Since my previous post was motivated by irritation, I might as well continue the trend and post another.
Thixotropy is a real poster child for an abused rheological term. Beside the previously mentioned issues, there is an even more fundamental misunderstanding of the word - that thixotropic materials are MORE viscous than non-thixotropic materials.
The best I can gather is this: many thixotropic agents, such as silica, when added to a polymer create a mixture that is more viscous than the initial polymer was (in an unstressed state). Since an unstressed state is the condition that most people view them (unless you are prone to sticking your hand into an operating mixer) and since the first syllable in the word sounds like "thick", I can see that one could make the conclusion that "thixotropic" is fancy jargon for "thicker". But this is clearly wrong, since if you start to shear the material, the viscosity will quickly drop.
For filled materials - such as when silica is added - the polymer chains, running willy-nilly are adsorbed onto the surface of the filler. This makes it more difficult for the polymer to relax and flow when a stress is applied, since the filler is effectively an entanglement preventing a chain from moving. The drop in viscosity over time is associated with a breakdown in this network. The strength of the adsorbtive bonds is low enough that the stress applied by the shear can remove them from the surface of the filler so that they are freer to move relative to each other. i.e., flow.