Preprint
Article

This version is not peer-reviewed.

Residual-Dimension Obstructions for Null-Projected Supersymmetric Yang–Mills Theories

Submitted:

14 November 2026

Posted:

16 November 2026

You are already at the latest version

Abstract
We classify null-projected supersymmetric Yang--Mills theories with a Majorana--Weyl gaugino in signature $(s,t)$, defined by five conditions on the gaugino constraint, the transformation rules, the residual gauge field, the translation bracket and the absorption of the cubic gaugino terms. Such a theory built on \(k\) null vectors exists if and only if \(s+t-2k=10\) and \(s-t\in\{8,0,-8\}\), or \(s+t-2k=2\) and \(s=t\). The key step is a converse of the cubic gamma-matrix identity: symmetric gamma matrices that satisfy it act on spinors of dimension \(2(d-2)\). Hence the signatures (9+8n+q)+(1+q) with n ≥ 1 and 0 ≤ q ≤ 3 proposed by Rios, Marrani and Chester, among them \(27+3\), carry no such theory, and several algebraic statements made alongside them, such as an embedding of \(\mathfrak{f}_{4(-20)}\) in \(\mathfrak{so}(25,1)\), are false. Neither another contraction of the source, nor any number of gauginos, mixed or not, nor a gauge bracket that is not a Lie bracket changes this. In the single-time signatures (9 + 8n) + 1 no unitary massless multiplet without tensor charges consists of scalars, spinors and vectors.
Keywords: 
;  ;  

MSC:  81T60; 81T13; 15A66; 17B25; 81R25; 83E15

1. Introduction

In more than two dimensions, conventional minimal supersymmetric Yang–Mills theory with only a gauge field and a gaugino exists for a single time in three, four, six and ten dimensions [1,2,3,4]. The reason is one algebraic fact. Varying the gauge field inside the covariant derivative of the gaugino kinetic term produces a term cubic in the gaugino, and no other term in the variation of the Lagrangian can cancel it. This term vanishes precisely when the symmetric gamma matrices obey the Fierz identity
γ α β μ γ μ γ δ + γ α γ μ γ μ β δ + γ α δ μ γ μ β γ = 0 ,
which holds for the Majorana spinors of three and four dimensions, the Weyl spinors of six and the Majorana–Weyl spinors of ten, the spinors attached to the four normed division algebras [2,3,4]. The same identity makes the Green–Schwarz superstring κ -symmetric [5], its analogues for forms of higher degree produce the scan of supersymmetric branes [6,7], and Blencowe and Duff extended that scan to spacetimes with several times [8]. Above ten dimensions the spinors are too large: a Majorana–Weyl spinor in D dimensions has 2 D / 2 − 1 components, while (1) allows 2 ( D − 2 ) of them (Theorem 3.3).
The null-frame constructions considered here extend this field content by introducing additional timelike as well as spacelike directions. Signatures with several times have been studied for several reasons. F-theory points to a twelve-dimensional origin of type IIB string theory [9]; Bars proposed an S-theory in 11 + 2 dimensions and made a case for fourteen [10,11]; in two-time physics an S p ( 2 , R ) gauge symmetry removes one time and one space direction [12,13,14]; and the dualities of string theory and M-theory relate theories of different signatures [15,16]. Superparticles beyond eleven dimensions were studied by Rudychev and Sezgin [17], and supergravity in 10 + 2 dimensions by Nishino [18]. Supersymmetric Yang–Mills equations in 10 + 2 dimensions were written by Nishino and Sezgin [19], a theory in 11 + 3 dimensions was given by Sezgin [20], and the mechanism behind these constructions was set out by Rudychev, Sezgin and Sundell [21]. Nishino then built supersymmetric Yang–Mills theories in every signature ( 9 + m ) + ( 1 + m )  [22]. His construction introduces m mutually orthogonal null vectors n ( 1 ) , … , n ( m ) , constrains the gaugino by * * n * * ( i ) λ = 0 , and lets the supersymmetry algebra close on translations contracted with n ( 1 ) ∧ … ∧ n ( m ) . The cubic gaugino terms then need not vanish. They only have to take a form that can be absorbed into the field equation of the gauge field, and in Nishino’s construction they do.
Rios, Marrani and Chester used the periodic reality and symmetry properties of spinors to propose that Nishino’s existence proofs could be transferred to new signatures [23]. Their proposal keeps D = s + t and s − t modulo eight fixed while moving 4 n timelike directions into space. The resulting four families are
( 9 + 8 n ) + 1 , ( 10 + 8 n ) + 2 , ( 11 + 8 n ) + 3 , ( 12 + 8 n ) + 4 ,
with the chiral superalgebras of their eqs. (3.5)–(3.8). They called these theories exceptional super Yang–Mills theories. The signature 27 + 3 , the case n = 2 of the third family, was then proposed as the home of a theory whose fields fill the exceptional Jordan algebra and whose worldvolume would carry M-theory [24]. The same theory was used to model dark matter [25], and it entered the construction of “monstrous” M-theory [26] and the programme of exceptional periodicity [27]. Here we test this proposed transfer within the explicit transformation, projection and absorption assumptions (N1)–(N5). Proposition 7.1 writes out what repeating Nishino’s proof in ( 11 + 8 n ) + 3 requires and shows that each choice of the null frame fails: Nishino’s rank does not fit, the translation bracket allows only rank two, and at rank two the cubic terms cannot be absorbed.
The transfer does not work under (N1)–(N5). Reality and symmetry of spinors are not the only data that enter this mechanism. The null vectors must be mutually orthogonal, and if they annihilate a real gaugino they lie in the complexification of a real totally isotropic subspace, whose dimension cannot exceed min ( s , t ) (Lemmas 2.2 and 2.3). Moving timelike directions into space lowers this number, and with it the number of null directions that can be projected out. What remains after the projection is a theory on the orthogonal complement of the null frame, of dimension D ′ = s + t − 2 k for k null vectors. We show that the cubic terms can be absorbed only if the residual spinors satisfy (1) themselves, and this forces D ′ = 10 whenever D ′ > 2 . The main result, proved in Section 5.4, reads as follows.
Main Theorem. 
Let a supersymmetric Yang–Mills theory with a Majorana–Weyl gaugino in signature ( s , t ) be built on a frame of k null vectors in the manner of Nishino, in the sense of the axioms (N1)–(N5) of Definition 5.3. Then k ≤ min ( s , t ) and s + t − 2 k ∈ { 2 , 10 } , so that s − t ∈ { 8 , 0 , − 8 } . In particular there is no such theory in any of the signatures (2) with n ≥ 1 , and none in 27 + 3 or in 19 + 3 . 
Figure 1 shows the geometry. Nishino’s signatures lie on the plane s + t − 2 k = 10 , while every signature of the families (2) with n ≥ 1 has residual dimension at least 16 for every admissible k. The necessary residual-dimension condition depends on s − t and the available Witt index, beyond their residue data.1 The conditions of the Main Theorem are sharp for residual dimension ten. A complex isometry that keeps D and the rank of the frame carries Nishino’s equations to every signature with s + t − 2 k = 10 and s − t ∈ { 8 , 0 , − 8 } , and the transported theory keeps a Majorana–Weyl gaugino and satisfies (N1)–(N5) (Theorem 5.12). For residual dimension at least three a theory therefore exists exactly when the Main Theorem allows it (Corollary 5.13). In particular 9 + 9 and 9 + 17 , two signatures of the second degenerate family of Ref. [23], do carry such theories, while its members 9 + ( 1 + 8 n ) with n ≥ 3 do not. The conditions are sharp for residual dimension two as well. In ( k + 1 ) + ( k + 1 ) Nishino’s transformations, written with a real frame of rank k, satisfy (N1)–(N5): the cubic gaugino terms vanish identically, and the supersymmetry algebra closes off shell on translations along one null direction of the residual plane (Theorem 5.14). The Main Theorem is therefore a classification. A theory of rank k exists if and only if s + t − 2 k = 10 and s − t ∈ { 8 , 0 , − 8 } , or s + t − 2 k = 2 and s = t (Corollary 5.15).
The exclusion of the families (2) does not depend on the null contraction in the translation bracket either. Whatever fixed tensor contracts the source in the field equation of the gauge field, with or without a constraint on the gaugino, the cubic gaugino terms cannot be absorbed once the residual dimension is sixteen or more, and every real frame in these families leaves at least sixteen (Theorem 7.3 and Corollary 7.4). This covers, for instance, the contraction of the eleven-form charge of 27 + 3 with ten spacelike vectors. Several gauginos do not help either. If each gaugino transforms into the field strength with its own parameter, the cubic terms that couple two different gauginos must vanish, and they do not once the residual dimension is three or more (Proposition 7.5). If the gauginos are mixed by their transformation rules, as in extended supersymmetry, the cross terms between two gauginos can be absorbed only when the Weyl spinors of the residual space have D ′ − 2 components, that is for D ′ = 4 or 6; with Majorana–Weyl gauginos this forces s = t , for any number of gauginos (Theorem 7.6 and Corollary 7.7). The Lie algebra is not needed: for any bilinear gauge bracket the conclusions of the Main Theorem hold, and a bracket that is not antisymmetric, such as the product of a Jordan algebra, leaves only residual dimension two (Theorem 7.11). Transformation rules and field equations nonlinear in the gaugino, and auxiliary fields, change nothing as long as they produce no new term with exactly two undifferentiated gauginos in the variation of the gaugino equation (Lemma 7.13 and Proposition 7.14).
These results do not exclude every theory in 27 + 3 , and no statement about the superalgebra alone could, since the superalgebra acts on superfields (Remark 7.16). An exclusion has to name the fields and their transformation rules. Auxiliary fields whose equations contain a product of two undifferentiated gauginos, the further mechanisms listed in Remark 7.15, and gauge structures in which the variation of the gauge field involves further fields lie outside the present results.
For a single time we can say more, without any assumption about how the theory is built. In ( 9 + 8 n ) + 1 the anticommutator of two supercharges contains the momentum, and in a unitary theory the supercharges that act nontrivially on a massless state generate a Clifford algebra with 2 3 + 4 n generators. A massless multiplet without tensor charges then has at least 2 2 2 + 4 n states, half of them bosonic (Theorem 6.1). For n = 1 that is 2 64 states, against the 16 bosonic states of a vector field in eighteen dimensions.
The paper is organized as follows. Section 2 collects what we need about spinors in signature ( s , t ) : the decomposition of the symmetric square of a Weyl spinor, which shows that the superalgebras of Ref. [23] are the most general ones and carry no information about dynamics, and two lemmas on isotropic subspaces. Section 3 proves the theorem on which everything else rests: if symmetric gamma matrices obey an identity of the type that absorption requires, the spinor dimension is 2 ( d − 2 ) . Section 4 applies it to ordinary super Yang–Mills theory. Section 5 describes null frames by fermionic oscillators, states the axioms, proves the reduction to ten residual dimensions and the Main Theorem, and shows that its conditions are also sufficient. Section 6 treats the single-time family by representation theory, and Section 7 spells out what the results mean for 27 + 3 and 19 + 3 , including the transfer of Nishino’s construction (Proposition 7.1), arbitrary contractions of the source (Theorem 7.3), several gauginos (Proposition 7.5), gauginos mixed by their transformation rules (Theorem 7.6), gauge brackets that are not Lie brackets (Theorem 7.11), nonlinear transformation rules and auxiliary fields (Proposition 7.14) and the scope of the axioms. Appendix A proves the oscillator decomposition, Appendix B describes numerical checks in explicit gamma matrices, Appendix C lists the symmetric squares for the families (2), Appendix D checks Proposition 7.1 in explicit gamma matrices, Appendix E does the same for Theorems 5.12 and 7.3, Appendix F for Theorem 5.14 and Proposition 7.5, and Appendix G for Theorems 7.6 and 7.11. Appendix H applies the same tools, the signature of an invariant form, the Witt index, the branching of spinors and the symmetric square of a Weyl spinor, to algebraic statements made alongside the 27 + 3 proposal and in a related grand-unified model, and shows that they are false. Figure 2 shows how the results depend on one another.

2. Spinors in Signature ( s , t )

2.1. Conventions

Let V = R s , t with metric
η = diag ( + 1 , … , + 1 , − 1 , … , − 1 ) ,
with s plus signs and t minus signs, and D = s + t . Except in Remark 5.11 we take D = 2 r with r ≥ 1 . The Clifford relations are γ μ γ ν + γ ν γ μ = 2 η μ ν . The Dirac module Δ of the complex Clifford algebra Cl ( V C ) has dimension 2 r and splits into the Weyl modules Δ ± of dimension 2 r − 1 , the eigenspaces of the chirality operator Γ * , normalized so that Γ * 2 = 1 . For a vector u we write u = u μ γ μ , and γ μ 1 … μ p is the antisymmetrized product of unit weight, so that γ μ 1 … μ p = γ μ 1 … γ μ p when the indices are distinct and the basis is orthonormal.
There is a nondegenerate bilinear form B on Δ , unique up to a factor once a sign θ = ± 1 is chosen, such that
B ( γ μ ψ , χ ) = θ B ( ψ , γ μ χ ) , B ( ψ , χ ) = ϵ B B ( χ , ψ ) ,
with ϵ B = ± 1 determined by r and θ  [29,30,31]. We write ψ ¯ χ = B ( ψ , χ ) and ψ ¯ γ μ 1 … μ p χ = B ( ψ , γ μ 1 … μ p χ ) . The form B pairs Δ + with Δ + when r is even and with Δ − when r is odd.
A real structure on Δ is an antilinear map C : Δ → Δ with C 2 = 1 that commutes with every γ μ . It exists exactly when s − t ≡ 0 , 1 , 2 ( mod 8 ) , and it preserves Δ ± exactly when s − t ≡ 0 ( mod 8 )  [29,30,32]. A Majorana spinor is a spinor with C ψ = ψ . Majorana–Weyl spinors, which are the spinors of all the theories in this paper, therefore exist if and only if s − t ≡ 0 ( mod 8 ) , and every signature in (2) and in Nishino’s list has this property.

2.2. The Symmetric Square of a Weyl Spinor

The anticommutator of two supercharges in the Weyl module Δ + is a symmetric element of Δ + ⊗ Δ + , so the superalgebras that can be written down in signature ( s , t ) are classified by the decomposition of Sym 2 Δ +  [31,35]. We write Λ p = Λ p V C , and Λ r stands for the half Λ + r of the middle-dimensional forms selected by Δ + .
Proposition 2.1. 
For D = 2 r ,
Sym 2 Δ + ≅ ⨁ j ≥ 0 , r − 4 j ≥ 0 Λ r − 4 j , Λ 2 Δ + ≅ ⨁ j ≥ 0 , r − 2 − 4 j ≥ 0 Λ r − 2 − 4 j .
In particular Λ 1 ⊂ Sym 2 Δ + if and only if r ≡ 1 ( mod 4 ) , that is D ≡ 2 ( mod 8 ) , and Λ 1 ⊂ Λ 2 Δ + if and only if r ≡ 3 ( mod 4 ) . 
Proof. 
The forms ( ψ , χ ) ↦ ψ ¯ γ μ 1 … μ p χ restricted to Δ + × Δ + vanish unless p ≡ r ( mod 2 ) , and those with p ≡ r ( mod 2 ) , 0 ≤ p ≤ r , give an isomorphism Δ + ⊗ Δ + ≅ ⨁ p Λ p  [29,30]. (As a check, ∑ p ≡ r ( 2 ) , p < r 2 r p + 1 2 2 r r = 2 2 r − 2 = ( dim Δ + ) 2 .) It remains to decide which summands are symmetric. By (3),
ψ ¯ γ μ 1 … μ p χ = θ p B ( γ μ p … μ 1 ψ , χ ) = θ p ( − 1 ) p ( p − 1 ) / 2 ϵ B χ ¯ γ μ 1 … μ p ψ ,
so the p-th form has symmetry σ p = ϵ B θ p ( − 1 ) p ( p − 1 ) / 2 , and σ p + 2 / σ p = θ 2 ( − 1 ) 2 p + 1 = − 1 . The symmetric summands are therefore either those with p ≡ r ( mod 4 ) or those with p ≡ r − 2 ( mod 4 ) . Using 2 r p = 2 r 2 r − p and summing over the fourth roots of unity, the first set has total dimension
1 2 ∑ 0 ≤ p ≤ 2 r p ≡ r ( 4 ) 2 r p = 1 8 ∑ ω 4 = 1 ω − r ( 1 + ω ) 2 r = 1 8 2 2 r + 2 r + 1 = 2 2 r − 3 + 2 r − 2 ,
because i − r ( 1 + i ) 2 r = i r ( 1 − i ) 2 r = 2 r and the term with ω = − 1 vanishes. This is N ( N + 1 ) / 2 for N = 2 r − 1 , while the second set has dimension N ( N − 1 ) / 2 . Hence the symmetric summands are those with p ≡ r ( mod 4 ) .    □
Appendix C lists the decomposition for the families (2). For ( 9 + 8 n ) + 1 one has r = 5 + 4 n , and Sym 2 Δ + consists of the forms of degrees 1 , 5 , … , 5 + 4 n ; for ( 10 + 8 n ) + 2 the degrees are 2 , 6 , … , 6 + 4 n ; for ( 11 + 8 n ) + 3 they are 3 , 7 , … , 7 + 4 n ; and for ( 12 + 8 n ) + 4 they are 0 , 4 , … , 8 + 4 n . These are exactly the right-hand sides of eqs. (3.5)–(3.8) of Ref. [23]. The superalgebras proposed there are therefore maximal central supertranslation algebras: the anticommutator of two supercharges is an arbitrary element of Sym 2 Δ + , with a central charge for every summand. Such an algebra exists for every Weyl spinor in every signature, and its existence alone does not establish a field-theoretic realization.2 It contains the momentum only for r ≡ 1 ( mod 4 ) .

2.3. Isotropic Subspaces

A subspace U of V or of V C is totally isotropic if u · u ′ = 0 for all u , u ′ ∈ U .
Lemma 2.2. 
A real totally isotropic subspace U ⊂ R s , t has dim U ≤ min ( s , t ) , and the bound is attained.3 
Proof. 
Write V = V + ⊕ V − with V + positive definite of dimension s and V − negative definite of dimension t. If u ∈ U ∩ V − , then u · u = 0 forces u = 0 , so the projection U → V + is injective and dim U ≤ s . Exchanging the roles gives dim U ≤ t . With orthonormal bases e 1 , … , e s of V + and f 1 , … , f t of V − , the vectors e i + f i , i ≤ min ( s , t ) , span a totally isotropic subspace of dimension min ( s , t ) .    □
The next lemma says that a Majorana spinor cannot distinguish a complex null vector from its complex conjugate.
Lemma 2.3. 
Let s − t ≡ 0 , 1 , 2 ( mod 8 ) and let λ ≠ 0 be a Majorana spinor. The set U λ = { u ∈ V C : u λ = 0 } is a complex totally isotropic subspace invariant under complex conjugation. Hence U λ = U R ⊗ C with U R ⊂ V real and totally isotropic, and dim C U λ ≤ min ( s , t ) . 
Proof. 
If u λ = 0 and * * u * * ′ λ = 0 , then 2 ( u · u ′ ) λ = ( u * * u * * ′ + * * u * * ′ u ) λ = 0 , so u · u ′ = 0 . Since C is antilinear and commutes with the γ μ , C ( u λ ) = u ¯ μ γ μ C λ = u ¯ λ , so u λ = 0 implies u ¯ λ = 0 . A conjugation-invariant complex subspace is the complexification of its real points, and the real points of a totally isotropic subspace form a totally isotropic subspace of V. Lemma 2.2 finishes the proof.    □
Figure 3 illustrates the lemma in R 2 , 1 , where min ( s , t ) = 1 : two independent real null vectors are never orthogonal.

3. Vector Structures and the Cubic Identity

Every condition on cubic fermion terms that appears in this paper is a statement about the following structure.
Definition 3.1. 
Let W be a complex vector space of dimension d with a nondegenerate symmetric bilinear form η, and let Σ be a complex vector space of dimension N ≥ 1 with dual Σ * . Avector structureon Σ is a pair of linear maps, γ from W to the symmetric maps Σ → Σ * and γ ˜ from W to the maps Σ * → Σ , such that in a basis of W
γ ˜ μ γ ν + γ ˜ ν γ μ = 2 η μ ν 1 Σ , γ μ γ ˜ ν + γ ν γ ˜ μ = 2 η μ ν 1 Σ * .
In a basis e α of Σ we write γ α β μ and γ ˜ μ α β for the matrix entries; γ α β μ = γ β α μ . 
The standard example is Σ = Δ + in D = 2 r with r odd, Σ * ≅ Δ − through B, γ α β μ = B ( e α , γ μ e β ) , and γ ˜ μ the Clifford action Δ − → Δ + ; by Proposition 2.1 the forms are symmetric when r ≡ 1 ( mod 4 ) . In ten dimensions these are the familiar 16 × 16 matrices σ α β μ and σ ¯ μ α β  [41].
Lemma 3.2. 
Let ( Σ , γ , γ ˜ ) be a vector structure with d ≥ 3 . Then
(a)
each γ ˜ μ is symmetric, and in an orthonormal basis γ ˜ μ = η μ μ ( γ μ ) − 1 (no sum); 
(b)
γ μ γ ˜ μ = d 1 Σ * and γ μ γ ˜ ν γ μ = − ( d − 2 ) γ ν ; 
(c)
tr Σ * ( γ μ γ ˜ ν ) = γ α β μ γ ˜ ν β α = N η μ ν . 
Proof. (a) Setting μ = ν in (5) gives γ μ γ ˜ μ = η μ μ 1 Σ * and γ ˜ μ γ μ = η μ μ 1 Σ , and the inverse of a symmetric map is symmetric. (b) Contracting the second relation in (5) gives the first identity, and then γ μ γ ˜ ν γ μ = γ μ ( 2 η ν μ − γ ˜ μ γ ν ) = 2 γ ν − d γ ν . (c) On Σ ⊕ Σ * the operators Γ μ = 0 γ ˜ μ γ μ 0 obey Γ μ Γ ν + Γ ν Γ μ = 2 η μ ν , and Γ * = diag ( 1 , − 1 ) anticommutes with every Γ μ . In an orthonormal basis take μ ≠ ν and choose ρ ∉ { μ , ν } , which is possible because d ≥ 3 . The invertible operator Γ ρ commutes with Γ μ Γ ν and anticommutes with Γ * Γ μ Γ ν , so tr ( Γ * Γ μ Γ ν ) = 0 ; and tr ( Γ μ Γ ν ) = − tr ( Γ ν Γ μ ) = 0 . The restriction of Γ μ Γ ν = diag ( γ ˜ μ γ ν , γ μ γ ˜ ν ) to Σ * is γ μ γ ˜ ν , and its trace 1 2 tr ( 1 − Γ * ) Γ μ Γ ν therefore vanishes for μ ≠ ν . For μ = ν it is N η μ μ by (a).    □
The condition that absorption of cubic terms will produce is the following. For a constant j ∈ C consider
1 2 γ α β μ γ μ γ δ + γ α γ μ γ μ β δ = j γ α δ σ γ σ β γ for all α , β , γ , δ .
For j = − 1 2 this is the identity (1).
Theorem 3.3. 
Let ( Σ , γ , γ ˜ ) be a vector structure with d ≥ 3 , and suppose that (6) holds for some j ∈ C . Then
N = 2 ( d − 2 ) , j = − 1 2 ,
so the identity is (1). Moreover N is divisible by 2 ⌈ d / 2 ⌉ − 1 , and therefore ( d , N ) ∈ { ( 3 , 2 ) , ( 4 , 4 ) , ( 6 , 8 ) , ( 10 , 16 ) } . 
Proof. 
Contract (6) with γ ˜ ν δ γ , which is symmetric by Lemma 3.2(a), and sum over γ , δ . Inside the bracket on the left, the first term gives γ α β μ tr ( γ μ γ ˜ ν ) = N γ α β ν and the second gives ( γ μ γ ˜ ν γ μ ) α β = − ( d − 2 ) γ α β ν , by Lemma 3.2(b),(c). The right-hand side gives j ( γ σ γ ˜ ν γ σ ) α β = − j ( d − 2 ) γ α β ν . Since γ ν ≠ 0 ,
N − ( d − 2 ) = − 2 j ( d − 2 ) .
Next contract with γ ˜ ν β γ and sum over β , γ . Both terms in the bracket give ( γ μ γ ˜ ν γ μ ) α δ = − ( d − 2 ) γ α δ ν , and the right-hand side gives j γ α δ σ tr ( γ σ γ ˜ ν ) = j N γ α δ ν , so
j N = − ( d − 2 ) .
Substituting j = − ( d − 2 ) / N into (8) gives N 2 − ( d − 2 ) N − 2 ( d − 2 ) 2 = 0 , that is ( N − 2 ( d − 2 ) ) ( N + d − 2 ) = 0 . As N + d − 2 > 0 , we get N = 2 ( d − 2 ) and j = − 1 2 .
For the divisibility, the operators Γ μ and Γ * of the proof of Lemma 3.2(c) make Σ ⊕ Σ * a module over the complex Clifford algebra on d + 1 generators. This algebra is semisimple and its irreducible modules have dimension 2 ⌊ ( d + 1 ) / 2 ⌋ = 2 ⌈ d / 2 ⌉ , so 2 ⌈ d / 2 ⌉ divides 2 N = 4 ( d − 2 ) . For d = 5 , 7 , 8 , 9 it does not, and for d ≥ 11 one has 4 ( d − 2 ) < 2 ⌈ d / 2 ⌉ . This leaves d = 3 , 4 , 6 , 10 .    □
Remark 3.4. 
The four solutions are the Majorana spinors of three and four dimensions, the Weyl spinors of six (with their S p ( 1 ) doublet) and the Majorana–Weyl spinors of ten, and in each case (1) holds [2,3,4]. Theorem 3.3 is a converse that makes no reference to division algebras and allows a general constant j: the dimension count alone fixes both N and j. The constant j is left free because absorbing the cubic terms into a modified field equation, the mechanism of Nishino’s construction, produces (6) with a constant that is not known in advance (Section 5). Figure 4 compares 2 ( d − 2 ) with the smallest admissible N. 

