7
$\begingroup$

I'm not a topologist and I just saw the definition of prime knot a while ago.

Today I'm somewhat supprised to realize that I don't even know if there are infinitely many prime knots? If this question is not completely trivial then I'm hoping to see a "proof from the book".

(A related question is, is the decomposition of a composite knot into prime knots unique? I would hope so but I don't have a strong reason to support that.)

$\endgroup$
4
  • 11
    $\begingroup$ Yes; in fact there are infinitely many distinct torus knots, all of which are prime (en.wikipedia.org/wiki/Torus_knot). $\endgroup$ Nov 22, 2011 at 19:16
  • 4
    $\begingroup$ Knots do decompose uniquely into a sum of primes, see e.g. chapter 2 of Lickorish's "An introduction to knot theory." Also, since knot genus is additive under connected sum it follows that every genus 1 knot is prime, so take your favorite knot and consider all of its twisted Whitehead doubles; these have genus 1 and are distinguished by their Alexander polynomials. $\endgroup$ Nov 22, 2011 at 20:07
  • $\begingroup$ It's a theorem of Schubert's from the 1930's that oriented knots under the connect-sum operation are a free commutative monoid on infinitely many generators. $\endgroup$ Nov 23, 2011 at 2:01
  • 2
    $\begingroup$ These notes give a basic exposition of the infinity of prime knots, and Schubert's proof of uniqueness of prime factorization: math.ucla.edu/~radko/191.1.05w/marcos.pdf $\endgroup$ Nov 23, 2011 at 3:28

2 Answers 2

14
$\begingroup$

Yes there are infinitely many prime knots.

Not a "proof from the book", but all (p,2) torus knots for p prime are prime, and they have different Alexander polynomials.

$\endgroup$
2
$\begingroup$

Yes.

Kauffman and Lopes found some very elementary families: https://arxiv.org/pdf/1604.02510.pdf .

$\endgroup$

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.