Polynomials

Henry Liu, 25 November 2004

[email protected]

Introduction
This brief set of notes contains some basic ideas and the most well-known theorems about
polynomials. I have not gone into deep details on the concepts that are too easy in practice (eg:
long division, synthetic division). I want you to be able to understand these ideas yourself. If
you have any problems on anything in this set of notes, let me know (or pester your teacher!)

Definition 1 A polynomial of degree n is a function f (x) of the form

f (x) = an xn + an−1 xn−1 + · · · + a2 x2 + a1 x + a0 ,

where an , an−1 , . . . , a1 , a0 are constants and an 6= 0.

For each i, 0 ≤ i ≤ n, ai is the coefficient of xi , and an is the leading coefficient of f (x).

The ai s may be in Z, Q, R or C. We denote the set of all polynomials in the variable x whose
coefficients are in Z by Z[x]. If f (x) ∈ Z[x], then we say that f (x) is a polynomial over Z.
Similar definitions hold when Z is replaced by Q, R or C. Note that Z[x] ⊂ Q[x] ⊂ R[x] ⊂ C[x].
The variable x is usually taken in R or C (if x is taken to be in C, then z may be used instead
of x).
f (x) is called monic if an = 1.
We write deg(f (x)) for the degree of f (x).
f (x) is a constant polynomial if n = 0, a linear polynomial if n = 1, a quadratic if n = 2,
a cubic if n = 3, a quartic if n = 4, and a quintic if n = 5.

Polynomial Division
If f (x) and g(x) are polynomials and deg(g(x)) ≤ deg(f (x)), then we may divide g(x)
into f (x) and get a remainder polynomial r(x) with deg(r(x)) < deg(g(x)). This works
quite like dividing an integer into another integer to get a remainder. More precisely, if
deg(g(x)) ≤ deg(f (x)) and g(x) 6= 0, then we can find polynomials q(x) and r(x) such that

f (x) = q(x)g(x) + r(x), where deg(r(x)) < deg(g(x)).

g(x) is the divisor, q(x) is the quotient and r(x) is the remainder.
If r(x) ≡ 0, then f (x) = q(x)g(x), and we say that g(x) divides f (x), or that f (x) is

1
divisible by g(x).
We can usually find q(x) and r(x) by long division.
Example. Divide g(x) = x2 − 1 into f (x) = x5 + 3x4 + 2x3 + x2 + x − 3.
Using long division,

x3 + 3x2 + 3x + 4
x2 − 1 x5 + 3x4 + 2x3 + x2 + x − 3
x5 − x3
3x4 + 3x3 + x2
3x4 − 3x2
3x3 + 4x2 + x
3x3 − 3x
4x2 + 4x − 3
4x2 −4
4x + 1

The quotient is q(x) = x3 + 3x2 + 3x + 4 and the remainder is r(x) = 4x + 1. So we have

f (x) = (x3 + 3x2 + 3x + 4)(x2 − 1) + (4x + 1).

The algorithm goes as follows. We find whatever we need to multiply x2 (the leading term
of g(x)) by to get x5 (the leading term of f (x)). That is x3 . We get x3 (x2 − 1) = x5 − x3 .
Put this below f (x), aligning the same powers. Put x3 on the top, aligning it with the −x3
term of x5 − x3 . Then, subtract x5 − x3 from x5 + 3x4 + 2x3 (the part of f (x) that sits above
x5 − x3 ) to get 3x4 + 3x3 . Then, bring down +x2 from f (x). Repeat the whole procedure until
we get a remainder term (4x + 1 in this example) which has degree less than g(x) (= 2 in this
example).
Note that if f (x) misses some powers, then we should put a coefficient of 0 for the missing
powers before carrying out the long division. For example, we should write 3x4 − 2x + 3 as
3x4 + 0x3 + 0x2 − 2x + 3. This helps us to preserve the alignment of the terms with the same
powers. Also, sometimes we may have to bring down more than one term of f (x) in one step.
Now try exercises 1 and 2.

The Remainder Theorem

Now, when we want to divide g(x) = x − a into f (x), then the remainder must have
degree at most 1, ie: it is a constant. We have the following very useful theorem.

2
Theorem 1 (The Remainder Theorem) Let f (x) ∈ C[x] and deg(f (x)) = n. Then we
have f (x) = q(x)(x − a) + r, where deg(q(x)) = n − 1 and r = f (a).

