30.11.2014 Views

Lecture notes of Matilde Marcolli - Vanderbilt University

Lecture notes of Matilde Marcolli - Vanderbilt University

Lecture notes of Matilde Marcolli - Vanderbilt University

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Noncommutative geometry, the<br />

field with one element, and<br />

3-manifolds<br />

<strong>Matilde</strong> <strong>Marcolli</strong><br />

2009


The field with one element<br />

Finite geometries (q = p k , p prime)<br />

#P n−1 (F q ) = #(An (F q ) {0})<br />

#G m (F q )<br />

= qn − 1<br />

q − 1 = [n] q<br />

#Gr(n, j)(F q ) = #{P j (F q ) ⊂ P n (F q )}<br />

=<br />

[n] q !<br />

[j] q ![n − j] q ! = ( n)<br />

j q<br />

[n] q ! = [n] q [n − 1] q · · ·[1] q , [0] q ! = 1<br />

J. Tits: take q = 1<br />

P n−1 (F 1 ) := finite set <strong>of</strong> cardinality n<br />

Gr(n, j)(F 1 ) := set <strong>of</strong> subsets <strong>of</strong> cardinality j<br />

Algebraic geometry over F 1 ?<br />

1


Extensions F 1 n (Kapranov-Smirnov)<br />

Monoid {0} ∪ µ n (n-th roots <strong>of</strong> unity)<br />

- Vector space over F 1 n: pointed set (V, v) with<br />

free action <strong>of</strong> µ n on V {v}<br />

- Linear maps: permutations compatible with<br />

the action<br />

F 1 n ⊗ F1 Z := Z[t, t −1 ]/(t n − 1)<br />

Zeta functions (Manin) as variety over F 1 :<br />

Z(T n , s) = s − n<br />

2π<br />

with T n = G n m<br />

⇒ F 1 -geometry and infinite primes (Arakelov)<br />

Γ C (s) −1 = ((2π) −s Γ(s)) −1 = ∏ s + n<br />

2π<br />

n≥0<br />

like a “dual” infinite projective space ⊕ ∞ n=0 T−n over F 1<br />

Note: Spec(Z) not a (fin. type) scheme over<br />

F 1 : what is Z ⊗ F1 Z??<br />

More recently: Soulé, Haran, Dourov, Toen–<br />

Vaquie, Connes–Consani, Borger<br />

2


Focus on two approaches:<br />

Descent data for rings from Z to F 1 :<br />

• by cyclotomic points (Soulé)<br />

• by Λ-ring structure (Borger)<br />

Gadgets over F 1<br />

C. Soulé, Les variétés sur le corps à un élément, Mosc.<br />

Math. J. Vol.4 (2004) N.1, 217–244.<br />

(X, A X , e x,σ )<br />

• X : R → Sets covariant functor,<br />

R finitely generated flat rings<br />

• A X complex algebra<br />

• evaluation maps: for all x ∈ X(R), σ : R → C<br />

⇒ e x,σ : A X → C algebra homomorphism<br />

e f(y),σ = e y,σ◦f<br />

for f : R ′ → R ring homomorphism<br />

Affine varieties V Z ⇒ gadget X = G(V Z ) with<br />

X(R) = Hom(O(V ), R) and A X = O(V ) ⊗ C<br />

3


Affine variety over F 1 (Soulé)<br />

Gadget with X(R) finite; variety X Z and morphism<br />

<strong>of</strong> gadgets<br />

X → G(X Z )<br />

such that all X → G(V Z ) come from X Z → V Z<br />

Should think: R = Z[Z/nZ] ⇒ X(R) cyclotomic points<br />

Λ-rings, endomotives, and F 1<br />

J. Borger, Λ-rings and the field with one element, preprint<br />

2009<br />

- Grothendieck: characteristic classes, Riemann–<br />

Roch<br />

- Λ-ring structure: “descent data” for a ring<br />

