The Effective Optical Metric of the Local Complex Structure
Introduction
The local complex structure of the framework identifies the temporal direction with the imaginary scalar generator, $r'_t\mapsto i r'_t\,e_t$, with $c$ the local speed of light in the medium. It induces at each point the quadratic form $$ \langle d\tilde{Q}\rangle_{\natural}=(i\,dr'_t)^2+dr_x^2+dr_y^2+dr_z^2=-d{r'_{t}}^2+dr_x^2+dr_y^2+dr_z^2, $$ of inertia $(3,1)$, so the framework reads a Lorentzian metric off the algebra rather than postulating it. The form is the unit form: in the length coordinate the temporal coefficient is $1$ and the cone is the $45^\circ$ cone. The medium enters through the conversion $t'=r'_t/c$ with $c=1/\sqrt{\epsilon\mu}$, and in seconds the same form is $-c^2\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2$ (The Local Complex Structure and the Speed of Light).
That construction is announced as a metric, and the word invites a specific identification that the parent articles leave open. A medium with refractive index $n$ already has a well-known effective geometry: the optical metric of Gordon (1923), in which light rays are null geodesics and the medium's index and motion are absorbed into the geometry. The parent article states the caution explicitly: the local structure of the framework "is isotropic and is written in the local rest frame of the material; the constitutive relations of a moving or anisotropic medium carry more data than the local complex structure, and their geometric interpretation is a separate question" (Electromagnetism in Media — The Local Complex Structure at Work).
This article takes up that separate question. The result is short and sharp: in a static medium the framework's local metric is exactly the optical metric, to machine precision and with no extra assumption, and the two independent boundaries — motion and anisotropy — show exactly where the framework's single scalar $c$ runs out. A third boundary is more interesting than either: when $\epsilon\mu$ changes sign the interval changes its sign pattern, from one minus on the time to none, and that change is the Wick rotation of the local structure.
Throughout, the algebra is $\mathbb{B}=\mathbb{C}\otimes_\mathbb{R}\mathbb{H}$ with basis $e_t=1,\vec e_x,\vec e_y,\vec e_z$, $\vec e_\eta^2=-e_t$, and central $i$. The material sector is $\mathbb{B}_{\Phi}$ with $\tilde R_\Phi=i r'_{t}\,e_t+r_x\,\vec e_x+r_y\,\vec e_y+r_z\,\vec e_z$ and norm $\langle\tilde R_\Phi\rangle_{\natural}=\tilde R_\Phi\tilde{R}_\Phi^{\natural}=\sum_\mu Q_\mu^2$. The framework's local metric is the unit metric $\delta_{\mu\nu}=(1,1,1,1)$ on the four $Q_\mu$, and its interval is $(i\,dr'_t)^2+dr_x^2+dr_y^2+dr_z^2=-d{r'_{t}}^2+dr_x^2+dr_y^2+dr_z^2$, whose one minus is the square of the $i$ in $Q_t=i q'_t$. Read in seconds, with $r'_t=c\,t'$, it is $-c^2\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2$. The medium speed is $c=1/\sqrt{\epsilon\mu}$, the vacuum speed $c_0=1/\sqrt{\epsilon_0\mu_0}$, and the refractive index $n=c_0/c$.
The Optical Metric of a Medium
Gordon's construction
Maxwell's equations in a medium with permittivity $\epsilon$ and permeability $\mu$ have the same form as in vacuum with a different metric. Gordon's observation is that for a dielectric of index $n$ moving with four-velocity $v_\mu$, the field equations in the medium are equivalent to the vacuum field equations in the effective metric $$ g^{\mathrm{J}}_{\mu\nu}=\delta_{\mu\nu}+\left(1-\frac{1}{n^2}\right)v_\mu v_\nu , \qquad \sum_\mu v_\mu v_\mu=-1 . $$ Here $\delta_{\mu\nu}=(1,1,1,1)$ is the unit metric on the four $Q_\mu$, and the metric is a quadratic form in those, $\sum_{\mu,\nu}g^{\mathrm J}_{\mu\nu}Q_\mu Q_\nu$; the intervals are written below in the real coordinates $(r'_t,r_x,r_y,r_z)$, with $Q_t=i q'_t$ and $c=c_0=1$ in this section. The light rays of the medium are the null geodesics of $g^{\mathrm{J}}$, and the medium's index and velocity are read off from the geometry. The two features that matter here are the conformal factor attached to the velocity direction and the off-diagonal drag terms produced when $v_i\neq0$.
