We saw in Theorem 2.3.10 Combinations of Derivatives that when we want to find the derivative of a sum or difference of functions, we can just find the derivatives of each function separately, and then combine the derivatives back together (by adding or subtracting). This, hopefully, is pretty intuitive: of course a slope of a sum of things is just the slopes of each thing added together!
In this section, we want to think about derivatives of product and quotients of functions. What happens when we differentiate a function that is really just two functions multiplied together?
Let’s just think about \(\Ddx{x^3}\) for a moment. What is\(x^3\text{?}\) Can you write this as a product? Call one of your functions \(f(x)\) and the other \(g(x)\text{.}\) You should have that \(x^3 = f(x)g(x)\) for whatever you used.
Use all of this to reassure yourself that even though the derivative of a sum of functions is the sum of the derivatives of the functions, we will need to develop a better understanding of how the derivatives of products of functions work.
We’ll use an area model for multiplication here: we can consider a rectangle where the side lengths are functions \(f(x)\) and \(g(x)\) that change for different values of \(x\text{.}\) Maybe \(x\) is representative of some type of time component, and the side lengths change size based on time, but it could be anything.
Find the area of the two rectangles. The second rectangle is just one where the input variable for the side length has changed by some amount, leading to a different side length.
In order to do this, we’ve broken the region up into three pieces. Calculate the areas of the three pieces. Use this to fill in the following equation:
We do not want to calculate the total change in area: a derivative is a rate of change, so in order to think about the derivative we need to divide by the change in input, \(\Delta x\text{.}\)
We’ll also want to think about this derivative as an instantaneous rate of change, meaning we will consider a limit as \(\Delta x \to 0\text{.}\) Fill in the following:
In any limit with \(f(x)\) or \(g(x)\) in it, notice that we can factor part out of the limit, since these functions do not rely on \(\Delta x\text{,}\) the part that changes in the limit. Factor these out.
In the third limit, factor out either \(\displaystyle \lim_{\Delta x\to 0} \left(f(x+\Delta x) - f(x)\right)\) or \(\displaystyle \lim_{\Delta x\to 0} \left(g(x+\Delta x) - g(x)\right)\text{.}\)
\begin{align*}
\ddx{f(x)g(x)} \amp = f(x)\lim_{\Delta x \to 0} \left(\frac{\fillinmath{XXXXXXXXXX}}{\Delta x}\right)\\
\amp + g(x)\lim_{\Delta x \to 0} \left(\frac{\fillinmath{XXXXXXXXXX}}{\Delta x}\right) \\
\amp +\lim_{\Delta x\to 0}\left(\fillinmath{XXXXXXXXXX}\right)\left( \lim_{\Delta x \to 0} \left(\frac{\fillinmath{XXXXXXXXXX}}{\Delta x}\right)\right)
\end{align*}
We can investigate this visual a bit more dynamically: see the differences in area as \(\Delta x\to 0\text{.}\) What parts are left, when \(\Delta x \to 0\text{?}\) What areas aren’t visible?
Since we can differentiate a product of functions, then the obvious next question should be about division: if we needed to build a reasonable way of differentiating a product, shouldn’t we also need to build a new way of thinking about derivatives of quotients?
A nice thing that we can do is think about division as really just multiplication. For instance, if we want to differentiate \(\Ddx{\dfrac{\sin(x)}{x^2}}\text{,}\) then we can just think about this quotient as really a product:
This works great! We can always think about quotients as just products of reciprocals! But the problem is that we can’t always differentiate these reciprocals (for now). We were able to differentiate \(\dfrac{1}{x^2}\) by rewriting this as just a power function (with a negative exponent).
This is a strange formula: we have a formula for \(f'(x)\) written in terms of \(f(x)\text{!}\) But we said earlier that \(f(x) = \dfrac{u(x)}{v(x)}\text{.}\)
Once you have this derivative, confirm that it is the same as \(\dfrac{\cos(x)}{x^2} - \dfrac{2\sin(x)}{x^3}\text{,}\) the way that we found it using the Product Rule.
SubsectionDerivatives of (the Rest of the) Trigonometric Functions
You might remember that of the six main trigonometric functions, we can write four of them in terms of the other two: here are the different trigonometric functions written in terms of sine and cosine functions:
A student encounters the problem \(\ddx{e^x(x^2+1)}\) and notes that since \(\ddx{e^x} = e^x\) and \(\ddx{x^2+1}=2x\text{,}\) then \(\ddx{e^x(x^2+1)} = 2xe^x\text{.}\) Explain not only why the student is incorrect, but also what kinds of pitfalls can occur if we use this “the derivative of a product is the product of the derivatives” method. Differentiate correctly, explaining your process for the student’s benefit.