The Fourier Transform and the Coupled Klein–Gordon Equation in Biquaternionic Form
Introduction
The coupled Klein–Gordon article states two equations side by side and leaves the choice between them open. The Coupled Klein–Gordon Equation in Biquaternionic Form writes (E1) $(\tilde\nabla_{\mathrm{gco},\Phi})^{\natural}\tilde\nabla_{\mathrm{gco},\Phi}\tilde\Xi=k_m^{2}\tilde\Xi$ and (E2) $\Box_{\mathrm{gco},\Phi}\tilde\Xi=k_m^{2}\tilde\Xi$, records that they coincide in the free limit and differ by the field strength, adopts E2, and puts E1's admissibility in its Open Questions. It also states, at §The Two Coupled Equations, that E2 is the equation of the corpus's action for "a connection whose coefficients are purely imaginary in every slot", and calls the electric slot's bookkeeping "the one addition". Those two sentences are the door this article walks through.
The transform turns the choice into a comparison of two polynomials. Wherever the connection is constant both equations have constant coefficients, so both have a symbol, and the symbols are central scalars differing by one term. Momentum space does not decide which equation the physics should use. It decides what each equation is, exactly, and it shows that the difference the parent defers to experiment is a difference between two real numbers rather than a difference between two kinds of number.
The finding is that the corpus's own writing is the reality condition. A symbol is real in a slot exactly when the connection's coefficient in that slot has the same reality as that slot's gradient symbol. The temporal gradient symbol is real and the three spatial ones are imaginary, so the condition is "the temporal coefficient real, the three spatial coefficients imaginary", and that is the anti-Hermitian writing $\tilde A_\Phi^{*}=-\tilde A_\Phi$ which the corpus already uses. The corpus's convention is therefore not a convention about signs. It is the condition for a real dispersion, and it is the unique assignment that satisfies the condition in all four slots.
The second finding is that the parent's two shells are correct, and that both are real. For a constant scalar potential the squared operator's roots are displaced by the same real number on both branches and the paired operator's are not displaced at all; the first is a linear shift of the energy, the second a shift of the mass squared. The two differ at first order in the potential, so the fork is a fork about the power of the response. It is not a fork about reality, and no pole in either reading leaves the real axis.
The complex number the path-integral companion measured is a coefficient of the wrong parity. The Path Integral and the Coupled Klein–Gordon Equation in Biquaternionic Form, at §The Quadratic Form and the Operator It Inverts, reports $-8.8876+1.86i$ for a real coefficient $0.31$ and $-9.0798$ for the paired form. Those three numbers are reproduced exactly by a real coefficient in a spatial slot, where the gradient symbol is imaginary: the real coefficient is the wrong parity for that slot, and the material connection cannot supply one, because its spatial coefficients are imaginary. The number is right and the reading of it was one slot off.
This is a clarification and the article says so once. There is no new solution, no new prediction and no new equation. The symbols are the standard ones with the corpus's signs, and what the article adds is the class statement, the corrected pair of shells and the diagnosis of the parent's complex number. The polar companion reads the same two symbols in the word, and the variational companion reads them from the action; this is the main article and the other two are register-specific readings of it.
