Skip to main content
Published

November 18, 2026

1 Vector Spaces

1.1 Real Vector Spaces

Definition 1 Real vector space

A real vector space is a set \(\mc{V}\) over the real numbers \(\b{R}\) on which the two operations of \(\cdot + \cdot : \mc{V}\times \mc{V} \to \mc{V}\) (addition) and \(\cdot * \cdot: \mc{V}\times \b{R}\to \mc{V}\) (scalar multiplication) are defined and satisfies the following properties:

  1. Additive commutativity
    \(\bs{u}+\bs{v}=\bs{v}+\bs{u}\) for all \(\bs{u},\bs{v} \in \mc{V}\)
  2. Additive associativity
    \(\bs{u}+(\bs{v}+\bs{z})= (\bs{u}+\bs{v})+\bs{z}\) for all \(\bs{u}, \bs{v}, \bs{z} \in \mc{V}\)
  3. Scalar multiplicative associativity
    \((ab)\bs{v} = a(b\bs{v})\) for all \(v \in \mc{V}\) and all \(a,b \in \b{R}\)
  4. Zero (null) element
    There exists a zero element \(\bs{0}_{\mc{V}}\) or simply \(\bs{0}\) when there is no risk of confusion, such that \(\bs{0} + \bs{v} = \bs{v}\) for all \(\bs{v} \in \mc{V}\)
  5. Additive inverse
    For every \(\bs{v} \in \mc{V}\), there exists \(\bs{v}' \in \mc{V}\) such that \(\bs{v} + \bs{v}' = \bs{0}\)
  6. Scalar multiplicative identity
    \(1\bs{v} = \bs{v}\) for all \(\bs{v} \in \mc{V}\)
  7. Distributivity of scalar over vector addition
    \(a(\bs{u}+\bs{v}) = a\bs{u} + a\bs{v}\) for all \(a \in \b{R}\) and all \(\bs{u},\bs{v} \in \mc{V}\)
  8. Distributivity of scalar addition over vector
    \((a+b)\bs{v} = a\bs{v} + b\bs{v}\) for all \(a,b \in \b{R}\) and all \(\bs{v} \in \mc{V}\)

We can show identities such as \(-1 * \bs{v}=\bs{v}'\) and \(0 * \bs{v}=\bs{0}\) from the above axioms.

\[ \begin{align} \pr{1+(-1)}\bs{v}&=1*\bs{v}+(-1)*\bs{v}\\ \pr{0}\bs{v}&=1*\bs{v}+(-1)*\bs{v}\\ \bs{0}&=1*\bs{v}+(-1)*\bs{v}\\ \bs{0}&=\bs{v}+(-1)*\bs{v}\\ (-1) * \bs{v}&=\bs{v}' \end{align} \]

\[ \begin{align} \pr{1+0} * \bs{v}&=1 * \bs{v}+(0) * \bs{v}\\ \pr{1} * \bs{v}&=1 * \bs{v}+(0) * \bs{v}\\ \bs{v}&=\bs{v}+(0) * \bs{v}\\ \bs{0}&=(0)*\bs{v}\\ \end{align} \]

When there is no confusion we will skip explicitly writing \(*\), so expressions such as \(a * \pr{b \bs{v}}=\pr{ab} * \bs{v}\) will appear as \(a\pr{b \bs{v}}=\pr{ab} \bs{v}\)

Examples:

  1. \(\b{R}^{1} = \b{R}\)
  2. \(\b{R}^{2} = \cpr{\pr{a, b} : a, b \in \b{R}}\)
  3. \(\b{P}^{k}= \cpr{p : \b{R} \ra \b{R} ~\vert~ p(x) = a_{0}+a_{1}x+\dots+a_{k}x^{k}}\)

Definition 2 Subspace

A subset \(\mc{U}\) of \(\mc{V}\) is a subspace of \(\mc{V}\) if it is also a vector space using the same addition and scalar operation as on \(\mc{V}\). Namely, \(\mc{U}\) is a subspace only if it satisfies, \(\bs{0} \in \mc{U}\), \(\bs{u} + \bs{v} \in \mc{U}\) for all \(\bs{u} \in \mc{U}\) and \(a \bs{u} \in \mc{U}\) for all \(a \in \b{R}\) and \(\bs{u} \in \mc{U}\).

1.2 Finite Dimensional Vector Spaces

A list of length \(n\) is an ordered collection of \(n\) elements. We denote lists without parenthesis as \(\bs{v}_{1},...,\bs{v}_{n}\). By definition, a list has finite length.

A linear combination of a list of vectors \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) is a vector of the form \(a_{1} \bs{v}_{1} + ... + a_{n} \bs{v}_{n}\) where \(a_{1},...,a_{n} \in \b{R}\).

The span of a list of vectors \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) is the set of all linear combinations of the form \(a_{1} \bs{v}_{1} + ... + a_{n} \bs{v}_{n}\) where \(a_{1},...,a_{n} \in \b{R}\) and is denoted by \(\text{span}\pr{\bs{v}_{1},...,\bs{v}_{n}}\). If \(\bs{v}_{1},...,\bs{v}_{n}\) spans \(\mc{V}\), then \(\text{span}\pr{\bs{v}_{1},...,\bs{v}_{n}} = \mc{V}\)

Definition 3 Finite dimensional vector space

A vector space \(\mc{V}\) is finite dimensional if some list of vectors \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) spans the space.

