Indecomposability parameters in LCFT

Transcription

1 Indecomposability parameters in LCFT Romain Vasseur Joint work with J.L. Jacobsen and H. Saleur at IPhT CEA Saclay and LPTENS (Nucl. Phys. B 851, (2011), arxiv : ) ACFTA (Institut Henri Poincaré)

2 1 Indecomposability parameters β r,s : a few definitions 2 A simple general formula from OPEs 3 Lattice models, indecomposability and scaling limit 4 Measure of indecomposability parameters from lattice models 5 Conclusion

3 Indecomposability parameters β r,s : a few definitions

4 Definition Original claim [Gurarie 93 99, Gurarie & Ludwig 02] : SUSY approach to disordered systems yields c = 0 CFT with a logarithmic partner for the stress energy tensor T

5 Definition Original claim [Gurarie 93 99, Gurarie & Ludwig 02] : SUSY approach to disordered systems yields c = 0 CFT with a logarithmic partner for the stress energy tensor T Chiral( case only. ) In the basis (T, t), L 0 has a Jordan cell 2 1 L 0 = 0 2

6 Definition Original claim [Gurarie 93 99, Gurarie & Ludwig 02] : SUSY approach to disordered systems yields c = 0 CFT with a logarithmic partner for the stress energy tensor T Chiral( case only. ) In the basis (T, t), L 0 has a Jordan cell 2 1 L 0 = 0 2 Global conformal invariance then implies T(z)T(0) = 0 T(z)t(0) = b z 4 t(z)t(0) = θ 2b logz z 4,

7 Definition Original claim [Gurarie 93 99, Gurarie & Ludwig 02] : SUSY approach to disordered systems yields c = 0 CFT with a logarithmic partner for the stress energy tensor T Chiral( case only. ) In the basis (T, t), L 0 has a Jordan cell 2 1 L 0 = 0 2 Global conformal invariance then implies T(z)T(0) = 0 T(z)t(0) = b z 4 t(z)t(0) = θ 2b logz z 4, θ is irrelevant whereas b is a fundamental number that characterizes the pair (T, t)

8 Definition Original claim [Gurarie 93 99, Gurarie & Ludwig 02] : SUSY approach to disordered systems yields c = 0 CFT with a logarithmic partner for the stress energy tensor T Chiral( case only. ) In the basis (T, t), L 0 has a Jordan cell 2 1 L 0 = 0 2 Global conformal invariance then implies T(z)T(0) = 0 T(z)t(0) = b z 4 t(z)t(0) = θ 2b logz z 4, θ is irrelevant whereas b is a fundamental number that characterizes the pair (T, t) At the time, b was thought of as an effective central charge

9 Definition LCFTs are actually characterized by a complicated structure of indecomposable Virasoro modules

10 Definition LCFTs are actually characterized by a complicated structure of indecomposable Virasoro modules Focus on rank 2 Jordan cells L 0 φ = hφ L 0 ψ = hψ +φ

11 Definition LCFTs are actually characterized by a complicated structure of indecomposable Virasoro modules Focus on rank 2 Jordan cells L 0 φ = hφ L 0 ψ = hψ +φ The associated correlation functions then read φ(z)φ(0) = 0 φ(z)ψ(0) = β ψ(z)ψ(0) = z 2h θ 2β logz z 2h

12 Definition LCFTs are actually characterized by a complicated structure of indecomposable Virasoro modules Focus on rank 2 Jordan cells L 0 φ = hφ L 0 ψ = hψ +φ The associated correlation functions then read φ(z)φ(0) = 0 φ(z)ψ(0) = β ψ(z)ψ(0) = z 2h θ 2β logz z 2h Using Virasoro bilinear form L n = L n, we get β = ψ φ

13 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture)

14 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture) φ must be singular with φ(z) = Aξ(z), A = L n +α (1) L n+1 L , n = h h ξ (PBW order).

