Machines + Society #15: Amorous Indecision; Therapy Thoughts; Between Us
A newsletter by Mako Shen
machines + society
Mako Shen | Aug 31, 2020
Fleabag and Decision Theory: How (not) to Make Romantic Decisions
“Do I dare to eat a peach? I shall wear white flannel trousers, and walk upon the beach.”
The final scene of the British comedy-drama Fleabag ends with a pathos as potent as anything I've seen (note: if you are inclined to watch this 2-hour feast of a TV show and don't like spoilers, you should skip this essay). Fleabag, the nameless main character, sits next to the similarly nameless priest at an empty bus station. Fleabag is, in her own words, "greedy, perverted, selfish, apathetic, cynical, depraved, morally bankrupt" — yet inevitably endearing. The priest smokes, drinks, and blasphemes, but is an otherwise good Catholic.
Somewhere along the line, they fell in love. She stood up in the silent Quaker meeting, bid by the Holy Spirit, and said, "I sometimes think I wouldn't be such a feminist if I had bigger tits." The priest coughed and paused, trying hard to stifle his laugh. Part of their shared wit is surely a product of the fact that both have suffered the slings and arrows of outrageous fortune, to borrow the analogy of another reviewer. Is it such a surprise then, to see the passion behind their romance?
Now here is an interesting dilemma. If the priest chooses God, he will continue on his upward spiritual trajectory, leaving behind his tormented past. He also knows that if he chooses (if choosing is possible) to love Fleabag, he will be transformed; the piety he has cultivated will be discarded for a more sensuous flavor of life. The uniqueness of their connection and Fleabag's apparent resolve to change make it a plausible difficult choice.
The question I'm interested in is not 'What should the priest (or nun) do?' per se, but rather 'How should the priest decide?'. This question of how we can systematically go about making decisions is the enterprise of decision theory. Note that this is different from much of the typical discussion surrounding standard theory in behavioral economics or psychology: we are not trying to describe what the priest actually does, but rather what he ought to do.
You may scoff at applying a framework to decisions of the heart, but I wager that simply going with the heart is also a framework, albeit a naive one.
Let's look at how we could tackle the priest's dilemma with decision theory.
Idea I: Expected Utility
The standard model, as proposed by Bernoulli (and later expanded by von Neumann and Morgenstern) begins with a utility function. When you put in two options A and B, the function will assign a higher value to option A if and only if you prefer A to B.
In math it looks like this:
p(E_i) is the subjective probability for some event E and u(x_i) is the utility of the outcome.
So, according to the standard theory (which I'll refer to as expected utility theory), all the priest has to do is:
Assign values to outcomes, which, from a naive perspective, are: i) being happy with Fleabag, ii) being not happy with Fleabag, iii) being happy as a priest, and iv) being unhappy as a priest.
Calculate the probabilities associated with each of the four outcomes.
Multiply the probabilities by the values
Choose the highest resulting option.
Clearly, there are a number of problems here. Putting aside the difficulties associated with subjective probabilities (if they are even well-defined when one can't list all the possible outcomes), the priest should be concerned about the very first step: assigning values to outcomes. He should be skeptical that a mental procedure for assigning numbers to outcomes exists at all, much less whether it would be useful. "How could I possibly put a number on a life with God or even a relationship with a woman that I love?", the priest might reasonably think.
At this point, a defender of the standard theory would stand up and retort something along the lines of "OK, in this case we might not be able to use this exact framework, but no one said that you're always supposed to calculate the expected utility. It's just supposed to be the right answer if you had unlimited brainpower and time to figure it out."
I have two responses: First, a utility function does not even exist in this case. It can't, because of the priest's preferences. Second, your theory is not very useful if you can only use it in environments that resemble a casino.
(For those who are curious, the reason that a utility function can't exist is because the priest, in my accounting of the situation, has preferences that are considered negatively intransitive. He may prefer Fleabag + $1000 dollars to Fleabag, but still not prefer Fleabag + $1000 to a life with God. For more details, see page 3 of this paper).
I might add that although the famous duo Kahneman and Tversky may have challenged the standard model to wide fanfare, they still use the utility function in their alternative. A large swath of work on human modeling relies in large part on the same assumptions that justified the standard theory.
So how should the priest presume?
Idea II: Compare internally
Come with an internal set of priorities for a person (some might call these internal values, but I'm going to stick with priorities to avoid confusion).
The priest could establish a set of values that he cares most about in a relationship with a Significant Other, and compare Fleabag and God along these axes. Some conceivable priorities, in order of importance, could include: fostering spiritual growth, becoming a better person, having shared humor, and sexiness.
The rundown:
Fosters spiritual growth (God > Fleabag)
Makes me a better person (God > Fleabag)
Shared humor (Fleabag > God)
Sexiness (Fleabag > God)
Since God wins out in the first two most important categories, the priority comparison framework would suggest he go with God (sorry, Fleabag).
