I want to start getting up in the morning, rather then in the afternoon. It seems like a pretty clear choice:

  1. Sleep in all morning and be grumpy
  2. Get up early and be happy

So afternoon-me makes a plan to get up early every morning, sets an alarm and calls it a day.

And that would be the end of it, except that when morning rolls around morning-me discovers a third option.

  1. Sleep in all morning and be grumpy
  2. Get up early and be happy
  3. Sleep five more minutes, then get up early and be happy

The best part about option 3 is that morning-me gets to choose it every five minutes. He never actually makes the decision to sleep in all morning. It just spontaneously coalesces in a totally blame-free fashion.

But afternoon-me is wily too, and has figured out a way to close this loophole by specifying a specific moment in time at which the decision has to be made. I set a particular song as the alarm and plan to get up on the first line of the first chorus.

  1. Sleep in all morning and be grumpy
  2. Get up early and be happy
  3. Sleep five more minutes, feel a bit conflicted about not sticking to the plan, then get up early and be happy

This actually works! Morning-me concedes and gets out of bed at the agreed point in time.

This is not a new idea and I’m not writing a self-help article here. What I’m actually interested in talking about is this:

Option 3 is a lie.

I have never, ever slept in for five minutes and then gotten up. I am well aware that this is a vanishingly unlikely possibility, at every time of day except for the five minutes after my alarm goes off. So the actual options are:

  1. Sleep in all morning and be grumpy
  2. Get up early and be happy
  3. Sleep in all morning but believe I’m going to get up early, because I’m either deliberately lying to myself or temporarily insane

All I’m doing by sharpening the decision point to a specific moment in time is penalizing option 3 to make it slightly less attractive.

Why is this interesting?

The dominant theory of decision making in behavioral economics is utility theory, which basically asserts that we have some inner function that assigns utility to every future outcome and we pick the outcome that maximizes utility. Reversals of preferences are often explained by temporal discounting - afternoon-me wants to get up early, but morning-me prefers the guaranteed reward of sleeping in now to the uncertain reward to getting up and feeling good about it ten minutes from now.

Even if we put aside the prospect of a temporal horizon so short that it won’t carry me down the stairs to breakfast, the only ways I see to explain my revealed preferences (sleep-but-pretend > get-up > sleep) with utility theory are to either assign utility to lying to myself or accept that my utility-predicting module is so detached from reality that it really believes I’m going to get up in five minutes.

Both explanations work - all the epicycles line up - but it seems to challenge the prospect of utility as a predictive theory. How do you predict the actions of a utility-maximizer that is allowed to invent impossible futures to justify its decisions?