15 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture) φ must be singular with φ(z) = Aξ(z), A = L n +α (1) L n+1 L , n = h h ξ (PBW order). e.g. φ = T, ψ = t, ξ = I and A = L 2

16 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture) φ must be singular with φ(z) = Aξ(z), A = L n +α (1) L n+1 L , n = h h ξ (PBW order). e.g. φ = T, ψ = t, ξ = I and A = L 2 Define β through A ψ(z) = βξ(z)

17 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture) φ must be singular with φ(z) = Aξ(z), A = L n +α (1) L n+1 L , n = h h ξ (PBW order). e.g. φ = T, ψ = t, ξ = I and A = L 2 Define β through A ψ(z) = βξ(z) β characterizes uniquely the staggered module

18 Staggered Virasoro modules and indecomposability parameters Crucial role in staggered Virasoro modules theory [ Rohsiepe 96, Kytola & Ridout 09 ] ψ P = ξ φ ρ Quotient of glueing of two Verma modules (cf. D. Ridout s lecture) φ must be singular with φ(z) = Aξ(z), A = L n +α (1) L n+1 L , n = h h ξ (PBW order). e.g. φ = T, ψ = t, ξ = I and A = L 2 Define β through A ψ(z) = βξ(z) β characterizes uniquely the staggered module Convenient to work with Virasoro bilinear form L n = L n β = ψ φ (given that ξ ξ = 1)

19 Questions we d like to answer How to compute these (hopefully physically relevant) numbers? It would also be nice to get a physical intuition of their meaning! Can we measure some of them on concrete lattice models (as for the central charge and the critical exponents)? Which observables are they related to?

20 A simple general formula from OPEs : log CFTs as limits of ordinary CFTs

21 Motivations Indecomposability parameters are most likely relevant for physics ( logarithmic structure constants). How can we calculate these numbers?

22 Motivations Indecomposability parameters are most likely relevant for physics ( logarithmic structure constants). How can we calculate these numbers? There exist many methods to compute indecomposability parameters (Fusion algorithm, differential equations, null vector equations, SLE...).

23 Motivations Indecomposability parameters are most likely relevant for physics ( logarithmic structure constants). How can we calculate these numbers? There exist many methods to compute indecomposability parameters (Fusion algorithm, differential equations, null vector equations, SLE...). Unfortunately, these methods may be quite cumbersome...

24 Motivations Indecomposability parameters are most likely relevant for physics ( logarithmic structure constants). How can we calculate these numbers? There exist many methods to compute indecomposability parameters (Fusion algorithm, differential equations, null vector equations, SLE...). Unfortunately, these methods may be quite cumbersome... Need of a simple explicit formula for the values of β

25 Motivations Indecomposability parameters are most likely relevant for physics ( logarithmic structure constants). How can we calculate these numbers? There exist many methods to compute indecomposability parameters (Fusion algorithm, differential equations, null vector equations, SLE...). Unfortunately, these methods may be quite cumbersome... Need of a simple explicit formula for the values of β Idea : LCFT = limit of usual (non log) CFTs c = 1 6 x(x + 1) h r,s = [r(x + 1) sx]2 1 4x(x + 1)

26 c 0 catastrophe c 0 catastrophe and b number [Cardy 01, Gurarie & Ludwig 02,...] Conformal invariance enforces Φ h (z)φ h (0) CI ΦΦ z 2h [ 1+ 2h ] c z2 T(0) Suppose there is another field with conformal weight h 2 as c 0 (φ 1,5 for percolation and φ 3,1 for polymers) Φ h (z)φ h (0) CI ΦΦ z 2h Then introduce a new field t(z) through [ 1+ 2h c z2 T(0)+... ]+ CΦ15 ΦΦ z [Φ 1,5(0)+...]. 2h h1,5 ( Φ 1,5 (z) = CI ΦΦ 2h T T t(z) 2h ) C Φ15 cb(c) c T(z) ΦΦ Where b(c) = T T h 1,5 2 and T T = c 2.

