1

Consider:

$$ f(x) = \frac{\sin(\frac{\pi}{x})}{\sin(\frac{\pi}{x})} $$

This function is undefined at $x=0$ and at $ x = \frac{1}{n} $, $\forall n \in \mathbb{N}.$

I thought I could show that the limit of f does not exist at 0, but it does not hold under the appropriate definition of a limit.

Erroneous "proof":

$$ \text{For any arbitrarily small } \delta. $$

$$ M =\bigg \lceil\frac{1}{\delta}\bigg\rceil \in \mathbb{N} $$

$$ \frac{1}{M+1} < \delta $$

$$ f(\frac{1}{M+1}) \text{ is undefined. } $$

$$ \text{Therefore, for any real number L that we choose as our candidate for the limit:} $$

$$ |\frac{1}{M+1}-0| < \delta \text{, but: } $$

$$ |f(\frac{1}{M+1})-L| \nless \epsilon $$

$$ \text{ because it is undefined.} $$

Again, this "proof" is incorrect. According to the definition of the limit for a real-valued function with domain $S\subset\mathbb R$, at a limit point of $S$ (here $S=\{x\in\mathbb R^*\mid\frac1x\notin\mathbb N\}$): $$\text{the limit of our function }f\text{ at }0\text{ is }1\text{, i.e.}$$ $$\forall\varepsilon>0\quad\exists\delta>0\quad\forall x\in S\quad(|x|<\delta\Rightarrow|f(x)-1|<\varepsilon)$$ (indeed, any $\delta>0$ will do, since $\forall x\in S\quad|f(x)-1|=0$).

What I find interesting is that the function becomes "infinitely porous" as we approach 0. It's almost like the line is disintegrating. It reminds me of the function:

$$ g(x) = 1, x \in \mathbb{Q} $$

$$ g(x) = 0, x \not \in \mathbb{Q} $$

I am having trouble wrapping my head around this function. It "behaves" like a normal line everywhere, except infinitesimally close to 0. I want to learn more but I don't know where to begin.

Q1) I'm sure there must be other discussions about functions with these properties, but I cannot find any. Are there any key words I could use to start my research?

Q2) Does this type of discontinuity have a name? I don't think it would be accurate to call it a hole.

Anne Bauval
  • 34,650
  • 1
    the name is removable singularity – Kurt G. Oct 01 '22 at 05:19
  • 1
    Your "proof" that $f$ has no limit at $0$ does not hold if you take the appropriate definition of limit. See e.g. https://en.wikipedia.org/wiki/Limit_of_a_function#More_general_subsets – Anne Bauval Oct 01 '22 at 06:00
  • 1
    Pardon me, but IMHO, this kind of question is very artificial. "Solving" it will not make a fundamental jump forward in analysis. – Jean Marie Oct 01 '22 at 06:44
  • 1
    Some ideas, about that. I think it will bring large insight looking the complex analog of your function. Also see for some information Meromorphic functions. – kolobokish Oct 01 '22 at 09:30
  • @JeanMarie, I am not sure I understand your comment. I merely wanted more information about the function because I found it interesting. – Matthew Gagnon Oct 01 '22 at 12:28
  • @AnneBauval, thank you, I will edit my post. So what would we say the limit as x approaches 0 is? – Matthew Gagnon Oct 01 '22 at 12:33
  • 1
    $f$ is the constant function $1$, just restricted to a smaller domain (the real line with a particular countable set of points removed). – Daniel Hast Oct 01 '22 at 12:43
  • 1
    You might be interested in investigating functions such as $;\sin \left(\frac{1}{\sin (1/x)} \right);$ *AND* $;\sin(1/x) \cdot \sin \left(\frac{1}{\sin (1/x)} \right);$ *AND* $;\sin \left( \frac{1}{\sin \left( \frac{1}{\sin (1/x)}\right)}\right);$ *AND* $;\sin \left(\frac{1}{\sin (1/x)} \right) \cdot \sin \left( \frac{1}{\sin \left( \frac{1}{\sin (1/x)}\right)}\right);$ *AND* $;\sin \left( \frac{1}{\sin \left( \frac{1}{\sin (1/x)}\right)}\right) \cdot \sin\left( \frac{1}{\sin \left( \frac{1}{\sin \left( \frac{1}{\sin (1/x)}\right)}\right)} \right).$ – Dave L. Renfro Oct 01 '22 at 13:48
  • 1
    I don't want to discourage you. All the contrary. What I wanted to say (IMHO, once agin) is that this kind of question has no mathematical "flesh", it is almost a notational question that becomes void is you consider, right from the beginning, that the quotient is "1"... I would say you the same thing if you were telling me that you are doing a PhD thesis on Chivalry in Gibraltar between 1500 and 1501 :). Briefly said: I consider it is not worth the value to spend/loose time on this kind of "problems". – Jean Marie Oct 01 '22 at 14:24
  • 1
    @JeanMarie, thank you. I suppose I should clarify that I find the example function in my question to be quite trivial. It was my first time seeing such behavior. I am not so much interested in the function itself or its notation, but rather I am intrigued by the idea of infinitesimally close discontinuities, especially those on a function that has continuous portions as well (unlike g(x) in my original post). I imagine there must be some worthwhile pursuits related to these ideas, no? In the future I will be more accurate with how I ask my questions on here. – Matthew Gagnon Oct 01 '22 at 17:42

0 Answers0