The intuitive definition of a limit and the crucial fact that it ignores the value at the point, estimating limits from tables and graphs and the ways a table can mislead, the three ways a limit fails to exist, one-sided limits and the theorem connecting them to the two-sided limit, and infinite limits with vertical asymptotes.
Subject: Calculus I · 67 slides · symbolic lesson
Open the interactive version of this deck
Title
Calculus I · Section 2.2
Describing where a function is heading, whatever happens at the point itself
Objectives
Section 2.1 used limits informally, for slopes of tangent lines and areas. This lesson gives the word a meaning you can use: what a limit is, how to estimate one, and the three ways one can fail to exist.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 116 — learning objectives 2.2.1 to 2.2.7
In Section 2.1 you met two problems, the slope of a tangent line and the area under a curve, and in both the answer came from a process that got closer and closer to something without ever arriving. That process is a limit, and every idea in calculus from here on is built on it.
This lesson makes the word precise enough to use, while staying intuitive: the careful epsilon-delta definition waits until Section 2.5. You will learn what a limit claims, how to estimate one numerically and from a picture, and, just as important, how to recognise when there is no limit and describe exactly why.
Keep one sentence in mind throughout: a limit is about where a function is heading, not where it is. Almost every mistake in this chapter comes from forgetting that.
Warm-up
\[ f(x) = \frac{x^2 - 4}{x - 2} \]
Discussion prompt
Evaluate f at 2, at 1.99 and at 2.01. What goes wrong at 2, and what do the other two values suggest? Then factor the numerator and simplify.
Do the three evaluations before you reveal the answer. The first one should feel uncomfortable: at x equal to 2 you get zero over zero, which is not a number at all. That is not a trick question; it is the situation in which limits matter most.
Now look at the other two values. Just below 2 you get 3.99, just above you get 4.01. They are telling you something the formula at 2 refused to say: the function is heading toward 4.
The factoring explains it. Away from 2, the factor x minus 2 appears on top and bottom and cancels, leaving x plus 2. So the graph is just a straight line with a single point punched out of it. The rest of this lesson gives you the words and the notation for the height of that missing point.
Section
Part 1
Picture it
Figure (svg): Three small graphs side by side. Left: the line y equals x plus 2 with a hollow circle at the point 2, 4. Middle: a horizontal segment at height minus 1 left of x equals 2 and at height 1 right of it, with hollow circles at both ends. Right: the curve one over x minus 2 squared, shooting upward on both sides of a dashed line at x equals 2.
\[ f(x) = \frac{x^2-4}{x-2}, \qquad g(x) = \frac{|x-2|}{x-2}, \qquad h(x) = \frac{1}{(x-2)^2} \]
Each formula divides by zero at 2, so none of the three has a value there. But near 2 they could hardly be more different: one heads for a single height, one splits in two, one explodes. The limit is the language for describing that difference.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 117 — Figure 2.12
These are the three functions the book uses to open the section, and between them they show almost everything that can happen near a point. Each formula divides by zero at 2, so if you only say the function is undefined there, you have said the same thing about all three.
But look at the pictures. In the first, the graph is a line with a hole: nearby outputs crowd around a single height. In the second, the graph jumps: outputs to the left are all minus one and outputs to the right are all plus one. In the third, the outputs grow enormous on both sides.
The limit is the tool that tells these three apart. By the end of the lesson you will describe the first as having limit 4, the second as having different one-sided limits, and the third as having an infinite limit with a vertical asymptote.
Picture it
Figure (svg): A close-up of the line y equals x plus 2 between x equals 1 and 3. Four orange dots on the line approach x equals 2 from the left and four purple dots approach it from the right; all of them crowd toward a hollow circle at height 4, marked by a dashed horizontal line.
| x | 1.9 | 1.99 | 1.999 | 2.001 | 2.01 | 2.1 |
|---|---|---|---|---|---|---|
| f(x) | 3.9 | 3.99 | 3.999 | 4.001 | 4.01 | 4.1 |
Whichever side you come from, the outputs home in on 4. The point x equal to 2 itself is never used: every dot is a genuine input, and none of them is 2.
Follow the orange dots from the left: each one is a real input, 1.2, then 1.5, then 1.75, then 1.9, and each output is a little closer to 4. The purple dots do the same from the right. The table underneath takes the same idea further, down to a thousandth away from 2.
Two features of this picture are the whole content of the definition you are about to see. First, the dots come from both sides, and both families head for the same height. Second, none of the dots is at 2. The hollow circle is never visited; it is only approached.
If you remember this picture, you will remember what a limit is: the height that the outputs close in on as the inputs close in on a, from either side, without ever using a itself.
Concept
Limit (intuitive definition) — Let f be defined on an open interval containing a, except possibly at a itself. If all the values f(x) approach a real number L as x approaches a, with x never equal to a, then the limit of f(x) as x approaches a is L.
\[ \lim_{x\to a} f(x) = L \]
The book's shorter version: as x gets closer to a, f(x) gets closer to L and stays close. For the function of the warm-up:
\[ \lim_{x\to 2}\frac{x^2-4}{x-2} = 4 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 117 — Definition and equation 2.3
Read the definition slowly, because each phrase is doing work. The function must be defined on an open interval around a, so that there are inputs on both sides to approach with. It may or may not be defined at a itself; the definition explicitly allows a hole.
The word all matters too. It is not enough for some inputs near a to give outputs near L; every input close enough to a must. That is exactly what will go wrong in the oscillating examples later, where some nearby inputs give one height and others give another.
The book's short version, gets closer and stays close, is worth memorising. It is the seed of the precise definition in Section 2.5, where close will be measured with numbers. For now, the warm-up function is your model: its limit at 2 is 4, even though its value at 2 does not exist.
Notation
Annotate
On: \( \lim_{x\to a} f(x) = L \)
Step through the annotations and connect each to the picture of dots closing in on a hole. The arrow under lim describes the motion of the inputs; the f of x describes what you watch; the L is where the watching leads.
The most important annotation is the second one. The inputs approach a but are never equal to a. That single rule is why the definition can handle zero over zero, and why a function's value can be changed at one point without affecting its limit.
Notice also that L must be a real number. Later in the lesson you will write statements like the limit equals plus infinity, but those will be descriptions of a particular way of failing, not limits that exist. The equation on this slide, with a real L, is the only kind that says a limit exists.
Concept
Figure (svg): Three copies of the line y equals x plus 2 near x equals 2. First: a hollow circle at height 4 and nothing else. Second: a hollow circle at height 4 and a solid dot below it at height 1. Third: the line unbroken, with a solid dot at height 4.
\[ \lim_{x\to 2} f(x) = 4 \quad\text{in all three panels} \]
The definition says x approaches a but never equals it. So you may change, delete or invent the value at a, and the limit does not notice. It is a statement about the neighbourhood, not about the point.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 121 — the observation after Example 2.6
Compare the three panels. In each one, the line is the same away from 2, so the neighbours of 2 all behave identically. What differs is only the single point at x equal to 2: missing in the first, moved down to height 1 in the second, sitting on the line in the third.
The limit is 4 in all three, because the limit only asks about the neighbours. That is not a technicality; it is the reason limits are useful. Derivatives in Chapter 3 will be defined as limits of expressions that are always zero over zero at the point, and the limit is precisely the tool that sees past that.
When the value and the limit do agree, as in the third panel, the function is called continuous there. That is the subject of Section 2.4, and it is a special property, not the general rule.
Prediction
\[ g(x) = \begin{cases} x + 2 & x \ne 2 \\ 1 & x = 2 \end{cases} \]
Predict first
What is the limit of g(x) as x approaches 2?
Correct: 4
Why: Near 2, but not at 2, g agrees with x plus 2, whose outputs approach 4. The single moved point at x equal to 2 is invisible to the limit. The value is 1; the limit is 4; they are allowed to differ.
Commit to an option before revealing. This is the idea of the previous slide in its most direct form, and the option that says 1 is there because it is the most natural mistake.
The function agrees with x plus 2 everywhere except at 2 itself. Every input close to 2, but not equal to it, gives an output close to 4. The limit therefore is 4, and the single redefined point is invisible to it.
If you chose the option saying the limit does not exist, notice that a removed or moved point is not a break in the sense of a jump. The graph still heads for one height from both sides. It is only a jump when the two sides head for different heights, which you will see in Part 4.
Trap
A common first attempt, substituting straight in:
\[ \lim_{x\to 2}\frac{x^2-4}{x-2} = \frac{0}{0} \]
\[ \Longrightarrow \text{ no limit?} \]
Wrong. Zero over zero says the formula failed, not the limit.
The limit never uses the value at 2, so a formula that breaks at 2 says nothing about it. Cancel the factor, which is legal because x is not 2, and look at the neighbours:
\[ \frac{(x-2)(x+2)}{x-2} = x + 2 \]
\[ \lim_{x\to 2}(x+2) = 4 \]
This is the most common error of the whole chapter, and it is worth understanding why it is tempting. For polynomials and most everyday functions, substituting gives the right answer, so substitution feels like what a limit is. It is only a shortcut that works when the function is continuous.
Zero over zero is the signal that the shortcut has failed, not that the limit has. It means the formula cannot be evaluated at the point, which the definition never asked you to do. Cancel the common factor, which is legal because x is never equal to 2 in a limit, and the simplified expression heads straight to 4.
Section 2.3 will turn this cancel-then-substitute move into a standard technique. For now, the lesson is simply that an undefined value and a nonexistent limit are different things.
