defns_of_continuity_updated
defns_of_continuity_updated
defns_of_continuity_updated
Aditya Ghosh
July 2019
Warm Up
√ √
1. Suppose that xn ≥ 0 for all n ≥ 1 and xn converges to x. Show that limn→∞ xn = x.
2. Suppose that xn converges to x. Show that sin xn converges to sin x and cos xn converges
to cos x. Is it necessary that tan xn converges to tan x?
Solution:
√ √ |xn − x| |x − x|
| xn − x| = √ √ ≤ n√ . (∗)
xn + x x
√
Since xn → x, we have an N ∈ N such that |xn − x| < ε x holds for all n ≥ N . Hence,
√ √
(∗) implies that | xn − x| < ε holds for every n ≥ N.
2. We shall use the fact that | sin t| ≤ |t| which holds for all t ∈ R. (It can be proved using
a diagram of a unit circle.) Using this, we obtain
xn − x xn + x xn − x
| sin xn − sin x| = 2 sin cos ≤ 2 sin ≤ |xn − x|.
2 2 2
Since |xn − x| < ε holds for all but finitely many n, we get the same for | sin xn − sin x|.
Hence we conclude that sin xn converges to sin x. For cos, we can do as follows: since
π π
yn = 2
− xn converges to y = 2
− x, we can say that sin yn converges to sin y. This is
same as saying that cos xn converges to cos x. Next, for tan, observe that if x 6= (2k+1) π2
sin xn sin x
for k ∈ Z then we get tan xn = → = tan x as n → ∞. But if x = (2k + 1) π2
cos xn cos x
for some k ∈ Z then it might happen that tan xn does not converge at all. Here is a
π
counter-example: xn = 2
− n1 if n is odd and π
2
+ n1 if n is even. Convince yourself that
although xn converges to π2 , tan xn does not converge at all.
Suppose xn is a sequence that converges to x (in symbols, xn → x). We have seen earlier
that axn + b → ax + b, |xn | → |x|, xkn → xk (k ∈ N), sin xn → sin x, cos xn → cos x,
√ √
xn → x (provided xn ≥ 0), tan xn → tan x (provided cos x 6= 0). You can also show
1
that for any polynomial P (·), we have P (xn ) → P (x). This idea generalizes to the notion of
continuity:
As per our discussion above, it follows that ax + b, |x|, xk (k ∈ N), sin x, cos x are con-
√
tinuous at every a ∈ R, x is continuous at every a ≥ 0 and tan x is continuous at every
/ {(2k + 1) π2 : k ∈ Z}. Next, we shall see a few properties of continuous functions.
a∈
Theorem. Suppose that f (x) and g(x) are continuous at x = a. Then, the following
functions are also continuous at x = a: f (x) + g(x), f (x) − g(x), cf (x) (where c is a
constant), f (x)g(x), |f (x)|, max{f (x), g(x)}, min{f (x), g(x)}.
Proof. Take any sequence xn that converges to a. Since f, g are continuous at x = a, we have
lim f (xn ) = f (a), lim g(xn ) = g(a). Hence we get lim f (xn ) + g(xn ) = f (a) + g(a). This
n→∞ n→∞ n→∞
shows that f +g is continuous at x = a. In a similar manner we can show that f −g, cf, f g, |f |
are continuous at x = a. Next, let us consider h(x) = max{f (x), g(x)}. Observe that for
any y, z ∈ R, we have
y + z + |y − z|
max{y, z} = .
2
Hence,
1
h(x) = f (x) + g(x) + |f (x) − g(x)| .
2
Complete the proof yourself. Do the same for min{f (x), g(x)}.
Note that the last theorem allows us to construct a large class of continuous functions
√ x + cos x
from the elementary ones, e.g. any polynomial, | sin x|, 1 − x2 , 2 , etc.
x +1
Next, we shall see another definition of continuity. Suppose f (x) is continuous at x = a.
When we draw the graph of f (x), what do you expect (for the function to be continuous at
x = a)? It is expected that as x gets closer and closer to a, f (x) gets closer and closer to
f (a). This intuition is formalised in the following definition:
Definition. We say that f (x) is continuous at x = a if, for every ε > 0 there exists δ > 0
such that for all x ∈ (a − δ, a + δ) it holds that f (x) ∈ (f (a) − ε, f (a) + ε).
In words, it says that f (x) will be sufficiently close to f (a) if we take x to be sufficiently
close to a. To have f (x) to be at a distance less than ε from f (a), we need x to be at a
distance less than δ from a. The definition says that for every ε > 0, there exists such a
δ > 0.
2
One important point to note here is the order of the quantifiers ‘for all’ and ‘there exists’.
Why does their ordering matter? You should get an answer by observing the following
statements:
(i) ∀ city C in India, ∃ a 6-digit number n such that n is the pincode of city C.
(ii) ∃ a 6-digit number n such that ∀ city C in India, n is the pincode of city C.
Can you give a real-life example of continuous function? There are plently of examples,
e.g. your age as a function of time, your height as a function of your age (not rounded
off to ‘years’), speed of a car. Can you give a real-life example of a function which is not
continuous? One example is the fare of any public transport (bus, metro) as a function of
the distance from the place you boarded. (From one stoppage to another it jumps in discrete
steps, it does not change continuously.)
One question that might come to your mind: how can we have two seemingly different
definitions for the same thing (continuity)? We shall address this question shortly; before
that let us do a few problems using each of the two approaches: sequential definition and
ε − δ definition.
Problem. Suppose f : R → R is a function that satisfies
(Here λ > 0 is fixed.) Show that f (x) is continuous (at x = a for every a ∈ R).
Proof. (i) Using sequential definition: Take any sequence xn that converges to a. Since f (x)
is continuous at x = a, we deduce that f (xn ) converges to f (a). Now, since yn = f (xn ) is a
sequence that converges to f (a), the continuity of g(y) at y = f (a) yields that g(yn ) must
3
converge to g(f (a)). This is same as saying that g(f (xn )) = h(xn ) converges to g(f (a)) =
h(a). Since this holds for every sequence xn that converges to a, we conclude that h(x) is
continuous at x = a.
(ii) Using ε − δ definition: Fix any ε > 0. Since g(y) is continuous at y = f (a), there
exists a δ > 0 such that for every y ∈ (f (a) − δ, f (a) + δ) we have |g(y) − g(f (a))| < ε, i.e.,
Again, since f (x) is continuous at x = a, we have a δ 0 > 0 such that for every x ∈ (a−δ 0 , a+δ 0 )
we have |f (x) − f (a)| < δ. Combining this with (?) we obtain
(We used (?) for y = f (x).) Thus, we have shown that for every ε > 0 there exists δ 0 > 0
such that |g(f (x)) − g(f (a))| < ε holds whenever |x − a| < δ 0 .
Problem. Consider the function f (x) = bxc = largest integer less than or equal to x. Show
that f (x) is continuous at x = a if and only if a is not an integer.
∀ε > 0, ∃δ > 0 such that ∀x ∈ (a − δ, a + δ), it holds that |f (x) − f (a)| < ε.
To negate this definition, start from the left and change each ∀ to ∃ and vice-versa. The
correct negation of the above will be: f (x) is not continuous at x = a if
4
Problem. Consider the following properties:
(1) For every ε > 0 there exists δ > 0 such that |f (x)−f (a)| < ε holds whenever |x−a| < δ.
(2) For every sequence xn that converges to a it holds that f (xn ) converges to f (a).
Show that a function f has property (1) if and only if it has property (2).
Proof. First we shall show that (1) ⇒ (2). Suppose a function has property (1). Take any
sequence that converges to a. We need to show that f (xn ) converges to f (a). Fix any ε > 0.
By property (1), there exists δ > 0 such that |f (x) − f (a)| < ε holds whenever |x − a| < δ.
Since xn → a, there exists N ∈ N such that |xn − a| < δ holds for every n ≥ N. Then, for
every n ≥ N we obtain |f (xn ) − f (a)| < ε. This completes the proof.
Next we show that (2) ⇒ (1). Suppose a function has property (2). It is actually hard
to give a δ > 0 for any fixed ε > 0 where δ has the desired property. So we shall use the
method of contradiction. Let, if possible, f be a function with property (2), which does not
satisfy (1). The negation of (1) reads: there exists an ε > 0 such that for every δ > 0, there
1
exists an x ∈ (a − δ, a + δ) for which |f (x) − f (a)| ≥ ε. So, for each n ∈ N, we set δ = n
to get a ‘bad’ x in (a − n1 , a + n1 ), which we call xn . Therefore, xn is a sequence such that
xn ∈ (a − n1 , a + n1 ) and |f (xn ) − f (a)| ≥ ε for every n ∈ N. Since a − 1
n
< xn < a + n1 ,
Sandwich tells us that xn converges to a. But f (xn ) being at least ε away from f (a), we see
that f (xn ) can not converge to f (a). This contradicts the fact that f has property (2) and
hence completes the proof.
Exercises
2. Suppose that f (x) and g(x) are not continuous at x = a. Is it possible that f (x) + g(x)
is continuous at x = a?
4. Suppose that f (x) and g(x) are not continuous at x = a. Is it possible that f (x)g(x) is
continuous at x = a?
5
(a) f (x) is discontinuous at x = a but g(y) is continuous at y = f (a).
(b) f (x) is continuous at x = a but g(y) is discontinuous at y = f (a).
(c) f (x) is discontinuous at x = a and g(y) is discontinuous at y = f (a).
Hints/Answers
1. If f is continuous at x = a and f (a) 6= 0 then we can show that 1/f is also continuous
at x = a. (Note that it is also a necessary condition.) To prove this, you can use the
sequential definition.
2. Yes. Take f (x) = [x] and g(x) = −[x]. Both f and g are discontinuous at any a ∈ Z,
but f + g is continuous everywhere.
6
6. (a) Yes. Take any function f which is discontinuous at x = a and set g ≡ 0 (i.e.
g(x) = 0 for all x).
1 if x is rational √
(b) Yes. Take g(x) = and f (x) = 2 for all x. You can show
0 if x is irrational
that g is not continuous anywhere. But, g(f (x)) = 0 for all x, hence g ◦ f is
continuous everywhere.
(c) Yes. Set a = 0. Take f (x) = g(x) = 1/x for x 6= 0 and define f (0) = g(0) = 0.
Observe that g(f (x)) = x for all x.
7. When a 6= 0, its easy to see that f (x) is continuous x = a. We claim that f is discon-
tinuous at x = 0. To prove our claim, it sufficies to give two sequences xn and yn such
that both of them converge to 0, but f (xn ) and f (yn ) do not converge to the same
limit. One possible example is the followng:
1 1
Take xn = and yn = , for all n ≥ 1.
2nπ + π/2 2nπ − π/2
8. When a 6= 0, its easy to see that f (x) is continuous x = a. We claim that f is continuous
at x = 0 as well. Observe that for every x, we have |f (x)−f (0)| = | sin x| ≤ |x| = |x−0|.
So, we can take δ = ε to meet the requirement in the ε − δ definition.
9. It should be clear that f is continuous at x = a for every a ∈ [1, 2) ∪ (2, 3]. Only
what can confuse you is the question of continuity at x = 2. However, since f is not
defined at x = 2, this question does not arise, because f (2) is not defined!. (This is
somewhat controversial. For f to be continuous at x = a, it is universally accepted that
x must belong to the domain of f. But for discontinuity, some authors (e.g. Rudin)
define f to be discontinuous at x = a only when a belongs to the domain and f is not
continuous at a. I also prefer this definition and hence say that f is neither continuous
nor discontinuous at x = 2.)
Note, though the function f is continuous on its entire domain, one cannot draw the
graph without lifting the pen. (Who told you that the graph of a continuous function
must be drawn without lifting the pen!).
10. The function is continuous on its domain A. As pointed out above, the question of f
being continuous at x = a does not arise if a ∈
/ A.
11. Carry out the steps given in the hint. Also see the next exercise set.