from Z to F 1<br />

Torsion free R with action <strong>of</strong> semigroup N by<br />

endomorphisms lifting Frobenius<br />

s p (x) − x p ∈ pR,<br />

Morphisms: f ◦ s k = s k ◦ f<br />

∀x ∈ R<br />

Q-algebra A ⇒ Λ-ring<br />

iff action <strong>of</strong> Gal(¯Q/Q) × N on X = Hom(A, ¯Q) factors<br />

through an action <strong>of</strong> Ẑ<br />

4


NCG <strong>of</strong> Q-lattices<br />

Alain Connes, M.M., Quantum statistical mechanics <strong>of</strong><br />

Q-lattices in “Frontiers in number theory, physics, and<br />

geometry. I”, 269–347, Springer, Berlin, 2006.<br />

Definition: (Λ, φ) Q-lattice in R n<br />

lattice Λ ⊂ R n + labels <strong>of</strong> torsion points<br />

φ : Q n /Z n −→ QΛ/Λ<br />

group homomorphism (invertible Q-lat if isom)<br />

φ<br />

general case<br />

invertible case<br />

φ<br />

Commensurability (Λ 1 , φ 1 ) ∼ (Λ 2 , φ 2 )<br />

iff QΛ 1 = QΛ 2 and φ 1 = φ 2 mod Λ 1 + Λ 2<br />

Q-lattices / Commensurability ⇒ NC space<br />

5


1-dimensional case<br />

(Λ, φ) = (λ Z, λ ρ) λ > 0<br />

ρ ∈ Hom(Q/Z, Q/Z) = lim ←−n<br />

Z/nZ = Ẑ<br />

Up to scaling λ: algebra C(Ẑ)<br />

Commensurability Action <strong>of</strong> N = Z >0<br />

α n (f)(ρ) = f(n −1 ρ)<br />

zero otherwise<br />

(partially defined action <strong>of</strong> Q ∗ + )<br />

1-dimensional Q-lattices up to scale / Commens.<br />

⇒ NC space C(Ẑ) ⋊ N<br />

Crossed product algebra<br />

f 1 ∗ f 2 (r, ρ) =<br />

f ∗ (r, ρ) = f(r −1 , rρ)<br />

∑<br />

s∈Q ∗ + ,sρ∈Ẑ<br />

f 1 (rs −1 , sρ)f 2 (s, ρ)<br />

6


Bost–Connes algebra<br />

J.B. Bost, A. Connes, Hecke algebras, type III factors<br />

and phase transitions with spontaneous symmetry<br />

breaking in number theory. Selecta Math. (N.S.) Vol.1<br />

(1995) N.3, 411–457.<br />

Algebra A Q,BC = Q[Q/Z] ⋊ N generators and<br />

relations<br />

µ n µ m = µ nm<br />

µ n µ ∗ m = µ∗ m µ n when (n, m) = 1<br />

µ ∗ nµ n = 1<br />

e(r + s) = e(r)e(s), e(0) = 1<br />

ρ n (e(r)) = µ n e(r)µ ∗ n = 1 n<br />

∑<br />

ns=r<br />

e(s)<br />

C ∗ -algebra C ∗ (Q/Z) ⋊ N = C(Ẑ) ⋊ N<br />

Time evolution<br />

σ t (e(r)) = e(r),<br />

σ t (µ n ) = n it µ n<br />

Hamiltonian Tr(e −βH ) = ζ(β)<br />

7


Quantum statistical mechanics<br />

(A, σ t ) C ∗ -algebra and time evolution<br />

State: ϕ : A → C linear ϕ(1) = 1, ϕ(a ∗ a) ≥ 0<br />

Time evolution σ t ∈ Aut(A)<br />

rep. on Hilbert space H ⇒ Hamiltonian H =<br />

d<br />

dt σ t| t=0<br />

Equilibrium states (inverse temperature β = 1/kT)<br />

1<br />

Z(β) Tr ( a e −βH)<br />

Classical points <strong>of</strong> NC space<br />

Z(β) = Tr ( e −βH)<br />

KMS states ϕ ∈ KMS β (0 < β < ∞)<br />

∀a, b ∈ A ∃ holom function F a,b (z) on strip: ∀t ∈ R<br />

F a,b (t) = ϕ(aσ t (b))<br />

F a,b (t + iβ) = ϕ(σ t (b)a)<br />

8


More on Bost–Connes<br />

• Representations π ρ on l 2 (N):<br />

µ n ǫ m = ǫ nm , π ρ (e(r))ǫ m = ζ m r ǫ m<br />

