I have a paper to write. Where do I start? The first time I asked this question, it was easy: just sit down and start typing. I wrote a few hundred words, then got stuck; I needed to think some more, so I took a break and did something else. Over the next few days, my thoughts on the subject settled, and I was ready to write again. So I sat down, and asked: What do I do next? Fortunately, my brain had a cached response ready to answer this question: Solitaire!
So I procrastinated, and every time I asked my brain what to write, I got back an answer like "Don't bother!". Now a deadline's approaching, and I still don't have much written, so I sit down to write again. This time, I'm determined: using my willpower, I will not allow myself to think about anything except for the paper and its topic. So I ask again: Where do I start? (Solitaire!) What thoughts come to mind? I should've started a week ago. Every time I think about this topic I get stuck. Maybe I shouldn't write this paper after all. These, too, are cached thoughts, generated during previous failed attempts to get started. These thoughts are much harder to clear, both because there are more of them, because of their emotional content, but I'm determined to do so anyways; I think through all the cached thoughts, return to the original question (Where do I start?), get my text editor open, start planning a section and... Ping! I have a new e-mail to read, I get distracted, and when I return half an hour later I have to clear those same cached thoughts again.
Many authors say to stop in the middle of a thought when you leave off, so that "Where do I start?" will always have an easy answer. This sounds like a solution, but it ignores the fact that you'll get stuck eventually, so that you have to stop, at a spot that won't be easy to come back to.
In order to stop procrastinating, there are two obstacles to overcome: A question to answer, and a cached answer to clear. The question is "What do I do first?" and the cached answer is "procrastinate more". Knowing that "procrastinate" was a cached answer makes it easier to get past, but the original question is still a problem. Why is deciding what to do first so often difficult?
When I'm programming, I make a long, unorded to-do list for each project, listing all of the features I plan to implement. When I finish one, I go back to the list to pick something to work on next. Sometimes, I can't decide; I just stare at the list for awhile, weighing the costs and benefits of each, until eventually something happens to distract me. Most of the items on that list are harmful options, which serve only to induce analysis paralysis. It's the same problem some people have ordering off restaurant menus, and the same solution works. Instead of considering a series of options and deciding for each whether it's good enough to settle on, choose one option as the current-best without considering it at all, and compare options against the current-best.
Usually, choosing where to start, or what to do next, requires generating options, not picking one off a menu. When choosing, say, the topic of the next chapter, it's easy to convince ourselves that we'll come up with the perfect answer, if only we think about it a little more. If we take the outside view, we can see that this is probably not the case; and if we let thinking about one decision crowd out everything else, and think about it long enough without reaching an answer, then eventually we will settle on Solitaire as the best choice. When deciding how much thought to apply, remember: The utility we get from thinking about a decision is the cost of deciding incorrectly times the probability that we'll change our mind from incorrect to correct, minus the probability that we'll change our mind from correct to incorrect; and the longer we have gone without changing our mind, the less likely we are to do so in the future.
Procrastination is not a single problem, at least two: cached thought, and analysis paralysis, working together to stop us from getting work done. If we miss the distinction, then any attempts to find solutions will be doomed to confusion and failure; we must recognize and address each underlying problem, separately.