Lecture 18: Connectedness

Now that we have properly defined open, closed sets and limit points in a
topological space, we can proceed to define the properties of connectedness and
compactness for arbitrary topological space. The properties of connectedness and
compactness are useful in proving three basic theorems about continuous functions
which lie at the core of calculus. These are the following:

i) Intermediate Value Theorem. If f : [a, b] → R is a continuous function and if r

is a real number between f (a) and f (b), then there exists an element c ∈ [a, b]
such that f (c) = r.

ii) Maximum Value Theorem. If f : [a, b] → R is continuous, then there exists

an element c ∈ [a, b] such that f (x) ≤ f (c) for every x ∈ [a, b].

iii) Uniform Continuity Theorem. If f : [a, b] → R is continuous, then given

epsilon > 0, there exists δ > 0 such that |f (x1 ) − fx2 | ≤  for every pair of
numbers x1 , x2 of [a, b] for which |x1 − x2 | < δ.

1 Connected Spaces
Coarsely speaking, a topological space is said to be connected if it is not separable
which implies that the space cannot be broken up into two disjoint open sets. A
more formal definition follows.

Definition 1.1. Let X be a topological space. A separation of X is a pair U, V

of disjoint nonempty open sets of X whose union is X. The space X is connected
if there does not exist a separation of X.

Connectedness is a topological property, since it is formulated entirely in terms

of the collection of open sets in X.
Remark 1. If the topological space X is connected, then so is any space homeo-
morphic to X. [Show proof]

1
 An alternate definition of connectedness is the following:
A space X is connected if and only if the only subsets of X that are both open and
closed in X are empty set and X itself.
Proof. Let X be not connected but separable. Then, there exists open sets U and
V which form a separation of X. The, by definition of separability, U is an open
set which is neither empty nor equal to X. Also since U = X\V , U is closed as
well. Likewise, V is also both open and closed. Conversely, let us assume that
there exists a set A which is neither empty nor equal to X which is both open and
closed in X. Then set X\A is an open subset of X. Also, A and X\A, together
form a separation for X and hence set X is not connected. Thus, we have proved
the contrapositive of the reverse statement.
For a subspace Y of X, we now define of connectedness as follows.
Theorem 1.2. If Y is a subspace of X, a separation of Y is a pair of disjoint
nonempty sets A and B whose union is Y , neither of which contain a limit point
of the other. The space Y is connected if there exists no separation of Y .
Proof. Let A and B form the separation of Y . We need to show that A and B
do not contain each other’s limit points. We first show that B does not contain
any limit points of A. Since Ā ∩ Y is the closure of A in Y , we need to show that
its intersection with B is an empty set. Here, Ā is the closure of A in X. Since
A is also closed in Y , we have A = Ā ∩ Y . But since A and B are disjoint by
hypothesis, Ā ∩ Y is also disjoint with B. Hence, we are done. Similarly, we can
show that A does not contain any limit points of B.
Conversely, suppose that A and B are disjoint nonempty sets whose union is Y ,
neither of which contains a limit point of the other, which means Ā∩B = φ. Along
with the facts that A ∩ B = φ and (A ∩ B) ⊂ (Ā ∩ B), we conclude that A = Ā
i.e, A is closed. Likewise, we can show that B is also closed. Since B = Y \A and
A = Y \B, both A and B are open in Y as well.
Example 1.3. Let X denote a two-point space in in-discrete topology T {φ, X}.
Clearly, there is no separation of X, so X is connected.
Example 1.4. Let Y denote the subspace [−10) ∪ (0, 1] on the real line R. Each
of the sets [−1, 0) and (0, 1] are nonempty and open in Y (although not in R);
therefore, they form a separation of Y . Alternatively, note that neither of these
sets contains a limit point of the other.
Example 1.5. Let X be the subspace [−1, 1] of the real line. The sets [−1, 0] and
(0, 1] are disjoint and nonempty, but they do not form a separation of X, because
the first set is not open in X. Alternatively, first set contains the limit point 0, of
the second set. Indeed, there exists no separation of the space [−1, 1].

