There’s is only so much that you can say in a journal article. But how much? And how do you decide what to say?

I usually answer the first question with a simple “back of the envelope” calculation. How many words do you have available? This will often be determined by the journal you are planning to submit to, but it is useful for you to have your own sense of how long the various kinds of papers you are writing will be. (You can then choose the right journal based on the sort of paper you want to publish, rather than the other way around.) Whatever number of words you settle on, divide it by 200 to get a lower bound on the amount of paragraphs you will be writing. Since paragraphs can be as short as 100 words, the upper bound will be roughly twice that. An 8000-word paper will consists of at least (I would actually say ideally) 40 paragraphs and at most 80.

The paragraph, however, is not just an arbitrary way of carving a larger text into 1-200-word units. The paragraph is the “unit of composition”, delivering a single point that it supports, elaborates or defends. That is, each paragraph says one thing you know and then tells us how you know it. When you are trying to decide what your paper will say, you are ultimately trying to decide what each paragraph will say. And after you have done your back-of-the-envelope calculation you know roughly how many decisions you have to make. Let each decision produce a “key sentence” that belongs somewhere in the outline of your paper.

I often suggest that people make these decisions in four stages. First, decide what you will say in paragraph 1, 2, 3 (the introduction) and “39” (the first paragraph of the conclusion). Then spend two hours writing those paragraphs, one moment at a time. Next, decide what you’ll say say in two paragraphs of each of the background, theory, methods and discussion sections, as well six things you’ll say in your analysis. Spend seven hours writing those 14 paragraphs, again taking a well-defined writing moment to produce each one. You know what you are trying to say; think carefully about how you’re going to say it. Craft the paragraph so that it tells your reader how you know and, ideally, imparts that knowledge to the reader. That is, at the end of the paragraph the reader, too, knows what you know about some particular thing.

You now have 18 paragraphs written and, if you’re writing a 40 paragraph paper, 22 paragraphs to go. In the third stage, read through what you have written, mindful of what you have not yet said but would like to say to your reader. Write down the key sentences of these paragraphs as you think of them. The  reading itself should take 18 minutes. (A paragraph should take about a minute to read. It should be possible to “get it” in that time; that is, after a minute the reader should understand what you are trying to say and why you think it is true.) If you add in the time to write down the “missing” ideas, the whole exercise should take no more than an hour. After that hour is over, take a look at what you’ve got. Put the key sentences in the right sections and in the most logical position relative to the 18 paragraphs you’ve already written.

Now take 11 hours to write those remaining 22 paragraphs. (Don’t work more than three hours per day on this, please. Two is plenty.) When you’re done, take a one or two day break from this paper. Now, give yourself an hour to read the whole text through, slowly and out loud. For each paragraph, identify the key sentence and put it in a separate document, numbering the paragraphs as you go. Don’t “construct” these sentences. Just pick the one sentence in each paragraph that expresses its point and put this in your after-the-fact outline. Don’t use any headings in this document. It should just be a list of 40 sentences, numbered to correspond to each paragraph in the paper.

Take a break. Go for a walk. Clear your head. You might even want to call it a day.

The fourth stage begins by looking at your after-the-fact outline. Read each key sentence out loud. Feel free to polish it and sharpen it so that it makes the strongest possible impression. Get it to say exactly what you think about the matter. (Remember that it appears in a paragraph that supports and qualifies it. Trust in that as you sharpen its edge.) Now look at the sentences in sequence. Do they make sense as assertions made one after the other? Remember that you’ve left out all the relevant arguments and evidence. The question, in a sense, is “If I persuade you of this, will the next claim make sense too?” Do your key sentences indicate the “line” of your argument? Move them around until they do. Add key sentences as needed. You might also have to remove some points that interfere with your purpose in this paper.

You now have a list of crisp, declarative sentences that say the roughly forty things that you want to get across to the reader in this paper in an order that makes logical and rhetorical sense. You have finally decided what you want to say. Now, take 20 hours to rewrite the paragraphs that support, elaborate or defend each of those 40 claims as best as you know how. When you’re done, the entire process will have taken about 50 hours: five five-day weeks working on average 2 hours per day on this paper. (Including reading and deciding, stage one took about five hours. Stage two 2 took about eight. Stage three took about 13 hours and stage four took about 24.) Whether or not it was worth it depends on what you’ve got to say, of course. I’ll leave that judgment entirely up to you.