I don't understand why we can't have a hierarchical knowledge/understanding/model scheme that defines an asymptotic unreachable bottom?
Now 'encouraging clarity' is a bit ambiguous. Clarity depends on purpose and context.
Imagine we are building a model of a house for the purpose of enabling someone to recognise the house when they see it from the front yard. A suitable model would be a photograph... this encourages abstraction by removing unnecessary detail like how many bathrooms the house has etc. We can encourage clarity in our photograph model by making important properties more visible (and using abstraction)... so if the side of the house isn't as important since we're looking at the house from the front yard, the photo should emphasize the front of the house.
This model has a high degree of clarity in the sense that it easily allows a person to use it to recognise a house and hence satisfies its purpose. But it wouldn't have a high amount of clarity for a thief who wants to find out whether the back door has an easily breakable lock.
I know this is slightly different to the OP in that I'm talking about modelling a real object with another object rather than modelling reality by scientific knowledge or general perception. But I think similar principles apply - the difference is that science holistically tries to eliminate abstraction as much as possible. Anyway I do agree that clarity for most people and purposes is essentially a filter for reality.
|