ζ r = ρ(e(r)) root <strong>of</strong> 1, for ρ ∈ Ẑ ∗<br />

• Low temperature extremal KMS (β > 1)<br />

ϕ β,ρ (a) = Tr(π ρ(a)e −βH )<br />

Tr(e −βH , ρ ∈ Ẑ ∗<br />

)<br />

High temperature: unique KMS state<br />

• Zero temperature: evaluations ϕ ∞,ρ (e(r)) = ζ r<br />

ϕ ∞,ρ (a) = 〈ǫ 1 , π ρ (a)ǫ 1 〉<br />

Intertwining: a ∈ A Q,BC , γ ∈ Ẑ ∗<br />

ϕ ∞,ρ (γa) = θ γ (ϕ ∞,ρ (a))<br />

θ : Ẑ ∗ ≃ → Gal(Q<br />

ab /Q)<br />

Class field theory isomorphism<br />

9


Bost–Connes endomotive<br />

A.Connes, C.Consani, M.M., Noncommutative geometry<br />

and motives: the thermodynamics <strong>of</strong> endomotives,<br />

Adv. in Math. 214 (2) (2007), 761–831<br />

A = lim −→n<br />

A n with A n = Q[Z/nZ]<br />

abelian semigroup action S = N on A = Q[Q/Z]<br />

Endomotives (A, S) from self maps <strong>of</strong> algebraic<br />

varieties s : Y → Y , s(y 0 ) = y 0 unbranched,<br />

X s = s −1 (y 0 ), X = lim ←−<br />

X s = Spec(A)<br />

ξ s,s ′ : X s ′ → X s , ξ s,s ′(y) = r(y), s ′ = rs ∈ S<br />

Bost–Connes endomotive:<br />

G m with self maps u ↦→ u k<br />

s k : P(t, t −1 ) ↦→ P(t k , t −k ), k ∈ N, P ∈ Q[t, t −1 ]<br />

ξ k,l (u(l)) = u(l) k/l , u(l) = t mod t l − 1<br />

X k = Spec(Q[t, t −1 ]/(t k − 1)) = s −1<br />

k<br />

(1) and<br />

X = lim ←−k<br />

X k<br />

u(l) ↦→ e(1/l) ∈ Q[Q/Z],<br />

C(X(¯Q)) = C(Ẑ)<br />

10


Integer model <strong>of</strong> the Bost–Connes algebra<br />

A.Connes, C.Consani, M.M., Fun with F 1 , to appear in<br />

J.Number Theory, arXiv:0806.2401<br />

A Z,BC generated by Z[Q/Z] and µ ∗ n , ˜µ n<br />

˜µ n˜µ m = ˜µ nm<br />

µ ∗ nµ ∗ m = µ ∗ nm<br />

µ ∗ n˜µ n = n<br />

˜µ n µ ∗ m = µ∗ m˜µ n (n, m) = 1.<br />

µ ∗ nx = σ n (x)µ ∗ n and x˜µ n = ˜µ n σ n (x)<br />

where σ n (e(r)) = e(nr) for r ∈ Q/Z<br />

Note: ρ n (x) = µ n xµ ∗ n ring homomorphism but<br />

not ˜ρ n (x) = ˜µ n xµ ∗ n (correspondences “crossed<br />

product” A Z,BC = Z[Q/Z] ⋊˜ρ N)<br />

11


NCG and Soulé’s F 1 -geometry:<br />

Roots <strong>of</strong> unity as varieties over F 1 :<br />

µ (k) (R) = {x ∈ R | x k = 1} = Hom Z (A k , R)<br />

A k = Z[t, t −1 ]/(t k − 1)<br />

• Inductive system G m over F 1 :<br />

µ (n) (R) ⊂ µ (m) (R), n|m, A m ։ A n<br />

A X = C(S 1 )<br />

• Projective system (BC): ξ m,n : X n ։ X m<br />

ξ m,n : µ (n) (R) ։ µ (m) (R),<br />

n|m<br />

µ ∞ (R) = Hom Z (Z[Q/Z], R)<br />

⇒ projective system <strong>of</strong> affine varieties over F 1<br />

ξ m,n : F 1 n ⊗ F1 Z → F 1 m ⊗ F1 Z<br />

A X = C[Q/Z]<br />

12


Bost–Connes and F 1<br />

A.Connes, C.Consani, M.M., Fun with F 1 , arXiv:0806.2401<br />

Affine varieties µ (n) over F 1 defined by gadgets<br />

G(Spec(Q[Z/nZ])); projective system<br />

Endomorphisms σ n (<strong>of</strong> varieties over Z, <strong>of</strong> gadgets,<br />

<strong>of</strong> F 1 -varieties)<br />

Extensions F 1 n: free actions <strong>of</strong> roots <strong>of</strong> 1<br />

