I've heard this concern raised before, and the real question is whether the thermal expansion of the fuel salt acts as a strong enough negative coefficient of reactivity.
Answer, yes it does. This was proven in the Molten Salt Reactor Experiment.
"Since the neutron flux in the reactor is the main variable which controls the power output it seems, that a Thorium reactor is significantly harder to control than a Uranium one." This is nonsense. LFTRs are self controlling with a significant negative coefficient. It really makes no more sense to think about putting Thorium in a solid fuel reactor than it does about using diesel in a spark ignition engine.
[1] .6 delayed neutrons per 100 fissions for U233 compared to 1.6 for U235. http://www.reak.bme.hu/Wigner_Course/WignerManuals/Budapest/...
[2] Especially https://en.wikipedia.org/wiki/Liquid_fluoride_thorium_reacto...