Linear-Algebra-Done-Right Β· Unit 5 Β· Video 4 Β· Interactive Practice

The Fundamental Theorem of Linear Maps and Linear Systems

IKey Formulas

FormulaNameHypothesis
dim⁑V=dim⁑null T+dim⁑range T\dim V = \dim \text{null}\, T + \dim \text{range}\, TFundamental theorem of linear maps (rank-nullity)VV finite-dimensional, T∈L(V,W)T \in \mathcal{L}(V, W)
dim⁑null Tβ‰₯dim⁑Vβˆ’dim⁑W>0\dim \text{null}\, T \geq \dim V - \dim W > 0Map to a smaller space is not injectivedim⁑V>dim⁑W\dim V > \dim W
dim⁑range T≀dim⁑V<dim⁑W\dim \text{range}\, T \leq \dim V < \dim WMap to a bigger space is not surjectivedim⁑V<dim⁑W\dim V < \dim W
T(x1,…,xn)=(βˆ‘k=1nA1,kxk,…,βˆ‘k=1nAm,kxk)T(x_1, \dots, x_n) = \left(\sum_{k=1}^{n} A_{1,k} x_k, \dots, \sum_{k=1}^{n} A_{m,k} x_k\right)System of mm equations in nn variables as T ⁣:Fnβ†’FmT\colon \mathbf{F}^n \to \mathbf{F}^mn>mn > m: nonzero homogeneous solution; n<mn < m: some constants unsolvable

Key Insight: Extend a basis u1,…,umu_1, \dots, u_m of null T\text{null}\, T to a basis u1,…,um,v1,…,vnu_1, \dots, u_m, v_1, \dots, v_n of VV; then Tv1,…,TvnTv_1, \dots, Tv_n is a basis of range T\text{range}\, T. Only the counts matter, never the coefficients, and dim⁑W\dim W never enters the formula.

IIBoth Sides of the Count

For every choice of coefficients Aj,kA_{j,k}, dim⁑null T+dim⁑range T\dim \text{null}\, T + \dim \text{range}\, T equals the number of variables n=dim⁑Fnn = \dim \mathbf{F}^n.

πŸ’‘ Challenge: with m=3m = 3 and n=4n = 4, try to make TT injective. The theorem says no entries can do it, because dim⁑null Tβ‰₯4βˆ’3=1\dim \text{null}\, T \geq 4 - 3 = 1.

IIIThe Proof in One Picture

A basis of null T\text{null}\, T goes to 00; the vectors added to reach a basis of VV go to a basis of range T\text{range}\, T.

Step 1 Β· Basis of null T for T(x1,…,x5)=(x1+x2,Β x3,Β x4+x5,Β 0)T(x_1, \dots, x_5) = (x_1 + x_2,\ x_3,\ x_4 + x_5,\ 0)

Tx=0β€…β€ŠβŸΊβ€…β€Šx2=βˆ’x1,Β x3=0,Β x5=βˆ’x4Tx = 0 \iff x_2 = -x_1,\ x_3 = 0,\ x_5 = -x_4 u1=(1,βˆ’1,0,0,0)u_1 = (1, -1, 0, 0, 0) u2=(0,0,0,1,βˆ’1)u_2 = (0, 0, 0, 1, -1)

dim⁑null T=2\dim \text{null}\, T = 2

IVCounting Equations Against Variables

Fewer equations than variables forces a nonzero homogeneous solution; more equations than variables leaves some constants unreachable.

πŸ’‘ With two equations in one variable, c=(0,0)c = (0, 0) is always solvable by x=0x = 0: the theorem promises only that some choice of constants fails.

VQuiz Questions

Problem 1 Β· Apply the Fundamental Theorem

Given: T∈L(F7,F4)T \in \mathcal{L}(\mathbf{F}^7, \mathbf{F}^4) with dim⁑null T=3\dim \text{null}\, T = 3. Find dim⁑range T\dim \text{range}\, T.