(Kapranov–Smirnov)<br />

ζ ↦→ ζ n , n ∈ N<br />

and ζ ↦→ ζ α ↔ e(α(r)), α ∈ Ẑ<br />

Frobenius action on F 1 ∞<br />

In reductions mod p <strong>of</strong> integral Bost–Connes<br />

endomotive ⇒ Frobenius<br />

Bost–Connes = extensions F 1 n plus Frobenius<br />

13


Characteristic p versions <strong>of</strong> the BC endomotive<br />

Q/Z = Q p /Z p × (Q/Z) (p)<br />

denom = power <strong>of</strong> p; denom = prime to p<br />

K[Q p /Z p ] ⋊ p Z+<br />

endomorphisms σ n for n = p l , l ∈ Z +<br />

ϕ Fp (x) = x p Frobenius <strong>of</strong> K char p<br />

f ∈ K[Q/Z]<br />

(σ p l ⊗ ϕ l F p<br />

)(f) = f pl<br />

(σ p l ⊗ϕ l F p<br />

)(e(r)⊗x) = e(p l r)⊗x pl = (e(r)⊗x) pl<br />

⇒ BC endomorphisms restrict to Frobenius on<br />

mod p reductions: σ p l Frobenius correspondence<br />

on pro-variety µ ∞ ⊗ Z K<br />

14


NCG and Borger’s F 1 -geometry<br />

M.M., Cyclotomy and endomotives, preprint arXiv:0901.3167<br />

Multivariable Bost–Connes endomotives<br />

Variety T n = (G m ) n endomorphisms α ∈ M n (Z) +<br />

X α = {t = (t 1 , . . . , t n ) ∈ T n | s α (t) = t 0 }<br />

ξ α,β : X β → X α , t ↦→ t γ , α = βγ ∈ M n (Z) +<br />

t ↦→ t γ = σ γ (t) = (t γ 11<br />

1 tγ 12<br />

2 · · · tγ 1n<br />

n , . . . , tγ n1<br />

1 tγ n2<br />

2 · · · tγ nn<br />

n )<br />

X = lim ←−α<br />

X α with semigroup action<br />

C(X(¯Q)) ∼ = Q[Q/Z] ⊗n generators e(r 1 ) ⊗ · · · ⊗ e(r n )<br />

A n = Q[Q/Z] ⊗n ⋊ ρ M n (Z) +<br />

generated by e(r) and µ α , µ ∗ α<br />

ρ α (e(r)) = µ α e(r)µ ∗ α = 1<br />

det α<br />

∑<br />

α(s)=r<br />

σ α (e(r)) = µ ∗ α e(r)µ α = e(α(r))<br />

Endomorphisms:<br />

σ α (e(r)) = µ ∗ α e(r)µ α<br />

e(s)<br />

15


The Bost–Connes endomotive is a direct limit<br />

<strong>of</strong> Λ-rings<br />

R n = Z[t, t −1 ]/(t n −1) s k (P)(t, t −1 ) = P(t k , t −k )<br />

Action <strong>of</strong> Ẑ:<br />

α ∈ Ẑ : (ζ : x ↦→ ζx) ↦→ (ζ : x ↦→ ζ α x)<br />

Ẑ = Hom(Q/Z, Q/Z)<br />

⇒ Frobenius over F 1 ∞ (Haran)<br />

Embeddings <strong>of</strong> Λ-rings (Borger–de Smit)<br />

Every torsion free finite rank Λ-ring embeds<br />

in Z[Q/Z] ⊗n with action <strong>of</strong> N compatible with<br />

S n,diag ⊂ M n (Z) +<br />

BC systems as universal Λ-rings<br />

⇒ multivariable Bost–Connes endomotives as<br />

universal Λ-rings<br />

16


Quantum statistical mechanics<br />

<strong>of</strong> multivariable Bost–Connes endomotives<br />

Representations on l 2 (M n (Z) + ): ζ α i<br />

= ∏ j ζα ij<br />

i<br />

ζ = (ζ 1 , . . . , ζ n ) = ρ(r),<br />

ρ ∈ GL n (Ẑ)<br />

µ β ǫ α = ǫ βα<br />

π ρ (e(r))ǫ α = ∏ i<br />

ζ ατ<br />

i ǫ α<br />

Time evolution:<br />

σ t (e(r)) = e(r),<br />

σ t (µ α ) = det(α) it µ α<br />

Hamiltonian:<br />

Hǫ α = logdet(α) ǫ α<br />

Problem: Infinite multiplicities in the spectrum:<br />

SL n (Z)-symmetry<br />

17


Groupoid and convolution algebra<br />

Similar to: A.Connes, M.M. Quantum statistical mechanics<br />

<strong>of</strong> Q-lattices, in “Frontiers in Number Theory,<br />

Physics, and Geometry, I” pp.269–350, Springer, 2006.<br />

U Γ = {(α, ρ) ∈ Γ\GL n (Q) + × Γ Ẑ n | α(ρ) ∈ Ẑ n }<br />

Quotient U Γ by SL n (Z) × SL n (Z)<br />

(γ 1 , γ 2 ) : (α, ρ) ↦→ (γ 1 αγ2 −1 , γ 2(ρ))<br />

Convolution algebra<br />

(f 1 ⋆ f 2 )(α, ρ) =<br />

∑<br />

(α,ρ)=(α 1 ,ρ 1 )◦(α 2 ,ρ 2 )∈U Γ<br />

f 1 (α 1 , ρ 1 )f 2 (α 2 , ρ 2 )<br />

f ∗ (α, ρ) = f(α −1 , α(ρ)) and σ t (f)(α, ρ) = det(α) it f(α, ρ)<br />

∑<br />

(π ρ (f)ξ)(α) =<br />

f(αβ −1 , β(ρ))ξ(β)<br />

β∈Γ\GL n (Q) + : βρ∈Ẑ ∗<br />

On l 2 (Γ\G ρ ). If ρ ∈ (Ẑ ∗ ) n :<br />

G ρ = {α ∈ GL n (Q) + | α(ρ) ∈ Ẑ n } = M n (Z) +<br />

Z(β) =<br />

∑<br />

m∈Γ\M n (Z) + det(m) −β 18


Why QSM?<br />

A.Connes, C.Consani, M.M., The Weil pro<strong>of</strong> and the geometry<br />

