Contents

1 Tachyons by Facu 2
1.1 Path integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Character of SO(1, 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1
Chapter 1

Tachyons by Facu

1.1 Path integral

We will be interested in calculating the partition function in EdS2
Z
Z = Dϕ e−SE [ϕ] (1.1.1)

√
Z
SE [ϕ] = d2 x gϕ[−2 + m2 ]ϕ (1.1.2)

We decompose ϕ in terms of harmonical spherics

∞ X
X m=l ∞ Y
Y l
ϕ= clm Ylm (Ω) =⇒ Dϕ = dclm (1.1.3)
l=0 m=−l l=0 m=−l

Which have the following properties

√
Z
2Ylm = −l(l + 1)Ylm , d2 x g Ylm Yl′ m′ = δll′ δmm′ (1.1.4)
S2

Puttign this in the Euclidean action we get

√ XX
Z
1
SE = d2 x g clm Ylm [l′ (l′ + 1) + m2 ]cl′ m′ Yl′ m′ (1.1.5)
2 lm l′ m′
√
Z
1 X X
= d2 x g Ylm Yl′ m′ clm cl′ m′ [l′ (l′ + 1) + m2 ] (1.1.6)
2 lm l′ m′
1 XX
= δll′ δmm′ clm cl′ m′ [l′ (l′ + 1) + m2 ] (1.1.7)
2 lm l′ m′
1
= c2lm [l(l + 1) + m2 ] (1.1.8)
2
Z Y
2 2
=⇒ Z = clm e−clm [l(l+1)+m ] (1.1.9)
lm

2
In [1] a tachyonic scalar field is defined by the following mass

m2 = −n(n + D − 1), n = 0, 1, 2, 3, ... (1.1.10)

Where D denotes the dimension of de Sitter space-time dSD . For now we will be interested
in two dimensions so
m2 = −n(n + 1), n = 0, 1, 2, 3, ... (1.1.11)
Reeplacing this mass in the partition function we get
Z Y Z Y
−c2lm [l(l+1)−n(n+1)] 2
Z= clm e = clm eclm [n(n+1)−l(l+1)] (1.1.12)
lm lm

From this we can deduce the following

l > n =⇒ Well defined Z. These are physical states. (1.1.13)

l < n =⇒ Ill defined Z. (1.1.14)

In this case we have a non-decaying exponential. Nevertheless, the contributions of these
states can be computing by Polchinski regularization (citar).
Z Y
l = n =⇒ Z = = V ol(G) (1.1.15)
lm

We may call these gauge states (donde le dicen ası́ en la literatura?). With this in mind we
propose to calculate the character (explicar bien que es esto). For the character only the
physical states will contribute. For example, take l = 0. We need to take n = 1 and so
on. A state with n = 0 will be ”gauge”. Take l = 1. A ”non physical polcho” state comes
from n = 0. A gauge state from n = 1. Take l = 2. Gauge state at n = 2, polcho states at
n = 0, n = 1.

1.2 Character of SO(1, 2)

(A)dS2 is defined as the hyperboloid

(X + )2 + (X − )2 − (X 1 )2 = R2

It is preserved under transformations Λ ∈ SO(2, 1). The equation that defines these trans-
formations is:
Λµ ρ ηµν Λν σ = ηρσ , η = diag(−1, −1, 1) (1.2.1) definin lam

Below, we will find the generators of the algebra so(2, 1). We define the elements of the
group near the identity as Λ = eϵJ . To find the generators, we expand infinitesimally close
to the identity.
Λµ ρ ≈ δ µ ρ + ϵJ µ ρ

3
We substitute this transformation into (1.2.1). Working with O(ϵ) we have

(δ µ ρ + ϵJ µ ρ )ηµν (δ ν σ + ϵJ ν σ ) = ηρσ

 + ϵJ µ ρ ηµσ
ηρσ + ϵJ ν σ ηρν ϵ2 J µ ρ ηµν J ν σ = 
ηρσ

| {z }
O(ϵ2 )≈0

J µ ρ ηµσ + J ν σ ηρν = 0
or in matrix notation
ηJ = −J T η (1.2.2) def J
reeplacing
 
