Well, I do an Arts subject so this maybe of some help, unless anyone else comes along who does the exact same subject.
-
You may not need an MA to get on the course, but you can only apply for funding if you have one - something to bear in mind. TBA, I'd really recommend it - I think 'exceptional circumstances' really means people who've published already because they're working in proximity to the subject they want to study.
-
Structure: You'll start off (whether or not you do an MA) with a provisional place: you'll be what's called a probationary research student. After a year or so, you will need to prepare a piece of work from your thesis (usually a short version of a chapter or an introduction, maybe 8000 or 10,000 words). You'll submit this to a panel of academics, and have an oral examination with them. If you pass, your PhD status will be confirmed and you'll continue to work for another two years. If you fail, you may end up re-submitting or you may be told you must stop now and take an MPhil instead of a PhD. Or, in the worst case scenario, you'll simply fail.
Assuming you end up doing the full 3/4 year course, you'll find you meet with your supervisor somewhere between every few weeks and every few months, depending on how your work is going, how much you need guidance, and how busy you both are. Lots of people find they spend 6 months or even a year feeling a bit aimless, reading lots and making lots of notes and trying to draft chapters that are small sections of their thesis. Eventually, you get a sense of what your topic will entail and how to plan it all out into chapters. Once you've got to that stage and you feel secure(ish) with your plan, you'll alternate between researching and writing. Your university will have occasional lectures or courses on how to research, which you may want to go to. You'll also be writing short papers to present at conferences, and trying to get short papers published.
At the end, you should have a thesis (around 90,000-100,000 words), which might have 4-6 chapters. Ideally, you'd have been to several conferences and presented a paper at two or three, and ideally, you'd have published at least one paper. These papers would probably be short versions, or early drafts, of one of your chapters. The conferences and publishers would have given you feedback on how to make your work better, so you have their imput as well as your supervisor's regular imput.
- To write a proposal you need a good general knowledge of your subject, to MA level (even if you've not got the MA). You need to demonstrate that you know what the current state of scholarship is in your area, and that your PhD topic is going to make a new and original contribution to that scholarship. You also need to show your topic has enough depth for a three year project of 90,000 words.
You don't need to know already exactly what you hope to conclude - you want to show you have important questions to ask and some sense of where to look for the answers, but you don't have to have all the actual answers yet. And, when you start your PhD, you can make big changes to your topic if it's necessary.
I think this is already a really long post so will stop, but ask if you want to know other things and think I can help. Good luck! 