This is appealing because it resolves the impossibility of assigning a number to each outcome, and instead compares alternatives along some important dimensions.
This framework also has some problems, though. Besides assuming that you can come up with a reliable list of your priorities (a task which I can personally attest is difficult), it also assumes that you can rank your priorities: what if the priest can't say whether spiritual growth is more important than shared humor?
The priest could designate priorities as all equal, and just count the number of instances that one alternative is better than the others (or use one of any number of voting systems— ideally the Borda count). More realistically, though, he can use this priority comparison procedure to better inform his gut instinct. I think this solution is the most workable. Some other interesting considerations can be found in this paper.
Idea III: Create your own reasons
The philosopher Ruth Chang has a somewhat different proposal: create a reason for you to commit.
Traditionally, people are considered rational if they are able to recognize and respond to reasons appropriately. In the case of the priest, this means recognizing that he shares humor with Fleabag (Fleabag +1), but also that God will foster more spiritual growth (God +1), and then carefully considering these reasons.
Chang suggests that, while it is of course important to take these external reasons into account, the priest can create internal reasons to justify the choice. So, for instance, the priest might think, "I want to be the type of person that is more distrustful of his emotions"— through his will, he creates a reason to choose God.
This is compatible with using internal comparisons, and the priest would do well to combine both.
In all honesty, if the priest were to ask me for advice, I'd suggest that he talk to his friends and family, sit with his decision for a while to see if his mind changes, and try to write about what he's feeling. I can't in good conscience suggest that he use any formal framework.
I find it interesting that there are such few established frameworks for making romantic decisions, or for that matter, major life decisions. Perhaps this is because if there were a framework for your problem (e.g. what type of wood varnish to use on an outdoor deck) then it would cease to be a major decision— it would just be obvious what to do. Yet I think this explanation ignores the uncertainty and internal confusion involved in major life decisions.
Like most people, I don't really know what I want, and I keep looking around for someone who can tell me what I want, what I should want; but these priorities, values— whatever you want to call them— can't stem wholly from without. To me, there is some solace in the knowledge that the darkness is broad, but not lonely. Everybody is slowly groping their way towards some edifice they can grasp onto.
"Let us go, through certain half-deserted streets,
The muttering retreats
Of restless nights in one-night cheap hotels
And sawdust restaurants with oyster-shells:
Streets that follow like a tedious argument
Of insidious intent
To lead you to an overwhelming question ...
Oh, do not ask, “What is it?”
Let us go and make our visit."
- T.S. Eliot
Questions:
What, precisely, is the problem with using utility functions to make decisions?
What are some alternatives to expected utility for deciding between alternatives?
📰 Assorted Links 📰
"We need a new approach to teaching modern Chinese history... these “modern China” textbooks — like most China histories aimed at general readerships — also embed an unexamined yet problematic conception of “China” itself. The political history of continental East Asia involves a multiplicity of kingdoms, some but not all now considered part of “Chinese” history. Their rulers were of various ethnicities and spoke different languages. Their states occupied many different areas of what is now the People’s Republic of China (PRC), often concurrently and inimically with other states. This drama is not so different from that of the past 2500 years of monarchies and empires in Europe, yet because in the twentieth century the Chinese Communist Party managed to assemble a territory of imperial scale, while Europe consists presently of distinct polities, we now discuss “China” as if it were a single continuous political entity with “five thousand years of history,” and see “modern China” as immanent in the parade of “Chinese” dynasties: an imperial chrysalis from which the People’s Republic of China, the CCP party-state, unfolded like a butterfly after milleniums of moltings." [Source]
Criticism of the MacArthur Fellowship: "I argue that the program functions basically as a PR vehicle for the MacArthur Foundation and for the foundation world in general. I also suggest that it could achieve better results by not awarding grants to people who already have jobs at elite institutions."
A helpful breakdown of the different philosophies towards psychiatric therapy.
How I developed perfect pitch in 30 days at 24 years old. Credible, but am still trying to personally verify.
Guess emails in your own backyard. Email permutator.
Poem of the month
Between Us
لماذا ؟
تشطبینَ كلَّ الأزمنة
وتوقفین حركةَ العصور
وتغتالین في داخلي
جمیعَ نساء العشیرة
..واحدة .. واحدة
ولا أعترض
عیناك مثل اللیلة الماطرة
مراكبي غارقة فیهما..
كتابتي منسیّة فیهما..
إنّ المرایا ما لها ذاكرة.
"...There is nothing I can do
nothing you can do
what can the wound do
with the knife on its way to it?
Your eyes are like a night of rain
in which ships are sinking
and all I wrote is forgotten.
In mirrors there is no memory..."
- Nizar Qabbani (H/t JL)
🎧 Music 🎧
Time — Angelo de Augustine
Poolside — THE ANXIETY. Infectious R&B.
The Frenchman — Maxence Cyrin.