1

d4 x(-g)1/2 i ¯ψγa V µ a (∂µ + Σbc ˜ωbc µ )ψ + H.c. (20) Integration parts, use of properties of the Dirac gamma matrices, and introduction of a coupling to Aµ yields [6] ˜ID = d4 x(-g)1/2 i ¯ψγa V µ a (∂µ + Σbcωbc µ - iAµ - iγ5 Sµ)ψ, (21) where Sµ = 1 8 (-g)-1/2 ǫµαβγ Qαβγ

3/2 (x)ψ(x) since, just like the vector potential Aµ, the equally minimally coupled Sµ also has zero conformal weight [8]. The ˜ID action thus has a remarkably rich local invariance structure, as it is invariant under local translations, local Lorentz transformations, local gauge transformations, local axial gauge transformations, and local conformal transformations. With Sµ having a structure identical to the Faraday Law structure given in Eqs. (5) and (6), and with Sµ precisely being an axial

4-vector, Sµ is thus the natural quantity to act as the second potential that appears in Xµν [9], and thus the natural axial vector needed to set up a variational procedure for Faraday’s Law of electromagnetism [10]. However, in order to set up a variational procedure we will need to construct a kinetic energy term for it. To generate such a kinetic energy term we appeal to the Dirac action. Specifically, we recall [11], [6] that when one does a path integration D ¯ψDψ exp(i˜ID) over the fermions (equivalent to a one fermion loop Feynman graph) one generates an effective action of the form [12] IEFF = d4 x(-g)1/2 C 1 20 Rµν Rµν - 1 3 (Rα α)2 + 1

β(x) ψ, we can introduce Aµ and Sµ by minimal coupling or by changing the geometry. The two potentials needed for electromagnetism can thus be put on a completely equal footing. Now a drawback in using a Bµ-dependent Wλ µν is that with it parallel transport is path dependent, with the geometry being a Weyl geometry rather than a Riemannian one. However, with iAµ the geometry associated with IEFF is a regular Riemannian one that uses only the connections given in Eq. (16). Thus by using iAµ instead of Bµ we convert a Weyl geometry into a Riemannian one. IV. THE NATURE OF TORSION While we have seen that the axial 4-vector Sµ gives electromagnetism a chiral structure, we need to comment on the fact that experimentally there is no apparent sign of Sµ. Moreover, since Sµ is associated with torsion it is not simply a typical spacetime axial vector field. To underscore the special nature of torsion, we note that even if the standard (torsionless) Riemann tensor is zero, torsion is not obliged to vanish. Torsion could thus exist in a spacetime with no Riemann curvature at all. In a space that is flat as far as the geometry of its four space-

with Eq. (22), we can consider the possibility that the torsion can written as Qλµν = ∇λAµν, where Aµν is an antisymmetric rank two tensor. In [10] it was suggested that this Aµν could be the antisymmetric part of a 16-component metric tensor. To this end, we now note that if we introduce a set of Grassmann vierbeins ξa µ, then the quantity Aµν = ηabξa µξb ν will be antisymmetric since ξa µξb ν + ξb µξa ν = 0. Thus we can envisage spacetime being enlarged to encompass both ordinary coordinates and Grassmann coordinates, and spaces of this type were constructed in e.g. [16], where it was shown that a canonical quantization in which vanishing anticommutators were replaced by non-vanishing ones led to the Dirac equation. As also noted in [16], because of the Pauli principle finite degree of freedom Grassmann coordinates ξb µ (as opposed to infinite degree of freedom Grassmann fields ψ(x)) could not be macroscopically oc-

Under a local conformal transformation the torsion transforms as [6] Qλ µν → Qλ µν + qΩ-1 (x)(δλ µ∂ν - δλ ν ∂µ)Ω(x) where q is its conformal weight. No matter what the value of q, under a conformal transformation Sµ transforms into itself with all derivatives of Ω(x) dropping out

Since Xµν is invariant under Aµ → Aµ + ∂µα(x) and Sµ → Sµ + ∂µβ(x), Xµν is recognized as the chiral generalization of Fµν. On allowing for these two transformations, Aµ and Sµ together contain 6 degrees of freedom, exactly the same number as ETOT and BTOT combined

The Riemann and Maxwell sector terms are given in [11] and references therein. The Sµν Sµν term is given in [6]

As an alternative to Grassmann coordinates one could construct the antisymmetric Aµν via Aµν = CabV a µ V b ν, where Cab is the charge conjugation matrix. Torsion would then be macroscopic, but could still have escaped detection if the bridge between the Aµ and Sµ sectors is purely quantum-mechanical. Another possibility is for Sµ to acquire a large Higgs mechanism-type mass