Congruence Twistors and the Fields They Generate

Introduction

The corpus holds the three pieces of one construction, in three different articles, and never joins them. Twistor Theory and Biquaternions states the incidence relation $\omega=i\,x\pi$ and the null quadric. Null Electromagnetic Fields and the Shear-Free Congruence states that a null field's principal direction is geodesic and shear-free. The Algebrodynamical Programme states that the equation of the shear-free congruence is solved by an arbitrary holomorphic function $\Pi$ of the twistor arguments, and that Maxwell, Yang–Mills and the effective geometry all descend from that one function.

This article joins them. The join is a single geometric statement and a single algebraic statement. Geometrically, a congruence is a two-surface in projective twistor space, a spacetime point is a line in that space, and the congruence's spinor at the point is the intersection of the two. Algebraically, every function of the twistor components satisfies one pair of first-order relations, and that pair is at once the complex eikonal and the constraint that the programme derives for its gauge parameter, with the wave equation following as well for the functions of $G$.

The source is Kassandrov and Rizcallah, Maxwell, Yang–Mills, Weyl and eikonal fields defined by any null shear-free congruence (arXiv:1612.06718, 2016). The article records the twistor side of that paper. The field algebra it belongs to — the nonlinear Cauchy–Riemann conditions, the generating system, the weak self-duality, the charge quantisation, the particle reading — is recorded once in The Algebrodynamical Programme and is not restated here; where the two overlap, the reader is sent there.

The overlap is real and is stated at the outset so that the division is visible. The programme article already records the two first-order relations for the principal ratio, and already draws from them the complex eikonal, the wave equation and the harmonicity of $\ln G$, with a recomputation on one explicit solution of the quadratic Kerr equation. It records the electromagnetic field as the second derivatives of $\ln G$ in the source's component assignment, and the representation of that field through the derivatives of $\Pi$ alone. It records the coincidence of the singular loci. What this article adds is seven things and no more.

  • The relations are shown to hold for an arbitrary function of all three twistor components and not only of $G$, which is what makes the rank statement work and, with one further reading of the parameter, the gauge group close.
  • That pair is identified with the constraint the source imposes on its rescaling parameter, so that the gauge condition and the eikonal are one equation.
  • The reach of the pair is mapped: the eikonal holds for every twistor function, the wave equation only for the functions of $G$, and the difference is computed rather than asserted, then reduced to a single Jacobian in the two twistor components.
  • The gauge group is described as a group acting on each point's twistor, with the projective twistor space as its orbit set — a proper subgroup of the naive one, since the parameter must be a function of the twistor and must be read on the projective twistor for the inverse to exist.
  • The multivaluedness of the congruence solution is identified as Bézout's theorem: the branch count at a point is the degree of the twistor surface, so the programme's many-valued $G$ is the intersection number of a line with a surface and not a defect of the ansatz.
  • The paper's two other field statements are recorded: the spinor inside the potential, in the form the paper derives it from its (13) and (67), which the corpus does not otherwise have, and the closed-form Yang–Mills triplet $\{-GF,iGF,-F\}$, whose modulus relation the programme article records but whose components in this form it does not — those components being what identifies the $\tfrac{1}{16}$ that separates the corpus's normalisation from the source's as the square of the trace coefficient, and so closes a factor the programme article leaves open.
  • The static generator is exhibited as a twistor function, by the identity $\Pi=G\kappa_0-\kappa_1+2ia\,G$, the member of the source's (70) family that its own quoted root requires — the display at (71) carries $2ia$ as a constant instead, and the difference, with the residuals that settle it, is recorded below — so that the programme's ring sits inside the surface picture and not beside it, and its two branches are the degree-two count in one line.

The recomputations are these, and each is reported where it is used. The null-twistor identity for a Hermitian coordinate matrix, on a hundred random Hermitian matrices. The pair of first-order relations and the eikonal, symbolically, for an arbitrary generating function, with vanishing residual — the programme article's check was numeric and on one quadratic generator, so the symbolic form is a strengthening rather than a repetition. The rank of the twistor field's Jacobian, on a hundred random complex points for each of three generating functions — quadratic, cubic and transcendental — checking that every three-by-three minor vanishes and that some two-by-two minor does not. The triplet's modulus identity, on a hundred random complex pairs. The closed form of the obstruction, on ten test functions of the twistor and on two generating functions, one quadratic and one cubic, to thirty digits. The transcription of the field formula from $\Pi$ — its normalisation and its index placement — against the directly computed components of (64) on the static generator, to thirty digits. And the tangency condition of the static Kerr generator, which reduces to $r_*^2=0$ exactly — together with the two residuals that decide which polynomial the source's quoted root belongs to, a discrepancy recorded below and not repaired in silence.

Two claims are quoted and not recomputed, because they belong to the programme's own machinery rather than to the twistor geometry: that the field built from the second derivatives of $\ln G$ satisfies the source-free Maxwell equations, and the explicit representation of that field through the derivatives of $\Pi$ alone. Both are recorded in The Algebrodynamical Programme. Of the first, only the harmonicity of $\ln G$ that carries it was checked here, and the Maxwell equations themselves are quoted and not re-derived, because the field-theoretic content is not what this article is about. Of the second, the field-theoretic derivation is likewise quoted — but the two choices the corpus made in transcribing the display, the normalisation of the second term and the placement of the indices, are checked against the source's own (64) from the static generator, where the formula collapses to a single term and the check is sharp; that is a check of the transcription and not of the derivation.

The conventions are those of the read-list articles and of the companion mathematics article The Two-Sided Derivative over the Biquaternion Algebra, which owns the coordinate chart. The complexified Minkowski coordinates are the four entries of the matrix

$$ Z=\begin{pmatrix}\upsilon & w\ p & v\end{pmatrix}, $$

with $\upsilon,v,w,p$ the complex coordinates and $z_0=\tfrac12(\upsilon+v)$, $z_3=\tfrac12(\upsilon-v)$, $z_1=\tfrac12(w+p)$, $z_2=\tfrac{i}{2}(w-p)$; on the real slice $\upsilon=r'_t+r_z$, $v=r'_t-r_z$, $w=x-iy$, $p=x+iy=\bar w$. The derivative with respect to a slot of $Z$ is the derivative with respect to its entry, $\partial/\partial Z_{00'}=\partial/\partial\upsilon$, $\partial/\partial Z_{01'}=\partial/\partial w$, $\partial/\partial Z_{10'}=\partial/\partial p$, $\partial/\partial Z_{11'}=\partial/\partial v$, and that is the only derivative the twistor construction uses. The wave operator is $4\left(\partial^2/\partial\upsilon\,\partial v-\partial^2/\partial w\,\partial p\right)$, written out wherever it acts and never abbreviated, and the complex eikonal is

$$ \frac{\partial\Sigma}{\partial\upsilon}\frac{\partial\Sigma}{\partial v}=\frac{\partial\Sigma}{\partial w}\frac{\partial\Sigma}{\partial p}. $$

The principal spinor is $\zeta=(\zeta_0,\zeta_1)$, its ratio is $G=\zeta_1/\zeta_0$, and the gauge $\zeta_0=1$ is used wherever a component is written out. Every partial derivative is written with the variable named after it, $\partial G/\partial\upsilon$, in the spelling the conventions article requires, and the two articles it leans on — the programme article and the mathematics article on the two-sided derivative — are written in the same spelling. The two twistor components attached to $G$ are

$$ \kappa_0=wG+\upsilon,\qquad \kappa_1=vG+p , $$

and the generating function is an arbitrary holomorphic $\Pi(G,\kappa_0,\kappa_1)$. Its total derivative along the line, $P=d\Pi/dG=\Pi_G+w\,\Pi_0+v\,\Pi_1$ with $\Pi_0,\Pi_1$ the derivatives in the second and third arguments — the source's counting, which is fixed once at (65) and used throughout, so that the symbol $\Pi_1$ means one thing in this article — is the object the source defines at (60), and the programme article calls it $P$ as well, so the symbol needs no translation; it is a total derivative and not the partial $\Pi_G$, and the two differ by $w\Pi_0+v\Pi_1$.