A list of vectors \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) is said to be linearly independent if its linear combination is \(\bs{0}\) iff the scalars \(a_{1},...,a_{n}\) are all \(0\). On the contrary, a list is linearly dependent if a linear combination is \(\bs{0}\) for some list of scalars \(a_{1},...,a_{n}\) where not all are \(0\). In a finite-dimensional vector space, the length of each linearly independent list of vectors is less than or equal to the length of every spanning list of vectors.

Definition 4 Basis

A basis of \(\mc{V}\) is a linearly independent list of vectors \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) that spans \(\mc{V}\). An equivalent definition is that the list \(\bs{v}_{1},...,\bs{v}_{n} \in \mc{V}\) is basis of \(\mc{V}\) iff every \(\bs{v} \in \mc{V}\) can be written as a unique linear combination of the list.

Every finite dimensional vector space \(\mc{V}\) has a basis. The dimension of \(\mc{V}\) is given by the length of any of its basis. All bases of a finite dimensional vector space have the same length. A proper subspace always has a smaller dimension than the larger space. We will denote finite dimensional vector spaces with a superscript to indicate the dimension, for example, \(\mc{V}^{n}\) or \(P^{k}\).

1.3 Real Inner Products and Norms

Definition 5 Inner product

An inner product on \(\mc{V}\) is an operation \(\apr{\cdot, \cdot} : \mc{V} \times \mc{V} \rightarrow \b{R}\) such that for all \(\bs{u}, \bs{v}, \bs{z} \in \mc{V}\) and for \(a \in \b{R}\), we have

  1. \(\apr{\bs{u}, \bs v} = \apr{\bs v, \bs u}\)
  2. \(\apr{\bs u, \bs v + a \bs z} = \apr{\bs u, \bs v} + a \apr{\bs u, \bs z}\)
  3. \(\inpd{\bs u}{\bs u} \geq 0\)
  4. \(\inpd{\bs u}{\bs u} = 0\) iff \(\bs{u} = \bs{0}\)

An real inner product space is a real vector space on which an inner product is defined. An inner product is not a requirement of a vector space.

Definition 6 Norm

A norm on \(\mc{V}\) is an operation \(\norm{\cdot} : \mc{V} \rightarrow \b{R}_{\geq 0}\) such that such that for all \(\bs{u}, \bs{v} \in \mc{V}\) and for \(a \in \b{R}\), we have

  1. \(\norm{a \bs u} = \abs{a} \norm{\bs u}\)
  2. \(\norm{\bs u + \bs v} \leq \norm{\bs u} + \norm{\bs v}\)
  3. \(\norm{\bs u} = 0\) iff \(\bs u = \bs 0\)

A vector space equipped with a norm is called a normed vector space. To be more precise, we denote a norm of vector space \(\mc{V}\) as \(\norm{\cdot}_{\mc{V}}\) and similarly for norms of other vector spaces. Every inner product induces a norm. The induced norm is given by \(\norm{\bs u} = \sqrt{\inpd{\bs u}{\bs u}}\).

1.4 Euclidean Space

Definition 7 An Euclidean vector space \(\b{E}\) is a finite dimensional real inner product space.

We denote an \(n\) dimensional Euclidean space by \(\b{E}^{n}\). We denote the Euclidean inner product between two vectors \(\bs{a}, \bs{b}\) by \(\bs{a} \cdot \bs{b}\).

Definition 8 An affine point space \(\mc{E}\) is a set of points on which the addition of two points is not defined, there is no unique zero element, and the difference between two points belongs to the associated Euclidean vector space.

We denote an \(n\) dimensional affine point space by \(\mc{E}^{n}\). The term point is reserved for elements of \(\mc{E}\) and the term vector for elements of \(\b{E}\).

Definition 9 Euclidean space

We denote the pair \(\pr{\mc{E}^{n}, \b{E}^{n}}\) as an Euclidean space, to mean an affine point space such that the associated vector space is an Euclidean vector space. Choosing a point \(o \in \mc{E}^{n}\) which we call the origin in \(\mc{E}^{n}\), a point \(x \in \mc{E}^{n}\) corresponds to a vector \(\bs{x} \in \b{E}^{n}\) iff \(o + \bs{x} = x\).

1.5 Components of Vectors

To manipulate and work with vectors, we would like to assign them components. Let \(\mc{F}= (1,\dots,n)\). Let a list of vectors \(\bs{e}_{i}, \dots, \bs{e}_{n} \in \b{E}^{n}\) form a basis for \(\b{E}^{n}\). Then for a vector \(\bs{x} \in \b{E}^{n}\), we can write \(\bs{x} = \sum_{i \in \mc{F}} x_{i} \bs{e}_{i}\). The components of \(\bs{x}\) w.r.t. to the basis \(\bs{e}_{i}, \dots, \bs{e}_{n}\) are the real numbers \(x_{i},\dots,x_{n}\). We give the components a column matrix representation. For example, the components of a vector \(\bs{x} \in \b{E}^{3}\) w.r.t a basis \(\bs{e}_{1},\bs{e}_{2},\bs{e}_{3}\) may be written as \[ \begin{bmatrix} a \\ b \\ c \end{bmatrix} \]

where \(a, b, c \in \b{R}\). The following is not a meaningful statement: \[ \bs{x} = \begin{bmatrix} a \\ b \\ c \end{bmatrix} \]

Instead, we write

\[ \bs{x} = \sum_{i=1}^{3} x_{i} \bs{e}_{i} \]

The column matrix representation of the components is only meaningful after an appropriate basis has been chosen, not otherwise.