4. Ordinary Super Yang–Mills Theory

Before turning to null projections we record what Theorem 3.3 gives without them. Let the gauge algebra g be a compact non-abelian Lie algebra with structure constants f I J K , totally antisymmetric in an orthonormal basis. Let the gaugino λ I and the parameter ϵ be Grassmann-odd Weyl spinors in Δ + (Majorana–Weyl when s − t ≡ 0 ( mod 8 ) ), and consider
L = − 1 4 F μ ν I F I μ ν + c 2 λ ¯ I γ μ D μ λ I , δ A μ I = a ϵ ¯ γ μ λ I , δ λ I = b γ μ ν F μ ν I ϵ ,
with D μ λ I = ∂ μ λ I + g f I J K A μ J λ K and nonzero constants a , b , c , g .
Proposition 4.1. 
Assume D > 2 and that the gaugino term in (10) is a nondegenerate kinetic term, rather than zero or a total derivative. If δ L is a total derivative, then the forms γ α β μ = B ( e α , γ μ e β ) on Δ + define a vector structure satisfying (1). Consequently D = 10 , and with Majorana–Weyl spinors the signature is ( 9 , 1 ) , ( 5 , 5 ) or ( 1 , 9 ) . 
Proof. 
The kinetic term is nonzero only if B pairs Δ + with γ μ Δ + = Δ − , that is for r odd, and it is not a total derivative only if the form γ α β μ is symmetric: for Grassmann-odd λ and antisymmetric G, λ α G α β ∂ λ β = 1 2 ∂ ( λ α G α β λ β ) . With Σ = Δ + , Σ * ≅ Δ − and γ ˜ μ the Clifford action Δ − → Δ + , the relations (5) are those of Cl ( V ) , so ( Δ + , γ , γ ˜ ) is a vector structure with d = D and N = 2 r − 1 .
The terms of δ L cubic in λ come only from δ A inside the covariant derivative, since δ λ is linear in F and δ F is linear in λ . They are
c 2 a g f I J K ( λ ¯ I γ μ λ K ) ( ϵ ¯ γ μ λ J ) = ± c 2 a g γ α γ μ γ μ δ β ϵ δ T α β γ , T α β γ = f I J K λ I α λ J β λ K γ .
where the sign depends only on the ordering convention for Grassmann variables. They contain no derivative, so they must vanish. The tensor T is totally symmetric: exchanging two of its indices and relabelling the gauge indices produces one sign from f and one from the anticommuting λ . It spans Sym 3 Σ . Indeed g contains a subalgebra su ( 2 ) , and for λ with values in it T α β γ c α β γ = 6 c α β γ λ 1 α λ 2 β λ 3 γ for every symmetric c, after normalizing f 123 = 1 . Since the monomials λ 1 α λ 2 β λ 3 γ are linearly independent, only c = 0 annihilates every T. Hence, for every δ , the part of γ α γ μ γ μ δ β symmetric in α , β , γ vanishes. Because each factor is symmetric in its own pair of indices, that symmetrization is one third of the sum over the three ways of pairing δ with α , β or γ , which is (1). Theorem 3.3 with N = 2 r − 1 and d = 2 r gives 2 r − 1 = 4 ( r − 1 ) , whose only solution with r ≥ 2 is r = 5 . Majorana–Weyl spinors in D = 10 require s − t ∈ { 8 , 0 , − 8 } .    □
The proposition is the familiar statement that ten is the largest dimension of super Yang–Mills theory [1], obtained here from the dimension count of Theorem 3.3 rather than from an explicit Fierz rearrangement. For D > 2 , changing the signature does not evade this conventional nondegenerate-kinetic-term ansatz, because the proof uses only D. The degenerate two-dimensional case is not covered by the proposition. The higher-dimensional constructions examined here therefore constrain the gaugino.

5. Null-Projected Super Yang–Mills Theories

5.1. Null Frames as Fermionic Oscillators

Definition 5.1. 
Anull frame of rank kis a set of vectors n 1 , … , n k , m 1 , … , m k ∈ V C with
n i · n j = 0 , m i · m j = 0 , n i · m j = δ i j .
It isrealif all the vectors lie in V. Itsresidual spaceis the orthogonal complement V ′ of span ( n i , m i ) , of dimension D ′ = D − 2 k , and we set v = n 1 ∧ … ∧ n k . 
The span of the n i and m i is a sum of k hyperbolic planes, so for a real frame V ′ is real of signature ( s − k , t − k ) . By Lemma 2.2, a real null frame of rank k exists exactly when k ≤ min ( s , t ) . In Nishino’s notation n i = n ( i ) and m i = m ( i ) ; his signatures ( 9 + m ) + ( 1 + m ) carry real frames of rank m with residual space R 9 , 1 .
The relations (12) say that the operators
b i = 1 2 * * m * * i , b i † = 1 2 * * n * * i
obey the canonical anticommutation relations { b i , b j † } = δ i j , { b i , b j } = { b i † , b j † } = 0 . (The dagger is notation; no adjoint is implied.) The number operators
P i ↑ = b i † b i = 1 2 * * n * * i * * m * * i , P i ↓ = b i b i † = 1 2 * * m * * i * * n * * i = 1 − P i ↑
are Nishino’s projectors [22], and they commute with each other. For a ∈ { 0 , 1 } k let Δ a be the joint eigenspace on which P i ↑ = a i ; mode i is occupied in Δ a if a i = 1 . Write 1 = ( 1 , … , 1 ) , 0 = ( 0 , … , 0 ) , a ¯ = 1 − a , and e i for the i-th unit vector of { 0 , 1 } k . For vectors of V ′ we use primed indices, γ μ ′ .
Lemma 5.2. 
(i)
Δ = ⨁ a Δ a . Each γ μ ′ preserves every Δ a , and each Δ a is a Dirac module of Cl ( V C ′ ) , of dimension 2 r − k . 
(ii)
* * n * * i maps Δ a with a i = 0 isomorphically onto Δ a + e i and annihilates Δ a with a i = 1 ; * * m * * i does the reverse. Hence
K : = ⋂ i ker * * n * * i = Δ 1 , K ^ : = ⋂ i ker * * m * * i = Δ 0 .
(iii)
T = * * n * * 1 * * n * * 2 … * * n * * k maps K ^ isomorphically onto K, annihilates Δ a for a ≠ 0 , and satisfies T γ μ ′ = ( − 1 ) k γ μ ′ T ; in particular T commutes with Spin ( V ′ ) . 
(iv)
B ( Δ a , Δ b ) = 0 unless b = a ¯ . 
(v)
On each Δ a the chirality Γ * equals ± ( − 1 ) | a | Γ * ′ , where Γ * ′ is the chirality of V ′ , | a | is the number of occupied modes, and the overall sign does not depend on a. Thus Δ a ∩ Δ + and Δ a ∩ Δ − are the two Weyl modules of V ′ . 
The proof is in Appendix A. Figure 5 shows the case k = 3 . The kernel K of all the * * n * * i is the fully occupied corner, the kernel K ^ of all the * * m * * i is the empty corner, T climbs from one to the other, and B pairs opposite corners.

5.2. The Axioms

Definition 5.3. 
Anull-projected super Yang–Mills theory of rank kin signature ( s , t ) , D = 2 r , consists of a null frame of rank k, a compact non-abelian gauge algebra g , a gauge field A μ I , a Grassmann-odd gaugino λ I and Grassmann-odd supersymmetry parameters ϵ, subject to the following conditions.
(N1)
Constraint.The complexified gaugino value space is the full corner K χ = K ∩ Δ χ for a fixed chirality χ; λ I takes values there, that is, * * n * * i λ I = 0 for all i. The parameter ϵ ranges over a Weyl module Δ χ ′ .4 
(N2)
Transformations. δ A μ I = a ϵ ¯ γ μ λ I with a ≠ 0 , and δ λ I = Φ ( F I ) ϵ for a linear map Φ from two-forms to End Δ that does not depend on λ. The covariant derivative is D μ λ I = ∂ μ λ I + g f I J K A μ J λ K with g ≠ 0 . 
(N3)
Residual gauge field.The components δ A μ ′ I along V ′ do not vanish identically. 
(N4)
Closure.The commutator of two supersymmetries contains the translation by
ξ μ = c ϵ ¯ 2 γ μ ν 1 … ν k ϵ 1 v ν 1 … ν k , c ≠ 0 ,
and its components ξ μ ′ along V ′ do not vanish identically. The displayed translation bracket is the coefficient in the supersymmetry anticommutator, so its matrix in the commuting spinor indices is symmetric. This graded symmetry is part of the closure requirement. 
(N)
Absorption.For every λ with values in K χ ⊗ g there are J σ I , depending on λ but not on ϵ, such that for every ϵ the part of δ ( γ μ D μ λ I ) quadratic in λ equals
γ σ ν 1 … ν k ϵ v ν 1 … ν k J σ I .
Axiom e is what makes the theory consistent: a term of the form (17) is cancelled by a source term in the field equation of the gauge field, which becomes D μ F μ [ σ v ν 1 … ν k ] ∝ J [ σ v ν 1 … ν k ] . For k = 0 the axiom says that the cubic terms have the form γ σ ϵ J σ .
Example 5.4. 
Nishino’s theories [22] in ( 9 + m ) + ( 1 + m ) satisfy a–e with k = m . His constraint (2.12), the contraction of v with γ μ λ , is equivalent to * * n * * ( i ) λ = 0 for every i, and he rewrites it as λ = P ↑ λ in Equation (2.22). The transformations b are his Equation (2.3), with Φ ( F ) ϵ = 1 4 γ μ ν ρ 1 … ρ m ϵ F μ ν v ρ 1 … ρ m , and the closure d is his eqs. (2.16)–(2.18). Axiom e is his Equation (2.29): the surviving cubic term is γ [ m + 1 ] ϵ contracted with the bilinear λ ¯ K γ [ m + 1 ] λ J , and for gauginos in K this ( m + 1 ) -form is the wedge product of v with a residual vector (Lemma A1), so the term has the form (17); it is absorbed by the source in his field equation (2.15).5 The residual space is R 9 , 1 , so D ′ = 10 . 
The variation of the gaugino equation follows directly from the axioms.
Lemma 5.5. 
Under b, the part of δ ( γ μ D μ λ I ) quadratic in λ is a g f I J K γ μ λ K ( ϵ ¯ γ μ λ J ) . 
Proof. 
δ ( D μ λ I ) = D μ ( δ λ I ) + g f I J K δ A μ J λ K . The first term is independent of λ because Φ is, and the second is a g f I J K ( ϵ ¯ γ μ λ J ) λ K .    □

5.3. Reduction to the Residual Space

Theorem 5.6. 
In a null-projected super Yang–Mills theory of rank k the residual dimension D ′ = D − 2 k equals 10 or 2. 
Proof. 
D ′ is even, and D ′ ≥ 1 because c requires residual directions. Assume D ′ ≥ 3 ; we prove D ′ = 10 . The proof has three steps.
Step 1: projection onto the occupied corner. Since n i · m j = δ i j , the inverse metric decomposes as η − 1 = η ′ − 1 + ∑ i ( n i ⊗ m i + m i ⊗ n i ) , so
γ μ λ ( ϵ ¯ γ μ λ ′ ) = γ μ ′ λ ( ϵ ¯ γ μ ′ λ ′ ) + ∑ i * * n * * i λ ( ϵ ¯ * * m * * i λ ′ ) + * * m * * i λ ( ϵ ¯ * * n * * i λ ′ ) .
For λ , λ ′ ∈ K the bracket vanishes, because * * n * * i annihilates K. By Lemma 5.5 the left-hand side of e is therefore
a g f I J K γ μ ′ λ K ( ϵ ¯ γ μ ′ λ J ) ,
which lies in the occupied corner Δ 1 because γ μ ′ preserves every Δ a . On the right-hand side of (17), write γ ( ω ) for the antisymmetrized product of gamma matrices attached to a ( k + 1 ) -vector ω , normalized so that γ ( u 0 ∧ … ∧ u k ) = * * u * * 0 … * * u * * k for mutually orthogonal u i . Up to a nonzero factor, which we absorb into J, the right-hand side is γ ( J ∧ v ) ϵ . Decompose J = J ′ + ∑ i [ ( J · n i ) m i + ( J · m i ) n i ] with J ′ ∈ V C ′ . Since n i ∧ v = 0 , the vector u ′ ∈ V C ′ is orthogonal to all the n l , and m i is orthogonal to all n l with l ≠ i ,
γ ( u ′ ∧ v ) = * * u * * ′ T , γ ( n i ∧ v ) = 0 , γ ( m i ∧ v ) = ± ∏ l ≠ i * * n * * l ( P i ↓ − P i ↑ ) ,
where the last formula uses 1 2 ( * * m * * i * * n * * i − * * n * * i * * m * * i ) = P i ↓ − P i ↑ . Now take ϵ ∈ K ^ χ ′ = K ^ ∩ Δ χ ′ . Then γ μ ′ T ϵ ∈ Δ 1 , while ( ∏ l ≠ i * * n * * l ) ( P i ↓ − P i ↑ ) ϵ lies in Δ 1 − e i , in which mode i is empty. Applying the projection ∏ i P i ↑ onto Δ 1 to both sides of e, we obtain for all ϵ ∈ K ^ χ ′
a g f I J K γ μ ′ λ K ( ϵ ¯ γ μ ′ λ J ) = γ μ ′ T ϵ J μ ′ I .
Step 2: the residual vector structure. By Lemma 5.2(iii), x = T ϵ ranges over K χ ′ ′ with χ ′ ′ = ( − 1 ) k χ ′ as ϵ ranges over K ^ χ ′ . Define
β ( x , y ) = B ( T − 1 x , y ) ( x ∈ K χ ′ ′ , y ∈ Δ 1 ) .
By Lemma 5.2(iv) and the nondegeneracy of B, β is a nondegenerate pairing of K χ ′ ′ with the Weyl half of Δ 1 that B pairs with K ^ χ ′ , and it is Spin ( V ′ ) -invariant by Lemma 5.2(iii). In this notation
δ A μ ′ I = a β ( T ϵ , γ μ ′ λ I ) , ξ μ ′ = c β ( T ϵ 2 , γ μ ′ T ϵ 1 ) ( ϵ , ϵ 1 , ϵ 2 ∈ K ^ χ ′ ) ,
the first because only the K ^ -component of ϵ pairs with γ μ ′ λ ∈ Δ 1 , the second by (20), with the normalization of v absorbed into c. Both residual bilinears depend only on the K ^ components of their supersymmetry parameters: T kills every other corner and B pairs opposite corners. Their nonvanishing in (N3) and (N4) therefore survives this restriction. Thus G μ ′ ( x , y ) = β ( x , γ μ ′ y ) is a nonzero equivariant pairing K χ ′ ′ × K χ ′ ′ → V C ′ by d, and a nonzero equivariant pairing K χ ′ ′ × K χ → V C ′ by c. By Lemma 5.2(v), K χ ′ ′ and K χ are Weyl modules of V ′ . With D ′ = 2 r ′ , Proposition 2.1 and its analogue for opposite chiralities, Δ + ′ ⊗ Δ − ′ ≅ ⨁ j Λ r ′ − 1 − 2 j , show that the tensor square of a Weyl module contains Λ 1 only for r ′ odd, and the product of the two Weyl modules only for r ′ even. Hence r ′ is odd and K χ ′ ′ and K χ have the same V ′ -chirality. Both lie in Δ 1 , so by Lemma 5.2(v) they coincide: K χ ′ ′ = K χ = : Σ .
Next, G μ ′ is symmetric on Σ . The commutator of two supersymmetries is antisymmetric under ϵ 1 ↔ ϵ 2 , and so is the parameter ξ of the translation in it. For Grassmann-odd x i = T ϵ i one has x 2 α G α β x 1 β = − x 1 β G α β x 2 α , so in (23) the symmetric part of G contributes a term that is odd under the exchange and the antisymmetric part a term that is even. The even term must vanish, and since x 1 , x 2 are independent Grassmann variables the antisymmetric part of G is zero.6 By d, G ≠ 0 . So G is a nonzero symmetric equivariant map Sym 2 Σ → V C ′ , and Proposition 2.1 gives r ′ ≡ 1 ( mod 4 ) . Let Σ opp be the opposite Weyl half of Δ 1 . The nondegenerate pairing defines an isomorphism
ι : Σ opp ⟶ Σ * , ι ( z ) ( x ) = β ( x , z ) .
Writing γ amb μ ′ for the ambient Clifford action, define
γ μ ′ = ι ∘ γ amb μ ′ | Σ , γ ˜ μ ′ = γ amb μ ′ | Σ opp ∘ ι − 1 .
The matrix of the first map is G μ ′ , hence symmetric. In both compositions of (5), ι − 1 ι cancels and the ambient Clifford relations apply. Thus ( Σ , γ , γ ˜ ) is a vector structure on W = V C ′ with d = D ′ and N = dim Σ = 2 r ′ − 1 .
Step 3: the identity. Write x = T ϵ and S I β δ = f I J K λ J β λ K δ . Exchanging β and δ and relabelling J ↔ K gives one sign from f and one from the anticommuting λ , so S I is symmetric. In components, (21) reads
κ γ α δ μ ′ γ μ ′ γ β x γ S I β δ = γ α γ μ ′ x γ J μ ′ I ,
with κ = ± a g ≠ 0 , the sign fixed by the ordering convention. Since (24) holds for all x, it is an identity between the matrices that multiply x γ . Contracting with γ ˜ ν γ α and using Lemma 3.2(c),
J I ν = κ N γ ˜ ν γ α γ α δ μ ′ γ μ ′ γ β S I β δ .
The current is determined explicitly, without a multiplicity argument. By the symmetry of γ μ ′ and Lemma 3.2(b),
γ ˜ ν γ α γ α δ μ ′ γ μ ′ γ β = ( γ μ ′ γ ˜ ν γ μ ′ ) β δ = − ( D ′ − 2 ) γ β δ ν .
Consequently
J I ν = κ j γ β δ ν S I β δ , j = − D ′ − 2 N .
This is a linear function of S I fixed by the absorption equation itself. Substituting into (24),
γ α δ μ ′ γ μ ′ γ β − j γ α γ μ ′ γ μ ′ β δ S I β δ = 0 .
The tensors S I span Sym 2 Σ : for λ with values in an su ( 2 ) ⊂ g , S 3 = λ 1 ⊗ λ 2 + λ 2 ⊗ λ 1 up to normalization, and the monomials λ 1 β λ 2 δ are independent.7 Hence the part of the bracket in (26) symmetric in β , δ vanishes:
1 2 γ α δ μ ′ γ μ ′ γ β + γ α β μ ′ γ μ ′ γ δ = j γ α γ μ ′ γ μ ′ β δ .
Exchanging the names of γ and δ and using the symmetry of γ μ ′ gives exactly (6). Theorem 3.3 now gives 2 r ′ − 1 = N = 2 ( D ′ − 2 ) = 4 ( r ′ − 1 ) , whose only solution is r ′ = 5 , that is D ′ = 10 .    □
Step 1 is where the transplanted constructions fail. In Nishino’s signatures the occupied corner K is a sixteen-dimensional Weyl module of the residual R 9 , 1 , and (21) holds by the ten-dimensional identity (1): contracting (1) with a symmetric S β δ gives γ α δ μ ′ γ μ ′ γ β S β δ = − 1 2 γ α γ μ ′ γ μ ′ β δ S β δ , so J μ ′ is a multiple of the residual fermion current, the source that Nishino places in the field equation (2.15) of the gauge field. Setting the non-residual components of J to zero extends this absorption solution from ϵ ∈ K ^ to every parameter: the left-hand side depends only on its K ^ component, and the right-hand side is γ ( J ∧ v ) ϵ = J T ϵ , which has the same dependence. This solves the absorption condition itself; it supplies no independent construction of the transformation Φ or of full closure. In ( 9 + 8 n + q ) + ( 1 + q ) with n ≥ 1 the frame has at most 1 + q null vectors, so the residual space has dimension at least s − t = 8 + 8 n ≥ 16 , and its Weyl spinors are too large for any identity (6).
Remark 5.7. 
The case D ′ = 2 escapes the argument because Lemma 3.2(c) needs d ≥ 3 . There Σ is one-dimensional, γ μ ′ is a null vector, and (6) holds trivially because both sides vanish. The residual gauge field in two dimensions has no local transverse gauge-boson polarizations; this does not by itself determine the propagation of every other field. By Theorem 5.8(iii) below it can occur only for s = t . 

5.4. The Main Theorem

Theorem 5.8 
(Main Theorem). Let a null-projected super Yang–Mills theory of rank k in signature ( s , t ) have a Majorana–Weyl gaugino, so that s − t ≡ 0 ( mod 8 ) . Then
(i)
k ≤ min ( s , t ) , and n 1 , … , n k lie in the complexification of a real totally isotropic subspace of V; 
(ii)
s + t − 2 k ∈ { 2 , 10 } ; 
(iii)
s − t ∈ { 8 , 0 , − 8 } , and s = t if s + t − 2 k = 2 . 
Proof. (i) The values of a Majorana–Weyl gaugino are real spinors in K χ , and by c the gaugino does not vanish identically, so K χ contains a Majorana spinor ψ ≠ 0 . By Lemma 2.3, U ψ ⊃ span ( n 1 , … , n k ) is the complexification of a real totally isotropic subspace U R , so k ≤ dim U R ≤ min ( s , t ) . (ii) is Theorem 5.6. (iii) From (i) and (ii), | s − t | = s + t − 2 min ( s , t ) ≤ s + t − 2 k ≤ 10 , and s − t ≡ 0 ( mod 8 ) leaves s − t ∈ { 8 , 0 , − 8 } ; if s + t − 2 k = 2 , then | s − t | ≤ 2 forces s = t .    □
Corollary 5.9. 
There is no null-projected super Yang–Mills theory with a Majorana–Weyl gaugino in any of the signatures (2) with n ≥ 1 . In particular there is none in 17 + 1 , 19 + 3 , 25 + 1 or 27 + 3 . 
Proof. 
In the families (2), s − t = 8 + 8 n ≥ 16 for n ≥ 1 .    □
Remark 5.10. 
The reality of the gaugino is essential for (i). With a complex Weyl gaugino a complex frame of rank k > min ( s , t ) is allowed, and Theorem 5.6 still gives D ′ = 10 when D ′ > 2 . For instance ( 17 , 1 ) admits a complex frame of rank four with residual dimension ten, and the absorption equation (21) can then be solved, as the numerical check of Appendix B confirms. But the occupied corner of such a frame contains no Majorana spinor (Lemma 2.3), so this projected sector has no Majorana real form compatible with the original real signature and frame. Its residual Clifford algebra is the ten-dimensional complex one. The superalgebras of Ref. [23] are built on Majorana–Weyl generators. 
Remark 5.11. 
For odd D there are no Weyl spinors, so there are no Majorana–Weyl gauginos and the Main Theorem has no odd-dimensional counterpart; the families (2) all have even D. Theorem 3.3 itself does not assume that d is even. 
Figure 6 summarizes the numerical test of Step 1 in explicit gamma matrices, described in Appendix B: the absorption equation (21) has a solution to machine precision in the tested cases with D ′ = 10 , and large residuals in the other tested cases.

