Yes, I agree, scientific theories have their value from being new/confabulated. Though this isn't mutually exclusive from solving today's problems of confabulations. The proposed methodology could be used to mark semantic coherence, but it doesn't mean we have to hide confabulations.