The Vacuum and the Nilpotent Ladder
Introduction
This entry reads the last of the five parts of the tenth section: the stratum on which both predicates fail.
$$ \Sigma_{00}=\{N=0,\ B^{2}=0\},\qquad N=Q_t^{2}+B^{2},\qquad B^{2}=Q_x^{2}+Q_y^{2}+Q_z^{2}. $$
Both predicates are false, and this is the only stratum of which that is true. The element has no inverse: it is a zero divisor of the plain product. The two eigenvalues coincide: they are equal and they are both zero. The first section of the series separated the two questions — can the element be inverted, can its two eigenvalues be told apart? — and this is the single stratum on which the two answers are the same answer.
The first claim is that the failure is total and simultaneous. Invertibility, the eigenbasis, the Peirce decomposition attached to the element, the spectral decomposition, the polar word and the logarithm all fail here, and they fail for one reason: the element is square zero. Nothing about this stratum is a partial failure, and nothing about it is a failure of the algebra. The algebra keeps working; it is the element that has collapsed.
The second claim is that the collapse is not an absence of structure. A square-zero element still carries a flag $0\subset\mathbb{C}\vec\Upsilon\subset\ker L_{\tilde Q}\subset\mathbb{B}$, still has a Jordan block, still generates a flow, and still has a left and a right annihilator. The structure is thinner than on the other three strata and it is not empty. What is lost is the spectrum, not the geometry.
The third claim is that the origin is the one point of the stratum that is not a nilpotent. The zero element is square zero and semisimple; every other element of the stratum is square zero and defective. So the stratum mixes a single semisimple point with a four-real-dimensional family of defective ones, and the semisimple point is the apex of both cones. That is why the reading of the origin is a separate reading from the reading of the nilpotent.
The fourth claim is that the origin is the whole of the stratum's material content. On the four-momentum chart the vector square is positive definite, so a nilpotent would need a real isotropic spatial vector, and there is none. The intersection with the material sector is the single point $\{0\}$. The archive article read the region as the vacuum and the ladder together; the ladder is not material, and this entry states the correction at the place where it is due.
The fifth claim is that the nilpotent is the null electromagnetic field. The field strength of the corpus is a pure element, so the stratum's test applies to it as it stands, and $\tilde F^{2}=0$ is exactly $|\vec E|=c|\vec B|$ with $\vec E\perp\vec B$. Six real data, two real conditions, four left: the amplitude, the direction of propagation and one polarisation angle. The complex scale of the element is not idle either: its modulus is the amplitude and its phase is the duality rotation, which turns $\vec E$ into $\vec B$ and back.
The sixth claim is that the ladder needs a pair, and that the pair is a two-rung system and nothing more. A single nilpotent is a square-zero direction with no state attached to it. Two nilpotents with the anticommutator $e_t$ produce the two Hermitian rungs, and the rungs are what a state is read on. The algebra is a full matrix algebra on a two-dimensional module, so there are exactly two rungs and no third.
The interpretation labels continue the section's sequence. The partition entry carries I.10.a to I.10.l, the mass-in-motion entry I.10.m to I.10.x, the rest-frame entry I.10.y to I.10.ah, and the massless entry I.10.ai to I.10.ar. This entry carries I.10.as to I.10.bd.
The readings are readings and not a model. The stratum is cut on the algebra by two equations. Every name below is a reading of the part of it that the material and the field charts reach, and §The Limits states what the readings do not carry. In particular no dynamics, no field equation and no statistics are supplied by the stratum, and the word vacuum names a role inside the algebra and not a theory of the vacuum.
The Source Table
| Mathematical Content | Source Mathematical Article |
|---|---|
| the sector, its three descriptions, the geometry, the elements, the annihilators, the Peirce data, the origin and the products | The N-Degenerate, Spectrally Degenerate Sector |
| the canonical Jordan basis of a defective element, and the pair selected by the Hermitian form | The Jordan Basis of a Spectrally Degenerate Biquaternion |
| the characteristic polynomial, the two eigenvalues and the discriminant | The Spectrum |
| the two predicates, the independence of the cones and the single point where both degenerations coincide | Relations between the N Norm and Spectral Regularity and Degeneracy |
| the corner rule, the off-diagonal nilpotents and the dimensions of the four corners | The Peirce Decomposition of the Biquaternion Algebra |
| the quaternionic norm, the null cone, its pure stratum and the group of invertibles | N-Regularity and N-Degeneracy in the Biquaternion Algebra |
| the minimal left and right ideals, the idempotents and the projective structure of the module | The Polarized Module of the Biquaternion Algebra |
| the exponential, its degenerate boundary and the truncation of the series | Biquaternion Elementary Functions |
| the place of $B^{2}$ outside the four norms, and the multiplicativity of the quaternionic one | The Four Forms and the Four Algebraic Norms over the Biquaternion Algebra |
| the rank of the two one-sided multiplications, the annihilators and the two-sided operator | The Four General Products and Their Relations over the Biquaternion Algebra |
The first row owns everything structural. The remaining rows supply what the entry uses and does not build: the Jordan basis, the spectral criterion, the Peirce corners, the ideal calculus of the module, and the exponential with its truncated boundary.
The Results in Brief
The Sector
The stratum is the intersection of the two cones, and it has three descriptions that coincide.
$$ \Sigma_{00}=\mathcal N\cap\mathcal S =\{Q_t=0,\ \vec Q^{2}=0\} =\{\tilde Q:\tilde Q^{2}=0\}. $$
The three descriptions are one theorem. An element is in the sector exactly when it is null and simultaneously spectrally degenerate, exactly when its scalar part vanishes and its vector part is isotropic, and exactly when its square is zero. The proof is the separation of $\tilde Q^{2}=(Q_t^{2}-B^{2})e_t+2Q_t\vec Q$ into central and vector parts, and it is the shortest proof in the section.
The origin is the only element that is square zero without being a nilpotent of index two. The word nilpotent is reserved for non-zero elements. The sector is therefore the nilpotent cone with its apex adjoined, and the apex behaves differently from every other point.
| question | answer on the sector |
|---|---|
| complex dimension | $2$ |
| real dimension | $4$ |
| smooth? | smooth off the origin, singular at the origin only |
| projective base | a smooth conic, hence $\mathbb{P}^{1}$ |
| fibre | $\mathbb{C}^{\times}$ |
| elements | the origin, and the non-zero nilpotents of index two |
The Element
Off the origin the element is a pure isotropic vector, and it is defective.
$$ \vec\Upsilon=Q_x\vec e_x+Q_y\vec e_y+Q_z\vec e_z,\qquad Q_x^{2}+Q_y^{2}+Q_z^{2}=0,\qquad \vec\Upsilon^{2}=0 . $$
Its spectrum is $\{0\}$ and its minimal polynomial is $X^{2}$. A minimal polynomial of degree two over a spectrum of one point is the definition of a defect: the element is a single Jordan block of size two, the eigenspace does not span, and no eigenbasis exists.
Its centraliser is the dual numbers.
$$ \mathbb{C}[\vec\Upsilon]=\{a\,e_t+b\,\vec\Upsilon\}\cong\mathbb{C}[\varepsilon]/(\varepsilon^{2}), $$
so its only idempotents are $0$ and $e_t$, and no non-trivial idempotent commutes with the element. That is the sharpest single contrast with the massless stratum, where the centraliser was $\mathbb{C}\oplus\mathbb{C}$ and carried four idempotents.
The Operator
Both one-sided multiplications are square zero of rank two, and the image equals the kernel.
$$ L_{\tilde Q}^{2}=R_{\tilde Q}^{2}=0,\qquad \operatorname{rank}L_{\tilde Q}=\operatorname{rank}R_{\tilde Q}=2,\qquad \operatorname{im}L_{\tilde Q}=\ker L_{\tilde Q}. $$
The plane that is annihilated is the plane that is reached. On the massless stratum the image and the kernel were two different planes, so a live plane survived beside a dead one. Here there is one plane, and the operator shifts the whole algebra into it.