Concept
Figure (svg): Two panels. Left: the line y equals x with dashed guides from x equals 3 up to the line and across to y equals 3. Right: a horizontal line at height 2, with a dashed guide at x equals 3 meeting it at height 2.
\[ \text{(i) } \lim_{x\to a} x = a \qquad \text{(ii) } \lim_{x\to a} c = c \]
For the identity function the output is the input, so as x approaches a the output approaches a. For a constant function the output is c at every input, so there is nothing for it to approach except c. These two facts are the seeds of every algebraic limit law in Section 2.3.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 122 — Theorem 2.1, equations 2.4 and 2.5
These two limits look almost too simple to state, but everything algebraic in Section 2.3 is built from them. Once you know the limit of x and the limit of a constant, the limit laws let you assemble the limit of any polynomial from sums and products of these two pieces.
The left panel shows the first. On the line y equals x, the output is the input, so as the input slides toward a, the output slides toward a too. The dashed guides at x equal to 3 show the output arriving at height 3.
The right panel shows the second. A constant function never moves, so whatever x does, the output is c. There is no approach to speak of; the outputs are already at c. The book confirms it with a table in which every entry is c.
Fill the middle
Fill in the blanks
\lim_-3 x = 7, \qquad \lim_\pi 7 = ___, \qquad \lim____ x = ___
Why: The identity function heads to wherever x is heading: minus 3, then pi. The constant 7 ignores x entirely, so its limit at 5, or anywhere else, is 7.
Fill in the three blanks before checking. Each one is a direct use of Theorem 2.1, and the only thing to get right is which half of the theorem applies.
In the first and third, the function is the identity, so the limit is wherever x is heading: minus 3, and then pi. It makes no difference that pi is irrational; the theorem holds at every real number.
In the second, the function is the constant 7. The number 5 under the limit sign is a distraction: a constant does not care where x is heading. If you wrote 5, go back to the right-hand picture on the previous slide, where the height never changes.
Section
Part 2
Concept
| x below a | f(x) | x above a | f(x) |
|---|---|---|---|
| a − 0.1 | f(a − 0.1) | a + 0.1 | f(a + 0.1) |
| a − 0.01 | f(a − 0.01) | a + 0.01 | f(a + 0.01) |
| a − 0.001 | f(a − 0.001) | a + 0.001 | f(a + 0.001) |
| a − 0.0001 | f(a − 0.0001) | a + 0.0001 | f(a + 0.0001) |
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 118 — Problem-Solving Strategy and Table 2.1
This is the book's problem-solving strategy, and the most important word in it is both. You build one column of inputs creeping up toward a from below and a second column creeping down toward a from above, and you read down each column separately.
The inputs a tenth, a hundredth, a thousandth away from a are the usual choice, and they work nearly every time. The book is careful to say nearly: you will see an example shortly where exactly these inputs give a completely wrong impression.
The last step matters as much as the first. A table is evidence, not proof. You can only ever compute finitely many rows, and a limit is a claim about every input near a. Confirming with a graph helps; confirming with algebra, from Section 2.3 onward, settles it.
Estimation
\[ \lim_{x\to 0}\frac{\sin x}{x} \]
Predict first
Both the top and the bottom go to 0. With x in radians, what do you expect the ratio to approach?
Correct: 1
Why: Near 0, in radians, sin x and x are almost the same number, so their ratio is almost 1. Zero over zero is not an answer; it only says the formula cannot be evaluated at the point. The next slide's table shows the ratio approaching 1 from both sides.
Make a genuine guess before revealing. At zero, both sine x and x are zero, so the formula gives zero over zero, and it is natural to think that means the limit is zero, or does not exist.
Here is a way to guess well. Think of the unit circle: for a small angle x in radians, the arc of length x and the height sine x are almost the same length. So their ratio should be almost 1. The smaller the angle, the better the match.
This limit matters far beyond this lesson. It is the reason the derivative of sine is cosine, which you will prove in Chapter 3, and it only comes out as 1 because the angle is measured in radians. In degrees the same limit is pi over 180.
Worked example
Estimate the limit using a table of functional values.
\[ \lim_{x\to 0}\frac{\sin x}{x} \]
Note why substitution fails
Why: At zero the formula divides zero by zero, so a table is needed.
\[ \frac{\sin 0}{0} = \frac{0}{0} \]
Evaluate at inputs from the right
Why: Radian mode, keeping all the digits.
\[ x = 0.1: \; 0.998334166468, \quad x = 0.01: \; 0.999983333417 \]
Evaluate closer still
Why: Each step adds about two more nines.
\[ x = 0.001: \; 0.999999833333, \quad x = 0.0001: \; 0.999999998333 \]
Evaluate from the left
Why: Sine is odd and x is odd, so the ratio is even: the left column repeats the right.
\[ \frac{\sin(-x)}{-x} = \frac{-\sin x}{-x} = \frac{\sin x}{x} \]
Read the trend
Why: Both columns climb toward the same number.
\[ \lim_{x\to 0}\frac{\sin x}{x} = 1 \]
Figure (svg): The graph of sine x over x from x equals minus 8 to 8: a central hump rising to a hollow circle at the point 0, 1, with smaller ripples on each side that die away.
Check the pattern of the digits
Why: The gap below 1 shrinks like x squared over 6, which matches every row of the table.
\[ 1 - \frac{(0.1)^2}{6} = 0.998333, \qquad 1 - \frac{(0.01)^2}{6} = 0.9999833 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 118-119 — Example 2.4, Table 2.2, Figure 2.13
The table does exactly what the strategy asks: inputs a tenth, a hundredth, a thousandth and a ten-thousandth from 0, on both sides. On the right the outputs are 0.998, then 0.99998, then a string of nines that keeps growing. Each step toward 0 adds about two more nines.
The left column needs no new work. Sine is an odd function, and so is x itself, so replacing x by its negative changes the sign of both top and bottom, and the ratio is unchanged. The two columns are identical, and they approach the same number, 1.
The final check is a good habit to copy: look for a pattern in the digits. The gap below 1 is about x squared over 6, which is where two nines per row come from, since squaring a tenth gives a hundredth. That pattern fits every row, which makes the estimate much more convincing than just eyeballing the last line.
Worked example
Estimate the limit using a table of functional values.
\[ \lim_{x\to 4}\frac{\sqrt{x} - 2}{x - 4} \]
Check the point itself
Why: Both the top and the bottom vanish at 4.
\[ \frac{\sqrt4 - 2}{4 - 4} = \frac{0}{0} \]
Tabulate from the left
Why: Inputs below 4.
\[ 3.9: \; 0.251582, \quad 3.99: \; 0.250156, \quad 3.999: \; 0.250016 \]
Tabulate from the right
Why: Inputs above 4.
\[ 4.1: \; 0.248457, \quad 4.01: \; 0.249844, \quad 4.001: \; 0.249984 \]
Read the trend
Why: The left column decreases toward one quarter, the right column increases toward it.
\[ \lim_{x\to 4}\frac{\sqrt{x}-2}{x-4} = 0.25 \]
Figure (svg): The graph of root x minus 2, over x minus 4, for x from 0 to 9: a curve falling from height 0.5 near x equals 0, passing through a hollow circle at the point 4, 0.25, and flattening out to the right.
Check with algebra
Why: Factor x minus 4 as a difference of squares in root x; the factor that caused zero over zero cancels.
\[ \frac{\sqrt{x}-2}{(\sqrt{x}-2)(\sqrt{x}+2)} = \frac{1}{\sqrt{x}+2} \to \frac{1}{2+2} = \frac14 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 119-120 — Example 2.5, Table 2.3, Figure 2.14
Again substitution gives zero over zero, so you build the table. This time the two columns are not identical: from the left the values are a little above one quarter and decreasing, from the right a little below and increasing. They squeeze toward 0.25 from opposite sides, which you can see in the graph as the curve passing through the hole.
The check step goes further than the book does at this point, and it is worth seeing now. The denominator x minus 4 is a difference of squares in disguise, root x squared minus 2 squared. Factor it, and the troublesome factor cancels, leaving one over root x plus 2.
That simplified expression has no trouble at 4: it gives one over 4, exactly the value the table suggested. This is the pattern of the next section. The table told you what to expect; the algebra proves it.
Worked example
Estimate the limit with a table, then confirm it.
\[ \lim_{x\to 1}\frac{\frac{1}{x} - 1}{x - 1} \]
Tabulate from the left
Why: Inputs 0.9, 0.99, 0.999.
\[ -1.111111, \quad -1.010101, \quad -1.001001 \]
Tabulate from the right
Why: Inputs 1.1, 1.01, 1.001.
\[ -0.909091, \quad -0.990099, \quad -0.999001 \]
Read the trend
Why: Both columns close in on minus one, from opposite sides.
\[ \lim_{x\to 1}\frac{1/x - 1}{x-1} = -1 \]
Rewrite the numerator over x
Why: Put the numerator over a common denominator.
\[ \frac{1}{x} - 1 = \frac{1 - x}{x} \]
Check by simplifying
Why: Since one minus x is the negative of x minus 1, the quotient is minus one over x, which at 1 is minus 1.
\[ \frac{1-x}{x(x-1)} = -\frac{1}{x} \to -1 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 120 — Checkpoint 2.4
The table shows the left column coming down toward minus 1, with values like minus 1.11 and minus 1.01, and the right column coming up toward it, with values like minus 0.91 and minus 0.99. Both head for minus 1, so that is the estimate.
The confirmation needs one algebraic move, a common denominator in the numerator. One over x minus 1 becomes one minus x, all over x. Now the expression is a fraction over x minus 1, and one minus x is exactly the negative of x minus 1.
After cancelling, all that is left is minus one over x. At x equal to 1 that is minus 1, matching the table. Notice how cleanly the table rows fit this formula: at 0.9, minus one over 0.9 is minus 1.111, exactly the first entry.
Worked example
For the function g drawn below, find the limit as x approaches minus 1.
