Solution (source code)

= Solution

Start with the <Clifford algebra> $\{\gamma^\mu,\gamma^\nu\}=2\eta^{\mu\nu}$. For the dimension count, complexify and rescale the timelike matrix so that the generators obey the Euclidean complex relation $\{\gamma_a,\gamma_b\}=2\delta_{ab}$. This changes neither the <complex representation> dimension nor its decomposition under the complexified even algebra. In $D=2k$, pair the generators into
$$
a_j=\frac{\gamma_{2j-1}+i\gamma_{2j}}2,\qquad
b_j=\frac{\gamma_{2j-1}-i\gamma_{2j}}2\quad(j=1,\ldots,k).
$$
Direct substitution gives $\{a_i,a_j\}=\{b_i,b_j\}=0$ and $\{a_i,b_j\}=\delta_{ij}$. The commuting operators $n_j=b_ja_j$ satisfy $n_j^2=n_j$, so their occupation <eigenvalues> are zero or one. An irreducible module contains an all-zero occupation state $|0\rangle$, obtained by lowering the occupied entries of a simultaneous <eigenvector>. The states
$$
b_1^{e_1}\cdots b_k^{e_k}|0\rangle,\qquad(e_1,\ldots,e_k)\in\{0,1\}^k,
$$
are nonzero and independent by their distinct joint occupation <eigenvalues>. They span an invariant module; irreducibility makes this the full module. Conversely this Fock construction supplies a representation on exactly these $2^k$ states. Products of occupation projectors and raising/lowering operators give its matrix units, so the generated algebra is the full <matrix algebra>. Therefore the minimal complex <Dirac spinor> has
$$
\boxed{\dim_{\mathbb C}S_{\mathrm{Dirac}}=2^k\quad(D=2k).}
$$

This module is irreducible under the full <Clifford algebra>, but the <Spinor representation of the Lorentz group> is generated by the even bilinears $\Sigma^{\mu\nu}=\tfrac14[\gamma^\mu,\gamma^\nu]$. The <chirality matrix> in the Euclideanized basis is
$$
\gamma_*=i^{-k}\gamma_1\cdots\gamma_{2k},\qquad
\gamma_*^2=1,\quad\{\gamma_*,\gamma_a\}=0,\quad[\gamma_*,\Sigma^{\mu\nu}]=0.
$$
Its two eigenspaces therefore remain invariant under <Lorentz transformations>. Each has half the Fock states, since any <gamma matrix> bijects them. The even algebra connects every pair of occupation states of the same parity and has no further invariant subspaces within either one. Hence
$$
\boxed{S_{\mathrm{Dirac}}=S_+\oplus S_-,\qquad
\dim_{\mathbb C}S_\pm=2^{k-1}.}
$$
These are the irreducible <Weyl spinors>. Globally they are ordinary representations of $\operatorname{Spin}(D-1,1)$, the double cover; calling them representations of $SO(D-1,1)$ uses the conventional projective spinor terminology.

For $D=2k+1$, adjoin $\gamma_{2k+1}=\pm\gamma_*$. It anticommutes with the old matrices and gives a Clifford module of the same dimension $2^k$. The new even generators $[\gamma_a,\gamma_{2k+1}]$ interchange the two old chirality eigenspaces, so the odd-dimensional Lorentz spinor is \b[irreducible of complex dimension $2^k$], with no Weyl splitting. The product of all odd-dimensional <gamma matrices> is central and acts as a scalar; it cannot be a new anticommuting chirality operator. Its two signs distinguish the two complex odd-Clifford modules, whose restrictions to <Spin group> are equivalent. <Majorana spinor> reality conditions, when the signature permits them, are additional conditions and should not be confused with complex chirality.

For the reduction, take the massless zero modes on a flat seven-torus with no flux or projections. The eleven-dimensional fields are the metric $G_{MN}$, a three-form $A_{MNP}$ and a Majorana <gravitino> $\Psi_M$. Split $M=(\mu,i)$ with $\mu=0,\ldots,3$ and $i=1,\ldots,7$. The metric supplies $G_{\mu\nu}$, seven vectors $G_{\mu i}$ and 28 <scalar fields> $G_{ij}=G_{ji}$. The three-form supplies a four-dimensional three-form $A_{\mu\nu\rho}$, seven two-forms $A_{\mu\nu i}$, $\binom72=21$ vectors $A_{\mu ij}$ and $\binom73=35$ scalars $A_{ijk}$.

A massless two-form in four dimensions has one physical <particle polarization> and can be dualized to a scalar by $H_3=*d\sigma$; the seven two-forms thus supply seven more scalars. The four-dimensional three-form has no local <particle polarization>, since its four-form field strength is locally a constant times the volume form. The total bosonic content is consequently
$$
\boxed{1\text{ graviton},\qquad28\text{ vectors},\qquad70\text{ real scalars}.}
$$
The eleven-dimensional <Majorana spinor> has 32 real components, decomposing into eight four-dimensional <Majorana spinors>. Accordingly $\Psi_\mu$ supplies eight <gravitini>, while $\Psi_i$ supplies $7\times8=56$ spin-$1/2$ fields after diagonalizing the <gravitino>/matter kinetic terms. The resulting theory has 32 real <supercharges>, hence is <four-dimensional N=8 supergravity>, with
$$
\boxed{8\text{ gravitini},\qquad56\text{ spin-}\tfrac12\text{ fermions}.}
$$
The on-shell counts check the reduction: the metric has $11(11-3)/2=44$ polarizations and the three-form has $\binom93=84$, totaling 128. In four dimensions $2+28\cdot2+70=128$ bosonic polarizations match $8\cdot2+56\cdot2=128$ fermionic ones. Higher Kaluza–Klein levels, fluxes or projections would be additional choices beyond this massless toroidal reduction.