Linear-Algebra-Done-Right · Unit 10 · Video 3 · Interactive Practice

Fundamental Theorem of Algebra and Complete Factorization

IKey Formulas

FormulaNameWhat it says
p∈P(C)p \in \mathcal{P}(\mathbf{C}) nonconstant ⇒p(ζ)=0\Rightarrow p(\zeta) = 0 for some ζ∈C\zeta \in \mathbf{C}Fundamental theorem of algebra, first versionA zero exists; the proof gives no method for finding it
p(z)=c(z−λ1)⋯(z−λm)p(z) = c(z - \lambda_1)\cdots(z - \lambda_m)Fundamental theorem of algebra, second versionFor nonconstant p∈P(C)p \in \mathcal{P}(\mathbf{C}) of degree mm, with c,λ1,…,λm∈Cc, \lambda_1, \dots, \lambda_m \in \mathbf{C}; unique up to the order of the factors; cc is the coefficient of zmz^m and the zeros of pp are exactly λ1,…,λm\lambda_1, \dots, \lambda_m
(r1/k(cos⁡θk+isin⁡θk))k=r(cos⁡θ+isin⁡θ)\left(r^{1/k}\left(\cos\tfrac{\theta}{k} + i\sin\tfrac{\theta}{k}\right)\right)^k = r(\cos\theta + i\sin\theta)kk-th roots from De MoivreFor a positive integer kk, every w=r(cos⁡θ+isin⁡θ)∈Cw = r(\cos\theta + i\sin\theta) \in \mathbf{C} (r≥0r \ge 0) has a kk-th root: take the kk-th root of the length and divide the angle by kk
βk=−1ak\beta^k = -\dfrac{1}{a_k},  ∣q(tβ)∣≤1−tk(1−tB)\ |q(t\beta)| \le 1 - t^k(1 - tB)The descent stepFor q(z)=1+akzk+⋯+amzmq(z) = 1 + a_k z^k + \cdots + a_m z^m with ak≠0a_k \neq 0 and 0<t<10 < t < 1, where B>1B > 1 is a constant with ∣ak+1(tβ)k+1+⋯+am(tβ)m∣≤B tk+1|a_{k+1}(t\beta)^{k+1} + \cdots + a_m(t\beta)^m| \le B\,t^{k+1}; at t=12Bt = \tfrac{1}{2B} the bound is 1−12tk<11 - \tfrac12 t^k < 1

Key Insight: The theorem is pure existence. x5−5x4−6x3+17x2+4x−7x^5 - 5x^4 - 6x^3 + 17x^2 + 4x - 7 has five zeros, approximately −1.87-1.87, −0.74-0.74, 0.620.62, 1.471.47, 5.515.51, yet no formula built from the coefficients will ever give them exactly.

IIEvery Complex Number Has a kk-th Root

If w=r(cos⁡θ+isin⁡θ)w = r(\cos\theta + i\sin\theta), the number of length r1/kr^{1/k} at angle θ/k\theta/k raised to the kk-th power is ww.

IIIThe Descent Step: ∣q∣|q| Drops Below 11

For q(z)=1+z2+z3q(z) = 1 + z^2 + z^3, moving along tβt\beta with β2=−1a2=−1\beta^2 = -\tfrac{1}{a_2} = -1 pushes ∣q∣|q| below 11.

💡 The proof's qq has minimum value 11 at 00; this step shows that no polynomial with q(0)=1q(0) = 1 and some ak≠0a_k \neq 0 can, so the assumption p(ζ)≠0p(\zeta) \neq 0 fails.

IVThe Zeros Are Exactly the λj\lambda_j

Shading darkens as ∣p(z)∣|p(z)| shrinks; it reaches 00 only at the λj\lambda_j, and each repeated λj\lambda_j still adds one to deg⁡p\deg p.

VQuiz Questions

Problem 1 · Factor z4−16z^4 - 16 Completely

Given: p(z)=z4−16∈P(C)p(z) = z^4 - 16 \in \mathcal{P}(\mathbf{C}) — find its factorization c(z−λ1)⋯(z−λ4)c(z - \lambda_1)\cdots(z - \lambda_4) from the second version of the fundamental theorem of algebra.

