That's getting to the core of the question. Both morphology and liveliness are abstractions that we're trying to layer onto natural phenomena. I think the best

That's getting to the core of the question. Both morphology and liveliness are abstractions that we're trying to layer onto natural phenomena. I think the best thing to do is to recognize the abstractions as abstractions, and not try to take them too far. Here, we're asking the AI to reconcile these loose notions and come up with concrete answers. If a person were asked to do this task, they would turn right around and say "Your question sucks". Shades of liveliness helps, but we're still hindered by some arbitrary biases that are hampering thought. These biases are brought on by the fact that we are human, and we want to evaluate things as they relate to us. So we need to impart those biases onto our AI, and it sucks, because the biases don't make underlying sense.