How does one formulate a minimal set of axioms? [closed]

Suppose the study of some kind of mathematical object has evolved in an "organic" fashion, until it reaches the level of maturity that one wants to take an axiomatic approach to the study of those objects.

How does one go about finding a "minimal" set of axioms so that the axiomatically-constructed object has all the desired properties of the "organic" object, and no more?

To give some context to the question: in my math class, we spent the first half of the semester studying the (real and sometimes the complex) projective plane. Now we're looking at the axiomatic approach (begun by Hilbert?) to projective geometry, starting by defining the axioms for an "incidence geometry".

-
I don't think there is an answer to your question... You might just as well ask "How does one do mathematical work?". –  Mariano Suárez-Alvarez Apr 5 '12 at 22:43
One discards redundant axioms by finding proofs of them from the others. One shows that an axiom $\phi$ is not redundant by finding a model $M$ of the remaining axioms such that $\phi$ is false in $M$, and such that $M$ is not something we want to allow. –  André Nicolas Apr 5 '12 at 22:44
@AndréNicolas, Thank you, that's a great way to put it. –  Will Apr 5 '12 at 22:50
@Will: There are complications. The Hilbert axiomatization of geometry was during the prehistory of logic, and was second-order. Nowadays, there is a preference for first-order formulations, which cannot be categorical, though they can be complete. Thus one gives up the notion of capturing all the intuition one has about a particular structure. But the general idea (prove, or find a model that exhibits the independence) works in principle for second-order. However, funny things can happen. For instance, the Pasch Axiom is a theorem if one denies Axiom of Choice sufficient strongly. –  André Nicolas Apr 5 '12 at 23:31

closed as not a real question by lhf, LVK, tomasz, Quixotic, Austin MohrSep 4 '12 at 16:15

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

One shows that an axiom $\varphi$ is not redundant by finding a model $M$ of the remaining axioms such that $\varphi$ is false in $M$, and such that $M$ is not something we want to allow.