The two annihilators are different planes and they are exchanged by the conjugation.
$$ \ker L_{\tilde Q}=\tilde Q\mathbb{B},\qquad \ker R_{\tilde Q}=\mathbb{B}\tilde Q,\qquad \ker L_{\tilde Q}\cap\ker R_{\tilde Q}=\mathbb{C}\tilde Q,\qquad (\ker L_{\tilde Q})^{\natural}=\ker R_{\tilde Q}. $$
The Origin
The origin is scalar, central, semisimple and of rank zero. Its minimal polynomial is $X$, its spectrum is $\{0\}$ with no defect, both multiplications are the zero map, and every idempotent of the algebra commutes with it. It is the only element of the stratum for which a spectral decomposition exists, and the decomposition is the one-term writing $0=0\cdot e_t$.
It lies in the closure of every stratum. Every stratum is invariant under $\tilde Q\mapsto\lambda\tilde Q$ for $\lambda\neq0$, so every punctured neighbourhood of the origin meets all four. The origin is the point at which the four regimes touch.
The Flow
The exponential truncates after one term.
$$ e^{\theta\tilde Q}=e_t+\theta\tilde Q,\qquad N\bigl(e^{\theta\tilde Q}\bigr)=1,\qquad \bigl\langle e^{\theta\tilde Q}\bigr\rangle_{*}=1+\theta^{2}\lVert\vec\Upsilon\rVert^{2}. $$
The flow is a shear and not a rotation. Its matrix is unipotent: every eigenvalue is $1$, the determinant is $1$, and the matrix is not the identity. The Hermitian norm grows without bound in $\theta$, polynomially and not exponentially. So the element generates a flow that is invertible at every parameter and that is never a bounded one.
No nilpotent is an exponential. The norm of every exponential is $e^{2X_t}$ and the norm of every element of the stratum is zero, so the stratum is disjoint from the image of the exponential and no nilpotent has a logarithm.
The Products
The stratum is not closed under multiplication, and where the product goes is decided by one scalar.
$$ \vec\Upsilon_2\vec\Upsilon=-\vec\Upsilon_2\cdot\vec\Upsilon+\vec\Upsilon_2\times\vec\Upsilon, \qquad \Sigma_{00}\cdot\Sigma_{00}\subseteq\{0\}\cup\Sigma_{01}. $$
The product is zero exactly when the two isotropic vectors are proportional, and otherwise it is a light-ray element. The scalar part is the negative of the bilinear pairing, and the vector square of the product is the negative square of that same pairing. So the product of two nilpotents is null in every case and spectrally degenerate in none but the proportional case.
The Ladder
The physical reading of the nilpotent needs a pair.
$$ \vec\Upsilon_x=\tfrac12\bigl(i\vec e_x-\vec e_y\bigr),\qquad \vec\Upsilon_y=\tfrac12\bigl(i\vec e_x+\vec e_y\bigr),\qquad \{\vec\Upsilon_x,\vec\Upsilon_y\}=e_t . $$
Their products are the two rungs.
$$ \tilde\Pi_{+}=\vec\Upsilon_x\vec\Upsilon_y=\tfrac12\bigl(e_t+i\vec e_z\bigr),\qquad \tilde\Pi_{-}=\vec\Upsilon_y\vec\Upsilon_x=\tfrac12\bigl(e_t-i\vec e_z\bigr), $$
and the rungs are not in the stratum. They are Hermitian, idempotent, and of vector square $-\tfrac14$, so they lie in the massless stratum. The ladder straddles the two degenerate strata: its operators are here and its states are next door.
The Field
The null electromagnetic field is the stratum's most direct physical occupant.
$$ \tilde F=i\sqrt{\epsilon}\,\vec E-\sqrt{\mu}\,\vec H,\qquad \tilde F^{2}=\epsilon\bigl(I_1+2icI_2\bigr)e_t,\qquad I_1=\vec E^{2}-c^{2}\vec B^{2},\quad I_2=\vec E\cdot\vec B . $$
So the stratum's condition on the field is the classical nullness condition. Two real equations on six real data, and the dictionary between the field pair and the pure element is a bijection.
The Sector
The three descriptions are proved by one identity.
$$ \tilde Q^{2}=(Q_t^{2}-B^{2})e_t+2Q_t\vec Q . $$
The square separates into a central part and a vector part, and the two must vanish separately. If $Q_t\neq0$ then $\vec Q=0$, hence $B^{2}=0$, hence $Q_t^{2}=0$, a contradiction; so $Q_t=0$, and then $B^{2}=0$. Conversely $Q_t=0$ with $B^{2}=0$ gives $\tilde Q^{2}=-B^{2}e_t=0$. The equivalence of the second and third descriptions follows, and the first follows from $N=Q_t^{2}+B^{2}$ together with $B^{2}=0$.
The variety is a complex cone of dimension two.
$$ \dim_{\mathbb{C}}\Sigma_{00}=2,\qquad \dim_{\mathbb{R}}\Sigma_{00}=4,\qquad \Sigma_{00}\subset\mathbb{B}\cong\mathbb{C}^{4}. $$
It is the common zero set of $Q_t$ and $B^{2}$, and those two gradients are independent off the apex. They are $(1,0,0,0)$ and $(0,2Q_x,2Q_y,2Q_z)$, so off the origin the intersection is smooth of complex dimension $4-2=2$. At the origin the second gradient vanishes: the Zariski tangent space there is $\{Q_t=0\}$, of dimension three, one more than the variety. So the apex is the only singular point, and the singularity is the ordinary one of a cone at its vertex.
The projective base is a conic, and a conic in the plane is a sphere.
$$ \Sigma_{00}\setminus\{0\}\ \longrightarrow\ \mathbb{P}^{1},\qquad \{Q_x^{2}+Q_y^{2}+Q_z^{2}=0\}\subset\mathbb{P}^{2}. $$
So the stratum is a $\mathbb{C}^{\times}$-bundle over a sphere. The base is two real dimensions and the fibre is two real dimensions, which is the real dimension four. The base is the same sphere that the massless stratum carried as its material base, and here it is the base of the whole complex cone and not only of a real slice.
Every point of the stratum is a real scale times a complex isotropic direction.
$$ \tilde Q=r\bigl(\vec u+i\vec u_\perp\bigr),\qquad r\ge0,\quad \vec u\ \text{a real unit vector},\quad \vec u_\perp\ \text{a real unit vector perpendicular to}\ \vec u, $$
and the count is $1+2+1=4$. The scale is one real parameter, the direction $\vec u$ is two, and the angle of $\vec u_\perp$ inside the plane perpendicular to $\vec u$ is one. On the bivector split $\vec Q=\vec\rho+i\vec\rho'$ the condition is $\lvert\vec\rho\rvert=\lvert\vec\rho'\rvert$ with $\vec\rho\perp\vec\rho'$, which is the same statement in real coordinates. Over fifty parametrised points the square, the two moduli and the orthogonality held to $4.6\times10^{-14}$.
The two cones are tangent along the whole stratum, and that is a different statement from the smoothness.
$$ \vec\nabla_\Phi N=2\bigl(Q_t,\vec Q\bigr),\qquad \vec\nabla_\Phi B^{2}=2\bigl(0,\vec Q\bigr),\qquad \vec\nabla_\Phi N=\vec\nabla_\Phi B^{2}\ \text{whenever }Q_t=0 . $$
So the two hypersurfaces meet without crossing. As point sets the stratum is their intersection, and along it the two normal directions coincide: a first-order displacement that keeps one condition keeps the other. The defining equations of the stratum are nevertheless $Q_t$ and $B^{2}$ and not $N$ and $B^{2}$, and those two gradients are independent off the apex, which is why the variety is smooth there even though the two cones are tangent. The stratum is a complete intersection of codimension two and not a tangency locus with a doubled equation.
As a scheme the intersection is non-reduced.
$$ (Q_t^{2}+B^{2},\ B^{2})=(Q_t^{2},\ B^{2}),\qquad \text{reduction }(Q_t,\ B). $$
The ideal is generated by two squares. The point set is the reduced variety and the multiplicity sits in the two normal directions, which is the algebraic form of the tangency: the intersection has the expected dimension and twice the expected multiplicity.
