The Roots of a General Element: Mass, Frames and the Mass Shell
Introduction
The root problem of the biquaternion algebra asks for the elements that square to a given element. For a central value $\lambda e_t$ the question is the one the two companion articles answer at $\lambda=-1,0,+1$; for a general element $\tilde Q$ it is the question of this article,
$$ \boxed{\;\tilde\Xi^2=\tilde Q\;}. $$
The answer depends on the product, and the first job of the article is to keep the products apart. The plain square $\tilde\Xi^2$ has a finite root set for a non-scalar radicand, generically four elements, falling to two or to none. The natural square $\tilde\Xi\tilde\Xi^{\natural}=\langle\tilde\Xi\rangle_{\natural}e_t$ has the level set of the interval as its root set, and on the material sector that level set is the mass shell. The Hermitian square $\tilde\Xi\tilde\Xi^{*}$ produces a positive semidefinite value; the eigenvalues of the mass operator are the mass-squareds of a doublet. The three squares are three root problems, not one.
The article interprets five physical consequences. The mass shell is a propagator pole; the mass can be a field; the mass can be generated by a condensate; the root can be a geometric map; and the roots are frames. Each interpretation is labelled, and each carries its boundary.
The article owns no theorem. It states the algebraic classification, with its source, and reads it. The proofs, the case distinctions and the examples are in the mathematics article.
The Results It Uses
The Root Problem by Band
The Plain Square
Write the radicand and the root in scalar and vector parts,
$$ \tilde Q=Q_te_t+\vec Q, \qquad \tilde\Xi=\Xi_te_t+\vec\Xi . $$
The plain square is
$$ \tilde\Xi^2=\bigl(\Xi_t^2-(\vec\Xi,\vec\Xi)\bigr)e_t+2\Xi_t\vec\Xi . $$
For a radicand with $\vec Q\neq0$ a root has $\Xi_t\neq0$ and $\vec\Xi=\vec Q/(2\Xi_t)$, and $x=\Xi_t^2$ solves the reduced quadratic
$$ 4x^2-4Q_tx-(\vec Q,\vec Q)=0 . $$
The roots come in opposite pairs $\pm\tilde\Xi$. There are four roots when $(\vec Q,\vec Q)\neq0$ and $n=Q_t^2+(\vec Q,\vec Q)\neq0$; two roots in the two degenerate rows; no roots when $(\vec Q,\vec Q)=0$, $n=0$ and $Q_t=0$; and a continuous family for a complex scalar radicand $\tilde Q=Q_te_t$, whose two scalar roots $\pm\sqrt{Q_t}e_t$ are joined by a four-real-parameter family of pure $\vec\Xi$ with $(\vec\Xi,\vec\Xi)=-Q_t$.
That count is the unit-case count of Biquaternion Elementary Functions, §The Roots. The table there reads the same multiplicity by regime. There are four for a unit with $(\vec Q,\vec Q)\neq0$, two when the norm vanishes with $Q_t\neq0$, two in the Jordan row, none for a nonzero nilpotent, and a continuous family on the scalar axis. The four plain roots are the four combinations of the two eigenvalue logarithms, modulo the global sign. The shifts that leave the root fixed remove the rest of the lattice.
The root set of $0$ is the nilpotent cone together with the origin. In the natural band the root set of a nonzero value is a level set, not a cone; in the plain band it is the set of the classification — finite for a non-scalar radicand, and a finite set together with a four-parameter family for a central scalar. For a central nonzero value, every root is a unit: the scalar roots are units, and a pure root has $(\vec\Xi,\vec\Xi)\neq0$, so its norm does not vanish. A non-trivial idempotent is its own root and is therefore a zero divisor; the unit statement is a statement about central values.
The Natural Square and the Shell
The natural square of an element is its interval,
$$ \tilde\Xi\tilde\Xi^{\natural}=\langle\tilde\Xi\rangle_{\natural}e_t, \qquad \langle\tilde\Xi\rangle_{\natural}=\Xi_t^2+(\vec\Xi,\vec\Xi). $$
The root set of a central value $\lambda e_t$ is the level set $\langle\tilde\Xi\rangle_{\natural}=\lambda$. On the material sector $\mathbb{M}_-$, write the four-momentum as
$$ \tilde P=i\frac{E}{c}e_t+\vec p . $$
Its norm is
$$ \langle\tilde P\rangle_{\natural}=-\frac{E^2}{c^2}+\vec p^{\,2}, $$
and the level set $\langle\tilde P\rangle_{\natural}=-m^2c^2$ is the mass shell,
$$ E^2=\vec p^{\,2}c^2+m^2c^4 . $$
The root set of the value $-m^2c^2$ in the natural band is the whole shell, a three-dimensional two-sheeted hyperboloid, not a finite set. Every point of the shell is a four-momentum; the root is the four-momentum itself. This is the precise sense in which the root problem is the mass-shell problem: it is the natural root problem, not the plain one.
