I've read in some book that Aristotle was really not into checking the conclusions with practice after the conclusion was made. When you're working on an idea, according to Aristotle, you use observations which are available - but you don't experiment. The result of reasoning is going to be best as can be argued, and passive observations are ok, but (active) experiments to check just fall out of picture - nobody is supposed to do them, either before or after working on the idea.
That was some millenia and a half before Galilei.