27 c 0 catastrophe The OPE then becomes well defined at c = 0 Φ h (z)φ h (0) CI ΦΦ z 2h [ 1+ h ] b z2 (T(0)log z + t(0)) , One can show that L 0 t = 2t + T at c = 0, correlation functions can also be calculated in a similar fashion. This yields b percolation b polymers c/2 = lim c 0 h 1,5 2 = 5 8 c/2 = lim c 0 h 3,1 2 = 5 6

28 Generalization to a more complicated case One can get rid of similar ill-defined terms in more complicated cases with the very same argument

29 Generalization to a more complicated case One can get rid of similar ill-defined terms in more complicated cases with the very same argument Logarithms come out naturally in this approach, and this yields explicit formulas for β

30 Generalization to a more complicated case One can get rid of similar ill-defined terms in more complicated cases with the very same argument Logarithms come out naturally in this approach, and this yields explicit formulas for β For a LCFT with c = 1 6 x 0(x 0+1). Let x = x 0 +ε, n = (h ψ h ξ ) ε=0 d φ φ β = lim ε 0 h ψ h ξ n = dε φ φ ε=0 d dε (h ψ h ξ ) ε=0

31 Generalization to a more complicated case One can get rid of similar ill-defined terms in more complicated cases with the very same argument Logarithms come out naturally in this approach, and this yields explicit formulas for β For a LCFT with c = 1 6 x 0(x 0+1). Let x = x 0 +ε, n = (h ψ h ξ ) ε=0 This is a 0 0 limit d φ φ β = lim ε 0 h ψ h ξ n = dε φ φ ε=0 d dε (h ψ h ξ ) ε=0

32 Generalization to a more complicated case One can get rid of similar ill-defined terms in more complicated cases with the very same argument Logarithms come out naturally in this approach, and this yields explicit formulas for β For a LCFT with c = 1 6 x 0(x 0+1). Let x = x 0 +ε, n = (h ψ h ξ ) ε=0 This is a 0 0 limit d φ φ β = lim ε 0 h ψ h ξ n = dε φ φ ε=0 d dε (h ψ h ξ ) ε=0 One needs to properly identify the conformal weights h ξ and h ψ in the spectrum (quite easy)

33 Example Application to Symplectic fermion theory [Kausch 95] General structure of the Jordan cells L 0 φ (j) = h 1,1+2j φ (j) L 0 ψ (j) = h 1,1+2j ψ (j) +φ (j) φ (j) = A j ξ (j) A j ψ(j) = β 1,1+2j ξ (j) associated with the staggered structure P j = ξ (j) ψ (j) ρ (j) φ (j) This OPE formula allows us to conjecture the general expression ξ A A ξ β 1,1+2j = lim x 1 h 1,1+2j h 1,1+2(j 1) (j 1) = [(2j 3)!]2 4 j 2 (j 1)

34 Lattice models, indecomposability and scaling limit

35 Temperley-Lieb algebra The Temperley-Lieb algebra is behind all the models we shall consider [e i, e j ] = 0 ( i j 2) e 2 i e i e i±1 e i = (q + q 1 )e i = e i

36 Temperley-Lieb algebra The Temperley-Lieb algebra is behind all the models we shall consider [e i, e j ] = 0 ( i j 2) e 2 i e i e i±1 e i = (q + q 1 )e i = e i Algebra of diagrams with e i =... i i+1....

37 Temperley-Lieb algebra The Temperley-Lieb algebra is behind all the models we shall consider [e i, e j ] = 0 ( i j 2) e 2 i e i e i±1 e i = (q + q 1 )e i = e i Algebra of diagrams with e i =... i i Standard (cell) modules V j : V 2 = { }, V 1 = {,, }, and V 0 = {, }.

38 Temperley-Lieb algebra The Temperley-Lieb algebra is behind all the models we shall consider [e i, e j ] = 0 ( i j 2) e 2 i e i e i±1 e i = (q + q 1 )e i = e i Algebra of diagrams with e i =... i i Standard (cell) modules V j : V 2 = { }, V 1 = {,, }, and V 0 = {, }. In a given module, we define our models by Transfer Matrix T = i even (1+e i) i odd (1+e i) Hamiltonian H = i e i

39 XXZ Spin chain 6-Vertex or XXZ representation H XXZ = (C 2 ) 2N Representation of the TL algebra on this space e i = I I 0 q q 0 I, The Hamiltonian reads H = 1 2 2N 1 i=1 (σ xi σ xi+1 +σ U q (sl 2 ) symmetry [H, U q (sl 2 )] = 0 yi q + q 1 σyi+1 + σi z σi+1 )+ z q q (σ z 1 σ z 2N)

