This definition of science, which uses 'experiment, observation, and deduction' to distinguish it from the more general quest for knowledge, philosophy, is very clear; but it might benefit from emphasizing science's predictions rather than explanations, for it is continuously being judged and changed by predictions alone.

My personal opinion is that science requires two languages: that which you see in journals and a metalanguage that you don't. Prediction is part of its language, and explanation its metalanguage. My own opinion is that by casting sciences, philosophies, and religions literature into deductive theories (for some writings on faith attempt to prove what is supposed to be illogical), one can see how they differ.

The axioms of a philosophical theory or of some religious literatures are 'very' true, and the logical constant 'truth' does, of course, not change its value. Most scientists express theories in axioms that are assumed true; but I like to use a whole new definition of truth and define science's axioms as possessing the logical variable 'truth', not constant.

Scientists therefore speak of statements deduced from the axioms of a scientific theory as 'true'. Scientific theories are always being tested, for the activity of science is to create theories, modify or replace those in existence. In science, it is possible to awake every morning and find 'true' has become 'false' in some theory. All this activity is based upon prediction and testing.

Explanation is perhaps the driving force of scientific activity, and is part of the metalanguage of science. Explanations fall out of deductive theories when the axioms are intuitive. Reasoned deductions then connect scientific statements with intuitive scientific statements, and there is a psychological satisfaction that comes from this, which we call 'explanation'.

 out of  found this valuable. Do you find this valuable?