Pointwise vs Uniform Continuity

Understanding the subtle but crucial distinction between pointwise and uniform continuity is a milestone in advanced calculus and real analysis. It marks the transition from viewing a function's behavior at isolated points to grasping its global behavior across an entire domain. This distinction underpins critical theorems about integration, approximation, and the convergence of function sequences, making it indispensable for rigorous mathematical work.

The Foundational Concept of Continuity at a Point

Before contrasting the two types, you must solidify the standard definition of continuity. A function $f$ is continuous at a point $c$ in its domain if, for every arbitrarily small positive distance $ϵ$, you can find a corresponding tolerance $\delta$ such that all points within $\delta$ of $c$ are mapped within $ϵ$ of $f(c)$. Formally: for every $ϵ>0$, there exists a $\delta >0$ such that $|x-c|<\delta$ implies $|f(x)-f(c)|<ϵ$.

This is a pointwise definition. The chosen $\delta$ can depend on both the challenge $ϵ$ and the specific point $c$. For a function to be pointwise continuous on a set $A$, it must satisfy this condition separately at every individual point $c\in A$. The $\delta$ for one point might be large, while for another, more sensitive point, it might be vanishingly small. Consider $f(x)=1/x$ on $(0,1]$. It is continuous at every point in its domain. However, as you choose points $c$ closer and closer to 0, the function becomes steeper, forcing you to pick smaller and smaller $\delta$ values to meet the same $ϵ$ challenge, even though 0 itself is not in the domain.

The Stronger Condition of Uniform Continuity

Uniform continuity strengthens the pointwise condition by removing the dependence of $\delta$ on the specific point. A function $f$ is uniformly continuous on a set $A$ if, for every $ϵ>0$, there exists a single $\delta >0$ that works for all points $c$ in $A$. The formal statement changes subtly but profoundly: for every $ϵ>0$, there exists a $\delta >0$ such that for all $x,y\in A$, if $|x-y|<\delta$, then $|f(x)-f(y)|<ϵ$.

Notice the universal quantifiers: "for all $x,y$." This means the function's output values cannot change too rapidly anywhere in the domain if the inputs are sufficiently close. The $\delta$ is a global property of the function on the set $A$. If a function is uniformly continuous on a set, it is automatically continuous at every point of that set. However, the converse is not always true. Uniform continuity is a global property of a function on a set, while pointwise continuity is a local property collected at each point.

Illustrating the Difference with Key Examples

The classic example that separates the two concepts is $f(x)=x^2$ on different domains.

1. On a closed, bounded interval like $[0,3]$: The function is uniformly continuous. You can find one $\delta$ (depending on $ϵ$) that works for every point in the interval. The function's rate of change is bounded here.
2. On an unbounded domain like $[0,\infty )$: Here, $f(x)=x^2$ is pointwise continuous but not uniformly continuous. Why? Imagine an $ϵ$ challenge of 1. For a point at $x=M$, a very large number, the function's slope is $2M$. To ensure $f(x)$ stays within 1 of $f(M)$, you'd need $|x-M|<\delta \approx 1/(2M)$. As $M$ grows, the required $\delta$ shrinks to zero, making it impossible to choose one $\delta$ that works for all points $M$ on the infinite domain.

Another revealing example is $g(x)=1/x$ on $(0,1)$, an open, bounded interval. It is continuous at every point but not uniformly continuous. The function becomes infinitely steep as $x\to 0^{+}$, so no single $\delta$ can control the output variation near the boundary for a fixed $ϵ$.

The Heine-Cantor Theorem: A Key Guarantee

A profound result that resolves when you can promote pointwise continuity to uniform continuity is the Heine-Cantor theorem. It states: If a function is continuous on a compact set, then it is uniformly continuous on that set.

In the context of real numbers, a compact set is one that is both closed and bounded (by the Heine-Borel theorem). This theorem explains why $x^2$ was uniformly continuous on $[0,3]$ (compact) but not on $[0,\infty )$ (not bounded) or $(0,1)$ (not closed). The proof, often sketched using sequences or open covers, leverages the fact that every open cover of a compact set has a finite subcover. This finiteness allows you to select a single, globally effective $\delta$ from the local $\delta$'s that exist at each point. The Heine-Cantor theorem is a powerful tool because it provides a clean, topological condition (compactness) that ensures uniform continuity without needing to wrestle with epsilon-delta arguments directly.

Counterexamples on Non-Compact Domains

The necessity of compactness in the Heine-Cantor theorem is shown by the counterexamples we've already touched on. They fall into two main categories related to failures of compactness:

1. Unbounded Domains: Functions whose derivative or rate of change becomes unbounded, like $x^2$ on $[0,\infty )$ or $x$ on $\mathbb{R}$, can fail uniform continuity. However, note that $f(x)=x$ is actually uniformly continuous on $\mathbb{R}$ (choose $\delta =ϵ$), showing that unboundedness alone doesn't guarantee failure—it's the unbounded rate of change that matters.
2. Open or Non-Closed Domains: The primary issue here is behavior near a boundary point that is not included in the domain. The function $1/x$ on $(0,1)$ is the canonical example. It is continuous at every point inside, but its asymptote at $x=0$ (a boundary point not in the set) prevents the selection of a uniform $\delta$ for points arbitrarily close to 0. The domain is not closed, hence not compact.

Common Pitfalls

1. Confusing the Order of Quantifiers: The most common error is misremembering the logical structure. Pointwise continuity: $\forall ϵ>0,\forall c,\exists \delta ...$. Uniform continuity: $\forall ϵ>0,\exists \delta ,\forall c...$. Swapping the "$\exists \delta$" and "$\forall c$" statements changes the meaning entirely.
2. Assuming Continuity Implies Uniform Continuity on Any Domain: After seeing the Heine-Cantor theorem, it's easy to forget its specific condition. A function continuous on $(0,1)$ or $[1,\infty )$ is not guaranteed to be uniformly continuous there. Always check the domain's properties.
3. Misapplying Visual Intuition: A function with a vertical asymptote clearly fails uniform continuity, but a function with a continuously increasing slope (like $x^2$ on $[0,\infty )$) can also fail. The graphical intuition is that for a uniformly continuous function, if you draw a rectangle of height $2ϵ$ and slide it along the graph, you can always find a width $2\delta$ such that the graph enters the rectangle on one side and doesn't leave until the other side, for all starting points.
4. Overlooking Simple Uniformly Continuous Functions on Unbounded Domains: Do not assume an unbounded domain always breaks uniform continuity. The linear function $f(x)=ax+b$ is uniformly continuous on $\mathbb{R}$ ($\delta =ϵ/|a|$ works globally). The condition is about controlling the difference $|f(x)-f(y)|$, not the values themselves.

Summary

• Pointwise continuity is a local property defined at each individual point, where the chosen $\delta$ can depend on both $ϵ$ and the point $c$.
• Uniform continuity is a stronger, global property on a set where a single $\delta$ must work for a given $ϵ$ for all possible point pairs in the set.
• The Heine-Cantor theorem provides a critical guarantee: continuity on a compact (closed and bounded) set implies uniform continuity on that set.
• Counterexamples of continuous but non-uniformly continuous functions typically occur on domains that are either unbounded (allowing for unbounded rates of change) or not closed (allowing for asymptotic behavior at missing boundary points).
• The logical order of quantifiers—$\forall c,\exists \delta$ (pointwise) versus $\exists \delta ,\forall c$ (uniform)—encapsulates the core conceptual difference. Mastering this distinction is fundamental to advanced analysis.