The Coordinate Matrix and Its Two Columns

A point of complexified Minkowski space is a two-by-two complex matrix $Z$, and the incidence relation

$$ \tau=Z\zeta,\qquad \tau_A=\sum_B Z_{AB}\,\zeta_B , $$

attaches to a spinor $\zeta$ a second spinor $\tau$. Read as a map, $\zeta\mapsto(\zeta,Z\zeta)$ is complex linear from $\mathbb{C}^2$ into $\mathbb{C}^4$, its image is two-dimensional, and the pair $(\zeta,\tau)$ is the twistor of the point. Twistor Theory and Biquaternions writes the same relation as $\omega=i\,x\pi$, with $x$ the Hermitian coordinate matrix and the $i$ the complex unit of twistor space; this article follows the source and writes $Z$, absorbing that $i$ into the coordinates, so that the Hermitian character of $Z$ on the real slice is simply the reality of $r'_t,r_x,r_y,r_z$. Two facts follow at once, and both are used below.

The image is a plane, so a spacetime point is a plane in twistor space. Conversely every two-plane of $\mathbb{C}^4$ that is the graph of a linear map is the plane of exactly one point: the four entries of $Z$ are read off the plane, and the map between points and planes is a bijection. Projectively the plane becomes a line in $\mathbb{PT}$, and the correspondence is the classical one — a point of spacetime is a line, a point of twistor space is a two-surface.

On the real slice the plane is null. A twistor is null when $\tau^{*}\zeta-\zeta^{*}\tau=0$. With $Z$ Hermitian this holds identically, because

$$ \tau^{*}\zeta=\zeta^{*}Z^{*}\zeta=\zeta^{*}Z\zeta=\zeta^{*}\tau , $$

the middle step being the Hermiticity. The identity was checked on a hundred random Hermitian matrices, the worst relative residual sitting at the double-precision floor, of order $10^{-15}$, as it must: the residual is a difference of two separately rounded products and nothing else. So the planes that represent real spacetime points lie inside the null quadric of twistor space, and the correspondence restricted to the real slice is a correspondence between real points and null lines. The null quadric of twistor space and the zero-divisor cone of the algebra are the same cone, which is the agreement recorded in Twistor Theory and Biquaternions; what the computation above adds is that the equivalence is not merely at the level of the defining equations but holds for the full incidence data.

The Twistor Surface

Fix a holomorphic function $\Pi$ of the three projective components. The equation

$$ \Pi(G,\kappa_0,\kappa_1)=0 $$

defines a two-complex-dimensional surface $S$ in $\mathbb{PT}$ — three projective coordinates, one equation — and the congruence's spinor at the point $x$ is the intersection of $S$ with the line $L_x$ attached to $x$. That single sentence is the geometric content of the whole construction, and everything else in this article is a consequence of it.

The intersection explains the multivaluedness. When $\Pi$ is a polynomial of total degree $d$ in its three arguments, its restriction to the line is a polynomial in $G$ of degree at most $d$ — exactly $d$ unless the top part vanishes on the line — so the line meets the surface in $d$ points, counted with multiplicity. (That hypothesis matters and is not always met: a transcendental $\Pi$ restricts to a transcendental equation with infinitely many roots in general, so there is no branch count and no algebraic surface at all. Bézout applies to the polynomial generators, which is the class the programme's explicit solutions belong to; the transcendental generator used in the rank check below is outside it, the rank being a local differential statement that needs no algebraic surface.) A quadratic generator gives two roots, a cubic three; the corpus's The Algebrodynamical Programme records the resulting multivaluedness and the programme's insistence that it is required rather than tolerated, together with the topological route to charge quantisation that it supplies. The addition here is the reason: the multivaluedness is Bézout's theorem, and the count of a line against a surface reduces in one step to the plane form that Algebraic Curves states: any plane through the line cuts the surface in a plane curve of the same degree $d$ — the line is not a component of it, since the line meets the surface in finitely many points — and the plane Bézout gives $d$ intersections of that curve with the line, the same $d$. It is not a defect of the eikonal equation and not a failure of the ansatz; it is the count of intersections of a line with a surface, and it disappears only when the surface is a plane.

The singular locus is a tangency. When the line is tangent to the surface the roots coincide, the implicit resolution of $\Pi=0$ for $G$ fails, and $P=d\Pi/dG$ vanishes beside $\Pi$. This is the condition the programme uses as its singular set, and in the twistor picture it is one thing only: the set of spacetime points whose line is tangent to the twistor surface. The point-like, string-like and surface-like singular sets of the programme's catalogue are then the various ways a two-surface can be touched by a family of lines.

The Pair of Relations

The algebraic content of the construction is one pair of equations, and it is worth isolating because two separate results of the source follow from it, with a third that follows only in part. The programme article records the pair for the principal ratio and draws the eikonal and the wave equation from it; what is done here is to derive the pair once for an arbitrary generating function, extend it to every twistor function, and identify it with the gauge constraint.

The relations. For the principal ratio $G$ obtained from any $\Pi$,

$$ \frac{\partial G}{\partial w}=G\,\frac{\partial G}{\partial\upsilon}, \qquad \frac{\partial G}{\partial v}=G\,\frac{\partial G}{\partial p}. $$

Where they come from, and why they hold for every generating function. Differentiating $\Pi(G,\kappa_0,\kappa_1)=0$ at fixed coordinates, and using $\partial\Pi/\partial\upsilon=\partial\Pi/\partial\kappa_0$, $\partial\Pi/\partial w=G\,\partial\Pi/\partial\kappa_0$, $\partial\Pi/\partial p=\partial\Pi/\partial\kappa_1$ and $\partial\Pi/\partial v=G\,\partial\Pi/\partial\kappa_1$, the implicit derivatives are

$$ \frac{\partial G}{\partial\upsilon}=-\frac{1}{P}\frac{\partial\Pi}{\partial\kappa_0},\qquad \frac{\partial G}{\partial w}=-\frac{G}{P}\frac{\partial\Pi}{\partial\kappa_0},\qquad \frac{\partial G}{\partial v}=-\frac{G}{P}\frac{\partial\Pi}{\partial\kappa_1},\qquad \frac{\partial G}{\partial p}=-\frac{1}{P}\frac{\partial\Pi}{\partial\kappa_1}, $$

and the two relations are read off by inspection. They are therefore not properties of a solution but properties of the ansatz: no generating function is special, and no solution has to be exhibited to know that they hold. Verified symbolically with the three partials of $\Pi$ kept as free symbols, the residuals are exactly zero, and they are zero for the same reason in every degree.

Every twistor function obeys them. Let $\lambda$ be any differentiable function of the three twistor components, $\lambda=\lambda(G,\kappa_0,\kappa_1)$. Since $\partial\kappa_0/\partial\upsilon=1+w\,\partial G/\partial\upsilon$, $\partial\kappa_0/\partial w=G+w\,\partial G/\partial w$, and similarly in the other column, the relations for $G$ give

$$ \frac{\partial\kappa_0}{\partial w}=G\,\frac{\partial\kappa_0}{\partial\upsilon}, \qquad \frac{\partial\kappa_1}{\partial w}=G\,\frac{\partial\kappa_1}{\partial\upsilon}, \qquad \frac{\partial\kappa_0}{\partial v}=G\,\frac{\partial\kappa_0}{\partial p}, \qquad \frac{\partial\kappa_1}{\partial v}=G\,\frac{\partial\kappa_1}{\partial p}, $$

and the chain rule carries the pair to $\lambda$ unchanged,

$$ \frac{\partial\lambda}{\partial w}=G\,\frac{\partial\lambda}{\partial\upsilon}, \qquad \frac{\partial\lambda}{\partial v}=G\,\frac{\partial\lambda}{\partial p}. $$

This was verified symbolically for an arbitrary $\lambda$, with its three partials free, and the residual is zero.