Definition 2 If f (x) is a function (not necessarily a polynomial in this definition) and f (a) =
0, then we say that a is a zero, or a root of f (x).

Corollary 2 Let f (x) ∈ C[x]. Then a is a zero of f (x) if and only if x − a divides f (x).

Proof. If a is a zero of f (x), then by the Remainder Theorem, f (x) = q(x)(x − a) + r =

q(x)(x − a), since r = f (a) = 0. So x − a divides f (x).
Conversely, If x − a divides f (x), then f (x) = q(x)(x − a) for some polynomial q(x). But
then f (a) = 0. 
Corollary 2 itself may also be called the Remainder Theorem. It is a very useful and
simple result that can be extremely useful in many Olympiad problems. This is because
if we know that a is a zero of a polynomial f (x), then we may immediately conclude that
f (x) = q(x)(x − a) for some polynomial q(x).

The Fundamental Theorem of Algebra and Synthetic Division

What do we know about the zeros of a polynomial f (x)? The following theorem is the
most important result.

Theorem 3 (The Fundamental Theorem of Algebra) Every polynomial f (x) ∈ C[x]

has a zero in C.

Theorem 3 is difficult to prove. There is more than one way to prove Theorem 3, but
every known proof uses advanced techniques which we will not go into here.
By an easy induction argument on the degree of f (x), and using the Remainder Theorem,
we have:

Corollary 4 If f (x) ∈ C[x] and deg(f (x)) = n, then f (x) may be factorized into the form

f (x) = a(x − α1 )(x − α2 ) · · · (x − αn ),

where a is the leading coefficient of f (x), and α1 , . . . , αn ∈ C are the zeros of f (x), not
necessarily distinct.

So, Corollary 4 tells us the following.

(a) If deg(f (x)) = n, then f (x) can have at most n distinct zeros in C.

3
 (b) If f (x) is a polynomial and has infinitely many zeros, then f (x) ≡ 0.

Corollary 4 also tells us that any polynomial can be factorized into a product of linear
polynomials in C[x]. But if f (x) ∈ R[x], and we want to only factorize f (x) into a product of
polynomials, each in R[x], then, we may have factors which are not linear. For example, we
know that, if f (x) is a quadratic over R, then we may not even be able to factorize f (x) into
two linear factors in R[x] if the discriminant of f (x) is less than 0.
In particular, factorizing a polynomial f (x) ∈ Z[x] into a product of polynomials, each
over Z, R or C, as far as possible (ie: the factors may not be factorized further; these factors
are called irreducible), is something very standard, and very important to know.
A very useful method in doing so is synthetic division.
Example. Factorize f (x) = x3 − 5x2 − 2x + 24 into a product of polynomials over Z.
We do the following. If x − a is a factor of f (x), then, by considering the constant terms,
we see that a divides 24. We make a guess that a = 3 is a zero, and do the following.

3 1 −5 −2 24
3 −6 −24
1 −2 −8 0

The algorithm goes as follows. Place the 3 and the coefficients of f (x) along the top row
as shown. Then, bring the 1 down to the third row. Then, multiply the 3 by the 1 in the
bottom to get 3, and place this 3 under the −5. Then, add the −5 and 3 to get −2 and place
this under the 3. Then multiply −2 by 3 to get −6, and repeat the previous procedure until
we get to the bottom right.
If we used a in place of 3, then the number in the bottom right is actually f (a). We see
that when we use 3, we get f (3) = 0. So, this means that 3 is a zero of f (x), and x − 3 is a
factor of f (x).
Furthermore, the numbers 1, −2, −8 that we have along the bottom row are the coeffi-
cients of the quotient polynomial. So, we have

f (x) = (x − 3)(x2 − 2x − 8).

We can factorize the quotient easily. The full factorization is f (x) = (x − 3)(x − 4)(x + 2).
Note that using 4 or −2 instead of 3 at the start would work as well, but for example,
if we used 1, then it would not have worked, since we would have f (1) 6= 0. It is a trial and
error procedure.
Example. Factorize f (x) = x3 − 2x2 + 5x + 26 into a product of polynomials (a) over Z,

4
and (b) over C.
We try synthetic division several times, using factors of 26 as possible zeros each time,
until we find one. We have

1 1 −2 5 26 −1 1 −2 5 26
1 −1 4 −1 3 −8
1 −1 4 30 1 −3 8 18

2 1 −2 5 26 −2 1 −2 5 26
2 0 10 −2 8 −26
1 0 5 36 1 −4 13 0

We see that 1, −1 and 2 are not zeros of f (x), but −2 is. We have

f (x) = (x + 2)(x2 − 4x + 13).

