There could be no more perfect timing than the posting of this PhD comic: "That's the test of a true Ph.D...To take 5 years of marginally related work and pretend you knew what you were doing the whole time". I read it just as I was sitting down to finally finish my dissertation proposal. This was definitely an exercise in pretending I meant to look at this theory all along, when what really happened was that I had randomly attached myself to some studies that produced interesting results while all my own research failed. There's nothing like trying to turn a completely unexpected result into a justified prediction.
The dissertation proposal process also fits in well with another PhD comic - take it out; put it in; take it out again. The proposal I turned in bears a striking similarity to the first draft I sent to my advisor - "striking" because at some point in the revisions process she had me substantially alter the last two experiments, and then had me change them back to what they used to be. The proposal is certainly better for her input and my rewriting, but I could have done without adding and then scrapping some experiment manipulations. It certainly doesn't bode well for my proposal meeting next week, when the rest of the committee will weigh in on how my proposed experiments could be "improved". At some level, I just want to tell them that I have a better idea than they do about what we're doing and what will work, and to let me do it. I don't think that would go over very well.