Lawrence

Nice work. It's a nice exercise to prove eigenvalues of the exceptional Jordan eigenvalue problem are real. You have enough content in your paper to show this. Essentially, one can use the trace to define an inner product on J(3,O) and use positive definiteness to get your result. However, once you study the eigenvalue problem in the split-octonion case, you can no longer argue by such means. In general, the eigenvalue problem one must study is over the full exceptional Jordan C*-algebra.

See Sparling's paper A Primordial Theory for insight on "fat points" of the split Cayley plane.

    Michael,

    The full C*-algebra problem means split Cayley plane is in effect a form of twistor space. The question I have is whether or not I can still have my idea about Morse theory, or maybe in this case Floer cohomology. I think that these large numbers of SLOCC entanglements with 1/2, 1/4, 1/8 supersymmetry are stable points in this general manifold. There may be quantum transitions between these, and the black hole horizon induces a type of uncertainty in the nature of entanglement. This might be a way around quantum monogamy that results in the firewall.

    I'll take a look at the paper you referenced.

    Cheers LC

    Lawrence

    Your Morse index tunneling picture looks promising. You might want to take a look at the attractor mechanism in supergravity for further insight.

    Michael,

    The link does not seem to work.

    The J3(O) matrix is diagonalized by F4 into real eigenvalues. The F4 is the automorphism group of the Jordan 3x3. The sequence

    F4: 0 --- > A_4 --- > F_{52/16} --- > OP^2 -- > 0

    Means the eigenvalues found with F4 can be found in part with the subgroup A_4 = SO(9) or SO(8,1).

    The E6 contains SO(10) or SO(9,1) . As with spacetime and the Lorentz group SO(3,1) the rotations SO(3) on a spatial surface that in turn foliate spacetime. These foliations define a thickened small region of spacetime with some lapse function, and the two spatial boundaries define two SO(3)s. In this way the SO(9,1) embeds SO(9) in a pairing. This is a physical reason why the E6 ÔèéSO(9,1) is over the CxO. Similarly the E7 is over HxO, the quarter-octonions.

    For E6 the eigenvalues are not necessarily real. These complex numbers may appear perplexing, and further violate unitarity. However, in this C* algebraic form of the Weyl-Heisenberg algebra (a, a^ÔÇá, N, c):

    [a, a^ÔÇá] = 2c

    [N,a] = Ôê'a

    [N,a^ÔÇá] =aÔÇá

    [c,Ôï...] = 0,

    where the standard theory has c = ›. We can see this does not result in a larger phase space volume or demolish the Hessian transformation principles of conservative physics. The Hopf algebra coproduct is introduced:

    ╬"a = aÔè--I + IÔè--a,

    which is a coproduct "duplication" of the operator into left and right forms of the operator. Now the coproduct can be deformed forming a quantum group:

    ╬"aq = aqÔè--qc + qÔê'cÔè--aq

    where q is the deformation parameter related to a geometric-like series:

    [x]q = q^x Ôê' q^{-x}/(qÔê'q^{Ôê'1})

    The elements of the q-deformation can be functions of any form. In standard QM these would be complex variables or functions, but they can also be real valued.

    The deformation ╬"a(q) = a(q)Ôè--q^c + q^{Ôê'c}Ôè--a(q) is related to Bogoliubov operators. We may write

    q^c = Ie^{2¤Çi¤å} = I(cos(2¤Ç¤å) + i sin(2¤Ç¤å)),

    for ¤å a parameter related to acceleration. Of course in addition

    (q^c)^{-1} = Ie^{-2¤Çi¤å} = I(cos(2¤Ç¤å) - i sin(2¤Ç¤å)).