Figure (svg): A smooth rising curve crossing the x-axis at minus 3, levelling off near height 3, then rising again. At x equals minus 1 the curve has a hollow circle at height 3, and a separate solid dot sits above it at height 4.
Read the value
Why: The solid dot sits at height 4, so that is g at minus 1.
\[ g(-1) = 4 \]
Follow the curve from the left
Why: Trace the curve toward x equal to minus 1: its height approaches the hollow circle.
\[ \lim_{x\to -1^-} g(x) = 3 \]
Follow the curve from the right
Why: The same hollow circle, approached from the other side.
\[ \lim_{x\to -1^+} g(x) = 3 \]
Conclude
Why: Both sides agree; the value at the point is ignored.
\[ \lim_{x\to -1} g(x) = 3 \ne 4 = g(-1) \]
Check against the drawn curve
Why: Away from minus 1 the curve drawn here is 3 plus x squared times x plus 1, over 6; just either side of minus 1 it gives heights near 3, not 4.
\[ g(-1.1) \approx 2.980, \qquad g(-0.9) \approx 3.014 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 120-121 — Example 2.6, Figure 2.15
This example removes the formula entirely: the book gives only a picture. That is the point. A limit is a property of the graph near a point, and you can read it without knowing any algebra.
Two different marks sit at x equal to minus 1. The solid dot at height 4 is the value of g there. The hollow circle at height 3 is where the curve itself is heading from both sides. Trace the curve with your finger from the left and from the right, and your finger arrives at the hollow circle, not the dot.
So the limit is 3 and the value is 4, and the definition is perfectly happy with that. The check uses the formula for the curve drawn on this slide, which was built to match the book's figure: just either side of minus 1 it gives heights of about 2.98 and 3.01, close to 3 and nowhere near 4.
Prediction
Figure (svg): A parabola-shaped curve opening upward, passing through the origin and through x equals 4, with its lowest point at x equals 2, height minus 1, drawn as a hollow circle.
Predict first
Use the graph of h to find the limit of h(x) as x approaches 2, if possible.
Correct: −1
Why: From both sides the curve heads for the bottom of the bowl at height minus 1. The hollow circle means h has no value at 2, but a limit never needs one. The curve drawn is (x minus 2) squared over 4, minus 1, which is minus 1 at x equal to 2.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 121 — Checkpoint 2.5, Figure 2.16
Commit to an answer before revealing. The graph is a smooth bowl with its lowest point removed, so the only question is whether a missing point prevents a limit.
It does not. From the left the curve slides down toward the bottom of the bowl, and from the right it slides down toward the same place. Both approaches lead to height minus 1, so that is the limit.
If you chose the option saying the limit does not exist because h is undefined at 2, go back to the definition: the function only has to be defined near a, with the possible exception of a itself. A hole is the exception the definition was written to allow.
Tweak it
Parameter explorer
Exercises 35 to 37 tabulate sin(2x)/x and sin(3x)/x near 0. Slide a and watch where the curve crosses x = 0. What is the limit of sin(ax)/x as x approaches 0?
\[ y = \frac{\sin({a}x)}{x} \]
\[ \lim_{x\to 0}\frac{\sin(ax)}{x} = a \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 133 — Exercises 35 to 37
Start at a equal to 2 and look at where the curve crosses the vertical axis. That height is the value the function is heading for at 0, even though the formula itself is zero over zero there. Now slide a and watch the crossing height follow.
Exercises 35 and 36 ask you to find this numerically for a equal to 2 and 3, and Exercise 37 asks you to guess the pattern. The slider makes the pattern visible: the limit is a itself.
The reason connects to Example 2.4. Near 0, sine of a x is almost equal to a x, just as sine x was almost x. So the ratio is almost a x over x, which is a. You will prove this properly with the limit laws and a substitution in the next section.
Section
Part 3
Concept
For the limit to exist, the outputs must approach a single real number. There are exactly three ways that can go wrong near a point.
Figure (svg): Three small graphs. Left, a jump: a horizontal segment at minus 1 for negative x and at plus 1 for positive x. Middle, a blow-up: the curve one over x squared rising without bound on both sides of the y-axis. Right, an oscillation: the curve sine of one over x, swinging between minus 1 and 1 faster and faster near x equals 0.
| failure | what the outputs do | example at 0 |
|---|---|---|
| jump | each side settles, but at different heights | |x|/x |
| blow-up | grow without bound | 1/x² |
| oscillation | never settle at any height | sin(1/x) |
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 122 — The Existence of a Limit
Everything so far has been about limits that exist. The book now turns to the requirement that makes existence possible: the outputs must approach a single real number. Take that phrase apart and you get the three ways it can fail.
The left panel is a jump. Each side settles down, but at a different height, so there are two candidates instead of one. The middle panel is a blow-up: the outputs grow without bound, so there is no finite number at all. The right panel is an oscillation: the outputs never settle anywhere, however close you get.
Every limit that fails to exist in this course fails in one of these three ways, or a combination, such as a blow-up on one side and a jump on the other. The rest of the lesson takes each in turn and gives you notation to describe it precisely.
Worked example
Evaluate the limit using a table of values, if it exists.
\[ \lim_{x\to 0}\sin\left(\frac1x\right) \]
Tabulate from the right
Why: The outputs wander with no visible trend.
\[ 0.1: \; -0.544, \quad 0.01: \; -0.506, \quad 0.001: \; 0.827, \quad 0.0001: \; -0.306 \]
Tabulate from the left
Why: Sine is odd, so each entry is the negative of the one on the right.
\[ -0.1: \; 0.544, \quad -0.01: \; 0.506, \quad -0.001: \; -0.827 \]
Choose inputs systematically
Why: Pick x so that one over x lands on the odd multiples of pi over 2.
\[ x = \frac{2}{\pi}, \frac{2}{3\pi}, \frac{2}{5\pi}, \frac{2}{7\pi}, \ldots \to 0 \]
Evaluate at those inputs
Why: One over x is pi over 2, 3 pi over 2, and so on, where sine alternates.
\[ \sin\frac{\pi}{2}, \; \sin\frac{3\pi}{2}, \; \sin\frac{5\pi}{2}, \ldots = 1, -1, 1, -1, \ldots \]
Figure (svg): The graph of sine of one over x for x between minus 0.4 and 0.4, oscillating between minus 1 and 1 ever faster near 0. Yellow dots mark the points x equals 2 over pi, 2 over 3 pi, 2 over 5 pi and so on, alternating between the top line y equals 1 and the bottom line y equals minus 1.
Conclude
Why: Inputs arbitrarily close to 0 give outputs of both 1 and minus 1, so no single L is approached.
\[ \lim_{x\to 0}\sin\left(\frac1x\right) \text{ DNE} \]
Check the sequence formula
Why: At x equal to 2 over (2k plus 1) pi, sine of one over x is minus 1 to the k; for k equal to 0, 1, 2 that is 1, minus 1, 1. The book's last table row, at a millionth, should read 0.349993502 in the ninth place, not 0.349993504, a rounding slip that changes nothing.
\[ \sin\frac{(2k+1)\pi}{2} = (-1)^k, \qquad \sin(10^6) = -0.349993502 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 122-123 — Example 2.7, Table 2.5, Figure 2.17
The table for this function is a mess: minus 0.54, minus 0.51, then 0.83, then minus 0.31. No trend is visible, and the book's first reaction is that the limit appears not to exist. But appearing is not enough, so it takes a more systematic approach.
The trick is to choose the inputs rather than accept the usual ones. If one over x is pi over 2, sine is 1; if it is 3 pi over 2, sine is minus 1; and so on. Solving for x gives a sequence of inputs marching toward 0, whose outputs alternate between 1 and minus 1 forever. The yellow dots on the graph are exactly those inputs.
That settles it. However close to 0 you look, there are inputs giving 1 and inputs giving minus 1, so no single number is approached. This is the oscillation failure. The check confirms the alternating formula, and notes a tiny rounding slip in the book's last table entry, which does not affect anything.
Worked example
Set up the book's table, make a guess, then test it.
\[ \lim_{\theta\to 0}\sin\left(\frac{\pi}{\theta}\right) \]
Evaluate at the table's inputs
Why: At theta equal to 0.1, pi over theta is 10 pi, a whole multiple of pi.
\[ \sin(10\pi) = \sin(100\pi) = \sin(1000\pi) = \sin(10000\pi) = 0 \]
Note the tempting guess
Why: Every row, on both sides, reads 0.
\[ \text{table: } 0, 0, 0, 0 \;\Longrightarrow\; \text{guess } 0? \]
Choose different inputs
Why: Pick theta so that pi over theta is (4k plus 1) times pi over 2, where sine is 1.
\[ \theta = \frac{2}{4k+1}: \quad \frac{2}{5}, \frac{2}{9}, \frac{2}{13}, \ldots \to 0 \]
Evaluate there
Why: These inputs also approach 0, yet every output is 1.
\[ \sin\left(\frac{(4k+1)\pi}{2}\right) = 1 \]
Figure (svg): The graph of sine of pi over theta for theta between 0.05 and 0.55, oscillating between minus 1 and 1. Red dots sit on the axis at theta equals one half, one third, and so on down to one tenth, all at height 0. Green dots sit at the tops of the waves, at theta equals 2 over 5, 2 over 9, 2 over 13 and so on, all at height 1.
Conclude
Why: Outputs 0 and 1 both occur arbitrarily close to 0.
\[ \lim_{\theta\to 0}\sin\left(\frac{\pi}{\theta}\right) \text{ DNE} \]
Check with a single input
Why: Theta equal to 2 over 401 is less than 0.005, closer to 0 than the table's second row, and the output is exactly 1.
\[ \theta = \frac{2}{401} \approx 0.00499: \quad \sin(200.5\pi) = 1 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 134 — Exercise 44
This is the example promised on the table-strategy slide, and it is the most important warning in the lesson. Use the usual inputs, a tenth, a hundredth, and so on, and pi over theta becomes 10 pi, 100 pi, 1000 pi: whole multiples of pi, where sine is exactly zero. The table reads zero in every row, on both sides.