The static case
For a medium at rest, $v_\mu=(i,0,0,0)$ and $v_tv_t=-1$, so the interval of $g^{\mathrm J}$ is $$ -\frac{1}{n^2}\,d{r'_{t}}^2+dr_x^2+dr_y^2+dr_z^2 . $$ In units $c_0=1$, its null cone is $$ -\frac{d{r'_{t}}^2}{n^2}+dr_x^2+dr_y^2+dr_z^2=0 \;\Longleftrightarrow\; \left|\frac{d(r_x\,\vec e_x+r_y\,\vec e_y+r_z\,\vec e_z)}{dr'_t}\right|=\frac{1}{n}=c , $$ so the coordinate speed of light in the medium is $1/n$: the optical metric reproduces the medium's light speed by construction. This is the standard object, and nothing about it is in question.
The Framework's Local Metric as an Optical Metric
The exact agreement
The interval of the framework's metric is $-c^2\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2$ with $c=1/\sqrt{\epsilon\mu}$ and $t'=r'_t/c$, and the interval of the static Gordon metric is $-\frac{c_0^2}{n^2}d{t'}^2+dr_x^2+dr_y^2+dr_z^2$. With the index $n=c_0/c$, so $c_0^2/n^2=c^2$, the two intervals are $$ \boxed{\;-c^2\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2=-\frac{c_0^2}{n^2}\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2\;}. $$
The two metrics are not merely similar or conformally related; in the seconds coordinate they are the same quadratic form, and the agreement was checked numerically for $n=1.00$, $1.50$, $2.00$ and $3.30$, to $10^{-14}$ or better. The identification of the framework's scale with the index is exact: $$ \boxed{\;n=\frac{c_0}{c}=\frac{1}{c}\;}\qquad(c_0=1),\qquad c=\frac{1}{n}. $$
The reason is structural and not a coincidence. Gordon's conformal factor in the static case is $1/n^2=1/(\epsilon\mu)$ in vacuum units, and the framework's coefficient in the $d{t'}^2$ slot is $c^2=1/(\epsilon\mu)$; the two expressions are the same, so they coincide once $n$ is identified with $c_0/c$. The framework's algebra supplies the sign — the minus is $(i)^2$ — and the medium supplies the magnitude through the conversion, which is exactly the division of labour that The Local Complex Structure and the Speed of Light states. In the length coordinate the framework's form carries no coefficient at all, $-d{r'_{t}}^2+d\vec r^2$, and the medium is invisible in it; the correspondence is one of geometry, read in seconds, and not of one shared coordinate.
The null cone
The null cone of the framework's form is the zero-divisor cone of the algebra at the point (The Material Sector: Four-Vectors, Interval and Lorentz Action). In the length coordinate that cone is the fixed $45^\circ$ cone, $\lvert d\vec r/dr'_t\rvert=1$, and it carries no $c$: it is the cone of the algebra at every point of every medium. Read in seconds it has aperture $c$. The null cone of the optical metric has aperture $1/n$ in vacuum units. The two apertures are one number, $c=c_0/n$, which is the statement that the seconds reading of the algebraic cone and the optical cone are one cone. For $n=1.50$ and $c_0=1$ the aperture is $0.666667$ on both sides.
