Fundamental Theorem of Algebra

Lecture notes from the Reading Classics

(Euler) Working Group, Autumn 2003
Lecture by Dan File∗
Notes by Steven Miller† with Dan File

Abstract
These notes are based on a lecture given by Dan File Lecturer on Wednes-
day, November 12, 2003, and were LaTeX-ed in real time by Steven J. Miller
and additions were made by Dan File.

1 History of the Fundamental Theorem of Algebra

D’Alembert (1746) observed that if p(x) is a polynomial with real coefficients
and z is a solution, then z is also a solution. His intent was to integrate rational
functions by means of partial fractions. His observation permitted him to separate
any real polynomial into linear and quadratic terms and hence find antiderivatives.
Euler became interested in this problem: Euler worked on the quartic and
quintic. For the quartic, Euler showed that there was an x-intercept. He was
relying on the fact that if you have roots βi (i ∈ {1, 2, 3, 4}), then −(β1 · · · β4 )2
is negative. This is fine if the βs are real or in complex conjugate pairs, but had
some trouble with the quntic.
Nicolas Bernouli claimed that x4 − 4x3 + 2x2 + 4x + 4 is irreducible over R,
but Euler factored it.
∗
E-mail:[email protected]
†
E-mail:[email protected]

1
 r
√ √ √
q q
4 3 2 2
x − 4x + 2x + 4x + 4 = (x − 2± 4 + 2 7)x + (1 ± 4 + 2 7 + 7)),
(1)
where above the two factors come from taking the + sign each time, or the −
sign each time. Note factoring a quartic into two real quadratics is different than
trying to find four complex roots.
Definition: A function f is analytic on an open subset R ⊂ C if f is complex
differentiable everywhere on R; f is entire if it is analytic on all of C.

2 Proof of the Fundamental Theorem via Liouville

Theorem 2.1 (Liouville). If f (z) is analytic and bounded in the complex plane,
then f (z) is constant.

We now prove

Theorem 2.2 (Fundamental Theorem of Algebra). Let p(z) be a polynomial

with complex coefficients of degree n. Then p(z) has n roots.

Proof. It is sufficient to show any p(z) has one root, for by division we can then
write p(z) = (z − z0 )g(z), with g of lower degree.
Note that if

p(z) = an z n + an−1 z n−1 + · · · + a0 , (2)

then as |z| → ∞, |p(z)| → ∞. This follows as
 an−1 a0 
p(z) = z n · an + + ··· + n. (3)

z z
1
Assume p(z) is non-zero everywhere. Then p(z) is bounded when |z| ≥ R.
1 1
Also, p(z) 6= 0, so p(z) is bounded for |z| ≤ R by continuity. Thus, p(z) is
a bounded, entire function, which must be constant. Thus, p(z) is constant, a
contradiction which implies p(z) must have a zero (our assumption).
[Lev]

2
3 Proof of the Fundamental Theorem via Rouche
Theorem 3.1 (Rouche). If f and h are each analytic functions inside and on a
domain C with bounding curve ∂C, and |h(z)| < |f (z)| on ∂C, then f and f + h
have the same number of zeros in C.
We now prove the Fundamental Theorem of Algebra:
Proof. Let

p(z) = an z n + an−1 z n−1 + · · · + a0

f (z) = an z n
h(z) = p(z) − f (z) = an−1 z n−1 + · · · + a0 . (4)

On the circle |z| = R, |f (z)| = |an |Rn , and

|h(z)| ≤ |an−1 |Rn−1 + ... + |a1 |R + |a0 |.

Make R large enough so that

|an−1 | + ... + |a1 | + |a0 |
< R.
|an |

Then |h(z)| < |f (z)| holds on the boundary of the circle centered at the origin of
radius R. Since f clearly has n zeros, we are done.
[Saff]

4 Proof of the Fundamental Theorem via Picard’s

Theorem
Theorem 4.1. If there are two distinct points that are not in the image of an entire
function p(z) (ie, ∃z1 6= z2 such that for all z ∈ C, p(z) 6= z1 or z2 ), then p(z) is
constant.
We now prove the Fundamental Theorem of Algebra:
Let p(z) be a non-constant polynomial, and assume p(z) is never 0.
Claim 4.2. If p(z) is as above, p(z) does not take on one of the values k1 for k ∈ N.