Conventions
The algebra and its units are those of Algebra and Element Conventions, the gradient, the d'Alembertian and the symbol rule those of Operator and Calculus Conventions, the four products those of Product and Form Conventions, the two forms those of Norm and Pairing Conventions, the invertibles and the material time those of Physics Units and Common Elements Conventions. The transform is the one of Derivatives, Fields and Transforms, whose symbol rule is quoted in its derived form and not re-derived:
$$ \tilde\nabla_\Phi\longmapsto i\tilde K_\Phi, \qquad \Box_\Phi\longmapsto-\langle\tilde K_\Phi\rangle_\natural, \qquad \tilde K_\Phi=ik_t'e_t+\vec k_\Phi, \qquad k_t'=\frac{\omega}{c}, \qquad \langle\tilde K_\Phi\rangle_\natural=-k_t'^{2}+k_\Phi^{2}. $$
The temporal slot carries a real symbol, and that is the whole article. The coordinate is $R_t=i\,r_t'$, so $\partial/\partial R_t=-i\,\partial/\partial r_t'$ and the temporal slot of the gradient's symbol is $+i(ik_t')=-k_t'$, real; the three spatial slots are $i\kappa_\eta$, imaginary. Write $\sigma_\mu$ for those four numbers, so $\sigma_t=-k_t'$ and $\sigma_\eta=i\kappa_\eta$. A reader who takes the metric signs from the slot list instead of from $\partial/\partial R_t=-i\,\partial/\partial r_t'$ will put an extra $-1$ on $\sigma_t$ and will then find an imaginary cross term that does not exist. The corpus has already spent that sign twice, once in $R_t=ir_t'$ and once in $\Box_\Phi=\Delta_\Phi-\partial^{2}/\partial{r_t'}^{2}$, and there is none left to spend.
The connection is written anti-Hermitially, and the coefficient carries an index. With $c_\mu=\frac{i\,ϙ}{\hbar}A_\mu$ and $\tilde A_\Phi=i a_t' e_t+\sum_\eta A_\eta\vec e_\eta$ for real $a_t'$ and real $A_\eta$, the four coefficients are
$$ c_t=\frac{i\,ϙ}{\hbar}\,i a_t'=-\frac{ϙ a_t'}{\hbar}\ \text{real}, \qquad c_\eta=\frac{i\,ϙ}{\hbar}A_\eta=i\,a_\eta\ \text{imaginary}, \qquad a_\eta:=\frac{ϙ}{\hbar}A_\eta . $$
The coefficient $c_\mu$ always carries an index and the speed of light $c$ never does; the corpus's $r_t'=ct'$ is the only place the letter appears unindexed. The writing $A_t=ia_t'$ is the parent's and it is the one the anti-Hermiticity of $\tilde A_\Phi$ produces: $\tilde A_\Phi^{*}=-\tilde A_\Phi$ holds exactly when $a_t'$ and $\vec A$ are both real.
The covariant wave number and the three denials, stated once. The mechanical wave number is $\tilde K_{\Phi,\mathrm{gco}}=\tilde K_\Phi+\frac{ϙ}{\hbar}\tilde A_\Phi$, whose temporal slot is $-k_t'+c_t$ in the gradient's units. First, the transform carries $\mathbb{B}$-valued functions to $\mathbb{B}$-valued functions and never carries a material-sector field to an informational-sector one; the fork below is about classes of scalars and not about sectors. Second, a symbol exists only where the coefficients are constant, which is the licence used throughout and the first entry in Limits. Third, a Legendre transform between a field and its conjugate momentum grants no Fourier reading, and nothing below is one.
The Two Symbols
Both symbols are central scalars, and that is why the fork is decidable at all. The connection's coefficients are central and the field's carrier is not involved, so the symbol of either second-order operator is one complex number per wave number. For a constant connection with $\vec A=0$ the two are
$$ \text{(E2)}\qquad \bigl(k_t'-c_t\bigr)^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2}, \qquad\qquad \text{(E1)}\qquad k_t'^{2}-c_t^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2}. $$
The first is $\sum_\mu(\sigma_\mu+c_\mu)^{2}$, the square of the covariant derivative summed slot by slot. The second is $\sum_\mu(\sigma_\mu-\bar c_\mu)(\sigma_\mu+c_\mu)$, the two first-order symbols with opposite coefficients on the two factors. They differ by
$$ \bigl(k_t'-c_t\bigr)^{2}-k_t'^{2}+c_t^{2}=2c_t\bigl(c_t-k_t'\bigr), $$
which is real for every real $c_t$. The difference of the two symbols is a real number and it is first order in the potential; at the values used below it is $2.0522$.
