Is the following argument correct?

Proposition. If $\{x_n\}$ is Cauchy sequence such that $x_n = c$ for infinitely many $n$, then $\lim_{n\to\infty}x_n = c$.

Proof. Let $\epsilon>0$. Since $\{x_n\}$ is a Cauchy sequence, there exists an $M\in\mathbb{N}$ such that $\forall \, j,k\ge M$, we have $|x_j-x_k|<\epsilon$. Now, since $x_n = c$ for infinitely many $c$, then surely $x_r = c$ for some $r \ge M,$ implying that $|x_j-c|<\epsilon \,\,\forall j\ge M$, completing the argument.


  • 2
    $\begingroup$ Yes, it is correct. In similar way you can show that if there is a subsequence which converges to $c$ then the whole sequence converges to $c$. $\endgroup$ – Robert Z Nov 5 '18 at 7:59
  • $\begingroup$ @RobertZ Thanks for you help $\endgroup$ – Atif Farooq Nov 5 '18 at 8:01
  • 2
    $\begingroup$ "... implying that $|x_j-c|<\epsilon, \forall j\ge M$". Since $\epsilon$ was arbitrary, this completes the argument. $\endgroup$ – Jimmy R. Nov 5 '18 at 8:02

As told in the comments, your proof is fine.

Usually when I am asked to prove a statement that seems rather obvious, I tend to prove it by contradiction: this usually allows me to explore the intuitive feel I have for the result.

Assume the sequence converged to $l \neq c$. Then for $\epsilon = \frac{|c - l|}{2} $ $\exists N $ s.t. $i > N \implies |x_i - l| < \epsilon $, by definition of convergence. But this is absurd, because the sequence had infinitely many terms equal to $c $ and thus infinitely many $x_k = c $ with $k > N $.

  • $\begingroup$ This argument presupposes that the context is $\mathbb R$ or some other complete metric space, so that the sequence $(x_n)$ is guaranteed to converge and you only need to check that it can't converge to a wrong limit. But in fact the OP's proof works in any metric space (if you write $d(x,y)$ instead of $|x-y|$), even if the space isn't complete. $\endgroup$ – Andreas Blass Nov 6 '18 at 0:27
  • $\begingroup$ @AndreasBlass Indeed the OPs proof is more general, but I didn't think it was wrong to assume I was in $\mathbb R $, as the tag real-analysis suggests. $\endgroup$ – RGS Nov 6 '18 at 7:00
  • 1
    $\begingroup$ I appreciate your counsel RGS. $\endgroup$ – Atif Farooq Nov 6 '18 at 11:59

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.