Two consequences, and they differ. Multiply the pair:

$$ \frac{\partial\lambda}{\partial\upsilon}\frac{\partial\lambda}{\partial v} =\frac{\partial\lambda}{\partial w}\frac{\partial\lambda}{\partial p}, $$

which is the complex eikonal, and it holds for every function of the twistor. Cross-differentiate the pair instead. Differentiating $\partial\lambda/\partial w=G\,\partial\lambda/\partial\upsilon$ in $v$ and $\partial\lambda/\partial v=G\,\partial\lambda/\partial p$ in $w$, and equating the two mixed derivatives, gives

$$ \frac{\partial G}{\partial v}\frac{\partial\lambda}{\partial\upsilon}+G\,\frac{\partial^2\lambda}{\partial v\,\partial\upsilon} =\frac{\partial G}{\partial w}\frac{\partial\lambda}{\partial p}+G\,\frac{\partial^2\lambda}{\partial w\,\partial p}, \qquad\text{hence}\qquad \frac{\partial^2\lambda}{\partial\upsilon\,\partial v}-\frac{\partial^2\lambda}{\partial w\,\partial p} =\frac{\partial G}{\partial\upsilon}\frac{\partial\lambda}{\partial p}-\frac{\partial G}{\partial p}\frac{\partial\lambda}{\partial\upsilon}, $$

on using $\partial G/\partial v=G\,\partial G/\partial p$ and $\partial G/\partial w=G\,\partial G/\partial\upsilon$ and dividing by $G$.

The wave equation is therefore not a consequence of the pair alone, and this is the one place where the article must part company with a natural guess. The bracket of the last display vanishes identically when $\lambda$ depends on $G$ alone, because then $\partial\lambda/\partial p=(d\lambda/dG)\,\partial G/\partial p$ and $\partial\lambda/\partial\upsilon=(d\lambda/dG)\,\partial G/\partial\upsilon$; it does not vanish for a general twistor function. At $\lambda=G$ the display reads $\partial^2G/\partial\upsilon\,\partial v-\partial^2G/\partial w\,\partial p=0$, which is the wave equation the source derives, and that is the sense in which the source's statement is about $G$. Checked on the quadratic generator at forty working digits, with $G$ taken from the closed-form root and the derivatives taken in closed form rather than by differences: the combination $\partial^2G/\partial\upsilon\,\partial v-\partial^2G/\partial w\,\partial p$ and the same combination of $\ln G$ vanish to the working precision, below $10^{-40}$, while those of $\kappa_0$ and $\kappa_1$ do not — at $\upsilon=1.13$, $v=0.71$, $w=0.43+0.29i$, $p=\bar w$ and $a=0.7$ their moduli, times the $4$ that stands on the display below, are $3.41$ and $1.35$ — and the identity above reproduces both values to forty digits. A finite-difference run of the same check agrees with those two values to twelve digits, which is the accuracy of the difference scheme and not of the identity.

The obstruction has a closed form, and it is a Jacobian. Substituting $\partial G/\partial\upsilon=-\Pi_0/P$ and $\partial G/\partial p=-\Pi_1/P$ into the bracket of the display turns it into

$$ 4\left(\frac{\partial^2\lambda}{\partial\upsilon\,\partial v}-\frac{\partial^2\lambda}{\partial w\,\partial p}\right) =-\frac{4}{P}\left(\Pi_0\,\frac{\partial\lambda}{\partial\kappa_1}-\Pi_1\,\frac{\partial\lambda}{\partial\kappa_0}\right), $$

the derivatives of $\lambda$ taken at fixed $G$; the bracket on the right is the Jacobian of the pair $(\Pi,\lambda)$ in the two twistor components. It was checked to thirty digits on ten test functions of the twistor — $\kappa_0$, $\kappa_1$, $\kappa_0^2$, $\kappa_0\kappa_1$, $\kappa_1^3$, $\kappa_0^3\kappa_1$, $G\kappa_0$, $e^{\kappa_0/2}$, $G$ and $\kappa_0+\kappa_1^2$ — on two generators, one quadratic and one cubic, with every residual at the working floor. It also reproduces the two numbers above, which are its cases $\lambda=\kappa_0$ and $\lambda=\kappa_1$: for the quadratic generator $\Pi_0=G$ and $\Pi_1=-1$, so the two moduli are $4/|P|=3.414$ and $4|G|/|P|=1.353$, as printed.

Three things follow, and the first concerns the caustic. For the twistor's own two components the bracket collapses to one derivative of $\Pi$, because each component is blind to one of the two arguments:

$$ 4\left(\frac{\partial^2\kappa_0}{\partial\upsilon\,\partial v}-\frac{\partial^2\kappa_0}{\partial w\,\partial p}\right)=\frac{4\Pi_1}{P}, \qquad 4\left(\frac{\partial^2\kappa_1}{\partial\upsilon\,\partial v}-\frac{\partial^2\kappa_1}{\partial w\,\partial p}\right)=-\frac{4\Pi_0}{P}, $$

so the combination for $\kappa_0$ vanishes identically exactly when $\Pi$ does not depend on $\kappa_1$, and symmetrically — a generator that ignores one argument leaves the component of the other argument harmonic, which is why the statement is made per generator and not in general. Where neither derivative vanishes, the failure carries the factor $1/P$: the same $P$ whose vanishing is the tangency condition. So the non-harmonicity of the twistor's components is not spread evenly over spacetime — finite away from the caustic, unbounded on it — and that is the sharpest form the link between the algebra and the singular locus takes in this article, stated about the components the construction writes down before it writes any field. The obstruction is also blind to $\lambda_G$, which does not appear in the closed form at all: only the two twistor dependencies matter, and a function of $G$ alone has none to contribute.

The second is a sharpening of the section's own claim. The bracket vanishes if and only if $\lambda$ is functionally dependent on $\Pi$ at fixed $G$ — that is, $\lambda=\lambda(G,\Pi)$ — and on the surface, where $\Pi=0$, such a $\lambda$ is a function of $G$ alone. So the wave equation holds for exactly those twistor functions whose restriction to the congruence is a function of $G$: the class the section named, now characterized rather than merely exhibited, with the converse supplied. The two descriptions agree on the congruence and differ off it, which is why nothing above changes; what the Jacobian adds is that harmonicity is a condition of dependence, not a differential equation to be solved.

The third is a reading and not a result. The bracket is the standard Jacobian of two functions on the $(\kappa_0,\kappa_1)$ plane, so $\Pi$ enters it as a Hamiltonian and the harmonic twistor functions are those that Poisson-commute with it. Nothing in the construction uses this, and no identification with any physical Hamiltonian is intended; it is recorded because it is the compact form of the second point — the obstruction is a bracket with $\Pi$, so the harmonic twistor functions are the ones $\Pi$ does not move.

So the pair gives two things of different reach. The eikonal holds for all twistor functions — every one of them is a null gradient — and the wave equation holds for the functions of $G$ alone, which is exactly the boundary the programme article draws when it says that every $C^2$ function of $G$ is harmonic on the solutions. The asymmetry is what the field construction needs and no more: the field is built from $\ln G$, a function of $G$, so harmonicity is available where it is used, and the gauge parameter, which is a general twistor function, is required only to be null.