The Element
Off the apex the element is a pure isotropic vector.
$$ \tilde Q=\vec\Upsilon,\qquad \vec\Upsilon=Q_x\vec e_x+Q_y\vec e_y+Q_z\vec e_z,\qquad \vec\Upsilon^{2}=0,\qquad \vec\Upsilon\neq0 . $$
The spectrum is one point and the minimal polynomial is its square.
$$ \operatorname{spec}\tilde Q=\{0\},\qquad m_{\tilde Q}(X)=X^{2},\qquad \tilde Q^{n}=0\ \ (n\ge2). $$
A minimal polynomial of degree two with a single root is a defect, and it is the exact opposite of what the massless stratum had. There the minimal polynomial was $X(X-\kappa)$ with distinct roots, so the element was semisimple and the spectral idempotents were unique. Here the two roots have merged, the polynomial is not squarefree, and there is no idempotent attached to the element.
The centraliser is thin.
$$ \mathbb{C}[\vec\Upsilon]=\{a\,e_t+b\,\vec\Upsilon\ :\ a,b\in\mathbb{C}\}\cong\mathbb{C}[\varepsilon]/(\varepsilon^{2}). $$
And it holds no idempotent but the two trivial ones. Solving $a^{2}=a$ and $2ab=b$ gives $(a,b)=(0,0)$ and $(a,b)=(1,0)$. So the centraliser's idempotents are $0$ and $e_t$, and the element commutes with no rung. Computed: no random element in one hundred draws commuted with a nilpotent, and no primitive idempotent in fifty draws did.
The Peirce form exists relative to an adapted idempotent and is purely off-diagonal.
$$ e=\tfrac12\bigl(e_t+i\vec e_z\bigr),\qquad f=e_t-e,\qquad \vec X=\tfrac12\bigl(\vec e_x-i\vec e_y\bigr), $$
$$ e\tilde Qe=0,\qquad e\tilde Qf=\tilde Q,\qquad f\tilde Qe=0,\qquad f\tilde Qf=0 . $$
The element lies wholly in the mixed corner. In the basis $(e,\tilde Q,\vec X,f)$ the Peirce matrix is the single nilpotent block
$$ \mathsf M_2(\tilde Q)=\begin{pmatrix}0&1\\0&0\end{pmatrix}=J_2(0), $$
and the four corner identities together with $\vec X^{2}=0$ held with residuals of order $10^{-16}$ or exactly zero.
No Peirce decomposition relative to the element itself can exist. Such a decomposition would place $\tilde Q$ in the diagonal corners $\mathbb{B}_{++}\oplus\mathbb{B}_{--}$, and the mixed-corner computation says it is nowhere else. The Peirce data therefore survive only as a decomposition of the algebra relative to somebody else's idempotent, and that idempotent is a gauge and not a function of the element.
The eigenbasis fails for a reason that is visible in one formula.
$$ \vec{\mathrm Z}=-i\frac{\vec Q}{B}\ \text{needs}\ B\neq0, \qquad B^{2}=0\ \text{here}. $$
So both roots that the regular construction divides by do not exist. The massless stratum built its eigenbasis from $\vec\zeta=\vec Q/Q_t$ and from the root $\vec{\mathrm Z}$; the first is undefined because $Q_t=0$ and the second because $B=0$. Both normalisations vanish at once, and that is the whole of the eigenbasis failure.
What survives is a flag.
$$ 0\subset\mathbb{C}\vec\Upsilon\subset\ker L_{\tilde Q}\subset\mathbb{B}, \qquad \dim_{\mathbb{C}}\ker L_{\tilde Q}=2 . $$
A flag is thinner than a decomposition and it is not nothing. A decomposition names two complementary subspaces; a flag names a nested pair. The element supplies the first two steps of the flag and its own annihilator supplies the third, and no choice is made in the construction.
The Operator
Left and right multiplication are square zero of rank two.
$$ L_{\tilde Q}^{2}=L_{\tilde Q^{2}}=0,\qquad R_{\tilde Q}^{2}=R_{\tilde Q^{2}}=0,\qquad \operatorname{rank}L_{\tilde Q}=\operatorname{rank}R_{\tilde Q}=2 . $$
Neither is injective and neither is surjective, and the two failures are the same failure. A square-zero operator has its image inside its kernel; with both of rank two in a four-dimensional space the inclusion is an equality.
$$ \operatorname{im}L_{\tilde Q}=\ker L_{\tilde Q}=\tilde Q\mathbb{B},\qquad \operatorname{im}R_{\tilde Q}=\ker R_{\tilde Q}=\mathbb{B}\tilde Q . $$
So the plane the operator kills is the plane it reaches. The image is a minimal right ideal of complex dimension two; over fifty nilpotents the sum of image and kernel had dimension exactly two. On the massless stratum the same rank appeared with a complementary image, so a live plane survived beside the annihilated one; here there is one plane and the map is a shift into it.
The two annihilator planes are different and the conjugation exchanges them.
$$ \ker L_{\tilde Q}\cap\ker R_{\tilde Q}=\mathbb{C}\tilde Q,\qquad \bigl(\ker L_{\tilde Q}\bigr)^{\natural}=\ker R_{\tilde Q}, $$
and the identity behind the second statement is one line. From $\tilde Q^{\natural}=-\tilde Q$ one gets $X\tilde Q=0$ exactly when $\tilde QX^{\natural}=0$. The left annihilator is a right ideal, the right annihilator is a left ideal, and a common two-dimensional subspace would be a two-sided ideal of a simple algebra, which a plane is not. Their intersection is therefore a line, and the line is the element itself.
On the algebra the operator has two Jordan blocks of size two.
$$ \operatorname{rank}L_{\tilde Q}=2,\qquad \operatorname{rank}L_{\tilde Q}^{2}=0,\qquad \text{blocks of size }2:\ 2,\qquad L_{\tilde Q}\sim J_2(0)\oplus J_2(0). $$
So the element is one block and the operator is two. The number of blocks of size two is the drop in rank from $L$ to $L^{2}$, which is two; the element itself, read in its own Peirce basis, is one block of size two. The doubling is the same doubling that appeared on the massless stratum as a rank that doubled on each side: multiplication acts on a four-dimensional space, the element lives in a two-dimensional one, and every rank and block count doubles when the element is turned into an operator.
The trace and the determinant of both operators vanish. That is the operator form of the same statement: the spectrum is $\{0\}$ with multiplicity four and the characteristic polynomial is $X^{4}$. There is no trace to normalise by, which is one more reason the stratum has no invariant of modulus type.
| question | on $\Sigma_{01}$ | on $\Sigma_{00}$ |
|---|---|---|
| $L^{2}$ | $\kappa L$ | $0$ |
| rank | $2$ | $2$ |
| image | a plane complementary to the kernel | equal to the kernel |
| the blocks of $L$ | two of size one and two of size two | two of size two |
| what survives | a live plane beside a dead one | nothing at any scale |
The last row is the reading. A square-zero operator is the algebraic form of a one-way coupling: one direction feeds another and nothing returns. The reading is exact here because the image and the kernel coincide, so there is no frequency at which the fed direction comes back, and no power of the operator produces anything the operator did not already produce.
The Flow
The exponential is exact after one term, and the verification is the truncation itself.
$$ e^{\theta\tilde Q}=e_t+\theta\tilde Q,\qquad \bigl\lvert e^{\theta\tilde Q}-(e_t+\theta\tilde Q)\bigr\rvert\le7.4\times10^{-16} $$
over fifty random elements and parameters. The series terminates because $\tilde Q^{2}=0$, so no approximation is involved: the exponential of a nilpotent is its own first-order Taylor polynomial, and that is a statement about the algebra and not about a parameter being small.
The flow is unipotent.
$$ \operatorname{spec}\bigl(e^{\theta\tilde Q}\bigr)=\{1\},\qquad \det\bigl(e^{\theta\tilde Q}\bigr)=1,\qquad \operatorname{tr}\bigl(e^{\theta\tilde Q}\bigr)=4,\qquad e^{\theta\tilde Q}\neq e_t\ \text{for }\theta\neq0, $$
checked over fifty elements to $2.1\times10^{-15}$. Four eigenvalues equal to one, a determinant of one, a trace equal to the dimension, and a matrix that is not the identity: that is a transvection, a shear with no scale, no rotation and no eigendirection beyond the ones it shifts.