The temporal slot is the only slot where the two operators differ. In a spatial slot $c_\eta$ is imaginary and $\bar c_\eta=-c_\eta$, so $(\sigma_\eta-\bar c_\eta)(\sigma_\eta+c_\eta)=(\sigma_\eta+c_\eta)^{2}$: the two readings coincide there identically, for every vector potential. A purely magnetic connection therefore gives one operator and not two, which is the parent's remark at §The Quadratic Form and the Operator It Inverts, read as the statement that the fork needs an electric slot. The vector potential enters both readings through the one combination $\vec\kappa+\vec a$, which is the mechanical wave number, and that absorption is the whole of the magnetic case.
The fork is therefore about a single slot and a single ordering. Everything below concerns the temporal slot with a constant scalar potential. The claim is not that the two operators differ everywhere; the claim is that where they differ they differ by a term that momentum space displays in one line and position space can only reach by an integration by parts.
The Parity Lemma
The lemma is one line of expansion. With $c_\mu$ constant, the symbol of the covariant d'Alembertian is real in slot $\mu$ exactly when $c_\mu$ has the reality of $\sigma_\mu$:
$$ \boxed{\; \bigl(\sigma_\mu+c_\mu\bigr)^{2}\in\mathbb{R} \quad\Longleftrightarrow\quad \operatorname{parity}\bigl(c_\mu\bigr)=\operatorname{parity}\bigl(\sigma_\mu\bigr) \;} $$
where parity means real or purely imaginary. The proof is the expansion: the square of $\sigma+c$ has imaginary part $2\,\mathrm{Re}(\sigma)\,\mathrm{Im}(c)+2\,\mathrm{Im}(\sigma)\,\mathrm{Re}(c)$, and a real $\sigma$ demands a real $c$ while an imaginary $\sigma$ demands an imaginary $c$.
Applied to the corpus's slots the lemma reads as one sentence about the connection. The temporal symbol is real and the three spatial symbols are imaginary, so the condition is "$c_t$ real and $c_\eta$ imaginary", which by the bookkeeping of §Conventions is "$A_t$ imaginary and $\vec A$ real", which is $\tilde A_\Phi^{*}=-\tilde A_\Phi$. The anti-Hermitian writing is the reality condition, and the corpus's convention is doing physical work that its own introductions describe as bookkeeping.
The lemma is measured, slot by slot. On a grid of $256$ points, with $k_t'=\lvert\vec\kappa\rvert=3$ and coefficients of magnitude $0.31$, the four cases are:
| slot | coefficient | squared symbol | paired symbol |
|---|---|---|---|
| temporal | real, $-0.31$ | $10.956100+0.000000i$ | $8.903900+0.000000i$ |
| temporal | imaginary, $0.31i$ | $8.903900-1.860000i$ | $8.903900-1.860000i$ |
| spatial | imaginary, $0.31i$ | $-10.956100+0.000000i$ | $-10.956100+0.000000i$ |
| spatial | real, $-0.31$ | $-8.903900-1.860000i$ | $-9.096100+0.000000i$ |
Two rows of that table are the article. The first row is the material writing and both symbols are real. The fourth row is a real coefficient in a spatial slot, the wrong parity, and the squared symbol is complex with imaginary part $-1.86=2\kappa c$; that row is the parent's number, as §The Parent's Three Numbers shows. The third row is the material vector potential, where both symbols are real and the two readings coincide.
The paired form is more forgiving than the squared form, and by exactly one slot. Expanding the paired symbol, $\sigma^{2}+\sigma(c-\bar c)-\lvert c\rvert^{2}$, the middle term is $2i\,\mathrm{Re}(\sigma)\,\mathrm{Im}(c)$: it vanishes for an imaginary $\sigma$, whatever $c$ is. So the paired symbol is real in every spatial slot for every complex coefficient, and its only condition is a real temporal coefficient; the squared symbol needs the condition in all four slots. Verified on $500$ random draws per class: with a real temporal coefficient and arbitrary complex spatial coefficients the paired symbol's largest imaginary part over the draws was $8.3\times10^{-13}$ while the squared symbol's was $34.78$; with both slots arbitrary the two were $17.04$ and $30.09$.
