A bit of background: I'm an engineer, not a mathematician, and I need to review and improve my calculus. In college, I never liked how they said $dy/dx$ was a single symbol, not a ratio; and then proceeded to write things like $dy = f(x) dx$ and integrate. So I'm trying a different angle this time, and reading the textbook Elementary Calculus: An Infinitesimal Approach, by H. Jerome Keisler, which is available online.

I'm at the part (p.55) that discusses the Increment Theorem. It says:

Let $y = f(x)$. Suppose $f'(x)$ exists at a certain point x, and $\Delta x$ is infinitesimal. Then $\Delta y$ is infinitesimal, and

$\Delta y = f'(x)\Delta x + \epsilon\Delta x$

for some infinitesimal $\epsilon$, which depends on $x$ and $\Delta x$.

And then he works some examples, finding $\epsilon$. For example, with $y = x^3$...

$$ y' = 3x^2 \\ \Delta y = (x + \Delta x)^3 - x^3 \\ \epsilon = \Delta y / \Delta x - y' \\ ...\\ \epsilon = 3x \Delta x + (\Delta x)^2 $$

I'm left wondering... what is the point? Where are we going with this? We seem to be revisiting the definition of the derivative, where $\epsilon$ is the part of the equation that we were able to discard because it was infinitesimal. For example, to get the derivative of $y = x^3$

$$ st( \frac{\Delta y}{\Delta x} ) = st(\frac{(x + \Delta x)^3 - x^3}{\Delta x}) \\ = st(3x^2 + 3x\Delta x + (\Delta x)^2) \\ = 3x^2 $$

  • $\begingroup$ The point is we can now write $$\frac{\Delta y}{\Delta x} = f'(x) + \varepsilon$$ where the fraction is legitimately an actual fraction, and doesn't just look like one. $\endgroup$ – sperners lemma Nov 3 '12 at 15:01
  • $\begingroup$ @spernerslemma It seems that you didn't read the source, since the inference is in the other direction - hence the OP's query as to why this simple rescaling is singled out as a new named theorem. $\endgroup$ – Bill Dubuque Nov 3 '12 at 15:38
  • $\begingroup$ @spernerslemma I don't understand why I couldn't legitimately write the fraction before, or why this theorem would change that. $\endgroup$ – Rob N Nov 5 '12 at 1:22
  • $\begingroup$ @RobN Because, $\Delta x and \Delta y$ are individually not numbers, since no number can be infinitely close to 0 but not 0 itself. But they can be thought of as a ratio when they get closer and closer to $0$. Or, as done in nonstandard-analysis, you could say "that's stupid" and declare numbers can be infinitely close to $0$ without being it. $\endgroup$ – PyRulez May 17 '13 at 20:08

The increment theorem is sort of a restatement of the usual procedure for derivatives, but it is more convenient because $\Delta x$ can be 0. Keisler uses it on page 86 to prove the chain rule.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.