The Hermitian norm grows polynomially and without bound.
$$ \bigl\langle e^{\theta\tilde Q}\bigr\rangle_{*} =\operatorname{Sc}\bigl(e^{\theta\tilde Q}\,(e^{\theta\tilde Q})^{*}\bigr) =1+\theta^{2}\lVert\vec\Upsilon\rVert^{2}, \qquad \lVert\vec\Upsilon\rVert^{2}=\sum_k\lvert\Upsilon_k\rvert^{2}, $$
checked to $1.8\times10^{-15}$ over fifty elements. The norm is a quadratic polynomial in the parameter with a positive quadratic coefficient, so the flow is invertible at every $\theta$ — its quaternionic norm is one — and it is never a bounded unit. The two notions of unit separate here, and this is the stratum where they separate furthest: the quaternionic norm is constant and the Hermitian norm is unbounded.
Conjugation by the flow preserves the norm of every element.
$$ N\bigl(e^{\theta\tilde Q}\tilde X e^{-\theta\tilde Q}\bigr)=N(\tilde X), \qquad \bigl\lvert\text{residual}\bigr\rvert\le1.4\times10^{-14} $$
over fifty random pairs. So the shear is a symmetry of the null cone and not a motion inside it: it moves nilpotents among nilpotents, and it moves every other element along its own level of the norm.
No nilpotent is the exponential of anything.
$$ N\bigl(e^{X}\bigr)=e^{2X_t}\ \text{for every }X\in\mathbb{B}, \qquad N(\tilde Q)=0, \qquad e^{2X_t}\neq0, $$
so the stratum is disjoint from the image of the exponential. The identity was checked over one hundred random elements to $2.1\times10^{-12}$, the residual being the truncation of the series and not a discrepancy. It is the identity that excluded the massless stratum, and the argument is the same argument: the norm of an exponential is never zero, and the norm of the element is zero.
The four strata carry the four behaviours of a one-parameter group.
| stratum | flow | type |
|---|---|---|
| $\Sigma_{11}$, the open stratum | $e^{\theta Q_t}\bigl(\cos(\theta B)e_t+\sin(\theta B)\vec U\bigr)$ | elliptic with a scale |
| $\Sigma_{10}$, the rest frame | a repeated eigenvalue with a defect linear in $\theta$ | parabolic with a scale |
| $\Sigma_{01}$, the light ray | the same formula, with a root and no defect | elliptic with a scale |
| $\Sigma_{00}$, the nilpotent | $e_t+\theta\tilde Q$, every eigenvalue $1$ | parabolic, unipotent |
The fourth row is the pure case of the third type. On the other strata a scale multiplies the group, so no eigenvalue is exactly one; here the scalar part vanishes and every eigenvalue is exactly one, which is unipotence and nothing else. A one-parameter group whose every element has all eigenvalues equal to one is a shear, and a shear is the algebraic form of a transformation that translates along a direction without turning and without scaling.
The Null Field
The field strength of the corpus is a pure element, so the stratum's test applies to it as it stands.
$$ \tilde F=i\sqrt{\epsilon}\,\vec E-\sqrt{\mu}\,\vec H,\qquad \vec B=\mu\vec H,\qquad c=\frac{1}{\sqrt{\epsilon\mu}} . $$
Its square is one line of algebra and it separates the two classical invariants.
$$ \tilde F^{2}=\epsilon\bigl(I_1+2icI_2\bigr)e_t, \qquad I_1=\vec E^{2}-c^{2}\vec B^{2}, \qquad I_2=\vec E\cdot\vec B, $$
checked over fifty random real fields to $4.0\times10^{-15}$. The computation is the product of two pure vectors: with $\vec V=\sqrt{\epsilon}\vec E$ and $\vec W=\sqrt{\mu}\vec H$ one has $\tilde F=i\vec V-\vec W$ and $\tilde F^{2}=-\vec V^{2}+\vec W^{2}+2i\vec V\cdot\vec W$.
The stratum's condition on a real field is therefore the classical nullness condition.
$$ \tilde F^{2}=0\iff \lvert\vec E\rvert=c\lvert\vec B\rvert \ \text{and}\ \vec E\cdot\vec B=0 . $$
Both directions hold and both were checked. One hundred constructed null pairs — equal magnitudes in the units of $c$ and perpendicular halves — gave a vanishing square in all one hundred cases, and since $\epsilon\neq0$ the equation $\tilde F^{2}=0$ is the pair of real equations $I_1=0$ and $I_2=0$, one from each part of the complex number.
The dictionary is a bijection.
$$ \vec E=\frac{\operatorname{Im}\tilde F}{\sqrt{\epsilon}},\qquad \vec H=-\frac{\operatorname{Re}\tilde F}{\sqrt{\mu}}, $$
so the real null fields and the non-zero elements of the stratum are the same four real data read in two alphabets. Nothing is lost and nothing is added, and the map is inverted by the same two lines.
A member, checked. Take $\tilde F=iV(\vec e_x+i\vec e_y)$ with $V>0$ real. Then $\vec E=(V/\sqrt{\epsilon})\vec e_x$ and $\vec B=\sqrt{\mu}V\vec e_y$, the halves are perpendicular, the magnitudes satisfy $c\lvert\vec B\rvert=\lvert\vec E\rvert$ because $c\sqrt{\mu}=1/\sqrt{\epsilon}$, and the square vanishes because $(\vec e_x+i\vec e_y)^{2}=0$.
The count is the stratum's dimension in the letters of the field. Six real components of $(\vec E,\vec B)$, two real conditions, four left: the amplitude, the direction of propagation which is two angles, and one angle that places the pair inside the transverse plane. That is $1+2+1$, the same count the parametrisation gave.
The complex scale has two physical readings and both are exact.
$$ \tilde F=\lambda\,\tilde F_0,\qquad \lvert\lambda\rvert\ \text{is the amplitude},\qquad \arg\lambda\ \text{is the duality angle}. $$
The modulus is the amplitude because the Hermitian norm of the field is proportional to it. For a null field $\lVert\tilde F\rVert^{2}=\epsilon\lvert\vec E\rvert^{2}+\lvert\vec B\rvert^{2}/\mu=2\epsilon\lvert\vec E\rvert^{2}$, so the modulus and the amplitude differ by the fixed factor $\sqrt{2\epsilon}$.
The phase turns $\vec E$ into $\vec B$ and back.
$$ e^{i\varphi}\tilde F\quad\longleftrightarrow\quad \vec E'=\vec E\cos\varphi-c\vec B\sin\varphi,\qquad \vec B'=\vec B\cos\varphi+\frac{\vec E}{c}\sin\varphi, $$
checked over fifty random null fields and angles to $4.4\times10^{-16}$. The transformation is the duality rotation, it preserves $\lvert\vec E\rvert=c\lvert\vec B\rvert$ and $\vec E\cdot\vec B=0$, and it therefore acts on the stratum as a symmetry of the fibre. The $\mathbb{C}^{\times}$ of the bundle is not an abstract scale: its modulus is an amplitude and its argument is a duality angle, and both are classical operations on a field.
The field straddles the two elementary sectors, which is why the stratum's reading is a field and not a momentum. The real part $-\sqrt{\mu}\vec H$ is anti-Hermitian and the imaginary part $i\sqrt{\epsilon}\vec E$ is Hermitian, so $\tilde F$ lies in neither $\mathbb{B}_{\Phi}$ nor $\mathbb{B}_{\Psi}$. Neither of those sectors contains a non-zero nilpotent, for the same reason in both cases: a real vector part has a positive-definite vector square and an imaginary one has a negative-definite square, and the stratum needs an isotropic one.
The Ladder
A single nilpotent is a direction and not yet a state, and the pair is what makes the state.
$$ \vec\Upsilon_x=\tfrac12\bigl(i\vec e_x-\vec e_y\bigr),\qquad \vec\Upsilon_y=\tfrac12\bigl(i\vec e_x+\vec e_y\bigr), $$
$$ \vec\Upsilon_x^{2}=\vec\Upsilon_y^{2}=0,\qquad \{\vec\Upsilon_x,\vec\Upsilon_y\}=e_t, $$
with the anticommutator exact. The two elements are nilpotents of the stratum, they are not proportional, and their symmetrised product is the identity.