The mass shell is not the plain root set of a central value. The plain roots of $-m^2c^2$ include the scalar elements $\pm imc\,e_t$ and a four-parameter family of pure roots with $(\vec\Xi,\vec\Xi)=m^2c^2$; they are not the four-momenta of the shell. The two statements are different and both are correct in their own band.
The Hermitian Square and the Mass Operator
The Hermitian square is
$$ \tilde\Xi\tilde\Xi^{*}, $$
a positive semidefinite element of $\mathbb{M}_+$. Its eigenvalues are non-negative, and its root set is a family of amplitudes of fixed Born weight. This is the band in which the mass-squared operator of a doublet lives. The three squares are the three places a root can be read: the plain square for the algebraic multiplicity, the natural square for the interval, the Hermitian square for the mass operator.
The Sign and the Causal Class
On the material slice the sign of a real central value is the causal class of its roots. The value $-m^2c^2$ is negative and its roots are timelike; the value $0$ is null and its roots lie on the cone; the value $+m^2c^2$ is positive and its roots are spacelike. The magnitude sets the size of the shell: if the unit shell is $\langle\tilde P\rangle_{\natural}=-1$, the shell of mass $m$ is its image under $\tilde P\mapsto mc\,\tilde P$, because the norm is quadratic in the scale.
The unit statement follows the same split. A root of a nonzero central value is a unit, so a tachyon is invertible; among the central values, only $0$ has non-invertible roots. The cone is the wall of the units, and the mass shell never touches it for $m\neq0$.
The sign is not decoration. A mass term built with the wrong conjugation places the shell on the spacelike side; the choice of conjugation, not a numerical sign, is what puts the physical mass on the timelike side. The algebra supplies both shells; the physics selects one by the conjugation it uses.
The Mass Is the Modulus
The natural root set is the level set of the norm; the Hamilton pairing names the invariant that the level set carries. For a four-momentum,
$$ \boxed{\;\tilde P=\Rho e^{\vec U\Theta}, \qquad \Rho=imc, \qquad \vec U=\hat p, \qquad \Theta=-i\psi\;}, $$
and the mass is $m=\lvert\Rho\rvert/c$. The root set of the value $-m^{2}c^{2}$ is the orbit of the element $\Rho e^{\vec U\Theta}$ under the frame factor: every point of the shell has the same $\lvert\Rho\rvert$ and a different angle. The level set is the geometric form of the statement, and the modulus is the invariant that defines it.
The mass is the modulus and not the energy. The energy is $mc^{2}\cosh\psi$, the modulus times the frame factor; a boost changes the energy and leaves the modulus. The same division appears in the spectrum: $\lambda_+\lambda_-=\Rho^{2}$ is the modulus, and $\lambda_+/\lambda_-=e^{2\psi}$ is the frame.
The boundary. The identification is licensed for the four-momentum chart and the principal branch. For a general element the Hamilton modulus is a complex amplitude, and the mass reading is an extra identification. The modulus is also not additive: it is multiplicative under composition, and the mass of a sum is not the sum of the masses.
The Isospectral Family
Two elements are isospectral when their eigenvalues agree with multiplicity. The spectrum fixes $Q_t$ and $B^{2}$, so the isospectral family of a generic element is
$$ \mathcal I(\lambda_+,\lambda_-) =\left\{Q_te_t+\vec Q:\ \vec Q\in\mathbb{C}^{3},\ (\vec Q,\vec Q)=B^{2}\right\}, $$
a complex quadric of four real dimensions. On the material slice $\vec Q\in\mathbb{R}^{3}$ and the family is the sphere $\lvert\vec p\rvert=B$: the directions of motion at fixed energy and momentum magnitude.
The family is the set that a spectrum-only measurement leaves open. It is also the common level set of the two invariants, the trace and the norm. A dynamics that conserves both moves inside the family and does not leave it.
The family is larger than a physical frame orbit. On the material slice a rotation acts transitively on the sphere, so the family is one rotation orbit. On the whole algebra the family contains complex vector parts that no real Lorentz transformation reaches: the spectrum is invariant under the full conjugation group, the physical frame group is a subgroup, and the difference is the part of the family that is not a frame.