The natural guess is that the limit is zero, and it is completely wrong. The inputs in the table are special: they all sit exactly on the zeros of the curve, the red dots in the picture. Choose instead inputs that land on the peaks, like 2 over 5, 2 over 9, 2 over 13, and every output is 1. Those inputs approach 0 just as surely.
So the limit does not exist, and the table missed it because it only ever sampled one kind of point. The check makes it concrete: an input of about 0.005, closer to 0 than the table's second row, gives exactly 1.
Counterexample
Discussion prompt
Build a function whose table at x = ±0.1, ±0.01, ±0.001, ±0.0001 reads exactly 1 in every row, but whose limit at 0 does not exist. Then name one input near 0 that exposes it.
Try to build your own example before revealing. The idea is to copy Exercise 44 but aim at a different wrong answer. You need a function that is exactly 1 at every table input, but that also takes some other value arbitrarily close to 0.
Cosine of pi over x does it. At the table's inputs, pi over x is 10 pi, 100 pi, and so on, even multiples of pi, where cosine is 1. But at x equal to one over 1001, or one over any odd number, pi over x is an odd multiple of pi and cosine is minus 1.
The lesson is not that tables are useless. They are an excellent way to form a guess. The lesson is that a table samples finitely many points, while a limit makes a claim about all points near a, and no finite sample can guarantee that claim.
Section
Part 4
Concept
Figure (svg): The graph of absolute value of x minus 2, over x minus 2: height minus 1 to the left of x equals 2 and height 1 to the right, with hollow circles at both ends. An orange arrow below the left piece points right toward x equals 2; a purple arrow above the right piece points left toward x equals 2.
\[ g(x) = \frac{|x-2|}{x-2} = \begin{cases} -1 & x < 2 \\ 1 & x > 2 \end{cases} \]
Saying the limit at 2 does not exist is true but vague. Each side, taken alone, has a perfectly definite destination, and naming both describes the break exactly.
\[ \lim_{x\to 2^-} g(x) = -1, \qquad \lim_{x\to 2^+} g(x) = 1 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 123-124 — One-Sided Limits
This is the middle function from the start of the lesson. The absolute value makes the fraction minus 1 whenever x is less than 2 and plus 1 whenever x is greater. So the graph is two flat pieces with a gap at 2.
Saying the limit at 2 does not exist is correct, but it throws away information. Look at each side separately, following the arrows. From the left, the outputs are all minus 1, so they certainly approach minus 1. From the right, they are all 1. Each side has a perfectly good destination.
One-sided limits record those destinations. The small minus sign after the 2 means approach from the left, through smaller numbers; the small plus sign means approach from the right. Together the two statements describe the jump completely: where it starts, where it ends, and how big it is.
Concept
Limit from the left — Let f be defined on an open interval (c, a). If f(x) approaches L as x approaches a through values less than a, then L is the limit of f(x) as x approaches a from the left.
\[ \lim_{x\to a^-} f(x) = L \]
Limit from the right — Let f be defined on an open interval (a, c). If f(x) approaches L as x approaches a through values greater than a, then L is the limit from the right.
\[ \lim_{x\to a^+} f(x) = L \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 124 — Definition, equations 2.6 and 2.7
These are the same idea as the two-sided definition, with one change: the inputs are restricted to one side of a. For the limit from the left, you only look at inputs smaller than a, so the function only needs to be defined on an interval that ends at a from below.
That last point is why one-sided limits are genuinely useful and not just a curiosity. Some functions only exist on one side of a point. The square root of x, for instance, has a limit from the right at 0, but no left side at all to approach from. The two-sided limit at 0 does not even make sense, while the one-sided one does.
Example 2.12, at the end of the lesson, is a physical case: speeds can only approach the speed of light from below, so the natural limit there is one-sided.
Notation
Annotate
On: \( \lim_{x\to a^-} f(x) = L \qquad \lim_{x\to a^+} f(x) = M \)
Step through the annotations. The first catches a very common misreading: the minus sign after a does not mean a is negative, and it does not mean minus a. It is a direction, the side where the numbers are smaller than a.
The third annotation is the reason one-sided limits are more flexible than two-sided ones. Each only needs the function on its own side of a. That is why a left-hand limit can exist even when the function is not defined anywhere to the right.
The last annotation sets up the next theorem. The two one-sided limits are independent: they are computed separately and may come out different. Only when they agree do you get a two-sided limit, and that is precisely Theorem 2.2.
Step zero
\[ f(x) = \begin{cases} x + 1 & x < 2 \\ x^2 - 4 & x \ge 2 \end{cases} \]
Discussion prompt
Before computing anything: for the limit from the left at 2, and for the limit from the right at 2, which formula governs the inputs you will use? And does the value f(2) matter for either?
Write your answer before revealing. This thirty-second pause is the step that prevents most errors with piecewise functions.
The rule is that a one-sided limit only sees inputs from its own side. Approaching 2 from the left, every input is less than 2, so the first formula, x plus 1, is the only one in play. Approaching from the right, every input is greater than 2, so only the second formula matters.
The value at 2 is decided by the second formula because of the greater-than-or-equal condition, but that value is not used by either one-sided limit. It is tempting to think it belongs to the right-hand limit. It does not: the right-hand limit uses inputs above 2, never 2 itself, and just happens to agree with the value here.
Worked example
\[ f(x) = \begin{cases} x + 1 & x < 2 \\ x^2 - 4 & x \ge 2 \end{cases} \]
Tabulate from the left, using x plus 1
Why: Inputs 1.9, 1.99, 1.999.
\[ 2.9, \quad 2.99, \quad 2.999 \quad\Longrightarrow\quad \lim_{x\to 2^-} f(x) = 3 \]
Tabulate from the right, using the square minus 4
Why: Inputs 2.1, 2.01, 2.001.
\[ 0.41, \quad 0.0401, \quad 0.004001 \quad\Longrightarrow\quad \lim_{x\to 2^+} f(x) = 0 \]
Figure (svg): A line y equals x plus 1 rising from the lower left and ending in a hollow circle at the point 2, 3; then the parabola y equals x squared minus 4 starting at a solid dot at the point 2, 0 and rising steeply to the right.
Compare the two sides
Why: Different one-sided limits mean no single L.
\[ 3 \ne 0 \;\Longrightarrow\; \lim_{x\to 2} f(x) \text{ DNE} \]
Check by substituting into each piece
Why: Each piece is a polynomial with no trouble at 2, so its own value at 2 is where it is heading.
\[ (2) + 1 = 3, \qquad (2)^2 - 4 = 0 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 124-125 — Example 2.8, Table 2.6, Figure 2.18
With the pieces sorted out, each side is a routine table. From the left, x plus 1 gives 2.9, 2.99, 2.999, heading for 3. From the right, x squared minus 4 gives 0.41, 0.0401, 0.004001, heading for 0.
The picture shows the same thing: the left piece ends at a hollow circle at height 3 and the right piece starts at a solid dot at height 0. Three and zero are different numbers, so the two-sided limit does not exist. That is the jump failure, with a jump of size 3.
The check is a shortcut worth knowing. Each piece on its own is a polynomial, and a polynomial has no holes or jumps, so the limit of each piece is just its value at 2. Substituting 2 into each formula gives 3 and 0, matching the tables. Section 2.3 makes this shortcut official.
Concept
Relating one-sided and two-sided limits (Theorem 2.2) — Let f be defined on an open interval containing a, except possibly at a. Then the limit of f at a equals L exactly when both one-sided limits equal L.
\[ \lim_{x\to a} f(x) = L \iff \lim_{x\to a^-} f(x) = L \text{ and } \lim_{x\to a^+} f(x) = L \]
Read in one direction, the theorem gives you a limit when both sides agree. Read the other way, it proves non-existence: two different one-sided limits, or one side that does not exist, and the two-sided limit is gone.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 125 — Theorem 2.2
This theorem is what makes one-sided limits a working tool rather than just extra notation. It says the two-sided limit is completely determined by the two one-sided limits: it exists and equals L exactly when both sides exist and equal L.
Use it in both directions. Forwards, it lets you compute a two-sided limit by splitting it, which is often easier, especially for piecewise functions and absolute values. Backwards, it is the standard way to prove that a limit does not exist: show that the two sides give different numbers.
The if-and-only-if is important. A common mistake is to find that one side exists and conclude the limit exists. The theorem needs both sides, and needs them to agree.
Worked example
\[ \lim_{x\to 2}\frac{|x^2-4|}{x-2} \]
Factor inside the absolute value
Why: The absolute value of a product is the product of absolute values.
\[ |x^2 - 4| = |x-2|\,|x+2| \]
Simplify from the left
Why: Near 2 and below it, x minus 2 is negative and x plus 2 is positive.
\[ x < 2: \quad \frac{-(x-2)(x+2)}{x-2} = -(x+2) \to -4 \]
Simplify from the right
Why: Now x minus 2 is positive, so the absolute value changes nothing.
\[ x > 2: \quad \frac{(x-2)(x+2)}{x-2} = x+2 \to 4 \]
Figure (svg): The graph of absolute value of x squared minus 4, over x minus 2. Left of x equals 2 it is a V shape pointing up with its tip on the axis at x equals minus 2, falling to a hollow circle at the point 2, minus 4. Right of 2 it is the line y equals x plus 2, starting from a hollow circle at the point 2, 4.
