0

Let $f(x)=$$x-1 |x \in \mathbb{Q} \brace 5-x| x \in \mathbb{Q}^c$

Show that $\lim_{x \to a}f(x)$ does not exists for any $a \not= 3$

I first showed that $lim_{x \to 3}f(x)=2$.

I don't know how to approach this part. Can anyone please guide? I was thinking of using density theorem at first to show that there will exist sequences of rationals and irrationals that will approach a but their limits would not equal but that is problematic as I don't really know the limit.

thank you

Alea
  • 946

3 Answers3

3

Let $a \ne 3$.

If $(x_n)$ is a sequence in $ \mathbb Q$ with $x_n \to a$, then $f(x_n)=x_n-1 \to a-1$.

If $(y_n)$ is a sequence in $ \mathbb Q^c$ with $y_n \to a$, then $f(x_n)=5-y_n \to 5-a$.

Since $a \ne 3$ we have $a-1 \ne 5-a$, hence $\lim_{x \to a}f(x)$ does not exist.

Fred
  • 77,394
1

Of course, you must use the fact that for any point, there is a sequence of rationals and another sequence of irrationals which converge to that point.

Here is the conventional limit definition :

Given $f : D \to \mathbb R$ and $x \in D$, suppose there is a number $L$ such that for every $\epsilon > 0$ there is a $\delta > 0$ such that if $|x - y| < \delta$ then $|f(x) - f(y)| < \epsilon$. Then, $\lim_{t \to x}f(t)$ is said to exist and equal $L$.

Using a usual trick, we bring this down to convergence of sequences, and then get a sequential definition :

Given $f : D \to \mathbb R$ and $x \in D$, if there exists a number $L$ such that for every sequence $x_n$ converging to $x$, the sequence $f(x_n)$ converges to $L$. Then $\lim_{t \to x}f(t)$ is said to exist and equal $L$.

Now, if the limit does not exist, then we must negate the above statement:

Given $f : D \to \mathbb R$ and $x \in D$, if for all numbers $L$, there exists a sequence $x_n$ converging to $x$ such that $f(x_n)$ does not converge to $L$, then the limit $\lim_{x \to a} f(x)$ is said to not exist.

Now, we are in shape to attack the problem : all we need to do, is given any $a \neq 3$ and a candidate $L$ for the limit, find a sequence of points which converge to $a$ such that the function values don't go to $L$. To do this, we use the piecewise definition of our function.

Let $a \neq 3$. Suppose that $L$ is a candidate limit.

EDITED : At this stage, we want a sequence of numbers which does converge to $a$, but whose function values don't converge to $L$.

The answer to your question below is this : The pieces of $f$, when treated as functions from the real line to itself, are continuous, and this fact can be used to prove that $f$ is not continuous at any point, other than $3$.

First, let $p_n$ is a sequence of rationals converging to $a$. We claim that $f(p_n) \to a-1$.

This follows from the fact that $p_n$ lies inside one piece of $f$, so we may use continuity of the function which defines $f$ on that piece. But if you want to argue by basics, then : $f(p_n) = p_n - 1$ from the fact that $p_n$ is rational. Now, since $p_n \to a$ and (the constant sequence)$-1 \to -1$, we can add limits to get $p_n -1 \to a-1$, and therefore $f(p_n) \to a-1$.

By uniqueness of limits, if $L \neq a-1$, then $p_n$ serves as a candidate for the sequence of values not converging to $L$, since $p_n \to a$ but $f(p_n) \not\to L$.

On the other hand, if $q_n$ is a sequence of irrationals converging to $a$, then we may repeat the above argument for the irrational piece to get $f(q_n) \to 5-a$. So $q_n$ now serves as a candidate for the sequence of values not converging to $L$, whenever $5-a \neq L$.

Now, if $a-1 \neq 5-a$, which happens precisely when $a \neq 3$, we see that for any $L$, of course $L$ can't be equal to both $a-1$ and $5-a$, so take the case which it does not equal, and that sequence works.

Hence, the limit of the given function does not exist, for any $a \neq 3$. Of course, if $f$ has to be continuous at $a$, then $\lim_{x\to a} f(x)$ has to exist and equal $f(a)$, but the limit doesn't exist if $a \neq 3$, so $f$ is not continuous at any such point.

  • So, are your exams over? If so, how were they? – Sarvesh Ravichandran Iyer May 17 '18 at 09:45
  • Thank you for the answer! However, isn't this additionally assuming continuity of the function as only in that case, we can say that the limit approaches to the value of the function at that point? – Alea May 17 '18 at 10:36
  • And my exams aren't over yet. Have one tomorrow and another one on 19th and then summers! I'm assuming you must be having a summer break at the moment as well. Hope its fun and productive. – Alea May 17 '18 at 10:40
  • You are right, I am assuming it, but it's also easy to prove from first principles of the limit. Yes, my summer is going on, so thanks for the compliment! – Sarvesh Ravichandran Iyer May 17 '18 at 15:33
  • But then the function is not continuous at any point except 3, no? That has been proved in your answer as well as you assumed continuity. I don't know If I'm making sense anymore. It's just that limit can exist even if your function is not continuous at a certain point. – Alea May 17 '18 at 15:58
  • I will edit the answer shortly to provide clarifications, but am a little busy right now. – Sarvesh Ravichandran Iyer May 18 '18 at 03:20
  • 1
    Edit complete. You may view the new post. – Sarvesh Ravichandran Iyer May 18 '18 at 04:05
  • Thank you for the answer! Apologies for such a late response. Had been travelling all this while! – Alea May 31 '18 at 10:19
  • 1
    You are welcome! Also, I have recently written an answer which generalizes this "rational-irrational" situation, recently. I will find the link and post it shortly. Good to travel after exams, a nice relaxation. – Sarvesh Ravichandran Iyer Jun 01 '18 at 03:52
  • Thank you! Yes, it's good to be back with family after so long and spend time with them! But need to get back to work now, lol. Can you please have a look at this question? https://math.stackexchange.com/questions/2812743/self-studying-analysis-direction It's subjective, I need a bit of direction on how to go about self-studying. Thank you. – Alea Jun 08 '18 at 17:08
  • 1
    I'll get back to you. Also, hee is the link to the rational - irrational answer of mine : https://math.stackexchange.com/questions/2787111/continuity-of-un-plottable-piece-wise-defined-function/2787259#2787259 . More specifically, it tells you precisely at which points your function is continuous. – Sarvesh Ravichandran Iyer Jun 09 '18 at 09:21
  • Thank you! I saw the answer but I’m also interested in the proof of it to see where/how it’s coming about! Will look it up on my own. – Alea Jun 12 '18 at 13:39
0

As $(x-1)-(5-x)=2x-6$, you will always find values in the neighborhood of $x$ that have their images $2x-6$ apart so that not all $\epsilon$ can be met. Unless $x=3$.