The degenerate case splits. At $B=0$ the eigenvalues coincide at $Q_t$, and the family is the isotropic quadric $(\vec Q,\vec Q)=0$, the cone through the origin. Within the matrix image a repeated eigenvalue has two conjugation classes, the scalar $Q_te_t$ and the Jordan block $Q_t(e_t+N_0)$. They have the same spectrum and the same norm, and they are not conjugate. The minimal polynomial, not the spectrum, separates them.
The boundary. The family is a set of elements, not a set of states and not a set of particles. Two isospectral elements need not be related by a physical transformation, and the article does not claim that they are. The reading is epistemic: it names what the spectrum leaves undetermined, and the algebra supplies no principle that selects one member.
The Mass Shell Is a Propagator Pole
The two-point function of a scalar of mass $m$ has the momentum-space denominator
$$ \langle\tilde P\rangle_{\natural}+m^2c^2 , $$
up to a normalisation. The propagator is singular where the denominator vanishes, and that locus is the mass shell. The root set of the natural band and the pole set of the propagator are the same set, read once as an algebraic level set and once as an analytic singularity.
The Feynman prescription deforms the shell,
$$ m^2c^2\longmapsto m^2c^2-i\epsilon , $$
and the deformation is a displacement into the complex plane. The algebra supplies the complex plane in which the displacement lives; it does not select the contour. The retarded, advanced and Feynman propagators differ by which side of the real axis the poles are displaced to, and that choice is a boundary condition, not an algebraic identity.
Boundary. The pole is a property of the propagator, not of a single root. The root is a four-momentum; the propagator is a correlation function, and its pole is where the correlation function diverges. The $i\epsilon$ is a prescription, not a derivation; the algebra gives a home for the $\epsilon$ but no principle that chooses the Feynman contour over the retarded or the advanced one.
The Mass Can Be a Field
The scalar mass can be promoted to a biquaternion field. In the Einstein–Mayer promotion of the Lanczos system the mass $m$ becomes $\tilde E$, and the second-order equations for the two field components $\tilde Q_1,\tilde Q_2$ become eigenvalue equations,
$$ \Box\tilde Q_1=\tilde Q_1\bigl(\tilde E\tilde E^{*}\bigr), \qquad \Box\tilde Q_2=\tilde Q_2\bigl(\tilde E^{*}\tilde E\bigr). $$
The two masses of the doublet are the eigenvalues of $\tilde E\tilde E^{*}$, equivalently of $\tilde E^{*}\tilde E$ for the nonzero spectrum. The mass is no longer a number multiplying the field; it is an operator acting on the field from the right.
Three cases show the range the corpus uses. A real scalar $\tilde E=s e_t$ gives $\tilde E\tilde E^{*}=s^2e_t$, a degenerate doublet of mass $|s|$. A Hermitian idempotent $\tilde E=\tilde\Pi$ gives $\tilde\Pi\tilde\Pi^{*}=\tilde\Pi$, with eigenvalues $1$ and $0$: one massive mode and one massless mode. A singular mass field leaves a massless mode whatever else it does, because
$$ \det\mathsf{M}_2\bigl(\tilde E\tilde E^{*}\bigr)=\bigl|\langle\tilde E\rangle_{\natural}\bigr|^2, $$
and the vanishing norm forces a zero eigenvalue. The example $\tilde E=\vec e_x+i\vec e_y$ has mass-squareds $4$ and $0$. A general mass field gives two distinct mass-squareds, the case the three ends of the range do not cover.
Boundary. The framework derives no numerical mass. The algebra is scale-free, and the mass field is put in by hand in the Einstein–Mayer system; the Petiau system closes it with a nonlinear equation, at the cost of terms the corpus has not written for a field-valued mass. The construction shows that the algebra can carry a mass that is a field with its own equation, and that the masses of a doublet are the eigenvalues of a biquaternion. It does not produce the electron mass or any other.
The Mass Can Be Generated by a Condensate
The massless biquaternion equation carries the axial symmetry $\tilde\Psi\mapsto\exp(i\alpha\gamma_5)\tilde\Psi$. A mass term breaks it: the vector phase passes through the mass, the axial phase does not, and the axial current obeys a divergence proportional to the mass, which vanishes at $m=0$; the axial anomaly is a separate quantum effect. A nonzero vacuum expectation of the fermion bilinear,
$$ \langle\bar\Psi\Psi\rangle_{*}\neq0, $$
is the order parameter of the breaking. The mass term and the condensate multiply the same bilinear, so the dynamically generated coefficient is
$$ m=-2G\langle\bar\Psi\Psi\rangle, $$
with $G$ the coupling of the four-fermion interaction. The condensate is a central scalar, Lorentz scalar and neutral under the vector phase; the vector symmetry survives and the axial one is broken. The Goldstone modes follow from the broken generator.