A B C
J = D E F 
G H I
we get

−1 0 0 −1 0 0
     
A B C A D G
 0 −1 0   D E F  = −  B E H   0 −1 0 
0 0 1 G H I C F I 0 0 1

−A −B −C A D −G
   
 −D −E −F  =  B E −H 
G H I C F −I
=⇒ A = 0, E = 0, I = 0, y H = F, C = G, D = −B.
Then, a general matrix J ∈ so(2, 1) can be written as
 
0 B C
J =  −B 0 F 
C F 0
     
0 1 0 0 0 1 0 0 0
=B  −1 0 0  +C  0 0 0 +F  0 0 1 
0 0 0 1 0 0 0 1 0
= aµ Jµ (1.2.3)

Thus, a possible basis for so(2, 1) consists of the following generators

     
0 1 0 0 0 0 0 0 1
J0 =  −1 0 0  , J1 =  0 0 1  , J2 =  0 0 0 
0 0 0 0 1 0 1 0 0

Where we find the algebra

[J0 , J1 ] = J2 , [J0 , J2 ] = −J1 , [J1 , J2 ] = −J0 (1.2.4) lie alg

4
By exponentiating the generators, we obtain the following group elements.

cos(θ) − sin(θ) 0
 

eθJ0 =  sin(θ) cos(θ) 0 

0 0 1
 
1 0 0
eα1 J1 =  0 cosh(α1 ) − sinh(α1 ) 
0 − sinh(α1 ) cosh(α1 )
cosh(α2 ) 0 − sinh(α2 )
 

eα2 J2 = 0 1 0 
− sinh(α2 ) 0 cosh(α2 )
From g(θ) = eθJ0 it is clear that g(0) = g(2π) = I. Then, θ ∈ [0, 2π]. The compact nature
of the generator J0 : is then recognized: starting from the group element g(0) I arrive at the
same element with g(2π). The case of the generators J1 , J2 is different; the group element
does not recur, so we see that α1,2 ∈ R. We conclude that the topology of the group SO(2, 1)
is R2 × S 1 . The casimir is
C = J02 − J12 − J22 (1.2.5)
importante: el casimir que toma vasilio es menos el anterior. Tomemos

C = −J02 + J12 + J22 (1.2.6)

We can indeed check that [C, Ji ] = 0.

Ladders: We define
J ± = J2 ± iJ1 (1.2.7)
=⇒ [J0 , J ± ] = ±iJ ± , (J + )∗ = −J − , [J + , J − ] = −2iJ0 (1.2.8)
1
C = −J02 + (J + J − + J − J + ) (1.2.9)
2
i 1
J1 = − (J + − J − ), J2 = (J + + J − ) (1.2.10) J1J2
2 2
Since [C, Jµ ] = 0 we will work with simultaneous eigenvectors of C y J0 . It will be useful
in what follows to rewrite the eigenvalue of the Casimir as

λ = ∆(∆ − 1)

then
C |∆, m⟩ = ∆(∆ − 1) |∆, m⟩ , J0 |∆, m⟩ = im |∆, m⟩ . (1.2.11)
En cuanto a m, deberá toma valores enteros para el grupo es SO(2, 1) o semi enteros para
SL(2, R). Los operadores J ± aumentan/bajan el autovalor de J0 en una unidad al actuar

5
sobre autofuncioines de J0

J0 (J ± |∆, m⟩) =(J ± J0 ± iJ ± ) |∆, m⟩

=J ± (J0 ± i) |∆, m⟩
=J ± (im ± i) |∆, m⟩
=i(m ± 1)(J ± |∆, m⟩).

