...why was the auto slat function of the stall warning computer/protection on the McDonald MD80/83 series not designed to deploy from a clean wing?
Because, according to a design engineer who worked on the type, natural stall warning was noticeable/sufficient in the clean configuration, whereas, with partial slats, it was not.
Don't shoot the messanger...that's what he said.
This sort of arrangement (natural stall buffet, not aircraft systems-generated warning) is present on other types as well.
Example.
With the L1011, natural stall buffet will be noticed well before stick-shaker activation in the clean configuration, whereas, with slats extended, it will not...IE: shaker first, then stall buffet.
I know this because, I have completed full stalls in the type, long ago....and not something I desire to repeat.
