MHB Assume the Limit Exists: Proving an Impossibility

  • Thread starter Thread starter Amad27
  • Start date Start date
  • Tags Tags
    Limit
Click For Summary
The discussion revolves around proving by contradiction that the limit of 1/x as x approaches 0 does not exist. Participants suggest assuming the limit equals a real number M and exploring the implications of this assumption through sided limits. They emphasize that if the limit existed, the behavior approaching from the left and right would yield the same result, which is not the case here. The conversation also touches on proving that for any real number M, there exists a delta such that the limit diverges, reinforcing that both one-sided limits approach infinity but are not finite. Ultimately, the participants aim to clarify the method for proving the non-existence of the limit across all delta values.
Amad27
Messages
409
Reaction score
1
Hello,

Prove that

$$\lim_{{x}\to{0}} \frac{1}{x}$$

Does not exist by contradiction. So the obvious step:Assume:$$\lim_{{x}\to{0}} \frac{1}{x} = M$$

$| 1/x - M| < \epsilon$ for $|x| <\delta_1$

Any ideas? PLEASE DO NOT SOLVE.
 
Physics news on Phys.org
If this limit exists, then it wouldn't matter if you go to zero from positive or from negative numbers.
 
Fallen Angel said:
If this limit exists, then it wouldn't matter if you go to zero from positive or from negative numbers.

So I suppose what you are suggesting is that we reach a contradiction using sided limits? So we get:

$$x< \delta \implies \left| 1/x - L \right| < \epsilon$$

$$-x < \delta_2 \implies \left| 1/x - L \right| < \epsilon$$

Let $\delta' = \min(\delta_1, \delta_2)$

Let $\epsilon = 1$

$$x< \delta \implies \left| 1/x - L \right| < 1$$

$$-x < \delta_2 \implies \left| 1/x - L \right| < 1$$

------------------------------------------------------------------

$$0 < \delta_1 + \delta_2 \implies 2|1/x - L| < 2$$

This is hard, since I cannot reach a contradiction.
 
EDIT:
In which sense do you want to prove the limit does not exist? Do you want to prove the two-sides limit does not exist? Or do you want to prove that $\lim_{x \to 0^{+}} \frac{1}{x} = +\infty$ and $\lim_{x \to 0^{-}} \frac{1}{x} = - \infty$ which also means that both limits do not exist as they're not finite.

If it's the first case then post #2 gives the answer. If not, you could prove it this way. Suppose you want to prove $\lim_{x \to 0^{+}} \frac{1}{x}$ does not exists (in the sense that it's value is not finite) then it's equivalent with proving $\forall M \in \mathbb{R}: \lim_{x \to 0^{+}} \frac{1}{x} \neq M$. This last statement is equivalent with
$$\forall M \in \mathbb{R}: \exists \varepsilon>0, \forall \delta>0, \exists x: 0<x<\delta \ \mbox{and} \left |\frac{1}{x}-M \right| \geq \varepsilon$$

To prove the above statement. Let $\epsilon = 1$ for example and try to find an $x$ such that the above statement holds.
 
Last edited:
Siron said:
EDIT:
In which sense do you want to prove the limit does not exist? Do you want to prove the two-sides limit does not exist? Or do you want to prove that $\lim_{x \to 0^{+}} \frac{1}{x} = +\infty$ and $\lim_{x \to 0^{-}} \frac{1}{x} = - \infty$ which also means that both limits do not exist as they're not finite.

If it's the first case then post #2 gives the answer. If not, you could prove it this way. Suppose you want to prove $\lim_{x \to 0^{+}} \frac{1}{x}$ does not exists (in the sense that it's value is not finite) then it's equivalent with proving $\forall M \in \mathbb{R}: \lim_{x \to 0^{+}} \frac{1}{x} \neq M$. This last statement is equivalent with
$$\forall M \in \mathbb{R}: \exists \varepsilon>0, \forall \delta>0, \exists x: 0<x-a<\delta \ \mbox{and} \left |\frac{1}{x}-M \right| \geq \varepsilon$$

To prove the above statement. Let $\epsilon = 1$ for example and try to find an $x$ such that the above statement holds.
Thanks Siron.

Let $\epsilon = 1$ so then we get:

$|\frac{1}{x} - M| > 1$

$|\frac{1}{x}| > 1 + |M| $ [by the triangle inequality]

$ x < \frac{1}{1 + |M|}$

So if $ x < \frac{1}{1 + |M|}$ then $|\frac{1}{x} - M| > 1$

So for $\delta = \frac{1}{1 + |M|}$ the limit as $x \to 0^{+}$ does not exist.

But this is only one $\delta$

How do you prove for all $\delta$?

In limit proofs, generally, how do you prove $|f(x) - L| > \epsilon$?

Do you have to prove for a SPECIFIC $\delta$ or ALL $\delta$?

Thanks!
 
Olok said:
Thanks Siron.

Let $\epsilon = 1$ so then we get:

$|\frac{1}{x} - M| > 1$

$|\frac{1}{x}| > 1 + |M| $ [by the triangle inequality]

$ x < \frac{1}{1 + |M|}$

So if $ x < \frac{1}{1 + |M|}$ then $|\frac{1}{x} - M| > 1$

So for $\delta = \frac{1}{1 + |M|}$ the limit as $x \to 0^{+}$ does not exist.

But this is only one $\delta$

How do you prove for all $\delta$?

In limit proofs, generally, how do you prove $|f(x) - L| > \epsilon$?

Do you have to prove for a SPECIFIC $\delta$ or ALL $\delta$?

Thanks!

Remark: note that I wrote $0<x-a<\delta$ in my previous post but this has to be $0<x<\delta$ of course.

It's true that we want to prove it $\forall \delta>0$. More precisely, we want to find $\forall \delta>0$ a suitable $x$ such that the statement holds.

Let $M \in \mathbb{R}$. Choose $\epsilon = 1$ and let $\delta>0$ be arbitrary. We'll translate the statement '$\forall \delta>0, \exists x$' as follows:

Let $N>0$ a real number such that $\delta > \frac{1}{N}$ (it's always possible to find such $\delta$) and suppose $x \in \left ]0,\frac{1}{N} \right[$. Clearly, $0<x<\frac{1}{N}$ by construction.

Knowing the above we get the following
$$\left|\frac{1}{x}-M \right | \geq \frac{1}{|x|}-|M| \geq N - |M|$$

Note that $N - |M|$ can be made greater than or equal to $1$. To see this, since $N>0$ we can set $N \geq |M|+\epsilon = |M|+1$ at the start of the proof.
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 23 ·
Replies
23
Views
3K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 15 ·
Replies
15
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K