(1, 1)-knot
のunknotting tunnel
東京農工大学 合田洋
(
林忠一郎氏(日本女子大学)との共同研究)
M : a closed orientable 3-manifold, K (⊂ M ) : a knot
Definition. A knot K is called a (1, 1)-knot if the pair (M, K ) can be decomposed into (H 1 , K 1 ) ∪ (H 2 ∪ K 2 ) where H i is a solid torus and K i is a boundary parallel arc properly embedded in H i .
H 2
1
K 1 K 2
H (1,1)-decomposition
Definition. Let τ be an arc with τ ∩K = ∂τ . τ is called an unknotting tunnel of K if M − IntN (K ∪ τ ) ia s genus 2 handlebody.
K
τ
V V
1 2
unknotting tunnel
A knot which has an unknotting tunnel is called a tunnel number
one knot.
Proposition. (1, 1)-knots are tunnel number one.
Proposition. (1, 1)-knots are tunnel number one.
H 2
1
K 1 K 2
H
Proposition. (1, 1)-knots are tunnel number one.
K
1K
2K
1K
2Proposition. (1, 1)-knots are tunnel number one.
K
1K
2K
1K
2I I ~ I I ~
K
Proposition. (1, 1)-knots are tunnel number one.
K
1K
2K
1K
2I I ~ I I ~
τ
1Proposition. (1, 1)-knots are tunnel number one.
K K
1
2
K
1K
2I I ~ I I ~
τ
1τ
1Definition. The unknotting tunnel obtained from a (1, 1)-decomposition is called a (1, 1)-tunnel. It has a dual tunnel.
K
1K
2τ
1τ
2In what follows, we consider the case of M = S 3 .
Theorem.[Boileau-Rost-Zieschang (88),
森元(89)] A torus knot has
at most 3 types unknotting tunnels. Two of them are (1, 1)-tunnels.
Torus knot T (5, 8)
Torus knot T (5, 8)
Torus knot T (5, 8)
τ
τ
τ
0
1 2
Torus knot T (5, 8)
Theorem.[
小林(99)] A two birdge knot has at most 4 unknotting tunnels (up to homeo.). All of them are (1, 1)-tunnels.
Theorem.[
森元−作間(91)] A satellite knot has at most 4 unknotting tunnels (up to homeo.). All of them are (1, 1)-tunnels.
Problem. Is there a tunnel number one knot which does not have a (1, 1)-tunnel ?
Answer. Yes.
森元−作間−横田knot (The twisted torus knot, (96)).
MSY(7, 17, 10m − 4) (m ∈ N ) is not (1, 1).
Theorem.[Scharlemann
−Thompson -
合(00)] Any unknotting tun-
( p,q ) - torus knot
r - half twists unknotting tunnel
p,q r ( ; ) MSY
From the observation of the torus knots and the previous theorem,
Question. Let K be a (1, 1)-knot. If K has an unknotting tunnel τ
§ Results of Rubinstein-Scharlemann and
小林−佐伯Let M = H 1 ∪ H H 2 = V 1 ∪ V V 2 be two Heegaard splittings of M . Roughly speaking, Rubinstein-Scharlemann proved that H and V may intersect in a non-empty collection of simple closed curves which are essential in both H and V .
小林−佐伯
studied the case of 3-manifolds with links. Applying
their results to our case, we may assume that:
Let (M, K ) = (H 1 , K 1 ) ∪ H (H 2 , K 2 ) : a (1, 1)-decomposition
= (V 1 , K ) ∪ V (V 2 , ∅) : a genus 2 Heegaard split.
Then, H and V may intersect in a non-empty collection of simple closed curves which are K -essential in both H and V .
Let ` be the minimum number of such simple closed curves, and τ be the unknotting tunnel associated with (V 1 , K ) ∪ V (V 2 , ∅).
Theorem.[
林−合] If ` 6= 2, one of the following conditions holds:
(1) τ is a (1, 1)-tunnel;
(2) τ may be isotoped into H fixing the knot K ;
Examples by Song : MSY(5,7,2).
τ
τ
A criterion for deciding ‘not (1, 1)-tunnel’
Let θ be a θ-graph obtained from K ∪ τ . We denote by θ 1 and θ 2 the consituent knots of θ.
K
τ
θ θ
θ
12
Proposition.
Examples by Song : MSY(5,7,2).
τ
τ
Examples by Song : MSY(5,7,2).
τ
ǫ
ǫ
ǫ
ǫ
Both θ 1 and θ 2 are not the trivial knots. Hence, τ 2 is not a (1, 1)-
Question.[Scharlemann] If we slide τ enough, both θ 1 and θ 2 become non-trivial knots although τ is a (1, 1)-tunnel ?
Answer. Yes. So, we need some conditions.
Suggestion.[Song] How about a thin position for τ ?
このあと,進展なし,,,
論文,ホームページに置きっぱ,,,
In the recent 10 years, some invariants for the (unknotting) tunnels were defined: ‘distance’ (Johnson), ‘depth’ (Cho-McCullough),
‘t-distance’ (
古宇田).
Theorem.[Scharlemann-Tomova, Johnson (06)]
dist(τ ) ≥ 6 = ⇒ τ is the unique unknotting tunnel.
Theorem.[Johnson] For any n, there is an unknotting tunnel τ such that dist(τ ) > n.
(印象)えーっつ! ほとんどユニーク!?
Problem. dist(τ ) ≤ 5
のunknotting tunnel
ってどのようなものか?分類可能か?
このうちユニークなものは? ユニークでない場合,そのうち一 つが
(1, 1)-tunnel
ならばどんなものか?Fact. (1) depth(τ ) = 1 ⇐⇒ τ is a (1, 1)-tunnel.
(2)
石原found an algorithm for calculating the depth of unknotting tunnels.
Theorem.[
石原] The depth of the unknotting tunnel τ 2 for MSY(5,7,2)
MSY(5,7,2)
とτ 2
の組はある操作によって一般化出来るが,こ れらについても同様のことが言えることを石原は示した.謝辞
本講演をするにあたり古宇田悠哉さん,石原海さんに様々な事 を教えて頂きました.感謝いたします.
ご静聴ありがとうございました.