Limits & the Epsilon-Delta Definition

⚡ Key Takeaways

One-line statement: $\lim_{x\to a} f(x) = L$ means $f(x)$ can be forced arbitrarily close to $L$ by taking $x$ sufficiently close (but not equal) to $a$ , formalized as: for every $\varepsilon>0$ there exists $\delta>0$ such that $0<|x-a|<\delta \implies |f(x)-L|<\varepsilon$ . Key result in words: The limit converts the vague notion of "approaching" into a rigorous logical game between a challenger (who picks tolerance $\varepsilon$ ) and a defender (who must find a neighborhood $\delta$ ). Where it applies: The foundation of continuity, derivatives, integrals, infinite series — all of real and complex analysis. Who/when: Augustin-Louis Cauchy (informal, 1820s) and Karl Weierstrass (the precise $\varepsilon$ – $\delta$ formulation, 1850s–1870s).

💡 The Core Insight

A limit is a promise that can always be kept. The challenger names a vertical tolerance band of half-width $\varepsilon$ around $L$ ; the defender must respond with a horizontal punctured interval of half-width $\delta$ around $a$ such that the entire graph over that interval lies inside the band. Crucially, $\delta$ is allowed to depend on $\varepsilon$ (smaller $\varepsilon$ usually forces smaller $\delta$ ), and the point $x=a$ is always excluded — what happens at $a$ is irrelevant. The limit is a statement about the neighborhood-deleted-of-its-center, not about the center.

DefinitionLimit of a Function — Epsilon-Delta

Let $f$ be a real-valued function defined on a punctured neighborhood of $a$ (i.e., on some set containing $(a-r,a)\cup(a,a+r)$ for some $r>0$ ). We say the limit of $f(x)$ as $x$ approaches $a$ is $L$ , written $\lim_{x\to a} f(x)=L$ , if:

1

\quad \forall\,\varepsilon>0\;\; \exists\,\delta>0 \;\;\text{such that}\;\; 0<|x-a|<\delta \implies |f(x)-L|<\varepsilon.

Derivation hint: read left to right as a quantifier game — $\varepsilon$ is chosen first (the adversary), then $\delta$ may depend on it; the inner implication is the contract that must hold. Here:

$\varepsilon>0$ is the vertical tolerance (how close $f(x)$ must be to $L$ ),
$\delta>0$ is the horizontal radius (how close $x$ must be to $a$ ),
$0<|x-a|$ enforces $x\ne a$ (the punctured condition),
$L$ is the proposed limit value.

The condition $0<|x-a|<\delta$ describes the punctured interval $(a-\delta,a)\cup(a,a+\delta)$ , and $|f(x)-L|<\varepsilon$ describes the open band $(L-\varepsilon,L+\varepsilon)$ .

DefinitionOne-Sided Limits

The right-hand limit $\lim_{x\to a^+} f(x)=L$ requires:

2

\quad \forall\,\varepsilon>0\;\;\exists\,\delta>0:\;\; a<x<a+\delta \implies |f(x)-L|<\varepsilon.

Derivation hint: replace the symmetric punctured interval with the one-sided interval $0<x-a<\delta$ . The left-hand limit $\lim_{x\to a^-} f(x)=L$ uses $a-\delta<x<a$ . The two-sided limit exists iff both one-sided limits exist and are equal.

DefinitionLimits Involving Infinity

Limit at infinity: $\lim_{x\to\infty} f(x)=L$ means

3

\quad \forall\,\varepsilon>0\;\;\exists\,M>0:\;\; x>M \implies |f(x)-L|<\varepsilon.

Derivation hint: " $x$ close to $a$ " becomes " $x$ large"; the radius $\delta$ becomes a threshold $M$ .

Infinite limit: $\lim_{x\to a} f(x)=\infty$ means

4

\quad \forall\,N>0\;\;\exists\,\delta>0:\;\; 0<|x-a|<\delta \implies f(x)>N.

