Surely a simpler explanation for that is that business people don't usually work with crisp, binary data. Instead they're (implicitly) doing some kind of statistical inference on noisy data.
Propositional logic might be an acceptable approximation to this kind of inference in certain narrow situations, but the approximation breaks down quicker when the chain of inference gets longer.
That could certainly be a simpler (and therefore better) explanation. :) When your "axioms" are fuzzy so to speak, and not real axioms, it's of course dangerous to draw conclusions from several of them (because the risk of one of them being false increases exponentially).
But I still have a feeling there's something there... :) For example, I remember reading here on hacker news about this CS professor who had found a way to predict performance in entry level programing courses: http://www.eis.mdx.ac.uk/research/PhDArea/saeed/. Basically, they just tested their students ability to form a self-consistent model of how programming works. If they could they did well in the course. If they couldn't they did poorly, and it was very difficult to help them.
That experiment leads me to believe that about 50% of the population is in the habit of constructing self-inconsistent systems of hypothesis (provisional axioms you could say). :) If that's true I'm not sure yours is a simpler explanation... ;)
Propositional logic might be an acceptable approximation to this kind of inference in certain narrow situations, but the approximation breaks down quicker when the chain of inference gets longer.