Section14.5The Other General Differentiation Rules
Theorem14.5.1.The Quotient Rule for Differentiation.
We assume that \(\alpha(x)\text{,}\)\(\beta(x)\text{,}\) and \(f(x)=\frac{\alpha(x)}{\beta(x)}\) are all differentiable functions Assume further that \(\beta(x)\neq0\text{.}\) Then
Proving this directly by using limits would be unpleasant, but as we observed in Chapter 4 the Quotient Rule can be viewed as a rearranged version of the Product Rule.
With the Product Rule for Differentiation in place we now have the tools needed to prove the Power Rule for Positive Integer Exponents. The method of proof we outline in the following problem is called Mathematical Induction and it can be used in other contexts as well. In fact, most of the “Find the Pattern” problems in this text require an Induction argument for full rigor.
Assume that \(n=1\text{.}\) Use the limit definition to show that \(\alpha^\prime(x) = nx^{n-1}.\) (This says, “The Power Rule holds for \(k=1\text{.}\)”)
Now assume that the Power Rule for Positive Integer Exponents holds for \(n=k\text{,}\) where \(k\) is an arbitrary, fixed positive integer. Let \(\beta(x)=x^{k+1}\) and show that \(\beta^\prime(x)=(k+1)x^k.\) (This says, “If the Power Rule holds for \(k\) then it must also hold for \(k+1\text{.}\)”)
Part 14.5.3.a and part 14.5.3.b of this problem constitute a rigorous proof of the Power Rule for Positive Integer Exponents by Mathematical Induction, and part 14.5.3.c asks you to explain the underlying logic. Although we didn’t mention it at the time, Mathematical Induction was the underlying idea in Problem 4.3.14 as well.
With the Power Rule for Positive Integer Exponents in place we can extend it to both negative and rational exponents in the same way we did it in Chapter 4. The following problem is essentially a repeat of Problem 4.3.32and 4.3.34, using Lagrange’s prime notation, and function notation, rather than differentials.
In the statement of Theorem 14.5.1 we explicitly assumed that the quotient, \(\frac{\alpha(x)}{\beta(x)}\text{,}\) is differentiable at \(x\text{.}\) This has the effect that the theorem does not necessarily apply to all possible quotients, in the same way that when we add \(\Delta\beta\neq0\) to the statement of the Chain Rule, the theorem applies to fewer compositions. And just like the Chain Rule the functions that Theorem 14.5.1 does not apply to are mostly pathological, and of no use to us right now.
If leaving these theorems incomplete in this way is troubling to you then you are almost certainly a mathematician by temperament. If you haven’t decided on a major yet, consider mathematics. You obviously like it. Why not learn more?
If you find that you simply don’t care about completing all of the details and you are not majoring in mathematics, congratulations! You’ve made the right choice.
Problem 14.5.6 will lead you through the steps necessary to prove the Quotient Rule for Differentiation without the assumption that \(\frac{\alpha(x)}{\beta(x)}\) is differentiable. Have fun!
Assume that \(\alpha(x)\) and \(\beta(x)\) are differentiable and that \(\beta(x)\neq0\text{,}\) but we make no assumption about the differentiability of \(f(x)=\frac{\alpha(x)}{\beta(x)}\text{.}\)
Use the Product Rule for Differentiation and the Chain Rule (along with the result of part a) to show that \(f(x)=\frac{\alpha(x)}{\beta(x)}\) is differentiable at \(x\) and that
Problem 14.5.8 will lead you through the steps necessary to prove the Product Rule for Rational Exponents without the assumption that \(x^{\frac{p}{q}}\) is differentiable. It relies on the result of Problem 14.5.7. Have fun!