Esta última ecuación nos dice que (J ± |∆, m⟩) es autovector de J0 con autovalor i(m ± 1).
Escribiremos
J ± |∆, m⟩ = cm |∆, m ± 1⟩
donde los kets |∆, m⟩ se asumen normalizados.
Teniendo en cuenta que
J − J + = J0 (J0 + i) + C (1.2.12)
J + J − = J0 (J0 − i) + C (1.2.13)
se concluye que
2
J + |∆, m⟩ = − ⟨∆, m| J − J + |∆, m⟩ (1.2.14)
= − ⟨∆, m| J02 + iJ0 + C |∆, m⟩ (1.2.15)
=m(m + 1) − ∆(∆ − 1) (1.2.16)

We conclude that
p
J + |∆, m⟩ = m(m + 1) − ∆(∆ − 1) |∆, m + 1⟩ (1.2.17)

Similarly
2
J − |∆, m⟩ = − ⟨∆, m| J + J − |∆, m⟩ (1.2.18)
= − ⟨∆, m| J02 − iJ0 + C |∆, m⟩ (1.2.19)
=m(m − 1) − ∆(∆ − 1) (1.2.20)

We conclude that
p
J − |∆, m⟩ = m(m − 1) − ∆(∆ − 1) |∆, m − 1⟩ (1.2.21)

Now we can write the non-compact generators in term of the ladders operators using 1.2.10
i
J1 |∆, m⟩ = − (J + − J − ) |∆, m⟩ (1.2.22)
2
i hp p i
J1 |∆, m⟩ = − m(m + 1) − ∆(∆ − 1) |∆, m + 1⟩ − m(m − 1) − ∆(∆ − 1) |∆, m − 1⟩
2
(1.2.23) J1

6
We aim to compute the character for a non-compact (hyperbolic/spacelike) direction

χ∆ (x) = tr eJ2 x (1.2.24)

To compute this character we start with

fm′ ,m (x) = ⟨m′ | exJ2 |m⟩ (1.2.25)

The idea is to compute this function by finding a differential equation for it. The trick is to
insert J0 inside the bracket and manipulate. The next equation holds

e−xJ2 J0 exJ2 = cosh x J0 − sinh x J1 (1.2.26) trick

We do the following trick

⟨m′ | J0 exJ2 |m⟩ = − im′ ⟨m′ | exJ2 |m⟩ (1.2.27)

′ xJ2
= ⟨m | e (cosh x J0 − sinh x J1 ) |m⟩ (1.2.28)
′ xJ2 ′ xJ2
=im cosh x ⟨m | e |m⟩ − sinh x ⟨m | e J1 |m⟩ (1.2.29)

From the first and third line we arrive at the following equation
(m′ + m cosh x)
i ⟨m′ | exJ2 |m⟩ = ⟨m′ | exJ2 J1 |m⟩ (1.2.30)
sinh x
Now we act with 1.2.23 on the rhs of the equation. We arrive at
m cosh x + m′ 1 p p 
fm′ ,m (x) = m(m − 1) − ∆(∆ − 1)fm′ ,m−1 (x) − m(m + 1) − ∆(∆ − 1)fm′ ,m+1 (x)
sinh x 2
(1.2.31) rec
Now taking a derivative
d d
fm′ ,m (x) = ⟨m′ | exJ2 |m⟩ = ⟨m′ | exJ2 J2 |m⟩ (1.2.32)
dx dx
Now using 1.2.10
d 1
fm′ ,m (x) = ⟨m′ | exJ2 (J + + J − ) |m⟩ (1.2.33)
dx 2
1 p p 
= m(m + 1) − ∆(∆ − 1)fm′ ,m+1 (x) + m(m − 1) − ∆(∆ − 1)fm′ ,m−1 (x)
2
(1.2.34)

