The epsilon-delta definition of a limit, why the order of its quantifiers matters, how to choose delta in a proof, and how to show that a number is not the limit.
“As close as we like” and “close enough” are fine for building intuition, but they cannot settle a disagreement. Think of a machinist who is told a part must be within 0.01 millimeters of spec. The machine has to be set precisely enough to hit that tolerance, and a tighter tolerance needs a tighter setting. The epsilon-delta definition turns this kind of promise into exact mathematics.
Key terms
Epsilon-delta definition
Definition Let f be defined on an open interval containing a, except possibly at a, and let L be a real number. Then limx→af(x)=L means that for every ε>0 there exists δ>0 such that for all x,
0<∣x−a∣<δ⟹∣f(x)−L∣<ε.
In plain words The number ε is the output tolerance someone demands, and δ is an input tolerance you supply. The limit is L when you can always meet the demand, whatever ε is.
Example To prove limx→3(2x−1)=5, let ε>0 and choose δ=ε/2. If 0<∣x−3∣<δ, then
∣(2x−1)−5∣=2∣x−3∣<2δ=ε.
Every epsilon before there exists delta
Definition The statement has the form “for every ε>0 there exists δ>0”. The order matters. The number δ is chosen after ε is given, so δ may depend on ε and on a. If the order were swapped to “there exists δ>0 such that for every ε>0”, the condition would force ∣f(x)−L∣<ε for all ε>0, that is f(x)=L for every x in a punctured neighborhood of a. That is a much stronger and different statement.
In plain words First the challenge, then the answer. Each tolerance gets its own δ, and smaller tolerances usually need smaller δ.
Example For limx→3(2x−1)=5 the rule δ=ε/2 gives
ε
1
0.1
0.01
δ
0.5
0.05
0.005
No single δ works for every ε here, which is fine, because the definition never asks for one.
The condition on x
Definition The hypothesis 0<∣x−a∣<δ has two parts. The right inequality ∣x−a∣<δ keeps x within δ of a. The left inequality 0<∣x−a∣ says x=a. So the condition describes exactly the punctured neighborhood (a−δ,a)∪(a,a+δ), and the value f(a) is never tested.
In plain words Only inputs near a count, and a itself never counts.
Example Let g(x)=x+1 for x=1 and g(1)=5. We claim limx→1g(x)=2. Given ε>0, take δ=ε. If 0<∣x−1∣<δ, then x=1, so
∣g(x)−2∣=∣(x+1)−2∣=∣x−1∣<ε.
At x=1 itself ∣g(1)−2∣=3, but that input is excluded, so it does no harm.
All nearby domain points
Definition In the definition, the implication
0<∣x−a∣<δ⟹∣f(x)−L∣<ε
must hold for every x in the domain of f that satisfies 0<∣x−a∣<δ, not merely for some of them. Under the standard convention of this course, every point of the punctured neighborhood is such an x once δ is small, because f is defined on an open interval around a except possibly at a.
In plain words It is not enough to find a few inputs near a that land close to L. Every input of the function in the punctured neighborhood has to land close to L.
Example Take h(x)=sin(π/x) near 0. For every positive whole number n the input x=1/n gives h(x)=sin(nπ)=0, so some inputs in every punctured neighborhood of 0 land exactly on 0. But the inputs x=2/(4n+1) give
h(4n+12)=sin(2nπ+2π)=1,
and they also come arbitrarily close to 0. With ε=1/2 no δ works for L=0, so 0 is not the limit.
Accumulation point
Definition A number a is an accumulation point of a set D if every interval (a−δ,a+δ) with δ>0 contains a point of D other than a. The point a itself may or may not belong to D. Some advanced texts extend the definition of limx→af(x)=L to a function with any domain D by testing only inputs x∈D, and then they take limits only at accumulation points of D. Under the convention of this course this condition holds automatically, since D contains a whole punctured neighborhood of a.
In plain words The function must have inputs arbitrarily close to a. Otherwise there is nothing to approach a through, and every number L would pass the test for no reason, which would break uniqueness.
Example The number 0 is an accumulation point of (0,1], even though 0 is not in the set, and also of the set {1,21,31,…}. The number 5 is not an accumulation point of [0,1]∪{5}, since the interval (4,6) contains no point of the set other than 5.
Choosing delta
Definition A proof of limx→af(x)=L needs a rule that turns each ε>0 into some δ>0 satisfying the definition. A standard method is to factor ∣f(x)−L∣ as ∣x−a∣ times another factor, bound that factor by first requiring δ≤1 (or some other fixed number), and then take δ to be the minimum of the two requirements. Any smaller positive δ also works.
In plain words Do the scratch work backwards from ∣f(x)−L∣<ε, then write the proof forwards. When a factor depends on x, trap x in a small interval first.
Example To prove limx→3x2=9, start from the factorization
∣x2−9∣=∣x−3∣∣x+3∣.
If ∣x−3∣<1, then 2<x<4 and ∣x+3∣<7. So take
δ=min(1,7ε).
If 0<∣x−3∣<δ, then both requirements hold, so
∣x2−9∣=∣x−3∣∣x+3∣<7ε⋅7=ε.
Negating the definition
Definition The statement limx→af(x)=L is false exactly when there exists ε>0 such that for every δ>0 there is an x with
0<∣x−a∣<δand∣f(x)−L∣≥ε.
To negate, each “for every” becomes “there exists”, each “there exists” becomes “for every”, and the implication “if 0<∣x−a∣<δ then ∣f(x)−L∣<ε” becomes ”0<∣x−a∣<δ and ∣f(x)−L∣≥ε”.
In plain words To show L is not the limit, find one tolerance that cannot be met. However small δ is, some input inside it lands at least ε away from L.
Example Let H(x)=0 for x<0 and H(x)=1 for x≥0. To show that 1 is not the limit at 0, take ε=1/2. For any δ>0 the input x=−δ/2 satisfies 0<∣x∣<δ and ∣H(x)−1∣=1≥1/2. In fact no L works. Take ε=1/2 again. If some δ worked, the inputs ±δ/2 would give the outputs 0 and 1, both within 1/2 of L, and the triangle inequality would give
1=∣1−0∣≤∣1−L∣+∣L−0∣<21+21=1,
which is impossible.
Check yourself
For limx→2(5x+3)=13 and ε=0.01, find a δ that works.
Show the answer
Since ∣(5x+3)−13∣=5∣x−2∣, any positive δ≤0.002 works, for example δ=0.002.
Prove that limx→1(4−3x)=1.
Show the answer
Let ε>0 and take δ=ε/3. If 0<∣x−1∣<δ, then
∣(4−3x)−1∣=3∣x−1∣<3δ=ε.
For limx→3x2=9 and ε=1, does δ=0.2 work?
Show the answer
No. The input x=3.19 satisfies ∣x−3∣<0.2, but x2=10.1761 is more than 1 away from 9. The rule from the lesson gives δ=min(1,1/7)=1/7, which does work.
May we count page visits to understand which courses and learning features are useful? Cloudflare collects visit and performance statistics, without your account details or answers. Your choice does not affect access. Details