dispersion relations and the analytic S-matrix

In physics, causes precede effects — a system cannot respond before it is poked. This innocent-sounding fact, causality, has a sharp mathematical shadow: the response function of any causal system, viewed as a function of frequency extended into the complex plane, must be holomorphic in a half-plane. Once that is known, complex analysis forces the real and imaginary parts of the response to be locked together. Dispersion relations are exactly those locking equations — they let you compute one measurable quantity (say absorption) from an integral over another (say refraction) across all frequencies.

The prototype is the Kramers-Kronig relations in optics. Because a medium's response is causal, the complex refractive index n(omega) is holomorphic in the upper half of the complex frequency plane and decays at infinity, so by Cauchy's integral formula with a closing arc its real and imaginary parts are Hilbert transforms of each other: the real part (refraction) at one frequency is a principal-value integral of the imaginary part (absorption) over all frequencies, and vice versa. In particle physics the same logic is applied to the scattering amplitude, the entries of the S-matrix. The analytic S-matrix program treats the amplitude as a holomorphic function of complex energy and momentum-transfer, whose singularities are not accidents but physics: poles on the real axis are bound states and resonances, and branch cuts are thresholds where new particles can be produced. Dispersion relations then relate the amplitude at one energy to an integral over its imaginary part (which the optical theorem ties to total cross-sections) elsewhere.

This is where complex analysis reaches into the frontier of theoretical physics. The same circle of ideas leads to Regge theory (treating angular momentum itself as a complex variable, so that families of particles trace out 'Regge trajectories'), and analyticity plus unitarity are load-bearing assumptions in the modern S-matrix bootstrap and in scattering amplitudes for high-energy physics. The honest framing: these relations are extraordinarily powerful precisely because they follow from very general principles — causality, unitarity, and assumed analyticity — rather than from any specific model, but that assumed analyticity (where exactly the amplitude is holomorphic, and how fast it grows) is a physical hypothesis that must be argued, not a theorem handed down for free.