βœ… Correct! 7=3+dim⁑range T7 = 3 + \dim \text{range}\, T gives 44; and a 44-dimensional subspace of F4\mathbf{F}^4 is all of F4\mathbf{F}^4, so TT is surjective.
❌ Not quite. You computed dim⁑Wβˆ’dim⁑null T\dim W - \dim \text{null}\, T. The theorem is about the domain: dim⁑V=dim⁑null T+dim⁑range T\dim V = \dim \text{null}\, T + \dim \text{range}\, T, and dim⁑W\dim W does not appear.
❌ Not quite. Adding 7+37 + 3 treats the null space as extra dimensions. The null space is part of VV: dim⁑range T=dim⁑Vβˆ’dim⁑null T\dim \text{range}\, T = \dim V - \dim \text{null}\, T. Also, range TβŠ†F4\text{range}\, T \subseteq \mathbf{F}^4 cannot have dimension 1010.
❌ Not quite. Solve dim⁑V=dim⁑null T+dim⁑range T\dim V = \dim \text{null}\, T + \dim \text{range}\, T with dim⁑V=7\dim V = 7 for the range.
Show solution

Here V=F7V = \mathbf{F}^7 is finite-dimensional, so the fundamental theorem applies:

dim⁑F7=dim⁑null T+dim⁑range T⟹7=3+dim⁑range T\dim \mathbf{F}^7 = \dim \text{null}\, T + \dim \text{range}\, T \quad\Longrightarrow\quad 7 = 3 + \dim \text{range}\, T dim⁑range T=4\dim \text{range}\, T = 4

Since range T\text{range}\, T is a subspace of F4\mathbf{F}^4 with dimension 4=dim⁑F44 = \dim \mathbf{F}^4, it equals F4\mathbf{F}^4: this TT is surjective.

Problem 2 Β· What the Dimensions Force

Given: an arbitrary T∈L(F3,F5)T \in \mathcal{L}(\mathbf{F}^3, \mathbf{F}^5). Which statement must be true?

βœ… Correct! dim⁑range T≀dim⁑F3=3<5\dim \text{range}\, T \leq \dim \mathbf{F}^3 = 3 < 5, so range Tβ‰ F5\text{range}\, T \neq \mathbf{F}^5.
❌ Not forced. A bigger target does not make a map injective: the zero map F3β†’F5\mathbf{F}^3 \to \mathbf{F}^5 has null T=F3\text{null}\, T = \mathbf{F}^3 and dim⁑range T=0\dim \text{range}\, T = 0. Neither theorem has a converse.
❌ Not forced. The bound dim⁑null Tβ‰₯dim⁑Vβˆ’dim⁑W\dim \text{null}\, T \geq \dim V - \dim W here reads dim⁑null Tβ‰₯3βˆ’5=βˆ’2\dim \text{null}\, T \geq 3 - 5 = -2, which says nothing. The inclusion (x1,x2,x3)↦(x1,x2,x3,0,0)(x_1, x_2, x_3) \mapsto (x_1, x_2, x_3, 0, 0) has null T={0}\text{null}\, T = \{0\}.
❌ Not quite. Compare dim⁑V=3\dim V = 3 with dim⁑W=5\dim W = 5: a map to a bigger space can never reach all of it.
Show solution

By the fundamental theorem and dim⁑null Tβ‰₯0\dim \text{null}\, T \geq 0:

dim⁑range T=3βˆ’dim⁑null T≀3<5=dim⁑F5\dim \text{range}\, T = 3 - \dim \text{null}\, T \leq 3 < 5 = \dim \mathbf{F}^5

Equal spaces have equal dimensions, so range Tβ‰ F5\text{range}\, T \neq \mathbf{F}^5 and TT is not surjective, whatever TT is.

The other statements hold for some maps and fail for others: the zero map is not injective and has dim⁑range T=0β‰ 3\dim \text{range}\, T = 0 \neq 3; the inclusion (x1,x2,x3)↦(x1,x2,x3,0,0)(x_1, x_2, x_3) \mapsto (x_1, x_2, x_3, 0, 0) has dim⁑null T=0<2\dim \text{null}\, T = 0 < 2.

Problem 3 Β· Size of a Solution Space

Given: a homogeneous system of 44 linear equations in 66 variables, written as T ⁣:F6β†’F4T\colon \mathbf{F}^6 \to \mathbf{F}^4 so that its solution set is null T\text{null}\, T.

Smallest possible dimension of the solution set?

If dim⁑range T=3\dim \text{range}\, T = 3, what is the dimension of the solution set?

