MATH 436 Notes: Functions and Inverses.

Jonathan Pakianathan
September 12, 2003

1 Functions
Definition 1.1. Formally, a function f : A B is a subset f of A B with
the property that for every a A, there is a unique element b B such that
(a, b) f . The set A is called the domain of f and the set B is the codomain
of f .

While the above definition provides a definition of a function purely in

terms of set theory, it is usually not a useful picture to work with. However
it does emphasize the important point that the domain and codomain of a
function are an intrinsic part of any function f .
Less formally, we usually think of a function f : A B as a rule of
assignment which assigns a unique output f (a) B for every input a A.
The graph of f , denoted Graph(f ) = {(a, f (a))|a A} A B then
recovers the more formal set theoretic definition of the function.

Definition 1.2. Let f : A B be a function.

Given S A we define f (S) = {f (s)|s S}. Note that f (S) B. f (S)
is called the image of the set S under f .
f (A) is called the image of f , and is denoted Im(f ).
Given T B we define f 1 (T ) = {a A|f (a) T }. Note that f 1 (T )
A. f 1 (T ) is called the preimage of the set T under f .

Fix a function f : A B, then it is easy to see that for all S A,

S f 1 (f (S)) and for all T B, we have f (f 1(T )) T . The next
example shows that these inclusions do not have to be equalities in general.

1
Example 1.3. Define f : Z Z via f (n) = n2 for all n Z. Then
Im(f ) = {0, 1, 4, 9, 16, . . . }, f ({2}) = {4} and f 1 ({0, 1, 2}) = {0, 1, 1}. It
follows that f (f 1 ({0, 1, 2})) = {0, 1} and f 1 (f ({2})) = {2, 2}.
The following are among the most important concepts involving functions,
we shall see why shortly.
Definition 1.4. Given a function f : A B we say that:
(a) f is surjective (equivalently onto) if Im(f ) = B.
(b) f is injective (equivalently one-to-one) if for all a, a A,
(f (a) = f (a )) = (a = a ).
Equivalently, f is injective if for all a, a A, (a 6= a ) = (f (a) 6= f (a )).
Thus an injective function is one that takes distinct inputs to distinct outputs.
(c) f is bijective (equivalently a one-to-one correspondence) if it has both
properties (a) and (b) above.
These properties are important as they allow f to be inverted in a certain
sense to be made clear soon.
First let us recall the definition of the composition of functions:
Definition 1.5. If we have two functions f : A B and g : B C then
we may form the composition g f : A C defined as (g f )(a) = g(f (a))
for all a A.
It is fundamental that the composition of functions is associative:
Proposition 1.6 (Associativity of composition). Let f : A B, g :
B C, h : C D be functions. Then (h g) f = h (g f ).
Proof. The two expressions give functions from A to C. To show they are
equal we only need to show they give the same output for every input a A.
Computing we find:
((h g) f )(a) = (h g)(f (a)) = h(g(f (a))).
(h (g f ))(a) = h((g f )(a)) = h(g(f (a))).
Since the two expressions agree, this completes the proof.
Definition 1.7. For any set S, the identity function on S, denoted by 1S :
S S is the function defined by 1S (s) = s for all s S.
It is easy to check that if we have a function f : A B then
f 1A = f = 1B f.

2
 Now we are ready to introduce the concept of inverses:
Definition 1.8. Suppose we have a pair of functions f : A B and g :
B A such that g f = 1A . Then we say that f is a right inverse for g and
equivalently that g is a left inverse for f .
The following is fundamental:
Theorem 1.9. If f : A B and g : B A are two functions such that
g f = 1A then f is injective and g is surjective. Hence a function with
a left inverse must be injective and a function with a right inverse must be
surjective.
Proof. g f = 1A is equivalent to g(f (a)) = a for all a A.
Showing f is injective: Suppose a, a A and f (a) = f (a ) B. Then
we may apply g to both sides of this last equation and use that g f = 1A
to conclude that a = a . Thus f is injective.
Showing g is surjective: Let a A. Then f (a) B and g(f (a)) = a.
Thus a g(B) = Im(g) for all a A showing A = Im(g) so g is surjective.

The following example shows that left (right) inverses need not be unique:
Example 1.10. Let A = {1, 2} and B = {a, b, c}. Define f1 , f2 : A B
by f1 (1) = a, f1 (2) = c, f2 (1) = b, f2 (2) = c. Define g1 , g2 : B A by
g1 (a) = g1 (b) = 1, g1 (c) = 2 and g2 (a) = 1, g2 (b) = g2 (c) = 2. Then it is an
easy exercise to show that g1 f1 = 1A = g1 f2 so that g1 has two distinct
right inverses f1 and f2 . Furthermore since g1 is not injective, it has no left
inverse.
Similarly one can compute, g2 f1 = 1A so that f1 has two distinct left
inverses g1 and g2 . Furthermore since f1 is not surjective, it has no right
inverse.
From this example we see that even when they exist, one-sided inverses
need not be unique.
However we will now see that when a function has both a left inverse and
a right inverse, then all inverses for the function must agree:
Lemma 1.11. Let f : A B be a function with a left inverse h : B A
and a right inverse g : B A. Then h = g and in fact any other left or
right inverse for f also equals h.

3
Proof. We have that h f = 1A and f g = 1B by assumption. Using
associativity of function composition we have:

h = h 1B = h (f g) = (h f ) g = 1A g = g.

So h equals g. Since this argument holds for any right inverse g of f , they
all must equal h. Since this argument holds for any left inverse h of f , they
all must equal g and hence h. So all inverses for f are equal.
We finish this section with complete characterizations of when a function
has a left, right or two-sided inverse.

Proposition 1.12. A function f : A B has a left inverse if and only if it

is injective.

Proof. = : Follows from Theorem 1.9. =: If f : A B is injective then

we can construct a left inverse g : B A as follows. Fix some a0 A and
define (
a if b Im(f ) and f (a) = b
g(b) =
a0 otherwise
Note this defines a function only because there is at most one a with f (a) = b.
It is an easy computation now to show g f = 1A and so g is a left inverse
for f .

Proposition 1.13. A function f : A B has a right inverse if and only if

it is surjective.

Proof. = : Follows from Theorem 1.9. =: Suppose f : A B is

surjective. Then for each b B, f 1 ({b}) is a nonempty subset of A. Thus
by the Axiom of Choice we may construct a choice function g : B A
such that g(b) is a choice of element from the nonempty set f 1 ({b}) for all
b B. It is easy now to compute that f g = 1B and so g is a right inverse
for f .

Proposition 1.14. A function f : A B has a two-sided inverse if and

only if it is bijective. In this case, the two-sided inverse will be unique and is
usually denoted by f 1 : B A.

4
Proof. First note that a two sided inverse is a function g : B A such
that f g = 1B and g f = 1A . = : Theorem 1.9 shows that if f has
a two-sided inverse, it is both surjective and injective and hence bijective.
=: Now suppose f is bijective. From the previous two propositions, we
may conclude that f has a left inverse and a right inverse. By Lemma 1.11
we may conclude that these two inverses agree and are a two-sided inverse
for f which is unique. Alternatively we may construct the two-sided inverse
directly via f 1 (b) = a whenever f (a) = b.