Fix a knot type $K \subset S^3$, and consider the set $$Y_K = \{ \mbox{Diagrams of }K \} / \mbox{planar isotopy}.$$ We can turn $Y_K$ into a metric space by considering the distance induced by Reidemeister moves: $d(D_0,D_1) = $ minimum length of a sequence of Reidemeister moves (each move is possibly followed by a planar isotopy), and $D_0,D_1$ are two diagrams of $K$.

We can build a graph $\mathcal{G}_K$ as follows: the vertices are given by the points of $Y_K$, while the edges are given by the pairs $(D_0,D_1)$ such that $d(D_0,D_1)=1$.

This construction parallels Hirasawa and Yoshiaki's in "The Gordian Complex of Knots" for the Gordian metric space of knots, but unlike the Gordian complex, each vertex of $\mathcal{G}_K$ has only finitely many edges (the number of edges at a vertex $D$ corresponds to the possible inequivalent diagrams - up to planar isotopy - that can be reached from $D$ by a single Reidemeister move).

For each knot $K$ we can thus define a function $$\varphi_K : Y_K \longrightarrow \mathbb{N}$$ called the simplicity, as $\varphi_K (D) = \#\{$edges of $\mathcal{G}_K$ that contain $D$ in their boundary$\}$.

As an example the "standard" diagram of a trefoil trefoil-1 has a lower simplicity than trefoil-2 (you need to take into account that the former is 3-symmetric).

Taking the minimum of the simplicity over all diagrams of a knot yields an invariant $\mathcal{S}(K)$.

Is the invariant $\mathcal{S}(K)$, or any possible application/connection, known?

  • 2
    $\begingroup$ Possibly stupid question: in the definition of simplicity, what does it mean for an edge to contain a vertex in its boundary? (What is the boundary of an edge?) At first I thought you just meant the two vertices on which the edge is incident, but then simplicity would be the same as degree. $\endgroup$ Feb 11, 2014 at 23:35
  • 1
    $\begingroup$ I didn't understand your example- the diagrams of trefoil-1 and of trefoil-2 are the same up to reparametrization. $\endgroup$ Feb 12, 2014 at 2:03
  • 3
    $\begingroup$ Other than by knowing a complete set of diagrams for K, is there any way to compute S(K)? $\endgroup$ Feb 12, 2014 at 8:16
  • 1
    $\begingroup$ @NateEldredge you are right, it is the degree, i just didn't know it was called that way. $\endgroup$ Feb 12, 2014 at 9:13
  • 1
    $\begingroup$ @DanieleCeloria Yeah, I didn't read carefully enough and was imagining we were talking about the graph where there would be multiple edges for multiple Reidemeister links between the same ambient-isotopic diagrams. $\endgroup$ Feb 13, 2014 at 16:26

1 Answer 1


I think the minimal degree of "Reidemeister complex" has some meaning.

I know one paper about this topic.

A distance for diagrams of a knot

  • $\begingroup$ Thanks, this paper is really useful, but it actually focuses on prime Reidemeister moves, and doesn't seem to care for planar isotopies, so the invariants defined should be different from $\mahtcal{S}$. $\endgroup$ Feb 13, 2014 at 15:01

Your Answer

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

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