Are there known examples of knot groups that do not surject onto any alternating group? I would be a little surprised if the answer is negative.

Update: To reflect the interesting part of the question, I'd prefer to avoid the unknot and the "abelianization" in the obvious way. So, the target alternating group should be considered as $A_n$ $(n > 3)$. That is, as suggested by YCor, does there exist some nonabelian knot group $G$ such that for every $n > 3$, the alternating group $A_n$ is not a quotient of $G$?

  • $\begingroup$ Since the meaning of "any" might be ambiguous, the question is, if I'm correct: does there exist some know group $G$ such that for every $n\ge 3$, the alternating group $\mathrm{Alt}_n$ is not a quotient of $G$. Right? $\endgroup$
    – YCor
    Commented May 17 at 16:51
  • 3
    $\begingroup$ The heuristic model of Dunfield-Thurston suggests that the probability a fundamental group of "random" $3$-manfold surjects onto $A_n$ tends to a (positive) limiting probability as $n \rightarrow \infty$ (arxiv.org/pdf/math/0502567), so it seems exceedingly unlikely for hyperbolic knots. OTOH Torus knots have fundamental group $\langle x,y |x^p=y^q \rangle$ and I suspect it should be elementary to prove that this group has many alternating quotients outside the degenerate cases. $\endgroup$
    – user523984
    Commented May 17 at 18:48
  • 2
    $\begingroup$ Re: the previous comment: it's not hard to show a $p$-cycle and a $q$-cycle (where $p<q$ are both odd) generate $A_q$. $\endgroup$
    – Steve D
    Commented May 17 at 22:42
  • $\begingroup$ @YCor There might be some misunderstandings when I said “abelianization” I was referring to the exclusion of the unknot. Similar to the nontrivial knot…I’ll use the phrase “unknot” then. $\endgroup$
    – Shijie Gu
    Commented May 18 at 10:35
  • 1
    $\begingroup$ Probably you should say "non-abelian knot group", because "nontrivial group" usually means not reduced to the unit. $\endgroup$
    – YCor
    Commented May 18 at 15:15

1 Answer 1


The unknot and the hopf link have abelian knot groups. See here. Hence they can't surject onto an $A_n$ for $n\gt3.$

  • 10
    $\begingroup$ :D Of course this doesn’t really answer the interesting version of the question. $\endgroup$
    – HJRW
    Commented May 17 at 17:52

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.