I wrote that post.
Take a crystal sitting on your desk. More often than not, the inter-lattice electric potential (Leonard-Jones) between the molecules on each lattice site sums to 0 in each unit cell,* which means that the electric fields in the whole crystal sum to 0 and each of the unit cells are in electrostatic equilibrium with each other.
If the inclusion do indeed exert electric fields on the crystal lattice (so that they can change the lattice spacing and move the gems), they must do so in a manner that keeps the sum of the electric field 0 over the whole gem. When the gem is not moving, then, this implies that they must first satisfy the NCA, namely "settle down". Otherwise the gems would accumulate charge on their head, or worse, break apart spontaneously. The former would explain why Phos's doritos deem to stick on to his head actually, since, over the course of several years when he's sleeping and the Phos + Lapis inclusions are settling down, the out-of-equilibrium electric fields accumulated charge on his head so that they're able to attract the doritos.
* This is actually the neutral-charge approximation (NCA), which is true for most stable crystals. It goes NCA ? electrostatic equilibrium but we (i.e. condensed matter theorists) assume the converse for convenience.