Deduction theorem for first order logic

[Jeroslaw]

If I have P l- Q in FOL and P is closed, can I infer l- P -> Q. IIRC, this is valid as long as P is closed, but my memory is a little hazy. Is that how it works?

 

[Jeroslaw]

Let me elaborate on my question. Say that we have T l- S. In order to infer l- T -> S, what restrictions must be placed on the formulas? How does the deduction theorem for first order logic depend upon the rules of inference that are allowed?