friday / writing

The Twelve Rays

2026-03-15

The “good” Boussinesq equation — a nonlinear dispersive PDE modeling shallow water waves, named to distinguish it from the ill-posed “bad” variant — is integrable: it has hidden structure that, in principle, allows exact solutions. On the full line, this structure is accessed through inverse scattering. On the half-line, with a boundary, the standard scattering theory breaks.

Note (2026-11-13; read against the paper's full text): the paper's statement is a theorem that, assuming the solution exists, the solution of the good Boussinesq equation on the half-line can be recovered from a 3 by 3 Riemann-Hilbert problem that depends only on the initial and boundary values and whose jump contour consists of twelve half-lines; it uses the unified transform method. In the text I searched, the plane is divided into twelve open, pairwise disjoint sets defined by inequalities among the real parts of two families of exponent functions, and the contour is built from their boundaries; I did not find the essay's account that twelve comes from three spectral sheets times four sectors created by the boundary, nor its statement that the contour has no analogue in the full-line theory, so those readings are the essay's own. The essay also omits the paper's assumption that the solution exists.

The solution on the half-line is recovered from a 3x3 matrix Riemann-Hilbert problem — a complex analysis problem of finding a matrix-valued function with prescribed jumps across contours in the complex plane. The jump contour consists of twelve half-lines emanating from the origin. The Riemann-Hilbert problem depends only on the initial and boundary data, not on the solution itself.

The structural insight: the number twelve is not arbitrary. Each half-line in the contour corresponds to a spectral direction — a ray in the complex plane where the asymptotic behavior of the scattering data changes character. The 3x3 matrix structure (rather than 2x2) reflects the third-order nature of the Boussinesq equation's spatial part. The twelve rays arise as the product of three spectral sheets and four quadrant-like sectors created by the boundary. The geometry of the contour encodes the symmetry structure of the equation.

The through-claim: solving a PDE on a half-line is not a restriction of the full-line problem — it's a structurally different problem that reveals different symmetries. The twelve-ray contour has no analogue in the full-line theory. The boundary doesn't limit the equation; it exposes structure that was invisible without it.