of vector-valued tensor mapping

We have derived for T = T ij g i ⊗ g j
grad T (u) = T ij g i ⊗ g j ⊗ g k .
k
Then
div T (u) = grad T (u) : I = T ij g i ⊗ g j ⊗ g k : δnm g n ⊗ g m

k
ij
= T g i (δnm )(δjn )(δm
k
)
k
= T ij g i δjk
k
ik
= T gi
k
Note that we consider the form I = δnm g n ⊗ g m just to have simplified contraction
operations. You can consider any form among the four possible options. You will
get the same final results. However, intermediate steps will be different.

AE 738 Tensors

Uploaded by

Copyright:

Available Formats

AE 738 Tensors

Uploaded by

Document Information

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

AE 738 Tensors

Uploaded by

Copyright:

Available Formats

AE 738 : Tensors for Engineers

3 Homomorphism of vector spaces or linear transformation 36

3.1 Range space R(T ) . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4 Homomorphism as vector space 50

5 Linear functional or linear form 56

7 Inner product space 74

AE 738 2 Dr. Krishnendu Haldar

8 Symmetric and skew/anti-symmetric tensor 95

10 Orthogonal tensors 116

11 Skew-symmetric tensors 118

12 Spectral decomposition of a second order tensor 119

13 Representation of a rigid rotation tensor 125

14 Representation of a skew-symmetric tensor 126

15 Tensor calculus 128

AE 738 3 Dr. Krishnendu Haldar

15.8 Product rule for any mapping . . . . . . . . . . . . . . . . . . . . . 142

16 Covariant derivative in Riemannian space 142

AE 738 4 Dr. Krishnendu Haldar

a11 , a22 , a12 , a21

a11 , a22 , a12 , a21

a11 , a22 , a12 , a22

1.1 Summation convention

This expression will be written ai xi by introducing the summation convention of

AE 738 5 Dr. Krishnendu Haldar

1.2 Kronecker delta

δ1j xj = δ11 x1 + δ12 x2 + ... + δ1n xn = x1

Thus if we multiply xk by δik , we simply replace the index k of xk by i. Moreover,

δii = δ11 + δ22 + δ33 = 3 (3)

1.3 Dot product

êi · êj = δij (4)

For a Cartesian system, you can think as

a · b = (ai êi ) · (bj êj ) = ai bj (êi · êj ) = ai bj δij = ai bi (5)

AE 738 6 Dr. Krishnendu Haldar

Figure 1: Even(left) and odd(right) permutations.

1.4 Permutation symbol

and we can write

1.5 Cross product

ê1 ê2 ê3

(a × b)i = ϵijk aj bk (8)

If we replace a and b, then

b × a = ϵikj bk aj êi = −ϵijk aj bk êi = −a × b

AE 738 7 Dr. Krishnendu Haldar

Let c = ar êr , then

AE 738 8 Dr. Krishnendu Haldar

Example 1 Show that

Figure 2: Even(left) and odd(right) permutations with i, j, and k.

Example 2 Show that

AE 738 9 Dr. Krishnendu Haldar

1.6 Differentiation notation

where ∇ = e1 ∂x∂ 1 + e2 ∂x∂ 2 + e3 ∂x∂ 3 = ei ∂i , ϕ(x1 , x2 , x3 ) is a scalar function, and

AE 738 10 Dr. Krishnendu Haldar

∇ × ϕu = ej ∂j × ϕuk ek = ϵijk (∂j (ϕuk ))ei

1.7.2 Surjective or onto mapping

AE 738 11 Dr. Krishnendu Haldar

Figure 3: An injective mapping T : X → Y.

Figure 4: A surjective mapping T : X → Y.

1.7.3 Bijective or one-to-one correspondence mapping

AE 738 12 Dr. Krishnendu Haldar

Figure 5: A bijective mapping T : X → Y.

1.7.4 Composition of two mappings