Some of you arrive knowing well what you'd like to research. But for those who don't, we have some advice.
The best topics are almost always much more specific than you think they should be.
For a history essay you should identify a set of primary sources pretty early, but you need to know enough about the general topic first in order to know what's interesting or novel about that set of primary sources that is worth devoting your time and attention to it. So start big – look at Cambridge Encyclopedia of the History of Science chapters, History Compass articles, and other review essays – and then start narrowing in on times and places you are interested in.
Especially with the shorter essays, do not try and take on a topic in which you will follow a subject through an entire century or half-century. It almost always results in an over-general topic – and you end up making the process much harder for yourself by committing to such a big intervention covering lots and lots of secondary literature. Wider scope becomes more feasible for the longer essay and the dissertation, but even then listen to your supervisor if they encourage you to tackle a less ambitious goal.
For a philosophy essay, you want to identify a literature to make your intervention. There is almost certainly more of it than you expect. The best way to get up to speed is to read the relevant entry of the Stanford Encyclopedia of Philosophy. You can also attend the Part IB or Part II lectures and examine their reading lists on Moodle. The list of supervision questions provided for each course (or even the past exam papers) are a great place to start if you are not sure about how to formulate a question for your essay. Do not be shy about adopting a ready-made question. Your supervisor will help you to make it your own.
To see what is a realistic goal to achieve in a 5,000-word essay you can check out the Supplement issues of the journal Philosophy of Science, which publishes papers presented at the PSA conference. The 8,000-word essay and the dissertation are similar in length to normal journal articles. Browse those or the examples of essays in the Whipple Library to get a feel for how to develop a sufficiently focused question.
Before making a final decision, ask yourself: what do you most want to read for hours and hours every day for the next 4–6 weeks?