The same pair is the gauge constraint. This is the observation that ties the algebra to the geometry. The programme requires its rescaling parameter to depend on the coordinates only through the twistor, and the differential form of that requirement is that $\alpha$ be annihilated by both ray derivations, $\partial/\partial w-G\,\partial/\partial\upsilon$ and $\partial/\partial v-G\,\partial/\partial p$. Those two span the tangent two-plane of the set of points that share the twistor of $x$ — the displacements with $\delta Z\,\zeta=0$, that is $\delta Z=\eta\otimes(G,-1)$ — so the condition is exactly that $\alpha$ be constant on that two-plane, which is what depending on the twistor means. In components it is

$$ -G\,\frac{\partial\alpha}{\partial\upsilon}+\frac{\partial\alpha}{\partial w}=0, \qquad -G\,\frac{\partial\alpha}{\partial p}+\frac{\partial\alpha}{\partial v}=0, $$

which is the pair again. The minus sign is not decoration: contracting the spinor into the gradient on a single index gives the plus-signed pair $\partial\alpha/\partial\upsilon+G\,\partial\alpha/\partial p$ and $\partial\alpha/\partial w+G\,\partial\alpha/\partial v$, and no function of the twistor satisfies those — checked on $\kappa_0$, where the first is $1$. The operators must be tangential to the two-plane, and only the minus signs make them so. The condition that a rescaling be a symmetry and the condition that a function be a solution of the complex eikonal are the same equation, and its solutions are exactly the functions of the twistor. Harmonicity, by the preceding section, is a further requirement, met when the twistor function depends on $G$ alone. This is why the gauge group of the next section is what it is, and it is the sharpest single statement in the paper.

Two Independent Components

A four-component twistor whose components are functions of four coordinates could depend on all four. It does not: exactly two of its components are functionally independent, and the count is what makes the twistor field a two-surface rather than a four-dimensional spread.

The statement is that the map $x\mapsto(G,\kappa_0,\kappa_1)$ has a Jacobian of rank exactly two, so that all three-by-three minors vanish and some two-by-two minor does not. This was checked numerically at forty working digits, on a hundred random complex points for each of three generating functions — a quadratic, a cubic, and a transcendental one — with the derivatives taken in closed form from the implicit expressions above. The largest three-by-three minor, scaled by the cube of the Jacobian's largest entry, came out of order $10^{-41}$ for all three families — at the forty-digit floor, and forty orders below the largest entry — while the smallest of the largest two-by-two minors, scaled by the square of the same entry, came out between a tenth and a half. The two figures are orders of magnitude and not digits: the run is seeded, and a rerun moves the last digits of both without moving either conclusion. The vanishing minors sit at the rounding floor of forty-digit arithmetic and the non-vanishing ones at $O(0.1)$, so the rank is two with forty orders of margin.

Physical interpretation. Rank two is the statement that the twistor field is a surface, and a surface is what the line has to meet. The count of components, the count of intersections, and the count of dimensions of $S$ are the same count seen three ways, and this is why the construction is rigid: the data are a surface, and a surface is fixed by one holomorphic function of three arguments, which is where the freedom sits and where it stops.

The Weakened Gauge Group

A rescaling of the spinor $\zeta\mapsto\alpha(x)\zeta$ is not a symmetry of the congruence equations unless $\alpha$ is restricted, and the restriction is the pair of the preceding section. The source's reading is that the symmetry group is therefore not the full group of complex functions but the group of functions of the twistor,

$$ \zeta\mapsto\alpha(T)\,\zeta,\qquad \tau\mapsto\alpha(T)\,\tau,\qquad T\mapsto\alpha(T)\,T . $$

It is a group, and the argument is the rank statement together with a reading of the parameter. Composition of two such rescalings is multiplication of two functions of $T$, hence of the same type, and associativity is inherited. The inverse needs the rank-two result, and it needs one thing more. On a fixed projective class the transformation is a map of the line, $T\mapsto\alpha(T)T$, whose natural inverse is $T'\mapsto T'/\alpha(T')$; composing the two returns $T$ exactly when $\alpha(\alpha(T)T)=\alpha(T)$, that is when $\alpha$ is invariant under the rescaling it defines. A parameter with explicit dependence on the twistor components fails this, and the failure is not exotic: $\alpha=1+\kappa_0$ obeys the pair, since every function of the twistor does, and it acts on the component it is built from as $\kappa_0\mapsto\alpha\kappa_0=\kappa_0+\kappa_0^2$ — a quadratic map, with two preimages $\tfrac12(-1\pm\sqrt{1+4\kappa_0'})$ for each image and hence no inverse, the two merging at the critical value $\kappa_0'=-\tfrac14$. So the rank-two count is necessary and not sufficient: the group is the group of projective twistor functions, the parameters that depend on the ratios and not on the components. On that reading the step is exact — $T$ has two independent components, so the image determines their ratios, and $\alpha^{-1}$ read on those ratios is again such a function. The source states the rank-two step and not the second, and the qualification is recorded here because the paragraph that follows uses it.

The orbits are the projective twistors, and this is the point. Since $\alpha$ is a function of $T$ and not of $x$, the ratio of any two twistor components is invariant, and two twistors related by a rescaling differ by nothing the geometry can see: the quotient by the group is the projective twistor space $\mathbb{PT}$. The corpus's twistor article writes the projective space as $\mathbb{PT}=\mathbb{CP}^3$, the complex lines of $\mathbb{C}^4$ — the orbit set of the full $\mathbb{C}^{\times}$ action on the four components. Here the projective structure is derived instead: it is the orbit set of the symmetry group that the congruence actually has, and the group is a proper subgroup of the naive one. A rescaling with a parameter that is not a twistor function is not a symmetry, and the source states this explicitly.

The potential shifts, and the field does not. The trace of the matrix one-form $Ξ\,dZ$, which is the potential one-form $\mathcal{A}:=2\,\mathrm{Sc}(Ξ\,dZ)$ (the source's $A$), transforms as

$$ \mathcal{A}\mapsto\mathcal{A}+d\ln\alpha , $$

which is the transformation of an electromagnetic potential under an abelian gauge change, and the two-form $F=d\mathcal{A}$ is invariant because $d\ln\alpha$ is closed. The invariance is thus exact and does not depend on the parameter being a twistor function; what depends on it is that the transformation is a symmetry of the primary system at all, and that the shift is a gradient of a function which the congruence itself defines.

Physical interpretation. The corpus's gauge articles take the gauge principle as a postulate: a symmetry is demanded, a connection is introduced to make a derivative covariant, and the field strength is the curvature. Here the order is reversed. The symmetry is a residual freedom of an over-determined system; the parameter is constrained by the system's own solution space; the connection is not introduced but read off; and the field strength's invariance is automatic. Whether that is an advance or a bookkeeping change is the judgement the corpus does not make, and the same reversal is stated in The Algebrodynamical Programme for the Maxwell and Yang–Mills equations.

The Three Fields from One Function

Three distinct field equations hold on the solutions of the congruence equation, and all three are read from the same $\Pi$. The electromagnetic and Yang–Mills statements are recorded in The Algebrodynamical Programme and are repeated here only in the form the twistor makes available; the spinor statement is not recorded there and is the reason this section exists.

