Dielectric polarization in nonconductors has mostly two causes, namely electronic polarizability (electrons shifting within molecules) and orientational polarizability (dipolar molecules rotating or flipping, or ions changing places). The first effect has a rather weak temperature dependence, the second a rather pronounced one.

In a *polar* medium, high temperature disturbs the alignment of dipoles in an outer field: the dielectric constant decreases with temperature. In a *ferroelectric* medium, the dipoles are already aligned and, unless the outer field is very strong, thus do not contribute to the orientational polarizability. High temperature disturbs the ferroelectric alignment and produces molecules that are free to reorient in an external field, and these contribute to the dielectric constant.

Well, this is a kind of very simplistic explanation, but it may give you an idea which forces are at work.

Dielectric polarization in nonconductors has mostly two causes, namely electronic polarizability (electrons shifting within molecules) and orientational polarizability (dipolar molecules rotating or flipping, or ions changing places). The first effect has a rather weak temperature dependence, the second a rather pronounced one.

In a *polar* medium, high temperature disturbs the alignment of dipoles in an outer field: the dielectric constant decreases with temperature. In a *ferroelectric* medium, the dipoles are already aligned and, unless the outer field is very strong, thus do not contribute to the orientational polarizability. High temperature disturbs the ferroelectric alignment and produces molecules that are free to reorient in an external field, and these contribute to the dielectric constant.

Well, this is a kind of very simplistic explanation, but it may give you an idea which forces are at work.

You had a microscopic explanation already given, let me add a thermodynamic view. The dielectric constant is a susceptibility. For a ferroelectric the polarization is the order parameter and the electric field is the conjugate field. On approaching the phase transition fluctuations increase which is related to an increase in the susceptibility. You mentioned magnetism and indeed this is very similar here where the magnetic susceptibility increases on approaching the critical temperature from below and above. In a ferromagnet the order parameter is the magnetization and the conjugate field is the magnetic field. Of course one has to be careful. If I refer for example to an Ising model with ferromagnetic interaction I have to add the spatial dimension. In one dimension the Ising model has no finite temperature phase transition. Perhaps this is what got you confused? If you look to the 2D Ising ferromagnet you find a phase transition with critical behavior and thus a diverging susceptibility. Briefly back to the ferroelectric and the microscopic reason for an increase in the response function on approaching the transition. In one class of ferroelectrics the transition is the result of the softening of a vibrational mode. The system changes its structure, e.g., from cubic to tetragonal in BaTiO3. This is reflected in the phonon spectrum by a mode having a frequency which approaches zero on approaching the transition. There is a celebrated relation called the Lyddane Sachs Teller relations which connects the mode softening with the divergence of the susceptibility.