Their products are the two rungs.
$$ \tilde\Pi_{+}=\vec\Upsilon_x\vec\Upsilon_y=\tfrac12\bigl(e_t+i\vec e_z\bigr),\qquad \tilde\Pi_{-}=\vec\Upsilon_y\vec\Upsilon_x=\tfrac12\bigl(e_t-i\vec e_z\bigr), $$
$$ \tilde\Pi_{\pm}^{2}=\tilde\Pi_{\pm},\qquad \tilde\Pi_{+}\tilde\Pi_{-}=\tilde\Pi_{-}\tilde\Pi_{+}=0,\qquad \tilde\Pi_{+}+\tilde\Pi_{-}=e_t, $$
and the rungs are Hermitian. All of these held with zero residual. The two products are not equal, because the multiplication is not commutative, so the order of the pair matters and the two rungs are the two orders.
The rungs are in the other degenerate stratum.
$$ N(\tilde\Pi_{\pm})=0,\qquad B^{2}(\tilde\Pi_{\pm})=-\tfrac14, $$
so an idempotent is null and spectrally regular, and the ladder straddles the two degenerate strata. Its operators are in $\Sigma_{00}$ and its states are in $\Sigma_{01}$, which is the same straddling the massless entry recorded from the other side.
The two sides of the multiplication sort the pair differently.
| product | value | product | value |
|---|---|---|---|
| $\vec\Upsilon_y\tilde\Pi_{+}$ | $\vec\Upsilon_y$ | $\tilde\Pi_{+}\vec\Upsilon_y$ | $0$ |
| $\vec\Upsilon_x\tilde\Pi_{+}$ | $0$ | $\tilde\Pi_{+}\vec\Upsilon_x$ | $\vec\Upsilon_x$ |
| $\vec\Upsilon_x\tilde\Pi_{-}$ | $\vec\Upsilon_x$ | $\tilde\Pi_{-}\vec\Upsilon_x$ | $0$ |
| $\vec\Upsilon_y\tilde\Pi_{-}$ | $0$ | $\tilde\Pi_{-}\vec\Upsilon_y$ | $\vec\Upsilon_y$ |
Right multiplication by a rung returns one nilpotent and kills the other; left multiplication returns the other and kills the first. The two readings of the same rung are therefore opposite, and the exactness of the eight products is what makes the pair a ladder rather than two unrelated operators.
There are two rungs and no third.
$$ \operatorname{Sc}\tilde\Pi_{\pm}=\tfrac12,\qquad \tilde\Pi_{+}+\tilde\Pi_{-}=e_t,\qquad \operatorname{Sc}e_t=1, $$
and the scalar part is the obstruction. Every primitive idempotent of this algebra has scalar part $\tfrac12$ — checked exactly over fifty of them — so a family of pairwise orthogonal non-zero idempotents summing to $e_t$ has at most two members. The bound is the dimension of the module and not a feature of this pair.
Nothing in the algebra supplies a third rung. The algebra is a full matrix algebra on a two-dimensional module, the module has two independent directions, and every maximal family of orthogonal idempotents has two members. So the ladder is a two-level system and not a truncation of a larger one.
The ladder is the bridge between two objects the algebra keeps apart. The nilpotents are not Hermitian, so they are not states in the Hermitian reading; their products are Hermitian, and the rungs are what the Hermitian form can weigh. The passage from an operator to a state is the product of two of them, and that passage is the ladder.
The Material Part
The stratum's material share is a single point.
$$ \tilde Q=i\,q'_t\,e_t+\vec r\ \text{real} \ \Longrightarrow\ B^{2}=\lvert\vec r\rvert^{2},\qquad N=-q_t'^{2}+\lvert\vec r\rvert^{2}, $$
so both conditions together force $\vec r=0$ and $q_t'=0$.
$$ \Sigma_{00}\cap\mathbb{B}_{\Phi}=\{0\}. $$
A definite form has no non-zero isotropic vector, and that is the whole proof. One hundred random material elements gave none in the stratum, and the only material elements with $B^{2}=0$ are on the scalar line, where the norm vanishes only at the origin.
So no four-momentum is a nilpotent, and the ladder is a statement about a carrier and not about a momentum. The archive article already carries this in its §The Material Case; what its title still obscures is that the region meets the material chart at the apex and at no other point.
The centre contributes the apex and nothing else, and so does the informational sector.
$$ \Sigma_{00}\cap\mathbb{B}_{C}=\{0\},\qquad \Sigma_{00}\cap\mathbb{B}_{\Psi}=\{0\}, $$
because on the centre the vector square vanishes identically and the norm is $q^{2}$, and on the informational sector the vector part is imaginary so its square is negative definite. A scalar has no ladder, and neither has an element of the informational sector.
Only the minimal left ideals reach the stratum, and each reaches it on one line.
$$ \Sigma_{00}\cap S_{\pm}=\mathbb{C}\bigl(\vec e_x\mp i\vec e_y\bigr), \qquad \dim_{\mathbb{R}}=2 . $$
So the two ideals contribute two lines and the four real dimensions are reached by the field chart alone. The readings separate by sector, and that separation is the reason the entry has three readings rather than one.
| reading | where it lives | share of the stratum |
|---|---|---|
| the vacuum | $\mathbb{B}_{\Phi}$, and the centre | the apex, real dimension $0$ |
| the ladder | the two minimal left ideals | two complex lines, real dimension $2$ each |
| the null field | neither elementary sector | the whole stratum, real dimension $4$ |
The third row is the only row that reaches the stratum completely. The field strength is the one object of the corpus that lies in the stratum by construction and fills it, and that is structural and not a choice of vocabulary: the field is the only one of the three that is neither Hermitian nor anti-Hermitian, and therefore the only one whose vector part is free to be isotropic.
The apex is the only element of the stratum that a four-momentum chart reaches. That is the exact content of the word vacuum here: the zero four-momentum, the element that carries neither mass nor motion, and the single point at which the material chart touches the stratum.
The Scalar Line
The scalar line is the one line of the algebra on which the vector square vanishes identically.
$$ \mathbb{C}e_t=\{q\,e_t\},\qquad B^{2}(q\,e_t)=0\ \text{for every }q,\qquad N(q\,e_t)=q^{2}. $$
So the scalar line is entirely inside the spectral stratum and meets this stratum at one point. Every scalar has $B^{2}=0$, and the scalar is null only at $q=0$. The line is therefore $\Sigma_{01}\cap\mathbb{B}_{C}=\mathbb{C}e_t$ as a set with its origin attached here: the two degenerate strata share exactly the zero element.
The identity is not the vacuum, and the scalar line is where the confusion is possible. Along the line the two conditions separate cleanly: the vector square vanishes everywhere and the norm vanishes at one point. The identity $e_t$ sits at $q=1$ with norm one and is invertible; the origin sits at $q=0$ with norm zero and is not. The line contains both, and the stratum's material share is the second and not the first.
The reading is that a four-momentum chart's own scalar axis is the boundary of the stratum and not part of it. The axis is where the vector square cannot fail; the stratum is where it does. A chart that followed the scalar axis to find the vacuum would find $e_t$ on the way, and $e_t$ is invertible in the algebra and not an element of the stratum.
The boundary. The line is a line of scalars and not a physical trajectory. Its points have no direction and no mode, and the reading is a statement about where the two strata meet.
The Spin
An element has no spin; a carrier does. The question is well posed only after an action is named, and here, as on the massless stratum, it has three answers.
Under the adjoint action the algebra is a singlet and a triplet, and the stratum removes the singlet.
$$ \mathbb{B}=Z\oplus V,\qquad Z=\mathbb{C}e_t\ \text{of spin }0,\qquad V=\operatorname{span}_{\mathbb{C}}\{\vec e_x,\vec e_y,\vec e_z\}\ \text{of spin }1, $$
$$ Q_t=0\ \Longrightarrow\ \Sigma_{00}\subset V . $$
So the stratum is a spin-one object and the condition $Q_t=0$ is the removal of the singlet. Every element of the stratum lies in the triplet, and the triplet is the carrier of which a nilpotent is a vector.
Under left multiplication the algebra is two spin-half amplitudes.
$$ \mathbb{B}=S_{+}\oplus S_{-},\qquad \text{weights }+\tfrac12\ \text{and}\ -\tfrac12,\ \text{each twice}, $$
and a nilpotent is a weight vector only in a special case. The condition $i\vec e_z\tilde Q=\tilde Q$ holds exactly for $\tilde Q=\vec a+i\vec e_z\vec a$ with $\vec a\perp\vec e_z$, a two-real-parameter family. A general nilpotent is a sum of both weights, as $\vec e_x+i\vec e_y$ is, and its weight string is not a spin label.
