That's a very uninteresting and unscientific definition of truth. But even less interesting is arguing over definitions.
As a scientific person, I'm only concerned with truth that derives directly from observation. "True but not demonstrable" is an oxymoron.
It's provable that things that are true are not demonstrable(see Gödel's incompleteness theorem). It follows that things that are not testable are true, and for some d definition of truth, it means they have an effect in the word. So we absolutely can, in a meaningful way, wonder about the posible consequences of something that while not testable may be true. Just because something it solely potentially meaningful does not mean it it's entirely meaningless.