9.2 Continuity

Continuity is a fundamental concept in mathematics that describes functions without breaks or jumps. It forms the basis for many important theorems in calculus and analysis, and understanding it well is essential before moving into derivatives, integrals, and beyond.

This topic covers the precise definition of continuity, the different types and their relative strengths, key theorems about continuous functions, how to identify and classify discontinuities, and how continuity connects to differentiability.

Definition of continuity

Continuity captures a simple idea: a function behaves predictably near every point in its domain. Small changes in input produce small changes in output. Getting from the intuitive picture to a precise definition is one of the key moves in mathematical thinking.

Informal vs formal definitions

The informal definition says a continuous function has no breaks, gaps, or jumps in its graph. You could draw it without lifting your pencil. This is great for building intuition, but it's not precise enough to prove anything.

The formal definition restates this idea using limits: a function is continuous at a point if the limit as you approach that point equals the actual function value there. The shift from "no breaks" to a limit-based statement is a classic example of how mathematicians turn intuition into something rigorous enough for proofs.

Epsilon-delta definition

The $\epsilon$-$\delta$ definition is the fully rigorous version. Here's what it says:

A function $f$ is continuous at $x = a$ if for every $\epsilon > 0$, there exists a $\delta > 0$ such that whenever $|x - a| < \delta$, we have $|f(x) - f(a)| < \epsilon$.

Think of it this way: you pick any tolerance $\epsilon$ for how close the outputs need to be. Then you need to find a window $\delta$ around the input $a$ that guarantees all outputs land within that tolerance. If you can always find such a $\delta$ no matter how small $\epsilon$ gets, the function is continuous at $a$.

This definition forms the backbone of rigorous analysis and is used in most formal proofs about continuous functions.

Continuity at a point

A function $f$ is continuous at a specific point $x = a$ when three conditions are all satisfied:

1. $f(a)$ is defined (the function has a value at that point)
2. $\lim_{x \to a} f(x)$ exists (the function approaches a single value from both sides)
3. $\lim_{x \to a} f(x) = f(a)$ (the limit matches the actual function value)

This is expressed compactly as $\lim_{x \to a} f(x) = f(a)$. If any one of these three conditions fails, the function is discontinuous at $a$. Checking all three gives you a systematic way to test continuity at any point.

Types of continuity

Not all forms of continuity are created equal. Mathematicians distinguish several types, each progressively stronger. Understanding the hierarchy helps you see why certain theorems require specific types of continuity.

Pointwise continuity

A function $f$ is pointwise continuous on a set $S$ if it's continuous at each individual point in $S$. The key detail: the $\delta$ you find in the $\epsilon$-$\delta$ definition can depend on which point you're looking at.

This means a function can be continuous at every point in its domain, but the "difficulty" of achieving a given tolerance can vary wildly from point to point. Most functions you encounter in introductory calculus are pointwise continuous on their domains.

Uniform continuity

Uniform continuity is a stronger requirement. Instead of allowing $\delta$ to depend on the specific point, you need a single $\delta$ that works for all points in the domain simultaneously.

Formally: for every $\epsilon > 0$, there exists a $\delta > 0$ such that for all $x, y$ in the domain, if $|x - y| < \delta$, then $|f(x) - f(y)| < \epsilon$.

For example, $f(x) = x^2$ on all of $\mathbb{R}$ is continuous but not uniformly continuous, because as $x$ gets larger, the function changes faster and faster, so no single $\delta$ works everywhere. However, $f(x) = x^2$ restricted to any closed bounded interval is uniformly continuous. In fact, any continuous function on a closed bounded interval is uniformly continuous (this is the Heine-Cantor theorem).

Absolute continuity

Absolute continuity is stronger still and connects continuity to integration. A function $f$ is absolutely continuous on $[a, b]$ if for every $\epsilon > 0$, there exists a $\delta > 0$ such that for any finite collection of disjoint intervals $(a_i, b_i)$ whose total length is less than $\delta$, the sum $\sum |f(b_i) - f(a_i)| < \epsilon$.