The rotation group acts transitively on the non-zero nilpotents. The stratum is a single orbit once the complex scale is allowed, so no rotation-invariant label separates two of its points: the invariant data are nullity and the scale, and the scale is free because the stratum is a cone.
What the stratum fixes is the carrier's spin and never a value inside the carrier. In the letters of the field: a null field is a spin-one field, its two circular polarisations are both null, and the stratum holds both because $\tilde F^{2}=0$ is blind to the dual. The sector fixes nullity and not helicity, and that boundary is stated again in §The Limits.
The Interpretations
I.10.as — Everything Fails at Once
Read: the three descriptions, the vanishing of $N$ and $B^{2}$ together, and the equivalence $\Sigma_{00}=\{\tilde Q^{2}=0\}$.
This is the only stratum on which the two independent questions of the series get the same answer. Invertibility fails and spectral separation fails, and they fail together because both are consequences of $\tilde Q^{2}=0$. One predicate does not cause the other; a common cause produces both, and that cause is the vanishing of the square.
The reading is that there is such a thing as a total failure, and that it is a stratum and not a defect of the analysis. A reader who meets the sector after the three others is prepared to see one description fail; this is the place where all of them fail, and they fail for a single reason that can be written in one line.
The comparison with the two mixed strata is what makes the point. There the failures were opposite and independent: the eigenbasis without the modulus, and the modulus without the eigenbasis. Here they coincide, and the coincidence is the definition of the strata's arrangement: the two walls cross nowhere, they touch along this one locus, and the touch is where both conditions hold.
The boundary. The statement is about the two predicates and not about physics. A physical object whose two descriptions fail at once is not thereby simpler or more fundamental; the algebra says only that the descriptions fail together.
I.10.at — The Vacuum Is the Additive Zero
Read: the origin as the unique semisimple element of the stratum, with minimal polynomial $X$, both multiplications zero, and every idempotent commuting with it.
The zero element is not the identity, and the difference is the whole content of the word. The identity $e_t$ is invertible with norm one, vector square zero and spectrum $\{1\}$; it lives on the scalar line of the other degenerate stratum. The origin has norm zero, rank zero and spectrum $\{0\}$. Calling the origin the vacuum and calling $e_t$ the vacuum are two different statements, and only the first is on the stratum.
The reading is that the vacuum is the element with no scale, no mode and no direction. It is the additive zero of the chart, so it is the value of a field that vanishes, and it is the only element of the stratum that a four-momentum can take.
The weakness of the reading is worth stating plainly. Nothing in the algebra distinguishes the additive zero of a momentum chart from the additive zero of any other chart, so vacuum names a role that the zero element plays in a chart and not a structure that the algebra supplies. The stratum's real content is the nilpotent cone, and the apex is a point of it added for completeness.
The boundary. The word is a reading of the zero element inside a chart. It does not name a state, an energy, a density or a stability, and no theorem here concerns the stability of anything.
I.10.au — A Square Zero Is Exclusion
Read: $\tilde Q^{2}=0$ on the whole stratum, and $\tilde Q^{n}=0$ for every $n\ge2$.
Applying the element twice gives zero, and that is the algebraic skeleton of a rule that a mode cannot be occupied twice.
$$ \tilde Q^{2}=0,\qquad \tilde Q^{n}=0\ \ (n\ge2). $$
The reading is one sentence long and it is worth keeping separate from the derivations. A single nilpotent is a direction whose second application is empty; a pair of them, with the anticommutator $e_t$, is a two-rung ladder; and the ladder is the structure in which a mode is either empty or filled and never doubly filled. The square zero is the empty second application.
What the reading does not do is derive anything. The Pauli principle is a statement about the antisymmetry of a many-particle state, and the stratum supplies neither a many-particle space nor a sign. What it supplies is the algebraic shape that a two-rung ladder has, and the shape is the reason the ladder can be read as a mode at all.
The contrast with the idempotent is exact and it is the sharpest pair in the section. An idempotent has $\tilde\Pi^{2}=\tilde\Pi$: applying it twice changes nothing, and it is a state. A nilpotent has $\tilde Q^{2}=0$: applying it twice gives nothing, and it is a transition. The two classes meet only at zero, and the ladder is the product that passes from one class to the other.
The boundary. Exclusion is a reading of the square zero. Statistics, anticommutation relations and the spin-statistics theorem are not in the stratum, and §The Limits repeats the point.
I.10.av — The Image Is the Kernel
Read: $L_{\tilde Q}^{2}=0$ with rank two, so $\operatorname{im}L_{\tilde Q}=\ker L_{\tilde Q}$.
Nothing that enters comes back, and nothing that is produced was not already there. The image is contained in the kernel because the operator is square zero, and the two have the same dimension, so they are the same plane. A map with that property is the algebraic form of a one-way coupling: it feeds a plane into itself and it produces no output outside the plane it already occupies.
The reading is that the stratum carries no oscillation. An oscillation needs an operator whose square is minus a multiple of the identity, or at least a spectrum with two different values; here the square is zero and the spectrum is one point, so there is no frequency, no period and no return.
The comparison with the light ray makes the point quantitative. There the operator satisfied $L^{2}=\kappa L$: it was a projection up to a scale, its image and kernel were complementary, and a live plane survived beside the annihilated one. Here the scale is zero and the two planes have merged. The single number $\kappa$ separates a projection from a shift.
The boundary. The statement is about the linear map on the algebra. It is not a statement about a scattering process, a decay or an irreversibility, and no arrow of time is supplied.
I.10.aw — The Flow Is a Shear, Not a Rotation
Read: $e^{\theta\tilde Q}=e_t+\theta\tilde Q$, with every eigenvalue $1$, norm $1$, and $\langle e^{\theta\tilde Q}\rangle_{*}=1+\theta^{2}\lVert\vec\Upsilon\rVert^{2}$.
The one-parameter group generated by a nilpotent is unipotent. It fixes every direction of its own plane, shifts the rest by a multiple of the parameter, and has no scale and no angle. The Hermitian norm grows as a quadratic polynomial, so the group is invertible at every parameter and is never a bounded one.
The reading is that the stratum supplies a translation and not a rotation. A rotation has a period and a bounded norm; a boost has an exponential norm; a shear has a polynomial norm and no period. The three behaviours are three different words, and the stratum settles which one it is by a single computation.
The separation of the two notions of unit is the sharpest statement in the section. The quaternionic norm of the flow is exactly one at every parameter, and its Hermitian norm is unbounded. A reader who uses one word unitary for both will conclude that the stratum is unitary; a reader who keeps them apart will see that it is invertible in the multiplicative group and not in the Hermitian one.
The shear preserves the norm, so it moves the stratum within itself. Conjugation by $e^{\theta\tilde Q}$ leaves $N$ of every element unchanged and maps nilpotents to nilpotents. It is a symmetry of the cone and not a motion off it, and it is the reason the stratum's own geometry is thin: the group that preserves it acts on it transitively up to scale.
The boundary. A shear on the algebra is not a physical transformation of a system. The word parabolic names a conjugacy type and not a process, and nothing here is a statement about a physical flow.
I.10.ax — The Null Field Is the Object of the Stratum
Read: $\tilde F=i\sqrt{\epsilon}\vec E-\sqrt{\mu}\vec H$, $\tilde F^{2}=\epsilon(I_1+2icI_2)e_t$, and the equivalence $\tilde F^{2}=0\iff\lvert\vec E\rvert=c\lvert\vec B\rvert$, $\vec E\cdot\vec B=0$.
The field strength is the one object of the corpus that the stratum names directly. It is pure, so the test applies to it without a projection; its square is a single central number; and the vanishing of that number is the classical nullness condition on a real field, in both directions.
The reading is exact and it is a dictionary, not an analogy. The six real components of the pair and the six real components of the pure element are the same six numbers, the two conditions are the same two conditions, and the map between them is invertible. A real null field and a point of the stratum are the same object in two alphabets.
