Why continuity is a necessary but not sufficient condition for a derivative to exist, using a sharp-corner function as the classic counterexample, and what this means for correctly applying derivative rules.
Why This Distinction Is Worth Making Explicit
It is tempting to assume that any function without breaks or jumps — any function that can be drawn "without lifting the pen," an informal description of continuity — automatically has a well-defined derivative everywhere. This assumption is incorrect, and understanding exactly why reveals something important about what the limit definition of the derivative, which underlies the power rule this site's Derivative & Limit Evaluator applies, actually requires beyond mere continuity.
Why Continuity Is a Necessary Condition for Differentiability
It is true, and provable directly from the limit definitions involved, that if a function is differentiable at a point, it must also be continuous there — differentiability is a strictly stronger condition that implies continuity as a consequence. This one-directional implication is why continuity is correctly described as a necessary condition for differentiability: a function cannot have a derivative at a point unless it is at least continuous there.
Why Necessary Does Not Mean Sufficient
The relationship, however, only runs in one direction — continuity does not, by itself, guarantee that a derivative also exists. A condition being necessary for a result (required for the result to be possible) is a logically different, weaker statement than the condition being sufficient for that result (guaranteeing the result on its own) — this is a distinction from formal logic that shows up constantly across calculus, and the continuity/differentiability relationship is one of the clearest, most commonly cited examples of it.
The Classic Counterexample: f(x) = |x| at x = 0
The absolute value function f(x) = |x| is continuous everywhere, including at x = 0 — there is no jump, no gap, no break in the graph at that point. However, f(x) = |x| is not differentiable at x = 0, because it has a sharp corner there rather than a smooth, well-defined tangent line — this single function is the standard textbook counterexample precisely because it demonstrates continuity without differentiability in the simplest possible way.
Why the Sharp Corner Breaks the Limit Definition Specifically
Examining the one-sided derivative limits at x = 0 for f(x) = |x| reveals exactly why: approaching from the right (positive h), the difference quotient [f(0+h) − f(0)] / h simplifies to h/h = 1 for every positive h; approaching from the left (negative h), it simplifies to −h/h = −1 for every negative h. These two one-sided derivative limits — 1 from the right, −1 from the left — do not agree, so, exactly as with the two-sided limit of a function requiring both directional approaches to agree, the two-sided derivative limit at x = 0 does not exist, even though the function itself remains perfectly continuous there.
Why This Connects Directly Back to the Two-Sided Limit Concept
This is precisely the same "both directional approaches must agree" requirement that governs whether a two-sided limit of a function exists — except applied here to the difference quotient itself, rather than to the function directly. A sharp corner produces two different slopes approaching from each direction, so the derivative's own limiting process fails the agreement test even while the underlying function's limiting process (continuity) passes it without issue.
Why This Matters for Correctly Applying Differentiation Rules in Practice
Power functions of the form this site's calculator evaluates, f(x) = a·xⁿ for real exponents, are smooth (differentiable) everywhere in their domain except at genuine singularities like x = 0 with a negative exponent — they do not have sharp corners, so the power rule applies without this particular caveat across their normal domain. But recognizing that continuity alone is not sufficient for differentiability is exactly the reasoning needed when working with piecewise-defined functions, absolute values, or other non-smooth functions elsewhere in an engineering or applied-math context, where blindly applying a differentiation rule without checking for corners, cusps, or vertical tangents can produce a nonsensical or simply incorrect result.