The crucial consequence: absolutely continuous functions are differentiable almost everywhere, and you can recover them from their derivatives via integration. This property is central to Lebesgue integration theory. Every absolutely continuous function is uniformly continuous, but not vice versa.

Properties of continuous functions

Continuous functions on closed bounded intervals have remarkably strong properties. These theorems are workhorses throughout calculus and analysis.

Intermediate value theorem

If $f$ is continuous on $[a, b]$ and $k$ is any value between $f(a)$ and $f(b)$, then there exists some $c \in [a, b]$ such that $f(c) = k$.

In plain terms: a continuous function hits every value between its endpoints. If $f(0) = -3$ and $f(5) = 7$, then $f$ must equal zero somewhere between 0 and 5. The graph can't get from $-3$ to $7$ without crossing every value in between.

This theorem is the foundation of root-finding methods like the bisection algorithm, and it's used to prove that solutions to equations exist even when you can't find them explicitly.

Extreme value theorem

If $f$ is continuous on a closed bounded interval $[a, b]$, then $f$ attains both a maximum and a minimum value somewhere on that interval. That is, there exist points $c, d \in [a, b]$ such that $f(c) \leq f(x) \leq f(d)$ for all $x \in [a, b]$.

Both conditions matter: the interval must be closed and bounded. The function $f(x) = 1/x$ on the open interval $(0, 1)$ is continuous but has no maximum (it blows up near 0). And $f(x) = x$ on $[0, \infty)$ is continuous on a closed but unbounded interval and has no maximum.

Mean value theorem

If $f$ is continuous on $[a, b]$ and differentiable on $(a, b)$, then there exists a point $c \in (a, b)$ such that:

$f'(c) = \frac{f(b) - f(a)}{b - a}$

This says there's at least one point where the instantaneous rate of change equals the average rate of change over the whole interval. Geometrically, there's a point where the tangent line is parallel to the secant line connecting $(a, f(a))$ and $(b, f(b))$.

The MVT is fundamental for proving many other results in calculus, including Taylor's theorem and various error bounds in numerical analysis.

Discontinuities

When a function fails to be continuous at a point, the type of failure matters. Classifying discontinuities tells you a lot about the function's behavior and whether the problem can be "fixed."

Removable discontinuities

A removable discontinuity occurs when the limit exists at a point, but either the function is undefined there or its value doesn't match the limit. The graph has a "hole" that could be filled in.

For example, $f(x) = \frac{x^2 - 1}{x - 1}$ simplifies to $x + 1$ for all $x \neq 1$, so $\lim_{x \to 1} f(x) = 2$. But $f(1)$ is undefined because the original expression has $0/0$. You can remove this discontinuity by defining $f(1) = 2$.

Jump discontinuities

A jump discontinuity happens when the left-hand and right-hand limits both exist but aren't equal. The function makes a sudden "jump" at that point.

The classic example is the Heaviside step function, which equals 0 for $x < 0$ and 1 for $x \geq 0$. The left-hand limit at 0 is 0, the right-hand limit is 1, and no single redefinition at $x = 0$ can fix this. Jump discontinuities are common in piecewise-defined functions.

Infinite discontinuities

An infinite discontinuity occurs when the function grows without bound as it approaches a point, typically associated with a vertical asymptote.

For instance, $f(x) = \frac{1}{x}$ has an infinite discontinuity at $x = 0$: the function heads toward $+\infty$ from the right and $-\infty$ from the left. These can be one-sided (both sides go to the same infinity) or two-sided (they go in opposite directions), and they appear frequently in rational functions where the denominator approaches zero.

Continuity in different contexts

The concept of continuity extends well beyond single-variable functions on the real line. Mathematicians adapt the core idea to fit different settings.