    This gives

    ╬"a(q) = a(q)Ôè--I(cos(2¤Ç¤å) + i sin(2¤Ç¤å)) + (cos(2¤Ç¤å) - i sin(2¤Ç¤å))IÔè--a(q)

    = (a(q)Ôè--I + IÔè--a(q))cos(2¤Ç¤å) + i(a(q)Ôè--I - IÔè--a(q))sin(2¤Ç¤å)

    We may of course form the Hermitian conjugate

    ╬"a^ÔÇá(q) = a^ÔÇá(q)Ôè--q^{-c} + qÔè--a^ÔÇá(q)

    = (a^ÔÇá(q)Ôè--I + IÔè--a^ÔÇá(q))cos(2¤Ç¤å) + i(a^ÔÇá(q)Ôè--I - IÔè--a^ÔÇá(q))sin(2¤Ç¤å).

    The commutator [╬"a(q), ╬"a^ÔÇá(q)] is

    [╬"a(q), ╬"a^ÔÇá(q)] = ([a(q), a^ÔÇá(q)] Ôè--I + IÔè--[a(q), a^ÔÇá(q)])(cos^2(2¤Ç¤å) + sin^2(2¤Ç¤å)),

    which gives a unit product.

    The hyperbolic trigonometric functions obtained from ¤å --- > i¤å are squeeze parameters and we have a C* form of the Bogoliubov operators

    b = au + a^ÔÇáv, b^ÔÇá = a^ÔÇáu + av

    for u = cosh(2¤Ç¤å) and v = sinh(2¤Ç¤å). The commutator of b and b^ÔÇá is

    [b, b^ÔÇá] = u^2 - v^2 = 1.

    The q-deformed algebra has the same u^2 - v^2 = 1. This then means we have the operators employed in computing Hawking and Unruh radiation.

    Cheers LC

    Indeed E6(C) contains the SO(9,1) and SO(5,5) determinant preserving groups of the octonion and split-octonion 2x2 Jordan algebras, respectively. From a twistor space perspective, the SO(9,1) transforms points on curves in OP^2, while not preserving distance, but merely preserving collinearity. This is sufficient to attempt to extend Witten's twistor amplitudes to the Cayley plane. (Picture 8-sphere D-instantons here.)

      I got the ppt file to work. For extremal BHs this will work with the 28 charges and 28 magnetic monopoles in the 57.

      LC

      Excellent. And remember, the norm for the 57 admits non-real solutions so the complete picture is of the extended Freudenthal triple system over the bi-octonions, i.e., the 57-dimensional charge-entropy space contains complex charges acted on by E8(C). Charge quantization gives charges that are Gaussian integers. The relation to Shimura varieties enters here, where instead of considering elliptic curves we consider the coset spaces of the U-duality groups modulo their maximal subgroups in integral form. The coset spaces are the moduli spaces of the extremal black holes, and in general we have very many copies of such spaces, as we consider systems of such black holes. The Shimura datum encodes these. Once one has the L-functions for these varieties, we can define their (mixed) motives and have clear knowledge of their geometrical and topological properties.

      I think this has to do with gauge theoretic aspects of the quaternions. I attached some time back a post on how quaternions naturally give rise to YM gauge fields. For the momentum p_{aa'} = О»_aО»_a' and the polarization Оµ^{aa'} we have the transversality condition p_{aa'}Оµ^{aa'} = 0. In addition there is the gauge transformation Оµ_{aa'} --- > Оµ_{aa'} + k p_{aa'}, k = constant so that p_{aa'}Оµ^{aa'} = 0 is invariant for a massless field. We can then find the gauge field according to these as

      F^{ОјОЅ} = Пѓ^Ој_{aa'}Пѓ^ОЅ_{bb'}( p_{aa'}Оµ^{bb'} - p_{bb'}Оµ^{aa'})

      The H^2 space constructs the gauge field and its dual to give the intersection form ∫F/\F = 8πk. Working in C^4 does not bring that structure in.

      The 8-dim spacetime enters into this picture as the E8 Cartan matrix. The intersection form is an invariant with respect to that. This is wrapped up in the theory of gauge field in four dimensions. There is underlying this a four dimensions of C^4 or H^2 in complex variables, but underlying this are 8 dimensions in pairs of real variables.

      LC

      From a projective space description the degree one, genus zero curves in OP^2 are 8-spheres. If Witten's formalism carries over to the Cayley plane, MHV amplitudes should localize on these curves. But first, I agree, it is better to develop the H^2 case with SO(5,1) symmetry.