The lemma does not depend on the model and does not need the transform. It is an observation about the square of a complex number, and it would hold for any operator whose slots carry one real and three imaginary symbols. What the transform adds is the display: in position space the two symbols are two operators related by an integration by parts, and the parity of each coefficient is not visible.
The Fork Is a Class Distinction
The census is three rows and it settles the class question. Over $500$ random draws per class, with the temporal and one spatial slot combined, the largest imaginary part of each symbol is:
| class | squared | paired |
|---|---|---|
| material: $c_t$ real, $c_\eta$ imaginary | $8.1\times10^{-13}$ | $8.1\times10^{-13}$ |
| $c_t$ real, $c_\eta$ arbitrary complex | $34.78$ | $8.3\times10^{-13}$ |
| $c_t$ imaginary, $c_\eta$ imaginary | $19.69$ | $19.69$ |
| both slots arbitrary complex | $30.09$ | $17.04$ |
The first row is the corpus's writing and the residual is roundoff. Two independently computed symbols, one of them a composition of two operators, agree on being real to eight parts in $10^{13}$ over five hundred draws. The corpus's anti-Hermitian writing is not merely compatible with a real dispersion; it is the only writing in the table that delivers one for both operators at once.
The second row is the asymmetry, and it is the corrected form of a true statement. The paired symbol survives an arbitrary spatial coefficient; the squared symbol does not. So the honest version of "one symbol is real always and the other only conditionally" is: the paired symbol is real whenever the temporal coefficient is real, and the squared symbol is real only when every coefficient matches its slot. The polarity of the usual statement is reversed, and the reason is that the corpus's temporal symbol is real rather than imaginary.
The fork that remains is the ordering, and it is not a fork about reality. Since both symbols are real in the corpus's writing, no measurement of reality and no criterion of self-adjointness can separate them; both operators are self-adjoint there. What separates them is the cross term $2c_t(c_t-k_t')$, which is a difference of two real functions of the potential. A fork about the class of the symbol would be a fork about which numbers are real; this one is a fork about which real function the potential enters.
What the article does not do is choose. The census says what each symbol is. The action says which operator the corpus's own functional produces, and that is the variational companion's subject. Neither statement is a verdict on which equation the physics should use, and the article keeps the two apart deliberately.
The Parent's Three Numbers
The parent's measurement is exact, and its three numbers are one arithmetic identity. The Coupled Klein–Gordon Equation in Biquaternionic Form is not the article that prints $-8.8876+1.86i$; that display is in The Path Integral and the Coupled Klein–Gordon Equation in Biquaternionic Form, at §The Quadratic Form and the Operator It Inverts, and it reports a free value $-8.9837$, a squared value $-8.8876+1.86i$ and a paired value $-9.0798$. The three are related by one real coefficient $c=0.31$ and one imaginary part:
$$ -8.9837+0.31^{2}=-8.8876, \qquad -8.9837-0.31^{2}=-9.0798, \qquad 2\times3\times0.31=1.8600 . $$
All three hold to the printed digits. The parent's arithmetic is right and its numbers are internally consistent; the free value carries a common offset of $0.0163$ that its letters do not name, and the offset cancels in both differences.
Those three numbers are a real coefficient in a spatial slot. A real $c$ in a slot with $\sigma=i\kappa$ gives the squared symbol $-\kappa^{2}+c^{2}+2i\kappa c$ and the paired symbol $-\kappa^{2}-c^{2}$: the real part moves by $+c^{2}$, the imaginary part is $2\kappa c$, and the paired form moves by $-c^{2}$. The parent's three numbers have exactly that shape, with $\kappa=3$ and $c=0.31$. The complex symbol is the mark of a coefficient whose parity is wrong for its slot.