Continuity on intervals

The type of interval affects how continuity is defined at the endpoints:

• On an open interval $(a, b)$: $f$ is continuous if it's continuous at every point in the interval.
• On a closed interval $[a, b]$: $f$ must be continuous on $(a, b)$, continuous from the right at $a$ (i.e., $\lim_{x \to a^+} f(x) = f(a)$), and continuous from the left at $b$ (i.e., $\lim_{x \to b^-} f(x) = f(b)$).
• Half-open intervals like $[a, b)$ or $(a, b]$ use the appropriate one-sided condition at the closed endpoint.

This distinction matters because the major theorems (IVT, EVT) specifically require continuity on closed intervals.

Continuity of composite functions

If $g$ is continuous at $a$ and $f$ is continuous at $g(a)$, then the composite function $f \circ g$ is continuous at $a$. In other words, $(f \circ g)(x) = f(g(x))$ is continuous at $a$.

This is powerful because it lets you build complex continuous functions from simpler ones. For example, since $\sin(x)$ and $x^2$ are both continuous everywhere, $\sin(x^2)$ is automatically continuous everywhere.

Continuity in multiple dimensions

For a function $f(x_1, x_2, \ldots, x_n)$ of several variables, continuity at a point means the function's value approaches $f(\mathbf{a})$ as the input approaches $\mathbf{a}$ from any direction.

A subtle point: being continuous in each variable separately (holding the others fixed) does not guarantee joint continuity. You need the limit to hold along every possible path of approach, not just along the coordinate axes. This is a common source of mistakes in multivariable calculus.

Preservation of continuity

One of the most useful features of continuity is that it's preserved under standard mathematical operations. This gives you a toolkit for determining continuity of complex functions without going back to the $\epsilon$-$\delta$ definition every time.

Under arithmetic operations

If $f$ and $g$ are both continuous at $a$, then:

• $f + g$ is continuous at $a$
• $f - g$ is continuous at $a$
• $f \cdot g$ is continuous at $a$
• $\frac{f}{g}$ is continuous at $a$, provided $g(a) \neq 0$

This means you can combine continuous functions freely using arithmetic, and the result stays continuous (just watch out for division by zero).

Under composition

As noted above, the composition of continuous functions is continuous. If $g$ is continuous at $a$ and $f$ is continuous at $g(a)$, then $f \circ g$ is continuous at $a$.

Combined with the arithmetic rules, this means that any function built from standard continuous functions (polynomials, trig functions, exponentials, etc.) using addition, multiplication, division, and composition is continuous wherever it's defined.

Under limits

If $f$ is continuous at $a$, then $\lim_{x \to a} f(x) = f(a)$. This is really just a restatement of the definition, but it has a practical payoff: for continuous functions, you can evaluate limits by direct substitution.

This is why, in a calculus course, you can often just "plug in" to find a limit. That shortcut works precisely because the functions involved are continuous at the point in question.

Applications of continuity

In calculus

Continuity is woven into nearly every major result in calculus:

• The Fundamental Theorem of Calculus requires the integrand to be continuous to guarantee that the antiderivative exists and is differentiable.
• Optimization relies on the Extreme Value Theorem, which requires continuity on a closed interval.
• Taylor series approximate functions locally, and continuity (along with differentiability) ensures the approximation is meaningful.
• Differential equations often require continuous coefficient functions to guarantee existence and uniqueness of solutions.

In topology

In topology, continuity is defined more generally using open sets rather than $\epsilon$-$\delta$. A function between topological spaces is continuous if the preimage of every open set is open.

This generalized definition leads to the concept of homeomorphism (a continuous bijection with a continuous inverse), which captures when two spaces are "the same" topologically. Continuity is also central to defining connected spaces, compact spaces, and studying which properties are preserved under continuous deformation.

