As a final (I think) follow up to my previous discussion, I think it worth noting that writing is often a process of discovery: Sometimes we don't have a clearly defined sense of purpose for our writing when we start.
Or, at the least, we don't have a clearly defined specific thesis for the work that we are attempting to create. And this is something slightly different than a sense of purpose.
A sense of purpose, as I'm thinking of it here, is a sense of something to be accomplished, a sense that there is something important going on that needs to be addressed. A clearly defined specific thesis on the other hand, is just that: clearly defined.
It is to be hoped that we can start with a sense of purpose and move towards a clearly defined thesis. This is, in a way, the ideal dynamic because it means that our writing is motivated, but it also means that we have not simply imposed our ideas on a project with no openness to discovery. And discovery is important--it is a crucial part of the academic process, and is, in fact, the very reason for the academic process.
But a sense of purpose need not be clearly defined: it can still usefully motivate coherent work. And the internal, personal truth of having a sense of purpose cannot be replaced. We benefit from having some idea that our work is important and that it is striving towards a goal. The fact that we will want to define that sense of purpose more clearly as we proceed does not eliminate its importance as a starting place.
A dissertation is not simply a repetition of the woks of others. It is intended to bring new ideas to the world--new information, new interpretations, new responses--something new. And that, in itself, is an abstract sense of purpose that can be defined in the process of writing--which is, as I've said above, a process of discovery.