      The symmetry is really SO(4,2) ~ SU(2,2). The group on the left is the isometry group of AdS_5 = SO(4,2)/SO(4,1). The group on the right is the group for twistor space with PT^ = SU(2,2)/(SU(2,1)xU(1)). We may think of the quaternion form of this as a transition from C^4 to H^2 with SL(4,C) --- > SL(2,H). This has four complex dimensions, eight real dimensions. We might of course be so bold as to double down, with SU(2,2,C) so that SL(4,C) --- > SL(8,C), but with the above identification of pairs we have that this is SL(4,H) ~ SL(2,O).

      LC

      Using J(2,H), gives determinant preserving group SL(2,H)=SO(5,1) in Coll(HP^2)=Str_0(J(3,H))=SL(3,H)=SU*(6). Like the SO(3,1) case, where one represents the momentum vectors as bi-spinors in J(2,C), a given vector is lightlike if and only if its determinant is zero. Hence, in the quaternion case, SL(2,H)=SO(5,1) preserves the lightlike property of each quaternionic bi-spinor.

      If one considers a Freudenthal triple system (FTS) over J(3,H), its automorphism group Aut(F(3,H))=SO*(12) contains SU*(6), as the collineation subgroup over each HP^2 component. The FTS contains two copies of HP^2, considered as electric and magnetic projective planes in N=2, D=4 symmetric SUGRA for dyonic extremal black holes.

      Moving up to the octonionic case, we have SL(2,H)=SO(9,1) in Coll(OP^2)=E6(-26). This gives FTS automorphism group Aut(F(3,O))=E7(-25) which contains an SO(2,10) subgroup. Your SO(2,4) is recovered if we restrict to a complex sub-algebra. Hence, in the octonion case we essentially have an AdS_11, which upon reduction to a complex subalgebra, we recover your AdS_5.

      Extending the FTS to 57-dimensions gives E8(-24) with an SO(3,11) subgroup.

      For division algebras, we have the inclusions:

      R: SO(4,3)->SO(3,2)->SO(2,1)

      C: SO(5,3)->SO(4,2)->SO(3,1)

      H: SO(7,3)->SO(6,2)->SO(5,1)

      O: SO(11,3)->SO(10,2)->SO(9,1)

      In the split division algebra cases:

      Cs: SO(4,4)->SO(3,3)->SO(2,2)

      Hs: SO(5,5)->SO(4,4)->SO(3,3)

      Os: SO(7,7)->SO(6,6)->SO(5,5)

      There is a super Yang-Mills theory formulated with SO(11,3) gauge symmetry, as given by Sezgin in hep-th/9703123. Bars extended this to supergravity in hep-th/9704054. Also recall that Lisi has interest in SO(11,3) based on Nesti and Percacci's GraviGUT theory [link:arxiv.org/abs/0909.4537]arXiv:0909.4537 [hep-th][/link]. With the insight from extremal black holes, it is tempting to conjecture the SO(11,3) super Yang-Mills theory is contained in a larger theory with E8(-24) symmetry. An (11,3)-signature theory would be a 14-dimensional generalization of M-theory in D=11.

        The inclusions

        R: SO(4,3)->SO(3,2)->SO(2,1)

        C: SO(5,3)->SO(4,2)->SO(3,1)

        H: SO(7,3)->SO(6,2)->SO(5,1)

        Have as their middle group the isometry group for AdS_4, AdS_5 and AdS_7. These all share a certain relationship with each other. The AdS_5 gives as 10 dimensional theory for AdS_5xS^5, while the other two give 11-dimensional theory for AdS_4xS^7 and AdS_7xS^4. These last two are dual theories and these have a relationship with the 10 dimension theory in sense that SO(9) is a subset of SO(9,1).

