The drawing on the last page of Yoshida's notes make me puzzle, perhaps you can help? It shows a "landscape" featuring the monodromy-weight conj., the general Ramanujan-conj., the Langlands correspondence and some connections between them. However, I guess the Ramanujan-conj. "is" the monodromy-weight-conj. expressed via Langlands correpondence and the Rapoport-Zink spectral sequence is a tool for showing the MWC in some cases and reducing the computation of L-functions to semi-simple L-functions. That seems not to fit to the drawing. Can you can clearify that?
Edit: A fascinating new survey on the "philosophy of weights" with many conjectures by Jannses. (Posted here because I don't know an entry where it would fit better)