This optional section explains why the notation looks like a fraction squared. The second derivative can be understood as the limit of a second-increment quotient, which motivated Leibniz's symbolic choice.
| Symbol | Meaning |
|---|---|
| First increment of | |
| Second increment of | |
| Second-increment quotient | |
| Limit of second-increment quotient |
The First Derivative as a Limit of Increment Quotients
Let . Recall:
where with and .
The Second Derivative as a Limit of Second-Increment Quotients
Consider three equally spaced points:
with corresponding values , , .
The second increment quotient is the "increment of the increment quotient":
\frac{1}{\Delta x}\left(\frac{y_2-y_1}{\Delta x}-\frac{y_1-y_0}{\Delta x}\right)=\frac{y_2-2y_1+y_0}{(\Delta x)^2}\tag{a}We call the second increment of and write it (the symbol is not a square; it denotes the difference of differences).
The second-increment quotient is therefore . The second derivative is the limit:
f''(x)=\lim_{\Delta x\to0}\frac{\Delta^2 y}{(\Delta x)^2}(If the second derivative is continuous, this can be proved equivalent to the definition f''=(f')'.)
Leibniz's Notation
This symbolism motivated Leibniz to introduce:
y''=f''(x)=\frac{d^2y}{dx^2},\qquad y'''=f'''(x)=\frac{d^3y}{dx^3},\qquad y^{(n)}=f^{(n)}(x)=\frac{d^ny}{dx^n}The numerator suggests "increment of the increment of " and the denominator is the square of the spacing.
Connection to Numerical Methods
Formula (a) is the basis of the finite difference method, a numerical technique for approximating differential equations. In that method:
f''(x)\approx\frac{f(x+h)-2f(x)+f(x-h)}{h^2}for small step size . This formula replaces the second derivative with computable arithmetic on function values and is widely used in engineering and scientific computing.
Notes
- The fact that the second derivative equals the limit of the second-increment quotient (not just the derivative of the first derivative) requires proof when the second derivative is continuous.
- The two definitions (limit of first derivative of first derivative, vs. limit of second-increment quotient) are equivalent whenever the second derivative is continuous.