Apply Theorem 2.2
Why: The sides disagree.
\[ \lim_{x\to 2^-} = -4 \ne 4 = \lim_{x\to 2^+} \;\Longrightarrow\; \text{DNE} \]
Check with a table
Why: The book asks for tables; they match the two formulas.
\[ f(1.99) = -3.99, \quad f(1.999) = -3.999, \quad f(2.01) = 4.01, \quad f(2.001) = 4.001 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 123 and 125 — Checkpoints 2.6 and 2.7
Checkpoint 2.6 asks for the two-sided limit and Checkpoint 2.7 for the two one-sided limits, so they belong together. The absolute value is the whole difficulty, and the way to handle an absolute value is to split by sign.
Factor first: x squared minus 4 is x minus 2 times x plus 2, and the absolute value of a product is the product of absolute values. Near 2, x plus 2 is about 4, positive, so its absolute value changes nothing. Only the factor x minus 2 changes sign, and that decides everything.
On the left, the absolute value of x minus 2 is its negative, so after cancelling you get minus x plus 2, heading for minus 4. On the right it is just x minus 2, and you get x plus 2, heading for 4. Different sides, so no two-sided limit. The table check matches the two simplified formulas exactly.
Worked example
\[ \lim_{x\to 0}\frac{5}{1 - e^{1/x}} \]
From the left, follow the exponent
Why: For small negative x, one over x is a large negative number.
\[ x \to 0^-: \quad \frac{1}{x} \to -\infty, \quad e^{1/x} \to 0 \]
Evaluate the left-hand limit
Why: The denominator approaches 1.
\[ \lim_{x\to 0^-}\frac{5}{1 - e^{1/x}} = \frac{5}{1 - 0} = 5 \]
From the right, follow the exponent
Why: For small positive x, one over x is a large positive number.
\[ x \to 0^+: \quad \frac{1}{x} \to \infty, \quad e^{1/x} \to \infty \]
Evaluate the right-hand limit
Why: The denominator grows without bound in size, so the fraction shrinks to 0.
\[ \lim_{x\to 0^+}\frac{5}{1 - e^{1/x}} = 0 \]
Figure (svg): The graph of 5 over 1 minus e to the one over x, for x from minus 1.5 to 1.5. Left of 0 the curve falls toward a hollow circle at height 5; right of 0 it rises from a hollow circle at height 0 and heads down to negative values as x grows.
Check with the table's first rows
Why: At plus and minus 0.1 the values already sit near 5 and near 0, and the sides disagree, so the two-sided limit does not exist.
\[ f(-0.1) \approx 5.000227, \qquad f(0.1) \approx -0.000227 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 133 — Exercise 40
This exercise shows that jumps are not only made by piecewise formulas. A single, smooth-looking expression can behave completely differently on the two sides of a point, because a piece of it, here one over x, does.
Follow the exponent. For small negative x, one over x is a huge negative number, and e raised to a huge negative power is almost zero, so the denominator is almost 1 and the fraction is almost 5. For small positive x, one over x is huge and positive, e to that power is enormous, and 5 divided by an enormous negative number is almost zero.
So the left-hand limit is 5 and the right-hand limit is 0, and the two-sided limit does not exist. The graph shows the two hollow circles, and the table values at plus and minus a tenth already sit within a thousandth of those limits.
Sorting
Sort into buckets
Sort each situation at x = a by whether the two-sided limit exists.
For each card, ask one question: do both one-sided limits exist and agree? Ignore any information about the value at a, which some cards include on purpose as a distraction.
Three cards pass. Left and right both 3, with a value of 5, is fine: the value does not matter. Left and right both 7 with no value at all is fine too. And the redefined line from earlier has limit 4, whatever it does at 2.
Three cards fail, each for a different reason. Sides of 3 and 0 disagree, which is a jump. Outputs swinging forever is an oscillation. And a side that is not even defined near a means the two-sided limit has nothing to work with; only the left-hand limit makes sense there.
Trap
For the piecewise function of Example 2.8:
\[ f(2) = 2^2 - 4 = 0 \]
\[ \Longrightarrow\; \lim_{x\to 2} f(x) = 0 ? \]
Wrong. The filled dot is only the value.
The value comes from the right-hand formula, so it can only speak for the right side. The left side arrives at 3. Check both sides every time a formula changes at a:
\[ \lim_{x\to 2^-} f(x) = 3 \]
\[ \lim_{x\to 2^+} f(x) = 0 \]
\[ 3 \ne 0 \;\Longrightarrow\; \text{DNE} \]
On a graph of a piecewise function, the filled dot is the most eye-catching thing at the break, and it is very tempting to read the limit from it. For Example 2.8 that gives 0, and it is wrong.
The filled dot is the value, and the value comes from whichever formula owns the endpoint, here the right-hand one. At best it can tell you about the right side. The left side arrives at 3, a different height, and so the two-sided limit does not exist.
The habit that prevents this is simple: whenever a formula changes at a, compute both one-sided limits explicitly, each from its own piece, before saying anything about the two-sided limit. Look at the value last, if at all.
Real world
\[ P(w) = \begin{cases} 1.00 & 0 < w \le 1 \\ 1.20 & 1 < w \le 2 \end{cases} \quad \text{(dollars, weight in ounces)} \]
Discussion prompt
A postal rule charges one dollar up to one ounce and one dollar twenty above it. Find both one-sided limits of the price at one ounce, decide whether the limit exists, and say what the difference means for a letter.
Write your answer before revealing. Step functions like this one are everywhere in pricing: postage, tax brackets, parking fees, phone plans. They are the natural home of the jump failure.
Approaching one ounce from below, every weight costs one dollar, so the left-hand limit is 1.00. Approaching from above, every weight costs one dollar twenty, so the right-hand limit is 1.20. The sides disagree, so there is no limit at one ounce.
The meaning is practical: a tiny change in weight near one ounce can change the price by twenty cents. That is what a jump is, a place where the output is not controlled by keeping the input close. In Section 2.4 you will call such a point a jump discontinuity.
Section
Part 5
Concept
Figure (svg): The curve one over x minus 2 squared near x equals 2, with a dashed horizontal line at height 25. Between x equals 1.8 and 2.2 the curve is above the line, and that region is shaded.
| x | 1.9 | 1.99 | 1.999 | 2.001 | 2.01 | 2.1 |
|---|---|---|---|---|---|---|
| h(x) = 1/(x − 2)² | 100 | 10,000 | 1,000,000 | 1,000,000 | 10,000 | 100 |
As x approaches 2, the values of h become larger than any number you name: past 25 within 0.2 of 2, past a million within 0.001. The book describes this by writing:
\[ \lim_{x\to 2}\frac{1}{(x-2)^2} = +\infty \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 125 — Infinite Limits
This is the third function from the start of the lesson. The table shows the outputs at a tenth from 2 are 100, at a hundredth they are ten thousand, at a thousandth a million. Each step closer multiplies the output by a hundred.
The picture shows what grow without bound means precisely. Draw a horizontal line at any height, here 25. Close enough to 2, within 0.2 in this case, every point of the graph is above that line. Choose a higher line, a million, and you need to be within a thousandth, but you can still do it. No height is safe.
The notation for this is the limit equals plus infinity. Be careful what it does and does not say. It says the outputs eventually exceed every bound. It does not say the function reaches infinity, which is not a number, and it does not say the limit exists.
Concept
| notation | meaning | equation |
|---|---|---|
| lim x→a⁻ f(x) = +∞ | increases without bound as x approaches a from the left | (2.8) |
| lim x→a⁻ f(x) = −∞ | decreases without bound from the left | (2.9) |
| lim x→a⁺ f(x) = +∞ | increases without bound from the right | (2.10) |
| lim x→a⁺ f(x) = −∞ | decreases without bound from the right | (2.11) |
| lim x→a f(x) = +∞ | increases without bound from both sides | (2.12) |
| lim x→a f(x) = −∞ | decreases without bound from both sides | (2.13) |
\[ \lim_{x\to a} f(x) = +\infty \iff \lim_{x\to a^-} f(x) = +\infty \text{ and } \lim_{x\to a^+} f(x) = +\infty \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 126 — Definition, equations 2.8 to 2.13
The book lists six separate definitions, and they look like a lot to learn. They are really one idea with two choices: which direction the outputs go, up without bound or down without bound, and which side the inputs come from, left, right or both.
The table lines them up with their equation numbers so you can find them in the book. The two-sided versions at the bottom need both sides to go the same way, which the displayed equivalence at the bottom of the slide makes explicit. It is Theorem 2.2 again, with infinity in place of L.
That equivalence will matter on the next few slides. When the two sides blow up in opposite directions, as they do for one over x, there is no two-sided infinite limit, and you must describe each side separately.
Intuition
When you write that a limit equals plus infinity, you are not saying the limit exists and happens to be a very large number. Infinity is not a real number, so the limit does not exist. The notation records HOW it fails to exist: by growing without bound in one direction.
\[ \lim_{x\to 2}\frac{1}{(x-2)^2} = +\infty \quad \text{is more informative than} \quad \text{DNE} \]
So when a limit is infinite, the book's convention is to write the infinity rather than DNE. Save DNE for the jump and the oscillation, and for sides that blow up in opposite directions.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 126 — the remark after the definition
This slide is about reading the notation honestly. When you write that the limit equals plus infinity, the equals sign is being used loosely. The limit, strictly, does not exist, because a limit must be a real number and infinity is not one.
So why not just write DNE? Because the infinity says much more. DNE could mean a jump, an oscillation, or a blow-up. Plus infinity tells the reader exactly which failure it is and which way the function goes. The book's convention is therefore to write the infinity whenever it applies.
Where it does not apply is when the two sides blow up in opposite directions. Then there is no single infinity to write for the two-sided limit, and DNE is the correct report, with the two one-sided infinities given separately.