In real-world modeling

Continuity is a natural assumption in many physical models because most real-world quantities change smoothly:

• Physics: position, velocity, and temperature typically vary continuously over time
• Finance: the Black-Scholes model for option pricing assumes continuous price paths
• Engineering: signal processing and control theory rely on continuous functions to model and design systems
• Computer graphics: smooth curves and surfaces are constructed using continuous (often differentiable) functions

Testing for continuity

Graphical methods

Looking at a graph gives you a quick, intuitive check. A continuous function has an unbroken curve with no holes, jumps, or vertical asymptotes. However, graphical methods have limitations: very small discontinuities might be invisible at normal zoom levels, and some functions (like the Weierstrass function) look continuous on a graph but have surprising properties. Graphs are good for building intuition but aren't sufficient for rigorous arguments.

Analytical methods

To test continuity analytically, examine the function's formula and domain:

1. Identify points where the function might have problems (where the formula changes in a piecewise definition, where a denominator could be zero, where a square root argument could be negative, etc.)
2. At each suspicious point, check whether the function is defined there.
3. Compute the left-hand and right-hand limits at that point.
4. Verify that both one-sided limits exist, are equal, and match the function value.

If all checks pass, the function is continuous at that point. If any check fails, you've found a discontinuity, and the way it fails tells you which type.

Limit-based approaches

For rigorous proofs, you use the $\epsilon$-$\delta$ definition directly:

1. Start with an arbitrary $\epsilon > 0$.

2. Find a $\delta > 0$ (which may depend on $\epsilon$ and the point $a$) such that $|x - a| < \delta$ implies $|f(x) - f(a)| < \epsilon$.

3. Prove that your chosen $\delta$ actually works by showing the implication holds.

This approach is standard in analysis courses and is the method used to prove uniform continuity (where $\delta$ must work for all points simultaneously).

Continuity vs differentiability

Relationship between concepts

The key relationship: every differentiable function is continuous, but not every continuous function is differentiable. Differentiability is strictly stronger than continuity.

Continuity means no breaks or jumps. Differentiability means the function also has a well-defined tangent line (a unique slope) at each point. A function can be perfectly continuous yet have a sharp corner or cusp where no single tangent line exists.

Examples of continuous non-differentiable functions

• Absolute value function $f(x) = |x|$: continuous everywhere, but has a sharp corner at $x = 0$ where the derivative doesn't exist (the left-hand derivative is $-1$, the right-hand derivative is $+1$).
• Cube root function $f(x) = \sqrt[3]{x}$: continuous everywhere, but has a vertical tangent at $x = 0$, so the derivative is undefined there.
• Weierstrass function: the famous example of a function that is continuous everywhere but differentiable nowhere. It's built as an infinite sum of cosine functions and shows that continuous functions can be far more "jagged" than intuition suggests.

Note: step functions are not continuous at their jump points, so they don't belong in a list of continuous but non-differentiable functions. The non-differentiability there comes from the discontinuity itself.

Historical development

Early concepts of continuity

Intuitive ideas about continuity go back to antiquity. Aristotle discussed continuous motion and magnitude, though without formal definitions. In the medieval period, scholars like Nicole Oresme studied continuous change and even graphed functional relationships. When Newton and Leibniz developed calculus in the 17th century, they relied on an intuitive sense that the functions they worked with behaved "nicely," but they never pinned down exactly what that meant. Euler and Lagrange pushed the understanding of functions further in the 18th century, but a rigorous definition of continuity was still missing.

Modern formalization

The push for rigor came in the 19th century. Bolzano (around 1817) and Cauchy (in the 1820s) gave early precise definitions of continuity using limit-based language. Weierstrass then developed the $\epsilon$-$\delta$ formulation that we still use today, putting continuity on a completely solid foundation. His construction of a continuous, nowhere-differentiable function in 1872 shocked the mathematical community and showed that continuous functions could behave in deeply counterintuitive ways. Cantor's set theory and the 20th-century development of topology and functional analysis extended the concept of continuity to abstract spaces far beyond the real number line.