A function $f:\mathbb{R} \to \mathbb{R}$ is discontinuous at $x_0 \in \mathbb{R}$ if there exists a sequence $a_n$ such that $a_n \to x_0$ but $f(a_n) \not \to f(x_0)$.

I thought of a way of strengthening this by saying that a function $f:\mathbb{R} \to \mathbb{R}$ is strongly discontinuous at $x_0 \in \mathbb{R}$ if for every injective sequence $a_n \to x_0$, $f(a_n)$ diverges.

But I can't think of an example of a function with this property. Does such a function exist?

  • $\begingroup$ $f(x) = 1/x$ with $f(0)=0$? $\endgroup$ – bitesizebo Apr 3 '18 at 3:32
  • $\begingroup$ By injective sequence do you mean a sequence all of whose terms are distinct? $\endgroup$ – carmichael561 Apr 3 '18 at 3:33
  • 1
    $\begingroup$ @Z.Xie Ah, of course. That works. Is there an example where each sequence $f(a_n)$ doesn't converge at all (i.e., not even to infinity), or alternatively an example where $f$ is bounded? $\endgroup$ – MathematicsStudent1122 Apr 3 '18 at 3:35
  • 2
    $\begingroup$ If bounded we could always apply Bolzano Weierstrass $\endgroup$ – bitesizebo Apr 3 '18 at 3:41
  • $\begingroup$ @Z.Xie Right, perfect. Thanks. $\endgroup$ – MathematicsStudent1122 Apr 3 '18 at 3:43

Maybe I'm misinterpreting this, but doesn't $f(x)=1/x$ and $x_0 = 0$ work? Any sequence tending towards zero will correspondingly tend towards infinity.

  • $\begingroup$ This answers the Q as it is stated. In the comments the proposer asks whether it is possible for all $x_0\in \Bbb R,$ which is a different Q, to which I have given an answer.....+1 $\endgroup$ – DanielWainfleet Apr 3 '18 at 17:20

The Q as written has been answered by ASKASK. The A to the Q in the proposer's comments, which is whether it can hold for every $x_0\in \Bbb R$, is NO.

For $n\in \Bbb N$ let $S(n)=\{y\in \Bbb R: |f(y)|>n\}.$

Suppose that for all $x\in \Bbb R$ and for every $(a_n)_n$ sequence in $\Bbb R\setminus \{x\}$ converging to $x,$ the sequence $(f(a_n))_n$ does not converge.

Then for $x\in \Bbb R$ and $n\in \Bbb N$ there must exist $r_{x,n}>0$ such that $$A(x,n)=(-r_{x,n}+x,\;r_{x,n}+x)\setminus \{x\} \subset S(n).$$ Otherwise $\forall s\in \Bbb N \;\exists a_s\; (0<|a_s-x|<1/s \land |f(a_s)|\leq n)$ but then $(f(a_s))_s$ has a convergent sub-sequence.

Let $U(n)=\cup_{x\in \Bbb R}A(x,n).$ Then $U(n)$ is open and dense.

The Baire Category Theorem implies that $\cap_{n\in \Bbb N}U(n)\ne \emptyset.$ But if $z\in U(n)$ for every $n$ then $z\in S(n)$ for every $n,$ so $|f(z)|>n$ for every $n\in \Bbb N,$ which is absurd.


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.