Worked example
Evaluate each limit, if possible.
\[ \text{(a) } \lim_{x\to 0^-}\frac1x \qquad \text{(b) } \lim_{x\to 0^+}\frac1x \qquad \text{(c) } \lim_{x\to 0}\frac1x \]
Tabulate from the left
Why: Each step toward 0 multiplies the output by ten, with a minus sign.
\[ -0.1: \; -10, \quad -0.01: \; -100, \quad -0.001: \; -1000 \]
Conclude (a)
Why: The values decrease without bound.
\[ \lim_{x\to 0^-}\frac1x = -\infty \]
Tabulate from the right and conclude (b)
Why: Now the outputs are positive and grow without bound.
\[ 0.1: \; 10, \quad 0.01: \; 100, \quad 0.001: \; 1000 \;\Longrightarrow\; \lim_{x\to 0^+}\frac1x = +\infty \]
Figure (svg): The graph of one over x: a branch in the lower left falling to minus infinity as x approaches 0 from the left, and a branch in the upper right rising to plus infinity as x approaches 0 from the right.
Conclude (c)
Why: The sides run off in opposite directions, so there is no single description.
\[ \lim_{x\to 0}\frac1x \text{ DNE} \]
Check the sign of each side
Why: A negative input gives a negative reciprocal and a positive one a positive reciprocal; at a millionth the outputs are minus and plus a million.
\[ \frac{1}{-0.000001} = -1{,}000{,}000, \qquad \frac{1}{0.000001} = 1{,}000{,}000 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 126-127 — Example 2.9, Table 2.7, Figure 2.19
The table here is striking in its simplicity: every step toward 0 multiplies the output by ten. From the left the outputs are negative and grow in size, minus 10, minus 100, minus 1000, so the left-hand limit is minus infinity. From the right they are positive and grow: plus infinity.
Part (c) is where you have to be careful. Both sides are infinite, but in opposite directions, so there is no single description of the two-sided behaviour. The limit does not exist, and you write DNE rather than an infinity.
The graph shows the two branches running off in opposite directions along the vertical axis. The check is really just the sign rule: a negative number has a negative reciprocal. That is the whole reason the two sides disagree, and it will be the key to Theorem 2.3.
Worked example
\[ \text{(a) } \lim_{x\to 0^-}\frac{1}{x^2} \qquad \text{(b) } \lim_{x\to 0^+}\frac{1}{x^2} \qquad \text{(c) } \lim_{x\to 0}\frac{1}{x^2} \]
Tabulate from the left
Why: Squaring removes the minus sign.
\[ -0.1: \; 100, \quad -0.01: \; 10{,}000, \quad -0.001: \; 1{,}000{,}000 \]
Conclude (a)
Why: The values increase without bound.
\[ \lim_{x\to 0^-}\frac{1}{x^2} = +\infty \]
Tabulate from the right and conclude (b)
Why: The same outputs as on the left.
\[ 0.1: \; 100, \quad 0.01: \; 10{,}000 \;\Longrightarrow\; \lim_{x\to 0^+}\frac{1}{x^2} = +\infty \]
Conclude (c)
Why: Both sides grow without bound in the same direction.
\[ \lim_{x\to 0}\frac{1}{x^2} = +\infty \]
Check the symmetry
Why: A number and its negative have the same square, so the two columns must match row for row.
\[ \frac{1}{(-x)^2} = \frac{1}{x^2} \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 127-128 — Checkpoint 2.8
This is the companion to Example 2.9, and the only change is the square. Squaring destroys the sign, so the left column is identical to the right column: 100, ten thousand, a million on both sides.
Both one-sided limits are plus infinity, and because they agree, the two-sided limit is also plus infinity. Compare this with one over x, where the sides disagreed. The difference between the two examples is exactly whether the power in the denominator is odd or even.
The check states the symmetry that the table displays. A number and its negative have the same square, so the function takes the same value at x and at minus x. Its graph is mirror-symmetric about the vertical axis, and both branches shoot up.
Error analysis
Annotate
On: \( \lim_{x\to 0}\frac1x = \infty \)
\[ \lim_{x\to 0^-}\frac1x = -\infty, \quad \lim_{x\to 0^+}\frac1x = +\infty, \quad \lim_{x\to 0}\frac1x \text{ DNE} \]
Look at the statement and decide what is wrong before stepping through the annotations. The tempting thought is that one over x obviously blows up at 0, so the limit is infinity.
The error is treating the two sides as one. On the left the function heads down without bound, on the right up without bound. An infinite limit, just like a finite one, needs both sides to agree, and here they point in opposite directions.
The corrected statement at the bottom is what a complete answer looks like: each one-sided limit written with its own infinity, then the conclusion that the two-sided limit does not exist. Writing only the plus infinity loses half the picture and states something false.
Concept
Figure (svg): Two panels with a dashed vertical asymptote at x equals 1. Left, n equals 3: the curve falls to minus infinity on the left of the line and rises to plus infinity on the right. Right, n equals 2: the curve rises to plus infinity on both sides.
Near a, the base x minus a is a tiny number, negative on the left and positive on the right. Raising it to a positive integer power keeps it tiny, so its reciprocal is huge; only the sign is in question.
\[ n \text{ even}: \quad \lim_{x\to a}\frac{1}{(x-a)^n} = +\infty \]
\[ n \text{ odd}: \quad \lim_{x\to a^+}\frac{1}{(x-a)^n} = +\infty, \quad \lim_{x\to a^-}\frac{1}{(x-a)^n} = -\infty \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 128 — Theorem 2.3, Figure 2.20
This theorem packages Example 2.9 and Checkpoint 2.8 into a rule for a whole family. Near a, the base x minus a is a tiny number, and its reciprocal raised to any positive power is huge. The only question is the sign, and that depends on whether the power is even or odd.
An even power always gives a positive result, so both sides go to plus infinity, as in the right-hand panel. An odd power keeps the sign of the base: negative on the left, where x is less than a, and positive on the right. That gives minus infinity on the left and plus infinity on the right, as in the left-hand panel.
In both cases the line x equals a is a vertical asymptote. You will use this theorem constantly, because near a zero of a denominator, most rational functions behave like one of these two pictures.
Tweak it
Parameter explorer
The curve is 1/(x − 1)ⁿ. Slide n through whole numbers. Which values of n send the left branch down, and which send it up?
\[ y = \frac{1}{(x-1)^{{n}}} \]
Start at n equal to 1, which is just one over x shifted to the right by 1. The left branch goes down and the right branch goes up. Now step n to 2 and watch the left branch flip upward.
Keep stepping. Odd values of n all look like the first picture, and even values all look like the second. The branches also hug the asymptote more tightly and flatten out faster away from it as n grows, because higher powers make small numbers smaller and large numbers larger.
Notice what never changes: the right branch always goes up. To the right of the asymptote the base x minus 1 is positive, and a positive number raised to any power is positive. Only the left side, where the base is negative, depends on the parity of n.
Concept
Vertical asymptote — The line x equals a is a vertical asymptote of f if at least one of the one-sided limits at a, or the two-sided limit, is plus infinity or minus infinity.
\[ \lim_{x\to a^-} f(x) = \pm\infty \quad\text{or}\quad \lim_{x\to a^+} f(x) = \pm\infty \quad\text{or}\quad \lim_{x\to a} f(x) = \pm\infty \]
One side is enough. Near an asymptote the points of the graph hug the vertical line, getting closer to it as they climb or fall.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 128-129 — Definition of a vertical asymptote
The definition is short, and the key word is or. The line x equals a is a vertical asymptote if any one of the listed limits is plus or minus infinity. It does not need both sides, and it does not need the two sides to agree.
That matters for functions like one over x, where the two-sided limit does not exist. Each one-sided limit is infinite, so x equals 0 is still an asymptote. It also matters for one-sided functions, like the Einstein example, where only the left side exists at all.
Geometrically, near an asymptote the graph climbs or falls alongside the vertical line and gets closer and closer to it. A graph can never cross its vertical asymptote at the asymptote's own x value, since the function would have to be defined and finite there.
Worked example
Evaluate each limit with Theorem 2.3 and identify any vertical asymptotes.
\[ f(x) = \frac{1}{(x+3)^4} \]
Match the theorem's form
Why: Write x plus 3 as x minus a.
\[ x + 3 = x - (-3): \quad a = -3, \; n = 4 \]
Apply the even case
Why: The power 4 is even, so both sides go up.
\[ \lim_{x\to -3^-} f(x) = +\infty, \qquad \lim_{x\to -3^+} f(x) = +\infty \]
State the two-sided limit
Why: Both sides agree in direction.
\[ \lim_{x\to -3} f(x) = +\infty \]
Figure (svg): The graph of one over x plus 3 to the fourth, for x from minus 6 to 0: two branches hugging the axis far from x equals minus 3 and both shooting up to plus infinity beside a dashed vertical line at x equals minus 3.
Name the asymptote
Why: An infinite limit at minus 3.
\[ x = -3 \text{ is a vertical asymptote} \]
Check with a number on each side
Why: A hundredth either side of minus 3 gives the same huge positive value.
\[ f(-3.01) = \frac{1}{(-0.01)^4} = 10^{8} = f(-2.99) \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 129 — Example 2.10
The first step is matching the formula to the form in Theorem 2.3, which is written with x minus a. Here the base is x plus 3, which is x minus minus 3, so a is minus 3 and the power n is 4.
Four is even, so the theorem gives plus infinity from both sides, and therefore plus infinity for the two-sided limit. The graph shows both branches shooting up beside the dashed line at minus 3, which is the vertical asymptote.
The check picks one point a hundredth away on each side. The base is minus a hundredth on the left and plus a hundredth on the right, and the fourth power of either is one hundred-millionth, so both give a hundred million. The even power has erased the sign, just as the theorem said.
