We write limx→x0f(x)=K if f(x) can be made arbitrarily close to K by taking x sufficiently close to x0.
Key point: This is about behaviour nearx0, not atx0. The limit can exist even if f(x0) is undefined or different from K.
The (ε, δ) Definition (Weierstrass)
limx→x0f(x)=K if and only if: for any ε>0, there exists δ>0 such that ∣f(x)−K∣<ε for all x satisfying 0<∣x−x0∣<δ.
Key features:
The order matters: for anyε, we must find someδ
The condition 0<∣x−x0∣ excludes x0 itself
We control x (through δ) to control f(x) (through ε)
Analogy: A game where an opponent picks ε, and you must respond with a valid δ.
Example: Proving limx→0xsinx=1
We need: for any ε>0, find δ such that ∣x∣<δ⇒xsinx−1<ε.
Using the inequality ∣sinx∣<∣x∣<∣tanx∣ for 0<x<π/2:
From sinx<x<tanx, we get cosx<xsinx<1
As x→0, cosx→1, so xsinx is squeezed to 1
This uses the squeeze theorem: If g(x)≤f(x)≤h(x) and limg=limh=L, then limf=L.
One-Sided Limits
Right-hand limit:limx→a+f(x)
Left-hand limit:limx→a−f(x)
The (two-sided) limit exists if and only if both one-sided limits exist and are equal.
Example: The Heaviside step function:
H(x)={01x<0x≥0
has limx→0+H(x)=1 and limx→0−H(x)=0. Since these differ, limx→0H(x) does not exist.
Limits at Infinity
limx→∞f(x)=K if for any ε>0, there exists X such that ∣f(x)−K∣<ε for all x>X.
Example:limx→∞x1=0
Given ε>0, choose X=1/ε. Then for x>X:
x1−0=x1<X1=ε
Continuity
A function f is continuous atx=a if:
f(a) exists (the function is defined there)
limx→af(x)=f(a)
(ε, δ) form: For any ε>0, there exists δ>0 such that ∣f(x)−f(a)∣<ε for all ∣x−a∣<δ.
Note: unlike the limit definition, this includesx=a.
Types of discontinuity:
Removable: limit exists but ≠ function value
Jump: one-sided limits exist but differ
Essential: limit does not exist (e.g., oscillation)
Differentiability
A function f is differentiable atx=a if the limit:
f′(a)=limh→0hf(a+h)−f(a)
exists.
Key relationship:
Differentiable ⇒ continuous
Continuous ⇒ differentiable
Example where continuous but not differentiable:f(x)=∣x∣ at x=0. The left and right derivatives differ.
Big-O Notation
We write f(x)=O(g(x)) as x→a if there exist constants K>0 and δ>0 such that:
∣f(x)∣≤K∣g(x)∣for all 0<∣x−a∣<δ
Intuition:f is bounded by a multiple of g near a.
Examples:
x2+x=O(x2) as x→∞ (the x2 term dominates)
sinx=O(x) as x→0 (since sinx/x→1)
Notation convention: We usually state the tightest bound.
Why Rigorous Analysis Matters
The (ε, δ) framework:
Makes precise what “close” means
Avoids circular reasoning in limits
Handles pathological cases (discontinuous functions, functions that oscillate)
Provides foundations for calculus proofs
Historically, calculus was developed before rigorous foundations. The (ε, δ) definition (Weierstrass, 19th century) resolved centuries of confusion about infinitesimals, convergence, and continuity.