✅ Correct! c=1c = 1 and the four zeros are 22, −2-2, 2i2i, −2i-2i: each satisfies z4=16z^4 = 16, and a degree-4 polynomial has four factors.
❌ True, but not complete. The product does equal z4−16z^4 - 16, but the second version demands degree-one factors, and over C\mathbf{C} the factor z2+4z^2 + 4 splits as (z−2i)(z+2i)(z - 2i)(z + 2i).
❌ Not quite. (z−2)2(z+2)2=(z2−4)2=z4−8z2+16(z - 2)^2(z + 2)^2 = (z^2 - 4)^2 = z^4 - 8z^2 + 16. Also (2i)4=16(2i)^4 = 16, so 2i2i must be a zero, and it is missing here.
❌ Check the imaginary zeros. (z−4i)(z+4i)=z2+16(z - 4i)(z + 4i) = z^2 + 16, so this product is (z2−4)(z2+16)=z4+12z2−64(z^2 - 4)(z^2 + 16) = z^4 + 12z^2 - 64. The zeros of z2+4z^2 + 4 are ±2i\pm 2i.
Show solution

Difference of squares twice:

z4−16=(z2−4)(z2+4)=(z−2)(z+2)(z2+4).z^4 - 16 = (z^2 - 4)(z^2 + 4) = (z - 2)(z + 2)(z^2 + 4).

Over C\mathbf{C}, z2+4=z2−(2i)2=(z−2i)(z+2i)z^2 + 4 = z^2 - (2i)^2 = (z - 2i)(z + 2i), so

z4−16=(z−2)(z+2)(z−2i)(z+2i).z^4 - 16 = (z - 2)(z + 2)(z - 2i)(z + 2i).

The coefficient of z4z^4 is 11, so c=1c = 1, and by uniqueness this is the only factorization apart from the order of the factors. The zeros of pp are exactly 22, −2-2, 2i2i, −2i-2i.

Problem 2 · Counting Zeros with Repeats

Given: p(z)=2(z−i)(z+1)2(z−i)(z−3)p(z) = 2(z - i)(z + 1)^2(z - i)(z - 3) — find deg⁡p\deg p and the number of distinct zeros of pp.

What is deg p?

How many distinct zeros does p have?

✅ Correct! In the form c(z−λ1)⋯(z−λ5)c(z - \lambda_1)\cdots(z - \lambda_5) the list is i,i,−1,−1,3i, i, -1, -1, 3: five factors, three distinct zeros.
❌ The constant is not a factor of degree one. c=2c = 2 contributes nothing to the degree; only the factors z−λjz - \lambda_j count.
❌ That counts distinct zeros. The degree counts every factor z−λjz - \lambda_j, repeats included: z−iz - i appears twice and z+1z + 1 appears twice.
❌ Count the square. (z+1)2=(z+1)(z+1)(z + 1)^2 = (z + 1)(z + 1) is two factors, so the factors are z−iz - i, z+1z + 1, z+1z + 1, z−iz - i, z−3z - 3.
❌ Not quite. Multiplying out, the highest power of zz is z1+2+1+1z^{1 + 2 + 1 + 1}.
❌ That counts with repeats. Five is deg⁡p\deg p; the values ii and −1-1 each occur twice in the list λ1,…,λ5\lambda_1, \dots, \lambda_5.
❌ The two factors z−iz - i give the same zero. Separated or not, both vanish at z=iz = i; the distinct zeros are ii, −1-1, 33.
❌ A zero is missing. pp is zero exactly when some factor is zero: at ii, at −1-1, and at 33.
❌ Not quite. The zeros of pp are exactly the numbers λj\lambda_j in the factorization; list them without repeats.
Show solution

Collect equal factors:

p(z)=2(z−i)2(z+1)2(z−3)=2(z−i)(z−i)(z+1)(z+1)(z−3).p(z) = 2(z - i)^2(z + 1)^2(z - 3) = 2(z - i)(z - i)(z + 1)(z + 1)(z - 3).

This is the form c(z−λ1)⋯(z−λm)c(z - \lambda_1)\cdots(z - \lambda_m) with c=2c = 2 and the list λ1,…,λ5=i,i,−1,−1,3\lambda_1, \dots, \lambda_5 = i, i, -1, -1, 3. The degree is the number of degree-one factors, m=5m = 5; the constant cc adds nothing.

Since c≠0c \neq 0, p(z)=0p(z) = 0 exactly when one factor is zero, so the zeros of pp are ii, −1-1, 33: three distinct zeros, with ii and −1-1 each appearing twice in the list.

Problem 3 · A Quadratic with a Complex Coefficient

Given: p(z)=z2+2iz+3p(z) = z^2 + 2iz + 3 — find its factorization into degree-one factors.

✅ Correct! (z−i)(z+3i)=z2+3iz−iz−3i2=z2+2iz+3(z - i)(z + 3i) = z^2 + 3iz - iz - 3i^2 = z^2 + 2iz + 3; the zeros are ii and −3i-3i.
❌ The signs are swapped. (z+i)(z−3i)=z2−2iz+3(z + i)(z - 3i) = z^2 - 2iz + 3. Its zeros −i-i, 3i3i are the negatives of the zeros of pp.
❌ Check (2i)2(2i)^2. (2i)2=4i2=−4(2i)^2 = 4i^2 = -4, not 44. With 44 the discriminant becomes 4−12=−84 - 12 = -8, and this product expands to z2+2iz+1z^2 + 2iz + 1, not pp.
❌ Not quite. (z−i)(z−3i)=z2−4iz+3i2=z2−4iz−3(z - i)(z - 3i) = z^2 - 4iz + 3i^2 = z^2 - 4iz - 3. The two zeros must add to −2i-2i and multiply to 33.
Show solution

Complete the square, using i2=−1i^2 = -1:

z2+2iz+3=(z+i)2−i2+3=(z+i)2+4.z^2 + 2iz + 3 = (z + i)^2 - i^2 + 3 = (z + i)^2 + 4.

