How Do You Know If A Function Is Differentiable

Ever sat in a calculus lecture, staring at a curve on a chalkboard, and felt that sudden, sharp disconnect? You see a line that looks smooth, or maybe a curve that looks perfectly fine, and then the professor asks if it's differentiable. Suddenly, the graph isn't just a shape anymore; it's a puzzle of limits, slopes, and hidden traps.

It’s one of those concepts that feels intuitive until you actually have to prove it. In practice, you know a function is "smooth," right? But in math, "smooth" has a very specific, very strict meaning that doesn't care about how your eyes perceive the graph.

What Is Differentiability

If you want to understand differentiability, forget the textbook definition for a second. Now, think about motion. Imagine you are driving a car along a winding road. If you can determine your exact speed at any single, precise moment, that's differentiability. You aren't just looking at your average speed over a minute; you're looking at the instantaneous rate of change.

In plain language, a function is differentiable at a certain point if it has a well-defined tangent line at that point. In real terms, this means the slope doesn't suddenly jump, disappear, or become infinite. It means the function is behaving predictably.

The Connection to Continuity

Here is the first big rule you need to wrap your head around: continuity is a prerequisite. If a function has a hole, a jump, or a vertical asymptote at a specific point, it is absolutely not differentiable there. You can't find the slope of a point that doesn't even exist or a point where the graph suddenly teleports from one height to another.

But—and this is the part that trips everyone up—being continuous doesn't guarantee differentiability. Because of that, you can have a perfectly connected line that is still "undifferentiable. " This is where the real work begins.

The Limit Definition

When we get technical, we look at the limit of the difference quotient. We are essentially asking: as we zoom in closer and closer to a point, does the slope of the secant line (the line connecting two points) settle down into one single, specific number? If that limit exists and is a finite number, you've got differentiability. If the limit fails to settle—if it goes to infinity or if the left side doesn't match the right side—the function fails the test.

Why It Matters

Why do we spend so much time worrying about this? Because calculus is the study of change, and change is measured through derivatives.

If you are an engineer designing a roller coaster, you need to know that the tracks are differentiable. If a track has a "sharp" point—a cusp or a corner—the sudden change in direction would create an infinite force at that exact moment. The coaster wouldn't just bump; it would theoretically experience a mathematical impossibility that would tear a real vehicle apart.

In data science and machine learning, differentiability is the engine under the hood. Most optimization algorithms, like gradient descent, rely on derivatives to find the lowest point of a loss function. If the function isn't differentiable, the algorithm might get stuck or simply fail to know which direction to move. Understanding where these "breaks" in differentiability occur helps researchers design better models.

How to Determine If a Function Is Differentiable

So, how do you actually do it? You can't just look at a graph and guess, especially when the functions get complex. You need a systematic approach.

Check for Continuity First

Before you do anything else, check if the function is continuous at the point in question. Now, if you're looking at a piecewise function, check if the pieces meet at the same $y$-value. If they don't, stop right there. Think about it: you're done. The function is not differentiable. It’s a quick win that saves you from doing unnecessary heavy lifting.

Look for Sharp Corners (Cusps)

This is the most common way a continuous function fails to be differentiable. " It is continuous—the lines meet. At $x = 0$, the graph looks like a perfect "V.Now, imagine the absolute value function, $f(x) = |x|$. But if you try to find the slope at the very bottom, you run into a problem.

From the left, the slope is $-1$. From the right, the slope is $+1$. Because the slope from the left doesn't match the slope from the right, the limit doesn't exist. The function is "pointy." In math terms, we say the derivative is undefined at a corner or a cusp.

Watch Out for Vertical Tangents

Sometimes a function is smooth and doesn't have any sharp corners, but it still fails the test. This happens when the graph becomes perfectly vertical for a split second.

Think about the cube root function, $f(x) = \sqrt[3]{x}$, at $x = 0$. Since the slope of a vertical line is undefined (it's essentially "infinite"), the derivative does not exist at that point. As you approach zero, the curve gets steeper and steeper until, for one infinitesimal moment, it is a vertical line. The function is continuous, but it isn't differentiable.

