The Definition of the Indicative Conditional From Wikipedia, we have:

In natural languages, an indicative conditional is the logical operation given by statements of the form "If A then B". Unlike the material conditional, an indicative conditional does not have a stipulated definition. The philosophical literature on this operation is broad, and no clear consensus has been reached.

http://en.wikipedia.org/wiki/Indicative_conditional
Is it not just a case of disallowing certain rules of inference in propositional logic, e.g. disallowing $\neg A \to A \implies B$ or some other rule(s) of inference that allows you do make such an inference?
I routinely use such rules (as above) in writing formal proofs. Would mathematics as we know it even be possible if we implemented such restrictions? 
 A: It's not that simple. For starters, $\neg A\to A\implies B$ isn't a rule of inference in standard propositional logic. You may be thinking of $\neg A \to A\implies A$.
[EDIT after below exchange of comments: I thought the OP meant $\neg A \to A\vdash B$, but in fact the OP meant $\neg A \vdash A\to B$, which indeed is a rule.]
But also, it's not just a matter of disallowing certain rules. One also has to consider whether certain other rules should be allowed that aren't supported by the material conditional. For example, are $\neg(A\to B)$ and $A\to \neg B$ equivalent? Maybe; note that $1-P(B|A) = P(\neg B|A)$. But the material conditional doesn't support that equivalence.
Finally, there's the whole problem of the semantics for indicative conditionals. Two-valued truth-functional logic simply won't work. Should the logic be three-valued? Four-valued? Modal? No consensus exists.
I guess math as currently done would have to be pretty thoroughly reexamined if one were to replace the familiar (material) construal of if-then with something else. But to say for sure, one would have to have a definite "something else" proposal in hand.
For more on this, see also http://plato.stanford.edu/entries/conditionals/.
A: The indicative conditional describes an ordering of the truth values of statements: B is not less true than A. The various paradoxes and problems of the material conditional become perfectly reasonable with this interpretation, and no more mysterious than claiming that A <= B. 
For classical two valued logic, the material conditional happens to describe the same relationship, and it is perfectly adequate for the job.  But this success is misleading and even tends to obscure understanding of it. 
In three valued logic, it turns out that doubts about the adequacy of the material conditional are well justified. It does not have the same useful, valuable properties in the three valued case that it does in the two valued case: and a conditional corresponding to an ordering of truth values has much better correspondence with the indicative conditional. 
