Contrapositive Epsilon-Delta Limits?

1.1k Views Asked by At

Is the contrapositive form of the epsilon-delta limit definition valid, or am I missing something?
-- If valid, are there any examples where it is easier in a proof?


\begin{align*}\scriptsize \lim_{x\to c} \; f(x)=L\\ &\scriptsize\equiv\quad \forall\epsilon\; (\;\exists\delta\; (\, 0< \; \mid x-c \mid \; < \delta \implies \mid f(x)-L \mid \; < \; \epsilon \,)) &\tiny(Defintion)\\ &\scriptsize\equiv\quad \forall\epsilon\; (\;\exists\delta\; (\, \neg( \; \mid f(x)-L \mid \; < \; \epsilon ) \implies \neg(0< \; \mid x-c \mid \; < \delta ) \, )) &\tiny(Contrapositive)\\ &\scriptsize\equiv \quad \forall \epsilon \; ( \; \exists \delta \; ( \, \mid f(x)-L \mid \; \ge \; \epsilon \implies (\;0\ge\;\mid x-c\;\mid\;)\; \lor \;(\;\mid x-c\,\mid\;\ge \delta\;) \, )) &\tiny(a\lt b\lt c \ \equiv\ ((a<b)\land(b<c)),\;DM)\\ &\scriptsize\equiv \quad \forall \epsilon \; ( \; \exists \delta \; ( \, \mid f(x)-L \mid \; \ge \; \epsilon \implies \mid x-c\,\mid\;\ge \delta \, )) &\tiny(\ (0\ \not\gt\ \mid a \mid)\ \land\ (x\ \not=\ c)\ )\\ \end{align*}
I'm still trying to conceptualize this, so as to posit tests and extremal cases for validation... any help here would be greatly appreciated!

1

There are 1 best solutions below

0
On

Of course the contrapositive form is valid, the logic you carried out was correct. Although right now, I can't think of any examples where the contrapositive form would be substantially easier than the standard definition (and I don't think I ever could), I'll convince you that it works. For example, consider $$\lim_{x \to 5} 3x-3 = 12$$

Using the standard definition, for all $\epsilon >0$ we must find a $\delta>0$ such that $$0<|x-5|<\delta \Rightarrow |3x-15|<\epsilon$$

We would do some rearranging to get $$0<|x-5|<\delta \Rightarrow 3|x-5|<\epsilon$$ $$0<|x-5|<\delta \Rightarrow |x-5|<\frac{\epsilon}{3}$$ and we would conclude that the implication is satisfied for $\delta=\frac{\epsilon}{3}$

Now consider the contrapositive form $$|3x-15|\geq\epsilon \Rightarrow |x-5|\geq\delta$$

We can do the same rearranging to get $$3|x-5|\geq\epsilon \Rightarrow |x-5|\geq\delta$$ $$|x-5|\geq\frac{\epsilon}{3} \Rightarrow |x-5|\geq\delta$$

and we would also conclude that the implication holds for $\delta=\frac{\epsilon}{3}$