Most people like pizza more than they like salad; I don’t think that’s an unfounded statement. But why don’t we ask this instead: how much more do you like pizza than salad?
That might be a tough question. Most people would say, “A lot.” I would take them at their word. But that answer doesn’t satisfy all the questions that one may have. If you like sushi “a lot” more than salad, then the question is, how much more do you like pizza more than sushi (or the other way)? How do we know if you like them exactly the same, or if there is some slight preference?
I have been studying introductory inductive logic, and there is a method (that is disputed) to somewhat quantify this. It involves prospects A, B, and C. You could have prospect B, or you could have a gamble between prospect A and C. This assumes that you prefer B less than A and more than C. That is, you like A the most, C the least, and B somewhat. Then, you have to choose. At what probability of getting A would you choose the gamble over getting B?
Importantly, this has to be lowest possible probability of getting A. You can’t choose a 99% chance for getting A if you would also accept a 50% chance of getting A. Mathematically, the utility of B is equal to that probability. This is, of course, presuming that the utility of A is 1 and the utility of C is 0. Voila, you have a mathematical scale of measuring how much you like something.
However, I’ve left out a crucial detail. That is, what is utility?
I don’t mean the pocketknife. There is likely a more professional and standardized definition, but it boils down to this: utility is preference. It is the overall preference that you have towards something based off of many factors that give you satisfaction. That is the general meaning.
It seems a little bit strange that we can measure preference and satisfaction in a numeric form that isn’t completely arbitrary. After all, this system does use the mathematical concepts of expected value and probability. This method does implicate another question, however: What other mental concepts can we measure using probability?
This is where utility begins to have problems. While it can generally represent satisfaction, there are certain things it can’t represent. For example, caring. I’ve talked about caring in a previous post here, but essentially, caring cannot simply be evaluated by just behavior or just your emotions. What about love? If you rank someone below another person in utility (I guess in terms of which person you get) does that mean that you have that much love in that order?
You see, this is where utility becomes an issue. While utility is useful in certain fields (such as economics, where it can accurately represent the satisfaction gained from a product), it does not hold the same power in measuring everything. This is, I suppose, not a problem with utility itself, but simply the definition of preference. It is itself a good question–how directly correlated are preference, love, care, gratitude, etc.?
In my limited knowledge of Expected Utility Theory, I do not know exactly how this method of applying utility to real world preference was created. I know how it works; it uses the expected value of a probability experiment and uses the individual’s perception of probability (though flawed as shown by the Allais problem) to come up with a probability founded scale of preference. However, from this group of facts, I infer one thing, and possibly the most important thing: the context of the experiment matters.
If I’m asking you to choose between foods, then the context of this gamble is, of course, eating. If I give you a gamble between maybe some treasures that are the same monetary value, but each have a certain look to them… Let us suppose that these are raw quantities of silver, a rare crystal from the Himalayas, and a golden idol from the Holy Roman Empire. I understand these are weirdly specific–don’t question it. In my opinion, I would value the gold idol the most, and the raw silver the least. After all, they all have the same monetary value. The gold idol has history behind it, and supposedly looks pretty, whereas the raw silver has little aesthetic appeal.
Doesn’t that look… somewhat underwhelming in appearance?
In any case, you might notice a very important detail. This gamble is inherently how valuable these objects are to me. This is NOT the satisfaction I get from consuming said objects, like the first scenario. If I was getting an object from the food gamble not for consumption purposes, but simply for visual appeal, my rankings are subject to change drastically. Who knows if I think pizza is more visually appealing than salad? And so, I believe the context of the experiment can also be used to more accurately measure other human feelings, such as love. While the line denoting love and not love is translucent, and not so much a line as an impressionist painting, there are obvious outliers. Intentionally hurting someone shows you do not love them. Inconveniencing yourself greatly to help someone is a clear sign you love and care about them.
So, let’s change an experiment to this: If you have specifically 3 people A, B, and C, and you love A more than C, with B in the middle, let’s have a gamble. You could either spend your time (talking, playing games, vacationing, anything) with person B, or have a gamble between doing so with person A or C. At what probability of getting person A would either choice be completely fair?
Note how this is a much harder question to answer than simply consumption. However, with proper consideration, it may very well be an answerable and viable question for some individuals. Whatever value you get from this between 0 and 1, does it truly now measure utility? How could it measure utility when the question has nothing to do with utility? And that is my point–by using this system practically with other scenarios, it may represent something that isn’t quite utility. Of course, it is definitely still measuring preference. However, the context of this preference can have many meanings that can be used to figure out other feelings.
I have to admit that it still isn’t directly measuring love. After all, love probably cannot just be boiled down to how much time you want to spend with a person. After all, maybe you do want to spend more time with a friend than a sibling you see very often–but it does not mean you love the sibling less. However, if we do this experiment within different scenarios, each of which test a behavior that insinuates love? Then, we may have a good experimental indicator to measuring these previously unmeasurable human concepts. It is likely this has been thought of before, and it likely has many flaws.
Regardless, one should always be allowed to dream.