Inferring P from P v Q': Is it Valid?

[EvLer]

So, there is this "inference rule" in my textbook:
from P v Q, we can assume P

but does this hold for P v Q', i.e. given this, can I say we have P?

My intuition says "no", but I'd rather have a reasoned answer...
Thanks in advance.

 

[AKG]

From P v Q, you may not infer P. You can assume P anytime you like, so long as you discharge all assumptions by the end of your proof. By "v", you do mean "OR", right? From P & Q you certainly may infer P. And if, by Q' you mean not-Q, then you may infer P from P & Q' just as you may infer it from P & Q.