Last month I asked about a 27-dimensional origin of the heterotic string. Now I'm looking at Witten's "Three-dimensional gravity revisited", where he proposes that pure gravity on AdS3 is dual to the "monstrous moonshine" CFT, which is bosonic string theory on a "Leech lattice" orbifold. Wouldn't that mean we were also talking about bosonic M-theory (the 27-dimensional theory) on AdS3 x R2424? But that would mean that on this occasion, the emergence of the "M" dimension was the same thing as the emergence of the radial AdS dimension. Does that even make sense?


Mitchell, if you search for "Josh Grey" at


you will see a comment that we had a related idea about the AdS compactifications of the bosonic string/M-theory in 2000. The advantage of having the bosonic string itself is that the one-loop beta function of $AdS_5\times S^{21}$ actually cancels just like for the $N=4$ supersymmetric background: the $11/3$ coefficient from the beta-function knows about the $2\times 11=22$ dimensions in which the $S^{21}$ sphere resides.

None of those backgrounds of the bosonic string/M-theory may be justified by any solid arguments, and they probably don't exist. In particular, there are no RR fluxes that could support the curvature of the sphere. That's true for your $AdS_3$ case of bosonic M-theory, too. Witten's CFT used in the monstrous moonshine has the Leech lattice instead of a sphere but the lattice cannot be decompactified so it's somewhat physically meaningless to link it to the decompactified bosonic string.

Witten's CFT itself, despite the name of its author, has its problem and probably can't be generalized to a larger curvature radius of the $AdS_3$, as shown by Davide Gaiotto using group-theoretical arguments. And bosonic M-theory itself is the least justified among all those building blocks. The lack of supersymmetry guarantees that all numerical comparisons between the two sides of the dualities fail. Moreover, there are no chiral fermions and no anomalies, so there's no nontrivial counterpart of the $E_8$ anomaly cancelation on the supersymmetric M-theory domain walls.

Some of the analogies are suggestive but supersymmetry prevents you from seeing that the backgrounds are stable, dual, and well-behaved, and if you ask me, I think that it's more likely that this is not just an illusion: the lack of supersymmetry simply makes these backgrounds genuinely pathological (unstable and other things). Supersymmetry plays a crucial role for consistent string/M-theory solutions and all these bosonic models should be viewed as mere toy models to learn the physics that really makes sense, and it's the physics of (either spontaneously broken or unbroken) supersymmetric vacua.

| cite | improve this answer | |

I thought about this a little after reading Susskind/Horowitz. There the bosonic M-theory is on an interval. This is seen also in heterotic M-theory in the sense that M-theory on a interval is an $E_8\times E_8$ string theory. one $E_8$ for each boundary (Witten/Horava). The interesting thing is that on the boundary $\mathbb{M}$ gives the Bimonster. the Wreath would be related to the interval somehow. $\mathbb{M}\wr\mathbb{Z}_2$ this makes sense since right moving is 26. There are also the 2branes and 21branes with AdS$_3$/CFT as in the Witten case. I dont know how the $c=24k$ would arise perhaps from $k$-2 branes. It might be better to work with the Borcherd lifts.

| cite | improve this answer | |

It might be that the ultimate structure of the universe is monstrous-moonshine. The potential role is with the interior of that $\hbar,~i$ planar region with the various string theories connected by a region which is not well known. If the $27$ dimensional system is the Jordan matrix algebra $J^3({\cal O})$, there are three scalar elements on the diagonal and the $E_8$, or octonions, ${\cal O}_i,~i~-~1,~2,~3$. The infinite momentum constraint on the diagonals, with metric signature $[-,+,+]$ reduces this to $26$ dimensions.

The following paper:

Borcherds symmetries in M-theory

Dedicated to Pr. S. Hawking on his 60th birthday.

Pierre Henry-Labord`ere, Bernard Julia and Louis Paulot


should be of interest.

The connections with $n~=~2,~3,~4$ n-qubit systems is with the automorphism groups of the groups, or in $24$ dimensions $Aut(M_{24}),~{\mathbb Z}_n~=~\Lambda_{24}$, which is the Leech lattice. The bipartite, tri-partite (qutrite) and 4-qubit entanglements are associated with the particular sublattice systems, $D_4$, $E_8$ and $\Lambda_{16}$, where the last is the Barnes-Wall lattice. The above infinite momentum constraint on the $3$ scalars reduces this to symmetries on a $2$ dimensional hyperbolic space with the two dimensional group, $SL(2,R)$, which are the isometries of the $AdS_2$. Without the constraint this is the isometries of $AdS_3$. This then does segue into Witten’s three dimensional gravity.