Worked example
\[ f(x) = \frac{1}{(x-2)^3} \]
Match the form
Why: Here a is 2 and n is 3, an odd power.
\[ a = 2, \quad n = 3 \]
Apply the odd case
Why: The sign of the base survives the cube.
\[ \lim_{x\to 2^-} f(x) = -\infty, \qquad \lim_{x\to 2^+} f(x) = +\infty \]
State the two-sided limit
Why: Opposite directions.
\[ \lim_{x\to 2} f(x) \text{ DNE} \]
Name the asymptote
Why: One infinite side is already enough.
\[ x = 2 \text{ is a vertical asymptote} \]
Check with numbers
Why: A tenth either side of 2 gives minus and plus a thousand.
\[ f(1.9) = \frac{1}{(-0.1)^3} = -1000, \qquad f(2.1) = \frac{1}{(0.1)^3} = 1000 \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 129 — Checkpoint 2.9
This is Example 2.10 with an odd power, and the conclusions change accordingly. With a equal to 2 and n equal to 3, the cube keeps the sign of x minus 2.
On the left the base is negative, so the function goes to minus infinity; on the right it is positive, so plus infinity. The two-sided limit does not exist, because the sides disagree in direction. But the line x equals 2 is still a vertical asymptote: the definition only needs one infinite side, and here there are two.
The check with numbers makes the sign concrete. A tenth below 2, the base is minus a tenth, whose cube is minus a thousandth, and the reciprocal is minus a thousand. A tenth above, everything is positive.
Worked example
The denominator vanishes at 2 and at minus 2. Classify both points.
\[ f(x) = \frac{x-2}{x^2-4} \]
Factor the denominator
Why: A difference of squares.
\[ \frac{x-2}{(x-2)(x+2)} \]
Cancel for x not equal to 2
Why: Legal near 2, since the limit never uses x equal to 2.
\[ f(x) = \frac{1}{x+2} \quad (x \ne 2) \]
At 2: a finite limit, so a hole
Why: The simplified formula is harmless there.
\[ \lim_{x\to 2} f(x) = \frac{1}{2+2} = \frac14 \]
At minus 2: Theorem 2.3 with n equal to 1
Why: The zero survives cancelling, and the power is odd.
\[ \lim_{x\to -2^-} f(x) = -\infty, \quad \lim_{x\to -2^+} f(x) = +\infty \]
Figure (svg): The graph of x minus 2 over x squared minus 4. It has a dashed vertical asymptote at x equals minus 2, where the branches run to minus infinity on the left and plus infinity on the right, and a hollow circle on the right branch at the point 2, one quarter.
Check with numbers near each point
Why: Near 2 the values sit near one quarter; near minus 2 they are huge with opposite signs.
\[ f(2.001) \approx 0.249938, \qquad f(-2.01) = -100, \quad f(-1.99) = 100 \]
Stewart, Calculus: Early Transcendentals 8e, §2.2 The Limit of a Function §2.2, pp. 83-94 — vertical asymptotes and removable holes
This example is your own, taken from the Stewart treatment of the same section, and it answers a question the book leaves implicit: when the denominator of a rational function is zero, how do you tell a hole from an asymptote?
Factor and cancel. The denominator is x minus 2 times x plus 2. The factor x minus 2 cancels with the numerator, so near 2 the function is just one over x plus 2, which heads calmly to one quarter. That is a hole. The factor x plus 2 does not cancel, so near minus 2 the function behaves like one over x plus 2 with an odd power, one: a vertical asymptote with opposite infinities.
The graph shows both features on the same curve. The general rule: a zero of the denominator that survives cancelling gives an asymptote; one that cancels away gives a hole.
Trap
Spotting a zero in the denominator:
\[ \frac{x^2-4}{x-2}: \; x - 2 = 0 \text{ at } 2 \]
\[ \Longrightarrow \text{ asymptote } x = 2 ? \]
Wrong. The numerator vanishes there too.
Zero over zero is undecided. Cancel first; only a zero that survives in the denominator, over a nonzero numerator, gives an infinite limit.
\[ \frac{x^2-4}{x-2} = x + 2 \to 4 \]
A hole at the point 2, 4, and no asymptote at all.
This shortcut is tempting because it works so often: find where the denominator is zero, and draw an asymptote there. It fails exactly when the numerator is also zero at that point, which is the zero-over-zero situation this whole lesson began with.
Zero over zero is undecided. It can turn out to be any finite number, as it does here, where cancelling leaves x plus 2 heading to 4. It can also still blow up, if the denominator's zero is of higher order than the numerator's. The only way to know is to cancel and look at what survives.
So check the numerator too. A zero in the denominator with a nonzero numerator gives an asymptote, guaranteed. A zero in both needs cancelling before you decide.
Discrimination
Sort into buckets
At the named point, does each function have a vertical asymptote or a removable hole?
Sort each card using the rule from the last two slides: if the numerator is not zero at the point while the denominator is, it is an asymptote; if both are zero, cancel or use a known limit before deciding.
The asymptotes: x plus 1 over x squared minus 4 at 2 has numerator 3, not zero. The cancelled form of x minus 2 over x squared minus 4 is one over x plus 2, which still blows up at minus 2. And the fourth-power example has a numerator of 1, never zero.
The holes are all zero over zero cases that settle to a finite number: 4 after cancelling for the first, 1 for sine x over x from Example 2.4, and one quarter for the square-root example from Example 2.5. Every one of them has a finite limit, so none is an asymptote.
Worked example
Use the graph to find the one-sided limits, the limit and the value at each marked point.
Figure (svg): A graph with four features. At x equals minus 4 it crosses the axis smoothly. At x equals minus 2 there is a hollow circle at height 3 with the curve arriving from both sides. At x equals 1 the curve ends in a solid dot at height 6 and restarts from a hollow circle at height 3. At x equals 3 both neighbouring branches plunge down beside a dashed vertical line.
At minus 4, an ordinary point
Why: The curve passes smoothly through the axis.
\[ \lim_{x\to -4^-} f = \lim_{x\to -4^+} f = \lim_{x\to -4} f = 0 = f(-4) \]
At minus 2, a hole
Why: Both sides reach height 3; there is no dot.
\[ \lim_{x\to -2^-} f = \lim_{x\to -2^+} f = \lim_{x\to -2} f = 3, \quad f(-2) \text{ undefined} \]
At 1, a jump
Why: The left piece ends at the filled dot at 6; the right piece starts at the hollow circle at 3.
\[ \lim_{x\to 1^-} f = 6, \quad \lim_{x\to 1^+} f = 3, \quad \lim_{x\to 1} f \text{ DNE}, \quad f(1) = 6 \]
At 3, an asymptote
Why: Both neighbouring branches plunge.
\[ \lim_{x\to 3^-} f = \lim_{x\to 3^+} f = \lim_{x\to 3} f = -\infty, \quad f(3) \text{ undefined} \]
Check each row with Theorem 2.2
Why: Wherever the two sides agree the two-sided entry copies them; only at 1, where 6 and 3 differ, is it DNE.
\[ 6 \ne 3 \;\Longrightarrow\; \text{DNE at } 1 \text{ only} \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 129-130 — Example 2.11, Figure 2.21
This example uses everything in the lesson at once. At each of four points, the book asks for four things: the left-hand limit, the right-hand limit, the two-sided limit, and the value. The graph drawn here was built to match the book's figure point for point.
At minus 4 nothing unusual happens: the curve passes through, and all four answers are 0. At minus 2 there is a hole, so the three limits are all 3 but the value is undefined. At 1 there is a jump with the filled dot on the left piece, so the left limit and the value are both 6, the right limit is 3, and the two-sided limit does not exist. At 3 both sides plunge, so all three limits are minus infinity and there is an asymptote.
The check runs Theorem 2.2 down the rows. Only at 1 do the two sides disagree, and that is the only DNE. That kind of sweep is a good habit on any multi-part graph question.
Prediction
Figure (svg): A curve rising from the lower left to a hollow circle at the point 1, 3, and a second curve starting at a solid dot at the point 1, 2 and falling gently through the x-axis at x equals 4.
Predict first
For the function drawn, what is the limit of f(x) as x approaches 1?
Correct: It does not exist
Why: From the left the curve rises to the hollow circle at 3; from the right it starts from the filled dot at 2. The one-sided limits are 3 and 2, which differ, so by Theorem 2.2 the two-sided limit does not exist. The filled dot gives the value, f(1) = 2, which is a separate fact.
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, p. 130 — Checkpoint 2.10
Commit to an answer before revealing. The graph has one feature at x equal to 1: the left curve ends at a hollow circle, and the right curve begins at a filled dot at a different height.
Follow each side separately. From the left, the curve rises to the hollow circle at height 3, so the left-hand limit is 3. From the right, the curve starts at the filled dot at height 2 and the nearby points are close to 2, so the right-hand limit is 2. Three and two differ, so the two-sided limit does not exist.
The option saying 2 is the filled-dot trap from a few slides ago. The filled dot is the value, and it belongs to the right-hand piece only.
Worked example
Einstein's equation gives the mass of an object moving at speed v, where m zero is its rest mass and c the speed of light. What bound does it put on speed?
\[ m = \frac{m_0}{\sqrt{1 - \frac{v^2}{c^2}}} \]
Divide by the rest mass
Why: Write the ratio of masses in terms of the speed ratio r, which is v over c.
\[ \frac{m}{m_0} = \frac{1}{\sqrt{1 - r^2}}, \quad r = \frac{v}{c} \]
Evaluate at r equal to 0.99
Why: First the root, then its reciprocal.
\[ \sqrt{1 - 0.99^2} = \sqrt{0.0199} \approx 0.1411, \quad \frac{m}{m_0} \approx 7.089 \]
Evaluate closer to 1
Why: Each extra nine shrinks the root by about a factor of the square root of 10.
\[ r = 0.999: \; 22.37, \qquad r = 0.9999: \; 70.71 \]
State the limit
Why: Speeds can only approach c from below, so this is a one-sided infinite limit.
\[ \lim_{r\to 1^-}\frac{1}{\sqrt{1 - r^2}} = +\infty \]
Figure (svg): The ratio of masses, one over the square root of 1 minus r squared, plotted against the speed ratio r from 0 to just below 1. It starts at 1, stays nearly flat until about 0.7, then shoots upward beside a dashed vertical line at r equals 1. Dots mark r equals 0.9 at about 2.29 and r equals 0.99 at about 7.09.
