It should be possible to write a journal article in about forty paragraphs. What this means is that writing down forty (carefully chosen) things you know (in the right order) should suffice to present a discovery you have made to your peers. When I say it should “suffice”, I mean exactly that: forty paragraphs should be enough to get your point across. It should give your peers an adequate basis on which to decide whether you are right or wrong. It should either impart your knowledge to them, or give them an occasion to correct your errors. That is, the writing and reading of those forty paragraphs will help make someone smarter.

Now, imagine presenting the same result in exactly 18 paragraphs. Why might this not be sufficient when forty paragraphs was sufficient? What would you be forced to leave out? When you think about it, you’re actually just reducing the amount of people that might engage meaningfully with your text. They will still (if you write it well enough) be able to discern what you are trying to say, but they will not be persuaded, either because you don’t present enough evidence or don’t tell them enough about your methods. Or perhaps you just leave out the discussion section and therefore give no indication of what the consequences of your discovery might be, whether for theory or practice. But a smaller group of very like-minded readers might not need any of this information. They would perhaps just have skipped over the paragraphs you’ve left out anyway.

Next, imagine 11 paragraphs. Then 5. In each case, think about what you would be leaving out.  And, more importantly, who you are leaving out. Who are you now not trying to convince. Who is becoming more of a spectator of than a participant in your discourse. Actually, at five paragraphs you are essentially writing an abstract (albeit an extended one) or synopsis. You are presenting a distillation of your discovery and implicitly telling the reader to ask you for more information.

So think about this text now as one that raises a series of questions and make sure you know the answers. (Don’t make the text a series of questions, though. Make it a series of facts that naturally stimulate questions.)

Now reduce your text by one paragraph. Then do it again. 4, 3, 2 …

Finally, imagine your paper in a single paragraph of at least six sentences and at most 200 words.  A good idea can be presented at these different levels of abstraction. I encourage you to develop your ability to scale the paper in these ways. Impose a set of constraints that are, from the point of the view of the idea itself, arbitrary.  Then make the most of the conditions you’ve given yourself.

But don’t let your time investment be unconstrained. Always think 27 minutes per paragraph. You don’t want to get good at something that takes forever to do well. You want to become good at writing down what you know in the moment.