This concept is explained in 'The Economic Way of Thinking', by Paul Heyne:

When you pass through the cafeteria line, pick up the tuna lasagna, and pay the cashier $1.90, you incur a cost: the value of whatever opportunity you will have to forgo because you've spent $1.90. Then you take your first bite and suddenly wish you hadn't selected this item. What will be the cost to you of leaving the lasagna on your plate?

It will not be $1.90 - or $1.80, if we assume your first bite consumed about ten cents' worth of tuna lasagna. That cost is history. The cost of leaving the lasagna on your plate will be the value of whatever opportunity you forgo by doing so. Do you have a dog that would enjoy pasta with tuna? If so, the cost of leaving the lasagna is the opportunity you forgo (by not asking for a doggie bag) to see your dog's eyes light up and its tail wag.

The price you paid is what economists dismiss as a sunk cost. Sunk costs are irrelevant to economic decisions. Bygones are bygones. The proper stance for making cost calculations is not looking back to the past, but forward to the future.

Of course, we must be certain that a cost is really sunk, or fully sunk, before we decide to regard it as irrelevant to decision making.

See interactive ebook on iBookstore