The Electromagnetic Field

Let the field-generating function be $\ln G$; the source writes it $\Lambda:=\ln G$, and this article spells it out to avoid the letter, which the programme article uses for $d^2\Pi/dG^2$ and, separately, for the conformal factor of its quaternionic maps. It is a function of the single twistor component $G$, so it satisfies the pair, the eikonal and the wave equation, by the argument of the section above. The programme article writes the three independent components of the antiself-dual field — antiself-dual in the convention $F+iF^{*}=0$, which is the source's (51), and which the source itself calls self-duality — as the second derivatives of $\ln G$,

$$ F_{00}=\frac{\partial^2\ln G}{\partial\upsilon\,\partial p}, \qquad F_{01}=\frac{\partial^2\ln G}{\partial w\,\partial p}, \qquad F_{11}=\frac{\partial^2\ln G}{\partial v\,\partial w}, $$

and the Maxwell equations then hold for it because $\ln G$ is harmonic. That was checked directly: with $G$ solved at forty digits and the derivatives taken in closed form, both the eikonal residual of $\ln G$ and the difference $\partial^2\ln G/\partial\upsilon\,\partial v-\partial^2\ln G/\partial w\,\partial p$ vanish to the working precision, below $10^{-40}$, for the quadratic generator and for a transcendental one; a finite-difference cross-check reproduces the same quantities to twelve digits. The mixed-derivative identity is $\partial^2\ln G/\partial\upsilon\,\partial v-\partial^2\ln G/\partial w\,\partial p=0$ written out.

A second normalisation of the same field exists, and it must not be confused with this one. The paper writes only (64), with $\Lambda=\ln G$; the second derivatives of $G$ itself are the corpus's, not the paper's, and they come from the other spinor of the construction. The Algebrodynamical Programme records that family: for every holomorphic function $F(G)$ of the twistor ratio — the letter names a scalar there and not the field strength, which is a different object in this article — the spinor taken from the pair $(\upsilon,w)$ has components

$$ \xi_{\mathrm{scl},\upsilon\upsilon}=\frac{\partial^2F}{\partial\upsilon^2}, \qquad \xi_{\mathrm{scl},\upsilon w}=\frac{\partial^2F}{\partial w\,\partial\upsilon}, \qquad \xi_{\mathrm{scl},ww}=\frac{\partial^2F}{\partial w^2}, $$

and the alignment condition is $\mathcal{D}^2F=0$ for the ray derivative $\mathcal{D}:=\partial/\partial w-G\,\partial/\partial\upsilon$, which holds because $\mathcal{D}F=F'(G)\,\mathcal{D}G$ and $\mathcal{D}G=0$ by the pair. Those are components in the $(\upsilon,w)$ pair; the paper's (64) uses $(\upsilon,p)$, $(w,p)$ and $(v,w)$. The two sets therefore have different index structure and are not the same components — $\partial\ln G=\partial G/G$ leaves behind a term quadratic in the first derivatives — and this article keeps them apart instead of presenting one as the other. What the two share is the singular locus, not the normalisation.

The sharpest form of the result is the source's. The field can be written through the derivatives of $\Pi$ alone,

$$ F_{AB}=\frac{1}{P}\,\Pi_{AB}-\frac{1}{P}\,\frac{d}{dG}\left(\frac{\Pi_A\Pi_B}{P}\right), $$

with $\Pi_A,\Pi_{AB}$ the first and second derivatives of $\Pi$ in its two twistor arguments, $A,B\in\{0,1\}$ indexing the second and third arguments as in the source (the companion article writes $\Pi_1,\Pi_2$ for the same pair, its indices shifted by one), and $P=d\Pi/dG$ its total derivative along the line, so that the field of a congruence is computed from the generating function without ever solving the Kerr equation for $G$. The corpus records this formula in The Algebrodynamical Programme; what this article adds is the reading of why it can exist at all, and a check of its two transcription choices, below. The field is a second derivative of a function of the twistor, the twistor is constrained to a surface, and a function on a surface has its derivatives fixed by the surface's defining equation — which is exactly what a formula in $\Pi$ and its derivatives expresses.

The indices on the display are raised, and that was checked. The display cannot be read with every index low, and the raising is the reason: with $\varepsilon_{01}=1$ the inverse symbol is $\varepsilon^{01}=-1$, and raising both indices gives $\Pi^{00}=\Pi_{11}$, $\Pi^{11}=\Pi_{00}$ and $\Pi^{01}=-\Pi_{01}$ — the last of these being where the sign on the mixed component comes from. So the display is

$$ F_{AB}=\frac{1}{P}\,\Pi^{AB}-\frac{1}{P}\,\frac{d}{dG}\left(\frac{\Pi^{A}\Pi^{B}}{P}\right), $$

indices up throughout. The static generator tests this exactly, because there $\Pi$ is linear in the twistor arguments and $\Pi_{AB}$ vanishes identically: only the second term survives, so it alone must produce all three components of (64). It does, to thirty digits — $F_{00}$ from $\Pi_1\Pi_1$, $F_{11}$ from $\Pi_0\Pi_0$, and $F_{01}$ from $\Pi_0\Pi_1$ with the minus sign that the raising supplies, residuals of order $10^{-31}$. Two things follow. The corpus's $1/P$ on the second term, which the source's display does not print, is the correct normalisation, and the check is what settles it. And the diagonal correspondence runs crossed — $\Pi_1\Pi_1$ to $F_{00}$ and $\Pi_0\Pi_0$ to $F_{11}$ — which is a property of the raising and not of the formula, so a reader comparing the display with (64) in all-low spelling will find the diagonal inverted unless the indices are restored.

The Spinor inside the Potential

This is the part of the paper the corpus does not otherwise have, and it is the reason the construction reaches the fermionic sector at all.

The potential is a complex four-vector, as the source calls it, and in the spinor basis that is one statement and not two: a four-vector is an object of type $(\tfrac12,\tfrac12)$, which is exactly one primed and one unprimed spinor index. Concretely it is the array $Ξ_{B'A}$ of the source's (13), built from the spinor by $Ξ_{B'A}=\sum_{A}\iota_{A}\frac{\partial\zeta_{B'}}{\partial Z_{AA'}}$ with a conjugate spinor $\iota$ normalised by $\zeta_{0'}\iota_{0}+\zeta_{1'}\iota_{1}=1$. Every spinor index is written low, as the corpus requires, so the sum that the source leaves to the index convention is named: $\sum_{A}$ runs over the two values, and the raised index the source puts on $\iota$ in its (13) and (14) is written low here — a renaming of the source's two components and not a lowering, so no antisymmetric symbol enters and the normalisation displayed is what defines them. Where an index really is lowered, as in the Weyl equation below, the cost is paid explicitly. The two indices are of different kinds: by the transformation law (10), $Ξ\mapsto\bar L^{\dagger}Ξ\bar L$, the group acts by $\bar L^{\dagger}$ on the primed index and by $\bar L$ on the unprimed one. (The source prints a plus where the corpus prints the dagger, the mark of the adjoint.) Fixing either index leaves a two-component spinor, and that is where the fermionic reading comes from — not from the four-vector, which is the bosonic one, but from the half of it that a choice of gauge leaves behind.

In the gauge $\zeta_{0'}=1$ half of it vanishes, and the reason is (13). (The source writes the spinor with a primed index, $\zeta_{A'}$; this is this article's $\zeta=(1,G)$, so the gauge is the $\zeta_0=1$ used above, and only the spelling differs.) There $Ξ_{0'A}=\sum_{A}\iota_{A}\frac{\partial\zeta_{0'}}{\partial Z_{AA'}}$ is the derivative of the constant $\zeta_{0'}=1$ and therefore zero, so the array is

$$ \begin{pmatrix}Ξ_{0'0} & Ξ_{0'1}\ Ξ_{1'0} & Ξ_{1'1}\end{pmatrix} =\begin{pmatrix}0 & 0\ \xi_{\mathrm{scl},0} & \xi_{\mathrm{scl},1}\end{pmatrix}, \qquad \xi_{\mathrm{scl},0}=\frac{\partial\ln G}{\partial w}=\frac{\partial G}{\partial\upsilon}, \qquad \xi_{\mathrm{scl},1}=\frac{\partial\ln G}{\partial v}=\frac{\partial G}{\partial p}, $$

the second equality in each being the pair of relations. Two of the four components of the potential are zero by the choice of gauge, and the two that remain form a chiral two-spinor. The vanishing is not an extra assumption: it is what the projective freedom $\zeta\mapsto\alpha(T)\zeta$ is spent on, and the source draws it from (13) and not from a postulate about the potential.