Three Boundaries
The moving medium and the drag that the algebra does not supply
Gordon's metric for a moving medium has off-diagonal terms. For a medium of index $n=1.5$ moving at $v=0.3\,c_0$ the computed off-diagonal entries are $$ g^{\mathrm{J}}_{ti}=(-0.183150,\,0,\,0)\qquad(i=x,y,z), $$ and these are the terms responsible for the Fresnel drag of light in moving matter (Fizeau, 1851). The framework's local metric has $g^{\mathrm{J}}_{ti}=0$ identically: the local complex structure is defined in the medium's rest frame and rescales the imaginary axis by the real positive scale $c$ without rotating or shifting it, so it cannot produce a cross term.
The boundary is therefore exact, and it is the sharper form of the caution in the parent article: the framework's local complex structure supplies the conformal half of the optical metric and not the drag half. A moving medium carries one more datum than the structure does, the four-velocity $v_\mu$ of the material, and the framework must import it. This is not a defect of the construction — the rest-frame metric is the object the algebra can define — but it is the precise sense in which the framework does not derive the optical metric of a moving dielectric. The corpus's own treatment of Fizeau drag belongs to The Lorentz Transformation as a Biquaternionic Rotation, where the drag is a kinematic consequence of the boost and not an algebraic one.
Anisotropy and birefringence: one scale cannot carry two indices
A birefringent medium has two refractive indices. The corpus's chiral medium is exactly of this kind: the Drude–Born–Fedorov relations give two indices $n_\pm(\omega)$, and the two circular polarisations propagate with different wave numbers (Electromagnetism in Media — The Local Complex Structure at Work). The framework's local complex structure carries one scalar $c$, hence one index, hence one optical metric.
The same loss appears for an anisotropic crystal, whose dielectric tensor $\epsilon_{ij}$ has three principal values: no single conformal factor reproduces a metric whose spatial part is not isotropic. The local complex structure is a rescaling of the time axis, so it can always be written as a conformal factor of the time direction and never as an anisotropy of the spatial directions. The interpretation is therefore one-metric optics: it covers the isotropic static medium and nothing beyond it, and the two-index case is owned by the chiral article rather than by a metric.
The sign pattern: $\epsilon\mu<0$ is a Wick rotation of the local structure
The conversion of the local complex structure is $c^2=1/(\epsilon\mu)$, and its sign is the sign of $\epsilon\mu$. Two regimes follow, and the difference is a change of the sign pattern of the interval read in seconds:
| Regime | $\epsilon\mu$ | $c^2$ | the interval $-c^2\,d{t'}^2+dr_x^2+dr_y^2+dr_z^2$ | Inertia |
|---|---|---|---|---|
| Transparent dielectric | $>0$ | $>0$ | one minus on $t'^2$; spatial part positive | $(3,1)$ |
| Plasma or metal below cutoff | $<0$ | $<0$ | no minus, all four terms positive | $(4,0)$ |
In an ideal plasma with $\epsilon(\omega)=\epsilon_0(1-\omega_p^2/\omega^2)$ the permittivity is negative below the plasma frequency, and at $\omega=\omega_p/\sqrt2$ one has $\epsilon_r=-1$ exactly. There $c^2<0$, the index $n=1/c$ is imaginary, the wave is evanescent, and the interval has inertia $(4,0)$: there is no real null cone, which is the geometric statement that no propagating wave exists below cutoff.
The interpretive content is exact. With $c$ purely imaginary, $c=i\lvert c\rvert$, the temporal coordinate of the material sector becomes $$ i r'_t=i c\,t'=i\,(i\lvert c\rvert)\,t'=-\lvert c\rvert\,t' , $$ real. The imaginary time direction has turned into a real one, which is precisely what the Wick rotation does: the framework's Euclidean continuation is the transfer from $\mathbb{B}_{\Phi}$ to the real-quaternion subspace $\mathbb{B}_{H}$ (Interpretation of the Wick Rotation in the Biquaternion Universe), and the same transfer occurs here when the medium's $\epsilon\mu$ changes sign. The corpus notes the boundary for a complex $c$ near an absorption line — the rescaling becomes a complex rotation and the scale can no longer be read as the aperture of a real null cone (The Local Complex Structure and the Speed of Light). The dielectric-to-plasma crossing is the clean lossless case of that boundary: it is the slice on which $c$ is exactly imaginary, and it is the local structure's Wick rotation.