40 Scaling limit TL Virasoro This XXZ spin chain at q = e iπ x+1 corresponds to a (L)CFT with central charge c = 1 6 x(x+1) in the scaling limit

41 Scaling limit TL Virasoro This XXZ spin chain at q = e iπ x+1 corresponds to a (L)CFT with central charge c = 1 6 x(x+1) in the scaling limit Excitations over the groundstate (vacuum) of the Hamiltonian correspond to critical exponents. In the scaling limit, we have the correspondence (TL) V j V h1,1+2j /V h1, 1 2j (Virasoro), which is generically irreducible

42 Scaling limit TL Virasoro This XXZ spin chain at q = e iπ x+1 corresponds to a (L)CFT with central charge c = 1 6 x(x+1) in the scaling limit Excitations over the groundstate (vacuum) of the Hamiltonian correspond to critical exponents. In the scaling limit, we have the correspondence (TL) V j V h1,1+2j /V h1, 1 2j (Virasoro), which is generically irreducible The standard TL module V j has a structure that mimics that of V h1,1+2j /V h1, 1 2j on the Virasoro side

43 Schur-Weyl duality and algebraic structure To understand the Jordan cell structure of the Hamiltonian H, it turns out to be convenient to study the representation theory of the symmetry algebra of our model.

44 Schur-Weyl duality and algebraic structure To understand the Jordan cell structure of the Hamiltonian H, it turns out to be convenient to study the representation theory of the symmetry algebra of our model. Let us focus on the XXZ case. When q is generic, we have the decomposition on 2N sites H Uqsl 2 = H TLq(2N) = N d j V j j=0 N (2j + 1)V j j=0

45 Schur-Weyl duality and algebraic structure To understand the Jordan cell structure of the Hamiltonian H, it turns out to be convenient to study the representation theory of the symmetry algebra of our model. Let us focus on the XXZ case. When q is generic, we have the decomposition on 2N sites H Uqsl 2 = H TLq(2N) = N d j V j j=0 N (2j + 1)V j j=0 For q root of unity, we can deduce which indecomposable TL q (2N) modules occur in the decomposition using the indecomposable structure of V 2N 1 for U q sl 2 (cf. lectures of A. Gaynutdinov, G. 2 Lehrer...)

46 Bimodule Vir(TL q ) 5 4 [Read & Saleur 07] U q sl 2

47 Scaling limit The indecomposable diamond TL modules tend to several staggered Virasoro modules in the limit χ j χ h1,1+2j χ j1 χ j2 = χ h1,1+2j1 χ h1,1+2j1 χ j χ h1,1+2j

48 Scaling limit The indecomposable diamond TL modules tend to several staggered Virasoro modules in the limit χ j χ h1,1+2j χ j1 χ j2 = χ h1,1+2j1 χ h1,1+2j1 χ j χ h1,1+2j In particular the lattice Jordan cells for H become L 0 logarithmic pairs

49 Scaling limit The indecomposable diamond TL modules tend to several staggered Virasoro modules in the limit χ j χ h1,1+2j χ j1 χ j2 = χ h1,1+2j1 χ h1,1+2j1 χ j χ h1,1+2j In particular the lattice Jordan cells for H become L 0 logarithmic pairs We know exactly from the representation theory of associative algebras the structure of the TL modules (and of the Jordan cells) on the lattice (finite size). The TL indecomposable modules mimic the staggered Virasoro modules that occur in the continuum limit. We can thus study the lattice models as regularizations of LCFTs built out of such staggered modules. Most of the continuum features can be recovered in finite size (subquotient structure, fusion rules,...) and we ll see that this is also true for indecomposability parameters.

50 Measure of indecomposability parameters from lattice models

51 From lattice Jordan cells to the continuum Finite size Jordan cells in the Hamiltonian become L 0 Jordan cells in the scaling limit.

52 From lattice Jordan cells to the continuum Finite size Jordan cells in the Hamiltonian become L 0 Jordan cells in the scaling limit. Identify Jordan cells in the spectrum E 0 E 1 E 2 1 H = E 2 E 3...