Boundary. The framework has no colour group and no strong gauge dynamics. The compact gauge algebra inside $\mathbb{B}$ is at most $\mathrm{U}(2)$ of real dimension four, and $SU(3)$ is absent, so a QCD-like condensate cannot be derived. The algebra supplies the carrier of the condensate, the broken-symmetry pattern and the sector placement of the order parameter; the coupling, the cutoff, the scale and the numerical value are empirical input. The condensate is also not the real structure $\flat$: it is the linear chiral pairing, not a Majorana-type pairing, and reading it through $\flat$ gives the wrong object.
The Root Can Be a Geometric Map
The Kustaanheimo–Stiefel map carries four coordinates $\varsigma=(\varsigma_1,\varsigma_2,\varsigma_3,\varsigma_0)$ to three,
$$ x_1=2(\varsigma_1\varsigma_3-\varsigma_2\varsigma_0),\qquad x_2=2(\varsigma_1\varsigma_0+\varsigma_2\varsigma_3),\qquad x_3=\varsigma_1^2+\varsigma_2^2-\varsigma_3^2-\varsigma_0^2 . $$
The radius is the quaternion norm,
$$ \boxed{\;r=\sqrt{x_1^2+x_2^2+x_3^2}=\varsigma_1^2+\varsigma_2^2+\varsigma_3^2+\varsigma_0^2=\langle\varsigma\rangle_{\natural}\;}, $$
and the Coulomb singularity is the reciprocal of the norm,
$$ \frac1r=\frac1{\langle\varsigma\rangle_{\natural}} . $$
The map is a circle bundle away from the origin: the fiber over a nonzero $x$ is a circle, and $\varsigma$ and $-\varsigma$ are two points of it. The sign pair is the discrete part of the ambiguity; the circle is the gauge freedom of the KS constraint, not a second physical position. The map is a square root of the position in the sense that the three position coordinates are quadratic in the four KS coordinates, while the radius is the algebra's norm of $\varsigma$.
Boundary. The transformation, the pseudo-time $dt'=r\,ds$, the reduction to the four-dimensional oscillator and the Mehler kernel are standard and imported. The algebra supplies the interpretation: the Coulomb potential is a central scalar, the radius is the norm, and the singularity is the inverse norm. It does not cure the analytic failure of the time-sliced kernel. The sign pair is not two physical positions either: it is one discrete ambiguity inside a circle of representatives, and the circle is the constraint's gauge freedom.
The Roots Are Frames
The plain root set of a generic element has four elements, two opposite pairs. The central scalar has a continuum: two scalar roots and a four-parameter family of pure roots. The algebra does not select one element of the set. A calculation chooses a root, and the choice fixes the sign and the branch of the square. That choice is what the article calls a frame.
The word is not a Lorentz frame and not a gauge orbit. It names the fact that the root is a representative: the value $\tilde Q$ is fixed, the root is one of its square roots, and the passage from one root to another is a discrete sign or a continuous family, not a symmetry of the value. In the natural band the interpretation is sharper. The root set of $-m^2c^2$ is the mass shell, and every point of the shell is a frame in this sense: a representative four-momentum, a choice of energy and momentum that realizes the same invariant mass. The word names the representative, not an observer's basis. The shell is the set of representatives, not a set of particles; a particle is a state, and the shell is its kinematical arena.
Boundary. The roots are not all related by a physical transformation. The four plain roots are two opposite pairs, not one orbit; the central family is a continuum, not a gauge orbit. The frame interpretation is an interpretation of the multiplicity, not a derivation of a gauge symmetry. The algebra gives the multiplicity; it does not give a principle for choosing one root over another.