5.5. Sufficiency for Residual Dimension Ten

Theorem 5.8 gives necessary conditions. When the residual dimension is ten they are also sufficient. Nishino’s theories carry over to every signature with the same total dimension and the same rank of the frame, as long as the frame stays real, and these are exactly the signatures with residual dimension ten that Theorem 5.8 leaves.
Theorem 5.12. 
Let s + t = 10 + 2 k with k ≥ 0 and s − t ∈ { 8 , 0 , − 8 } . Then R s , t carries a real null frame of rank k with residual space R 9 , 1 , R 5 , 5 or R 1 , 9 . Nishino’s equations for ( 9 + k ) + ( 1 + k ) (Example 5.4), written with this frame, define a null-projected super Yang–Mills theory of rank k in ( s , t ) with a Majorana–Weyl gaugino. It satisfies a–e, and its supersymmetry transformations close as in Nishino’s eqs. (2.16)–(2.18). For s − t = 8 it is Nishino’s theory, and for k = 0 it is ordinary super Yang–Mills theory in ( 9 , 1 ) , ( 5 , 5 ) or ( 1 , 9 ) . 
Proof. 
Since | s − t | ≤ 8 , min ( s , t ) = 1 2 ( s + t − | s − t | ) ≥ k + 1 , so a real null frame of rank k exists by Lemma 2.2, and its residual space has signature ( s − k , t − k ) , with s − k + t − k = 10 and ( s − k ) − ( t − k ) = s − t .
Let V 0 = R 9 + k , 1 + k with Nishino’s frame ( n i 0 , m i 0 ) and V = R s , t with a real frame ( n i , m i ) . The complexified residual spaces are nondegenerate quadratic spaces of dimension ten, so there is a complex isometry between them, and sending n i 0 ↦ n i , m i 0 ↦ m i extends it to a complex isometry φ : V 0 , C → V C , by (12). It induces an isomorphism of the complex Clifford algebras and, since the Dirac module is unique, a linear isomorphism Φ : Δ 0 → Δ with Φ γ ( u ) = γ ( φ u ) Φ . Labelling the chiralities of V so that Φ preserves them, Φ maps Weyl modules to Weyl modules and the corners K, K ^ of the first frame onto those of the second, and it carries B 0 to a nonzero multiple of B, because a form with the property (3) for a given θ is unique up to a factor. Nishino’s constraint, transformation rules, field equations and closure relations are polynomial identities in the gamma matrices, the metric, the frame, the structure constants and the components of the fields and parameters. The maps φ and Φ preserve all of them. Hence the same equations, written in V with the frame ( n i , m i ) , satisfy every identity that they satisfy in V 0 , and a–e hold for them over C , with D ′ = 10 .
It remains to impose reality. Since s − t ≡ 0 ( mod 8 ) , Δ carries a real structure C that commutes with every γ μ and preserves Δ ± (Section 2). As the n i are real, C commutes with every * * n * * i and preserves K, so the Majorana elements of K χ form a real form of K χ ; this is a with a Majorana–Weyl gaugino. The form B ′ ( ψ , χ ) = B ( C ψ , C χ ) ¯ again satisfies (3) with the same θ , so B ′ = c B for a constant c, and applying the construction twice gives | c | = 1 . Replacing B by c 1 / 2 B , we may assume that B ( ψ , χ ) is real for Majorana spinors ψ , χ . Then δ A = a ϵ ¯ γ λ is real for real a, and δ λ is a Majorana spinor because the gamma matrices commute with C and F and v are real. The real fields therefore form a consistent sector of the transported theory.    □
Corollary 5.13. 
Let s − t ≡ 0 ( mod 8 ) and let a real null frame of rank k have residual dimension D ′ = s + t − 2 k ≥ 3 . A null-projected super Yang–Mills theory of rank k with a Majorana–Weyl gaugino exists in ( s , t ) if and only if D ′ = 10 and s − t ∈ { 8 , 0 , − 8 } . In particular such theories exist in the signatures 9 + 9 (with k = 4 ) and 9 + 17 (with k = 8 ) of the second degenerate family 9 + ( 1 + 8 n ) of Ref. [23], and in no member of that family with n ≥ 3 . 
Proof. 
Necessity is Theorem 5.8, since D ′ ≥ 3 excludes D ′ = 2 . Sufficiency is Theorem 5.12. In 9 + ( 1 + 8 n ) one has s − t = 8 − 8 n , which lies in { 0 , − 8 } for n = 1 , 2 , and D ′ = 10 gives k = ( s + t − 10 ) / 2 .    □
The theorem shows what a transfer of Nishino’s construction must keep. A complex isometry keeps the total dimension D and the rank k, and the transported theory stays real when the frame can be chosen real, k ≤ min ( s , t ) , and Majorana–Weyl spinors exist, s − t ≡ 0 ( mod 8 ) . With D ′ = 10 these conditions force s − t ∈ { 8 , 0 , − 8 } . In thirty dimensions, for example, Nishino’s rank k = 10 in 19 + 11 carries over to 15 + 15 and 11 + 19 , and to no signature with three times. Figure 7 shows the signatures that carry these theories. For residual dimension two, which Theorem 5.8 allows when s = t , no ten-dimensional theory can be transported, and Theorem 5.14 below constructs the theory directly.

5.6. Residual Dimension Two

The Main Theorem leaves a second case, s + t − 2 k = 2 with s = t , so the signatures ( k + 1 ) + ( k + 1 ) with a frame of rank k. Here no ten-dimensional theory can be transported, and we build the theory directly. The residual plane V ′ ≅ R 1 , 1 is spanned by two real null vectors ℓ + , ℓ − with η ( ℓ + , ℓ − ) = 1 . Since γ ( ℓ ± ) 2 = 0 and γ ( ℓ + ) γ ( ℓ − ) + γ ( ℓ − ) γ ( ℓ + ) = 2 , the two Weyl modules of a Dirac module of Cl ( V C ′ ) are ker γ ( ℓ + ) and ker γ ( ℓ − ) , and each is one-dimensional. We label ℓ ± so that the gaugino space Σ = K χ is ker γ ( ℓ + ) (Lemma 5.2(i),(v)), and we take ϵ of the chirality χ ′ for which T K ^ χ ′ = Σ , so that x = T ϵ ranges over Σ as ϵ ranges over K ^ χ ′ (Lemma 5.2(iii)). The pairing β of (22) is Spin ( V ′ ) -invariant and nondegenerate, and Spin ( 1 , 1 ) acts on the two Weyl modules with opposite weights, so β pairs Σ with γ ( ℓ − ) Σ = ker γ ( ℓ − ) ; hence b = β ( e , γ ( ℓ − ) e ) ≠ 0 for a basis vector e of Σ .
Theorem 5.14. 
Let k ≥ 0 , let n 1 , … , n k be a real null frame of rank k in R k + 1 , k + 1 with v = n 1 ∧ … ∧ n k , and let g be a compact non-abelian Lie algebra. Take Nishino’s transformations
δ A μ I = a ϵ ¯ γ μ λ I , δ λ I = 1 4 γ μ ν ρ 1 … ρ k ϵ F μ ν I v ρ 1 … ρ k ,
with a ≠ 0 , a Majorana–Weyl gaugino with values in K χ = Σ , the constraints n i μ F μ ν I = 0 and n i μ D μ λ I = 0 of Nishino’s eqs. (2.11)–(2.13), and the gaugino equation γ μ D μ λ I = 0 . Then:
(i)
the constraints are preserved by (27), and a–e hold with residual space R 1 , 1 ; 
(ii)
the part of δ ( γ μ D μ λ I ) quadratic in λ vanishes for every ϵ, so e holds with J = 0 ; 
(iii)
the commutator of two supersymmetries is, off shell, the translation by ξ on λ, and the translation by ξ, a gauge transformation and a shift of the components m i μ A μ on A; 
(iv)
the gaugino equation is D ℓ + λ I = 0 , and ξ is a multiple of ℓ + plus a combination of the n i . 
For k = 0 these are the transformations of two-dimensional super Yang–Mills theory with one Majorana–Weyl supercharge. 
Proof. 
A real null frame of rank k exists because k < min ( s , t ) = k + 1 , its residual space has signature ( 1 , 1 ) , and Majorana–Weyl spinors exist because s − t = 0 . Reality of the transformations follows as in the proof of Theorem 5.12, since the frame is real.
Reduction to the residual plane. The constraint n i μ F μ ν = 0 says that F lies in Λ 2 of W = span ( n 1 , … , n k ) ⊕ V ′ . Raising its indices, the legs along W that are not residual are multiples of the n i , and these die in the wedge product with v. Hence γ μ ν ρ 1 … ρ k F μ ν v ρ 1 … ρ k is a nonzero multiple of γ ( F ′ ) T , with F ′ the residual part of F and T = * * n * * 1 … * * n * * k . On Σ = ker γ ( ℓ + ) the operator γ ( ℓ + ∧ ℓ − ) = 1 2 ( γ ( ℓ + ) γ ( ℓ − ) − γ ( ℓ − ) γ ( ℓ + ) ) acts as 1, and γ ( F ′ ) = F ( ℓ − , ℓ + ) γ ( ℓ + ∧ ℓ − ) , so
δ λ = c F ( ℓ + , ℓ − ) x , x = T ϵ ∈ Σ ,
for a constant c ≠ 0 ; components of ϵ outside K ^ are annihilated by T (Lemma 5.2(iii)). In particular δ λ ∈ K , and the gaugino constraint is preserved. For a vector Y, δ A ( Y ) = a B ( ϵ , γ ( Y ) λ ) . It vanishes for Y = n i , because * * n * * i λ = 0 , and for Y = ℓ + , because γ ( ℓ + ) λ = 0 . For Y = ℓ − it is a β ( x , γ ( ℓ − ) λ ) , since B pairs Δ 1 only with K ^ (Lemma 5.2(iv)). With λ = ψ e and x = θ e , ψ and θ Grassmann-odd, this is a b θ ψ ≠ 0 , which gives c. The constraint n i μ F μ ν = 0 is preserved, since n i μ δ F μ ν = D n i δ A ν − D ν δ A ( n i ) = a ϵ ¯ γ ν D n i λ = 0 , and so is D n i λ = 0 , since D n i F = 0 by the Bianchi identity and δ A ( n i ) = 0 .
The gaugino equation and the cubic terms. In the basis ℓ ± , n i , m i , whose dual basis is ℓ ∓ , m i , n i ,
γ μ D μ λ = γ ( ℓ − ) D ℓ + λ + γ ( ℓ + ) D ℓ − λ + ∑ i * * m * * i D n i λ + * * n * * i D m i λ = γ ( ℓ − ) D ℓ + λ ,
because D μ λ takes values in Σ and D n i λ = 0 ; as γ ( ℓ − ) is injective on Σ , the gaugino equation is D ℓ + λ = 0 . By Lemma 5.5 the part of δ ( γ μ D μ λ I ) quadratic in λ is a g f I J K γ μ λ K B ( ϵ , γ μ λ J ) . In the same basis the frame terms are * * n * * i λ K B ( ϵ , * * m * * i λ J ) + * * m * * i λ K B ( ϵ , * * n * * i λ J ) = 0 , and the residual terms are γ ( ℓ + ) λ K B ( ϵ , γ ( ℓ − ) λ J ) + γ ( ℓ − ) λ K B ( ϵ , γ ( ℓ + ) λ J ) = 0 , since γ ( ℓ + ) annihilates Σ . This proves (ii), for every ϵ and not only for ϵ ∈ K ^ .
Closure. By d, ξ ( Y ) = c ′ ϵ ¯ 2 γ ( Y ∧ v ) ϵ 1 . For Y = n i it vanishes, and for Y ∈ V ′ it is a multiple of β ( x 2 , γ ( Y ) x 1 ) , which vanishes for Y = ℓ + . Hence ξ = ξ + ℓ + + ∑ i c i n i with ξ + a nonzero multiple of b θ 2 θ 1 ; this is (iv), and it gives d, because the bracket β ( x 2 , γ ( ℓ − ) x 1 ) is symmetric in the commuting spinor indices of the one-dimensional space Σ . Since F ( n i , · ) = 0 and D n i λ = 0 , the translation by ξ acts as the translation by ξ + ℓ + . On λ , (28) and δ F ( ℓ + , ℓ − ) = D ℓ + δ A ( ℓ − ) − D ℓ − δ A ( ℓ + ) = a b θ D ℓ + ψ give
[ δ 1 , δ 2 ] λ = c a b θ 1 ( D ℓ + ψ ) θ 2 − θ 2 ( D ℓ + ψ ) θ 1 e = 2 a b c θ 2 θ 1 D ℓ + λ ,
the translation along ℓ + , without use of any field equation. On A, the components A ( n i ) are invariant and F ( ξ , n i ) = 0 ; the component along ℓ + is invariant and F ( ξ + ℓ + , ℓ + ) = 0 ; and the component along ℓ − changes by
[ δ 1 , δ 2 ] A ( ℓ − ) = a β ( x 2 , γ ( ℓ − ) δ 1 λ ) − β ( x 1 , γ ( ℓ − ) δ 2 λ ) = 2 a b c θ 2 θ 1 F ( ℓ + , ℓ − ) ,
which is F ( ξ + ℓ + , ℓ − ) with the same coefficient. The translation of a gauge field by ξ is ξ ν F ν μ up to a gauge transformation with parameter − ξ ν A ν , and the remaining components A ( m i ) change by the extra transformation δ E A μ = v μ ν 1 … ν k − 1 Ω ν 1 … ν k − 1 of Nishino’s Equation (2.10), which shifts each A ( m i ) independently. This proves (iii), and (i) follows from the preceding steps. For k = 0 the frame is empty, T = 1 , and (27) is δ A μ = a ϵ ¯ γ μ λ , δ λ = 1 4 γ μ ν F μ ν ϵ .    □
Corollary 5.15. 
Let s − t ≡ 0 ( mod 8 ) and let R s , t carry a real null frame of rank k. A null-projected super Yang–Mills theory of rank k with a Majorana–Weyl gaugino exists if and only if either s + t − 2 k = 10 and s − t ∈ { 8 , 0 , − 8 } , or s + t − 2 k = 2 and s = t . Thus the conditions of the Main Theorem are necessary and sufficient. 
Proof. 
Necessity is Theorem 5.8. Sufficiency is Theorem 5.12 for residual dimension ten and Theorem 5.14 for residual dimension two.    □
The theories of Theorem 5.14 are chiral: the gaugino moves along ℓ + only, and supersymmetry translates along the same direction, so on shell the commutator acts trivially on λ . Their cubic gaugino terms vanish for the same reason as in four, six and ten dimensions: the residual spinors satisfy the identity (1), here trivially, because the only nonzero component of the vector bilinear is along ℓ + and γ ( ℓ + ) annihilates them. None of the families (2) reaches residual dimension two, since there s − t ≥ 16 .

6. A Single Time: Counting States

For the first family in (2) the time is single, and representation theory gives a bound that does not depend on how the theory is constructed. Consider a unitary, Poincaré-invariant quantum theory in R 9 + 8 n , 1 with n ≥ 1 , whose supercharges Q α form one Majorana–Weyl spinor, so that there are S = 2 4 + 4 n real supercharges. By Proposition 2.1 the most general anticommutator is
{ Q α , Q β } = ( γ μ C ) α β P μ + ∑ p ∈ { 5 , 9 , … , 5 + 4 n } ( γ μ 1 … μ p C ) α β Z μ 1 … μ p ,
with C the charge conjugation matrix and Z [ p ] tensor charges, in accordance with the general structure of supersymmetry algebras [35,42]. Tensor charges of this kind are carried by extended objects [43]; in ten dimensions the five-form charge is carried by five-branes. We state their absence on the massless multiplet as hypothesis (H).
Theorem 6.1. 
Assume (H) that the tensor charges vanish on a massless supermultiplet. Then the multiplet has at least 2 S / 4 = 2 2 2 + 4 n states for each momentum, half of them bosonic. In particular the theory has no massless vector multiplet: such a multiplet would have D − 2 = 8 + 8 n bosonic states for each generator of the gauge algebra. 
Proof. 
On the multiplet, with momentum p, p 2 = 0 , p 0 > 0 , the anticommutator is M α β = ( p C ) α β , a real symmetric matrix, positive semidefinite because the Q α are Hermitian and the norm is positive. Its rank is S / 2 : choosing a real null frame of rank one with n 1 = p , Lemma 5.2 shows that p annihilates exactly the half K of each Weyl module. A supercharge Q in the kernel of M has { Q , Q } = 0 , so ∥ Q ψ ∥ 2 = 0 and Q vanishes on every state ψ of the multiplet. The remaining S / 2 = 2 3 + 4 n combinations, rescaled, satisfy { Q ^ a , Q ^ b } = 2 δ a b and generate a complex Clifford algebra on 2 3 + 4 n generators, whose irreducible representations have dimension 2 2 2 + 4 n and are split evenly by ( − 1 ) F . The multiplet is a direct sum of such representations.    □
For n = 0 the bound is 2 4 = 16 : the eight bosonic and eight fermionic states of ten-dimensional super Yang–Mills theory. For n = 1 , D = 18 , it is 2 64 , against sixteen bosonic and 128 fermionic states of a vector and a Majorana–Weyl gaugino. Figure 8 shows the gap for n ≤ 3 . The argument is the one that bounds the helicity of massless supermultiplets in four dimensions by one for sixteen supercharges and by two for thirty-two [44,45]. To exclude arbitrary multiplicities of vector, spinor and scalar fields, a state count alone is insufficient. The following helicity argument supplies the needed bound.
Lemma 6.2 
(Helicity span). Under the hypotheses of Theorem 6.1, suppose the massless multiplet has a finite helicity spectrum for a rotation J in a transverse two-plane. If h min and h max are its least and greatest helicities, then
h max − h min ≥ S / 8 .
In particular a multiplet containing only scalar, spinor and vector polarizations has S ≤ 16 , independently of their multiplicities. 
Proof. 
Choose the null momentum in a four-dimensional Lorentzian subspace and let J generate rotations in its transverse two-plane. A transverse spinor has J-weights + 1 2 and − 1 2 with equal multiplicities. In its occupation basis the first occupation fixes this weight, while the chirality condition fixes only the parity of the remaining occupations; there are equally many choices for the two weights. On the S / 2 real active supercharges, complexification therefore gives S / 4 operators of each weight. The positive anticommutator of Theorem 6.1, invariant under J, pairs the two weights nondegenerately. After normalization the operators may be written a i † , a i , 1 ≤ i ≤ S / 4 , with
{ a i , a j † } = δ i j , { a i , a j } = { a i † , a j † } = 0 , [ J , a i † ] = 1 2 a i † .
Take a unit vector Ω of helicity h min . Each a i lowers helicity by 1 2 , so a i Ω = 0 . The canonical anticommutation relations give
a 1 † … a S / 4 † Ω 2 = 1 .
This nonzero vector has helicity h min + S / 8 , proving the span bound. A scalar has helicity 0, a spinor has helicities ± 1 2 , and a vector has helicities 0 , ± 1 under this rotation. Arbitrary multiplicities of these polarizations still have their spectrum in [ − 1 , 1 ] . Thus S / 8 ≤ 2 , or S ≤ 16 .    □
For n ≥ 1 , S = 2 4 + 4 n > 16 , so a massless multiplet in this unitary, tensor-charge-free class cannot be assembled from any finite number of those low-spin fields. The mechanism is the familiar helicity bound on extended supersymmetry [44,45]; it is written out here to distinguish it from the separate dimension count.

7. What This Means for 27 + 3 and 19 + 3

In 27 + 3 the Weyl spinor has 2 14 = 16 384 components, and its symmetric square contains forms of degrees 3, 7, 11 and 15 only. The anticommutator of two supercharges contains no momentum. In the constructions of Refs. [19,20,22] translations are obtained by contracting a form charge with null vectors, as in d; the decomposition of the symmetric square does not by itself force this mechanism, and Sec. Section 7.2 treats the other contractions. With three times k ≤ 3 , so the residual space has dimension at least 24, and Corollary 5.9 applies: the cubic gaugino terms cannot be absorbed. The same holds in 19 + 3 , where the residual dimension is at least 16, and in every other signature of the families (2) with n ≥ 1 . For any compact non-abelian Lie gauge algebra, the obstruction (21) already arises on an su ( 2 ) subalgebra and is unaffected by additional representation labels. A non-Lie extension or other field content would require assumptions beyond the present definition.

7.1. The Transfer of Nishino’s Construction

The theories in ( 11 + 8 n ) + 3 rest on one argument of Ref. [23]. Moving 4 n timelike directions into space keeps D = s + t and s − t modulo eight, so the reality and symmetry properties of spinors and gamma matrices are unchanged, and the authors conclude that Nishino’s existence proofs can be repeated in the new signatures.8 The paper on 27 + 3 takes this theory as given: its superalgebra (2.1) is (3.7) at n = 2 ,
{ Q α , Q β } = ( γ μ ν ρ ) α β Z μ ν ρ + ( γ μ 1 … μ 7 ) α β Z μ 1 … μ 7 + ( γ μ 1 … μ 11 ) α β Z μ 1 … μ 11 + ( γ μ 1 … μ 15 ) α β Z μ 1 … μ 15 ,
and it gives no transformation rules or field equations.9 The quoted sentence names the proofs for (3.1) and (3.2), whose transfers are the first two families of (2); for those, Section 5 already shows that the repetition fails. The signatures ( 11 + 8 n ) + 3 come from (3.3), the superalgebra of Nishino’s Section 4, which treats ( 11 + 4 n ) + ( 3 + 4 n ) . That section is the proof to be repeated, and we write it out.
Nishino’s coordinates in ( 11 + 4 n ) + ( 3 + 4 n ) are x 0 , … , x 9 , x 11 , … , x 14 + 8 n , with x 0 and x 12 , x 14 , … , x 14 + 8 n timelike [22]. His null vectors, eqs. (4.5) and (4.6), are
n ( i ) = 1 2 e 9 + 2 i + e 10 + 2 i , m ( i ) = 1 2 e 9 + 2 i − e 10 + 2 i , i = 1 , … , 4 n + 2 ,
each pairing one spacelike with one timelike unit vector, so the 4 n + 2 vectors n ( i ) use 4 n + 2 different times. With v = n ( 1 ) ∧ … ∧ n ( 4 n + 2 ) his theory consists of
δ A μ I = ϵ ¯ γ μ λ I , δ λ I = − 1 4 γ μ ν [ 4 n + 2 ] ϵ F μ ν I v [ 4 n + 2 ] , Equation ( 4.3 )
v [ 4 n + 1 ] μ γ μ λ I = 0 , Equation ( 4.9 )
D μ F μ [ ν 1 I v ν 2 … ν 4 n + 3 ] = 1 2 4 n + 2 ( 4 n + 3 ) ! f I J K λ ¯ J γ ν 1 … ν 4 n + 3 λ K , Equation ( 4.12 )
{ Q α , Q β } = ( γ μ [ 4 n + 2 ] ) α β v [ 4 n + 2 ] P μ , Equation ( 4.15 )
together with the closure (4.13), (4.14), whose translation parameter is ξ μ = − ( ϵ ¯ 1 γ μ [ 4 n + 2 ] ϵ 2 ) v [ 4 n + 2 ] . Here [ p ] denotes p antisymmetrized indices. The right-hand side of () is the source that absorbs the cubic gaugino terms, and () shows where the momentum comes from: the ( 4 n + 3 ) -form charge contracted with v. Repeating the proof in ( 11 + 8 n ) + 3 means keeping these equations and choosing a null frame in the new signature. Since the rank 4 n + 2 may be too large there, we allow any rank k.
Proposition 7.1. 
Let n ≥ 1 and let g be a compact non-abelian Lie algebra. In R 11 + 8 n , 3 consider a gauge field A μ I , a Majorana–Weyl gaugino λ I and a Majorana–Weyl parameter ϵ of the same chirality, as in (3.7) of Ref. [23] and (30), a real null frame of rank k ≥ 0 with v = n 1 ∧ … ∧ n k ( v = 1 if k = 0 ), a coupling g ≠ 0 in D μ λ I , and Nishino’s equations with 4 n + 2 replaced by k, as identities for all gaugino values allowed by (b) and all ϵ:
(a)
δ A μ I = ϵ ¯ γ μ λ I and δ λ I = − 1 4 γ μ ν ρ 1 … ρ k ϵ F μ ν I v ρ 1 … ρ k , as in (32); 
(b)
v ρ 1 … ρ k − 1 μ γ μ λ I = 0 if k ≥ 1 , as in (), and the gaugino values are all Majorana spinors satisfying it; 
(c)
the commutator of two supersymmetries contains the translation by ξ μ = − ( ϵ ¯ 1 γ μ ρ 1 … ρ k ϵ 2 ) v ρ 1 … ρ k , the coefficient of P μ in { Q α , Q β } = ( γ μ ρ 1 … ρ k ) α β v ρ 1 … ρ k P μ as in (), and its residual components ξ μ ′ do not vanish identically; 
(d)
the terms quadratic in λ in δ ( γ μ D μ λ I ) are cancelled by a source J ∧ v in the field equation of the gauge field, as in (). 
Then:
(i)
k ≤ 3 , so Nishino’s rank 4 n + 2 cannot be kept. 
(ii)
Condition (c) holds only for k = 2 . 
(iii)
For k = 2 conditions (a)–(c) imply a–d, condition (d) is e, the residual space is R 9 + 8 n , 1 , and (d) fails: absorption would force the identity (6) on the residual Weyl module, of dimension N = 2 4 + 4 n , while Theorem 3.3 requires N = 2 ( 8 + 8 n ) = 16 + 16 n . 
Hence no field theory in ( 11 + 8 n ) + 3 , in particular none in 19 + 3 or 27 + 3 , satisfies (a)–(d) for any rank of the null frame. 
Proof. 
We write D = 14 + 8 n = 2 r , so r = 7 + 4 n , and use the oscillator notation of Section 5: K is the joint kernel of the * * n * * i and K ^ the joint kernel of the * * m * * i . The gaugino does not vanish identically, since otherwise δ A = 0 and no field transforms.
Step 1: the rank is at most three. First, (b) says exactly that λ I takes values in K. Indeed v ρ 1 … ρ k − 1 μ γ μ is, for each choice of ρ 1 , … , ρ k − 1 , a linear combination of * * n * * 1 , … , * * n * * k , so * * n * * i λ = 0 for all i implies (b). Conversely, contract (b) with m j ρ in the slots ρ 1 , … , ρ k − 1 for all j ≠ i ; since n l · m j = δ l j , only the term with n i in the last slot survives, and (b) becomes a nonzero multiple of * * n * * i λ I = 0 . Next, the n i are real, mutually orthogonal null vectors, linearly independent because n i · m j = δ i j , so they span a real totally isotropic subspace of dimension k, and Lemma 2.2 gives k ≤ min ( 11 + 8 n , 3 ) = 3 . Lemma 2.3 gives the same bound for any frame, real or complex, once the values of λ are Majorana spinors. Nishino’s rank is 4 n + 2 ≥ 6 , which proves (i). The same count holds in the source signature, where min ( 11 + 4 n , 3 + 4 n ) = 3 + 4 n ≥ 4 n + 2 leaves room for his frame.
Step 2: only rank two gives a translation. By Proposition 2.1, the forms ϵ ¯ 1 γ μ 1 … μ p ϵ 2 on Δ + × Δ + vanish for even p, are symmetric in ϵ 1 , ϵ 2 for p ≡ r ≡ 3 ( mod 4 ) , and antisymmetric for p ≡ 1 ( mod 4 ) . For 27 + 3 the degrees are
symmetric : 3 , 7 , 11 , 15 , antisymmetric : 1 , 5 , 9 , 13 , zero : 0 , 2 , 4 , … , 14 ,
and the symmetric ones are the charges of (30). The parameter ξ μ of (c) is the component of ϵ 1 ⊗ ϵ 2 in degree k + 1 . For k = 1 and k = 3 the degree is even and ξ vanishes identically. For k = 0 the degree is one, the form is antisymmetric, and the momentum is absent from the anticommutator; this is why, among the superalgebras (3.1)–(3.8) of Ref. [23], only (3.1) and (3.5) contain P μ . For k = 2 the degree is three and the form is symmetric. With k ≤ 3 this leaves k = 2 , which proves (ii). Nishino’s rank 4 n + 2 would draw the momentum from the ( 4 n + 3 ) -form charge, symmetric as well, contracted with a null form of rank 4 n + 2 ; in 27 + 3 this is the eleven-form charge of Ref. [24], and his contraction needs ten mutually orthogonal real null vectors, whereas by Step 1 there are at most three.
Step 3: at rank two the axioms (N1)–(N4) hold. Let k = 2 . Axiom a holds because, by hypothesis, the gaugino values are all spinors allowed by (b), and by Step 1 these form the full corner K χ ; ϵ ranges over the full module, as in Nishino’s construction. Axiom b is (a), with a = 1 and Φ ( F ) ϵ = − 1 4 γ μ ν ρ 1 ρ 2 ϵ F μ ν v ρ 1 ρ 2 , which is linear in F and independent of λ , and g ≠ 0 by hypothesis. For c and d we use Lemma 5.2. Since k is even and λ and ϵ have the same chirality, T ϵ ranges over K χ = : Σ as ϵ ranges over K ^ χ , and Σ is a Weyl module of the residual space V ′ = R 9 + 8 n , 1 , of dimension 2 r − k − 1 = 2 4 + 4 n . As r is odd, B pairs Δ + with Δ − , so β of (22) pairs Σ nondegenerately with the opposite Weyl half Σ opp of Δ 1 , and γ μ ′ maps Σ injectively into Σ opp . Therefore δ A μ ′ I = β ( T ϵ , γ μ ′ λ I ) and ξ μ ′ = ± β ( T ϵ 2 , γ μ ′ T ϵ 1 ) , as in (23), do not vanish identically, and the pairing in ξ μ ′ is symmetric because r ′ = r − 2 ≡ 1 ( mod 4 ) . Condition (d) asserts that the quadratic terms have the form (17), which is e.
Step 4: absorption fails. Theorem 5.6 now applies and gives D ′ ∈ { 2 , 10 } , whereas D ′ = 14 + 8 n − 4 = 10 + 8 n . Concretely, Steps 1–3 of its proof turn (d) into the identity (6) on Σ with d = D ′ = 10 + 8 n and N = 2 4 + 4 n . The second contraction (9) fixes the constant, j = − ( d − 2 ) / N = − ( 8 + 8 n ) / 2 4 + 4 n , and the first contraction (8) then requires N = 2 ( d − 2 ) , which is false for every n ≥ 1 because 2 4 + 4 n = 16 n + 1 > 16 ( n + 1 ) for n ≥ 1 . This proves (iii).    □
For 27 + 3 the numbers are D ′ = 26 , N = 4096 and j = − 24 / 4096 = − 3 / 512 , and (8) would require 4096 − 24 = − 2 j · 24 , that is 4072 = 9 / 32 . In Nishino’s own signatures the same count gives N = 16 = 2 ( 10 − 2 ) , because his frame has rank t − 1 and leaves the residual space R 9 , 1 . Compare Sezgin’s theory: in 11 + 3 a rank-two frame leaves ten residual dimensions and everything closes; in 27 + 3 the same rank, the only one the translation bracket allows, leaves twenty-six. Table 1 and Figure 9 and Figure 10 collect the two cases singled out in Refs. [23,24], and Appendix D checks each step in explicit gamma matrices of size 2 15 .
Remark 7.2. 
The transfer becomes consistent only by giving up the reality of the gaugino. The complexified Clifford algebra of R s , t depends only on D, so Nishino’s equations in 19 + 11 can be written with a complex null frame of rank 10 in C 30 and read in the coordinates of R 27 , 3 . The residual space is then ten-dimensional and the absorption holds, as in Remark 5.10. But the frame spans a complex totally isotropic subspace of dimension 10 > 3 , so by Lemma 2.3 its occupied corner contains no nonzero Majorana spinor: the gaugino, and through δ A = ϵ ¯ γ λ the gauge field, are complex. These are the complexified equations of the theory in 19 + 11 , whose real form there carries a real frame. The data that Ref. [23] keeps fixed, D and s − t modulo eight, fix the complex Clifford algebra and the reality type of its spinors, but not the Witt index min ( s , t ) that a real null frame needs.10 