3
Proof. Assume not. Thus, ∃zk ∈ C such that p(zk ) = k1 . If we take a circle
D centered at the origin with sufficiently large radius, then |p(z)| > 1 for all z
outside D. Thus, each zi ∈ D. By Bolzano-Weierstrasss, as all the points zk ∈ D,
we have a convergent subsequence. Thus, we have zni → z 0 . But

p(z 0 ) = lim p(zni ) = 0. (5)

ni →∞

Thus, there must be some k such that p(z) 6= k1 . Since there are two distinct
values not in the image of p, by Picard’s Theorem it is now constant. This con-
tradicts our assumption that p(z) is non-constant. Therefore, p(z0 ) = 0 for some
z0 .

Remark 4.3. One can use a finite or countable version of Picard. Rather than
missing just two points, we can modify the above to work if Picard instead stated
that if we miss finitely many (or even countably many) points, we are constant.
Just look at the method above, gives k11 . We can then find another larger one,
say k12 . And so on. We can even get uncountably many such points by looking at
numbers such as πk (using now the transcendence of C is 1).

[Boas, 1935]

5 Proof of the Fundamental Theorem via Cauchy’s

Integral Theorem
Theorem 5.1 (Cauchy Integral Theorem). Let f (z) be analytic inside on on the
boundary of some region C. Then
Z
f (z)dz = 0. (6)
∂C

We now prove the Fundamental Theorem of Algebra:

Proof. Without loss of generality let p(z) be a non-constant polynomial and as-
sume p(z) = 0. For z ∈ R, assume p(z) ∈ R. (Otherwise, consider p(z)p(z)).
Therefore, p(z) doesn’t change signs for z ∈ R, or by the Intermediate Value
Theorem it would have a zero.
Z 2π
dθ
6= 0. (7)
0 p(2 cos θ)

4
 This follows from our assumption that p(z) is of constant sign for real argu-
ments, bounded above from 0. This integral equals the contour integral

z n−1
Z Z
1 dz 1
= , (8)
i |z|=1 zp(z + z −1 ) i |z|=1 Q(z)
where

Q(z) = z n p(z + z −1 ). (9)

If z 6= 0, Q(z) 6= 0.
If z = 0, then Q(z) 6= 0 since

p(z + z −1 ) = an (z + z −1 )n + · · ·
z n p(z + z −1 ) z n · · · an z −n + · · ·


=
= an + z(· · · ). (10)

Thus, Q(z) = an , which is non-zero. Hence, Q(z)

z n−1 1
R 6= z0, and consequently
n−1
Qz
is analytic. By the Cauchy Integral Formula i |z|=1 Q(z) 6= 0. Thus, a
contradiction!

[Boas 1964]

6 Proof of the Fundamental Theorem via Maximum

Modulus Principle
Theorem 6.1 (Maximum (Minimum) Modulus Principle). No entire function
attains its maximum in the interior.

We now prove the Fundamental Theorem of Algebra:

Proof. Assume p(z) is non-constant and never zero. ∃M such that |p(z)| ≥ |a0 | = 6
0 if |z| > M . Since |p(z)| is continuous, it achieves its minimum on a closed in-
terval. Let z0 be the value in the circle of radius M where p(z) takes its minimum
value.

So |p(z0 )| ≤ |p(z)| for all z ∈ C, and in particular |p(z0 )| ≤ |p(0)| = |a0 |.

5
 Translate the polynomial. Let p(z) = p((z − z0 ) + z0 ); let p(z) = Q(z − z0 ).
Note the minimum of Q occurs at z = 0: |Q(0)| ≤ |Q(z)| for all z ∈ C.

Q(z) = c0 + cj z j + · · · + cn z n , (11)
where j is such that cj is the first coefficient (after c0 ) that is non-zero. I must
show Q(0) = 0. Note if c0 = 0, we are done.
We may rewrite such that

Q(z) = c0 + cj z j + z j+1 R(z). (12)

We will extract roots. Let
c0
reiθ = − . (13)
cj
Further, let
1 iθ
z1 = r j e j . (14)
So,
cj z j = −c0 . (15)
Let  > 0 be a small real number. Then

Q(z1 ) = c0 + cj j z1j + j+1 z1j+1 R(z1 )

|Q(z1 )| ≤ |c0 + cj j zjj | + j+1 |z1 |j+1 |R(z1 )|
|c0 | − j |c0 | + j+1 |z1 |j+1 N, (16)

where N is chosen such that N > |R(z1 )|, and  is chosen so that

j+1 |z1 |j+1 < j |c0 |.