Derivation hint: " $f(x)$ close to $L$ " becomes " $f(x)$ arbitrarily large"; tolerance $\varepsilon$ becomes a height threshold $N$ .

TheoremUniqueness of Limits

If $\lim_{x\to a} f(x)=L_1$ and $\lim_{x\to a} f(x)=L_2$ , then $L_1=L_2$ .

TheoremAlgebraic Limit Laws

If $\lim_{x\to a} f(x)=L$ and $\lim_{x\to a} g(x)=M$ , then:

5

\quad \lim_{x\to a}\big(f(x)+g(x)\big)=L+M,\qquad \lim_{x\to a}\big(f(x)g(x)\big)=LM,

Derivation hint: bound $|fg-LM|\le|f||g-M|+|M||f-L|$ using a local bound on $|f|$ near $a$ . and, provided $M\ne0$ ,

6

\quad \lim_{x\to a}\frac{f(x)}{g(x)}=\frac{L}{M}.

Derivation hint: control the denominator away from zero by taking $\delta$ small enough that $|g(x)|>|M|/2$ .

TheoremSqueeze (Sandwich) Theorem

If $g(x)\le f(x)\le h(x)$ on a punctured neighborhood of $a$ and $\lim_{x\to a} g(x)=\lim_{x\to a} h(x)=L$ , then

7

\quad \lim_{x\to a} f(x)=L.

Derivation hint: from both outer limits within $\varepsilon$ of $L$ , deduce $L-\varepsilon<g(x)\le f(x)\le h(x)<L+\varepsilon$ .

ProofUniqueness of Limits

Suppose for contradiction that $\lim_{x\to a}f(x)=L_1$ and $=L_2$ with $L_1\ne L_2$ . 1. Let $\varepsilon=\dfrac{|L_1-L_2|}{2}>0$ . 2. By the limit hypothesis for $L_1$ , there is $\delta_1>0$ with $0<|x-a|<\delta_1 \implies |f(x)-L_1|<\varepsilon$ . 3. By the limit hypothesis for $L_2$ , there is $\delta_2>0$ with $0<|x-a|<\delta_2 \implies |f(x)-L_2|<\varepsilon$ . 4. Let $\delta=\min(\delta_1,\delta_2)$ . Since $f$ is defined on a punctured neighborhood, pick any $x$ with $0<|x-a|<\delta$ (such $x$ exists). 5. Then by the triangle inequality:

8

\quad |L_1-L_2|\le |L_1-f(x)|+|f(x)-L_2|<\varepsilon+\varepsilon=2\varepsilon=|L_1-L_2|.

Derivation hint: insert $\pm f(x)$ and apply $|a+b|\le|a|+|b|$ . 6. This gives $|L_1-L_2|<|L_1-L_2|$ , a contradiction. Hence $L_1=L_2$ . $\blacksquare$

∎

ProofSum Law:

\lim(f+g)=L+M

Let $\varepsilon>0$ be given. 1. Since $\lim_{x\to a}f=L$ , there is $\delta_1>0$ with $0<|x-a|<\delta_1 \implies |f(x)-L|<\varepsilon/2$ . 2. Since $\lim_{x\to a}g=M$ , there is $\delta_2>0$ with $0<|x-a|<\delta_2 \implies |g(x)-M|<\varepsilon/2$ . 3. Let $\delta=\min(\delta_1,\delta_2)>0$ . Then for $0<|x-a|<\delta$ , both bounds hold simultaneously. 4. By the triangle inequality:

9

\quad |(f+g)(x)-(L+M)|\le|f(x)-L|+|g(x)-M|<\tfrac{\varepsilon}{2}+\tfrac{\varepsilon}{2}=\varepsilon.

Derivation hint: the "split $\varepsilon$ in half" trick — distribute the tolerance budget across the two error sources. 5. This is exactly the definition with limit $L+M$ . $\blacksquare$

∎

ProofProduct Law: a representative argument

Let $\varepsilon>0$ . We estimate

10

\quad |f(x)g(x)-LM| = |f(x)g(x)-f(x)M + f(x)M-LM| \le |f(x)|\,|g(x)-M| + |M|\,|f(x)-L|.