So p(z)=0p(z) = 0 exactly when (z+i)2=−4=(2i)2(z + i)^2 = -4 = (2i)^2, that is z+i=±2iz + i = \pm 2i, giving z=iz = i or z=−3iz = -3i. (The quadratic formula gives the same: the discriminant is (2i)2−4⋅3=−16(2i)^2 - 4 \cdot 3 = -16, with square roots ±4i\pm 4i, and z=−2i±4i2z = \frac{-2i \pm 4i}{2}.)

p(z)=(z−i)(z+3i).p(z) = (z - i)(z + 3i).

Check: (z−i)(z+3i)=z2+3iz−iz−3i2=z2+2iz+3(z - i)(z + 3i) = z^2 + 3iz - iz - 3i^2 = z^2 + 2iz + 3 ✓

Problem 4 · Cube Roots of −8-8 by De Moivre

Given: p(z)=z3+8p(z) = z^3 + 8, whose zeros are the cube roots of −8-8 — find all of them, writing −8=8(cos⁡θ+isin⁡θ)-8 = 8(\cos\theta + i\sin\theta).

✅ Correct! Length 81/3=28^{1/3} = 2 and angles π3\tfrac{\pi}{3}, π\pi, 5π3\tfrac{5\pi}{3}, so z3+8=(z+2)(z−1−3 i)(z−1+3 i)z^3 + 8 = (z + 2)(z - 1 - \sqrt{3}\,i)(z - 1 + \sqrt{3}\,i).
❌ Two zeros are missing. z3+8z^3 + 8 has degree 33, so it has three factors z−λjz - \lambda_j; the other two zeros are not real.
❌ Those two are cube roots of 88. −1+3 i=2(cos⁡2π3+isin⁡2π3)-1 + \sqrt{3}\,i = 2\left(\cos\tfrac{2\pi}{3} + i\sin\tfrac{2\pi}{3}\right), and its cube is 8(cos⁡2π+isin⁡2π)=88(\cos 2\pi + i\sin 2\pi) = 8. Start from the angle θ=π\theta = \pi of −8-8.
❌ Not quite. (2i)3=8i3=−8i(2i)^3 = 8i^3 = -8i, not −8-8. Divide the angle π\pi of −8-8 (and π+2π\pi + 2\pi, π+4π\pi + 4\pi) by 33.
Show solution

In polar form −8=8(cos⁡θ+isin⁡θ)-8 = 8(\cos\theta + i\sin\theta) with θ=π\theta = \pi, and equally with θ=3π\theta = 3\pi or 5π5\pi, since cosine and sine have period 2π2\pi. De Moivre gives a cube root of length 81/3=28^{1/3} = 2 and angle θ/3\theta/3 for each choice:

θ=π: 2(cos⁡π3+isin⁡π3)=1+3 i,\theta = \pi:\ 2\left(\cos\tfrac{\pi}{3} + i\sin\tfrac{\pi}{3}\right) = 1 + \sqrt{3}\,i, θ=3π: 2(cos⁡π+isin⁡π)=−2,\theta = 3\pi:\ 2(\cos\pi + i\sin\pi) = -2, θ=5π: 2(cos⁡5π3+isin⁡5π3)=1−3 i.\theta = 5\pi:\ 2\left(\cos\tfrac{5\pi}{3} + i\sin\tfrac{5\pi}{3}\right) = 1 - \sqrt{3}\,i.

These are three distinct zeros of the degree-3 polynomial z3+8z^3 + 8, so by the second version

z3+8=(z+2)(z−(1+3 i))(z−(1−3 i)),z^3 + 8 = (z + 2)\big(z - (1 + \sqrt{3}\,i)\big)\big(z - (1 - \sqrt{3}\,i)\big),

and there are no other zeros. Check: (z−(1+3 i))(z−(1−3 i))=(z−1)2+3=z2−2z+4\big(z - (1 + \sqrt{3}\,i)\big)\big(z - (1 - \sqrt{3}\,i)\big) = (z - 1)^2 + 3 = z^2 - 2z + 4, and (z+2)(z2−2z+4)=z3+8(z + 2)(z^2 - 2z + 4) = z^3 + 8 ✓

Solved: 0 / 4