1

If we have a lengthening pendulum and the length $l$ of the pendulum at time $t$ is $$l=l_0+vt$$ where $l_0$ is the initial length of the pendulum and $v$ is the velocity for which the pendulum's length is increasing.

Then the differential change in the length of the pendulum is given by $$\mathrm{d}l=v\,\mathrm{d}t$$ this implies that $$\frac{1}{\mathrm{d}t}=v\frac{1}{\mathrm{d}l}\tag{1}$$ My question is why does it now follow that $$\frac{\mathrm{d}}{\mathrm{d}t}=v\frac{\mathrm{d}}{\mathrm{d}l}$$ Do we simply multiply both sides of $(1)$ by $\mathrm{d}$ or is there more to it than that?

BLAZE
  • 8,458

2 Answers2

3

The chain rule gives

$$\begin{align} \left(\frac{d}{dt}\right)f(l(t))&=\frac{df(l)}{dl}\frac{dl}{dt}\\\\ &=\left(v\frac{d}{dl}\right)f(l) \end{align}$$

Mark Viola
  • 179,405
  • Thanks for your answer. So taking the derivative of both sides with respect to time works even though $f(l(t))$ is equal to unity in this case? – BLAZE Jun 03 '16 at 20:02
  • 1
    You're welcome Blaze. My pleasure, The chain rule applies to any differentiable functions $f$ and $l$. So, if $f=1$, then its derivative with respect to $l$ is zero. – Mark Viola Jun 03 '16 at 20:06
  • Sorry just one more question; if the derivative with respect to $l$ is zero then doesn't that mean that we have $\dfrac{\mathrm{d}}{\mathrm{d}t}f(l(t))=v\dfrac{\mathrm{d}}{\mathrm{d}l}=0$? – BLAZE Jun 03 '16 at 20:11
  • 1
    If $\frac{df}{dl}=0$, then $\frac{df}{dt}=0$. However, $f$ can be any differentiable function, not just the constant $1$. So, the chain rule asserts $$\frac{d}{dt}=\frac{dl}{dt}\frac{d}{dl}$$If $\frac{dl}{dt}=v$, then $$\frac{d}{dt}=v\frac{d}{dl}$$ – Mark Viola Jun 03 '16 at 20:17
1

Here is an expansion on the above answer. Technically speaking, it doesn't make sense to "multiply" both sides by $d$ or "divide" by infintesimals (without proper justification). What you should really say is that a small change $\Delta l$ is related to a small change in $t$ via $$\Delta l=v\Delta t\;.$$ The quantities $\Delta l$ and $\Delta t$ are positive quantities which can be made arbitrarily small, but are nonzero. More precisely, we have not yet taken the limit $\Delta t\to 0$. To get the equation you are looking for, we should see how taking the derivative of a function $f$ of $l$ relates to taking the derivative with respect to $t$. Let us assume that $l$ is a function of $t$ and let us pick a point $t_0$. By definition, we have $$\frac{df}{dt}(l(t_0))=\lim_{\Delta t\to 0}\frac{f(l(t))-f(l(t_0))}{\Delta t}=\lim_{\Delta t\to 0}\frac{f(l(t))-f(l(t_0))}{\Delta l}\cdot \frac{\Delta l}{\Delta t}$$$$=\lim_{\Delta t\to 0}\frac{f(l(t))-f(l(t_0))}{\Delta l}\cdot v$$ Since $\Delta l\to 0$ as $\Delta t \to 0$, the RHS gives you exactly $v\dfrac{df}{dl}(l_0)$. Since the point $t_0$ was arbitrary, we get the desired equation $$\frac{df}{dt}=v\frac{df}{dl}\;.$$ Since the function $f$ was arbitrary, the differential operators are related via the equation $$\frac{d}{dt}=v\frac{d}{dl}\;.$$ I have sort of implicitly used the proof of the chain rule, but I wanted to provide a solution which more closely follows your thought process.

BLAZE
  • 8,458