What Lobos indicates with this being a bosonic theory is worthy of note. However, with string theory we start with the bosonic string in $26$ dimensions, which is the dimension where the Kac-Moody anomaly cancellation occurs. The structures above have strong connections with $26$ dimensional string. The dimensions of $24$, $26$ and $27$ are particularly interesting. $24$ dimensions is the dimension of the Leech lattice, $26$ is half of the $F_4$ group, $dim(F_4)~=~52$

The discrete group $PSL(2,~7)$ is the automorphism group of the octonionic Fano plane. For complex-octonions the automorphism is $F_4$ of 52 dimensions. $F_4$ is equivalent to adding 16 short roots vectors to the four roots of the group $SO(9)$. This is also the symmetry of the 24-cell. The short roots define the quotient $F_4/SO(9)$ which defines the sequence: $$ F_4/SO(9):~1~\rightarrow~spin(9)~\rightarrow~F_{4\setminus 52}~\rightarrow~{\cal O}P^2, $$ where $4\setminus 52$ means the group is restricted to $36$ of the $52$ dimensions. The ${\cal O}P^2$ defines the Cayley plane. $F_4$ is the automorphism group of the Jordan algebra with the symmetric product $X\odot Y~=~(1/2)(AY~+~YX)$. The group $F_4$ acts transitively on ${\cal O}P^2$ with isotropy subgroup $spin(9)$. For the octo-octonions the group is the exceptional $E_8\times E_8$ with heterotic structure.

This does ultimately connect to supergravity. It does so in much the same way that the bosonic string connects with supersymmetric string theory. The $F_4/B_4$ gadget fits into this. Suppose we have that $SO(16)~\rightarrow~~SO(14,2)$. Now let us restrict this to $SO(14,1)$ the Lorentz spacetime. We then have the further reduction due to the lightcone gauge or the “infinite momentum frame. This reduces dimensions further, yet we may not want to do that here. We have the $SO(16)\times SO(10)$ in the type II theory corresponding to the 26-dimensional bosonic string. This embeds into the larger $SO(26)$. So let us do our reductions on $SO(24,2)$, to $SO(24,1)$, then the infinite momentum or light cone gauge reduction is to SO(23,1).

The $SO(24)$ group (here Euclideanized for simplicity) is the string theory of quantum gravity. The $E_8\times E_8$ corresponds to the left and right moving modes on the string. Consider a heterotic string with $26$ dimension for left moving modes, which correspond to $10$ right moving modes. The k-metric is then $k^2~=~k_l^2~–~k_r^2$, which for $10$ of the $k_l$ modes $k_l~=~k_r$ component by component $1$ to $10$, then the k-metric is $16$ dimensional. This is the origin of the $E_8\times E_8$ or equivalently $SO(2*dim(k))~=~SO(32)$. The bosonic portion in $26$ dimensions contains two dimensions which are tachyonic vacuum modes and the remaining an $SO(24)$ field with tensor $\Omega_{ab}$. The string operators $\alpha^a_{-1}$ projects the vacuum $|0\rangle$ to $|\Omega_{ab}\rangle$ by $$ \alpha^a_{-1}\alpha^b_{-1}|0\rangle~=~|\Omega^{ab}\rangle $$ This field has a traceless portion which is the spin-2 graviton, the trace term a dilaton operator and the anti-commuting portion define a gauge field theory in the adjoint representation. This is a decomposition into $SO(8)$ and $SO(16)$, where the $SO(8)$ portion is conformal gravity and the $SO(16)$ gives the roots of $E_8$.

The $AdS_n$, $n~=~2,~3$ gravity then has a correlation with the standard $SO(24)$ bosonic string gravity in Green, Schwarz & Witten. I think there is a great prospect for the future of this, which should lead to a monstrous-moonshine realization of the universe.

| cite | improve this answer | |
  • 3
    $\begingroup$ There is a lot of interesting numerology mentioned here but I have a hard time following the logic. $\endgroup$ – pho Feb 7 '11 at 15:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.