Historical Context & Motivation
The relationship between differentiability and continuity was one of the most debated topics in the development of modern analysis. For much of the eighteenth century, mathematicians such as Euler and Lagrange implicitly assumed that any "reasonable" function could be differentiated, because the functions they encountered—polynomials, trigonometric expressions, exponentials—were all smooth. This comfortable assumption persisted until the nineteenth century, when a series of surprising counterexamples forced the mathematical community to rigorously define what it means for a function to be continuous and, separately, differentiable.
The central question this lesson addresses is deceptively simple: If a function is continuous at a point, can we always compute its derivative there? The answer is a definitive no, and understanding exactly when and why derivatives fail to exist is essential for success on the AP Calculus BC exam. Mastering this topic requires you to internalize both the logical relationship between continuity and differentiability and the geometric features of graphs where the derivative does not exist.
Core Principles & Definitions
Before exploring the interplay between differentiability and continuity, we must establish precise definitions. A function f is continuous at x = a if three conditions hold: f(a) is defined, lim(x→a) f(x) exists, and lim(x→a) f(x) = f(a). A function is differentiable at x = a if the limit of the difference quotient, lim(h→0) [f(a + h) − f(a)] / h, exists as a finite number. Differentiability at a point is a strictly stronger condition than continuity, meaning that every differentiable function is continuous, but many continuous functions are not differentiable at certain points.
Differentiability ⟹ Continuity
Continuity ⟹̸ Differentiability
Discontinuity ⟹ Non-Differentiability
Left and Right Derivatives Must Agree
Visual Explanation: Where Derivatives Fail
The diagram below illustrates four classic scenarios where a function is continuous at a point but fails to be differentiable. Each panel shows a different geometric pathology: a corner (where left and right derivatives exist but differ), a cusp (where one or both one-sided derivatives tend to ±∞), a vertical tangent (where the tangent line is vertical, giving an infinite slope), and a jump discontinuity (where continuity itself fails, automatically precluding differentiability).
In each of the first three panels, the function is continuous at the marked point—there is no break in the graph—yet the derivative fails to exist because the difference quotient does not converge to a single finite limit. Geometrically, the derivative at a point gives the slope of the unique tangent line. When the graph has a sharp corner, two competing tangent directions exist; when a cusp or vertical tangent appears, the slope tends to infinity. These distinctions are tested frequently on the AP exam, often in conjunction with piecewise-defined functions.
Mathematical Framework
The Derivative as a Limit
Proof: Differentiability Implies Continuity
This is one of the most elegant short proofs in introductory analysis and appears regularly on the AP exam as a conceptual question. Suppose f is differentiable at x = a, so f′(a) exists. We wish to show that lim(x→a) f(x) = f(a), which is the definition of continuity at a.
One-Sided Derivatives
Classifying Points of Non-Differentiability
Points where f′(a) fails to exist can be systematically classified based on the behavior of the difference quotient. This classification helps you quickly identify the type of failure on an exam, whether you are reading a graph, analyzing a formula, or inspecting tabular data. The SVG below organizes these types into a decision flowchart, and the table that follows provides a compact reference.
| Type | Continuity? | Left Derivative | Right Derivative | Example |
|---|---|---|---|---|
| Corner | Yes | Finite (e.g., −1) | Finite but different (e.g., +1) | f(x) = |x| at x = 0 |
| Cusp | Yes | → +∞ (or −∞) | → −∞ (or +∞) | f(x) = x²ᐟ³ at x = 0 |
| Vertical Tangent | Yes | → +∞ | → +∞ | f(x) = x¹ᐟ³ at x = 0 |
| Discontinuity | No | N/A | N/A | Piecewise with a jump |
Worked Example: Piecewise Differentiability
Consider the piecewise function f(x) = { x² + 1, for x ≤ 2 ; 4x − 3, for x > 2 }. Determine whether f is differentiable at x = 2.
Strengths & Limitations of the Differentiability–Continuity Link
| Aspect | Strengths | Limitations / Pitfalls |
|---|---|---|
| Logical implication | Differentiability ⟹ Continuity provides a fast discontinuity check: if f is not continuous, f is not differentiable. | The converse is false. Students often assume continuity guarantees differentiability, leading to errors on piecewise and absolute-value questions. |
| Graphical analysis | Corners, cusps, and vertical tangents are visually recognizable, making graph-based differentiability questions approachable. | Subtle cusps or near-corners can be hard to identify on low-resolution graph displays or in calculator-drawn plots. |
| One-sided derivatives | Comparing left and right derivatives gives a clean, mechanical procedure for piecewise functions—highly reliable. | This technique requires that you first establish continuity; skipping that step can produce misleading results when the pieces don't connect. |
| Algebraic verification | The difference-quotient definition provides a rigorous criterion that works even when graphing is impractical. | Computing the limit from the definition can be algebraically tedious; shortcuts (derivative rules) are faster once differentiability is established. |
Connection to Higher-Order Differentiability and Smoothness
The differentiability–continuity relationship is the first rung on a ladder of smoothness conditions that becomes increasingly important in BC topics such as Taylor and Maclaurin series, parametric and polar differentiation, and solving differential equations. A function that is differentiable once may still have a derivative that itself is discontinuous or non-differentiable. The class of functions whose first n derivatives all exist and are continuous is denoted Cⁿ, and functions that are infinitely differentiable are labeled C∞. Taylor series require C∞ behavior in a neighborhood of the expansion point, and the radius of convergence reflects how far that smoothness extends.
| Smoothness Class | Requirement | AP Calculus BC Relevance |
|---|---|---|
| C⁰ (continuous) | f is continuous on its domain | Intermediate Value Theorem, Extreme Value Theorem |
| C¹ (differentiable, cont. derivative) | f′ exists and is continuous | Mean Value Theorem, curve sketching, optimization |
| C² (twice differentiable) | f″ exists and is continuous | Concavity analysis, second derivative test, Euler's method |
| C∞ (smooth) | All derivatives exist and are continuous | Taylor/Maclaurin series, Lagrange error bound |
As you progress through the BC curriculum, keep in mind that every theorem involving derivatives—the Mean Value Theorem, L'Hôpital's Rule, integration by parts—carries implicit differentiability hypotheses. Failing to verify these hypotheses before applying a theorem is one of the most common errors on the AP exam's free-response section. The habit of checking differentiability and continuity conditions before applying calculus tools will serve you well not only on the exam but in any future mathematics or engineering coursework.
Practice Problems
Lesson Summary
This lesson established the fundamental one-way implication: differentiability implies continuity, but continuity does not imply differentiability. The proof relies on rewriting f(x) − f(a) as the product of the difference quotient and (x − a), then taking the limit. We identified four key failure modes for differentiability: corners (where left and right derivatives are finite but unequal), cusps (where one-sided derivatives tend to ±∞ in opposite directions), vertical tangents (where slopes tend to +∞ or −∞ from both sides), and discontinuities (where even continuity fails).
For piecewise functions, always follow a two-step protocol: first verify continuity at the boundary by checking that both one-sided limits equal the function value, then compute and compare the left-hand and right-hand derivatives. This framework extends naturally into higher-order smoothness classes (C¹, C², C∞) that underpin Taylor series, the Mean Value Theorem, and other advanced BC topics.