The material connection cannot supply such a coefficient. In the corpus's writing a real spatial coefficient would be $c_\eta$ real, that is $A_\eta$ imaginary, that is a non-material connection. The material potential has $A_\eta$ real and $c_\eta$ imaginary, and then the spatial symbol is $-(\kappa+a)^{2}$, real, with no imaginary part at all. So the parent's complex number is not a property of a scalar potential coupled to a scalar field. It is a property of a coefficient of the wrong parity, tested in a slot where the corpus never puts one.
What the parent's own text says, and what it does not. The same passage calls the imaginary coefficient "shifts the pair together" and the real one "separates them", and both phrases are right as descriptions of those two coefficients. The paragraph immediately before it is sharper still: it states that the two operators coincide exactly when $\bar c_\mu=-c_\mu$ in every slot, that the three spatial slots satisfy this for a real vector potential, and that the temporal slot does not, because $c_t$ is real. That paragraph has the slot structure right, and the measurement two paragraphs later uses a slot the paragraph rules out. The resolution is the one above: the number is a spatial-slot number, and the material connection does not produce one.
The correction is worth stating because the number is quoted. A reader who carries $-8.8876+1.86i$ away as "the symbol of a scalar potential" will look for a complex pole in a constant electric background and will not find one. The replacement is the pair of §The Two Shells: two real polynomials, no imaginary part, differing by a real cross term.
The Two Shells
The two shells are the parent's, they are real, and they differ by the power of the potential. Setting each symbol to zero with $\vec A=0$ gives
$$ \text{(E2)}\qquad \bigl(k_t'-c_t\bigr)^{2}=\lvert\vec\kappa\rvert^{2}+k_m^{2}, \qquad\qquad \text{(E1)}\qquad k_t'^{2}=\lvert\vec\kappa\rvert^{2}+k_m^{2}+c_t^{2}. $$
The first is the parent's $(\omega-c_t)^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2}=0$, the second its $\omega^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2}-c_t^{2}=0$, with $\omega$ standing for $k_t'$ in the parent's letters. Both displays in the parent are correct, and no transposition of a real and an imaginary part is needed to read them.
The first shell displaces both roots by the same real number. Its roots are $k_t'=\pm\sqrt{\lvert\vec\kappa\rvert^{2}+k_m^{2}}+c_t$: at $k_m=0.7$ and $\lvert\vec\kappa\rvert=3$ they are $+2.770584$ and $-3.390584$ against the free $\pm3.080584$. The pair translates by $c_t=-0.310000$ and its separation stays $6.161169$. A common real displacement of both branches is a shift of the energy, since $E=\hbar c\,k_t'$: $E\mapsto E+\hbar c\,c_t=E-ϙ c\,a_t'$, which is the parent's $E\mapsto E+ϙ c\,a_t'$ with the sign of $c_t$ carried through.
The second shell leaves the branches where they are and opens them. Its roots are $\pm\sqrt{\lvert\vec\kappa\rvert^{2}+k_m^{2}+c_t^{2}}$, here $\pm3.096143$: no displacement, and a separation of $6.192286$ against the free $6.161169$. A separation that grows with the square of the potential is a shift of the mass squared, $k_m^{2}\mapsto k_m^{2}+c_t^{2}$, and the mass shift is second order where the energy shift is first.
The two are distinguishable by the power, not by the sign of a complex part. At small $c_t$ the linear shell moves by $c_t$ and the quadratic shell by $c_t^{2}/2k_t'$: a factor $2k_t'/c_t$, which is twenty at the values above. That is the fork, stated as a measurement would meet it: a response linear in the potential against a response quadratic in it, in the same slot, from the same coupling.
Both shells are on the real axis and neither is a rate. Neither root pair has an imaginary part, so neither reading describes a state with a decay or growth rate. Whatever separates the two equations, it is not the existence of an asymptotic state, and an article that reported an off-axis pole here would be reporting the wrong-parity case of §The Parent's Three Numbers under the wrong name.