7.2. Other Contractions

Proposition 7.1 keeps the null contraction of Nishino’s translation bracket. The superalgebra (30) admits others. Contracting its eleven-form charge with ten orthonormal spacelike vectors e 1 , … , e 10 , for instance, gives { Q α , Q β } = ( γ a 1 … 10 ) α β P a with a orthogonal to the e i and the P a central, a consistent superalgebra with symmetry Spin ( 17 , 3 ) × Spin ( 10 ) . The next theorem shows that no choice of this kind rescues the field theory. Whatever fixed tensor contracts the source in the field equation of the gauge field, the cubic gaugino terms cannot be absorbed once the residual dimension is sixteen or more. The theorem does not use d, and it includes a gaugino without any constraint ( k = 0 ).
For a multivector ω ∈ Λ V C let γ ( ω ) be the antisymmetrized product of gamma matrices, as in Step 1 of the proof of Theorem 5.6. For a null frame, V C is the direct sum of span ( n i ) , span ( m i ) and V C ′ , and for ω ∈ Λ V C we write ω ∘ ∈ Λ V C ′ for the coefficient of n 1 ∧ … ∧ n k in the part of ω that contains no m i .
Theorem 7.3. 
Let D = 2 r , let a null frame of rank k ≥ 0 , real or complex, be given (for k = 0 there is no frame and K = K ^ = Δ ), and let D ′ = D − 2 k = 2 r ′ ≥ 2 . Suppose that the gaugino takes all values in K χ , that b and c hold, and that there is a fixed u ∈ Λ q V C such that for every λ there are J I , independent of ϵ, with
part of δ ( γ μ D μ λ I ) quadratic in λ = γ ( J I ∧ u ) ϵ for every ϵ ∈ K ^ χ ′ .
Let P ′ = { p : 0 ≤ p < r ′ , p ≡ r ′ ( mod 4 ) } . Then P ′ has at most one element p, and D ′ p ≤ D ′ + k . Consequently D ′ ≤ 14 , and D ′ ≤ 10 when k ≤ 53 . 
Proof. 
Projection. As in Step 1 of the proof of Theorem 5.6, for ϵ ∈ K ^ χ ′ the left-hand side of (36) is a g f I J K γ μ ′ λ K β ( x , γ μ ′ λ J ) with x = T ϵ and β as in (22), and it lies in K = Δ 1 . On the right, expand J ∧ u in the monomials n I ∧ m L ∧ ω ′ with ω ′ ∈ Λ V C ′ . The factors for different i and the residual factor span mutually orthogonal subspaces, so γ of such a monomial is, up to sign, the product of γ ( ω ′ ) , of * * n * * i for i ∈ I ∖ L , of * * m * * i for i ∈ L ∖ I , and of γ ( n i ∧ m i ) = P i ↓ − P i ↑ for i ∈ I ∩ L . By Lemma 5.2(ii) it maps Δ a into Δ a + e I ∖ L − e L ∖ I , so it maps K ^ = Δ 0 into K only if I = { 1 , … , k } and L = ∅ . Projecting both sides onto K therefore gives, for every x ∈ K χ ′ ′ ,
a g f I J K γ μ ′ λ K β ( x , γ μ ′ λ J ) = ± γ ( J I ∧ u ) ∘ x .
Write J = J ′ + ∑ i ( a i n i + b i m i ) with J ′ ∈ V C ′ . A term b i m i produces an m i , and a i n i contributes only through the coefficient u ( i ) ∈ Λ V C ′ of n 1 ∧ … ∧ n i ^ ∧ … ∧ n k in the part of u without m’s. Hence ( J ∧ u ) ∘ = J ′ ∧ u ∘ ± ∑ i a i u ( i ) lies in a subspace E ⊂ Λ p 0 V C ′ , p 0 = q − k + 1 , of dimension at most D ′ + k .
Degrees. By Lemma 5.2(i), K is a Dirac module of Cl ( V C ′ ) , and β is a nondegenerate form on it with β ( γ μ ′ x , y ) = ± β ( x , γ μ ′ y ) , by Lemma 5.2(iii) and (3). Such a form is unique up to a factor, so it is symmetric or antisymmetric, and the proof of Proposition 2.1 applies to it: with Σ = K χ , the map that sends S ∈ Sym 2 Σ to its components Y p ( S ) in the forms of degree p ≡ r ′ ( mod 4 ) , p ≤ r ′ , is an isomorphism onto ⨁ p Λ p V C ′ . The left-hand side of (37) is ± a g ∑ μ ′ γ μ ′ X I γ μ ′ x with X I = f I J K λ K β ( λ J , · ) , and in an orthonormal basis
∑ μ ′ γ μ ′ γ ( ω ) γ μ ′ = ( − 1 ) p ( D ′ − 2 p ) γ ( ω ) ( ω ∈ Λ p V C ′ ) ,
because γ μ ′ anticommutes with the factors of a basis monomial that differ from it and commutes with the one equal to it. On K χ ′ ′ the left-hand side is therefore ± a g ∑ p ( − 1 ) p ( D ′ − 2 p ) γ ( Y p ( S I ) ) , summed over p ≡ r ′ ( mod 4 ) , p ≤ r ′ , where the term with p = r ′ vanishes. The right-hand side is γ ( e ) with e ∈ E ⊂ Λ p 0 V C ′ , and on a Weyl module of V ′ the operators γ ( Λ p 0 ) and γ ( Λ D ′ − p 0 ) coincide by Hodge duality, while the operators γ ( Λ p ) with p < r ′ of different degrees are independent. As in Step 3 of the proof of Theorem 5.6, the tensors S I span Sym 2 Σ , so for every p ∈ P ′ the components Y p ( S I ) fill Λ p V C ′ , while D ′ − 2 p ≠ 0 . Hence every p ∈ P ′ equals min ( p 0 , D ′ − p 0 ) , and then Λ p V C ′ is the image of E under Hodge duality or E itself, so D ′ p ≤ dim E ≤ D ′ + k .
For r ′ ≥ 8 both r ′ − 4 and r ′ − 8 lie in P ′ , so D ′ ≤ 14 . For D ′ = 12 and D ′ = 14 one has P ′ = { 2 } and P ′ = { 3 } , and 12 2 = 66 , 14 3 = 364 exceed D ′ + k unless k ≥ 54 .    □
Corollary 7.4. 
Let ( s , t ) be one of the signatures (2) with n ≥ 1 . Let the Majorana–Weyl gaugino take all values in the joint kernel of k ≥ 0 mutually orthogonal null vectors, and let b and c hold. Then for no multivector u can the cubic gaugino terms be absorbed by a source of the form J ∧ u . In 27 + 3 this covers every k ≤ 3 , the spacelike contraction above, every contraction with a mixture of null and non-null vectors, and the gaugino without constraint. 
Proof. 
By Lemma 2.3 the null vectors lie in a real totally isotropic subspace, so k ≤ min ( s , t ) = t , and they extend to a real null frame of rank k. Its residual dimension is D ′ = s + t − 2 k ≥ s − t = 8 + 8 n ≥ 16 , which Theorem 7.3 excludes.    □
Theorem 7.3 is weaker than Theorem 5.6 in one respect and stronger in two. Without d it does not force D ′ = 10 : for D ′ ≤ 6 the projected cubic terms vanish identically, and for D ′ = 8 a source of degree seven in the residual directions absorbs them. But it allows any fixed tensor u, and its bound D ′ ≤ 14 holds for every frame. Figure 11 shows the degrees that decide it. Appendix E checks the theorem in explicit gamma matrices.

7.3. Several Gauginos

With several gauginos the natural extension of b lets each gaugino transform into the field strength with its own parameter. This is the structure of extended supersymmetry, where a pairing of the gaugino and parameter labels that is invariant under the R-symmetry makes both transformation rules diagonal. A source in the field equation of the gauge field enters the variation of the field equation of λ A only through δ λ A , so the cubic terms it can absorb are linear in ϵ A ; the cubic terms produced by the other parameters must vanish. The next proposition shows that they do not vanish once the residual dimension is at least three.
Proposition 7.5. 
Let a null frame of rank k ≥ 0 , real or complex, have residual dimension D ′ ≥ 3 , and let g be non-abelian. Consider M ≥ 2 gauginos λ 1 , … , λ M in the adjoint representation, where λ A takes all values in K χ A for a chirality χ A , with independent parameters ϵ 1 , … , ϵ M and transformations
δ A μ I = ∑ A = 1 M a A ϵ ¯ A γ μ λ A I , δ λ A I = Φ A ( F I ) ϵ A ,
where the linear maps Φ A do not depend on λ and D μ λ A I = ∂ μ λ A I + g f I J K A μ J λ A K with g ≠ 0 . Suppose that the residual components δ A μ ′ I do not vanish identically. Then for some A the part of δ ( γ μ D μ λ A I ) quadratic in λ does not vanish when ϵ A = 0 , so it cannot be absorbed by a source. 
Proof. 
For ϵ ∈ K ^ χ ′ and y ∈ K one has B ( ϵ , γ μ ′ y ) = β ( T ϵ , γ μ ′ y ) by (22), and B ( ϵ , γ μ ′ y ) = 0 for ϵ in the other corners, by Lemma 5.2(i) and (iv). Hence the residual components of δ A are ∑ A a A β ( T ϵ A , γ μ ′ λ A ) , and by hypothesis there is an index B with a B ≠ 0 for which the residual vector G ( x , y ) with components G μ ′ ( x , y ) = β ( x , γ μ ′ y ) is not identically zero for x ∈ K χ ′ ′ , y ∈ Σ B = K χ B .
Choose A ≠ B , which is possible as M ≥ 2 , and put ϵ C = 0 for C ≠ B . Then δ λ A = 0 , and the part of δ ( γ μ D μ λ A I ) quadratic in λ is a B g f I J K γ μ λ A K ( ϵ ¯ B γ μ λ B J ) . In the basis n i , m i and a basis of V ′ the frame terms are * * n * * i λ A B ( ϵ B , * * m * * i λ B ) + * * m * * i λ A B ( ϵ B , * * n * * i λ B ) = 0 , since * * n * * i annihilates K, and what remains is
a B g f I J K γ G ( x , λ B J ) λ A K , x = T ϵ B ,
with γ ( g ) = g μ ′ γ μ ′ . Take λ B J = y ψ for a single value of J and λ B = 0 otherwise, λ A K = σ φ for a single K with f I J K ≠ 0 for some I, and x = x ˜ ϑ , where x ˜ , y, σ are commuting spinors and ϑ , ψ , φ are Grassmann-odd. Since λ A and λ B are independent fields, the product ϑ ψ φ is not zero, so if the cubic terms vanished for ϵ A = 0 we would have
γ G ( x ˜ , y ) σ = 0 for all x ˜ ∈ K χ ′ ′ , y ∈ Σ B , σ ∈ Σ A .
Let W = { g ∈ V C ′ : γ ( g ) Σ A = 0 } . It contains a nonzero vector G ( x ˜ , y ) . As Σ A is invariant under Spin ( V ′ ) and s γ ( g ) s − 1 = γ ( s · g ) , the subspace W is invariant under Spin ( V ′ ) , hence under its Lie algebra and, being complex, under so ( V C ′ ) . For D ′ ≥ 3 the vector representation of so ( V C ′ ) is irreducible, so W = V C ′ . But V C ′ contains a vector g with η ( g , g ) ≠ 0 , and then γ ( g ) 2 = η ( g , g ) makes γ ( g ) invertible, so γ ( g ) Σ A ≠ 0 . This contradiction proves the proposition.    □
The proof uses neither d nor the form of the source, and it holds for every frame, real or complex, and for gauginos of either chirality. For D ′ = 2 it gives nothing: there G ( x , y ) is a multiple of ℓ + , which annihilates Σ , and the cross terms vanish, as in the proof of Theorem 5.14. Appendix F checks in explicit gamma matrices that the cross terms are nonzero for D ′ = 4 , 6 , 8 , 10 and vanish for D ′ = 2 .

7.4. Mixed Gauginos

Proposition 7.5 lets each gaugino transform with its own parameter. In extended supersymmetry the gauginos are in general mixed: the gauge field transforms into several gauginos with the same parameter, and a gaugino transforms into the field strength with several parameters. In six dimensions this is the structure of ( 1 , 0 ) super Yang–Mills theory, whose gaugino is a doublet paired by ε i j , and in four dimensions that of N = 1 theory written with the two Weyl halves of a Majorana gaugino. Mixing therefore changes the answer, and the next theorem says exactly how.
Let a null frame of rank k ≥ 0 , real or complex, be given (for k = 0 there is no frame and K = K ^ = Δ ), with residual dimension D ′ = 2 r ′ . Consider M ≥ 2 gauginos λ 1 , … , λ M , where λ A takes all values in K χ A ⊗ g , and parameters ϵ 1 , … , ϵ M ′ , where ϵ P ranges over Δ χ P ′ , with
δ A μ I = ∑ A , P a A P ϵ ¯ P γ μ λ A I , δ λ A I = ∑ P Φ A P ( F I ) ϵ P ,
where the a A P are constants, the linear maps Φ A P from two-forms to End Δ do not depend on the gauginos, and D μ λ A I = ∂ μ λ A I + g f I J K A μ J λ A K with g ≠ 0 and structure constants f I J K that are not all zero. We impose two conditions.
(M1)
The residual components δ A μ ′ I do not vanish identically.
(M2)
For all A and P and every value of the gauginos there are J A P σ I , independent of the parameters, such that for every ϵ P , with ϵ Q = 0 for Q ≠ P , the part of δ ( γ μ D μ λ A I ) quadratic in the gauginos equals γ σ ν 1 … ν k ϵ P v ν 1 … ν k J A P σ I .
Condition b is the absorption axiom e for each gaugino equation and each parameter. A source J in the field equation of the gauge field enters the variation of the field equation of λ A through Φ A P ( F ) ϵ P , and with Φ A P a multiple φ A P of Nishino’s map of Example 5.4 it produces b with J A P = φ A P J . Allowing an independent current for every pair ( A , P ) only weakens the hypothesis. If the gaugino equations are ∑ B c A B γ μ D μ λ B = 0 with an invertible matrix c, condition b for these combinations is equivalent to b as stated, since the right-hand side of b is preserved by invertible linear combinations over A.
Theorem 7.6. 
Under a and b, either D ′ ≤ 2 or D ′ ∈ { 4 , 6 } . If D ′ ∈ { 4 , 6 } , the dimension N = 2 r ′ − 1 of a Weyl module of V ′ equals D ′ − 2 . 
Proof. 
Assume D ′ ≥ 3 . We use the pairing β ( z , w ) = B ( T − 1 z , w ) of (22), now for all z , w ∈ K = Δ 1 ; it is nondegenerate and satisfies β ( γ μ ′ z , w ) = θ ′ β ( z , γ μ ′ w ) with θ ′ = ( − 1 ) k θ , by Lemma 5.2(iii) and (3). For z , w ∈ K let G ( z , w ) ∈ V C ′ be the vector with components G μ ′ ( z , w ) = β ( z , γ μ ′ w ) , and write γ ( u ) = u μ ′ γ μ ′ for u ∈ V C ′ .
The cross term. As in the proof of Proposition 7.5, the residual components of δ A are ∑ B , P a B P β ( T ϵ P , γ μ ′ λ B ) , and only the K ^ -components of the parameters enter. By a there are P and B with a B P ≠ 0 such that G ( x , y ) is not identically zero for x ∈ X : = K χ ′ ′ , χ ′ ′ = ( − 1 ) k χ P ′ , and y ∈ Σ B : = K χ B . Choose A ≠ B , possible as M ≥ 2 , put ϵ Q = 0 for Q ≠ P and take ϵ P ∈ K ^ χ P ′ . By Lemma 5.5, applied to each gaugino, the part of δ ( γ μ D μ λ A I ) quadratic in the gauginos is g f I J K γ μ λ A K ∑ C a C P ( ϵ ¯ P γ μ λ C J ) . The frame terms of (18) vanish because * * n * * i annihilates K, and projecting b onto Δ 1 as in Step 1 of the proof of Theorem 5.6 gives, with x = T ϵ P ,
g f I J K γ μ ′ λ A K ∑ C a C P β ( x , γ μ ′ λ C J ) = γ μ ′ x J A P μ ′ I ( x ∈ X ) .
Choose I , J , K with f I J K ≠ 0 , and take λ B J = y ψ , λ A K = σ φ , x = x ˜ ϑ , with commuting spinors y ∈ Σ B , σ ∈ Σ A : = K χ A , x ˜ ∈ X , Grassmann-odd ψ , φ , ϑ , and all other components of all gauginos zero. The term with C = A contains φ 2 = 0 , so the left-hand side of (40) is ± g a B P f I J K γ ( G ( x ˜ , y ) ) σ ϑ ψ φ . The current lies in the Grassmann algebra generated by ψ and φ , since it does not depend on the parameters, and comparing the coefficients of ϑ ψ φ shows that for all σ ∈ Σ A and y ∈ Σ B there is a vector u ( σ , y ) ∈ V C ′ with
γ G ( x , y ) σ = γ u ( σ , y ) x for all x ∈ X .
The cross term is not zero. If γ ( G ( x , y ) ) σ = 0 for all x ∈ X , y ∈ Σ B , σ ∈ Σ A , the argument of the last paragraph of the proof of Proposition 7.5 applies: the space of g ∈ V C ′ with γ ( g ) Σ A = 0 contains a nonzero vector G ( x , y ) and is invariant under so ( V C ′ ) , hence equals V C ′ , which contains a vector g with γ ( g ) 2 = η ( g , g ) ≠ 0 . Hence the left-hand side of (41) does not vanish identically.
Chirality. By Lemma 5.2(v), X and Σ A are Weyl modules of V ′ in Δ 1 , so either Σ A = X or Σ A is the other Weyl half X opp . In the second case the left-hand side of (41) lies in γ μ ′ X opp = X and the right-hand side in X opp , so both vanish, which was just excluded. Hence Σ A = X .
The current. Fix σ ∈ X and y ∈ Σ B , apply γ ν to both sides of (41) and take the trace over x ∈ X . On the left, the map x ↦ γ ν γ μ ′ σ β ( x , γ μ ′ y ) has image in X and trace β ( γ ν γ μ ′ σ , γ μ ′ y ) = θ ′ 2 β ( σ , γ μ ′ γ ν γ μ ′ y ) = − ( D ′ − 2 ) G ν ( σ , y ) . On the right, tr X ( γ ν γ μ ′ ) = N η ν μ ′ by the argument of Lemma 3.2(c), which needs D ′ ≥ 3 , so the trace is N u ν ( σ , y ) . Therefore
u ( σ , y ) = j G ( σ , y ) , j = − D ′ − 2 N ,
the same constant as in Step 3 of the proof of Theorem 5.6.
The swap. Substituting into (41),
γ G ( x , y ) σ = j γ G ( σ , y ) x for all x , σ ∈ X , y ∈ Σ B .
Both x and σ range over X, so the identity may be applied with their roles exchanged, and then γ ( G ( x , y ) ) σ = j 2 γ ( G ( x , y ) ) σ . The left-hand side is not identically zero, so j 2 = 1 , and as j < 0 , N = D ′ − 2 , that is 2 r ′ − 1 = 2 ( r ′ − 1 ) . This holds for r ′ = 2 , 3 , and for r ′ ≥ 4 the left-hand side exceeds the right. Hence D ′ ∈ { 4 , 6 } .    □
The proof uses neither d nor the Jacobi identity, and it holds for every frame, real or complex. It does not assume that the gauginos are paired with the parameters in any particular way: a single gaugino λ B in the variation of the gauge field and a single further gaugino λ A with the covariant derivative suffice. For D ′ = 10 it says that a second adjoint gaugino with rules of the form (39) is incompatible with absorption, and for k = 0 it says that ordinary super Yang–Mills theory with a gauge field and M ≥ 2 gauginos, no other fields and rules of the form (39) is possible in more than two dimensions only in four and six. Proposition 7.5 treats rules in which δ λ A contains only ϵ A . A source then enters the equation of λ A only through ϵ A , so that J A P = 0 for P ≠ A , and the conclusion is stronger: there is no such theory for D ′ ≥ 3 .
Corollary 7.7. 
Let a null-projected theory of rank k in signature ( s , t ) have M ≥ 2 Majorana–Weyl gauginos with the transformations (39) and satisfy a and b. Then k ≤ min ( s , t ) , s = t and s + t − 2 k ∈ { 2 , 4 , 6 } . In particular no such theory exists in any of the signatures (2) with n ≥ 1 , for any number of gauginos and any mixing; none exists in 27 + 3 or 19 + 3 . 
Proof. 
By a some gaugino does not vanish identically, and its values are Majorana spinors in some K χ A . The proof of Theorem 5.8(i) then gives k ≤ min ( s , t ) , so | s − t | ≤ s + t − 2 k = D ′ . The residual dimension is even and positive, so by Theorem 7.6 it lies in { 2 , 4 , 6 } . Then | s − t | ≤ 6 and s − t ≡ 0 ( mod 8 ) force s = t .    □
For D ′ ∈ { 4 , 6 } the identity that the proof of Theorem 7.6 extracts from b holds, and so does the vanishing of the term of a gaugino with itself.
Lemma 7.8. 
Let D ′ ∈ { 4 , 6 } , let X be a Weyl module of V ′ in Δ 1 , and let Y be the Weyl module of Δ 1 for which G does not vanish identically on X × Y ; thus Y = X opp for D ′ = 4 and Y = X for D ′ = 6 . Then for all x , σ ∈ X and y , y ′ ∈ Y
( a ) γ G ( x , y ) σ + γ G ( σ , y ) x = 0 , ( b ) γ G ( x , y ) y ′ + γ G ( x , y ′ ) y = 0 .
Proof. 
Both left-hand sides are trilinear and so ( V C ′ ) -equivariant, because β is Spin ( V ′ ) -invariant, and they are symmetric in ( x , σ ) and in ( y , y ′ ) . They are therefore equivariant linear maps Sym 2 X ⊗ Y → X opp and X ⊗ Sym 2 Y → Y opp , since γ μ ′ exchanges the two Weyl modules of Δ 1 , and it suffices to show that the targets do not occur in the sources [38]. For D ′ = 4 , so ( V C ′ ) ≅ sl 2 ⊕ sl 2 with Weyl modules X = ( 2 , 1 ) and Y = ( 1 , 2 ) , and the vector module ( 2 , 2 ) lies in X ⊗ Y and not in X ⊗ X , as stated. Then Sym 2 X ⊗ Y = ( 3 , 2 ) does not contain X opp = ( 1 , 2 ) , and X ⊗ Sym 2 Y = ( 2 , 3 ) does not contain Y opp = ( 2 , 1 ) . For D ′ = 6 , so ( V C ′ ) ≅ sl 4 with Weyl modules 4 and 4 ¯ and vector module Λ 2 4 ⊂ 4 ⊗ 4 , so Y = X , say X = 4 . By Pieri’s rule Sym 2 4 ⊗ 4 = Sym 3 4 ⊕ S ( 2 , 1 ) 4 , of dimensions 20 and 20, which does not contain 4 ¯ .    □
Proposition 7.9. 
Let D ′ ∈ { 4 , 6 } , take g, f and the rules (39) as above with f antisymmetric, and for each P let X P = K χ ′ ′ , χ ′ ′ = ( − 1 ) k χ P ′ , and Y P be as in Lemma 7.8. Suppose that for each P every gaugino λ C with a C P ≠ 0 takes values in Y P and every other gaugino in X P , and, for D ′ = 4 , that at most one C has a C P ≠ 0 . Then b holds. 
Proof. 
Fix A and P and take ϵ P ∈ K ^ χ P ′ ; by Step 1 of the proof of Theorem 5.6 it suffices to write the left-hand side of (40) as γ μ ′ x J A P μ ′ I with residual J, since both sides then depend only on the K ^ -component of the parameter. The term with C = A appears only if a A P ≠ 0 , so that λ A takes values in Y P , and as f is antisymmetric it is a sum of terms γ ( G ( x , y ) ) y ′ + γ ( G ( x , y ′ ) ) y , which vanish by Lemma 7.8(b). A term with C ≠ A and a C P ≠ 0 has λ C in Y P and, for D ′ = 4 , λ A in X P by the uniqueness of C; for D ′ = 6 also λ A takes values in X P = Y P . It is a sum of terms γ ( G ( x , y ) ) σ with σ ∈ X P , y ∈ Y P , and by Lemma 7.8(a) each equals γ ( u ) x with u = − G ( σ , y ) , which does not depend on x. Collecting these terms gives J A P .    □
Remark 7.10. 
For k = 0 both residual dimensions occur. Four-dimensional N = 1 theory, with the two Weyl halves of the gaugino as independent fields λ 1 ∈ Δ + , λ 2 ∈ Δ − , and six-dimensional ( 1 , 0 ) theory, with the doublet λ i written as two independent fields of one chirality and a A P proportional to ε A P , satisfy the hypotheses of Proposition 7.9. Transported to ( 2 , 2 ) and ( 3 , 3 ) by a complex isometry as in the proof of Theorem 5.12, they have real coefficients and hence real forms with Majorana–Weyl gauginos, by the last paragraph of that proof. In both, N = D ′ − 2 ( N = 2 and N = 4 ), the cross term (41) is absorbed with u = j G ( σ , y ) , j = − 1 , and the total spinor dimension M N = 2 ( D ′ − 2 ) is the one of Theorem 3.3. For k ≥ 1 Proposition 7.9 shows that the cubic terms can be absorbed in ( k + 2 ) + ( k + 2 ) and ( k + 3 ) + ( k + 3 ) , but a complete theory, with transformation rules Φ A P and closure, has not been constructed: no null-projected theory of residual dimension four or six is available to transport by the method of Theorem 5.12, and the direct construction of Theorem 5.14 has not been carried out for these residual dimensions. The number M is not fixed by a and b, because the currents J A P are independent; fixing it requires relating them to a single source. 

