When narrowing uncertainties, deductive reasoning can be more
helpful than inductive.
I think in the application of deductive reasoning, however,
parallel threads are important. I consider reasoning akin to
navigating a maze.
If one can follow several likes of deductive reasoning
simultaneously - or split up the task, then elimination of
improbabilities won't be necessary, as there will be some that
navigate the more narrow, less likely passageways.
The main line of deductive reasoning may well be correct in the
end, but I think of criminal work or solving mysteries for
example; as new facts enter into the case, very sophisticated
lines of deductive reasoning may have to take a lower peg in the
chain of priority, although still being followed. Sometimes
rejecting the entire line of reasoning based on a single piece
of evidence turns out to be hasty as well. It's why I believe in
parallel threads.
I don't believe in elimination of improbabilities entirely but
rather lowering them in priority, for sometimes the improbable
is simply missing a single datum.
