"backwards" Diffraction Grating with light source at maxima (1st, 2nd, nth order)? How will a light source with a specific $\lambda$ positioned right at the nth order maximum pointing towards a diffraction grating be affected? Will it come out "directed"?
 A: This is effectively just a diffraction problem.
I.E. you get out a diffraction pattern on the left similar to the one you expect to get on the right in the first case, and one of the peaks will be at the original light source.
Why will we get multiple peaks instead of just one corresponding to our original light source?
In the case of the first diffraction grating, if we took all of the photons in mid air and reversed time we would get them all to go back to the original source, this is because each of the maxima acts as its own source producing a diffraction patter but due to super position of all of the patterns all of the maxima (except for the central one) cancel out leaving only the source.
So in order to produce directed light as you suggest you need multiple coherent light sources positioned at each of the maxima and they have to be in phase with each other. This would be very difficult to achieve.
However one way to accomplish this may be to get a light source diffract it through a grating, then re-focus it using a parabolic mirror onto a different diffraction grating.
