Uniformity and directional symmetry are separate claims
Homogeneity means that the relevant statistical or geometric description does not prefer one location. Isotropy means that it does not prefer one direction. A single source or a single rule can generate a pattern with either property, both, or neither. The claim of common origin does not replace a proof of the relevant symmetry. [C04] [C07]
The laboratory gives an exact counterexample in a plane. Every location has the same positive quadratic response tensor Q=diag(1+ε,1−ε), with |ε|<1. Spatial translations do not change Q, so the model is homogeneous. Unless ε=0, its two principal directions respond differently, so it is anisotropic.
This tensor is an illustrative response law, not a measurement of cosmic light-speed anisotropy. If a particular wave equation used Q as its principal spatial coefficient, the directional quantity below could describe a phase speed. Without that additional equation, it is simply a well-defined positive directional readout.
Change the anisotropy without changing the angular mean
Move ε away from zero and rotate the direction marker. The angular plot displays the squared directional response and a flat isotropic reference. At ε=0 the result is independent of angle. Reversing ε exchanges the two principal axes.
Interactive calculation
Play scans Direction θ (degrees). Step advances it. No animation starts automatically.
Enable JavaScript for adjustable values. The static figure and worked calculation remain available.
For ε=0.6, the x direction has squared response 1.6 and the y direction has squared response 0.4. Their positive responses are about 1.2649 and 0.6325, a ratio of two. The displayed squared-response ratio is four. At 45° the squared response is exactly one, even though the model remains anisotropic.
The directional average of the squared response also remains one. Consequently, one mean value or a single measurement at 45° would miss the anisotropy. Measurements along several controlled directions reveal the difference that the average conceals.
Compare eigenvalues and rotated directions
The minimum and maximum squared responses are 1−|ε| and 1+|ε|. Positivity therefore requires |ε|<1 if every nonzero direction is to have a strictly positive response. The ratio of the largest to smallest positive response is √[(1+|ε|)/(1−|ε|)].
The trace of Q is fixed at two, so its angular mean is fixed. Its determinant is 1−ε², which records a change invisible to that trace. Equality of the eigenvalues, rather than the constancy of their sum, is what removes the preferred directions in this model.
A rotation of coordinate labels does not remove unequal eigenvalues. It changes the matrix representation but preserves their ratio. The polar graph of v(θ) should not automatically be called an elliptical wavefront: phase speed, group velocity, and a propagated wavefront are different constructions and depend on the specified wave equation.
A useful further distinction is scale. A microscopic lattice can have only a finite rotation symmetry while its leading long-wavelength quadratic term is isotropic. Higher-order corrections may retain directional dependence. An isotropic leading approximation is therefore not the same as exact rotation invariance at all resolutions.
What a generated cosmic geometry must establish
The rank-three reconstruction work addresses a conditional route to directional readout. Three independent directions alone do not prove that their calibrated metric coefficients are equal or that a large-scale matter distribution is homogeneous. Those are additional properties for the source dynamics and observation map to establish. [P01]
A constructive cosmology could test directional propagation, distance comparisons, and matter correlations separately, then show that their large-scale anisotropies vanish under a controlled limit. It should state the averaging ensemble, sampling rule, and error estimate. A finite observation window or a biased detector can also produce apparent anisotropy even when the underlying ensemble is symmetric.
The cosmological principle used in the following chapters is therefore an explicit baseline assumption. Its usefulness is that it reduces the geometry to a scale factor and a curvature parameter, making definite calculations possible. It does not follow merely from asserting that all events share one origin. [C08] [C04]
Continue to expansion. The original homogeneity-and-isotropy guide develops lattice examples, and Observation examines what an aggregate record can hide.