The surviving pair satisfies the massless Weyl equation, and its content is the wave equation for $G$. The source writes it as a divergence, one gradient with its index raised and contracted into the spinor. Lowering that index is not a formality — it costs the antisymmetric symbol, $\varepsilon_{01}=1$ — so in the corpus's all-low spelling, where nothing is raised, the equation is $\partial\xi_{\mathrm{scl},1}/\partial Z_{0B'}-\partial\xi_{\mathrm{scl},0}/\partial Z_{1B'}=0$, whose two component equations, at $B'=0'$ and $B'=1'$, are

$$ \frac{\partial\xi_{\mathrm{scl},0}}{\partial p}=\frac{\partial\xi_{\mathrm{scl},1}}{\partial\upsilon}, \qquad \frac{\partial\xi_{\mathrm{scl},0}}{\partial v}=\frac{\partial\xi_{\mathrm{scl},1}}{\partial w}, $$

with $p=x+iy=\bar w$ the fourth chart coordinate throughout, here as in $\kappa_1=vG+p$. Substituting the components, the first reads $\partial^2G/\partial\upsilon\,\partial p=\partial^2G/\partial p\,\partial\upsilon$ and is an identity of mixed partials; the second reads $\partial^2G/\partial\upsilon\,\partial v=\partial^2G/\partial w\,\partial p$ and is the wave equation $\partial^2G/\partial\upsilon\,\partial v-\partial^2G/\partial w\,\partial p=0$. So the pair carries the wave equation for $G$ beside an identity, and it adds no equation of its own: the Weyl relations hold because $G$ is null and harmonic, which is what the pair of the earlier section supplies. The source states the two equations as equivalent to the self-duality conditions together with the inhomogeneous Lorentz condition, which is what the first-order pair already encodes.

What that leaves is nonetheless the substantive point. A two-component spinor field, obeying the relations that a massless chiral fermion's field obeys, is what the complex four-potential becomes once its gauge freedom is spent, with no Dirac equation and no gamma matrices anywhere in the derivation. The potential is bigger than the electromagnetic field it carries, and the extra part is not gauge — it is a spinor. What the construction does not supply is a reason for the spinor to have a mass or to obey a massive equation, and the source claims neither.

Physical interpretation. The corpus derives its fermions from the Dirac equation and its bosons from the gauge principle, and treats the two as separate imports. Here one holomorphic function of the twistor yields the electromagnetic field, a complex Yang–Mills triplet, and a two-component spinor, the first two as its second derivatives and the third as the potential's own surviving components. The three are not unified in the sense of a group that contains them, and the algebra does not identify their couplings; what is unified is their origin. Since the source's construction is an external programme, the corpus records this as the programme's structure and not as a result of the framework.

The Triplet

The trace-free part of the same connection is a complex triplet whose strength the source writes in closed form through the electromagnetic one and the ratio $G$,

$$ F=\{-GF,\;iGF,\;-F\}, $$

with the repeated $F$ on the left the triplet and on the right the electromagnetic two-form. Its modulus in the complexified isotopic space equals the modulus of the electromagnetic field, and this is not an approximation: the three squares sum to

$$ (-GF)^2+(iGF)^2+(-F)^2=F^2\bigl(G^2-G^2+1\bigr)=F^2 , $$

which is an identity in $G$ and $F$ and was checked on a hundred random complex pairs, the worst relative residual at the double-precision floor, of order $10^{-15}$. The 1995 presentation prints the relation as an exact equality, $L_{a\mu\nu}L_{a\mu\nu}=(F_{\mu\nu})^2$ (its (50)), while the programme article records that a recomputation with the source's own connection normalisations gives $\tfrac{1}{16}$ of the right-hand side and leaves the factor unreconciled. The factor is now accounted for, and it is not a discrepancy at all. Both accounts are the determinant identity $\det(cF\,I+L)=(cF)^2-\sum_a(L_a)^2=0$ applied with different trace coefficients $c$: with the curvature split as $\mathrm{Ric}=cF\,I+L$ and $L$ traceless, the identity gives $\sum_a(L_a)^2=c^2F^2$. The source's (50) is the case $c=1$, where the triplet's own squares already sum to $F^2$. The programme's recomputation gets the trace part $\tfrac14F\,I$, and because $\det\mathrm{Ric}=0$ holds for the matrix it actually builds, the identity then forces $\sum_a(L_a)^2=\tfrac{1}{16}F^2$ — which is the value it reports. So the corpus's curvature is the source's scaled by $\tfrac14$ throughout, trace part and triplet alike, and $\tfrac{1}{16}=(\tfrac14)^2$ is that one scaling squared. What follows for a reader is a rule rather than a reconciliation: the identity ties the trace coefficient to the triplet, so the two must be compared as wholes and never mixed. Taking the source's $\{-GF,iGF,-F\}$ with the corpus's trace coefficient gives $\det\neq0$, which is not a third convention but an error.

The cancellation is the interesting part. The identity holds because the first two components of the multiplier $\{-G,iG,-1\}$ have squares $G^2$ and $-G^2$. Had the second component been $G$ rather than $iG$, the two would have added instead of cancelling and the modulus would have been $F^2(2G^2+1)$, so the abelian field would not have been the length of the non-abelian one. It is the minus sign from $i^2$ and nothing else that makes the length come out right. So the modulus relation is a statement about the complex structure and not about the algebra of the triplet index: the two components that cancel do so because one carries the central $i$ and the other does not, and the three-dimensional index is being used as a complex three-space rather than as a real one. This is the same point the source makes when it observes that the Yang–Mills field of the construction is necessarily complex — its real and imaginary parts separately do not obey free equations, because the equations are nonlinear.

And the internal direction is not free. The multiplier $\{-G,iG,-1\}$ is built from $G$, which is the projective coordinate of the congruence. So the direction in which the non-abelian field differs from the abelian one is a function of the congruence's own direction, and the triplet carries no internal parameter that the congruence does not already carry. The multiplier is not null — its squared modulus is one — so the reading is not that the internal direction is the null direction; the reading is that the two are not independent, one being the image of the other. That is a strong claim and an unusual one, and the corpus records it without endorsing it: it says that in this construction an internal index and a spacetime direction are distinguished by where they sit in the formula rather than by their origin.

The Caustic Is the Particle

For the effective curvature and for the electromagnetic field the singular locus is the same locus, and the twistor picture names it: the set of points whose line is tangent to the twistor surface. The coincidence is the source's, is recorded in The Algebrodynamical Programme, and is carried here with the qualification that belongs to it: the associated Yang–Mills fields, and the Weyl spinor beside them, carry generically additional string-like singularities, so the coincidence is exact for the curvature and the electromagnetic field and is not claimed for the other two. The source says this in a footnote and says it of both: the triplet's strings are those of $G$ itself and resemble the Dirac string, and the two-spinor built from the same potential has strings of the same type. The spinor's strings matter here more than the triplet's, because the spinor is this article's own addition to the corpus: its singular set is larger than the caustic, and the tangency picture describes only the part of it that the caustic accounts for. Naming the locus a tangency turns the coincidence into a theorem of intersection theory: a point is singular for the field exactly when the algebraic resolution of $\Pi$ for $G$ degenerates, and a resolution degenerates exactly when the line and the surface are tangent, and the field is a function of the resolved $G$ and its derivatives. For the two abelian objects nothing else can happen, so their coincidence is forced.