7.5. Gauge Brackets That Are Not Lie Brackets

The proofs of Theorem 5.6 and of the Main Theorem use the gauge algebra only through the structure constants in the cubic term of Lemma 5.5. We replace b by the following weaker condition.
(N2′)
The gaugino λ I takes values in K χ ⊗ R for a complex vector space R with basis e I , the gauge field A μ a in a complex vector space a with basis t a , and
δ A μ a = a h a J ϵ ¯ γ μ λ J , δ λ I = Φ ( F ) I ϵ , D μ λ I = ∂ μ λ I + g C I a K A μ a λ K ,
with a , g ≠ 0 , a linear map h : R → a , a bilinear map C : a × R → R , and Φ independent of λ . The bracket is f ( e J , e K ) = C ( h ( e J ) , e K ) = f I J K e I , and it is not zero.
No Jacobi identity, antisymmetry, invariant metric or compactness is assumed, and a need not equal R. For a Lie algebra in the adjoint representation, h = 1 and f is the Lie bracket. For the product of a Jordan algebra, f is symmetric.
Theorem 7.11. 
Let a theory satisfy (N2′), and a, c, d and e with g replaced by R. Then D ′ = D − 2 k ∈ { 2 , 10 } . If f ( ξ , ζ ) ≠ − f ( ζ , ξ ) for some ξ , ζ ∈ R , then D ′ = 2 . 
Proof. 
Lemma 5.5 becomes, with the same proof, the statement that the part of δ ( γ μ D μ λ I ) quadratic in λ is a g f I J K γ μ λ K ( ϵ ¯ γ μ λ J ) . Steps 1 and 2 of the proof of Theorem 5.6 do not involve the bracket, except that by c the vector δ A μ ′ a = a h a J β ( T ϵ , γ μ ′ λ J ) is not identically zero, so that G μ ′ ( x , y ) = β ( x , γ μ ′ y ) is not identically zero; they give, for D ′ ≥ 3 , a vector structure on Σ = K χ with d = D ′ and N = 2 r ′ − 1 . In Step 3 the tensor S I β δ = f I J K λ J β λ K δ is no longer symmetric, but (24), (25) and the computation that follows hold for an arbitrary S I and give
γ α δ μ ′ γ μ ′ γ β − j γ α γ μ ′ γ μ ′ β δ S I β δ = 0 , j = − D ′ − 2 N ,
for every value of the gaugino. Write f I J K = a I J K + s I J K with a antisymmetric and s symmetric in J , K . For J ≠ K , take λ J = y ψ , λ K = σ φ and all other components zero; for J = K take λ J = y ψ + σ φ . In both cases
S I β δ = a I J K ( y β σ δ + σ β y δ ) + s I J K ( y β σ δ − σ β y δ ) ψ φ .
Since (42) is linear in S I and holds for ( y , σ ) and for ( σ , y ) , it holds for both parts separately.
If s ≠ 0 , choose I , J , K with s I J K ≠ 0 . Then (42) holds for all antisymmetric S, on which the second term vanishes because γ μ ′ is symmetric. Hence γ α δ μ ′ γ μ ′ γ β is symmetric in β , δ :
γ α δ μ ′ γ μ ′ γ β = γ α β μ ′ γ μ ′ γ δ .
Contract with γ ˜ ν δ α . By Lemma 3.2(b),(c) the left-hand side gives N γ γ β ν and the right-hand side ( γ μ ′ γ ˜ ν γ μ ′ ) γ β = − ( D ′ − 2 ) γ γ β ν , so N = − ( D ′ − 2 ) , which is impossible. Hence D ′ ≥ 3 forces s = 0 .
If s = 0 , then f ≠ 0 gives I and J ≠ K with a I J K ≠ 0 , the tensors S I span Sym 2 Σ , and the rest of Step 3 gives (6) and, by Theorem 3.3, D ′ = 10 .    □
Corollary 7.12. 
The Main Theorem and Corollary 5.9 hold with b replaced by (N2′), for every nonzero bracket. If the bracket is not antisymmetric, a theory with a Majorana–Weyl gaugino has s = t and s + t − 2 k = 2 . In particular a gauge structure given by a commutative product, such as the product of the exceptional Jordan algebra, carries no such theory of residual dimension greater than two. 
Proof. 
Part (i) of the Main Theorem does not involve the gauge structure, and part (ii) is Theorem 7.11; part (iii) and Corollary 5.9 follow as before. If the bracket is not antisymmetric, D ′ = 2 and | s − t | ≤ 2 give s = t .    □
Theorem 7.11 covers every gauge structure that enters through a covariant derivative bilinear in the gauge field and the gaugino and a variation of the gauge field linear in the gaugino. Structures in which the variation of the gauge field involves further fields, as for three-algebras, where it contains scalars, lie outside it.

7.6. Nonlinear Transformation Rules and Auxiliary Fields

Let the fields be the gaugino λ I , the gauge field A μ I and finitely many Grassmann-even fields H r , the auxiliary fields, in any representation. A supersymmetry transformation is a derivation δ of the algebra of polynomials in the components of the fields and their derivatives, even, commuting with ∂ μ , and linear in the constant parameter ϵ . The gaugino equation E I is an odd polynomial with values in Δ . For a polynomial Q let π 2 ( Q ) be its part made of monomials that are products of exactly two undifferentiated components of λ , with constant coefficients that may contain ϵ . We assume:
(B1)
δ λ has no term independent of the fields;
(B2)
δ A μ = a ϵ ¯ γ μ λ + R μ , where no monomial of R μ is a constant times one undifferentiated component of λ ;
(B3)
no monomial of δ H r is a constant times one undifferentiated component of λ ;
(B4)
E I = γ μ D μ λ I + E ′ I , where E ′ contains no monomial of the form c λ , c A λ or c H λ with a constant c and undifferentiated factors.
Apart from these conditions δ λ , δ A , δ H and E ′ are arbitrary: δ λ may contain any products of λ , F, A, H and their derivatives, and E ′ any terms of order three or more in the gaugino.
Lemma 7.13. 
Under a–d, π 2 ( δ E I ) = a g f I J K γ μ λ K ( ϵ ¯ γ μ λ J ) . 
Proof. 
For a monomial m = ϕ 1 … ϕ n in components of fields and their derivatives, δ m = ∑ i ± ϕ 1 … δ ϕ i … ϕ n . A term contributes to π 2 only if the factors ϕ j with j ≠ i , which are unchanged, are undifferentiated components of λ , and the chosen term of δ ϕ i , multiplied by them, has exactly two undifferentiated components of λ and no other field. If ϕ i = ∂ α ϕ with | α | ≥ 1 , then δ ϕ i = ∂ α δ ϕ , and every term of it either vanishes or contains a derivative of a field, since ϵ is constant. So ϕ i is undifferentiated and m = c ϕ i λ p , a constant times ϕ i and p undifferentiated components of λ .
If ϕ i is a component of λ , then m has p + 1 gaugino factors and is odd, so p ∈ { 0 , 2 , … } . For p = 0 , m = c λ , which does not occur in E by d, since γ μ D μ λ contains λ only differentiated or multiplied by A. For p = 2 the term of δ λ must be independent of the fields, which a excludes, and p ≥ 4 is impossible. If ϕ i is a component of A or of H, oddness of m forces p odd, and the term of δ ϕ i must contain 2 − p ≥ 0 gaugino factors and no other field, so p = 1 and the term is a constant times one undifferentiated component of λ . For H this is excluded by c. For A, by b the term is a ϵ ¯ γ μ λ , and by d the monomials c A λ of E are those of g f I J K γ μ A μ J λ K . Their contribution is the stated expression.    □
The absorption axiom e says that the part of δ E quadratic in λ is γ ( J ∧ v ) ϵ . In a nonlinear theory the source J in the field equation of the gauge field is a polynomial in the fields, and since γ ( · ∧ v ) ϵ is linear with constant coefficients, π 2 ( γ ( J ∧ v ) ϵ ) = γ ( π 2 ( J ) ∧ v ) ϵ . We therefore use
  • for every λ with values in K χ ⊗ g there are J σ I , independent of ϵ , with π 2 ( δ E I ) = γ σ ν 1 … ν k ϵ v ν 1 … ν k J σ I for every ϵ .
Proposition 7.14. 
Let a theory satisfy a, d, a–d and a, and c for the term a ϵ ¯ γ μ λ of δ A μ , with δ λ , δ A , δ H and E otherwise arbitrary. Then D ′ ∈ { 2 , 10 } , and with a Majorana–Weyl gaugino the conclusions of the Main Theorem and of Corollary 5.9 hold. 
Proof. 
By Lemma 7.13, a is the equation (19) = γ ( J ∧ v ) ϵ of Step 1 of the proof of Theorem 5.6. Step 2 uses c only through the pairing G μ ′ ( x , y ) = β ( x , γ μ ′ y ) that appears in the term a ϵ ¯ γ μ λ , and d only through the translation bracket, which does not involve the fields. Step 3 and the proof of Theorem 5.8 then apply unchanged.    □
Remark 7.15. 
Conditions a–d mark the mechanisms that could change the result. A term c λ in the gaugino equation together with a term of δ λ quadratic in λ, a field-independent term in δ λ (a nonlinear realization of supersymmetry) together with a term cubic in λ in the gaugino equation, or a further boson whose variation contains ϵ ¯ λ without derivative and which couples to λ without derivative, all contribute to π 2 ( δ E ) . Auxiliary fields enter in a second way. If δ E is only required to vanish modulo all field equations, the field equations E H r of the auxiliary fields may appear. When these equations vanish at zero fields, they contribute to π 2 ( δ E ) only through constant multiples of their own parts π 2 ( E H r ) , with coefficients linear in ϵ. If no auxiliary field equation contains a product of two undifferentiated gauginos, this contribution vanishes and Proposition 7.14 applies. If one does, the absorption condition constrains nothing: an auxiliary field with values in Hom ( Δ χ ′ , Δ ) ⊗ g whose equation sets it equal to π 2 ( δ E ) , read as a linear map of ϵ, matches the cubic terms in every dimension. For such theories the obstruction moves to the supersymmetry variation of the auxiliary field equations and to the closure of the algebra on the auxiliary fields, which the present methods do not decide. 

7.7. Scope

Remark 7.16. 
The results of this section concern theories whose fields and transformation rules are written down, and no statement about the superalgebra alone can exclude every theory in 27 + 3 . The superalgebra (30) acts on functions of Grassmann coordinates θ α and of commuting coordinates y μ 1 … μ p , antisymmetric in their indices, for p = 3 , 7 , 11 , 15 : the operators
Q α = ∂ ∂ θ α + 1 2 ∑ p ∑ μ 1 < … < μ p ( γ μ 1 … μ p θ ) α ∂ ∂ y μ 1 … μ p
satisfy (30) with Z μ 1 … μ p = ∂ / ∂ y μ 1 … μ p , because the matrices γ α β μ 1 … μ p of these degrees are symmetric. Such a realization carries no dynamics, and nothing in this paper bears on it. What is excluded is a field content with specified transformation rules: one constrained gaugino with the null contraction of the translation bracket (Theorem 5.8), with any bilinear gauge bracket (Theorem 7.11) and with rules and field equations nonlinear in the gaugino under the conditions a–d (Proposition 7.14); one gaugino, constrained or not, with any contraction of the source (Theorem 7.3); and any number of gauginos, mixed by their transformation rules or not (Theorem 7.6 and Corollary 7.7). Outside these results lie auxiliary fields whose equations contain a product of two undifferentiated gauginos, the mechanisms that a–d exclude (Remark 7.15), gauge structures in which the variation of the gauge field involves further fields, and mechanisms carried by the 11-brane. 
The results leave the algebraic structures of Refs. [23,26,27] untouched. The superalgebras (3.5)–(3.8) of Ref. [23] exist by Proposition 2.1, and so do the cosets and Jordan-algebraic constructions built on them. Several statements made alongside them fail for the reasons used above, the signature of an invariant form and the symmetric square of a Weyl spinor: f 4 ( − 20 ) has no nonzero homomorphism into so ( 25 , 1 ) , no real form of so ( 15 ) in so ( 27 , 3 ) splits the Majorana–Weyl spinor into odd forms, and no Lie superalgebra with odd part the 128 of so ( 12 , 4 ) has an anticommutator with a component in so ( 12 , 4 ) (Appendix H). The obstruction concerns realization by the specific null-projected field theory defined here. Applications invoking a 27 + 3 theory by this particular transfer, including a worldvolume theory for M-theory [24] and the dark matter model of Ref. [25], require a construction outside the assumptions of this section, or a change to the claimed transfer; by Theorem 7.3, changing only the contraction of the source does not suffice. Every theory satisfying (N1)–(N5) obeys the necessary condition s − t ∈ { 8 , 0 , − 8 } . For residual dimension greater than two its cubic fermion structure satisfies the ten-dimensional identity, and by Theorem 5.12 a theory exists in every signature with residual dimension ten that the Main Theorem leaves. For residual dimension two, which the Main Theorem allows when s = t , Theorem 5.14 gives a theory in every such signature, with vanishing cubic terms. This agrees with the view, expressed already in Ref. [21], that the null projections of super Yang–Mills theories beyond ten dimensions encode ten-dimensional theories with a covariant presentation of the extra directions.
Two remarks fix the scope of the results. The Main Theorem concerns theories with a single gaugino of definite chirality and the transformation rules b. Theorem 7.11 allows any bilinear gauge bracket, Proposition 7.14 allows rules and field equations nonlinear in the gaugino and auxiliary fields, subject to a–d, and Theorem 7.6 allows any number of gauginos mixed by their transformation rules, which then forces residual dimension two, four or six and, for Majorana–Weyl gauginos, s = t ; what remains open is listed in Remark 7.16. Theorem 6.1 assumes unitarity and hypothesis (H), which is why it covers only the single-time family: we do not assume the positive-energy unitary representation used in the counting argument for the several-time signatures. The algebraic Theorem 5.8 is the result applied there.
Remark 7.17. 
The axioms a–e are conditions on transformation rules and field equations, and they can be checked only for a theory whose equations are written down. By Corollary 5.9 no theory in 27 + 3 satisfies them, by Corollary 7.4 none satisfies a–c and absorbs the cubic terms with any other contraction of the source, by Corollary 7.7 adding gauginos, mixed by their transformation rules or not, does not help, and by Corollary 7.12 neither does a gauge bracket that is not a Lie bracket. The content of Proposition 7.1 is that the one field-theoretic construction proposed for 27 + 3 , the repetition of Nishino’s proof in Ref. [23], would satisfy them, and therefore does not exist. The superalgebra (2.1) of Ref. [24] lists the possible central charges, but it contains no gaugino constraint, transformation rule or field equation, so by itself it can neither satisfy nor violate the axioms. A theory in 27 + 3 that realizes it by other means, with auxiliary fields whose equations contain gaugino bilinears, rules outside a–d, a gaugino constraint other than annihilation by null vectors, a gauge structure in which the variation of the gauge field involves further fields, or a mechanism carried by the 11-brane, is not decided here. The Magic Star algebras of Ref. [27], which are not Lie algebras, fall under Corollary 7.12 if they enter through a covariant derivative and a variation of the gauge field of the form (N2′); we have not checked whether the constructions there do so. The worldvolume theory invoked in Ref. [24] is the theory of Sezgin in 11 + 3  [20], Nishino’s case n = 0 with a frame of rank two. It satisfies a–e with residual space R 9 , 1 and s − t = 8 , in agreement with Theorem 5.8, and nothing in this paper bears on whether the 11-brane in 27 + 3 carries it. 

8. Conclusions

Supersymmetric Yang–Mills theories built with null projections in signatures with several times are governed by an invariant that the Clifford algebra does not see. The bound min ( s , t ) limits the number of null directions that can be projected out of a real gaugino, and a residual dimension greater than two must equal ten because the absorption equation forces the ten-dimensional Fierz identity. The dimension count N = 2 ( d − 2 ) of Theorem 3.3 makes the second statement rigid, and together the two leave only the signatures with s − t ∈ { 8 , 0 , − 8 } . The families ( 9 + 8 n ) + 1 , ( 10 + 8 n ) + 2 , ( 11 + 8 n ) + 3 and ( 12 + 8 n ) + 4 with n ≥ 1 , including 27 + 3 and 19 + 3 , are excluded; in particular the repetition of Nishino’s proof proposed for these signatures fails for every rank of the null frame (Proposition 7.1), and no other contraction of the source in the field equation of the gauge field repairs it (Theorem 7.3). The conditions are also sufficient. For residual dimension ten Nishino’s theories carry over to every signature with s − t ∈ { 8 , 0 , − 8 } and a real frame of the same rank, among them 9 + 9 and 9 + 17 (Theorem 5.12), and for residual dimension two his transformations give theories in every signature ( k + 1 ) + ( k + 1 ) , with vanishing cubic terms and off-shell closure (Theorem 5.14). Several gauginos that transform with their own parameters are impossible once the residual dimension is three or more (Proposition 7.5), and gauginos mixed by their transformation rules need residual dimension two, four or six, hence s = t (Corollary 7.7). The conclusions hold for every bilinear gauge bracket (Theorem 7.11) and for rules and field equations nonlinear in the gaugino under the conditions a–d (Proposition 7.14). For ( 9 + 8 n ) + 1 , Theorem 6.1 and Lemma 6.2 exclude massless vector multiplets and finite multiplets made only of scalar, spinor and vector fields under the stated unitarity and tensor-charge hypotheses. A supersymmetric gauge theory in these signatures would need auxiliary fields whose equations contain gaugino bilinears, one of the mechanisms excluded by a–d, or a gauge structure in which the variation of the gauge field involves further fields, and it would then no longer be a repetition of Nishino’s proof. Two questions remain open within the present framework: whether such auxiliary fields can close the algebra, and whether complete theories with mixed gauginos exist for residual dimension four or six when k ≥ 1 .

Funding

The author received no funding for this work.

Data Availability Statement

The general results are proved in the text; numerical samples are supplementary. The plotted and tabulated residuals of Appendix B retain the archived reference run of the 34 checks of verify.py, which a fresh run on 24 October 2026 reproduced. The 24 checks of Appendix D are made by transfer_check.py, the 32 checks of Appendix E by contraction_check.py, the 50 checks of Appendix F by residual_two_check.py, and the 122 checks of Appendix G by mixed_check.py; all five programs build their gamma matrices with clifford.py. The finite arithmetic of Appendix H is rechecked by f4_signature.py, spinor_checks.py and arithmetic_checks.py, which are in the folder null-sym-appendix of [47]. Figure 7 and Figure 11 are drawn by make_v4_figures.py. The complete source, the five numerical programs and their verification logs are archived with every earlier version of this manuscript under one concept record [46]. The schematics of Figure 2 and Figure 5 are drawn in the source with TikZ.

Conflicts of Interest

The author declares no competing interests.