Now, we want to try and factorize x2 − 4x + 13. By the quadratic formula, the solutions
of x2 − 4x + 13 = 0 are √
4± 16 − 4 · 13
x= = 2 ± 3i.
2
So, x2 − 4x + 13 = (x − (2 + 3i))(x − (2 − 3i)), and x2 − 4x + 13 does not factorize over
Z. So the answers are:

(a) f (x) = (x + 2)(x2 − 4x + 13), and

(b) f (x) = (x + 2)(x − (2 + 3i))(x − (2 − 3i)).

Now, try exercises 3 and 4.

Relations between the zeros of a polynomial

If α1 , α2 are the solutions of x2 + a1 x + a0 = 0, then

x2 + a1 x + a0 = (x − α1 )(x − α2 ) = x2 − (α1 + α2 )x + α1 α2 ,

so a1 = −(α1 + α2 ) and a0 = α1 α2 .
Similarly, if α1 , α2 , α3 are the solutions of x3 + a2 x2 + a1 x + a0 = 0, then

x3 + a2 x2 + a1 x + a0 = (x − α1 )(x − α2 )(x − α3 )
= x3 − (α1 + α2 + α3 )x2 + (α1 α2 + α2 α3 + α3 α1 )x − α1 α2 α3 ,

5
so a2 = −(α1 + α2 + α3 ), a1 = α1 α2 + α2 α3 + α3 α1 , and a0 = −α1 α2 α3 .
In general, if the solutions of xn + an−1 xn−1 + · · · + a1 x + a0 = 0 are α1 , . . . , αn , then:
X
an−1 = − αi ,
i
X
an−2 = αi αj ,
i<j
X
an−3 = − αi αj αk ,
i<j<k
..
.
a0 = (−1)n α1 · · · αn .

To get an−i , we sum the products of the i-tuples of the αj s, and multiply by (−1)i .
Example. If α, β, γ are the zeros of f (x) = 2x3 + x2 − 5x − 3, find a polynomial whose
zeros are α2 , β 2 , γ 2 .
If f (x) = 0, then x3 + 21 x2 − 52 x − 32 = 0. So we have −(α + β + γ) = 12 , αβ + βγ + γα = − 52
and −αβγ = − 23 . So
2 2 2 2
 1 2  5  21
α +β +γ = (α + β + γ) − 2(αβ + βγ + γα) = − −2 − = ,
2 2 4
α2 β 2 + β 2 γ 2 + γ 2 α2 = (αβ + βγ + γα)2 − 2αβγ(α + β + γ)
 5 2 3  1  31
= − −2· − = ,
2 2 2 4
 3 2 9
α2 β 2 γ 2 = = .
2 4
So a possible polynomial is g(x) = 4x3 − 21x2 + 31x − 9.
Now try exercise 5.

Some algebraic identities

The following identities are useful when we deal with polynomials, and also in other areas
as well.

(a) xn − y n = (x − y)(xn−1 + xn−2 y + · · · + xy n−2 + y n−1 ),

(b) xn + y n = (x + y)(xn−1 − xn−2 y + · · · − xy n−2 + y n−1 ) if n is odd,

(c) x3 + y 3 + z 3 − 3xyz = (x + y + z)(x2 + y 2 + z 2 − yz − zx − xy)

= 12 (x + y + z)((y − z)2 + (z − x)2 + (x − y)2 ).

So, for example, (a) gives

xn − 1 = (x − 1)(xn−1 + xn−2 + · · · + x + 1).

6
Roots of Unity
The complex number Reiθ has modulus R and argument θ. It has real part R cos θ and
imaginary part R sin θ. So, we have Reiθ = R cos θ + iR sin θ.
If ω = e2πi/n , then the polynomial f (x) = xn − 1 has (complex) zeros 1, ω, ω 2 , . . . , ω n−1 .
These are the nth roots of unity. So over C, f (x) factorizes into

f (x) = (x − 1)(x − ω)(x − ω 2 ) · · · (x − ω n−1 ).

