It's more or less the same way putting salt on ice causes it to melt. Diffusion still occurs with solids, just at a lower rate than with liquids. As for wheth

It's more or less the same way putting salt on ice causes it to melt. Diffusion still occurs with solids, just at a lower rate than with liquids. As for whether it's liquid or gaseous, I probably shouldn't have called it "water", and should have said "hydrogen and oxygen in silicate minerals". Silicates always have oxygen, but some hydrogen too (hydroxyl groups - i.e. OH). Basically, you have minerals that transform to other minerals at depth due to the pressures and temperatures involve. That transformation leads to the release of hydrogen and oxygen as they're in one crystal structure and not in the other. (Basically, minerals with hydroxyl groups transform into minerals that do not have hydrogen in their composition, releasing hydrogen and oxygen in the process.) That hydrogen and oxygen diffuses into adjacent mineral crystals and can cause them to melt. With that said, any time you have magma (i.e. melt), it's going to have H2O, CO2, halogens, etc dissolved in it, just like water has oxygen and carbon dioxide dissolved in it. We talk about these in simple terms of "H2O" and "CO2" etc, but often the details of how things are bonded are a bit different, just like CO2 dissolved in water isn't exactly CO2, but is carbonic acid (H2CO3) instead.