The Formal Limit Test

If you're dealing with a complex piecewise function and you can't visually "see" the problem, you have to go back to the math. You must calculate the left-hand derivative and the right-hand derivative.

For more on this topic, read our article on what are the least common multiples of 3 and 4 or check out acid and base combine to form.

Calculate the limit as you approach the point from the left.
Calculate the limit as you approach the point from the right.
Compare them.

If they are equal and finite, you are golden. If they are different, or if they blow up to infinity, the function is not differentiable at that point.

Common Mistakes / What Most People Get Wrong

I've seen students (and even seasoned pros) fall into the same traps. Here is what usually goes wrong.

One major mistake is assuming that continuity implies differentiability. It is a one-way street. If you see a graph that is connected, don't automatically assume you can take its derivative everywhere. All differentiable functions are continuous, but not all continuous functions are differentiable. Always check for those sharp turns or vertical climbs.

Another mistake is getting lost in the algebra of limits. People often try to jump straight into complex derivative rules (like the product rule or chain rule) without checking if the function is even continuous at the point of interest. That said, if the function isn't continuous, the rules don't apply. You're essentially trying to calculate the speed of a car that has just teleported from one side of the road to the other.

Lastly, people often forget to check the domain. A function can't be differentiable anywhere it doesn't exist. It sounds obvious, but when you're deep in a complex problem involving logarithms or square roots, it's easy to forget that the function might have "dead zones" where it simply isn't defined.

Practical Tips / What Actually Works

If you're working through a problem set or analyzing a real-world data model, here is my advice for staying sane.

Visualizing is a great start, but it's not a proof. Use a graphing tool to get a "feel" for the function. If you see a sharp point or a vertical section, you've likely found your answer. But if you're writing a formal proof, you must use the limit definition. Don't rely on "it looks smooth" as a mathematical argument.

Master the piecewise approach. Most "tricky" differentiability problems involve piecewise functions. When you see these, always check the "junction" points. That is almost always where the differentiability breaks. Check the $y$-values first (continuity), then check the slopes (differentiability).

Keep an eye on the "power" of your terms. When looking at functions like $x^n$, remember that the exponent matters. If you have a fractional exponent, you're much more likely to encounter vertical tangents or cusps. If you have an absolute value, you're looking for corners.

FAQ

Can a function be differentiable but not continuous?

No. This is a fundamental rule. If a function is differentiable at a point, it must be continuous there. If it's not continuous, it's impossible for it to be differentiable.

What is a "

What is a "cusp" and how does it affect differentiability?

A cusp is a specific type of non-differentiable point where the function's slope approaches infinity from one side and negative infinity from the other. Visually, it looks like a very sharp, pointed "V" shape, but the sides are curved. While the function might be continuous, the derivative fails to exist because the limit of the difference quotient does not result in a finite number.

Why does the limit definition matter if I know the shortcut rules?

Shortcut rules (like the power rule or quotient rule) are actually derived from the limit definition. These rules are "shortcuts" only because they assume the function is already behaving nicely. If you apply the power rule to a function at a point where the limit definition fails, you will get a numerical answer that is mathematically invalid. The limit definition is your safety net; it is the ultimate authority on whether a derivative actually exists.

Can a function be continuous everywhere but differentiable nowhere?

Yes. While it sounds impossible, functions like the Weierstrass function exist. These functions are so incredibly "wiggly" and jagged at every single scale that they never settle down into a smooth slope, even though you can draw them without lifting your pen. They are the ultimate counter-example to our intuition.

Conclusion

Understanding differentiability is less about memorizing formulas and more about developing a sense of "mathematical smoothness." It requires a disciplined, step-by-step approach: first, ensure the function is connected (continuity); second, ensure the transition is smooth (no sharp turns); and third, ensure the slope is finite (no vertical tangents).

By moving beyond simple algebraic manipulation and embracing the underlying logic of limits, you will stop falling into common traps and start seeing the true landscape of calculus. Remember: continuity is the foundation, but differentiability is the refinement. Master the relationship between the two, and you will master the core of calculus. Simple, but easy to overlook.