I forgot to say something important. The action of non-conservative
(gas pressure-like) forces in a liquid dielectric is something
curious but, on the other hand, it is fatal for the present
electrostatic theory. Electrostatics is based on the explicit
assumption that all the acting forces are conservative. Needless to
say, when the basic assumption is wrong, the results are almost
certainly wrong as well. So I don't expect physics teachers to inform
students about the possible existence of non-conservative forces -
one does not question a theory when one has nothing to replace it
with. Still, in case someone really wants to create and then teach a
correct theory, I can give a clue. Perhaps the most important
difference between gas pressure-like forces and conservative forces
is that the former directly depend on the temperature whereas the
latter are either independent or their effect depends indirectly on
T - e.g. when the increased T destroys some polarization order. So,
if Panofsky's pressure is gas pressure-like, the T-dependence of the
dielectric constant in the Coulomb's law will be quite different from
the T-dependence of the dielectric constant decreasing the voltage.
Then the approach can be phenomenological - it is easy to deduce, for
a capacitor that is being immersed in a water pool, the equation
dA = -SdT + gdh + fdx
where A is the Helmholtz energy, S is the entropy, g is the net force
that pulls the capacitor downwards, h is the height of the capacitor
with respect to the ground, f is the force of attraction between the
plates and x is the distance between the plates. By cross-
differentiation, we obtain 3 Maxwell relations:
-dS/dh = dg/dT; -dS/dx = df/dT; dg/dx = df/dh
where all derivatives are partial, have physical meanings and are
measurable. If there were no non-conservative forces, the three
relations would be found, experimentally, to be equalities. However
the presence of non-conservative forces converts them into
inequalities and the difference between the two partial derivatives
in a Maxwell relation could give us a lot of information about the
mechanism of action of the non-conservative force. In other words,
instead of writing
-dS/dh = dg/dT
we could find, experimentally, the difference
-dS/dh - dg/dT,
then find its temperature dependence, interprete the findings etc. In
partucular, the difference
dg/dx - df/dh,
positive or negative, proves that the system can act as an isothermal
heat engine.