Introduction

Most of us first encounter the exponential function \( e^x \) in one of two ways:

  1. As the unique function that is its own derivative: \( f’(x) = f(x) \)
  2. As the limit: \( e^x = \lim_{n \to \infty} (1 + x/n)^n \)

But there’s another, perhaps more intuitive way to think about exponentials: they are the functions that look the same at every scale.

What does this mean? Imagine zooming into different parts of the exponential curve. No matter where you look, no matter how much you zoom in or out, the shape you see is always the same—just stretched or compressed vertically. This remarkable property turns out to be completely equivalent to the derivative definition, and we can prove it using only high school mathematics.

What Does “Self-Similar” Mean?

Let’s make this precise. We say a function \( f(x) \) is self-similar under horizontal shifts if:

$$f(x + c) = k \cdot f(x)$$

for some constant \(k\) (that may depend on \(c\)).

In other words, when you shift the graph horizontally by some amount \(c\), you get exactly the same shape, just scaled vertically by some factor \(k\).

Let’s check if this holds for \( f(x) = e^x \):

$$f(x + c) = e^{x+c} = e^x \cdot e^c = e^c \cdot f(x)$$

Yes! The exponential function satisfies this property perfectly, with \( k = e^c \).

Why Is This Special?

Let’s compare with other functions:

Quadratic functions: \( f(x) = x^2 \)

  • \( f(x + 1) = (x + 1)^2 = x^2 + 2x + 1 \)
  • This is NOT just a vertical scaling of \( x^2 \)

Linear functions: \( f(x) = mx + b \)

  • \( f(x + c) = m(x + c) + b = mx + (mc + b) \)
  • This shifts vertically by a constant, not a scaling

Sine function: \( f(x) = \sin(x) \)

  • \( f(x + \pi/2) = \sin(x + \pi/2) = \cos(x) \)
  • This gives us a completely different function!

The exponential function is unique in this regard (along with its scalar multiples).

The Connection to Derivatives

Here’s where it gets beautiful. This self-similarity property forces the function to be its own derivative. Let me show you why.

Suppose \( f(x + c) = k(c) \cdot f(x) \) for all \(x\) and \(c\), where \( k(c) \) is some function of \(c\).

Let’s use the definition of the derivative:

$$f’(x) = \lim_{h \to 0} \frac{f(x + h) - f(x)}{h}$$

Using our self-similarity property with \( c = h \):

$$\begin{align} f’(x) &= \lim_{h \to 0} \frac{k(h) \cdot f(x) - f(x)}{h} \ &= \lim_{h \to 0} f(x) \cdot \frac{k(h) - 1}{h} \ &= f(x) \cdot \lim_{h \to 0} \frac{k(h) - 1}{h} \end{align}$$

Notice that the limit \( \lim_{h \to 0} \frac{k(h) - 1}{h} \) is just a number (it doesn’t depend on \(x\)). Let’s call this number \( \lambda \):

$$f’(x) = \lambda \cdot f(x)$$

So self-similarity implies that the derivative is proportional to the function itself!

Finding the Constant

What is this constant \( \lambda \)? Well, we need one more piece of information. Notice that:

$$k(0) = \frac{f(x + 0)}{f(x)} = \frac{f(x)}{f(x)} = 1$$

So \( k(h) \) starts at 1 when \( h = 0 \). Therefore:

$$\lambda = \lim_{h \to 0} \frac{k(h) - 1}{h} = k’(0)$$

This is the derivative of \(k\) at 0. For \( f(x) = e^x \), we found \( k(c) = e^c \), so:

$$\lambda = \left.\frac{d}{dc}[e^c]\right|_{c=0} = e^0 = 1$$

This is why \( e^x \) is its own derivative! The base \(e\) is special because it makes \( \lambda = 1 \).

Other Bases

What about other exponential functions like \( 2^x \) or \( 10^x \)?

For \( f(x) = a^x \): $$ f(x + c) = a^{x+c} = a^c \cdot a^x = a^c \cdot f(x) $$

So \( k(c) = a^c \) and

$$ \lambda = \left.\frac{d}{dc}[a^c]\right|_{c=0} = a^c \cdot \ln(a)\bigg|_{c=0} = \ln(a) $$

This gives us:

$$f’(x) = \ln(a) \cdot f(x) \quad \text{for } f(x) = a^x$$

The base \(e\) is special because \( \ln(e) = 1 \), making the function exactly equal to its own derivative.

Conclusion

The exponential function’s self-similarity—the fact that it looks the same at every scale—is not just a curiosity. It’s the fundamental property that makes exponentials so important in mathematics and nature.

When something grows proportionally to its current size (like populations, compound interest, or radioactive decay), it must follow an exponential law. This is why exponentials appear everywhere: they are the mathematical expression of self-similar growth.

The next time you see \( e^x \), don’t just think of it as a function with a special derivative. Think of it as the function that never changes its shape—only its scale.

Note: If you want to be even more rigorous, you can show that if \( f(x+c) = k(c) \cdot f(x) \), then \(k\) must satisfy \( k(a+b) = k(a) \cdot k(b) \), which forces \( k(c) = e^{\lambda c} \) for some constant \( \lambda \). This is a beautiful result but requires a bit more work to prove carefully.