<strong>of</strong> the adeles class space, arXiv:math/0703392<br />

For original BC system ⇒ dual system<br />

 = A ⋊ σ R. Restriction map:<br />

δ : Â ♮ β → S♮ (Ω β )<br />

Cyclic module: cokernel HC 0 (D(A, ϕ))<br />

⇒ Spectral realization <strong>of</strong> zeros <strong>of</strong> Riemann<br />

zeta function and trace formula<br />

Multivariable case:<br />

Manin (1994): problem <strong>of</strong> Z ⊗ F1 Z =???<br />

cf. Weil’s pro<strong>of</strong> for function fields C × Fq C<br />

Zeta function viewpoint: Kurokawa<br />

tensor product zeta function: summing over zeros<br />

∏<br />

λ∈Ξ(s − λ) m λ<br />

⊗ ∏ µ∈Θ<br />

∏<br />

(s − µ) n µ<br />

:= (s − λ − µ) m λ+n µ<br />

(λ,µ)<br />

as zeta-regularized infinite products<br />

Question: Kurokawa zeta functions from multivariable<br />

dual systems?<br />

19


Thermodynamics <strong>of</strong> endomotives and RH<br />

A.Connes, C.Consani, M.M., Noncommutative geometry<br />

and motives: the thermodynamics <strong>of</strong> endomotives,<br />

Adv. in Math. 214 (2) (2007), 761–831<br />

• Dual system: Â = A BC ⋊ σ R<br />

• Scaling action: θ λ ( ∫ x(t) U t dt) = ∫ λ it x(t) U t dt<br />

• Classical points: ˜Ω β ≃ Ω β × R ∗ +<br />

• Restriction map: (morphism <strong>of</strong> cyclic modules)<br />

δ : Â β<br />

low T KMS states<br />

π<br />

−→ C(˜Ω β , L 1 ) −→ Tr C(˜Ω β )<br />

• Cokernel (abelian category) D(A, ϕ) = Coker(δ)<br />

• Cyclic homology (with scaling action) HC 0 (D(A, ϕ))<br />

• Galois action+ scaling C Q = Ẑ ∗ × R ∗ +<br />

Weil explicit formula as trace formula<br />

on H 1 = HC 0 (D(A, ϕ)):<br />

Tr(ϑ(f)| H 1) = ˆf(0)+ˆf(1)−∆•∆f(1)− ∑ v<br />

ϑ(f) =<br />

∫<br />

∫ ′<br />

C Q<br />

f(g)ϑ g d ∗ g f ∈ S(C Q )<br />

(K ∗ v,e Kv )<br />

Self inters <strong>of</strong> diagonal ∆ • ∆ = log |a| = −log |D|<br />

f(u −1 )<br />

|1 − u| d∗ u<br />

(D = discriminant for #-field, Euler char χ(C) for F q (C))<br />

20


Note:<br />

• Connes’ 1997 RH paper: Tr(R Λ U(f)):<br />

- only critical line zeros<br />

- Trace formula (global) ⇔ RH<br />

• Tr(ϑ(f)| H 1):<br />

- all zeros involved<br />

- RH ⇔ positivity<br />

Tr ( ϑ(f ⋆ f ♯ )| H 1)<br />

≥ 0 ∀f ∈ S(CQ )<br />

where<br />

(f 1 ⋆ f 2 )(g) =<br />

∫<br />

f 1 (k)f 2 (k −1 g)d ∗ g<br />

multiplicative Haar measure d ∗ g and adjoint<br />

f ♯ (g) = |g| −1 f(g −1 )<br />

⇒ Better for comparing with Weil’s pro<strong>of</strong> for<br />

function fields<br />

21


Weil’s pro<strong>of</strong> in a nutshell<br />

K = F q (C) function field, Σ K = places deg n v = # orbit<br />

<strong>of</strong> Fr on fiber C(¯F q ) → Σ K<br />

ζ K (s) = ∏ Σ K<br />

(1 − q −n vs ) −1 =<br />

P(q −s )<br />

(1 − q −s )(1 − q 1−s )<br />

P(T) = ∏ (1−λ n T) char polynomial <strong>of</strong> Fr ∗ on H 1 et (¯C, Q l )<br />

