Homotopy invariants of Gauss words (Q532538)

The author proves that the number of homotopy classes of Gauss words is infinite. This result answers negatively a conjecture due to \textit{V. Turaev} [``Lectures on topology of words'', Jpn. J. Math. (3) 2, No. 1, 1--39 (2007; Zbl 1162.68034)] that open homotopy of Gauss words is trivial, in other words, the number of open homotopy classes of Gauss words is one. It is not only the negative answer but also a strengthened result. Precisely, the negative answer should be that the number of open homotopy classes of Gauss words is at least two. The author's result removes `open', and changes `at least two' into `infinite' in the statement. \textit{V. O. Manturov} [``On free knots'', \url{math.GT/0901.2214}] obtains the same result. Note that the term `homotopy' in nanoword theory is not equal to that in algebraic topology, it just means one of the equivalence relations for nanowords. A Gauss word is a finite sequence of letters (i.e. a word) with the condition that every letter in the sequence appears exactly twice. The basic homotopy for Gauss words comes from the Reidemeister moves for virtual knots. When we introduce an additional structure to Gauss words, we obtain nanowords. Virtual knots, flat virtual knots, Legendrian knots, and so on are examples of nanowords. We can also define the multi-component cases for Gauss words and nanowords. They are called Gauss phrases and nanophrases, respectively. Suppose that the basic homotopy above allows the shift move of the start point. If we do not allow the shift move, then we obtain open Gauss words and open nanowords. Hence the equivalence classes of open nanowords are finer than those of nanowords. Usually, an open virtual knot is called a long virtual knot. The proof of the author's result above is divided into two parts: Corollary 4.4 states that there exist Gauss words that are not homotopically trivial. Proposition 5.9 states that there is an infinite number of homotopy classes of Gauss words. To prove Corollary 4.4, the author defines a homotopy invariant \(z\) which comes from the mod 2 intersection number of two curves (Theorem 4.1). To prove Proposition 5.9, the author defines a covering for a Gauss word, and shows that the homotopy class of the covering is a homotopy invariant of the Gauss word (Proposition 5.2). Since \textit{V. Turaev} [``Topology of words'', Proc. Lond. Math. Soc. (3) 95, No. 2, 360--412 (2007; Zbl 1145.57018)] defined a covering only for open nanowords, Proposition 5.2 is a new result. The author defines the height of a Gauss word to be the number of hierarchies of coverings of the Gauss word. He proves Proposition 5.9. by a construction that yields an infinite family of Gauss words with different heights, thus being mutually non-homotopic. In Section 6, the author shows that the homotopy invariants of Gauss words in the paper can be modified into open homotopy invariants (Theorem 6.1). He gives an example of a Gauss word which is homotopically trivial but open homotopically non-trivial (Example 6.2). In Section 7, it is remarked that the set of homotopy classes of Gauss words is equivalent to the set of virtual knots modulo crossing changes and the virtual switch.