The Field Strength and the Three Benches
The field strength is a multiplication operator, and that fixes the ladder. With $\tilde F=i[D_\mu,D_\nu]$ and $D_\mu$ the per-slot covariant derivative, the commutator has no derivative left: it is pointwise multiplication by $\frac{iϙ}{\hbar}F_{\mu\nu}$. The transform of a multiplication is the convolution of the transform of its coefficient with the transform of the field, so the coefficient's own transform decides the bench.
The first bench is $\tilde F=0$ and the operator is a multiplier. A constant connection gives a symbol and the plane wave stays a plane wave. Measured on $256$ points: the covariant derivative of a plane wave of wave number $3$ occupies one mode, and the free case occupies one mode. The count is one because the coefficient is a constant and its transform is a single coefficient.
The second bench is $\tilde F$ constant and the operator is affine. An affine coefficient is not periodic, so the periodic test cannot carry it; on a line of $4096$ points with central differences, an affine $A_y=Bx$ and a Gaussian probe, the commutator identity $[D_x,D_y]g=iBg$ holds to $1.1\times10^{-5}$ against a term of scale $0.731$, which is the second-order scheme's own error. In the transform an affine coefficient is $i\,\mathrm{d}/\mathrm{d}k$, so a single wave number is carried to a derivative of a delta: still one point of support, now a jet at that point. The Heisenberg algebra of the covariant momenta, $[\tilde\pi_\mu,\tilde\pi_\nu]=i\hbarϙF_{\mu\nu}$ with a central constant on the right, is the reason the spectrum is a ladder there, and Landau Levels in Biquaternionic Form computes it; this article cites and does not recompute.
The third bench is $\tilde F$ varying and the operator is a convolution. The coefficient's transform is then a spread, the plane wave is carried into a band of modes, and there is no ladder: a convolution is an integral operator of no finite order and it has no eigenvalue problem in the algebraic sense at all. Measured with a single sinusoid in the coefficient the derivative of a plane wave occupies $3$ modes; with a random coefficient it occupies all $256$ above the threshold. The third bench is the end of the classification, and the article says so rather than reaching for an approximation to it.
The ladder is the parent's three cases and one criterion. The parent records the constant background at §A Constant Background, the Landau spectrum as a separate computation, and the operator setting as an open question. The criterion that joins them is the order of the coefficient's transform: a delta is a multiplier, a derivative of a delta is a first-order differential operator, and a function is a convolution.
The ladder does not depend on the parity lemma. The lemma decides whether a symbol is real; the ladder decides whether a symbol exists. A constant connection of the wrong parity has a complex symbol and is still a multiplier, and a varying connection of the right parity has a real symbol nowhere and is a convolution. The two classifications are independent, and the article states them as two because they are two.
The Prohibition in Symbol Language
The parent's prohibition on central fields becomes a statement about which elements a scalar symbol can be. The covariant square of a central field splits into the scalar d'Alembertian plus $\frac{iϙ}{\hbar}\xi\tilde F$, and the second term is a pure vector. A symbol is a central scalar, and a central scalar cannot absorb a vector part; so wherever $\tilde F\ne0$ the symbol of the covariant square is not the symbol of any scalar equation. The prohibition is a mismatch of types between the symbol and the curvature term.
The transform makes the mismatch visible in one line. The curvature term's symbol is $\frac{iϙ}{\hbar}$ times the transform of $\xi\tilde F$, which is a convolution with the transform of $\tilde F$; a scalar symbol is a single number at a single wave number. A convolution cannot equal a product, so no choice of the central field makes the equation hold in a region of nonzero curvature. That is the parent's conclusion with its reason attached.
What the prohibition does not say is anything about the fork. The prohibition concerns E1 in a varying field strength; the fork of §The Two Shells concerns a constant one, where $\tilde F=0$ and E1 and E2 coincide as operators. The two statements sit in different regimes, and confusing them would turn a constant-potential bookkeeping question into a curvature question it is not.
