I am interested in the following situation: I have two codimension-2 knots $K_1$ and $K_2$ in $S^n$ and they are not isotopic. Furthermore, $K_1$ is not isotopic to the mirror image of $K_2$ and vice versa. Could it be that $K_1$ and $K_2$ become isotopic after connect summing (away from the knots) another $n$-manifold $X$ (so that $K_1$ and $K_2$ are then isotopic inside of $X$)?

(In the case where say $K_2$ is the mirror image of $K_1$, we could connect sum on a nonorientable $X$ and going around a nonorientable loop would do the trick -- I don't know any other tricks.)

I would like to know the answer to this question specifically in low-dimensions (<5). My guess: this can't happen in dimension 2 (here changing the codimension to consider curves on the surface -- say for some geometric reasons), this can't happen in dimension 3 (say by Gordon-Lueke and the fact that the fundamental group of the complement says a lot), this can happen in dimension 4 (since it's a jungle out there).

  • $\begingroup$ What do you mean when you say “this can’t happen in dimension 2 (say for some geometric reason)”? If I understand your question, this would be about embedded $S^0$’s in $S^2$, which for silly reasons are always isotopic. $\endgroup$ Commented Oct 16, 2022 at 14:54
  • $\begingroup$ @AndyPutman Oh yes -- I'm sorry I was totally incorrect in my phrasing there. I was thinking of curves on a surface (just as fun low low dimensional analogy). I'll edit - thank you. $\endgroup$
    – Sprotte
    Commented Oct 16, 2022 at 16:57

1 Answer 1


I believe can happen in dimension $4$, and probably in all higher dimensions. Take two inequivalent knots $K, K'$ with the same exterior (Cappell-Shaneson; Gordon). Then $K'$ is obtained from $K$ by a "Gluck twist", in other words trivialize the normal bundle of $K$ (there's only one way to do this), remove it, and glue back in using the diffeomorphism $f: S^1 \times S^2$ given by $f(\theta,z) = (\theta,$ rotation of $z$ by $\theta$).

There's a general principle that connect summing with $\mathbb{C}P^2$ undoes Gluck twists, and I think that this should prove that $K$ and $K'$ are now isotopic in $S^4 \# \mathbb{C}P^2$. I'd have to think about the proof (or maybe someone could point to a reference).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.