53 From lattice Jordan cells to the continuum Finite size Jordan cells in the Hamiltonian become L 0 Jordan cells in the scaling limit. Identify Jordan cells in the spectrum E 0 E 1 E 2 1 H = E 2 E 3... We normalize our states { φ (L), ψ (L) } to prepare the comparison with CFT H (L) E 0 (L)1 = πv ( ) F h (L) 1 L 0 h (L), with h (L) = L πv F (E(L) E 0 (L)).

54 Lattice scalar product We need somewhat to measure ψ (L) φ (L)...

55 Lattice scalar product We need somewhat to measure ψ (L) φ (L)......but we need a scalar product which goes to Virasoro bilinear form in the continuum limit

56 Lattice scalar product We need somewhat to measure ψ (L) φ (L)......but we need a scalar product which goes to Virasoro bilinear form in the continuum limit In our XXZ case, treat q as a formal parameter φ = + q has norm φ φ = 1+q 2.

57 Lattice scalar product We need somewhat to measure ψ (L) φ (L)......but we need a scalar product which goes to Virasoro bilinear form in the continuum limit In our XXZ case, treat q as a formal parameter φ = + q has norm φ φ = 1+q 2. H is hermitian for this scalar product : L 0 = L 0

58 Lattice scalar product We need somewhat to measure ψ (L) φ (L)......but we need a scalar product which goes to Virasoro bilinear form in the continuum limit In our XXZ case, treat q as a formal parameter φ = + q has norm φ φ = 1+q 2. H is hermitian for this scalar product : L 0 = L 0 Note that H is obviously not hermitian for the canonical inner product over C

59 Lattice scalar product We need somewhat to measure ψ (L) φ (L)......but we need a scalar product which goes to Virasoro bilinear form in the continuum limit In our XXZ case, treat q as a formal parameter φ = + q has norm φ φ = 1+q 2. H is hermitian for this scalar product : L 0 = L 0 Note that H is obviously not hermitian for the canonical inner product over C OK, so let s compute ψ (L) φ (L) then...

60 Virasoro algebra regularization on the lattice...but the Jordan cell is invariant under a global rescaling of the basis states φ (L) α φ (L) and ψ (L) α ψ (L) while ψ (L) φ (L) α 2 ψ (L) φ (L)!

61 Virasoro algebra regularization on the lattice...but the Jordan cell is invariant under a global rescaling of the basis states φ (L) α φ (L) and ψ (L) α ψ (L) while ψ (L) φ (L) α 2 ψ (L) φ (L)! We thus need to normalize things so that φ (L) corresponds precisely to φ = Aξ in the continuum limit.this is non trivial because φ φ = 0.

62 Virasoro algebra regularization on the lattice...but the Jordan cell is invariant under a global rescaling of the basis states φ (L) α φ (L) and ψ (L) α ψ (L) while ψ (L) φ (L) α 2 ψ (L) φ (L)! We thus need to normalize things so that φ (L) corresponds precisely to φ = Aξ in the continuum limit.this is non trivial because φ φ = 0. For c = 0 and the stress energy tensor, trousers trick [Dubail, Jacobsen, Saleur 10]

63 Virasoro algebra regularization on the lattice...but the Jordan cell is invariant under a global rescaling of the basis states φ (L) α φ (L) and ψ (L) α ψ (L) while ψ (L) φ (L) α 2 ψ (L) φ (L)! We thus need to normalize things so that φ (L) corresponds precisely to φ = Aξ in the continuum limit.this is non trivial because φ φ = 0. For c = 0 and the stress energy tensor, trousers trick [Dubail, Jacobsen, Saleur 10] In the general case, this is achieved using a Virasoro algebra regularization on the lattice [Koo & Saleur 93] L (2N) n =0 = L π [ 1 v F L 1 (e i e )cos i=1 ( ) niπ + 1 L 2 L vf 2 i=1 [e i, e i+1 ]sin ( ) ] niπ L