Derivation hint: add and subtract the "mixed" term $f(x)M$ to bridge between the two factors. 1. Bound $|f(x)|$ locally. Since $f\to L$ , choose $\delta_0$ so that $|f(x)-L|<1$ , hence $|f(x)|<|L|+1$ for $0<|x-a|<\delta_0$ . 2. Control the first term. Choose $\delta_1$ so that $|g(x)-M|<\dfrac{\varepsilon}{2(|L|+1)}$ . 3. Control the second term. Choose $\delta_2$ so that $|f(x)-L|<\dfrac{\varepsilon}{2(|M|+1)}$ . 4. Let $\delta=\min(\delta_0,\delta_1,\delta_2)$ . Then for $0<|x-a|<\delta$ :

11

\quad |f(x)g(x)-LM| < (|L|+1)\cdot\frac{\varepsilon}{2(|L|+1)} + |M|\cdot\frac{\varepsilon}{2(|M|+1)} < \frac{\varepsilon}{2}+\frac{\varepsilon}{2}=\varepsilon.

Derivation hint: the $+1$ in denominators guards against $L=0$ or $M=0$ making the fraction undefined. $\blacksquare$

∎

KW

Karl Weierstrass

1815–1897

Often called the "father of modern analysis," Weierstrass spent fourteen years as a secondary-school teacher before his groundbreaking lectures at the University of Berlin made rigor the gold standard of analysis. He distrusted the geometric intuition and "infinitely small quantities" pervading earlier calculus. Through his lectures (transcribed and circulated, as he published little) he crystallized the static, inequality-based $\varepsilon$ – $\delta$ definitions of limit and continuity that bear no reference to motion or infinitesimals. His 1872 construction of a continuous-but-nowhere-differentiable function shattered the intuition that continuity implies smoothness, proving that only airtight definitions — not pictures — could be trusted. His students (Cantor, Schwarz, Frobenius, Kovalevskaya, Mittag-Leffler) carried this rigor across mathematics.

🧠 Intuition

Think of the limit as a manufacturing tolerance problem. You operate a machine whose dial setting is $x$ and whose output is $f(x)$ ; you want output near a target $L$ . A customer specifies the maximum allowable output error $\varepsilon$ . Your job: find a dial tolerance $\delta$ such that any setting within $\delta$ of the ideal $a$ (but you needn't hit $a$ exactly — perhaps the dial is broken precisely at $a$ ) yields acceptable output. The limit "exists and equals $L$ " precisely when, for every conceivable customer demand $\varepsilon$ (no matter how strict), you can find a working tolerance $\delta$ . The logical order is everything: $\varepsilon$ comes first, $\delta$ responds. If you could fix one $\delta$ that works for all $\varepsilon$ , the function would be locally constant — far too strong. The dependence $\delta=\delta(\varepsilon)$ is the whole content. Geometrically: thin the horizontal band ( $\delta\to0$ ) until the graph — viewed only on the punctured interval — fits inside the vertical band of height $2\varepsilon$ .

Figure

The epsilon-delta band game: the graph over the punctured

\delta

-interval must lie inside the

\varepsilon

-band around

L

.

ExampleLinear limit by epsilon-delta

Claim: $\lim_{x\to 3}(2x+1)=7$ .

Given: $f(x)=2x+1$ , $a=3$ , proposed $L=7$ .

Scratch work (find $\delta$ ): We need $|f(x)-L|<\varepsilon$ , i.e.

12

\quad |(2x+1)-7| = |2x-6| = 2|x-3|<\varepsilon \iff |x-3|<\frac{\varepsilon}{2}.

Derivation hint: solve the target inequality for $|x-a|$ to read off the required $\delta$ .

