Differentiable function

In mathematics, a differentiable function is a type of function that is remarkably smooth.

If you look at the graph of a differentiable function, you will notice it has no breaks, sharp angles, or sudden points. At every interior point in its domain, a derivative exists. This derivative tells us the rate of change at that specific location. Visually, this means the graph has a non-vertical tangent line at every point. A tangent line is a straight line that just touches the curve at a single point.

Because these lines exist everywhere, we can use them to approximate the function locally. This means that if you zoom in very closely on a point, the curve begins to look like a simple straight line.

To understand the mechanism of differentiability, we must look at the relationship between the function and its derivative. For a function of one real variable, differentiability at a specific point implies that the function is also continuous at that point. Continuity means there are no gaps or jumps in the graph. However, the reverse is not always true. A function can be continuous without being differentiable. For example, the absolute value function is continuous because it has no gaps.

Yet, it is not differentiable at the point where it crosses the axis. This is because the graph makes a sharp bend there. At such a bend, you cannot define a single, unique tangent line. Other shapes, like a cusp, also prevent a function from being differentiable.

At a cusp, the function is continuous but fails to be differentiable due to the sharp point.

Mathematicians organize these functions into different levels called differentiability classes. These classes are defined by how many times you can take a derivative and still have a continuous result. A function is said to be continuously differentiable if its first derivative exists and is itself a continuous function. We use the term $C^1$ to describe these functions. If both the first and second derivatives exist and are continuous, the function is of class $C^2$. We can continue this pattern for more derivatives. If the derivatives exist for all positive integers, the function is called a smooth function, or a function of class $C^{\infty}$. Some functions are differentiable but fail to be continuously differentiable.

An example is a function where the derivative exists but has an essential discontinuity, meaning it does not behave smoothly.

While we often work with smooth functions in science, they are actually quite rare in the mathematical sense. A mathematician named Stefan Banach provided a surprising result regarding this. He stated that the set of functions that have a derivative at some point is a meagre set within the space of all continuous functions. In simpler terms, most continuous functions are not differentiable at any point. The most famous example of this extreme case is the Weierstrass function. This function is continuous everywhere, meaning you can draw it without lifting your pencil. However, it is differentiable nowhere. It is so jagged that it never looks like a straight line, no matter how much you zoom in.

We can also extend the idea of differentiability to functions with many variables. For a multivariable function, differentiability is more complex than just finding partial derivatives. A partial derivative measures how a function changes as you move along one specific axis. Even if all partial derivatives exist, the function might not be differentiable. For a multivariable function to be truly differentiable, there must be a linear map that approximates the function. This map is represented by the Jacobian matrix, which is an $n \times m$ matrix. If all the partial derivatives exist in a neighborhood and are continuous at a point, then the function is guaranteed to be differentiable at that point.

Another specialized area is complex analysis, which uses complex numbers. A function that is complex-differentiable at a point is called holomorphic. This is a much stricter requirement than being differentiable in real numbers. If a function is holomorphic in a neighborhood, it is automatically infinitely differentiable and analytic. This means it is incredibly well-behaved. Interestingly, a function can be differentiable as a multi-variable real function without being complex-differentiable. This happens if the limit used to find the derivative gives different values depending on how you approach the point.

Finally, these concepts apply to even more advanced structures called manifolds. If we have a differentiable manifold, a function on that manifold is considered differentiable if it is differentiable according to its coordinate charts. This allows mathematicians to apply the rules of calculus to curved, complex spaces. Whether studying simple lines or complex multidimensional shapes, differentiability remains a core tool for understanding how the world changes.