I have decided on the area I want to contribute to, and have found a basic but huge problem for my research. Now my professor told me to redefine the area, so to break the, at the moment, huge area down into subquestions to create an interesting closed problem, which is addressed by my thesis.

At the moment I have come up with several different questions I have found. However, I do not see any pattern where these questions are related to each other. My questions also seem quite loose and do not define the closed problem I am aiming for.

A related question on this site is how to come up with research ideas. My question is not about how to come up with ideas in the first place, but how to define these ideas so that they can practically become a thesis.

Hence, any suggestions how to come up with several subproblems, which define a closed corpus for my thesis?

  • 4
    The usual algorithm for my students and colleagues' students is first get some research results, then write some papers, then choose a subset to bundle into the thesis, and only then decide/deduce what the thesis is about.
    – JeffE
    Commented Jul 22, 2014 at 23:05
  • Depending on the local rules (and subject), it may not be strictly necessary for the material included to all fit one theme.
    – Jessica B
    Commented Nov 18, 2014 at 6:45

1 Answer 1


This sort of thing only evolves as the product of your exhaustive reading, and discussions with your supervisors and peers.

After that, at some point, your supervisor will say something like: "before you can really get into your research, just resolve this little thing - it should be fairly trivial."

That's your thesis, right there.

  • +1 just resolve this little thing - it should be fairly trivial; one could only hope! Commented Jul 21, 2014 at 17:55
  • 3
    +1 The core idea for my doctoral dissertation was the result of a lunch time conversation with my adviser, during which he asked me an apparently simple little question. Answering it took me several years of reading, thinking, and computer simulations. Commented Jul 21, 2014 at 20:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .