You are on page 1of 6

2 Limits and Continuity

2.1 Definition • 2.2 Computing Limits • 2.3 Continuity • 2.4 Examples of limits

2.1 Definition
Our main definition is the following.

lim f (x) = L
x→a

means that however x approaches the fixed number a, the


value of the function at x, f (x), approaches the number L.
If such an L does not exist, we say that the limit does not
exist.

Note that in this definition the value of f at a itself does not matter.
For example, the following picture illustrates this for limx→2 (x2 − x + 2) = 4:

The tricky part of this definition is the different ways that x can approach a. For instance, x
can approach a from the left or the right, and the result for f (x) might be different. Consider
the following graph, which uses the graphing convention that an open circle means that that
point is not on the graph; here f (1) = 1, not 2.

Here as x goes to 1 from the left, f (x) goes to 1, but if x comes from the right, f (x) goes to
2. So in this case, the definition above tells us that limx→1 f (x) does not exist, because two
different ways of approach give different approached values. Nevertheless, we would still like
to have notation for one-sided limits, which only involve the x on one side of a.

1
We write lim f (x) = L
x→a−

if as x approaches a from the left, f (x) approaches L.


Similarly, we write
lim+ f (x) = L
x→a

if as x approaches a from the right, f (x) approaches L.

So be warned, a tiny change of notation (a− instead of a) makes a big difference in meaning.
In the second graph above, where the two-sided limit did not exist, we have

lim f (x) = 1 and lim f (x) = 2.


x→1− x→1+

It is still possible for a one-sided limit to not exist. For example, limx→0 x12 does not exist,
because as x gets smaller and smaller, x12 gets bigger and bigger, so does not approach any
number.
This gives a way to check if a two-sided limit exists, by first checking if the two one-sided
limits exist, and then seeing if their values are equal.

We have lim f (x) = L


x→a

if both
lim f (x) = L and lim f (x) = L.
x→a− x→a+

If the left and right limit are not equal, or if one of the
one-sided limits does not exist, then the (two-sided) limit
does not exist.

So in the first graph above, we have limx→2 f (x) = 4 since the left and right limits are both
equal to 4. In the second graph, limx→1 f (x) does not exist, since the left and right limits
are different.
This left/right approach is especially convenient (and necessary) for piecewise defined func-
tions, when the number a is on the border between two of the pieces. The most common
example of this is the absolute value function, so let’s look at
lim |x|.
x→0
To determine the existence and value of this limit, we need to look at the one-sided limits
first. For the left-hand limit, we only need to consider x with x < 0, for which the definition
of |x| says |x| = −x. Hence
lim− |x| = lim− (−x) = 0.
x→0 x→0

The last step is true because as x approaches 0, so does −x.


Similarly,
lim+ |x| = lim+ x = 0.
x→0 x→0

So indeed the left and right limit are equal, and we can conclude that limx→0 |x| = 0.
Of course the above limit would have been easy to see from the graph of |x|, but we want to
be able to determine limits without looking at the graph of the function, especially because
later on we will have to use limits to draw graphs.

2
As another example, consider
½
2x − 1 if x ≤ 1
f (x) =
x2 if x > 1.

Then to the left of x = 1, the function is gives by 2x − 1, and on the right by x2 , so if we


calculated the two one-sided limits separately, we get
µ ¶
lim− f (x) = lim− (2x − 1) = 2 lim− x − 1 = 2 · 1 − 1 = 1,
x→1 x→1 x→1

µ ¶2
2
lim+ f (x) = lim+ x = lim+ x = 12 = 1.
x→1 x→1 x→1

So the left and right limits exist and both equal, hence we can conclude that limx→1 f (x) = 1.

Remark that when we say ’x approaches a’, we don’t care about what happens when x equals
a, since that is not part of the approach. If instead of |x| we considered the (weird) function

 −x if x < 0
h(x) = 13 if x = 0

x if x > 0,

the left and right limits would still be 0, so the limit exists and limx→0 h(x) = 0, even though
h(0) = 13.

2.2 Computing Limits


The following example illustrates the basic approach to computing limits of functions given
by nice formulas.

lim (2x2 − 3x + 4) = lim (2x2 ) + lim (−3x) + lim 4


x→5 x→5 x→5 x→5
2
= 2(lim x) − 3(lim x) + (lim 4)
x→5 x→5 x→5
2
= 2 · 5 − 3 · 5 + 4 = 50 − 15 + 4 = 39.

The idea is to reduce the big limit to several small limits which are easy to determine, usually
one of the types
lim x = a or lim b = b.
x→a x→a

Since the left/right issue is not a problem for these two easy limits, we do not have to worry
about that in this approach.
The question is if all the steps above are actually valid: is it true for instance that
³ ´2
lim x2 = lim x ?
x→5 x→5

We call this ’interchanging the order of the operations’: on the left you square x first and
then consider the limit of that, on the right you first take the limit and then square that.
This is what the ’limit laws’ (Stewart 2.3) answer: can we interchange the limit operation
with the basic arithmetic operations? I like to summarize these laws as follows:
Limits interchange perfectly with addition, multiplication, division and root taking, except
that we can’t divide by 0, and we can’t take even roots of anything negative.

3
So for polynomials like the one above, everything works out. The same is true for rational
and algebraic functions, as long as neither of the exceptions occurs. For instance
√ √ √
3
x+6 3
limx→2 x + lim 6 3
2+6 1
lim 2 = 2
= = .
x→2 x + 1 (limx→2 x) + lim 1 4+1 5

Two examples that do not work are

x2 + 1 √
lim and lim 3 4 x + x2 ;
x→2 x − 2 x→−1

in the first we would be dividing by limx→2 x − 2 = 0, in the second we’d get a 4 −1.
So now we can compute any limit of an algebraic function (or find that the limit doesn’t
exist), by interchanging limits and operations until we have an easy limit, or until we run
into one of the exceptions.
Writing this down step-by-step will get annoying quickly, as you’ll realize that you’re basically
plugging in the number that x approaches. We can simplify this, but we still have to keep
in mind when it’s allowed and when not. For that we will need the notion of continuity.

2.3 Continuity
A function is continuous at a if
lim f (x) = f (a),
x→a

and it is continuous on the interval I if it is continuous at all numbers a in the interval I. If


we just say that a function is continuous, it means that it is continuous on all of its domain.
This means that for x close to a, the function value f (x) is also close to f (x). An intuitive
way to think of it is that you can draw the graph of f (above I) without lifting your pen
from the paper (though this is incorrect in some cases, e.g. if there is a gap in the domain).
Yet another way to think of it is that ’applying f ’ is an operation that we can interchange
limits with, because another way to write the equation above is (just plug in a = limx→a x):

lim f (x) = f (lim x).


x→a x→a

So if f is continuous, applying f then the limit is the same as the other way around. For
example, when we wondered if limx→5 x2 = (limx→5 x)2 , we could just as well have asked if
the function f (x) = x2 is continuous at x = 5. The limit laws then said that

Polynomials, rational and algebraic functions are continuous, except when there is
division by zero, or an even root of a negative number.

To check from the definition if a given function is continuous, we need to check the three
things that are implicit in the definition:

• f (a) is defined (that is, a is in the domain of f )

• lim f (x) exists


x→a

• lim f (x) = f (a).


x→a

4
If any of these fails, we say that the function is discontinuous at a, or has a discontinuity.
To see some examples, it’s best to look at the examples of discontinuities on p.120 of Stewart.

The game is now to determine which functions are continuous where on their domains, and
especially if the basic functions are continuous on all of their domain.
Of course, outside of its domain a function is not defined, so the first of the three conditions
above fails, and the function is certainly not continuous. By the limit laws, all algebraic
functions are continuous on their domains; the two exceptions correspond to points outside
of the domains. For instance, the function x1 is continuous on its domain, which consists of
all real numbers except for 0. Since the function is not defined at 0, it doesn’t make much
sense to say that it’s continuous there or not.
The same is true for exponential and trigonometric functions, and their inverses (which we’ll
meet later): they are all continuous on all of their domain. But note that some of these
functions do not have all of R as their domain; for instance, integer multiples of π/2 are not
in the domain of tan x = sin x/ cos x, since there cos x is 0.
So in summary, all the elementary functions above are continuous on their domains. Of the
three conditions, only the first one ever fails for these functions, mostly because of division
by zero or even roots of negative numbers. When the second and third condition fail, it
is most often for piecewise defined functions (including |x|), which is why most continuity
questions in the book or the homework are about such functions.
The final thing to know is:

A composition of two continuous functions is continuous as well.

With that fact we can figure out if any function given by some sort of formula is continuous
or not, by breaking it up as a composition of simple functions, and using what we know
about those. The only hard part may be when piecewise defined functions are involved, in
which case we will have to analyze the left and right limits.
For example, the function
1
f (x) =
2 + sin x
is continuous on its domain (which is R: we never divide by zero because −1 ≤ sin x ≤ 1),
1
because it is a composition f = g ◦ h of the functions g(x) = 2+x and h(x) = sin x, a rational
and a trigonometric function, both of which we know to be continuous.
On the other hand, the function
 2
 x + 1 if x < 1
f (x) = 1 if x = 1
 2
x
if x ≥ 1.

is not continuous at x = 1. We check the three conditions for continuity:

• The function is clearly defined at x = 1.

• To find limx→1 f (x), we need to check the left and right limits (because it’s piecewise
defined):
2
lim f (x) = lim x2 + 1 = 2, lim f (x) = lim = 2.
x→1− x→1− x→1+ x→1+ x

Since the two one-sided limits are equal, the two-sided limit exists, and limx→1 f (x) = 2.

• But limx→1 f (x) = 2 6= f (1)!

5
So one of the conditions fails, and the function is not continuous at x = 1. At all other x,
however, the function is given by x2 + 1 or x2 , which we know to be continuous, except for x2
at x = 0, but that doesn’t happen. So this f (x) is continuous at all x except 1.

2.4 Examples of Limits


Right now we have two main methods for computing limits:

• We can interchange limits and continuous functions, until we end up with easy limits,
or we run into one of the exceptions. In practice, we do this by saying why the function
is continuous (if it is), and then just plugging in.

• We can compute the left and right limits, and see if they are equal. Especially useful
for piecewise defined functions.

In later chapters we will add to this two harder approaches, the Squeeze Theorem and Taylor
series.

Many of the limit questions on exams are however of a slightly different nature: they require
some algebraic trick before one of the methods above can be applied. We give some examples
of these.
The most common kind is the following type of limit calculation:

x2 − 1 (x − 1)(x + 1)
lim = lim = lim x + 1 = 2.
x→1 x − 1 x→1 x−1 x→1

Here the function is a rational function that seems to be undefined at the number that x
approaches, so we couldn’t just plug in. But after a cancellation there is no problem at all.
What makes it hard is that the factors that cancel each other may be hidden, and you wil
have to factor a polynomial first. Another example is

x2 + 5x + 4 (x + 4)(x + 1) x+1 3
lim 2
= lim = lim = .
x→−4 x + 3x − 4 x→−4 (x + 4)(x − 1) x→−4 x − 1 5

Another very common kind looks like this:


√ √ √
x−3 x−3 x+3 x−9 1 1
lim = lim ·√ = lim √ = lim √ = .
x→9 x − 9 x→9 x − 9 x + 3 x→9 (x − 9)( x + 3) x→9 x + 3 6

The trick is that when you see something√of the form a+ b in √ a rational√function, multiplying
top and bottom by the conjugate a − b will lead to (a + b)(a − b) = a2 − b2 , which
might clear up some cancellation.
But note that the second √ kind is actually
√ a lot like the first kind, because for instance we
could have used x − 9 = ( x − 3)( x + 3), although such a factorization may be harder to
spot.
Another algebraic step that you might have to take is clearing denominators, like in
1
a
− x1 x−a
x−a 1 1
lim = lim ax = lim = lim = 2.
x→a x − a x→a x − a x→a ax(x − a) x→a ax a

You might also like