What the Reading Does and Does Not Add
What it adds. The identification of the framework's local metric with Gordon's optical metric, exact in the static case; the interpretation of the null cone of the algebra as the optical-metric cone; and the three boundaries, of which the change of the sign pattern is the substantive one, since it connects the medium's dielectric sign to the framework's own Euclidean continuation.
What it does not add. No new prediction: the optical metric is standard, and the framework's contribution is a repackaging of it in which the conformal factor is the algebra's own scale. In particular the interpretation does not deliver gravity. For a constant medium the interval $-d{r'_{t}}^2+dr_x^2+dr_y^2+dr_z^2$, which in the clock time is $-c^2d{t'}^2+dr_x^2+dr_y^2+dr_z^2$, is the flat Minkowski interval in the complex time coordinate $Q_t=i q'_t$, so it carries no curvature; and for a varying one the framework's class of metrics is the class with flat spatial slices, which admits no non-flat vacuum geometry — no Weyl curvature, no gravitational waves, no black-hole exteriors (Curved Spacetime and the Biquaternion Framework). The optical metric is a causal structure, not a gravitational one, and this article claims no more.
Physical Readings
The local complex structure reads as an optical metric: the medium's refractive index is the scale of the algebra's imaginary-time axis, $n=1/c$, and the light cone of the algebra is the cone in which the medium's light propagates. Read on the medium, the framework's scale is the conformal factor of the optical geometry, so the interpretation is the algebraic form of the analogue-gravity statement that the wave cone is a property of the medium and not of a fixed background (Curved Spacetime and the Biquaternion Framework). Read on the sign of $\epsilon\mu$, the local structure has two phases: a Lorentzian phase of propagating waves and a Euclidean phase below the plasma cutoff in which the imaginary time direction has become real, so the medium's dielectric sign selects between the Lorentzian and the Wick-rotated interpretations of the same algebra (Interpretation of the Wick Rotation in the Biquaternion Universe). Two boundaries belong with the interpretation: the framework supplies the conformal factor and not the drag, so a moving medium needs its four-velocity as an imported datum, and the framework supplies one index and not two, so birefringence is outside the metric interpretation (Electromagnetism in Media — The Local Complex Structure at Work).
Read further, the conformal factor is a dilaton shape: the scene at each point carries a scale $\log c$, and a change of the local complex structure is a change of that scale rather than of the algebra, so the framework's version of the scalar partner of a metric is the logarithm of the local light speed and not a new field (Algebra and Element Conventions).
Summary
The framework's local complex structure produces at each point the unit interval $-d{r'_{t}}^2+dr_x^2+dr_y^2+dr_z^2$, which in the clock time $t'=r'_t/c$ reads $-c^2d{t'}^2+dr_x^2+dr_y^2+dr_z^2$ with $c=1/\sqrt{\epsilon\mu}$. This article identifies it with the optical metric of the medium: in the static case $g^{\mathrm J}$ is exactly Gordon's metric $g^{\mathrm{J}}_{\mu\nu}=\delta_{\mu\nu}+(1-1/n^2)v_\mu v_\nu$ with $v$ at rest, the identification being $n=c_0/c$, and and the null cone of the algebra, read in the clock time, is the optical-metric cone of aperture $c=1/n$.