The Root Set in One Table
The three bands answer three different root questions. The table states the difference.
| band | root equation | root set of a value | physical reading |
|---|---|---|---|
| plain | $\tilde\Xi^2=\tilde Q$ | four, two or none for a non-scalar radicand; two scalars and a four-parameter family for a central scalar | the algebraic multiplicity; roots as frames |
| natural | $\tilde\Xi\tilde\Xi^{\natural}=\lambda e_t$ | the level set $\langle\tilde\Xi\rangle_{\natural}=\lambda$ | the mass shell; the causal class |
| Hermitian | $\tilde\Xi\tilde\Xi^{*}=\tilde Q$ | a positive semidefinite family | the mass-squared operator; the Born weight |
Limits
The article derives no mass, no coupling and no dynamics. The mass field is an input; the condensate needs a dynamics the algebra does not contain; the propagator contour is a boundary condition; the Kustaanheimo–Stiefel transformation is imported. The root set is not a particle spectrum: the mass shell is a set of four-momenta, and the eigenvalues of the mass field are masses, not roots. The frame interpretation is not a gauge symmetry.
The bands must be kept apart. The plain root set of a central value is a finite set together with a four-parameter family; the natural root set is the shell; the Hermitian root set is the mass operator's family. A reader who applies the plain classification to the mass shell will look for a finite root set where there is a hyperboloid, and a reader who applies the natural level set to the plain problem will lose the multiplicity. The distinction is the first result of the article and the boundary of the other four.
Summary
The root problem $\tilde\Xi^2=\tilde Q$ is band-dependent. In the plain band the roots are generically four, falling to two or to none, with a four-parameter family for a central scalar; the root of $0$ is the nilpotent cone. In the natural band the root set of a central value is the level set of the interval, and on the material sector the level set $-m^2c^2$ is the mass shell. In the Hermitian band the root set is the mass-squared family of a doublet. The five interpretations use these sets: the mass shell is a propagator pole; the mass can be a field whose eigenvalues are the doublet masses; the mass can be generated by a chiral condensate; the root can be the Kustaanheimo–Stiefel square root of the position; and the roots are frames, the representatives the algebra does not choose among. The mass itself is the Hamilton modulus, $m=\lvert\Rho\rvert/c$, and the shell is its level set. The isospectral family of a generic element is the complex quadric of fixed trace and fixed vector square; on the material slice it is the sphere of directions, and at a repeated eigenvalue it splits into a scalar and a Jordan class. The algebra supplies the sets and the identities; it supplies no numerical mass, no dynamics and no choice of contour.
Summary of Notation
| notation | meaning |
|---|---|
| $\mathbb{B}$ | the biquaternion algebra $\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$ |
| $\tilde\Xi$ | a root |
| $\tilde Q$ | a radicand |
| $\langle\tilde Q\rangle_{\natural}$ | the natural norm $\sum_\mu Q_\mu^2$ |
| $\tilde P$ | a material four-momentum, $\tilde P=i(E/c)e_t+\vec p$ |
| $\mathbb{M}_-$ | the material sector, the anti-Hermitian subspace |
| $\mathbb{M}_+$ | the informational sector, the Hermitian subspace |
| $\tilde E$ | the mass field, a biquaternion |
| $\langle\bar\Psi\Psi\rangle$ | the chiral condensate |
| $\gamma_5$ | the chirality involution of the spinor module, $\gamma_5=\mathrm{diag}(-I_2,I_2)$ |
| $\varsigma$ | the Kustaanheimo–Stiefel coordinates |
| $\Rho=imc$ | the Hamilton modulus of a four-momentum, the mass |
| $\mathcal I(\lambda_+,\lambda_-)$ | the isospectral family of a pair of eigenvalues |
Further Reading
- The Physical Interpretation of the 12 Products, for the root sets of the twelve products and the existence classes.
- Biquaternion Elementary Functions, for the root table by regime and the logarithm the roots come from.
- Roots, Idempotents and Observables, for the roots of $+1$ and the idempotents as states.
- Roots, Idempotents and Quantum Information, for the idempotent structure of the qubit.
- The Zero-Divisor Cone and the Massless Spectrum, for the cone as the massless shell.
- The Material Sector: Four-Vectors, Interval and Lorentz Action, for the four-momentum and the interval.
- The Four Forms and the Four Algebraic Norms over the Biquaternion Algebra, for the norm as the mass shell.
- The Klein–Gordon Propagator and Its Green's Functions in Biquaternionic Form, for the pole and the four prescriptions.
- The Petiau System and the Quantisation of Mass in Biquaternionic Form, for the mass field and the eigenvalue equation.
- The Chiral Condensate and Dynamical Symmetry Breaking in Biquaternionic Form, for the condensate.
- The Coulomb Path Integral and the Duru–Kleinert Transformation in Biquaternionic Form, for the Kustaanheimo–Stiefel map.
- Conformal Invariance and the Massless Limit in Biquaternionic Form, for the branch point at $m=0$.
- Zitterbewegung in Biquaternionic Form, for the doubled phase of the mass shell.