Linear Algebra · Undergraduate

Abstract Vector Spaces: Polynomials, Matrices and Functions

Quick answer

A vector space is any collection that can be added and scaled by the familiar rules: arrows, lists in ℝⁿ, but also polynomials, matrices and functions. The ideas from ℝⁿ carry over unchanged. A subset is a subspace when it contains zero and is closed under addition and scaling. A basis is an independent spanning set, and its size is the dimension: polynomials of degree at most 2 form a 3-dimensional space with basis 1, t, t². Coordinates in a basis turn any finite-dimensional space into ℝⁿ.

What you'll learn

  • Recognize polynomials, matrices and functions as vector spaces
  • Apply the subspace test
  • Find a basis and the dimension of a space of polynomials or matrices
  • Write coordinates and matrices of maps such as differentiation

Vectors that are not arrows

Everything done with vectors in Rn\mathbb{R}^n used only two operations: adding and scaling. Polynomials can be added and scaled, and so can matrices and functions. Any set with these operations, obeying the usual rules, is a vector space:

SpaceA typical vectorZero vector
P2\mathbb{P}_2, polynomials of degree at most 223−2t+t23 - 2t + t^2the zero polynomial
M2×2M_{2\times 2}, 2×22 \times 2 matrices[140−2]\begin{bmatrix} 1 & 4\\ 0 & -2 \end{bmatrix}the zero matrix
functions on R\mathbb{R}sin⁡t\sin tthe function 00

Subspace test. A subset HH of a vector space is a subspace when

  1. the zero vector is in HH;
  2. u+vu + v is in HH whenever uu and vv are;
  3. cucu is in HH whenever uu is.

Span, linear independence, basis and dimension keep their meanings word for word.

Why the ideas carry over

The theorems about Rn\mathbb{R}^n never used that vectors were lists of numbers, only the rules of addition and scaling. So they hold in every vector space. Better still, a basis turns any nn-dimensional space into Rn\mathbb{R}^n: with the basis 1,t,t21, t, t^2, the polynomial 3−2t+t23 - 2t + t^2 becomes the coordinate vector (3,−2,1)(3, -2, 1). Adding polynomials adds their coordinates, and scaling scales them. Once a basis is chosen, a question about polynomials or matrices becomes a question about columns of numbers, and row reduction answers it.

Worked examples

Common mistakes

Practice problems

  1. Is the set of 2×22 \times 2 matrices with trace 00 a subspace? If so, what is its dimension?

    Answer

    Yes; dimension 33

    Full solution

    Trace is linear, so sums and multiples of trace-zero matrices have trace 00. The matrices [abc−a]\begin{bmatrix} a & b\\ c & -a \end{bmatrix} have three free entries.

  2. Is {p∈P2:p(0)=1}\{p \in \mathbb{P}_2 : p(0) = 1\} a subspace?

    Answer

    No

    Full solution

    It does not contain the zero polynomial.

  3. What is the dimension of P3\mathbb{P}_3?

    Answer

    44

    Full solution

    A basis is 1,t,t2,t31, t, t^2, t^3.

  4. Find the coordinates of 2+5t−t22 + 5t - t^2 in the basis 1,t,t21, t, t^2.

    Answer

    (2,5,−1)(2, 5, -1)

    Full solution

    Read off the coefficients in order.

  5. What is the dimension of the space of symmetric 2×22 \times 2 matrices?

    Answer

    33

    Full solution

    [abbc]\begin{bmatrix} a & b\\ b & c \end{bmatrix} has three free entries.

  6. Find the coordinates of 3+t3 + t in the basis 1+t1 + t, 1−t1 - t.

    Answer

    (2,1)(2, 1)

    Full solution

    a(1+t)+b(1−t)=3+ta(1 + t) + b(1 - t) = 3 + t gives a+b=3a + b = 3 and a−b=1a - b = 1.

  7. The solutions of y′′+y=0y'' + y = 0 form a vector space. Give a basis and its dimension.

    Answer

    cos⁡t\cos t, sin⁡t\sin t; dimension 22

    Full solution

    Every solution is c1cos⁡t+c2sin⁡tc_1\cos t + c_2\sin t, and neither function is a multiple of the other.

  8. A student says the invertible 2×22 \times 2 matrices form a subspace. Give a reason they do not.

    Answer

    The zero matrix is not invertible.

    Full solution

    Also, I+(−I)=0I + (-I) = 0 shows the set is not closed under addition.

Frequently asked questions

What is a vector space?

A set with an addition and a scalar multiplication that obey the usual rules, such as u + v = v + u, a zero vector, and c(u + v) = cu + cv. ℝⁿ, polynomials, matrices and functions all qualify.

How do you test whether a subset is a subspace?

Check three things: it contains the zero vector, the sum of two members is a member, and every scalar multiple of a member is a member.

What is the dimension of the polynomials of degree at most n?

n + 1, with basis 1, t, t², …, tⁿ.

What is the dimension of the 2 × 2 matrices?

4. The four matrices with a single 1 and three 0s form a basis.

Why are coordinates useful?

Writing each vector by its coordinates in a basis turns an n-dimensional space into ℝⁿ, so row reduction and matrices apply to polynomials and functions.

What to learn next