C(¯F q ) ⊃ Fix(Fr j ) = ∑ k<br />

(−1) k Tr(Fr ∗j |H k et(¯C, Q l ))<br />

RH ⇔ eigenvalues λ n with |λ j | = q 1/2<br />

Correspondences: divisors Z ⊂ C × C; degree, codegree,<br />

trace:<br />

d(Z) = Z • (P × C) d ′ (Z) = Z • (C × P)<br />

Tr(Z) = d(Z) + d ′ (Z) − Z • ∆<br />

RH ⇔ Weil positivity Tr(Z ⋆ Z ′ ) > 0<br />

⇒ Dictionary between NCG and Alg Geom:<br />

C × Fq C ⇔ NCG version <strong>of</strong> Spec(Z)× F1 Spec(Z)<br />

22


A possible intermediate step: Witt vectors<br />

Affine ring scheme <strong>of</strong> big Witt vectors<br />

Z[λ 1 , λ 2 , . . . , λ n , . . .]<br />

symmetric functions λ i , ghost coordinates<br />

ψ n = ∑ d|n<br />

d u n/d<br />

d<br />

, with ψ n = x n 1 + xn 2 + · · ·<br />

W(R) = ∏ n≥1 R addition/multipl <strong>of</strong> ghost coords<br />

Truncated Witt schemes with Z[u 1 , . . . , u N ] ⇒<br />

gadgets W (N)<br />

Z<br />

over F 1 (Manin)<br />

Also Λ-ring structure given by the ψ n<br />

Borger’s suggestion: Witt schemes related to<br />

Spec(Z) × F1 · · · × F1 Spec(Z)<br />

⇒ Realize inside multivariable BC endomotives<br />

23


Other aspects <strong>of</strong> F 1 -geometry:<br />

Analytic geometry<br />

Yu.I. Manin, Cyclotomy and analytic geometry over F 1 ,<br />

arXiv:0809.1564.<br />

The Habiro ring<br />

K. Habiro, Cyclotomic completions <strong>of</strong> polynomial rings,<br />

Publ. RIMS Kyoto Univ. (2004) Vol.40, 1127–1146.<br />

Ẑ[q] = lim ←−n<br />

Z[q]/((q) n )<br />

(q) n = (1 − q)(1 − q 2 ) · · ·(1 − q n )<br />

Z[q]/((q) n ) ։ Z[q]/((q) k ) for k ≤ n since (q) k |(q) n<br />

Evaluation maps at roots <strong>of</strong> 1: surj ring homom<br />

ev ζ :<br />

Ẑ[q] → Z[ζ]<br />

give an injective homomorphism:<br />

ev : Ẑ[q] → ∏<br />

Z[ζ]<br />

Taylor series expansions<br />

ζ∈Z<br />

τ ζ :<br />

Ẑ[q] → Z[ζ][[q − ζ]]<br />

injective ring homomorphism<br />

24


Ring <strong>of</strong> “analytic functions on roots <strong>of</strong> unity”<br />

⇒ Another model for the NCG <strong>of</strong> the cyclotomic<br />

tower, replacing Q[Q/Z] with Ẑ[q]<br />

Multivariable Habiro rings (Manin)<br />

̂ Z[q 1 , . . . , q n ] = lim ←−N<br />

Z[q 1 , . . . , q n ]/I n,N<br />

where I n,N is the ideal<br />

( (q1 − 1)(q 2 1 − 1) · · ·(qN 1 − 1), . . . ,(q n − 1)(q 2 n − 1) · · ·(qN n − 1))<br />

Evaluations at roots <strong>of</strong> 1<br />

ev (ζ1 ,...,ζ n ) :<br />

Taylor expansions<br />

T Z :<br />

̂ Z[q1 , . . . , q n ] → Z[ζ 1 , . . . , ζ n ]<br />

̂ Z[q1 , . . . , q n ] → Z[ζ 1 , . . . , ζ n ][[q 1 −ζ 1 , . . . , q n −ζ n ]]<br />

