Do you promise?

I googled for "sunk costs" and came across an argument by Robert Bass of Coastal Carolina University that it's sometimes rational to honor sunk costs. He says that economic theory is incomplete because of these cases. He gives an example: let's say you make a promise, and later realize you'll see greater benefits by breaking the promise. Should you honor that sunk cost and keep the promise, or break the promise? Bass says it's beneficial to be considered a promise-keeper, so it can be rational to keep the promise because of that benefit.

If I'm reading his argument right, I think he's wrong. There's nothing wrong with economic theory -- it's right and rational to ignore sunk costs. Assuming a right and rational world. Humans aren't always right and rational. Making promises is not rational.

So how do I answer his argument that there's cases when you should consider sunk costs? That it's irrational to assume that humans are rational. Back to the promise-keeper -- it's beneficial for me to make a deal with a promise-keeper. It minimizes risk to deal with a promise-keeper. Wanting to deal with a promise-keeper is rational. You're banking on his irrationality.

On a planet of robots, no one makes promises.