How does it work? Well, you feed in a parameter that specifies how much to suppress the smaller terms.
-- define a superposition: sa: |list> => |a> + 5.2|b> + 23|c> + 13|d> + 17|e> + 17|f> + 15|g> -- suppress everything except the biggest term: sa: inhibition "" |list> 0|a> + 0|b> + 23|c> + 0|d> + 0|e> + 0|f> + 0|g> -- inhibit at half strength: sa: inhibition[0.5] "" |list> 0.5|a> + 2.6|b> + 23|c> + 6.5|d> + 8.5|e> + 8.5|f> + 7.5|g> -- negative inhibit (ie, suppress the biggest term): sa: inhibition[-1] "" |list> 2|a> + 10.4|b> + 23|c> + 26|d> + 34|e> + 34|f> + 30|g> -- and again: sa: inhibition[-1]^2 "" |list> 4|a> + 20.8|b> + 46|c> + 52|d> + 34|e> + 34|f> + 60|g> -- and again: sa: inhibition[-1]^3 "" |list> 8|a> + 41.6|b> + 92|c> + 104|d> + 68|e> + 68|f> + 60|g>Hopefully that makes some sense.
Update: inhibition[-1]^k may be one way to approach the idea of creativity. When thinking of a problem, the obvious, boring, answer will have higher coefficient. But if we suppress the highest few coefficients, then maybe we have something more creative. Of course, we need some way to test the proposed solution satisfies the required properties. I'm not yet sure the way to implement this, but it seems to be something that the brain makes heavy use of. Posing questions, and then hunting for answers with the right properties.