        The unification of these implies they are physically an aspect of the octonion. The groups in these inclusions SO(2,1), SO(3,1), SO(5,1) and SO(9,1), dimensions = 3, 4, 6, 10 correspond to physics in these respective dimensions, with 4 for gravity, 6 the CY manifold of compactification. This seems to hinge on the collination of the Moufang OP^2 by E6(-26) which contains SL(2,H) ~ SO(9,1).

        There seem to be interesting connections here, and I have this suspicion that an aspect of this is related to the 5-dim moduli space for SU(2) or SO(4) bundles on 4-manifolds. The Lorentzian case gives the hyperbolic SO(4,2)/SO(4,1) ~ AdS_5.

        More later after I try to bend some more metal on this.

        Cheers LC

        What I have is a couple of questions. We have that a group G decomposes into H and K or G = HвЉ--KвЉ--M with the algebraic g = hвЉ•kвЉ•m. This Cartan decomposition for the exceptional groups E6(C) and E7(C) are H = E6 and K = SO(10)xSO(2) and H = E7 and K = E6xSO(2). We also have for G = E8(C) that H = E8 and K = E7xSU(2). These are symmetric spaces and are moduli of the form E6/SO(10)xU(1), E7/E6xU(1) and E8/E7xSU(2). The question is whether these are related to SLOCC groups, such as

        G_4/H_4 = E_{7(-25)}/(E_{6(-78)}xU(1))

        G_3/H_3 = E_{8(-24)}/(E_{7(-25)}xSL(2,R))

        The first of these is D = 4, N = 2 SUSY and the second D = 3, N = 2 SUSY. These moduli spaces with E6 have SU(3) holonomy. For the static CY these are cases of the static Hodge diamond for K3 or T^6. These groups are then moduli for the CY compactification. There is then some relationship between moduli spaces. The question is, what is the nature of these moduli spaces?

        The analysis I presented last week concerned characteristic determinant equation

        -det(A - О»I) = О»^3 - Tr(A)О»^2 + Пѓ(A)О» - det(A) = x

        for x the root of an equation involving the commutator of the elements z \in O and the associator. I worked this out for σ(A) = R(χ. χ') = ω[χ. iχ'] = symplectic form. This equation looks very similar to an elliptic curve. Of course this is for real eigenvalues diagonalized by F4. For the complex case we have a similar system diagonalized by E6. It then seems plausible that an analogous equation exists for E6 that is an elliptic curve. This means that our moduli space is a system of modular forms that are also an algebraic variety. This means the spaces above have to be taken with a quotient of some finite discrete group. The question I whether this makes any sense?

        Some aspect of conformal completion of AdS spacetimes involves Kleinian groups, and I have an interest in the prospect this system would work. Heegner prime numbers are from the discriminant of elliptic curves, and these might give a quotient that would satisfy this requirement. I am not sure it this would work or not, but it seems plausible in some way.

        Cheers LC

        The symmetric spaces you mention are for the compact (complex) cases. One recovers these groups using the complexified octonions. The SLOCC groups you listed are produced by the octonions. You can use the split-octonions for E6(6), E7(7) and E8(8). You'll notice E7(7) corresponds to the symmetry for D=4, N=8 SUGRA.

        For the eigenvalue problem, recall that E6(-26) corresponds to the reduced structure group of the exceptional Jordan algebra. Apply its determinant preserving property to the Jordan eigenvalue problem to see how it acts on eigenvalues and eigenmatrices. This will show you how to proceed with your elliptic curve analogy.

          So I am not completely bat shit crazy here. What you say about E6(-26) appears workable since it has F4 as the mcs, which in turn diagonalizes the J3(O) in the reals. As Einstein once said, "I need a little think." This stuff is not easy for me, particularly since I have been primarily employed in applied work.

          Cheers LC

          The E6(-26) scales the eigenmatrices, while the F4 fixes their lengths, as measured by the Frobenius norm (trace norm) over J(3,O). There is another geometrical property preserved, as you can check for yourself. Hint: imagine each eigenmatrix assigned to a vertex.

          I will probably have to communicate by email, as I think this contest is ending soon.

