I think "understanding" is not well defined often and this causes confusion. Some people mean the phenomenology of it; what it feels like subjectively. For me, this is unnecessarily ambiguous. Understanding can perhaps be better described in terms of evidence for its existence. If an agent appears to confidently and deftly move. If it demonstrates dexterity, for instance, we may conclude that the ensemble of units (not necessarily biological neurons) activating in concert represent understanding. We know subjectively that we can have both conscious and unconscious understanding using this definition. When you first are learning piano, any progress made is felt subjectively; you feel that you understand something. On the unconscious side of it, you are constantly refining motor skill behind the scenes. So, we know understanding should probably be defined as a process wholly independent of consciousness. A grasshopper may not be conscious, but could reasonably be said to "understand" how to chew and eat a leaf. What do you think?