Things to think about when you are putting together your PhD dissertation proposal.

If you are considering asking me to be on your PhD defense committee, I will expect you to have a plan for your PhD proposal defense. The requirements for this are quite clearly explained in the CSE Graduate Handbook's relevant section, and is also explained in the following piece of advice taken (modified) from UPenn's linguistics department (thanks to Georgia Green):

  1. The purpose of the proposal is to convince your committee that there is a tractable question which is worth pursuing and that you are in a position to do a good job of pursuing it.
  2. Therefore, the proposal should demonstrate that you:
    • [a.] have defined and delimited an interesting research question
    • [b.] can explain the importance of the question to a computer scientist not intimately familiar with it
    • [c.] can formulate testable hypotheses
    • [d.] have a detailed plan for testing the most promising hypotheses
  3. You do not have to read everything that was ever written about anything that might conceivably be relevant to a full understanding of the phenomenon you are interested in addressing before you write the proposal, but you do need to be familiar with material that you know is germane to your approach to the problem. You are expected to make an effort to locate such material.
  4. Whether your proposal contains a Literature Survey summarizing the history of relevant research on your topic, and if so, how extensive it must be, should be settled early between you and your advisor. In any case, you should situate your proposed dissertation within the context of what is known and/or generally believed about the phenomena you will investigate, and you should discuss both the lasting contributions and the shortcomings of previous research.
  5. Do not attempt to satisfy (2a-d) by doing the dissertation research before you write the proposal. Do not write the dissertation before the prelim. If you do, you will be treating your hypotheses like conclusions, and your prelim will turn into a defense of those propositions. Since that is the role of the dissertation defense, scheduled after a year or so of testing, writing, reviewing, revising, retesting, and rewriting, you can expect to fail if you try to do it at this point. A proposal is supposed to describe what you propose to do, and why and how you propose to do it.
  6. Questions your proposal should answer directly:
    • What problem are you going to tackle?
    • Why is it a problem?
    • Why is it important to solve it?
    • Where are you going to look for answers?
    • Why are you going to look there?
Nothing in any of the above implies any particular structural format that a dissertation proposal must have. When you plan your proposal, it should be with the purpose of the proposal (as indicated above) in mind. For each section, it should be transparently clear what that section has to with which purpose.

A note on exposition

  1. Don't put the footnotes at the end of the document; put them at the foot of the page.
  2. It is not enough to say what you believe to be true. You need to be clear and explicit about how your (tentative) conclusions follow from the assumptions you make, and then make a big deal of them.