Since xn − 1 = (x − 1)(xn−1 + xn−2 + · · · + x + 1), it follows that

xn−1 + xn−2 + · · · + x + 1 = (x − ω)(x − ω 2 ) · · · (x − ω n−1 ),

that is, the zeros of g(x) = xn−1 + xn−2 + · · · + x + 1 are ω, ω 2 , . . . , ω n−1 .

On the Argand diagram, the complex numbers 1, ω, ω 2 , . . . , ω n−1 are the vertices of a
regular n-gon. We have the relation 1 + ω + ω 2 + · · · + ω n−1 = 0. Also, note that ω n = 1, and
hence ω j = ω k ⇐⇒ j − k ≡ 0(mod n).

Exercises and Problems

Standard exercises.

1. Use long division to divide f (x) = x5 + 2x3 − 3x2 − 5 by g(x) = x3 + 3x. Hence write
f (x) in the form q(x)g(x) + r(x), where deg(r(x)) < deg(g(x)).

2. Repeat exercise 1 with f (x) = x4 − 3x3 + 2 and g(x) = x2 + 3x − 6. What can you say?

3. Factorize f (x) = x4 + 6x3 − 12x2 − 88x − 96 into a product of polynomials over Z.

4. Factorize f (x) = x4 − 12x3 + x2 + 18x − 360 into a product of polynomials (a) over Z
and (b) over C.

5. Suppose that f (x) = x3 + ax2 + cx + d has zeros α, β and γ.

(a) Find a polynomial whose zeros are α2 , β 2 and γ 2 .

(b) Find a polynomial whose zeros are α3 , β 3 and γ 3 .

Problems.
Hint. Look through the set of notes carefully. At least one of the ideas discussed will be
relevant for each problem.
There are some easy problems here and some of medium difficulty.

7
 6. Find all polynomials f (x) such that f (1) = 1 and f (x2 + 2004) = (f (x))2 + 2004 for all
x ∈ R.

7. Let α, β, γ and δ be the zeros of f (x) = x4 + ax3 + bx2 + cx + d. Prove that if αβ = γδ,
then a2 d = c2 .

8. Suppose that f (x) = x2 + ax + 1 has zeros α, β, and g(x) = x2 + bx + 1 has zeros γ, δ.

Prove that (α − γ)(β − γ)(α + δ)(β + δ) = b2 − a2 .

9. Let f (x) ∈ Z[x]. Suppose that a, b, c, d ∈ Z are distinct and we have f (a) = f (b) =
f (c) = f (d) = −4. If r ∈ Z is a zero of f (x), prove that 4r = a + b + c + d.

10. Suppose that f (x) ∈ Z[x] satisfies f (21) = 17, f (32) = −247 and f (37) = 33. Prove
that if f (n) = n + 51 for some n ∈ Z, then n = 26.

11. Suppose that f (x) ∈ Z[x], and f (2) is divisible by 5, and f (5) is divisible by 2. Prove
that f (7) is divisible by 10.
1
12. Suppose that f (x) is a polynomial of degree n such that f (k) = for k = 1, 2, 4, 8, . . . , 2n .
k
Find the value of f (0).

13. Suppose that f (x) ∈ C[x] and deg(f (x)) < k. Let ω = e2πi/k . Prove that
k−1
1X
f (ω i ) = f (0).
k i=0

14. Find all polynomials f : R → R such that 2(1 + f (x)) = f (x − 1) + f (x + 1) for all
x ∈ R.

15. Find all polynomials f : R → R such that (x − 16)f (2x) = 16(x − 1)f (x) for all x ∈ R.

16. Find all polynomials f : C → C such that f (x)f (x + 1) = f (x2 ) for all x ∈ C.

17. Let f (x) = xk−1 + xk−2 + · · · + x + 1. Find the remainder when f (xk ) is divided by f (x).

18. Let f (x) = x4 − 3x3 + 5x2 − 9x. Find all pairs of integers (a, b), where a < b, such that
f (a) = f (b).

19. Find all pairs (n, k), where n ∈ Z, n > 0 and k ∈ R, such that f (x) = (x + 1)n − k is
divisible by g(x) = 2x2 + 2x + 1.