The Denial That Survives
The transform does not make the two sectors dual, and the fork is not about sectors. Every object above is a central scalar multiplying a $\mathbb{B}$-valued field, and the transform returns the same algebra. The four coefficients of $\tilde\Xi$ decouple because the operator is central, not because a sector has been exchanged. The fork is a fork about the class of a scalar and about the ordering of two factors, and neither is a statement about $\mathbb{B}_\Phi$ and $\mathbb{B}_\Psi$.
What survives is the licence, and it is the one thing the transform genuinely needs. A symbol exists where the coefficients are constant. Every display above is a constant-coefficient display, and the article does not claim a symbol where the connection varies — there the ladder's third bench applies and the correct object is a convolution.
Limits
The constant-coefficient licence bounds every statement here. The symbols of §The Two Symbols and the shells of §The Two Shells are constant-coefficient objects. For a varying scalar potential the operator is not a multiplier, the symbol is not a function of the wave number alone, and nothing in the article is a statement about that case.
The affine case is the first non-constant one and it is covered only by the ladder. A constant field strength makes the covariant derivative a first-order differential operator in the transform, whose eigenfunctions are not plane waves; the Landau ladder is cited and not derived.
The fork is a temporal-slot statement. The spatial slots carry no fork, because there $\bar c_\eta=-c_\eta$ and the two readings coincide identically. A vector potential therefore cannot separate the two equations at all, and an article that reported a magnetic fork would be reporting the parity lemma backwards.
Open Questions
Which of the three the corpus should drop is not decided here, and the corrected triplet is narrower than the usual one. The three assumptions are: the temporal coordinate $R_t=i\,r_t'$, the anti-Hermitian connection $\tilde A_\Phi^{*}=-\tilde A_\Phi$, and a real dispersion for a scalar potential. Any two hold and the third fails — dropping $R_t=ir_t'$ puts an extra $-i$ on $\sigma_t$ and turns the material $c_t$ into the wrong parity, dropping the anti-Hermitian writing does the same thing directly, and keeping both forces the dispersion to be real, which is what the corpus does. Whether the first two are negotiable is a question about the algebra's presentation and not about physics, and the article records the three without ranking them.
The paired operator's functional is not identified. Both operators of §The Two Symbols are self-adjoint in the corpus's writing, so both are the Euler–Lagrange operators of some real quadratic functional; the article exhibits the difference between the two functionals as the cross term $2c_t(c_t-k_t')$ and shows which one the corpus's action is, in the variational companion. What the paired functional is — whether it is a cross bilinear between two charge-conjugate factors, and whether it has a name in the standard literature — is open.
Whether the first-order response is measurable in the corpus's own settings is open. The linear shell moves the energy by $ϙc\,a_t'$ and the quadratic shell by $c_t^{2}/2k_t'$, and nothing in the article says which of the two a physical scalar in a constant potential follows. That is the parent's question, restated as a question about a power rather than about a real part, and it is not answered here.
Summary
The transform of the coupled Klein–Gordon operator is a central scalar per wave number, and the two readings of the covariant square give two of them:
$$ \bigl(k_t'-c_t\bigr)^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2} \qquad\text{and}\qquad k_t'^{2}-c_t^{2}-\lvert\vec\kappa\rvert^{2}-k_m^{2}, $$
differing by the real cross term $2c_t(c_t-k_t')$, which is $2.0522$ at the values used throughout.
A symbol is real in a slot exactly when the connection's coefficient has the reality of that slot's gradient symbol. The temporal symbol is real and the three spatial ones are imaginary, so the condition is the anti-Hermitian writing $\tilde A_\Phi^{*}=-\tilde A_\Phi$, and the corpus's convention is the reality condition rather than a convention about signs. Measured over $500$ random material connections, both symbols are real to $8.1\times10^{-13}$; over $500$ draws with a real temporal coefficient and arbitrary complex spatial coefficients the paired symbol stays real to $8.3\times10^{-13}$ while the squared symbol reaches an imaginary part of $34.78$.
The parent's shells are correct. The first displaces both roots by the same real $c_t$ and leaves their separation at $6.161169$: a linear shift of the energy, $E\mapsto E+ϙc\,a_t'$. The second leaves the roots where they are and opens the separation to $6.192286$: a mass-squared shift by $c_t^{2}$. The fork is a fork about the power of the response, not about the reality of a number, and neither reading has a pole off the real axis.
The parent's measured complex symbol, $-8.8876+1.86i$, is reproduced exactly by a real coefficient $0.31$ in a spatial slot, since $-8.9837+c^{2}=-8.8876$ and $2\kappa c=1.86$. That is a coefficient of the wrong parity, which the material connection never supplies. The field strength then classifies the operator's order: a delta is a multiplier, a derivative of a delta is first order, a function is a convolution — one, three and $256$ occupied modes in the three cases measured.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\tilde K_\Phi=ik_t'e_t+\vec k_\Phi$ | The four-wave number; $k_t'=\omega/c$ |
| $\sigma_\mu$ | The free slot symbol: $\sigma_t=-k_t'$ real, $\sigma_\eta=i\kappa_\eta$ imaginary |
| $\tilde A_\Phi=i a_t'e_t+\sum_\eta A_\eta\vec e_\eta$ | The connection, anti-Hermitian in the material writing |
| $c_\mu=\frac{iϙ}{\hbar}A_\mu$ | The per-slot connection coefficient; $c_t=-ϙa_t'/\hbar$ real, $c_\eta=ia_\eta$ imaginary |
| $a_\eta=\frac{ϙ}{\hbar}A_\eta$ | The reduced vector potential, real |
| $k_m=mc/\hbar$ | The mass wave number |
| (E2), the squared reading | $\sum_\mu(\sigma_\mu+c_\mu)^{2}$; shell $(k_t'-c_t)^{2}=\lvert\vec\kappa\rvert^{2}+k_m^{2}$ |
| (E1), the paired reading | $\sum_\mu(\sigma_\mu-\bar c_\mu)(\sigma_\mu+c_\mu)$; shell $k_t'^{2}=\lvert\vec\kappa\rvert^{2}+k_m^{2}+c_t^{2}$ |
| $2c_t(c_t-k_t')$ | The difference of the two symbols; real, first order in the potential |
| $\tilde F=i[D_\mu,D_\nu]$ | The field strength; a multiplication by $\frac{iϙ}{\hbar}F_{\mu\nu}$ |
| $\tilde K_{\Phi,\mathrm{gco}}$ | The mechanical wave number, $\tilde K_\Phi+\frac{ϙ}{\hbar}\tilde A_\Phi$ |
Further Reading
- The Coupled Klein–Gordon Equation in Biquaternionic Form, for the two equations, the anti-Hermitian writing, the covariant square and the prohibition.
- The Fourier Transform and the Coupled Klein–Gordon Equation in Biquaternionic Polar Variables, for the same two symbols read in the word and for the ordering term the chart cannot write.
- The Lagrangian and the Coupled Klein–Gordon Equation in Momentum Space, for the action's Euler–Lagrange operator and the reading of the fork from the functional.
- The Gauge Principle in Biquaternionic Form, for the localization of the central phase and the sign of the coupling.
- The Covariant Derivative and Gauge Connection in Biquaternionic Form, for $\tilde\nabla_{\mathrm{gco},\Phi}$, the commutator and $\tilde F$.
- Landau Levels in Biquaternionic Form, for the ladder that a constant field strength produces.
- Derivatives, Fields and Transforms, for the symbol rule and the convolution theorem.
- Operator and Calculus Conventions, for $\partial/\partial R_t=-i\,\partial/\partial r_t'$ and the sign of the d'Alembertian.
- R. N. Bracewell, The Fourier Transform and Its Applications (McGraw-Hill, 3rd ed., 1999), for the convolution theorem and the transform of a product.
- C. Itzykson and J.-B. Zuber, Quantum Field Theory (McGraw-Hill, 1980), Chapter 2, for the minimal substitution and the constant-potential shift in the standard language.