The count is the stratum's dimension and it is also the field's. Amplitude, direction and one polarisation angle: one, two and one real data, four in total, which is the real dimension of the cone. Nothing is left over on either side, and that agreement is the reason the reading is not an analogy.
The reading also explains why the stratum's physical occupant is a field and not a particle. No four-momentum is in the stratum, so no particle's momentum chart reaches it; the field strength is the only construction of the corpus that lands in the stratum by its own definition. The stratum is therefore the algebra's account of a null field, and the other two readings are its two limiting contacts with the rest of the theory.
The boundary. A null field is a value of the field at a point. Whether a field is null everywhere, what congruence its principal direction follows, and whether it satisfies a field equation are further questions that the stratum does not answer.
I.10.ay — The Phase of the Scale Is a Duality Rotation
Read: $\tilde F=\lambda\tilde F_0$ with $\lvert\lambda\rvert$ the amplitude, and $e^{i\varphi}\tilde F$ acting as $\vec E\mapsto\vec E\cos\varphi-c\vec B\sin\varphi$, $\vec B\mapsto\vec B\cos\varphi+(\vec E/c)\sin\varphi$.
The two real dimensions of the fibre are two classical operations. The modulus scales both halves of the field and the argument rotates one into the other. Neither is a phase in the quantum sense, and neither is a gauge: they are the two things one may do to a null field and stay on the stratum.
The reading gives the bundle's fibre a physical content and closes the entry's count. Two real dimensions of base, two of fibre, four in total, and every one of the four is a named physical datum: direction, polarisation, amplitude, duality angle. Nothing in the description of a null field is left over, and nothing is double-counted.
The duality rotation is a symmetry and not a change of element. The transformation preserves both nullness conditions, so it stays on the stratum; and it changes the element, so it is not a redundancy. That combination is exactly what a fibre coordinate is: a parameter that moves within a set of distinct elements which the stratum treats alike.
The boundary. The duality angle is a rotation of the field pair and not a phase of a wave function. Nothing here is a $U(1)$ gauge freedom, and no charge is associated with it.
I.10.az — Nullness Is Not Radiation
Read: the stratum's condition $\tilde F^{2}=0$ and nothing else about the field.
A static field can satisfy it. The condition is algebraic and pointwise: two equations on the six components of the pair at one point. It says nothing about time, nothing about propagation and nothing about a source, so a configuration that satisfies it and does not vary in time is a null field that radiates nothing.
The reading is a warning about a natural over-reading. Because the null field is the classical field of radiation, it is tempting to read the stratum as the algebra's account of radiation. The stratum supplies the algebraic form of a null field and supplies no dynamics, so the identification stops at the value and does not extend to the process.
What is missing is exactly what a wave carries beyond its value at a point. A potential from which the field descends, a current to which it couples, a congruence of principal directions along which the field is transported, and an equation that makes the configuration a solution: all four are extra data, and the corpus treats each of them in its own article.
The boundary. The sector is not a theory of radiation, not an equation of motion and not a statement about energy transport. It is the locus on which the value of a field is null.
I.10.ba — The Ladder Needs a Pair
Read: $\{\vec\Upsilon_x,\vec\Upsilon_y\}=e_t$ with both nilpotent, and $\tilde\Pi_{\pm}=\vec\Upsilon_x\vec\Upsilon_y$, $\vec\Upsilon_y\vec\Upsilon_x$ the two rungs.
One nilpotent is a direction and two are a system. The single element is square zero, has no state attached to it, and its operator's image is its kernel. The pair anticommutes to the identity, and the two orders of the product are the two orthogonal Hermitian rungs. Nothing is added to the algebra between the one and the two: the pair is a choice of two directions inside the same stratum.
The reading is that a fermionic mode is a pair of nilpotents and not one. The reason is structural and it is visible in the operator: a single square-zero element annihilates everything it produces, so a state cannot be read off it. The product of two of them lands in a Hermitian idempotent, and a Hermitian idempotent is what a state is read on.
The two rungs are the two orders and the multiplication is not commutative, so the pair is ordered. Interchanging the two nilpotents interchanges the rungs, and the interchange is the only freedom the construction has. That is the whole of the two-valued structure: two directions, two orders, two rungs.
The boundary. The pair is a pair of elements and not a pair of particles. Fermionic is a reading of the anticommutator and of the two-rung product; the stratum supplies neither a field operator nor a statistics.
I.10.bb — Two Rungs and No Third
Read: $\operatorname{Sc}\tilde\Pi_{\pm}=\tfrac12$ for every primitive idempotent, and $\tilde\Pi_{+}+\tilde\Pi_{-}=e_t$.
The trace caps the ladder at two. Every primitive idempotent of the algebra has scalar part one half, and a family of orthogonal ones summing to the identity must therefore have exactly two members. The cap is the dimension of the module, and the module here has dimension two.
The reading is that the algebra's fermionic content is a single mode and not a field. A two-level system, two raising and lowering operators, and two rungs summing to the identity: that is one mode, and a fermionic field would need a family of modes indexed by something the stratum does not have. The distinction matters because the algebra is a matrix algebra of fixed size, and its size is what fixes the cap.
The cap is also why the ladder is a pair of directions and not a chain. A longer ladder would need a larger module, and a larger module would need a larger algebra. Nothing in this algebra can be extended to three rungs, and no limit of the stratum produces a third.
The boundary. The count is a statement about idempotents in this algebra. A physical fermion has infinitely many modes, and the stratum's two rungs are the algebraic form of one mode and not a truncation of a field.
I.10.bc — The Stratum Carries Spin One and No Weight
Read: $\Sigma_{00}\subset V$ under the adjoint action, the two spin-half amplitudes under left multiplication, and transitivity of the rotation group on the non-zero nilpotents.
The carrier's spin is one and the element's weight is not defined. The removal of the singlet is the condition $Q_t=0$, so every element of the stratum lies in the triplet. Inside the triplet no rotation-invariant label separates two points, so the stratum fixes the carrier and no value in it.
The reading is that the stratum can say what kind of object a nilpotent is and cannot say which state of that object it is. A spin-one carrier with no projection selected is a vector with no axis; the stratum's elements are such vectors, and the axis is supplied by a frame and not by the element.
In the letters of the field the same statement is about helicity. A null field is a spin-one field, its two circular polarisations are both null, and the stratum holds both because its condition is blind to the dual. So the sector fixes nullity and not helicity, and the helicity is a datum of the module and not of the element.
The boundary. The reading is a statement about carriers. A null field is spin one is a statement about the carrier it belongs to, not a measurement on an element, and no polarisation is selected by the stratum.
I.10.bd — The Apex Is the Whole Material Share
Read: $\Sigma_{00}\cap\mathbb{B}_{\Phi}=\{0\}$ and $\Sigma_{00}\cap\mathbb{B}_{\Psi}=\{0\}$, with $B^{2}$ definite on both.
The correction the entry exists to make is this one line. The archive article's title pairs the vacuum with the ladder, and its §The Material Case already says that the region's material content is the apex. What the pairing obscures is that the ladder is not a material object at all: no four-momentum is a nilpotent, and the stratum's real four dimensions are reached by the field chart and by nothing else.
The reading is that the section's four strata do not all meet the material chart in the same way. The open stratum is the material bulk, the rest frame is a material wall, the light ray is a material cone, and this stratum is a single material point. The point is where all four touch, and it is the only material object the stratum contains.
The correction does not weaken the fermionic reading; it relocates it. The ladder is a statement about the algebra's carriers and about the field that fills the stratum, and it is not a statement about a momentum chart. Once it is relocated, the reading is stronger: it no longer has to compete with the kinematics of a particle, and the two can be stated separately.
The boundary. Not material is a statement about the four-momentum chart. It is not a statement that the stratum is unphysical, and the field reading of §The Null Field is the counterexample.
The Limits
The stratum is a locus and the readings are readings of a locus. The whole mathematical content is two equations on four complex numbers. Everything the entry names — vacuum, null field, ladder, mode — is a name for what those equations mean when the chart is supplied with a physical vocabulary, and the names carry no content beyond the equations and the chart.
No dynamics is supplied. The stratum says which elements are square zero and says nothing about how an element changes. A field equation, a propagation law, a Lagrangian and a variational principle are all extra structure, and the corpus treats each of them separately. A null field that satisfies the stratum's condition and does not satisfy any field equation is a legitimate point of the stratum.