          The f4 algebra is the set of isometries of C\otimes OP^2, which is in a sense the "complexification" of the OP^2 that has F4 as its isometry group. The 78 dimensions of E6 is 52 26 so that E6 = F4x26, where the 26 are vectors corresponding to the Killing form = -26. The collineations e6 correspond to isometries of OP^2 by F4 as SO(9,1) and SO(9). The points that are collinear in OP^2 correspond in C^3 to SL(4,C). These are points collinear in C^3, This complex conformal group defines the set of lines connected to a loop. This is the standard system of two vertices and and two vertices _ and -.

          This will then define collineations of points on algebraic curves in CP^3. I suspect in some way this may lead to the elliptic curve.

          More later,

          LC

          If we consider the Jordan-Wigner QM as diagonalized by F4, this seems to fit with the F4 scheme for the Kochen-Specker contextuality result. If we shift to E6 are we then generalizing the nature of quantum mechanics? Would the occurrence of real and imaginary eigenvalues correspond to the Bogoliubov coefficients in Hawking radiation.

          I am curious about the physical implications of this.

          Cheers LC

            Indeed F4 is the group of unitary transformations for a Jordan formulation of three-dimensional octonion quantum mechanics. E6(-26) is the group of collineations for OP^2, which includes F4 as a subgroup. In other words, one can consider F4 as a rotation group and E6(-26) as a Lorentz group, as is typically done in the extremal black hole literature.

            As I look at E6, this seems to be the case. F4 ~ SU(3,O) and E6 ~ SL(3,O), which implies there are now an additional set of operations. These are analogous to the case with SL(2,C) ~ SU(2)xU(1,1) or SU(2)xSL(2,R). The E6 is then SL(3,O) ~ SU(3,O)xSL(3,H). These SL(3,H) transformations are rapidities similar to boosts in relativity.

            Kochen-Specker proved that the measurement of can't be described by a nonlocal hidden variable that has no dependency on the context of the measurement. The Kochen-Specker theorem depends upon 48 root vectors of the F4. These are the vertices of the 24 cell. The 24 vertices have lines that emanate from them. These are basis elements for 24 quantum states. These lines define projective rays in four dimensions. For any projective ray there are three others that are orthogonal to it. There are 36 quartets of these rays that are mutually orthogonal. However, there is only need for 9 of these, where the other 24 amount to over counting. These orthogonal conditions are a condition for vanishing commutators. There are further 18 rays that exist in two sets.

            If each of these 18 in two sets is determined by a hidden variables. This means there are 9 variables set to one state, say 0 or 1, and the other 3 set to 1 or 0. Yet since each appears twice this means there is an even number of these, and 9 is and odd number. It is also the case that these 9 correspond to the vanishing of 9 commutators in the BFSS holographic theory of the string. The fibration

            F4:A_4 ---> F_{52/36} ---> OP^2

            concerns the "36" that are ker/im of the maps, and form a cohomology.

            I think we can take this further with E7. With E6 we are moving from R to C, and we have "double downed" on the number roots. We can move to H and do this again. We now have that E7 is dim = 133. We can have the quotients E7(-7)/(F4(-4)xR^{26} ~ E7(-25)/(E6(-78)xU(1)) for 2 sets of 27 gauge potential plus a graviton. Physically this "double down" amounts to pairing up the boson field above. The gauge/gravity duality, by Dixon, Bern et al, illustrates how the graviton has the same quantum numbers as a pair of gluons in a colorless entanglement. This construction appears to suit this pretty well.

            The twistor approach to this is to move from C^4 to H^4, which with twistor space CP^3 from C^4 gives us PH^3. This does contain CP^3xCP^3, which means a form of twistor theory is contained in this. Of course this may be generalized I think for supermanifolds CP{3|4} into PH^{3|4}. However, the universal bundle theorem for the G_{2,2} will involve a generalization form

            G_{2,2} = SU(2,2)/(SU(2)xSU(2)

            into

            G_{2,2,C} = SU(2,2,C)/(SU(2,C)xSU(2,C)

            or

            G_{4,4} = SU(4,4)/SU(4)xSU(4).

            Cheers LC