A utility function is quantitative, not qualitative.
How would you go about transforming these vague statements into precise mathematical definition?
(I’ll grant you “black box rights”; you can use terms—anger, doubt, etc… - that humans can understand, without having to define them mathematically. So if you come up with a scale of anger with generally understandable anecdotes attached to each level, that will be enough to classify the “anger” term in your overall utility function. Which we will need when we start talking quantitatively about trading anger off against pain, love, pleasure, embarrassment...). Indirect ways of measuring utility—“utility is money” being the most trivial—are also valid if you don’t want to wade into the mess of human psychology, but they come with their own drawbacks (instrumental versus terminal goal, eg).
A utility function is quantitative, not qualitative.
How would you go about transforming these vague statements into precise mathematical definition?
(I’ll grant you “black box rights”; you can use terms—anger, doubt, etc… - that humans can understand, without having to define them mathematically. So if you come up with a scale of anger with generally understandable anecdotes attached to each level, that will be enough to classify the “anger” term in your overall utility function. Which we will need when we start talking quantitatively about trading anger off against pain, love, pleasure, embarrassment...). Indirect ways of measuring utility—“utility is money” being the most trivial—are also valid if you don’t want to wade into the mess of human psychology, but they come with their own drawbacks (instrumental versus terminal goal, eg).
Utility is the dot product of the derivative of desires and the observations. Desires are what you attempt to make happen.
If you start trying to make what’s currently happening happen more often, then you’re happy.