Z = (ζ 1 , . . . , ζ n ) in Z n 25


Endomotives: Habiro ring version<br />

M.M., Cyclotomy and endomotives, preprint arXiv:0901.3167<br />

One variable:<br />

σ n (f)(q) = f(q n )<br />

lifts P(ζ) ↦→ P(ζ n ) in Z[ζ] through ev ζ<br />

Action <strong>of</strong> N ⇒ group crossed product<br />

A Z,q = Ẑ[q] ∞ ⋊ Q ∗ +<br />

A Z,q = generated by Ẑ[q] and µ n, µ ∗ n<br />

µ n σ n (f) = fµ n , µ ∗ nf = σ n (f)µ ∗ n<br />

Ẑ[q] ∞ = ∪ N A N with A N gen by µ N fµ ∗ N<br />

Ẑ[q] ∞ = lim −→n<br />

(σ n :<br />

Ẑ[q] → Ẑ[q])<br />

injective ⇒ automorphisms σ n : Ẑ[q] ∞ → Ẑ[q] ∞<br />

26


P Z = polyn in Q powers q r<br />

ˆP Z = lim ←−N<br />

P Z /J N<br />

J N = ideal gen by (q r ) N = (1 − q r ) · · ·(1 − q rN ), r ∈ Q ∗ +<br />

ρ r (f)(q) = f(q r )<br />

Multivariable:<br />

Ẑ[q] ∞ ≃ ˆP Z , µ n fµ ∗ n ↦→ f(q 1/n )<br />

̂ Z[q 1 , . . . , q n ] = lim ←−N<br />

Z[q 1 , . . . , q n , q −1<br />

1 , . . . , q−1 n ]/J n,N<br />

J n,N ideal generated by the (q i − 1) · · ·(q N i<br />

− 1), for i =<br />

1, . . . , n and the (q −1<br />

i<br />

− 1) · · ·(q −N<br />

i<br />

− 1)<br />

Torus T n = (G m ) n , algebra Q[t i , t −1<br />

i<br />

]<br />

t α = (t α i ) i=1,...,n with t α i = ∏ j<br />

t α ij<br />

j<br />

Semigroup action α ∈ M n (Z) + :<br />

q ↦→ σ α (q) = σ α (q 1 , . . . , q n ) =<br />

(q α 11<br />

1 qα 12<br />

2 · · · qα 1n<br />

n , . . . , qα n1<br />

1 qα n2<br />

2 · · · qα nn<br />

n ) = qα 27


Question: 3-manifolds and F 1 -geometry?<br />

Universal Witten–Reshetikhin–Turaev invariant<br />

K.Habiro, A unified Witten–Reshetikhin–Turaev invariant<br />

for integral homology spheres, Invent. Math. 171<br />

(2008) 1–81<br />

- Chern–Simons path integral (Witten)<br />

- quantum groups at roots <strong>of</strong> 1 (Reshetikhin–Turaev)<br />

- Ohtsuki series<br />

τ(M) : Z → C,<br />

τ O (M) = 1 +<br />

τ ζ (M)<br />

∞∑<br />

λ n (M)(q − 1) n<br />

n=1<br />

Unified view (Habiro): J M (q) = J L (q)<br />

J M (q) ∈ Ẑ[q]<br />

function in the (one-variable) Habiro ring:<br />

ev ζ (J M (q)) = τ ζ (M)<br />

τ 1 (J M (q)) = τ O (M)<br />

Using: 3-dimensional integral homology sphere M<br />

surgery presentation M = SL 3 , algebraically split link<br />

L = L 1 ∪ · · · ∪ L l in S 3 framing ±1<br />

SL 3 = ∼ SL 3 ⇔ L ∼ ′ L′ Fenn–Rourke moves<br />

28


Integral homology 3-spheres<br />

Zhs = free ab group generated by orientationpreserving<br />

homeomorphism classes <strong>of</strong> integral<br />

homology 3-spheres<br />

Ring with product M 1 #M 2 connected sum<br />

J M1 #M 2<br />

(q) = J M1 (q)J M2 (q), J S 3(q) = 1<br />

J −M (q) = J M (q −1 )<br />

⇒ WRT ring homomorphism<br />

Ohtsuki filtration<br />

J : Zhs → Ẑ[q]<br />

Zhs = F 0 ⊃ F 1 ⊃ · · · F k ⊃ · · ·<br />

F k Z-submodule spanned by<br />

[M, L 1 , . . . , L k ] =<br />

∑<br />

L ′ ⊂{L 1 ,...,L k }<br />

L i = alg split links ±1-framed<br />

Habiro conjecture<br />

J :<br />

Ẑhs → Ẑ[q]<br />

(−1) |L′ | ML ′<br />

Ẑhs = lim ←−d<br />

Zhs/F d with d : N → Z ≥0<br />

d(n)-components <strong>of</strong> link have framing ±n<br />

29


Integral homology spheres and F 1 ?<br />

X Zhs (R) := {φ : Zhs → R | ∃˜φ : Ẑ[q] → R, φ = ˜φ◦J}<br />

J : Zhs → Ẑ[q] WRT invariant;<br />

X Zhs (R) = set <strong>of</strong> “coarser” R-valued invariants<br />

A X := Zhs⊗C WRT evaluations as cyclotomic<br />

points: σ ◦ φ : Zhs → C factors through some<br />

evaluation ev ζ<br />

⇒ X Zhs gadget over F 1<br />

Question: Using Habiro conjecture XẐhs inductive<br />

limit <strong>of</strong> affine varieties over F 1 ? (finiteness)<br />

30


Question: 3-manifolds and endomotives?<br />

Semigroup action? Is it possible to construct<br />