Formal proof: Let $\varepsilon>0$ . Choose $\delta=\dfrac{\varepsilon}{2}$ . If $0<|x-3|<\delta$ , then

|(2x+1)-7| = 2|x-3| < 2\cdot\frac{\varepsilon}{2}=\varepsilon.

\boxed{\;\lim_{x\to 3}(2x+1)=7,\quad \delta=\varepsilon/2\;}

ExampleQuadratic limit requiring a bounding step

Claim: $\lim_{x\to 2} x^2 = 4$ .

Given: $f(x)=x^2$ , $a=2$ , $L=4$ .

Scratch work: $|x^2-4| = |x-2|\,|x+2|$ . The factor $|x+2|$ is not constant, so we pre-bound it by restricting $\delta\le1$ : if $|x-2|<1$ then $1<x<3$ , so $3<x+2<5$ , giving $|x+2|<5$ .

13

\quad |x^2-4| = |x-2|\,|x+2| < 5\,|x-2| \quad\text{whenever}\quad |x-2|<1.

Derivation hint: cap $\delta$ at 1 first to make the "messy" factor bounded, then choose the rest of $\delta$ to finish.

Choose $\delta$ : Take $\delta=\min\!\left(1,\dfrac{\varepsilon}{5}\right)$ . Then for $0<|x-2|<\delta$ :

|x^2-4| < 5|x-2| < 5\cdot\frac{\varepsilon}{5}=\varepsilon.

Numerical check: For $\varepsilon=0.1$ , $\delta=\min(1,0.02)=0.02$ . At $x=2.02$ : $x^2=4.0804$ , error $=0.0804<0.1$ . ✓

\boxed{\;\lim_{x\to 2}x^2=4,\quad \delta=\min(1,\varepsilon/5)\;}

⚠️ Common mistake

The limit depends on $f(a)$ . False. The condition $0<|x-a|$ explicitly excludes $x=a$ . A function can be undefined at $a$ , or defined with a value different from $L$ , and still have $\lim_{x\to a}f(x)=L$ . For example $\dfrac{\sin x}{x}$ is undefined at $0$ yet has limit $1$ .

$\delta$ must be unique or "the largest possible." False. Any smaller positive $\delta$ also works; if $\delta_0$ succeeds, so does $\delta_0/2$ . We only need existence.

You may choose $\delta$ before $\varepsilon$ . False — that reverses the quantifiers and produces a much weaker (usually trivial) statement. The order $\forall\varepsilon\,\exists\delta$ is essential.

" $f(x)$ gets close to $L$ " is enough. Vague closeness fails; the definition demands an arbitrarily-prescribable closeness controllable by $x$ . Oscillating functions can get close to $L$ infinitely often without having limit $L$ .

CounterexampleA limit that does not exist

Consider $f(x)=\sin(1/x)$ as $x\to0$ . As $x\to0^+$ , $1/x\to\infty$ and $\sin(1/x)$ oscillates between $-1$ and $1$ infinitely many times in every punctured neighborhood of $0$ . No candidate $L$ can satisfy the definition: take $\varepsilon=1/2$ . For any $\delta>0$ , the interval $(0,\delta)$ contains points where $f=1$ and points where $f=-1$ (since $1/x$ ranges over an entire interval of length $>2\pi$ ). At least one of these is more than $1/2$ from $L$ . Hence

14

\quad \lim_{x\to 0}\sin(1/x)\ \text{does not exist}.