2
Example 1.6. The set of rational Q on the real line is also not connected. The
only connected subspaces of Q are the one point sets: If Y is a subspace of Q,
containing two points p and q, one can choose an irrational number a lying between
p and q, and write Y as a union of open sets
Y ∩ (−∞, a) and Y ∩ (a, ∞).
Theorem 1.7. If the sets C and D form a separation of X, and if Y is a connected
subspace of X, then Y lies entirely within either C or D.
Proof. Since C and D form a separation of X, the sets C and D are open in X and
consequently the sets C ∩ Y and D ∩ Y are open in Y . Since C and D are disjoint,
so are C ∩ Y and D ∩ Y . Also, (C ∩ Y ) ∪ (D ∩ Y ) = (C ∪ D) ∩ Y = X ∩ Y = Y , the
pair {(C ∩ Y ), (D ∩ Y )} constitute a valid separation for Y . From connectedness
of Y , one of them must be empty and therefore, the other one must contain the
entire set Y . Hence, Y has to lie completely inside either C or D.
Now, we will try to build more sophisticated connected sets by combining
smaller connected sets. Following theorem shows us how to do so.
Theorem 1.8. The union of collection of connected subspaces of X that have a
point in common is connected.
Proof. Let {Aα } be a collection of connected subspaces of space X. Let p be a
point which is common to all Aα i.e., p ∈ Aα . We need to show that Y = ∪α Aα is
a connected space. We prove this via contradiction. Let us assume that C and D
form the separation of subspace Y in X. Since Aα ⊆ Y , and Aα is connected, by
Theorem 1.7, Aα must lie completely inside either C or D. Same goes for all α. Let
there be α1 and α2 such that Aα1 and Aα2 belong to C and D respectively. Since
C ∩ D = φ, Aα1 and Aα2 cannot have a common element, which is a contradiction.
Therefore, said α1 and α2 do not exist and there is no non-trivial separation of Y ,
and hence it is connected.
Theorem 1.9. Let A be a connected subspace of X. If A ⊂ B ⊂ Ā, then B is
also connected.
Proof. Suppose that C and D constitute the separation of B. Since A is a con-
nected subset of B, A must lie completely inside either C or D. Without loss of
generality, let A ⊂ C. This implies that Ā ⊂ C̄. Since D does not contain any
limit points of C, we have C̄ ∩ D = φ. Consequently, Ā ∩ D = φ. Since B ⊂ Ā, B
and D do not intersect, implying that D is empty. This contradicts our assumption
that C and D form a separation for B. Thus, B has to be connected.
Theorem 1.10. The image of a connected space under a continuous map is con-
nected.

3
Proof. Let f : X → Y be a continuous map; let X be connected. Then, we need
to show that the image space Z = f (X) is also connected. Let g : X → Z be a
map obtained from f by restricting its range to Z, Then, g is also a continuous
and surjective map. Let Z = A ∪ B be a separation of Z into disjoint nonempty
open sets in Z. Then, g −1 (A) and g −1 (B) are also disjoint open sets in X, since
g is a continuous map. Also g −1 (A) and g −1 (B) are nonempty as g is surjective.
Therefore, g −1 (A) and g −1 (B) form a nontrivial separation of X, contradicting the
assumption that X is connected.

Theorem 1.11. A finite cartesian product of connected spaces is connected.

Proof. We start by proving the theorem for two connected spaces X and Y . Choose
a “base point” a × b in the product space X × Y . Note that the horizontal slice
X × b is connected as a consequence of it being homeomorphic with X via an
extension map f : X → X × R, which maps x ∈ X to (x, b). Likewise, each
vertical slice a × Y is homeomorphic with Y . As a result, the “T-shaped” space
Tx (X × b) ∪ (a × Y ) is also connected (Theorem 1.8). Now, proceed to form the
union ∪x Tx , which is also connected as it is a union of connected spaces with a
common point a × b. Since this union equals X × Y , the space X × Y is also
connected.
The proof for any finite number of connected spaces follows similarly by induc-
tion, using that fact that X1 ×X2 . . . Xn is homeomorphic with (X1 ×X2 . . . Xn−1 )×
Xn
It is natural to ask whether the above theorem extends to arbitrary products of
connected spaces. The answer depends on which topology is used for the product,
as the following example shows.

Example 1.12. Consider the Cartesian product R∞ in the box topology. We can
write R∞ as the union of the set A consisting of all bounded sequences of real
numbers, and the set B of all the unbounded sequences. These sets are disjoint,
and each is open in the box topology.
If a is a point of R∞ , the open set

U = (a1 − 1, a1 + 1) × (a1 − 1, a1 + 1) × . . .

consists entirely of bounded sequences if a is bounded, and of unbounded sequences

if a is unbounded. Thus, even through R is connected, R∞ is clearly not connected
as we have demonstrated its nontrivial separation.