Thus,

|Q(z1 )| < |c0 |, (17)

but this was supposed to be our minimum. Thus, a contradiction!
[Fefferman]

6
7 Proof of the Fundamental Theorem via Radius of
Convergence
We now prove the Fundamental Theorem of Algebra: As always, p(z) is a non-
constant polynomial. Consider
1
f (z) = = b0 + b1 z + · · · , (18)
p(z)
and

p(z) = an z n + · · · + a0 , a0 6= 0. (19)
Lemma 7.1. ∃c, r ∈ C such that |bk | > crk for infinitely many k.
Now, 1 = p(z)f (z). Thus, a0 b0 = 1. This is our basis step. Assume we
have some coefficient such that |bk | > crk . We claim we can always find another.
Suppose there are no more. Then the coefficient of z k+n in p(z)f (z) is

a0 bk+n + a1 bk+n−1 + · · · + an bk = 0. (20)

Thus, as we have |bj | > crj in this range, we have the coefficient satisfies

|a0 |rn + |a1 |rn−1 + · · · + |an−1 |r ≤ |an | (21)

if

|an |
r ≤ min{1, . (22)
|a0 | + · · · + |an−1 |
This will give that

|a0 bk+n + · · · + an−1 bk+1 |

|bk | =
|an |
|a0 bk+n | + · · · + |an−1 bk+1 |
≤ ≤ crk (23)
|an |
for sufficiently small.
Let z = 1r . Then

bk z  = |bk | > c.
 k
(24)
rk
7
 This is true for infinitely many k, hence the power series diverges, contradict-
ing the assumption that the function is analytic and its power series converges
everywhere.
[Velleman]

References
[1] Open Mappings and the Fundamental Theorem of Algebra
R. L. Thompson
Mathematics Magazine, Vol. 43, No. 1. (Jan., 1970), pp. 39-40.

[2] What! Another Note Just on the Fundamental Theorem of Algebra?

R. M. Redheffer
The American Mathematical Monthly, Vol. 71, No. 2. (Feb., 1964), pp.
180-185.

[3] Another Proof of the Fundamental Theorem of Algebra

Daniel J. Velleman
Mathematics Magazine, Vol. 70, No. 3. (Jun., 1997), pp. 216-217.

[4] Euler and the Fundamental Theorem of Algebra

William Dunham
The College Mathematics Journal, Vol. 22, No. 4. (Sep., 1991), pp. 282-293.

[5] An Elementary Constructive Proof of the Fundamental Theorem of Algebra

P. C. Rosenbloom
The American Mathematical Monthly, Vol. 52, No. 10. (Dec., 1945), pp.
562-570.

[6] Proof of the Fundamental Theorem of Algebra

J. L. Brenner; R. C. Lyndon
The American Mathematical Monthly, Vol. 88, No. 4. (Apr., 1981), pp.
253-256.

8
[7] Yet Another Proof of the Fundamental Theorem of Algebra
R. P. Boas, Jr.
The American Mathematical Monthly, Vol. 71, No. 2. (Feb., 1964), p. 180.

[8] The Fundamental Theorem of Algebra

Raymond Redheffer
The American Mathematical Monthly, Vol. 64, No. 8. (Oct., 1957), pp.
582-585.

[9] A Proof of the Fundamental Theorem of Algebra

R. P. Boas, Jr.
The American Mathematical Monthly, Vol. 42, No. 8. (Oct., 1935), pp.
501-502.

[10] An Easy Proof of the Fundamental Theorem of Algebra

Charles Fefferman
The American Mathematical Monthly, Vol. 74, No. 7. (Aug. - Sep., 1967),
pp. 854-855.

[11] Euler: The Master of Us All

William Dunham
The Mathematica Association of America
Dolciani Mathematical Expositions No. 22, 1999.