The static family is the test case, and it works exactly. For the generator of the ring solution,

$$ \Pi=wG^2+(\upsilon-v+2ia)G-p , $$

with $z:=z_3=\tfrac12(\upsilon-v)$ and $a$ the ring's radius, the tangency condition $P=0$ gives $G=-(z+ia)/w$, and substituting it back into $\Pi=0$ leaves

$$ \Pi\big|_{P=0}=-\frac{r_x^2+r_y^2+(r_z+ia)^2}{w}=-\frac{r_*^2}{w}, $$

checked symbolically with $w=x-iy$ and $p=x+iy$. The tangency locus is therefore $r_*^2=0$ — the ring of radius $a$ in the plane $z=0$, with its two real conditions $r_x^2+r_y^2+r_z^2=a^2$ and $2az=0$. The particle of the programme is the set of lines tangent to the twistor surface, and for the Kerr generator that set is the Kerr ring, which is as direct a confirmation of the geometric reading as the construction admits.

The generator is a twistor function, and the count of roots is its degree. Written in the coordinates, the display does not look like a function of $(G,\kappa_0,\kappa_1)$, and the definition of the surface requires that it be one. It is, since $G\kappa_0-\kappa_1=wG^2+(\upsilon-v)G-p$, and one correction to the source belongs at exactly this point. The source prints (71) as $\Pi=G\tau_0-\tau_1+2ia$, with $2ia$ a constant, beside the root $G=\bar w/[(z+ia)\pm r_*]$; the two do not fit. Read with the constant, the polynomial takes at that root the value $2ia(1-G)$, which vanishes only at $G=1$, and its own tangency elimination gives $-(r_x^2+r_y^2+r_z^2-2iaw)/w$ where the ring identity gives $-r_*^2/w$. Both residuals were computed, and they identify the intended generator as the member of the source's own (70) family with $\xi_{\mathrm{scl}}=2ia\,G$:

$$ \Pi=G\,\kappa_0-\kappa_1+2ia\,G , $$

which vanishes at the printed root identically and gives exactly $-r_*^2/w$ at tangency. This article uses that member, and records the difference between it and the printed polynomial, $2ia(G-1)$, rather than correcting the display in silence. In the (70) family, $\xi_{\mathrm{scl}}(G)+(qG+s)\,\tau_1-(\gamma G+\bar q)\,\tau_0$ — the source's constant $p$ is written $\gamma$ here, the letter being spent on the chart coordinate — the parameters are $q=\bar q=0$ and $s=\gamma=-1$, so the Kerr–Schild statement of the source attaches to it, and the two roots counted in the section on the twistor surface are the two roots of a polynomial of degree two in $G$ — the degree of the term $G\kappa_0$ and of nothing else.

What the picture does and does not settle. It settles why the singularities coincide, and it gives their dimension count. Tangency is one complex condition, hence two real conditions, so the singular set has codimension two: a surface generically in four coordinates, a curve when one coordinate is ignorable and the two conditions fall on three, and a point at a further degeneracy. That is the source's catalogue of points, strings and surfaces, and it is a count rather than an observation. What the picture does not settle is what the singularities are physically. The corpus records the programme's particle reading — a common singularity of the fields, with quantised charge — in the article that owns it, and does not adopt it here.

What the Construction Is and Is Not

It is not the Penrose transform. The incidence geometry is the classical one, and the null quadric is the same quadric; but the object attached to a twistor here is the general solution of a nonlinear first-order system, not a solution of a linear massless field equation. There is no cohomology, no $\bar\partial$-closed form, no contour integral, and no transform in the sense of the twistor programme: the field is an algebraic function of the same $\Pi$ that defines the congruence, and the "transform" is differentiation. Twistor Theory and Biquaternions records the three senses in which the corpus's literature uses the word; this article is the sense it calls the congruence twistor, taken past the incidence relation to the fields.

It bears on two of the open questions of Introduction to the Biquaternion Universe. Its question 1 asks whether the algebra's structure produces consequences or only reformulates known physics, to be judged by the richness of what it yields; its question 4 asks for a relativistic theory of spinor fields whose classical precursor is the biquaternion Dirac equation. The construction here does something that neither of those programmes does: it yields a chiral two-component spinor, a Maxwell field and a non-abelian triplet from one holomorphic function, with no gauge principle imposed and no Dirac equation written down. It does so, however, on an algebra of coordinates rather than on the corpus's algebra of fields, so the answer is given in a neighbouring setting and does not transfer without a translation the corpus does not possess. Whether the construction is a derivation of the linear equations or only a very economical bookkeeping is the question it leaves open, and it is the right question to leave open.

What it does not give. No action, hence no canonical quantisation of the kind the corpus's quantisation articles presuppose. No superposition: a sum of two solutions satisfies the linear equations but not the congruence equation, which the programme article records as a selection rule and a cost. No dynamics beyond the algebraic constraint: the "time evolution" of a singular locus is the motion of a tangency, and it is determined, not postulated, which is either the construction's strength or its limitation depending on what one wanted. And no numerical prediction except the one the programme derives elsewhere.

Summary

A congruence of null rays is a two-surface $S$ in projective twistor space; a spacetime point is a line $L_x$ in that space; the congruence's principal spinor at the point is the intersection $L_x\cap S$, and the number of its branches is the degree of the surface, by Bézout, counted with multiplicity and generically that many. The whole construction is that one correspondence, and the algebra that makes it computable is one pair of first-order relations, $\partial\lambda/\partial w=G\,\partial\lambda/\partial\upsilon$ and $\partial\lambda/\partial v=G\,\partial\lambda/\partial p$, satisfied by every function of the twistor because it is satisfied by the twistor's own components. Multiplying the pair gives the complex eikonal; cross-differentiating it gives a wave equation for the functions of $G$ and not for the general twistor function, with the obstruction computed and reduced to a single Jacobian in the two twistor components, the quantity that vanishes exactly for the harmonic ones and that leaves the twistor's own components carrying the caustic factor $1/P$; and written on the spinor $\zeta=(1,G)$ the pair is the constraint that the programme imposes on its rescaling parameter. The condition for a rescaling to be a symmetry and the condition for a function to solve the eikonal are therefore one condition, whose solutions are the functions of the twistor.

From that pair everything else follows. The four-component twistor field has rank two — verified as an exact vanishing of every three-by-three minor at the rounding floor of forty-digit arithmetic, with the non-vanishing minors between a tenth and a half, over three generating functions and a hundred points each — which is why it is a surface and why the inverse of a rescaling is available at all, granted the parameter is read on the projective twistor as the section on the group requires. The group's orbits are the projective twistors, so the projective structure is derived from the restricted gauge symmetry rather than assumed, and the potential shifts by a gradient while the field strength stays fixed. Then three fields come out of one function: the electromagnetic field as the second derivatives of $\ln G$, guaranteed to obey the free Maxwell equations by the harmonicity that the pair supplies for a function of $G$; a complex Yang–Mills triplet whose modulus equals the electromagnetic field because the squares of its first two components cancel between $G^2$ and $(iG)^2$; and a two-component spinor inside the four-potential: the gauge $\zeta_{0'}=1$ annihilates the first row of the potential array $Ξ_{B'A}$ by the source's (13), and the surviving pair $\xi_{\mathrm{scl},0}=\partial\ln G/\partial w=\partial G/\partial\upsilon$, $\xi_{\mathrm{scl},1}=\partial\ln G/\partial v=\partial G/\partial p$ satisfies the Weyl relations. Those relations carry the wave equation $\partial^2G/\partial\upsilon\,\partial v-\partial^2G/\partial w\,\partial p=0$ beside an identity of mixed partials, so the reduction is recorded as a representation of the harmonicity and not as a new equation.

