Chapter 1: Fundamentals of Vector Spaces
Axiomatic Definition and Concrete Examples
Goals
A "vector" is not just an arrow. Anything that satisfies the axioms is a vector. Understand that polynomials, functions, matrices, and more can all form vector spaces.
1. Why Do We Need Abstraction?
1.1 Arrows Are Not Enough
In introductory courses, we learn "vector = arrow." But consider:
- What happens when you add polynomials $p(x) = 2x^2 + 3x + 1$ and $q(x) = x^2 - x$?
- What happens when you multiply the function $f(x) = \sin x$ by 2?
- What is the sum of two matrices $A$ and $B$?
All of these support "addition" and "scalar multiplication." They share the same structure as arrows!
1.2 Extracting the Common Structure
The concept of a vector space extracts the "rules of addition and scalar multiplication" common to these different objects.
Once we verify the axioms are satisfied, every theorem proved for $\mathbb{R}^n$ applies immediately!
2. Axioms of a Vector Space
2.1 Definition
Definition: A set $V$ is a vector space over a field $\mathbb{F}$ (usually $\mathbb{R}$ or $\mathbb{C}$) if it is equipped with "addition" and "scalar multiplication" satisfying the following axioms.
2.2 Addition Axioms (4)
For all $\boldsymbol{u}, \boldsymbol{v}, \boldsymbol{w} \in V$:
- Associativity: $(\boldsymbol{u} + \boldsymbol{v}) + \boldsymbol{w} = \boldsymbol{u} + (\boldsymbol{v} + \boldsymbol{w})$
- Commutativity: $\boldsymbol{u} + \boldsymbol{v} = \boldsymbol{v} + \boldsymbol{u}$
- Zero vector: There exists $\boldsymbol{0} \in V$ such that $\boldsymbol{v} + \boldsymbol{0} = \boldsymbol{v}$
- Additive inverse: For each $\boldsymbol{v}$, there exists $-\boldsymbol{v} \in V$ such that $\boldsymbol{v} + (-\boldsymbol{v}) = \boldsymbol{0}$
2.3 Scalar Multiplication Axioms (4)
For all $\boldsymbol{u}, \boldsymbol{v} \in V$ and $a, b \in \mathbb{F}$:
- Distributivity (vector): $a(\boldsymbol{u} + \boldsymbol{v}) = a\boldsymbol{u} + a\boldsymbol{v}$
- Distributivity (scalar): $(a + b)\boldsymbol{v} = a\boldsymbol{v} + b\boldsymbol{v}$
- Associativity: $(ab)\boldsymbol{v} = a(b\boldsymbol{v})$
- Identity element: $1 \cdot \boldsymbol{v} = \boldsymbol{v}$
2.4 Meaning of the Axioms
These eight axioms abstract the "obvious computation rules in $\mathbb{R}^n$." Any set satisfying them can be analyzed using the same methods as $\mathbb{R}^n$.
3. Concrete Examples of Vector Spaces
3.1 $\mathbb{R}^n$ (Coordinate Space)
The most fundamental example. Ordered $n$-tuples of real numbers:
$$\mathbb{R}^n = \left\{ \begin{pmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{pmatrix} \,\middle|\, x_i \in \mathbb{R} \right\}$$Component-wise addition and scalar multiplication satisfy all eight axioms.
3.2 Polynomial Space $P_n$
All polynomials of degree at most $n$:
$$P_n = \{a_0 + a_1 x + a_2 x^2 + \cdots + a_n x^n \mid a_i \in \mathbb{R}\}$$Polynomial addition and scalar multiplication satisfy the axioms.
Example: For $p(x) = 2x^2 + 3x + 1$ and $q(x) = x^2 - x$:
$p + q = 3x^2 + 2x + 1$, $\quad 2p = 4x^2 + 6x + 2$
Zero vector: $0$ (the zero polynomial)
Dimension: $\dim(P_n) = n + 1$ (basis: $1, x, x^2, \ldots, x^n$)
3.3 Matrix Space $M_{m \times n}$
All $m \times n$ matrices:
$$M_{m \times n} = \{A \mid A \text{ is an } m \times n \text{ matrix}\}$$Matrix addition and scalar multiplication satisfy the axioms.
Dimension: $\dim(M_{m \times n}) = mn$
3.4 Function Space
All continuous functions on $[0, 1]$:
$$C[0, 1] = \{f: [0, 1] \to \mathbb{R} \mid f \text{ is continuous}\}$$Pointwise addition and scalar multiplication satisfy the axioms.
Zero vector: $f(x) = 0$ (the zero function)
Dimension: Infinite-dimensional! (cannot be spanned by finitely many functions)
3.5 Solution Space
The set of all solutions to a homogeneous linear system $A\boldsymbol{x} = \boldsymbol{0}$:
$$\ker(A) = \{\boldsymbol{x} \in \mathbb{R}^n \mid A\boldsymbol{x} = \boldsymbol{0}\}$$This is a subspace of $\mathbb{R}^n$.
