Continuous Function

Overview

Continuity is the first precise bridge between pictures of functions and arguments about limits. It asserts that nearby inputs produce nearby outputs at a specified point, with a tolerance chosen to match the desired accuracy. The statement concerns the function together with its domain: the same formula can behave differently when the allowed inputs change.

Continuity neither requires a tangent nor bounds the speed of change. Its value is that it allows limits to pass through a function, turning information about an approximating sequence into information about its image.

Prerequisites

Understand absolute value, open balls, and the order of quantifiers in an error estimate. A metric $d$ measures distance and satisfies the triangle inequality. On the real line it is $d(x,y)=|x-y|$. Sequence limits provide a useful alternative formulation.

Definition and conventions

Let $(X,d_X)$ and $(Y,d_Y)$ be metric spaces and $f:X\to Y$. Continuity at $a\in X$ means that for every $\varepsilon>0$ there exists $\delta>0$ such that, for every $x\in X$,

The map is continuous if this holds at every point. The choice of $\delta$ may depend on both $a$ and $\varepsilon$. At an endpoint of an interval, inputs approach only through that interval. A missing endpoint is not a point at which continuity is tested.

Intuition

Think of $\varepsilon$ as an allowed output error. Continuity promises an input accuracy sufficient to meet it. It does not promise a universal calibration across the domain. A continuous graph may have a corner or extremely steep stretches; drawing without lifting a pencil is therefore only a limited mnemonic, especially for disconnected domains.

Examples and non-examples

For $f(x)=x^2$, fix $a$ and require $|x-a|<1$. Then $|x+a|\le 2|a|+1$, so

Choosing $\delta=\min(1,\varepsilon/(2|a|+1))$ proves continuity at $a$.

The function equal to $1$ on $[0,\infty)$ and $0$ on $(-\infty,0)$ is discontinuous at zero: negative inputs arbitrarily close to zero have output distance $1$ from its value there. In contrast, $1/x$ is continuous on $(0,\infty)$ despite being unbounded near the excluded endpoint.

Important properties

For maps between metric spaces, continuity is equivalent to preservation of convergent sequences. To prove the converse, if the error condition fails, choose $x_n$ within $1/n$ of $a$ whose output error stays above one fixed positive threshold. This contradicts sequence preservation.

Continuous images of compact sets are compact. Consequently a real continuous function on a nonempty compact domain attains its maximum and minimum. Connectedness supplies the intermediate value property; a general disconnected domain does not. Sums, products, and compositions preserve continuity wherever the operations are defined.

Connections and further reading

Uniform continuity strengthens the quantifier order, while differentiability adds a linear approximation. Uniform convergence preserves continuity of a sequence of continuous functions. Taylor, §3.1 develops the metric formulation and its compactness consequences. Read these results with their domain hypotheses visible.