Now we want a way to figure out the speed for different position functions. One important rule for doing that is the power rule. In Part 2, the example x(t)=t2 suggested that the speed of tn should be ntn−1. Now we can prove that pattern in general.
The short-time statement we want is
(t+h)n=tn+ntn−1h+o(h).
If this holds, then the coefficient of h is ntn−1, so the speed of x(t)=tn is ntn−1. If you want to recall the setup from Part 3, open the or the below.
We will prove the formula by induction on n. When n=1, we have
t+h=t+1⋅h+o(h),
because 0 is certainly o(h).
Now assume that for some n we already know
(t+h)n=tn+ntn−1h+o(h).
We want to prove the same kind of formula for n+1. Start with
(t+h)n+1=(t+h)n(t+h).
The only tools we need from Part 3 are the , the , and the .
Now look at the bracketed terms. By the , both h2 and ho(h) are o(h). By the , ntn−1h2 and to(h) are both o(h). Then the says that the whole bracket is still o(h).
That gives
(t+h)n+1=tn+1+(n+1)tnh+o(h).
So if the formula is true for n, then it is also true for n+1. Since it is true for n=1, it follows that it is true for every positive integer n.
Therefore the speed of x(t)=tn is
v(t)=ntn−1.
That is the power rule.
Definition
A notation
A(c) means some quantity whose absolute value is at most c.
If you are used to seeing measurements written with a plus/minus, this is the same basic idea. We are keeping track of how large the error can be without needing to know exactly what it is.
Definition
o notation
f(h)=o(h) means f(h)=hε(h) for some function ε(h) with ε(h)→0 as h→0.
We do not know exactly what ε(h) is, and we do not need to. We only need to know that it goes to 0.
Definition
Exact local speed statement
x(t+h)=x(t)+v(t)h+o(h) is the precise way to say that the position looks linear for small h.
The main term is v(t)h, and everything else is smaller than that in the h→0 limit.
Rule 1
Sum rule
If f(h)=o(h) and g(h)=o(h), then f(h)+g(h)=o(h).
Write f(h)=hε(h) and g(h)=hδ(h). Then f(h)+g(h)=h(ε(h)+δ(h)), and the bracket still goes to 0.
Rule 2
Constant multiple rule
If f(h)=o(h) and c is fixed, then cf(h)=o(h).
Multiplying the small error factor by a fixed constant does not stop it from going to 0.
Rule 3
Extra-factor rule
Multiplying by an extra factor of h gives a term of size o(h). In particular, h2=o(h), and if f(h)=o(h), then hf(h)=o(h).
For h2, just write h2=h⋅h. For hf(h), write f(h)=hε(h), so hf(h)=h(hε(h)).