Appendix A. The Oscillator Decomposition

We prove Lemma 5.2. The P i ↑ are commuting projections, so the products Π a = ∏ i a i P i ↑ + ( 1 − a i ) P i ↓ are orthogonal idempotents with ∑ a Π a = 1 , and Δ = ⨁ a Δ a with Δ a = Π a Δ . For j ≠ i , * * n * * i anticommutes with * * n * * j and * * m * * j and therefore commutes with P j ↑ . From * * n * * i 2 = 0 and * * n * * i * * m * * i + * * m * * i * * n * * i = 2 we get * * n * * i * * m * * i * * n * * i = 2 * * n * * i , that is P i ↑ * * n * * i = * * n * * i and * * n * * i P i ↑ = 0 . Hence * * n * * i annihilates Δ a when a i = 1 and maps Δ a into Δ a + e i when a i = 0 ; on the latter * * m * * i * * n * * i = 2 P i ↓ = 2 , so the map is injective. The same argument with * * m * * i gives an injection Δ a + e i → Δ a , so both maps are isomorphisms. This proves (ii), and all the Δ a have the same dimension 2 r / 2 k .
Each γ μ ′ anticommutes with every b i and b i † , hence commutes with every P i ↑ and preserves every Δ a . Therefore Δ a is a module over Cl ( V C ′ ) , which for even D ′ is a simple algebra of dimension 2 2 r − 2 k with a unique irreducible module, of dimension 2 r − k . As dim Δ a = 2 r − k , Δ a is that module. This proves (i).
For (iii), up to the sign ( − 1 ) k ( k − 1 ) / 2 , T is the composition of the isomorphisms of (ii) along the path 0 → e 1 → e 1 + e 2 → … → 1 , it annihilates every Δ a with a ≠ 0 by (ii), and γ μ ′ anticommutes with each of its factors. For (iv), by (3), B ( P i ↑ ψ , χ ) = 1 2 θ 2 B ( ψ , * * m * * i * * n * * i χ ) = B ( ψ , P i ↓ χ ) . So if ψ ∈ Δ a , χ ∈ Δ b and a i = b i = 1 , then B ( ψ , χ ) = B ( P i ↑ ψ , χ ) = B ( ψ , P i ↓ χ ) = 0 , and similarly if a i = b i = 0 . For (v), choose an orthonormal basis of V C consisting of a basis of V C ′ and the vectors e i = ( n i + m i ) / 2 , f i = ( n i − m i ) / 2 , with e i 2 = 1 and f i 2 = − 1 . Then * * e * * i * * f * * i = 1 2 ( * * m * * i * * n * * i − * * n * * i * * m * * i ) = P i ↓ − P i ↑ , and the chirality operator, proportional to the product of all basis gamma matrices, is Γ * = ± Γ * ′ ∏ i ( P i ↓ − P i ↑ ) , with the sign fixed by Γ * 2 = 1 . On Δ a the product equals ( − 1 ) | a | .
.1 (Null-frame current).Lemma Let λ 1 , λ 2 ∈ K . After identifying vectors and covectors with the metric, the ( k + 1 ) -form
ω μ 0 … μ k = B ( λ 1 , γ μ 0 … μ k λ 2 )
has the form v ∧ j for a residual one-form j ∈ ( V ′ ) C * . 
Proof. 
Evaluate ω on the null-frame basis n i , m i and a basis of V ′ . Both spinors are in the fully occupied corner. By Lemma 5.2(iv), a Clifford word pairing nontrivially with the first spinor must take the second to the empty corner. It must therefore contain every lowering operator * * m * * i . A word of degree k + 1 can do this only by using exactly these k operators and one residual generator: an additional raising operator would require another lowering operator and hence at least k + 2 factors. Antisymmetrization preserves this occupancy count. Thus a nonzero component has one argument m i for every i and one residual argument. Since n i · m j = δ i j , these are exactly the components of v ∧ j .    □

Appendix B. Numerical Checks in Explicit Gamma Matrices

The program verify.py in the repository builds the gamma matrices of R s , t as Kronecker products of Pauli matrices, multiplying the first t Euclidean generators by i, and checks the Clifford relations. It finds a charge conjugation matrix C with C γ μ C − 1 = ± ( γ μ ) T among the products of the real or of the imaginary Euclidean generators, and builds the vector forms of Definition 3.1 from it. It then runs the following checks, all of which pass (34 checks, no failures; runtime depends on the numerical libraries and hardware).
(a)
Proposition 2.1: the dimension count for r = 2 , … , 20 , and the form degrees of the four families (2) for n = 0 , 1 , 2 .
(b)
The identity (1) on random spinors: it holds in ( 9 , 1 ) , ( 5 , 5 ) and ( 1 , 9 ) (relative residual at most 1.3 × 10 − 16 ) and fails in ( 17 , 1 ) , ( 13 , 5 ) , ( 9 , 9 ) and ( 25 , 1 ) (relative residual between 0.09 and 0.36 ).
(c)
The identity (6) with j fitted by least squares: j = − 0.5000 in ( 9 , 1 ) and ( 5 , 5 ) ; no j fits in ( 17 , 1 ) and ( 25 , 1 ) (residual at least 0.99 ).
(d)
Lemma 3.2(b),(c) in ( 9 , 1 ) and ( 17 , 1 ) , exactly.
(e)
The absorption equation (21), solved by least squares for J over random λ ∈ K and ϵ ∈ K ^ with explicit real frames * * n * * i = ( γ space + γ time ) / 2 ; see Table A1 and Figure 6.
(f)
Lemma 2.3 in ( 17 , 1 ) with a complex frame of rank four: the complex conjugate of each n i is m i , and * * n * * i * * m * * i + * * m * * i * * n * * i = 2 , so the occupied corner contains no Majorana spinor.
(g)
Lemma 2.2: real null frames of rank min ( s , t ) satisfying (12).
(h)
Theorem 6.1: the counts of Figure 8 for n = 0 , 1 , 2 , 3 .
Table A1. The absorption equation (21) in explicit gamma matrices. The residual is the relative norm of the least-squares remainder over three random draws. These samples are consistent with the analytic obstruction of Theorem 5.6 for D ′ > 2 ; the complex frame in ( 17 , 1 ) is the case of Remark 5.10.
Table A1. The absorption equation (21) in explicit gamma matrices. The residual is the relative norm of the least-squares remainder over three random draws. These samples are consistent with the analytic obstruction of Theorem 5.6 for D ′ > 2 ; the complex frame in ( 17 , 1 ) is the case of Remark 5.10.
( s , t ) k frame D ′ dim K χ residual solvable
( 9 , 1 ) 0 – 10 16 3.0 × 10 − 15 yes
( 10 , 2 ) 1 real 10 16 1.6 × 10 − 15 yes
( 11 , 3 ) 2 real 10 16 1.5 × 10 − 15 yes
( 12 , 4 ) 3 real 10 16 1.4 × 10 − 15 yes
( 13 , 5 ) 4 real 10 16 2.6 × 10 − 15 yes
( 6 , 6 ) 1 real 10 16 1.4 × 10 − 15 yes
( 7 , 7 ) 2 real 10 16 2.8 × 10 − 15 yes
( 17 , 1 ) 0 – 18 256 0.98 no
( 18 , 2 ) 1 real 18 256 0.99 no
( 19 , 3 ) 0 – 22 1024 1.0 no
( 19 , 3 ) 2 real 18 256 0.99 no
( 20 , 4 ) 1 real 22 1024 1.0 no
( 20 , 4 ) 3 real 18 256 0.98 no
( 17 , 1 ) 4 complex 10 16 1.7 × 10 − 15 yes (complex sector)
( 19 , 3 ) 2 complex 18 256 0.97 no

Appendix C. Symmetric Squares for the Four Families

Table A2 lists Sym 2 Δ + , by Proposition 2.1, for the families (2) with n = 0 , 1 , 2 , together with the residual dimension s − t of a real null frame of the largest rank k = t . The momentum Λ 1 occurs only in the first family. Since D ′ decreases as k grows, a residual dimension of ten is reachable only when s − t ≤ 10 , that is only for n = 0 .
Table A2. Symmetric squares of the Weyl spinor for the families (2), n = 0 , 1 , 2 . The last column is s − t = s + t − 2 min ( s , t ) , the residual dimension of a real null frame of rank k = t . For n = 0 Nishino’s frames have k = t − 1 and D ′ = 10 ; a frame with k = t gives D ′ = 8 , which Theorem 5.6 excludes. For n ≥ 1 every admissible frame has D ′ ≥ 16 .
Table A2. Symmetric squares of the Weyl spinor for the families (2), n = 0 , 1 , 2 . The last column is s − t = s + t − 2 min ( s , t ) , the residual dimension of a real null frame of rank k = t . For n = 0 Nishino’s frames have k = t − 1 and D ′ = 10 ; a frame with k = t gives D ′ = 8 , which Theorem 5.6 excludes. For n ≥ 1 every admissible frame has D ′ ≥ 16 .
( s , t ) D dim Δ + degrees in Sym 2 Δ + dim Sym 2 Δ + D ′ at k = t
( 9 , 1 ) 10 16 5, 1 136 8
( 10 , 2 ) 12 32 6, 2 528 8
( 11 , 3 ) 14 64 7, 3 2 080 8
( 12 , 4 ) 16 128 8, 4, 0 8 256 8
( 17 , 1 ) 18 256 9, 5, 1 32 896 16
( 18 , 2 ) 20 512 10, 6, 2 131 328 16
( 19 , 3 ) 22 1 024 11, 7, 3 524 800 16
( 20 , 4 ) 24 2 048 12, 8, 4, 0 2 098 176 16
( 25 , 1 ) 26 4 096 13, 9, 5, 1 8 390 656 24
( 26 , 2 ) 28 8 192 14, 10, 6, 2 33 558 528 24
( 27 , 3 ) 30 16 384 15, 11, 7, 3 134 225 920 24
( 28 , 4 ) 32 32 768 16, 12, 8, 4, 0 536 887 296 24

Appendix D. The Transfer in Explicit Gamma Matrices

The program transfer_check.py in the repository checks Proposition 7.1 in the gamma matrices of Appendix B, which in thirty dimensions have size 2 15 = 32 768 . Real frames pair the i-th timelike generator with a spacelike one, * * n * * i = ( γ space + γ time ) / 2 , as in Nishino’s eqs. (4.5) and (4.6). Dimensions of corners are traces of the projections ∏ i P i ↑ restricted to a chirality, so they are exact integers. For the cubic terms we take x = T ϵ with a random ϵ ∈ K ^ + , use the residual pairing G μ ′ ( T ϵ , T ϵ ′ ) = B ( ϵ , γ μ ′ T ϵ ′ ) of Step 2 of the proof of Theorem 5.6, and evaluate Y = ∑ μ ′ η μ ′ μ ′ G μ ′ ( x , x ) γ μ ′ x , which vanishes for every x exactly when the identity (1) holds; the identity (6) is tested with j fitted by least squares. Sezgin’s 11 + 3 and Nishino’s source signature 19 + 11 serve as controls. All 24 checks pass; Table A3 lists them. They confirm the written proof and do not replace it.
Table A3. Checks of Proposition 7.1 in explicit gamma matrices. Relative errors of the identities that hold are below 3 × 10 − 14 .
Table A3. Checks of Proposition 7.1 in explicit gamma matrices. Relative errors of the identities that hold are below 3 × 10 − 14 .
signature check result
all four Clifford relations, spinor dimension 2 D / 2 hold
( 27 , 3 ) , ( 19 , 11 ) degrees symmetric on Δ + × Δ + 3 , 7 , 11 , 15 ; antisymmetric 1 , 5 , 9 , 13 ; even degrees zero
( 19 , 3 ) degrees symmetric on Δ + × Δ + 3 , 7 , 11 ; antisymmetric 1 , 5 , 9 ; even degrees zero
( 27 , 3 ) , ( 19 , 3 ) translation bracket for real frames of rank 1 and 3 vanishes identically
( 27 , 3 ) , k = 3 real frame, dim K + , D ′ exists; 2 048 ; 24
( 27 , 3 ) , k = 2 real frame, dim K + , D ′ exists; 4 096 ; 26
( 27 , 3 ) , k = 2 δ A μ ′ , ξ μ ′ nonzero, ξ symmetric yes (asymmetry 8 × 10 − 16 )
( 27 , 3 ) , k = 2 identity (1); identity (6) for some j fail ( | Y | relative 0.11 ; residual 0.999 )
( 19 , 3 ) , k = 2 dim K + , D ′ ; δ A μ ′ , ξ μ ′ 256, 18; nonzero, symmetric
( 19 , 3 ) , k = 2 identity (1); identity (6) for some j fail ( | Y | relative 0.38 ; residual 0.99 )
( 27 , 3 ) , complex k = 10 B n ¯ i B − 1 = * * m * * i ; Majorana spinors in the corner yes; none (corner of dimension 32)
( 11 , 3 ) , k = 2 dim K + , D ′ ; identities (1), (6) 16, 10; hold, j = − 0.5000
( 19 , 11 ) , k = 10 dim K + , D ′ ; identities (1), (6) 16, 10; hold, j = − 0.5000

Appendix E. Sufficiency and Other Contractions in Explicit Gamma Matrices

The program contraction_check.py in the repository checks Theorems 5.12 and 7.3 in the gamma matrices of Appendix B. For a real frame of rank k and a fixed multivector u it takes two random gauginos λ 1 , λ 2 in the occupied corner, forms the operator ϵ ↦ ∑ μ η μ μ γ μ λ 2 B ( ϵ , γ μ λ 1 ) + γ μ λ 1 B ( ϵ , γ μ λ 2 ) on K ^ χ ′ , which is the polarisation of the left-hand side of (36), and fits it by least squares with the operators γ ( e a ∧ u ) , a = 1 , … , D . A relative residual at machine precision means that a source J ∧ u absorbs the cubic terms for this pair of gauginos; a residual of order one means that no source of this form does. The contractions are u = 1 , products of spacelike basis vectors, and u = n 1 ∧ … ∧ n k ∧ w with w a product of residual basis vectors; indices refer to the orthonormal basis e 0 , … , e D − 1 in which e 0 , … , e t − 1 are timelike, and the frame is n i = ( e t + i − 1 + e i − 1 ) / 2 , m i = ( e t + i − 1 − e i − 1 ) / 2 . For Theorem 5.12 the program also evaluates the residual bilinears δ A μ ′ and ξ μ ′ of c and d for random parameters and checks that ξ is symmetric. All 32 checks pass; Table A4 lists them. They confirm the written proofs and do not replace them.
Table A4. Checks of Theorems 7.3 (upper part) and 5.12 (lower part) in explicit gamma matrices. In the lower part the asymmetry of ξ is below 4 × 10 − 14 and the residuals of absorption below 3 × 10 − 15 .
Table A4. Checks of Theorems 7.3 (upper part) and 5.12 (lower part) in explicit gamma matrices. In the lower part the asymmetry of ξ is below 4 × 10 − 14 and the residuals of absorption below 3 × 10 − 15 .
signature, rank contraction u D ′ result (relative residual)
( 9 , 1 ) , ( 5 , 5 ) , k = 0 1 10 absorbed ( 2 × 10 − 16 )
( 9 , 1 ) , k = 0 e 9 10 not absorbed ( 1.00 )
( 7 , 1 ) , k = 0 e 1 ∧ … ∧ e 7 ; 1 8 absorbed ( 3 × 10 − 16 ); not absorbed ( 1.00 )
( 11 , 1 ) , k = 0 1; e 1 ; e 1 ∧ … ∧ e 9 12 not absorbed ( 1.00 ; 0.93 ; 0.98 )
( 11 , 3 ) , k = 0 e 3 ∧ … ∧ e 12 14 not absorbed ( 1.00 )
( 9 , 7 ) , k = 0 e 7 16 not absorbed ( 1.00 )
( 11 , 3 ) , k = 2 n 1 ∧ n 2 (Sezgin); n 1 ∧ n 2 ∧ e 13 10 absorbed ( 1 × 10 − 15 ); not absorbed ( 1.00 )
( 11 , 3 ) , k = 1 n 1 ; n 1 ∧ e 12 12 not absorbed ( 1.00 ; 0.91 )
( 19 , 3 ) , k = 2 n 1 ∧ n 2 ; n 1 ∧ n 2 ∧ e 21 18 not absorbed ( 0.99 ; 1.00 )
( 19 , 3 ) , k = 3 n 1 ∧ n 2 ∧ n 3 ∧ e 12 ∧ … ∧ e 21 16 not absorbed ( 1.00 )
( 19 , 3 ) , k = 1 n 1 ∧ e 14 ∧ … ∧ e 21 20 not absorbed ( 1.00 )
( 9 , 1 ) , ( 5 , 5 ) , ( 1 , 9 ) , k = 0 δ A μ ′ , ξ μ ′ ; u = 1 10 nonzero, symmetric; absorbed
( 7 , 7 ) , k = 2 ; ( 3 , 11 ) , k = 2 δ A μ ′ , ξ μ ′ ; u = n 1 ∧ n 2 10 nonzero, symmetric; absorbed
( 2 , 10 ) , k = 1 δ A μ ′ , ξ μ ′ ; u = n 1 10 nonzero, symmetric; absorbed
( 9 , 9 ) , k = 4 δ A μ ′ , ξ μ ′ ; u = n 1 ∧ … ∧ n 4 10 nonzero, symmetric; absorbed

Appendix F. Checks for Residual Dimension Two and Several Gauginos

The program residual_two_check.py checks Theorem 5.14 and Proposition 7.5 in the gamma matrices of Appendix B, with the real frame n i = ( e t + i − 1 + e i − 1 ) / 2 , m i = ( e t + i − 1 − e i − 1 ) / 2 of Appendix E. For Nishino’s transformations (27) with random field strengths obeying F ( n i , · ) = 0 it checks that δ λ stays in K, that the residual components of δ A and ξ are nonzero and that ξ is symmetric, and that the commutator of two supersymmetries with Grassmann parameters is the translation by ξ : on A for every vector in V ′ and in the span of the n i , and on λ for random derivatives D μ λ in K with D n i λ = 0 , both on shell and, for D ′ = 2 , off shell. The commutators are written as symmetric parts of explicit matrices, and the coefficient of the translation, which is fixed by the closure on A, is the same on λ . It then evaluates the cubic term ∑ μ η μ μ γ μ λ 2 B ( ϵ , γ μ λ 1 ) + γ μ λ 1 B ( ϵ , γ μ λ 2 ) for every ϵ in the Weyl module of the parameter, and the cross term ∑ μ η μ μ γ μ σ B ( ϵ , γ μ y ) of two independent gauginos σ , y ∈ K . Nishino’s theories in ( 9 , 1 ) with k = 0 and ( 10 , 2 ) with k = 1 serve as controls. All 50 checks pass; Table A5 lists them. They confirm the written proofs and do not replace them.
Table A5. Checks of Theorem 5.14 (upper part, with Nishino’s theories as controls) and of Proposition 7.5 (lower part) in explicit gamma matrices. A nonzero cross term is of the size of the product of the two gauginos, at least 0.9 times it in every random sample.
Table A5. Checks of Theorem 5.14 (upper part, with Nishino’s theories as controls) and of Proposition 7.5 (lower part) in explicit gamma matrices. A nonzero cross term is of the size of the product of the two gauginos, at least 0.9 times it in every random sample.
signature, rank D ′ check result
( 1 , 1 ) , k = 0 ; ( 2 , 2 ) , k = 1 2 δ λ ∈ K ; (N3), (N4) holds; nonzero, symmetric
( 3 , 3 ) , k = 2 ; ( 4 , 4 ) , k = 3 2 closure on A translation by ξ (residual ≤ 3 × 10 − 16 )
closure on λ , on and off shell translation by ξ (residual ≤ 3 × 10 − 16 )
cubic term; cross term both vanish ( ≤ 4 × 10 − 16 )
( 9 , 1 ) , k = 0 ; ( 10 , 2 ) , k = 1 10 δ λ ∈ K ; (N3), (N4); closure on A and λ hold (residual ≤ 3 × 10 − 16 )
cubic term; cross term both nonzero
( 3 , 1 ) , ( 2 , 2 ) , k = 0 ; ( 3 , 3 ) , k = 1 4 cross term nonzero
( 5 , 1 ) , ( 3 , 3 ) , k = 0 ; ( 4 , 4 ) , k = 1 6 cross term nonzero
( 7 , 1 ) , k = 0 ; ( 5 , 5 ) , k = 1 ; ( 6 , 6 ) , k = 2 8 cross term nonzero
( 9 , 1 ) , k = 0 10 cross term nonzero

Appendix G. Checks for Mixed Gauginos and General Brackets

The program mixed_check.py uses the gamma matrices of Appendix B and the real frames of Appendix E. For two independent gauginos σ ∈ K χ A , y ∈ K χ B and a parameter of chirality χ ′ it fits the cross term ∑ μ η μ μ γ μ σ B ( ϵ , γ μ y ) by the least-squares source γ ( J ∧ v ) ϵ , J ∈ V C , simultaneously for a basis of parameters, in every chirality configuration with B ( ϵ , γ μ ′ y ) ≠ 0 . For a single gaugino it separates the parts of the cubic term that come from the antisymmetric and from the symmetric part of the bracket, and for residual dimension four and six it evaluates the term of a gaugino with itself. It also confirms that N = D ′ − 2 with N = 2 D ′ / 2 − 1 has only the solutions D ′ = 4 , 6 among the even D ′ ≤ 64 . All 122 checks pass; Table A6 lists them. They confirm the written proofs and do not replace them.
Table A6. Checks of Theorem 7.6, Lemma 7.8 and Proposition 7.9 (upper and middle parts) and of Theorem 7.11 (lower part) in explicit gamma matrices. In the upper part the parameters run over K ^ χ ′ and, in a second run for D ′ = 4 , 6 , over the whole Weyl module, with the same results; J ′ is the residual part of the fitted current.
Table A6. Checks of Theorem 7.6, Lemma 7.8 and Proposition 7.9 (upper and middle parts) and of Theorem 7.11 (lower part) in explicit gamma matrices. In the upper part the parameters run over K ^ χ ′ and, in a second run for D ′ = 4 , 6 , over the whole Weyl module, with the same results; J ′ is the residual part of the fitted current.
signature, rank D ′ check result
( 2 , 2 ) , ( 3 , 1 ) , k = 0 ; ( 3 , 3 ) , k = 1 ; ( 4 , 4 ) , k = 2 4 cross term, χ A = ( − 1 ) k χ ′ absorbed ( ≤ 10 − 15 ), J ′ ‖ G ( σ , y )
cross term, other chirality not absorbed (residual 1.000 )
( 3 , 3 ) , ( 5 , 1 ) , k = 0 ; ( 4 , 4 ) , k = 1 6 cross term, χ A = ( − 1 ) k χ ′ absorbed ( ≤ 10 − 15 ), J ′ ‖ G ( σ , y )
cross term, other chirality not absorbed (residual 1.000 )
( 2 , 2 ) , k = 0 ; ( 3 , 3 ) , k = 1 ; ( 4 , 4 ) , k = 2 4 term of a gaugino with itself vanishes ( ≤ 10 − 15 )
( 3 , 3 ) , k = 0 ; ( 4 , 4 ) , k = 1 ; ( 5 , 5 ) , k = 2 6 term of a gaugino with itself vanishes ( ≤ 3 × 10 − 14 )
( 4 , 4 ) , k = 0 ; ( 5 , 5 ) , k = 1 8 cross term not absorbed (residual ≥ 0.55 )
( 9 , 1 ) , ( 5 , 5 ) , k = 0 ; ( 10 , 2 ) , k = 1 10 cross term not absorbed (residual ≥ 0.55 )
( 6 , 6 ) , k = 0 ; ( 9 , 9 ) , k = 2 ; ( 8 , 8 ) , k = 0 12, 14, 16 cross term not absorbed (residual ≥ 0.55 )
( 9 , 1 ) , ( 5 , 5 ) , k = 0 ; ( 10 , 2 ) , ( 6 , 6 ) , k = 1 10 antisymmetric bracket; symmetric bracket absorbed ( ≤ 10 − 15 ); not absorbed ( 1.000 )
( 1 , 1 ) , k = 0 ; ( 2 , 2 ) , k = 1 ; ( 3 , 3 ) , k = 2 2 antisymmetric bracket; symmetric bracket both vanish