The singular locus is the tangency locus: a point is singular for the field exactly when the line and the surface are tangent, because the field is a function of the resolved $G$ and a resolution degenerates exactly there. For the static ring generator the condition reduces exactly to $r_*^2=0$, so the ring of the programme is literally the set of lines tangent to its twistor surface — for the generator that the source's own root requires, which is a correction of one term in the display the source prints, recorded above. What this settles is the coincidence of the two abelian singular loci and the dimension count of the tangency set; what it does not settle is what the singularities are, and the corpus leaves the programme's particle reading with the programme.

Summary of Notation

Symbol Meaning
$Z$, $Z_{AB}$ Complex coordinate matrix of complexified Minkowski space
$\upsilon,v,w,p$ The entries $Z_{00},Z_{11},Z_{01},Z_{10}$; $w=x-iy$, $p=x+iy$ on the real slice
$\zeta=(\zeta_0,\zeta_1)$, $\tau=Z\zeta$ Principal spinor and the incident twistor
$\tau^{*}\zeta-\zeta^{*}\tau=0$ Null-twistor condition; automatic for Hermitian $Z$
$G=\zeta_1/\zeta_0$ Principal ratio, the congruence's projective coordinate; gauge $\zeta_0=1$
$\kappa_0=wG+\upsilon$, $\kappa_1=vG+p$ The two twistor components attached to $G$
$T=(\zeta,\tau)$, projectively $(G,\kappa_0,\kappa_1)$ The twistor of the point: four components, whose projective class is the three coordinates; one symbol where the group acts on it
$\Pi(G,\kappa_0,\kappa_1)$ Generating (world) function; $S=\{\Pi=0\}$ is the twistor surface
$P=d\Pi/dG=\Pi_G+w\,\Pi_0+v\,\Pi_1$ Total derivative along the line; the source's (60) and the programme article's letter alike. Not the spatial Laplacian, which is another operator under another letter throughout the corpus. The tangency condition is $P=0$ with $\Pi=0$
$\Pi_G$, $\Pi_A$, $\Pi_{AB}$ Derivatives of $\Pi$: $\Pi_G$ in $G$, and $\Pi_A,\Pi_{AB}$ in the twistor arguments with $A,B\in\{0,1\}$ over $(\kappa_0,\kappa_1)$, so that $\Pi_0=\partial\Pi/\partial\kappa_0$ and $\Pi_1=\partial\Pi/\partial\kappa_1$ — the source's counting, which differs by one from the companion article's
$\partial\lambda/\partial w=G\,\partial\lambda/\partial\upsilon$, $\partial\lambda/\partial v=G\,\partial\lambda/\partial p$ The pair of relations, obeyed by every twistor function $\lambda$
$\Sigma$ Generic symbol for the eikonal function: $\Sigma=G$ is the congruence's own solution, and every twistor function is one
$\mathcal{D}=\partial/\partial w-G\,\partial/\partial\upsilon$ The ray derivative; it annihilates $G$ by the pair, and with it every function of $G$
$\alpha(T)$ Rescaling parameter; a symmetry exactly when it is a twistor function, and the group law needs it read projectively
$\mathcal{A}\mapsto\mathcal{A}+d\ln\alpha$ Shift of the potential one-form $\mathcal{A}=2\,\mathrm{Sc}(Ξ\,dZ)$; $F=d\mathcal{A}$ invariant
$\ln G$ Field-generating function; the source writes it $\Lambda$, the programme article writes it in full
$F_{00},F_{01},F_{11}$ Independent components of the antiself-dual field, as second derivatives of $\ln G$; the source's (64)
$Ξ_{B'A}$ The potential array of the source's (13), one primed and one unprimed index; a complex four-vector
$\iota_{A}$ The conjugate spinor of (14), $\zeta_{0'}\iota_{0}+\zeta_{1'}\iota_{1}=1$; all spinor indices low, sums named
$\bar L^{\dagger}$, $\bar L$ The two factors of the transformation (10), $Ξ\mapsto\bar L^{\dagger}Ξ\bar L$; the source prints a plus for the dagger
$\xi_{\mathrm{scl},0},\xi_{\mathrm{scl},1}$ The two surviving components of the potential, $\xi_{\mathrm{scl},0}=\partial\ln G/\partial w$ and $\xi_{\mathrm{scl},1}=\partial\ln G/\partial v$ in the gauge $\zeta_{0'}=1$
$\{-GF,iGF,-F\}$ Yang–Mills triplet; its modulus is $F$
$\xi_{\mathrm{scl}}(G)$, $q$, $s$, $\gamma$ Constants of the source's (70) family, to which the static generator belongs; the source's real constant $p$ is written $\gamma$ here because $p$ is a chart coordinate
$r_*^2=r_x^2+r_y^2+(r_z+ia)^2$ Ring condition; the tangency locus of the static generator is $r_*^2=0$

Further Reading

  • Null Electromagnetic Fields and the Shear-Free Congruence, for the null field, Bateman's representation, the principal null direction and the Mariot–Robinson theorem, which is the classical statement that a null field and a shear-free congruence determine one another.
  • The Algebrodynamical Programme: Nonlinear Cauchy–Riemann Conditions, Self-Quantized Charge, and Induced Causal Geometry, which owns the programme this article takes its twistor from: the generating system, the weak self-duality, the Yang–Mills triplet, the field from the derivatives of $\Pi$ alone, the charge quantisation and the particle reading.
  • Twistor Theory and Biquaternions, for the incidence relation, the two spaces $S\oplus\bar S$, the Klein quadric, and the three senses in which the corpus uses the word twistor.
  • The Two-Sided Derivative over the Biquaternion Algebra, which owns the coordinate chart $\upsilon,v,w,p$, the determinant identity behind the complex eikonal, and the proof that the twistor-generated class is both null and harmonic.
  • Algebraic Curves, for Bézout's theorem in its projective-plane form and for the intersection multiplicity that the branch count of the congruence is an instance of.
  • Helicity in Biquaternionic Form, for the massless momentum as a zero divisor, the rank-one factorisation of a null object, and the little group — the structures that the incidence relation's two-column decomposition is the geometric counterpart of.
  • The Yang–Mills Equation in Biquaternionic Form, for the differential Yang–Mills equation and its Bianchi identity, against which the algebraic triplet of this article should be read.
  • The Self-Dual and Anti-Self-Dual Split: Spin 1 from the Bivector Plane, for the two halves of a complexified two-form and the naming of the two signs, which this article's $F_{01}$ and its partners presuppose.
  • V. V. Kassandrov and J. A. Rizcallah, "Maxwell, Yang–Mills, Weyl and eikonal fields defined by any null shear-free congruence", arXiv:1612.06718 [gr-qc] (2016), the paper this article was written around: the twistor form of the shear-free equations, the weakened gauge group and its projective orbits, the complex eikonal equation for an arbitrary function of the spinor components, the field through the derivatives of $\Pi$ alone, the Weyl spinor inside the potential, and the closed-form Yang–Mills triplet with its modulus. The paper is cited nowhere else in the corpus; the field algebra it belongs to is recorded under the programme article above.
  • R. Penrose and W. Rindler, Spinors and Space-Time, Vol. 1 and Vol. 2 (Cambridge University Press, 1984 and 1986), for the two-spinor calculus, the incidence relation, and the Kerr theorem in its standard form.
  • G. C. Debney, R. P. Kerr and A. B. Schild, "Solutions of the Einstein and Einstein–Maxwell Equations", Journal of Mathematical Physics 10 (1969) 1842–1854, for the electrovacuum Kerr–Schild family that the linear generating functions produce.
  • R. P. Kerr and W. B. Wilson, "Singularities in the Kerr–Schild metrics", General Relativity and Gravitation 10 (1979) 273, for the theorem that the static solutions of the congruence equation with a singular set bounded in three-space are exhausted by the Kerr solution, up to translations and rotations.