M = S(L,m) 3 ↦→ σ n(M) = S(L 3 n ,m n )<br />

(inv. under Fenn–Rourke moves)<br />

σ n (M 1 #M 2 ) = σ n (M 1 )σ n (M 2 )<br />

Question: ∃ ? Υ M ∈ Ẑ[q]<br />

Υ σn (M) = σ n(Υ M )<br />

Is it possible to use endomotives (eg multivariable<br />

BC) to construct 3-manifold invariants ?<br />

Zhs ⋊ N → A = A ⋊ N<br />

Quantum statistical mechanics?<br />

31


Categorical approach:<br />

- Reidemeister moves<br />

(link diagrams: same ambient isotopy class)<br />

- Hoste moves (admissible links)<br />

(same Zhs: trivial linking #s and frames 1/m)<br />

component 1/m-framing replaced by −m full<br />

twists<br />

1/m<br />

−m full twists<br />

32


2-category:<br />

- Objects: (D, r) admissible framed link diagrams<br />

- 1-morphisms: (D, r) H → (D ′ , r ′ ) Hoste moves<br />

- 2-morphisms: R = (R, R ′ ) Reidemeister moves<br />

(D 1 , r 1 ) H 1 <br />

(D<br />

1 ′ , r′ 1 )<br />

R<br />

R ′<br />

(D 2 , r 2 ) H 2 <br />

(D ′ 2 , r′ 2 )<br />

Convolution algebra: A CL (2) functions with finite<br />

support on 2-Mor<br />

- vertical convolution product<br />

(f 1 ◦ f 2 )(R) = ∑<br />

R=R 1 ◦R 2<br />

f 1 (R 1 )f 2 (R 2 )<br />

- horizontal convolution product<br />

(f 1 • f 2 )(R) = ∑<br />

R=R 1 •R 2<br />

f 1 (R 1 )f 2 (R 2 )<br />

33


For compositions <strong>of</strong> 2-morphisms:<br />

- horizontal R 1 • R 2 with R 1 = (R, R ′ ) and R 2 = (R ′ , R ′′ )<br />

(D 1 , r 1 ) H 1 <br />

(D<br />

1 ′ , r′ 1 )<br />

R<br />

R ′<br />

H ′ 1 <br />

(D ′′<br />

1 , r′′ 1 )<br />

R ′′<br />

(D 2 , r 2 ) H 2 <br />

(D<br />

2 ′ , r′ 2 ) H′ 2 <br />

(D<br />

2 ′′,<br />

r′′ 2 )<br />

- vertical R 1 ◦ R 2 with R 1 = (R 1 , R ′ 1 ) and R 2 = (R 2 , R ′ 2 )<br />

(D 1 , r 1 ) H 1 <br />

(D<br />

1 ′ , r′ 1 )<br />

R 1<br />

R ′ 1<br />

(D 2 , r 2 ) H 2 <br />

(D 2 ′ , r′ 2 )<br />

R 2<br />

R ′ 2<br />

(D 3 , r 3 ) H 3 <br />

(D ′ 3 , r′ 3 ) 34


Involution: f ∗ (R) = f(R −1 )<br />

(D ′ 2 , r′ 2 ) H −1<br />

2 <br />

R ′−1<br />

(D 2 , r 2 )<br />

R −1<br />

(D<br />

1 ′ , r′ 1 ) H −1<br />

1<br />

(D 1 , r 1 )<br />

- Horizontal time evolution: σ f (f 1 ◦ f 2 ) = σ t (f 1 ) ◦<br />

σ t (f 2 )<br />

σ t (f)(H) = e imt f(H), with m = l − l ′<br />

Hoste move between L = L 1 ∪· · ·∪L l and L ′ = L ′ 1 ∪· · ·∪L′ l ′<br />

- Vertical time evolution: σ f (f 1 •f 2 ) = σ t (f 1 )•σ t (f 2 )<br />

self-linking number with respect to the blackboard<br />

framing <strong>of</strong> the link diagram<br />

(changed by Reidemeister moves)<br />

35


Variant: only oriented Hoste moves (semigroupoid)<br />

Cabling: L n = c n (L) with n = (n, . . . , n)<br />

(L, r) admissible and n|m i : if oriented Hoste<br />

move<br />

(L, r) H → (L ′ , r ′ )<br />

then n|m i also for L ′ and cabled oriented Hoste<br />

move<br />

(L n , nr) H n<br />

→ (L ′ n , nr′ )<br />

Action:<br />

ρ n (f)(H) =<br />

{<br />

f(Hn ) n|m i , ∀i,<br />

0 otherwise,<br />

endomorphisms <strong>of</strong> A CL (1)<br />

(after collapsing 2-morphisms to equiv rel)<br />

36


Similar to case <strong>of</strong><br />

M.M., A.Zainy al-Yasry, Coverings, correspondences, and<br />

noncommutative geometry, J.Geom.Phys. Vol.58 (2008)<br />

N.12, 1639–1661<br />

2-category with Obj=embedded graphs,<br />

1-Mor=3-manifolds as branched coverings,<br />

2-Mor=4-dim cobordisms<br />

Horizontal and vertical time evolutions<br />

(multiplicity <strong>of</strong> coverings, index theorems, Hartle–Hawking<br />

gravity)<br />

Further questions: Relation <strong>of</strong> the Habiro<br />

rings and (semi)group actions to Zagier’s quantum<br />

modular forms ?<br />

37

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!