Appendix H. Further Obstructions in the 27+3 and Grand-Unified Constructions

Several algebraic statements made alongside the proposals of Section 7 can be decided with the tools of Section 2: the signature of an invariant form, the Witt index, the branching of spinors and the symmetric square of a Weyl spinor. We take up five of them, from Refs. [24,25,28]; equation numbers refer to the arXiv versions listed in the bibliography. All five are statements about Lie algebras, Lie superalgebras and their modules. They do not involve the axioms a–e, and the scope fixed in Remark 7.16 is unchanged: theories outside those axioms are not excluded by anything below. A real Lie algebra h and its modules are compared through their complexifications h C ; for a complex orthogonal space W of dimension 2 m or 2 m + 1 we write Δ ( W ) for its Dirac module and Δ ± ( W ) for its Weyl modules when dim W is even, as in Section 2. For so ( 8 ) the labels 8 s , 8 c denote Δ + , Δ − and 8 v the vector module, up to triality.
Lemma A.1. 
Let V = V 1 ⊥ … ⊥ V m be a complex orthogonal space with dim V i = 2 r i even.
(i)
As a module of so ( V 1 ) ⊕ … ⊕ so ( V m ) , Δ ( V ) ≅ Δ ( V 1 ) ⊗ … ⊗ Δ ( V m ) , and the chirality operators can be normalized so that
Δ ± ( V ) ≅ ⨁ ε 1 … ε m = ± 1 Δ ε 1 ( V 1 ) ⊗ … ⊗ Δ ε m ( V m ) .
(ii)
If V = W ⊥ C u with u · u = 1 , then each of Δ ± ( V ) restricts to so ( W ) as the spinor module Δ ( W ) . 
Proof. (i) It suffices to treat m = 2 . Let γ 1 , γ 2 act on Δ ( V 1 ) , Δ ( V 2 ) , and let Γ 1 be the chirality operator of V 1 , with Γ 1 2 = 1 . On Δ ( V 1 ) ⊗ Δ ( V 2 ) put γ ( v 1 ) = γ 1 ( v 1 ) ⊗ 1 and γ ( v 2 ) = Γ 1 ⊗ γ 2 ( v 2 ) . Since Γ 1 anticommutes with every γ 1 ( v 1 ) , these operators satisfy the Clifford relations of V; the module has dimension 2 r 1 + r 2 , so it is the Dirac module of the simple algebra Cl ( V ) . The quadratic elements are γ ( v 1 ) γ ( v 1 ′ ) = γ 1 ( v 1 ) γ 1 ( v 1 ′ ) ⊗ 1 and γ ( v 2 ) γ ( v 2 ′ ) = 1 ⊗ γ 2 ( v 2 ) γ 2 ( v 2 ′ ) , so so ( V 1 ) ⊕ so ( V 2 ) acts by the tensor product action. The product of the γ of an orthonormal basis of V is, up to a nonzero factor, ( Γ 1 ⊗ 1 ) ( Γ 1 2 r 2 ⊗ Γ 2 ) = Γ 1 ⊗ Γ 2 , so Γ * = ± Γ 1 ⊗ Γ 2 , and the sign can be absorbed into Γ 1 .
(ii) For w ∈ W put c ( w ) = i γ ( w ) γ ( u ) . As γ ( u ) anticommutes with γ ( w ) and squares to 1, c ( w ) c ( w ′ ) = γ ( w ) γ ( w ′ ) , hence c ( w ) c ( w ′ ) + c ( w ′ ) c ( w ) = 2 w · w ′ . The c ( w ) are even, so they preserve Δ ± ( V ) , which thereby become modules of Cl ( W ) of dimension 2 r − 1 , where 2 r = dim V . This is the dimension of the irreducible modules of Cl ( W ) , and so ( W ) acts through the quadratic elements c ( w ) c ( w ′ ) = γ ( w ) γ ( w ′ ) , so each Δ ± ( V ) is the spinor module of so ( W )  [38].    □

The Real form f 4(-20) and so(25,1)

Section 4 of Ref. [24] states: “A non-compact real form of a theorem by Dynkin [28] yields the maximal and non-symmetric `anomalous’ embedding [29]
f 4 ( − 20 ) ⊂ so 25 , 1 , 26 = 26 ,
under which the vector representation of so 25 , 1 stays irreducible, providing the fundamental representation of the minimally non-compact real form f 4 ( − 20 ) of f 4 .” This is Equation (4.4) there, and Equation (4.6) identifies the vector of so 25 , 1 with the traceless part of the Lorentzian octonionic Jordan algebra, 26 ≃ ( J 1 , 2 O ) 0 .11
Proposition H.2. 
Let φ : f 4 ( − 20 ) → so ( p , q ) with p + q = 26 be a homomorphism of real Lie algebras. If φ ≠ 0 , then { p , q } = { 10 , 16 } . In particular every homomorphism f 4 ( − 20 ) → so ( 25 , 1 ) is zero, and (A1) does not hold. 
Proof. 
Write f = f 4 ( − 20 ) . Since f is simple, φ ≠ 0 is injective, and R p , q becomes a real f -module M of dimension 26 with an invariant nondegenerate symmetric form η . The nontrivial irreducible modules of f C = f 4 ( C ) have dimension at least 26, and there is exactly one of dimension 26 [38,39]. Hence M C is that module, and M is irreducible.
Let f = k ⊕ p be a Cartan decomposition. Here k ≅ so ( 9 ) and p C is the spinor module Δ 9 of dimension 16 [39,40], and under k the module M C decomposes as C ⊕ C 9 ⊕ Δ 9  [39], the branching 26 = 1 ⊕ 9 ⊕ 16 quoted in Equation (4.7) of Ref. [24]. Complex conjugation of M C maps k C -submodules to k C -submodules of the same dimension, and the three summands have different dimensions, so it preserves each of them. Their real points are irreducible k -modules M 1 , M 9 , M 16 with M = M 1 ⊕ M 9 ⊕ M 16 .
The compact real form u = k ⊕ i p of f C acts on M C , and since the corresponding simply connected group is compact, M C carries a u -invariant Hermitian product h. The elements of k and of i p act by skew-Hermitian operators, so those of p act by Hermitian ones. On M the real part 〈 x , y 〉 = Re h ( x , y ) is an inner product for which k acts by antisymmetric and p by symmetric operators. Inequivalent irreducible summands of an orthogonal module are orthogonal, so M 1 , M 9 , M 16 are mutually orthogonal.
Write η ( x , y ) = 〈 J x , y 〉 with J symmetric and invertible. Invariance of η reads X T J + J X = 0 , so J X = X J for X ∈ k and J X = − X J for X ∈ p . Thus J preserves each M d , and on M d it is a symmetric operator commuting with an irreducible action, so J | M d = λ d with λ d ≠ 0 .
The action map p ⊗ M → M is k -equivariant, hence equivariant for the simply connected group Spin ( 9 ) , whose central element − 1 acts by + 1 on C and C 9 and by − 1 on Δ 9 and on p C . Therefore every X ∈ p maps M 1 ⊕ M 9 into M 16 . If p M 1 = 0 , then M 1 is an f -submodule, contradicting irreducibility; so some X ∈ p has a nonzero component X 16 , 1 : M 1 → M 16 , and the component of J X = − X J from M 1 to M 16 gives ( λ 16 + λ 1 ) X 16 , 1 = 0 , that is λ 1 = − λ 16 . The same argument applied to M 9 gives λ 9 = − λ 16 . Hence η is definite of one sign on the 10-dimensional space M 1 ⊕ M 9 and of the opposite sign on the orthogonal 16-dimensional space M 16 , and its signature is ( 10 , 16 ) or ( 16 , 10 ) .    □
The proof also shows where the Peirce decomposition places its pieces: the summands 1 and 9 have one sign and 16 the other. The 26 of f 4 ( − 20 ) is realized on ( J 1 , 2 O ) 0 , as eqs. (4.5) and (4.6) of Ref. [24] say, and by Proposition H.2 its invariant form has signature ( 10 , 16 ) , in agreement with the computation described at the end of this appendix. The correct form of (A1) is therefore f 4 ( − 20 ) ⊂ so ( 10 , 16 ) .

The so(15)-Covariant Splitting of the Spinor of so(27,3)

Equation (3.4) of Ref. [24] reads
16384 so 27 , 3 MW spinor = ⨁ i = 0 15 ∧ 2 i + 1 15 = 15 ⊕ 455 ⊕ 3 , 003 ⊕ 6 , 435 ⊕ 5 , 005 ′ ⊕ 1 , 365 ′ ⊕ 105 ′ ⊕ 1 so 15 - cov . odd ( co ) homology of the 15 - brane ,
and Equation (3.5) gives the conjugate spinor as ⨁ i ∧ 2 i 15 .12 The dimensions agree, ∑ i 15 2 i + 1 = 2 14 . The question is whether some so ( 15 ) inside so ( 27 , 3 ) realizes the splitting.
Proposition H.3. 
Let h ⊂ so ( 27 , 3 ) be a subalgebra isomorphic to so ( p , q ) for some p + q = 15 , the compact so ( 15 ) included, and let S be a Majorana–Weyl module of so ( 27 , 3 ) of either chirality. Then S C , restricted to h C ≅ so ( 15 , C ) , is the sum of 128 copies of the spinor module of so ( 15 , C ) . In particular it is isomorphic neither to ⨁ i Λ 2 i + 1 C 15 nor to ⨁ i Λ 2 i C 15 , and (A2) holds for no such h . 
Proof. 
Let V = R 27 , 3 , a faithful h -module with invariant form η , and let W be the vector module of h C . The irreducible modules of so ( 15 , C ) of dimension at most 30 are C and W; the next smallest has dimension 105 [38,39]. As h acts faithfully, V C ≅ W ⊕ C 15 with trivial action on C 15 , or V C ≅ W ⊕ W .
The case V C ≅ W ⊕ W does not occur. The vector module of h ≅ so ( p , q ) is the real form W R = R p , q of W, with invariant form g of signature ( p , q ) , unique up to a factor because W is irreducible. Since Hom h ( W R , V ) ⊗ C ≅ Hom h C ( W , V C ) ≠ 0 , V contains a copy of W R , whose invariant complement is a real form of W, hence again isomorphic to W R . So V ≅ W R ⊗ R 2 , and an invariant bilinear form on it is an element of Hom h ( W R , W R * ) ⊗ ( R 2 ⊗ R 2 ) * , that is η = g ⊗ c for a bilinear form c on R 2 , which is symmetric and nondegenerate because η and g are. Diagonalizing c shows that η has signature ( 2 p , 2 q ) , ( 15 , 15 ) or ( 2 q , 2 p ) . Since 27 is odd, none of these is ( 27 , 3 ) .
The case V C ≅ W ⊕ C 15 . Call the trivial summand U. For u ∈ U , w ∈ W and X ∈ h C , η ( u , X w ) = − η ( X u , w ) = 0 , and h C W = W , so U ⊥ W and both are nondegenerate. Hence h C embeds in so ( W ) , and since both have dimension 105, h C = so ( W ) . Choose a unit vector u ∈ U and put W ′ = W ⊥ C u and U ′ = u ⊥ ∩ U , of dimensions 16 and 14. By Lemma H.1(i), S C ≅ Δ ± ( V C ) restricts to so ( W ′ ) ⊕ so ( U ′ ) as Δ + ( W ′ ) ⊗ Δ ± ( U ′ ) ⊕ Δ − ( W ′ ) ⊗ Δ ∓ ( U ′ ) , and by Lemma H.1(ii) both Δ ± ( W ′ ) restrict to so ( W ) as the spinor module Δ ( W ) . As dim Δ ± ( U ′ ) = 64 , S C | h C ≅ Δ ( W ) ⊕ 128 .
The module Δ ( W ) is irreducible of dimension 128, while ⨁ i Λ 2 i + 1 W contains Λ 1 W = W and ⨁ i Λ 2 i W contains Λ 0 W = C . Neither is a summand of Δ ( W ) ⊕ 128 .    □
In the excluded case the claimed splitting would indeed follow: V C = W ⊗ C 2 is then the sum of the two h C -invariant maximal isotropic subspaces L = W ⊗ ℓ and L ′ = W ⊗ ℓ ′ , where ℓ , ℓ ′ are the isotropic lines of c, and Chevalley’s construction identifies the Weyl modules with Λ even L and Λ odd L , twisted by a character of gl ( L ) that is trivial on h C  [33]. For the compact so ( 15 ) the form g is definite and the only indefinite option is ( 15 , 15 ) , with real Witt index 15, whereas R 27 , 3 has Witt index 3 (Lemma 2.2). This is the obstruction of Section 5 in another guise: the spinors see D and s − t modulo eight, while the splitting needs isotropic subspaces that the signature does not provide.

The Magic Star Algebra e 7 (2) and e 11

Section 2 of Ref. [24] states: “It can also be shown that e 11 contains so 20 ⊕ su 2 ⊕ ( 512 , 2 ) , which is precisely e 7 ( 2 ) .” We read the sentence as saying that e 7 ( 2 ) , with its bracket, is a Lie subalgebra of e 11 .
Proposition H.4. 
There is no complex Lie algebra g = g 0 ⊕ g 1 in which g 0 ≅ so ( 20 , C ) ⊕ sl ( 2 , C ) is a subalgebra, g 1 is a g 0 -submodule isomorphic to Δ + ( C 20 ) ⊗ C 2 or Δ − ( C 20 ) ⊗ C 2 , and [ g 1 , g 1 ] ≠ 0 . The same holds for every real form, since complexification preserves these properties. 
Proof. 
Let h 0 be a Cartan subalgebra of g 0 , of dimension 11. The so ( 20 , C ) -weights of g 1 are the vectors ( ± 1 2 , … , ± 1 2 ) , so those of Λ 2 g 1 have integer coordinates, and the equivariant component Λ 2 g 1 → g 1 of the bracket vanishes. Hence [ g 1 , g 1 ] ⊂ g 0 , and the map θ equal to 1 on g 0 and − 1 on g 1 is an automorphism of g .
The radical r of g is invariant under automorphisms, so r = r 0 ⊕ r 1 with r i = r ∩ g i . Here r 0 is a solvable ideal of the semisimple g 0 , so r 0 = 0 , and r 1 is a g 0 -submodule of the irreducible g 1 . If r 1 = g 1 , then [ g 1 , g 1 ] ⊂ r ∩ g 0 = 0 , which is excluded. So r = 0 and g is semisimple.
The elements of h 0 act diagonalizably on g 0 and on g 1 , and no weight of g 1 vanishes, so the centralizer of h 0 in g is h 0 . A toral subalgebra equal to its own centralizer is maximal toral, so g has rank 11. For a simple Lie algebra of rank n ≤ 11 the ratio of dimension to rank is n + 2 , 2 n + 1 , 2 n + 1 , 2 n − 1 for A n , B n , C n , D n and 7 , 13 , 13 , 19 , 31 for G 2 , F 4 , E 6 , E 7 , E 8  [38], never more than 31. A semisimple Lie algebra of rank 11 therefore has dimension at most 341, while dim g = 190 + 3 + 1024 = 1217 .    □
A Lie subalgebra of e 11 of the stated module type therefore has [ g 1 , g 1 ] = 0 . In e 7 ( 2 ) the bracket of two elements of ( 512 , 2 ) does not vanish: so 20 ⊕ su 2 is a Lie algebra acting on ( 512 , 2 ) , so with [ g 1 , g 1 ] = 0 the Jacobi identity would hold, whereas Ref. [24] attributes the failure of the Jacobi identity in the Magic Star algebras to their spinor generators and calls these algebras non-Lie. Read as a statement about subalgebras, the quoted sentence is therefore false.13

Appendix A Superconformal Algebra in 11+3 Dimensions

Section 1 of Ref. [28] states: “This Lie superalgebra, denoted as Γ ( e 8 ( − 24 ) ) , is the N = 1 D = 11 + 3 superconformal algebra”. Here Γ ( g ) = g 0 ⊗ G 0 ⊕ g 1 ⊗ G 1 is the Grassmann envelope of e 8 ( − 24 ) = so ( 12 , 4 ) ⊕ 128 , and its generators M a b and Q α are required to satisfy, in Equation (6) there,
{ Q α , Q β } = γ a γ b − γ b γ a α β M a b .
Proposition H.5. 
Let g = g 0 ¯ ⊕ g 1 ¯ be a real Lie superalgebra, of any dimension, with so ( 12 , 4 ) ⊂ g 0 ¯ and g 1 ¯ isomorphic, under so ( 12 , 4 ) , to a Majorana–Weyl module 128 . Then π ( [ g 1 ¯ , g 1 ¯ ] ) = 0 for every so ( 12 , 4 ) -equivariant linear map π : g 0 ¯ → so ( 12 , 4 ) . In particular, elements M a b ∈ g 0 ¯ and a basis Q α of g 1 ¯ satisfy (A3) only if every M a b is zero. 
Proof. 
The map β ( Q , Q ′ ) = π ( [ Q , Q ′ ] ) is symmetric, because the bracket of two odd elements is, and so ( 12 , 4 ) -equivariant, by the super Jacobi identity [ M , [ Q , Q ′ ] ] = [ [ M , Q ] , Q ′ ] + [ Q , [ M , Q ′ ] ] for M ∈ so ( 12 , 4 ) . Since 12 − 4 ≡ 0 ( mod 8 ) , the complexification of 128 is a Weyl module of so ( 16 , C ) , say Δ + , and β C is an equivariant map Sym 2 Δ + → Λ 2 . By Proposition 2.1 with r = 8 , Sym 2 Δ + ≅ Λ 0 ⊕ Λ 4 ⊕ Λ + 8 , of dimensions 1, 1820 and 6435, and Λ 2 Δ + ≅ Λ 2 ⊕ Λ 6 . These forms are irreducible so ( 16 , C ) -modules [38], and Λ 2 , of dimension 120, is isomorphic to none of the summands of Sym 2 Δ + . By Schur’s lemma β C = 0 .
For the second statement, the right-hand side of (A3), with the index of Q β lowered by an invariant pairing of the 128 with itself, has the form κ a b ( Q , Q ′ ) M a b , where κ = ( κ a b ) is a nonzero equivariant bilinear map from 128 × 128 to Λ 2 R 12 , 4 . By Proposition 2.1, Λ 2 occurs in Δ + ⊗ Δ + only inside Λ 2 Δ + , so κ is antisymmetric, while the bracket is symmetric. Hence κ a b ( Q , Q ′ ) M a b = 0 for all Q , Q ′ . As κ is surjective onto the irreducible Λ 2 , the forms κ a b , a < b , are linearly independent, and all M a b vanish.    □
In a superconformal algebra the anticommutator of the supercharges contains the translations of the conformal algebra, so its projection to the conformal algebra is not zero. By Proposition H.5, no Lie superalgebra whose even part is finite dimensional and contains the conformal algebra so ( 12 , 4 ) of R 11 , 3 , and whose odd part is a single 128 , has this property; for finite-dimensional g 0 ¯ the projection π onto so ( 12 , 4 ) along an invariant complement exists by complete reducibility. The Grassmann envelope is a different object. It is a Lie superalgebra over the even part G 0 of the Grassmann algebra, with even part so ( 12 , 4 ) ⊗ G 0 and odd part 128 ⊗ G 1 , infinite dimensional over R when the Grassmann algebra is, and the bracket of two of its odd elements, [ T α ⊗ ϑ , T β ⊗ ϑ ′ ] = [ T α , T β ] ⊗ ϑ ϑ ′ , carries the antisymmetric bracket [ T α , T β ] of e 8 ( − 24 ) , in agreement with Λ 2 Δ + ⊃ Λ 2 . It is not a Lie superalgebra with even part so ( 12 , 4 ) and odd part a single 128 , and by Proposition H.5 no such superalgebra satisfies (A3) with M a b ≠ 0 .

Brane Branchings in 28+4

Ref. [25] restricts the semispinor 32 , 768 of so 28 , 4 along so 28 , 4 → so 12 , 4 ⊕ so 16 → so 4 , 4 ⊕ so 8 ⊕ so 16 and along so 28 , 4 → so 20 , 4 ⊕ so 8 → so 4 , 4 ⊕ so 16 ⊕ so 8 . Eqs. (34) and (37) there read
32 , 768 = ( 128 , 128 ′ ) ⊕ ( 128 ′ , 128 )
= ( 8 s , 8 c , 128 ′ ) ⊕ ( 8 c , 8 s , 128 ′ ) ⊕ ( 8 c , 8 s , 128 ) ⊕ ( 8 s , 8 c , 128 ) , 32 , 768 = ( 2048 , 8 c ) ⊕ ( 2048 ′ , 8 s )
= ( 8 s , 128 ′ , 8 c ) ⊕ ( 8 c , 128 , 8 c ) ⊕ ( 8 c , 128 , 8 s ) ⊕ ( 8 s , 128 ′ , 8 s ) .
Proposition H.6. 
Let S be either Majorana–Weyl module of so ( 28 , 4 ) and decompose R 28 , 4 = R 4 , 4 ⊥ R 8 , 0 ⊥ R 16 , 0 . Then S C restricted to so ( 4 , 4 ) ⊕ so ( 8 ) ⊕ so ( 16 ) is the sum of four irreducible modules A ⊗ B ⊗ C , with A , B Weyl modules of the two so ( 8 , C ) and C one of so ( 16 , C ) , and for any two of the three factors the four pairs of labels are distinct. Neither the second line of (A4) nor that of () has this property, and none acquires it after applying automorphisms of the factors, such as triality of so ( 4 , 4 ) or so ( 8 ) or the exchange 128 ↔ 128 ′ . Moreover the two semispinors of so ( 12 , 4 ) restrict to non-isomorphic modules of so ( 4 , 4 ) ⊕ so ( 8 ) , and those of so ( 20 , 4 ) to non-isomorphic modules of so ( 4 , 4 ) ⊕ so ( 16 ) . 
Proof. 
All signatures involved have s − t ≡ 0 ( mod 8 ) , so the real Weyl modules are real forms of the complex ones and it suffices to work over C . By Lemma H.1(i) with m = 3 , S C ≅ ⨁ Δ ε 1 ⊗ Δ ε 2 ⊗ Δ ε 3 over the four sign triples with a fixed product ε 1 ε 2 ε 3 . Any two signs determine the third, so for any two factors the four pairs are distinct; since Δ + ¬ ≅ Δ − for so ( 8 , C ) and so ( 16 , C ) , distinct sign pairs are distinct pairs of isomorphism classes. In the second line of (A4) the pair of so ( 4 , 4 ) ⊕ so ( 8 ) labels ( 8 s , 8 c ) occurs twice, and so does ( 8 c , 8 s ) ; in that of () the pair of so ( 4 , 4 ) ⊕ so ( 16 ) labels ( 8 c , 128 ) occurs twice, and so does ( 8 s , 128 ′ ) . An automorphism of a factor permutes the isomorphism classes of its modules, so a repeated pair stays repeated after any relabeling. For the last statement, Lemma H.1(i) with m = 2 gives Δ ± ( C 16 ) → Δ + ⊗ Δ ± ⊕ Δ − ⊗ Δ ∓ under so ( 8 , C ) ⊕ so ( 8 , C ) , and Δ ± ( C 24 ) → Δ + ⊗ Δ ± ⊕ Δ − ⊗ Δ ∓ under so ( 8 , C ) ⊕ so ( 16 , C ) ; the two restrictions differ in each case.    □
The first lines of (A4) and () are correct. The second lines contradict them: in (A4) the 128 of so ( 12 , 4 ) , paired with 128 ′ , and the 128 ′ , paired with 128 , are both assigned ( 8 s , 8 c ) ⊕ ( 8 c , 8 s ) , and in () the 2048 and 2048 ′ of so ( 20 , 4 ) are both assigned ( 8 s , 128 ′ ) ⊕ ( 8 c , 128 ) . With labels chosen so that the first lines hold, 128 → ( 8 s , 8 s ) ⊕ ( 8 c , 8 c ) and 2048 → ( 8 s , 128 ) ⊕ ( 8 c , 128 ′ ) , the restrictions are
32 , 768 = ( 8 s , 8 s , 128 ′ ) ⊕ ( 8 c , 8 c , 128 ′ ) ⊕ ( 8 s , 8 c , 128 ) ⊕ ( 8 c , 8 s , 128 ) , 32 , 768 = ( 8 s , 128 , 8 c ) ⊕ ( 8 c , 128 ′ , 8 c ) ⊕ ( 8 s , 128 ′ , 8 s ) ⊕ ( 8 c , 128 , 8 s ) .

Computational Checks

The finite arithmetic of this appendix is rechecked by three programs, each run as a single Python process with python3 -I. The program f4_signature.py builds J 1 , 2 O with the product X ∘ Y = 1 2 ( X η Y + Y η X ) , η = diag ( 1 , 1 , − 1 ) , verifies the Jordan identity, and solves for the derivations: it finds a 52-dimensional algebra whose trace form on the 27-dimensional algebra has 16 positive and 36 negative eigenvalues, a 26-dimensional span of the images of the derivations, a one-dimensional space of invariant symmetric forms on it, and signature ( 10 , 16 ) both for that form and for the Jordan trace form. The program spinor_checks.py works with weights. It restricts the semispinors of so ( 32 , C ) to so ( 8 , C ) ⊕ so ( 8 , C ) ⊕ so ( 16 , C ) in the orders of (A4) and () and finds that none of the 72 relabelings by triality of each so ( 8 ) factor and exchange of the semispinors of so ( 16 ) reproduces either equation; it restricts the semispinors of so ( 30 , C ) to so ( 15 , C ) in the two cases of the proof of Proposition H.3 and lists the signatures ( 2 p , 2 q ) and ( 15 , 15 ) ; and with explicit gamma matrices of Cl ( 16 ) it finds the bilinear forms ψ ¯ γ a 1 … a k χ on Δ + × Δ + symmetric for k = 0 , 4 , 8 and antisymmetric for k = 2 , 6 . Among other counts, arithmetic_checks.py recomputes the dimensions in (A2) and compares the 1206 roots that the algebra of Proposition H.4 would need with the largest number, 258, of roots in rank 11. The outputs agree with the statements above; they confirm the proofs and do not replace them.