4. Subspaces
4.1 Definition
Definition: A subset $W$ of a vector space $V$ is a subspace if:
- $\boldsymbol{0} \in W$ (contains the zero vector)
- $\boldsymbol{u}, \boldsymbol{v} \in W \Rightarrow \boldsymbol{u} + \boldsymbol{v} \in W$ (closed under addition)
- $\boldsymbol{v} \in W, a \in \mathbb{F} \Rightarrow a\boldsymbol{v} \in W$ (closed under scalar multiplication)
4.2 Examples of Subspaces
- A plane through the origin in $\mathbb{R}^3$
- A line through the origin in $\mathbb{R}^3$
- All diagonal matrices (subspace of $M_{n \times n}$)
- All even functions (subspace of $C[-1, 1]$)
4.3 Non-Examples
- A plane not through the origin (does not contain the zero vector)
- Points on the unit circle (not closed under addition)
5. Basis and Dimension
5.1 Linear Independence
Vectors $\boldsymbol{v}_1, \ldots, \boldsymbol{v}_k$ are linearly independent if:
$$c_1\boldsymbol{v}_1 + c_2\boldsymbol{v}_2 + \cdots + c_k\boldsymbol{v}_k = \boldsymbol{0} \quad \Rightarrow \quad c_1 = c_2 = \cdots = c_k = 0$$5.2 Basis
Definition: A basis of a vector space $V$ is a set of vectors $\{\boldsymbol{e}_1, \ldots, \boldsymbol{e}_n\}$ satisfying both:
- Linearly independent — none of the $\boldsymbol{e}_i$ can be built from the others (no waste)
- Spans $V$ — every vector in $V$ can be written as a linear combination of $\boldsymbol{e}_1, \ldots, \boldsymbol{e}_n$ (reaches everywhere)
Intuitively, "a minimal toolkit that covers the whole space". Too few and it fails to cover; too many and there is redundancy (linear dependence).
Examples: Basis or not?
| Set of vectors | Lin. indep.? | Spans $\mathbb{R}^2$? | Basis? |
|---|---|---|---|
| $\{(1,0), (0,1)\}$ — standard basis | Yes | Yes | Yes |
| $\{(1,0), (1,1)\}$ | Yes | Yes | Yes (non-standard bases are also bases) |
| $\{(1,0)\}$ | Yes | No (cannot reach $y$) | No (too few) |
| $\{(1,0), (2,0)\}$ | No (on the same line) | No | No |
| $\{(1,0), (0,1), (1,1)\}$ | No (third is redundant) | Yes | No (too many) |
Bases in Various Vector Spaces
- $\mathbb{R}^n$: standard basis $\{\boldsymbol{e}_1, \ldots, \boldsymbol{e}_n\}$ ($\boldsymbol{e}_i$ has a 1 in the $i$-th slot, 0 elsewhere) — $n$ vectors
- $P_n$ (polynomials of degree $\le n$): $\{1, x, x^2, \ldots, x^n\}$ — $n+1$ vectors
- $M_{m \times n}$ ($m \times n$ matrices): $\{E_{ij}\}$ (1 at position $(i, j)$, 0 elsewhere) — $mn$ vectors
- $C[0, 1]$ (continuous functions): has no finite basis (infinite-dimensional)
Note: a basis is not unique. The same space admits many different bases. For example, on $\mathbb{R}^2$, each of $\{(1,0), (0,1)\}$, $\{(1,1), (1,-1)\}$, $\{(3,1), (5,2)\}$ is a basis. Which one to use depends on the problem.
5.3 Dimension
The number of elements in a basis is called the dimension $\dim(V)$ of $V$.
Key Theorem
The number of elements in a basis of a finite-dimensional vector space is the same regardless of the choice of basis.
5.4 Concrete Examples
| Vector Space | Standard Basis | Dimension |
|---|---|---|
| $\mathbb{R}^n$ | $\boldsymbol{e}_1, \ldots, \boldsymbol{e}_n$ | $n$ |
| $P_n$ | $1, x, x^2, \ldots, x^n$ | $n + 1$ |
| $M_{m \times n}$ | $E_{ij}$ (1 in position $(i,j)$, 0 elsewhere) | $mn$ |
| $C[0,1]$ | None (infinite-dimensional) | $\infty$ |
6. Why Abstraction Is Useful
6.1 Reusing Theorems
A theorem proved once applies to every vector space:
- An $n$-dimensional space requires exactly $n$ basis vectors
- Any $n+1$ vectors must be linearly dependent
- The dimension of a subspace is at most that of the whole space
6.2 Unifying Different Fields
- Differential equations: The solution space is a vector space
- Fourier analysis: Functions treated as "vectors"
- Quantum mechanics: State vectors (wave functions) live in infinite-dimensional vector spaces
6.3 Example: Differential Equations
The solution set of $y'' + y = 0$ is a 2-dimensional vector space with basis $\sin x$ and $\cos x$.
The general solution $y = c_1 \sin x + c_2 \cos x$ is a "linear combination of the basis" — the standard form.
7. Summary
Key Takeaways
- Vector space: A set satisfying eight axioms
- Examples: $\mathbb{R}^n$, polynomials, matrices, functions
- Subspace: A subset closed under addition and scalar multiplication
- Dimension: The number of basis vectors (invariant under change of basis)
- Power of abstraction: Treat different objects with a unified theory