We can use this equation to reeplace the term fm′ ,m−1 (x) in 1.2.31. We get the following
d (m cosh x + m′ ) p
fm′ ,m (x) = fm′ ,m (x) + m(m + 1) − ∆(∆ − 1)fm′ ,m+1 (x) (1.2.35)
dx sinh x
Or we can reeplace the term fm′ ,m+1 (x) in 1.2.31 . We get
d (m cosh x + m′ ) p
fm′ ,m (x) = − fm′ ,m (x) + m(m − 1) − ∆(∆ − 1)fm′ ,m−1 (x) (1.2.36)
dx sinh x
7
Now we define the following differential operator

(m cosh x + m′ )
 
± d p
L fm′ ,m (x) = ∓ fm′ ,m (x) = m(m ± 1) − ∆(∆ − 1)fm′ ,m±1 (x)
dx sinh x
(1.2.37)
p
L+ L− fm′ ,m (x) = m(m − 1) − ∆(∆ − 1)L+ fm′ ,m−1 (x) (1.2.38)
p p
= m(m − 1) − ∆(∆ − 1) m(m − 1) − ∆(∆ − 1)fm′ ,m (x) (1.2.39)
=(m(m − 1) − ∆(∆ − 1))fm′ ,m (x) (1.2.40)

p
L− L+ fm′ ,m (x) = m(m + 1) − ∆(∆ − 1)L− fm′ ,m+1 (x) (1.2.41)
p p
= m(m + 1) − ∆(∆ − 1) m(m + 1) − ∆(∆ − 1)fm′ ,m (x) (1.2.42)
=(m(m + 1) − ∆(∆ − 1))fm′ ,m (x) (1.2.43)

Now acting on the lhs

m cosh x + m′
 
+ − + d
L L fm′ ,m (x) =L + fm′ ,m (x) (1.2.44)
dx sinh x
(m − 1) cosh x + m′ m cosh x + m′
  
d d
= − + fm′ ,m (x) (1.2.45)
dx sinh x dx sinh x

We also do
m cosh x + m′
 
− + − d
L L fm′ ,m (x) =L − fm′ ,m (x) (1.2.46)
dx sinh x
(m + 1) cosh x + m′ m cosh x + m′
  
d d
= + − fm′ ,m (x) (1.2.47)
dx sinh x dx sinh x

And we add the two expressions.

 
′ ′′ ′ 2
−2 − coth(x)f (x) − f (x) + f (x) (m csch(x) + m coth(x)) = 2(m2 − ∆(∆ − 1))f
(1.2.48)
′
We put m = m and we get
  x 
coth(x)f ′ (x) + f ′′ (x) − f (x) −∆2 + ∆ + m2 coth2 + m2 =0 (1.2.49)
2
 x 
f ′′ (x) + coth x f ′ (x) + ∆(∆ − 1) − m2 (1 + coth2 ( ) f (x) = 0 (1.2.50)
2
We use
cosh x + 1 cosh x − 1
cosh2 (x/2) = , sinh2 (x/2) = (1.2.51)
2 2
1 + cosh x 2 cosh x
=⇒ coth2 (x/2) = =⇒ 1 + coth2 (x/2) = (1.2.52)
cosh x − 1 cosh x − 1

8
  
′′ ′ 2 cosh x
f (x) + coth x f (x) + ∆(∆ − 1) − 2m f (x) = 0 (1.2.53)
cosh x − 1
Same expression guille notes, different than vasilio’s notes

f (x) = fm,m (x) = ⟨m| exJ2 |m⟩ (1.2.54)

We need to solve this differential equation, and once we do that, we obtain the character as
the sum of all posible m’s.

9
 Bibliography

epstein [1] H. Epstein and U. Moschella, de Sitter tachyons and related topics, Commun. Math.
Phys. 336 (2015) 381 [1403.3319].

dio [2] D. Anninos, T. Anous, B. Pethybridge and G. S¸eng¨or, The discreet charm of the
discrete series in dS2, J. Phys. A 57 (2024) 025401 [2307.15832].

rosen [3] J. Bonifacio, K. Hinterbichler, A. Joyce and R.A. Rosen, Shift Symmetries in (Anti) de
Sitter Space, JHEP 02 (2019) 178 [1812.08167].

10