Suppose you love Pokemon cards. You spend your first 50K on Pokemon cards, and your second 50K on Pokemon cards. You can't infer, then, that you value the second 50K of Pokemon cards less. 1/

Even if the goods purchased are different, maybe I get the same or more value from the second 50K, but I had to combine it with the goods purchased first? (Maybe I buy a record player first, but it's the records I get the most utility from?) There are lots of possibilities! 2/

The main evidence for declining marginal utility when economics is taken as a positive science is in risk aversion. Absent declining marginal utility, you should be indifferent between doing nothing and flipping a coin, heads you will $1000, tails you lose the same $1000. 3/

Most people refuse that bet, absent some sweetener, suggesting the prospect of gaining an amount, not in terms of any particular good but the purchasing power with which we purchase goods, is less valuable than the cost of losing the same. /fin