No statistics is supplied. The two-rung ladder is the algebraic form of a mode with two values, and the reading of the ladder as a fermionic mode is a reading. The stratum has no many-particle space, no sign under interchange and no anticommutation relation among field operators. The word exclusion is a name for the square zero and not a theorem about particles.
No arrow is supplied. The image-equals-kernel property is a statement about one linear map. It is not a decay, not a dissipation and not a direction of time, and the shear generated by the element is reversible at every parameter because $e_t-\theta\tilde Q$ inverts $e_t+\theta\tilde Q$ exactly.
No radiation is supplied. The stratum's condition is algebraic and pointwise, so it constrains the value of a field at one point and not a process. Null everywhere, null at one point, and null along a congruence are three different conditions, and only the first of the three is a statement one would make about a wave.
No topology is supplied beyond the base. The stratum has a sphere for a base and a contractible fibre, and that is the whole of its topology. There is no winding number, no instanton charge and no topological invariant attached to an element, and the single orbit under the rotation group is a statement about transitivity and not a classification.
No scale is fixed. The stratum is a cone, so every non-zero element has a one-parameter family of rescalings that stay on it. The amplitude of a null field is therefore free, and nothing in the stratum selects one.
The four strata are strata of one algebra and not four algebras. The partition is by two predicates, the predicates are functions of the element, and the entry's readings are attached to one cell of that partition. A statement that holds here and not on another stratum is a statement about the cell and is not a contradiction anywhere.
Summary
| result | content |
|---|---|
| the three descriptions | $\Sigma_{00}=\mathcal N\cap\mathcal S=\{Q_t=0,\vec Q^{2}=0\}=\{\tilde Q^{2}=0\}$ |
| dimension and singularity | complex dimension $2$, real dimension $4$; smooth off the apex, Zariski tangent space of dimension $3$ at the apex |
| the base and the fibre | a smooth conic, hence $\mathbb{P}^{1}$; fibre $\mathbb{C}^{\times}$; parametrisation $r(\vec u+i\vec u_\perp)$ |
| the two cones | tangent along the whole stratum, $\vec\nabla_\Phi N=\vec\nabla_\Phi B^{2}$; ideal generated by two squares |
| the element | a pure isotropic vector; spectrum $\{0\}$; minimal polynomial $X^{2}$; every power from the second vanishes |
| the centraliser | the dual numbers $\mathbb{C}[\varepsilon]/(\varepsilon^{2})$; no idempotent beyond $0$ and $e_t$ |
| the Peirce data | adapted idempotent $e$, the element in the mixed corner, block $J_2(0)$; no Peirce data relative to the element |
| the eigenbasis | both normalisations undefined, $Q_t=0$ and $B=0$; no eigenbasis |
| the flag | $0\subset\mathbb{C}\vec\Upsilon\subset\ker L_{\tilde Q}\subset\mathbb{B}$, of length two |
| the operator | $L^{2}=R^{2}=0$, rank $2$, two blocks $J_2(0)$, trace and determinant zero |
| image and kernel | $\operatorname{im}L=\ker L=\tilde Q\mathbb{B}$, a minimal right ideal; the plane killed is the plane reached |
| the two annihilators | $\ker L\cap\ker R=\mathbb{C}\tilde Q$; exchanged by the conjugation |
| the exponential | $e^{\theta\tilde Q}=e_t+\theta\tilde Q$ exactly; unipotent; determinant $1$; trace $4$ |
| the Hermitian norm | $1+\theta^{2}\lVert\vec\Upsilon\rVert^{2}$; an invertible element at every parameter and never a bounded one |
| no logarithm | $N(e^{X})=e^{2X_t}\neq0$; the stratum is disjoint from the image of the exponential |
| the products | $\operatorname{Sc}(\vec\Upsilon_2\vec\Upsilon)=-\vec\Upsilon_2\cdot\vec\Upsilon$, $B^{2}(\vec\Upsilon_2\vec\Upsilon)=-(\vec\Upsilon_2\cdot\vec\Upsilon)^{2}$; the product lies in $\{0\}\cup\Sigma_{01}$ |
| the null field | $\tilde F^{2}=\epsilon(I_1+2icI_2)e_t$; $\tilde F^{2}=0$ exactly when $\lvert\vec E\rvert=c\lvert\vec B\rvert$ and $\vec E\cdot\vec B=0$ |
| the count | six real data minus two conditions; amplitude, direction, polarisation |
| the scale | modulus is the amplitude; argument is the duality angle $\vec E\mapsto\vec E\cos\varphi-c\vec B\sin\varphi$ |
| the ladder | $\vec\Upsilon_x,\vec\Upsilon_y$ nilpotent with $\{\vec\Upsilon_x,\vec\Upsilon_y\}=e_t$; rungs $\tilde\Pi_{\pm}$ in $\Sigma_{01}$ |
| the rung count | $\operatorname{Sc}\tilde\Pi_{\pm}=\tfrac12$; at most two orthogonal idempotents sum to $e_t$ |
| the material share | $\Sigma_{00}\cap\mathbb{B}_{\Phi}=\{0\}$; the apex and nothing else |
| the informational share | $\Sigma_{00}\cap\mathbb{B}_{\Psi}=\{0\}$ |
| the ideals | $\Sigma_{00}\cap S_{\pm}=\mathbb{C}(\vec e_x\mp i\vec e_y)$, a line in each |
| the scalar line | $B^{2}\equiv0$ on $\mathbb{C}e_t$; the two degenerate strata share the apex only |
| the spin | $\Sigma_{00}\subset V$; carrier spin one; no weight, no helicity |
Summary of Notation
| symbol | meaning |
|---|---|
| $\tilde Q$, $\vec Q$, $Q_t$ | a biquaternion and its vector and scalar parts |
| $N$, $B^{2}$ | the quaternionic norm $Q_t^{2}+B^{2}$ and the vector square $Q_x^{2}+Q_y^{2}+Q_z^{2}$ |
| $\Sigma_{00}$, $\Sigma_{01}$, $\Sigma_{10}$, $\Sigma_{11}$ | the four strata, by the truth values of $N=0$ and $B^{2}=0$ |
| $\mathcal N$, $\mathcal S$ | the null cone and the spectral-degeneracy cone |
| $\vec\Upsilon$ | a non-zero nilpotent of the stratum |
| $\vec\Upsilon_x$, $\vec\Upsilon_y$ | the canonical ladder pair, with the anticommutator $e_t$ |
| $\tilde\Pi_{\pm}$ | the two Hermitian rungs $\tfrac12(e_t\pm i\vec e_z)$ |
| $L_{\tilde Q}$, $R_{\tilde Q}$ | left and right multiplication by $\tilde Q$ |
| $\natural$ | the conjugation $\tilde Q\mapsto\tilde Q^{*}$ followed by the negation, an algebra anti-automorphism |
| $\langle\cdot\rangle_{*}$ | the Hermitian norm $\operatorname{Sc}(XX^{*})$ |
| $e_t$ | the identity of the algebra |
| $\tilde F$, $\vec E$, $\vec H$, $\vec B$ | the field strength, the electric field, the magnetic excitation and the magnetic field, with $\vec B=\mu\vec H$ |
| $I_1$, $I_2$ | the two classical invariants $\vec E^{2}-c^{2}\vec B^{2}$ and $\vec E\cdot\vec B$ |
| $\lambda$ | the complex scale of a null field, modulus the amplitude and argument the duality angle |
Further Reading
- The Massless Regime, for the stratum on which the eigenbasis exists and the modulus does not, and for the rungs whose operators are here.
- The Rest Frame and the Exceptional Point, for the stratum on which the modulus exists and the eigenbasis does not.
- Mass in Motion: The Region of Both Invariants, for the open stratum and the two-mode picture.
- The Partition: Two Invariants and Four Regimes, for the arrangement of the four strata and the statement that the two conditions are independent.
- Rotor, Spin, Pair and Carrier: A Survey of the Closure Block, for the reach of each carrier on each stratum.
- The Group of Invertibles, the Cone and the Orders, for the cone whose apex is this stratum's material share.
- Norm and Pairing Conventions, for the letters, the inertias and the two norms used throughout.