People who claim non-existence are, by default, gnostics, since it is impossible to logically demonstrate non-existence. They must assert non-existence through some magical knowledge. If they do not claim such then they are agnostic by default. Agnosticism actually doesn't specify a position on the existence question, it is a qualifier for one's epistemological position. However, logic suggests that you start with an empty set and only add statements whose truth can be verified. So, barring evidence, I begin without any existence statements and only add agents to my model when they can be demonstrated.
Having a working hypothesis about Harvey the Rabbit just strikes me as silly. I can propose a never ending list of entities that cannot be detected by any test. Will you fill your models with non-existence clauses for each of this infinite set? Have I not exercised power over you in making you assign importance to a word of my choosing and? Have I not forced ideas into your model that do not help to explain anything that you observe? Thanks, but I prefer to keep my slate clean and deal with ideas that I find useful rather than go around arguing with people about words that appear meaningless.
I think you mean homonyms.
I think you mean homophones.
It's not so hard to lift yourself by your bootstraps once you're off the ground. -- Daniel B. Luten