βœ… Correct! dim⁑null T=6βˆ’dim⁑range Tβ‰₯6βˆ’4=2\dim \text{null}\, T = 6 - \dim \text{range}\, T \geq 6 - 4 = 2, and with dim⁑range T=3\dim \text{range}\, T = 3 it is exactly 6βˆ’3=36 - 3 = 3.
❌ Not quite. The solution set cannot be just {0}\{0\}: with more variables than equations, dim⁑null Tβ‰₯6βˆ’4=2>0\dim \text{null}\, T \geq 6 - 4 = 2 > 0.
❌ Check the lower bound. dim⁑range T≀dim⁑F4=4\dim \text{range}\, T \leq \dim \mathbf{F}^4 = 4, so dim⁑null T=6βˆ’dim⁑range Tβ‰₯6βˆ’4\dim \text{null}\, T = 6 - \dim \text{range}\, T \geq 6 - 4.
❌ Not quite. 4βˆ’34 - 3 subtracts from the number of equations. The theorem subtracts from the dimension of the domain: dim⁑null T=6βˆ’3\dim \text{null}\, T = 6 - 3.
❌ Check the exact count. Use dim⁑F6=dim⁑null T+dim⁑range T\dim \mathbf{F}^6 = \dim \text{null}\, T + \dim \text{range}\, T with dim⁑range T=3\dim \text{range}\, T = 3.
Show solution

Lower bound. range T\text{range}\, T is a subspace of F4\mathbf{F}^4, so dim⁑range T≀4\dim \text{range}\, T \leq 4. The fundamental theorem gives

dim⁑null T=6βˆ’dim⁑range Tβ‰₯6βˆ’4=2.\dim \text{null}\, T = 6 - \dim \text{range}\, T \geq 6 - 4 = 2.

The bound is attained, for example by T(x1,…,x6)=(x1,x2,x3,x4)T(x_1, \dots, x_6) = (x_1, x_2, x_3, x_4), whose null space is spanned by e5,e6e_5, e_6.

Exact count. If dim⁑range T=3\dim \text{range}\, T = 3, then

dim⁑null T=6βˆ’3=3.\dim \text{null}\, T = 6 - 3 = 3.

The solution set is a 33-dimensional subspace of F6\mathbf{F}^6; and since range Tβ‰ F4\text{range}\, T \neq \mathbf{F}^4, some choice of constants c1,…,c4c_1, \dots, c_4 makes the non-homogeneous system unsolvable as well.

Problem 4 · Beyond Fⁿ: Polynomials

Given: T∈L(P4(R),R3)T \in \mathcal{L}(\mathcal{P}_4(\mathbf{R}), \mathbf{R}^3) defined by Tp=(p(1),p(2),p(3))Tp = \big(p(1), p(2), p(3)\big), and TT is surjective. Find dim⁑null T\dim \text{null}\, T.

βœ… Correct! dim⁑P4(R)=5\dim \mathcal{P}_4(\mathbf{R}) = 5 and dim⁑range T=3\dim \text{range}\, T = 3, so dim⁑null T=5βˆ’3=2\dim \text{null}\, T = 5 - 3 = 2.
❌ Close, but… P4(R)\mathcal{P}_4(\mathbf{R}) has basis 1,x,x2,x3,x41, x, x^2, x^3, x^4, so its dimension is 55, not 44.
❌ Not quite. A map from a 55-dimensional space to R3\mathbf{R}^3 is never injective: dim⁑null Tβ‰₯5βˆ’3=2\dim \text{null}\, T \geq 5 - 3 = 2.
❌ Not quite. Surjective means range T=R3\text{range}\, T = \mathbf{R}^3; then subtract its dimension from dim⁑P4(R)\dim \mathcal{P}_4(\mathbf{R}).
Show solution

P4(R)\mathcal{P}_4(\mathbf{R}) has basis 1,x,x2,x3,x41, x, x^2, x^3, x^4, so it is finite-dimensional with dimension 55. Surjectivity gives range T=R3\text{range}\, T = \mathbf{R}^3, so dim⁑range T=3\dim \text{range}\, T = 3. The fundamental theorem:

5=dim⁑null T+3⟹dim⁑null T=2.5 = \dim \text{null}\, T + 3 \quad\Longrightarrow\quad \dim \text{null}\, T = 2.

Check directly. p∈null Tp \in \text{null}\, T means p(1)=p(2)=p(3)=0p(1) = p(2) = p(3) = 0, so

p(x)=(xβˆ’1)(xβˆ’2)(xβˆ’3)(a+bx),a,b∈R,p(x) = (x - 1)(x - 2)(x - 3)(a + bx), \qquad a, b \in \mathbf{R},

and the two polynomials (xβˆ’1)(xβˆ’2)(xβˆ’3)(x-1)(x-2)(x-3) and x(xβˆ’1)(xβˆ’2)(xβˆ’3)x(x-1)(x-2)(x-3) form a basis of null T\text{null}\, T: dimension 22.

Solved: 0 / 4