A limit is not the value of a function. It is where the function is headed, which is a different question and often has a different answer.
By the end of this chapter you can
State what it means for a limit to exist
Distinguish the limit at a point from the value at that point
Use one-sided limits to show a limit does not exist
Read a limit off a graph and off a table
Almost every idea in this course is a limit wearing a costume. Derivatives are
limits of slopes, integrals are limits of sums, and series are limits of partial
sums. Getting this chapter right pays for itself four times over.
Here is the whole idea in one sentence: a limit asks where a function is
headed as the input approaches some point, and deliberately refuses to look at
what happens when it arrives.
That refusal is the useful part, not a technicality.
The function that cannot answer for itself
Consider
f(x)=x−1x2−1
At x=1 this is 00, which is not a number. The function has
nothing to say there. But for every other x, the numerator factors and the
fraction cancels:
x−1x2−1=x+1
x→1lim(x−1x2−1)=2
checked by the build
So f agrees with x+1 everywhere except at x=1, where f is undefined
and x+1 is 2. Approach 1 from either side and f gets arbitrarily close
to 2. It never is2.
We write
x→1limf(x)=2
and read it “the limit of f(x) as x approaches 1 is 2”. The statement is
about the neighbourhood of 1, not about 1.
From both sides, or not at all
A limit exists only if the function approaches the same value from the left
and from the right. Those two are the one-sided limits:
x→a−limf(x)from below,x→a+limf(x)from above
and the two-sided limit exists exactly when both exist and agree.
The standard counterexample is x∣x∣. For positive x it is 1;
for negative x it is −1. It never settles.
{
"kind": "secant",
"title": "abs(x)/x — drag the second point across zero and watch the slope refuse to settle",
"fn": "abs(x)/x",
"domain": [-2, 2],
"range": [-2.5, 2.5],
"at": -0.6,
"controls": [
{ "name": "h", "label": "h", "min": 0.05, "max": 1.8, "step": 0.01, "value": 1.4 }
]
}
The jump at zero is what “no limit” looks like. There is no single value the
function is heading towards, so there is nothing for the limit to be — and
saying “the limit is both 1 and −1” is not an answer, because a limit is a
single number or it does not exist.
Three ways a limit fails
Worth knowing by name, because a question that asks why a limit fails is
asking which of these it is:
The one-sided limits disagree — a jump, like x∣x∣ at 0.
The function grows without bound — x21 at 0. It is heading
somewhere, but ∞ is not a number, so the limit does not exist. Writing
lim=∞ describes how it fails; it is not a value.
The function oscillates forever — sin(x1) at
0 crosses every value between −1 and 1 infinitely often in any
neighbourhood of zero, so it is never eventually close to anything.
x→0limx21=∞
x→2lim(x−2x2−4)=4
x→0lim(xx+1−1)=21
checked by the build
The second and third of those are the shape you will meet most: a fraction that
is 00 at the point, rescued by algebra. The third needs multiplying
by the conjugate rather than factoring, which is chapter 1.2’s business.
Reading a limit off a table
When the algebra is not obvious, a table narrows things down. For
x1+x−1 near zero:
x
value
0.1
0.4881
0.01
0.4988
0.001
0.4999
−0.001
0.5001
−0.01
0.5013
−0.1
0.5132
Both sides converge on 0.5, which is what the verified claim above says
exactly — the limit is 21.
{
"question": "f(x) = (x² − 1)/(x − 1) is undefined at x = 1, yet the limit as x approaches 1 is 2. What best explains this?",
"options": [
{
"text": "The limit ignores what happens at x = 1 and describes only what happens near it",
"correct": true,
"why": "Exactly the definition. f agrees with x + 1 at every x other than 1, and the limit is determined entirely by those other points."
},
{
"text": "The 0/0 cancels, so f(1) is really 2 after simplification",
"correct": false,
"why": "The cancellation is valid only where x ≠ 1, because that is the only place you may divide by x − 1. f(1) remains undefined; the simplified expression is a different function that happens to agree everywhere else."
},
{
"text": "The limit is an approximation of f(1), accurate to within a small error",
"correct": false,
"why": "It is not an approximation of anything — it is an exact statement about the neighbourhood. f(1) does not exist, so there is nothing to approximate."
},
{
"text": "Limits only exist where a function is undefined",
"correct": false,
"why": "Backwards. Most limits are taken where the function is perfectly well defined and equal to the limit — that situation has a name, continuity, and it is the usual case rather than the exception."
}
]
}