One way to focus your mind on the problem of writing a particular paragraph is to ask yourself what its central claim is. What fact are you claiming is the case? At the end of the day, what state of affairs in world makes it true? The paragraph is ultimately trying to persuade the reader on exactly that point.
A couple of weeks ago, I used Wittgenstein's Tractatus as an example. "The world is everything that is the case," he wrote, and left it at that. But he could have written a whole paragraph to establish this point in a more prosaic way. (I'm not going to say that this would have been more "effective". It's hard to criticize the effectiveness of what is arguably the most influential book of philosophy in the twentieth century.) What, then, would support for this claim look like? Well, why would it need support?
The answer here comes from thinking about the reader. What are we expecting the reader to do with the claim that the world is everything that is the case? Some claims are hard to believe, and we would have to provide solid evidence. Other claims are hard to understand, and we would have to help the reader to do so by unpacking it. I think this claim is a special case of the second kind. It's not hard to believe because it is essentially a truism. The rhetorical problem is showing the reader why it is important to say it. Why do we have to be reminded of this elementary truth? Because philosophy is always trying to draw it into question. This is called skepticism.
Wittgenstein eventually rejected his own rhetoric here, taking it to be an example of metaphysical magic—a kind of trick. "When I began [the Tractatus] to talk about the 'world' (and not about this tree or table)," he said, "was I trying to do anything except conjure up something of a higher order by my words?" But conjuring up higher orders (of abstraction or generality) is a perfectly respectable task in writing. The whole idea is to get the arrangement of particulars to suggest something of a higher order.
A paragraph is a great place to arrange particulars in support of universals, specificities in support of generalities, the concrete in support of the abstract. Something like this, for example:
There is a tree outside my window. It is swaying gently in the breeze and casting a shadow along the ground. Yesterday, I walked past the tree and saw a squirrel scurrying up its trunk, scaring off a bird that had been sitting in its branches. Today, it seems to be abandoned in the parking lot, growing out of its little island of earth and cobblestones. But there is more: there are buildings all around, a hundred windows that look out on the tree. The tree is outside each of those windows. There is a tree outside each of those windows. The world is everything that is the case.
As with any literary experiment, there's some editing to be done here. I'll return to it on Friday (I have something else in mind for Wednesday). The effect I'm trying to achieve is the passage from an isolated (lonely) fact to a hundred views of that fact to the objectivity of the fact-in-the-world. It is a fact that the tree is outside my window, but not just my window. There are a hundred facts of this kind (one for each window). This suggests that the tree out there is a fact "in itself". And this suggests a "real" world.
Each paragraph in your paper states something that is the case. Together they position what you know to be the case in the larger of whole of everything that is case. They find a place in the world for your knowledge.