Three boundaries mark where the framework's single scalar runs out, and one of them is more than a boundary. The moving medium has off-diagonal drag terms $g^{\mathrm{J}}_{ti}\neq0$ that the framework's metric cannot carry, because the local structure rescales the imaginary axis without rotating it; the four-velocity must be imported. A birefringent or anisotropic medium has two indices, and one scalar cannot carry them, so those cases belong to the chiral article and not to the metric interpretation. And $\epsilon\mu<0$ — a plasma or a metal below cutoff — flips the sign of $c^2$, turns the metric Riemannian of inertia $(4,0)$, and turns the imaginary time direction real: the dielectric sign selects between the Lorentzian and the Wick-rotated phases of the local complex structure.
The interpretation adds no prediction and no gravity. It is a repackaging of Gordon's optical metric in which the conformal factor is the scale of the algebra, and its content is that the framework's light cone is the medium's light cone, its one free number is the medium's index, and its Euclidean phase is the medium's $\epsilon\mu<0$.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\mathbb{B}=\mathbb{C}\otimes_\mathbb{R}\mathbb{H}$ | Biquaternion algebra |
| $e_t=1,\vec e_x,\vec e_y,\vec e_z$ | Quaternion basis, $\vec e_\eta^2=-e_t$ |
| $i$ | Central scalar imaginary, $i^2=-1$ |
| $\mathbb{B}_{\Phi}$ | Material sector, $\tilde R_\Phi=i r'_{t}\,e_t+r_x\,\vec e_x+r_y\,\vec e_y+r_z\,\vec e_z$ |
| $\langle\tilde{Q}\rangle_{\natural}=\tilde{Q}\tilde{Q}^{\natural}=\sum_\mu Q_\mu^2$ | Biquaternion norm; null cone $N=0$ |
| $c=1/\sqrt{\epsilon\mu}$, $c_0=1/\sqrt{\epsilon_0\mu_0}$ | Medium and vacuum speeds |
| $n=c_0/c$ | Refractive index |
| $\delta_{\mu\nu}=(1,1,1,1)$, $Q_t=i q'_t$ | The one metric; the framework's interval is $-{r'_{t}}^2+r_x^2+r_y^2+r_z^2$ |
| $g^{\mathrm{J}}_{\mu\nu}=\delta_{\mu\nu}+(1-1/n^2)v_\mu v_\nu$, $\sum_\mu v_\mu v_\mu=-1$ | Gordon's optical metric |
| $g^{\mathrm{J}}_{ti}$ | Drag terms; zero for a static medium, nonzero for a moving one |
| $\epsilon(\omega)=\epsilon_0(1-\omega_p^2/\omega^2)$ | Ideal plasma permittivity; negative below $\omega_p$ |
| $\epsilon\mu<0$, $g^{\mathrm J}$ of inertia $(4,0)$ | Euclidean phase; $c$ imaginary, imaginary time real |
| $n_\pm(\omega)$ | Two indices of a chiral (birefringent) medium |
Further Reading
- W. Gordon, "Zur Lichtfortpflanzung nach der Relativitätstheorie," Annalen der Physik 72 (1923) 421–456, for the effective metric of a dielectric medium.
- Ulf Leonhardt and Thomas G. Philbin, "General relativity in electrical engineering," New Journal of Physics 8 (2006) 247, for the modern transformation-optics interpretation of a medium's effective geometry.
- W. G. Unruh, "Experimental black-hole evaporation?," Physical Review Letters 46 (1981) 1351–1353, for the analogue-gravity programme in which a moving medium's effective metric mimics a gravitational field.
- L. D. Landau and E. M. Lifshitz, Electrodynamics of Continuous Media (Pergamon, 2nd ed., 1984), for the optical properties of a plasma below the cutoff and for birefringence.
- H. Fizeau, "Sur les hypothèses relatives à l'éther lumineux," Comptes Rendus de l'Académie des Sciences 33 (1851) 349–355, for the drag of light in moving water.
- The companion articles: The Local Complex Structure and the Speed of Light, Electromagnetism in Media — The Local Complex Structure at Work, Curved Spacetime and the Biquaternion Framework, Interpretation of the Wick Rotation in the Biquaternion Universe, and Conventions in the Biquaternion Universe.