Check the book's closing number
Why: A 100 kilogram object at 0.9999 c would have mass 100 times the ratio, matching Table 2.8. An infinite mass is impossible, so no object reaches c: the line r equal to 1 is a vertical asymptote.
\[ 100 \times 70.712 \approx 7071 \text{ kg} \]
OpenStax Calculus Volume 1, §2.2 The Limit of a Function §2.2, pp. 131-132 — Example 2.12, Figure 2.23, Table 2.8
This is the chapter-opening example: a physical law that contains a vertical asymptote. Einstein's formula says a moving object's mass is its rest mass divided by the square root of one minus the speed ratio squared. Dividing by the rest mass leaves a function of the speed ratio alone.
The table shows what happens as the speed ratio approaches 1. At 0.99 the mass is about seven times the rest mass; at 0.999, about 22 times; at 0.9999, about 71 times. Each extra nine divides the square root by about 3.16, the square root of ten, so the ratio has no ceiling.
Because speeds approach the speed of light only from below, this is a one-sided infinite limit, and the line where the speed ratio is 1 is a vertical asymptote. The book's conclusion follows: reaching the speed of light would require infinite mass, so no object with mass can do it. The picture makes the asymptote visible, flat for everyday speeds and vertical near c.
Section
Part 6
Pattern
Figure (svg): A flow diagram. Start: find the left-hand limit and the right-hand limit separately. If both are the same real number L, the limit is L. If they are different real numbers, the limit does not exist: a jump. If a side grows without bound, write plus or minus infinity and mark a vertical asymptote. If a side never settles, the limit does not exist: an oscillation. A side note: the value f(a) is never part of the decision.
This is the procedure the whole lesson has been building, in the order to use it. The flow diagram shows the key design choice: you always split into two sides first, and every possible outcome is one of four branches.
The first branch, both sides agreeing on a real number, is the only one where the limit exists. The second is the jump, the third the blow-up, and the fourth the oscillation, matching the three failure modes plus the success case. For the blow-up, remember that a two-sided infinity needs both sides to go the same way.
The last line is the one most often ignored. The value f of a is looked at only at the end, if at all. It never decides the limit, but comparing it with the limit afterwards is what Section 2.4 will call testing for continuity.
Ranking
Put in order
Order the steps of estimating a limit from a table, as in Examples 2.4 and 2.5.
Why: Substitution first tells you why a table is needed. Both columns are built before any conclusion, and the conclusion from a table is only an estimate until a graph or, better, algebra confirms it, as Exercise 44 showed.
Drag the steps into order before checking. Most of the order is natural; the step people most often leave out or put last is trying direct substitution first, which tells you whether a table is needed at all and why.
The final step, confirming with a graph or algebra, is not optional. Exercise 44 showed that a table built exactly by the book's strategy can point confidently at the wrong answer. Treat the table as a way of forming a conjecture, and the confirmation as the way of believing it.
Comparison
Comparison matrix
| function at the point | left limit | right limit | two-sided limit | value |
|---|---|---|---|---|
| (x² − 4)/(x − 2) at 2 | 4 | 4 | 4 | undefined |
| |x − 2|/(x − 2) at 2 | -1 | 1 | DNE | undefined |
| 1/(x − 2)² at 2 | +∞ | +∞ | +∞ | undefined |
| Example 2.8 at 2 | 3 | 0 | DNE | 0 |
Fill in each blank from what you have seen in the lesson. The four rows are the three opening functions plus the piecewise function of Example 2.8, so each should be familiar.
Pay attention to how independent the columns are. The first row has a limit but no value. The second has two one-sided limits and neither a two-sided limit nor a value. The third is infinite on both sides, so its two-sided entry is plus infinity. The last has a value of 0 that agrees with the right side but not the left, and no limit.
If you can fill a table like this for any function and point, you have mastered this section.
Check
Check your understanding
A function has f(3) = 8. What can you conclude about the limit of f(x) as x approaches 3?
Answer: C
Why: The limit is determined by the values near 3, never by the value at 3. The limit could be 8, could be some other number, or could fail to exist, and f(3) = 8 is consistent with all three.
The question tests the central idea of the lesson: the value at a point and the limit at that point are independent pieces of information. Knowing one tells you nothing certain about the other.
The tempting answer is that the limit is 8, and for a continuous function it would be. But nothing says the function is continuous. Example 2.6 had a value of 4 and a limit of 3; Example 2.8 had a value and no limit at all. Both are consistent with knowing only the value.
Check
Check your understanding
For f(x) = 2x + 1 when x ≤ 3 and f(x) = x + 4 when x > 3, what is the limit as x approaches 3?
Answer: A
Why: From the left, 2x + 1 approaches 2(3) + 1 = 7; from the right, x + 4 approaches 3 + 4 = 7. Both sides agree, so by Theorem 2.2 the limit is 7.
Split into the two sides and substitute into the right piece for each. The left piece gives 2 times 3 plus 1, which is 7; the right piece gives 3 plus 4, which is also 7.
Because the two sides agree, Theorem 2.2 gives a two-sided limit of 7. A change of formula at a point creates a jump only when the pieces arrive at different heights, and here they are arranged to meet. You will see in Section 2.4 that this function is continuous at 3.
Check
Check your understanding
Which statement about f(x) = 1/(x + 1)³ is correct?
Answer: B
Why: The power 3 is odd, so by Theorem 2.3 the left side goes to −∞ and the right side to +∞. The two-sided limit does not exist, but one infinite side is already enough for x = −1 to be a vertical asymptote.
Match the function to Theorem 2.3 first. The base is x plus 1, so a is minus 1, and the power is 3, which is odd.
Odd powers keep the sign, so the left side goes to minus infinity and the right side to plus infinity. The two-sided limit therefore does not exist. The trap is option D: an asymptote needs only one infinite side, so the failure of the two-sided limit does not remove the asymptote.
Explain it to yourself
Discussion prompt
In two or three sentences, explain why the definition of a limit insists that x approach a but never equal it. Use the warm-up function to show what would be lost without that rule.
Write your explanation before revealing. The aim is to see that excluding the point is not an arbitrary rule but the thing that makes the definition useful.
Think about where limits are needed most. The warm-up function breaks at 2 with zero over zero, and yet its graph is obviously heading for 4. A definition that looked at the value at 2 would inherit the breakdown and report nothing. The same will be true of every derivative in Chapter 3, which is a limit of a difference quotient that is always zero over zero at the point.
The second benefit is robustness. Change a function at a single point, as in Example 2.6, and its limit does not move. The limit describes the trend of the whole neighbourhood, so one odd value cannot distort it.
Exit ticket
\[ f(x) = \frac{x+1}{x^2 - 4} \]
Discussion prompt
Find the one-sided limits of f at x = 2 and at x = −2, decide whether each two-sided limit exists, and name every vertical asymptote.
This one problem uses the factoring of the denominator, the sign analysis behind Theorem 2.3, and the definition of a vertical asymptote. Work it through before revealing.
Near 2, the numerator is about 3 and the factor x plus 2 is about 4, both positive, so the sign of the whole fraction follows the sign of x minus 2: negative on the left, positive on the right. That gives minus infinity and plus infinity.
Near minus 2, the numerator is about minus 1 and x minus 2 is about minus 4. Their quotient is positive, so the sign follows x plus 2: again minus infinity on the left and plus infinity on the right. Neither two-sided limit exists, both lines are vertical asymptotes, and a quick table confirms the signs.
Recap
| idea | statement | remember |
|---|---|---|
| limit | lim x→a f(x) = L: outputs approach one real L | the value f(a) is never used |
| tables and graphs | two columns closing in from both sides | evidence, not proof (Exercise 44) |
| one-sided limits | x → a⁻ from below, x → a⁺ from above | limit exists iff both agree (Theorem 2.2) |
| failure modes | jump, blow-up, oscillation | a limit must be ONE real number |
| infinite limits | grow without bound: write +∞ or −∞ | describes the failure; not a value |
| vertical asymptote | some one-sided limit is ±∞ | 1/(x − a)ⁿ: even up-up, odd down-up |
\[ \lim_{x\to a} f(x) = L \iff \lim_{x\to a^-} f(x) = L = \lim_{x\to a^+} f(x) \]
Next, Section 2.3 replaces tables with the limit laws, algebraic rules that compute limits exactly and turn every estimate in this lesson into a proof.
Stewart, Calculus: Early Transcendentals 8e, §2.2 The Limit of a Function §2.2, pp. 83-94 — the same material in Stewart
The whole lesson in one table. A limit is the single real number the outputs approach as the inputs approach a, and the value at a plays no part. Tables and graphs estimate limits but cannot prove them, as the table that lied in Exercise 44 showed.
One-sided limits split the approach in two, and Theorem 2.2 says the limit exists exactly when both sides agree. A limit can fail in three ways: a jump, a blow-up and an oscillation. Blow-ups are described with infinite limits, which record how a limit fails rather than giving it a value, and they locate vertical asymptotes, with Theorem 2.3 handling every power of one over x minus a.
Everything here was estimated. The next section introduces the limit laws, which let you compute limits exactly by algebra, and they are how you will actually evaluate limits from now on.
Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.