Degrees of Freedom in case of multiple regression?

#1
What does dof of each predictor show in case of multiple regression?

Instead of explaining mathematically, could anyone explain the physical reasoning or explain in words what does it actually mean?
 

Dason

Ambassador to the humans
#3
Are you talking about the degrees of freedom associated with the predictor parameters? If you're doing an actual regression then those should just be 1 for all of them. If you have df greater than one my guess is you have some categorical variables as predictors as well. Is that the case?