64 Virasoro algebra regularization on the lattice...but the Jordan cell is invariant under a global rescaling of the basis states φ (L) α φ (L) and ψ (L) α ψ (L) while ψ (L) φ (L) α 2 ψ (L) φ (L)! We thus need to normalize things so that φ (L) corresponds precisely to φ = Aξ in the continuum limit.this is non trivial because φ φ = 0. For c = 0 and the stress energy tensor, trousers trick [Dubail, Jacobsen, Saleur 10] In the general case, this is achieved using a Virasoro algebra regularization on the lattice [Koo & Saleur 93] L (2N) n =0 = L π [ 1 v F L 1 (e i e )cos i=1 ( ) niπ + 1 L 2 L vf 2 i=1 [e i, e i+1 ]sin We can construct A (L) on the lattice. We expect the state A (L) ξ (L) to correspond to the operator φ = Aξ in the limit L. ( ) ] niπ L

65 Lattice β (L) s Let s sum up : Using exact diagonalization methods, find a Jordan basis for the first few excitations of H on L = 2N sites.

66 Lattice β (L) s Let s sum up : Using exact diagonalization methods, find a Jordan basis for the first few excitations of H on L = 2N sites. Identify a Jordan cell in the spectrum of H and normalize the states to compare with CFT.

67 Lattice β (L) s Let s sum up : Using exact diagonalization methods, find a Jordan basis for the first few excitations of H on L = 2N sites. Identify a Jordan cell in the spectrum of H and normalize the states to compare with CFT. Also identify the state ξ (L) and normalize it such that ξ (L) ξ (L) = 1 for the lattice scalar product.

68 Lattice β (L) s Let s sum up : Using exact diagonalization methods, find a Jordan basis for the first few excitations of H on L = 2N sites. Identify a Jordan cell in the spectrum of H and normalize the states to compare with CFT. Also identify the state ξ (L) and normalize it such that ξ (L) ξ (L) = 1 for the lattice scalar product. Using Virasoro generators on the lattice, construct the operator A (L).

69 Lattice β (L) s Let s sum up : Using exact diagonalization methods, find a Jordan basis for the first few excitations of H on L = 2N sites. Identify a Jordan cell in the spectrum of H and normalize the states to compare with CFT. Also identify the state ξ (L) and normalize it such that ξ (L) ξ (L) = 1 for the lattice scalar product. Using Virasoro generators on the lattice, construct the operator A (L). Compute β (L) ψ (L) A (L) ξ (L) 2 = ψ (L) φ (L)

70 Some results Numerics for the Symplectic fermion theory L = 2N β 1,5 β 1,7 β 1, ± (0) ± (00) ± 25 Exact

71 Some results Some results (cont d) Percolation (q = e iπ/3 ) L β 1,4 L β 1, ± ± Exact -1/2 Exact -5/8 = 0.625

72 Some results A example of new result : Logarithmic Ising model (q = e iπ/4 ) ( ) 5/2 1 L 0 =, φ(z) = (L 0 5/ L2 1)ξ(z). L = 2N β 1, (8) ± Exact 35/ β 1,5 = lim ε 0 φ (2) φ (2) h 1,5 h 1,3 2 = 35 24

73 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory

74 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory and they should also be relevant for physics!

75 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory and they should also be relevant for physics! We found a simple physical way to understand how they are fixed in a given theory

76 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory and they should also be relevant for physics! We found a simple physical way to understand how they are fixed in a given theory Concrete lattice models mimic the indecomposability that occurs in the continuum, and indecomposability parameters can be measured directly on finite size systems Open questions :

77 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory and they should also be relevant for physics! We found a simple physical way to understand how they are fixed in a given theory Concrete lattice models mimic the indecomposability that occurs in the continuum, and indecomposability parameters can be measured directly on finite size systems Open questions : It would be very interesting to extend all these results to the bulk (Vir Vir)

78 Conclusion Indecomposability parameters are crucial in the Virasoro staggered modules theory and they should also be relevant for physics! We found a simple physical way to understand how they are fixed in a given theory Concrete lattice models mimic the indecomposability that occurs in the continuum, and indecomposability parameters can be measured directly on finite size systems Open questions : It would be very interesting to extend all these results to the bulk (Vir Vir) Geometrical/physical observables involving these numbers???

79 Conclusion Thank you!