Derivation hint: exhibit two sequences $x_n\to0$ with $f(x_n)\to1$ and $f(y_n)\to-1$ ; differing subsequential limits forbid a limit. Contrast: $x\sin(1/x)\to0$ by the Squeeze Theorem, since $|x\sin(1/x)|\le|x|$ . The damping factor tames the oscillation.

connection

The $\varepsilon$ – $\delta$ limit is the seed from which the calculus grows:

<a class="wiki-link" href="/wiki/ndli-isc-12-mathematics-continuity">Continuity</a> is the special case $\lim_{x\to a}f(x)=f(a)$ — the limit equals the function value, removing the "puncture" exclusion's effect.
The Derivative is a limit of difference quotients, $f'(a)=\lim_{h\to0}\frac{f(a+h)-f(a)}{h}$ .
The Riemann Integral is a limit of Riemann sums as the mesh tends to zero.
Sequences and Their Limits use an analogous $\varepsilon$ – $N$ definition; the Sequential Criterion states $\lim_{x\to a}f=L$ iff $f(x_n)\to L$ for every sequence $x_n\to a$ with $x_n\ne a$ .
Topology and Open Sets generalizes $\varepsilon$ and $\delta$ neighborhoods to open sets, yielding the topological definition: preimages of open sets are open.
<a class="wiki-link" href="/wiki/math-topgeo-metric-spaces">Metric Spaces</a> replace $|x-a|$ by a distance $d(x,a)$ , extending limits to abstract settings.

🔬 Where this appears

⏳ History

- c. 1665–1684 — Newton ("ultimate ratios") and Leibniz (differentials

dx

) develop calculus using infinitesimals, without rigorous foundations.

- 1734 — Bishop George Berkeley's The Analyst derides infinitesimals as "ghosts of departed quantities," exposing the logical gaps.

- 1765–1797 — d'Alembert and Lagrange attempt limit- and algebra-based foundations; Lagrange tries (unsuccessfully) to ground calculus in Taylor series.

- 1817 — Bernard Bolzano gives an essentially modern definition of continuity using inequalities, in Rein analytischer Beweis, but his work goes largely unnoticed.

- 1821 — Augustin-Louis Cauchy's Cours d'analyse defines limits, continuity, and convergence verbally ("indefinitely approaches"), introducing

\varepsilon

and

\delta

as quantities but not yet fully formalized inequalities.

- 1823 — Cauchy bases the derivative and integral on limits in his lectures, an enormous step toward rigor.

- 1850s–1870s — Karl Weierstrass, in his Berlin lectures, supplies the fully static, quantifier-precise

\varepsilon

–

\delta

definitions still used today, eliminating any appeal to motion or infinitesimals.

- 1872 — Weierstrass exhibits a continuous nowhere-differentiable function, demonstrating that only rigorous definitions can be trusted; Heine publishes related

\varepsilon

–

\delta

formulations crediting Weierstrass.

- 1960 — Abraham Robinson's nonstandard analysis rehabilitates infinitesimals on a rigorous logical basis, showing they were not inherently flawed — but

\varepsilon

–

\delta

remains the standard framework.

open-problem

The $\varepsilon$ – $\delta$ definition, while logically airtight, raises deeper questions and has known limitations:

Uniform vs. pointwise. Ordinary limits/continuity allow $\delta$ to depend on the point $a$ . Uniform continuity demands a single $\delta(\varepsilon)$ working for all points — a strictly stronger, often subtle condition (e.g., $1/x$ on $(0,1)$ is continuous but not uniformly so). Distinguishing these correctly trips up many students and matters greatly in integration and approximation theory.
Order of limits. For functions of two variables, $\lim_{x\to0}\lim_{y\to0}f$ and $\lim_{y\to0}\lim_{x\to0}f$ can differ from each other and from the joint limit. The conditions under which limits and operations (integration, summation, differentiation) may be interchanged form the substance of advanced analysis (dominated/monotone convergence, uniform convergence theorems).
Constructive content. Classically the definition asserts existence of $\delta$ ; constructive mathematics demands an explicit algorithm producing $\delta$ from $\varepsilon$ . Not all classically valid limit statements are constructively provable, raising foundational questions about what "the limit exists" means computationally.
Generalization frontiers. Extending limits to nets and filters (for general topological spaces) and to nonstandard analysis (with genuine infinitesimals) shows the $\varepsilon$ – $\delta$ formulation is one of several equivalent — but not always interchangeable — viewpoints, each illuminating different aspects.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also