References

  1. Brink, L.; Schwarz, J. H.; Scherk, J. Supersymmetric Yang–Mills theories. Nucl. Phys. B 1977, 121 77. [Google Scholar] [CrossRef]
  2. Kugo, T.; Townsend, P. K. Supersymmetry and the division algebras. Nucl. Phys. B 1983, 221, 357. [Google Scholar] [CrossRef]
  3. Evans, J. M. Supersymmetric Yang–Mills theories and division algebras. Nucl. Phys. B 1988, 298 92. [Google Scholar] [CrossRef]
  4. Baez, J. C.; Huerta, J. Division algebras and supersymmetry I, in Superstrings, Geometry, Topology, and C*-algebras. In Proc. Sympos. Pure Math. 81; American Mathematical Society: Providence, 2010; p. 65. [Google Scholar] [CrossRef]
  5. Green, M. B.; Schwarz, J. H. Covariant description of superstrings. Phys. Lett. B 1984, 136, 367. [Google Scholar] [CrossRef]
  6. Achúcarro, A.; Evans, J. M.; Townsend, P. K.; Wiltshire, D. L. Super p-branes. Phys. Lett. B 1987, 198 441. [Google Scholar] [CrossRef]
  7. Bergshoeff, E.; Sezgin, E.; Townsend, P. K. Supermembranes and eleven-dimensional supergravity. Phys. Lett. B 1987, 189 75. [Google Scholar] [CrossRef]
  8. Blencowe, M. P.; Duff, M. J. Supermembranes and the signature of space-time. Nucl. Phys. B 1988, 310 387. [Google Scholar] [CrossRef]
  9. . Vafa, Evidence for F-theory. Nucl. Phys. B 1996, arXiv:hep-th/9602022469 403. [CrossRef]
  10. . Bars, S-theory. Phys. Rev. D. 1997, arXiv:hep55, 2373. [CrossRef]
  11. . Bars, A case for 14 dimensions. Phys. Lett. B 1997, arXiv:hep-th/9704054403 257. [CrossRef]
  12. Bars, I.; Kounnas, C. Theories with two times. Phys. Lett. B 1997, arXiv:hep402 25. [Google Scholar] [CrossRef]
  13. Bars, I.; Deliduman, C.; Andreev, O. Gauged duality, conformal symmetry and spacetime with two times. Phys. Rev. D. 1998, arXiv:hep-th/980318858, 066004. [Google Scholar] [CrossRef]
  14. . Bars, Two-time physics in field theory. Phys. Rev. D. 2000, arXiv:hep-th/000310062, 046007. [CrossRef]
  15. C. M. Hull, Duality and the signature of space-time. JHEP 1998, arXiv:hep11 017. [CrossRef]
  16. Hull, C. M. Timelike T-duality, de Sitter space, large N gauge theories and topological field theory. JHEP 1998, arXiv:hep07 021. [Google Scholar] [CrossRef]
  17. Rudychev, I.; Sezgin, E. Superparticles in. Phys. Lett. B 1997, arXiv:hep415, 363. [Google Scholar] [CrossRef]
  18. Nishino, H. Supergravity in 10+2 dimensions as consistent background for superstring. Phys. Lett. B 1998, arXiv:hep428 85. [Google Scholar] [CrossRef]
  19. Nishino, H.; Sezgin, E. Supersymmetric Yang–Mills equations in 10+2 dimensions. Phys. Lett. B 1996, arXiv:hep-th/9607185388, 569. [Google Scholar] [CrossRef]
  20. Sezgin, E. Super Yang–Mills in (11,3) dimensions. Phys. Lett. B 1997, arXiv:hep403 265. [Google Scholar] [CrossRef]
  21. Rudychev, I.; Sezgin, E.; Sundell, P. Supersymmetry in dimensions beyond eleven. Nucl. Phys. B Proc. Suppl. 1998, arXiv:hep-th/971112768 285. [Google Scholar] [CrossRef]
  22. Nishino, H. Supersymmetric Yang–Mills theories in D≥12. Nucl. Phys. B 1998, arXiv:hep523 450. [Google Scholar] [CrossRef]
  23. Rios, M.; Marrani, A.; Chester, D. Geometry of exceptional super Yang–Mills theories. Phys. Rev. D. 2019, arXiv:1811.0610199, 046004. [Google Scholar] [CrossRef]
  24. Rios, M.; Marrani, A.; Chester, D. Exceptional super Yang–Mills in 27+3 and worldvolume M-theory. Phys. Lett. B 2020, arXiv:1906.10709808, 135674. [Google Scholar] [CrossRef]
  25. Paduroiu, S.; Rios, M.; Marrani, A.; Chester, D. Warm dark matter from higher-dimensional gauge theories. Universe 2021, arXiv:2202.084597 462. [Google Scholar] [CrossRef]
  26. Marrani, A.; Rios, M.; Chester, D. Monstrous M-theory. Symmetry 2023, arXiv:2008.0674215, 490. [Google Scholar] [CrossRef]
  27. Truini, P.; Marrani, A.; Rios, M.; de Graaf, W. Exceptional periodicity and Magic Star algebras. Expo. Math. 2025, 43, 125621. [Google Scholar] [CrossRef]
  28. Chester, D.; Marrani, A.; Rios, M.; Irwin, K. Flipped SU(5) GUT with conformal gravity from a single supermultiplet. J. Phys. Conf. Ser. 2024, arXiv:2412.201912912, 012040. [Google Scholar] [CrossRef]
  29. Deligne, P. Notes on spinors. In Quantum Fields and Strings: A Course for Mathematicians; American Mathematical Society: Providence, 1999; vol. 1, pp. 99–135. [Google Scholar]
  30. Van Proeyen, A. Tools for supersymmetry. Ann. Univ. Craiova Phys. 1999, arXiv:hep9 1. [Google Scholar]
  31. D’Auria, R.; Ferrara, S.; Lledó, M. A.; Varadarajan, V. S. Spinor algebras. J. Geom. Phys. 2001, arXiv:hep-th/001012440 101. [Google Scholar] [CrossRef]
  32. Lawson, H. B.; Michelsohn, M.-L. Spin Geometry; Princeton University Press: Princeton, 1989. [Google Scholar] [CrossRef]
  33. . Chevalley, The Algebraic Theory of Spinors; Columbia University Press: New York, 1954. [CrossRef]
  34. Budinich, P.; Trautman, A. The Spinorial Chessboard; Springer: Berlin, 1988. [Google Scholar] [CrossRef]
  35. van Holten, J. W.; Van Proeyen, A. N=1 supersymmetry algebras in d=2,3,4 mod 8. J. Phys. A 1982, 15, 3763. [Google Scholar] [CrossRef]
  36. . Witt, Theorie der quadratischen Formen in beliebigen Körpern. J. Reine Angew. Math. 1937, 176 31. [CrossRef]
  37. Lam, T. Y. Introduction to Quadratic Forms over Fields. In Grad. Stud. Math. 67; American Mathematical Society: Providence, 2005. [Google Scholar] [CrossRef]
  38. Fulton, W.; Harris, J. Representation Theory: A First Course, Grad. In Texts in Math. 129; Springer: New York, 1991. [Google Scholar] [CrossRef]
  39. Slansky, R. Group theory for unified model building. Phys. Rep. 1981, 79 1. [Google Scholar] [CrossRef]
  40. S. Helgason, Differential Geometry, Lie Groups, and Symmetric Spaces. In Grad. Stud. Math. 34; American Mathematical Society: Providence, 2001. [CrossRef]
  41. Freedman, D. Z.; Van Proeyen, A. Supergravity; Cambridge University Press: Cambridge, 2012. [Google Scholar] [CrossRef]
  42. Haag, R.; opuszański, J. T.; Sohnius, M. All possible generators of supersymmetries of the S-matrix. Nucl. Phys. B 1975, 88 257. [Google Scholar] [CrossRef]
  43. de Azcárraga, J. A.; Gauntlett, J. P.; Izquierdo, J. M.; Townsend, P. K. Topological extensions of the supersymmetry algebra for extended objects. Phys. Rev. Lett. 1989, 63, 2443. [Google Scholar] [CrossRef]
  44. . Nahm, Supersymmetries and their representations. Nucl. Phys. B 1978, 135 149. [CrossRef]
  45. Strathdee, J. A. Extended Poincaré supersymmetry. Int. J. Mod. Phys. A 1987, 2 273. [Google Scholar] [CrossRef]
  46. Bhattacharjee, D. Source and verification files for residual-dimension obstructions for null-projected supersymmetric Yang–Mills theories. Zenodo Version 2.0.0 All. Earlier Versions 2026. [Google Scholar] [CrossRef]
  47. Bhattacharjee, D. creelie/ff-2, verification programs, GitHub. 2026. Available online: https://github.com/creelie/ff-2.
1
Bott periodicity makes the real Clifford algebras of R s , t and R s + 4 , t − 4 isomorphic, but no isomorphism carries the space of generators R s , t onto R s + 4 , t − 4 . Null vectors, their orthogonality and the bound min ( s , t ) are properties of the generators.
2
For 27 + 3 one has r = 15 and Sym 2 Δ + = Λ + 15 ⊕ Λ 11 ⊕ Λ 7 ⊕ Λ 3 , of dimension 77 558 760 + 54 627 300 + 2 035 800 + 4 060 = 134 225 920 = 2 13 ( 2 14 + 1 ) . These are the 3-, 7-, 11- and 15-form charges of Equation (2.1) of Ref. [24].
3
The number min ( s , t ) is the Witt index of η  [36,37].
4
As in Nishino’s construction, ϵ is not constrained. The proof uses only the component of ϵ in K ^ , which by Lemma 5.2(iv) is the only component that enters the residual components of δ A . Any family of parameters whose K ^ -components form a nonzero Spin ( V ′ ) -invariant subspace leads to the same conclusions.
5
Nishino writes v [ 4 n ] for the class m = 4 n and treats m = 4 n + 1 , 4 n + 2 , 4 n + 3 in the same way. We write k for the number of null vectors in all cases.
6
This is the usual reason why the anticommutator of supercharges is symmetric: with Grassmann parameters, [ δ ( ϵ 1 ) , δ ( ϵ 2 ) ] = ϵ 2 α ϵ 1 β { Q α , Q β } up to convention-dependent factors.
7
For Majorana gauginos the components λ J β are real Grassmann generators in a real basis of Σ . The identity (26) is linear in S I , so it extends to all complex symmetric tensors.
8
Ref. [23]: “The simple but crucial observation is that, by switching 4 n timelike dimensions into 4 n spacelike dimensions in the above space-time signatures, one keeps D and ρ unchanged, and so all the properties of spinors and γ -matrices relevant for the definition of the above super-Poincaré algebras are left unchanged. Thus, one can check that the proofs given in [6] (see also [5]) for the existence of SYM’s with superalgebras (3.1) and (3.2) can be successfully repeated to show the existence of SYM’s based on global minimal, chiral (1,0) superalgebras in the following space-time signatures”. Here ρ = s − t modulo eight, [6] is Ref. [22] and [5] is Ref. [20]; the list that follows is eqs. (3.5)–(3.8), and the third entry is D = ( 11 + 8 n ) + 3 with the superalgebra (3.7).
9
Ref. [24]: “In D = 27 + 3 , the electric 11-brane has a multi-time worldvolume, with signature 11 + 3 , which can be used to provide a worldvolume realization for the 11 + 3 SYM of Bars and Sezgin [7–9].” Its footnote 2 reads: “It should be recalled that the Bars-Sezgin SYM theories in D > 10 are not Lorentz covariant, and so are the generalizations presented here.” The lack of Lorentz covariance is the choice of the null vectors.
10
The annihilator of a spinor is a totally isotropic subspace; spinors whose annihilator has the largest possible dimension are Cartan’s pure spinors [33,34]. Lemma 2.3 adds that for a Majorana spinor this subspace is the complexification of a real one.
11
Dynkin’s theorem concerns complex Lie algebras, where it gives f 4 ( C ) ⊂ so ( 26 , C ) ; for the compact form the invariant form on the 26 is definite and f 4 ⊂ so ( 26 ) . For a noncompact real form the signature of the ambient algebra has to be determined, and this is done below.
12
The list in Equation (3.5) ends with 25 ′ , which should read 15 ′ .
13
Whether e 11 contains a subalgebra so 20 ⊕ su 2 together with a subspace transforming as ( 512 , 2 ) , without closing into a subalgebra, is not decided here.
Figure 1. Signatures ( s , t ) and numbers k of null vectors. Blue points: Nishino’s signatures ( 9 + m ) + ( 1 + m ) with k = m , on the plane s + t − 2 k = 10 (blue). Red segments: the signatures ( 9 + 8 n + q ) + ( 1 + q ) of the families (2) for n = 1 , 2 and q = 0 , 1 , 2 , 3 , with every k ≤ t ; the star marks 27 + 3 . The grey surface is the wall k = t , beyond which real null frames do not exist. No red point reaches the blue plane.
Figure 1. Signatures ( s , t ) and numbers k of null vectors. Blue points: Nishino’s signatures ( 9 + m ) + ( 1 + m ) with k = m , on the plane s + t − 2 k = 10 (blue). Red segments: the signatures ( 9 + 8 n + q ) + ( 1 + q ) of the families (2) for n = 1 , 2 and q = 0 , 1 , 2 , 3 , with every k ≤ t ; the star marks 27 + 3 . The grey surface is the wall k = t , beyond which real null frames do not exist. No red point reaches the blue plane.
Preprints 244177 g001
Figure 2. The logical structure of the paper. The axioms pass through the oscillator decomposition and Steps 1–3 to the identity (6) on the residual spinors; Theorem 3.3 turns that identity into D ′ ∈ { 2 , 10 } , and the bound on the number of null vectors converts this into a condition on s − t . Proposition 2.1 enters Step 2 (dashed) and the single-time theorem. Proposition 4.1 is the case without null vectors. In the fourth row, Theorems 5.14 and 5.12 show that the conditions of the Main Theorem are sufficient when D ′ = 2 and D ′ = 10 . Theorem 7.3 drops (N4) and allows any contraction of the source; it uses Proposition 2.1 (dashed) and, for real gauginos, the bound on the number of null vectors. Proposition 7.5 uses the same projection (dashed) and excludes several gauginos. In the last row, Proposition 7.14 and Theorem 7.11 rerun Steps 1–3 of Theorem 5.6 for nonlinear rules and for any bilinear bracket, and Theorem 7.6 extends Proposition 7.5 to gauginos mixed by their transformation rules.
Figure 2. The logical structure of the paper. The axioms pass through the oscillator decomposition and Steps 1–3 to the identity (6) on the residual spinors; Theorem 3.3 turns that identity into D ′ ∈ { 2 , 10 } , and the bound on the number of null vectors converts this into a condition on s − t . Proposition 2.1 enters Step 2 (dashed) and the single-time theorem. Proposition 4.1 is the case without null vectors. In the fourth row, Theorems 5.14 and 5.12 show that the conditions of the Main Theorem are sufficient when D ′ = 2 and D ′ = 10 . Theorem 7.3 drops (N4) and allows any contraction of the source; it uses Proposition 2.1 (dashed) and, for real gauginos, the bound on the number of null vectors. Proposition 7.5 uses the same projection (dashed) and excludes several gauginos. In the last row, Proposition 7.14 and Theorem 7.11 rerun Steps 1–3 of Theorem 5.6 for nonlinear rules and for any bilinear bracket, and Theorem 7.6 extends Proposition 7.5 to gauginos mixed by their transformation rules.
Preprints 244177 g002
Figure 3. The light cone of R 2 , 1 . Two independent real null vectors n = ( cos θ , sin θ , 1 ) and n ′ = ( cos θ ′ , sin θ ′ , 1 ) satisfy n · n ′ = cos ( θ − θ ′ ) − 1 < 0 , so the plane they span (grey) is not isotropic. In R s , t the largest isotropic subspaces have dimension min ( s , t ) (Lemma 2.2).
Figure 3. The light cone of R 2 , 1 . Two independent real null vectors n = ( cos θ , sin θ , 1 ) and n ′ = ( cos θ ′ , sin θ ′ , 1 ) satisfy n · n ′ = cos ( θ − θ ′ ) − 1 < 0 , so the plane they span (grey) is not isotropic. In R s , t the largest isotropic subspaces have dimension min ( s , t ) (Lemma 2.2).
Preprints 244177 g003
Figure 4. The spinor dimension N = 2 ( d − 2 ) forced by the cubic identity (blue) against the smallest dimension 2 ⌈ d / 2 ⌉ − 1 that a vector structure can have (red). A solution requires the blue value to be a multiple of the red one. This happens for d = 3 , 4 , 6 , 10 (circles) and never again, because the red curve grows exponentially.
Figure 4. The spinor dimension N = 2 ( d − 2 ) forced by the cubic identity (blue) against the smallest dimension 2 ⌈ d / 2 ⌉ − 1 that a vector structure can have (red). A solution requires the blue value to be a multiple of the red one. This happens for d = 3 , 4 , 6 , 10 (circles) and never again, because the red curve grows exponentially.
Preprints 244177 g004
Figure 5. The decomposition of the Dirac module for a null frame of rank k = 3 . The vertices are the joint eigenspaces Δ a of the number operators P i ↑ , each a Dirac module of the residual space V ′ . The gaugino lives in the fully occupied corner K; the supersymmetry parameters that act on the residual components of the gauge field live in the opposite corner K ^ .
Figure 5. The decomposition of the Dirac module for a null frame of rank k = 3 . The vertices are the joint eigenspaces Δ a of the number operators P i ↑ , each a Dirac module of the residual space V ′ . The gaugino lives in the fully occupied corner K; the supersymmetry parameters that act on the residual components of the gauge field live in the opposite corner K ^ .
Preprints 244177 g005
Figure 6. Relative residual of the least-squares solution of the absorption equation (21) for real frames, plotted as 16 + log 10 ( residual ) against s and k. Blue: ( 9 , 1 ) , ( 10 , 2 ) , ( 11 , 3 ) , ( 12 , 4 ) , ( 13 , 5 ) with k = t − 1 , and ( 6 , 6 ) , ( 7 , 7 ) with k = 1 , 2 , all with D ′ = 10 ; the residuals are below 3 × 10 − 15 . Red: ( 17 , 1 ) , ( 18 , 2 ) , ( 19 , 3 ) , ( 20 , 4 ) with every tested k, D ′ ∈ { 16 , 18 , 20 , 22 } ; the residuals lie between 0.97 and 1, so no solution is found in these samples; the exact exclusion follows from the analytic theorem.
Figure 6. Relative residual of the least-squares solution of the absorption equation (21) for real frames, plotted as 16 + log 10 ( residual ) against s and k. Blue: ( 9 , 1 ) , ( 10 , 2 ) , ( 11 , 3 ) , ( 12 , 4 ) , ( 13 , 5 ) with k = t − 1 , and ( 6 , 6 ) , ( 7 , 7 ) with k = 1 , 2 , all with D ′ = 10 ; the residuals are below 3 × 10 − 15 . Red: ( 17 , 1 ) , ( 18 , 2 ) , ( 19 , 3 ) , ( 20 , 4 ) with every tested k, D ′ ∈ { 16 , 18 , 20 , 22 } ; the residuals lie between 0.97 and 1, so no solution is found in these samples; the exact exclusion follows from the analytic theorem.
Preprints 244177 g006
Figure 7. The signatures ( s , t ) and ranks k that carry a null-projected super Yang–Mills theory (Corollary 5.15). Blue: Nishino’s signatures ( 9 + k ) + ( 1 + k ) . Green: the signatures ( 5 + k ) + ( 5 + k ) and ( 1 + k ) + ( 9 + k ) reached from them by complex isometries that keep D and k (Theorem 5.12); the large green points are 9 + 9 and 9 + 17 of the second degenerate family. Purple: residual dimension two, ( 1 + k ) + ( 1 + k ) (Theorem 5.14). Red: the signatures ( 11 + 8 n ) + 3 , n = 0 , 1 , 2 , with every rank k ≤ 3 ; only Sezgin’s 11 + 3 with k = 2 meets the plane s + t − 2 k = 10 .
Figure 7. The signatures ( s , t ) and ranks k that carry a null-projected super Yang–Mills theory (Corollary 5.15). Blue: Nishino’s signatures ( 9 + k ) + ( 1 + k ) . Green: the signatures ( 5 + k ) + ( 5 + k ) and ( 1 + k ) + ( 9 + k ) reached from them by complex isometries that keep D and k (Theorem 5.12); the large green points are 9 + 9 and 9 + 17 of the second degenerate family. Purple: residual dimension two, ( 1 + k ) + ( 1 + k ) (Theorem 5.14). Red: the signatures ( 11 + 8 n ) + 3 , n = 0 , 1 , 2 , with every rank k ≤ 3 ; only Sezgin’s 11 + 3 with k = 2 meets the plane s + t − 2 k = 10 .
Preprints 244177 g007
Figure 8. Massless multiplets in ( 9 + 8 n ) + 1 , plotted as log 2 log 2 of the number of states. Blue: a vector field and a Majorana–Weyl gaugino, ( 8 + 8 n ) + 2 3 + 4 n states. Red: the least number 2 2 2 + 4 n of states of a massless multiplet without tensor charges (Theorem 6.1). The two agree only for D = 10 .
Figure 8. Massless multiplets in ( 9 + 8 n ) + 1 , plotted as log 2 log 2 of the number of states. Blue: a vector field and a Majorana–Weyl gaugino, ( 8 + 8 n ) + 2 3 + 4 n states. Red: the least number 2 2 2 + 4 n of states of a massless multiplet without tensor charges (Theorem 6.1). The two agree only for D = 10 .
Preprints 244177 g008
Figure 9. The transfer in the space of signatures and ranks ( s , t , k ) . Blue: Nishino’s real frames of rank 4 n + 2 in ( 11 + 4 n ) + ( 3 + 4 n ) for n = 0 , 1 , 2 , all below the wall k = min ( s , t ) (grey). Dashed: moving 4 n times into space. Red circles: the same ranks in ( 11 + 8 n ) + 3 , above the wall, since a real frame there has at most three null vectors. Green stars: the rank k = 2 , the only one compatible with the translation bracket; at n = 0 it is Sezgin’s theory in 11 + 3 .
Figure 9. The transfer in the space of signatures and ranks ( s , t , k ) . Blue: Nishino’s real frames of rank 4 n + 2 in ( 11 + 4 n ) + ( 3 + 4 n ) for n = 0 , 1 , 2 , all below the wall k = min ( s , t ) (grey). Dashed: moving 4 n times into space. Red circles: the same ranks in ( 11 + 8 n ) + 3 , above the wall, since a real frame there has at most three null vectors. Green stars: the rank k = 2 , the only one compatible with the translation bracket; at n = 0 it is Sezgin’s theory in 11 + 3 .
Preprints 244177 g009
Figure 10. At rank two in ( 11 + 8 n ) + 3 , the dimension N = 2 4 + 4 n of the residual Weyl module (red) against the dimension 2 ( D ′ − 2 ) = 16 + 16 n that absorption of the cubic terms requires by Theorem 3.3 (blue), on a logarithmic scale. They agree only for 11 + 3 , Sezgin’s theory.
Figure 10. At rank two in ( 11 + 8 n ) + 3 , the dimension N = 2 4 + 4 n of the residual Weyl module (red) against the dimension 2 ( D ′ − 2 ) = 16 + 16 n that absorption of the cubic terms requires by Theorem 3.3 (blue), on a logarithmic scale. They agree only for 11 + 3 , Sezgin’s theory.
Preprints 244177 g010
Figure 11. The degrees p ∈ P ′ in which the projected cubic terms have nonzero components (Theorem 7.3), for residual dimensions D ′ = 2 , … , 30 . Bar heights are log 10 D ′ p , the dimension of the forms that a source must supply in degree p. Grey: P ′ is empty and the cubic terms vanish after projection. Green: one degree of dimension at most D ′ , so a source can absorb them ( D ′ = 8 and D ′ = 10 ). Orange: one degree, too large unless the frame has rank at least 54 ( D ′ = 12 , 14). Red: two or more degrees, which no source of a single degree can supply ( D ′ ≥ 16 , where every signature of (2) with n ≥ 1 lies).
Figure 11. The degrees p ∈ P ′ in which the projected cubic terms have nonzero components (Theorem 7.3), for residual dimensions D ′ = 2 , … , 30 . Bar heights are log 10 D ′ p , the dimension of the forms that a source must supply in degree p. Grey: P ′ is empty and the cubic terms vanish after projection. Green: one degree of dimension at most D ′ , so a source can absorb them ( D ′ = 8 and D ′ = 10 ). Orange: one degree, too large unless the frame has rank at least 54 ( D ′ = 12 , 14). Red: two or more degrees, which no source of a single degree can supply ( D ′ ≥ 16 , where every signature of (2) with n ≥ 1 lies).
Preprints 244177 g011
Table 1. The transfer of Nishino’s construction from ( 11 + 4 n ) + ( 3 + 4 n ) to ( 11 + 8 n ) + 3 for n = 1 , 2 (Proposition 7.1). The rank of Nishino’s frame exceeds the Witt index 3 of the target; the only rank compatible with the translation bracket is 2; and at rank 2 the residual Weyl module has N components where absorption needs 2 ( D ′ − 2 ) .
Table 1. The transfer of Nishino’s construction from ( 11 + 4 n ) + ( 3 + 4 n ) to ( 11 + 8 n ) + 3 for n = 1 , 2 (Proposition 7.1). The rank of Nishino’s frame exceeds the Witt index 3 of the target; the only rank compatible with the translation bracket is 2; and at rank 2 the residual Weyl module has N components where absorption needs 2 ( D ′ − 2 ) .
target source Nishino’s k largest real k k allowed by (c) D ′ N 2 ( D ′ − 2 )
( 19 , 3 ) ( 15 , 7 ) 6 3 2 18 256 32
( 27 , 3 ) ( 19 , 11 ) 10 3 2 26 4 096 48
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.