Skip to main content

Gap Lagrangian models: a selected action and its scope

T-75. Dynamics and action

A phase action and a GKSL generator each require a domain, coefficients and an observation model. Relating their solutions requires a derivation from shared dynamics.

1. Coordinates and constraints​

In a fixed frame, γij=rijeiθij\gamma_{ij}=r_{ij}e^{i\theta_{ij}} for i<ji<j. A local phase chart needs rij>0r_{ij}>0; estimation stability needs an amplitude floor. Seven populations have six independent coordinates;21complex coherences have42real coordinates. Phases alone do not determine a state. Varying phases at fixed amplitudes can violate positivity, so the admissible domain must be checked.

Diagonal rephasing sends θij\theta_{ij} to θij+αi−αj\theta_{ij}+\alpha_i-\alpha_j. Triangle holonomy is invariant under this operation, but selected real-G2-frame observables need not share that gauge symmetry. G2 does not generally act on a fixed-amplitude phase torus as a closed state domain.

2. Explicit model ansatz​

Select a local phase domain, a constant symmetric positive definite matrix MM, smooth potential V(θ)V(\theta) and action

S[θ]=∫t0t1(12θ˙TMθ˙−V(θ))dt.S[\theta]=\int_{t_0}^{t_1}\left(\tfrac12\dot\theta^TM\dot\theta-V(\theta)\right)dt.

MM has units consistent with action and the selected time; coefficients do not follow from seven axes. The potential and its symmetries are separate inputs. Select a force Q=−Bθ˙+uQ=-B\dot\theta+u with symmetric B⪰0B\succeq0. Lagrange–d’Alembert gives

Mθ¨+∇V+Bθ˙=u.M\ddot\theta+\nabla V+B\dot\theta=u.

This follows exactly from the stated action and forces, not from arbitrary density evolution.

Choose a kinetic term, potential and dissipative dynamics. A positive mass matrix MM and well-defined generator are independent requirements. Additional Lindblad channels are admissible.

3. What a matrix ODE supplies​

For specified ρ˙=F(ρ,t)\dot\rho=F(\rho,t) where γij≠0\gamma_{ij}\ne0,

r˙ij=Re(e−iθijFij),θ˙ij=Im(Fij/γij).\dot r_{ij}=\mathrm{Re}(e^{-i\theta_{ij}}F_{ij}),\qquad\dot\theta_{ij}=\mathrm{Im}(F_{ij}/\gamma_{ij}).

Differentiate reiθr e^{i\theta} to obtain these identities. They give first-order dynamics. A second-order inertial equation needs velocity state or an independent closure; relabeling a dephasing rate as friction does not create it. At zero amplitude the phase and division are undefined.

4. Keldysh and the physical bridge​

Closed-time-path methods are tools for a specified microscopic model. Deriving an effective action needs degrees of freedom, initial state, reservoir interaction, eliminated variables and controlled approximations. The method itself does not select MM, a potential or UHM’s six terms. See the primary review Sieberer–Buchhold–Diehl.

The trace of the product of two Hermitian matrices is real. A nonzero Berry term therefore needs data such as an eigenline and its connection. Exact equivalence of the phase action and the generator would require matching trajectories and observables.

5. Symmetries and interpretation​

If specified M,V,B,uM,V,B,u respect a group, apply Noether’s theorem to the conservative part and compute charge changes from QQ. Equal projector dimensions do not supply intertwiners with irreducible G2 representations. An odd term of a selected potential does not prove universal PT breaking: transformation of the state and all parameters must be specified.

Habit, qualia and the direction of experienced time are interpretations [I]. Phase coordinates, physical energy units and experimental readout need independent calibration.

See variational derivation, Gap, Gap thermodynamics and Noether charges.